The Proteasome α-Subunit XAPC7 Interacts Specifically with Rab7 and Late Endosomes

Rab7 is a key regulatory protein governing early to late endocytic membrane transport. In this study the proteasome α-subunit XAPC7 (also known as PSMA7, RC6-1, and HSPC in mammals) was identified to interact specifically with Rab7 and was recruited to multivesicular late endosomes through this interaction. The protein interaction domains were localized to the C terminus of XAPC7 and the N terminus of Rab7. XAPC7 was not found on early or recycling endosomes, but could be recruited to recycling endosomes by expression of a Rab7-(1-174)Rab11-(160-202) chimera, establishing a central role for Rab7 in the membrane recruitment of XAPC7. Although XAPC7 could be shown to associate with membranes bearing ubiquitinated cargo, overexpression had no impact on steady-state ubiquitinated protein levels. Most notably, overexpression of XAPC7 was found to impair late endocytic transport of two different membrane proteins, including EGFR known to be highly dependent on ubiquitination and proteasome activity for proper endocytic sorting and lysosomal transport. Decreased late endocytic transport caused by XAPC7 overexpression was partially rescued by coexpression of wild-type Rab7, suggesting a negative regulatory role for XAPC7. Nevertheless, Rab7 itself was not subject to XAPC7-dependent proteasomal degradation. Together the data establish the first direct molecular link between the endocytic trafficking and cytosolic degradative machineries.

Rab7 is a key regulatory protein governing early to late endocytic membrane transport. In this study the proteasome ␣-subunit XAPC7 (also known as PSMA7, RC6 -1, and HSPC in mammals) was identified to interact specifically with Rab7 and was recruited to multivesicular late endosomes through this interaction. The protein interaction domains were localized to the C terminus of XAPC7 and the N terminus of Rab7. XAPC7 was not found on early or recycling endosomes, but could be recruited to recycling endosomes by expression of a Rab7-(1-174)Rab11-(160 -202) chimera, establishing a central role for Rab7 in the membrane recruitment of XAPC7. Although XAPC7 could be shown to associate with membranes bearing ubiquitinated cargo, overexpression had no impact on steady-state ubiquitinated protein levels. Most notably, overexpression of XAPC7 was found to impair late endocytic transport of two different membrane proteins, including EGFR known to be highly dependent on ubiquitination and proteasome activity for proper endocytic sorting and lysosomal transport. Decreased late endocytic transport caused by XAPC7 overexpression was partially rescued by coexpression of wild-type Rab7, suggesting a negative regulatory role for XAPC7. Nevertheless, Rab7 itself was not subject to XAPC7-dependent proteasomal degradation. Together the data establish the first direct molecular link between the endocytic trafficking and cytosolic degradative machineries.
Receptor-mediated endocytosis involves the sequential internalization and passage of internalized receptors through a series of endosomal compartments. Initially receptors pass through early endosomes from which plasma membrane recycling may take place or transfer to late endosomes can occur. Transfer to late endosomes, generally exhibiting a distinctive multivesicular morphology, often is accompanied by sorting from the limiting membrane of these structures to internal membrane vesicles. Sorting to the luminal vesicles of multivesicular bodies is a process demonstrated to be important in the down-regulation of receptor signaling and membrane protein sorting (1). Following sorting in late endosomes, membrane proteins may have at least two fates, including recycling to the trans-Golgi or transfer to lysosomes. The existence of a direct plasma membrane recycling route from late endosomes is suggested by the release of exosomes following the fusion of multivesicular bodies in antigen-presenting cells (2). Although receptor degradation may be initiated in late endosomes, complete degradation typically occurs only after lysosomal delivery. Thus the endocytic pathway, whose overall function is to correctly process and target incoming cargo to the appropriate destinations, is comprised of an intricate series of networked compartments, each specialized to execute specific activities.
The proper transfer of internalized receptors from one endosomal compartment to the next is closely regulated and subject to control by recognition of sorting signals, inclusion within select membrane microdomains and specific targeting of endocytic carriers to their appropriate destination. Over the last decade a great deal of attention has focused on delineating the components of the endocytic transport and sorting machineries that ensure transport specificity. As a result the Rab GTPases have emerged as key regulators of endosomal transport, with 13 of over 60 identified family members having defined functions in endocytosis (3). Among the endosomal Rab proteins, seven have been studied in detail, and results have provided insights into the control of individual endocytic transport steps. Rab5, Rab4, Rab11, and Rab15 are documented to control early endocytic and plasma membrane recycling, while Rab7, Rab9, and Rab34 control late endocytic events. The general consensus that has emerged from the study of these Rab GTPases and the identification of numerous interacting partners and effectors can be summarized as follows. The Rab GTPases serve as molecular switches that are activated on GTP binding and inactivated following nucleotide hydrolysis. In their active form, Rab proteins serve to provide temporal and spatial control over vesicle budding, cytoskeletal transport and membrane fusion (4). Rab proteins accomplish this task by interacting with numerous partner proteins in a sequential and spatially defined manner. For example, Rab5 localized at the plasma membrane facilitates clathrin-coated vesicle budding and the subsequent interaction with the cytoskeleton mediates the transport of the small vesicles to early endosomes (5,6). On early endosomes, Rab5 recruits endosome tethering and fusion factors including the rabaptins and EEA1 to promote homotypic fusion. The latter processes are orchestrated by the formation of phosphorylated phosphoinositide lipid domains in response to the recruitment and activation of the PI 3Ј-kinase hVPS34/p150 complex by active early endosomal Rab5 (7). Some Rab5 interacting proteins are shared in common between Rab4 and Rab7 that are operative on adjacent or overlapping branches of the endocytic pathway. However as increasing numbers of Rab-interacting proteins are identified, many are proving to be unique partners of individual Rab proteins. The importance and uniqueness of each of these processes is further underscored by the existence of distinct human diseases caused by mutations in different Rab proteins or associated transport machinery (3). Thus, the endocytic Rab proteins collectively govern cargo selection, cytoskeletal transport and organization of specific membrane domains, but do so in a manner that appears uniquely tailored to each endocytic compartment and transport step.
