Ypc1p (yeast phyto-ceramidase 1) and Ydc1p (yeast dihydroceramidase 1) are alkaline ceramide hydrolases that reside in the ER (endoplasmic reticulum). Ypc1p can catalyse the reverse reaction, i.e. the condensation of non-esterified fatty acids with phytosphingosine or dihydrosphingosine and overexpression of YPC1 or YDC1 can provide enough ceramide synthesis to rescue the viability of cells lacking the normal acyl-CoA-dependent ceramide synthases. To better understand the coexistence of acyl-CoA-dependent ceramide synthases and ceramidases in the ER we investigated the membrane topology of Ypc1p by probing the cysteine residue accessibility of natural and substituted cysteines with membrane non-permeating mass-tagged probes. The N- and C-terminal ends of Ypc1p are oriented towards the lumen and cytosol respectively. Two of the five natural cysteines, Cys27 and Cys219, are essential for enzymatic activity and form a disulfide bridge. The data allow the inference that all of the amino acids of Ypc1p that are conserved in the Pfam PF05875 ceramidase motif and the CREST {alkaline ceramidase, PAQR [progestin and adipoQ (adiponectin) receptor] receptor, Per1 (protein processing in the ER 1), SID-1 (sister disjunction 1) and TMEM8 (transmembrane protein 8)} superfamily are located in or near the ER lumen. Microsomal assays using a lysine residue-specific reagent show that the reverse ceramidase activity can only be blocked when the reagent has access to Ypc1p from the lumenal side. Overall the data suggest that the active site of Ypc1p resides at the lumenal side of the ER membrane.

INTRODUCTION

The mature sphingolipids of Saccharomyces cerevisiae, namely the IPCs (inositolphosphorylceramides), MIPCs (mannosyl-IPCs) and M(IP)2Cs (inositolphosphoryl-MIPCs) are major components of the yeast plasma membrane. Moreover, as in higher eukaryotes, the metabolic intermediates of sphingolipid metabolism have been advocated as signalling molecules in many cell biological phenomena and responses (for a review see [1]).

Ceramide is a central metabolic intermediate (Figure 1), which can result from breakdown of IPCs by Isc1p (inositol phosphosphingolipid phospholipase C 1) [2] or can be synthesized from acyl-CoA and LCBs (long-chain bases) by ceramide synthases consisting of Lag1p (longevity assurance gene 1) or Lac1p (Lag1 cognate 1) in complex with Lip1p (Lag1p/Lac1p-interacting protein 1) [37] (Figure 1). Yeast Lag1p and Lac1p are functionally redundant, but concomitant deletion of LAG1 and LAC1 causes a significant growth defect in the genetic background of W303 cells and is lethal in the YPK9 background [3,4,8]. Insertion of glycosylation and Factor Xa protease sites into Lag1p and Lac1p, the catalytically active subunits of the yeast ceramide synthase, generated a topological model comprising eight putative membrane-spanning domains, whereby the Lag motif, potentially containing the active site, has most of its conserved residues embedded inside the membrane [9]. The two substrates of Lag1p and Lac1p are a VLCFA (very-long-chain fatty acid) coupled to CoA and a LCB, either PHS (phytosphingosine) or DHS (dihydrosphingosine) [10]. VLCFA–CoA is generated at the cytosolic surface of the ER (endoplasmic reticulum) membrane [11], whereas endogenous and exogenous LCBs are generated at the cytosolic and lumenal surface of the ER membrane respectively [9]. On the basis of these data it has been suggested that LCBs could access the ceramide synthase active site from both sides of the membrane, whereas VLCFA–CoAs could access it only from the cytosolic side. However, although probable, it has not yet been proven that the resulting ceramide is released into the cytoplasmic leaflet of the ER membrane.

Ceramide biosynthesis and utilization in yeast

Figure 1
Ceramide biosynthesis and utilization in yeast

The various metabolic pathways generating and consuming ceramides are shown. Gene names are in italic.

Figure 1
Ceramide biosynthesis and utilization in yeast

The various metabolic pathways generating and consuming ceramides are shown. Gene names are in italic.

Natural and spin-labelled ceramides can rapidly flip-flop in liquid-ordered giant unilamellar vesicles [12,13]; however, the relevance of the data obtained with these highly artificial systems to natural membranes remains to be demonstrated. In particular, the flip rate of C26:0-containing ceramides, typical of yeast and presumed to be concentrated in rafts, i.e. in areas of the ER membrane where lipids are present in a liquid-ordered state, has not been measured so far.

Ypc1p (yeast phyto-ceramidase 1) and Ydc1p (yeast dihydroceramidase 1) are two highly homologous alkaline ceramidases with a 54% identity over their entire length [14,15] (Figure 1). They are predicted to have seven TMs (transmembrane helices; Supplementary Figure S1 at http://www.biochemj.org/bj/452/bj4520585add.htm). Overexpression of YPC1 or YDC1 suppresses the lethality of the lag1∆ lac1∆ double mutation [5,16] by allowing ceramide biosynthesis through the reverse ceramidase reaction, i.e. the condensation of fatty acids with PHS or DHS [14,15]. Indeed, the reverse reaction catalysed by Ypc1p can easily be demonstrated in vitro when microsomes or the purified enzyme are incubated with radiolabelled fatty acids and PHS or DHS [14,15].

The simultaneous presence of acyl-CoA-dependent ceramide synthases and alkaline ceramidases in the ER seems to potentially create a futile circle. Indeed, in wt (wild-type) cells, the overexpression of Ypc1p and Ydc1p proteins was shown to significantly increase the levels of free LCBs and LCB-phosphates and to reduce the biosynthetic flow of LCBs towards mature sphingolipids, whereas deletion of YPC1 caused a significant increase in mature sphingolipids as detected through metabolic labelling with [3H]palmitate or [3H]serine or through MS quantification of LCBs [14,17].

Human ACER (alkaline ceramidase) 1, ACER2 and ACER3 are homologues of YPC1. ACER1 is only present in skin keratinocytes, whereas ACER3, showing the highest homology with YPC1, is specific for ceramides carrying unsaturated fatty acids, a kind which is not commonly present in human cells [18]. The ubiquitous ACER2 is localized in the Golgi [18,19], whereas the mammalian acyl-CoA-dependent ceramide synthases are localized in the ER [20]. As ACER2 levels are subject to regulation and moderate overexpression of ACER2 increases sphingosine 1-phosphate levels, the potential for a futile circle also exists in mammalian cells [19].

Ypc1p possesses the Pfam PF05875 ceramidase motif containing seven predicted TMs. To date this gene family contains 374 members in 211 species ranging from bacteria to humans. A recent bioinformatics study revealed some distant similarity between alkaline ceramidases, the PAQR [progestin and adipoQ (adiponectin) receptor] family of receptors, the PER1 (protein processing in the ER 1) GPI (glycosylphosphatidylinositol) remodelase family, and several other gene families with known and unknown functions, all possessing conserved [SxxxH], [D] and [HxxxH] motifs being located at defined positions at the ends of predicted TMs 2, 3 and 7 [17,21]. The members of this so-called CREST [alkaline ceramidase, PAQR receptor, Per1, SID-1 (sister disjunction 1) and TMEM8 (transmembrane protein 8)] superfamily may all be metal-binding hydrolases [21].

In the present study we tried to further our understanding of sphingolipid homoeostasis by determining the membrane topology of the conserved residues of Ypc1p.

EXPERIMENTAL

Yeast strains, media and reagents

The strains used in the present study are listed in Supplementary Table S1 (at http://www.biochemj.org/bj/452/bj4520585add.htm). Cells were grown at 30°C on rich medium {YPD medium [1% (w/v) yeast extract, 2% (w/v) peptone and 2% (w/v) glucose], supplemented with uracil and adenine} or synthetic complete medium (YNB-yeast nitrogen base; United States Biological) containing 2% glucose (D) or Gal (galactose) as a carbon source [22]. Radiolabelled palmitic acid ([3H]palmitate, 60 Ci/mmol) was obtained from ANAWA Trading SA, PHS from Avanti Polar Lipids and NEM (N-ethylmaleimide) from Fluka. Vivaspin 2 ultrafiltration devices having a 10 kDa molecular mass cut-off were from Millipore. All other reagents used were from suppliers described previously [23].

Bioinformatics tools

The conserved amino acids and motifs in Ypc1p were obtained from Pfam and the Conserved Domain Database (http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml). Blast was used for the protein similarity searches. The sequence alignments were performed using ClustalW. Topology predictions were obtained from TOPCONS [46]. Calculations of ΔGmi values for the TMs were obtained from the ΔG prediction server (full protein scan with length correction turned off, allowing TMs 15–30 amino acids in length) [23b].

