Activity of neutral and alkaline ceramidases on fluorogenic N-acylated coumarin-containing aminodiols.

Ceramidases catalyze the cleavage of ceramides into sphingosine and fatty acids. Previously, we reported on the use of the RBM14 fluorogenic ceramide analogs to determine acidic ceramidase activity. In this work, we investigated the activity of other amidohydrolases on RBM14 compounds. Both bacterial and human purified neutral ceramidases (NCs), as well as ectopically expressed mouse neutral ceramidase hydrolyzed RBM14 with different selectivity, depending on the N-acyl chain length. On the other hand, microsomes from alkaline ceramidase (ACER)3 knockdown cells were less competent at hydrolyzing RBM14C12, RBM12C14, and RBM14C16 than controls, while microsomes from ACER2 and ACER3 overexpressing cells showed no activity toward the RBM14 substrates. Conversely, N-acylethanolamine-hydrolyzing acid amidase (NAAA) overexpressing cells hydrolyzed RBM14C14 and RBM14C16 at acidic pH. Overall, NC, ACER3, and, to a lesser extent, NAAA hydrolyze fluorogenic RBM14 compounds. Although the selectivity of the substrates toward ceramidases can be modulated by the length of the N-acyl chain, none of them was specific for a particular enzyme. Despite the lack of specificity, these substrates should prove useful in library screening programs aimed at identifying potent and selective inhibitors for NC and ACER3.

In this work, we show that RBM14 compounds are hydrolyzed by NC, ACER3, and NAAA, but not by ACER1 and ACER2 . Each amidohydrolase exhibits different substrate selectivity depending on the nature of the N -acyl chain. The results presented underscore the usefulness of these compounds in massive screening programs aimed at identifying amidohydrolase inhibitors.

Overexpression of hACER2 and mACER1
To overexpress ACER2, we used the stable cell line HeLa T-Rex/ACER2 that we constructed in our previous study ( 17 ). To overexpress ACER1, we applied the same expression system. The coding sequence of the mACER1 was cloned into the empty vector, pcDNA4, to construct the expression vector, pcDNA4-ACER1, which was used to generate the stable cell line, HeLa-T-Rex/ Acer1, as described in our previous study ( 17 ).

