Molecular characterization and gene disruption of mouse lysosomal putative serine carboxypeptidase 1 Katrin Kollmann 1 , Markus Damme 1 , Florian Deuschl 1 ,Jo ¨ rg Kahle 2 , Rudi D’Hooge 3 , Renate Lu ¨ llmann-Rauch 4 and Torben Lu ¨ bke 1 1 Abteilung Biochemie II, Georg-August Universita ¨ tGo ¨ ttingen, Germany 2 Abteilung Molekularbiologie, Georg-August Universita ¨ tGo ¨ ttingen, Germany 3 Laboratory of Biological Psychology, KU Leuven, Belgium 4 Anatomisches Institut, Universita ¨ t Kiel, Germany The lysosomal compartment plays a pivotal role in the degradation of macromolecules within the cell. To date, over 60 soluble lysosomal hydrolases and acces- sory proteins and 25 lysosomal membrane proteins have been identified [1–3]. Defects in the lysosomal proteins mostly result in one of about 50 lysosomal storage diseases (LSDs) which are characterized by the accumulation of undigested materials in the lysosomes. As a result of the clinical relevance of soluble lyso- somal proteins in LSDs and a notable number of LSD-like diseases of unknown etiology, there is a com- mon interest in the identification of the proteome of the lysosomal compartment and of the soluble luminal lysosomal mannose 6-phosphate (M6P)-containing Keywords gene disruption; lysosomes; processing; Scpep1; serine carboxypeptidase Correspondence T. Lu ¨ bke, Zentrum Biochemie und Molekulare Zellbiologie, Abteilung Biochemie II, Georg-August Universita ¨ t Go ¨ ttingen, Heinrich-Du ¨ ker-Weg 12, D-37073 Go ¨ ttingen, Germany Fax: +49 551 395979 Tel: +49 551 395932 E-mail: tluebke@gwdg.de (Received 4 July 2008, revised 18 December 2008, accepted 23 December 2008) doi:10.1111/j.1742-4658.2009.06877.x The retinoid-inducible serine carboxypeptidase 1 (Scpep1; formerly RISC) is a lysosomal matrix protein that was initially identified in a screen for genes induced by retinoic acid. Recently, it has been spotlighted by several proteome analyses of the lysosomal compartment, but its cellular function and properties remain unknown to date. In this study, Scpep1 from mice was analysed with regard to its intracellular processing into a mature dimer consisting of a 35 kDa N-terminal fragment and a so far unknown 18 kDa C-terminal fragment and the glycosylation status of the mature Scpep1 fragment. Although Scpep1 shares notable homology and a number of structural hallmarks with the well-described lysosomal carboxypeptidase protective protein ⁄ cathepsin A, the purified recombinant 55 kDa precursor and the homogenates of Scpep1-overexpressing cells do not show proteo- lytic activity or increased serine carboxypeptidase activity towards artificial serine carboxypeptidase substrates. Hence, we disrupted the Scpep1 gene in mice by a gene trap cassette, resulting in a Scpep1 ⁄ b-galactosidase ⁄ neo- mycin phosphotransferase fusion protein. The fusion protein is devoid of the C-terminal half of Scpep1, including two amino acids of the assumed catalytic triad which is indispensable for its predicted serine carboxypepti- dase activity. However, Scpep1-deficient mice were viable and fertile, and did not exhibit either lysosomal storage or reduced lysosomal SC activity under any tested condition. Abbreviations AEBSF, 4-(2-aminoethyl)benzenesulfonyl fluoride; CBZ, benzyloxycarbonyl; Cpvl, carboxypeptidase vitellogenic-like; CPY, carboxypeptidase Y; Ctsa, protective protein ⁄ cathepsin A; FA, furylacryloyl; geo, b-galactosidase ⁄ neomycin phosphotransferase; Lamp1, lysosomal associated membrane protein 1; LSD, lysosomal storage disease; M6P, mannose 6-phosphate; MEFs, mouse embryonic fibroblasts; MPR, mannose 6-phosphate receptor; PNGase F, peptide N-glycosidase F; RISC, retinoid-inducible serine carboxypeptidase; SC, serine carboxypeptidase; Scpep1, serine carboxypeptidase 1; Scpep1-gt, Scpep1 gene trap. 1356 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS proteins in particular. Soluble lysosomal proteins receive M6P residues on their N-linked oligosaccha- rides [4], which are recognized in the trans-Golgi net- work by M6P receptors (MPRs) required for lysosomal transport [5]. By exploiting the M6P recog- nition marker of the soluble lysosomal proteins for MPR-dependent affinity chromatography, followed by their identification by mass spectrometry, we and others have identified a novel putative serine carboxy- peptidase 1 (Scpep1) [6–9]. Originally, Scpep1 was identified in rat aortic smooth muscle cells by a screen- ing for retinoid-inducible genes, as reflected by its ini- tial name, ‘retinoid-inducible serine carboxypeptidase’ (RISC) [10]. Northern blot analyses demonstrated high transcript levels in kidney and aorta in rat and lower levels in heart, spleen and lung, whereas the human transcript was detected strongly in the kidney and heart but at a low level in a number of other tissues [10]. In mice, Scpep1 is expressed in embryonic heart