Construction of strains and plasmids

Plasmids and PCR primers are listed in Supplementary Tables S2 and S3 (at http://www.biochemj.org/bj/452/bj4520585add.htm). The YPK9 lag1∆ lac1∆ strain rescued by overexpression of YPC1 (2∆.YPC1) was constructed by amplifying the ORF (open reading frame) of YPC1 by PCR using genomic DNA as a template and primers 872 and 873 (Supplementary Table S3). The PCR product was inserted into the pGREG533 vector by homologous recombination [24] yielding the plasmid pBF842 which has YPC1 behind the GAL1 promoter, N-terminally tagged by a 7× HA (haemagglutinin) tag. This plasmid was transfected into 2∆.LAC1 cells and the plasmid pBM150-LAC1 was shuffled out by plating the cells on medium containing FOA (5-fluoro-orotic acid). The resulting strain grew on YNBGal. PCR analysis confirmed that LAC1 was absent, indicating that the plasmid pBM150-LAC1 had been shuffled out successfully. In the pBF842 vector the sequence linking the N-terminal 7× HA tag and the YPC1 ORF contains two cysteine residues, which in some of the SCAM™ experiments served as a lumenal control, but also complicated the interpretation of the results. Therefore we replaced the 7× HA tag and this linker sequence in the empty pGREG533 vector by a cysteine free His6–V5 tag. For this, the GAL1 promoter of the pGREG533 vector was amplified with primers 1493 and 1496 resulting in a fragment containing the GAL1 promoter followed by a cysteine-free His6–V5 epitope between Sal1 and EcoR1 sites. This fragment was used to replace the Sal1–EcoR1 fragment of the original pGREG533 vector that contains the GAL1 promoter, the linker and the 7× HA tag. This replacement yielded the (empty) pNAG3 plasmid containing the GAL1 promoter directly followed by the His6–V5 tag. Some YPC1 alleles in the pGREG533 vector were transferred into pNAG3 by cutting with EcoR1/Xho1 and ligating the fragment into similarly digested pNAG3. Mutations of the natural cysteine residues or Ypc1p to alanine or glycine and the replacement of natural amino acids by cysteine were done using the QuikChange site-directed mutagenesis kit (Stratagene). All mutations and constructs were verified by sequencing at Microsynth.

Preparation of microsomes

The microsomes used for protease accessibility, SCAM™ assays and TNBS (2,4,6-trinitrobenzenesulfonic acid) derivatization were generated from cells cultured in Gal at 30°C overnight to an A600 of 3 using needle disruption of spheroplasts as described previously [23]. At the end of the incubation, the microsomes were resuspended in small volumes of buffers A, B or C, depending on the experiment: for the protease accessibility assays buffer A (0.2 M sorbitol, 3 mM EDTA, 5 mM MgCl2 and 0.1 M potassium phosphate, pH 7.4) was used, for the SCAM™ studies buffer B (0.2 M sorbitol, 5 mM MgCl2 and 0.1 M sodium phosphate, pH 7.2) was used and for the lysine derivatization using TNBS the microsomes were resuspended in buffer C (0.2 M sorbitol, 3 mM EDTA, 5 mM MgCl2 and 0.1 M sodium bicarbonate, pH 8.5). Protein concentrations were determined using the Bradford reagent.

Protease protection experiments

The 2∆.YPC1 cells were grown in YNBGal to stationary phase and microsomes were prepared. Aliquots (100 μg of protein/sample) were treated with proteinase K in buffer A for 30 min at room temperature (25°C). Reactions were stopped by the addition of 1× Roche complete EDTA-free protease inhibitor cocktail and Pefabloc, EGTA and PMSF to final concentrations of 20, 5 and 20 mM respectively. After 10 min the proteins were precipitated by TCA (trichloroacetic acid) and processed for Western blotting.

Gel electrophoresis and Western blotting

Samples were incubated for 60 min at room temperature in reducing Laemmli sample buffer and separated by SDS/PAGE (8% gel). Proteins were transferred on to a PVDF membrane using a transfer buffer (25 mM Tris, 192 mM glycine, pH 8.3, and 10% methanol).

Cysteine accessibility (SCAM™) assays

Plasmid-born cysteine-mutated epitope-tagged alleles of Ypc1p were expressed in the yy∆∆ strain (FBY1224) or the yy∆∆ strain containing FLAG–Gpi8p (GPI anchor biosynthesis 8; FBY8224), Gpi8p being a better lumenal control than Kar2p (karyogamy 2) [25]. Microsomes were analysed using UBI-mal [ubiquitin–EMCS (N-ϵ-malemidocaproyl-sulfoxysuccinimide ester)] and PEG-mal [poly(ethylene glycol) 5000-maleimide] in the presence or absence of detergents as described recently [26]. After incubation in presence of UBI-mal (0.1 mM) or PEG-mal (0.5 mM) for 30 min on iced water, 40 mM DTT (dithiothreitol) was added. After 10 min on iced water the derivatization of the target proteins was visualized by SDS/PAGE (10% gel) and Western blotting using anti-V5, anti-FLAG and anti-Kar2p antibodies for detection of Ypc1p, Gpi8p and Kar2p respectively. Alternatively, we performed SCAM™ assays using the NEM adapting protocols described previously [27,28]. For this, microsomes (100 μg of protein) were incubated with NEM for 30 min in iced water unless otherwise indicated. Thereafter, the residual NEM was removed by washing the membranes twice in buffer B by sedimentation. Microsomes were then incubated on iced water with 1 mM PEG-mal, except for the negative controls or where indicated otherwise. After 10 min SDS (0.5% final) or DDM (dodecyl maltoside) were added, and the samples were further incubated with PEG-mal for another 25 min, at room temperature for the samples containing SDS and at 0°C for the others. The samples were then quenched with DTT and processed for Western blotting as described above.

Microsome preparation for reverse ceramidase assay

Microsomes were generated roughly as described previously [14]. In short, cells were grown overnight at 30°C in Gal to an A600 of 2. Microsomes were prepared by glass-bead disruption of cells in buffer D (25 mM Tris/HCl, pH 7.4, and 5 mM CaCl2) containing 1 mM PMSF and 1× Roche complete protease inhibitor cocktail. After removal of the cell wall debris by low speed centrifugation, microsomes were sedimented at 16000 g for 30 min at 4°C and resuspended in buffer E (25 mM Tris/HCl, pH 8.0, and 5 mM CaCl2). Solubilization was performed with 0.5% Triton X-100 on ice for 60 min, the lysates were centrifuged at 16000 g for 30 min at 4°C and the supernatant was used for the assays.

Reverse ceramidase assay

For in vitro assessment of the enzymatic activity of ypc1 alleles, we chose to measure reverse ceramidase activity, which can be assayed using commercially available and relatively water-soluble substrates. We used a method described previously [14] with minor modifications. Assays using microsomes from 2∆.YPC1 cells contained 5 nmol PHS and 2 μCi [3H]palmitic acid (0.3 nmol total). Assays of microsomes from yy∆∆ cells having different Ypc1p alleles contained 10 nmol PHS and 10 μCi [3H]palmitic acid (0.23 nmol total). These ingredients were dried under vacuum in a rotary evaporator and the dried lipids were dissolved in buffer E by water bath sonication. Reactions were started by the addition of microsomes or microsomal detergent extract. Microsomal protein from 2∆.YPC1 and yy∆∆.YPC1 cells, 100 and 10 μg respectively, were used for an assay based on preliminary experiments (Supplementary Figure S2 at http://www.biochemj.org/bj/452/bj4520585add.htm). Volumes were increased to 200 μl and the samples were incubated at 30°C for 120 min at 300 rev./min on a tube shaker. The reactions were terminated by adding 780 μl of chloroform/methanol [2:1 (v/v)]. Lipids were extracted, desalted and mild-base treated as described recently [29]. Aliquots of each sample were applied to Silica gel 60 plates, which were developed by ascending TLC in CHCl3/CH3OH/25% ammonium hydroxide (9:2:0.5). Radioactivity was visualized and quantified by two-dimensional radioscanning using a Berthold® radioscanner. Radioactivity contained in N-[3H]palmitoyl-PHS was expressed as a percentage of the total radioactivity in the TLC lane. Means and standard deviations of all values obtained from two to four independent experiments performed in duplicate were calculated.

DTR (dual topology reporter) analysis

DTRs were added to the C-terminus of C-terminally truncated versions of YPC1 using homologous recombination and the constructs were analysed as described previously [26,30].

