1

Investigating the activation and regulation of theproteasome, an essential proteolytic complex

Patrick Masson

Institutionen för Molekylärbiologi & FunktionsgenomikStockholm Universitet

Stockholm 2004

2

Abstract

The proteasome is a major non-lysosomal proteolytic complex present in eukaryoticcells and has a central role in regulating many protein levels. The complex has beenshown to participate in various intracellular pathways including cell cycle regulationor quality control of newly synthesized proteins and many other key pathways. Thisamazing range of substrates would not be possible without the help of regulators thatare able to bind to the 20S proteasome and modulate its activity. Among those, thePA700 or 19S regulator and the PA28 family are the best characterized in highereukaryotes. The 19S regulatory particle is involved in the recognition of ubiquitinatedproteins, targeted for degradation by the proteasome. The PA28 (also termed 11SREG) family is composed of two members the PA28αβ and PA28γ. The function ofPA28αβ is related to the adaptive immune response with a proposed contribution inMHC class I peptide presentation whereas the biological role PA28γ remainsunknown. The main objectives of the laboratory, and subsequently of this thesis are touse Drosophila melanogaster model system and its advantages to better understandthe precise contribution of these different activators in the regulation of theproteasome. Using genomic resources, a unique Drosophila PA28 member wasidentified, characterized and was shown to be a proteasome regulator with all theproperties of PA28γ. Through site-directed mutagenesis we identified a functionalnuclear localization signal in the homolog-specific insert region. Study of thepromoter region revealed that transcription of Drosophila PA28γ (dPA28γ) gene isunder control of DREF, a transcription factor involved in the regulation of genesrelated to DNA synthesis and cell proliferation. To confirm that dPA28γ has a role incell cycle progression, the effect of removing dPA28γ from S2 cells was tested usingRNA interference. Drosophila cells depleted of dPA28γ showed partial arrest in G1/Scell cycle transition confirming a conserved function between Drosophila andmammalian forms of PA28γ. Finally, characterization of the Dictyostelium regulator,an evolutionarily distant member of the PA28γ, was carried out using fluorogenicdegradation assays. We are currently knocking-out the gene in order to determine thebiological function of the activator. A second part of my work consisted in thegeneration of a Drosophila assay used to identify in vivo substrates of the 19Sregulator, an assay system that has been originally engineered by Dantuma andcoworkers in human cell lines. This was achieved by cloning of GFP behind a seriesof modified ubiquitins that create substrates degraded through different pathwaysinvolving the proteasome pathways. The last project of my thesis concerns thecharacterization of the mechanism for upregulation of proteasomal gene mRNA afterMG132 (proteasome inhibitor) treatment. So far, we found that the 5´-UTR of thegenes is responsible for this induction. We are now looking for the precise motifinvolved in this regulation.

© Patrick MassonISBN 91-7265-837-1

3

TABLE OF CONTENT

LIST OF PUBLICATIONS 4

ABBREVIATIONS 5

INTRODUCTION 6

PROTEASES: HISTORICAL ASPECTS 6THE 20S PROTEASOME 8STRUCTURAL PROPERTIES 9PROTEASOME ACTIVITY, FUNCTION AND LOCALIZATION 10THE 26S PROTEASOME AND THE UBIQUITIN PATHWAY 13STRUCTURE OF THE 19S REGULATORY PARTICLE 13THE UBIQUITIN-PROTEASOME PATHWAY 15THE PA28 ACTIVATOR OR 11S REG COMPLEX 17STRUCTURE AND ACTIVATION OF THE PROTEASOME 17PA28, IMMUNOPROTEASOME AND IMMUNITY 19THE PA28GAMMA 22STRUCTURE AND ACTIVATION OF THE PROTEASOME 22PA28γ FUNCTION 23PA28γ INTERACTING PROTEINS 24ADVANTAGES OF DROSOPHILA MODEL IN OUR INVESTIGATION: 26DROSOPHILA MELANOGASTER AS MODEL SYSTEM 26DROSOPHILA EMBRYOGENESIS 26RNAI TECHNIQUES 26DROSOPHILA MODEL FOR TRANSCRIPTIONAL REGULATION 26

AIM OF THE WORK 29

RESULTS AND DISCUSSION 30

STUDY OF THE INVERTEBRATE PA28 HOMOLOG 30THE DROSOPHILA PA28 : A HOMOLOG OF HUMAN PA28GAMMA (PAPER I) 30THE DICTYOSTELIUM PA28 HOMOLOG (PAPER IV) 35INVESTIGATING THE FUNCTION OF PA28GAMMA 37AN IN VIVO ASSAY TO STUDY THE DROSOPHILA 26S PROTEASOME (PAPER III) 40STUDY OF THE PROTEASOMAL REGULATION AFTER MG132 TREATMENT (PAPER V) 41

CONCLUSIONS AND PERSPECTIVES 43

ACKNOWLEDGMENTS 44

REFERENCES 45

4

LIST OF PUBLICATIONS

This thesis is based on the following publications:

I. Masson, P., Andersson, O., Petersen, U.-M., Young, P.Identification and characterization of a Drosophila nuclearproteasome regulator: a homolog of human 11S REGγ (PA28γ).J. Biol. Chem. 2001 Jan 12; 276(2):1383-90.

II. Masson P., Lundgren J., Young P.Drosophila proteasome regulator REGgamma: transcriptionalactivation by DNA replication-related factor DREF and evidencefor a role in cell cycle progression.J. Mol. Biol. 2003 Apr 11; 327(5):1001-12.

III. Lundgren J, Masson P, Realini CA, Young P.Use of RNA interference and complementation to study thefunction of the Drosophila and human 26S proteasome subunitS13.Mol. Cell. Biol. 2003 Aug; 23(15):5320-30.

IV. Masson, P, Söderbom F, Young, P. (2004)Identification and characterization of a Metazoan ProteasomeRegulator 11 S REG (PA28) in Dictyostelium discoideum.(Manuscript)

V. Lundgren J, Masson P, Zahra Mirzaei, Young P. (2004)Microarray analysis to identify genes of the Drosophilaproteasome regulatory network. (Submitted)

5

ABBREVIATIONS

ATP Adenosine-tri-phosphateCSN2 Signalosome subunit 2COP9 constitutive photomorphogenesis signalosomeDPE Downstream promoter elementDRE Drosophila replication elementDREF Drosophila replication element factorER Endoplasmic reticulumERAD Endoplasmic reticulum associated degradationHCV Hepatitis c virusHECT homologous to E6-AP C terminusMHC Major histocompatibility complexGFP Green fluorescence proteinRNAi RNA interferenceMG 132 Cbz-Leu-Leu-leucinalMCA MethylcouraminLLVY Leu-Leu-Val-TyrLPS LipopolysaccharideLRR Leu-Arg-ArgODC Ornithine decarboxylasePA28 Proteasome activatorPCNA Proliferating cell nuclear antigen11S REG 11S RegulatorRNP RibonucleoproteinRpn Regulatory particle non-ATPaseRpt Regulatory particle triple-A proteinSLE Systemic lupus erythrematosusSNO Strawberry notchTAP Transporter associated with antigen processingUTR Untranslated regionVCP vasolin-containing protein

6

INTRODUCTION

Proteases: historical aspects

Proteases or proteolytic enzymes constitute the largest proteinfamily and their function is to catalyze the cleavage of peptide bonds.This reaction is one of the most frequent protein modification occurringin nature. The initial interest of scientists towards understanding digestionwas the foundation for our current knowledge on enzymology in generaland proteases in particular. During the early 18th century, Reaumurdiscovered that gastric juices were able to degrade meat chemically andnot mechanically which was the current idea held at that time (Reaumur,1761). Continuing the work of Reaumur, Spallanzani found that theisolated gastric juice was more efficient at body temperature rather thanroom temperature (Spallanzani, 1780). In 1836, while also investigatingthe digestive process, the German physiologist Theodor Schwann isolateda substance responsible for digestion in the stomach, and named it pepsin,which is the first enzyme prepared from animal tissue. In 1877, Kühneproposed the general term enzyme (Kühne, 1877). Whereas catalyticproperties and specificity of enzymes were discovered shortly after, ittook until mid-twenties for the scientific community to accept thatenzymes were proteins. In early 1930s Northrop and his collaboratorscrystallized pepsin, trypsin and chymotrypsin and demonstrated the purityof their obtained crystals (Northrop, 1946). Rapidly improving proteinpurification techniques as well as other major discoveries during the lastcentury including the DNA double-helix structure in 1953 (Watson andCrick, 1953) and the genetic code (Crick, 1968), allowed researchers tocharacterize an impressive amount of new proteases. Today severalthousand proteolytic enzymes have been described and characterized.

The importance of proteases in living organisms as well as thegrowing number of publications concerning them directed scientists toclassify them into families. In Drosophila for instance 564 out of the13600 known genes are estimated to be proteases, which represents4,15% of the total amount (Rawlings et al., 2002). In order to achieve aclassification, the protease catalytic residue or the structure of the activesite involved in the cleavage has been used to establish the main groups(Barrett and Rawlings, 1991) and so far, six main groups have beencategorized according to the specific active site. Within a group, familieshave been created based upon common evolutionary origins and common

7

structures. The six groups and examples of proteases within each groupare described below:

- aspartic proteases including pepsin A and pepsin B- metalloproteases including aminopeptidase A- cysteine proteases including caspase-1 to -14- serine proteases including trypsin and chymotrypsin- threonine proteases including the proteasome- unknown proteases

Recently a detailed classification has been realized and a comprehensivedatabase termed MEROPS has been created (Rawlings and Barrett,1999). A second nomenclature termed the EC nomenclature of proteaseslists a total of 13 subgroups.

Cellular pathways are tightly regulated and significantly dependupon the rate of synthesis and degradation of the involved proteins.Proteases are thus required to precisely regulate the crucial pathwaysneeded by the cell in order to survive, both in a qualitative andquantitative way.

In 1953, Simpson established the surprising requirement for energy(ATP) in protein breakdown. Several lines of research were laterundertaken to determine the different functions of ATP served inproteolytic processes. This led in 1980 to the identification of the energy-dependent formation of a covalent linkage between a protein substrateand one or more ATP-dependent proteolysis factor 1 (APF-1) molecules,termed today as ubiquitin (Hershko et al., 1980). Three years later,Tanaka and coworkers identified an additional step requiring ATP for theubiquitin dependent protein degradation to function properly (Tanaka etal., 1983). This second step led to the identification of a large proteasetermed the proteasome whose function is to degrade the ubiquitinatedsubstrate in an ATP-dependent manner (Hough et al., 1986).

Now, the functions of the proteasome are well documented and thiscomplex has been shown to degrade a wide range of cellular proteins.Cell cycle progression for example requires the degradation of manyshort-lived proteins in order to proceed; and the vast majority of them aretargeted to the proteasome. The impressive range of participations of theproteasome would not be possible without the help of different activatorsthat greatly increase its performances within the cell. Studying the precisecontribution of these different activators will provide much moreunderstanding on the overall proteasome activity and functions.

8

The 20S proteasome

The proteasome is a large multicatalytic protease involved in manyintracellular pathways including degradation of aberrant or damagedproteins and inactivation of proteins that play key roles in regulatoryprocesses. The proteasome is composed of four stacked rings, each onecontaining seven subunits, which form a barrel-like structure with a α1-7 β1-7 β1-7 α1-7 stoichiometry (Puhler et al., 1992). The catalytic sitesare located inside a central cavity, which restricts access to specificsubstrates targeted for degradation. Other proteases such as ClpAP orHslU share this quaternary structure, which is very convenient to avoidundesired and uncontrolled peptide hydrolysis throughout the cell (Kesselet al., 1995), (Sousa et al., 2000). After its discovery in humanerythrocytes (Harris, 1968), the 20 S proteasome has been characterizedin a wide variety of organisms, including all the eukaryotes and part ofprokaryotes. Sequenced genome analysis indicates however, that mosteubacteria such as Esherichia coli do not possess proteasomes but insteadshare a proteasome-like system termed HslUV. The HslUV complex hasa diameter of 110 Å and a length of 75Å whereas the diameter and lengthof the archaeal Thermoplasma acidophilum proteasome are 100Å and175Å respectively. Structurally, HslUV consists of only two homo-hexameric rings, explaining the shorter length (fig. 1).

Fig. 1. Crystal structure of E. coli HslUV, and proteasomes from Thermoplasmaacidophilum and yeast respectively. Adapted from (Groll and Clausen, 2003).

9

Structural propertiesTypically, two different genes encode the archaeal proteasome

subunits alpha and beta (Zwickl et al., 1992). The active sites are locatedat the N-terminal threonine of the beta subunits and face the centralchamber. Two narrow openings of approximately 13 Å reside at bothends of the cylinder (Lowe et al., 1995). The eukaryotic proteasome ismade of 14 different subunits, 7 alpha and 7 beta, encoded by 14 distinctgenes that confer a more complex organization (Heinemeyer et al., 1994).The catalytic subunits are also located on the beta subunits but only threeof them are active (β1, β2 and β5 respectively Y, Z, X in the humannomenclature). Despite this difference, they share their catalyticmechanism with an N-terminal nucleophilic threonine, which places theproteasome in the family of Ntn (N terminal nucleophile) hydrolases(Seemuller et al., 1995), (Heinemeyer et al., 1997).

Upon assembly of the eukaryotic complex, the N-terminalextensions located on the active subunits, also termed propeptides, areautocatalytically removed to allow generation of the N-terminal threonineinside the cylinder (Schmidt and Kloetzel, 1997). In addition to thisfunction, Arendt and coworkers have shown that these propeptides alsoplay a role in protecting the N-terminal threonine against N-alphaacetylation (Arendt and Hochstrasser, 1999). In the inactive betasubunits, this propeptide is never cleaved, leading to a structure where thethreonine is masked by the prosequence.

While the Thermoplasma proteasome only possesses one activityas a result of the identical beta subunits, the eukaryotic complex has threedistinct proteolytic activities for each active beta subunits: the β1 subunithas a peptidylglutamyl-hydrolyzing activity also termed postacidic orcaspase-like activity, meaning that the peptide bond hydrolysis occursdirectly after acidic residues. The β2 subunit carries a trypsin-like activitycleaving after basic residues. Finally, a chymotrypsin-like activity hasbeen assigned to the β5 subunit. This increased number of activities ispartially due to the differences in the structural architecture of theproteolytic chambers. Studies using calpain I inhibitors together with theyeast proteasome crystal structure identified essential residues for thedifferent activities (Groll et al., 1997). The S1 pocket (specificity) ismainly shaped by five conserved residues with the residue 45 forming thebottom of the pocket and largely contributing to its character. The residue45 is an arginine in the β1 subunit, a positively charged amino acid wellsuited for the peptidylglutamyl-hydrolazing activity. Concerning β2, theresidue is a glycine forming a spacious S1 pocket, which is convenientfor large residues, and glutamine 53 at the bottom of the pocket alsocontributes to the trypsin-like activity. Finally, β5 possess a methionine at

10

position 45, an apolar residue ideal for chymotrypsin-like activity. Inaddition, inhibitor studies using two analogue compounds MB1 and MB2(fig. 2) differing only in their P3 (position 3) and P4 residues showed thatresidues other than P1 also influence the degradation (Groll et al., 2002).This result highlights the importance of the whole inner chamberimposing a physical constraint to the substrate rather than the sole S1pocket.

