Until recently, a modest number of approx. 40 lysosomal membrane proteins had been identified and even fewer were characterized in their function. In a proteomic study, using lysosomal membranes from human placenta we identified several candidate lysosomal membrane proteins and proved the lysosomal localization of two of them. In the present study, we demonstrate the lysosomal localization of the mouse orthologue of the human C1orf85 protein, which has been termed kidney-predominant protein NCU-G1 (GenBank® accession number: AB027141). NCU-G1 encodes a 404 amino acid protein with a calculated molecular mass of 39 kDa. The bioinformatics analysis of its amino acid sequence suggests it is a type I transmembrane protein containing a single tyrosine-based consensus lysosomal sorting motif at position 400 within the 12-residue C-terminal tail. Its lysosomal localization was confirmed using immunofluorescence with a C-terminally His-tagged NCU-G1 and the lysosomal marker LAMP-1 (lysosome-associated membrane protein-1) as a reference, and by subcellular fractionation of mouse liver after a tyloxapol-induced density shift of the lysosomal fraction using an anti-NCU-G1 antiserum. In transiently transfected HT1080 and HeLa cells, the His-tagged NCU-G1 was detected in two molecular forms with apparent protein sizes of 70 and 80 kDa, and in mouse liver the endogenous wild-type NCU-G1 was detected as a 75 kDa protein. The remarkable difference between the apparent and the calculated molecular masses of NCU-G1 was shown, by digesting the protein with N-glycosidase F, to be due to an extensive glycosylation. The lysosomal localization was impaired by mutational replacement of an alanine residue for the tyrosine residue within the putative sorting motif.
Lysosomes are membrane-limited organelles of eukaryotic cells which degrade intracellular and extracellular materials they obtain by autophagy and endocytosis respectively. Within the lysosomes, more than 60 acid hydrolases and associated cofactors degrade the imported macromolecules . The products of this breakdown are exported into the cytoplasm where they are metabolized. The metabolite export through the lysosomal membrane is less understood than the macromolecular degradation within the lysosomal matrix. Although, the lysosomal v-ATPase (vacuolar-type H+-ATPase), which is required for the acidification of the matrix, is well-characterized, the number of transporters is less than expected from that of the solutes which must be released from lysosomes [2,3]. The importance of these transporters is underscored by the occurrence of lysosomal storage diseases . Thus a loss of the cystine transporter cystinosin results in cystinosis , loss of the sialic acid transporter sialin manifests as sialic acid storage diseases , and loss of mucolipin manifests as mucolipidosis IV [7–9]. Recently, the lysosomal cobalamin transporter has been identified, which is mutated in the cblF defect of vitamin B12 metabolism . Inherited defects of the prominent LAMP-2 (lysosome-associated membrane protein-2) causes Danon disease which is characterized by a (cardio)myopathy due to an abnormal accumulation of phagosomes and autophagosomes .
The vast majority of soluble lysosomal proteins are transported from the TGN (trans-Golgi network) to endosomes after receiving a M6P (mannose 6-phosphate) tag at their N-glycans that is recognized by specific MPRs (M6P receptors) . As an exception, the lysosomal soluble hydrolase β-glucocerebrosidase is directed to the lysosomes by a piggy-back transport along with LIMP-2 (lysosomal integral membrane protein type-2) .
The lysosomal sorting of most integral lysosomal membrane proteins is mediated by short amino acid sequence motifs in their cytosolic loops and tails. Two types of lysosomal sorting motifs are known. On the one hand, the tyrosine-based sorting signals NPXY and YXXΦ and, on the other hand, the dileucine-based sorting signals [DE]XXXL[LI] and DXXLL [14,15]. These motifs resemble endocytic and TGN sorting signals; however, they differ from the latter by their close proximity (mostly 6–13 amino acids) to the transmembrane segment [16,17]. They are recognized by the heterotetrameric adaptor proteins AP-1 to AP-4, or by the multidomain GGAs (Golgi-localized, gamma-ear-containing, ADP-ribosylation factor-binding proteins) (reviewed in [14,15]).
