Selective binding and uptake of ribonuclease A and glyceraldehyde-3-phosphate dehydrogenase by isolated rat liver lysosomes.

Ribonuclease A (RNase A) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) are selectively taken up and degraded by isolated rat liver lysosomes by very similar processes. The uptake and degradation of both of these proteins are stimulated by the heat shock cognate protein of 73 kDa and ATP/Mg2+. Both binding and uptake of RNase A and GAPDH by lysosomes are saturable, and uptake of RNase A and GAPDH requires a protease-sensitive component within the lysosomal membrane. GAPDH competes for binding and uptake of RNase A by lysosomes and vice versa while another protein, ovalbumin, does not compete. RNase S-peptide (amino acids 1-20 of RNase A) also competes for RNase A binding and uptake by lysosomes, while RNase S-protein (amino acids 21-124 of RNase A) does not compete. The uptake of RNase A by lysosomes appears to involve an intermediate step in which approximately 2 kDa of the polypeptide's COOH terminus remains outside lysosomes while the remainder is inside the lysosomal lumen.


of RNase A by lysosomes appears to involve an intermediate step in which approximately 2 kDa of the polypeptide's COOH terminus remains outside lysosomes while the remainder is inside the lysosomal lumen.
Both nonlysosomal and lysosomal pathways of proteolysis operate in eukaryotic cells (13). Lysosomes are especially important in the degradation of long-lived proteins, and there are a variety of pathways of delivery of intracellular proteins to lysosomes for subsequent digestion. These pathways include endocytosis (4, 51, crinophagy (61, microautophagy (7-91, macr o a u t o~~a g y (10-14), and direct protein transport mediated by the heat shock cognate protein o f 73 kDa (hsc73)* (1,15).
Endocytosis is responsible for delivery of exogenous polypeptides and certain plasma membrane proteins to the lysosomes for degradation (4, 5). Crinophagy diverts secreted proteins to lysosomes for degradation when requirements for their secretion are reduced (6). Microautophagy and macroautophagy appear to internalize different cytosolic proteins at similar rates * This work was supported in part by Direccion General de Investigacion Cientifica y Tecnica del Ministerio de Educacidn y Ciencia Grants pTR91-001 and PB91-0830 (to E. K.) and by Grant AGO6116 from the National Institutes of Health (to J. E D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ll Current address: Dept. of Biology, University of California at San // To whom correspondence should be addressed.
M i c~a u t o p h a~ has been reproduced in isolated lysosomes (71, but macroautophagy has not. The selective uptake and degradation of ribonuclease A (RNase A) and ribonuclease Speptide (RNase S-peptide; residues 1-20 of RNase A) has been achieved using isolated lysosomes from human fibroblasts (18-20). and the selective uptake and degradation of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) by isolated rat liver lysosomes have been reported (21). Here we show selective uptake and degradation of RNase Aand GAPDH by isolated rat liver lysosomes and demonstrate that these processes are mechanistically similar. Furthermore, a transport intermediate in the uptake o f RNase A by lysosomes is identified.

MATERIALS AND METHODS
Animals-Male Wistar rats weighing 200-250 g were fasted for 20 h prior to use.
hsc73-This constitutive member of the heat shock protein 70-kDa (hsp70) family was purified from a bovine brain extract by ATP-agarose affinity chromatography as described previously (22).
Antibodies-The primary antibodies used were a rabbit IgG against bovine pancreatic RNase A (Rockland Inc., Gilbertsville, PA) and a mouse IgG against GAPDH (Dr. Amelia Martinez-Ramh, Instituto de Investigaciones Citologicas, Valencia, Spain). Secondary antibodies for Western blots were goat IgGs raised against rabbit or mouse IgG and conjugated to alkaline phosphatase (Promega, Madison, WI). ~s o~a~i o n of Lysosomes and M~t~~ondr~-Lysosomes were isolated from a light mi~chondrial fraction in a discontinuous metrizamide density gradient by the shorter method reported previously 1211. Rat liver mitochondria were prepared as described (23).
