Functional characterization of the lysosomal membrane protein TMEM192 in mice

Abstract

The Transmembrane protein 192 (TMEM192) is a lysosomal/late endosomal protein initially discovered by organellar proteomics. TMEM192 exhibits four transmembrane segments with cytosolic N- and C-termini and forms homodimers. Devoid of significant homologies, the molecular function of TMEM192 is currently unknown. Upon TMEM192 knockdown in hepatoma cells, a dysregulation of autophagy and increased apoptosis were reported. Here, we aimed to define the physiological role of TMEM192 by analysing consequences of TMEM192 ablation in mice. Therefore, we compared the biochemical properties of murine TMEM192 to those of the human orthologue. We reveal lysosomal residence of murine TMEM192 and demonstrate ubiquitous tissue expression. In brain, TMEM192 expression was pronounced in the hippocampus but also present in the cortex and cerebellum, as analysed based on a lacZ reporter allele. Murine TMEM192 undergoes proteolytic processing in a tissue-specific manner. Thereby, a 17 kDa fragment is generated which was detected in most murine tissues except liver. TMEM192 processing occurs after lysosomal targeting by pH-dependent lysosomal proteases. TMEM192-/- murine embryonic fibroblasts (MEFs) exhibited a regular morphology of endo-/lysosomes and were capable of performing autophagy and lysosomal exocytosis. Histopathological, ultrastructural and biochemical analyses of all major tissues of TMEM192-/- mice demonstrated normal lysosomal functions without apparent lysosomal storage. Furthermore, the abundance of the major immune cells was comparable in TMEM192-/- and wild type mice. Based on this, we conclude that under basal conditions in vivo the loss of TMEM192 can be efficiently compensated by alternative pathways. Further studies will be required to decipher its molecular function.

Keywords: autophagy; lysosomal membrane; lysosome; proteolytic processing; transmembrane protein 192.

Figure 1. Biochemical properties of murine TMEM192

A. Schematic drawing of membrane topology and orientation of murine TMEM192. The position of the epitope used to generate polyclonal antisera for the detection of murine TMEM192 is indicated. B. Western blot analysis of HeLa cells transiently expressing human or murine TMEM192, both fused to an HA epitope, was performed in order to validate functionality and species specificity of the generated TMEM192 antibody. In order to confirm protein expression and equal protein loading the membrane was also stained with antibodies against the HA epitope and β-Tubulin. C. HeLa cells were transiently transfected with murine TMEM192 fused to an HA tag at its C-terminus. In order to validate the newly generated antibodies, the heterologously expressed murine TMEM192 was visualized by indirect immunofluorescence using the polyclonal TMEM192 antibody or the monoclonal antibody 3F10 binding to the HA epitope followed by appropriate fluorophore-conjugated secondary antibodies. As indicated, lysosomal localization was confirmed by detecting the lysosomal marker protein LAMP-2. Scale bar, 10 µm. D. To analyse disulphide-dependent dimerization of murine TMEM192, HeLa cells were transiently transfected with human or murine TMEM192. Aliquots of total lysates were denatured for 5 minutes at 95°C in the presence or absence of DTT and subjected to Western blotting. TMEM192 monomers and dimers were detected with monoclonal antibody 3F10 (HA) and are labelled with open and closed arrow-heads, respectively. Equal loading was confirmed by re-staining the membrane with anti-GAPDH. E.Alignment of the C-termini of human (aa251 - 271) and murine (aa 251 - 266) TMEM192. The cysteine residue C266 (bold, underline) previously identified to mediate dimerization of human TMEM is not present in the murine protein.

Figure 2. Murine TMEM192 is ubiquitously expressed and proteolytically processed

A. Targeting strategy for disruption of the TMEM192 gene. In the initially generated knockout-first allele (tm1a), expression of the TMEM192 gene is disabled by a genetrap cassette which contains a β-Galactosidase reporter that is controlled by the endogenous TMEM192 promotor. This cassette was removed by breeding with Flip deleter mice leaving behind an allele with a loxP-flanked exon 3 (tm1c). Upon breeding with Cre deleter mice, this exon was excised generating a constitutive knockout allele (tm1d, TMEM192-/-). Positions of the three genotyping primers (TMEM192-PostCre-Fw, TMEM192-PostCre-Rv, TMEM192-Ex3-Rv1) that were used to distinguish mice with the tm1d allele and wild type mice are indicated in the scheme. B. Total organ lysates from wild type and TMEM192-/- mice were generated and analysed by Western blotting with antibodies against TMEM192. As a control for equal loading, the membrane was re-probed with anti-GAPDH. In addition to the monomeric TMEM192 full-length protein (FL, closed arrow-head), an N-terminal fragment (NTF, open arrowhead) derived from this protein was detected. *, unspecific band. C. N-glycosylation of endogenous murine TMEM192 was examined in total organ lysates from spleen, thymus, brain and liver from a wild type mouse. Prior to Western blotting, aliquots of lysates were incubated in absence or presence of PNGase F. Detection of TMEM192 was achieved with polyclonal TMEM192 antibody. As a control for glycosidase activity, deglycosylation of the lysosomal protein LIMP-2 was confirmed. D. Organ lysates of wild type mice were analysed by Western blotting for endogenous TMEM192 protein as in (B). Bands of the TMEM192 FL protein as well as of the NTF were quantified densitometrically and the ratio (NTF/FL) was calculated. Data are depicted as mean + SD and are based on the analysis of n = 5 mice.

