A molecular chaperone complex at the lysosomal membrane is required for protein translocation

Many intracellular proteins must be translocated across a membrane before they reach their proper intracellular location. Mechanisms of protein secretion across the plasma membrane of bacteria (Nishiyama et al., 1999; Bernstein, 2000) as well as across organelle membranes in eukaryotic cells (Brodsky, 1998; Teter and Klionsky, 1999; Bauer et al., 2000; Herrmann and Neupert, 2000) have been the subjects of many recent reviews.

Certain proteins must also be translocated across lysosomal membranes for their degradation. Proteins can be taken up and degraded by lysosomes following several different pathways (Dunn, 1994; Dice, 2000). One of these lysosomal proteolytic pathways is activated in confluent monolayers of cultured cells by withdrawal of serum growth factors (Auteri et al., 1983; Backer et al., 1983), in certain tissues from intact animals during prolonged starvation (Wing et al., 1991; Cuervo et al., 1995), and in yeast in response to nitrogen deprivation (Horst et al., 1999). In this pathway of proteolysis particular cytosolic proteins are transported into lysosomes by mechanisms similar to protein import for residence in organelles such as mitochondria, the endoplasmic reticulum and chloroplasts (Brodsky, 1998; Bauer et al., 2000; Herrmann and Neupert, 2000). For example, transport of protein substrates into lysosomes requires ATP/MgCl₂ (Chiang et al., 1989), the cytosolic constitutively expressed protein of the heat shock family of 70 kDa (Hsc70) (Chiang et al., 1989), and Hsc70 in the lysosomal lumen (lyHsc70) (Agarraberes et al., 1997; Cuervo et al., 1997). Substrate proteins contain a pentapeptide motif related to KFERQ (Dice, 1990), and a receptor for substrate proteins resides in the lysosomal membrane and has been identified as the lysosomal-associated membrane protein type 2a (lamp2a) (Cuervo and Dice, 1996). Increasing the levels of lamp2a in the lysosomal membrane increases the activity of the entire proteolytic pathway (Cuervo and Dice, 1996; Cuervo and Dice, 2000), suggesting that lamp2a levels can be a rate-limiting step for chaperone-mediated autophagy.

Several different substrates of this proteolytic pathway bind to Hsc70 and also to lamp2a (Cuervo and Dice, 1996; Dice, 2000). One substrate, RNase S-peptide, which consists of amino acids 1-20 of bovine pancreatic ribonuclease A (RNase A) (Backer et al., 1983), binds to Hsc70 (Terlecky et al., 1992) but not to lamp2a (A. M. Cuervo and J. F. Dice, unpublished). Nevertheless, RNase S-peptide specifically binds to the lysosomal membrane and blocks the binding and uptake of other protein substrates (Terlecky and Dice, 1993; Cuervo et al., 1994). These results suggest that components of the protein binding and import machinery other than lamp2a are also critical for operation of this proteolytic pathway.

The interactions of cytosolic Hsc70 and protein substrates are mediated by cycles of ATP binding and hydrolysis, with the ADP-bound form of Hsc70 having high affinity for protein substrates (Hightower and Leung, 1997). Several molecular chaperones interact with cytosolic Hsc70 and modulate its ATPase activity and the stability of the Hsc70-polypeptide substrate complex (Frydman and Hohfeld, 1997). The heat-inducible protein of the heat shock family of 70 kDa (Hsp70)-interacting protein (Hip) and the heat shock protein of 40 kDa (Hsp40), a DnaJ homologue, act as chaperone enhancers. Hip stimulates the assembly of Hsc70 with Hsp40 and the polypeptide substrate (Hohfeld et al., 1995), whereas Hsp40 stimulates the ATPase activity of Hsc70 (Suh et al., 1999) leading to increased rates of binding and release of polypeptides. The heat shock protein of 90 kDa (Hsp90) recognizes flexible regions of proteins and prevents the proteins from aggregating (Buchner, 1999). Hsp90-Hsp70 organizing protein (Hop) binds to both Hsp90 and Hsc70 and acts as an adapter between the two molecular chaperones (Pratt and Toft, 1997). Hop may also function as a nucleotide exchanger (Gross and Hessefort, 1996). The Bcl2-associated athanogene 1 protein (BAG-1) uncouples the binding of the protein substrate from the ATPase activity of Hsc70 (Bimston et al., 1998) and may act as a postive (Terada and Mori, 2000) or a negative (Luders et al., 2000; Nollen et al., 2000) regulator of Hsc70, perhaps depending upon the BAG-1 isoform (Luders et al., 2000).

