Clathrin light chains function in mannose phosphate receptor trafficking via regulation of actin assembly

Abstract

Clathrin-coated vesicles (CCVs) are major carriers for endocytic cargo and mediate important intracellular trafficking events at the trans-Golgi network (TGN) and endosomes. Whereas clathrin heavy chain provides the structural backbone of the clathrin coat, the role of clathrin light chains (CLCs) is poorly understood. We now demonstrate that CLCs are not required for clathrin-mediated endocytosis but are critical for clathrin-mediated trafficking between the TGN and the endosomal system. Specifically, CLC knockdown (KD) causes the cation-independent mannose-6 phosphate receptor (CI-MPR) to cluster near the TGN leading to a delay in processing of the lysosomal hydrolase cathepsin D. A recently identified binding partner for CLCs is huntingtin-interacting protein 1-related (HIP1R), which is required for productive interactions of CCVs with the actin cytoskeleton. CLC KD causes mislocalization of HIP1R and overassembly of actin, which accumulates in patches around the clustered CI-MPR. A dominant-negative CLC construct that disrupts HIP1R/CLC interactions causes similar alterations in CI-MPR trafficking and actin assembly. Thus, in mammalian cells CLCs function in intracellular membrane trafficking by acting as recruitment proteins for HIP1R, enabling HIP1R to regulate actin assembly on clathrin-coated structures.

Clathrin-coated vesicles (CCVs) are major carriers for protein and lipid cargo at the plasma membrane (PM), the trans-Golgi network (TGN), and the endosomal system. The structural unit of clathrin, the triskelion, is composed of three copies of clathrin heavy chain (CHC) and up to three clathrin light chains (CLCs) (1–4). CLCs consist of two related proteins, CLCa and CLCb, which are expressed in all tissues at varying relative levels (5, 6) and are heterogeneously distributed in clathrin triskelia (7). However, the functional role(s) of CLCs remains poorly defined. CLCs are stably associated with CHC, and in brain the proteins are found with a 1:1 stoichiometry (1, 2). In in vitro assembly assays, CLCs inhibit the assembly of clathrin triskelia into cages with maximal inhibition occurring at ratios of CLCs to CHC close to 1:1 (8, 9). Thus, one prevalent model is that CLCs function as negative regulators of CHC assembly. However, this model has been challenged with the observation that knockdown (KD) of both CLCa and CLCb has no effect on the kinetics of EGF receptor or transferrin receptor endocytosis, suggesting that clathrin assembly is normal (10). Moreover, in non-brain tissues there is a deficit of CLCs relative to CHC (3, 4), and at the observed CLC to CHC ratio (≈0.2:1) the inhibitory effect of CLCs upon assembly should be minimal (8).

Another potential role for CLCs is to regulate huntingtin-interacting proteins (HIPs). Recently, HIP1 and HIP1-related (HIP1R), which are expressed in neuronal and nonneuronal tissues and cells (11–13), were identified as binding partners for CLCs (13–16). HIP1 was originally discovered in a two-hybrid screen with huntingtin, the protein product of the Huntington's disease gene, and has thus been implicated in the pathophysiology of this inherited neurodegenerative disorder (17). HIP1R was identified based on sequence similarity with HIP1 but does not bind huntingtin (11). HIP1 and HIP1R share common features: an N-terminal ANTH domain for phospholipid interaction (18), a central helical domain for dimerization and CLC interaction, and a C-terminal actin-binding THATCH domain (13–16, 19). Both are coat components of clathrin-coated structures (CCS) at the PM and the TGN (13, 20–24) but display significant differences in binding partners as HIP1 binds strongly to CHC and adaptor protein 2 (AP-2), whereas HIP1R binds weakly to CHC and does not bind AP-2 (13, 16, 22–24). Moreover, only HIP1R binds strongly to actin (13, 16, 21), and it was also recently shown to bind the actin regulatory protein cortactin (25). In fact, HIP1R and the yeast HIP orthologue Sla2p function in membrane trafficking by linking CCS to the actin network (13, 15, 20, 26–28). Thus, another potential function for CLCs is to regulate the association of CCS with the actin cytoskeleton by scaffolding HIP1R.