and vasculature, as well as in a broad range of adult tissues [10]. The mouse Scpep1 gene (GeneID: 74617; cDNA Accession No. NM_029023) encodes a product of 452 amino acids (Protein Accession No. NP_0832299) that localizes to the lysosomes [11]. Fur- thermore, it has been demonstrated that, in mice, a 55 kDa Scpep1 precursor is processed into a 35 kDa form [11]. Although no peptidase activity has been demonstrated so far, Scpep1 has been assigned to the serine carboxypeptidase (SC) family S10 because of reasonable sequence homology to members of this family, such as the lysosomal protective pro- tein ⁄ cathepsin A (official gene name Ctsa; 35% simi- larity) and four conserved domains that are predicted to constitute the substrate-binding site and three cata- lytic sites. Each of these catalytic sites accounts for one amino acid of the catalytic triad Ser-Asp-His [10,12]. To obtain an insight into the physiological and cellular function of the putative lysosomal SC Scpep1, we analysed the molecular properties of Scpep1 and generated an Scpep1 gene trap (Scpep1-gt) mouse model. Results Molecular forms of Scpep1 In order to generate Scpep1-specific antisera, we purified a C-terminally His-tagged version of full- length mouse Scpep1 from secretions of stably expressing HT1080 cells (HT1080-Scpep1). Coomassie and silver staining after SDS-PAGE revealed an apparent molecular size of 55 kDa for the secreted and purified His-tagged Scpep1 starting with Ile29, as identified by N-terminal sequencing. We derived Scpep1-specific antisera from rat and rabbit. Both antisera were suitable for confirming the lysosomal localization of endogenous Scpep1 by immunofluores- cence (see Fig. S1). Western blot analysis using an antibody directed against the C-terminal His-tag detected the 55 kDa Scpep1 in cell extracts and in the medium of HT1080- Scpep1 cells (Fig. 1A, lanes 3 and 4), but not in untransfected HT1080 cells (lanes 1 and 2). Unexpectedly, an additional 18 kDa form of Scpep1 was detectable by the anti-His IgG1 in homogenates of HT1080-Scpep1 cells (lane 3), thus representing the C-terminal fragment of processed Scpep1. Both of our Scpep1-specific antisera detected the 55 kDa precursor in homogenates and secretions of HT1080-Scpep1 cells (lanes 7, 8 and 11, 12), as well as a strong signal at 35 kDa in the homogenates (lanes 7 and 11), representing the N-terminal moiety of processed Scpep1. Rabbit antiserum showed low B α α 1pe pcS- tibba r Standard (kDa) Retention time (min) 20 25 30 40 35 18 mn082DO 160 67 43 13.7 α -His α -Scpep1 α -Scpep1 rabbit rat 150 75 50 37 25 20 15 10 C M C M C M C M C M C M 55 35 18 kDa kDa 0801TH - 0801TH siH-1 p epcS 08 01 T H -0801TH s i H-1 p e p cS 0 8 0 1T H -080 1 TH siH-1pepcS A Lane 1 2 3 4 5 6 7 8 9 10 11 12 Fig. 1. Molecular forms of Scpep1. (A) Analysis of molecular forms of Scpep1: 100 lg of cell lysates (C) and 50 lL of medium (M) of HT1080 and HT1080-Scpep1 were separated by SDS-PAGE, blotted and probed with the a-His antibody and the a-Scpep1 antisera from rabbit and rat, respectively. (B) Gel filtration analysis of a lysosome- enriched fraction (F2): 50 lg of F2 were buffered in 20 m M Mes (pH 4.5) containing 150 m M NaCl, loaded onto a Superdex 75 col- umn on an analytic SMARTÔ system (Pharmacia) and eluted in 20 lL fractions at a flow rate of 40 lLÆmin )1 , which were analysed by western blot using the rabbit a-Scpep1 antibody. A mixture of molecular mass standard proteins, including IgG (160 kDa), albumin (67 kDa), ovalbumin (43 kDa) and ribonuclease A (13.7 kDa), was applied to gel filtration under the same conditions. K. Kollmann et al. Functional characterization of lysosomal Scpep1 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS 1357 cross-reactivity with polypeptides in homogenates of untransfected human HT1080 (lane 5), whereas no specific signal could be detected in the medium (lane 6) or with rat antiserum in HT1080 cells (lanes 9 and 10). Omitting the reducing agents did not alter the mobility properties in SDS-PAGE (data not shown). However, gel filtration chromatography with a lysosome-enriched fraction from mouse liver showed that the 35 and 18 kDa subunits of Scpep1 co-eluted in the same fractions with an apparent size between 43 and 67 kDa (Fig. 1B), suggesting that both fragments are linked to each other noncovalently. Processing of Scpep1 To investigate the processing of the endogenous Scpep1 precursor in mice, mouse embryonic fibroblasts (MEFs) were pulse labelled with [ 35 S]methionine for 1 h and chased for up to 72 h. After immunoprecipita- tion, Scpep1 was separated by SDS-PAGE and analy- sed by autoradiography (Fig. 2A). In MEFs which had been labelled for 1 h (0 h chase), only the  55 kDa precursor was detected. After 2 h of chase, half of the 55 kDa precursor had been processed into 37 and 20 kDa intermediates. Between 4 and 6 h of