Decreased turnover of the CNS myelin protein Opalin in a mouse model of hereditary spastic paraplegia 35

Abstract

Spastic paraplegia 35 (SPG35) (OMIM: 612319) or fatty acid hydroxylase-associated neurodegeneration (FAHN) is caused by deficiency of fatty acid 2-hydroxylase (FA2H). This enzyme synthesizes sphingolipids containing 2-hydroxylated fatty acids, which are particularly abundant in myelin. Fa2h-deficient (Fa2h^−/−) mice develop symptoms reminiscent of the human disease and therefore serve as animal model of SPG35. In order to understand further the pathogenesis of SPG35, we compared the proteome of purified CNS myelin isolated from wild type and Fa2h^−/− mice at different time points of disease progression using tandem mass tag labeling. Data analysis with a focus on myelin membrane proteins revealed a significant increase of the oligodendrocytic myelin paranodal and inner loop protein (Opalin) in Fa2h^−/− mice, whereas the concentration of other major myelin proteins was not significantly changed. Western blot analysis revealed an almost 6-fold increase of Opalin in myelin of Fa2h^−/− mice aged 21–23 months. A concurrent unaltered Opalin gene expression suggested a decreased turnover of the Opalin protein in Fa2h^−/− mice. Supporting this hypothesis, Opalin protein half-life was reduced significantly when expressed in CHO cells synthesizing 2-hydroxylated sulfatide, compared to cells synthesizing only non-hydroxylated sulfatide. Degradation of Opalin was inhibited by inhibitors of lysosomal degradation but unaffected by proteasome inhibitors. Taken together, these results reveal a new function of 2-hydroxylated sphingolipids namely affecting the turnover of a myelin membrane protein. This may play a role in the pathogenesis of SPG35.

Introduction

Hereditary spastic paraplegias (HSP) are characterized by degeneration of long axons of the CNS, in particular those of the corticospinal tracts (1). Mutations in more than 80 genes have been identified as a cause for HSP (2–4). According to Harding (5), HSP are divided into pure forms, displaying primarily weakness of lower limbs and spasticity, and complicated forms showing various additional neurological symptoms. Spastic paraplegia 35 (SPG35) (OMIM: 612319), which is caused by deficiency in fatty acid 2-hydroxylase (FA2H), is a complicated HSP associated with leukodystrophy (also known as fatty acid hydroxylase-associated neurodegeneration, FAHN) (6–9). More than 50 mutations in the FA2H gene have been identified in SPG35 cases (see overview in reference 10). FAHN is also regarded as a subtype of neurodegeneration with brain iron accumulation (NBIA) (9). Except for the CNS pathology, a unique feature of SPG35 is also an abnormal structure of the hair shaft (10).

FA2H is essential for synthesis of 2-hydroxylated sphingolipids (HFA-SL) in the nervous system (11–16). HFA-SL [primarily 2-hydroxylated galactosylceramide (HFA-GalC) and 2-hydroxylated sulfatide (HFA-sulfatide)] are abundant lipids in CNS and PNS myelin (17,18). Fa2h-deficient (Fa2h^−/−) mice develop late-onset axonal degeneration and myelin loss with behavioral symptoms reminiscent of spastic paraplegia and serve as an animal model for SPG35/FAHN (15,16). HFA-SL can be synthesized by other enzymes besides FA2H, which follows from their presence in skin and other organs of Fa2h^−/− mice (19) and is also suggested by their normal levels in lymphocytes from SPG35 patients (20). FA2H is, however, presumably the sole enzyme performing the 2-hydroxylation reaction in myelinating cells, because HFA-SL are undetectable in Fa2h^−/− CNS and PNS (15,16).

Figure 1

Myelin proteome analysis of Fa2h^−/− mice. (A) TLC analysis of sphingolipids (after alkaline hydrolysis of glycerolipids) in purified myelin from 6- to 17-month-old Fa2h^+/+ and Fa2h^−/− mice. Position of sphingomyelin (SM), non-hydroxylated (NFA) and 2-hydroxylated (HFA) GalC and sulfatide are indicated. (B) Densitometric quantification (a.u., arbitrary units) of sulfatide bands (NFA and HFA combined) revealed no significant (ns) differences between genotypes in both age groups. Shown are the mean ± SD (n = 2/3/5/3; from left to right). (C) Representative silver stained SDS-polyacrylamide gel of purified myelin isolated from Fa2h^−/− and Fa2h^+/+ mice of all age groups analyzed. (D) Scheme of the sample processing for LC–MS/MS analysis. (E) Venn diagrams indicating the number of protein groups identified, quantified and significantly changed and a classification of the latter (the 83 proteins identified in previous studies refers to the myelin proteome studies summarized in reference 32). (F) Age-dependence of log(2)-fold change (mean of three biological replicates) of selected canonical myelin membrane proteins in Fa2h^−/− versus wild type. Dotted lines indicate the threshold selected for up- (>1.5-fold) and down-regulated (<0.66-fold) protein groups.

Galactosylceramide (GalC) and its sulphated derivative sulfatide are abundant sphingolipids in myelin (21,22) and several studies documented that they are essential for correct formation and maintenance of glia–axon interaction at the paranodes (23–25). Particularly, sulfatide containing membrane microdomains (or rafts) in CNS myelin in conjunction with axonal ganglioside rafts are essential to maintain neurofascin 155 (NF155) and myelin associated glycoprotein at the paranodes (26). Sulfatide is also involved in transcytotic sorting of proteolipid protein 1 (PLP1) in oligodendrocytes (27). However, the specific role of sulfatide and GalC 2-hydroxylation and how the inability to synthesize HFA-SL in the nervous system leads to SPG35 is largely unknown. In particular, it is unclear how an apparently myelin-specific lipid modification can lead to axonal degeneration. We hypothesized that HFA-SL may affect transport or localization of membrane proteins either by direct interaction or by changing their distribution between specific membrane subdomains. According to this hypothesis, we assumed that the abundance of some myelin membrane proteins might be altered in Fa2h^−/− myelin. We therefore examined the CNS myelin proteome of Fa2h^−/− and wild type mice by mass spectrometry using tandem mass tags (TMT) (28). We observed a significant increase of one major myelin protein, oligodendrocytic myelin paranodal and inner loop protein (Opalin, also known as Tmem10) (29–31), in Fa2h^−/− mice. In contrast, the levels of other canonical myelin membrane proteins were not affected. Further experiments suggested that the turnover of Opalin might be influenced by HFA-SL, particularly HFA-sulfatide.

