Carboxypeptidase E is efficiently secreted and internalized via lysosomes

Carboxypeptidase E (CPE) a key factor in the biosynthesis of most peptide hormones and neuropeptides, is predominantly expressed in endocrine tissues and the nervous system. This highly conserved enzyme cleaves the C-terminal basic residues of the peptide precursors to generate their bioactive form. CPE is a secreted protein; however, the Intracellular pathways leading to its secretion are still obscure. We combined live-cell microscopy and molecular analysis to examine the intracellular distribution and secretion dynamics of fluorescently tagged CPE. CPE was found to be a soluble luminal protein as it traffics from the ER via the Golgi apparatus to lysosomes. Moreover, CPE is efficiently secreted and reinternalized to lysosomes of neighboring cells. The C-terminal amphipathic helix of CPE is essential for its efficient targeting to, and secretion from lysosomes. Fluorescence resonance energy transfer demonstrated that CPE and its substrate neuropeptide Y (NPY) interact in the Golgi apparatus and Immunoprecipitation analysis demonstrated that both CPE and NPY are co-secreted. The implications of the well-defined CPE intra and extracellular routes are discussed.


INTRODUCTION
CPE is a 35kDa metal-binding exopeptidase that removes one or more basic amino acids (Lys or Arg) from the C-terminal of its substrates (Fricker & Snyder, 1983). CPE is a soluble enzyme that is co-translationally inserted into the ER lumen while its conserved signal sequence is cleaved. The last 25 amino acids at its C-terminal, considered to act as a membrane binding amphiphilic helix (AH) were associated with sorting of substrates into secretory granules via its membrane binding properties (Manser et al, 1991). Besides being a biosynthetic enzyme, CPE was shown to have several other functions such as a sorting protein to the regulated secretory pathway, facilitating endocytic and secretory transport processes and regulating signaling pathways (Ji et al, 2017;Skalka et al, 2016). CPE was shown to expedite the sorting of several of its substrates from the Golgi complex/TGN to the regulated secretory granules. The enzymatic activity sites of CPE are in organelles with pH=5.0 suggesting that it may be active in the Golgi complex and in lysosomes. CPE interacts shortly with its substrates as a protease but may also bind additional proteins for longer time periods to facilitate sorting from the TGN to lysosomes or secretory granules.
Here, we combined biochemical analysis with live-cell microscopy to analyze the intra-and extracellular transport pathways of CPE. Tagged at the C-terminal with a fluorescent protein (FP), CPE was expressed and analyzed in intact living cells, demonstrating that the latter is a soluble luminal protein which is efficiently secreted via lysosomes. The secretion of several truncated CPE-constructs demonstrated that the C-terminal AH is essential for its efficient targeting to and secretion from lysosomes. Fluorescence recovery after photobleaching (FRAP) revealed that CPE is rapidly exported from the ER. Interaction of CPE with its substrate NPY in the Golgi complex was shown using fluorescence resonance energy transfer (FRET) and immunoprecipitation assays. Finally, we found that secreted CPE is reinternalized and apparently concentrated in lysosomes.
(FRAP) to quantitatively analyze the ER to Golgi transport rates of CPE-GFP and ss-mCherry (Fig. 3E). COS7 cells were co-transfected with both CPE-GFP and ss-mCherry. A region of interest (ROI) over the Golgi was photobleached and images were captured for about 30 min to analyze the ER to Golgi traffic-mediated recovery. Figure 3E demonstrates a typical experiment with two cells in the field of view. The graph shows ROI fluorescence intensity plotted against time and fitted to an exponential equation. The time scale (1/k) for Golgi recovery in the two cells examined is 1.35 and 2-fold faster for CPE-GFP compared to ss-mCherry, supporting the notion of a rapid and efficient secretion of the former.