A recent study shows that protein substrates must be unfolded to be translocated across the lysosomal membrane during chaperone-mediated autophagy (Salvador et al., 2000). This unfolding is not required for binding of substrates to the lysosomal membrane, so unfolding must occur at the lysosomal surface (Salvador et al., 2000). We now show that a multi-molecular chaperone complex exists at the lysosomal membrane and is required for the translocation of substrate proteins.

IMR-90 diploid fetal human lung fibroblasts (Corriel Cell Repositories, Camden, NJ) were cultured as described (Neff et al., 1981). Cell fractionation, purification of cytosol and lysosomes (Storrie and Madden, 1990), and isolation of lysosomal membranes and matrix (Ohsumi et al., 1983) was carried out as previously described. Culture and treatment of cells for immunolocalization were performed as described previously (Agarraberes et al., 1997). Lysosomal membranes were resuspended twice in 0.5 M NaCl, unless otherwise stated, and spun down at 130,000 g for 15 minutes.

Purified Hsp90 and Hsp40 proteins were purchased from Stressgen (Victoria, BC, Canada). Hsc70 was purified from bovine brain cytosol as described (Welch and Feramisco, 1985). The monoclonal antibody (mAb) 13D3 against Hsc70 was a generous gift of Joseph Chandler (Maine Biotechnology Services Inc., Portland, ME). The mAbs against Hsp70 and Hsp90 and rabbit polyclonal antibodies against Hsp40 and against the γ subunit of the chaperonin containing t-complex polypeptide 1 (CCTγ) were purchased from Stressgen. The mAb against Hip used for immunoprecipitation was a gift of David Smith (Mayo Clinic, Scottsdale, AZ), and the anti-Hip antibodies used for western blot analysis and transport blocking assays were purchased from Affinity Bioreagents Inc. (Golden, CO). The human antibody against DNA was obtained from the Center for Disease Control (Atlanta, GA). The mAb against BAG-1 was a gift of Shinichi Takayama and John C. Reed (The Burham Institute, La Jolla, CA). The rabbit antibody against Hop was a gift of Michael Lassle (Massachusetts Institute of Technology, Cambridge, MA). The mAb against the p19 protein from the avian myoblastosis virus (AMV-3C2) was obtained from the Developmental Studies Hybridoma Bank (Iowa City, IA). The antibody against lamp2a was generated as described (Cuervo and Dice, 1996). Antibodies against RNase A were from Rockland Inc. (Gilbertsville, PA), and those against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were from Amelia Martinez-Ramon, Instituto de Investigaciones Citologicas, Valencia, Spain. Rabbit serum was purchased from Sigma Chemical Co. (St Louis, MO). Rhodamine-labeled goat anti-rat IgG was purchased from Chemicon (Temecula, CA). All other fluorescence- and horseradish peroxidase-labeled secondary antibodies were purchased from Jackson ImmunoResearch Laboratories Inc. (West Grove, PA). Protein determinations were performed using the Lowry assay (Lowry et al., 1951) or the BioRad protein assay reagent (BioRad Laboratories, Hercules, CA).

One dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) was performed as described (Terlecky and Dice, 1993). Two-dimensional electrophoresis was carried out as described (O’Farrell, 1975) with the following modification: samples were boiled for 3 minutes in 0.5% SDS, 10 mM DTT, 10 mM Tris-HCl buffer, pH 6.8, before incubation in isoelectric focusing solubilization buffer to inactivate lysosomal proteases. Isoelectric focusing was carried out as previously described (Agarraberes et al., 1997). For two-dimensional analysis lysosomal membranes were purified as described above, except that they were not washed in 0.5 M NaCl.

Proteins were electrotransferred onto Immobilon-P membranes (Millipore Corp., Bedford, MA). Immunoblotting was carried out as previously described (Agarraberes et al., 1997). Antibodies and dilutions were as follows: mouse mAb 13D3 (1:5000), mouse anti-Hsp70 (1:1000), rabbit anti-lamp2a (1:5000), rat anti-Hsp90 (1:5000), rabbit anti-Hsp40 (1:5000), rabbit anti-Hop (1: 3000), mouse anti-Hip (1:5000), mouse anti-BAG-1 (1:300), mouse anti-CCTγ (1:1000), rabbit anti-GAPDH (1:2000), and rabbit anti-RNase A (1:2000)

The ratio among Hsc70, Hsp90 and Hsp40 was determined by western blot analysis of serial dilutions of the purified proteins and different concentrations of purified lysosomal membranes. Western blots were developed with chemiluminescense methods (Renaissacence^©, NEN-Life Science Products, Boston, MA), and the signals were quantified by densitometry. The lowest dilution of each lysosomal membrane protein presenting a chemiluminescense signal that closely matched the value of one of the dilutions of its corresponding purified protein was used for calculation.