Lysine derivatization with TNBS

Microsomes from 2∆.YPC1 (FBY8183) cells (100 μg/sample) in buffer C were left in iced water for 15 min and centrifuged at 16000 g for 30 min at 4°C. After a second wash, the microsomes were again resuspended in buffer C and incubated with or without Triton X-100 (0.1%) for 15 min in iced water. After the addition of TNBS (0.2%) or buffer the tubes were further incubated on iced water for 30 min. Thereafter TNBS was quenched by the addition of 50 mM Tris (pH 8.8), 50 mM imidazole and 100 mM lysine. After 40 min, TNBS-treated membranes or lysates were either used for a reverse ceramidase assay as described above or treated with proteinase K for 30 min at 0°C. Thereafter proteinase K was blocked with inhibitors and TCA precipitation and proteins processed for Western blotting as described above.

RESULTS

The cysteine residues at positions 27 and 219 are required for reverse ceramidase activity

SCAM™ is a variation of the SCAM (substituted cysteine accessibility method) adapted for determining the transmembrane segment orientation of polytopic membrane proteins [3134]. When polytopic membrane proteins of microsomes are studied by SCAM™, the reaction of cysteine residues with an non-permeating reagent indicate that they are in contact with water and located at the cytosolic side. To reduce ambiguities it is ideal to analyse alleles that have a single or only few cysteine residues and still preserve good functionality. Ypc1p contains five cysteines. Removing the N-terminal 35 or C-terminal 74 amino acids of Ypc1p, thereby removing Cys27 or Cys271, yielded non-functional ypc1 alleles that could not rescue YPK9 lag1∆ lac1∆ ypc1∆ ydc1∆ (4∆) cells (Figure 2A). All of the 42 members of the Pfam PF05875 ceramidase family conserve Cys219 of Ypc1p. Indeed, replacing Cys219, as well as Cys27, of Ypc1p with an alanine residue created non-functional alleles (Figure 2B). On the other hand, replacement of the three remaining cysteine residues (Cys107, Cys115 and Cys271) created ypc1 alleles that still could rescue YPK9 4∆ cells. Microsomal reverse ceramidase activity assays using such alleles confirmed that only Cys27 and Cys219 are required for enzymatic activity (Figure 2C). Treating microsomes with the cysteine-residue-specific membrane-permeating alkylating agent NEM abolished the reverse ceramidase activity of wt Ypc1p, whereas the activity of an allele retaining only the two essential cysteines Cys27 and Cys219 was NEM resistant. This indicates that derivatization of one or several of the three non-essential cysteine residues interferes with the activity.

The cysteine residues at positions 27 and 219 are required for reverse ceramidase activity

Figure 2
The cysteine residues at positions 27 and 219 are required for reverse ceramidase activity

(A) The YPK9 4∆.LAG1 (FBY2171) strain was transfected with wt YPC1 or ypc1 alleles having N- or C-terminal truncations up to the indicated amino acids. Dilutions of cells (10-fold) were plated on to YNBGal medium with or without FOA. In the presence of FOA, only cells having lost the pBM150-LAG1 (URA3) plasmid can survive. Plates were incubated for 3 days at 30°C. (B) The same strain was transfected with ypc1 alleles in which part of the five natural cysteine residues were replaced by alanine or glycine. Their exact positions are indicated at the top with positions containing conserved cysteine residues being in bold. The functionality of the alleles was tested as in (A). (C) Some of the alleles tested in (B) were transfected into the yy∆∆ strain (FBY1224) and microsomes were used for an in vitro microsomal reverse ceramidase assay. Samples were run on TLC, radioactivity was detected by radioscanning and N-[3H]palmitoyl–PHS expressed as the percentage of the total counts per lane. Results are means±S.D. for two independent assays performed in duplicate. (D) Microsomes (100 μg) from yy∆∆ cells harbouring the wt CCCCC (FBY8203) or the CGGCG allele (FBY8204) (bold indicates conserved cysteine residues) of YPC1 were incubated with 15 mM NEM or ethanol as a control for 30 min in iced water. Microsomes were sedimented (16000 g for 30 min) at 4°C. The pellet was resuspended in 200 μl of buffer E containing 20 mM DTT and further incubated for 10 min in iced water. A total of 10 μg per sample was used for Ypc1p reverse activity assays. The data were analysed as in (C) and shown as means±S.D. for three independent assays performed in duplicate.

Figure 2
The cysteine residues at positions 27 and 219 are required for reverse ceramidase activity

(A) The YPK9 4∆.LAG1 (FBY2171) strain was transfected with wt YPC1 or ypc1 alleles having N- or C-terminal truncations up to the indicated amino acids. Dilutions of cells (10-fold) were plated on to YNBGal medium with or without FOA. In the presence of FOA, only cells having lost the pBM150-LAG1 (URA3) plasmid can survive. Plates were incubated for 3 days at 30°C. (B) The same strain was transfected with ypc1 alleles in which part of the five natural cysteine residues were replaced by alanine or glycine. Their exact positions are indicated at the top with positions containing conserved cysteine residues being in bold. The functionality of the alleles was tested as in (A). (C) Some of the alleles tested in (B) were transfected into the yy∆∆ strain (FBY1224) and microsomes were used for an in vitro microsomal reverse ceramidase assay. Samples were run on TLC, radioactivity was detected by radioscanning and N-[3H]palmitoyl–PHS expressed as the percentage of the total counts per lane. Results are means±S.D. for two independent assays performed in duplicate. (D) Microsomes (100 μg) from yy∆∆ cells harbouring the wt CCCCC (FBY8203) or the CGGCG allele (FBY8204) (bold indicates conserved cysteine residues) of YPC1 were incubated with 15 mM NEM or ethanol as a control for 30 min in iced water. Microsomes were sedimented (16000 g for 30 min) at 4°C. The pellet was resuspended in 200 μl of buffer E containing 20 mM DTT and further incubated for 10 min in iced water. A total of 10 μg per sample was used for Ypc1p reverse activity assays. The data were analysed as in (C) and shown as means±S.D. for three independent assays performed in duplicate.

The N-terminus of Ypc1p is in the ER lumen

In the absence of detergent, microsomal HA–Ypc1p was quite resistant to proteinase K (Figure 3A), although treatment with very high concentrations of proteinase K led to the appearance of a degradation product of lower molecular mass (results not shown). As N-terminal tags often are not folded into the protein core, these results suggested that the N-terminus of this fully functional allele resides in the ER lumen. To confirm this notion, using SCAM™ we analysed a cysteine residue-free HA–Ypc1p version carrying two cysteines in the linker region between the N-terminal 7× HA tag and the Ypc1p sequence. For the SCAM™ analysis we used a novel mass-tagged maleimide derivative consisting of ubiquitin coupled to the lysine/cysteine heterobifunctional cross-linker EMCS, UBI-mal [26,34]. The two cysteine residues in the linker region of HA–Ypc1p were accessible to UBI-mal only in the presence of mild detergent and not in its absence (Figure 3B, lanes 2 and 3). This result also supports the notion that the N-terminus of Ypc1p is ER lumenal.

The N-terminus of Ypc1p is in the lumen of the ER

Figure 3
The N-terminus of Ypc1p is in the lumen of the ER

(A) Microsomes from 2∆.YPC1 cells containing HA–Ypc1p with a 7× HA tag at the N-terminus were incubated without or with the indicated amounts of proteinase K (PK) in the presence or absence of Triton X-100 (TX100; 0.5%) for 30 min at room temperature and the products were analysed by Western blotting using anti-HA and anti-Kar2p antibodies. Molecular mass is shown on the right-hand side in kDa. (B) Microsomes from yy∆∆ cells harbouring the pBF849 vector were used. pBF849 carries a cysteine residue-free (AGGAG) HA–Ypc1p allele with two non-natural cysteines residing in the linker between the 7× HA tag and YPC1. These non-natural cysteine residues were localized by SCAM™, incubating samples with UBI-mal (UBIm) in the presence of 0.66% DDM (D) or 0.16% SDS (S). Molecular mass is shown on the left-hand side in kDa.

Figure 3
The N-terminus of Ypc1p is in the lumen of the ER

(A) Microsomes from 2∆.YPC1 cells containing HA–Ypc1p with a 7× HA tag at the N-terminus were incubated without or with the indicated amounts of proteinase K (PK) in the presence or absence of Triton X-100 (TX100; 0.5%) for 30 min at room temperature and the products were analysed by Western blotting using anti-HA and anti-Kar2p antibodies. Molecular mass is shown on the right-hand side in kDa. (B) Microsomes from yy∆∆ cells harbouring the pBF849 vector were used. pBF849 carries a cysteine residue-free (AGGAG) HA–Ypc1p allele with two non-natural cysteines residing in the linker between the 7× HA tag and YPC1. These non-natural cysteine residues were localized by SCAM™, incubating samples with UBI-mal (UBIm) in the presence of 0.66% DDM (D) or 0.16% SDS (S). Molecular mass is shown on the left-hand side in kDa.