Fig 2. Chemical structures of the twoPeptide Vinyl Sulfone Proteasomeinhibitors MB1 Ac-PRLN-VS (top)and MB2 Ac-YLLN-VS (bottom)differing on their P3 and P4. Adaptedfrom Groll et al. 2002.

Finally, mutational studies revealed that neighboring subunits interferewith the functions of the catalytic subunits (Groll et al., 1999).

Proteasome activity, function and localizationThe cleavage fragments that are generated in vitro by the

mammalian 20S proteasome vary in length between 3 and 22 amino acidswith an average of 7-8 amino acids (Kisselev et al., 1999). In other terms,the products leave the catalytic chamber before complete degradation toindividual amino acids. Additional proteases have been suggested tofunction downstream of the 20S proteasome to further process fragmentsand produce single amino acids, such as the tricorn protease (Tamura etal., 1996). By itself, the 20S proteasome only poorly degradesoligopeptides and unfolded proteins, and the purified complex hasnegligible activity towards native proteins (Tanaka et al., 1986). The 20Stypically requires binding of additional activators to be able to degradefolded proteins or long polypeptides. Among these, the 19S regulatorycomplex binds to the 20S core and induces protein degradation through

11

the ubiquitin pathway that will be discussed later. Other types ofactivators exist in the cell with one member of the PA28 family being thefocus of this thesis. The proportions of these different activatingcomplexes is believed to depend on the cell type, the cell cycle phase andmetabolic conditions (Brooks et al., 2000). The free 20S core particlehowever has been reported to be in large excess in the cell (Brooks et al.,2000), possibly suggesting either a specific role in proteolysis in absenceof any bound regulatory particles, or the excess of 20S prevents unwantedcompetition between the different regulators in gaining access to the 20Sactivity.

In HeLa cells, the 20S proteasome is very abundant and constitutesapproximately 0.6% of the total cellular protein (Hendil, 1988). Anumber of studies indicate that the proteasome is widespread both in thecytoplasm and the nucleus. The precise distribution is very difficult todetermine, and probably depends on cell type, growth conditions, andmethods used for detection (Wojcik and DeMartino, 2003). Electronmicroscopy studies have determined that 14% of cytoplasmicproteasomes is associated to the outer surface of the endoplasmicreticulum in rat hepatocytes (Rivett et al., 1992). The importance of thisfraction is probably due, at least in part, to the involvement of theproteasome in the ERAD process (Endoplasmic Reticulum-associateddegradation) consisting of the degradation of abnormal or misfoldedproteins from the ER. Subfractionation of the microsomes shows that theproteasomes are associated with the smooth endoplasmic reticulum andwith the cis-golgi but are practically absent from the rough ER (Palmer etal., 1996). Another explanation for the presence of the proteasome in theER comes from its involvement in antigen presentation. Indeed, it canassociate with the ER membrane where it could in theory come in contactwith the TAP transporter (transporter associated with antigen processing),an essential complex required for the MHC class I peptide presentation(Rechsteiner et al., 2000).

The proteasome is also present in the nucleus. The crystal structureof the mammalian proteasome identified four putative nuclearlocalization signals on the alpha subunits 1, 2, 3, and 4 with theconsensus sequence K-K(R)-X-K(R) that could in principle interact withthe importin-α for nuclear import (Unno et al., 2002). In the nucleus, theproteasome is localized in various regions including the euchromatin andthe periphery of the heterochromatin and nucleoli. However, this overalldistribution is very dynamic and is very cell specific. Indeed, during cellcycle progression, there is an increase of nuclear proteasome from G1 toG2 phases. During prophase, 20S proteasomes accumulate around thecondensing chromosomes whereas in late anaphase they colocalize withα-tubulin of the spindle fibers. In telophase and early interphase of the

12

daughter cells, an intensive nucleus staining was still observed(Amsterdam et al., 1993). Alternatively, during programmed cell death,proteasomes from rat ovarian granulosa cells are removed from thenucleus and accumulate within the apoptotic blebs at the periphery of thecell (Pitzer et al., 1996).

The great dynamic distribution of proteasomes together with theirability to form various complexes with different regulators greatlyincreases the complexity of this critical complex and may also increasethe number of essential proteolytic roles that the 20S proteasome canfulfill within the cell. In the next part I will focus on the two most studiedregulatory particle families, the 19S and the PA28 family and I willdescribe the reasons why they render the 20S core particle a much moreefficient and specific destruction machinery.

13

The 26S proteasome and the ubiquitin pathway

The 19S regulatory particle is composed of at least 18 subunits and canattach to one or both ends of the 20S core particle in the presence of ATP.Binding on both ends leads to the structure called the 26S proteasome.One problem in the field is the multiple names given to each subunit indifferent organisms. For instance, one subunit is termed S5a in human,Rpn10 or Mcb-1 in Yeast, and p54 in Drosophila.

Structure of the 19S regulatory particleUnlike the 20S core, the 19S complex structure is not resolved.

Indeed, only a low-resolution structure of the complexes was provided byelectro-microscopy whereas precise positioning of individual subunitsremains largely unknown (Walz et al., 1998). Purified yeast proteasomemutants (Glickman et al., 1998) instead provided the data concerning thestructure of the complex. The 19S complex is composed of a baseimmediately adjacent to the 20S core and a distal lid (fig. 3).

Fig. 3. Subunit distribution of the 19S proteasome activator. The 19S represented as17 subunits subdivided into the base and the lid (derived from Ferrell et al. 2000).

The base of the 19S is composed of 8 subunits including 6 ATPases plusthe two largest subunits S1 and S2. The six ATPases (termed S4, S6, S6´,S7, S8, and S10b in human and Rpt1 to Rpt6 in yeast) belong to thefamily of AAA-ATPases (ATPases associated with a variety of cellular

14

activities) with a shared conserved region of approximately 400 aminoacid residues. Despite the unknown structure, several lines of evidencesuggest that these ATPases subunits form a heterohexameric ring. Severalcomplexes such as VCP, CDC48 from the AAA-ATPase family areknown to also form hexameric rings and this may be a conserved feature.Sequence alignments show a similar architecture of the six genes with acentral nucleotide-binding domain about 60% identical betweenmembers, and a 40% identity concerning the last 150 amino acids.

Several regulatory functions have been attributed to the base.Among these, the base is thought to perform two essential steps forprotein degradation, unfolding and translocation of the protein substratesto the 20S proteasome. The unfolding occurs at the surface of the ATPasering and the translocation follows after the ATP-dependent unfolding(Navon and Goldberg, 2001). The base is also able to open the gate of the20S core in order to allow access to the catalytic chamber byrearrangements of the alpha subunits. Furthermore, some specificfunctions have been identified for certain subunits of the base. Forinstance, the subunit S6´ has been proposed to associate withpolyubiquitin chains (Lam et al., 2002), which is the signal that iscovalently bound to the substrate and targets it to the 26S proteasome fordegradation, see ubiquitin section below. In conclusion, the base iscomposed of a hexameric ring where subunits cooperatively work tounfold and translocate substrates to the 20S cylinder, but also mayfunction individually or in smaller complexes for different purposes.

The S5a subunit located between the base and the lid is believed tostabilize the two subcomplexes (Glickman et al., 1998). Indeed, deletionof S5a in yeast causes dissociation of the lid and the base. The entire 26Sproteasome displays a strong and specific activity for polyubiquitinchains, but S5a is one of the few individual subunits identified to be ableto bind ubiquitin chains through two binding sites termed PUbs (Young etal., 1998) or UIM for ubiquitin interacting motif (Hofmann and Falquet,2001).

The lid of the metazoan 19S particle is composed of at least eightsubunits: S3, S9, S10a, S11, S12, S13, S14, p55 and shares strongsimilarities with another multiprotein complex, the COP9 (constitutivephotomorphogenesis) signalosome (Henke et al., 1999). Several proteinsof both complexes contain a characteristic PCI domain, a structural motifimportant for complex assembly. The subunit S9 physically interacts withsubunit CSN2 of the signalosome suggesting a potential link betweenboth complexes (Lier and Paululat, 2002). Besides this homology,specific functions have been assigned to several of the subunits formingthe lid. For instance, deletion of Rpn3 subunit in budding yeast (S3)caused a metaphase arrest. The Rpn3 mutants strongly altered

15

degradation of several substrates by the 26S proteasome such as G1-phase cyclin Cln2, S-phase cyclin Clb5, and the anaphase inhibitor Pds1that are essential for cell cycle progression (Bailly and Reed, 1999).Additionally, mutational studies of another lid subunit Rpn9 (S11) alsoshowed defects in cell cycle progression in yeast by delaying substratesdegradation by the 26S and increasing multiubiquitinated protein pool at37°C (Takeuchi and Toh-e, 2001). Finally, a mutation in the Rpn11 (S13)gene in yeast results in a cell cycle arrest, an over replication of nuclearand mitochondrial DNA, as well as an altered mitochondrial morphology(Rinaldi et al., 1998). These studies show a fundamental role for the lid ofthe 19 regulatory particle, especially during the cell cycle progression.However, additional studies are needed to understand the range offunctions or specific interactions that the 19S probably possesses.

The ubiquitin-proteasome pathwayDespites the reported ability of the proteasome to degrade several

substrates without ubiquitin, such as oxidized proteins (Grune et al.,1997) or ornithine decarboxylase ODC (Gandre and Kahana, 2002), mostof the substrates are degraded by the proteasome/ubiquitin pathway.Ubiquitin is a very conserved protein consisting of 76 amino acids (8kDa) found to be involved in the destruction of a large number ofproteins, by covalently attaching to its substrate. Hershko and coworkersproposed a model where ubiquitin acts as a destruction signal in an ATP-dependent manner (Hershko et al., 1980). Once proposed, this model ledto an extensive search for the enzymes or complexes responsible for theactual degradation of these ubiquitin tagged proteins. The 26Sproteasome was shown shortly after to be involved in this conservedpathway (Hough et al., 1986). Today, ubiquitin conjugation is recognizedas a multifunctional signaling mechanism with regulatory significancecomparable to phosphorylation. Most of the substrates are efficientlytargeted to the 26S proteasome when they are attached to a polyubiquitinchain consisting of at least 4 ubiquitin molecules (Pickart, 2004). Recentstudies show that monoubiquitination might serve others purposes, suchas re-localization (Stelter and Ulrich, 2003). The ubiquitination of asubstrate has been extensively studied and is well understood. It requiresthree types of enzymes, named E1s, E2s and E3s. The first ubiquitin isactivated on its C-terminal by E1s (Ub activating enzymes). The E2 (Ubconjugating enzymes) transiently carries the activated ubiquitin to itscorresponding E3 (Ub ligating enzyme). Finally, the E3 ligases arethought to be most directly involved in substrate recognition. The E3s canbe divided in two main families: the HECT domain E3s and the RINGE3s. Studies in yeast showed a physical association of certains ubiquitin

16

ligases with the 26S proteasome suggesting a participation in the deliveryof substrates to the proteasome (Xie and Varshavsky, 2000).

Polyubiquitinated proteins are then recognized by the 26Sproteasome and degraded. The recognition by the 26S is still not knownprecisely, but the S5a subunit was originally suspected to play a role inubiquitin chain recognition since it contains two UIM, ubiquitininteracting motifs, characterized by alternative leucine and alanineresidues LALAL (Young et al., 1998). This model is in accordance withdeletion study of the Drosophila S5a that showed lethality and multiplemitotic defects (Szlanka et al., 2003). However, knockout of S5a subunitin yeast displays a viable phenotype with minor defects in proteindegradation (Wilkinson et al., 2000). In that particular case, it wasproposed that several subunits might be involved in the recognition, andthus deletion of S5a is rescued by other cellular components. Indeed, theS5a mutant is lethal in combination with mutations of the otherproteasomal component encoded-genes Rpn1, Rpn11 and Rpn12(Wilkinson et al., 2000). Additionally, chemical cross-linking studies inmammalian suggested that one ATPase subunit, namely S6´, interactswith polyubiquitin chain and may thus be involved in the degradationsignal recognition (Lam et al., 2002).

In conclusion, despites the large number of papers publishedconcerning the structure and function of the 26S proteasome, manyquestions remain unanswered. Particularly, the precise mechanisminvolved in the polyubiquitin chain recognition is not yet wellunderstood, the exact role of ATP in the hydrolysis is still unclear, andspecific functions for each 19S subunit remain to be discovered. Anumber of questions should be resolved with the determination of thewhole 26S proteasome structure.

17

The PA28 activator or 11S REG complex

In 1992, two groups purified and biochemically characterized acomplex able to enhance the proteasomal degradation of fluorogenicpeptides in an ATP-independent manner (Dubiel et al., 1992; Ma et al.,1992). However, this activator was unable to enhance proteasomaldegradation of native proteins, such as bovine serum albumin orubiquitin-lyzozyme conjugates by the 20S proteasome. Native gelelectrophoresis of the human activator showed a complex ofapproximately 200 kDa, and two-dimensional electrophoresis revealedthat this complex is composed of two different subunits. Cloning andsequencing of the corresponding genes revealed that the two subunitsshare about 50% similarities in their primary sequence, and have amolecular mass of 28 and 29 kDa respectively (Mott et al., 1994). Twodifferent names have been attributed to the complex, De martino and hiscolleagues designated it proteasome activator PA28, whereas Rechsteinerand colleagues named it 11S REG since the regulator has a sedimentationcoefficient around 11S. Electron-microscopy studies have shown that theactivator occurs as an oligomeric ring able to bind to both ends of the 20Sproteasome (Gray et al., 1994), or at only one end of the 20S proteasomewhereas the other end is associated to the 19S regulatory particle, and thiscombined structure has been termed the hybrid proteasome.