chase, the 55 kDa precursor disappeared, whereas the 37 kDa intermediate and the final 18 kDa fragment which was derived from the 20 kDa fragment showed constant signals. At 24 h of chase, the 35 kDa fragment was finally processed, and both fragments remained detect- able even after 72 h of chase (Fig. 2A). The medium did not show any specific Scpep1 signal (data not shown). Some lysosomal peptidases, such as cathepsin B and cathepsin D, undergo autoproteolytic activation [13,14]. To investigate the autoproteolytic activation of Scpep1, we incubated 0.16–3.2 lm of recombinant 55 kDa Scpep1 precursor at varying pH (pH 4.5 and pH 7.5), temperature (4 °C, room temperature, 37 °C) and incubation time (1–16 h), but could not detect any processing of the precursor into mature forms by western blot analysis (data not shown). In order to further define the Scpep1-processing pro- tease, MEFs were pulse labelled in the presence or absence of various protease inhibitors. The conversion of the 55 kDa Scpep1 precursor into the mature form was sensitive to the serine protease inhibitor 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF) (Fig. 2B). However, two other serine protease inhibi- tors, aprotinin and antipain, had no effect, although the latter interferes with cathepsin A activity [15]. Pep- statin A, an aspartic protease inhibitor, and the cyste- ine protease inhibitor E-64, as well as the metal 0 2 4 6 24 48 72 A B C Chase (h) 55 35 150 100 75 50 37 kDa 18 ∗ ∗ ∗ 250 25 20 15 MEF kDa ∗ MEF Chase (h) –– ATDE 46-E FSBEA xiM-IP 150 100 75 50 37 kDa 250 25 20 15 kDa 0 4 ∗ ∗ 55 35 18 Chase (h) 0 6 0 6 55 35 NH 4 Cl –+ 150 100 75 50 37 kDa HT1080-Scpep1 C M C M C M C M % of t 0 100 – 15 45 100 – 20 55 Antipain Aprotinin Pepstatin A Fig. 2. Processing of Scpep1. (A) Immunoprecipitation of Scpep1 from MEFs using the rat-derived Scpep antiserum after pulse label- ling with [ 35 S]methionine for 1 h. Cells were chased for up to 72 h. Nonspecific signals are marked with asterisks (*) at chase time 0. (B) Effects of protease inhibitors on the processing of Scpep1. MEFs were pulse labelled and chased for 0 or 4 h in the absence ()) or presence of the following inhibitors: EDTA (2 m M final con- centration), E-64 (10 l M), pepstatin A (100 lM), AEBSF (1 mM), aprotinin (0.3 l M), antipain (75 lM) and a mixture containing all inhibitors in the assigned concentrations. Scpep1 was immunopre- cipitated and visualized by autoradiography. (C) Processing of Scpep1 in NH 4 Cl-treated HT1080-Scpep1 cells. Intracellular (C) or secreted (M) Scpep1 was immunoprecipitated after pulse labelling and 0 or 6 h of chase in the absence ()) and presence (+) of the lysosomotropic agent NH 4 Cl. Functional characterization of lysosomal Scpep1 K. Kollmann et al. 1358 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS chelator EDTA, had no effect on Scpep1 maturation. These results suggest that, under our test conditions, a serine protease different from cathepsin A mediates Scpep1 processing. In order to investigate whether Scpep1 maturation occurs in the endosomal–lysosomal compartment, HT1080-Scpep1 cells were pulse labelled in the absence or presence of NH 4 Cl (Fig. 2C). In untreated HT1080- Scpep1 cells, the 55 kDa precursor was processed into the 35 kDa fragment, but the 18 kDa fragment was not detected. In addition, HT1080-Scpep1 cells secreted large amounts of the 55 kDa precursor after 6 h of chase (45% of the total Scpep1 signal at chase 0 h). NH 4 Cl interfered with the intracellular processing of the precursor to the 35 kDa form, but only moder- ately enhanced the secretion of the precursor (55% of the Scpep1 signal at chase 0 h). These results indicate that Scpep1 matures in late endosomes or lysosomes and could be targeted in an M6P-dependent manner. Glycosylation of Scpep1 in MEFs The amino acid sequence of Scpep1 contains five puta- tive N-glycosylation sites, four of which are located within the 35 kDa N-terminal fragment (Asn64, Asn102, Asn126, Asn192) and one within the 18 kDa C-terminal fragment (Asn362). MEFs were pulse labelled for 1 h and chased for 4 h, and immunopre- cipitated Scpep1 was subjected to peptide N-glycosi- dase F (PNGase F) treatment for 1 h and separated by SDS-PAGE (Fig. 3). The major form of the Scpep1 precursor from MEFs migrated at an apparent molec- ular mass of 55 kDa (lane 1). In addition, a minor signal was detected with a slightly reduced molecular mass of  52 kDa that was partially covered by the 55 kDa form and most probably represents a less gly- cosylated Scpep1. PNGase F treatment of the precur- sor resulted in a shift towards  40 kDa in size (lane 2). After 4 h of chase, the  55 kDa precursor was completely converted into two major mature subunits of 35 and 18 kDa in size and a minor signal of  30 kDa that might arise from the 52 kDa precur- sor form (lane 3). The limited