Mutation at the Processing Site of Chicken Low Density Lipoprotein Receptor-related Protein Impairs Efficient Endoplasmic Reticulum Exit, but Proteolytic Cleavage Is Not Essential for Its Endocytic Functions*

The low density lipoprotein receptor-related protein (LRP) is synthesized as a proreceptor that undergoes post-translational proteolytic processing, yielding a noncovalently associated αβ dimer as the mature LRP. We tested the role of processing by creating a mutant in which the P1 residue (Arg3942) of the consensus site for furin cleavage (Arg-Asn-Arg-Arg3942↓) was replaced with Ser in chicken LRP. Transfection of the mutant LRP (designated LRP-RS) into a Chinese hamster ovary cell line lacking endogenous LRP resulted in expression of the unprocessed full-length proreceptor. Comparison of cell lines stably expressing either the wild-type LRP (LRP-wt) or the unprocessed LRP-RS showed that at comparable expression levels, both receptors restored the sensitivity of cellular protein synthesis toPseudomonas exotoxin A (IC50 = 25 ng/ml). Subcellular fractionation and neuraminidase treatment showed that both LRP forms were transported to the plasma membrane. In addition, LRP-RS exhibited kinetics of binding, endocytosis, and degradation of methylamine-activated α2-macroglobulin that were identical to those of LRP-wt. The internalization rate constant was similar for LRP-wt (K e = 0.259 min−1) and mutant LRP-RS (K e = 0.252 min−1), suggesting that it takes about 4 min for the entire surface LRP pool to be internalized. Sorting of LRP from the endosomal compartment to lysosomes or recycling to the plasma membrane were also unaltered in mutant LRP-RS. Pulse-chase analysis showed that the lack of processing of LRP had no effect on the stability of its post-endoplasmic reticulum form or on the rate of its intracellular transit from the endoplasmic reticulum to the Golgi apparatus. However, the exit of mutant LRP from the endoplasmic reticulum was retarded by the Arg3942-to-Ser substitution, as evidenced by prolonged retention within the endoplasmic reticulum (t½ = 4 h for LRP-wt and t½ > 13 h for LRP-RS).

The low density lipoprotein (LDL) 1 receptor-related protein (LRP), a member of the LDL receptor gene family, adopts the topology of a type I integral membrane protein (1). A major structural feature of this family of receptor proteins is the arrangement of clusters of ligand-binding repeats (designated class A repeats) interspersed between clusters of epidermal growth factor precursor repeats within the extracellular domains (2). Variations in the makeup of the modular extracellular domains appear to dictate the ligand binding characteristics of the various family members, which in the case of LRP shows multifunctionality for a wide range of proteins, including ␣ 2 -macroglobulin (␣ 2 M) and Pseudomonas exotoxin A (PEA). However, all of the members of this receptor family share a capacity to bind a 39-kDa receptor-associated protein, or RAP (3). Furthermore, the cytoplasmic domain of LRP contains two copies of the NPXY motif that is important for the clathrincoated pit-mediated internalization of the LDL receptor (4).
LRP is translated as a single polypeptide (600 kDa) consisting of 4525 amino acids in humans (5) and 4522 amino acids in chickens (6). A distinct structural feature that sets LRP apart from the other LDL receptor family members is that LRP undergoes post-translational proteolytic processing (7). Processing occurs in the trans-Golgi compartment 1-2 h after LRP translation (7) and is presumably catalyzed by the endopeptidase furin (8), which recognizes the RXRR consensus sequence, RHRR in human LRP (5) and RNRR in chicken LRP (6). This post-translational processing results in the formation of mature LRP as a noncovalently associated heterodimer, consisting of the extracellular 515-kDa ␣-chain and the transmembrane 85-kDa ␤-chain (7). Proteolytic processing by furin also occurs in other membrane proteins (e.g. the insulin receptor), and sometimes a crucial function for processing becomes evident. For instance, a mutation at the consensus sequence of the human insulin receptor (i.e. RNRR-to-RNRS substitution) that results in defective processing leads to impaired receptor function and severe insulin resistance (9,10).