Structure and activation of the proteasomeHaving a size of approximately 200 kDa, work was undertaken to

determine the exact number of subunits forming the complex. Usingdensitometric quantification of immunoprecipitates obtained with eitheranti-PA28α or anti-PA28β antibodies, Ahn and coworkers observed a 1:1stoichiometry and suggested a hexameric structure with 3 alpha and 3beta subunits (Ahn et al., 1996). This result was confirmed in anotherstudy where the pattern of crosslinked PA28 using bis(sulfosuccinimidyl) substrate indicated a hexamer, probably composed ofalternating alpha and beta subunits (Song et al., 1996). However, laterinvestigations differed from these initial results. Indeed, coexpression ofwild type PA28α and PA28β in bacterial cells revealed a large majorityof alpha3-beta4 heptamers using ESI-TOF mass spectrometry (Zhang etal., 1999). The human recombinant PA28α crystal structure determined ata 2,8 Å resolution also confirms a heptameric structure (fig. 4) (Knowltonet al., 1997). The monomer is predominantly helical with four long α-helices. The sequence between the helices 1 and 2 is made of 39 residuesforming a flexible loop not resolved on the structure. This disordered

18

region has been termed homolog-specific insert since it constitutes themost divergent region between PA28α and PA28β in term of amino acidsequence. Deletion studies showed that this insert is not required forbinding and activating the proteasome by PA28α, whereas removal of thePA28β insert reduces binding of this subunit as well as PA28αβ bindingto proteasomes (Zhang et al., 1998b). The PA28α insert might play a rolein interacting with others protein complexes since it contains a KEKEmotif, a stretch of alternating glutamate and lysine residues suggested tobe important in protein-protein interactions. This motif was also found inother proteins such as hsp90, hsp70, calnexin and two subunits of theproteasome (Realini et al., 1994). In addition to this insert, the PA28complex contains two essential parts allowing interaction and activationof the 20S proteasome. Screening single-site PA28α mutants for alteredactivity identified a region between Arg-141 and Gly-149 critical forproteasome activation. The last ten residues were also found to beimportant for binding/activation because Pro240 together with Met247and the C-terminal residue Tyr249 produced inactive heterooligomerswhen mutated (Zhang et al., 1998a). This result was confirmed by thecrystal structure showing that the two regions were immediately adjacent.

Fig. 4. Crystal structure of recombinant human PA28α (Knowlton et al., 1997).A. side view.B. top view.

One challenge in the field is to determine the precise mechanismresponsible for proteasome activation by PA28 association. As mentionedearlier, the PA28 is able to selectively enhance short peptide degradationbut not native proteins. For instance, LLVY-MCA fluorogenic peptidewas cleaved approximately 100 fold, whereas the peptide LLE-MCAcleavage was only enhanced 22 fold. The crystallisation of PA28 togetherwith the 20S proteasome would give a clear view of the activationmechanism, but unfortunately no group has succeeded in this task.Instead, Whitby and coworkers managed to crystallise the PA26 togetherwith yeast 20 S proteasome (Whitby et al., 2000). The proteasome

A B

19

activator PA26 has been isolated from Trypanosoma brucei (Yao et al.,1999), and its structure ressembles very much the PA28α structure evenif the two sequences only share 14% similarity. The result of this PA26-proteasome structure suggested a conformational change of the outeralpha subunits of the 20S particle without affecting the beta subunitsconformations. In contrast, several biochemical studies have suggested achange in the beta subunit conformation after PA28 binding. Amongthese, a study on peptide cleavage of the 20S together with PA28 showedthat addition of PA28 does not change the cleavage pattern but ratherincrease dual cleavage of 25 mer peptides, instead of single cleavageobserved with 20S proteasome alone (Dick et al., 1996). The authorsproposed a model where the beta subunits were affected by PA28allowing dual-cleavage to occur. These results however are notincompatible with the model where PA28 affects the degradation only byselective peptide entrance onto its channel. It may also increase the speedof the cleaved product to exit from the proteasome in order to rapidlyempty the catalytic centers for faster degradation.

PA28, immunoproteasome and immunityShortly after the characterization of PA28, several studies

suggested a role in immunity. The first evidence appeared when bothmRNA and protein levels were found to be strongly induced upontreatment with the immune cytokine γ-interferon (Ahn et al., 1995),(Groettrup et al., 1995). γ-interferon is well known to induce genesinvolved in MHC class I presentation such as the transporter associatedwith antigen processing, TAP. Furthermore, the following findingsconcerning PA28 reinforce this idea of a functional link to immunity:• The complex has been identified only in animals possessing an adaptiveimmune system, from zebra fish to mammals.• Within the cell, the complex is located both in the cytoplasm and thenucleus but upon γ-interferon treatment, the complex is relocalized tospecific structures called PML bodies, these bodies have been shown toparticipate in the immune response (Fabunmi et al., 2001).

In addition to PA28, organisms with adaptive immune systempossess three additional proteasomal beta subunits LMP2, LMP7 andMECL1 (also called LMP10). These subunits are able to replace the threeconstitutive active subunits beta1, beta5, and beta2 after γ-interferoninduction (Groettrup et al., 1996). The newly incorporated subunitschange the structural and catalytic properties of the proteasome, leadingto a new complex termed the immunoproteasome (Aki et al., 1994).Studies focusing on the immunoproteasome assembly revealed acooperative incorporation of the subunits: MECL-1 requires LMP2 for

20

efficient incorporation, and preproteasomes containing LMP2 andMECL-1 require LMP7 for efficient maturation (Griffin et al., 1998),(Groettrup et al., 1996) suggesting an important role of all three subunits.The enzymatic activity is also modified by the three subunits but theprecise contribution of each catalytic site and the overall modificationsremain controversial. The incorporation of LMP2 and LMP7 subunits hasbeen reported to down-regulate cleavage C-terminal of acidic residuesand increase the cleavage C-terminal of hydrophobic residues (Driscoll etal., 1993), (Aki et al., 1994). However, other groups did not observe asignificant change of the chymotrypsin-like activity (Ustrell et al., 1995).Ehring and coworkers favor a model where the cleavage site was notdetermined by the P1 position alone, corresponding to the residue situateddirectly in front of the cleavage site (Ehring et al., 1996). Finally, somereports mention contradictory results with a decrease in thechymotrypsin-like activity of the 20S after γ-interferon treatment (Boes etal., 1994), (Kuckelkorn et al., 2002). These contradictory results may beexplained by the differences in proteasome preparations and the differentorganisms used during the various studies. However, the use offluorogenic peptides and synthetic polypeptides clearly indicates adifference in activity between constitutive proteasomes andimmunoproteasomes.

The fact that the proteasome releases short peptides led severalgroups to study the proteasomal antigen generation capacity. Severalstudies established that purified 20S proteasomes were able to produceantigenic peptides (8 to 9 residues) from polypeptides or intact proteins(Dick et al., 1994), (Niedermann et al., 1995). Furthermore, the use ofpeptide aldehyde inhibitors of the proteasome coupled with introductionof ovalbumin into lymphoblasts demonstrated that the proteasome isessential for the presentation on MHC class I molecules of an ovalbumin-derived (Rock et al., 1994). Many groups evaluated theimmunoproteasome contribution in antigen presentation but contradictoryresults emerged from these studies. First, work on humanlymphoblastoid cell lines LMP2 and LMP7 deficient revealed that thesetwo subunits are not essential for antigen presentation (Arnold et al.,1992). Yewdell and coworkers made the same type of observation andshowed that cells lacking LMP2 and LMP7 could still present antigenicpeptides derived from viral proteins (Yewdell et al., 1994). Nevertheless,LMP2 depleted mice have reduced levels of CD8+ T lymphocytes andgenerate fewer influenza nucleoprotein-specific cytotoxic T lymphocytesprecursors (Van Kaer et al., 1994). In agreement to this last finding, micewith a deletion of the gene encoding LMP7 have reduced level of MHCclass I cell-surface expression and present the endogenous antigen HYinefficiently (Fehling et al., 1994). So, despite the lack of marked effects

21

in cell lines, the data supports an essential role of the immunoproteasomein MHC class I peptide presentation.

The PA28αβ is also strongly induced after γ-interferon treatmentand several groups studied the influence of the activator in antigenprocessing. In order to test this hypothesis, the proteasomal digestion of a25-mer from murine cytomegalovirus pp89 was assessed in presence orabsence of PA28 and the results showed a change both in quality and inquantity of peptides produced (Groettrup et al., 1995). Furthermore, adetailed study on the mechanism of action revealed that proteasomalgeneration of MHC class I epitopes was optimized in presence of PA28by inducing a coordinated double cleavage mechanism (Dick et al.,1996). Another clear example of the role of PA28 in antigen processingwas obtained by Sun and coworkers who found that melanoma cellslacking both PA28 and immunoproteasomes did not display a specificepitope to CTLs derived from the melanoma antigen tyrosinase-relatedprotein 2 (TRP2). However, PA28 expression completely rescued theepitope presentation (Sun et al., 2002).

These studies mentioned above demonstrate an important functionof immunoproteasomes and PA28 in the antigen presentation pathway,but the exact contribution of both complexes and the link between themduring immune response remains difficult to appreciate. For instance,mice lacking both PA28α and PA28β showed normal processing ofovalbumin and influenza A virus derived antigens (Murata et al., 2001),(Sun et al., 2002). However, the knockout mice completely loose theability to process the TRP2-derived peptide as observed by Sun andcoworkers (Murata et al., 2001). In addition, the PA28 has been shown inseveral occasions to be able to enhance antigen presentation of some CTLepitopes independently of changes in proteasome composition (van Hallet al., 2000), (Schwarz et al., 2000).

In conclusion, despite the increasing understanding of PA28 andimmunoproteasomes roles in antigen presentation, several aspects of theirbiological functions remain to be discovered.

22

The PA28gamma

Systemic Lupus Erythrematosus (SLE) is an autoimmune diseaseaffecting a large number of organs, particularly joints, skin, kidneys,heart and lungs. This disease is also characterized by an array ofautoantibodies directed against the nucleosome, its DNA and histonecomponents as well as other self-proteins such as PCNA (proliferatingcell nuclear antigen) or small nuclear RNPs. One of these proteins calledKi autoantigen was identified in 1981 as a soluble nuclear antigen from aJapanese SLE patient (Tojo et al., 1981). In a clinical and serologicalstudy carried out by Riboldi and coworkers on 516 subjects, anti-Kiautoantibodies were found in 12% of patients having SLE. Interestingly,this study also showed a significant correlation between Ki and PCNAantibodies (Riboldi et al., 1987).

The use of anti-Ki antibody as a probe led to isolation of thebovine and human corresponding cDNAs. The two proteins share 95%sequence homology and the gene expression is correlated withproliferative conditions, since the mRNA abundance in mouse fibroblastcells greatly increased 4 hours after serum stimulation whereas inquiescent cells mRNA level was almost non-detectable (Nikaido et al.,1989).

By sequence comparison, the Ki autoantigen was found to sharestrong homology with both subunits of the PA28 activator describedearlier. Ki shares 41% amino acid sequence identity with PA28α and34% with PA28β (Ahn et al., 1995) and was found tocoimmunoprecipitate as a homopolymer together with the 20Sproteasome and reversibly associate with the 20S on glycerol gradients(Tanahashi et al., 1997). In accordance with these properties, the Kiautoantigen has been renamed PA28γ or 11 S REGγ.

Structure and activation of the proteasomeThe PA28γ was first purified by Sakamoto and coworkers from

rabbit thymus (Sakamoto et al., 1989). They found a molecular mass of32 kDa for the monomer and 224 kDa for the native protein, so thecomplex is likely to be heptameric. These results have been confirmed bypurification of a 6-his-PA28γ fusion protein and analysis the resultingfractions on SDS-page. In that case, the molecular mass was calculated as215 kDa confirming the heptameric structure of the ring (Wilk et al.,2000).

23

Even if the structure of the complex is unknown, the primarysequence possesses all the features for being a PA28 family member andsupposes a similar shape as PA28. Indeed, the activation loop isconserved between human PA28α and PA28γ and the essential prolineand tyrosine located in the C-terminal region are also conserved.Concerning the homolog-specific insert region, the KEKE motif is notpresent but there is, around the middle of the loop, an amino acidsequence homologous to a nuclear localization signal (Takasaki et al.,1996) suggesting a distinct function from the hypothetical protein-proteininteraction domain.

The enhancement of the proteasomal activity by PA28γ for smallpeptides is very different compared to PA28αβ. Unlike PA28, binding ofthe mammalian recombinant PA28γ strongly stimulates the trypsin-likeactivity of the proteasome whereas the chymotrypsin-like activity wasoriginally reported to be almost unaffected by the presence of PA28γcompared to the activation pattern observed with PA28αβ (Realini et al.,1997). My recent work with Drosophila indicates a more complicatedpicture with PA28γ functioning as an actual inhibitor to thechymotrypsin-like activity. Surprisingly, a substitution of Lys188,completely changes the activation pattern of PA28γ that becomes similarto PA28 pattern (Li et al., 2001). The fact that the lysine is located almoston the top of the ring implies that either the channel promotes or inhibitspeptide transfer to the 20S proteasome or there are long-rangeconformational changes along the complex that result in changing thebeta subunits structure.

Overall, the mammalian PA28γ is able to activate the proteasomebut to a lower extent than the PA28 complex (Realini et al., 1997).

PA28gamma functionIn 1989, a first hint of the biological function came with a study

performed by Nikaido and coworkers that showed that the expression ofPA28γ (termed the Ki antigen by the authors in relation to the formationof autoantibodies in SLE) was growth regulated with an expressionpattern similar to c-myc. The level of mRNA was barely detectable inquiescent cells whereas an increase was observed in the cell by 4 hoursafter serum stimulation. They suggested a role related to cell growthregulation and proliferation (Nikaido et al., 1989). Once the homologybetween the Ki antigen and the PA28 established, several groups studiedthe effect of γ-interferon on the level of PA28γ mRNA or protein, inorder to evaluate the putative role of PA28γ in immunity. Firstinvestigations showed a very slight increase in mRNA levels (Ahn et al.,

24

1995), (Jiang and Monaco, 1997), but later results contradicted thesefindings, and two reports concluded a complete loss of PA28γ proteinlevel after γ-interferon without affecting the mRNA level (Tanahashi etal., 1997) and a reduced protein level in liver of lymphocyticchoriomeningitis virus-infected mice respectively (Khan et al., 2001)suggesting a function rather independent of the immune response.The most valuable study in the search for the activator function wasprobably undertaken by Murata and coworkers that generated PA28γdeficient mice. The lack of the complex did not give any appreciableabnormalities to the newborn mice, but resulted in smaller body sizewhen compared to wild-type mice (Murata et al., 1999). The analysis ofthe PA28γ depleted cultured embryonic fibroblasts performed in the samestudy revealed growth retardation when compared to wild type cells.Interestingly, PA28γ depleted cells were slightly larger than wild typecells. To examine the cause of growth retardation, Murata and colleaguesstudied cell cycle progression by flow cytometry. The results showed anincrease of cells in G1 and a decrease in S phase for PA28γ -/- cells,suggesting an important function for the entry into S phase (Murata et al.,1999). This apparent role in cell cycle progression agrees well with theobservation that cells in various thyroid neoplasms show an abnormalhigh expression of PA28γ especially where the cells are rapidly growing(Okamura et al., 2003).

