Ceramidase (CDase, EC18.104.22.168) is an enzyme that catalyses the cleavage of the N-acyl linkage of ceramide (Cer) to generate free fatty acids and sphingosine (Sph). Sph can be further phosphorylated by Sph kinase to give the subsequent product sphingosine-1-phosphate (S1P). Several lines of investigation have clearly established Cer as one of the prime mediators of eukaryotic stress response.1 Accumulating evidence suggests that Cer plays an important role in differentiation,2 cell cycle arrest,3 apoptosis,4 senescence5 and autophagy.6 Recently, enzymes that regulate Cer metabolism have been established as a potential target for cancer therapy.7
Based on their pH optima and primary sequences, CDases are classified into three groups: acid, neutral and alkaline CDases.8 Acid ceramidase (acidCDase), which hydrolyses Cer in lysosomes, was first identified and purified from rat brain by Gatt9 in 1963. AcidCDase deficiency was identified as being responsible for Faber disease, in which Cer accumulates in lysosomes.10 For the first time, Yada et al.11 were the first to introduce alkaline ceramidases (alkCDases) by purifying two membrane-bound enzymes, with a molecular mass of 60 and 148 kDa, from guinea pig skin. These alkCDases prefer phytoceramide over normal Cer-containing sphinganine as a substrate and are found in yeast and humans. Recently, several alkCDases were cloned and characterized, including phytoceramidase (phytoCDase) and dihydroceramidase from the yeast Saccharomyces cerevisiae12,13 and the human homologue of phytoCDase.14 Recently, human alkCDase 2 (ACER2), a Golgi enzyme that regulates the maturation of the integrin-β1 subunit by controlling the generation of Sph, has been reported.15
Neutral ceramidases (nCDases) have also been purified and biochemically characterized from various tissues such as mouse liver,16 rat kidney,17 rat brain,18,19 rat intestine20 and human intestine.21 In addition, nCDases have been cloned from Pseudomonas aeruginosa,22 slime moulds,23 Drosophila,24 zebrafish,25 rats,17 mice26 and humans.27 Pata et al.28 have recently reported the molecular cloning and characterization of a phytoCDase from rice (Oryza sativa). Moreover, a novel amidase motif containing a serine residue has been identified as crucial for the catalytic activity of nCDase.29 The molecular mechanism of hydrolysis and synthesis of Cer by nCDase enzyme has been well characterized and demonstrates the breakdown or synthesis of the N-acyl linkage of Cer through a zinc (Zn2+)-dependent mechanism.30
The involvement of nCDase in the metabolism of Cer, and regulation of sphingolipid-mediated signalling, in the plasma membrane and extracellular milieu has been reported in nCDase overexpressing Chinese hamster ovary (CHOP) cells.30,31 It has been shown that rat and mouse nCDases are mainly localized in the plasma membrane, whereas the human homologue expression was detected in mitochondria of human embryonic kidney (HEK)293 cells.27 The knockdown of nCDase enzyme, using small interfering RNA (siRNA), increased cellular Cer levels and caused cell cycle arrest, which was similar to the effects caused by gemcitabine, a chemotherapeutic agent, when used to treat murine epithelial cells.32 Similarly, a number of studies have reported that nCDase can be regulated by cytokines and growth factors.33–35 A protective role for nCDase against inflammation was suggested in a previous study in which the loss of nCDase resulted in protection against inflammation by an unexpected increase in S1P generation.36 These findings indicate a major role for nCDases in the regulation of these bioactive lipids. However, the major cellular molecules that regulate the activation or inactivation of these enzymes are still unclear. Therefore, detailed enzymological studies are required for the characterization of the physiological function of these enzymes.
Very recently, we have purified and characterized a 100-kDa nCDase from camel brain.37 In the present study, we purified and biochemically characterized a second, more hydrophobic protein, ≈ 120 kDa, with neutral to alkaline CDase activity (CBCDase II) from camel brain. The findings from this study imply that this purified CBCDase II is different from the CBCDase I in several aspects such as pH optimum, molecular mass, extent of glycosylation and phospholipid dependence.
