Lysosomal TMEM106B interacts with galactosylceramidase to regulate myelin lipid metabolism.
Free full text in Europe PMC
Abstract
Free full text
Lysosomal TMEM106B interacts with galactosylceramidase to regulate myelin lipid metabolism
Abstract
TMEM106B is an endolysosomal transmembrane protein not only associated with multiple neurological disorders including frontotemporal dementia, Alzheimer’s disease, and hypomyelinating leukodystrophy but also potentially involved in COVID-19. Additionally, recent studies have identified amyloid fibrils of C-terminal TMEM106B in both aged healthy and neurodegenerative brains. However, so far little is known about physiological functions of TMEM106B in the endolysosome and how TMEM106B is involved in a wide range of human conditions at molecular levels. Here, we performed lipidomic analysis of the brain of TMEM106B-deficient mice. We found that TMEM106B deficiency significantly decreases levels of two major classes of myelin lipids, galactosylceramide and its sulfated derivative sulfatide. Subsequent co-immunoprecipitation assay showed that TMEM106B physically interacts with galactosylceramidase. We also found that galactosylceramidase activity was significantly increased in TMEM106B-deficient brains. Thus, our results suggest that TMEM106B interacts with galactosylceramidase to regulate myelin lipid metabolism and have implications for TMEM106B-associated diseases.
Introduction
TMEM106B, which encodes TMEM106B, a highly glycosylated type II transmembrane protein, was initially identified as a risk modifier of frontotemporal lobar degeneration (FTLD) especially in patients with mutations in the GRN (FTLD-GRN)1–3. Subsequently, TMEM106B has also been linked to many other neurodegenerative diseases, such as Alzheimer’s disease (AD)4–6, chronic traumatic encephalopathy7, and limbic-predominant age-related TDP-43 encephalopathy8. In addition, a de novo mutation in TMEM106B was found in patients with hypomyelinating leukodystrophy (HLD), a neurodevelopmental disorder characterized by myelin deficits9. Outside the brain, previous studies using an in vivo genetic screen have identified TMEM106B as a primary robust driver of lung cancer metastasis10,11. Furthermore, three different studies using genome-wide CRISPR screens independently identified TMEM106B as a host factor in certain cell lines for SARS-CoV-2, a strain of coronavirus that caused coronavirus disease 2019 (COVID-19)12–14.
While TMEM106B has been shown to be associated with a great variety of human conditions, little is known about molecular mechanism by which TMEM106B is involved in those conditions. TMEM106B is localized at the endolysosome and thought to regulate several aspects of lysosomal function, but so far, a limited number of the interacting proteins have been reported1,2. Previous studies using cryo-electron microscopy have shown that truncated C-terminal TMEM106B forms amyloid fibrils in the brain15–19. However, the TMEM106B fibrils have been found not only in the brains of people with neurodegenerative diseases but also in neurologically normal ageing brains. Additionally, a disease-protective variant of TMEM106B also forms the amyloid fibrils. Although an increase in C-terminal TMEM106B fragment was biochemically detected in the sarkosyl-insoluble fraction of the brain of TMEM106B FTLD-risk SNP carriers, such an increase has not been observed in histological analysis20–22. It thus remains unclear whether the TMEM106B amyloid fibrils are involved in and responsible for the progression of TMEM106B-associated neurological disorders.
Dysregulation of lipid metabolism has been reported in the brains of FTLD-GRN patients as well as mouse models of FTLD-GRN23–26. Given the role of TMEM106B in FTLD-GRN, these findings raise a possibility that TMEM106B also regulates lipid metabolism under physiological and/or pathological conditions. Here, we performed lipidomic analysis of the brain of TMEM106B-deficient mice. We found that levels of two major classes of myelin lipids galactosylceramide (GalCer) and sulfatide (ST) are significantly reduced in TMEM106B-deficient brains. Subsequent unbiased proteomics and co-immunoprecipitation (co-IP) assay showed that TMEM106B physically interacts with galactosylceramidase (GALC), a lysosomal enzyme that hydrolyzes GalCer. We also found that GALC activity was significantly increased in TMEM106B-deficient brains. These results suggest that TMEM106B interacts with GALC to regulate myelin lipid metabolism particularly GalCer and ST levels. Thus, GALC inhibition may be a therapeutic target for TMEM106B-associated diseases.
Results
Levels of GalCer and ST are decreased in TMEM106B-deficient brains
Given the potential roles of TMEM106B in regulating lipid metabolism, we performed multidimensional mass spectrometry-based shotgun lipidomic analysis27 using the brain of 12-month-old TMEM106B-deficient mice and wild-type (WT) littermates (n=
6, cohort 1). Lipid classes quantified in this discovery analysis include phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), phosphatidylglycerol (PG), phosphatidylinositol (PI), phosphatidic acid (PA), cardiolipin (CL), lyso-phosphatidylcholine (LPC), lyso-phosphatidylethanolamine (LPE), lyso-cardiolipin (LCL), sphingomyelin (SM), ceramide (Cer), sulfatide (ST), hexosylceramide (HexCer), free cholesterol (CHL), cholesterol esters (CE), bis(monoacylglycerol)phosphate (BMP), and the total number of lipid species analyzed is 294 (Fig. 1a). Strikingly, with p
<
0.05 cut off, we found that levels of multiple species of HexCer and ST were significantly decreased in TMEM106B-deficient brains (Fig. 1b, c), although TMEM106B deficiency affected levels of several species of other lipid classes such as PA, PE, and PC as well (Fig. 1c). Importantly, total levels of ST, HexCer, and PA but not the other classes were also found to be decreased in TMEM106B-deficient brains (Fig. 1d).