PA28gamma interacting proteinsA fair amount of information concerning the in vivo function of the

complex arose from different studies involving protein-proteininteractions. Indeed, two-hybrid screen experiments have in severaloccasions found PA28γ as major binding protein to the bait used,supposing that PA28γ interacts with many cellular factor other than theproteasome, and all these data might give some hints on the intracellularfunction of the complex. Here is a brief summary of these findings:

-Interaction with Caspases -3 and -7: screening of a human brainadult cDNA library with a constitutive active mutant of caspase 7 led tothe isolation of PA28γ. Subsequent work showed that human PA28γ iscleaved by both caspase –3 and –7 through a DGLD cleavage site locatedin the homolog-specific insert region of human PA28γ. This studyshowed that human PA28γ is an endogenous substrate for these caspases(Araya et al., 2002).

- Interaction with Emerin: using a high-stringency yeast two hybridmethod to screen a human heart cDNA library, Wilkinson and coworkersidentified PA28γ as one of 5 candidates interacting with Emerin. Emerinis a nuclear membrane protein that interacts with lamin A/C at the nuclear

25

envelope. The function of Emerin is largely unknown (Wilkinson et al.,2003).

-Interaction with MEKK3: MEKK3 is a MAPK kinase kinase thatacts on the JNK and p38 activation pathways that are induced by cellularstress including UV and γ irradiations. The PA28γ was isolated in a twohybrid screen using MEKK3 as bait. Once the interaction identified, theauthors confirmed the relevance of the binding by GST pull-down assay,showed that MEKK3 was able to increase PA28γ protein expression incos-7 cells, as well as phosphorylate this one. These results suggest a rolefor PA28γ in response to various stresses (Hagemann et al., 2003). Thephosphorylation is of particular interest because the activation of 20Sproteasome by PA28 in rabbit reticulocytes has been proposed to requirephosphorylation (Li et al., 1996).

-Interaction with the hepatitis virus C core protein: once again, theuse of two-hybrid screen identified PA28γ as binding factor, this time ofthe hepatitis virus C core protein. This interaction was later demonstratedin cell culture as well as in the liver of HCV core transgenic mice.Interestingly, knock-out of PA28γ led to the export of the HCV coreprotein from the nucleus to the cytoplasm, whereas overexpression of thegene increased degradation of the HCV core protein. These resultssuggest the indirect role or “undesired” function of PA28γ is to interactwith the HCV core protein and direct it to the nucleus where it can exertits pathogenesis (Moriishi et al., 2003).

-Additional interactions: recently, a large-scale study wasperformed on 10,623 predicted transcripts from Drosophila (Giot et al.,2003). Using two-hybrid based protein interaction-map, a total of 4679proteins were found to share 4780 interactions in a high confidence map.Concerning the Drosophila homolog of PA28γ, a total of ten interactionswere identified, mainly with unknown genes, but only one with a highconfidence. This last one was the interaction with CG3162 (snRNP U2component), a gene that encodes a product with pre-mRNA splicingfactor activity involved in nuclear mRNA splicing, via the spliceosome.

All these various data obtained by yeast two-hybrid method seemto indicate that the PA28γ plays different functions in addition to theregulation of the G1 to S phase transition and give insights in what couldbe these functions.

26

Advantages of Drosophila model in our investigation:

Drosophila melanogaster as model systemThe Drosophila genome size is approximately 180Mb whereas the humangenome consists of 3200Mb. Although the two genomes have a vast sizedifference, the number of genes identified is much closer between the twogenomes. Indeed, the number of genes in human is estimated to varybetween 30000 and 40000 whereas Drosophila may possess 13600 ormore genes.

Drosophila embryogenesisThe fruit-fly Drosophila melanogaster is probably the most popularmodel system for the study of developmental processes for severalreasons. First, Drosophila is very convenient and powerful as a modelorganism because it has short life cycle of two weeks, making it possibleto study numerous generations in a relative short time. Also, it is largeenough that many attributes can be seen with the naked eye or under low-power magnification. Moreover, it has a very long history in biologicalresearch (since the early 1900s) and there are many useful tools tofacilitate genetic study, including a detailed literature available describingthe various stages of the Drosophila development.

RNAi techniquesRNA interference (RNAi) has emerged as a powerful tool for thesilencing of gene expression in animals. RNAi is mediated by smallinterfering RNAs (siRNAs) 21 to 25 nucleotides long originally producedfrom larger double stranded RNAs in vivo through the action of Dicer(Agrawal et al., 2003). In the second step, the siRNAs join an RNasecomplex, RISC (RNA-induced silencing complex), which acts on thecognate mRNA and degrades it. Because of its specificity and efficiency,RNAi is being considered as a simple and rapid method not only forfunctional genomics, but also for gene-specific therapeutic activities thattarget the mRNAs of disease-related genes.

Drosophila model for transcriptional regulationIn this section, I will present the opportunities of working in transcriptionregulation using Drosophila as model system, and I will mainly focus ontwo aspects using examples relevant to my work:

-First, transcription factors are well characterized in this organism,and a lot of available information is unique and has not yet beendiscovered in others species, or only partially. The well-definedDRE/DREF system is one example,

27

-second, the pathways are typically simpler than in mammals withfewer genes involved making Drosophila a powerful tool to studytranscription. The E2F transcription factor family is well suited as anexample of this characteristic.

The DRE/DREF SystemIn 1993, Hirose and coworkers compared the promoters of two

genes involved in DNA replication, PCNA (proliferating cell nuclearantigen) and DNA polymerase α. This sequence analysis led to theidentification of a conserved palindromic sequence of 8 nucleotides,TATCGATA, present in both promoters (Hirose et al., 1993). This motifwas termed DRE standing for DNA replication related element based onthe important roles that PCNA and polymerase α have in DNA synthesis.The efficiency of DRE elements has been evaluated and confirmed bothin transgenic flies and cultured cells (Hirose et al., 1999). Gel mobilityshift assays detected the presence of a protein factor in nuclear extracts ofcultured Drosophila Kc cells able to specifically bind to DRE sequences.The factor, named DREF, is a transcription factor regulatingproliferation-related genes in Drosophila by binding to DRE regulatoryelements. DREF is a protein of 701 amino acid residues with a molecularweight of 80 kDa (Hirose et al., 1996). Deletion analysis indicated thatthe DNA binding domain was located in the basic amino acid-rich regionbetween residues 16 and 105. Immunocytochemical analysisdemonstrated the presence of DREF polypeptides in the nuclei after theeighth nuclear division cycle, suggesting that a nuclear accumulation ofDREF was important for the coordinate zygotic expression of DNAreplication-related genes carrying DRE sequences (Hirose et al., 1996).

The E2F pathwayThe E2F transcription factors play various roles in controlling

entry and progression through the S phase of cell cycle, apoptosis anddifferentiation. The S phase, where DNA synthesis occurs, requires theactivation of a large set of genes. The E2F transcription factor familyregulates many of these genes.

Human E2F family:The human family of E2F transcription factors is composed of at least 6members, E2F1-6, and each one heterodimerizes with one of the three DPproteins (DP1-3). All combinations of E2F/DP complexes can exist in-vitro (Dyson, 1998). Based on their transcriptional properties as well astheir sequence homologies, the E2F family can be divided into threesubgroups. E2F-1, E2F-2, and E2F-3 are bound to their target genes invivo when expressed during late G1 and S-phase, activating them and

28

driving cellular proliferation (DeGregori, 2002). Overexpression of eachprotein is sufficient to drive resting cells into cell cycle (DeGregori et al.,1997) and depletion of all three at the same time prevents entry into S-phase. For these properties this subgroup has been classified as activatorE2Fs and their expression is maximal as cell approach G1-S phaseboundary. On the contrary, E2F4 and E2F5 are uniformly expressedthroughout the cell cycle with significant levels during G0, and theypreferably bind to their target genes in resting cells and during early G1when transcription is turned off. These observations suggest that they aremainly involved in repression during early cell cycle progression(Moberg et al., 1996) and for that reason have been termed repressorE2Fs. Finally the E2F6 shares only the core DNA binding domain anddimerization domain but lacks the C-terminal transactivation and thepocket protein binding domains (Morkel et al., 1997). Recent studiesrevealed that E2F6 is able to recruit multiple factor to form a complextargeting the E2F responsive elements. Among these factors, chromatin-modifying proteins can be recruited primarily in quiescent cells (Ogawaet al., 2002). This result suggests a role in transcription silencing inquiescent cells by modifying the structure of the chromatin.

Fly E2F familyThe Drosophila E2F family is composed of only two members E2F1 andE2F2 and only one gene coding for the DP complex has been identified.The E2F1 transcription factor shows considerable homology to humanE2F1 with over 65% identity in the DNA binding region (Ohtani andNevins, 1994). The factor was shown to bind specifically to the E2Frecognition site, and to activate transcription of a reporter gene locateddownstream of the polymerase α gene promoter that contains E2Frecognition sequences. In addition, the E2F1 is required for S phaseduring Drosophila embryogenesis (Duronio et al., 1995), reinforcing theidea of a true homolog of the human counterpart. In contrast, DrosE2F2represses the transcription of E2F reporters and the loss of DrosE2F2results in increase of gene expression (Frolov et al., 2001). Therefore, thetwo types of complexes seem to play antagonistic roles in Drosophila andacts on the cell cycle the same way the E2F family in human regulatescell proliferation, but with fewer members.

29

AIM OF THE WORK

The proteasome field has expanded dramatically the last few years,especially because of its involvement in many intracellular pathways.Furthermore, the use of proteasome inhibitors for cancer patients or otherdiseases is an approaching treatment, and has already started in severalcases. In order to use the proteasome as a target in the treatment ofvarious diseases, the specific functions must be investigated and shouldbe known or at least evaluated to minimize possible side effects. Themain objectives of our laboratory are to use the advantages of a well-established model system Drosophila to investigate the functions of thedifferent proteasome regulators identified in these organisms. The mainfocus of my project was the study of the Drosophila PA28 (or REG)homolog, a unique gene coding for a putative PA28 (REG) memberisolated from a Drosophila cDNA library. The secondary objectivesinvolved the generation of a Drosophila in vivo assay to monitor the 26Sproteasome degradation, and the genomic search for global proteasomalregulation mechanisms using sequenced promoters available from theFlyBase database of the Drosophila genome projects (2003).

30

RESULTS AND DISCUSSION

Study of the invertebrate PA28 homolog

The Drosophila PA28 : A homolog of human PA28gamma (Paper I)

The first part of the project consisted in characterizing theDrosophila homolog of the PA28 gene family identified in mammals. Atthat time, only one group (Paesen and coworkers) had identified a PA28homolog in invertebrates, Rhipicephalus appendiculatus, sharing 55%identity with the human PA28γ primary structure (Paesen and Nuttall,1996). No biochemical study had been performed and the work onlymentioned the identification of a cDNA. In Drosophila, the sequencing ofthe genome was advancing but no clear sequence shared significantsimilarity with PA28 genes. However, a cDNA from an adult cDNAlibrary was identified as a PA28 homolog. We subcloned it and startedour investigation at that point.

Characterization of Drosophila PA28

The purification of the recombinant protein using DEAE and gelfiltration chromatography techniques revealed a rather large complex ofaround 200 kDA, in accordance with previous reports (Wilk et al., 2000),suggesting that the bacterial recombinant was able to form a polymer. Inorder to perform biochemical assays, Drosophila 20S proteasome was alsopurified from a large batch of fly embryos. Fluorogenic peptide assaysdemonstrate that our activator shares similar preference for activating theproteasome trypsin-like activity in a similar manner as the human PA28γ.The proteasomes and PA28 are well conserved between species, especiallyin their interaction domains. This finding probably motivated Whitby andcoworkers to resolve the structure of yeast proteasome together with theTrypanosoma brucei PA26 (Whitby et al., 2000). In order to investigatethe properties of the Drosophila activator, a mammalian proteasome wasalso tested in our assays, and the trypsin-like activity was stimulated but toa lower extent (5 fold compared to 22 fold for Drosophila 20S).Surprisingly, the chymotrypsin-like activity monitored by the degradationof LLVY-MCA fluorogenic peptide was decreased in presence of theDrosophila activator. This inhibition was the first example where a PA28member could function as a positive and negative regulator of proteasomal

31

activity. This is the reason why we named it dREG for Drosophilaproteasome regulator. For clarity reasons, the term PA28 will still be usedhere, since PA28 nomenclature is more common in the scientificcommunity, and dREG will be employed for the Drosophila complex.Interestingly, the chymotrypsin-like activity of the mammalian proteasomewas also inhibited by dREG. This inhibitory effect was later observed byLi and coworkers with recombinant human PA28γ (Li et al., 2001).Surprisingly, they also observed that substitution of Lys188 to Asp188 orGlu188 abolished this effect, and the mutant showed the exact sameactivation pattern as PA28. They suggested a change in the 20Sproteasome beta-subunit conformation induced by the mutation.One of the main challenges concerning the PA28 family will be tounderstand precisely the way PA28 activates the proteasome. So far,contradictory reports do not put forward a clear mechanism, but rathersuggest a contribution of both the regulator inner channel andconformational changes in the 20 S proteasome complex.

Homolog-specific insert regions and PA28 localisation

Based on the fluorogenic assays, the dREG activation patternresembled the PA28γ rather than the PA28, with a major activation of thetrypsin-like activity. In order to better characterize the Drosophilahomolog, we performed immunofluorescence with a polyclonalantibodies raised from the recombinant protein injected in a rabbit. Theresults clearly showed a major nuclear localization whereas none or verylittle staining was observed in the cytoplasm. This observation reinforcesthe homology to vertebrate PA28γ that also localizes in the nucleuswhereas PA28αβ is mainly found in the cytoplasm (Soza et al., 1997).Interestingly, a putative monopartite nuclear localization signal (NLS)was found within the insert specific region forming the flexible loop (seeintroduction) in Drosophila. The signal is usually composed of 3 or 4amino-acids, with lysine or arginine at positions 1, 2, and 4, that arerecognized by importin-α. The mutation of the signal from KRQR toSSQS was done by site-directed mutagenesis, and the obtained plasmidwas transfected into Drosophila mbn-2 cells. Our results showed a strongcytoplasmic staining supporting an important role of the signal in thelocalization of the complex. Primary sequence alignments showed thatthe signal is conserved in all organisms possessing a PA28γ atapproximately the same position (always within the flexible loop).