Materials and methods
Fresh camel brains were obtained from a local slaughterhouse. Diethylaminoethyl (DEAE)–Sepharose high-performance (HP), Phenyl Sepharose HP, Superdex 200 high-resolution (HR) 10/30, Mono S HR 5/5 and pre-packed, disposable (PD)-10 columns were purchased from GE Healthcare (Uppsala, Sweden). Centriprep and Centricon sample concentrators were supplied by Amicon, Inc. (Beverly, MA, USA). Pre-coated silica gel 60 thin-layer chromatography (TLC) plates were obtained from Whatmann (GE Healthcare). d-Erythro-carbon-12-nitrobenzoxadiazole-ceramide (d-erythro-C12-NBD-Cer) was kindly provided by the Lipidomics Core Facility at the Medical University of South Carolina. All other lipids were from Avanti Polar Lipid (Alabaster, AL, USA). Rabbit polyclonal anti-nCDase antibody was made previously.19 Goat anti-rabbit horseradish peroxidase-conjugated secondary antibody was obtained from Sigma-Aldrich (St. Louis, MO, USA). All sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) reagents were purchased from Bio-Rad (Hercules, CA, USA). Triton X-100, Extracti-Gel D detergent-removing gel, bicinchoninic acid (BCA) protein assay kits and enhanced chemiluminescence were purchased from Pierce (Rockford, IL, USA). Glycosidase F was purchased from Calbiochem–Merck (Millipore Darmstadt, Germany). All other chemicals used were purchased from Sigma-Aldrich.
Neutral ceramidase enzyme assay and biochemical characterization
Ceramidase (CDase) activity was measured using C12-NBD-Cer as a substrate, as described previously.37 Briefly, 25 μl of 100-μmol d-erythro-C12-NBD-Cer was incubated at 37°C for 1 hour with an appropriate amount of the enzyme (25 μl). The reaction was stopped by adding 100 μl of chloroform/methanol (1:1). After drying in a speed vacuum concentrator (Savant Instruments, Inc., New York, NY, USA), the sample was dissolved in 25 μl of chloroform/methanol (2:1) and applied to a TLC plate which was developed with chloroform/methanol/ammonia (75:15:0.9). The spots corresponding to NBD-dodecanoic acid and C12-NBD-Cer were scraped and then incubated with ethanol at 37°C for 5 minutes to extract the compounds. Their fluorescence was measured at (485/535 nm) excitation/emission wavelength in a Victor X3 Perkin-Elmer (Waltham, MA, USA) multilabel plate reader. The compounds were quantified using a standard curve of known amounts of C12-NBD-Cer and NBD-dodecanoic acid. One enzyme unit is defined as the amount capable of catalysing the release of 1 μmol of NBD-dodecanoic acid/minute from C12-NBD-Cer. For determination of the optimum pH, the substrate was dissolved in the following buffers: pH 3–5, 100 mM acetate buffer; pH 6–7, 100 mM phosphate buffer; pH 7–8, 100 mM Tris or 4-(2-hydroxyethyl)-1-piperazineethanesulphonic acid (HEPES) buffer and pH 8–10, 100 mM glycine buffer.
Fractionations and Triton X-100 extraction
Tissue fractionation and Triton X-100 extraction was carried out as previously described.37 Briefly, fresh camel brain was homogenized in homogenization buffer [500 ml of 20-mM cold phosphate buffer of pH 7.4, containing 0.25 M sucrose, 1 mM ethylenediaminetetraacetic acid (EDTA) and 0.2 mM phenylmethylsulphonyl fluoride] using a Dounce homogenizer. The homogenate was centrifuged at 1000 × g for 10 minutes, and the pellet was further homogenized using 100 ml of homogenization buffer. After centrifugation at 1000 × g for 10 minutes, the pellet was washed twice with 100 ml of homogenization buffer. All supernatants were combined and designated as the post-nuclear supernatant fraction. The post-nuclear supernatant fraction was then centrifuged at 10 000 × g for 30 minutes and the pellet of this centrifugation was resuspended in solubilization buffer (150 ml of 20-mM Tris buffer of pH 7.4, 1 mM EDTA, 0.2 mM phenylmethylsulphonyl fluoride, 0.5% Triton X-100). After mixing for 1 hour, the Triton X-100 solubilized fraction was obtained by centrifugation of the mixture at 10 000 × g for 30 minutes. The supernatant (Triton X-100 extract) was used as a source for CDase purification. All the steps were carried out at 4°C.