The current knowledge of the biological significance of posttranslational LRP processing is limited. In several studies, expression of membrane-anchored LRP minireceptors, which contained a portion of the extracellular domain plus the transmembrane and intracellular sequences but lacked the process-ing site, has yielded functional receptors. Willnow et al. (11) have expressed such a LRP minireceptor that bound and degraded the 39-kDa RAP and the tissue-type plasminogen activator-plasminogen activator inhibitor-1 complex. Recently, we have expressed two membrane-anchored chicken LRP minireceptors and found that both of the unprocessed receptors restored the toxicity of PEA and degradation of ␣ 2 M in transfected LRP-null cells (12). However, the role of posttranslational processing in the function of the full-length LRP has not been analyzed. In the current study, we created a full-length chicken LRP processing mutant in which Arg 3942 within the furin cleavage site (RNRR 3942 2) (6) was replaced by Ser and examined the requirement for processing in the endocytic functions and intracellular trafficking of LRP.

EXPERIMENTAL PROCEDURES
Materials-DNA restriction and modification enzymes, endoglycosidase H (Endo H), and peptide:N-glycosidase F were purchased from New England Biolabs. Neuraminidase was purchased from Boehringer Mannheim. ProMix TM (a mix of [ 35 S]methionine and [ 35 S]cysteine; 1000 Ci/mmol), carrier-free Na 125 I, horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G antibody, and the enhanced chemiluminescence (ECL) reagents for immunoblotting were obtained from Amersham Pharmacia Biotech. Bicinchoninic acid protein assay reagent and EZ-Link TM NHS-SS-Biotin (sulfosuccinimidyl-2-(biotinamido)ethyl-1,3-dithiopropionate) were obtained from Pierce. Bacterially expressed recombinant PEA was prepared as described (13). The glutathione S-transferase/RAP expression plasmid containing the human RAP cDNA was obtained from D. Strickland (American Red Cross, Rockville, MD). The purification of glutathione S-transferase/RAP and subsequent thrombin cleavage of the fusion protein were carried out as described previously (14).
Preparation of Mutant LRP Expression Plasmid-The parental expression plasmid pcLRP-wt that encoded the full-length chicken LRP (originally designated pcLRP-100) was constructed in the pCMV5 vector using cDNA fragments as described previously (12). A fragment that extended from nucleotide 9112 (an XhoI site) of the LRP cDNA to another XhoI site (in the pCMV5 vector) was subcloned into pBluescript SK ϩ (Stratagene) to create pLRP3Ј for mutagenesis purposes. To abolish cLRP proteolytic processing, the cDNA sequence encoding the furin cleavage site, RNRR 3942 (6), was altered to encode RNRS by polymerase chain reaction-based mutagenesis. Two mutagenic primers, cla1 (5Ј-GAGAACGTGCGCATCGATGCC-3Ј) and sac2rs (5Ј-ACCCCGCGGCA-TCTTCAGCCCGGAGATGTTCAGGTGGGTGACGCCGCCGTCGATC-TGCGATCGGTTGCGG-3Ј), were used to amplify the ClaI-SacII (nucleotides 12969 -13148 of the chicken LRP cDNA) fragment that encompassed the RNRR 3942 coding sequence. The sac2rs primer changed the encoded RNRR into RNRS and also introduced a PvuI site (underlined). Amplification of the ClaI-SacII fragment was achieved by polymerase chain reaction using Vent R ™ polymerase (New England Biolabs). The mutant ClaI-SacII polymerase chain reaction product was inserted into pLRP3Ј that had been digested with ClaI and SacII, and finally a XhoI-XhoI fragment of the mutant pLRP3Ј was inserted into pcLRP-wt that had been digested with XhoI, to create the fulllength mutant plasmid pcLRP-RS (Fig. 1A). The nucleotide sequences of the resulting mutant constructs were verified by double-stranded DNA sequencing with T7 DNA polymerase (Amersham Pharmacia Biotech) according to the manufacturer's instructions.
Cell Culture and Transfection-Wild type CHO-K1 and the LRP-null CHO cell line (15) were cultured in Ham's F-12 medium containing 10% fetal bovine serum as described previously (12). Stable transfection into LRP-null CHO cells (in 100-mm dishes) was achieved by cotransfection of 15 g of pcLRP-wt or pcLRP-RS with 0.15 g of pSV2neo (12). Stable transformants were selected using 700 M G418 and maintained with 500 M G418.