Here we have used siRNA to knock down CLCs. Consistent with previous results (10), we found no effect on clathrin-mediated endocytosis (CME) or the formation of clathrin-coated pits (CCPs) at the PM. In contrast, we found alterations in protein trafficking at the TGN resulting from disruption of HIP1R recruitment to CCS and disorganization of the actin cytoskeleton. Our results demonstrate that CLCs contribute to intracellular membrane trafficking via regulation of actin assembly.

Results

CLC KD Disrupts Trafficking of TGN-Derived Cargo.

Simultaneous KD of CLCa and CLCb has no influence on CME of transferrin or EGF (10). A second major site of clathrin-mediated trafficking is the TGN, where CCV budding generates carrier vesicles to transport cargo such as the cation-independent mannose-6 phosphate receptor (CI-MPR) to the endosomal system (6). To examine for potential defects in this pathway, we generated siRNA duplexes specific for CLCs (10). COS-7 and HeLa cells had strongly reduced levels of CLCa or CLCb after transfection with siRNA specific to each isoform but were unaffected by a nonspecific control siRNA or by mock transfection [supporting information (SI) Fig. 6A]. When siRNAs for each CLC isoform were used in combination, the KD efficiency was 95% or greater for both isoforms in both cell types (SI Fig. 6 A and B) and the proteins were nearly undetectable by immunofluorescence (see for example Figs. 4 and 5).

Fig. 4.

Disruption of CLC function alters the actin cytoskeleton. (A) Mock- or CLCa and CLCb siRNA-transfected HeLa cells were processed for immunofluorescence with antibodies for CLC, and actin distribution was assessed by using phalloidin-TRITC. Higher-magnification images correspond to the areas marked by the white squares in the lower-magnification images. (Scale bar: 10 μm and 2 μm for the low- and high-magnification images, respectively.) (B) HeLa cells were transfected with GFP-CLCb or GFP-CLCb EED to QQN mutant (CLCb QQN). At 48 h after transfection, cells were fixed and processed to detect actin by using phalloidin-TRITC (red). Higher-magnification images correspond to areas marked by white squares in the lower-magnification images. (Scale bar: 10 μm and 2 μm for the low- and high-magnification images, respectively.) (C) CLCa and CLCb siRNA-transfected HeLa cells were processed for immunofluorescence with antibodies for CLC (green) and CI-MPR (blue), and actin distribution was assessed by using phalloidin-TRITC (red). Higher-magnification images correspond to the area marked by white squares in the lower-magnification images. (Scale bar: 10 μm and 4 μm for the low- and high-magnification images, respectively.)

Fig. 5.

CLC KD disrupts the levels and distribution of HIP1R. (A) Cell lysates from mock- or CLCa and CLCb-transfected cells were processed for Western blot with antibodies for the indicated proteins. Quantification reveals a 25 ± 3% (mean ± SEM, n = 12) decrease in HIP1R levels in CLC KD versus mock-transfected cells. (B) Immunofluorescence analysis of the localization of HIP1R in mock-transfected COS-7 cells and cells transfected with CLCa and CLCb siRNA. (Scale bar: 10 μm.) (C) Mock- or CLCa and CLCb siRNA-transfected HeLa cells were subsequently transfected with plasmid encoding HIP1R and processed for immunofluorescence with antibodies for HIP1R (green) and CLCs (blue), and actin distribution was assessed by using phalloidin-TRITC (red).