CPE is secreted via lysosomes.
We and others demonstrated that CPE is found in lysosomes [Fig. 2, supplementary movie 2, and (Saito et al, 2011)]. Based on our results with BFA ( Fig. 3D) we tested the hypothesis that CPE may be secreted from lysosomes. To this end we analyzed the effects of chloroquine (CQ), an inhibitor of lysosomal acidification and function (Hamano et al, 2008) (Fedele & Proud, 2020) on CPE-GFP secretion. CQ was reported to block protein trafficking in and out of lysosomes by raising their intraluminal pH (Andrei et al, 1999;Ling et al, 1998;Luo et al, 2011;Tapper & Sundler, 1990). Moreover, lysosomal secretion of FABP4 was demonstrated using CQ (Villeneuve et al, 2018). In the presence of CQ, CPE-GFP accumulated in large peripheral punctate organelles (Fig. 4A). Moreover, WB analysis demonstrated that CQ as well as Bafilomycin A1 (BafA1, a reagent known to disrupt autophagosome-lysosome fusion and lysosome acidification (Mauvezin et al, 2015) inhibited CPE-mCherry secretion. In accordance, CPE secretion was enhanced in the presence of ionomycin, a calcium ionophore that triggers calcium release thereby inducing lysosomal secretion (Bennett et al, 1979;Carvalho et al, 1982). As shown in Figure 4D the secretion of CPE was increased following ionomycin treatment, further supporting the existence of a CPE lysosomal exocytosis pathway. The WB in Fig. S1 demonstrates that the effect of CQ and BafA1 can be reproduced in PC12 cells that secrete endogenous CPE ( Fig. S1).
Mass spectrophotometric analysis of proteins that co-immunoprecipitated with CPE-GFP, identified GRP78/BIP. GRP78 is a well-characterized secreted ER chaperone that mediates the correct folding and assembly of newly synthesized proteins (Hurtley et al, 1989). It was also shown to target aberrant proteins for proteasomal degradation (Corrigall et al, 2004;Delpino & Castelli, 2002;Kern et al, 2009;Marin-Briggiler et al, 2010). Interestingly, a recent study demonstrated trafficking and co-release of GRP78/BIP with β-coronaviruses through lysosomal secretion (Ghosh et al, 2020). Thus, we co-expressed CPE-GFP or GFP-tagged CPE N'-terminal domain with GRP78-RFP. Coimmunoprecipitation analysis demonstrated that CPE interacts with GRP78 via its N terminal domain (Fig. 4E). Moreover, the two proteins colocalized in LGP120-GFP positive lysosomes (Fig. 4F, upper panels). Coexpression of CPE-mTag-BFP with an inefficiently secreted GRP78 K633Q mutant (Li et al, 2016) significantly reduced targeting of CPE to lysosomes, supporting the premise that GRP78 is involved in the lysosomal secretion of CPE.

The C'-terminal amphipathic helix is involved in CPE secretion.
The 25 C'-terminal amino acids of CPE encode an amphipathic α-helix (AH) domain that has been shown to associate with membranes under acidic pH (5.5-6.5) and plays a role in targeting CPE into the regulated secretory vesicles (Dhanvantari & Loh, 2000;Varlamov & Fricker, 1996). The CPE AH shows high conservation throughout evolution (Fig. S2). It contains of several conserved clusters of alternating acidic and basic amino acids. The secretion of CPE lacking the 33 C'-terminal amino acids was shown to be blocked while CPE that lacks only 23 C'-terminal residues is still secreted (Varlamov & Fricker, 1996). We generated several truncations at the C'-terminal of CPE as shown in Fig. 5A. Contrary to previous reports, we found that CPE lacking the C'-terminal 25 amino acids (CPE-451) is the only construct that gained TGN localization (Fig. 5B). All the other larger truncations were transport incompetent and thus retained in the ER. Interestingly, CPE-451 was detected in lysosomes however to a lesser extent compared to the wildtype CPE (Fig. 5C). WB analysis demonstrated that CPE-451 was also secreted although inefficiently (Fig. 5D). These data strongly support the hypothesis that CPE secretion depends on lysosomal targeting.