Standard lncubatiun Conditions to Study the Uptake of RNase A and GAPDH into Lysosomes-Lysosomes (100 pg of protein) were incubated in a final volume of 30 pl with RNase A (25 pg for Western blots or 50 pg for Coomassie Brilliant Blue R-250 staining) for 20 min at 37 "C in 0.3 M sucrose and 10 m M MOPS buffer (pH 7.2). The integrity of the lysosomal membranes, as judged by the latency of the lysosomal enzymes ~-hexofiaminidase, measured as N-acetyl-~-glucosaminidase described previously (20,211, was >93% at the end of all incubations. Additional experimental details can be found in the table and figure legends. bated in a final volume of 125 pl with protein substrates. The sub-Proteolysis Measurements-Lysosomes (25 pg of protein) were incustrates used were GAPDH radiolabeled by reductive methylation (21) with ['*Clformaldehyde (Amersham International, Bucks, UK) to a specific radioactivity of 1.2 x lo6 dpdnmol), t3H1RNase S-peptide radiolabeled by reductive methylation with NaB'H, (Dupont NEN (24)) to a specific radioactivity of 1.1 x lo7 dpdnmol, and t3Hlleucine-labeled soluble proteins from L-132 human cells (5,000 d p d p g of protein). These soluble proteins were labeled and prepared as described previously (25).
Incubations were stopped by the addition of trichloroacetic acid to a final concentration of 10% or, in the RNase S-peptide experiments, phosphotungstic acid in HCl to final concentrations of 3.25 and 5% of Proteins by Isolated Rat Liver Lysosomes   A B C respectively. The acid-soluble material was collected by filtration in the Multiscreen Assay System (Millipore Corp., Bedford, M A ) using a 0.45-pm pore membrane. Proteolysis was expressed as the percentage of the initial acid-insoluble radioactivity converted to acid-soluble radioactivity. The energy-regenerating system used in some experiments was as described previously (18).
Protein Sequencing-Determination of the amino acid sequence of the low molecular weight RNase A-derived band was carried out after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transfer to Immobilon-S in an automated protein sequencer (Applied Biosystems, Foster City, CA).
General Methods-SDS-PAGE (17% gels for studies of RNase A and 12% gels for studies of GAPDH) and immunoblotting procedures were carried out as described previously (21). Quantitation of the immunoblots was performed with an LKB Ultroscan laser densitometer (LKB-Pharmacia, Uppsala, Sweden) with a Hewlett-Packard (Palo Alto, CA) 3396 Series I1 integrator. The linearity of the method was established using different amounts of RNase A and GAF'DH as standards in the immunoblotting assays. Statistical analyses were performed by the Student's t test. Best straight lines were calculated by linear regression, and best curved lines and K,,, values were calculated using Enzfitter software (Elsevier-Biosoft, Cambridge, UK).

RESULTS
Uptake of RNase A by Rat Liver Lysosomes-RNase A was incubated with freshly isolated rat liver lysosomes under conditions similar to those described to study the selective uptake of GAPDH (21). They included (a) incubation for 20 min at 37 "C with or without chymostatin to inhibit lysosomal proteolysis, ( b ) treatment or not with proteinase K to remove the protein outside of lysosomes, and (c) collection of lysosomes and analysis of pellets and supernatants by SDS-PAGE followed by Coomassie staining and/or immunoblotting using anti-RNase A As was the case for GAPDH (21), a portion of RNase A was associated with the lysosomal pellets, and part of this RNase A was resistant to proteinase K digestion (not shown). The RNase A that was found associated with lysosomes in the absence of protease inhibitors represented primarily (95%) RNase A bound to the external face of the lysosomal membrane since it was susceptible to digestion by proteinase K. In chymostatintreated lysosomes, the RNase A that was associated with lysosomes represented both surface-bound and internalized molecules since only part of it was susceptible to digestion by

IgGs.
proteinase K the RNase A remaining after proteinase K treatment represents RNase A inside lysosomes. Several other experiments similar to those carried out with GAPDH (21) but using RNase A as a substrate yielded equivalent results (not shown): (a) chymostatin and NH,Cl were the most effective protease inhibitors, ( b ) binding and uptake were time-and temperature-dependent with a 10-min lag in uptake compared with binding, and (c) some binding (one-third that seen for lysosomes) but no import could be detected using isolated mitochondria under similar incubation conditions.