Cys271 is the only accessible cysteine residue in wt Ypc1p

SCAM™ analysis of microsomes containing His6–V5-tagged wt Ypc1p showed that amongst its five cysteine residues only one reacts with UBI-mal in the absence of detergent (Figure 4A, lane 2). This cysteine residue must be located in a cytosolic loop. In the presence of the mild detergent DDM, a second cysteine residue was also slightly derivatized. Similarly, the use of PEG-mal as a mass tag showed that no more than one cysteine residue of wt Ypc1p was readily accessible, whether or not DDM was added (Figure 4A, lanes 6–8). The addition of SDS revealed that a maximum of three out of the five cysteine residues of Ypc1p could be derivatized (Figure 4A, lane 5). Kar2p, a soluble Hsp70 (heat-shock protein 70) of the ER lumen has a single cysteine residue and served as a control. It showed that lumenal cysteine residues became accessible only in the presence of DDM or SDS, but not of digitonin (Figure 4A, lower panel). As reported previously, the single cysteine residue of Kar2p is poorly accessible in the native protein, but becomes more accessible in the presence of the denaturing detergent SDS (Figure 4A, lanes 3, 5 and 8) [25,34]. We also used an alternative approach which avoids the use of mild detergents as a means to remove the membrane barrier. For this approach, microsomes (containing the proteins in their native state) were exposed to the membrane-permeating NEM so that the cysteine residues on both sides of the membrane were alkylated. Residual NEM was the removed and a mass-tagged maleimide derivative (UBI-mal or PEG-mal) added in the presence of the denaturating detergent SDS [27,28]. Those cysteine residues that are not alkylated by NEM when the protein is in the native state, but which get derivatized by a mass-tagged maleimide after denaturation, are classified as ‘buried’ in the native protein structure. Only those cysteine residues that in the native structure are accessible to both NEM and water will not score as buried, since the alkylation reaction of cysteines by NEM also requires the presence of water. As for the wt Ypc1p, this alternative approach again showed that UBI-mal derivatized up to three cysteine residues in the presence of SDS, but only a single cysteine in the absence of detergent (Figure 4B, lanes 1, 2 and 6). Pre-treatment with 1 mM NEM readily abolished the reactivity of this latter cysteine residue (Figure 4B, lane 3), but not of the two other cysteines (Figure 4B, lane 4). These two buried cysteines also remained inaccessible to higher concentrations of NEM (Figure 4B, lanes 5 and 8). Thus the two SCAM™ approaches used in Figures 4(A) and 4(B) gave the same result: of the five cysteine residues of Ypc1p, one is in a cytosolic loop, two are buried and the last two are derivatized under no circumstance.

Cys27 and Cys219 of Ypc1p form a disulfide bridge

Figure 4
Cys27 and Cys219 of Ypc1p form a disulfide bridge

(A) Microsomes from yy∆∆ cells transfected with wt His6–V5–YPC1 were treated with UBI-mal (UBIm) or PEG-mal (PEGm) in the presence of buffer (−), digitonin (Dt; 0.66%), DDM (D; 0.66%) or SDS (S; 0.16%). (B) The same microsomes were pre-treated with 1, 5 or 10 mM NEM at 0°C and then washed twice before being treated with UBI-mal (0.1 mM) in the presence or absence of SDS (0.5%). (C and D) As (A), but using microsomes containing the cysteine residue-free His6–V5–Ypc1p (C) or a His6–V5–Ypc1p lacking Cys271 (D). (E) Microsomes from yy∆∆ cells having His6–V5–Ypc1p-27-219 (pBF85) were solubilized in 0.5% SDS in the presence and absence of DTT (25 mM) for 60 min at room temperature. Upon the removal of DTT by ultrafiltration the lysates were treated with 0, 1, 3 or 5 mM PEG-mal. The protein extracts were processed as above. Molecular mass is shown on the left-hand side in kDa.

Figure 4
Cys27 and Cys219 of Ypc1p form a disulfide bridge

(A) Microsomes from yy∆∆ cells transfected with wt His6–V5–YPC1 were treated with UBI-mal (UBIm) or PEG-mal (PEGm) in the presence of buffer (−), digitonin (Dt; 0.66%), DDM (D; 0.66%) or SDS (S; 0.16%). (B) The same microsomes were pre-treated with 1, 5 or 10 mM NEM at 0°C and then washed twice before being treated with UBI-mal (0.1 mM) in the presence or absence of SDS (0.5%). (C and D) As (A), but using microsomes containing the cysteine residue-free His6–V5–Ypc1p (C) or a His6–V5–Ypc1p lacking Cys271 (D). (E) Microsomes from yy∆∆ cells having His6–V5–Ypc1p-27-219 (pBF85) were solubilized in 0.5% SDS in the presence and absence of DTT (25 mM) for 60 min at room temperature. Upon the removal of DTT by ultrafiltration the lysates were treated with 0, 1, 3 or 5 mM PEG-mal. The protein extracts were processed as above. Molecular mass is shown on the left-hand side in kDa.

As expected, a non-functional cysteine-free version of Ypc1p was derivatized neither by UBI-mal nor by PEG-mal, and the presence of detergents also did not uncover any reactive residues (Figure 4C). A ypc1 allele in which a glycine residue replaced Cys271 could not be derivatized in the absence or presence of a mild detergent (DDM; Figure 4D, lanes 1–3). Denaturation with SDS, however, rendered up to two cysteine residues accessible (Figure 4D, lane 4). The result suggested that the one cytosolic cysteine residue of wt Ypc1p, which is accessible to UBI-mal or PEG-mal in the absence of any detergent (Figure 4A, lanes 1, 2 and 7), is Cys271.

Cys27 and Cys219 form a disulfide link

Two cysteine residues remained inaccessible both in the wt and the C271G allele even in presence of SDS (Figures 2A, 2B and 2D) and we guessed that this may be owing to the formation of a disulfide bond between those two cysteines that are essential for protein function, Cys27 and Cys219. Indeed, a ypc1-27-219 allele retaining only these two cysteines (CGGCG) could not be derivatized with PEG-mal even after denaturation with SDS (Figure 4E, lanes 1, 3, 5 and 7). On the other hand, after prior reduction with DDT these two cysteine residues became accessible (Figure 4E, lanes 2, 4, 6 and 8). This strongly suggests that a disulfide bond links these two cysteine residues in Ypc1p-(27–219) and also in wt Ypc1p and supports the notion that the N-terminus of Ypc1p is located in the ER lumen. Generally speaking, Cys27 and Cys219 both are absolutely conserved in the 42 founding members of the pfam05875 motif (Figure 9), suggesting that the disulfide bridge is conserved throughout the ceramidase family.

Topology of loops L1–2 and L2–3 of Ypc1p

For further topological studies we decided to substitute individual residues in the predicted cytosolic or lumenal loops (Supplementary Figure S1) with a cysteine in the Ypc1p allele retaining only the two disulfide-linked cysteines, Cys27 and Cys219 (ypc1-27-219). For this we preferentially chose non-conserved residues in regions for which the secondary structure was predicted to be a random coil. The functionality of the constructs was tested by introducing them into 4∆.LAG1, a lag1∆ lac1∆ ypc1∆ ydc1∆ strain, which dies when it loses the LAG1-containing URA3 plasmid pBM150-LAG1. As shown in Supplementary Figure S3 (at http://www.biochemj.org/bj/452/bj4520585add.htm), all substitutions created functional alleles, which rescued growth of 4∆ cells on FOA with the exception of the K67C and G235C alleles (FOA kills cells retaining the covering pBM150-LAG1 plasmid). Moreover, the ypc1-27-91-219 allele carrying the K91C mutation provided only limited growth to 4∆ cells. It should be noted that the rescue of 4∆ cells requires massive overexpression of YPC1 [5,16], so that any reduction in the catalytic activity of an allele is expected to result in reduced growth of 4∆ cells.