deglycosylation of the 35 kDa subunit led to a deglycosylated  25 kDa frag- ment via the  30 kDa fragment and an intermediate of 28 kDa (lane 4). PNGase F treatment of the 18 kDa subunit resulted in a partial shift towards a lower molecular mass of  16 kDa (lane 4). Consider- ing an apparent molecular mass of about 2–2.5 kDa per N-linked oligosaccharide [16,17], the results suggest that, in MEFs, all N-glycosylation sites are utilized. SC activity of Scpep1 Like carboxypeptidase Y (CPY) and Ctsa, Scpep1 is classified as a member of the SC type C family (S10.013, MEROPS database) and should preferen- tially exhibit proteolytic activity at acidic pH towards hydrophobic amino acids in the P1¢ position [18]. Puri- fied Scpep1 protein, mainly consisting of the 55 kDa precursor, recombinant CPY as a positive control and BSA were tested for SC activity towards different N-terminal blocked peptides, such as CBZ-Phe-Leu and FA-Phe-Phe (CBZ, benzyloxycarbonyl; FA, furyl- acryloyl), representing SC type C substrates, FA-Ala- Lys, as an SC type D substrate, and the non-SC substrate CBZ-Gly-Leu [19]. Although CPY cleaved SC substrates such as FA-Phe-Phe in a pH-dependent manner, neither purified Scpep1 precursor nor BSA showed proteolytic activity under any test condition (Fig. 4). As most lysosomal hydrolases are not active as zym- ogens, we determined the SC activity in homogenates of HT1080 cells and HT1080-Scpep1 cells (data not shown). Although the latter mainly show Scpep1 in its processed form, we could not detect any differences in acid SC activity. To exclude cell line and vector- specific effects of His-tagged Scpep1, we assayed Chase (h) 0 4 55 PNGase F –+ –+ MEF 35 18 3 2 0 1 0 1 ∗ ∗ Lane 1 2 3 4 Fig. 3. Glycosylation of Scpep1. MEFs were pulse labelled and chased for 4 h. Scpep1 was immunoprecipitated from the lysates, treated with PNGase F and separated by SDS-PAGE. The filled arrowheads point to the fully glycosylated forms of Scpep1 with the number of their N-glycans, and the open arrowheads to degly- cosylated forms of the 35 kDa processed form and the 18 kDa pro- cessed form. Nonspecific signals are marked with asterisks (*) at chase time 0. K. Kollmann et al. Functional characterization of lysosomal Scpep1 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS 1359 HT1080 cells that had been transiently transfected with an untagged variant of Scpep1, His-tagged Scpep1 or the mock vector (pcDNA3.1-Hygro) (Fig. 5A), as well as COS-7 cells transfected with Scpep1-His6 and the pCI-neo mock vector (Fig. 5B). Although large amounts of processed Scpep1 were detected in the homogenates, no differences in acid SC activity could be measured regardless of the cell line and substrate used. Scpep1-gt mice In order to obtain an insight into Scpep1 function, we generated a gene trap mouse model. A blast of Scpep1 cDNA within the BayGenomics (San Fran- cisco, CA, USA) database identified the ES cell line RST426 as a Scpep1-gt cell line. The gene trap vector used by BayGenomics contains a splice acceptor site upstream of a b-galactosidase ⁄ neomycin phosphotrans- ferase (geo) fusion gene (Fig. S2), which inserted into intron 7 of the Scpep1 gene as confirmed by genomic sequencing. Hence, the downstream exons 8–13 were deleted from the gene trap transcript and were replaced by the promoterless geo cassette. Most importantly, two amino acids, Asp371 and His431, of the putative catalytic triad for SC activity were excluded from the resulting fusion product. As gene trapping does not consistently result in the inactivation of a gene, we checked for Scpep1 mRNA with a 3¢-specific probe by northern blot analyses (Fig. S3) and for Scpep1 protein by western blotting (Fig. 5A). In Scpep1-gt mice, the 35 kDa Scpep1 signal was absent from virtually all tissues tested. However, an antibody against the geo moiety of the gene trap fusion product detected a 200 kDa protein in the tissues of Scpep1-gt mice, corresponding to the 35 kDa Scpep1 expression pattern in wild-type mice, confirming the calculated size for the Scpep1-geo fusion protein of about 200 kDa (Fig. 6A). Subcellular fractionation of mouse liver after Triton WR-1339 (tyloxapol) injection, including differential centrifugation steps followed by a discontinuous sucrose gradient, enables the isolation of a fraction (F2) which is  50-fold enriched in lysosomal marker enzymes such as b-hexosaminidase. Western blot ana- lysis of each fraction of the lysosomal purification from wild-type mice showed co-fractionation of the pro- cessed 35 kDa Scpep1 and 18 kDa Scpep1 with lyso- somal proteins such as cathepsin D (Ctsd) and lysosomal associated membrane protein 1 (Lamp1) in fraction F2 (Fig. 6B). Western blot analyses from sub- cellular fractions derived from Scpep1-gt mice failed to detect Scpep1 in fraction F2 (Fig. 6C). In contrast, the geo antibody revealed a specific  