We tested the effect of hsc73 and an energy-regenerating system on the proteolysis of RNase S-peptide (which is degraded by the same pathway as RNase A but is a better substrate (20)). Both the energy-regenerating system and hsc73 were stimulatory, and both components added together were most effective ( Table I). A similar pattern of stimulation was evident for GAPDH except that the amount of stimulation was smaller, and the stimulation by addition of both components together was not statistically different from either one added alone (Table I). The stimulation of proteolysis by ATP was also evident using a mixture of metabolically labeled soluble pro- teins from human cells ( Table I). In this case the lack of an effect of added hsc73 alone can be explained because the soluble proteins contain abundant hsc73. However, in the presence of an energy-regenerating system, added hsc73 also stimulated degradation of these soluble proteins. ATP hydrolysis appeared to be required for most of the stimulation in proteolysis since €-ATP, a nonhydrolyzable ATP analog, was much less effective (not shown). Saturable Binding and Uptake of RNase A and GAPDH by Rat Liver Lysosomes-When increasing amounts of RNase A (Fig. 1) or GAPDH (Fig. 2) were incubated with the same amount of lysosomes, the binding (A) and uptake ( B ) of both proteins were saturable. The Kd values calculated for RNase A and GAPDH binding were 95 and 10 PM, respectively. The K,,, values for uptake of RNase A and GAPDH were 4 and 18 PM, respectively.
It is interesting to note that after proteinase K treatment of lysosomes, in addition to the intact RNase A band, a lower molecular weight band is consistently evident (Fig. lC, right  panel). This immunoreactive species migrates at the position of RNase S-protein suggesting the removal of approximately 2 kDa from RNase A. This result is further addressed below. A lower molecular weight fragment of GAPDH is also occasionally but inconsistently seen following proteinase K treatment (Fig. 2C, lane 14). The uptake of RNase A (Fig. 3 A ) and GAPDH (Fig. 3B) was reduced by trypsin in a dose-dependent manner. The integrity of the lysosomal membrane was not affected by treatment even at 250 pg/ml based on latency measurements for lysosomal enzymes as described under "Materials and Methods." In addition, compared with untreated lysosomes, the trypsin-treated lysosomes did not release detectable proteolytic activity, showed normal morphology by electron microscopy, and exhibited unaltered SDS-PAGE protein banding patterns (not shown). Incubation with trypsin and a 2-fold molar excess of soybean trypsin inhibitor did not affect uptake of RNase A or GAPDH, excluding a competition by trypsin for the binding or uptake machinery. Finally, similar results were obtained when lysosomes were pretreated with proteinase K or elastase instead of trypsin (not shown).
As expected, increasing amounts of RNase A also decreased the binding (Fig. 5A, lanes 2-6) and uptake (Fig. 5C, lanes 2-6) of GAPDH while ovalbumin had no effect (Fig. 5, A and C, lanes   7-10). In this case the Ki values, 122-152 p~, are in the same range as the Kd for RNase A binding but are higher than the K,,, for RNase A uptake.
We also tested the effects of increasing amounts of RNase S-protein (Fig. 6 , A and B, lanes 2-6) and RNase S-peptide (Fig.   6, C and D, lanes 2-6) on the binding (Fig. 6, A , C, and E) and uptake (Fig. 6, B, D, and F) of RNase A. RNase S-protein had no effect on binding or uptake of RNase A. In contrast, increasing amounts of RNase S-peptide reduced the binding (Ki = 205 p~) and uptake (Ki = 160 p~) of RNase A by lysosomes.