Results

Myelin proteome analysis of Fa2h^−/− mice

Myelin was isolated from Fa2h^+/+ and age-matched Fa2h^−/− mice. Thin-layer chromatography (TLC) lipid analysis confirmed absence of HFA-sulfatide and HFA-GalC in Fa2h^−/− myelin (Fig. 1A), but no significant reduction in total sulfatide content (Fig. 1B). Reproducibility of the myelin preparation was monitored by SDS-PAGE followed by silver staining (Fig. 1C). Myelin samples from three biological replicates per age group and genotype were analyzed by mass spectrometer as described in the Methods section (Fig. 1D). A total of 1682 protein groups were identified (see complete dataset in the PRIDE partner repository, dataset identifier PXD021021). In comparison with a published CNS myelin reference proteome list containing 1264 proteins (32), we found an overlap of 829 protein groups (66% of the proteins in the reference list; 49% of the protein groups identified in our study were present in the reference proteome). Filtering (criteria: presence in at least two out of three biological replicates) reduced the number of protein groups that could be quantitatively evaluated to 1205 (Supplementary Material, Table S1). Of these, 145 were significantly changed [i.e. |log(2)-fold change|≥0.585 with a significance level of P < 0.05 (one sample t-test)] in at least one time point (90 increased, 55 decreased) (Fig. 1E and Supplementary Material, Tables S2 and S3). From these candidate lists of significantly up- or down-regulated proteins, we selected all transmembrane proteins (excluding likely mitochondrial proteins), or glycosylphosphatidylinositol (GPI)-linked proteins (Table 1). Four proteins (group C in Table 1) are not present in the reference list (32) and have to our knowledge not been previously identified in myelin. It is possible that these proteins are contaminants derived from non-myelin membranes. Five proteins (group B in Table 1) have been identified in purified myelin in earlier proteome studies (reference 32 and references therein), but to our knowledge their presence in myelin has not been confirmed by other methods. Only two up-regulated proteins, cell adhesion molecule 4 (Cadm4) and Opalin, are well-established myelin membrane proteins. In particular, Opalin showed the highest relative increase (2-fold in 17-month-old Fa2h^−/− mice compared to wild type controls) among all significantly altered membrane proteins. Notably, many other well-known myelin proteins identified by mass spectrometry in our study were not changed significantly: 2′,3′-cyclic nucleotide 3′-phosphodiesterase (CNP), large isoform of myelin-associated glycoprotein (L-MAG), myelin basic protein (MBP), myelin oligodendrocyte glycoprotein (MOG), PLP1 and NF155 (Fig. 1F).

Table 1

Membrane proteins significantly altered in Fa2h^−/− myelin

	Gene symbol	Protein name	Accession number	number of TM helices or GPI anchor	log2 fold change (−log10  P-value):
					6 months	13 months	17 months
A	Cadm4	Cell adhesion molecule 4	Q8R464	1	0.42 (2.227)	0.51 (1.586)	0.59 (2.139)
	Opalin	Opalin	Q7M750	1	0.93 (1.152)	0.76 (1.045)	1.01 (1.957)
B	Cd59a	CD59A glycoprotein	O55186	GPI	−0.23 (1.041)	0.07 (0.085)	−1.60 (1.565)
	Igsf8	Immunoglobulin superfamily member 8	Q8R366	1	−0.22 (1.348)	−0.43 (2.315)	−0.59 (2.44)
	Scamp4	Secretory carrier-associated membrane protein 4	Q9JKV5	4	0.57 (1.966)	0.62 (2.091)	0.612 (1.700)
	Sirpa	Tyrosine-protein phosphatase non-receptor type substrate 1	P97797	1	0.62 (1.496)	0.27 (1.316)	0.48 (0.655)
	Sypl1	Synaptophysin-like protein 1	O09117	3	0.63 (1.732)	0.09 (0.517)	−0.09 (0.204)
C	Gprc5b	G-protein coupled receptor family C group 5 member B	Q923Z0	7	1.15 (1.124)	0.59 (1.619)	0.01 (0.024)
	Mcam	Cell surface glycoprotein MUC18	Q8R2Y2	1	0.72 (4.045)	0.79 (1.699)	0.43 (0.788)
	Slc38a3	Sodium-coupled neutral amino acid transporter 3	Q9DCP2	11	0.09 (0.038)	0.11 (0.231)	−1.56 (2.577)
	Unc5b	Netrin receptor Unc5b	Q8K1S3	1	0.33 (1.443)	0.63 (2.061)	0.47 (2.257)

From the significantly changed protein groups identified (see Supplementary Material, Table S1) only membrane proteins [transmembrane proteins or proteins with GPI-anchor] are listed here. The proteins were further divided into categories A: known myelin protein, as listed in reference (54), B: previously identified as myelin-associated proteins, as reported by de Monasterio-Schrader et al. (32), and C: newly identified in myelin in the present study. The number of transmembrane (TM) helices was predicted with TMHMM2.0 (71) (signal peptides were not considered). Gene names, protein names and accession numbers are given according to UniProt (www.uniprot.org). The mean log(2)-fold change (Fa2h^−/− versus Fa2h^+/+ of three biological replicates per genotype) and the −log(10) P-value (one sample t-test) are presented and depicted in bold if significant according to the applied criteria (|log(2)-fold change| ≥ 0.585 and P < 0.05).

Opalin concentration decreases in old wild type but not Fa2h^−/− mice

To confirm the specific upregulation of Opalin and Cadm4, as well as the unaltered levels of other myelin proteins, western blot analysis of independent myelin samples isolated from 5-week-, 6-month- and 17-month-old wild type and Fa2h^−/− mice was performed (Fig. 2A). In wild type mice, Opalin levels decreased at 17 months of age compared to younger mice Interestingly, Fa2h^−/− mice did not show this age-dependent decline, but instead a 2-fold higher Opalin level at 17 months of age compared to wild type controls (P = 0.018, two-sided t-test) (Fig. 2B). Although a similar age-dependent profile was found for Cadm4, its apparent higher level in Fa2h^−/− mice at 17 months did not reach statistical significance in western blot analysis (P = 0.059, two-sided t-test) (Fig. 2B). None of the canonical myelin proteins examined (MOG, L-MAG, NF155, CNP, PLP1 and MBP) showed significant alterations in Fa2h^−/− compared to Fa2h^+/+ myelin at any age group, in agreement with the mass spectrometry results (Fig. 1F and Supplementary Material, Table S1).

Figure 2

Western blot analysis of Opalin and other major myelin proteins. (A) Myelin isolated from brains of Fa2h^−/− and Fa2h^+/+ mice at different age groups (5 weeks, 6 months, 17 months) was analyzed by western blotting with antibodies directed against Opalin, Cadm4, PLP1, L-MAG, NF155, MOG, CNP and MBP. Representative blots are shown. (B) Densitometric quantification (a.u., arbitrary units) of western blots showed significantly elevated levels of Opalin in 17-month-old Fa2h^−/− mice compared to age-matched wild-type mice, but not of the other myelin proteins examined or at other age groups [P-values (two-sided t-test) for Cadm4 and Opalin in myelin from 17-month-old mice are specified]. Data shown are the mean + SD (n = 3 per age group and genotype).