To determine the topology of loops between TM1 and TM2 (loop L1–2) we analysed microsomes from yy∆∆ cells harbouring a ypc1-27-65-219 or ypc1-27-67-219 allele using two approaches, opening membranes with mild detergent as in Figure 4(A) or pre-treating them with NEM as in Figure 4(B). Cysteine residues in positions 65 and 67 were not directly accessible when the microsomes were incubated with PEG-mal, but became accessible in the presence of a mild detergent (Figure 5B, lane 2 compared with lane 7 and Figure 5C, lane 2 compared with lane 12). Also, NEM is able to prevent their subsequent derivatization with PEG-mal, suggesting that they are not buried, but exposed at the ER lumenal side of the membrane (Figure 5B, lane 4 and Figure 5C, lanes 3–10). As a control, in most of the SCAM™ experiments described below we used a FLAG-tagged Gpi8p allele. Gpi8p is the core protease of the 5-subunit GPI transamidase complex. It is a 50 kDa type I membrane glycoprotein with its N-terminus and all of its four cysteine residues on the lumenal side of the ER membrane and only 14 amino acids being displayed at the cytosolic side [35]. Up to four cysteine residues were accessible to PEG-mal when Gpi8p was denatured with SDS (at 25°C) and up to three when the membrane was permeabilized by DDM at 0°C (Figures 5B and 5E, lanes 6 and 7). Pre-treatment with NEM confirmed that only one cysteine residue of Gpi8p is buried inside the protein and cannot be reached by 10 mM NEM at 0°C, although at a higher temperature (25°C) even this cysteine becomes accessible to NEM (Figure 5B and 5E, lanes 4 and 5). As these controls indicate that NEM efficiently penetrated into the ER lumen, the data shown in Figures 5(B) and 5(C) suggest that L1–2 is lumenal. The same pattern, characteristic for lumenally exposed cysteine residues, was also observed for two cysteines introduced at positions 91 and 93 in loop L2–3, a very short loop comprising only amino acids 91–94 according to the current TOPCONS global prediction (Figures 5A, 5E and 5F). This prediction also places Thr89 at the end of TM2 and, indeed, we found the T89C mutant cysteine of ypc1-27-89-219 allele was buried, although it became partially accessible to NEM at 25°C (Figure 5D, lanes 4 and 5). Overall, the data clearly indicate that loops L1–2 and L2–3 are lumenal.

Topology of loops L1–2 and L2–3

Figure 5
Topology of loops L1–2 and L2–3

(A) TM predictions for the first 130 amino acids of Ypc1p by the TOPCONS global prediction and the one proposed by the present report (Supplementary Figure S1, line j and line i respectively at http://www.biochemj.org/bj/452/bj4520585add.htm). The positions of amino acids replaced by cysteines are indicated as open circles. (BF) Microsomes from FBY8228 (B), FBY8229 (C), FBY8263 (D), FBY8230 (E) and FBY8231 (F) were incubated for 30 min with 0–20 mM NEM at the indicated temperatures and processed for SCAM™. Molecular mass is shown on the left-hand side in kDa. D, DDM; PEGm, PEG-mal; S, SDS.

Figure 5
Topology of loops L1–2 and L2–3

(A) TM predictions for the first 130 amino acids of Ypc1p by the TOPCONS global prediction and the one proposed by the present report (Supplementary Figure S1, line j and line i respectively at http://www.biochemj.org/bj/452/bj4520585add.htm). The positions of amino acids replaced by cysteines are indicated as open circles. (BF) Microsomes from FBY8228 (B), FBY8229 (C), FBY8263 (D), FBY8230 (E) and FBY8231 (F) were incubated for 30 min with 0–20 mM NEM at the indicated temperatures and processed for SCAM™. Molecular mass is shown on the left-hand side in kDa. D, DDM; PEGm, PEG-mal; S, SDS.

Topology of loops L3–4, L4–5, L5–6 and L6–7 of Ypc1p

Using the same approach as described above we found that cysteine residues that replaced Asn123 and Gly124 were both immediately accessible to PEG-mal in the absence of any detergent (Figures 6B and 6C). These results show that L3–4 is cytosolic (Figure 6A).

Topology of loops L3–4 and L4–5

Figure 6
Topology of loops L3–4 and L4–5

(A) TM predictions for amino acids 90–180 of Ypc1p as in Figure 5(A). (BE) Microsomes from FBY8233 (B), FBY8234 (C), FBY8264 (D) and FBY8235 (E) cells were incubated for 30 min with 0 or 10 mM NEM at 0°C and processed for SCAM™. Molecular mass is shown on the left-hand side in kDa. D, DDM; PEGm, PEG-mal; S, SDS.

Figure 6
Topology of loops L3–4 and L4–5

(A) TM predictions for amino acids 90–180 of Ypc1p as in Figure 5(A). (BE) Microsomes from FBY8233 (B), FBY8234 (C), FBY8264 (D) and FBY8235 (E) cells were incubated for 30 min with 0 or 10 mM NEM at 0°C and processed for SCAM™. Molecular mass is shown on the left-hand side in kDa. D, DDM; PEGm, PEG-mal; S, SDS.

Loop L4–5 is very short and comprises only amino acids 157–159 according to TOPCONS (Figure 6A). It was probed by introducing cysteine residues at positions 157 and 158. These cysteine residues were not accessible in the absence of detergent, but were accessible to NEM, arguing that they are lumenally exposed (Figures 6D and 6E).

Loop L5–6 comprises amino acids 181–195 according to TOPCONS and was probed by substituting cysteine residues at positions 181, 182, 184 and 185, all of which seemed to be located at the lumenal side of the ER (Figures 7A–7E). Ser197 which is located at the beginning of the following TM6 proved to be completely buried (Figure 7F). Overall, the data indicate a lumenal orientation of L5–6.

Topology of loops L5–6 and L6–7

Figure 7
Topology of loops L5–6 and L6–7

(A) TM predictions for amino acids 160–270 of Ypc1p as in Figure 5(A). (BH) Microsomes from FBY8236 (B), FBY8237 (C), FBY8238 (D), FBY8239 (E), FBY8240 (F), FBY8241 (G) and FBY8242 (H) cells were incubated for 30 min with 0–20 mM NEM at the indicated temperatures and processed for SCAM™. Molecular mass is shown on the left-hand side in kDa. D, DDM; PEGm, PED-mal; S, SDS. (I) DTR constructs containing amino acids 1–233 (P233) 1–292 (P292) of Ypc1p followed by the DTR (3× HA–Suc2–His4C cassette) were transfected into his4∆ (STY50) cells. Left-hand panel, serial 10-fold dilutions of cells were plated on minimal medium containing either histidine or its precursor histidinol (His4∆ grow on histidinol, if the DTR cassette is on the cytosolic side of the ER membrane and not if it is lumenal). Right-hand panel, proteins of exponentially growing cells were extracted, incubated with (+) or without (−) endoglycosidase H (EndoH), separated by SDS/PAGE and processed for Western blotting using an anti-HA antibody.

Figure 7
Topology of loops L5–6 and L6–7

(A) TM predictions for amino acids 160–270 of Ypc1p as in Figure 5(A). (BH) Microsomes from FBY8236 (B), FBY8237 (C), FBY8238 (D), FBY8239 (E), FBY8240 (F), FBY8241 (G) and FBY8242 (H) cells were incubated for 30 min with 0–20 mM NEM at the indicated temperatures and processed for SCAM™. Molecular mass is shown on the left-hand side in kDa. D, DDM; PEGm, PED-mal; S, SDS. (I) DTR constructs containing amino acids 1–233 (P233) 1–292 (P292) of Ypc1p followed by the DTR (3× HA–Suc2–His4C cassette) were transfected into his4∆ (STY50) cells. Left-hand panel, serial 10-fold dilutions of cells were plated on minimal medium containing either histidine or its precursor histidinol (His4∆ grow on histidinol, if the DTR cassette is on the cytosolic side of the ER membrane and not if it is lumenal). Right-hand panel, proteins of exponentially growing cells were extracted, incubated with (+) or without (−) endoglycosidase H (EndoH), separated by SDS/PAGE and processed for Western blotting using an anti-HA antibody.

Loop L6–7 comprises, according to TOPCONS, amino acids 217–245. The ypc1-27-219-235 allele is non-functional, but SCAM™ analyses indicated position 235 to be lumenal (Figure 7G). The ypc1-27-219-236 allele was classified as partially buried, since part of the protein could still be derivatized with PEG-mal after NEM treatment (Figure 7H, lanes 4 and 6). NEM pre-treatment performed at 25 instead of 0°C, or a NEM concentration raised to 20 mM, did not increase the accessibility of Cys236 (Figure 7H, lanes 4–6, 8 and 9). Another way to look at the data is to say that the PEG-mal derivatization of Cys236 was substantially reduced by NEM pre-treatment, indicating that it is partially accessible to NEM (Figure 7H, lanes 2 and 4–6). Similarly Cys236 was also partially accessible to PEG-mal in the presence of DDM (Figure 7H, lane 7), whereas it was not accessible to PEG-mal in the absence of detergent (Figure 7H, lane 2). The result is compatible with the idea that there are two forms of Ypc1p, one in which Cys236 is accessible from the lumenal side of the ER and another where this cysteine residue is buried, e.g. because another protein interacts with Ypc1p. The lumenal location of L6–7 (Figure 7A) is supported also by a DTR inserted after Pro233 showing that the Ypc1p-P233-DTR construct is glycosylated and does not support growth of his4∆ cells on histidinol (Figure 7I). The DTR approach has been shown to be able to yield a correct topology prediction only when the DTR is inserted at a certain distance downstream of a TM with a negative ∆Gmi value and of sufficient length to span the membrane [34,36]. These conditions were met for the DTR cassettes truncating Ypc1p after TM6. The existence of a disulfide bond between Cys27 and Cys219 also supports a lumenal orientation of L6–7 (Figure 4E).