Intriguingly, trafficking of the CI-MPR was altered after CLC KD. Specifically, the endogenous receptor displayed a more clustered, perinuclear pattern in KD cells (Fig. 1A). The redistribution was also pronounced in HeLa cells stably expressing a CD8–CI-MPR chimera (29) (Fig. 1A) and was seen using another combination of siRNAs (SI Fig. 7). The distribution of adaptor protein 1 (AP-1), a major clathrin adaptor at the TGN and endosomes (6), was also altered, showing a more clustered, perinuclear pattern despite the fact that the levels of AP-1 were unchanged (Fig. 1 B and C). A significant portion of CI-MPR that clusters in the perinuclear area after CLC KD surrounds but does not colocalize with the TGN marker TGN46 (Fig. 1D), suggesting that the CI-MPR is not trapped directly in the TGN. Consistently, CCVs were readily detectable by EM analysis in the vicinity of the TGN after CLC KD, suggesting that CCVs still bud from TGN membranes (Fig. 1E). Together, these data are consistent with a functional role for CLCs in trafficking of the CI-MPR subsequent to CCV budding/formation.

Fig. 1.

CLC KD disrupts CI-MPR trafficking. (A) HeLa or HeLaM cells (expressing CD8–CI-MPR), transfected without (Mock) or with CLCa and CLCb siRNAs for 96 h, were processed for immunofluorescence with antibodies for CI-MPR or the CD8 tag. Higher-magnification images in Center correspond to areas marked by white squares in the lower-magnification images in Left. [Scale bar: 10 μm (Left and Right) and 2 μm (Center).] (B) HeLa and COS-7 cell lysates were prepared 96 h after transfection without (Mock) or with CLCa and CLCb siRNAs, and equal protein aliquots were processed for Western blot with antibodies for the indicated proteins. (C) Immunofluorescence analysis of AP-1 in mock-transfected COS-7 cells and cells transfected with CLCa and CLCb siRNA. Higher-magnification images in Right correspond to areas marked by white squares in the lower-magnification images in Left. [Scale bar: 10 μm (Left) and 2 μm (Right).] (D) Mock-transfected or CLCa and CLCb siRNA-transfected COS-7 cells were processed for immunofluorescence with antibodies for CI-MPR (red) and TGN46 (green). (Scale bar: 2 μm.) (E) HeLa cells, mock-transfected or transfected with CLCa and CLCb siRNA, were processed for conventional transmission EM. (Scale bar: 100 nm.) The arrows indicate CCVs. (F) HeLaM cells expressing CD8–CI-MPR chimera were transfected without siRNA (Mock) or with CLCa and CLCb siRNAs for 96 h. Live cells were then placed at 4°C and exposed to CD8 mAb for 30 min followed by a switch to 37°C for 3 h to allow receptor/antibody internalization. Cells were then fixed and processed for immunofluorescence with secondary antibody to detect the CD8 antibody. (Scale bar: 10 μm.)

A pool of CI-MPR is present at the PM from where it undergoes CME to enter endosomes with subsequent trafficking to the TGN (30, 31). To assess the effect of CLC KD on this pool of CI-MPR, CD8–CI-MPR chimera-expressing HeLa cells (29) were exposed to CD8 antibody at 4°C to label PM-localized receptor followed by a switch to 37°C to allow internalization. In control cells, after 3 h, the chimeric receptor was internalized and was localized to punctate structures, likely corresponding to TGN/endosomes (Fig. 1F). In CLC KD cells, the receptor displayed a stronger and denser signal in the perinuclear area (Fig. 1F). Thus, CI-MPR that has entered cells via CME is also retained in the altered TGN/endosome trafficking pathway.

The ability of CI-MPR to enter cells after CLC KD suggests that CLCs are not required for CME of this receptor. In fact, CME of anti-CD8 antibody recognizing chimeric CD8–low-density lipoprotein receptor (LDLR) and CD8–CI-MPR was unaffected by CLC KD but was blocked by CHC KD (Fig. 2A). Moreover, there was no influence of CLC KD on the formation of CCPs and CCVs at the PM as determined by morphometric analysis of EM data (Fig. 2 B and C), and the expression levels and distribution of AP-2, a key component of PM-associated CCPs, were normal (Fig. 1B and SI Fig. 8). Thus, consistent with earlier results (10, 32), CLCs do not appear to play a major role in CME or in the regulation of clathrin assembly in vivo.