The difference in Opalin concentration was more pronounced in older (21–23-month-old) mice, where we detected an almost 6-fold difference in purified myelin (Fig. 3A) and a 3-fold difference in total brain (Fig. 3B). The latter is in line with the overall reduced myelin content as demonstrated by the 50% reduction in MBP level in total brain (Fig. 3B). Immunofluorescence staining of brain sections (20 months) indicated elevated Opalin levels in the white matter of different brain regions, e.g. in the corpus callosum (Fig. 3C) and the cerebellum (Fig. 3D). Higher magnification of the cerebellar white matter showed presence of Opalin in cellular processes, myelin and the oligodendrocyte cell bodies (Fig. 3E). Compared to wild type, Opalin staining in Fa2h^−/− mice was clearly more intense in myelinated axons and cellular processes (arrows in Fig. 3E), but apparently similar in the cell bodies (arrow heads in Fig. 3E). This suggest that Opalin turnover is specifically altered in the myelin (and cellular processes).

Figure 3

Elevated Opalin level in old Fa2h^−/− mice. Western blot analysis of Opalin and MBP in purified myelin (A) and total brain homogenates (B) from 21- to 23-months-old Fa2h^+/+ and Fa2h^−/− mice. Representative blots are shown and data of the densitometric quantification are shown as mean + SD (n = 4) with mean of wild type set to one (P-value given for two-sided t-test). Immunofluorescence staining of Opalin and MBP (nuclei stained with DAPI) of brain sections from 20-month-old mice showed increased levels of Opalin in corpus callosum (C) and cerebellum (D). Higher magnification (deconvolution of z-stacks) of the cerebellar white matter showed higher levels of Opalin in MBP positive myelin from Fa2h^−/− mice compared to wild type, but comparable levels in oligodendrocyte cell bodies (arrow heads) in both genotypes (E). Scale bars, 100 μm (C, D), 20 μm (E).

Astrogliosis and microglia activation in Fa2h^−/− mice

Besides canonical myelin proteins, a large number of other proteins were identified in the proteomics data set. Several of them have also been identified in earlier myelin proteome screens (reference 29 and references therein; Fig. 1E), but whether these proteins are true myelin proteins or contaminants requires further studies. A significant upregulation starting at the earliest time point examined (6 months) was found for glial fibrillary acidic protein (GFAP), and the complement factors C1qb, and C4b (Supplementary Material, Table S2). The strong increase of the GFAP level in myelin and total brain was confirmed by Western blotting (Fig. 4A and B), and immunofluorescence, which showed progressive astrogliosis in Fa2h^−/− brains (Fig. 4C). We propose that GFAP signals in western blots of isolated myelin are likely caused by contamination, because immunofluorescence did not reveal any co-localization of GFAP with MBP (or Opalin) (data not shown). Elevated levels of C1qb and C4b possibly indicate microglia activation, which was confirmed by Iba1 immunofluorescence staining. Brain sections of Fa2h^−/− mice exhibited the typical morphological changes of activated microglia in different brain regions, i.e. the hippocampus and neocortex (Fig. 4D). Microglia activation was substantial at 20 months, but only weak at 6 months.

Figure 4

Astrogliosis and microglia activation in Fa2h^−/− mice. Western blot analysis of (A) purified myelin and (B) total brain homogenates showed strong increase of GFAP in 21–23-month-old Fa2h^−/− mice. Relative intensity (mean + SD; n = 4) (in total homogenates normalized to tubulin) with wild type set to one; P-values for two-sided t-test are indicated. (C) Immunofluorescence staining of GFAP in wild type and Fa2h^−/− brain sections (age: 20 months). Note the increased GFAP signal in cerebellum, cortex and hippocampus of Fa2h^−/− mice, indicating massive astrogliosis. (D) Iba1 immunofluorescence staining in hippocampus and neocortex indicated significant microglia activation in 20-months-old Fa2h^−/− mice. Nuclei were stained with DAPI. Scale bars, 50 μm (C), 100 μm (D).

Increase of Opalin in Fa2h^−/− mice results from a posttranscriptional process

To examine whether the increased Opalin concentration in Fa2h^−/− myelin is the result of altered gene expression (or mRNA stabilization) quantitative real-time PCR was performed. There was a significant reduction of Opalin expression in 6 as well as 17-month-old mice compared to mice at 2 months of age [main effects ANOVA, effect of age: F(2,20) = 15.5231, P < 0.0001; post hoc Tukey HSD test P < 0.05] (Fig. 5A). There was, however, no significant effect of the genotype [F(1,20) = 0.036; P = 0.85]. It is therefore unlikely that elevated Opalin levels reflects enhanced remyelination, which potentially could also result in higher Opalin concentration, because remyelination would be accompanied by increased expression of Opalin and other canonical myelin genes. However, expression levels of other myelin genes [Plp1, Mbp, myelin and lymphocyte protein (Mal)] showed a similar age-dependent profile in Fa2h^−/− and wild-type controls (Fig. 5B–D). Thus, unaltered Opalin mRNA levels in Fa2h^−/− mice suggests decreased protein turnover (or potentially more efficient translation) of Opalin in the absence of HFA-SL.

Figure 5

Opalin gene expression does not differ between Fa2h^−/− and wild type mice. Gene expression of Opalin (A), Plp1 (B), Mbp (C) and Mal (D) in total brain was examined by quantitative real-time RT-PCR at 2, 6 and 17 months of age in Fa2h^−/− (white bars) and wild type controls (black bars). Expression levels were normalized to ubiquitin C (Ubc) and β-actin (actb) (geometric mean) by the 2^−ΔCt method. Shown are the mean ± SD (n = 3) of three independent samples per age group and genotype (except Opalin, 17 months: n = 6). Main effects ANOVA revealed a significant effect of age for all genes except Mal [Opalin: F(2,20) = 15.5231, P < 0.0001; Plp1: F(2,14) = 25.8681, P < 0.0001; Mbp: F(2,14) = 8.561, P = 0.0037; Mal: F(2,14) = 2.7227, P = 0.10], but no effect of genotype (^*P < 0.05, post hoc Tukey HSD test).

Presence of Opalin in the CHAPS insoluble membrane fraction is not affected by the absence of HFA-SL

Potentially, changes in Opalin turnover in the absence of HFA-SL could be the result of altered localization of Opalin in specific membrane domains that can be biochemically characterized as detergent resistant membranes. Opalin was largely soluble in Triton X-100 (Fig. 6A). In contrast, Opalin was partially present in the CHAPS insoluble membrane fraction (CIMF), as shown by its partial insolubility in CHAPS at 4°C, increased solubility at 37°C (Fig. 6B), and presence of the insoluble Opalin in low-density fractions of a sucrose density gradient (Fig. 6C). MOG, which was analyzed as a control, was almost insoluble in CHAPS at both, 4 and 37°C, in line with a previous report (33). Because of its tendency towards higher concentration in Fa2h^−/− myelin, Cadm4 was also examined. Solubility of Cadm4 was to some extent similar to Opalin (Fig. 6B and C). There was, however, no difference in the solubility of Opalin or Cadm4 when myelin from wild type and Fa2h^−/− mice was compared.