Diethylaminoethyl–Sepharose: the Triton X-100 extract (150 ml) was applied to DEAE–Sepharose column (25 ml) equilibrated with buffer A1 (20 mM Tris, pH 7.4, 1 mM EDTA, 0.2 mM phenylmethylsulphonyl fluoride, 0.005% Triton X-100) at 1 ml/minute. The unbound proteins were eluted by washing the column with 200 ml linear gradient of sodium chloride (NaCl) from 0 to 0.3 M in buffer B1 (20 mM Tris, pH 7.4, 1 mM EDTA, 0.2 mM phenylmethylsulphonyl fluoride, 0.005% Triton X-100 and 1.5 M NaCl). The salt concentration was then increased to 1.5 M in 100 ml. After the B1 buffer, the column was washed again with 50 ml of A1 buffer and the tightly bound proteins were eluted using a linear gradient of buffer B2 (0–0.5% Triton X-100 in B1). The CDase activity was measured in the 5 ml fractions that were collected and fractions with second peak activity were pooled.
Phenyl Sepharose HP: the active fractions obtained from the combination of salt and Triton X-100 gradient of DEAE–Sepharose were collected and loaded on to a Phenyl Sepharose column which was equilibrated with buffer B1 (buffer A1 with 0.225 M NaCl) at a flow rate of 0.5 ml/minute. After the sample was applied, the flow rate was increased to 1 ml/minute. The column, on the other hand, was washed with decreasing concentrations of the buffer B1. A stepwise elution was applied using 100 ml of a 70% buffer B2 (1 mM Tris, pH 7.4, 1 mM EDTA, 0.2 mM phenylmethylsulphonyl fluoride), then 100 ml of 100% buffer B2. Finally, 100 ml gradient from 0–1% Triton X-100 in buffer B2 was applied. Fractions of 1 ml were collected, and CDase activity was measured. Fractions containing peak activity were pooled.
Superdex 200 HR 10/30: the pooled Phenyl Sepharose active fractions were concentrated using Centriprep and then loaded on to a Superdex 200 column (25 ml) equilibrated with buffer A1 at a flow rate of 0.1 ml/minute and 1-ml fractions were collected. Fractions containing peak activity were pooled.
Mono S: the pooled fractions from the Superdex column were diluted five times with buffer A1 and applied to a Mono S column (1 ml) equilibrated with buffer A1 at a flow rate of 1 ml/minute. After washing the column with 10 ml of buffer A1 to remove the unbound protein, the active fractions could elute only with a 20-ml linear gradient combination of salt and Triton X-100. Fractions of 1 ml were collected, and CDase activity was measured.
Protein assay and electrophoresis
The protein concentration was determined using the Bradford assay, and for samples containing Triton X-100 the BCA protein quantifying assay was used. The SDS-PAGE of the reducing and non-reducing condition was carried out as reported previously.19
Western blot analysis
The purified enzyme preparation was subjected to SDS-PAGE (10%). The separated protein was electrophoretically transferred on to a nitrocellulose membrane. After blocking with 5% non-fat milk in Tris buffer saline containing 0.1% Tween 20, the membrane was incubated with anti-nCDase antibody followed by horseradish peroxidase-conjugated secondary antibody. Proteins were visualized using enhanced chemiluminescence.
Glycosidase F treatment
The glycosylation treatment was performed according to manufacturer’s protocol. Briefly, CDase was denatured in SDS-PAGE 6× sample buffer for 3 minutes. The denatured enzyme was then incubated at 37°C for 18 hours with 0.5 milliunit of glycosidase F in the presence of 0.5% Triton X-100. After incubation, the samples were subjected to SDS-PAGE and western blot analysis using anti-nCDase antibody.
Purification of camel brain ceramidase II
Recently, we reported a second type of CDase activity detected during the purification and characterization of camel brain membrane-bound ceramidase I (CBCDase I).37 This second CDase activity was seen only when eluting with Triton X-100 (0.5%) in combination with NaCl in a DEAE–Sepharose column.37 This activity represented approximately 50–60% of the total enzyme activity which was applied to the DEAE–Sepharose column. However, it was not determined whether this second peak of CDase activity represented a more hydrophobic conformation of the salt-only eluted nCDase, or another isoenzyme. Therefore, we named this activity camel brain neutral ceramidase II (CBCDase II), and we decided to purify and characterize this CBCDase II activity.