Fig. 2.

Endocytic function is normal after CLC KD. (A) Mock- and siRNA-treated HeLaM cells, stably expressing CD8–LDLR or CD8–CI-MPR chimeras, were allowed to endocytose the anti-CD8 antibody for 8 min and were then fixed and processed for immunofluorescence with secondary antibody to detect the CD8 antibody. (Scale bar: 10 μm.) (B) Mock- and CLCa/CLCb siRNA-treated HeLa cells were processed for conventional transmission EM. (Scale bar: 100 nm.) (C) Multiple randomly selected EM images from mock- or CLCa and CLCb siRNA-transfected HeLa and COS-7 cells were overlaid with a 500-nm grid. The number of PM and CCP intersections, respectively, were as follows: HeLa (mock), 7,211 and 40; HeLa (CLCs siRNA), 6,819 and 38; COS-7 (mock), 8,056 and 55; COS-7 (CLCs siRNA), 7,211 and 58.

CLC KD Impairs Sorting of Cathepsin D.

To confirm that CLC KD impairs trafficking of TGN-derived cargo, we used pulse–chase and immunoprecipitation to follow the maturation of newly synthesized cathepsin D, a lysosomal hydrolase that is targeted to endosomes/lysosomes through its binding to MPR and targeting to CCVs at the TGN. Cathepsin D binds to MPR as a 52-kDa pro-form. A first proteolytic cleavage yields an intermediate 48-kDa single-chain molecule that is cleaved into a mature two-chain enzyme consisting of light (14-kDa) and heavy (34-kDa) domains once cathepsin D has reached lysosomes (ref. 33; reviewed in ref. 34). CLC KD slows proteolytic processing of pro-cathepsin D and leads to a decrease in the accumulation of the intermediate form (Fig. 3). Because pro-cathepsin D processing to the intermediate form occurs during its transport to lysosomes, these observations suggest that CLC KD leads to a functional impairment of CI-MPR trafficking between the TGN and lysosomes.

Fig. 3.

CLC KD delays processing of cathepsin D. (A) Mock- or CLCa and CLCb siRNA-transfected COS-7 cells were pulsed for 2 h at 20°C with ³⁵S-labeled methionine/cysteine followed by a chase with unlabeled methionine. At the indicated time points, cathepsin D was immunoprecipitated from cell lysates and analyzed by SDS/PAGE and autoradiography. (B) The percentage of the pro- and intermediate forms of cathepsin D compared with total cathepsin D signal is plotted for mock- and CLCa and CLCb siRNA-transfected cells at each time point (mean ± SEM of n = 8 experiments).

Overexpression of a CLC Mutant Disrupts CI-MPR Trafficking.

The conserved domain near the N terminus of CLCs is crucial for CLC regulation of CHC assembly in vitro (9) and binding to HIPs (35, 36). We thus generated a CLCb mutant in which three critical acid residues in the conserved domain, EED, were mutated to QQN (CLCb QQN). This mutant is deficient for binding to HIPs (35, 36) but binds CHC normally (35). Neither WT GFP-CLCb nor GFP-CLCb QQN influenced CME of transferrin (SI Fig. 9A) whereas a redistribution of CI-MPR similar to that seen after CLC KD was observed in a proportion of the cells overexpressing CLCb QQN but not in cells overexpressing WT CLCb (SI Fig. 9B). Thus, the disruption of TGN/endosome trafficking observed in both CLC KD cells and in cells overexpressing the CLC mutant may result from alterations in the association of HIPs with the clathrin machinery.

CLC KD Leads to Abnormal Actin Assembly.