Beads of concanavalin A (ConA) linked to sepharose (Sigma Chemical Co.) were washed three times in PBS (Terlecky et al., 1992). The mAbs 13D3, anti-Hip and anti-p19 were incubated with ConA sepharose beads for 2 hours at 25°C in PBS, washed three times in PBS, and resuspended in immunoprecipitation (IP) buffer containing 150 mM NaCl, 20 mM Tris-HCl, pH 8.2, 1% Nonidet P-40. Lysosomal membranes were solubilized in IP buffer and centrifuged at 130,000 g for 30 minutes. Supernatants containing solubilized lysosomal membrane proteins were incubated with ConA sepharose-antibody for 2 hours at 4°C, washed twice with IP buffer, and washed once with PBS. SDS-PAGE solubilization buffer (Agarraberes et al., 1997) was added, and samples were processed for electrophoresis.

For indirect immunofluorescence studies, cells were handled as previously described (Agarraberes et al., 1997). Briefly, cells were fixed in cold methanol (−20°C) for 1 minute, subjected to triple immunostaining (Agarraberes et al., 1997), and viewed by confocal microscopy (Odyssey XL, Noran Instruments, Middleton, WI). Primary antibodies were used as 1:25 dilutions, and fluorescent-labeled secondary antibodies were used in dilutions according to the manufacturer’s recommendations. Omission of primary antibodies resulted in no visible fluorescence (data not shown).

Samples were treated as previously described (Terlecky and Dice, 1993). Briefly, purified lysosomes (100 μg protein) were incubated in a final volume of 0.1 ml in an ice bath for 10 minutes in 100 μM chymostatin A (Sigma Chemical Co.) to inhibit digestion of translocated substrate proteins (Cuervo et al., 1994). Samples were then incubated for 20 minutes with no serum or increasing amounts of the following serums: mouse serum containing anti-Hip, nonimmune mouse serum, rabbit serum containing anti-Hsp40, rabbit serum containing anti-Hop, and nonimmune rabbit serum. Samples were incubated for 15 minutes at 37°C in an ATP-regenerating system (Terlecky and Dice, 1993) with either RNase A (Sigma Chemical Co.), GAPDH (Boehringer Mannheim, Indianapolis, IN), or no substrate. Finally, lysosomes were treated with 10 μg of Proteinase K and 1 μM CaCl₂ for 20 minutes in an ice bath. The reaction was stopped with 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride (Sigma Chemical Co.). Lysis buffer was added, and samples were processed for SDS-PAGE. All reagents were prepared in 0.25 M ultra pure grade sucrose (Sigma Chemical Co.), 5 mM Tris-HCl, pH 7.3.

Previous studies using two-dimensional SDS-PAGE have shown that cytosolic Hsc70 consists of a major isoform at pI 5.5 and a minor isoform at pI 5.3 (Agarraberes et al., 1997). Hsc70 was also present in lysosomes with the majority residing in the lysosomal lumen (Terlecky and Dice, 1993). The Hsc70 associated with the lysosomal membrane (lymHsc70) represents a small fraction (<10%) of total Hsc70 present in lysosomes (Fig. 1a). In order to determine the lymHsc70 topology, we treated purified intact lysosomes with trypsin (Fig. 1a). After protease treatment lysosomal membranes were purified as described in Materials and Methods, and samples were resolved by two-dimensional SDS-PAGE. Western blot analysis using the specific anti-Hsc70 antibody, 13D3, showed that the more acidic isoform of Hsc70 (pI 5.3) is lost after protease treatment, indicating that lymHsc70 is on the cytosolic side of the lysosomal membrane. LyHsc70 (pI 5.5) remains completely protected from degradation (Fig. 1a, middle). In addition, similar analysis of purified lysosomal membranes shows that lyHsc70 (pI 5.5) is entirely in the lysosomal lumen, since we were unable to detect any of this isoform associated with membranes (Fig. 1a, bottom).