From these experiments, it is clear that the homolog-specific insertregion plays other functions than contributing to the binding of PA28

32

complex to the proteasome as originally proposed by Rechsteiner andcolleagues (Zhang et al., 1998b). The sequence of PA28α and P28βpossess an additional KEKE motif that may serve other purposes. Thepresence of these nuclear signals, highly conserved between species attestto an important physiological function. The distribution of proteasomes inthe cell is very dynamic and varies during different processes, such as theincrease of nuclear proteasomes during the cell cycle from G1 to thenuclear envelope breakdown in early mitosis. One might speculate acontribution of the PA28 family in the overall 20S cellular distributionsince little is known about the mechanisms responsible for the balancebetween the nuclear and cytoplasmic pools of proteasomes. In that sense,the PA28 might bind the 20S proteasome directly after being assembledand then re-localised it where the proteasome should play its specificfunction. The observation of the capase-7 interaction with PA28γ agreeswith this hypothesis. Indeed, proteasomes are removed from the nucleusduring apoptosis and accumulate within apoptotic blebs. In vitro humanPA28γ subunits are cleaved by both caspase-3 and –7, in the specificsequence DGDL, showing that PA28γ is an endogenous substrate ofcaspases –3 and –7 (Araya et al., 2002). Interestingly, the specificcleavage site is located very close to the NLS, four amino acid before thefirst lysine of the NLS. The cleavage can thus destroy the ability of thePA28γ to be transported inside the nucleus, directing the proteasome outof this structure. The homolog specific insert region has been shown notto play any role in proteasome binding and activation for PA28α (Zhanget al., 1998b). We did test this hypothesis in Drosophila by looking atdREG protein levels in Drosophila cells undergoing apoptosis afterstaurosporin treatment, but we could not observe the specific cleavagefound in human described above (unpublished data), meaning that themechanism might not be conserved between species. This hypothesisconcerning a role for PA28s in directing the 20S proteasome to specificstructures has also been proposed by Zhang and coworkers who proposedthat the insert region of PA28 could serve to couple the calnexin TAP-MHC class I complexes in the endoplasmic reticulum, promoting transferof antigenic peptides to the MHC class I molecules (Zhang et al., 1998b).

Study of the dREG promoter (Paper II)

Identifying gene regulation at the transcriptional level by studyingpromoters is a powerful tool to understand the function of a gene. Thefully sequenced Drosophila genome allowed us to quickly analyze thepromoter region located upstream of dREG. The promoter region is

33

about 926 bp long and the two adjacent genes strawberry notch (sno) anddREG, are arranged in opposite directions with both promoter regionswithin a short intergenic region. This region includes two typical DREmotifs, an 8 basepair palindromic sequence where the Drosophilareplication related transcription factor is able to bind (see introduiction).In addition, the first intron possesses a typical E2F binding site (fig. 5).

Fig. 5. Schematic representation of dREG promoter region. The putativetranscription factor-binding sites are represented as well as their positions relative tothe longest 5´transcript found.

The first step in the promoter analysis was to determine thetranscriptional start of the gene. To our surprise, we identified 4 differenttranscriptional starts located close together within a 35 bp region. Thispromiscous transcriptional initiation may be explained by the absence ofa TATA-box or DPEs (downstream promoter elements) associated withTATA-less promoters upstream of dREG gene.

After this initial step, the two DRE elements as well as the E2Fbinding motif were tested by cloning the upstream region of dREG genein front of the LacZ reporter gene. Site-directed mutagenesis was carriedout on the proposed elements, and β-galactosidase was assayed aftertransfection of each of the constructs in mbn-2 cells. The results showed acomplete abolishment of the β-gal activity when the first DRE sequencewas mutated, whereas mutation of the downstream DRE was found not toaffect the expression of the reporter. These results demonstrated that thefirst DRE site is essential for transcription of dREG in mbn-2 cells. Wewere surprised to observe that the second site had absolutely no effect ontranscription. However, similar results have been reported in theliterature: Choi and coworkers identified three DRE motifs upstream ofthe Drosophila TBP gene and mutation of one of these motif did notdecrease luciferase activity whereas the two other did significantly (100%and 80% of reduced expression respectively) (Choi et al., 2000). Theother site might serve other purposes. Indeed, BEAF, a factor involved inchromatin remodeling, has been shown to compete with DREF and bind

ATGDRE 1 DRE 2 E2F

REG 5´UTRSno 5´UTRATG

-358 -22

100bp

tsp tsp

34

to the sequence CGATA which is also part of the DRE motifTATCGATA (Hart et al., 1999). However the BEAF activity, whichrequires three close CGATA motifs, was impossible to monitor bytransient transfection experiments (Hart et al., 1999), so the DRE motifpresent in dREG promoter whose mutation did not affect the β-galactivity might be used by such factors other than DREF.

Concerning the E2F element, we found a surprising increase of β-gal activity when the E2F site was mutated. The E2F family inDrosophila is composed of two members that play antagonists roles.Drosophila E2F1 is required for S phase progression whereas E2F2inhibits transcription of gene possessing E2F motif in their promoter(Frolov et al., 2001). The site present in dREG first intron may beoccupied in majority by the E2F member that inhibit the transcription ofthe gene. Thus, mutation of the site would rather show an increase of theβ-gal activity. The fact that several members of a family are capable tobind to the same exact site makes the result harder to interpret. Forinstance, the Drosophila polymerase α gene promoter possesses threeE2F binding sites that act as stimulators in Kc cells, but two of them actrather negatively on promoter activity during development (Yamaguchi etal., 1997). The E2F activity depends also highly on the cell cycle phase.E2F sites in mammalian promoters can act negatively in G0 and G1, butrelease of RB protein from the complex act as a positive effect increasingthe transcription. Therefore, the in vivo situation is more complex thanwhat we observed in cell lines. However, we can still conclude thatDREF and E2F both act on dREG promoter suggesting a role in G1 to Sphase transition and maybe related to DNA replication. Interestingly, tworecent publications suggest a similar transcriptional activation in otherorganisms. Microarray experiments on E2F1 in mice fibroblast cellsindicated PA28γ as positive candidate (Ma et al., 2002b) meaning that theE2F activation of PA28γ RNA is conserved between species. Anotherrecent study revealed the presence of a homolog to DREF in the humangenome, with a similar putative DRE element (Ohshima et al., 2003). ThePA28γ possess such a site in its promoter confirming a conservedactivation pathway from Drosphila to human.

In parallel with our attempts to characterize the dREG promoter,we also analyzed the promoters of the Drosophila proteasome genes forDRE motifs. Surprisingly, several but not all of the subunits possess oneor two DRE motifs in their proximal promoter regions. One of the fewgenes that contain multiple DRE sequence is the 20S β-2 gene. InDrosophila, two additional β-2 isoforms have been identified β2R1, andβ2R2. Their promoters do not contain any DRE motifs, and northern-blotanalysis revealed that the isoforms are male-specific expressed (Ma et al.,

35

2002a). In general, DREF might contribute at least in part, to varying theexpression levels of certain proteasomes genes thereby allowingproteasome structural heterogeneity. Alternatively, the DREF-inducedsubunits may serve a specific function during the G1 to S phase transitionindependent of the proteasome. Future work should help in understandingoverall proteasome regulation.

The Dictyostelium PA28 homolog (Paper IV)

Dictyostelium is a powerful system for research in cell anddevelopmental biology. The organism has unique advantages for studyingfundamental cellular processes including cytokinesis, motility,phagocytosis, chemotaxis, signal transduction, and different aspects ofdevelopment (Eichinger and Noegel, 2003). One of the main advantagesof this model system resides in the fact that Dictyostelium grows as aseparate, independent cell but interacts to form multicellular structureswhen challenged by conditions such as starvation. Furthermore, theseapproaches have benefited from bioinformatics resources including on-line databases. In a Dictyostelium cDNA developmental libraryrepresenting all the mRNAs expressed when the organism becomesmulticellular, we surprisingly identified a single putative gene that hadhigh similarities to PA28s. Cloning of this gene and overexpression in E.coli system led to purification of a large complex suggesting a similarsize to native PA28s. The use of fluorogenic peptides showed that thecomplex has the capacity of activating purified 20S proteasome fromDictyostelium, with a major activation of the trypsin-like activitymonitored by LLR-MCA fluorogenic peptide. We also performedimmunofluorescence experiments with rabbit polyclonal antibodies raisedagainst the protein. The staining was observable almost exclusively in thenucleus. Examination of different developmental stages (before or afterstarvation, meaning in both monocellular or multicellular forms) alsoshowed a nuclear staining of the Dictyostelium complex. These datasuggested a true homolog of PA28γ in accordance with what we observedin Drosophila (see results above), the nuclear complex PA28γ probablyarising first in the evolution before the PA28 (made of α and β subunits)complex itself only found in organisms with adaptive immune system(Fig. 6).

36

Diplomonads

Kinetoplasts

CryptosporidiumTheileria

Plasmodium species

Toxoplasma

PlantaeMammalsFishNematodesArthropods

Dictyostelium (Conosa)Proposedroot ofeukaryote tree

Apicomplexans

Fungi

γγ

γ?

γ

α/β

γ

γγ

α/βγ

Fig. 6. Schematic representation of the main features of a eukaryote supertreeconstructed from 100 genes by Bapteste et al. (2002), with the distribution of REG α,β and γ overlaid. The tree demonstrated the clustering of amoeboid lineages (hererepresented by Dictyostelium) as a monophyletic grouping, the Conosa, distinct fromplants, animals & fungi. The arrow indicates the root of the eukaryote tree assuggested from the distribution of DHFR-TS fusion genes (Stechmann & Cavalier-Smith 2002). The DHFR-TS fusion suggests the root is located between Opisthokontsand Amoebozoa and, if correct, suggests that REG γ was present in the ancestraleukaryote, and that it has been lost from lineages such as Fungi and Plants. Coloringindicates which eukaryote supergroup the represented lineages/groups belong to.

37

Investigating the function of PA28gamma

We decided to start our investigation by looking at the Drosophilacomplex after induction of the immune response. Since the PA28 presentin higher eukaryotes has been shown to be involved in the immuneresponse, we speculated that the ancient form might as well serve arelated function in invertebrate. In order to test this hypothesis, we treatedmbn-2 cells with lipopolysaccharides (LPS), a compound present in theouter cell membrane of gram negative bacteria that induces an immuneresponse. Our results did not show an increase in dREG expressionneither at the RNA or the protein level (Paper I), indicating that dREG isnot induced after immune response as PA28 does with γ-interferon.

The second part, done in relation to the promoter analysisdescribed above, was to knockdown dREG from Drosophila cells anddetermine the effects on the cells cycle by FACS analysis. Thisexperiment was motivated by the result from Murata and coworkers whoobserved a G1->S phase inhibition in mouse embryonic fibroblastslacking PA28γ (Murata et al., 1999). The results obtained from ourexperiments clearly showed the same change in the G1->S phasetransition in Drosophila but the effect was more pronounced in the insectsystem (fig. 7). The two studies agree on the important function of PA28γon the cell cycle progression.

0

10

20

30

40

50

60

G1 S G2/M 0

10

20

30

40

50

60

G1 S G2/M

Fig. 7. Left panel, comparison of the cell cycle transition defects betweenPA28γ knock-out in mouse embryos fibroblasts (adapted from Murata et al.) and rightpanel, knock-down experiment of the Drosophila homolog (our study) using RNAi.

Our most recent attempt to gain insight into the biologicalsignificance of the complex consisted of immunostaining of Drosophilaembryos with polyclonal antibodies raised against dREG in order tofollow the expression profile during embryogenesis. Indeed, genetranscriptional expression patterns have been measured at a genomic

38

scale during a complete time-course of Drosophila development anddREG transcriptional activity showed highest levels duringembryogenesis. Staining of syncytium early embryos revealed a disperselocalization with no apparent concentration around the chromosomessuggesting no direct interaction with DNA (Paper II). The appearance ofnuclear envelopes lead to the nuclear localization of dREG. Later,Drosophila embryos go through gastrulation, a process involving fourmajor morphogenic events: ventral furrow formation, posterior midgutinvagination, cephalic furrow formation and germband extension (Costaet al., 1994). The first three events are driven by cell shape changes.During this phase, the invaginating cells showed increase levels of dREG.Interestingly, a recent screen to identify gastrulation changes identifiedthree proteasome subunits which had altered abundance protein levels inventralized versus lateralized embryos and RNAi knockdown of thesesubunits caused ventral furrow defects, confirming the role of theseproteins in ventral furrow morphogenesis (Gong et al., 2004). Thecombination of our past work and this recent finding supports animportant role for the proteasome and dREG during gastrulation thatwould be worth examining in more detail. After this step, dREG wasfound uniformly in all nuclei except during germ band elongation where acytoplasmic localization was apparent in discrete domains. Thesedomains were identified to be mitotic domains that consist of specific cellclusters that undergo mitosis in a synchronous manner (Foe, 1989).Thereby, the cytoplasmic staining results of the nuclear envelopebreakdown, again demonstrating a function without direct interactionwith DNA.

Concluding remarksOverall, PA28γ seems to be the most ancient member of the PA28 family.PA28 arose probably from PA28γ by gene duplication and became anefficient activator recruited by the adaptive immune system. So what isrole of the ancient form? It is difficult to draw any picture from theactivation results in experiments using fluorogenic peptides, since thecompounds used in fluorogenic assays including MCA (methyl-couramin) are relatively large when compared to the small polypeptidesmade of three or four residues. That of course renders difficult anyinterpretation. Results involving more ”natural peptides” like use of 25-mer by Li et al. showed that PA28γ is helping the proteasome simply byincreasing the number of smaller peptides released by the complex.However, we still can describe the main properties of this complex basedon our study as well as the literature:

39

In the cell, PA28γ is nuclear (completely or in a very large percentage)and several experiments showed that:1- The complex is able to increase proteasomal degradation, with a

major effect on the trypsin-like activity, whereas it inhibits thechymotrypsin-like activity.

2 - The complex is induced and required for proper G1->S phasetransition, from Drosophila to human.

3 - The Drosophila possess a DRE site and may possess E2F siteimportant for transcription regulation, and the mechanism is likely tobe conserved since human E2F1 activates PA28γ as well.

4- The removal of PA28γ in mouse and in Drosophila induced a growthretardation from G1->S.