Knockdown of ACER3 in HCT116 cells
A lentiviral vector expressing a control shRNA (shCtrl) (CC-GGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTC-ATCTTGTTGTTTTT) or ACER3-specifi c shRNA (shACER3) (CCGGTATACAGCTGTTGCATATTTGCTCGAGCAAATATGCA-ACAGCTGTATATTTTTTG) from Sigma-Aldrich (St. Louis, MO) was packaged into lentiviral particles in HEK293T cells using a viral packaging kit (GeneCopoeia, Rockville, MD) and lentiviruses expressing shCtrl or shACER3 were then transduced into HCT116 cells according to the manufacturer's protocol. Forty-eight hours later, the transduced cells were replated and treated with 5 g/ml puromycin (Sigma-Aldrich) for 2 weeks. Puromycin-resistant The three ACERs include Asah3/Acer1, Asah3L/Acer2, and Acer3. Acer1 is encoded by the ASAH3/ACER1 gene and is exclusively expressed in the endoplasmic reticulum of skin cells. Acer2, encoded by the ASAH3L/ACER2 gene, is a Golgi resident enzyme particularly abundant in placenta. Finally, Acer3 is encoded by the PHCA /A CER3 gene and is localized to both the endoplasmic reticulum and the Golgi. An interesting feature of the three ACERs is their activation by Ca 2+ ions.
Validation of ceramidases as therapeutic targets has been carried out mainly for AC ( 5,6 ) and NC ( 7,8 ). This has triggered the development of ceramidase inhibitors as leads for the further development of potential drugs. Most ceramidase inhibitors have been discovered after either rational design or screening of a small series of compounds. Although a number of procedures for the determination of ceramidase activities have been reported ( 9 ), massive screening relies on the availability of high-throughput methods, only a few of which have been described (10)(11)(12). These include a fl uorescent sphingolipid fl uorescence resonance energy transfer (FRET) probe that allows homogeneous ratiometric determination of enzyme activity in real-time ( 12 ). In a previous article ( 10 ), we reported on the use of a coumarinic analog of ceramide, namely RBM14C16 ( Fig. 1 ), to determine AC activity. Other analogs of RBM14C16 with different N -acyl chain lengths were later reported as AC substrates and for use in diagnosis of Farber disease ( 11 ). Among the several analogs, RBM14C12 was the preferred substrate for AC. Lysates from Farber cells overexpressing AC ( 13 ) hydrolyzed RBM14C12 with apparent K m and V max values of 26 M and 334 pmol/min/mg protein .
Besides ceramidases, other amidohydrolases include those involved in the hydrolysis of bioactive N -acylethanolamines to fatty acids and ethanolamine. These amidohydrolases include N -acylethanolamine-hydrolyzing acid amidase (NAAA), which it is only active at acidic pH ( 14 ). Comparison of the primary structures of NAAA and AC revealed a high sequence homology. They also share functional features, and they belong to the choloylglycine hydrolase family ( 15 ). wavelengths of 355 and 460 nm, respectively. The same reaction mixtures without enzymes were used as blanks.
To determine ceramidase activity in intact cells, the day before the assay, 2 × 10 4 cells per well were seeded in a 96-well plate. Cells were incubated for 24 h at 37°C and 5% CO 2 . Medium was replaced by 100 l of fresh medium to which the required volume of a 4 mM stock solution of the RBM14 substrate in ethanol had been added to obtain a 40 M fi nal concentration (1% fi nal concentration of ethanol). The plate was incubated for 3 h at 37°C, 25 l/well of methanol were added and the assay was continued as described above for cell lysates.

NAAA activity
The reaction mixtures contained 25 l/well of protein (1 mg/ml, 25 g/well), 40 l/well of 0.1 mM substrate solution (40 M fi nal concentration) in reaction buffer [100 mM acetic/acetate buffer (pH 4.5)], and 35 l/well of reaction buffer. The reaction mixture was incubated at 37°C for 3 h, stopped with 25 l/well of methanol, and the assay was continued as described in the above section.

Protein concentration determination
Protein concentrations were determined with BSA as a standard using a BCA protein determination kit (Thermo Scientifi c) according to the manufacturer's instructions.

Western blot analysis
Thirty micrograms of microsomal protein were combined with Laemmli sample buffer and heated at 55°C for 30 min. Samples were loaded onto a 12% polyacrylamide gel, separated by electrophoresis at 140 V/1 h, and transferred onto a polyvinylidene difl uoride membrane (100 V/1 h). Unspecifi c binding sites were then blocked with 5% milk in TBS-Tween (TBST ) (ACER2) or 3% BSA in TBST (actin). Anti-ACER2 antibody was diluted 1:1,000 in 5% milk in TBST. Anti-actin antibody was diluted 1:2,000 in 3% BSA in TBST. Membranes were incubated overnight at 4°C under gentle agitation. After washing with TBST, membranes were probed with the correspondent secondary antibody for 1 h at room temperature (ACER2, anti-rabbit diluted 1:1,000 in 3% BSA in TBST; actin, anti-mouse diluted 1:10,000 in 5% milk in TBST). Antibody excess was eliminated by washing with TBST and protein detection was carried out using ECL and membrane scanning with LI-COR C-DiGit® blot scanner. Band intensities were quantifi ed by LI-COR Image Studio Lite software.
Overexpression of mouse NC . Twenty-four hours before transfection, Farber cells were plated in 6-well plates (2.5 × 10 5 cells per well). Then cells were transfected with 2.5 g/well of pcDNA5/ TO (empty vector) or pcDNA5/TO harboring the mouse ASAH2 gene (pcDNA5/TO-ASAH2) using opti-MEM/lipofectamine, following the manufacturer's instructions. Test compounds were added 24 h after transfection.