a Diagram showing the experimental procedures of lipidomic analysis using the brain of 12-month-old WT and Tmem106b−/− mice. First, targeted shotgun lipidomics were performed using 6 mice per genotype (cohort 1) to analyze 17 general lipid classes; phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), phosphatidylglycerol (PG), phosphatidylinositol (PI), phosphatidic acid (PA), cardiolipin (CL), lyso-phosphatidylcholine (LPC), lyso-phosphatidylethanolamine (LPE), lyso-cardiolipin (LCL), sphingomyelin (SM), ceramide (Cer), sulfatide (ST), hexosylceramide (HexCer), free cholesterol (CHL), cholesterol esters (CE), bis(monoacylglycerol)phosphate (BMP). The total number of lipid species analyzed in this study is 294. Second, ST and HexCer levels were analyzed using another cohort with 7 mice per genotype (cohort 2) to test replicability. Finally, SFC-MS/MS was performed using cohort 1 to analyze galactosylceramide (GalCer) and glucosylceramide (GlcCer) species separately. The diagram was created with BioRender.com. b Volcano plot of discovery analysis with cohort 1 showing all 294 molecular lipid species. n=
6 (3 males and 3 females) per genotype. HexCer and ST species (p
<
0.05) are highlighted in red and orange, respectively. The horizontal line indicates p
=
0.05. p Values were calculated using two-tailed unpaired t-test. c Heatmap of all lipid species significantly affected by TMEM106B deficiency at p
<
0.05 cutoff. d Scatter plots of the sum of all species within a lipid class for each class in WT versus Tmem106b−/−. Mean
±
SEM, n
=
6 (3 males and 3 females) per genotype. *p
=
0.0411 (PA), *p
=
0.0311 (ST), *p
=
0.0279 (HexCer); two-tailed unpaired t-test.
In our discovery analysis quantifying 294 lipid species, q-values were more than 0.05 even in the top differentially regulated lipid species (Supplementary Data 1). Therefore, to exclude the possibility that differences we observed in the discovery analysis are type I errors, we performed a replication analysis focusing on ST and HexCer using another cohort of 12-month-old TMEM106B-deficient mice and their WT littermates (n=
7, cohort 2). In the replication analysis, we observed a significant decrease in levels of most ST and HexCer species in TMEM106B-deficient brains (Fig. 2a, b), consistent with the discovery analysis. Total levels of ST and HexCer were also found to be significantly decreased in the cohort 2 of TMEM106B-deficient brains. Combining the cohorts 1 and 2, the p-values in differences between WT and TMEM106B-deficient brains in total ST and HexCer levels were less than 0.0001. Similar results were obtained between males and females (Fig. 2c, d). We thus concluded that a decrease in ST and HexCer is reproducibly observed irrespective of sex in TMEM106B-deficient brains.
a Levels of ST species in the brain of WT and Tmem106b−/− mice (cohort 2) at 12 months of age. Mean±
SEM, n
=
7 mice (5 males and 2 females) per genotype. p-Values were determined using two-tailed unpaired t-test and shown on the right side of the graph when p
<
0.05. b Levels of HexCer species in the brain of WT and Tmem106b−/− mice (cohort 2) at 12 months of age. Mean
±
SEM, n
=
7 mice (5 males and 2 females) per genotype. p-Values were determined using two-tailed unpaired t-test and shown on the right side of the graph when p
<
0.05. c Total ST levels (sum of all species) in the brain of WT and Tmem106b−/− mice at 12 months of age. Mean
±
SEM. Cohorts 1 and 2 are from the discovery and replication analyses, respectively. p-Values were determined using two-tailed unpaired t-test and were shown on the top of the graphs. d Total HexCer levels (sum of all species) in the brain of WT and Tmem106b−/− mice at 12 months of age. Mean
±
SEM. Cohorts 1 and 2 are from the discovery and replication analyses, respectively. p-Values were determined using two-tailed unpaired t-test and were shown on the top of the graphs.
HexCer comprises galactosylceramide (GalCer) and glucosylceramide (GlcCer). To determine levels of GalCer and GlcCer species separately, we next performed supercritical fluid chromatography-tandem mass spectrometry (SFC-MS/MS) using the brain homogenates from the cohort 1 and found that levels of multiple species of GalCer were significantly decreased in TMEM106B-deficient brains while none of GlcCer species was significantly altered (Figs. 1a and 3a, b). TMEM106B deficiency also significantly decreased total levels of GalCer but not GlcCer in the brain (Fig. 3c, d). Given that TMEM106B is an HLD gene, a decrease in GalCer and its sulfated derivative ST in TMEM106B-deficient brains was of particular interest as GalCer and ST are two major classes of myelin lipids28. Note that TMEM106B deficiency however had no effects on levels of Cer and SM (Fig. 1d), revealing a specific effect of TMEM106B deficiency on GalCer and ST levels. Together, our lipidomic data suggest that TMEM106B deficiency affects myelin lipid metabolism by decreasing levels of GalCer and ST in the brain.
a Levels of GalCer species in the brain of WT and Tmem106b−/− mice at 12 months of age. Mean±
SEM, n
=
6 mice (3 males and 3 females) per genotype. *p
=
0.0396 (C16-GalCer), *p
=
0.015 (C20-GalCer), **p
=
0.0034 (C20:1-GalCer), **p
=
0.0073 (C22-GalCer), **p
=
0.005 (C22:1-GalCer), *p
=
0.0423 (C24-GalCer), *p
=
0.0203 (C24:1-GalCer), *p
=
0.0285 (C25-GalCer), *p
=
0.0298 (C26-GalCer); Two-tailed unpaired t-test. b Levels of GlcCer species in the brain of WT and Tmem106b−/− mice at 12 months of age. Mean
±
SEM, n
=
6 mice (3 males and 3 females) per genotype. None of the species were significantly affected by two-tailed unpaired t-test. c Total GalCer levels (sum of all species) in the brain of WT and Tmem106b−/− mice at 12 months of age. Mean
±
SEM, n
=
6 mice (3 males and 3 females) per genotype. *p
=
0.0188; two-tailed unpaired t-test. d Total GlcCer levels (sum of all species) in the brain of WT and Tmem106b−/− mice at 12 months of age. Mean
±
SEM, n
=
6 mice (3 males and 3 females) per genotype. Two genotypes were not significantly different by two-tailed unpaired t-test.