The purification procedure of CBCDase II is summarized in Table 1. The enzyme was solubilized with 0.5% Triton X-100 and, in order to separate CDase I and II, this solubilized extract was applied to a DEAE–Sepharose column. The DEAE–Sepharose column was efficient in separating these two isoforms of CBCDases. First, the well-characterized CDase I, and other protein contaminants, were removed using a linear shallow gradient (Figure 1a). Second, the column was washed with 1.5 M NaCl in order to remove as many unwanted bound proteins as possible. After this wash, a linear gradient of detergent and salt buffer was used to finally elute CBCDase II, at a salt concentration of 1.5 M NaCl and detergent concentration of 0.5% Triton X-100 (Figure 1a).
The active fractions of CBCDase II obtained from the DEAE–Sepharose column were pooled, and loaded on to the Phenyl Sepharose column. After a stepwise gradient in order to remove unbound proteins, the active fractions were eluted with 0–1% Triton X-100 in buffer B2 (Figure 1b). The active fractions were then loaded on to a Superdex 200 column to remove other impurities and Triton X-100, followed by a Mono S column to provide the, almost, homogenous protein (Figures 1c and 1d). This purification procedure resulted in ≈ 3800-fold purification (Table 1) of neutral ceramidase II enzyme to give near homogeneity.
The purified fractions from the Mono S column were subjected to SDS-PAGE, followed by staining with GelCode Blue stain reagent (Figure 2a). On SDS-PAGE, the purified CDase protein appeared at ≈ 120 kDa, which corresponds to the peak activity obtained after gel filtration on the Superdex column (Figure 1c). Next we checked the relationship of CBCDase II to CBCDase I by immune blot analysis. As shown in Figure 2b, both enzymes reacted robustly with an antibody raised against rat brain CDase (RBCDase) (Figure 2b). To verify whether the purified enzyme is glycosylated or not, the purified enzyme was treated with N-glycosidase F and visualization of the deglycosylated sample was performed using western blot analysis. Interestingly, the 120-kDa protein band of CBCDase II had a very slight shift after treatment with glycosidase F (Figure 2c, lanes 3 and 4). Indeed, when the same experiment was repeated on purified CBCDase I, whose apparent molecular weight was found to be ≈ 100 kDa, the molecular weight was reduced to 80 kDa following deglycosylation (Figure 2c, lanes 1 and 2). The molecular mass of deglycosylated CBCDase II is clearly different from that of CBCDase I, suggesting that the CBCDase II is not a differentially glycosylated form of the previously purified CBCDase I.
Characterization of pH optimum and kinetics of camel brain ceramidase II
The aliquots of purified fractions obtained after Mono S were used to characterize the enzyme activity. The purified CBCDase II showed a broad pH activity ranging from 7 to 9 when assayed using C12-NBD-Cer as a substrate (Figure 3a). The optimal activity was observed at pH 8.0 (Figure 3a). Unlike CBCDase I, which showed significant activity at an acidic pH of 5, CBCDase II exhibited less activity at a pH in the acidic range.
The hydrolytic capacity of the purified CBCDase II in the 50 mM Tris buffer was examined with C12-NBD-Cer as a substrate. The enzyme showed classic Michaelis–Menten kinetics. Lineweaver–Burk plots with C12-NBD-Cer as a substrate derived a Km of 15.04 μmol and a Vmax of 2.208 μmol/minute/mg (Figures 3b and 3c).
Effect of cations on camel brain ceramidase II
The effect of different metal ions on CBCDase II enzyme activity was determined, and it was observed that both Zn2+ and copper (Cu2+) inhibited the enzyme in a dose-dependent manner (Figure 4a). The addition of magnesium chloride (MgCl2) and manganese chloride (MnCl2) had no effect on the CBCDase II activity. On the other hand, the addition of calcium chloride (CaCl2), in a dose-dependent manner, stimulated the enzyme activity.