200 kDa signal in the microsomal fraction P, indicating that the Scpep1- geo fusion product was retained in the endoplasmic reticulum and ⁄ or in the Golgi (Fig. 6C). Phenotype of Scpep1-gt mice Genotyping of 350 offspring from heterozygous bree- dings showed the expected Mendelian frequency with 23.6% homozygous Scpep1-gt mice, indicating that Scpep1 is not essential for correct embryonic develop- ment. Homozygous Scpep1-gt mice and wild-type mice showed comparable sizes and weight developments, as well as fertility and mortality (data not shown). Deter- minations of blood (full blood count), serum (e.g. aspartate aminotransferase, c-glutamyl transferase) and urine parameters showed no pathological findings. The activities of several lysosomal hydrolases were normal in various tissues and lysosomes from liver and kidney of Scpep1-gt mice, and were inconspicuous with regard to their distribution in a Percoll gradient, indi- cating that the density and size of the lysosomes were unaltered (data not shown). The following organs were regularly examined histo- logically using semithin sections: liver, lung, kidney, spleen, pancreas, retina, cornea, and spinal cord; in some instances, the inner ear (cochlea) and cerebellar cortex were also investigated. Ultrastructural examina- tion was performed on liver and spinal cord of two wild-type and two Scpep-gt mice. We were unable to find any consistent differences between wild-type and Scpep-gt mice. In particular, there was no evidence of lysosomal storage in any of the numerous cell types inspected. 3.5 4.5 5.5 6.5 7.5 8.5 0.0 2.5 5.0 7.5 10.0 pH ytivitca CS .ceps (U·mg –1 ) Fig. 4. SC activity determination. C-terminally His-tagged 55 kDa Scpep1 precursor (h) was purified from stably expressing HT1080 cells and incubated at different pH values ranging from 3.5 to 8.5 with FA-Phe-Phe as SC substrate. Yeast CPY (d) served as a posi- tive control and BSA (D) as a negative control. Functional characterization of lysosomal Scpep1 K. Kollmann et al. 1360 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS SC activity in Scpep1-gt mice Homogenates from various tissues and lysosome- enriched fractions from liver (F2) of control mice and Scpep1-gt mice showed equal levels of acid SC activity (data not shown). We further separated F2 fractions from control and Scpep1-gt mice by gel filtration and tested each fraction for Scpep1 and Ctsa by western blot analysis and acid SC activity. The Ctsa elution profiles were similar in both F2 fractions ranging from 1.0 2.0 3.0 4.0 CBZ-Phe-Leu CBZ-Leu-Phe CBZ-Gly-Leu α 1p e pcS - α -GAPDH 0801TH kcom + 1pepcS + A Substrate * 55 35 75 50 37 25 kDa 37 6siH-1pepcS + 0 8 01TH kcom + 1pepcS + 6siH- 1p epc S + 080 1TH kcom + 1 pepc S + 6siH-1pepcS + 080 1TH kcom + 1pepcS + 6s iH-1pepcS + 0.5 1.0 1.5 2.0 CBZ-Phe-Leu CBZ-Leu-Phe CBZ-Gly-Leu B α 1pepcS- α -GAPDH 7-SOC kco m + 6siH-1pepcS + 7-S O C kc om + 6siH-1pepcS + 7-SOC kcom + 6siH-1pepcS + Substrate 7- S O C kcom + 6siH-1 p epc S + * 55 35 75 50 37 25 kDa 37 Specific activity of acid carboxypeptidases (U·mg –1 ) Specific activity of acid carboxypeptidases (U·mg –1 ) Fig. 5. Expression and acid SC activity of Scpep1 in COS-7 and HT1080 cells. HT1080 (A) and COS (B) cells were transfected with either the appropriate mock construct (pCI-neo for COS-7; pcDNA3.1-Hgyro for HT1080), Scpep1 or Scpep-His6, as indicated, and assayed for acid SC activity using various artificial substrates. The columns represent the mean of three technical replicates for each cell line. Scpep1 expres- sion was monitored by western blot analysis using 100 lg of the cell homogenates, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as loading control. Nonspecific signals are marked with asterisks (*). K. Kollmann et al. Functional characterization of lysosomal Scpep1 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS 1361 fraction 16 to 18 (Fig. 7), whereas Scpep1 signals were solely detectable in fractions 15–17 from control mice (Fig. 7). However, the lysosomal SC activity distribu- tion was roughly identical in both elution profiles, regardless of the presence or absence of Scpep1 (Fig. 7). Thus, Scpep1 did not show proteolytic activity towards common lysosomal SC type C and D substrates. Discussion To date, four putative lysosomal SCs have been identified, but proteolytic activity has only been proven for Ctsa and the distantly related prolyl- carboxypeptidase [19,20]. The third putative SC, carboxypeptidase vitellogenic-like (Cpvl), has been reported to be a lysosomal SC restricted to T W t g T W t g T W t g T W t g T W t g T W t g T W t g T W t g T W t g T W t g 1 p e p c S l o r t n o c T W t g α α 1 p e p c S - α α -geo kDa 25 20 100 75 50 37 15 α α -GAPDH 34 250 150 0 8 0 1 T H F E M A α α -Scpep1 α α -Ctsd α α -Lamp1 kDa -37 -25 - 150 - 100 -20 -20 -50 -37 -25 α α -geo -150 -50 -75 N E M