Northern Blot Analysis-Total RNA was isolated from confluent cells (100-mm dish) using the TRI TM reagent (Molecular Research Center) and resolved (20 g) by electrophoresis on 1% agarose gel (6% formaldehyde). The samples were transferred onto Hybond-N membrane (Amersham Pharmacia Biotech) and probed with a 32 P-labeled ClaI-ClaI fragment (nucleotides 2136 -3039 of the chicken LRP cDNA).
Fractionation of Subcellular Membranes and Preparation of Membrane Extract-Transfected CHO cells were homogenized using a ball bearing homogenizer, and microsomal and plasma membrane fractions were prepared by sequential ultracentrifugation (12). The Triton X-100-soluble membrane extract was prepared either from total cell lysate or from fractionated membranes as described previously (12).
Neuraminidase, Endoglycosidase H, or Endoglycosidase F Digestion of LRP Proteins-Glycosylation status of LRP was analyzed by sensitivity to Endo H or neuraminidase using fractionated microsomes or plasma membranes or using whole cell lysates as described previously (12). In pulse-chase experiments, confluent cells (35-mm dishes) were labeled with ProMix TM [ 35 S]methionine and [ 35 S]cysteine (200 Ci/ml), and radiolabeled LRP was immunoprecipitated from whole cell lysates and subjected to Endo H or peptide:N-glycosidase F digestion according to the manufacturer's instructions. The samples were resolved on gradient polyacrylamide gels (3-8%) containing 0.1% SDS (SDS-PAGE), and LRP was visualized by fluorography.
Immunoblot Analysis-LRP was resolved by SDS-PAGE and transferred electrophoretically to polyvinylidene difluoride membranes at 125 V for 6 h under cooling to Ϫ10°C. The LRP ␤-chain and the mutant LRP-RS were detected by immunoblot analysis using an antiserum raised against the C-terminal 17 amino acids of the ␤-chain (designated anti-␤-chain). The LRP ␣-chain was detected using an antiserum against the C-terminal 15 amino acids of the chicken LRP ␣-chain (designated the anti-␣-chain). Immune complexes were visualized by enhanced chemiluminescence.
Endocytosis of 125 I-Labeled Methylamine-activated ␣ 2 M ( 125 I-␣ 2 M*)associated LRP-Confluent cells (35-mm dishes) were incubated with 1 nM 125 I-␣ 2 M* (on ice) for 4 h. After washing with ice-cold phosphatebuffered saline (PBS) containing 10 mg/ml bovine serum albumin, cells were incubated in PBS at 37°C for 0 -10 min. At the indicated times, the dishes were placed on ice, and PBS was removed and replaced with ice-cold F-12 containing pronase (0.3%, w/v). After a 1-h incubation on ice, cells were collected by centrifugation (14,000 ϫ g, 5 min) in a microcentrifuge. Radioactivity associated with the supernatant (pronase-released cell surface 125 I-␣ 2 M*) or cell pellet (internalized 125 I-␣ 2 M*) was quantified. Greater than 90% of the initial radioactivity was recovered during the 10-min experiment.
The endocytosis rate constant (K e ) was determined by In/Sur analysis as described by Wiley and Cunningham (16). Cells were incubated with 125 I-␣ 2 M* (2 nM) at 37°C (in 1 ml of F-12 containing CaCl 2 (5 mM) and bovine serum albumin (10 mg/ml)). At indicated times, the dishes were placed on ice, and cells were extensively washed (three times with PBS/bovine serum albumin and twice with PBS). Cells were treated with pronase (0.3% (w/v) in ice-cold F-12) for 1 h, and the supernatant (surface 125 I-␣ 2 M* (Sur)) and cell pellets (internalized 125 I-␣ 2 M* (In)) were separated by centrifugation as described above. The endocytosis rate K e was calculated as ⌬(In/Sur)/⌬t (the slope of In/Sur versus time).