Effect of phospholipids on camel brain ceramidase II
Next, the effects of various phospholipids on purified CBCDase II activity was investigated using C12-NBD-Cer as a substrate. These lipids were added at the indicated concentration with the substrate. As observed in Figures 4b and 4c, all phospholipids inhibited the hydrolytic activity of the CBCDase II enzyme. Among these, the addition of phosphatidic acid (PA), phosphatidylglycerol (PG), phosphatidylinositol (PI) and phosphatidylserine (PS) inhibited CBCDase II activity even at very low concentrations (0.25 mM). Phosphatidylethanolamine (PE) and phosphatidylcholine (PC) showed a less inhibitory effect on CBCDase II.
Effect of thiol groups on camel brain ceramidase II
The effect of thiol groups on the hydrolytic activity of purified CBCDase II was tested by adding thiol group-containing compounds such as GSH, GSSG, NAC and cysteine. These reducing agents were added at different concentrations, starting at 0 mM and increasing to 20 mM. At 10 mM, the entire thiol group-containing compounds that were tested inhibited the hydrolytic activity of the purified CBCDase II. Amongst these, GSSG and cysteine showed the most potent inhibitory action (Figure 4d).
Effect of nucleotides on camel brain ceramidase II
The effect of purine and pyrimidine nucleotides on the enzyme activity of this purified CBCDase II was measured. As shown in Figure 5a, adenosine monophosphate (AMP) showed an attenuated effect on CDase activity, whereas both adenosine triphosphate (ATP) and adenosine diphosphate (ADP) inhibited CDase activity in a concentration-dependent manner, and ADP showed more of an inhibitory effect on enzyme activity than ATP. Figure 5b represents the effect of the three guanosine nucleotides on purified CBCDase II activity. These three nucleotides did not have a recognizable effect on the enzyme activity. Among the uridine nucleotides, only uridine monophosphate (UMP) inhibited enzyme activity in a dose-dependent manner. Uridine diphosphate (UDP) and uridine triphosphate (UTP), on the other hand, did not have a recognizable effect on the hydrolytic activity of the purified enzyme (Figure 5c). Similarly, among the thymidine nucleotides, only thymidine monophosphate (TMP) completely inhibited the enzyme activity in a dose-dependent manner, while thymidine triphosphate (TTP) inhibited 40% of the activity at low concentrations and thymidine diphosphate (TDP) had a very low inhibitory effect on CBCDase II activity (Figure 5d).
This study demonstrated that, similar to RBCDases (RBCDase I and II),18,19 at least two CDases with different molecular masses and enzymatic properties are present in camel brain (CBCDase I and CBCDase II). This study focused on the purification and characterization of the membrane-associated second type of nCDase from camel brain, which we have named CBCDase II. The CBCDase II described in this study is clearly different from the previously described CBCDase I in several aspects, including molecular mass, the extent of glycosylation, pH optima and effect of other biomolecules, such as lipids.19
As described in the previous investigation of CBCDase I,37 the DEAE–Sepharose column chromatography was efficient in separating the two CDase types. With all other columns used, except Superdex, the CBCDase II peak activity could elute only with a Triton X-100 gradient. The need for Triton X-100 to elute CBCDase II enzyme indicates that this enzyme may have a more hydrophobic conformation, which is entirely different from CBCDase I. Given that the enzyme elutes only in the presence of Triton X-100, it was very difficult to separate the protein from other impurities. Even after passing through a series of chromatographic steps, the purified fractions still contain some of the contaminant proteins, which can be seen in the SDS-PAGE after staining. Before application of the Mono S column, the sample was passed through gel filtration chromatography in order to remove the high concentration of Triton X-100. This step also removed a large number of other protein impurities. In the next step, the combined fractions from the gel filtration column were applied to the Mono S column. The enzyme is tightly bound to the column, and it is eluted using 0–1% Triton X-100 (Figure 1d). This highlights the strong hydrophobic character of CBCDase II.