Figure 6

Opalin is partially present in CIMF independent of Fa2h expression. (A) Purified myelin from 6- and 17-month-old Fa2h^−/− and Fa2h^+/+ mice was extracted with 1% Triton X-100 for 30 min at 4°C, and detergent resistant and soluble membranes were separated by centrifugation at 20 000 × g. Equal amounts of supernatant (S) and pellet (P; resuspended in the original volume of extraction buffer), were analyzed by Western blotting with antibodies against Opalin, MOG and CNP. (B) Purified myelin from 17-month-old Fa2h^−/− and Fa2h^+/+ mice was extracted with 1% CHAPS for 30 min at 4°C or 37°C and further processed and analyzed as described in (A). In these experiments, western blots were also probed with anti-Cadm4 antibody. (C) CHAPS insoluble myelin membranes [prepared as described in (B) at 4°C] were subjected to sucrose density gradient centrifugation and aliquots of gradient fractions (1–10) together with a corresponding aliquot of the CHAPS soluble fraction (S) were analyzed by western blotting.

Opalin degradation in a sulfatide synthesizing cell line is accelerated by Fa2h co-expression

To explore whether Opalin protein turnover may be directly affected by the hydroxylation status of sphingolipids, we used stable CHO cell lines (that do not endogenously synthesize HFA-SL) overexpressing UDP-galactose:ceramide galactosyltransferase (Ugt8) (CHO-GalT) or Ugt8 in combination with galactose-3-O-sulfotransferase 1 ((Gal3st1) (CHO-Sulf). An Opalin expression plasmid was co-transfected with a wild type or mutant Fa2h plasmid, the latter coding for an enzymatically inactive FA2H mutant (His315/319Ala) that lacks a catalytically essential histidine motif (34) (Fig. 7A). Inactivity of this mutant was confirmed by lipid TLC (Fig. 7B). Beside the fully processed N-glycosylated Opalin (>40 kDa), two bands of 27 and 20 kDa were detected with the Opalin antiserum in Opalin expressing cells (Fig. 7C), in agreement with previous reports (29,31). The 27 kDa band most likely represents immature N-glycosylated Opalin (because of its Endo H sensitivity; data not shown), whereas the origin of the 20 kDa band is unclear. Cycloheximide (CHX) treatment revealed faster degradation of the fully processed N-glycosylated (mature) Opalin in CHO-Sulf cells expressing active Fa2h compared to CHO-Sulf cells expressing the inactive Fa2h mutant (Fig. 7C and D). In contrast, Opalin turnover was not influenced by FA2H activity in the parental CHO-GalT cell line. The time course of Opalin degradation in this cell line was similar to CHO-Sulf cells co-expressing inactive Fa2h. This suggests that possibly HFA-sulfatide but not other HFA-SL is responsible for the faster decline of Opalin in Fa2h expressing CHO-Sulf cells. The increase of (mature) Opalin after 4 h of CHX treatment can be explained by the N-glycan processing dependent molecular mass shift of the immature Opalin. The half-life of immature Opalin was independent of Fa2h expression in both, CHO-GalT and CHO-Sulf cells. This indicates that Opalin transport from the ER to Golgi apparatus (where maturation of the N-glycan occurs) was not affected by HFA-SL. This would again be in line with a specific role of HFA-sulfatide, because the sulfatide synthesizing sulfotransferase is located in the Golgi apparatus (35), whereas HFA-GalC is formed in the ER (36). Degradation of Opalin was inhibited by leupeptin (LEU) but not by the proteasome inhibitor bortezomib (BTZ), indicating lysosomal degradation (Fig. 7E and F). In contrast, degradation of the (unstable) inactive Fa2h mutant was partially inhibited by BTZ, in line with the expected proteasomal degradation of this ER resident protein (Fig. 7E). The faster degradation of Opalin in the presence of BTZ may result from autophagy induction (37). Under these conditions, degradation of Opalin occurred at a similar rate in CHO-Sulf cells co-expressing active or inactive Fa2h (Fig. 7F). This suggests that intra-lysosomal degradation of Opalin was not affected by HFA-SL/sulfatide.

Figure 7

Altered turnover of Opalin in CHO cells synthesizing 2-hydroxylated sulfatide. (A) Western blot analysis of FA2H in transiently transfected CHO-GalT and CHO-Sulf cells. An enzymatically inactive FA2H mutant was generated by replacing two histidine residues (His-315, His-319) by alanine that are required to coordinate one of the two catalytically essential zinc ions in the active center of the enzyme, as shown in the display detail of the homology modeled FA2H structure. (B) TLC analysis of sphingolipids (glycerolipids were removed by alkaline hydrolysis) isolated from Fa2h transfected CHO-GalT and CHO-Sulf cells. Position of lipid standards run on the same TLC are indicated on the left side (GalC, galactosylceramide; GlcC, glucosylceramide; SM, sphingomyelin). (C) CHO-GalT and CHO-Sulf cells were transiently co-transfected with expression plasmids coding for Opalin and active FA2H or the inactive FA2H mutant (His315/319Ala). After 16 h, cells were treated with cycloheximide (CHX) for 0, 4 or 8 h and cell lysates analyzed by western blotting with antibodies against Opalin, FA2H and actin. One representative experiment is shown. Note that the FA2H mutant is unstable and therefore almost undetectable after 4 and 8 h of CHX treatment. (D) Densitometric quantification of mature and immature Opalin polypeptides (relative intensity normalized to actin; intensity at 0 h CHX treatment set to 100%). Data shown are the mean ± SD (n = 3). White circles = co-expression of active Fa2h; black squares = co-expression of inactive Fa2h mutant. Intensity of mature Opalin (>40 kDa band) in CHO-Sulf cells co-expressing active Fa2h was significantly reduced at 8 h CHX treatment, compared to CHO-Sulf cells co-expressing inactive Fa2h (^*P < 0.05, t-test). (E) CHO-Sulf cells were transiently co-transfected with plasmids coding for Opalin and active (+) or inactive FA2H (−^*) as in (C), and treated for 8 h with CHX, leupeptin (LEU) and/or BTZ, as indicated. Western blot analysis of cell lysates was performed as in (C), except that blots were reprobed with anti-α-tubulin (instead of anti-actin) antibodies for loading control. One representative of four independent experiments is shown. (F) Densitometric quantification of the mature Opalin band (untreated cells set to 100%). Asterisks indicate significant differences (^*P < 0.05, t-test) when compared to cells treated with CHX and leupeptin (LEU). Data shown are the mean ± SD (n = 4).