Our laboratory has focused on defining the function of the late endocytic Rab7 GTPase. Using both morphological and biochemical assays we have delineated a function for Rab7 in the transfer of both internalized cargo and newly synthesized lysosomal enzymes from early to late endosomes (8,9,10). Lysosomal degradation has also been shown to depend on active Rab7, though it is unclear if Rab7 directly controls heterotypic fusion of late endosomes and lysosomes (11). The first known Rab7 effector, termed Rab-interacting lysosome protein (RILP), 1 is a novel protein shown to control endosomal and lysosomal clustering by mediating recruitment of the minusend directed dynein/dynactin motors (12,13,14,15). Recently, two further Rab7 effectors have been identified, one a novel protein called Rabring7 (Rab7-interacting ring finger protein) and the second, an effector shared in common with Rab5, hVPS34/p150 (16,17). Rabring7 contains a ring finger motif at its C terminus and interacts specifically with active Rab7 via its N-terminal sequences (16). Overexpression of Rabring7 caused perinuclear clustering of endosomes and lysosomes much like RILP, but in contrast to RILP, also interfered with receptor degradation (12,16). The hVPS34-p150 complex was shown to colocalize preferentially to Rab7-positive late endosomes and hVPS34 phosphatidylinositol 3Ј-kinase activity was dependent on nucleotide cycling of Rab7 (17). The cumulative data to date suggest Rab7 controls cargo transport to late endosomes and degradative compartments, in part through control of cytoskeletal associations and the generation of PI3P membrane domains. Furthermore, Rab7-interacting partners may have positive or negative regulatory properties on late endocytic transport.
PI3P membrane domains provide an important platform for the recruitment and assembly of multiple protein complexes. In the case of late endosomes, emerging data suggest that PI3P membrane domains provide a critical link to late endosomal receptor sorting that is driven by a cycle of receptor ubiquitination and deubiquitination and depends in part on active proteasome (18). Proteins containing FYVE domains are specifically recruited to membranes by virtue of PI3P recognition by the FYVE motif (19). Among the known FYVE-domain proteins, Hrs (hepatocyte growth factor-regulated tyrosine kinase substrate, also known as Vps27 in yeast) has been the focus of significant attention as a protein with dual binding specificities for PI3P and ubiquitin. Hrs also leads off the recruitment of the ESCRT protein complexes I-III that promote intraluminal vesicle formation and receptor sorting in an ubiquitin-dependent manner (20). Thus, Hrs serves as a key adaptor between PI3P domains and the recognition of ubiquitinated cargo via its ubiquitin-interacting motif. Because of the importance of Rab7 in regulating PI3P formation on late endosomes it was of significant interest to analyze whether or not Rab7 might also be linked to the machinery controlling receptor ubiquitination and late endosomal sorting. Here, we report the identification of an ␣-subunit of the 26 S proteasome as a new specific interacting partner of Rab7, demonstrating for the first time a direct molecular link between proteasomes and the endocytic sorting machinery.

EXPERIMENTAL PROCEDURES
Cells, Antibodies, and Reagents-BHK21 cells were obtained from American Type Culture Collection. BHK21 cells were grown in complete G-MEM (10% fetal calf serum, 2 mM glutamine, 50 units/ml penicillin, 50 g/ml streptomycin, and 2.6 mg/ml tryptose phosphate broth). All cell culture reagents were from Invitrogen. Reagents for SDS-PAGE were purchased from Bio-Rad, except acrylamide, which was obtained from Genomic Solutions (Ann Arbor, MI). Horseradish peroxidase-conjugated antibodies were obtained from Amersham Biosciences and chemiluminescence detection reagent (ECL TM ) was from Pierce Biochemical Corp. Reagents for immunofluorescence experiments included Mowiol 4 -88 from Calbiochem (La Jolla, CA), saponin, paraformaldehyde, and protease inhibitors MG132 from Sigma Chemical Co., and secondary antibodies and detection reagents from Vector Laboratories (Burlingame, CA) and Jackson ImmunoResearch (West Grove, PA). Monoclonal I1 (8G5F11), recognizing an ectoplasmic epitope of the vesicular stomatitis virus (VSV) G protein (21) was used for the transport assay. Monoclonal antibody P5D4 recognizing the VSVG protein C-terminal TDIEMNRLG sequence (22) was used to detect expression of the VSVG protein epitope tag in the G-XAPC7 construct. A chicken and a rabbit polyclonal antibody were raised against Rab7 C-terminal KQETEVELYNEFPEPIKLDKNDRAKT-SAESCSC sequence (23,8) and both were used for transport assays and immunostaining studies. A rabbit polyclonal antiserum reactive against XAPC7, an ␣-type subunit of the 26 S canine proteasome, was raised using clone C7 (108 C-terminal amino acids of canine XAPC7) expressed in pGEX and purified from Escherichia coli on glutathioneagarose as the antigen. The polyclonal XAPC7 antibody specifically recognized the XAPC7 ␣-subunit (tested against hamster, canine, and human XAPC7), but not other ␣-subunits in immunoprecipitation, immunoblotting, and immunofluorescence assays (data not shown). Monoclonal antibody FK2 demonstrated to recognize only mono-and multiubiquitinated proteins, but not free ubiquitin, as well as monoclonal antibodies reactive against proteasome subunits HC2 and XAPC7, and a rabbit polyclonal antibody directed against the 20 S core part of proteasome were all purchased from AFFINITI Research Products Ltd (Maunhead, UK). Rabbit anti-ubiquitin used in immunoblotting was purchased from DakoCytomation (Carpinteria, CA). Anti-actin monoclonal antibody was purchased from Chemicon International (Temecula, CA).
Plasmids-Plasmids encoding wild-type VSVG protein (pAR-G) (24), Rab5, Rab7 (23), Rab7N125I (25), Rab7T22N (8) under the control of the T7 promoter were used for all expression studies. Inserting an EcoR1 fragment containing the human transferrin receptor cDNA into pcDNA3 vector generated plasmid pcDNA3-hTfR. We replaced the 33 C-terminal residues of Rab7 with the 43 C-terminal residues of Rab11 to make the Rab7-(1-174)Rab11-(160 -202) chimera. This cDNA fragment was obtained using a polymerase chain reaction (PCR)-based approach (26), and inserted into the GFP-C1 vector (BD Biosciences, San Jose, CA). The open reading frame of Rab7-(1-174)Rab11-(160 -202) was completely confirmed by nucleotide sequencing. Wild-type canine Rab2, Rab5, and Rab9 cDNAs (23) were cloned into pEG202 (2, HIS3 ϩ , Amp r ) (27) such that the resulting bait proteins were fused to amino acids 1-202 of LexA and expressed under the control of the constitutive ADH1 promoter. To construct expression vectors encoding the full-length XAPC7 clone, we ordered human expressed sequence tag (EST) clone (IMAGE: 84457, Stratagene, La Jolla, CA) and confirmed that it encoded the full-length cDNA of proteasome ␣-subunit XAPC7 by nucleotide sequencing. We amplified the cDNA using PCR amplification and then subcloned it in pGEM3 by restriction digest with EcoRI and XhoI. The inserted fragment was released with EcoRI/HindIII digestion. An epitope-tagged variant of G-XAPC7 was generated in pCDNA3 by fusing the TDIEMNRLG sequence derived from VSVG protein in-frame onto the N terminus of XAPC7 using an appropriate primer and a PCR-based strategy.