Cell lysates
Cell pellets were resuspended in the appropriate volume of a 0.25 M saccharose solution. The suspension was submitted to three cycles of 5 s sonication (probe) at 10 watts/5 s resting on ice. The cell lysate was centrifuged at 600 g for 10 min. The supernatant was collected and protein concentration was determined as specifi ed below.

Microsomal preparations
The above 600 g centrifugation supernatants were transferred to 1.5 ml ultracentrifuge tubes and were spun at 100,000 g for 45 min to 1 h at 4°C. Pellets were either stored at Ϫ 80°C or protein concentration was measured, adjusted with a 0.25 M saccharose solution to the appropriate concentration, and used for activity determination or processed for quantitative (q)PCR or Western blot . . The reaction mixture was incubated at 37°C for 3 h, except for the determination of K m and V max (30 min). In all cases, reactions were stopped with 25 l/well of methanol and then 100 l/well of NaIO 4 [2.5 mg/ml in 100 mM glycine-NaOH buffer (pH 10.6)] was added . After incubation at 37°C for 1 h in the dark, 100 l/well of 100 mM glycine-NaOH buffer (pH 10.6) was added and fl uorescence was measured spectrophotometrically at excitation and emission substrates, the assay was optimized with this enzyme. Because an aldehyde is produced in the oxidation step of the development phase of the procedure ( 1 ), the presence of primary amines in the solution must be avoided. Therefore, Tris was substituted with HEPES, which was also compatible with the presence of Ca 2+ in the solution, and so it could be used for ACER. Although Triton X-100 has been previously used in ceramidase activity measurements, we found that it decreased the fl uorescence signal ( Fig. 2A ). Therefore, it was omitted in the assays. The higher hydrophilicity of the RBM14 compounds, as compared with other substrates used [i.e., 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD)-labeled ceramides], allows avoiding the use of Triton X-100 in activity determinations. As concluded from the kinetic parameters, the best substrates for pCDase were RBM14C14 ≈ RBM14C12 > RBM14C10 >> RBM14C8 ( Fig. 2B ; Table 1 ).

Ceramidase activities
Determination of enzyme activity in vitro was also carried out with human recombinant enzyme (hNC). As shown in Fig. 2C and Table 1 , as in the case of pCDase, the substrate affi nity for hNC increased directly with increasing the N -acyl chain length, as concluded from the decrease of the K m values. Kinetic parameters show that substrate preference is RBM14C16 ≈ RBM14C14 >> RBM14C12 >> RBM14C10 > RBM14C8.
The hydrolysis of RBM14 compounds by NC was also determined in intact cells. Farber cells, which do not hydrolyze RBM14, were transfected with mNC [Farber cells transfected with the pcDNA5/TO plasmid containing mouse ASAH2 (FD-pcDNA5/TO-ASAH2)]. In these experiments, in which all substrates were tested simultaneously at the same concentration (40 M), NC showed a slight preference for the N -C10 analog ( Fig. 2D ). This was also the case in cell lysates ( Fig. 2E ). Interestingly, the N -C8 analog was also hydrolyzed by mNC, in contrast to AC, which does not accept this compound as substrate ( 11 ), and the hNC and bacterial NC, for which the C8 and the C10 derivatives are poor substrates.
To confi rm the robustness of the assay, we calculated the Z factor, a statistical parameter used to evaluate the suitability of a HTS assay based on the range of the signal and the internal variation of the measurements. A score of 1 > Z у 0.5 indicates a good assay, and a score of Z = 1 indicates a perfect assay. Our calculated Z factor for the NC assay using the pure human protein was 0.69 ± 0.17 using the best substrate, RBM14C16. This result demonstrated that the assay is suitable for HTS applications.