Attempts to determine the endocytosis rate of Opalin in CHO-Sulf cells using cell surface biotinylation were unsuccessful, because we were unable to detect significant endocytosis of cell surface biotinylated Opalin (endocytosis rate of biotinylated Opalin was <1% within 8 h, independent of FA2H activity, which did not allow a reliable quantification; data not shown). This very low endocytosis rate may indicate that intracellular sorting of Opalin is affected by sphingolipid hydroxylation in CHO-Sulf cells. However, we cannot rule out the possibility that biotinylation of the only extracellular lysine residue of Opalin, which is close to the transmembrane domain, may interfere with endocytosis.

Discussion

The myelin proteome analysis presented here shows that proteins that are significantly altered in sulfatide deficient mice, such as NF155 and MAG (26,38), are not changed in myelin from Fa2h^−/− mice, suggesting that the formation of ‘sulfatide rafts’ occurs independent of sphingolipid hydroxylation. This is in line with normal localization of ion channels at the nodes of Ranvier in Fa2h^−/− mice (15), in contrast to the abnormal paranode structure in sulfatide deficient mice (39). Our data suggest a new role of HFA-sulfatide namely influencing the turnover of the myelin membrane protein Opalin.

A decrease of Opalin level in certain brain regions of adult mice (6 months) has been reported previously (40). We show here that this reduction is even more pronounced in older wild type but is not detectable in Fa2h^−/− mice. The absence of a significant upregulation of Opalin gene expression in Fa2h^−/− mice suggests that the specific absence of the age dependent decline in the Opalin protein level is caused by altered protein turnover. The faster degradation of Opalin in CHO-Sulf cells expressing Fa2h supported this hypothesis and further suggested that HFA-sulfatide but not other HFA-SL may enhance Opalin sorting towards lysosomal degradation. We have currently no evidence for an altered distribution of Opalin within the myelin membrane that could explain a selective retention of Opalin in HFA-SL deficient myelin. Alternatively, transport of newly synthesized Opalin into myelin or its sorting within recycling endosomes may be influenced by the hydroxylation status of sphingolipids. Differential endosomal sorting of myelin proteins has been documented in oligodendrocytes (41). In this context, it should be examined whether Opalin may directly bind to HFA-SL. The binding enhancement or binding-qualification of proteins to sphingolipids via 2-hydroxylation has been shown for example for the interaction of galectin-4 with HFA-GalC and HFA-sulfatide (42), and binding of HIV-1 gp120 envelope protein to HFA-GalC (43).

SPG35 belongs to the small group of HSP caused by mutations in genes mainly expressed in oligodendrocytes. Other HSP in this group are SPG2 (mutations in PLP1) (44), SPG44 (mutations in connexin 47) (45), and SPG75 (mutations in MAG) (46). Because there may be no truly oligodendrocyte specific gene (e.g. Plp1 is also expressed in neurons) (47), it was questioned whether loss of these genes in oligodendrocytes is responsible for pathogenesis of the corresponding HSP. For Plp1, however, it was shown that only an oligodendrocyte-specific knock-out leads to HSP symptoms in mice (47). Although a study by Potter et al. (16) suggested a role for Fa2h independent of its expression in oligodendrocytes, spastic paraplegia in Fa2h^−/− mice is most likely caused by Fa2h deficiency in oligodendrocytes, as suggested by the following observations. Absence of significant amounts of HFA-SL in neurons is suggested by the observation that transgenic mice overexpressing Gal3st1 and Ugt8 under control of a neuron specific promoter selectively increased the synthesis of C18:0 sulfatide [in line with the predominant neuronal expression of ceramide synthase CerS1 (48,49), which is highly selective for C18:0 acyl-CoA (50)] but not its 2-hydroxylated variant (51). Moreover, when Gal3st1 and Ugt8 were overexpressed in neurons of arylsulfatase A-deficient mice, which are unable to degrade sulfatide, storage of non-hydroxylated C18:0-sulfatide increased (selectively in neurons), but there was no obvious accumulation of 2-hydroxylated C18:0-sulfatide (52). This strongly suggests that synthesis of HFA-SL in neurons is marginal, if present at all. Moreover, comparable cerebellar dysfunction was observed in total and oligodendrocyte-specific Fa2h^−/− mice (16). It is therefore highly probable that SPG35 and axonal pathology in Fa2h^−/− mice are caused by FA2H deficiency in oligodendrocytes.

It remains to be determined whether the relative increase of Opalin in myelin does play a role in the pathogenesis of Fa2h deficiency in mice and possibly SPG35. Though, at least some observations suggest this possibility. Significant higher levels of Opalin in older Fa2h^−/− mice correlate with the late onset of the disease in Fa2h^−/− mice, as demonstrated by the time course of microglia activation shown here and the detection of cerebral dysfunction not before 7 months of age (16). A role of Opalin in the pathogenesis of SPG35 could also explain why primarily CNS axons are affected in SPG35, as in all HSP, although HFA-SL are equally abundant in PNS and CNS myelin, because Opalin is exclusively expressed by oligodendrocytes but not Schwann cells (29,31). Similarly, connexin 47 (SPG44) and PLP1 (SPG2) are characteristic for CNS (53,54) but not present or only at a low level in PNS myelin (55–58). To examine a direct link between increased Opalin levels and spastic paraplegia, studies of transgenic mice overexpressing Opalin and testing a possible rescue of the Fa2h^−/− mouse pathology by homozygous or heterozygous Opalin deficiency may prove particularly informative.

Material and Methods

Animals

Generation, breeding and genotyping of Fa2h^−/− mice (MGI:3829000; Fa2h^tm1Meck) have been described earlier (15). All animal experiments were in accordance with the national law and approved by the local authorities (Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, Germany).

Antibodies

Antibodies used in this study are listed in Table 2. An Opalin antiserum was generated by immunizing guinea pigs with the cytoplasmic domain of murine Opalin fused to maltose binding protein. The fusion protein was expressed in Escherichia coli SoluBL21 and purified by affinity chromatography via amylose resin (New England Biolabs, Frankfurt am Main, Germany). Guinea pigs were immunized and the antiserum obtained was used without further purification.