MDCK cDNA Library and Yeast Two-hybrid Assays-An MDCK cDNA library was constructed and screened as previously described (28). Briefly, the library plasmids in pJG4 -5 were prepared from a pool of unamplified 3.0 ϫ 10 6 primary transformants. More than 95% of the library plasmids contained MDCK cDNA inserts with an average size of 1.3 kb. The yeast reporter strain EGY48 (MATa trp1 ura3 his3 LEU2::pLexAop6-LEU2) was used as a host for all two-hybrid assays. The plasmid pLexA-Rab7Q67L (2, HIS3 ϩ , Amp r ) with expression under the control of the constitutive ADH1 promoter was prepared and used as the two-hybrid bait (28). EGY48 was sequentially transformed with the pLexA-Rab7Q67L plasmid and with the MDCK cDNA library. Transformants were plated on synthetic medium lacking tryptophan, histidine, and uracil and incubated at 30°C for 2 days. The transformants were isolated, and Leu ϩ and LacZ ϩ yeast were selected (29,27). Library plasmids from positive clones were rescued using E. coli XL1-Blue cells and subsequently analyzed by transformation tests and DNA sequencing. DNA databases were searched using GCG sequence analysis software package (Genetics Computer Group, Inc.). Four independent clones out of 123 positive clones encoded the C-terminal part of a proteasome ␣-subunit XAPC7. Clone C7 consisted of 600 bp and encoded the last 108 amino acids of canine XAPC7. C7 was used to test the interaction specificity with various wild-type Rab proteins or mutant Rab7 by growth and ␤-galactosidase assays (both liquid and agar plate assays).
Accession Number and BLAST Analyses-The partial canine XAPC7 sequence was deposited in GenBank TM (accession number AY526609). Over the length of the partial canine clone there is 98% amino acid sequence identity with the human XAPC7 protein. XAPC7 is also known as PSMA7 variant 1 and HSPC in humans; PSMA7, C6-I in mouse; PSMA7, RC6 -1 in rat; DD5 in Dictyostelium discoideum; and ␣4 in fish and amphibians. Identity between mammalian forms of XAPC7 or PSMA7 is greater than 98%. Rat RC6 -1, which contains a 6 amino acid insert, exhibits 95% identity. Goldfish and Xenopus laevis ␣4 exhibit 95% identity. Homologues with lower homologies exist in slime mold, insects, and plants.
Transient Infection/Transfection-Transient overexpression studies were performed using the T7 RNA polymerase recombinant vaccinia virus expression system as described (30,8). Briefly, BHK21 grown to less than 70ϳ80% confluent were infected with T7 RNA polymeraserecombinant vaccinia virus (vTF7.3). Cells were then transfected with plasmids containing cDNA under the control of the T7 promoter using LipofectAMINE (Invitrogen) reagent according to the manufacturer's instructions. For all experiments, transfected cells were incubated for 5ϳ6 h at 37°C in a 5% CO 2 incubator prior to analysis.
Immunoprecipitations-As described above, BHK21 cells grown on 6-well plates were transiently infected/transfected with plasmids encoding XAPC7 or different forms of Rab7 or Rab5. Mock transfections were performed using vectors lacking an insert. Under the experimental conditions used, transfection efficiencies are very high (Ͼ90%). After 4.5 h, transfected BHK cells were incubated for 30 min in medium without methionine and cysteine and were metabolically labeled using 100 Ci/ml Tran 35 S-label (ICN Biomedicals, Irvine, CA) for 30 min at 37°C. The cells were then washed twice with PBS. For analysis, cells were lysed in 500 l of radioimmune precipitation assay (RIPA) buffer (1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 7.4, and 150 mM NaCl) with PB/CLAP protease inhibitor mixture (1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 1 g/ml chymostatin, leupeptin, antipain, and pepstatin A) added freshly. RIPA lysates were clarified at 15,000 rpm in an Eppendorf microcentrifuge for 15 min at 4°C. The supernatant was precleared by incubation with Pansorbin for 1 h and subsequently clarified at 15,000 rpm for 15 min at 4°C. The primary antibodies and Pansorbin were added to the lysates with all incubations performed at 4°C. The immune complexes were collected and washed three times with RIPA buffer, and once more with 50 mM Tris-HCl (pH 7.4) by resuspension and centrifugation in a cold Eppendorf microcentrifuge at 2,500 rpm for 5 min. The washed immunoprecipitates were resuspended in SDS-PAGE sample buffer (50 mM Tris-HCl, pH 6.8, 2.5 mM EDTA, 2% SDS, 7% glycerol, and 0.01% bromphenol blue), heated to 95°C for 10 min and resolved on 12.5% SDS-polyacrylamide gels. Dried gels were analyzed on a Molecular Dynamics Storm PhosphorImager.
For IP/Westerns, unlabeled cell lysates were prepared as detailed above except 10-cm dishes of BHK cells were used for the initial transfection, lysates were prepared by addition of 300 l of RIPA buffer, and immunoprecipitates were isolated using a rabbit polyclonal antibody directed against Rab7 as the primary antibody.
Immunoelectron Microscopy-For immunoelectron microscopy, cells in suspension were lightly fixed, pelleted, then fixed in 4% paraformaldehyde and 0.01% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4 for 1 h. The pellets were washed and infiltrated in 25% polyvinylpyrrolidone and 1.86 M sucrose in 0.1 M phosphate buffer. The pellet was frozen in liquid nitrogen and sections were cut on a Leica EMFCS ultramicrotome. Cryosections were placed on formvar-coated grids and incubated for 1 h with an affinity-purified rabbit antibody directed against Rab7 to detect overexpressed Rab7, or purified chicken IgY directed against Rab7 to detect the endogenous protein. Overexpressed G-XAPC7 was detected with a monoclonal antibody P5D4 against the G-epitope tag or a rabbit polyclonal directed against XAPC7. Appropriate gold-conjugated secondary antibodies were from Jackson Immunoresearch Laboratories, Inc. (West Grove, PA). Grids were washed, fixed in 2% glutaraldehyde in PBS for 10 min, counterstained with uranyl acetate for 40 min and embedded in polyvinyl alcohol. The grids were examined in a Hitachi H7500 electron microscope at 80 kV.