TMEM106B physically interacts with GALC
To examine the mechanisms by which TMEM106B deficiency affects myelin lipid metabolism, we next sought to identify protein interaction partners of TMEM106B in the brain by using immunoprecipitation (IP) and liquid chromatography tandem mass spectrometry (LC-MS/MS) procedure (Fig. 4a). We confirmed that our anti-TMEM106B antibody cross-linked to protein A Sepharose beads reproducibly immunoprecipitated endogenous TMEM106B from mouse brains (Fig. 4b). Consistent with previous studies29–31, we observed a ~70kDa immunoreactive band of TMEM106B likely due to its glycosylation and homodimerization when the samples were run under non-reducing conditions (without 2-mercaptoethanol and heat). The ~70
kDa band was detected by two different anti-TMEM106B antibodies (Supplementary Fig. 1a).
a Diagram showing experimental procedures of IP and LC-MS/MS analysis using mouse brains. The diagram was created with BioRender.com. b Representative blots to validate immunoprecipitation of endogenous TMEM106B from mouse brains using anti-TMEM106B antibody linked to protein A Sepharose. Note that all samples were run without treating with 2-mercaptoethanol and heat. c Venn diagrams to identify potential lysosomal TMEM106B-binding proteins. Total 848 proteins were identified in immunoprecipitates from WT brains but 268 of them were also found in those from TMEM106B-deficient brains. Thus, 580 proteins were considered as potential TMEM106B-binding proteins. Among them, 22 proteins were lysosomal proteins based on The Mouse Lysosomal Gene Database (http://lysosome.unipg.it/mouse.php). d Biosynthesis and degradation of GalCer and ST. SMase sphingomyelinase, SMS sphingomyelin synthase, CGT cerebroside galactosyl transferase, CST cerebroside sulfotransferase, ASA arylsulfatase A. TMEM106B deficiency results in a decrease in levels of GalCer and ST (blue) while having no effects on levels of Cer and SM (green). e Schematic drawing of full-length (FL) TMEM106B-GFP and TMEM106B-GFP lacking amino acids (aa)123–275 (ΔC) and aa6–94 (ΔN). f Representative blots of co-IP assays using HEK293T cells expressing GFP, FL TMEM106B-GFP, ΔC TMEM106B-GFP, or ΔN TMEM106B-GFP, together with myc-DDK-tagged mouse GALC. Note that all samples were treated with 2-mercaptoethanol and heat (95°C) before running a gel. Monomeric TMEM106B-GFP fusion protein was detected at ~70
kDa (~43
kDa glycosylated TMEM106B + ~27
kDa GFP). g Quantification of co-IP in (f). Mean
±
SEM, n
=
3–4 experiments. p-Values were determined using one-way ANOVA with Tukey’s multiple comparisons test and were shown on the top of the graph.
LC-MS/MS analysis of elution of anti-TMEM106B immunoprecipitates identified 22 potential lysosomal TMEM106B-binding proteins. Of importance, some of them were previously reported to interact with TMEM106B, such as adapter protein subunits, vacuolar-ATPase V0 domain subunits, and cathepsin D29,31,32, which validates the specificity of our IP and LC-MS/MS analysis (Fig. 4c). We also confirmed interaction between TMEM106B and cathepsin D using co-IP using HEK293T cells transfected with TMEM106B and cathepsin D (Supplementary Fig. 1b, c). Interestingly, one of the TMEM106B-binding proteins identified was galactosylceramidase (GALC), an enzyme that hydrolyzes GalCer (Fig. 4c, d). To confirm the results of MS analysis, we first performed co-IP assay using HEK293T cells transiently expressing myc-tagged GALC and green fluorescent protein (GFP), full-length (FL) TMEM106B-GFP, ΔC-TMEM106B-GFP, or ΔN-TMEM106B-GFP (Fig. 4e). We found that myc-tagged GALC was significantly co-immunoprecipitated with FL TMEM106B-GFP or ΔN-TMEM106B-GFP, but not GFP or ΔC-TMEM106B-GFP (Fig. 4e–g). Immunofluorescent analysis showed that ΔC-TMEM106B-GFP was localized at LAMP1-positive lysosomes in HEK293T cells, indicating that a decrease in co-immunoprecipitation of GALC by ΔC-TMEM106B-GFP is not due to loss of lysosomal localization by deletion of the C-terminal region (Supplementary Fig. 2). These results suggest that TMEM106B interacts with GALC in the lysosomal lumen. We also detected co-immunoprecipitation of myc-tagged GALC with immunoprecipitated endogenous TMEM106B in HEK293T cells, excluding the requirement for overexpression of two proteins or TMEM106B-GFP fusion protein (Supplementary Fig. 1d, e). T185S and D252N mutations in TMEM106B are associated with human FTLD and HLD, respectively1. These mutations however have no significant effects on co-IP of GALC (Supplementary Fig. 3).
TMEM106B regulates GALC activity
To further examine whether TMEM106B regulates GALC activity, we performed GALC activity assay using several brain regions of WT and TMEM106B-deficient mice. GALC activity can be measured using the fluorogenic substrate 4-methylumbelliferyl β-D-galactopyranoside (MUGAL) in the presence of AgNO3, which specifically inhibits β-galactosidase (β-Gal)33. We found that GALC activity was significantly increased in the forebrain and brainstem but not cerebellum of TMEM106B-deficient mice (Fig. 5a), although TMEM106B expression was not markedly different between the brain regions and was slightly higher in the cerebellum of WT mouse brains (Supplementary Fig 4). We also performed GALC activity assay using 6-hexadecanoylamino-4-methylumbelliferyl-β-D-galactopyranoside (HMGal), a specific substrate for GALC34, and obtained a similar result using the same forebrain lysates (Fig. 5b, c).