We next analyzed the type of interaction by which lymHsc70 associated with the lysosomal membrane. Purified lysosomal membranes were treated with either 0.5 M NaCl (Fig. 1b, lane 1), 0.1M Na₂CO₃, pH 11.0 (Fig. 1b, lane 2), or 5.0 M urea (Fig. 1a, lane 3). Western blot analysis of these samples indicated that urea is the only treatment that released lymHsc70 from the lysosomal membrane (Fig. 1b). Lamp2a, a lysosomal transmembrane protein, was not affected by any of the treatments (Fig. 1b). We explored the possibility that Hsc70 interacted directly with lipids in the lysosomal membrane. The reagent [¹²⁵I]TID labels molecules exposed to hydrophobic environments (Otto and Smith, 1996). In extensive studies lamp2a was labeled by this reagent, but lymHsc70 was not labeled indicating that lymHsc70 does not interact directly with lipids in the membrane bilayer (data not shown). Taken together, these results indicate that lymHsc70 is a peripheral protein tightly associated with other proteins in the lysosomal membrane. Although resistance to alkaline extraction commonly signifies an integral membrane protein (YaDeau and Blobel, 1989; Wiemer et al., 1996; Kermorgant et al., 1997), a growing number of peripheral proteins are known to be resistant to alkaline extraction (Breyton et al., 1994; Kourtz and Ko, 1997; Lin et al., 1998). Specifically, a chloroplast outer membrane hsp70, Com70, is resistant to NaCl and Na₂CO₃ washes even though it is a peripheral membrane protein (Kourtz and Ko, 1997).

We analyzed the interaction of lymHsc70 with a protein substrate of this proteolytic pathway, RNase A (Backer et al., 1983; Dice, 2000), using a protease protection assay. Purified lysosomes were incubated either in the presence or absence of ADP/MgCl₂ and/or RNase A at 37°C to favor protein binding to lymHsc70 and transport initiation (Fig. 1c). Lysosomes were treated with proteinase K and membranes were purified and analyzed by SDS-PAGE and then immunoblotted with the Hsc70-specific antibody, 13D3. These results indicate that the protein substrate in the presence of ADP/MgCl₂ has a protective effect on Hsc70 and that a conformational change in Hsc70 might take place during binding and transport of the protein substrate.

We next explored whether or not lymHsc70 might be in a complex with other chaperones as has been described for cytosolic Hsc70 (Hohfeld et al., 1995; Gross and Hessefort, 1996; Frydman and Hohfeld, 1997; Demand et al., 1998; Stuart et al., 1998; Gross et al., 1999; King et al., 1999). Confluent cultures of cells were serum-deprived for 18 hours. Purified lysosomal membranes were detergent-solubilized and subjected to immunoprecipitation as described in Materials and Methods. Solubilized membranes were incubated with either anti-Hsc70, anti-Hip, or an unrelated antibody to the p19 protein from the avian myoblastosis virus. Western blot analysis of the immunoprecipitates was performed with specific antibodies against Hsp90, Hsc70, Hop, Hip, BAG-1, Hsp40, Hsp70 and CCTγ (Kubota et al., 1994). Fig. 2 shows the presence on the lysosomal membrane of multiple molecular chaperones and cochaperones known to interact with cytosolic Hsc70 (Hohfeld et al., 1995; Gross and Hessefort, 1996; Frydman and Hohfeld, 1997; Demand et al., 1998; Stuart et al., 1998; Gross et al., 1999; King et al., 1999). Interestingly, anti-Hip antibodies also immunoprecipitate the same set of chaperones/cochaperones, including BAG-1 (Fig. 2). Hip and BAG-1 have overlapping binding sites on the N-terminal domain of Hsc70 and cannot bind to the same molecule of Hsc70 (Demand et al., 1998). Therefore, multiple complexes of lymHsc70 must be present at the lysosomal membrane. Neither Hsp70 nor CCTγ was present in the immunoprecipitates even though they are readily detectable in cytosol (Fig. 2) and, to a more variable extent, associated with solubilized lysosomal membranes (data not shown). In addition, immunoprecipitation with an unrelated antibody to p19 failed to immunoprecipitate any of the proteins analyzed (Fig. 2). Finally, we also identified GAPDH, a natural substrate of the pathway, as part of this large complex of chaperones (data not shown). This result indicates that at least some of the chaperone complexes are likely to be functional.