Table 2

Antibodies used in this study

Primary antibodies
Name/antigen	Host	Supplier/reference	Prod. no.	Lot no.	Dilution (applicationa)
Opalin	Guinea pig	This paper	n/a	#2/70d	1:20 000 (WB) 1:1000 (IF)
Opalin (H12)	Mouse	Santa Cruz	sc-374490	#J1911	1:500 (WB)
MOG (clone 8-18C5)	Mouse	Merck Millipore	MAB 5680	2606334	1:10 000 (WB)
CNP (clone 11-5B)	Mouse	Merck Millipore	MAB 326	2373391	1:3000 (WB)
Neurofascin	Rabbit	Abcam	ab31457	GR264760-1	1:1000 (WB)
L-MAG (M4)	Rabbit	Reference (72)	n/a	n/a	1:10 000 (WB)
Cadm4 (clone N244/5)	Mouse	UC Davis/NIH NeuroMab facility, Antibodies Inc.	ABIN1304992	437-4VA-37	1:250 (WB)
MBP	Rabbit	Merck Millipore	AB980	2739500	1:10 000 (WB) 1:800 (IF)
PLP	Rabbit	Abcam	ab28486	GR3304474-1/GR2224399	1:5000 (WB)
FA2H	Rabbit	Reference (19)	n/a	#2/145d	1:2000 (WB)
α-tubulin	Mouse	Sigma-Aldrich	T5168	39M4769	1:20 000 (WB)
GFAP	Rabbit	Abcam	ab7260	GR47492-2, GR3284661-1	1:5000 (WB) 1:2000 (IF)
Iba1	Rabbit	Fujifilm Wako Chemicals	019-19741	CAJ3125	1:5000 (IF)
β-actin	Mouse	Sigma-Aldrich	A5316	018M4804V	1:20 000 (WB)
Secondary antibodies
Host/specificity/conjugation		Supplier	Prod. no.	Lot no.	Dilution (application  a)
Goat anti-mouse peroxidase		Dianova	115-035-044	132855	1:40 000 (WB)
Goat anti-rabbit peroxidase		Dianova	111-035-003	137093	1:10 000 (WB)
Goat anti-guinea pig peroxidase		Dianova	106-035-003	Unknown	1:5000 (WB)
Goat anti-rabbit Cy3		Dianova	111-165-144	143018	1:500 (IF)
Goat anti-guinea pig Cy3		Dianova	106-165-003	110228	1:500 (IF)
Goat anti-rabbit Alexa647		Thermo Fisher	A21244	1783430	1:500 (IF)
Goat anti-guinea pig Alexa647		Thermo Fisher	A21450	1711474	1:500 (IF)

^aIF, immunofluorescence; WB, western blotting

Purification of myelin

Myelin was purified by sucrose gradient centrifugation according to Larocca and Norton (59). Briefly, brains were homogenized in 10.5% isotonic sucrose solution supplemented with cOmplete protease inhibitor cocktail (Roche, Mannheim, Germany) using an Ultra-Turrax homogenizer (IKA, Staufen, Germany). Samples were centrifuged at 18 000 × g 4°C for 45 min to remove high-density constituents. Supernatants were dissolved in 30% sucrose solution, overlaid with 10.5% sucrose solution and centrifuged at 68 000 × g and 4°C for 50 min (SW 41 Ti rotor). The myelin containing interphase was collected and washed twice in 1 mm EDTA (68 000 × g, 4°C, 30 min). To further purify myelin, sucrose centrifugation and washing steps were repeated as mentioned above. Protein concentrations were determined using the DC protein assay (Bio-Rad, Feldkirchen, Germany) with bovine serum albumin as standard.

Detergent extraction of myelin and sucrose density gradient centrifugation

Purified myelin (500 μg protein) was extracted in 1 ml of 1%(v/v) Triton X-100 or 1%(w/v) CHAPS in TNE buffer [20 mm Tris–HCl pH 7.4, 150 mm NaCl, 5 mm EDTA, 1 × HALT protease inhibitor mix (Thermo Fisher)] for 30 min at 4 or 37°C (33). Detergent insoluble and soluble fractions were separated by centrifugation (20 000 × g, 10 min, 4°C) and analyzed by western blotting. Sucrose density gradient centrifugation of the CIMF was performed as described elsewhere (33), with minor modifications. Briefly, the CHAPS insoluble 20 000 × g membrane pellet was resuspended in 1 ml of CHAPS in TNE, mixed with 2 ml of 2 m sucrose in TNE, overlaid with 6 ml of 1 m sucrose and 3 ml of 0.2 m sucrose, and centrifuged for 21–23 h at 175 000 × g_max [32 000 rpm, SW 41 Ti swinging-bucket rotor (Beckman Coulter, Krefeld, Germany)]. Fractions of 1 ml were collected from the top and aliquots were analyzed by western blotting (loading aliquots of each fraction corresponding to 10 μg starting material).

Processing of myelin samples for mass spectrometric measurements

About 40 μg of purified myelin protein fractions per age and genotype were acetone precipitated to remove lipids, dissolved in 0.1% RapiGest (Waters, Milford, MA) in 0.1 m triethylammonium bicarbonate pH 8.5, reduced with 5 mm dithiothreitol for 45 min at 60°C, alkylated with 50 mm acrylamide for 20 min at room temperature and finally digested overnight at 37°C with trypsin (Promega, Fitchburg, WI) at an enzyme to protein ratio of 1:100. Peptides were labeled with a TMT-6plex isobaric label reagents set (Thermo Fisher, Waltham, MA) according to the manufacturer’s instruction. Reagents were dissolved in acetonitrile (ACN) and protein samples were labeled for 90 min at room temperature. In doing so samples of the same time point and genotype were labeled each with 40 μl of the same reagent. Labeling was stopped by the addition of 8 μl of 5% hydroxylamine in 0.1 M triethylammonium bicarbonate and incubation for 15 min. Differentially labeled samples were combined and RapiGest was removed by acidic cleavage with 5% trifluoroacetic acid (37°C, 800 rpm for 45 min) and a precipitation for 30 min at 20 000 × g. Supernatants were desalted using Oasis HLB 1cc (10 mg) columns (Waters) that were activated with 3 × 1 ml 70% ACN/0.1% formic acid (FA) and then equilibrated with 4 × 0.1% FA. Samples were diluted in 0.1% FA, loaded onto the column and washed with 10 × 1 ml 0.1% FA. Sequentially elution was performed with 500 μl 30% ACN and 300 μl 70% ACN and the combined eluates were dried in a vacuum centrifuge. Peptides were fractionated by isoelectric focusing via an Agilent G3100A OFFGEL system (Agilent Technologies, Santa Clara, CA) in combination with 12 well pH 3–10 IPG strips according to the manufacturer’s instructions but with the omission of glycerol from the rehydration buffer. The resulting 12 fractions were desalted using self-prepared C₁₈ stop-and-go extraction (Stage) tips (60). The acidified samples were loaded onto the Stage tips, washed with 100 μl 0.5% acetic acid and eluted twice with 10 μl of 80% ACN/0.5% acetic acid. Elution fractions were vacuum dried and stored at −80°C until further processed and analyzed.