One function for HIPs is to regulate clathrin assembly, and another is to regulate actin assembly, mainly through HIP1R, which binds directly to actin and cortactin (13, 16, 21, 25). Because HIP1R KD leads to actin assembly defects and impairs trafficking, primarily at the TGN (20), we looked for modification of the actin cytoskeleton in CLC KD cells and observed a dramatic change in actin organization with the prominent accumulation of actin rings and patches that were rarely seen in control cells (Fig. 4A). Similar actin structures were observed in cells overexpressing CLCb QQN but not WT CLCb (Fig. 4B). In triple labeling experiments, despite some peripheral patches, the actin patches observed after CLC KD were generally located in proximity to the clustered CI-MPR (Fig. 4C). Interestingly, CLC KD leads to a 25 ± 3% (mean ± SEM, n = 12) decrease in the levels of HIP1R (Fig. 5A). Immunofluorescence analysis with a newly developed HIP1R antibody that recognizes both the peripheral and perinuclear pools of the protein (SI Fig. 10) reveals that CLC KD causes alterations in the localization of HIP1R (Fig. 5B). In particular, there is less staining in the perinuclear compartment, and overall the staining is weaker and more diffuse (Fig. 5B). Thus, the defect in actin assembly occurring after CLC KD could be caused by decreased levels or altered localization of HIP1R. To address this, we transfected HIP1R in control and CLC KD cells. The presence of CLC was required for proper targeting of HIP1R to CCS (Fig. 5C and SI Fig. 11), and there was no rescue of the actin phenotype in HIP1R overexpressing cells in the absence of CLC (Fig. 5C), suggesting that the interaction of HIP1R with CLCs and its proper localization are necessary for actin regulation. Thus, a major function for CLCs is to recruit HIP1R to CCS to regulate actin assembly and trafficking of TGN-derived CCVs.

Discussion

One potential function for CLCs is to regulate the assembly of clathrin lattices, because addition of CLCs to CLC-free CHC triskelia or hub fragments inhibits spontaneous assembly in vitro (8, 37). However, CLC KD does not affect CME of transferrin, EGF (10, 32), LDLR, or CI-MPR. Moreover, CCPs and CCVs still form after CLC KD. In fact, most proteins that stimulate clathrin assembly including AP-1, AP-2, and a number of clathrin accessory proteins (AP180, auxilin 1, Dab2, epsin 1, and enthoprotin/epsinR) interact with the CHC terminal domain (38–43), whereas CLCs interact with the proximal leg segment. Moreover, CHCs separated from CLCs assemble with unimpaired efficiency in the presence of AP180, AP-1, or AP-2 (44). Thus, it seems unlikely that regulation of clathrin assembly represents a major functional role for CLCs in mammalian systems in vivo, although our data do not rule out the possibility of minor kinetic defects in CME resulting from subtle alterations in clathrin assembly.

Given that CLCs do not play a major role in clathrin assembly, how do they function in trafficking? Here we demonstrate that CLC KD disrupts the localization of HIP1R to CCS. HIP1R binds through a proline-rich sequence to the SH3 domain of cortactin (25). By forming a complex with cortactin, HIP1R inhibits actin assembly such that HIP1R KD leads to overly abundant actin assembly (25). The abnormal actin structures observed are similar to the actin patches and rings we observe after CLC KD. Interestingly, a 40–60% reduction in HIP1R levels has no effect on CME, but causes defects in trafficking of the MPR and the maturation of cathepsin D (20) whereas a 90% reduction in HIP1R is needed to cause defects in CME (27). This suggests that clathrin trafficking at the TGN is more sensitive to HIP1R-mediated actin regulation than at the PM and may explain the compartment-specific trafficking defects observed in our study. Thus, in mammalian cells CLCs appear to function as scaffolding proteins for HIP1R and are necessary for normal regulation of actin assembly. These results are consistent with recent studies in yeast in which CLC expression rescues defects in endocytosis after loss of CHC function dependent on the ability of CLC to bind Sla2p, the yeast HIP homologue (28). Thus, the ability of CLC to scaffold a HIP protein is also necessary for normal vesicle trafficking in yeast although, in this system, the CLC/Sla2p interaction is mandatory at the PM but not at the TGN (28). A second difference is that the ability of CLCs to scaffold HIP1R in mammalian systems depends on CHC. Specifically, CHC KD leads to a nearly complete loss in CLC expression (SI Fig. 12A) (45), and, whereas GFP-tagged CLC localizes normally in control cells, it remains predominantly soluble in CHC KD cells (SI Fig. 12B). GFP-CLC also remains soluble in control cells if it bears the point mutation W127R (SI Fig. 12C), a mutation known to abolish its interaction with CHC (46). The ability of CLCs to function as scaffolds for HIP1R in the presence of CHC is also consistent with a recent molecular model of clathrin cages based on cryo-EM that indicates that CLCs are on the outside of the cage where they are best positioned to interact with cytosolic factors (47).