So what could be the function of the activator in this transition?This section is based upon a speculative interpretation of the current data.One can speculate a role for PA28γ in the removal of the long fragmentsgenerated by the 26S proteasome that could potentially inhibit the cellcycle progression. This involvement of PA28γ in cell cycle regulatordegradation has also been proposed by Murata and colleagues thatsuggested a role in p21 or p27 degradation (Murata et al., 1999). Thesesuggestions need of course to be tested on the bench to be validated.An explanation could originate from the structure of certain activator orinhibitors of the cell cycle. For instance, p21 is a mammalian cell cycleinhibitor, involved in the inhibition of the G1 phase. Although thisprotein is not known in Drosophila, the same type of system probablyexists (Avedisov et al., 2001). Recent work suggested that p21 possessspecific regions that are essential for its function. For instance, PCNA isimportant for cell cycle progression and p21 can modulate its activitythrough its C-terminal domain (Zheleva et al., 2002). Furthermore, Chenand coworkers identified a 39 amino acid fragment of p21 that wassufficient to bind to PCNA and partially inhibit DNA replication in vivo(Chen et al., 1996). It is now clear that the proteasome is involved in thedegradation of many of these cell cycle regulators including p21 (Bloomet al., 2003). A putative role for PA28γ complex could be to activate thetrypsin-like activity of the 20S proteasome in order to further process thedegradation products. This mechanism would avoid cell cycle inhibitionas a result of release of longer polypeptides by the proteasome that couldstill inhibit cell cycle progression, such as p21 polypeptide that has beenshown to inhibit PCNA activity.

40

An in vivo assay to study the Drosophila 26S proteasome(Paper III)

My involvement in this project concerned the cloning of threedifferent constructs from mammalian to a constitutive Drosophila vectorpAc5.1 and generating stable Drosophila S2 cell lines. Originally,Dantuma and coworkers developed a convenient reporter system tomonitor proteasome activity in living cells by using short-lived substrates(Dantuma et al., 2000). The assay consists of a GFP molecule fused toubiquitin (fig. 8).

X

X=M for resistant control fusionX=R for N-end rule pathwayX=G76V for UFD pathway

Ubiquitin GFP

Fig. 8. Schematic representation of the fusion genes generated in order tomonitor the 26S proteasome activity in vivo. X represents the residue that differs fromthe three constructs.

GFP was originally derived from a jellyfish protein that has theproperty to emit autofluorescence when stimulated at the correct wavelenght. This stable protein with a half-life over 24 hours became apowerful tool to monitor cellular localization of substrates when fused tothis protein. The constructs presented here also possess ubiquitin, a 76residue conserved protein that is responsible for targeting substrates todegradation by the 26S proteasome. As shown on fig. 8, three differentplasmids have been generated according to the residue placed in front ofthe GFP protein:

1- the M construct generates a stable control fusion substrate thatdoes not contain a degradation signal upon ubiquitin cleavage andis stable in cells.

2 - the R construct generates an N-end rule pathway substrate.Indeed, once the ubiquitin is cleaved, the argine becomes the N-terminal residue and has been shown to be fastly degraded throughthe N-end rule pathway (approximately 2 minutes when R is infront of β-galactosidase) (Bachmair et al., 1986).

41

3- The G76V construct generates a mutated uncleavable ubiquitin. Inthis case the construct is designed to function as a ubiquitin fusiondegradation (UFD) substrate. In other words, the deubiquitinationof the Ub fusion is completely inhibited if the C-terminal Gly76 issubstitituted by another residue. Thus, the non-removable ubiquitinwill result in a fast multiubiquitination and will lead to rapiddegradation of the GFP by the 26S proteasome.

The three constructs generated upstream of a constitutive promoter andstably expressed in Drosophila cell lines have been very useful tomonitor the performance of the 26S proteasome when challenged bydifferent treatments. Indeed, the proteasome activity can easily bemeasured by the described in vivo assay after the use of proteasomeinhibitors including MG132 or RNAi interference against variousproteasomal subunits, or other cellular component that may compromisethe proper function of the proteasome when depleted from the cells.

Study of the proteasomal regulation after MG132 treatment(Paper V)

Paper III, study of the RNAi knockdown of S5a and S13 subunitsrevealed a strong induction of 20S and 19S proteasomal subunits. Thisinteresting result was also observed by DeMartino and colleagues whofound the same activation pattern (Wojcik and DeMartino, 2002). Inorder to investigate the mechanism responsible for this induction, weamplified by PCR and subsequently cloned three 26S proteasome subunitpromoters namely alpha2, beta2 (20S proteasome) and S2 (19Sproteasome), as well as Ter94 promoter (non proteasomal gene that isalso induced after MG132 or S5a depletion; see Paper V) upstream of aLacZ reporter vector in order to perform transient transfectionexperiments. The β-galactosidase levels monitoring the promoterefficiencies showed an increase after MG132 induction, as expected fromthe micro-array data. Interestingly, when the 5´UTR of the α2 and S2promoters were removed, this induction was lost suggesting a role of thisregion in the induction of the mRNA. The next step consisted in replacingthe dREG 5´UTR from the promoter construct generated in the study ofdREG promoter, by the S2 5´UTR, to observe if the induction was stillpresent. Preliminary results suggested that the dREG promoter acquiredthe ability to be induced after MG132 treatment strongly supposing thepresence of an essential regulatory motif in the S2 5´UTR. Sequencealignment using the MEME program revealed that the Drosophila genes

42

coding for the proteasomal subunits may have a conserved motif withintheir 5´ UTR (fig. 9).

Fig. 9. Result of the MEME motif search on twenty-two proteasomal 5´ UTRs. Thesearch program identified a consensus sequence that might be of importance for thetranscriptional up-regulation seeing on the micro-array experiments as well as thetransfection assays (see Paper V).

In conclusion, the putative motif present in the gene 5´ UTRs may inducetranscription of the genes after proteasome inhibitor treatment either onthe DNA level or the RNA level. Our group is currently investigating themechanism responsible for this transcriptional up-regulation.

43

CONCLUSIONS AND PERSPECTIVES

The proteasome has a wide range of activities within the cell, and thisthesis discussed some aspects of them. The identification of a Drosophilahomolog of PA28γ showed that this gene is conserved in invertebratesand seems to share the same function related to cell cycle progression.Currently, we are unable to point out the precise biological function ofthe regulator, but we have instead gained insights into its role. Thedeletion of the homolog gene in Dictyostelium should be of great help toreveal the biological relevance of the complex since we can easilymonitor developmental or chemotaxis defects produced by the removal ofa gene of interest in this organism. The study of the dREG promotershowed that the transcriptional control of the gene was governed at leastin part by DREF and E2F transcription factors and that the mechanism ofactivation is probably conserved in humans. Other aspects of theproteasomal regulation have been addressed in this thesis and future workin the laboratory will surely provide insight into the co-regulatedtranscriptional activation of all the proteasome genes.

44

ACKNOWLEDGMENTS

Thanks to all the persons who supported me directly or indirectly during this thesis.

I would like to especially thank:

-Patrick Young, my supervisor for always offering me help and guidance whennecessary, and for also letting me independent at other times. Thanks also for thegreat help during the preparation of this thesis.

-Josefin Lundgren, my coworker during most of the five years, thanks for your greatenthusiasm and our long discussions. I hope that I will be as sharp as you one day(and thanks for the organization lessons of course…).

I also would like to express my gratitude to the persons not from our lab that helpedme during this period:- Ylva and her group, especially Gunnel for all the Drosophila work, thanks forteaching me very useful Drosophila techniques.- Marie and her group, thanks for the help with the RNA work.- a special thanks to Britta for preparing all the buffers, you save us a lot of time!!

Special thanks for all the nice people I met at the Department, especially:- Alex, thanks for never refusing me a coffee at any time of the day, and for all thesediscussions where we never agreed, it would not have been so fun otherwise!- Eduard, it´s good that you brought some Mediterranean touch in the lab (or Catalantouch I should say…)- Rula, thanks for all the nice time we spent inside as well as outside the lab (merci àEyad aussi donc…)- Anthony thanks for your help and advices, and good luck with your complicatedtheories,-Susannah and Haleh, as Alex said, thanks for the very nice discussions at lunches wehad together (sometimes he can be right…).-thanks to all the other very nice persons I met during this time: David, Linus, Poppy,Ignasi, Mats, and Johan among others.

More personally, I would like to thank my parents and my brother Christophe for theirconstant support.I will always remember the persons that were always there for me here in Sweden:-Mounia parents, Amine, Nadia and Samira, thanks a lot for introducing me to bothAlgerian and Swedish cultures, and thanks all of you for your great generosity.-Hanna and Marek as well as Loïc, thanks for sharing a lot of excellent moments.

Finally, I dedicate this thesis for the two princesses of my life, my wife Mounia andmy daughter Ines. Thanks Mounia for your never-ending support, and for your love.

45

REFERENCES

(2003). The FlyBase database of the Drosophila genome projects and communityliterature. Nucleic Acids Res 31, 172-175.

Agrawal, N., Dasaradhi, P. V., Mohmmed, A., Malhotra, P., Bhatnagar, R. K., andMukherjee, S. K. (2003). RNA interference: biology, mechanism, and applications.Microbiol Mol Biol Rev 67, 657-685.

Ahn, J. Y., Tanahashi, N., Akiyama, K., Hisamatsu, H., Noda, C., Tanaka, K., Chung,C. H., Shibmara, N., Willy, P. J., Mott, J. D., and et al. (1995). Primary structures oftwo homologous subunits of PA28, a gamma-interferon-inducible protein activator ofthe 20S proteasome. FEBS Lett 366, 37-42.

Ahn, K., Erlander, M., Leturcq, D., Peterson, P. A., Fruh, K., and Yang, Y. (1996). Invivo characterization of the proteasome regulator PA28. J Biol Chem 271, 18237-18242.

Aki, M., Shimbara, N., Takashina, M., Akiyama, K., Kagawa, S., Tamura, T.,Tanahashi, N., Yoshimura, T., Tanaka, K., and Ichihara, A. (1994). Interferon-gammainduces different subunit organizations and functional diversity of proteasomes. JBiochem (Tokyo) 115, 257-269.

Amsterdam, A., Pitzer, F., and Baumeister, W. (1993). Changes in intracellularlocalization of proteasomes in immortalized ovarian granulosa cells during mitosisassociated with a role in cell cycle control. Proc Natl Acad Sci U S A 90, 99-103.

Araya, R., Takahashi, R., and Nomura, Y. (2002). Yeast two-hybrid screening usingconstitutive-active caspase-7 as bait in the identification of PA28gamma as aneffector caspase substrate. Cell Death Differ 9, 322-328.

Arendt, C. S., and Hochstrasser, M. (1999). Eukaryotic 20S proteasome catalyticsubunit propeptides prevent active site inactivation by N-terminal acetylation andpromote particle assembly. Embo J 18, 3575-3585.

Arnold, D., Driscoll, J., Androlewicz, M., Hughes, E., Cresswell, P., and Spies, T.(1992). Proteasome subunits encoded in the MHC are not generally required for theprocessing of peptides bound by MHC class I molecules. Nature 360, 171-174.

Avedisov, S. N., Rogozin, I. B., Koonin, E. V., and Thomas, B. J. (2001). Rapidevolution of a cyclin A inhibitor gene, roughex, in Drosophila. Mol Biol Evol 18,2110-2118.

Bachmair, A., Finley, D., and Varshavsky, A. (1986). In vivo half-life of a protein is afunction of its amino-terminal residue. Science 234, 179-186.

46

Bailly, E., and Reed, S. I. (1999). Functional characterization of rpn3 uncovers adistinct 19S proteasomal subunit requirement for ubiquitin-dependent proteolysis ofcell cycle regulatory proteins in budding yeast. Mol Cell Biol 19, 6872-6890.

Bapteste, E., Brinkmann, H., Lee, J. A., Moore, D. V., Sensen, C. W., Gordon, P.,Durufle, L., Gaasterland, T., Lopez, P., Muller, M., and Philippe, H. (2002). Theanalysis of 100 genes supports the grouping of three highly divergent amoebae:Dictyostelium, Entamoeba, and Mastigamoeba. Proc Natl Acad Sci U S A 99, 1414-1419.

Barrett, A. J., and Rawlings, N. D. (1991). Types and families of endopeptidases.Biochem Soc Trans 19, 707-715.

Bloom, J., Amador, V., Bartolini, F., DeMartino, G., and Pagano, M. (2003).Proteasome-mediated degradation of p21 via N-terminal ubiquitinylation. Cell 115,71-82.

Boes, B., Hengel, H., Ruppert, T., Multhaup, G., Koszinowski, U. H., and Kloetzel, P.M. (1994). Interferon gamma stimulation modulates the proteolytic activity andcleavage site preference of 20S mouse proteasomes. J Exp Med 179, 901-909.

Brooks, P., Fuertes, G., Murray, R. Z., Bose, S., Knecht, E., Rechsteiner, M. C.,Hendil, K. B., Tanaka, K., Dyson, J., and Rivett, J. (2000). Subcellular localization ofproteasomes and their regulatory complexes in mammalian cells. Biochem J 346 Pt 1,155-161.

Chen, J., Peters, R., Saha, P., Lee, P., Theodoras, A., Pagano, M., Wagner, G., andDutta, A. (1996). A 39 amino acid fragment of the cell cycle regulator p21 issufficient to bind PCNA and partially inhibit DNA replication in vivo. Nucleic AcidsRes 24, 1727-1733.

Choi, T., Cho, N., Oh, Y., Yoo, M., Matsukage, A., Ryu, Y., Han, K., Yoon, J., andBaek, K. (2000). The DNA replication-related element (DRE)-DRE-binding factor(DREF) system may be involved in the expression of the Drosophila melanogasterTBP gene. FEBS Lett 483, 71-77.

Costa, M., Wilson, E. T., and Wieschaus, E. (1994). A putative cell signal encoded bythe folded gastrulation gene coordinates cell shape changes during Drosophilagastrulation. Cell 76, 1075-1089.

Crick, F.H.C. (1968) The origin of the genetic code. J. Mol. Biol. 38, 367-379

Dantuma, N. P., Lindsten, K., Glas, R., Jellne, M., and Masucci, M. G. (2000). Short-lived green fluorescent proteins for quantifying ubiquitin/proteasome-dependentproteolysis in living cells. Nat Biotechnol 18, 538-543.

DeGregori, J. (2002). The genetics of the E2F family of transcription factors: sharedfunctions and unique roles. Biochim Biophys Acta 1602, 131-150.

47

DeGregori, J., Leone, G., Miron, A., Jakoi, L., and Nevins, J. R. (1997). Distinct rolesfor E2F proteins in cell growth control and apoptosis. Proc Natl Acad Sci U S A 94,7245-7250.Dick, L. R., Aldrich, C., Jameson, S. C., Moomaw, C. R., Pramanik, B. C., Doyle, C.K., DeMartino, G. N., Bevan, M. J., Forman, J. M., and Slaughter, C. A. (1994).Proteolytic processing of ovalbumin and beta-galactosidase by the proteasome to ayield antigenic peptides. J Immunol 152, 3884-3894.

Dick, T. P., Ruppert, T., Groettrup, M., Kloetzel, P. M., Kuehn, L., Koszinowski, U.H., Stevanovic, S., Schild, H., and Rammensee, H. G. (1996). Coordinated dualcleavages induced by the proteasome regulator PA28 lead to dominant MHC ligands.Cell 86, 253-262.