Hydrolysis of RBM14 by ACERs
As mentioned in the introduction, three different ACERs have been reported, all of them exhibiting an optimum activity at pH around 9.0. Because NC exhibits residual ceramidase activity at basic pH, the effectiveness of ACERs to hydrolyze RBMC14 was fi rst explored using MEFs defective in the ASAH2 gene ( ASAH2 ( Ϫ / Ϫ ) MEFs). As expected, lysates from these cells were signifi cantly less competent at hydrolyzing RBM14C12 than wild types at neutral pH values ( Fig. 3A ). As shown in Fig. 3B , all the analogs except for RBM14C8 were hydrolyzed in intact cells, with The fi nal step was a 1 min incubation at 60°C. All reactions were performed in triplicate. Relative mRNA abundances of the different genes were calculated from the second derivative maximum of their respective amplifi cation curves (Cp, calculated by duplicates). Cp values for target genes (Tg) were compared with the corresponding values for the GAPDH reference gene to obtain the ⌬ Cp values ( ⌬ Cp = CpGAPDH -CpTg).

LC/MS
Lipid extracts and LC/MS analysis were carried out as previously described ( 18,19 ).

Statistics
Comparison between two means has been carried out with the unpaired two-tailed t -test and statistical differences are marked with asterisks. For comparison of more than two means, data have been analyzed by one-way ANOVA test followed by Bonferroni's multiple comparison test. Statistical differences between means are pointed out with different letters atop each bar (same letter indicates no statistical difference).
The quality of the assays for NC and ACER3 has been assessed by calculating the Z factor, which is a valuable tool to evaluate the robustness and suitability of high-throughput screening (HTS) assays ( 20 ). This parameter is calculated using the equation:

Neutral ceramidases hydrolyze RBM14
Because different NCs have been reported, we tested and compared the activity of bacterial NC, mNC, and hNC on RBM14 compounds. The bacterial ceramidase from Pseudomonas aeruginosa was purifi ed as previously reported ( 16 ). Prior to screening the enzyme activity on the fl uorogenic We then determined the levels of mRNA for the different ACERs in ASAH2 ( Ϫ / Ϫ ) MEFs. As shown in Fig. 3D , these cells have high levels of ACER3 transcript. In contrast, levels of ACER2 mRNA are very low and ACER1 mRNA is undetectable. These results are in agreement with reported data (http://biogps.org) and suggest that ACER3 is the ACER involved in the hydrolysis of RBM14 at basic pH in ASAH2 ( Ϫ / Ϫ ) MEFs.
RBM14C14 producing the highest fluorescence values. Hydrolysis of RBM14 was also analyzed in ASAH2 ( Ϫ / Ϫ ) MEF lysates at different pH. As shown in Fig. 3C , the highest activity occurred at basic pH, at which RBM14C14 gave the highest hydrolysis rates. The activity observed at neutral pH is likely to correspond to residual activity of ACERs. These results suggested that RBM14s, mainly the C16 analog, were hydrolyzed by at least one of the three ACERs. In (E), compounds were used at the specifi ed concentrations using 5 ng of hNC and the same buffer as in (D). Data are mean ± SD of three to fi ve independent experiments with triplicates. In (A), asterisks indicate signifi cant difference between samples with and without Triton X-100 ( P < 0.001, unpaired twotailed t -test). In (C) and (D), data were analyzed by one-way ANOVA test [ P < 0.0001 (C); P < 0.0005 (D)] followed by Bonferroni's multiple comparison test. Different letters denote a statistically signifi cant difference between groups ( P < 0.05). detected in ACER3 knockdown cells than in mock cells, these increases were low ( Fig. 4C ). However, amounts of 16:0, 24:0, and 24:1 ceramide monohexosides and lactosylceramides were signifi cantly increased in ACER3 knockdown over controls, suggesting metabolization of the augmented ceramide via glycosylation. No difference was observed for sphingomyelins between the two cell lines (data not shown).
To evaluate whether ACER2 hydrolyzes the fl uorogenic substrates, we used a stable HeLa-based cell line expressing ACER2 in HeLa T-REx cells under the control of a tetracycline-inducible promoter system ( 3 ). This inducible expression system allows the expression of ACER2 to be turned on/off by adding/removing tetracycline to the medium . As expected, addition of tetracycline to the culture medium provoked a dramatic increase in the levels of ACER2 mRNA in HeLa T-REX cells stably transfected with To confi rm hydrolysis of RBM14 by ACER3, microsomal preparations from ACER3 knockdown of HCT116 (HCT116-shACER3) cells and the corresponding mock cells (HCT116-shRNA) were incubated with the fl uorogenic probes for 3 h in HEPES buffer with 1 mM Ca 2+ at pH 9.0. As shown in Fig. 4A , signifi cantly lower fl uorescence levels were released from RBM14C12, RBM14C14, and RBM14C16 by ACER3 knockdown than by wild-type cells, indicating that ACER3 hydrolyzes these three fl uorogenic RBM14 derivatives. As expected, ACER3 mRNA levels were about four times lower in ACER3 knockdown cells than in wild types ( Fig. 4B ). The activity of the ACER3 knockdown cells was evident by lipid analysis. Thus, microsomes from ACER3 knockdown cells had signifi cantly lower levels of sphingosine and dihydrosphingosine than mock cells ( Fig. 4C ). Furthermore, although significantly higher levels of 16:0, 24:1, and 24:2 ceramides were K m is given in micromoles and V max as nanomoles per minute per milligram protein. Data are mean ± SD of three to fi ve independent experiments with triplicates. Protein amounts were: pCDase, 50 ng; hNC, 5 ng. Experiments were carried out as detailed in the Experimental Procedures. In the case of pCDase, the assays were performed in the absence of Triton X-100. C16Cer, N -palmitoylsphingosine. n.d., not determined.