LC–MS measurements

Dried peptides were reconstituted in 5% ACN/5% FA and analyzed by liquid chromatography mass spectrometry (LC–MS). Peptides were separated on self-prepared nano-emitter columns (300 mm length × 100 μm i.d.) packed with C18 reversed phase material (Magic C₁₈, 5 μm, Bruker-Michrom, Germany) using an Easy-nLC 1000 UHPLC (Thermo Fisher). 0.1% FA/5% dimethyl sulfoxide/94.9% H₂O was used as solvent A and 0.1% FA/5% dimethyl sulfoxide/94.9% ACN as solvent B. Peptides were eluted over the course of 60 min by linearly increasing the concentration of buffer from 1 to 35% at a flow rate of 400 nl/min. Eluting peptides were analyzed by an Orbitrap Velos (Thermo Fisher) mass spectrometer in data-dependent acquisition mode using a TOP10 method. MS1-spectra were acquired at a resolution of 30 000 from 400 to 1200 m/z (AGC threshold: 1E⁶, max. Inject time: 300 ms). The 10 most intense precursors were selected for HCD-fragmentation (intensity threshold: 5E³, isolation width: 2 m/z, normalized collision energy: 42, charge state: >1) and resulting MS2-spectra measured in the Orbitrap at a resolution of 7500 with the first mass fixed at 100 m/z (AGC threshold: 5E⁴, max. Inject time: 250 ms). Dynamic exclusion was set to 30 s to avoid repeated precursor fragmentations.

Data analysis

All MS raw data files were searched together against a UniProtKB Mus musculus (51 544 entries, downloaded 11/2013) and a common contaminant protein database with MaxQuant 1.4.1.2 (www.maxquant.org) (61). The following search parameters were used: protease: trypsin; missed cleavages: 2; variable modifications: acetyl (protein N-terminal), oxidation; fixed modifications: propionamide; MS tolerance (first/main): 20/6 ppm; MS/MS tolerance: 20 ppm; peptide false discovery rate: 0.01; protein false discovery rate: 0.01; quantification method: TMT-6plex; reporter ion PIF filter: 0.75; peptides for quantification: unique + razor. The proteinGroups.txt result file was further processed with Perseus 1.4.1.3 (https://maxquant.net/perseus/) (62) to determine regulated proteins. Contaminant, reverse and ‘only identified by site’ protein hits were removed, protein abundances of each TMT channel was normalized by the channel mean and normalized protein abundances were further used to calculate Fa2h^−/− to Fa2h^+/+ abundance ratios for each time point and within each replicate. After log(2)-transformation, all proteins with less than two valid values at one time point were filtered out. To determine proteins significantly regulated at each time point, one-sample t-tests were conducted. Proteins exhibiting a log(2)-fold change ≥0.585 or ≤−0.585 and a P-value < 0.05 were considered as regulated. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository (63) with the dataset identifier PXD021021.

SDS-PAGE and western blot analysis

SDS-PAGE and semi-dry western blotting was performed as described elsewhere (64). Samples were denaturated at 95°C for 5 min in sample buffer (final concentration: 125 mm Tris–HCl pH 6.8, 2.5% SDS, 10% glycerol, 0.005% bromophenol blue, 2.5% 2-mercaptoethanol). In some experiments (immunodetection of PLP1), samples were mixed with 2 volumes of urea sample puffer (50 mm Tris–HCl, 5% SDS, 4 m urea, 2.5% glycerol, 10 mm dithiothreitol, 0.01% bromophenol blue, pH 6.8) (33). Proteins were separated in discontinuous SDS-polyacrylamide gels and transferred to nitrocellulose blotting membranes (Amersham Protran 0.1 μm; GE Healthcare) by semi-dry blotting with 2 mA/cm² for 70 min (or in some experiments with constant voltage of 25 V for 30 min). Membranes were blocked with 5% nonfat dry milk and 0.05% Tween-20 in 20 mm Tris–HCl pH 7.4, 150 mm NaCl (TBS) and primary and secondary antibodies (Table 2) were incubated in 3% milk, 0.05% Tween-20 in TBS. Bound secondary antibodies were detected using enhanced chemiluminescence and a CCD camera system (Fusion Solo; Vilber Lourmat, Eberhardzell, Germany). In some experiments secondary antibodies conjugated to infrared fluorescent dyes (Table 2) were used and bound antibodies detected using an Odyssey Infrared Imaging System (LI-COR, Bad Homburg, Germany).

Lipid extraction and TLC

Lipid extraction and alkaline hydrolysis of glycerolipids was performed as described previously (65). Lipids were separated by TLC using silica gel 60 HPTLC plates (Merck, Darmstadt, Germany) and chloroform/methanol/water (70:30:4) as mobile phase. Detection of lipids and densitometric quantification was done as described (65).

Immunofluorescence

Opalin and MBP staining was performed using paraffin sections. Immersion fixation of brain tissues in modified Bouin solution, embedding in paraffin, sectioning and antigen retrieval was performed as previously described (66). Blocking was done for 1 h in 5% normal goat serum in TBS. Incubation with primary antibodies (Table 2) in 5% normal goat serum was overnight at 4°C in a humid chamber. After washing in TBS, sections were incubated with the appropriate secondary antibodies (Table 2) in 5% normal goat serum in TBS supplemented with 300 nM DAPI (Thermo Fisher). Washed sections were then mounted in ProLong Gold antifade (Thermo Fisher). GFAP and Iba1 staining was performed using cryosections of liquid nitrogen snap-frozen brains that were post fixed after sectioning in 4% paraformaldehyde for 30 min, washed with TBS and blocked for 60 min in 2% BSA/2% NGS in TBS at room-temperature. Primary and secondary antibodies were applied as described above with the omission of DAPI in the staining solution. Sections were embedded in Prolong Diamond Antifade supplemented with DAPI (Thermo Fisher). Micrographs were taken with an Axiovert 200 M microscope (Carl Zeiss, Jena, Germany). Deconvolution of z-stacks was performed with the program Axiovision (Carl Zeiss) using the inverse filter method.

RNA isolation and quantitative real-time RT-PCR

Total RNA was isolated from brain tissues using RNeasy Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. cDNA was synthesized from 5 μg total RNA using 200 units superscript II reverse transcriptase (Invitrogen) and oligo (dT) primers in a final volume of 20 μl. After synthesis, cDNA was diluted with 80 μl of water and stored at −20°C. Real-time PCR was done as described previously (67) using an 7300 Real-Time PCR instrument (Life Technologies, Carlsbad, USA.) and SYBR Green PCR-Mastermix (Thermo Fisher). Relative expression levels were calculated by the 2^−ΔCt method using the geometric mean of the C_t values obtained for Atnb and Ubc for normalization. Oligonucleotides used in this study are listed in Table 3.