Driscoll, J., Brown, M. G., Finley, D., and Monaco, J. J. (1993). MHC-linked LMPgene products specifically alter peptidase activities of the proteasome. Nature 365,262-264.

Dubiel, W., Pratt, G., Ferrell, K., and Rechsteiner, M. (1992). Purification of an 11 Sregulator of the multicatalytic protease. J Biol Chem 267, 22369-22377.

Duronio, R. J., O'Farrell, P. H., Xie, J. E., Brook, A., and Dyson, N. (1995). Thetranscription factor E2F is required for S phase during Drosophila embryogenesis.Genes Dev 9, 1445-1455.

Dyson, N. (1998). The regulation of E2F by pRB-family proteins. Genes Dev 12,2245-2262.

Ehring, B., Meyer, T. H., Eckerskorn, C., Lottspeich, F., and Tampe, R. (1996).Effects of major-histocompatibility-complex-encoded subunits on the peptidase andproteolytic activities of human 20S proteasomes. Cleavage of proteins and antigenicpeptides. Eur J Biochem 235, 404-415.

Eichinger, L., and Noegel, A. A. (2003). Crawling into a new era-the Dictyosteliumgenome project. Embo J 22, 1941-1946.

Fabunmi, R. P., Wigley, W. C., Thomas, P. J., and DeMartino, G. N. (2001).Interferon gamma regulates accumulation of the proteasome activator PA28 andimmunoproteasomes at nuclear PML bodies. J Cell Sci 114, 29-36.

Fehling, H. J., Swat, W., Laplace, C., Kuhn, R., Rajewsky, K., Muller, U., and vonBoehmer, H. (1994). MHC class I expression in mice lacking the proteasome subunitLMP-7. Science 265, 1234-1237.

Foe, V. E. (1989). Mitotic domains reveal early commitment of cells in Drosophilaembryos. Development 107, 1-22.

Frolov, M. V., Huen, D. S., Stevaux, O., Dimova, D., Balczarek-Strang, K., Elsdon,M., and Dyson, N. J. (2001). Functional antagonism between E2F family members.Genes Dev 15, 2146-2160.

48

Gandre, S., and Kahana, C. (2002). Degradation of ornithine decarboxylase inSaccharomyces cerevisiae is ubiquitin independent. Biochem Biophys Res Commun293, 139-144.

Giot, L., Bader, J. S., Brouwer, C., Chaudhuri, A., Kuang, B., Li, Y., Hao, Y. L., Ooi,C. E., Godwin, B., Vitols, E., et al. (2003). A protein interaction map of Drosophilamelanogaster. Science 302, 1727-1736.

Glickman, M. H., Rubin, D. M., Coux, O., Wefes, I., Pfeifer, G., Cjeka, Z.,Baumeister, W., Fried, V. A., and Finley, D. (1998). A subcomplex of the proteasomeregulatory particle required for ubiquitin-conjugate degradation and related to theCOP9-signalosome and eIF3. Cell 94, 615-623.

Gong, L., Puri, M., Unlu, M., Young, M., Robertson, K., Viswanathan, S.,Krishnaswamy, A., Dowd, S. R., and Minden, J. S. (2004). Drosophila ventral furrowmorphogenesis: a proteomic analysis. Development 131, 643-656.

Gray, C. W., Slaughter, C. A., and DeMartino, G. N. (1994). PA28 activator proteinforms regulatory caps on proteasome stacked rings. J Mol Biol 236, 7-15.

Griffin, T. A., Nandi, D., Cruz, M., Fehling, H. J., Kaer, L. V., Monaco, J. J., andColbert, R. A. (1998). Immunoproteasome assembly: cooperative incorporation ofinterferon gamma (IFN-gamma)-inducible subunits. J Exp Med 187, 97-104.

Groettrup, M., Kraft, R., Kostka, S., Standera, S., Stohwasser, R., and Kloetzel, P. M.(1996). A third interferon-gamma-induced subunit exchange in the 20S proteasome.Eur J Immunol 26, 863-869.

Groettrup, M., Ruppert, T., Kuehn, L., Seeger, M., Standera, S., Koszinowski, U., andKloetzel, P. M. (1995). The interferon-gamma-inducible 11 S regulator (PA28) andthe LMP2/LMP7 subunits govern the peptide production by the 20 S proteasome invitro. J Biol Chem 270, 23808-23815.

Groll, M., and Clausen, T. (2003). Molecular shredders: how proteasomes fulfill theirrole. Curr Opin Struct Biol 13, 665-673.

Groll, M., Ditzel, L., Lowe, J., Stock, D., Bochtler, M., Bartunik, H. D., and Huber,R. (1997). Structure of 20S proteasome from yeast at 2.4 A resolution. Nature 386,463-471.

Groll, M., Heinemeyer, W., Jager, S., Ullrich, T., Bochtler, M., Wolf, D. H., andHuber, R. (1999). The catalytic sites of 20S proteasomes and their role in subunitmaturation: a mutational and crystallographic study. Proc Natl Acad Sci U S A 96,10976-10983.

Groll, M., Nazif, T., Huber, R., and Bogyo, M. (2002). Probing structuraldeterminants distal to the site of hydrolysis that control substrate specificity of the20S proteasome. Chem Biol 9, 655-662.

49

Grune, T., Reinheckel, T., and Davies, K. J. (1997). Degradation of oxidized proteinsin mammalian cells. Faseb J 11, 526-534.

Hagemann, C., Patel, R., and Blank, J. L. (2003). MEKK3 interacts with the PA28gamma regulatory subunit of the proteasome. Biochem J 373, 71-79.

Harris, J. R. (1968). Release of a macromolecular protein component from humanerythrocyte ghosts. Biochim Biophys Acta 150, 534-537.

Hart, C. M., Cuvier, O., and Laemmli, U. K. (1999). Evidence for an antagonisticrelationship between the boundary element-associated factor BEAF and thetranscription factor DREF. Chromosoma 108, 375-383.

Heinemeyer, W., Fischer, M., Krimmer, T., Stachon, U., and Wolf, D. H. (1997). Theactive sites of the eukaryotic 20 S proteasome and their involvement in subunitprecursor processing. J Biol Chem 272, 25200-25209.

Heinemeyer, W., Trondle, N., Albrecht, G., and Wolf, D. H. (1994). PRE5 and PRE6,the last missing genes encoding 20S proteasome subunits from yeast? Indication for aset of 14 different subunits in the eukaryotic proteasome core. Biochemistry 33,12229-12237.

Hendil, K. B. (1988). The 19 S multicatalytic "prosome" proteinase is a constitutiveenzyme in HeLa cells. Biochem Int 17, 471-477.

Henke, W., Ferrell, K., Bech-Otschir, D., Seeger, M., Schade, R., Jungblut, P.,Naumann, M., and Dubiel, W. (1999). Comparison of human COP9 signalsome and26S proteasome lid'. Mol Biol Rep 26, 29-34.

Hershko, A., Ciechanover, A., Heller, H., Haas, A. L., and Rose, I. A. (1980).Proposed role of ATP in protein breakdown: conjugation of protein with multiplechains of the polypeptide of ATP-dependent proteolysis. Proc Natl Acad Sci U S A77, 1783-1786.

Hirose, F., Yamaguchi, M., Handa, H., Inomata, Y., and Matsukage, A. (1993). Novel8-base pair sequence (Drosophila DNA replication-related element) and specificbinding factor involved in the expression of Drosophila genes for DNA polymerasealpha and proliferating cell nuclear antigen. J Biol Chem 268, 2092-2099.

Hirose, F., Yamaguchi, M., Kuroda, K., Omori, A., Hachiya, T., Ikeda, M.,Nishimoto, Y., and Matsukage, A. (1996). Isolation and characterization of cDNA forDREF, a promoter-activating factor for Drosophila DNA replication-related genes. JBiol Chem 271, 3930-3937.

Hirose, F., Yamaguchi, M., and Matsukage, A. (1999). Targeted expression of theDNA binding domain of DRE-binding factor, a Drosophila transcription factor,attenuates DNA replication of the salivary gland and eye imaginal disc. Mol Cell Biol19, 6020-6028.

50

Hofmann, K., and Falquet, L. (2001). A ubiquitin-interacting motif conserved incomponents of the proteasomal and lysosomal protein degradation systems. TrendsBiochem Sci 26, 347-350.

Hough, R., Pratt, G., and Rechsteiner, M. (1986). Ubiquitin-lysozyme conjugates.Identification and characterization of an ATP-dependent protease from rabbitreticulocyte lysates. J Biol Chem 261, 2400-2408.

Jiang, H., and Monaco, J. J. (1997). Sequence and expression of mouse proteasomeactivator PA28 and the related autoantigen Ki. Immunogenetics 46, 93-98.

Kessel, M., Maurizi, M. R., Kim, B., Kocsis, E., Trus, B. L., Singh, S. K., and Steven,A. C. (1995). Homology in structural organization between E. coli ClpAP proteaseand the eukaryotic 26 S proteasome. J Mol Biol 250, 587-594.Khan, S., van den Broek, M., Schwarz, K., de Giuli, R., Diener, P. A., and Groettrup,M. (2001). Immunoproteasomes largely replace constitutive proteasomes during anantiviral and antibacterial immune response in the liver. J Immunol 167, 6859-6868.

Kisselev, A. F., Akopian, T. N., Woo, K. M., and Goldberg, A. L. (1999). The sizesof peptides generated from protein by mammalian 26 and 20 S proteasomes.Implications for understanding the degradative mechanism and antigen presentation. JBiol Chem 274, 3363-3371.

Knowlton, J. R., Johnston, S. C., Whitby, F. G., Realini, C., Zhang, Z., Rechsteiner,M., and Hill, C. P. (1997). Structure of the proteasome activator REGalpha(PA28alpha). Nature 390, 639-643.

Kuckelkorn, U., Ferreira, E. A., Drung, I., Liewer, U., Kloetzel, P. M., and Theobald,M. (2002). The effect of the interferon-gamma-inducible processing machinery on thegeneration of a naturally tumor-associated human cytotoxic T lymphocyte epitopewithin a wild-type and mutant p53 sequence context. Eur J Immunol 32, 1368-1375.

Kühne, W. (1877b) “Erfahrungen und Bemerkungen über Enzyme und Fermente”Untersuchungen a. d. physiol. Institut der Universität Heidelberg 1, 291–324.

Lam, Y. A., Lawson, T. G., Velayutham, M., Zweier, J. L., and Pickart, C. M. (2002).A proteasomal ATPase subunit recognizes the polyubiquitin degradation signal.Nature 416, 763-767.

Li, J., Gao, X., Ortega, J., Nazif, T., Joss, L., Bogyo, M., Steven, A. C., andRechsteiner, M. (2001). Lysine 188 substitutions convert the pattern of proteasomeactivation by REGgamma to that of REGs alpha and beta. Embo J 20, 3359-3369.

Li, N., Lerea, K. M., and Etlinger, J. D. (1996). Phosphorylation of the proteasomeactivator PA28 is required for proteasome activation. Biochem Biophys Res Commun225, 855-860.

Lier, S., and Paululat, A. (2002). The proteasome regulatory particle subunit Rpn6 isrequired for Drosophila development and interacts physically with signalosomesubunit Alien/CSN2. Gene 298, 109-119.

51

Lowe, J., Stock, D., Jap, B., Zwickl, P., Baumeister, W., and Huber, R. (1995).Crystal structure of the 20S proteasome from the archaeon T. acidophilum at 3.4 Aresolution. Science 268, 533-539.

Ma, C. P., Slaughter, C. A., and DeMartino, G. N. (1992). Identification, purification,and characterization of a protein activator (PA28) of the 20 S proteasome(macropain). J Biol Chem 267, 10515-10523.

Ma, J., Katz, E., and Belote, J. M. (2002a). Expression of proteasome subunitisoforms during spermatogenesis in Drosophila melanogaster. Insect Mol Biol 11,627-639.

Ma, Y., Croxton, R., Moorer, R. L., Jr., and Cress, W. D. (2002b). Identification ofnovel E2F1-regulated genes by microarray. Arch Biochem Biophys 399, 212-224.

Moberg, K., Starz, M. A., and Lees, J. A. (1996). E2F-4 switches from p130 to p107and pRB in response to cell cycle reentry. Mol Cell Biol 16, 1436-1449.

Moriishi, K., Okabayashi, T., Nakai, K., Moriya, K., Koike, K., Murata, S., Chiba, T.,Tanaka, K., Suzuki, R., Suzuki, T., et al. (2003). Proteasome activator PA28gamma-dependent nuclear retention and degradation of hepatitis C virus core protein. J Virol77, 10237-10249.

Morkel, M., Wenkel, J., Bannister, A. J., Kouzarides, T., and Hagemeier, C. (1997).An E2F-like repressor of transcription. Nature 390, 567-568.

Mott, J. D., Pramanik, B. C., Moomaw, C. R., Afendis, S. J., DeMartino, G. N., andSlaughter, C. A. (1994). PA28, an activator of the 20 S proteasome, is composed oftwo nonidentical but homologous subunits. J Biol Chem 269, 31466-31471.

Murata, S., Kawahara, H., Tohma, S., Yamamoto, K., Kasahara, M., Nabeshima, Y.,Tanaka, K., and Chiba, T. (1999). Growth retardation in mice lacking the proteasomeactivator PA28gamma. J Biol Chem 274, 38211-38215.

Murata, S., Udono, H., Tanahashi, N., Hamada, N., Watanabe, K., Adachi, K.,Yamano, T., Yui, K., Kobayashi, N., Kasahara, M., et al. (2001). Immunoproteasomeassembly and antigen presentation in mice lacking both PA28alpha and PA28beta.Embo J 20, 5898-5907.

Navon, A., and Goldberg, A. L. (2001). Proteins are unfolded on the surface of theATPase ring before transport into the proteasome. Mol Cell 8, 1339-1349.

Niedermann, G., Butz, S., Ihlenfeldt, H. G., Grimm, R., Lucchiari, M., Hoschutzky,H., Jung, G., Maier, B., and Eichmann, K. (1995). Contribution of proteasome-mediated proteolysis to the hierarchy of epitopes presented by majorhistocompatibility complex class I molecules. Immunity 2, 289-299.

52

Nikaido, T., Shimada, K., Nishida, Y., Lee, R. S., Pardee, A. B., and Nishizuka, Y.(1989). Loss in transformed cells of cell cycle regulation of expression of a nuclearprotein recognized by SLE patient antisera. Exp Cell Res 182, 284-289.