Secreted CPE is endocytosed and delivered to the lysosomes but not to the Golgi.
Previously it has been reported that a chimeric protein containing CPE C'-terminal 25 amino acids could be recycled back to the TGN from the PM (Arnaoutova et al, 2003).The mechanism involved interaction with lipid rafts as well as with ARF6. In our hands, we could not find any evidence for membrane interaction of CPE as it consistently localized to the lumen of organelles. However, we tested whether secreted CPE can be internalized back into the cell. To this end, COS7 cells transfected with CPE-mCherry or ss-mCherry as a control, were re-plated together with cells expressing the lysosomal marker LGP120-GFP ( Fig. 6A). Figure 6B shows images captured 24h after re-plating the cells. Secreted CPE-mCherry was internalized by the LGP120-GFP expressing cells and localized to lysosomes ( Fig. 6B). A similar experiment, where the acceptor cells expressed the Golgi marker GalT-CFP demonstrated that CPE internalization apparently did not occur via the Golgi apparatus.
To verify the results, we repeated the experiment by adding only media from the CPE-mCherry secreting cells (Fig. 6D). Here CPE-mCherry was also detected in lysosomes after 24h as shown in Fig. 6E. These data demonstrate, for the first time to our knowledge that CPE secreted from lysosomes into the media can re-enter and localize to lysosomes of neighboring cells. Thus, the intra and extra cellular routes of CPE remarkably resemble that of corona viruses that are internalized and exit their host cell via lysosomes (Ghosh et al., 2020).

CPE and NPY interact and co-localize in lysosomes.
Next, we investigated the interaction of CPE with its substrates along its intracellular route.
Neuropeptide Y (NPY) is a conserved 36 amino acid peptide, expressed in the peripheral and central nervous systems and is a known substrate of CPE (Brakch et al, 1997;Fricker, 1988;Tang et al, 2009). Primarily we tested the colocalization of CPE with its NPY substrate.
Co-expression of CPE-mTagBFP, NPY-mRFP and LGP-120-GFP demonstrated that both CPE and NPY localized to the lumen of the Golgi apparatus (not shown) and lysosomes ( Fig.   7A). Next, we asked to determine if CPE interacts with its NPY substrate. To this end, we used stepwise acceptor photobleaching (Yaffe et al, 2015) to analyze FRET interaction between CPE-GFP and NPY-mRFP. The analysis was carried out in the Golgi as the CPE-GFP fluorescence was eliminated by the acidic environment of the lysosomes. Cells expressing both CPE-GFP and NPY-mRFP serving as donor and acceptor, respectively, were fixed as described in the ‫״‬Methods‫״‬ section prior to a series of photobleaching of an ROI over the Golgi apparatus. The images in Fig. 7B shows the diminishing fluorescence of the acceptor (upper panel) as well as the resulting increase in donor fluorescence (pseudocolor, lower panel). The FRET efficiency is plotted with the acceptor fluorescence and is compared with a negative control where the donor is LGP-120-GFP. The fluorescent proteins in this control pair were separated by the membrane as the GFP tagged to LGP-120 donor is cytosolic. Thus, the increase in donor fluorescence as a result of photobleaching the acceptor, is indicative of the close proximity of CPE and NPY and their potential interaction.
These results were confirmed by immunoprecipitation analysis of CPE-GFP and NPY-mRFP. Moreover, the interaction of the secreted proteins apparently persisted in the medium ( Fig. 7C). Figure 7D demonstrates that the transport incompetent CPE-350 truncation did not affect the lysosomal targeting of NYP-mRFP.
Taken together, combining live cell and molecular analysis we delineate for the first time the secretory route of CPE, a key enzyme in the biosynthetic pathway of numerous hormones and neuropeptides. CPE is secreted from-and is internalized directly to lysosomes.