Previously, several novel lysosomal membrane proteins were identified using different strategies such as subcellular fractionation (p40, ), genetic approaches (TMEM74 , TMEM76 , MFSD8 ) and proteome studies [22,23]. Among these, TMEM76 and MFSD8, have been shown to be associated with the lysosomal diseases mucopolysaccharidosis IIIC [20,22] and a late-infantile onset form of NCL (neuronal ceroid lipofuscinosis) . In a proteomics analysis of lysosomal membranes purified from human placenta we identified 124 lysosomally enriched proteins including 12 novel proteins of unknown function . Two of these, LOC201931 and LOC51622, were shown to localize mainly to the lysosomal compartment using their YFP (yellow fluorescent protein)-tagged versions and immunofluorescence . In the present study, we verify the lysosomal localization of a third candidate called kidney-predominant protein NCU-G1, and show that it represents a highly glycosylated integral membrane protein of the lysosomal compartment.
MATERIALS AND METHODS
Cell lines and cell culture
If not stated otherwise HT1080 cells, HeLa cells and MEFs (mouse embryonic fibroblasts) were grown in complete DMEM (Dulbecco's modified Eagle's medium; Gibco Life Technologies) supplemented with 10% FBS (fetal bovine serum; PAN Biotech), 1% penicillin/streptomycin (Gibco Life Technologies) and 1% L-glutamine (Gibco Life Technologies) at 37 °C with 5% CO2.
Antibodies used for the present study include monoclonal anti-RGS-His4-tag and anti-His5-Alexa Fluor® 488 conjugated (Qiagen), anti-porin HL31 (Calbiochem), anti-Gapdh (glyceraldehyde 3-phosphate dehydrogenase; Santa Cruz Biotechnologies), anti-cathepsin D , anti-human LAMP-2 (H4B4; Developmental Studies Hybridoma Bank), anti-mouse LAMP-1 (1D4B; Developmental Studies Hybridoma Bank) and anti-Scpep1 (serine carboxypeptidase 1)  antisera. HRP (horseradish peroxidase)- and fluorescence-conjugated secondary antibodies were supplied by Dianova and Invitrogen respectively.
A rabbit polyclonal antiserum (anti-NCU-G1) was generated against an internal, putative luminal peptide of mouse NCU-G1 (CPSVNERNSIDDEYAPAVF). The serum was affinity-purified using the immobilized peptide.
Cloning, transfection and expression of NCU-G1-RGS-His5
The NCU-G1 cDNA (GenBank® accession number: BC021547; 1563 bp) was received from the RZPD (Deutsche Ressourcenzentrum für Genomforschung; clone IRAVp968F0636D6) and was subcloned into the pcDNA3.1/Hygro(+) vector (Invitrogen) by add-on PCR using the following NheI-forward primer 5′-gtatgctagcATGTTTCGCTGTTGGGGACCTCAC-3′ (NheI site underlined) and the NotI-RGS-His5-reverse primer 5′-gtatgcggccgcTTATCCGTGATGGTGATGGTGCGATCCTCTTCCGTTTATGGACTGGTACTCAGAATACC-3′ (NotI site underlined) Lower-case letters represent non-translated sequences and upper-case letters represent coding sequences. Finally, all constructs were full-length sequenced to exclude unintentional mutations. HT1080 cells, HeLa cells and MEFs were transfected with FuGENE™ 6 reagent (Roche) according to the manufacturer's protocol.
MEF or HeLa cells were transiently transfected and cultivated for 24–36 h. After fixation with methanol, the cells were analysed for immunofluorescence as described previously  using anti-LAMP-1 (1D4B) and anti-LAMP-2 (H4B4) antibodies. His-tagged NCU-G1 protein was detected with an Alexa Fluor® 488-coupled monoclonal anti-His5 antibody (Qiagen). The LAMP antibodies were detected with Alexa Fluor® 546-conjugated goat anti-rat or goat anti-mouse IgG secondary antibodies (Invitrogen). Immunofluorescence images were obtained using a Leica TCS Sp2 AOBS laser-scanning microscope (Leica Microsystems).
Site-directed mutagenesis of NCU-G1
We inserted the Y400A mutation into the NCU-G1-RGS-His5 construct using the QuikChange® site-directed mutagenesis kit (Stratagene) with 5′-TATTCTGAGGCCCAGTCCATAAAC-3′ (forward primer) and 5′-GTTTATGGACTGGGCCTCAGAATA-3′ (reverse primer), according to the manufacturer's protocol.