Internalization and Recycling of Biotinylated LRP-Internalization studies of biotinylated LRP were performed according to a previously published method (17). Cells (60-mm dishes) were placed on ice and washed twice with ice-cold Krebs-Ringer solution (10 mM sodium phosphate, pH 7.4, 128 mM NaCl, 1.25 mM CaCl 2 , 1.25 mM MgSO 4 ). Cell surface proteins were labeled with 0.5 mg/ml of EZ-Link TM NHS-SS-Biotin (Pierce) at 0°C for 1 h. After labeling, cells were incubated with 1 ml of Krebs-Ringer solution (37°C) for the indicated times. (One dish was kept on ice as the untreated control.) After incubation, cells were treated with or without glutathione solution (15 mg/ml in 1 ml of ice cold Krebs-Ringer solution containing 10% FBS) for 30 min. After the removal of glutathione solution, cells were washed once with ice-cold PBS containing iodoacetamide (5 mg/ml) and twice with PBS alone. Cells were collected into a lysis buffer (Tris-maleate, pH 6.0, Triton X-100 (1.4%, v/v), and phenylmethylsulfonyl fluoride (0.015%, w/v)) in microcentrifuge tubes and mixed for 2 h at 4°C. The cell lysates were pretreated with protein A-bacterial adsorbent (ICN Canada) prior to immunoprecipitation of LRP with polyclonal antibodies raised against affinity-purified human LRP (see below). The immune complexes were absorbed onto protein A-Sepharose (Amersham Pharmacia Biotech) followed by washing with 1.4% Triton X-100 and dissolved in sample buffer (10 mM Tris-glycine, pH 8.3, 8 M urea, 2% (w/v) SDS, 10% (v/v) glycerol). Bound proteins were resolved by SDS-PAGE (3-8% gels) and transferred to polyvinylidene difluoride membranes. Biotinylated LRP was visualized using horseradish peroxidase-conjugated streptavidin (Amersham Pharmacia Biotech) according to the manufacturer's instructions.
For recycling studies, cells were biotinylated at 37°C for 2 h and treated with glutathione on ice (see above) to remove the biotin moiety of surface-located LRP while retaining the biotinylated LRP located within the endosomal compartments. After washing, the cells were incubated at 37°C to allow recycling of endosomal LRP back to the cell surface. At the indicated times, the cells were placed on ice and treated with glutathione to monitor the returning of biotinylated LRP to plasma membrane. Cells were lysed, and LRPs were immunoprecipitated and analyzed by a biotin-streptavidin assay as described above. One set of cells were lysed without the glutathione treatment and subjected to biotin-streptavidin analysis to monitor the recovery of total biotinylated endosomal LRP.
Preparation of Polyclonal Antibodies against LRP-Human LRP was affinity-purified from mammary adipose tissue. The tissue was homogenized with extraction buffer (1% octyl-␤-D-glucopyranoside in 50 mM Tris-HCl, pH 7.4, containing 150 mM NaCl, 2 mM CaCl 2 , 5 mM leupeptin, and 0.1 mg/ml phenylmethylsulfonyl fluoride) (18), and the homogenate supernatant was applied to a glutathione S-transferase/RAPglutathione Sepharose column (19). The column was washed with extraction buffer, and the bound LRP was eluted using 50 mM Tris-HCl, 150 mM NaCl, 20 mM EDTA, and 1% octyl-␤-D-glucopyranoside, pH 3.6. The purified LRP was used to immunize New Zealand White rabbits to raise antiserum according to standard procedures (20). The resulting antiserum reacts with both human and chicken LRP.
Other Methods-The PEA toxicity assay was performed using stably transfected cells as described previously (12). Proteins were determined by the bicinchoninic acid method (Pierce) according to the manufacturer's instructions. Purification, activation, iodination, and uptake/ degradation of ␣ 2 M* were performed according to procedures reported previously (12).

Generation and Expression of Processing Mutant LRP-RS-
The mutant LRP-RS was created by changing the furin cleavage site RNRR 3942 to RNRS (Fig. 1A), simulating a naturally occurring mutation found in the insulin receptor (10). Nucleotide sequences encoding the mutated furin cleavage site were verified by DNA sequencing (Fig. 1B). Stably transfected cells expressing either LRP-wt or LRP-RS were generated using LRP-null CHO cells (15). Respective stable cell lines were selected that exhibited similar level of LRP expression as judged by immunoblot analysis of total cell lysates (Fig. 1C) and by Northern blotting and metabolic labeling of LRP (data not shown). As expected, expression of LRP-wt led to detection of the processed 515-kDa ␣-chain and 85-kDa ␤-chain that reacted with respective anti-␣- (Fig. 1C, left) and anti-␤-chain (Fig. 1C, right) antibodies. In contrast, the expression of LRP-RS yielded the unprocessed 600-kDa full-length proreceptor that reacted with both anti-␣-and anti-␤-chain antibodies (Fig. 1C, right and left). A small amount of unprocessed LRP-wt (600-kDa) that reacted with anti-␤-chain antibody was detect-able in LRP-wt transfected cells (Fig. 1C, right). The slight difference in the molecular mass between LRP-RS and the full-length LRP-wt (Fig. 1C, right) was attributable to different glycosylation status between these two proteins (see below). The lack of any secondary cleavage of LRP indicated that no other proteolytic cleavage sites were introduced into LRP by mutagenesis. These results demonstrate that replacing Arg 3942 with Ser effectively abolishes furin-mediated post-translational endoproteolysis of LRP.