DISCUSSION
CPE was reported to function in the TGN and dense core secretory granules (Dikeakos & Reudelhuber, 2007;Kim et al, 2006;McGirr et al, 2013;Ramamoorthy & Whim, 2008;Topalidou et al, 2020).The results presented here indicate that CPE may be targeted to, and function in the acidic lumen of lysosomes. The majority of studies describing the secretion of CPE have been based on biochemical analyzes and fractionation assays. Here, for the first time to our knowledge, live cell microscopy combined with biochemical molecular analysis of CPE secretion was carried out. Primarily, we found that counter to previous reports, CPE is a soluble luminal protein that is efficiently targeted to lysosomes from where it is secreted. Direct secretion from the Golgi is demonstrated as well. These findings are in agreement with some reports (Fricker & Snyder, 1983) and contradict others such as those reporting a role for lipid rafts and the small GTPase ARF6 in its targeting (Arnaoutova et al., 2003). Moreover, we found that the C'-terminal AH, previously reported to mediate membrane anchoring and retention in the cell, is essential for both efficient lysosomal targeting and secretion. A possible interpretation of the quantitative FRAP analysis is that the AH serves as a sorting signal for CPE which recruits the secretory machinery for efficient transport. The ER to Golgi transport-mediated recovery is significantly faster compared to that of the ss-mCherry. Furthermore, the results of the FRAP experiment are an underestimation when considering a rapid Golgi export of CPE. The deletion of the AH resulted in a significant decrease in CPE secretion. However, its lysosomal targeting was at least in part preserved. The secretion of CPE via lysosomes is compatible with secretion from the regulated secretory pathway found in professional secreting cells. This pathway is regulated by calcium and thus can be activated using reagents such as ionomycin. The role of the co-transported GRP78 in CPE trafficking is not clear. However, GRP78 seems to be obligatory for lysosomal targeting. GRP78 was reported to travel with corona virus proteins that enter and exit cells from lysosomes (Ghosh et al., 2020). We hypothesize that the role of GRP78 in CPE trafficking may be to prevent its denaturation or even degradation in the acidic hostile environment of lysosomes. The finding of the re-internalization of secreted CPE to lysosomes is intriguing as CPE was reported to be internalized to the Golgi apparatus. It is yet to be determined whether CPE internalization is specific or rather mediated by bulk flow fluid phase endocytosis. The latter is less likely as the fluorescence of re-internalized CPE was easily visualized in lysosomes whereas in the cell media it was highly diluted to bellow detection levels. The role of CPE in the processing of NPY was described over two decades ago (Brakch et al., 1997). The presence of both CPE and NPY in secretory and lysosomal organelles, their interaction in the Golgi apparatus, and their co-secretion, are in agreement with their reported colocalization in dense-core granules of cortical astrocytes (Cheng et al, 2013). The interpretation of our FRET data demonstrates a direct interaction between CPE and its NPY substrate. As CPE was also reported to act as a lysosomal targeting chaperon for substrates (Saito et al., 2011) it is not clear whether NPY binding results in its processing or simply a persistent interaction aimed to facilitate their targeting to lysosomes. However, based on our results and previous studies, NPY does not require CPE to arrive at lysosomes.
To summarize, we demonstrate that CPE is an efficiently secreted soluble enzyme that is sorted to lysosomes from where it is secreted. The physiological significance of reinternalization of secreted CPE back to lysosomes is yet to be discovered. CPE is a key enzyme that processes numerous hormones and neuropeptides. Thus, it is possible that CPE re-internalization to lysosomes of neighboring cells allows efficient processing of its substrates in acidic organelles by equally spreading the enzyme activity within the entire secreting organ.
Ronit Sagi-Eisenberg (Department of Cell and Developmental Biology, Sackler Faculty of Medicine, Tel Aviv University). F-CPE-mCherry was constructed using the Gibson Chew Back and Anneal Assembly technique to subclone CPE into pmCherry-N1 (Clontech) vector using XhoI and BamHI restriction sites. LGP120-GFP was purchased from Addgene. CPE-mTAG BFP was constructed using the Gibson Chew Back and Anneal Assembly technique to subclone mTAGBFP into F-CPE GFP vector using BamHI and NotI restriction sites. CPE-mTAGBFP-Amphipathic Helix was constructed by Gibson assembly by sub-cloning CPE without the Amphipathic Helix, mTAGBFP, and the CPE Amphipathic Helix motif into CPE mCherry vector using BamHI and NotI restriction sites. CPE 451 mCherry constructed by Gibson assembly by sub-cloning the 451 amino acids of CPE (without the Amphipathic Helix) into pmCherry-N1 (Clontech) vector using XhoI and BamHI restriction sites.As control vectors we used pEGFP-N3 and pmCherry-N1.All constructs were sequenced to confirm the fidelity of the process. Brefeldin A, Cycloheximide, Chloroquine, BafA1 and Ionomycin were purchased from Sigma-Aldrich (Rehovot, Israel).

Antibodies
The following antibodies were used: mouse anti-GFP (IB: 1:1,000; IP: 1: 400, Santa Cruz Fluorescence emissions resulting from 405 nm excitation for CFP, 488 nm excitation for GFP, and 543 nm excitation for mCherry were detected using filter sets supplied by the manufacturer. The confocal and time-lapse images were captured using a Plan-Apochromat ×63 1.4-numerical-aperture (NA) objective (Carl Zeiss MicroImaging). The temperature on the microscope stage was held stable (37°C) during time-lapse sessions using an electronic temperature-controlled airstream incubator. Images and movies were generated and analyzed using the Zeiss LSM Zen software and NIH Image and ImageJ software (W. Rasband, National Institutes of Health, Bethesda, MD, USA).