Northern blot analysis
Northern blot analysis was performed as described previously , with the exception that total mouse RNA from various tissues was separated. NCU-G1 transcripts were detected with the full-length NCU-G1 cDNA as a probe. The Northern blot membrane was stripped and re-hybridized with the full-length mouse Gapdh cDNA as an RNA loading control.
Membrane extraction assay
HeLa cells were transiently transfected with the His-tagged NCU-G1 construct and harvested by scraping, extracted in lysis buffer [10 mM Tris/HCl (pH 7.5), with a mixture of protease inhibitors] by sonication and finally cleared by centrifugation (10000 g for 10 min at 4°C). The membranes, 200 μg of protein, were subjected to extraction with either 0.1 M sodium carbonate (pH 11.5) or 1% Triton X-100 in lysis buffer, for 30 min on ice. Subsequently, the soluble and solid fractions were separated by centrifugation at 45000 rev./min (Beckman TLA-55 rotor) and 4 °C for 90 min. The presence of different antigens in the fractions was examined by Western blot analysis.
Deglycosylation by PNGase F (peptide N-glycosidase F)
Cleared lysates from NCU-G1 transiently transfected HT1080 cells were subjected to PNGase F (Roche) treatment as described previously .
Lysosome isolation (tritosomes)
Mice were treated by a single injection of 0.75 mg of tyloxapol (Triton WR1339; Sigma) per g of body weight 4 days prior to liver withdrawal. The isolation of tritosomes from tyloxapol-treated and control mice included differential centrifugation and isopycnic centrifugation and resulted in a lysosome-enriched fraction F2 as described previously [29,30]. All experimental procedures on animals were carried out according to protocols that were approved by the faculty and by the appropriate authorities.
For the continuous sucrose gradient, the ‘light mitochondria’ fraction (L) as described by de Duve et al.  was layered on to a pre-formed 15–30% continuous sucrose gradient and centrifuged at 33000 rev./min for 150 min in a Beckman SW41Ti rotor. Fractions (1 ml each) were collected from the top and analysed by Western blotting. In addition, protein and the activity of β-hexosaminidase were determined using photometric assays .
NCU-G1 was analysed bioinformatically using ExPASy (http://au.expasy.org), SignalP prediction server (; http://www.cbs.dtu.dk/services/SignalP-2.0), PSORT II prediction server (http://psort.nibb.ac.jp/form2.html), ELM server (http://elm.eu.org) and TMHMM  prediction service (http://www.cbs.dtu.dk/services/TMHMM).
Standard techniques were used for Western blotting to PVDF membranes. Antibody–protein complexes were visualized by enhanced chemiluminescence (Pierce).
Bioinformatics analysis of mouse NCU-G1
Human C1orf85 protein, gi15079485, was identified in a comprehensive proteomics analysis of lysosomal integral and associated membrane proteins from human placenta . Its mouse orthologue NCU-G1 is encoded in chromosome 3 and its cDNA is composed of six exons. The mRNA (GenBank® accession number: BC021547) is at least 1563 nt long and consists of a 26 nt 5′-UTR (untranslated region), a 1215 nt coding sequence and a 3′-UTR. The deduced amino acid sequence of the NCU-G1 protein consists of 404 residues with a molecular mass of approx. 40 kDa. The bioinformatics analysis predicts a hydrophobic N-teminal signal peptide (residues 1–35) followed by the main part of the molecule (residues 36–369) containing nine putative N-glycosylation sites with the consensus sequence NxS/T, a single transmembrane segment (residues 370–392) and a C-terminal tail (residues 393–404) as illustrated in Figure 1(A). The C-terminal residue is preceded by a potential tyrosine-based (YxxΦ) lysosomal targeting signal with YQSI (amino acids 400–403). These data suggest that NCU-G1 is a type I integral membrane protein with a large luminal N-terminal domain, a single transmembrane segment and a short cytoplasmic tail (Figure 1B). The protein described here does not show any significant sequence similarity to other proteins except in the C-terminal tail, which resembles those in LAMP-1 and -2 proteins. The tails are short, contain several basic residues next to the transmembrane segment and a single tyrosine-based localization motif at or near their C-terminus.