We established the stoichiometry of Hsc70, Hsp40 and Hsp90 at the lysosomal membrane using defined amounts of purified proteins as described in Materials and Methods to be 15:10:1. The low amount of Hsp90 may indicate its presence in only a small fraction of the molecular chaperone complexes. Alternatively, Hsp90 may be lost from the chaperone complexes during the purification and manipulation of lysosomes perhaps due to its weaker interactions in the complex.

These results were in marked contrast to our findings in the lysosomal lumen, where only lyHsc70 was present; although lyHsc70 could be readily detected by immunoblotting, Hsp90, Hop, Hip, BAG-1 and Hsp40 were absent from the lysosomal lumen (data not shown). Hsc70 is known to enter the lysosomal lumen (Agarraberes et al., 1997; Cuervo et al., 1997). Whether the other proteins failed to enter the lysosomes with Hsc70 or were rapidly degraded within the lysosomes after entry is not known.

We fixed serum-deprived IMR-90 human fibroblasts with methanol at −20°C. to remove cytosolic proteins. Previous studies showed that Hsc70 colocalized with the majority (>90%) of vesicular structures containing lysosome-associated membrane protein 1 (Agarraberes et al., 1997), a lysosomal marker (Green et al., 1987). We examined the lysosomal localization of the chaperones and cochaperones by confocal microscopy. Triple immunostaining of fibroblasts was performed using Hsc70 as a lysosomal marker, anti-human DNA antibodies as a nuclear marker, and specific antibodies against the other proteins as shown in Fig. 3. Specific antibodies against the molecular chaperones and cochaperones Hsp90, Hop, BAG-1, Hsp40 and Hip (red signal) largely colocalized with Hsc70 (green signal). Colocalization in the merged images registers as orange/yellow. Quantitation of the merged images after magnification indicated the following degrees of colocalization of Hsc70 and the other chaperones: Hsp90 (97%), Hop (88%), BAG-1 (97%), Hsp40 (95%) and Hip (98%). These results confirm our conclusion that a complex of molecular chaperones exists at the lysosomal membrane.

We next determined whether or not activation of the proteolytic pathway modified the levels of the different proteins found in the complex with lymHsc70 (Fig. 4a). We quantified the levels of Hip, Hsp90, Hsc70, Hop, Hsp40 and BAG-1 in the lysosomal membranes from serum-fed (+S) and serum-starved (−S) cells. In three separate experiments we observed an increase in Hsp90, while Hip, Hsc70, Hop, Hsp40 and BAG-1 did not change significantly (Fig. 4a, top). These results suggest that the components of the molecular chaperone complex may be somewhat dynamic. Similarly, we determined the levels of these proteins in the cytosol from both serum-fed and serum-starved cells. Their levels did not significantly change in response to serum withdrawal (Fig. 4a, bottom).

We next asked whether or not the chaperones and cochaperones might be associated with the lysosomal membrane because they were substrates for this lysosomal proteolytic pathway. Purified lysosomes were incubated in an ATP-regenerating system such that protein substrates trapped in transport during cell disruption would probably complete translocation. Fig. 4b shows that there is no significant change in the levels of the chaperones and cochaperones analyzed, but the level of GAPDH decreased during the time course of the assay.

We directly assessed the requirement for Hip, Hop and Hsp40 in the process of binding and transport of substrate proteins using antibodies against these proteins. Previous studies showed that antibodies to the cytosolic tail of lamp2a inhibited binding and uptake of substrate proteins (Cuervo and Dice, 1996). In addition, incubation of isolated lysosomes with the antibody to Hsc70 blocked binding and uptake of substrate proteins (A. M. Cuervo and J. F. Dice, unpublished). Purified lysosomes were preincubated with increasing amounts of each specific antibody, and either RNase A or GAPDH was used as substrate. Fig. 5 shows the concentration-dependent inhibitory effect of serum containing antibodies against Hip, Hop and Hsp40 on substrate transport in a cell-free system. Control assays using appropriate nonimmune serum showed no effect on transport of protein substrates. The lower panels in Fig. 5 show the densitometric quantification of these results. These results indicate that Hip, Hop and Hsp40 are required for import of substrate proteins into lysosomes. The mAb against Hsc70 also inhibited translocation of RNase A (data not shown). The antibodies to Hsp90 and BAG-1 did not efficiently immunoprecipitate the native proteins, so they could not be used in this type of analysis. The requirement for multiple molecular chaperones in this lysosomal proteolytic pathway justifies the term ‘chaperone-mediated autophagy’ to distinguish this pathway from microautophagy and macroautophagy.