Fig. 3. Hydrolysis of RBM14 compounds by intact (B) and lysed (A, C) ASAH2
( Ϫ / Ϫ ) MEFs. In (A), experiments were performed at pH 7.4 and data correspond to the mean ± SD of two experiments with triplicates using 17 and 12 g of protein. The asterisk denotes statistical signifi cance at P < 0.00005 (unpaired two-tailed t -test, n = 6). In (B), data (mean ± SD) were obtained from three different experiments with triplicates (3.7, 5.1, and 6.5 × 10 4 cells/well); In (C), incubations were carried out with 25 g of protein at acid, neutral, or alkaline pH. Data correspond to the mean ± SD of three to fi ve experiments with triplicates. Data were analyzed by one-way ANOVA test [ P < 0.0001 (B); P < 0.0001 (C, pH 7.4); P < 0.0001 (C, pH 9.0)] followed by Bonferroni's multiple comparison test. Different letters denote a statistically signifi cant difference between groups ( P < 0.05). D: Agarose gel electrophoresis ACER analysis of the RT-PCR products obtained from mRNA isolated from ASAH2 ( Ϫ / Ϫ ) MEFs. Primers and PCR conditions are detailed in the Experimental Procedures. A representative image is shown. all substrates tested as compared with tetracycline-free control cells. That the expressed enzyme was active was proven by LC/MS analysis. Thus, microsomes from ACER2-TET-ON cells treated with tetracycline contained significantly lower levels of 16:0, 18:0, 22:1, 24:0, 24:1, and 24:2 ceramides than microsomes from cells cultured in the absence of tetracycline (controls) ( Fig. 5D ). The higher difference was observed for the 24:1 and 24:2 species. In agreement with the activity of the overexpressed ACER2, ACER2 (ACER2-TET-ON cells) over those of cells cultured in tetracycline-free medium ( Fig. 5A ). In agreement, Western blot analysis with antibodies against ACER2 showed that expression of ACER2 was induced with tetracycline (10 ng/ml, 24 h) in ACER2-TET-ON cells ( Fig. 5B ). In contrast, microsomes of ACER2-TET-ON cells cultured in the presence of tetracycline had no activity on the fl uorogenic substrates ( Fig. 5C ). Conversely, tetracycline provoked a low, but signifi cant, decrease in fl uorescence from Fig. 4. Hydrolysis of RBM14 compounds by ACER3. A: Activity of microsomes isolated from ACER3 knockdown and mock HCT116 cells over the fl uorogenic substrates. Incubations were carried out with 6 g of protein in 50 mM HEPES buffer with 1 mM CaCl 2 at pH 9.0. Data, shown as mean ± SD, correspond to three different experiments with duplicates. Data were analyzed by oneway ANOVA test [shCtrl, P < 0.0001 (A); shACER3, P < 0.0001 (A)] followed by Bonferroni's multiple comparison test. Different letters denote statistically signifi cant differences between means at P < 0.05 as found by this test. Lowercase letters correspond to the shCtrl group and capitalized letters correspond to the shACER3 group. B: ACER3 and ACER2 mRNA levels in ACER3 knockdown and mock HCT116 cells as determined by qPCR. C: Amounts of ceramide, sphingoid bases, ceramide monohexosides (CMH) and lactosylceramides (LacCer) in microsomes isolated from ACER3 knockdown and mock HCT116 cells as determined by LC/MS. Data correspond to the mean ± SD of two experiments with triplicates. Mean (±SD) sphingoid base phosphate levels were 2.2 ± 0.2 pmol/mg protein for shCtrl and 1.7 ± 0.2 pmol/mg protein for shACER3 . Asterisks indicate signifi cant difference between shCtrl and shACER3 at P < 0.05 (unpaired two-tailed t -test).