Subcellular Fractionation and Immunoblotting-BHK-21 cells were transiently infected/transfected with a plasmid encoding XAPC7. Mock transfection was performed using vector lacking an insert. For each sample, two 100-mm dishes were washed twice and scraped in PBS using a wiper blade. From this point on, all buffers contained a protease inhibitor mixture (PB/CLAP) consisting of 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, and 1 g/ml each of chymostatin, leupeptin, antipain, and pepstatin A. Cells were pelleted by centrifugation at 1,000 rpm in a rotor (model SA600; Sorvall) for 10 min at 4°C. The cell pellet was resuspended in 10 ml of homogenization buffer (HB; 20 mM Hepes, pH 7.4, and 2 mM CaCl 2 ). The cells were again collected by centrifugation and resuspended in three-pellet volumes of HB. The suspension was passed through a 27-gauge needle 12 times. Cell debris and nuclei were removed by two successive centrifugation steps at 1,000 rpm for 5 min in an Eppendorf microcentrifuge at 4°C. Mitochondria were removed by centrifugation at 1,950 ϫ g for 20 min, and a crude membrane fraction was collected by centrifugation at 105,000 ϫ g for 60 min. About 300 l of the total supernatant was collected as the cytosol fraction and the membrane pellet was resuspended in 150 l of RIPA buffer. Equal volumes of each sample were resolved by SDS-PAGE and subjected to immunoblot analysis. Endogenous actin was used as control for protein loading and as a cytosolic marker.
For IP/Westerns, immunoblot analysis of unlabeled immunoprecipitates was performed by resolving the immunoprecipitates and total lysate samples on 12% SDS-PAGE. The resolved proteins were transferred to Hybond-C membranes (Amersham Biosciences) using a plate electrode blotter at 500 mA for 30 min. The membranes were probed with a monoclonal antibody directed against to XAPC7.
Immunofluorescence Microscopy-Cells were grown and processed for immunofluorescence staining as described previously (8). Briefly, cells were pre-permeabilized with 0.05% (v/v) saponin in Pipes buffer (80 mM Pipes-KOH, pH 6.8, 5 mM EGTA, 1 mM MgCl 2 ) for 2-3 min and fixed with 3% (w/v) paraformaldehyde in PBS. All incubations with primary and detecting antibodies were conducted at room temperature for 30 min with antibodies diluted in PBS/0.05% saponin. The coverslips were washed four times with PBS/saponin and mounted on glass slides in Mowiol 4 -88. Cells were viewed on a Zeiss LSM510 confocal microscope equipped with a ϫ63 oil immersion lens. Images were transferred as JPEG files to Adobe Photoshop for compilation.
Morphological Assay of VSVG Protein Transport-The assay was performed as described previously (8,17). Briefly, BHK21 cells were infected/transfected with plasmids to induce overexpression of VSVG protein, XAPC7 and/or Rab7. After a 5-h incubation at 37°C cells were washed twice with ice-cold low bicarbonate, serum-free MEM and then incubated with monoclonal antibody 8G5 directed against the ectoplasmic domain of VSVG protein in serum-free MEM for 30 min at 4°C with occasional shaking. Cells were then washed twice with serum-free, low bicarbonate MEM and transferred to a 15°C water bath in the same medium. After 45 min at 15°C, the incubation medium was replaced with prewarmed, complete G-MEM and the cells were placed in 5% CO 2 incubator for 30 min. Finally, the cells were washed with PBS, permeabilized with 0.05% saponin for 2 min and fixed with 3% paraformaldehyde. Immunofluorescence staining monitored the transport of cell surface labeled VSVG and expression of XAPC7 or Rab7. Quantification of VSVG transport was undertaken by examining a minimum of 100 cells (N 0 ) coexpressing XAPC7 and VSVG and scoring the number (N) of cells showing VSVG transport to perinuclear late endosomes, identified as mannose 6-phosphate receptor positive. The ratio (N/N 0 ) reflects the percentage of cells displaying late endosomal transport of VSVG. The fractional VSVG protein transport to late endosomes was scored for samples transfected with a control vector (mock), wild-type Rab7 (Rab7WT) alone, or both Rab7WT and XAPC7 in the same experiment. The ratio for mock was assumed to be 100% for purposes of comparison with the other samples.
Morphological Assay of EGF Receptor Transport-BHK21 cells were transfected with plasmids encoding EGF receptor (EGFR), XAPC7, and/or Rab7WT. The culture medium was replaced with serum-free GMEM 18 h post-transfection. After 4 h of incubation, the cells were treated with 200 ng/ml of EGF tetramethylrhodamine conjugate (Molecular Probes) on ice for 10 min. Then the cells were washed twice with serum-free GMEM and left with complete GMEM in 5% CO 2 incubator for 30 min. Finally, the cells were washed with PBS, permeabilized with saponin, and fixed with 3% paraformaldehyde. The overexpression of XAPC7 was monitored by immunostaining and EGFR internalization and transport to late endosomes was evaluated by analysis of the EGF tetramethylrhodamine conjugate. Quantification was conducted as described above for VSVG protein transport.

RESULTS
Identification of the Proteasome ␣-Subunit XAPC7 as a Rab7binding Protein-Potential Rab7-binding partners were identified using a yeast two-hybrid assay as described (28). The dominant active Rab7Q67L was used as the bait to screen an MDCK library. 123 positive clones were initially selected based on their survival on selective medium (lacking tryptophan, histidine, and uracil) and strong ␤-galactosidase activity (28). Four unique clones were found to have homology to a human proteasome ␣-subunit called XAPC7 (PSMA7, HSPC, RC6 -1, and C6-I in mammals; see "Experimental Procedures" for results of BLAST analyses) (31). Clone C7 was 600-bp long and contained the 3Ј-portion of the coding sequence of the canine XAPC7 homologue, the 3Ј-untranslated region and the poly(A) tail. The shortest clone sufficient to reconstitute interaction with Rab7 encoded the C-terminal amino acids 165-248. Using C7 as the probe, the expression of XAPC7 in MDCK cells was confirmed by Northern blot analysis revealing a single 930 bp transcript (data not shown).