a Enzymatic activity of forebrain, brainstem, and cerebellar extracts of WT and Tmem106b−/− mice at 13 months of age toward MUGAL substrate in the absent or presence of 12.5μM AgNO3. n
=
6 mice (2 males and 4 females) per genotype. *p
=
0.0137 (forebrain, 0
μM), **p
=
0.0016 (forebrain, 12.5
μM), **p
=
0.0035 (brainstem, 0
μM), *p
=
0.0383 (brainstem, 12.5
μM); two-tailed unpaired t-test. b GALC activity of forebrain extracts of WT and Tmem106b−/− mice at 13 months of age measured using HMGal. n
=
6 mice (2 males and 4 females) per genotype. **p
=
0.0017; two-tailed unpaired t-test. c Correlation between the results of GALC activity assay using MUGAL and HMGal. d GALC activity of digitonin-soluble brainstem lysates of WT and Tmem106b−/− mice at 13 months of age measured with HMGal. n
=
6 mice (2 males and 4 females) per genotype. **p
=
0.0022; two-tailed unpaired t-test. e Representative blots of the digitonin-soluble brainstem lysates using anti-GALC and anti-GAPDH antibodies. The asterisks indicate non-specific bands. Whole brain lysates from ~40-day-old wild type and twitcher (GALC-deficient) mice were used as positive and negative controls (PC and NC). f Quantification of the blots in (e). Mean
±
SEM, n
=
6 mice (2 males and 4 females) per genotype. Two genotypes were not significantly different by two-tailed unpaired t-test.
To examine whether TMEM106B deficiency affects protein levels of GALC, we homogenized brainstems from WT and TMEM106B-deficient mice using Tris-buffered saline with 0.4% digitonin and performed immunoblot analysis using an anti-mouse GALC antibody (CL1021AP)35. Increased GALC activity was confirmed in the digitonin-soluble brainstem lysates from TMEM106B-deficient mice (Fig. 5d). In the immunoblot analysis, we observed a mature GALC band at ~50kDa in the brainstem lysates but not in the negative control brain lysate from the twitcher mouse, which does not express GALC protein due to the presence of homozygous p.W339X genotype36. However, TMEM106B deficiency had no significant effects on the level of the mature GALC (Fig. 5e, f). Together, these results suggest that TMEM106B regulates enzymatic activity of GALC but not protein levels of mature GALC in the forebrain and brainstem.
Discussion
The major finding of the present study is that endolysosomal protein TMEM106B regulates myelin lipid metabolism. Our shotgun lipidomics and subsequent SFC-MS/MS revealed that levels of two major classes of myelin lipids GalCer and ST, which together constitute ~27% of the myelin lipid28, were significantly decreased in the brain of TMEM106B-deficient mice. Our results could potentially explain how a mutation in lysosomal TMEM106B leads to HLD. Importantly, dysfunction of myelin and/or myelin lipid metabolism has also been reported in other TMEM106B-associated brain disorders such as FTLD and AD. A recent study has shown disrupted myelin lipid metabolism in FTLD-GRN patients25. White matter abnormality has long been known in AD37–42. In addition, several studies have reported a significant reduction in ST levels in preclinical AD cases43–46. Interestingly, a previous rodent study has revealed that adult-onset ST deficiency is sufficient to cause AD-like neuroinflammation and cognitive impairment47. ST has also been shown to bind with high affinity to TREM2, a microglial AD risk factor48. These studies raise an intriguing possibility that the TMEM106B variant may be involved in the brain disorders by altering levels of GalCer and ST. Outside the brain, TMEM106B has been identified as a driver of lung cancer metastasis10,11 and as a proviral host factor for SARS-CoV-2, a coronavirus that caused the COVID19 pandemic12–14. It will be interesting to examine whether a reduction in GalCer or ST is also observed in other TMEM106B-deficient tissues and protects against the lung cancer metastasis or SARS-CoV-2 infection.
In addition to GalCer and ST, TMEM106B deficiency has been shown to affect levels of other lipids such as PA and several species of PC and PE in our lipidomic analysis. How TMEM106B deficiency affects other lipids and whether those lipids are involved in TMEM106B-associated diseases remain unclear. Several previous studies including ours have found that PGRN deficiency results in alteration in the levels of BMP, glucosylshingosine (GlcSph), GlcCer, and/or gangliosides in human and/or mouse models of FTLD-GRN23,24,49. However, these lipids were not significantly altered in TMEM106B-deficient brains. On the other hand, our previous lipidomic analysis has also revealed a decrease in some species of ST in PGRN-deficient brains49, similar to the results of TMEM106B-deficient brains. Alteration in ST levels has been also reported in carriers of TMEM106B disease risk allele21 as well as FTD-GRN cases25. Further work will be required to better understand our lipidomic results and link them to previous findings and ultimately to understand how TMEM106B is involved in multiple brain disorders including FTD-GRN.
Our unbiased IP and LC-MS/MS analysis identified GALC, an enzyme that hydrolyzes GalCer, as an interaction partner for TMEM106B. We also found that TMEM106B deficiency causes an increase in enzymatic activity of GALC while having no significant effect on its protein levels. It thus appears that TMEM106B binds and regulates activity of GALC in the lysosomal lumen, which explains reduced levels of both GalCer and ST in TMEM106B-deficient brains as ST is produced by sulfation of GalCer through the action of cerebroside sulfotransferase50. Future studies will have to determine whether TMEM106B directly associates with GALC as it is possible that the interaction observed in the co-IP is mediated by other proteins in the same physical complex. Although the T185 variant and D252N mutation in TMEM106B have been associated with FTLD and HLD, respectively1, our co-IP assay showed no significant effects of the variant and mutation on TMEM106B binding to GALC. Whether the negative result is due to our assay system overexpressing both TMEM106B and GALC or whether the genetic variant and mutation influence activity of GALC rather than TMEM106B binding to GALC is currently unclear and requires further investigation. Recent cryo-electron microscopy studies found amyloid fibrils composed of C-terminal TMEM106B in human brains15–18. Our results suggest that the C-terminal fragment of TMEM106B could also interact with GALC. Effects of the C-terminal fragment on GALC activity in the brain will be an important avenue for future studies.
In the present study, we observed increased GALC activity in the forebrain and brainstem but not in the cerebellum of TMEM106B-deficient mice. The mechanism of the brain region-selective effect of TMEM106B on GALC activity is currently unclear. TMEM106B protein levels did not appear to be markedly different between these brain regions and are rather slightly increased in the cerebellum. Thus, lack of the effect from TMEM106B loss in the cerebellum is not simply explained by lower expression levels of TMEM106B in the cerebellum. Further investigation will be required to elucidate the exact mechanism.