Cell Surface Presentation of LRP and Restored PEA Sensitivity-Plasma membrane was isolated from transfected cells and treated with neuraminidase. Like ␣and ␤-chains of LRPwt, the unprocessed LRP-RS proreceptor was sensitive to neuraminidase treatment ( Fig. 2A). Thus, abolishing posttranslational processing does not alter trafficking of LRP from the ER, through the trans-Golgi compartment, to the plasma membrane.
Expression of LRP-RS proreceptor also effectively restores the cell's sensitivity to PEA (Fig. 2B). Similar to previous observations (21), the concentration of PEA required to inhibit protein synthesis to 50% of untreated cells (IC 50 ) was ϳ25 ng/ml in normal CHO-K1 cells and Ͼ500 ng/ml in the LRP-null cells. Cells transfected with either LRP-wt or LRP-RS exhibited an IC 50 value of ϳ25 ng/ml, comparable with that of normal CHO-K1 cells. Hence, the entrance of PEA via the LRP-mediated pathway is not compromised by the lack of receptor processing.
Endocytosis of ␣ 2 M in LRP-transfected Cells-Expression of either LRP-wt or LRP-RS in LRP-null cells also restored the ability of the cells to bind and degrade ␣ 2 M*. The RAP-inhibitable binding of 125 I-␣ 2 M* at 4°C was similar between LRPwt-and LRP-RS-transfected cell lines at three concentrations (i.e. 0.5, 1.0, and 2.5 nM) of the ligand (Fig. 3A). Binding of 125 I-␣ 2 M* to the transfected cells was about 2.5-fold greater than that to normal CHO-K1 cells (Fig. 3A). Degradation of 125 I-␣ 2 M* in cells expressing the unprocessed LRP-RS proreceptor was as efficient as that in LRP-wt-expressing cells (Fig.  3B). These results suggest that the LRP-mediated endocytosis and lysosomal degradation of ␣ 2 M* are independent of LRP processing. The Endocytic Rate of LRP-The endocytic rate of surfacebound ␣ 2 M* was determined using two approaches. In the first approach, cells were incubated with 125 I-␣ 2 M* at 0°C and then switched to 37°C. Internalization of surface-bound 125 I-␣ 2 M* was monitored by measuring the acquisition of resistance of 125 I-␣ 2 M* to exogenous pronase (Fig. 4A). At the beginning of temperature switch (time zero), ϳ70% of the total 125 I-␣ 2 M* was surface-bound (i.e. 30% pronase-resistant). After incubation at 37°C for 5 min, 80% of total 125 I-␣ 2 M* became pronaseresistant (Fig. 4A). There was no difference between LRP-wt and LRP-RS in gaining pronase resistance, suggesting that efficient internalization of 125 I-␣ 2 M* can be achieved through both receptors.
In the second approach, the endocytic rate of surface-bound 125 I-␣ 2 M* was measured using In/Sur analysis (16). The ratio of internalized 125 I-␣ 2 M* (pronase-resistant) to surface-bound 125 I-␣ 2 M* (pronase-sensitive) was plotted as a function of time. The slope of the resulting line is a measure of the endocytic rate constant (K e ) that defines the probability of an ␣ 2 M*-occupied receptor being internalized in 1 min at 37°C (16). Since K e can be influenced by ligand concentration (16), we performed the In/Sur analysis at two different ␣ 2 M* concentrations, 1 nM (data not shown) and 2 nM (Fig. 4B). During the entire period of the experiment, the amount of surface-bound 125 I-␣ 2 M* was relatively constant with respect to time (from 3 to 15 min) in both transfected cell lines. The calculated K e value for LRP-wt was 0.259 Ϯ 0.044 min Ϫ1 (correlation coefficient of 0.959), and that for LRP-RS was 0.252 Ϯ 0.050 min Ϫ1 (correlation coefficient of 0.946) (Fig. 4B). Therefore, there is an equal (ϳ25%) probability for both wild-type and mutant LRP that an occupied receptor would be internalized in 1 min at 37°C. These data again indicate that lack of processing does not impair the endocytic function of LRP.