The interaction specificity of clone C7 with various Rab proteins, as well as wild-type and mutant forms of Rab7 was scored by growth on selective medium (Fig. 1A). Analogous results were obtained scoring for ␤-galactosidase activity on plates or in liquid culture (Fig. 1B and data not shown). C7 interacted strongly with all forms of Rab7, including wild-type Rab7, the dominant active Rab7Q67L mutant, non-nucleotidebound Rab7N125I, and GDP-bound Rab7T22N. No interaction was observed with wild-type endoplasmic reticular Rab2, or the endosomal Rab5 and Rab9 used as controls. Interaction of XAPC7 with Rab7 was confined to the N-terminal two-thirds of Rab7, as XAPC7 interacted with Rab7 lacking the unique C terminus (Rab7⌬175-207), but not with a construct bearing only the C terminus (Rab7⌬1-186) (Fig. 1, A and B). We also tested the interaction between C7 and Rab7 lacking only the last three C-terminal amino acids (Rab7⌬CSC). Such a deletion prevents post-translational prenylation at the C terminus (32) and consequently decreases the background (or false positives) in the two-hybrid assay (12). Deletion of the prenylation sequence did not abolish the interaction confirming the specificity of the interaction. However, the reaction with Rab7⌬CSC was somewhat weaker than with the full-length Rab7 constructs, suggesting that proper prenylation may be important for the interaction, perhaps by dictating Rab7 protein stability or folding. The above results indicate that the XAPC7 proteasome ␣-subunit specifically binds the N-terminal regions of Rab7 via a sequence in its C-terminal domain. The interaction appeared insensitive to the nucleotide bound state of Rab7, but exhibited some dependence on prenylation at the extreme C terminus of Rab7.
Rab7 Forms a Complex with the Proteasome ␣-Subunit XAPC7-The interactions between Rab7 and XAPC7 observed by yeast two-hybrid assay were further substantiated by immunoprecipitation experiments in mammalian cells. Since all the canine clones of XAPC7 isolated in our two-hybrid screen proved to be incomplete, a full-length human XAPC7 clone was obtained and used for the experiments detailed in this and all subsequent sections of the article. The wild-type or mutant forms of Rab7 were expressed plus or minus human XAPC7 in BHK21 cells. As detailed under "Experimental Procedures," cells were metabolically labeled with Trans 35 S, and cell lysates were immunoprecipitated with anti-Rab7 antibodies. Endogenous and overexpressed XAPC7 was specifically and proportionately coimmunoprecipitated with all forms of Rab7 tested ( Fig. 2A). The reciprocal immunoprecipitation experiment was also performed with our anti-XAPC7 antibody, yielding analogous results (Fig. 2B). Quantification of these data showed that there is a much stronger interaction between XAPC7 and all forms of Rab7 than with Rab5 (Fig. 2, C-D). The strongest interaction occurred between XAPC7 and the Rab7N125I mutant analogous to what we have observed for another effector of Rab7, namely the hVPS34/p150 complex (17). As an additional control for interaction specificity, two groups of BHK21 cells were transfected with XAPC7 and Rab7WT individually. Following cell lysis, samples were mixed and aliquoted into three groups: one sample was kept on ice and subjected to immunoprecipitation as usual while the other samples were first incubated at 37°C for 30 min or 90 min, respectively, prior to immunoprecipitation. Mere mixing of separate cell lysates containing overexpressed XAPC7 or Rab7 did not result in the coimmunoprecipitation of the two proteins (Fig. 2E). However, incubation at 37°C for increasing time periods allowed the formation of the complex in vitro. IP/Westerns were used to confirm the identity of the coprecipitated proteins. For this purpose unlabeled lysates were first immunoprecipitated with a rabbit polyclonal antibody directed against Rab7 and subsequently immunoblotted with a monoclonal antibody to XAPC7 (Fig. 2F). The cumulative results demonstrate that Rab7 specifically complexes with the proteasome ␣-subunit XAPC7.
FIG. 1. Yeast two-hybrid assay identifies a specific interaction between the proteasome ␣-subunit XAPC7 and Rab7. Yeast strain EGY48 was transformed with the canine XAPC7 C7 clone together with vectors encoding the indicated Rab constructs. A, transformants were assayed for growth on Ura Ϫ His Ϫ TRP Ϫ LEU Ϫ conditioned media. B, transformants plated on Ura Ϫ His Ϫ TRP Ϫ X-gal and scored for ␤-galactosidase activity by the presence of blue colonies. PEG202 and RFHM-1 are LexA fusion control plasmids. pSH17-4 is plasmid expressing LexA fused to GAL4 activation domain and was used as a positive control for reporter activation. Rab7⌬ denotes deleted residues, WT denotes wildtype, and standard mutant nomenclature is used.
The Proteasome ␣-Subunit XAPC7 Colocalizes with Rab7 on Endosomal Membranes-The interactions between Rab7 and XAPC7 observed by yeast two-hybrid assay and coimmunoprecipitation raised the question as to whether the interaction between the XAPC7 proteasome ␣-subunit and Rab7 occurred on endosomal membranes or was limited to the cytoplasm.
At first, we analyzed the distribution of XAPC7 in BHK cells. Crude membrane and cytosolic fractions were prepared by differential centrifugation from mock-transfected cells, cells overexpressing XAPC7, or untransfected cells. The distribution of XAPC7 between membrane and cytosolic fractions was monitored by Western blotting (Fig. 3, lanes 1-3, respectively). A significant fraction of the total endogenous or overexpressed XAPC7 was detected in the membrane pool. Actin was used as a control for protein loading and as a marker for cytosolic contamination of the membrane fraction. The recovery of actin in each of the three membrane or cytosolic samples was similar, indicating the sample handling was equivalent. Based on quantification of the immunoblot, the fractional recovery of actin in the membrane versus the cytosolic pools was 1:6. In contrast, the fractional recovery of XAPC7 in the membrane pool was significantly higher at 1:2, and was taken as an initial indication of its preferential enrichment in the membrane fraction. Several higher molecular weight bands were specifically detected only in the membrane fractions with our XAPC7 antibody and were increased in the sample where XAPC7 was overexpressed. These may represent a post-translationally modified form of XAPC7. There is precedence for post-translational modification of XAPC7, resulting in proteasome inactivation in response to oxidative stress (33). Therefore, this issue will be of interest for further investigation.
A more detailed analysis of the membrane localization of XAPC7 was achieved through microscopic evaluation and colocalization studies. For these experiments, BHK21 cells were cotransfected with XAPC7 and Rab7WT or Rab11WT and were permeabilized with low levels of saponin before fixation to allow optimal visualization of membrane structures and minimize cytosolic protein staining (17). Under these conditions XAPC7 colocalized with Rab7WT on small punctate structures in the perinuclear area, characteristic of late endosomal localization (Fig. 4, upper panels). The colocalization between Rab7 and XAPC7 on membranes was identical even when cells were cotransfected with the epitope-tagged G-XAPC7 (detected with a specific monoclonal antibody directed against the VSVG protein epitope) and a GFP-Rab7 fusion protein (data not shown). Additionally, we found XAPC7 colocalized with both dominant active and dominant negative mutant forms of Rab7, including Rab7Q67L, Rab7T22N and Rab7N125I (data not shown). In contrast, no colocalization was evident when XAPC7 was coexpressed with the early endosomal wild-type Rab5 or wild-type Rab11 proteins present on recycling endosomes (34) (Fig. 4, lower panels and data not shown). Additional immunoelectron microscopic analyses showed that overexpressed human XAPC7 (white arrows) colocalized with both endogenous or overexpressed wild-type Rab7 (black arrows) on the limiting membrane of multivesicular late endosomes (Fig. 5). Together the data demonstrate that although a sizeable fraction of the proteasomal XAPC7 protein is cytosolic, as might be expected, a significant fraction of endogenous or overexpressed XAPC7 could also be found specifically localized on late endosomal membranes together with Rab7. Importantly, XAPC7 was not present on other endosomal membranes or the Golgi complex.