Predicted structure of NCU-G1
Cloning and expression of NCU-G1
The NCU-G1 cDNA was received in the pCMV-SPORT6 vector from the RZPD and was expressed as a C-terminal RGS-His-tagged derivative after subcloning by add-on PCR into the pcDNA3.1/Hygro(+) vector (see the Material and methods section). Expression of the C-terminal His-tagged NCU-G1 in various cell types such as HT1080 cells (HT1080-NCU-G1) or HeLa cells, and analysis of the cell lysates by Western blotting using an anti-His-antibody resulted in two bands with apparent molecular masses of approx. 70 and 80 kDa respectively (Figure 2A, left-hand panel). For further characterization of the endogenous NCU-G1 protein, we raised a polyclonal antiserum in rabbit against the peptide CPSVNERNSIDDEYAPAVF (residues 245–263) that detected two polypeptides at molecular masses comparable with the anti-His-tag antibody (Figure 2A, right-hand panel).
Molecular characterization of NCU-G1
N-glycosylation of NCU-G1
Incubation of HT1080 (Figure 2B, lanes 1 and 2) and HT1080-NCU-G1 cell lysates (Figure 2B, lanes 3 and 4) with PNGase F and analysis by Western blotting resulted in a major decrease in the apparent molecular mass by approx. 40% (Figure 2B). In the treated sample, two bands were observed at 37 and 44 kDa (Figure 2B, lane 4) as compared with the expected value of 43.8 kDa. This result suggests that the existence of two glycosylated molecular forms of NCU-G1 in HT1080 cells is due to differences in the polypeptide, rather than in the carbohydrate, moieties. Deglycosylation of endogenous NCU-G1 in a lysosome-enriched fraction (F2) that was derived from a tyloxapol-treated mouse liver resulted a similar shift from a 75 kDa glycosylated form of NCU-G1 (Figure 2B, lane 5) to 40 kDa (Figure 2B, lane 6).
NCU-G1 is an integral membrane protein
In order to demonstrate that mouse NCU-G1 is an integral membrane protein, we performed an extraction assay. Homogenates from NCU-G1 transfected and untransfected HeLa cells (50 μg of protein) were incubated for 30 min on ice with homogenization buffer as a control, 0.1 M sodium carbonate (pH 11.5) or 1% Triton X-100. After the incubation, supernatant and pellet fractions were separated by ultracentrifugation at 45000 rev./min (Beckman TLA-55 rotor) and analysed by immunoblotting for NCU-G1–His with the anti-His antibody and the mitochondrial integral membrane protein porin 31HL using a porin-specific monoclonal antibody (Figure 2C). In the Western blot analysis, NCU-G1–His was almost completely recovered in the pellet fraction after incubation in the homogenization buffer. However, NCU-G1–His was partially solubilized in the presence of sodium carbonate and largely solubilized in the presence of Triton X-100, thus resembling the integral membrane protein porin 31HL (Figure 2C).
Tissue distribution of NCU-G1
The transcription pattern of NCU-G1 was analysed by using a mouse MTN (multi-tissue Northern) blot containing 10 μg of total RNA derived from various tissues. The MTN was hybridized with the 1215 bp full-length cDNA probe that identified a single NCU-G1 transcript of 1.5 kb widely expressed in most mouse tissues (Figure 3). As expected, NCU-G1 (also termed kidney-predominant protein) expression was highest in the kidney. The signal was significantly less in samples from liver, brain, intestine, testis, spleen, heart and lung. Interestingly, no NCU-G1 transcript could be detected in skeletal muscle.
Expression pattern of NCU-G1
The Western blot analyses of various tissues with the NCU-G1 antiserum that was raised against the peptide from amino acids 245–263 were hampered by a number of probably non-specific signals (results not shown). As shown previously, NCU-G1 was detectable in a lysosomally enriched fraction from mouse liver with an apparent size of 75 kDa (Figure 2B, lane 5).
Subcellular localization of NCU-G1–His by immunofluorescence
The anti-NCU-G1 antiserum was not suitable for immunofluorescence since many structures in addition to lysosomes were stained. Therefore we confirmed the lysosomal localization of NCU-G1 by immunofluorescence with the C-terminal His-tagged NCU-G1 derivative in MEFs and HeLa cells (results not shown). For visualization we used a monoclonal anti-His5 antibody that was directly coupled with the chromophore Alexa Fluor® 488. NCU-G1 presented in punctuated vesicular structures throughout the cell. These were shown to represent lysosomes and late endosomes by co-staining with the lysosomal membrane protein LAMP-1 (Figure 4A).