In conclusion, accumulating evidence clearly suggests that lysosomal TMEM106B plays a key role in a wide range of human conditions. Our study revealed a physiological function of TMEM106B regulating myelin lipid metabolism by physically interacting with GALC. These findings provide mechanistic insights into TMEM106B-associated disorders and offer potential therapeutic strategies targeting TMEM106B-GALC interaction or modulating GalCer/ST levels.
Methods
Mouse strains
All animal protocols were approved by the Yale Institutional Animal Care and Use Committee. We have complied with all relevant ethical regulations for animal use. TMEM106B-deficient mice on C57BL/6 background were described previously29. This Tmem106b−/− line (Tmem106btm2a(KOMP)Wtsi) was generated by LacZ gene trap strategy and expresses 5–10% residual full-length TMEM106B protein51,52. All mice were housed in groups of 2–5 animals per cage with ad libitum access to standard mouse chow and water. Lighting was maintained at 12:12 light: dark cycle. Both male and female were used in this study and the numbers and ages are described in Figure legends.
Mouse brain tissue preparation
Mice were euthanized using CO2. After transcardiac perfusion using ice-cold DPBS, the brains were removed, and the hemispheres were divided and stored at −80°C until use. For GALC activity assay, forebrain and hindbrain were separated and the hindbrain was further dissected into cerebellum and brainstem. The dissected brain regions were snap-frozen in liquid nitrogen and stored at −80
°C until use.
Lipidomic analysis
Lipid species were analyzed using multidimensional mass spectrometry-based shotgun lipidomic analysis as previously described49,53. In brief, the brain hemispheres from 12-month-old TMEM106B-deficient mice and WT littermates were homogenized in 0.1× PBS and the brain homogenate containing 1.0mg of protein, which was determined with a Pierce™ BCA protein assay kit (ThermoFisher SCIENTIFIC #23225), was transferred to a disposable glass culture test tube. A premixture of lipid internal standards was added prior to conducting lipid extraction for quantification of the targeted lipid species. The following internal standards purchased from Avanti Polar Lipids were used: 26.25 nmol of PC(14:1/14:1), 23.58
nmol of PE(16:1/16:1), 19.724
nmol of PS(14:0/14:0), 3.08
nmol of PG(15:0/15:0), 0.482
nmol of PA(14:0/14:0), 1.331
nmol of CL(14:0/14:0/14:0/14:0), 1.232 nmol of LPC(17:0), 1.257
nmol of LPE(14:0), 2.732
nmol of SM(d18:1/12:0), 1.053
nmol of Cer(d18:1/17:0), 3.08
nmol of ST(d18:0/16:0), 8.008
nmol of GalCer(d18:1/15:0), 169.33
nmol of d7-Cholesterol, 0.247
nmol of CE(18:1(d7)). PG(15:0/15:0) was also used for BMP and PI. CL(14:0/14:0/14:0/14:0) was also used for LCL. Lipid extraction was performed using a modified Bligh and Dyer procedure53, and each lipid extract was reconstituted in chloroform:methanol (1:1, v:v) at a volume of 400
µL/mg protein. Phosphatidylethanolamine (PE) and cholesterol (CHL) were derivatized as described previously54,55 before lipidomic analysis. Quantitative analysis of isomeric bis(monoacylglycerol)phosphate (BMP) and phosphatidylglycerol (PG) was performed as reported56.
For shotgun lipidomics, lipid extract was further diluted to a final concentration of ~500=fmol total lipids per µL. Mass spectrometric analysis was performed on a triple quadrupole mass spectrometer (TSQ Altis, Thermo Fisher Scientific, San Jose, CA) and a Q Exactive mass spectrometer (Thermo Scientific, San Jose, CA), both of which were equipped with an automated nanospray device (TriVersa NanoMate, Advion Bioscience Ltd., Ithaca, NY) as previously described57. The triple quadrupole and orbitrap systems were run separately for mutual validation and identification. Identification and quantification of lipid species were performed using an automated software program as previously reported58,59. Data processing (e.g., ion peak selection, baseline correction, data transfer, peak intensity comparison and quantitation) was performed as previously described59. The result was normalized to the protein content (nmol lipid/mg protein or pmol lipid/mg protein for CE). Total levels of each lipid class were determined by adding up the levels of all species of the class. Full results of lipidomic results are provided as Supplementary Data files.
GlcCer and GalCer species analysis separated by SFC-MS/MS
Separation of GlcCer and GalCer species was performed at the Lipidomics Shared Resource at Medical University of South Carolina. Lipids were extracted from the 0.1× PBS homogenate containing 1mg of protein, and levels of GlcCer and GalCer species were measured with supercritical fluid chromatography-tandem mass spectrometry (SFC-MS/MS) analysis. The equipment consists of a Waters UPC2 system coupled to a Thermo Scientific Quantum Access Max triple quadrupole mass spectrometer equipped with an ESI probe operating in the multiple reaction monitoring (MRM) positive ion mode tuned and optimized for the Waters UPC2 system.
Chromatographic separations are obtained utilizing a carbon dioxide gas and methanol with 1mM ammonium formate in 0.2% formic acid. Prior to analysis, samples undergo an ethyl acetate/isopropanol liquid-liquid extraction as previously described60–62. The quantitative analyses of GalCer and GlcCer are based on eight-point calibration curves generated for each target analyte as previously described60–62. The synthetic standards along with a set of internal standards are spiked into an artificial matrix; they are then subjected to an identical extraction procedure as the biological samples. These extracted standards are then analyzed by the SFC-MS/MS system. Peaks for the target analytes and internal standards are recorded and processed using the instrument’s software. Plotting the analyte/internal standard peak area ratios against analyte concentrations generates the analyte specific calibration curves. Any glycosphingolipids for which no standards are available are quantitated using the calibration curve of its closest counterpart. Levels of GalCer and GlcCer species were expressed as picomoles per milligrams of protein. Total levels of GalCer or GlcCer were determined by adding up the levels of all species of GalCer or GlcCer. Full results of GalCer and GlcCer species analyses by SFC-MS/MS are provided as Supplementary Data files.