Chimeric Rab protein constructs were used to further evaluate if binding to Rab7 was principally responsible for the specific late endosomal membrane localization of XAPC7. It is  3) were collected separately and used to isolated cell membrane and cytosolic fractions by differential centrifugation. Samples were resolved by SDS-PAGE and then subjected to immunoblotting to detect XAPC7 and actin as a control.
known that the unique membrane localization of each Rab protein is dictated in part by their hypervariable C termini and the localization of Rab proteins can often be altered by swapping C-terminal sequences (35,36). Since the XAPC7 binding sequences within Rab7 are confined to the N terminus, the endosomal targeting of XAPC7 should be altered by replacing the C termini of Rab7 to alter its membrane targeting if indeed the membrane targeting of XAPC7 is principally dependent on Rab7 binding. To test this postulate, a chimera Rab7-(1-174)Rab11-(160 -202) was constructed where the C-terminal hypervariable domain of Rab7-(175-207) was replaced with the counterpart of Rab11-(160 -202). The Rab7-(1-174)Rab11-(160 -202) chimera was also GFP-tagged at the N terminus to allow unambiguous localization. The Rab7-(1-174)Rab11-(160 -202) chimera exhibited a reticular perinuclear distribution analogous to GFP-Rab11 (28) and distinct from the normally more disperse and punctate wild-type-Rab7 distribution (Fig. 6, G-I and compare with Fig. 4). The Rab7-(1-174)Rab11-(160 -202) chimera colocalized extensively with transferrin receptor in the perinuclear area just as would be expected for a recycling endosome localization dictated by the hypervariable Rab11 C terminus (Fig. 6, A-C). When the Rab7-(1-174)Rab11-(160 -202) chimera was coexpressed with G-XAPC7, the distribution of the epitope-tagged XAPC7 was also shifted away from disperse Rab7-positive vesicles and was instead extensively colocalized with Rab7-(1-174)Rab11-(160 -202) (Fig. 6, B-F and compare with Fig. 4). This result offered independent confirmation of the yeast two-hybrid assay results that XAPC7 associated with Rab7 through its N-terminal (1-174) domain and established that XAPC7 is primarily recruited to late endosomal membranes through an interaction with Rab7.
XAPC7 and Rab7 Colocalize with Ubiquitinated Proteins-Ubiquitination is emerging as an important signal for endocytosis and late endocytic protein sorting (18). It was therefore of interest to check if XAPC7 might be involved in the regulation of endocytic ubiquitination. Initially, ubiquitin conjugates were immunolocalized in BHK cells transfected with XAPC7 or Rab7. Both Rab7 and XAPC7 were extensively colocalized with ubiquitinated cargo (Fig. 7A).
Previous work using proteasome inhibitors indicates that proteasome activity is required for proper endosomal sorting and lysosomal degradation of a number of receptors on the FIG. 4. XAPC7 is present on Rab7-positive late endosomes. BHK-21 cells were co-transfected with plasmids encoding epitopetagged G-XAPC7 and Rab7WT or GFP-Rab11WT. The cells were washed, fixed, and processed for immunostaining. Samples were stained for G-XAPC7 with mAb P5D4 directed against the epitope tag and detected with a rhodamine-labeled secondary antibody (red). Rab7 was detected with a specific rabbit polyclonal antibody and a fluorescein isothiocyanate-conjugated secondary antibody. Colocalization of XAPC7 with Rab7 is seen as yellow in the merged image. GFP-Rab11WT was detected directly after fixation (green). Bars, 20 m. Representative images from four independent experiments are shown. 5. Rab7 and XAPC7 are present together on the membranes of multivesicular late endosomes. A, cryoimmunoelectron micrographs of BHK-21 cell expressing XAPC7. Samples were processed for immunoelectron microscopy and stained for endogenous Rab7 with a specific chicken antibody detected with an appropriate gold-labeled secondary antibody (12-nm gold) and for overexpressed human XAPC7 with a specific rabbit antibody directed against the C terminus detected with an appropriate gold-labeled secondary antibody (6-nm gold). B-D, BHK-21 cells were cotransfected with wild-type canine Rab7 and epitope-tagged G-XAPC7. Rab7 was detected with a specific rabbit antibody directed against hypervarible C terminus detected with an appropriate gold-labeled secondary antibody (6-nm gold) and G-XAPC7 was detected with mAb P5D4 directed against the epitope tag and an appropriate gold-labeled secondary antibody (12-nm gold). White arrows highlight XAPC7 and black ones denote Rab7 in all panels. Bars (A-D), 100 nm. Representative images from three independent experiments are shown. FIG. 6. Chimeric Rab7-(1-174)Rab11-(160 -202) relocalizes XAPC7 to recycling endosomes. BHK-21 cells were transfected with plasmids encoding GFP-Rab7-(1-174)Rab11-(160 -202) (denoted Nrab7Crab11 on figure panels) and transferrin receptor (TfR) or epitope-tagged G-XAPC7. A-C, samples were stained for TfR (red) and GFP-Rab7-(1-174)Rab11-(160 -202) was visualized directly (green). Colocalization between GFP-Rab7-(1-174)Rab11-(160 -202) and TfR is seen as yellow in the merged image (C). D-F, samples were stained for G-XAPC7 (blue) and GFP-Rab7-(1-174)Rab11-(160 -202) was visualized directly (green). Colocalization of G-XAPC7 with GFP-Rab7-(1-174)Rab11-(160 -202) is seen as mint green in the merged image (F). G-I, samples were stained for endogenous Rab7 (red) and GFP-Rab7-(1-174)Rab11-(160 -202) was visualized directly (green). Bars, 20 m. Representative images from three independent experiments are shown. endocytic pathway (37,38,39,40). One suggested mechanism proposes that the inhibition of proteasome activity interferes with transport by perturbing the requisite protein mono-ubiquitination cycle due to decreased availability of free ubiquitin and increased steady state accumulation of mono-and multiubiquitinated proteins. Therefore, the possibility that the overexpression of XAPC7 might alter cellular proteasome activity and either enhance or diminish the metabolism of ubiquitinated proteins was considered. To examine this possibility, BHK21 cells were transfected with XAPC7, and the level of total ubiquitinated proteins in the cell lysates was compared with relevant controls by Western blotting. The results showed that the levels of ubiquitinated proteins in samples transfected with XAPC7 were similar to those in control samples without infection/transfection or transfected with vector alone (mock) (Fig. 7B, lanes 1-3). In contrast, addition of MG132, a proteasome inhibitor, resulted in increased levels of ubiquitinated protein in all three samples (Fig. 7B, lanes 4 -6). The levels of endogenous actin were also detected and used as control of the amount of sample applied on SDS-PAGE. Together, these results suggest that overexpression of XAPC7 does not grossly alter overall cellular proteasome function or the overall metabolism of ubiquitinated proteins.