Subcellular localization of NCU-G1 and NCU-G1-Y400A
Assuming NCU-G1 is a type I integral membrane protein, its topology suggests a single putative tyrosine-based lysosomal sorting motif YQSI in close proximity of the C-terminus and the transmembrane segment. In order to analyse the role of this putative motif we replaced the tyrosyl residue in position 400 with an alanyl residue by site-directed mutagenesis of the His-tagged NCU-G1 construct (NCU-G1-Y400A). The corresponding expression vector was introduced into HeLa cells and the subcellular localization of NCU-G1-Y400A was analysed by immunofluorescence using the Alexa Fluor® 488-coupled anti-His antibody, and was compared with that of the lysosomal marker LAMP-2. An extensive co-localization was observed with the NCU-G1 wild-type construct. However, the NCU-G1-Y400A mutant was found in a diffuse peripheral network resembling the ER (endoplasmic reticulum; Figure 4B) and rarely in the LAMP-2-positive organelles.
Subcellular localization of endogenous NCU-G1 using density centrifugation
Tyloxapol (Triton WR1339) is a non-ionic detergent that, upon intravenous injection, inhibits the lipoprotein lipase and thus increases serum lipoprotein levels such as VLDL (very-low-density lipoprotein). Endocytosis of these lipoproteins and the detergent into hepatocytes causes a decrease in the buoyant density of the lysosomes. Upon fractionation in a density gradient the lysosomes of this cell population are shifted towards the top fraction and the resulting lysosome-enriched fraction (F2) is virtually devoid of other organelles, mitochondria in particular. We prepared a ‘light mitochondria fraction’ (L) from either tyloxapol-treated or non-treated mouse liver and applied it to a discontinuous sucrose gradient. After the separation we analysed the resulting fractions F1–F4 (50 μg of protein each) by immunoblotting (Figure 4D). In the control liver, in fractions of the sucrose gradient, very small amounts of endogenous NCU-G1 and LAMP-1 were detected (Figure 4D, left-hand panel). Both proteins were detected in fraction F4 which corresponds to the load of the gradient with the L fraction, although their signals are weaker because mitochondrial proteins represent a major part of fraction F4 and hence dilute the lysosomal proteins. In contrast, in the tyloxapol-treated liver, NCU-G1 (75 kDa), as well as LAMP-1, shifted into fraction F2 due to the selectively altered density of the lysosomes (Figure 4D, right-hand panel). The tyloxapol-induced shift of both NCU-G1 and LAMP-1 is considered a specific proof for the lysosomal localization of the former.
Furthermore, we loaded the L fraction from tyloxapol-treated and control mice on to a preformed continuous sucrose gradient ranging from 15–30% and observed a similar tyloxapol-induced shift specific to lysosomes. In the control mouse, NCU-G1, LAMP-1 and the soluble lysosomal protein Scpep1  were collected in the pellet of the sucrose gradient (Figure 5A). After tyloxapol treatment all three proteins, as well as the lysosomal marker β-hexosaminidase, entered the gradient (Figures 5B and 5C). In the control, 6% of the total activity of the marker was found in fractions 7–10, whereas, in tyloxapol-treated samples this proportion was increased to 47% of the total. These results, in combination with the immunofluorescence data, suggest clearly that NCU-G1 is a lysosomal or endosomal protein.
Tyloxapol-induced density shift of lysosomes in a continuous sucrose gradient
In a recent proteome analysis of the lysosomal membrane from human placenta, we identified 58 known lysosomal membrane proteins including 17 polypeptides of the v-ATPase and most interestingly twelve novel proteins of so far unknown function that were highly enriched in the lysosomal membrane fraction . For two of these twelve candidates (LOC201931 and LOC51622), we were able to confirm their lysosomal localization with the help of YFP-tagged variants of the proteins . In the present study, we validate the lysosomal localization and provide first insights into the molecular features of the third candidate derived from that particular subproteome analysis called NCU-G1 (C1orf85) in mouse. Using a proteomics approach, the human NCU-G1 orthologue (RGSV2553) was identified as a candidate lysosomal membrane protein. Although the sequence coverage has reached merely 20% of the 406 amino acid residues of the human precursor form, all of the recovered peptides were detected in similar enrichments as the LAMP-1 and -2 proteins . An inspection of the sequence indicated that most of the remaining peptides of NCU-G1 were not suited for detection either because of size limits or due to glycosylation.