An Intermediate in the Dunsport of RNase A into Rat Liver
Lysosomes-We examined further the possibility that the band smaller than RNase A produced by proteinase K treatment (Fig. lC, lanes 9-14; Fig. 3 A , lanes 2-6; Fig. 4C, lanes 2-9; Fig.  6, B and D, lanes 2-6) was a lysosomal import intermediate. This intermediate form represents 16 2 8% of total RNase A taken up by lysosomes after a 20-min incubation at 37 "C and is found only associated with lysosomes and not in the supernatant fractions.
The possibility that this smaller form is generated as an intermediate in the intralysosomal degradation of RNase A is unlikely since this form is never seen unless the lysosomes are treated with proteinase K. Also, such an intermediate is never detected when RNase A is incubated with broken lysosomes or when RNase A is incubated with proteinase K without lysosomes (not shown).
We compared the RNase A inside lysosomes immediately after incubation or after a second 5-or 10-min incubation in the absence of RNase A to allow complete entry of the RNase A. Fig.  7A shows Coomassie-stained gels while Fig. 7B shows two separate immunoblots. The clipped RNase A evident immediately after incubation was present a t reduced levels after a 5-min chase and was absent when 10 min were allowed for import. A likely interpretation of this result is that a kinetic block exists in the import of RNase A such that approximately 2 kDa of the protein remains susceptible to proteolysis when the rest of the molecule is within lysosomes. Amino-terminal sequence analysis of the lower molecular weight band produced by the proteinase K treatment revealed the mature RNase A sequence, KETAA. Therefore, the clipped portion of RNase A must be at the carboxyl terminus.
A model for RNase A import into lysosomes consistent with the results presented in this paper is shown in Fig. 7C. RNase A interacts with hsc73 in the cytosol (step 1) and then binds to a receptor or a polypeptide transporter on the lysosomal membrane (step 2). Most of the RNase A is then rapidly imported into the lysosome interior, but approximately 2 kDa of the COOH-terminal portion of the protein remains outside the lysosome for a short time (step 3). Finally, the entire molecule is transported into the lysosome (step 4 ) where it is rapidly degraded (step 5 ) unless lysosomal proteases are inhibited. DISCUSSION There are many similarities in the lysosomal uptake of GAPDH (211, RNase A (19,20), and RNase S-peptide (19,20) including their stimulation by hsc73 and ATP/Mg2* (Table I). Lysosomal uptake of RNase A and GAPDH is inhibited by prior proteolytic treatment of the lysosomes (Fig. 31, and binding of RNase S-peptide to fibroblast lysosomes is also inhibited by prior trypsin treatment (20). These results suggest the involve-  (lanes 4, 6, and 8) or not (lanes 3 , 5 , and 7) with proteinase K, ment of a proteinaceous receptor or transporter. More direct evidence for common elements in the lysosomal uptake mechanisms of these polypeptides includes competition for binding and uptake of RNase A by GAPDH (Fig. 41, competition for binding and uptake of GAPDH by RNase A (Fig. 5), and competition for binding and uptake of RNase A by RNase S-peptide (Fig. 6).
Quantitation of the binding and uptake of RNase A and GAPDH by rat liver lysosomes (Figs. 1 and 2) indicates that GAPDH binds more avidly than does RNase A (Kd values = 10 and 95 p~, respectively). However, uptake of RNase A by lysosomes saturates more readily than uptake of GAPDH (K, values = 4 and 18 w, respectively). Therefore, GAPDH binding and uptake by lysosomes occur in the same concentration range, but an excess of binding sites compared with uptake sites exists for RNase A. The initial RNase A molecules that bind to lysosomes must somehow be very efficiently coupled to the uptake process.