Considering these and previous results, our current working model of mechanisms of chaperone-mediated autophagy is shown in Fig. 6. Activation of the lysosomal proteolytic pathway results in cytosolic Hsc70 and its associated chaperones binding to substrate proteins at least in part at their KFERQ-like motif. This complex might then bind to the lysosomal membrane or the substrate protein might be passed to a second chaperone complex already docked at the lysosomal membrane. We believe that multiple Hsc70 molecules assemble together with various chaperones and cochaperones because both Hip and BAG-1 are present in immunoprecipitates, and they bind to the same location on Hsc70 (Demand et al., 1998). Our previous studies demonstrated that Hsc70, a protein substrate, and lamp2a can be isolated from solubilized lysosomal membranes in a ternary complex (Cuervo and Dice, 1996). Also, recent evidence shows that lamp2a forms homomultimers (Cuervo and Dice, 2000). We speculate that at the lysosomal membrane the multi-chaperone complex acts, at least in part, to unfold protein substrates because unfolding is required for transport of substrate proteins into the lysosomal lumen but not for binding of substrate proteins to the lysosomal membranec (Salvador et al., 2000). The protein substrate may then bind to lamp2a (Cuervo and Dice, 1996) and subsequently be transported across the lysosomal membrane. This transport requires the activity of the lyHsc70 in the lysosomal lumen (Agarraberes et al., 1997; Cuervo et al., 1997). Once the protein enters the lysosomal lumen, it is rapidly degraded by the battery of lysosomal proteases.

The different chaperones and cochaperones at the lysosomal membrane most likely regulate lymHsc70. Hsp90, Hip and Hop probably serve to stabilize the lymHsc70-substrate complex on the surface of the lysosome (Hohfeld et al., 1995; Frydman and Hohfeld, 1997; Buchner, 1999; Suh et al., 1999). This stabilization may be required to allow the complete unfolding of substrate proteins prior to initiation of import. BAG-1 may be present in certain complexes to uncouple the ATPase and protein binding activities. BAG-1 may thereby also contribute to the stabilization of lymHsc70-substrate interactions. Hsp40 should enhance the Hsc70 ATPase activity and promote both the release and rebinding of substrate proteins. In addition, Hsp40 promotes the multimerization of Hsc70 (King et al., 1999). If several molecules of Hsc70 are present in each complex of chaperones and cochaperones, they could interact with substrate proteins in a cooperative manner. The precise roles of the molecular chaperones and cochaperones for protein translocation across lysosomal membranes is likely to require an experimental system even more defined than purified lysosomes or lysosomal membranes.

Whether or not all lymHsc70 is in such large chaperone complexes remains to be determined. Our immunoprecipitation assays contained solubilized lysosomal membrane proteins in excess, so lack of immunoprecipitation did not necessarily reflect lack of a particular protein in the complex. Our limited stoichiometry results suggest that lymHsc70 is associated with substoichiometric levels of Hsp40 and Hsp90. More complete studies with cytosolic Hsc70 suggest that Hsp40 and Hop are present at one-tenth the amount of Hsc70 (Kosano et al., 1998), whereas Hip (Hohfeld et al., 1995) and BAG-1 (Stuart et al., 1998; Nollen et al., 2000) are approximately equimolar with Hsc70. The amount of Hsp90 in the molecular chaperone complex is highly variable depending on the tissue type and physiological status (Schumacher et al., 1994; Kimmins and MacRae, 2000). Whether or not additional regulators of cytosolic Hsc70 activity such as the C-terminus of Hsc70-interacting protein (Ballinger et al., 1999), Scythe (Kaye et al., 2000; Thress et al., 2001) or Reaper (Thress et al., 2001) are in the lymHsc70 complex remain to be determined

The requirement for rapid and dramatic changes in cellular protein composition during starvation has resulted in a mechanism for identifying protein substrates (the KFERQ motif) and a process of targeting a large amount of proteins to lysosomes for degradation. The involvement of cytosolic Hsc70, lymHsc70 and lamp2a may represent a triple checking mechanism for accurate identification of substrates. These cytosolic proteins are degraded in a membrane-confined compartment rich in hydrolases of rather low specificity. With this system a large number of proteins can be degraded to single amino acids. These amino acids can then be used for synthesis of new proteins or as an energy source.

We thank our colleagues for their generosity with antibodies as detailed in Materials and Methods. Special thanks to Michael Lassle for many valuable discussions. This work was supported by grant AG06116 from the National Institutes of Health.