Internalization and Recycling of LRP-Constitutive cycling of biotinylated LRP between the cell surface and the endosomal compartment was analyzed using two protocols. First, cell surface LRP was labeled with a thiol-cleavable biotin linker (sensitive to glutathione) at 0°C, and turnover of the surface LRP was followed at 37°C (Fig. 5A). Initially, nearly all of the biotinylated LRP proteins were on the cell surface, as shown by their complete sensitivity to glutathione (time 0). After incubation at 37°C for 5 min, approximately 60% of the surface LRP was internalized, therefore gaining glutathione resistance. No significant difference in the constitutive LRP internalization was observed between LRP-wt and LRP-RS. The internalized LRP remained glutathione-resistant for 30 min, suggesting Plasma membranes were isolated from cells expressing wild-type (LRP-wt) or the processing mutant LRP (LRP-RS) and treated with (ϩ) or without (Ϫ) neuraminidase as described under "Experimental Procedures." Samples were resolved by SDS-PAGE, and LRP proteins were detected by immunoblotting as in Fig. 1C. B, susceptibility to Pseudomonas exotoxin A of LRP-transfected cells. Parental cells (LRP-null) and cells expressing wild-type (LRP-wt) or processing mutant LRP (LRP-RS) were plated, treated with PEA, and assayed for [ 35 S]methionine/cysteine incorporation as described previously (12). Normal CHO-K1 cells expressing endogenous LRP were used as a control.  ). B, degradation of 125 I-␣ 2 M*. Confluent monolayers (six-well plates) were incubated with 1 nM 125 I-␣ 2 M* at 37°C. At the indicated times, the medium was collected and proteins were precipitated with 15% trichloroacetic acid. Trichloroacetic acid-soluble, noniodide radioactivity was determined in the aqueous phase following treatment of the 15% trichloroacetic acid supernatant with 40% KI and H 2 O 2 and extraction of liberated iodine with chloroform. Data points are RAP-inhibitable degradation, determined by subtracting total degradation counts from those determined in the presence of 100 nM RAP. that at least 30 min is required for recycling of LRP back to the plasma membrane. There was no change in the level of total cellular LRP-wt or LRP-RS during the entire course of the experiments, as demonstrated by immunoblotting using anti-␣-chain or anti-␤-chain antibodies (data not shown).
In the second approach, LRP associated with the endosomal compartment was biotinylated. This was achieved by incubating cells with thiol-cleavable biotin at 37°C to label both cell surface and endosomal LRP and subsequently removing surface biotin signal with glutathione at 0°C. The return of biotinylated LRP from endosomes back to the cell surface was followed by monitoring its sensitivity to glutathione in subsequent incubations at 37°C for up to 60 min (Fig. 5B). While no change in glutathione sensitivity of the endocytosed LRP was observed during the first 30-min incubation, there was a 50% decrease in the biotin label associated with both LRP-wt and LRP-RS (i.e. increase in glutathione sensitivity) at 60 min of chase (Fig. 5B, blots labeled ϩ glutathione). The acquisition of glutathione sensitivity resulted from recycling of the endosomal LRP back to the cell surface rather than from lysosomal degradation, since recovery of biotinylated LRP during the entire chase was undiminished (Fig. 5B, blots labeled Ϫ glutathione). These results suggest that approximately 50% of LRP cycled back to the plasma membrane between 30 and 60 min after endocytosis. Furthermore, the lack of processing did not affect the rate of LRP recycling.
Post-translational Stability and Intracellular Trafficking of LRP-The effect of abolishing processing on post-translational stability of LRP was determined by pulse-chase analysis. At the end of 1 h of pulse, only the proreceptor forms of LRP-wt and LRP-RS were detected (Fig. 6A, band a). Both of the proreceptors were located predominately within the ER, since the carbohydrate moieties of the proreceptors were sensitive to digestion with Endo H (Fig. 6B, bands a and aЈ). Conversion of the Endo H-sensitive proreceptors into Endo H-resistant species (from band a to band b) occurred between 1 and 2 h of

FIG. 5. Recycling of LRP-wt and mutant LRP-RS in transfected cells.