L P S F1 F2 F3 F4 B C Wild-type mice Fraction α α -Scpep1 α α -geo α α -Lamp1 -100 - 150 -37 -25 -20 -50 -75 N E M L P S F1 F2 F3 F4 Scpep1-gt mice Fraction Testis Liver Kidney Intestine Stomach Bladder Brain Lung Heart Spleen Fig. 6. Differential western blot analysis of Scpep1 expression in tissues from wild-type (WT) and Scpep1-gt mice. (A) Protein from various tissue extracts (200 lg per lane) and cell lysates (HT1080 and MEF, 50 lg per lane) were separated by SDS-PAGE, blotted onto poly(vinylidene difluoride) membrane and probed with the antibodies as indicated. Tyloxapol-filled lysosomes from mouse liver of control mice (B) and Scpep1-gt mice (C) were separated by differential centrifugation (corresponding to fractions N–S). Fraction L (light mitochondria) was loaded under a sucrose gradient (F1–F4), resulting in a codistribution of Scpep1 with the lysosomal marker proteins cathepsin D (Ctsd) and Lamp1 in F2, as shown by western blot analysis after SDS-PAGE loaded with 250 lg for each fraction of the differential centrifugation (N–S) and 50 lg of F1–F4 of the sucrose gradient. The blot membrane was additionally probed with antibodies against neomycin phosphotransferase (a-geo). E, postnuclear fraction; F1–F4, sucrose gradient fractions 1–4; L, light mitochondria fraction; M, heavy mitochondria fraction; N, nuclear fraction; P, microsomal fraction; S, cytosolic fraction. Functional characterization of lysosomal Scpep1 K. Kollmann et al. 1362 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS macrophages [21]. Recently, it has been demonstrated that Cpvl localizes to the endoplasmic reticulum rather than to lysosomes, and hence a role in major histocompatibility complex loading has been sug- gested [22], but proof for the enzymatic activity of Cpvl has not yet been served. In this study, we focused on the characterization of the fourth putative lysosomal SC, Scpep1, with regard to its localization, processing, glycosylation and presumed SC activity. Molecular forms of Scpep1 Our antisera raised against the 55 kDa Scpep1 precursor confirmed the lysosomal localization of endogenous Scpep1 by immunofluorescence and co-fractionation, as postulated previously by our group and others [7,11]. Tissue-specific expression analyses were performed according to a recent study [11], with highest Scpep1 levels found in visceral organs such as the liver and kidney. Western blot analyses of homogenates from HT1080-Scpep1 cells and from a 50-fold lysosome-enriched fraction revealed the presence of the expected 35 kDa pro- cessed fragment and an as yet unknown 18 kDa C- terminal fragment, in contrast with a recent publica- tion [11] in which the processing of the Scpep1 pre- cursor to a C-terminal 35 kDa fragment and a putative, but undetected, N-terminal 16 kDa peptide was postulated. The maturation from a zymogen into a two-chain form bears resemblance to a number of other SCs, such as barley SC [23] and lysosomal Ctsa, in particular. Ctsa is synthesized as a 54 kDa precursor and further processed into N-terminal 32 kDa and C-terminal 20 kDa polypeptides [24]. However, although both subunits of Ctsa are linked to each other by disulfide bonds to form the 54 kDa monomer [24], the two-chain form of Scpep1 does not form disulfide bridges. Moreover, although Ctsa dimerizes and, together with b-galactosidase and sialidase, forms a large multienzyme complex [25], Scpep1 from mouse liver elutes at  50 kDa in gel filtration assays (Fig. 1B), as predicted for the mono- mer, and in this regard resembles the yeast SCs CPY [26] and KEX1 [27], which are also active as monomers. The maturation of lysosomal hydrolases from a zymogen is essential for their functional activation, and hence must be tightly regulated in terms of pro- tecting the cell against self-digestion. We could not mimic autoprocessing, as described in vitro for lyso- somal cathepsin B, D or L [13,14,28]. The matura- tion of the Scpep1 precursor into the 35 kDa fragment is a multistep process (Fig. 2A), which is prevented by the addition of the serine proteinase inhibitor AEBSF (Fig. 2B), as well as by the NH 4 Cl- mediated uncoupling of MPR-dependent lysosomal transport (Fig. 2C), suggesting that Scpep1 process- ing is mediated by a lysosomal serine proteinase. Previously, we have demonstrated that partially puri- fied mouse Scpep1 derived from BHK cells binds on immobilized MPR46 and MPR300 and is internalized by I-cell fibroblasts in an M6P-dependent manner [7]. Limited deglycosylation by PNGase F digest demonstrated that all putative N-glycosylation sites are occupied in MEFs. Sleat et al. [29] identified M6P sites on 92 MPR-binding proteins derived from human and mouse brain, which were both of lyso- somal function or unknown function. Although a total of 135 M6P sites were identified in 69 proteins, M6P sites on Scpep1 escaped the analysis [29]. Most probably, these sites were missed because of the size of tryptic peptides containing the N -glycosylation sites, which range from 30 to 62 amino acids, and thus may exceed the preset mass range of the MS analysis [29]. 