FIG.
Overexpression of XAPC7 Interferes with the Transport of VSVG Protein from Early to Late Endosomes-Rab7 is a key regulator of endocytosis from early to late endosomes and proteasome inhibitors are known to interfere with receptor sorting on the endocytic pathway (38,39,17). Therefore, it was of interest to investigate if the XAPC7 might otherwise be involved in the control of late endocytic membrane transport. Both vesicular stomatitis virus (VSV) G protein and EGFR were used as endocytic tracers to monitor the impact of XAPC7 expression on their endocytic transport.
VSVG protein was selected as an endocytic marker because of the facility with which it can be transiently overexpressed together with other proteins of interest, and its demonstrated utility in monitoring the sequential transport between discrete endocytic compartments (10). The synchronous endocytosis of cell surface-expressed VSVG proteins was triggered by crosslinking with ectoplasmic domain-specific antibodies. Incubation at 15°C for 45 min allowed VSVG proteins to internalize and accumulate in small, disperse vesicles known to be early endosomes (8). Following transfer to 37°C, VSVG protein was rapidly delivered to large clustered, perinuclear late endosomes, resulting in significant colocalization with CI-M6PR within 15 min and reaching a maximum after 60 min (17). Transport to late endosomes was scored and quantified after 60 min at 37°C based on CI-M6PR colocalization (not shown) and a characteristic morphology (Fig. 8A). Relative to the mocktransfected samples, expression of wild-type Rab7 significantly enhanced late endosomal transport of VSVG (160%) (Fig. 8B). In contrast, expression of XAPC7 significantly impaired transport (28%), which could be partially rescued by coexpression of wild-type Rab7WT (60%). Analogous results were obtained by tracing EGF-stimulated endocytosis of EGFR (Fig. 8C). These somewhat surprising data raised the question as to whether or not Rab7 itself might be targeted for proteasomal degradation by XAPC7, however, experiments to test this possibility did not reveal any significant increase in Rab7 degradation upon overexpression of XAPC7 nor was there any stabilization of steady- FIG. 7. XAPC7 colocalizes with ubiquitinated proteins, but XAPC7 does not alter their metabolism. A, BHK-21 cells were transfected with plasmids encoding epitope tagged G-XAPC7 or wildtype GFP-Rab7. Both endogenous mono-and multi-ubiquitinated proteins were detected with monoclonal antibody FK2 (red), and G-XAPC7 was detected with mAb P5D4 directed against the epitope tag (green, upper panels). GFP-Rab7 was visualized directly (green, lower panels). Colocalization of ubiquitin conjugates with G-XAPC7 or GFP-Rab7 is seen as yellow in the merged image. Bars, 20 m. B, mock-infected/ transfected cells (lanes 2 and 5), cells infected/transfected with plasmid encoding XAPC7 (lanes 3 and 6) or uninfected BHK-21 cells (lanes 1 and  4). Samples were left untreated as controls (lanes 1-3) or were treated with 10 mM of the proteasome inhibitor MG132 (lanes 4 -6). The control and the proteasome inhibitor-treated samples were subjected to SDS-PAGE and Western blotting with rabbit anti-ubiquitin antibodies. Bars, 20 m.

FIG. 8.
Overexpression of XAPC7 interferes with late endocytic membrane transport of VSVG protein and EGFR. BHK-21 cells were infected/transfected with plasmids encoding (A and B) VSVG protein or (C) EGFR in combination with XAPC7 and/or wild-type Rab7. The cells were processed as described under "Experimental Procedures." Samples were stained for VSVG (red), XAPC7 (green). EGFR internalization was stimulated and monitored using rhodamine-labeled EGF. VSVG (B) and EGFR (C) transport was quantified from three independent experiments as detailed under "Experimental Procedures." The average percent transport and S.E. are plotted. Bars, 20 m. state Rab7 levels in the presence of proteasome inhibitors (data not shown). The cumulative data demonstrates the first direct molecular link between a bona fide transport machinery component, namely Rab7, and a known constituent of proteasomes, the XAPC7 ␣-subunit. Overexpressed XAPC7 appears to negatively regulate transport, while excess wild-type Rab7 activates transport.

DISCUSSION
Rab7 plays a key role in late endocytic transport. In this study, we identify XAPC7, a proteasome ␣-subunit also known as PSMA7 and HSPC in humans, as a novel Rab7-interacting protein. XAPC7 and Rab7 were initially shown to interact and form a complex using yeast two-hybrid assay. Coimmunoprecipitation studies confirmed the specificity of the interaction in mammalian cells and established that even endogenous XAPC7 could be found complexed to Rab7. The Rab7/XAPC7 complex was found specifically membrane localized on multivesicular late endosomes even in the absence of Rab7 overexpression. Overexpression of XAPC7 in BHK cells did not affect the metabolism of ubiquitinated proteins, but did interfere with the transport of membrane proteins from early to late endosome, suggesting a negative regulatory role. These data are novel and important in that they demonstrate a direct molecular link between the endocytic sorting and cytosolic degradative machineries and thereby offer a basis for the connections between proteasome activity and endocytic sorting described previously only at a functional level (40,41).