A, cells (60-mm dishes) were labeled with NHS-SS-Biotin on ice and then incubated at 37°C for the indicated times. After treatment with glutathione to cleave the biotin moiety of cell surface LRP, cells were lysed, and LRP was recovered by immunoprecipitation. Samples were resolved by SDS-PAGE (on a 3-8% gradient gel), and biotinylated LRP was visualized using horseradish peroxidase-conjugated streptavidin. Biotinylated LRP was quantified by scanning densitometry. The glutathioneresistant signal at each time point is expressed as a percentage of the untreated control (NT). B, cells were labeled with NHS-SS-biotin at 37°C for 2 h and treated with glutathione (0°C) to remove the biotin moiety from cell surface LRP. Monolayers were then incubated at 37°C for the indicated time, and the sensitivity of biotinylated LRP to glutathione was analyzed at each point as described above. Repetition of the experiments yielded essentially the same results.
chase. There was no difference between LRP-wt and LRP-RS in the time required to attain Endo H resistance, suggesting that the rate of transit between the ER and Golgi apparatus was independent of LRP processing. The half-life of the post-ER form of mutant LRP-RS (i.e. band b) as estimated between 4 h (at the peak) and 15 h of chase, was approximately 10 h, similar to that of LRP-wt (i.e. band c in Fig. 6A, Endo H-resistant in Fig. 6B).
However, a difference in the rate of ER exit was evident between LRP-wt and mutant LRP-RS. At between 2 and 4 h of chase, nearly all of the proreceptor form of LRP-wt was converted from Endo H-sensitive (band a) to Endo H-resistant (band b) species, and the latter was concomitantly processed into 515-kDa ␣-chain (Fig. 6, A and B, band c) and 85-kDa ␤-chain (not shown). The ER residence time (t1 ⁄2 ) of the proreceptor form of LRP-wt (band b) was approximately 4 h, as estimated by comparing its intensity between 2 h (at the peak) and 6 h of chase (Fig. 6B). In contrast, only a portion of the proreceptor form of mutant LRP-RS underwent additional carbohydrate modification to become Endo H-resistant (Fig. 6B,  bottom, bands b), whereas the remainder did not leave ER and remained Endo H-sensitive (band a) 12-15 h after chase. The ER residence time of LRP-RS was at least 13 h, since there was no apparent decrease in its intensity between 2 and 15 h of chase (Fig. 6A, bottom). Thus, mutation at the furin cleavage site drastically retarded the ER exit of LRP but had little effect on its post-translational stability. DISCUSSION Since the discovery of proteolytic cleavage of LRP in 1990 by Herz et al. (7), little has been learned about the biological function of this post-translational processing event. Here, using an expression system, we examine the effect of a single amino acid substitution at the processing site of chicken LRP on its endocytic functions and intracellular trafficking. We have shown that the proteolytic processing can be completely blocked by replacing Arg 3942 with Ser at the furin cleavage site (RHRR 3942 2), clearly indicating that LRP processing indeed occurs at the consensus furin cleavage site during its maturation and transport to the cell surface. Data obtained from the present studies have shown that the processing mutant LRP-RS is transported onto the plasma membrane ( Fig. 2A), presumably via the classical ER-to-Golgi trafficking. In addition, the mutant LRP-RS appears to be fully functional, as demonstrated by its ability to bind the three ligands examined (i.e. ␣ 2 M*, RAP, and PEA). Expression of LRP-RS can, as effectively as LRP-wt, restore the sensitivity of LRP-deficient cells to PEA (Fig. 2B) as well as reconstitute their ability to bind, internalize, and degrade ␣ 2 M* in a RAP-inhibitable fashion (Fig. 3). Furthermore, the rates of endocytosis and recycling of LRP-RS, either in ␣ 2 M*-bound form (Fig. 4) or else as an unoccupied receptor (Fig. 5), are nearly identical to those of LRP-wt. Therefore, the present results, together with previous observations of the truncated minireceptors (11,12), all indicate that the unprocessed receptors exhibit endocytic functions indistinguishable from LRP-wt, providing strong evidence that post-translational endoproteolysis is not essential for endocytosis.