In the model we propose for CLC function, cargo destined for endosomal membranes still sorts to CCS and buds off the TGN membrane after CLC KD. Consistent with this model, CI-MPR clusters around the TGN but is not localized directly at the TGN membrane after CLC KD, and CCVs are observed near the TGN by EM. Furthermore, the lack of an effect of CLC KD on CME suggests that there is not a general deficit in CCV budding or fission. However, this must be interpreted carefully because we do not measure budding/fission directly, and recent evidence suggests an important role for a dynamic actin cytoskeleton in budding from the TGN (48), possibly through a requirement in dynamin-dependent fission (49). We think that the phenotypes observed after CLC KD are more likely due to alterations in vesicle transport with accumulation of endosome-destined cargo in the vicinity of the TGN. This could result from vesicles becoming trapped in an overly stable actin meshwork. Alternatively, it has become clear that a dynamic interplay between the actin- and microtubule-based cytoskeletons is necessary for transport of TGN-derived vesicles, perhaps by capturing nascent vesicles and propelling them toward microtubule-based transport systems (50). In conclusion, our data reveal CLCs as important regulators of the actin cytoskeleton, laying the groundwork for future studies to examine its important roles in vesicle trafficking.

Materials and Methods

Antibodies, Fluorescent Probes, and cDNA Constructs.

mAb CON-1 against CLCa/CLCb was from Santa Cruz Biotechnology, and a CLCa/CLCb polyclonal antibody was generated as described (51). A mAb against CHC for immunofluorescence was generated from the hybridoma X22 obtained from American Type Culture Collection. Other mAbs were from the noted commercial sources: CHC and AP-1 (γ-adaptin) from BD Transduction Laboratories; CI-MPR and AP-2 (α-adaptin) from ABR; and actin from ICN Biomedicals. Rabbit polyclonal antibodies for tubulin and cathepsin D were from ICN Biomedicals and EMD Biosciences, respectively. Sheep anti-TGN46 was from Serotec. mAb against CD8 was a gift of Matthew Seaman (University of Cambridge, Cambridge, U.K.). Polyclonal antibody recognizing CI-MPR was a gift of Paul Luzio (University of Cambridge, Cambridge, U.K.). A rabbit polyclonal antibody recognizing HIP1R was generated against the peptide sequence (QDHQLDKKDGIYPAQLVNY) and was affinity-purified by using the same epitope. Transferrin-Cy3 was from Jackson Laboratories, and phalloidin-TRITC was from Sigma. The GFP-CLCb construct was a gift from Juan Bonifacino (NIH, Bethesda, MD). The GFP-CLCb (EED/QQN) mutant was generated by subcloning the GST-CLCb (EED/QQN) construct (35) into pEGFP-C1 vector from Clontech. The GFP-CLCb W127R mutant was generated by using the megaprimer technique as described in ref. 35, with subsequent cloning into pEGFPC1. The HIP1R construct cloned in pCI neo was a gift from Michael Hayden (University of British Columbia, Vancouver, BC, Canada). Constructs were transfected into cells by using Lipofectamine 2000 (Invitrogen), and cells were processed for immunofluorescence or transferrin uptake after 48 h as described below.

siRNA-Mediated KD.