Figure 3. Proteolytic processing of murine TMEM192 occurs in lysosomes

A. Abundance of the TMEM192 protein (closed arrow-head) and the derived TMEM192 N-terminal fragment (NTF, open arrow-head) was compared in murine embryonic fibroblasts (MEF), RAW 264.7 cells, N2A cells and primary Bone-marrow-derived macrophages from wild type mice (+/+). As a control for antibody specificity, macrophages from TMEM192-/- were included in the analysis. Total cell lysates of the depicted cell lines were subjected to Western blotting using the TMEM192 antibody. Equal loading was confirmed by re-probing the membrane with anti-GAPDH antibody. B. HeLa cells were transiently transfected with murine TMEM192. Cells were treated overnight with Bafilomycin (30 nM), NH₄Cl (25 mM), Leupeptin (100 µM), E64d (40 µM), Brefeldin A (1 µg/ml) or H₂O or DMSO as negative control. Aliquots of total cell lysates were analysed by Western blotting using the TMEM192 antibody. Due to a low processing efficiency of TMEM192 upon overexpression conditions, a longer exposure of the membrane was included to visualize the NTF and the effect of the different inhibitors on the generation of this fragment.

Figure 4. Loss of TMEM192 does not impair lysosomal function in TMEM192-deficient murine embryonic fibroblasts (MEFs)

A. Wild type and TMEM192-/- MEFs were analysed for possible differences in lysosomal morphology by indirect immunofluorescence. Therefore, the luminal lysosomal protease Cathepsin D, the lysosomal membrane protein LAMP-2 and the late endosomal/lysosomal lipid LBPA were detected with specific antibodies and visualized with fluorochrome-conjugated secondary antibodies. Scale bar, 10 µm. B. The specific β-Hexosaminidase activity of total lysates from wild type and TMEM192-/- MEF cells was calculated based on the spectrophotometrically measured enzyme activity and the protein concentration. Results are based on n = 3 independent cell lines per genotype and are depicted as mean + SD. ns non-significant, unpaired, two tailed Student's t test. C. Wild type and TMEM192-/- MEF cells were treated with 10 µM ionomycine or an equivalent volume of DMSO in serum-free DMEM for 10 min at 37°C. Immediately after the incubation the medium was separated from the cells and cells were detached and lysed. The activity of β-Hexosaminidase was measured in the media as well as the cell lysates. The proportion (%) of released enzyme in relation to the total cellular enzyme pool (medium + lysate) was calculated for each sample and is depicted as mean + SD. n = 3; ns non-significant; unpaired, two tailed Student's t test.

Figure 5. Comparable autophagic flux and regulation of mTOR activity in wild type and TMEM192-deficient murine embryonic fibroblasts (MEFs)

A., B. Wild type and TMEM192-/- MEFs were analysed for basal steady-state levels of LC3-I and LC3-II and p62. Based on a densitometric quantification LC3-II/I and p62/GAPDH ratios were calculated and are depicted without any normalisation. Means + SD from n = 7-8 values per genotype are shown. C.-E. The capability to induce autophagy was compared in wild type and TMEM192-/- MEFs. Cells were starved in EBSS for 1, 2 and 4 h or kept in regular medium. To concomitantly inhibit lysosomal proteolysis, bafilomycin A1 (Baf) was added as indicated. Total cell lysates were subjected to Western blotting employing antibodies against LC3 and p62 as depicted from a representative experiment. To confirm equal loading the membrane was re-probed with GAPDH antibodies. A densitometric quantification of n = 3 experiments was performed as shown in (D) and (E). LC3-II/I ratios were determined and normalised to the value of the Control (regular medium, no Bafilomycin) of the respective genotype within each experiment (D). p62 values were corrected for putative differences in protein loading by calculating p62/GAPDH ratios. These values were also normalised to that of untreated cells as described above for the LC3-II/I ratios and are depicted in (E). All values are means + SD from n = 3 independent experiments. F., G. Intact regulation of mTORC1 activity in TMEM192-/- MEF cells. Wild type and TMEM192-/- MEFs were either left untreated (Control) or starved for one hour in EBSS (EBSS pretreated) in order to induce mTORC1 inactivation. Subsequently, mTORC1 reactivation was triggered by re-introducing amino acids (aa) via incubation in DMEM for 60, 90 and 120 min. Treatment with 250 nM Torin 1 during the reactivation period was used to silence mTORC1 kinase activity even in the presence of amino acids. The activity of mTORC1 was assessed based on the phophorylation status of its target p70S6 kinase (S6K) which was analysed by Western blotting as shown from a representative experiment in (F). The position of the band representing pS6K and two unspecific bands are marked with an arrowhead and asterisks, respectively. Densitometric quantification of phosphorylated S6K (pS6K) versus total S6K (pS6K/S6K) from n = 3 independent experiments is depicted in (G).