Immunoprecipitation
Immunoprecipitation using mouse brain lysates was performed as previously reported52. WT and Tmem106b−/− brains at 6–7-months of age were homogenized in ice-cold TBST (50mM Tris-HCl, pH 7.4, 150
mM NaCl, 1% Triton X-100) supplemented with cOmplete-mini (Roche). After ultracentrifugation at 100,000 × g, the supernatant was pre-cleared using Protein A-Sepharose CL-4B (SIGMA #17-0780-01) for 3
h at 4
°C. Then, anti-TMEM106B antibody (Abcam #ab140185) covalently linked to Protein A-Sepharose CL-4B with BS3 (ThermoScientific #21580) was added to the pre-cleared lysates. After overnight incubation at 4
°C, the immunoprecipitates were washed six times with ice-cold TBST and proteins were eluted with 2× Laemmli buffer (Bio-Rad).
In solution protein digestion
A methanol-chloroform precipitation was performed according to standard protocols. The protein pellet was resuspended in 20µL 8
M urea, 0.4
M ammonium bicarbonate with vigorous vortexing. Proteins were reduced by the addition of 2.0
µL 45
mM dithiothreitol (ThermoFisher SCIENTIFIC #20290) and incubation at 37
°C for 30
min, and then alkylated with the addition of 2.0
µL 100
mM iodoacetamide (SIGMA #I1149) and incubation in the dark at room temperature for 30
min. The urea was diluted to 2
M by adding 55
µL of water. Samples were then enzymatically digested using 1.0
µL 0.5
mg/ml trypsin (Promega Seq. Grade Mod. Trypsin, # V5113) and incubation at 37
°C for 16
h. Digestion was halted by acidification with trifluoroacetic acid (TFA) to 0.1%. The mixtures were desalted using C18 Ultra microspin columns (The Nest Group, #SUM SS18V) following the manufacturer’s directions. Peptides were eluted with 2
×
160
µL 0.1% TFA, 80% acetonitrile, then speedvaced dry. Peptides were dissolved in 32
µL MS loading buffer (2% aceotonitrile, 0.2% trifluoroacetic acid). A nanodrop measurement (Thermo Scientific Nanodrop 2000 UV–Vis Spectrophotometer) determined protein concentrations (A260/A280). Each sample was then further diluted with MS loading buffer to 0.04
µg/µL, with 200
ng (5
µL) injected for LC–MS/MS analysis.
LC–MS/MS on the Thermo Scientific Q Exactive Plus
LC–MS/MS analysis was performed on a Thermo Scientific Q Exactive Plus equipped with a Waters nanoAcquity UPLC system utilizing a binary solvent system (A: 100% water, 0.1% formic acid; B: 100% acetonitrile, 0.1% formic acid). Trapping was performed at 5µL/min, 97% Buffer A for 3
min using a Waters Symmetry® C18 180
µm
×
20
mm trap column. Peptides were separated using an ACQUITY UPLC PST (BEH) C18 nanoACQUITY Column 1.7
µm, 75
µm
×
250
mm (37
°C) and eluted at 300
nL/min with the following gradient: 3% buffer B at initial conditions; 5% B at 1
min; 35% B at 90
min; 50% B at 105
min; 90% B at 110
min, 90% B at 115
min; return to initial conditions at 116
min. MS was acquired in profile mode over the 300-1,700
m/z range using 1 microscan, 70,000 resolution, AGC target of 3E6, and a maximum injection time of 45
ms. Data dependent MS/MS were acquired in centroid mode on the top 20 precursors per MS scan using 1 microscan, 17,500 resolution, AGC target of 1E5, maximum injection time of 100
ms, and an isolation window of 1.7
m/z. Precursors were fragmented by HCD activation with a collision energy of 28%. MS/MS were collected on species with an intensity threshold of 2E4, charge states 2-6, and peptide match preferred. Dynamic exclusion was set to 20
s.
Peptide identification
Data was analyzed using Proteome Discoverer (version 1.3) software and searched in-house using the Mascot algorithm (version 2.6.0) (Matrix Science). The data was searched against a SwissProtein database with taxonomy restricted to Mus musculus. Search parameters used were trypsin digestion with up to 2 missed cleavages; peptide mass tolerance of 10 ppm; MS/MS fragment tolerance of 0.02Da; and variable modifications of methionine oxidation and carbamidomethyl cysteine. Normal and decoy database searches were run, with the confidence level was set to 95% (p
<
0.05).
Co-immunoprecipitation (co-IP) using HEK293T cells
Co-IP was performed as previously reported29,49 with slight modifications. Plasmids encoding myc-DDK-tagged mouse GALC (#MR225749) and human GALC (#RC211578), mouse cathepsin D (#MR225301) were purchased from OriGene Technologies, Inc. Plasmids encoding human TMEM106B-mCherry was generated using pmCherry-N1 vector (Clontech #632523). T185S and D252N mutations were introduced using Quik change II XL site-directed mutagenesis kit (Agilent #200521). Two days after transfection using Lipofectamine™ 2000 (ThermoFisher SCIENTIFIC #11668019), HEK293T cells were harvested and lysed with ice-cold 50mM Tris-HCl pH 7.5, 150
mM NaCl, 1% Triton X-100 supplemented with cOmplete-mini and PhosSTOP (Roche). After ultracentrifugation at 100,000×g for 30
min at 4
°C, the supernatant was incubated with ChromoTek GFP- or RFP-trap® Agarose (proteintech #gta or #rta) for 3
h at 4
°C. For immunoprecipitation of endogenous TMEM106B, anti-TMEM106B (Cell Signaling Technology #93334) or Rabbit (DA1E) mAb IgG XP(R) Isotype Control (Cell Signaling Technology #3900) prebound to Protein A-Sepharose CL-4B was added to the supernatant and the samples were nutated for 5
h at 4
°C. The immunoprecipitates were washed five times with ice-cold lysis buffer and boiled with 2× Laemmli Buffer (Bio-Rad) with βME.