Competition for binding and uptake of GAPDH by RNase A (Fig. 5) is in the 100 1.1~ range consistent with its competition being through binding to the lysosomes. The lack of competition for GAPDH uptake by RNase A in the 4 1 " range suggests that the uptake machinery is different for the two proteins. Alternatively, the uptake machinery may be the same, but the efficiency of RNase A uptake may be decreased in the presence of GAPDH.
The Scatchard analyses ( Figs. 1 and 2, insets) indicate maximal binding per 100 pg of lysosomal protein of 408 pmol of RNase A or 63 pmol of GAPDH. Our unpublished results indicate that rat liver lysosomes have an average volume of 0.042 pm3 consistent with reports of others (26). Using the same criteria that have been used to calculate mitochondrial binding and import sites (27)(28)(29), we calculate that each lysosome has approximately 1,700 binding sites for GAPDH and 10,000 for RNase A. These values are in the range of receptor numbers for the import of mitochondrial proteins (27)(28)(29).
One of the most striking apparent differences between the rat liver lysosomes used in these studies and the fibroblast lysosomes used previously (18-20) is the affinity of binding for substrate proteins. The Kd value for binding of RNase S-peptide for fibroblast lysosomes was found to be 100 n~ (20) while binding of GAPDH and RNase A to rat liver lysosomes shows a Kd value of 10-95 1.1~. This difference between fibroblast and rat liver lysosomes may be due to the different experimental procedures used to measure binding and uptake in these different studies. For example, binding of RNase S-peptide to fibroblast lysosomes was carried out at 0 "C (20) while the binding to rat liver lysosomes reported here was carried out at 37 "C. Another consideration for this apparent difference is that the liver lysosomes may not be optimally active when derived from rats fasted for 20 h. We find that fasting for 96 h yields lysosomes that are 2-4 times more active: but whether this greater activity corresponds to altered K,,, or Kd values remains to be determined.
The selective uptake of RNase A and RNase S-peptide by lysosomes appears to require a peptide region biochemically related to KFERQ (15,19). Consistent with this idea, binding and uptake of RNase A can be competed with RNase S-peptide that contains the KFERQ sequence and not by RNase S-protein that does not contain a KFERQ-like region (Fig. 6). In addition, ovalbumin does not contain a KFERQ-like peptide region and does not compete for binding or uptake of RNase A (Fig. 4) or GAPDH (Fig. 5). GAPDH does not contain an exact KFERQ motif as previously defined (15), but it does contain a closely related sequence beginning at amino acid 258 in the human, hamster, and mouse sequence KWKQ. The three-dimensional structure of GAPDH shows that this sequence is in a solventexposed a-helical region of the tetramer (30). Whether this sequence is actually required for selective lysosomal uptake remains to be determined. Alternatively, KFERQ-like sequences might be formed by the tertiary structure of GAPDH.
The mechanisms by which RNase A and GAPDH are selectively taken up and degraded by lysosomes are not likely to be through microautophagy or macroautophagy since these proc-esses do not show substrate specificity (14, 16) and are not known to be stimulated by ATP and hsc73. The lysosomal uptake process appears to be very similar to pathways of uptake of precursor proteins into organelles such as the mitochondrion and endoplasmic reticulum. These similarities include stimulation by ATPMgZ' and a cytosolic hsp7O and the requirement for an hsp70 within the organelle to pull precursor proteins across the organelle membrane3 (31)(32)(33)(34)(35). In addition, the import process appears to occur in stages; we report an import intermediate with approximately 2 kDa of the COOH-terminal portion of RNase A still exposed outside the lysosome (Fig. 7). Approximately 2 kDa from the COOH terminus, there are three consecutive branched-chain amino acids, IN, that may stall the uptake process. In any event, intermediates have also been reported for the import of proteins into the mitochondrion and the endoplasmic reticulum (31)(32)(33)(34)(35). If these similarities extend further, there may be receptors for substrate proteins and/or hsc73 on the lysosome membrane as well as polypeptide translocation channels through the membrane as there are for proteins translocated into mitochondria and endoplasmic reticulum.