Fluorescence recovery after photobleaching (FRAP)
For FRAP measurements, a ×63 1.4-NA Plan-Apochromat objective was used on an inverted LSM800 system. Photobleaching of GFP and mCherry was performed using four to six rapid scans with the laser at full power. A region of interest (ROI) over the Golgi was photobleached at both 488 and 565 nm using a high-power laser, and images were captured for about 30 min to record the ER to Golgi traffic-mediated recovery. ROI fluorescence intensity was plotted against time. Data was fitted to a single exponential equation: Pre-and postbleach images were captured at 0.5 to 3 s intervals, using low laser intensity.
Fluorescence recovery in the bleached region during the time series was quantified using LSM Zen software (Carl Zeiss MicroImaging). For presentation purposes, 16-bit confocal images were exported as TIFFs, and their contrast and brightness were optimized in Adobe Photoshop software (San Jose, CA) or in ImageJ (Wayne Rasband, NIH, Bethesda MD).

FRET analysis
For acceptor photobleaching FRET, cells were grown on glass coverslips. The cells were fixed by addition of paraformaldehyde to the medium to a final concentration of 4% and incubation for 15 min at room temperature. The cells were then washed twice with PBS containing 1% FCS, and once with PBS, and were mounted for visualization on the microscope. The mCherry-tagged acceptor was photobleached in a ROI over the Golgi complex. The 488 nm or 514 nm laser line was used to assess the acceptor photobleaching.
FRET efficiency (E) was calculated from the CFP-channel images according to: EF = (Fpost-Fre)/Fpost, where F is the intensity of GFP fluorescence using the 488 nm laser before (prebleach) and after (post-bleach) photobleaching of mCherry using a high-power 543 nm laser.

CPE internalization assay
In order to evaluate internalization of CPE into cells, COS-7 cells were transfected with the indicated plasmids in a 6-well plate, and then, 24 h post transfection, the cells were combined and re-plated. Twenty-four hours later, the pooled cells were visualized by a confocal microscope. In some experiments, 24 h post transfection, the medium of LGP120-GFP transfected cells was replaced with medium from COS-7 cells transfected with F-CPE-mCherry or ss-mCherry. The cells were incubated for 30 min at 37 o C followed by visualization with a confocal laser microscope.

SDS-PAGE and Western blot analysis
Forty-eight hours following transfection, medium was collected and centrifuged at 11,000g for 5 min. The cells were then washed with PBS and solubilized in lysis buffer (100 mM NaCl, 50 mM Tris, pH7.5, 1% Triton X-100, 2 mM EDTA) containing protease inhibitor cocktail (Sigma-Aldrich). Cell lysates were incubated on ice for 20 min, then were homogenized and clarified by centrifugation at 14,000 g for 15 min at 4°C. The protein concentration was determined by Bradford reagent. Samples with equal amounts of protein were separated on 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE), and the proteins were then transferred onto nitrocellulose membranes. The membranes were blocked with 5% low-fat milk and incubated with specific primary antibodies at 4°C (as indicated), washed with PBS containing 0.001% Tween-20 (PBST), and incubated with the appropriate horseradish peroxidase-conjugated secondary antibody for 1 hour at room temperature. After washing in PBST, membranes were subjected to enhanced chemiluminescence (ECL) detection analysis.