NCU-G1 is a lysosomal protein
Using immunofluorescence and two different density-centrifugation techniques, we show that mouse NCU-G1 co-localizes and/or co-fractionates with the lysosomal marker proteins LAMP-1, Scpep1 and β-hexosaminidase. This conclusion is supported by the finding that NCU-G1 is subject to the tyloxapol (Triton WR1339)-induced density shift with a selective decrease in the buoyant density of lysosomes, which has been described by others [30,34]. Previously, other putative lysosomal proteins were confirmed to be of lysosomal origin by comparable density-shift techniques [18,35]. The present results, in conjunction with the enrichment of human NCU-G1 in a lysosomal membrane fraction derived from human placenta , strongly support the lysosomal localization of NCU-G1.
Our observation that a mutant of the single putative lysosomal-sorting motif (Y400QSI) of NCU-G1 only marginally co-localized with the lysosomal marker LAMP-1 indicates that this signal is sufficient to mediate transport towards the lysosomal compartment. Typically, the canonical tyrosine-based sorting signals (YxxΦ) are N-terminally flanked by amino acids with a small side chain, mostly a glycine residue as known in, e.g. the LAMP proteins (−1; −2A–C; −3), cystinosin or the acid phosphatase precursor . However, the mucin-like type I membrane protein endolyn exhibits a tyrosine-based sorting signal that is preceded by an asparagine residue instead of glycine . The NCU-G1 sorting signal is a novel variant of the YxxΦ-type since it has a glutamic acid residue in position −1. It should be of interest whether the 12 amino acid NCU-G1 cytoplasmic tail could mediate lysosomal transport of other type I transmembrane proteins such as has been shown for the 10 amino acid tail of endolyn .
NCU-G1 is widely expressed and is a highly glycosylated integral membrane protein
The bioinformatics analysis of mouse NCU-G1 predicted a 40 kDa type I transmembrane protein with a 35 residue N-terminal signal peptide, a large luminal domain with nine putative N-glycosylation sites and a single C-terminal transmembrane segment encompassing residues 370–392 and preceding a short cytosolic tail. Initially, the NCU-G1 cDNA was identified in a PCR-based approach to clone a potential prototype of the complement protein C3 from the mouse embryonic carcinoma cell line F9 . Although NCU-G1 cDNA was the only PCR product, it showed little homology with mouse C3. Northern blot analyses with a NCU-G1-specific probe revealed a single 1.5-kb transcript . The highest expression levels were detected in kidney so that NCU-G1 was also termed synonymously kidney-predominant protein 1 . Our Northern blot analysis resembles those observations regarding the ubiquitous expression of NCU-G1 with the exception that no transcript was detected in skeletal muscles. At protein level, our NCU-G1 antiserum hardly detected NCU-G1, even in mouse kidney homogenates in which high amounts of NCU-G1 transcripts were shown. However, endogenous NCU-G1 was detectable in a lysosome-enriched fraction from mouse liver, although at an unexpected molecular range of approx. 75 kDa. The expression of NCU-G1 in different cell lines came up with comparable apparent molecular masses of approx. 70–80 kDa. Our deglycosylation assays using PNGase F clearly demonstrate that the remarkable difference between the calculated and apparent size of NCU-G1 was due to an extensive glycosylation, suggesting that most of its nine N-glycosylation sites are used. This observation is not surprising since many lysosomal membrane proteins, including the type I LAMP and the LIMP proteins , as well as multipass transmembrane proteins NPC1 , cystinosin  and others, show significantly higher apparent molecular masses than calculated from their primary sequence. The function of the extensive glycosylation of many lysosomal proteins and of lysosomal membrane proteins is assigned to a protective function in two respects: it had been estimated that particularly the high amount of LAMP proteins in the lysosomal membrane generate a glycocalyx that lines the limiting membrane . In addition, N-glycans are known to protect lysosomal membrane proteins such as LAMP-1 and -2 from proteolysis .