While mutational analysis detects no involvement of LRP processing in endocytosis, the possibility that processing is involved in other cellular functions remains to be tested. Recently, Bu et al. (22) have reported that the cytosolic tail of LRP is phosphorylated in a neuronal cell line (GT1-1) transfected with the receptor for nerve growth factor and that the phosphorylation increases upon the growth factor treatment. The functional role of phosphorylation of the cytosolic tail of LRP is yet unclear. Early studies with a processing mutant form of insulin receptor showed that the unprocessed receptor expressed in transfected CHO cells bound insulin normally but was inefficiently phosphorylated (23,24). Therefore, it will be of interest to determine in future studies the effect of abolishing processing on phosphorylation of LRP. Since the cytoplasmic domain of integral membrane proteins plays a major role in sorting and targeting, phosphorylation may be part of signaling pathways responsible for acute regulation and distribution of LRP between different intracellular pools (22). The existence of an endosomal "storage" pool of LRP has been suggested in studies with rat adipocytes, in which a rapid and transient mobilization of LRP from the intracellular pool to cell surface was observed when the cells were treated with insulin (25). Recently, we have obtained preliminary experimental evidence showing that the insulin-induced cell surface presentation of LRP is sensitive to inhibitors of phosphatidylinositide 3-kinase in 3T3-L1 adipocytes. 2 Thus, it will also be of interest in future studies to determine if LRP processing plays a role in the acute redistribution of LRP between endosomes and plasma membranes.
An important observation made in this study is the delayed LRP transport out of ER caused by the single amino acid substitution at the furin cleavage site. The prolonged ER retention displayed by the mutant LRP-RS is not likely attributable to an artifact of overexpression of the mutant receptor, since the level of protein synthesis between LRP-RS and LRP-wt was similar (Fig. 6). The mechanism responsible for the delayed ER exit of LRP-RS is unclear, but some possibilities may be considered. It has been shown in human glioblastoma U87 cells (27,28) and in transgenic mice (29) that the molecular chaperone RAP confers proper folding of LRP and prevents ligand-induced aggregation of the receptor. Five RAP-binding sites have been identified within the class A repeats of human LRP (30), but it is not known if there are additional RAPbinding sites within sequences other than the class A repeats. It is possible that introducing the Arg 3942 -to-Ser substitution may have impaired the affinity of nascent LRP for RAP. Interaction of LRP with RAP within the secretory pathway was not determined in this study. However, we did not find the retained mutant LRP-RS to be aggregated or rapidly degraded within the ER. The half-life of mutant LRP-RS that was retained in the ER (i.e. remained sensitive to Endo H) was 12-15 h (Fig. 6). Therefore, it is unlikely that the retarded ER exit of mutant LRP-RS was a consequence of impaired association with RAP. Another possibility for the delayed mutant LRP transport out of the ER is that the sequence RXRR acts as a recognition site for the ER quality control system. Studies with renin, a secretory protein derived from proteolytic cleavage of the precursor prorenin, have shown that the presence of a protease processing site per se within prorenin is essential for its sorting and secretion (31). Exposure of the potential proteolytic cleavage site (RXRR) within LRP may be essential for efficient ER exit in addition to furin cleavage. Conclusive evidence of this functionality of the LRP processing site will require further definitive studies.
Recently, it was reported that the 515-kDa ␣-chain of human LRP could occur as a soluble form in the plasma (32), raising the possibility that processing may play a role in shedding. In CHO cells transfected with either wild type or the processing mutant LRP, we could not detect any ␣-chain or LRP-derived fragments in the media (data not shown). These results suggest that processing per se is not sufficient for the release of ␣-chain from cells in culture. However, when a cDNA construct containing an engineered premature stop codon at the carboxyl terminus of the RNRR sequence was expressed in CHO cells, we found that the encoded LRP ␣-chain accumulated in the culture medium. 3 Thus, soluble LRP fragments found in the plasma may arise from proteolysis at the cell surface (independent of furin cleavage) and subsequent dissociation of the ␣-chain from the transmembrane fragment.
In summary, post-translational processing of the chicken LRP does not grossly affect its endocytic functions in transfected CHO cells. However, transport of LRP out of the ER appears to be extremely sensitive to a single amino acid substitution at the endoproteolytic cleavage site. Mechanisms that are responsible for the retarded ER exit of the processing mutant form of LRP are currently under investigation.