13 16 17 18 1914 15 α-Ctsa α-Ctsa α-Scpep1 α-Scpep1 F2 WT F2 gt B A 13 14 15 16 17 18 19 0 1 2 3 4 Fraction Spec. SC activity (mU·mg –1 ) Fig. 7. SC activity profiling and western blot analyses of Scpep1 and Ctsa after gel filtration of lysosome-enriched fractions. F2 frac- tions derived from wild-type mice (s) and Scpep1-gt mice ( ) were separated on an FPLC Superdex 200 10 ⁄ 300 GL column in 20 m M Mes, pH 4.5, 150 mM NaCl. The collected fractions were assayed for SC activity (A) and analysed by western blot analyses (B), using the Scpep1 antiserum to detect the 35 kDa fragment and an anti- Ctsa rat IgG2B to detect the 32 kDa heavy chain of Ctsa. K. Kollmann et al. Functional characterization of lysosomal Scpep1 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS 1363 The lysosomal localization, processing and conserva- tion of critical domains are shared features of Scpep1 and Ctsa. However, under conditions adapted to SCs such as Ctsa or CPY, we failed to demonstrate acid SC activity of Scpep1 in any approach addressed so far, regardless of which molecular form of Scpep1 or substrate was assayed. Because of its zymogen status, it is reasonable that the purified Scpep1 precursor does not exhibit SC activity; however, surprisingly, homo- genates of HT1080 and COS-7 cells, which highly expressed and subsequently processed mouse Scpep1 in a tagged or untagged version, did not show any elevated acid SC activity. The lack of additional SC activity on Scpep1 overexpression may be ascribed to an unknown limiting factor of Scpep1 activation. As an example, lysosomal sulfatases are modified in the endoplasmic reticulum by the formylglycine-generating enzyme, which has been shown to be an essential and limiting factor for sulfatase activity [30,31]. Further- more, it has been reported that the activation of the cathepsin D precursor is accelerated when it is com- plexed with prosaposin [32]. The Scpep1-gt mouse did not exhibit an obvious phenotype and did not show any lysosomal storage, although we confirmed the loss of the lysosomal 35 kDa mature form of Scpep1. Despite the deletion of the entire C-terminus, including two critical amino acids of the putative catalytic triad, we were unable to show reduced acid SC activity in Scpep1-deficient mice. We would like to point out that the computational modelling of Scpep1 also predicts a Ctsa- (1ivyA) and CPY-like (1cpy_) folding (http://swissmodel. expasy.org/SWISS-MODEL.html; Fig. 8). In addition, the alignment of Scpep1 with several SCs from mouse, Saccharomyces cerevisiae and Trypanoso- ma cruzei identifies a highly conserved substrate binding site (I), as well as three conserved catalytic regions (II–IV), each embedding one amino acid of the catalytic triad, and hence strongly favouring acidic SC activity (Fig. 9). Consequently, the apparent lack of in vitro SC activity of Scpep1 in its processed form must be ascribed either to the selection of an inappropriate substrate or to a nonproteolytic function of Scpep1. It is worth mentioning that another study failed to demonstrate proteolytic activity of Scpep1 for Ctsa substrates such as endothelin-1 and, in combination with immunohistological studies, suggests a function in the homeostasis of the renal and reproductive Ctsa (1ivyA) Scpep1 (Model: 1ivyA) CPY (1cpy_) Scpep1 (Model: 1cpy_ ) Fig. 8. Predicted three-dimensional struc- ture of Scpep1. Scpep1 was homology modelled with Ctsa and CPY as templates to predict its three-dimensional structure using the SWISSMODEL alignment mode (http://swissmodel.expasy.org// SWISS-MODEL.html). Functional characterization of lysosomal Scpep1 K. Kollmann et al. 1364 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS systems [11]. A recent gene target Ctsa mouse model (Ctsa S190A ), in which the catalytic serine residue was substituted but the protective protein function was preserved [33], does not develop a secondary galacto- sialidosis like the ‘classic’ Ctsa-deficient mouse [34]. In Ctsa S190A mice, neither Scpep1 nor any other predicted lysosomal SC efficiently compensates for the loss of in vitro SC activity, resulting in residual activities of 5–10% in visceral organs [33], although Scpep1 and, moreover, Ctsa and Scpep1 are Fig. 9. Multiple alignment of the primary amino acid sequences of SCs. SCs were aligned according to the CLUSTALW algorithm. Sequences are highlighted in grey for two or three homologous sequences or in black for amino acids conserved in all four sequences. The substrate binding site (I) is defined by the line. The catalytic domains embedding the amino acids of the catalytic triad (*) are marked as II–IV. Saccha- romyces cerevisiae CPY (GI:115901); Trypanosoma cruzei SCP (GI:35181448); mouse Ctsa (GI:84042523); mouse Scpep1 (GI:13436038); mouse vitellogenic-like carboxypeptidase VLCP (GI:187952735). K. Kollmann et al. Functional characterization of lysosomal Scpep1 FEBS Journal 276 (2009) 1356–1369 ª 2009 The Authors Journal compilation ª 2009 FEBS 1365 [...]