Figure 5. The C'-terminal amphipathic helix of CPE is essential for efficient secretion from lysosomes. A.
A scheme listing the CPE truncation mutants generated. Light green bars is the cleavable signal peptide. AH -C'-terminal amphipathic helix. B. Intracellular localization of the truncation mutants in living cells. COS7 cells were co-transfected with the Golgi marker GalT-CFP (red) and one of the CPE-GFP truncation mutant (green). Confocal Images were obtained 24 hours post transfection. Bar = 10µm. C. The CPE451-mCherry truncation arrives at lysosomes. COS7 cells were co-transfected with LGP120-GFP (green) and either the full-length CPE-mCherry or CPE451-mCherry (red). The images in white square are enlarged 4-fold in images on right hand side. Bars= 10 µm and 1 µm in enlarged inserts. D. WB analysis of CPE truncation mutants. HEK293T cells were transfected with one of the CPE truncation mutants. Cells were harvested and culture media were collected 48 hours post transfection. Lysates and media with equal protein concentrations or volumes, respectively, were loaded onto SDS-PAGE gels for WB analysis using an anti-GFP antibody. Graph shows band intensity for secreted wildtype and CPE451-GFP. E. Scheme of CPE constructs with the FP upstream or downstream to the AH. F. The FP at the C'-terminal does not affect CPE localization via steric hindrance on the AH. COS7 cells were co-transfected with LGP120-GFP (green), CPE-mCherry (red), CPE-mTAG-BFP with the AH at the C'-terminus (blue). Images of living cells were captured 24 hours after transfection. Inserts I, II and III are enlarged 5-fold on the right-hand side. Bar = 10 µm. Figure 6. Secreted CPE is endocytosed, delivered, and concentrated in lysosomes but not in the Golgi. A. Schematic illustration of the experimental procedure. Cells were cultured together after separately transfecting each with LGP120-GFP or GalT-GFP, CPE-mCherry, or ss-mCherry as a control. Images were captured 24 hours after co-culturing of the cells. B. Lysosomal internalization of secreted CPE-mCherry. Images of the cells expressing LGP120-GFP (green) were captured 24 hours after coculturing with the CPE-mCherry or ss-mCherry (red) expressing cells (red). Inserts are regions containing lysosomes magnified 4fold. Bar = 2 µm C. CPE is not internalized to the Golgi apparatus. Same as in B accept that GalT-GFP was used instead of CPE-GFP. Bar = 2 µm. D. Schematic illustration of the experimental procedure where only the media from cells transfected with CPE-mCherry was added to cells transfected with LGP-120-GFP. E. COS-7 cells transfected with LGP120-GFP (green) were incubated for 30 minutes with media obtained from CPE-mCherry (red) transfected cells. Confocal live cell images are shown. Insert showing lysosomes is magnified 3-fold. Bar = 2 µm.

Figure 7. CPE and its substrate NPY interact and co-localize in lysosomes. A.
CPE and its substrate NPY colocalize to the lumen of lysosomes. COS-7 cells were co-transfected with CPE-mTAG-BFP (blue), NPY-mRFP (red), and LGP120-GFP (green). Images were captured 24 hours post transfection. Inserts are enlarged 5 -fold. Bar = 5µm B. FRET analysis demonstrates interaction of CPE and NPY. Cells were co-transfected with CPE-GFP and NPY-mCherry. Stepwise acceptor photobleaching was carried out after fixing the cells with 4% formaldehyde. Top panel shows a typical cell during the experiment. White square is the photobleached region over the Golgi. White square are magnified 4-fold and the green FRET donor channel is shown in pseudo-color with a look up table on the right hand side. Graphs show donor (green) and acceptor (red) fluorescence for the experiment (left) and for a negative control (right)where LGP-120-GFP was used as a donor. C. IP analysis confirms interaction of CPE and NPY. HEK293T cells were co-transfected with CPE-GFP and NPY-mRFP. Collected medium and harvested cells were harvested 48 hours post transfection and subjected to IP using anti-CPE antibody. Precipitants were separated on SDS-PAGE and analyzed with anti-GFP and anti-RFP antibodies to recognize the CPE and NYP, respectively. D. NPY does not require CPE for its lysosomal targeting. Confocal image of a typical cell co-expressing CPE350-GFP and NPY-mRFP captured 24 h after transfection. Insert is enlarged 3-fold. Bar = 10 µm. Figure S1. Endogenous CPE is secreted from lysosomes in PC12 cells. WB analysis of CPE secretion by PC12 cells. Cells were grown to confluency and at time 0 were incubated in the absence or presence of BAFA1 or chloroquine for the indicated times. obtaining sample 0. Media samples were analyzed by WB using anti-CPE antibody. Figure S2. The CPE AH is highly conserved throughout evolution. Human CPE C'-terminal AH was analyzed using ConSurf (Ashkenazy et al, 2016) alignment of mouse CPE AH (top) with primitive vertebrates