Immunoblot
Immunoblot was performed as previously reported29,49,63, with slight modification. For the brain samples (except for Fig. 5e and Supplementary Fig. 1a), the Triton X-100-soluble fraction was mixed with 2× Laemmli Sample Buffer (Bio-Rad #1610737) without adding βME and kept on ice until running SDS-PAGE to detect ~70kDa putative TMEM106B dimers. The protein samples were resolved by SDS-PAGE with precast 4–20% Tris-glycine gels (Bio-Rad #5671095) and transferred using an iBlot 2 Transfer Device onto nitrocellulose membranes (Invitrogen #IB23001). The membranes were incubated in blocking buffer (Rockland MB-070) for 1
h at room temperature and then incubated overnight at 4
°C in blocking buffer with primary antibodies: TMEM106B (Abcam #ab140185, 1:1000), TMEM106B (E7H7Z) (Cell Signaling Technology #93334, 1:1000), GFP (B-2) (Santa Cruz #sc-9996, 1:500), Myc-Tag (71D10) (Cell Signaling Technology #2278, 1:1000), Myc-Tag (9B11) (Cell Signaling Technology #2276, 1:1000), β-actin (8H10D10) (Cell Signaling Technology #3700, 1:2000), RFP Tag (RF5R) (ThermoFisher SCIENTIFIC #MA5-15257, 1:1000), β-tubulin (D3U1W) (Cell Signaling Technology #86298, 1:1000). The membranes were then washed three times with TBST for 3
min and incubated in secondary antibodies (Li-Cor, IR Dye 680 or 800, all 1:10,000) for 1
h at room temperature. After washing three times with TBST for 3
min, The immunoreactive protein bands were visualized with an Odyssey Infrared imaging system (Li-Cor) and quantified using Fiji (ImageJ) software.
Immunofluorescence
HEK293T cells were plated onto 8-well chamber slides (LAB-TEK, #154941) using DMEM (GIBCO #11965-092) supplemented with 10% fetal bovine serum (GIBCO #16000-044) and Pen Strep (GIBCO #15140-122) and cultured for one day before transfection. Two days after transfection using Lipofectamine™ 2000 (ThermoFisher SCIENTIFIC #11668019), transfected cells were fixed with 4% paraformaldehyde (SIGMA #158127) in DPBS for 1h. After washing with DPBS, the cells were permeabilized and blocked with 10% normal donkey serum, 0.2% Triton X-100 in DPBS for 30
min and then incubated overnight at 4
°C with primary antibodies diluted in 1% normal donkey serum and 0.2% Triton X-100 in DPBS. Primary antibodies used were rabbit anti-LAMP1 antibody (Cell Signaling Technology #9091, 1:75) and mouse anti-GFP antibody (SantaCruz #sc-9996, 1:200). The cells were then washed three times with DPBS and incubated for 1.5
h at room temperature in Alexa Fluor 488 donkey anti-mouse IgG (H
+
L) (Invitrogen #A21202, 1:500) and Alexa Fluor 568 donkey anti-rabbit IgG (H
+
L) (Invitrogen #A10042, 1:500) in 1% normal donkey serum, 0.2% Triton X-100 in DPBS. After washing three times with DPBS, the chamber slides were coverslipped with VECTASHIELD antifade mounting medium with DAPI (VECTOR laboratories, #H-1200). Images were taken using LSM800 confocal microscopy (Zeiss) with 63× magnification objective lens. The maximal intensity projection function was used with z-stack confocal images.
GALC activity assay
GALC activity assay was performed as previously reported with modifications33. Briefly, frozen dissected brain tissues (forebrain, brainstem, and cerebellum) were homogenized using a dounce homogenizer in citrate/phosphate (CP) buffer pH 5.2, 1% Triton X-100, supplemented with cOmplete-mini and phosSTOP (Roche) on ice and incubated for 30min at 4
°C. The supernatants after ultracentrifugation at 100,000 × g were used for GALC activity assay. Protein concentrations were determined by PierceTM BCA protein assay kit (ThermoFisher SCIENTIFIC #23225) and used for normalization. Five micro litters of the supernatants were mixed and preincubated with 10
μL of H2O (vehicle) or AgNO3 (SIGMA #204390) (final concentration 12.5
μM) in CP buffer pH 4 in 96-well black plate (Costar #3916) for 10
min at room temperature (RT) and then added 30
μL of 4-methylumbelliferyl β-D-galactopyranoside (MUGAL) (final concentration 0.8
mM) (SIGMA #M1633) and incubated for 1
h at 37
°C. Reaction was stopped by adding 200
μL of 0.5
M glycine-NaOH pH 10.5. Measurements were taken with VICTOR Nivo® multimode plate reader (PerkinElmer) at excitation of 355
nm and emission of 460
nm.
Brain lysate preparation for HMGal-GALC assay and immunoblot
Mouse brainstem was homogenized in cold TDI buffer (TBS, 0.4% digitonin, protease inhibitors cocktail, pH 7.4) as a 20% (w/v) homogenate followed by centrifugation at 16,000 × g for 1h at 4
°C. The supernatant was then separated from the detergent-insoluble fraction and analyzed for protein concentration by BCA protein assay.
HMGal-GALC activity assay
GALC activity in the brainstem supernatant was measured by a fluorescence substrate turnover assay as previously reported34,64 with modifications. Fifty micrograms of total protein were added to a reaction cocktail containing CP buffer pH 4.5, 7mg/ml sodium taurocholate, 2
mg/ml oleic acid and 0.5
mg/ml 6-hexadecanoylamino-4-methylumbelliferyl-β-D-galactopyranoside (HMGal; Biosynth Inc., Newbury, UK) and incubated for 2
h at 37
°C. The reaction was stopped by adding 2 volumes of 0.1
M glycine/0.1
M sodium hydroxide solution and 4 volumes of absolute ethanol. The concentration of reaction product was determined by fluorescence plate reader at excitation of 385
nm and emission of 450
nm. A 4-methylumbelliferone standard curve was ran in parallel for calculation of absolute GALC activity expressed in nmol per hour per milligram protein.