Four siRNAs matching selected regions of human CLCa (CLCa1–4), one siRNA matching CLCb, and one siRNA matching CHC were previously described (10). Unless specified, silencing of CLCs was performed with CLCa1 alone or in combination with CLCb siRNA. siRNAs were synthesized by Dharmacon as 21-mers with UU overhangs, and the antisense strand was chemically phosphorylated to ensure maximal activity. Nonsilencing control siRNA was from Qiagen. For transfection, COS-7 or HeLa cells were plated in DMEM without antibiotics and were transfected at 60% confluency with 80 nM siRNA by using Lipofectamine 2000. For combined KD of CLCa and CLCb, each siRNA was used at 40 nM. For control, cells were transfected with Lipofectamine 2000 only or with nonsilencing siRNA. Experiments were generally performed 96 h after transfection.

Quantification of Protein Levels After KD.

Cells were homogenized in buffer A [20 mM Hepes (pH 7.4) containing 0.83 mM benzamidine, 0.23 mM PMSF, 0.5 μg/ml aprotinin, and 0.5 μg/ml leupeptin] and were centrifuged at 800 × g for 10 min. Equal protein aliquots of the supernatants were analyzed by SDS/PAGE and Western blot. Blot signals were quantified by using NIH ImageJ software.

Morphological Studies.

For immunofluorescence, cells were fixed with 3% paraformaldehyde in PBS (20 mM NaH₂PO₄/0.9% NaCl, pH 7.4) for 15 min, permeabilized with 0.2% Triton X-100/PBS for 4 min, and processed with appropriate primary and secondary antibodies. For the polyclonal anti-CI-MPR, cells were fixed in 3% paraformaldehyde for 15 min followed by 5 min of fixation/permeabilization in methanol at −20°C. For actin staining, cells were incubated with phalloidin-TRITC concomitantly with secondary antibodies. All images were captured on a Zeiss 510 confocal microscope. When comparing KD cells to control cells, confocal settings were first established for the control and then kept constant for the analysis of the KD cells. HeLa and COS-7 cells were prepared for ultrastructural analysis by using transmission EM, and subsequent morphometric analysis was as described previously (52).

Anti-CD8 Uptake.

HeLaM cells, expressing the CD8 reporter constructs CD8–CI-MPR or CD8–LDLR (kind gift of Matthew Seaman), were grown to 70% confluency on 22-mm coverslips. After 96 h of siRNA transfection, cells were processed for anti-CD8 uptake as described previously (29). Briefly, after an ice-cold PBS wash, the coverslips were incubated for 15 min at 4°C in prechilled media. The coverslips were then washed in ice-cold PBS, blotted dry, and placed over a 100-μl drop of DME containing 1 μg of anti-CD8 antibody for 30 min at 4°C. After an ice-cold PBS wash, the coverslips were transferred to six-well dishes containing 3 ml of prewarmed media at 37°C. After incubation for 8 min, the coverslips were washed and fixed with ice-cold methanol/acetone. The continuous uptake assay was performed in a similar manner, except that the cells were incubated for 3 h continuously with the anti-CD8 antibody and there was no preincubation at 4°C.

Cathepsin D Sorting Assay.

COS-7 cells mock-transfected or transfected with CLCa and CLCb siRNA were washed and placed in cysteine/methionine-free DMEM (Invitrogen) for 1 h at 37°C. Cells were then pulsed for 2 h with 0.1 mCi/ml ³⁵S-proMix methionine/cysteine (Amersham) at 20°C, washed, and chased in DMEM with 1% FBS, 5 mM mannose-6-phosphate, and 1 mM methionine. At 1, 3, and 5 h, cells were washed with PBS and lysed in 50 mM Tris (pH 7.5) containing 150 mM NaCl, 10% glycerol, 5 mM EDTA, 1% Triton X-100, and protease inhibitors as described for buffer A. The lysates were microfuged for 5 min at 15,000 × g, and the supernatants were immunoprecipitated with polyclonal antibody against cathepsin D. Immunoprecipitated proteins were separated on SDS/PAGE and processed for autoradiography by using a STORM PhosphorImager (Amersham Biosciences) followed by exposure to x-ray film. Quantitation of eight independent experiments was performed by using ImageJ software.