XAPC7 is one of at least seven different ␣-subunits that comprise the 26 S proteasome, a multisubunit protein degradation machine (42). XAPC7 is a highly conserved ␣-proteasome subunit expressed in numerous organisms ranging from mammals to plants. In human and mouse tissue, XAPC7 is a bona fide component of the 26 S proteasome (43). XAPC7 also has regulatory functions in hepatitis and HIV viral infection (44,45,46,47) and some disease states including cancer and glaucoma (48,49). The 26 S proteasome consists of two 19 S regulatory units, flanking a central 20 S degradative unit. The 20 S core is composed of seven ␣-subunits and seven ␤-subunits, where the 7 ␤-subunits are responsible for the specific proteolytic degradation of substrates (50). The 19 S units are largely responsible for recognizing and unfolding various ubiquitin tagged substrates. The unfolded peptides are then delivered into the 20 S central barrel and degraded (51). This ubiquitin-proteasome system has classically been recognized for its important regulatory functions in cellular protein turnover and the destruction of misfolded proteins. In the course of the last decade it has been recognized that the proteasome is involved in functions besides the degradation of cytosolic, multiubiquitinated proteins. It has been shown to be crucial for the degradation of misfolded ER membrane proteins (52), to be important in endocytic transport and sorting of receptors (40,41) and to recognize some substrates in a ubiquitin-independent manner (53,54). It is of interest to consider our findings in the context of these more recently ascribed proteasome functions.
There is growing evidence implicating proteasome activity in receptor sorting and endocytic membrane trafficking (40,41,39,55). Using proteasome inhibitors, MG132 and lactacystin, proteasome activity was shown to be required for the sorting of LRP (low-density lipoprotein receptor-related protein) from recycling to degradation pathway (41). Immunoelectron microscopy analyses demonstrated a proteasomal inhibitor-dependent reduction in LRP minireceptor within both limiting membrane and internal vesicles of late endosome. Yet, the initial internalization of LRP minireceptor did not require a functional ubiquitin-proteasome system (41). Other studies also implicated proteasome activity in sorting of the interleukin 2 receptor ␤-chain and EGFR to late endosomes (39,40). In a very recent and exciting study, proteasomes were shown to be recruited to phagosomes in a time-dependent manner (56). The time of proteasome recruitment to phagosomes was analogous to the time-course for Rab7 recruitment to phagosomes (57). Despite the central role of the proteasome in protein turnover in a large and diverse set of events, its subcellular localization, regulation, and interactions with regulatory complexes and nonproteasomal proteins remain poorly understood and under intensive study (58,59). Therefore, the link between an ␣-subunit of the 26 S proteasome and Rab7 as a component of the endocytic transport machinery is extremely exciting and suggests that the proteasome bearing XAPC7 is directly and specifically recruited to late endosomes. In keeping with a possible role in the control of receptor sorting, overexpression of XAPC7 was found to impair the trafficking of both VSVG protein and EGFR. The inhibitory effect of XAPC7 expression suggests endosome associated proteasome has a negative regulatory in the transport process. For example, binding might inactivate Rab7. The mechanism was shown not to involve increased Rab7 degradation, but could result from impairment of Rab7 interaction with its other known effectors including p150/ hVPS34 (17), RILP or rabring7. At first this might appear contradictory to the finding that active proteasome is required to promote endocytic transport, based on inhibitor studies. In this regard it is important to note that the inclusion of proteasome inhibitors has a very significant impact on overall metabolism and leads to an accumulation of ubiquitinated proteins. Therefore, such treatments most likely have pleiotropic effects. Multiple endocytic transport steps including receptor internalization that is dependent on mono-ubiquitination are likely affected and make it difficult to evaluate a specific effect on late endocytic transport. It is also conceivable that overexpression of XAPC7 leads to an excess of the free subunit, which could bind to Rab7 and interfere with recruitment of 26 S proteasome and/or block proteasome function. We consider these possibilities less likely for two reasons. First, even when Rab7 and XAPC7 were coexpressed, endocytic transport was still significantly impaired relative to control levels. Second, there was no excess accumulation of ubiquitinated protein substrates in cells expressing XAPC7 suggesting total proteasome activity was not affected.
It is also of interest to consider whether or not endosome associated XAPC7 in the context of the proteasome might participate in the recognition and processing of ubiquitinated cargo on membranes. Interestingly, XAPC7 and Rab7 were prevalent on endosomes bearing ubiquitinated cargo. Although there was no change in total cellular ubiquitinated protein levels with or without XAPC7 expression and mono-ubiquitinated receptors are not typically proteasome substrates (18), it will be of interest to assess how XAPC7 affects the endocytic ubiquitination/deubiquitination cycle in future studies, in particular if there is any interaction with Hrs or other ubiquitinated or ubiquitin interacting regulatory proteins.
XAPC7 and HC8 are related ␣-subunits (33% identity) and have both been shown to function in the non-ubiquitin dependent recognition and proteasome recruitment of several proteins with distinct consequences. The HC8 ␣-subunit of 20 S proteasome interacts with the C terminus of the Cdk inhibitor p21 Cip1 and mediates its proteasomal degradation (60). XAPC7 or PSMA7 interacts with both viral and cellular proteins. Liang and co-workers (45,44,61) were the first to report and map an interaction between XAPC7 and the Hepatitis B virus X (HBX) protein. Functionally, the interaction of HBX with XAPC7 was shown to be crucial for virus replication in that it blocks proteasome activity and stimulates transcriptional transactivation by HBX. Interestingly, HBX was also a substrate and could be degraded by proteasome (61). Subsequently, XAPC7 was shown to be important in the control of hepatitis C virus translation and downstream replication (47). The hypoxia-inducible factor-1␣ (HIF-1␣) was identified as the first cellular target of XAPC7 and is an important transcription factor in the cellular response to oxygen tension (62). Binding of HIF-1␣ to XAPC7 was shown to lead to ubiquitin-independent proteasomal degradation of the transcription factor and was implicated as an important regulatory mechanism in HIF-1␣-dependent transactivation functions. The interaction between Rab7 and XAPC7 was mapped to the N-terminal domain of Rab7 and the Cterminal half of XAPC7. The same C-terminal sequences of XAPC7 were also shown to be important for its interaction with HIF-1␣ and HBX establishing the importance of this domain in proteasomal recruitment of non-ubiquitinated proteins. The XAPC7-binding regions in the target proteins were in different regions and mapped to two domains in HIF-1␣. BLAST, GCG, AND PROSITE (63) comparisons between the XAPC7-binding domains in Rab7, HIF-1␣ and HBX failed to reveal any characteristic binding motifs indicating that if there is a common substrate recognition mechanism, it may involve a conformational epitope.
Understanding the interfaces between phosphatidyl inositol based membrane microdomains and ubiquitin-mediated sorting is crucial to unraveling the mechanisms underlying late endocytic transport and sorting of internalized receptors, as well processes that intersect with the endocytic pathways such as phagocytosis. The identification of a proteasome ␣-subunit as an interacting partner of Rab7 suggests that this late endocytic GTPase might be centrally poised to coordinate phospholipid signaling and membrane domain organization through its interactions with p150/hVPS34 (17) and at the same time provide a critical link to the ubiquitination and cytosolic degradative machinery via an association with 26 S proteasome.