We also demonstrated using a membrane extraction assay that C-terminally His-tagged NCU-G1 from HeLa cells is most likely an integral membrane protein, although a significant amount of NCU-G1 was already solubilized from the lysosomal membrane upon carbonate treatment. It is worth mentioning that NCU-G1 in a lysosome-enriched fraction F2 from mouse liver showed similar extraction behaviour and was also partially solubilized by sodium carbonate (results not shown).
The present results are in contrast with a recently published characterization in which human NCU-G1 was reported to function as a transcription factor and as a nuclear receptor activator for the PPAR (peroxisome-proliferator-activated receptor)-α . The authors of that study expressed the human NCU-G1 in Drosophila Schneider S2 cells and detected a protein with an apparent molecular mass of approx. 45 kDa using a polyclonal antiserum against a peptide representing the last fifteen amino acids. The authors did not identify the N-terminal signal sequence nor a putative transmembrane segment but allocated two nuclear export signals and several potential SH3 (Src homology 3)-interacting domains to the NCU-G1 protein. The authors also showed using chloramphenicol acetyl-transferase assays, that NCU-G1 might bind to and activate transcription from a CRBPI (cellular retinol-binding protein type I) promotor . However, immunofluorescence studies with a C-terminal GFP (green fluorescent protein)-tagged version of NCU-G1 and an RFP (red fluorescent protein)-tagged version of PPAR showed negligible co-localization of these proteins. Whereas the nuclear protein PPAR was exclusively sorted to the nucleus, NCU-G1 largely localized to extranuclear compartments and this was attributed to a possible interference by the GFP moiety .
In the present study, we demonstrate that NCU-G1 is a highly glycosylated, integral membrane protein of the lysosomal compartment. Although, the function of NCU-G1 is unknown and the protein does not exhibit homology with known proteins, it may be predicted that NCU-G1 fulfils a function similar to other type I lysosomal membrane proteins such as LAMP-1 and -2. Furthermore, NCU-G1 could be involved in the transport of lysosomal hydrolases as described for the LIMP-2 that binds and directs the soluble hydrolase β-glucocerebrosidase towards the lysosome in a M6P-independent manner  or in forming larger complexes such as has been described for nicastrin in the γ-secretase complex . Unfortunately, our attempts to establish a stably NCU-G1-expressing cell line were not successful, although we tested the His-tagged NCU-G1 in several cell lines such as HeLa, HT1080 and HEK (human embryonic kidney) cells (results not shown). Initially, a few cells survived the hygromycin selection protocol; however, the subsequent loss of the cells suggested that a NCU-G1 gain-of-function somehow interfered with cell survival. It is well-described that the release of lysosomal proteases, e.g. cathepsins, is one prerequisite for apoptosis [45,46] and recently it was demonstrated that the lysosomal protease CLN2 plays a critical role in TNF (tumour necrosis factor)-induced Bid-cleavage .
The physiological role of NCU-G1 remains to be characterized. Working towards this goal we are generating a NCU-G1-deficient mouse model and are examining the expression of the luminal domain of NCU-G1 and searching for interacting proteins using immunoprecipitation and affinity chromatography.
Note added in proof (received 29 June 2009)
After this paper was accepted, a report has been published by Sardiello et al.  where, among other things, it was shown that the human orthologue of NCU-G1, C1orf85, is part of a gene network that regulates lysosomal viogenesis and function, and that C1orf85 localizes to lysosomes.
glyceraldehyde 3-phosphate dehydrogenase
green fluorescent protein
lysosome-associated membrane protein
lysosomal integral membrane protein
mouse embryonic fibroblast
- PNGase F
peptide N-glycosidase F
Deutsche Ressourcenzentrum für Genomforschung
serine carboxypeptidase 1
yellow fluorescent protein
Oliver Schieweck, Markus Damme, Bernd Schröder and Torben Lübke performed the research. Andrej Hasilik, Bernhard Schmidt and Torben Lübke designed the research and analysed the results. All authors were involved in the preparation of the manuscript and approved the final manuscript.
We thank Ellen Eckermann-Felkl, Nicole Eiselt and Klaus Neifer for excellent technical assistance as well as Karthikeyan Radhakrishnan and Sebastian Krol for their help and discussions.
This work was supported by theDeutsche Forschungsgemeinschaft [grant number LU1173/1-4 (to T. L.)]; and by the BMBF (Federal Ministry of Education and Research) [grant number 031U204A0].