Table 3

Oligonucleotides used in this study

Number	Sequences	Target gene	Application
1.	AGCCCGTAGAGGAGACTGAG, TATGTGCCCTTCTTGGCTGC	Opalin	Real time PCR
2.	TTTGTGAGTTTGTCTTTGGAGGC, CCGCCATGAGTACCAATTATGT	Mal	Real time PCR
3.	GACCATCCAAGAAGACCCCAC, GCCATAATGGGTAGTTCTCGTGT	Mbp	Real time PCR
4.	TGAGCGCAACGGTAACAGG, GGGAGAACACCATACATTCTGG	Plp1	Real time PCR
5.	AGGCAAGACCATCACCTTGGACG, CCATCACACCCAAGAACAAGCACA	Ubc	Real time PCR
6.	TCCATCATGAAGTGTGACGT, GAGCAATGATCTTGATCTTCAT	Actb	Real time PCR
7.	GACAGCTTACTACCTGGCCTTCGGC, GAAGGCCAGGTAGTAAGCTGTCATG	Fa2h	Site-directed mutagenesis
8.	GCGCAAGCTTACCATGAGTTTTTCACTGAACTTTACACTGCC, GCGCTCTAGACTCGAGTCATTCTAGGCTCAGGCTGGGTA	Opalin	cDNA cloning
9.	GTGCTGTACCTCAGCTGGTC, GCTCTTCTTCAAGAGCCATCC, ATTCGCAGCGCATCGCCTTCTATC	Fa2h/neomycin resistance gene	Mouse genotyping

Plasmids, site-directed mutagenesis and homology modeling

The expression plasmid pcDNA3-FA2H containing a mouse Fa2h cDNA has been described previously (12). A cDNA coding for an enzymatically inactive FA2H variant (His315/319Ala) was generated by site-directed mutagenesis using primer pair 7 (Table 3), as described previously (8). The resulting plasmid encodes an FA2H variant with two functionally essential histidine residues replaced by alanine. Mouse Opalin cDNA was amplified by PCR from mouse brain cDNA using primer pair 8 (Table 3) and subcloned into pcDNA3 vector (via restriction enzyme sites HindIII and XhoI). All plasmid constructs were verified by DNA sequencing (Eurofins Genomics, Ebersberg, Germany). Homology modeling of murine FA2H was performed using the SWISS-MODEL Workspace (68) with the yeast FA2H homolog Scs7p (PDB entry 4ZR0) (34) as template. Details of the homology modeled FA2H structure were drawn using Mol^* Viewer (https://molstar.org/).

Cell culture and transfection

All cells were grown in DMEM:F-12 (1:1), 2 mm l-glutamine, 2.5% fetal calf serum (37°C, 5% CO₂). CHO-GalT cells (69) (kind gift of Brian Popko, University of Chicago) and CHO-Sulf cells (70) were maintained in the presence of 700 μg/ml G418 (both cell lines) and 400 μg/ml zeocin (CHO-Sulf only) for selection (antibiotics were, however, omitted during transfection and subsequent experiments). Cells were transiently transfected using Lipofectamine LTX (Thermo Fisher) according to the manufacturer’s instructions (transfection mixture for one 35-mm-dish was: 2.5 μg plasmid DNA, 2.5 μl Plus reagent and 7 μl Lipofectamine LTX in 200 μl OptiMEM medium added to 2 ml of medium). To block protein synthesis, cycloheximide (CHX; 10 μg/ml) was added for up to 8 h. Cells were lysed in NETN buffer (50 mm Tris–HCl, pH 8.0, 100 mm NaCl, 2 mm EDTA, 0.5% NP-40). To inhibit lysosomal degradation, cells were treated with 100 μg/ml leupeptin (Carl Roth, Karlsruhe, Germany) and inhibition of proteasomal degradation was achieved by adding 250 nM bortezomib (Merck, Darmstadt, Germany).

Cell surface biotinylation and internalization assay

Transiently transfected cells in 35 mm dishes were biotinylated 16 h after transfection. Therefore, cells were cooled on ice, washed twice with cold (4°C) biotinylation buffer (BB: 128 mm NaCl, 4.7 mm KCl, 1.25 mm CaCl₂, 1.25 mm MgCl₂, 5 mm Na₂HPO₄, 20 mm HEPES pH 7.4), and treated with 0.3 mg/ml EZ-Link™ Sulfo-NHS-SS-Biotin (Thermo Fisher) in BB for 30 min at 4°C. Subsequently, cells were washed twice for 15 min at 4°C with 20 mm glycine in BB to block unreacted reagent. Biotin and glutathione (GSH) controls were washed twice with BB, and GSH samples were then incubated 3 × 15 min in GSH stripping buffer (SB: 50 mm reduced glutathione, 90 mm NaCl, 1.25 mm CaCl₂, 1.25 mm MgCl₂, 0.2% bovine serum albumin, pH 8.6), 2 × 5 min with 10 mm iodoacetamide, 20 mm glycine in BB and twice in BB. Endocytosis samples were washed three times with pre-warmed complete growth medium and incubated under standard cell culture conditions (37°C, 5% CO₂) for 8 h. To stop endocytosis, cells were placed on ice and wash twice with cold BB. Surface biotin label was removed by 3 × 15 min incubation with SB. Cells were then washed 2 × 5 min with 10 mm iodoacetamide, 20 mm glycine in BB and twice with BB buffer. After the final wash with BB, all samples were lysed in NETN buffer with 1 × HALT protease inhibitor cocktail (500 μl/35 mm dish). Lysates were centrifuged at 10 000 × g for 10 min (4°C) and the supernatant was used for precipitation of bioinylated proteins as follows. From the 500 μl sample, 25 μl were removed for total lysate and protein assay. Lysates were adjusted to identical protein concentration in a volume of 1.4 ml, 10% SDS was added to a final concentration of 0.1% SDS. Samples were subjected to biotin affinity enrichment with NeutrAvidin agarose beads (20 μl bead volume per sample) (Thermo Fisher) for 4 h at 4°C using an end over end tube rotator. Beads were washed four times with NETN containing 0.1% SDS and 1 × complete protease inhibitor cocktail (Merck). After adding 20 μl of 2 × Laemmli sample buffer with 5% 2-mercaptoethanol, samples were heated to 95°C for 5 min and proteins separated by SDS-PAGE, followed by western blotting. Blots were probed with guinea pig anti-Opalin antiserum and anti-guinea pig peroxidase antibodies and bound antibodies were detected by chemiluminescence using SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher).

Acknowledgements

We thank Dr Arthur M. Butt for the kind gift of antibodies. The excellent technical assistance of Ivonne Becker is highly acknowledged.

Conflict of Interest statement

The authors declare no conflict of interests.

Funding

Deutsche Forschungsgemeinschaft through SFB645 of the University of Bonn (SFB645/B5 to M.E., SFB645/Z4 to V.G.).

Author notes

Robert Hardt and Silvia Jordans contributed equally to this work