The C-terminal end of Ypc1p is cytosolic

The TOPCONS global prediction places amino acids 267–316 into the cytosol. Indeed, a global study to locate the C-terminal ends of yeast membrane proteins using DTR has mapped the C-terminal end of Ypc1p to the cytosol [37]. This conclusion is fully supported also by the results obtained with the Ypc1p-P292-DTR construct shown in Figure 7(I). It is further supported by SCAM™ data showing a cytosolic location of the natural Cys271 (Figure 4 and Supplementary Figure S4 at http://www.biochemj.org/bj/452/bj4520585add.htm). When generating the ypc1-27-184-219 allele the PCR mutagenesis wrongly introduced an additional mutation, S293C. This ypc1-27-184-219-293 allele contains one cysteine residue, which reacts with PEG-mal in the absence of detergent (Supplementary Figure S4A, lane 2), whereas the corrected ypc127-184-219 does not react in this condition (Figure 7D, lane 2) and thus Cys293 must be cytosolic. Similarly, probing the allele with the S297C mutation showed that position 297 is accessible from the cytosolic side of the membrane (Supplementary Figure S4B). Overall, all of the data argue that the C-terminus of Ypc1p is cytosolic, allowing its C-terminal KKEK motif to act as an ER retrieval signal.

Lysine derivatization in the absence of detergent does not interfere with Ypc1p activity

TNBS is a lysine-reactive reagent, which does not penetrate the membrane. When TNBS was added to intact microsomes, the reverse ceramidase activity remained unchanged (Figures 8A and 8B, lanes 1 and 3). The addition of TNBS to an assay containing Triton X-100, however, caused a drastic reduction in Ypc1p activity (Figures 8A and 8B lanes 2 and 4). When, before being added to microsomes, TNBS was quenched with a mix of lysine, imidazol and Tris, it did not inhibit reverse ceramidase activity (Figure 8A, lanes 3, 4, 6 and 7). This activity did not decrease even when microsomes were treated with TNBS for up to 2 h (results not shown). TNBS-derivatized proteins were detected by Western blotting using an anti-DNP antibody (Supplementary Figure S5A at http://www.biochemj.org/bj/452/bj4520585add.htm). To localize the TNBS derivatization on microsomal proteins, TNBS-treated microsomes were treated with proteinase K. When derivatization was performed in the presence of Triton X-100, all proteins were almost completely digested by the subsequent addition of proteinase K (Supplementary Figure S5A, lanes 3, 6 and 7). Even when TNBS derivatization was performed in the absence of detergent, a large part of the TNBS-derivatized proteins were destroyed by proteinase K (Supplementary Figure S5A, lanes 2, 4 and 5), whereas 74% of the ER lumenal Kar2p and 80% of the lumenally tagged HA–Ypc1p were preserved (Supplementary Figures S5B and S5C). This indicates that TNBS does not efficiently penetrate the ER membrane. The resistance of Kar2p and HA–Ypc1p to protease added from the cytosolic side is not owing to the protection of lysine residues by TNBS, since these proteins were even more resistant when TNBS treatment was omitted (Supplementary Figure S5A–S5C, lanes 4, 5 and 8–11). Incidentally, the results of the present study show again that Ypc1p is highly resistant to protease added from the cytosolic side of microsomes, whereas it is readily degraded when detergent is present (Figure 3A and Supplementary Figure S5C). Ypc1p contains 16 lysine residues, nine of which are in cytosolic loops, whereas the other seven are in lumenal loops, according to our model (Figure 9). The data argue that the derivatization of cytosolic lysines in the absence of detergent does not interfere with reverse ceramidase activity, whereas derivatization of some lysine in the presence of detergent does, although no single lysine residue of Ypc1p is conserved.

Ypc1p activity is blocked by TNBS in the presence of detergent

Figure 8
Ypc1p activity is blocked by TNBS in the presence of detergent

(A) Microsomes from 2∆.YPC1 cells were incubated with or without TNBS (0.2%) in the presence or absence of Triton X-100 (TX100; 0.1%) in iced water. Thereafter TNBS was quenched and ingredients for the reverse ceramidase assay were added. The microsomes in lane 5 were boiled (b). In lanes 6 and 7 TNBS had been quenched before being added to the microsomes. The Berthold radioscans are shown. (B) Quantification of [3H]C16 incorporated into N-[3H]palmitoyl–PHS (Cer) in lanes 1–4 of (A) as a percentage of total the counts per lane, and of three independent repeat assays each performed in duplicate±S.D.

Figure 8
Ypc1p activity is blocked by TNBS in the presence of detergent

(A) Microsomes from 2∆.YPC1 cells were incubated with or without TNBS (0.2%) in the presence or absence of Triton X-100 (TX100; 0.1%) in iced water. Thereafter TNBS was quenched and ingredients for the reverse ceramidase assay were added. The microsomes in lane 5 were boiled (b). In lanes 6 and 7 TNBS had been quenched before being added to the microsomes. The Berthold radioscans are shown. (B) Quantification of [3H]C16 incorporated into N-[3H]palmitoyl–PHS (Cer) in lanes 1–4 of (A) as a percentage of total the counts per lane, and of three independent repeat assays each performed in duplicate±S.D.

Topology model for Ypc1p based on the results of the present study

Figure 9
Topology model for Ypc1p based on the results of the present study

The membrane is shown as a grey box, the 24 residues considered to be highly conserved in Pfam motif pf05875 are in red font, the 13 residues strictly conserved in the same family according to the Conserved Domain Database are in red circles, the five residues conserved in the CREST family are filled in turquoise, and the residues which were substituted for SCAM are filled in yellow if the cysteine substitution resulted in a fully functional allele and violet when substitution compromised the functionality. Non-conserved cysteines are in orange lettering and residues where a DTR was inserted are filled in green.

Figure 9
Topology model for Ypc1p based on the results of the present study

The membrane is shown as a grey box, the 24 residues considered to be highly conserved in Pfam motif pf05875 are in red font, the 13 residues strictly conserved in the same family according to the Conserved Domain Database are in red circles, the five residues conserved in the CREST family are filled in turquoise, and the residues which were substituted for SCAM are filled in yellow if the cysteine substitution resulted in a fully functional allele and violet when substitution compromised the functionality. Non-conserved cysteines are in orange lettering and residues where a DTR was inserted are filled in green.

DISCUSSION

Ceramides are becoming apparent as intermediates in both the biosynthetic and catabolic pathways that insure the subcellular and global sphingolipid and lipid homoeostasis of eukaryotic cells. In yeast, ceramides are generated either through the acyl-CoA-dependent ceramide synthases Lag1p and Lac1p or through degradation of IPCs. Isc1p most certainly generates ceramides from IPCs in the cytosolic leaflet, whereas this has not been proven for the ceramide synthases Lag1p and Lac1p. Ceramides basically can enter four different pathways: (i) IPC biosynthesis [38], (ii) GPI anchor remodelling [39], (iii) 1-O-acylation generating acylceramides [29] and (iv) hydrolysis by ceramidases [14] (Figure 1). Pathways (i)–(iii) have been claimed to occur at the lumenal side of the secretory pathway on the basis of biochemical data and bioinformatic analysis of the location of the conserved motifs [27,29,4043]. The present study makes it probable that the degradation via alkaline ceramidases also occurs lumenally. Thus the ER-based anabolic pathways leading to the integration of ceramides into GPI anchors and acylceramides may compete with ceramidases for ceramides residing in the same leaflet of the bilayer.