After Mono S ion exchange chromatography, a single band with an apparent molecular mass of ≈ 120 kDa is identified using GelCode blue staining. This is further confirmed when western blot analysis of the purified protein with anti-CDase antibody is carried out. Moreover, the western blot analysis of CBCDase I and II clearly suggests that these are two different isoforms of CDases, having different molecular masses and a different glycosylation level. When CBCDase I was treated with N-glycosidase enzyme, the molecular mass of CBCDase I shifted from ≈ 100 to ≈ 80 kDa.37 This implies that CBCDase I is a highly N-glycosylated protein. On the other hand, treatment of CBCDase II with N-glycosidase enzyme resulted in only a very slight shift in molecular mass, suggesting that CBCDase II is an isoform of CBCDase I and not a hyperglycosylated version of it. It has been reported that the deglycosylation of RBCDase I and RBCDase II resulted in a shift in molecular weight from ≈ 95 kDa to ≈ 70 kDa and ≈ 110 kDa to ≈ 95 kDa, respectively.19 These data also support the conclusion that CBCDase I and II are different enzymes from the same origin. It has been suggested that carbohydrate moieties tend to make proteins more hydrophilic. This could explain the hydrophobic behaviour of CBCDase II, as it appears to be less glycosylated, whereas CBCDase I is more hydrophilic in nature, as it is attached to a high carbohydrate moiety (Figure 2c).
It is important to elucidate the role of these CBCDase isoforms in the regulation of Cer, Sph and S1P levels, as these sphingolipid metabolites do possess important physiological effects. Based on previous reports,19,24,37 C12-NBD-Cer was used as a substrate for the purified enzyme and all characterization experiments were conducted with this substrate. Unlike CBCDase I, the optimal pH of CBCDase II is not strictly neutral, but it seems slightly alkaline at pH 8.0. The difference in the pH optimum also supports the conclusion that these CBCDases are not same enzyme. These observations indicate that the purified CBCDase II enzyme is a member of neutral to alkaline CDases, such as RBCDase I,18 with less specificity in the acidic range and is clearly different from the acid/neutral/alkaline CDases isolated from other origins.
Ceramidases are composed of multiple isoforms and they can be distinguished based on their biochemical properties. The effect of metal ions on CBCDase II is similar to that on CBCDase I.37 Like RBCDase II19 and CBCDase I,37 CBCDase II activity is slightly stimulated by calcium (Ca2+) ions. Similarly, as seen in CBCDase I, PS and PA and PG totally inhibited CBCDase II activity. The major difference between these two CDases, with respect to the effect of phospholipids, is the effect of PI and PC. With CBCDase II, the inhibitory effect of PI is more significant compared with CBCDase I, whereas, with CBCDase I, PC behaves as a potent inhibitor even at low concentrations, but the inhibition ratio is moderately less with CBCDase II. Like CBCDase I, none of the phospholipids is seen to stimulate the enzyme activity of CBCDase II.
The entire thiol group-containing compounds, such as GSH, GSSG, NAC and cysteine, which were tested on CBCDase II activity, inhibited enzyme activity. This might be a mode of regulation for enzyme activity by these compounds in order to regulate the level of bioactive sphingolipids. Inhibition of another important sphingolipid-metabolizing enzyme-neutral sphingomyelinase (nSMase) by GSH has been previously reported.38 Moreover, the CBCDase I activity was shown to be more inhibited by the GSSG.37 In this study, however, both GSH and GSSG have similar effects on CBCDase II activity. In order to identify the significance of the inhibitory action of these reducing agents on CBCDase activity and the requirement of the sulphhydryl functional group for the inhibition of enzyme, further mechanistic studies are required.
Previous studies have shown that nucleotides can affect CDase activity.19,37 Similar to CBCDase I, ADP, ATP, UMP and TMP inhibit the activity of the purified CBCDase II enzyme in a dose-dependent manner. Only TTP shows a slight variation from CBCDase I in that it partially inhibited CBCDase II activity, while it did not have a significant effect on CBCDase I activity, even at higher concentrations. Further studies are required to identify the exact role of these nucleotides in the regulation of the CBCDase II enzyme.
For the first time, this study demonstrates the purification and characterization of a second type of nCDase enzyme from camel brain. At least two CDases, with different molecular masses and enzymatic properties, are present in camel brain. Based on the biochemical analysis, this purified CBCDase II enzyme differs from the recently identified CDase enzymes. The purification and identification of its biochemical behaviour with different biomolecules are important to understand the regulation of CDase enzyme activity, since CDase controls the balance between the bioactive lipids and, hence, their biological effects. Further mechanistic studies are required to identify how molecules such as phospholipids, thiol group-containing compounds and nucleotides regulate CDase activity.