... 3 010 –30 21 10 Chen J, Streb JW, Maltby KM, Kitchen CM & Miano JM (20 01) Cloning of a novel retinoid-inducible serine carboxypeptidase from vascular smooth muscle cells J Biol Chem 276, 3 417 5–3 418 1 11 Lee TH, Streb JW, Georger MA & Miano JM (2006) Tissue expression of the novel serine carboxypeptidase Scpep1 J Histochem Cytochem 54, 7 01 711 12 Jung G, Ueno H & Hayashi R (19 98) Proton-relay system of carboxypeptidase. .. FEBS Journal 276 (2009) 13 56 13 69 ª 2009 The Authors Journal compilation ª 2009 FEBS 13 67 Functional characterization of lysosomal Scpep1 13 14 15 16 17 18 19 20 21 22 23 24 K Kollmann et al studies on mutagenic replacement of his 397 J Biochem 12 4, 446–450 Rozman J, Stojan J, Kuhelj R, Turk V & Turk B (19 99) Autocatalytic processing of recombinant human procathepsin B is a bimolecular process FEBS... & Sandhoff K (19 82) Lysosomal enzyme precursors in human fibroblasts Activation of cathepsin D precursor in vitro and activity of beta-hexosaminidase A precursor towards ganglioside GM2 Eur J Biochem 12 5, 317 – 3 21 Leake DS & Peters TJ (19 81) Proteolytic degradation of low density lipoproteins by arterial smooth muscle cells: the role of individual cathepsins Biochim Biophys Acta 664, 10 8 11 6 Wendland... Calpha-formylglycine-generating enzyme J Biol Chem 280, 14 900 14 910 Wattiaux R, Wibo M & Baudhuin P (19 63) [Effect of the injection of Triton WR 13 39 on the hepatic lysosomes of the rat] Arch Int Physiol Biochim 71, 14 0 14 2 Gieselmann V, Pohlmann R, Hasilik A & von Figura K (19 83) Biosynthesis and transport of cathepsin D in cultured human fibroblasts J Cell Biol 97, 1 5 Taylor S & Tappel AL (19 73) Lysosomal peptidase... & Fincher GB (19 88) The A- and B-chains of carboxypeptidase I from germinated barley originate from a single precursor polypeptide J Biol Chem 263, 11 106 11 110 Galjart NJ, Gillemans N, Harris A, van der Horst GT, Verheijen FW, Galjaard H & d’Azzo A (19 88) Expression of cDNA encoding the human ‘protective protein’ associated with lysosomal beta-galactosidase and 13 68 25 26 27 28 29 30 31 32 33 34 35... available: Fig S1 Lysosomal localization of endogenous Scpep1 by immunofluorescence Fig S2 Disruption of the Scpep1 gene (A) Insertion of the b-geo gene trap cassette into the Scpep1 gene (B) Southern blot (C) Multiplex-PCR analysis of genomic DNA derived from three F2 mice Fig S3 Multitissue northern blot analysis with RNA derived from wild-type (+ ⁄ +) mouse tissues, Scpep1gt () ⁄ )) tissues and heterozygous... were determined according to [ 41] for CBZ substrates and [42] for FA substrates Cloning, transfection and expression of Scpep1 Generation of the Scpep1-gt mouse model The Scpep1 cDNA was subcloned into the pcDNA3 .1 ⁄ Hygro(+) vector and pCI-neo (Promega GmbH, Mannheim, Germany) by add-on PCR, as described previously [36] HT1080 cells and COS-7 cells were transfected with Fugene6 reagent (Roche GmbH, Mannheim,... Robinson D & Galjaard H (19 82) Molecular defect in combined betagalactosidase and neuraminidase deficiency in man Proc Natl Acad Sci USA 79, 4535–4539 Jung G, Ueno H & Hayashi R (19 99) Carboxypeptidase Y: structural basis for protein sorting and catalytic triad J Biochem 12 6, 1 6 Latchinian-Sadek L & Thomas DY (19 93) Expression, purification, and characterization of the yeast KEX1 gene product, a polypeptide... al (19 95) Mouse model for the lysosomal disorder galactosialidosis and correction of the phenotype with overexpressing erythroid precursor cells Genes Dev 9, 2623–2634 Pohlmann R, Boeker MW & von Figura K (19 95) The two mannose 6-phosphate receptors transport distinct complements of lysosomal proteins J Biol Chem 270, 27 311 –27 318 Deuschl F, Kollmann K, von Figura K & Lubke T ¨ (2006) Molecular characterization. .. function of lysosomal SCs, it could be insightful to crossbreed Scpep1-deficient mice with CtsaS190A mice to investigate the overlap and distinct functions of Scpep1 and Ctsa Purification of Scpep1-His6 from stably expressing HT1080 cells Materials and methods Deglycosylation by PNGase F Cell lines and cell culture Cell lysates of HT1080-Scpep1 were subjected to PNGase F (Roche) treatment as described previously . preset mass range of the MS analysis [29]. 13 16 17 18 19 14 15 α-Ctsa α-Ctsa α-Scpep1 α-Scpep1 F2 WT F2 gt B A 13 14 15 16 17 18 19 0 1 2 3 4 Fraction Spec M 55 35 18 kDa kDa 0801TH - 0801TH siH -1 p epcS 08 01 T H -0801TH s i H -1 p e p cS 0 8 0 1T H -080 1 TH siH-1pepcS A Lane 1 2 3 4 5 6 7 8 9 10 11 12 Fig. 1. Molecular forms of Scpep1. (A)