Figure 6. No obvious neurodegeneration and lysosomal pathology in the brain of TMEM192-deficient mice

A. Genotype distribution among off-spring (n = 203 mice) from heterozygous matings of mice carrying the Post-Cre knockout allele (tm1d). Nomenclature is as follows: +/+, wild type; +/-, heterozygous; -/-, TMEM192 knockout. B. Comparison of body weight of age- and sex-matched wild type and TMEM192-/- mice. Mean + SD, n = 16 per genotype, ns non-significant, unpaired, two tailed Student's t test. C. Distribution of TMEM192 expression in mouse brain. Brain sections from homozygous mice with the tm1a allele (TMEM192A/A) and wild type mice as negative control were subjected to X-gal staining in order to visualize activity of the β-galactosidase reporter which is part of the tm1a allele. Mice were transcardially perfused with 4% PFA and cryosections were prepared. Scale bars: 2000 µm, overview; 500 µm,cortex (CTX) and cerebellum (CBX); 200 µm, hippocampus (HP). D. Immunohistochemical staining of TMEM192 in cortical brain sections from wild type mice. Cryosections from perfusion-fixed brains (4% PFA) were used. Diaminobenzidine was employed for visualization. To control for unspecific labelling, sections from TMEM192-/- brains were stained in parallel. TMEM192-positive vesicular structures are marked with arrow-heads. Scale bar = 250 µm. E. Representative brain sections from wild type and TMEM192-/- mice stained with hematoxylin and eosin demonstrated regular architecture of TMEM192-/- cortex, cerebellum and hippocampus (CA1 region). Staining was performed on paraffin sections from immersion-fixed brains (4% PFA). Scale bar: 100 µm. F. Total brain lysates from wild type and TMEM192-/- mice were generated and analysed by Western blotting with antibodies against TMEM192, myelin basic protein (MBP), glial fibrillary acidic protein, and cathepsin D. As loading control, the membrane was re-detected with anti-GAPDH. In B, D, E and G a two tailed Student's t test was performed. No statistical significant difference between wild type and TMEM192-deficient cells was observed.

Figure 7. Regular liver and spleen architecture without signs of lysosomal dysfunction in TMEM192-deficient mice

A. Representative histological sections from livers of wild type and TMEM192-/- mice. Tissue was fixed by transcardial perfusion with 3% glutaraldehyde. Semi-thin sections were prepared from araldite-embedded tissue and stained with toluidine blue. Scale bar, 50 µm. B. To analyse the ultrastructure of lysosomes in hepatocytes, transmission electron microscopy of ultrathin sections from perfusion-fixed (3% glutaraldehyde) and araldite-embedded wild type and TMEM192-/- livers was performed. BC, bile canaliculus; Ly, lysosome. Scale bar, 1 µm. C. To screen for differences in autophagy activation, total liver lysates of each genotype were analysed by Western blotting for LC3 conversion. The bands representing the two forms LC3-I and LC3-II are indicated. As a control for correct genotypes and protein loading, detection of TMEM192 and GAPDH was performed. D. The activity of the lysosomal enzyme β-Hexosaminidase was measured spectrophotometrically in total lysates of livers from wild type and TMEM192-/- mice. Means + SD of the specific enzyme activities are depicted, which were calculated for each sample based on the respective protein concentrations. Data are based on n = 6; unpaired, two tailed Student's t test, ns non-significant. E. Representative spleen sections from wild type and TMEM192-deficient mice. Paraffin sections from immersion-fixed (4% PFA) spleens were stained with hematoxylin and eosin. Scale bar, 250 µm. F. Western blot analysis of total spleen lysates from wild type and TMEM192-/- mice. The abundance of the lysosomal membrane protein LIMP-2, the autophagy marker LC3, and the lysosomal protease Cathepsin D was compared. In addition, TMEM192 and GAPDH were detected as controls. G. Specific activity of β-Hexosaminidase in total spleen lysates from wild type and TMEM192-/- mice. Data are based on n = 15; unpaired, two tailed Student's t test, ns non-significant.