A previous study showed that the human alkaline ceramidase ACER2 has its N-terminal end in the lumen and the C-terminus in the cytosol and that the deletion of the N-terminal end, containing residues conserved between yeast and mammalian alkaline ceramidases, abolishes ceramidase activity [19]. The bioinformatic predictions of the membrane topology of ACER2 place the most highly conversed residues of all alkaline ceramidases into the lumen, whereas the C-terminus and three putative cytosolic loops do not contain highly conserved residues. Based on this it was proposed that the presumed active site of ACER2 may be located in the lumen of the Golgi complex [19]. Using bioinformatics tools, this prediction can be extrapolated to the other two human ceramidases [21] (Supplementary Figure S6 at http://www.biochemj.org/bj/452/bj4520585add.htm). Overall, almost all residues conserved between YPC1 and its closest fungal, as well as metazoan, homologues are in or near the ER lumen, according to both our topology model and the global TOPCONS prediction (Supplementary Figure S1). Interestingly, part of the residues conserved in the Pfam pf05875 ceramidase family are also conserved in the larger CREST superfamily encompassing approximately 3000 predicted proteins with seven TMs (Figure 9 and Supplementary Figure S1). Yeast harbours four other proteins belonging to the CREST superfamily, namely the Izh (implicated in zinc homoeostasis) 1–Izh4 proteins, three of which are localized in the ER [44]. Within the CREST superfamily, the Izh1–Izh4 proteins are most closely related to the PAQR family, a family that is predicted to have an inverted topology in that all five conserved residues are predicted to be cytosolic [21]. The global TOPCONS predictions for Izh1p–Izh4p indeed predict a cytosolic location of their conserved residues (Supplementary Figure S6), and these predictions have an uncommonly high reliability, since for each Izh protein all five algorithms (corresponding to Supplementary Figure S1, lines a–e) indicate the exact same topology (results not shown). Interestingly, Izh2p has been shown to be involved in a signalling pathway, which results in the elevation of LCBs, raising the possibility that also this group of proteins acts as ceramidases [17]. If this can be shown biochemically, it would imply that functionally and structurally similar yeast proteins are made with two opposite topologies.

Although the mammalian ceramidase ACER2 has been shown to have the potential of regulating N-glycan modifications and production of sphingosine and sphingosine 1-phosphates in studies where the proteins was overexpressed [19,45], the reasons for the preservation of alkaline ceramidases in fungi is, so far, not apparent and further efforts are required to understand their physiological roles.

Abbreviations

     
  • ACER

    alkaline ceramidase

  •  
  • CREST

    alkaline ceramidase, progestin and adiponectin receptor receptor, protein processing in the endoplasmic reticulum 1, sister disjunction 1 and transmembrane protein 8

  •  
  • DDM

    dodecyl maltoside

  •  
  • DHS

    dihydrosphingosine

  •  
  • DTR

    dual topology reporter

  •  
  • DTT

    dithiothreitol

  •  
  • EMCS

    N-ϵ-malemidocaproyl-sulfoxysuccinimide ester

  •  
  • ER

    endoplasmic reticulum

  •  
  • FOA

    5-fluoro-orotic acid

  •  
  • Gal

    galactose

  •  
  • GPI

    glycosylphosphatidylinositol

  •  
  • Gpi8

    GPI anchor biosynthesis 8

  •  
  • HA

    haemagglutinin

  •  
  • IPC

    inositolphosphorylceramides

  •  
  • Isc1

    inositol phosphosphingolipid phospholipase C1

  •  
  • Izh

    implicated in zinc homoeostasis

  •  
  • Kar2

    karyogamy 2

  •  
  • Lag1

    longevity assurance gene 1

  •  
  • Lac1

    Lag1 cognate 1

  •  
  • LCB

    long-chain base

  •  
  • MIPC

    mannosyl-IPC

  •  
  • NEM

    N-ethylmaleimide

  •  
  • ORF

    open reading frame

  •  
  • PAQR

    progestin and adiponectin receptor

  •  
  • PEG-mal

    poly(ethylene glycol) 5000-maleimide

  •  
  • Per1

    protein processing in the ER 1

  •  
  • PHS

    phytosphingosine

  •  
  • TCA

    trichloroacetic acid

  •  
  • TM

    transmembrane helix

  •  
  • TNBS

    2,4,6-trinitrobenzenesulfonic acid

  •  
  • UBI-mal

    ubiquitin–EMCS

  •  
  • VLCFA

    very-long-chain fatty acid

  •  
  • wt

    wild-type

  •  
  • Ydc1

    yeast dihydroceramidase 1

  •  
  • Ypc1

    yeast phytoceramidase 1

AUTHOR CONTRIBUTION

Nagaraju Ramachandra performed the experimental work, and Nagaraju Ramachandra and Andreas Conzelmann planned the experiments and wrote the paper.

We thank Dr Martin Pagac and Arlette Bochud for helpful discussions.

FUNDING

This work was supported by the Swiss National Science Foundation [grant numbers CRSI33_125232 and 31003A_131078].

References

References
1
Dickson
R. C.
Roles for sphingolipids in Saccharomyces cerevisiae
Adv. Exp. Med. Biol.
2010
, vol. 
688
 (pg. 
217
-
231
)
2
Sawai
H.
Okamoto
Y.
Luberto
C.
Mao
C.
Bielawska
A.
Domae
N.
Hannun
Y. A.
Identification of ISC1 (YER019w) as inositol phosphosphingolipid phospholipase C in Saccharomyces cerevisiae
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
39793
-
39798
)
3
Guillas
I.
Kirchman
P. A.
Chuard
R.
Pfefferli
M.
Jiang
J. C.
Jazwinski
S. M.
Conzelmann
A.
C26-CoA-dependent ceramide synthesis of Saccharomyces cerevisiae is operated by Lag1p and Lac1p
EMBO J.
2001
, vol. 
20
 (pg. 
2655
-
2665
)
4
Jiang
J. C.
Kirchman
P. A.
Zagulski
M.
Hunt
J.
Jazwinski
S. M.
Homologs of the yeast longevity gene LAG1 in Caenorhabditis elegans and human
Genome Res.
1998
, vol. 
8
 (pg. 
1259
-
1272
)
5
Schorling
S.
Vallee
B.
Barz
W. P.
Riezman
H.
Oesterhelt
D.
Lag1p and Lac1p are essential for the acyl-CoA-dependent ceramide synthase reaction in Saccharomyces cerevisae
Mol. Biol. Cell
2001
, vol. 
12
 (pg. 
3417
-
3427
)
6
Winter
E.
Ponting
C. P.
TRAM, LAG1 and CLN8: members of a novel family of lipid-sensing domains?
Trends Biochem. Sci.
2002
, vol. 
27
 (pg. 
381
-
383
)
7
Vallee
B.
Riezman
H.
Lip1p: a novel subunit of acyl-CoA ceramide synthase
EMBO J.
2005
, vol. 
24
 (pg. 
730
-
741
)
8
Barz
W. P.
Walter
P.
Two endoplasmic reticulum (ER) membrane proteins that facilitate ER-to-Golgi transport of glycosylphosphatidylinositol-anchored proteins
Mol. Biol. Cell
1999
, vol. 
10
 (pg. 
1043
-
1059
)
9
Kageyama-Yahara
N.
Riezman
H.
Transmembrane topology of ceramide synthase in yeast
Biochem. J.
2006
, vol. 
398
 (pg. 
585
-
593
)
10
Guillas
I.
Jiang
J. C.
Vionnet
C.
Roubaty
C.
Uldry
D.
Chuard
R.
Wang
J.
Jazwinski
S. M.
Conzelmann
A.
Human homologues of LAG1 reconstitute acyl-CoA-dependent ceramide synthesis in yeast
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
37083
-
37091
)
11
Denic
V.
Weissman
J. S.
A molecular caliper mechanism for determining very long-chain fatty acid length
Cell
2007
, vol. 
130
 (pg. 
663
-
677
)
12
Lopez-Montero
I.
Rodriguez
N.
Cribier
S.
Pohl
A.
Velez
M.
Devaux
P. F.
Rapid transbilayer movement of ceramides in phospholipid vesicles and in human erythrocytes
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
25811
-
25819
)
13
Pohl
A.
Lopez-Montero
I.
Rouviere
F.
Giusti
F.
Devaux
P. F.
Rapid transmembrane diffusion of ceramide and dihydroceramide spin-labelled analogues in the liquid ordered phase
Mol. Membr. Biol.
2009
, vol. 
26
 (pg. 
194
-
204
)
14
Mao
C.
Xu
R.
Bielawska
A.
Obeid
L. M.
Cloning of an alkaline ceramidase from Saccharomyces cerevisiae. An enzyme with reverse (CoA-independent) ceramide synthase activity
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
6876
-
6884
)
15
Mao
C.
Xu
R.
Bielawska
A.
Szulc
Z. M.
Obeid
L. M.
Cloning and characterization of a Saccharomyces cerevisiae alkaline ceramidase with specificity for dihydroceramide
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
31369
-
31378
)
16
Jiang
J. C.
Kirchman
P. A.
Allen
M.
Jazwinski
S. M.
Suppressor analysis points to the subtle role of the LAG1 ceramide synthase gene in determining yeast longevity
Exp. Gerontol.
2004
, vol. 
39
 (pg. 
999
-
1009
)
17
Villa
N. Y.
Kupchak
B. R.
Garitaonandia
I.
Smith
J. L.
Alonso
E.
Alford
C.
Cowart
L. A.
Hannun
Y. A.
Lyons
T. J.
Sphingolipids function as downstream effectors of a fungal PAQR
Mol. Pharmacol.
2009
, vol. 
75
 (pg. 
866
-
875
)
18
Mao
C.
Obeid
L. M.
Ceramidases: regulators of cellular responses mediated by ceramide, sphingosine, and sphingosine-1-phosphate
Biochim. Biophys. Acta
2008
, vol. 
1781
 (pg. 
424
-
434
)
19
Sun
W.
Jin
J.
Xu
R.
Hu
W.
Szulc
Z. M.
Bielawski
J.
Obeid
L. M.
Mao
C.
Substrate specificity, membrane topology, and activity regulation of human alkaline ceramidase 2 (ACER2)
J. Biol. Chem.
2010
, vol. 
285
 (pg. 
8995
-
9007
)
20
Pewzner-Jung
Y.
Ben-Dor
S.
Futerman
A. H.
When do Lasses (longevity assurance genes) become CerS (ceramide synthases)?: insights into the regulation of ceramide synthesis
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
25001
-
25005
)
21
Pei
J.
Millay
D. P.
Olson
E. N.
Grishin
N. V.
CREST: a large and diverse superfamily of putative transmembrane hydrolases
Biol. Direct
2011
, vol. 
6
 pg. 
37
 