NORTHROP J. H. The preparation of pure enzymes and virus proteins. NobelLecture, December 12, 1946

Ogawa, H., Ishiguro, K., Gaubatz, S., Livingston, D. M., and Nakatani, Y. (2002). Acomplex with chromatin modifiers that occupies E2F- and Myc-responsive genes inG0 cells. Science 296, 1132-1136.

Ohshima, N., Takahashi, M., and Hirose, F. (2003). Identification of a humanhomologue of the DREF transcription factor with a potential role in regulation of thehistone H1 gene. J Biol Chem 278, 22928-22938.

Ohtani, K., and Nevins, J. R. (1994). Functional properties of a Drosophila homologof the E2F1 gene. Mol Cell Biol 14, 1603-1612.

Okamura, T., Taniguchi, S., Ohkura, T., Yoshida, A., Shimizu, H., Sakai, M., Maeta,H., Fukui, H., Ueta, Y., Hisatome, I., and Shigemasa, C. (2003). Abnormally highexpression of proteasome activator-gamma in thyroid neoplasm. J Clin EndocrinolMetab 88, 1374-1383.

Paesen, G. C., and Nuttall, P. A. (1996). A tick homologue of the human Ki nuclearautoantigen. Biochim Biophys Acta 1309, 9-13.

Palmer, A., Rivett, A. J., Thomson, S., Hendil, K. B., Butcher, G. W., Fuertes, G., andKnecht, E. (1996). Subpopulations of proteasomes in rat liver nuclei, microsomes andcytosol. Biochem J 316 ( Pt 2), 401-407.

Pickart, C. M. (2004). Back to the future with ubiquitin. Cell 116, 181-190.Pitzer, F., Dantes, A., Fuchs, T., Baumeister, W., and Amsterdam, A. (1996).Removal of proteasomes from the nucleus and their accumulation in apoptotic blebsduring programmed cell death. FEBS Lett 394, 47-50.

Puhler, G., Weinkauf, S., Bachmann, L., Muller, S., Engel, A., Hegerl, R., andBaumeister, W. (1992). Subunit stoichiometry and three-dimensional arrangement inproteasomes from Thermoplasma acidophilum. Embo J 11, 1607-1616.

Rawlings, N. D., and Barrett, A. J. (1999). MEROPS: the peptidase database. NucleicAcids Res 27, 325-331.

Rawlings, N. D., O'Brien, E., and Barrett, A. J. (2002). MEROPS: the proteasedatabase. Nucleic Acids Res 30, 343-346.

Realini, C., Dubiel, W., Pratt, G., Ferrell, K., and Rechsteiner, M. (1994). Molecularcloning and expression of a gamma-interferon-inducible activator of the multicatalyticprotease. J Biol Chem 269, 20727-20732.

53

Realini, C., Jensen, C. C., Zhang, Z., Johnston, S. C., Knowlton, J. R., Hill, C. P., andRechsteiner, M. (1997). Characterization of recombinant REGalpha, REGbeta, andREGgamma proteasome activators. J Biol Chem 272, 25483-25492.

Reaumur, Mr. de (1761) “Sur la digestion des oiseaux, second mémoire” Tirés desRegistres de l’Academie Royale des Sciences, Mém. Math. Phys. ser. 2, 3, 701–752.

Rechsteiner, M., Realini, C., and Ustrell, V. (2000). The proteasome activator 11 SREG (PA28) and class I antigen presentation. Biochem J 345 Pt 1, 1-15.Riboldi, P., Asero, R., Origgi, L., and Crespi, S. (1987). The SL/Ki system inconnective tissue diseases: incidence and clinical associations. Clin Exp Rheumatol 5,29-33.

Rinaldi, T., Ricci, C., Porro, D., Bolotin-Fukuhara, M., and Frontali, L. (1998). Amutation in a novel yeast proteasomal gene, RPN11/MPR1, produces a cell cyclearrest, overreplication of nuclear and mitochondrial DNA, and an alteredmitochondrial morphology. Mol Biol Cell 9, 2917-2931.

Rivett, A. J., Palmer, A., and Knecht, E. (1992). Electron microscopic localization ofthe multicatalytic proteinase complex in rat liver and in cultured cells. J HistochemCytochem 40, 1165-1172.

Rock, K. L., Gramm, C., Rothstein, L., Clark, K., Stein, R., Dick, L., Hwang, D., andGoldberg, A. L. (1994). Inhibitors of the proteasome block the degradation of mostcell proteins and the generation of peptides presented on MHC class I molecules. Cell78, 761-771.

Sakamoto, M., Takasaki, Y., Yamanaka, K., Kodama, A., Hashimoto, H., and Hirose,S. (1989). Purification and characterization of Ki antigen and detection of anti-Kiantibody by enzyme-linked immunosorbent assay in patients with systemic lupuserythematosus. Arthritis Rheum 32, 1554-1562.

Schmidt, M., and Kloetzel, P. M. (1997). Biogenesis of eukaryotic 20S proteasomes:the complex maturation pathway of a complex enzyme. Faseb J 11, 1235-1243.

Schwarz, K., Eggers, M., Soza, A., Koszinowski, U. H., Kloetzel, P. M., andGroettrup, M. (2000). The proteasome regulator PA28alpha/beta can enhance antigenpresentation without affecting 20S proteasome subunit composition. Eur J Immunol30, 3672-3679.

Seemuller, E., Lupas, A., Stock, D., Lowe, J., Huber, R., and Baumeister, W. (1995).Proteasome from Thermoplasma acidophilum: a threonine protease. Science 268,579-582.

Song, X., Mott, J. D., von Kampen, J., Pramanik, B., Tanaka, K., Slaughter, C. A.,and DeMartino, G. N. (1996). A model for the quaternary structure of the proteasomeactivator PA28. J Biol Chem 271, 26410-26417.

54

Sousa, M. C., Trame, C. B., Tsuruta, H., Wilbanks, S. M., Reddy, V. S., and McKay,D. B. (2000). Crystal and solution structures of an HslUV protease-chaperonecomplex. Cell 103, 633-643.

Soza, A., Knuehl, C., Groettrup, M., Henklein, P., Tanaka, K., and Kloetzel, P. M.(1997). Expression and subcellular localization of mouse 20S proteasome activatorcomplex PA28. FEBS Lett 413, 27-34.

Spallanzani, A. L. (1780) Dissertazioni di Fisica Animale, e Vegetabile, Presso laSocietà Tipografica, Modena [Engl. Transl. (1789) by T. BEDDOES as DissertationsRelative to the Natural History of Animals, J. Murray, London].

Stechmann, A., and Cavalier-Smith, T. (2002). Rooting the eukaryote tree by using aderived gene fusion. Science 297, 89-91.

Stelter, P., and Ulrich, H. D. (2003). Control of spontaneous and damage-inducedmutagenesis by SUMO and ubiquitin conjugation. Nature 425, 188-191.

Sun, Y., Sijts, A. J., Song, M., Janek, K., Nussbaum, A. K., Kral, S., Schirle, M.,Stevanovic, S., Paschen, A., Schild, H., et al. (2002). Expression of the proteasomeactivator PA28 rescues the presentation of a cytotoxic T lymphocyte epitope onmelanoma cells. Cancer Res 62, 2875-2882.

Szlanka, T., Haracska, L., Kiss, I., Deak, P., Kurucz, E., Ando, I., Viragh, E., andUdvardy, A. (2003). Deletion of proteasomal subunit S5a/Rpn10/p54 causes lethality,multiple mitotic defects and overexpression of proteasomal genes in Drosophilamelanogaster. J Cell Sci 116, 1023-1033.

Takasaki, Y., Yano, T., Hirokawa, K., Takeuchi, K., Ando, S., Takahashi, T.,Shimada, K., and Hashimoto, H. (1996). An epitope on Ki antigen recognized byautoantibodies from lupus patients shows homology with the SV40 large T antigennuclear localization signal. Arthritis Rheum 39, 855-862.

Takeuchi, J., and Toh-e, A. (2001). Genetic dissection of the yeast 26S proteasome:cell cycle defects caused by the Deltarpn9 mutation. Biochimie 83, 333-340.

Tamura, T., Tamura, N., Cejka, Z., Hegerl, R., Lottspeich, F., and Baumeister, W.(1996). Tricorn protease--the core of a modular proteolytic system. Science 274,1385-1389.

Tanahashi, N., Yokota, K., Ahn, J. Y., Chung, C. H., Fujiwara, T., Takahashi, E.,DeMartino, G. N., Slaughter, C. A., Toyonaga, T., Yamamura, K., et al. (1997).Molecular properties of the proteasome activator PA28 family proteins and gamma-interferon regulation. Genes Cells 2, 195-211.

Tanaka, K., Ii, K., Ichihara, A., Waxman, L., and Goldberg, A. L. (1986). A highmolecular weight protease in the cytosol of rat liver. I. Purification, enzymologicalproperties, and tissue distribution. J Biol Chem 261, 15197-15203.

55

Tanaka, K., Waxman, L., and Goldberg, A. L. (1983). ATP serves two distinct rolesin protein degradation in reticulocytes, one requiring and one independent ofubiquitin. J Cell Biol 96, 1580-1585.

Tojo, T., Kaburaki, J., Hayakawa, M., Okamoto, T., Tomii, M., and Homma, M.(1981). Precipitating antibody to a soluble nuclear antigen "Ki" with specificity forsystemic lupus erythematosus. Ryumachi 21 Suppl, 129-140.

Unno, M., Mizushima, T., Morimoto, Y., Tomisugi, Y., Tanaka, K., Yasuoka, N., andTsukihara, T. (2002). The structure of the mammalian 20S proteasome at 2.75 Aresolution. Structure (Camb) 10, 609-618.

Ustrell, V., Pratt, G., and Rechsteiner, M. (1995). Effects of interferon gamma andmajor histocompatibility complex-encoded subunits on peptidase activities of humanmulticatalytic proteases. Proc Natl Acad Sci U S A 92, 584-588.

van Hall, T., Sijts, A., Camps, M., Offringa, R., Melief, C., Kloetzel, P. M., andOssendorp, F. (2000). Differential influence on cytotoxic T lymphocyte epitopepresentation by controlled expression of either proteasome immunosubunits or PA28.J Exp Med 192, 483-494.

Van Kaer, L., Ashton-Rickardt, P. G., Eichelberger, M., Gaczynska, M., Nagashima,K., Rock, K. L., Goldberg, A. L., Doherty, P. C., and Tonegawa, S. (1994). Alteredpeptidase and viral-specific T cell response in LMP2 mutant mice. Immunity 1, 533-541.

Walz, J., Erdmann, A., Kania, M., Typke, D., Koster, A. J., and Baumeister, W.(1998). 26S proteasome structure revealed by three-dimensional electron microscopy.J Struct Biol 121, 19-29.

Watson, J.D. & Crick, F.H.C. (1953) A structure for desoxyribose nucleic acid.Nature 171, 737-738.

Whitby, F. G., Masters, E. I., Kramer, L., Knowlton, J. R., Yao, Y., Wang, C. C., andHill, C. P. (2000). Structural basis for the activation of 20S proteasomes by 11Sregulators. Nature 408, 115-120.

Wilk, S., Chen, W. E., and Magnusson, R. P. (2000). Properties of the nuclearproteasome activator PA28gamma (REGgamma). Arch Biochem Biophys 383, 265-271.

Wilkinson, C. R., Ferrell, K., Penney, M., Wallace, M., Dubiel, W., and Gordon, C.(2000). Analysis of a gene encoding Rpn10 of the fission yeast proteasome revealsthat the polyubiquitin-binding site of this subunit is essential when Rpn12/Mts3activity is compromised. J Biol Chem 275, 15182-15192.

Wilkinson, F. L., Holaska, J. M., Zhang, Z., Sharma, A., Manilal, S., Holt, I., Stamm,S., Wilson, K. L., and Morris, G. E. (2003). Emerin interacts in vitro with thesplicing-associated factor, YT521-B. Eur J Biochem 270, 2459-2466.

56

Wojcik, C., and DeMartino, G. N. (2002). Analysis of Drosophila 26 S proteasomeusing RNA interference. J Biol Chem 277, 6188-6197.Wojcik, C., and DeMartino, G. N. (2003). Intracellular localization of proteasomes.Int J Biochem Cell Biol 35, 579-589.

Xie, Y., and Varshavsky, A. (2000). Physical association of ubiquitin ligases and the26S proteasome. Proc Natl Acad Sci U S A 97, 2497-2502.

Yamaguchi, M., Hayashi, Y., Hirose, F., Nishimoto, Y., and Matsukage, A. (1997).Distinct roles of E2F recognition sites as positive or negative elements in regulationof the DNA polymerase alpha 180 kDa catalytic subunit gene promoter duringDrosophila development. Nucleic Acids Res 25, 3847-3854.

Yao, Y., Huang, L., Krutchinsky, A., Wong, M. L., Standing, K. G., Burlingame, A.L., and Wang, C. C. (1999). Structural and functional characterizations of theproteasome-activating protein PA26 from Trypanosoma brucei. J Biol Chem 274,33921-33930.

Yewdell, J., Lapham, C., Bacik, I., Spies, T., and Bennink, J. (1994). MHC-encodedproteasome subunits LMP2 and LMP7 are not required for efficient antigenpresentation. J Immunol 152, 1163-1170.

Young, P., Deveraux, Q., Beal, R. E., Pickart, C. M., and Rechsteiner, M. (1998).Characterization of two polyubiquitin binding sites in the 26 S protease subunit 5a. JBiol Chem 273, 5461-5467.

Zhang, Z., Clawson, A., Realini, C., Jensen, C. C., Knowlton, J. R., Hill, C. P., andRechsteiner, M. (1998a). Identification of an activation region in the proteasomeactivator REGalpha. Proc Natl Acad Sci U S A 95, 2807-2811.

Zhang, Z., Krutchinsky, A., Endicott, S., Realini, C., Rechsteiner, M., and Standing,K. G. (1999). Proteasome activator 11S REG or PA28: recombinant REG alpha/REGbeta hetero-oligomers are heptamers. Biochemistry 38, 5651-5658.

Zhang, Z., Realini, C., Clawson, A., Endicott, S., and Rechsteiner, M. (1998b).Proteasome activation by REG molecules lacking homolog-specific inserts. J BiolChem 273, 9501-9509.

Zheleva, D. I., McInnes, C., Gavine, A. L., Zhelev, N. Z., Fischer, P. M., and Lane, D.P. (2002). Highly potent p21(WAF1)-derived peptide inhibitors of CDK-mediatedpRb phosphorylation: delineation and structural insight into their interactions withcyclin A. J Pept Res 60, 257-270.

Zwickl, P., Grziwa, A., Puhler, G., Dahlmann, B., Lottspeich, F., and Baumeister, W.(1992). Primary structure of the Thermoplasma proteasome and its implications forthe structure, function, and evolution of the multicatalytic proteinase. Biochemistry31, 964-972.