Immunoblot for GALC
Fifty micrograms of total protein from the brainstem supernatant were mixed with Tris-glycine sample buffer and reducing reagent followed by heating at 85°C for 2
min. The samples were resolved by SDS-PAGE with 10–20% Tris-glycine gel (Thermofisher Scientific, Waltham, MA). Gel proteins were transferred to PVDF membranes by Criterion blotter (Bio-Rad, Hercules, CA) according to manufacturer’s instruction. The protein blot was blocked in 5% non-fat milk for 2
h at room temperature followed by incubation with an anti-mouse GALC antibody (CL1021AP, 250
ng/ml)35 for 18
h at 4
°C. The blot was probed with an anti-chicken IgY horseradish peroxidase-conjugated antibody (Millipore Sigma, Burlington, MA; 500
ng/ml) for 1
h at room temperature, followed by signal development using the Immobilon Western Chemiluminescent HRP substrate (MilliporeSigma, Burlington, MA). Western blot images were obtained by the Image-quant LAS 4000 mini system. To obtain the GAPDH signals, the GALC blots were washed extensively in TBST buffer (TBS, 0.05% tween-20, pH 7.4) then reprobed with an anti-GAPDH antibody (Proteintech #60004-1-Ig, 1:100,000) for 30
min at room temperature. After incubation with an anti-mouse IgG horseradish peroxidase-conjugated antibody (Millipore Sigma, 500
ng/ml) for 30
min at room temperature, the signal images were obtained by the same procedures as above. GALC protein levels were quantified by densitometric measurement of target band intensity normalized to GAPDH levels.
Statistics and reproducibility
Two-tailed unpaired t test (for 2 groups) and one-way ANOVA (for >2 groups) were performed using GraphPad Prism (version 9.2.0). All data are shown as mean±
SEM and specific n values are reported in each Figure legend. Data are considered to be statistically significant if p
<
0.05. For lipidomic analysis, by using two independent mouse cohorts (i.e., discovery (n
=
6) and replication (n
=
7) cohorts), we confirmed that a decrease in HexCer and ST is consistently observed in TMEM106B-deficient brains. All lipidomic analyses were performed by investigators who were blinded to the genotypes. All co-IP experiments were independently repeated at least three time. All attempts at replication were successful. GALC activity of WT and TMEM106B-deficient brains was measured by two different assays (MUGAL and HMGal) and investigators and similar results were obtained.
Supplementary information
Acknowledgements
We thank Kristin DeLuca for assistance with mouse husbandry. We also thank Dr. Janet Deane of University of Cambridge for helpful discussion. This work was supported by National Institute on Aging of the National Institutes of Health under grant number R01AG034924 and R01AG066165 to S.M.S. We also thank the MS & Proteomics Resource at Yale University for providing the necessary mass spectrometers and the accompany biotechnology tools funded in part by the Yale School of Medicine and by the Office of The Director, National Institutes of Health (S10OD02365101A1, S10OD019967, and S10OD018034). This work was supported in part by the Lipidomics Shared Resource, Hollings Cancer Center, Medical University of South Carolina (P30 CA138313 and P30 GM103339). This work was also supported by a grant from National Institute on Aging (RF1 AG061729) to X.H. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Author contributions
H.T. and S.M.S. planned and oversaw all aspects of the study. H.T. performed IP using the mouse brain, subsequent co-IP, and GALC activity assay using MUGAL and analyzed the results of most of the experiments in this study. A.P.-C. performed mouse breeding and prepared all brain samples used in lipidomics. C.W.L. performed GALC activity assay using HMGal and immunoblot using anti-GALC antibody. H.Y. and X.H. performed discovery and replication lipidomic analyses. H.T. and S.M.S. wrote the paper with input and revisions from all authors.
Peer review
Peer review information
Communications Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editor: Manuel Breuer.
Data availability
Uncropped blot images are provided as Supplementary Fig. 5 in the Supplementary Information. LC–MS/MS raw data for Fig. 3 have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD043192. Full results of the discovery, replication, and HexCer analyses of lipidomics are provided in the Supplementary Data 1–3, respectively. Source data are provided in Supplementary Data 4.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
The online version contains supplementary material available at 10.1038/s42003-024-06810-5.
References
Articles from Communications Biology are provided here courtesy of Nature Publishing Group
Citations & impact
Impact metrics
Alternative metrics
Discover the attention surrounding your research
https://www.altmetric.com/details/167055539
Article citations
Deficiency of galactosyl-ceramidase in adult oligodendrocytes worsens disease severity during chronic experimental allergic encephalomyelitis.
Mol Ther, 32(9):3163-3176, 26 Jun 2024
Cited by: 0 articles | PMID: 38937968
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
A role of the frontotemporal lobar degeneration risk factor TMEM106B in myelination.
Brain, 143(7):2255-2271, 01 Jul 2020
Cited by: 29 articles | PMID: 32572497 | PMCID: PMC7363491
The major TMEM106B dementia risk allele affects TMEM106B protein levels, fibril formation, and myelin lipid homeostasis in the ageing human hippocampus.
Mol Neurodegener, 18(1):63, 19 Sep 2023
Cited by: 11 articles | PMID: 37726834 | PMCID: PMC10510131
Identification of TMEM106B amyloid fibrils provides an updated view of TMEM106B biology in health and disease.
Acta Neuropathol, 144(5):807-819, 02 Sep 2022
Cited by: 14 articles | PMID: 36056242 | PMCID: PMC9547799
Review
Free full text in Europe PMC
Funding
Funders who supported this work.
NIA NIH HHS (3)
Grant ID: R01 AG066165
Grant ID: R01 AG034924
Grant ID: P30 AG066508
U.S. Department of Health & Human Services | National Institutes of Health (NIH) (1)
Grant ID: R01AG066165
U.S. Department of Health & Human Services | National Institutes of Health (1)
Grant ID: R01AG066165