22
Sherman
F.
Getting started with yeast
Methods Enzymol.
2002
, vol. 
350
 (pg. 
3
-
41
)
23
Pagac
M.
Vazquez
H. M.
Bochud
A.
Roubaty
C.
Knopfli
C.
Vionnet
C.
Conzelmann
A.
Topology of the microsomal glycerol-3-phosphate acyltransferase Gpt2p/Gat1p of Saccharomyces cerevisiae
Mol. Microbiol.
2012
, vol. 
86
 (pg. 
1156
-
1166
)
23b
Hessa
T.
Meindl-Beinker
N.
Bernsel
A.
Kim
J.
Sato
Y.
Lerch
M.
Lundin
C.
Nilsson
I.
White
SH.
von Heijne
G.
Molecular code for transmembrane-helix recognition by the Sec61 translocon
Nature
2007
, vol. 
450
 (pg. 
1026
-
1030
)
24
Jansen
G.
Wu
C.
Schade
B.
Thomas
D. Y.
Whiteway
M.
Drag&Drop cloning in yeast
Gene
2005
, vol. 
344
 (pg. 
43
-
51
)
25
Bochud
A.
Ramachandra
N.
Conzelmann
A.
Adaptation of low-resolution methods for the study of yeast microsomal polytopic membrane proteins: a methodological review
Biochem. Soc. Trans.
2013
, vol. 
41
 (pg. 
35
-
42
)
26
Pagac
M.
de la Mora
H. V.
Duperrex
C.
Roubaty
C.
Vionnet
C.
Conzelmann
A.
Topology of 1-acyl-sn-glycerol-3-phosphate acyltransferases SLC1 and ALE1 and related membrane-bound O-acyltransferases (MBOATs) of Saccharomyces cerevisiae
J. Biol. Chem.
2011
, vol. 
286
 (pg. 
36438
-
36447
)
27
Liu
Q.
Siloto
R. M.
Snyder
C. L.
Weselake
R. J.
Functional and topological analysis of yeast acyl-CoA:diacylglycerol acyltransferase 2, an endoplasmic reticulum enzyme essential for triacylglycerol biosynthesis
J. Biol. Chem.
2011
, vol. 
286
 (pg. 
13115
-
13126
)
28
Wang
Y.
Toei
M.
Forgac
M.
Analysis of the membrane topology of transmembrane segments in the C-terminal hydrophobic domain of the yeast vacuolar ATPase subunit a (Vph1p) by chemical modification
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
20696
-
20702
)
29
Voynova
N. S.
Vionnet
C.
Ejsing
C. S.
Conzelmann
A.
A novel pathway of ceramide metabolism in Saccharomyces cerevisiae
Biochem. J.
2012
, vol. 
447
 (pg. 
103
-
114
)
30
Kreft
S. G.
Wang
L.
Hochstrasser
M.
Membrane topology of the yeast endoplasmic reticulum-localized ubiquitin ligase Doa10 and comparison with its human ortholog TEB4 (MARCH-VI)
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
4646
-
4653
)
31
Falke
J. J.
Dernburg
A. F.
Sternberg
D. A.
Zalkin
N.
Milligan
D. L.
Koshland
D. E. J.
Structure of a bacterial sensory receptor. A site-directed sulfhydryl study
J. Biol. Chem.
1988
, vol. 
263
 (pg. 
14850
-
14858
)
32
Karlin
A.
Akabas
M. H.
Substituted-cysteine accessibility method
Methods Enzymol.
1998
, vol. 
293
 (pg. 
123
-
145
)
33
Bogdanov
M.
Zhang
W.
Xie
J.
Dowhan
W.
Transmembrane protein topology mapping by the substituted cysteine accessibility method (SCAM™): application to lipid-specific membrane protein topogenesis
Methods
2005
, vol. 
36
 (pg. 
148
-
171
)
34
Bogdanov
M.
Dowhan
W.
Lipid-assisted protein folding
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
36827
-
36830
)
35
Benghezal
M.
Benachour
A.
Rusconi
S.
Aebi
M.
Conzelmann
A.
Yeast Gpi8p is essential for GPI anchor attachment onto proteins
EMBO J.
1996
, vol. 
15
 (pg. 
6575
-
6583
)
36
Cassel
M.
Seppala
S.
von Heijne
G.
Confronting fusion protein-based membrane protein topology mapping with reality: the Escherichia coli ClcA H+/Cl− exchange transporter
J. Mol. Biol.
2008
, vol. 
381
 (pg. 
860
-
866
)
37
Kim
H.
Melen
K.
Osterberg
M.
von Heijne
G.
A global topology map of the Saccharomyces cerevisiae membrane proteome
Proc. Natl. Acad. Sci. U.S.A.
2006
, vol. 
103
 (pg. 
11142
-
11147
)
38
Nagiec
M. M.
Nagiec
E. E.
Baltisberger
J. A.
Wells
G. B.
Lester
R. L.
Dickson
R. C.
Sphingolipid synthesis as a target for antifungal drugs. Complementation of the inositol phosphorylceramide synthase defect in a mutant strain of Saccharomyces cerevisiae by the AUR1 gene
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
9809
-
9817
)
39
Fankhauser
C.
Homans
S. W.
Thomas-Oates
J. E.
McConville
M. J.
Desponds
C.
Conzelmann
A.
Ferguson
M. A.
Structures of glycosylphosphatidylinositol membrane anchors from Saccharomyces cerevisiae
J. Biol. Chem.
1993
, vol. 
268
 (pg. 
26365
-
26374
)
40
Levine
T. P.
Wiggins
C. A.
Munro
S.
Inositol phosphorylceramide synthase is located in the Golgi apparatus of Saccharomyces cerevisiae
Mol. Biol. Cell
2000
, vol. 
11
 (pg. 
2267
-
2281
)
41
Umemura
M.
Fujita
M.
Yoko-O
T.
Fukamizu
A.
Jigami
Y.
Saccharomyces cerevisiae CWH43 is involved in the remodeling of the lipid moiety of GPI anchors to ceramides
Mol. Biol. Cell
2007
, vol. 
18
 (pg. 
4304
-
4316
)
42
Ghugtyal
V.
Vionnet
C.
Roubaty
C.
Conzelmann
A.
CWH43 is required for the introduction of ceramides into GPI anchors in Saccharomyces cerevisiae
Mol. Microbiol.
2007
, vol. 
65
 (pg. 
1493
-
1502
)
43
Choudhary
V.
Jacquier
N.
Schneiter
R.
The topology of the triacylglycerol synthesizing enzyme Lro1 indicates that neutral lipids can be produced within the luminal compartment of the endoplasmatic reticulum: implications for the biogenesis of lipid droplets
Commun. Integr. Biol.
2011
, vol. 
4
 (pg. 
781
-
784
)
44
Huh
W. K.
Falvo
J. V.
Gerke
L. C.
Carroll
A. S.
Howson
R. W.
Weissman
J. S.
O'shea
E. K.
Global analysis of protein localization in budding yeast
Nature
2003
, vol. 
425
 (pg. 
686
-
691
)
45
Xu
R.
Jin
J.
Hu
W.
Sun
W.
Bielawski
J.
Szulc
Z.
Taha
T.
Obeid
L. M.
Mao
C.
Golgi alkaline ceramidase regulates cell proliferation and survival by controlling levels of sphingosine and S1P
FASEB J.
2006
, vol. 
20
 (pg. 
1813
-
1825
)
46
Bernsel
A.
Viklund
H.
Hennerdal
A.
Elofsson
A.
TOPCONS: consensus prediction of membrane protein topology
Nucleic Acids Res.
2009
, vol. 
37
 (pg. 
W465
-
W468
)

Supplementary data