Fig. 5. Hydrolysis of RBM14 compounds by ACER2. HeLa T-REx
ACER2-TET-ON cells were grown in the absence ( Ϫ Tet) or presence (+Tet) of tetracycline (10 ng/ml) for 24 h. A: ACER2 mRNA levels as quantifi ed by qPCR. B: Representative Western blot showing the levels of ACER2 protein in microsomes (6 g of protein). C: Hydrolysis of the fl uorogenic substrates by microsomes isolated from HeLa T-REx ACER2-TET-ON cells grown in the absence ( Ϫ Tet) or presence (+Tet) of tetracycline (10 ng/ml) for 24 h. Incubations were carried out with 6 g of protein in 50 mM HEPES buffer with 1 mM CaCl 2 at pH 9.0. D: Amounts of ceramide and sphingoid bases in microsomes isolated from HeLa T-REx ACER2-TET-ON cells grown in the absence ( Ϫ Tet) or presence (+Tet) of tetracycline (10 ng/ml) for 24 h. Data correspond to the mean ± SD of three experiments with duplicates. In (C), data (mean ± SD) were analyzed by one-way ANOVA test ( P < 0.0001) followed by Bonferroni's multiple comparison test. Different letters denote a statistically signifi cant difference between means at P < 0.05 as found by this test. Lowercase letters correspond to the ACER2-TET-ON cells grown in the absence ( Ϫ Tet) of tetracycline group and capitalized letters correspond to the ACER2-TET-ON cells grown in the presence (+Tet) of tetracycline group. Asterisks indicate a signifi cant difference between samples from cells treated with and without tetracycline at P < 0.05 (unpaired two-tailed t -test). In (D), data correspond to the mean ± SD of two experiments with triplicates. Mean (±SD) sphingoid base phosphate levels were: -Tet, 5.1 ± 0.3 pmol/mg protein; +Tet, 16.9 ± 1.1 pmol/mg protein. 24:0, 24:1, and 24:2 ceramides and an increase in the amounts of free bases, mainly sphingosine ( Fig. 6C ).
Because RBM14C14 is the best substrate for ACER3, the Z factor was calculated using cell lysates from ASAH2 ( Ϫ / Ϫ ) MEFs. A Z factor of 0.74 ± 0.05 was found, demonstrating that the assay is suitable for HTS applications.

Hydrolysis of RBM14 by NAAA
Because NAAA has been shown to have a low, but signifi cant, ceramide hydrolyzing activity over N -lauroylsphingosine at pH 4.5 ( 15 ), we tested to determine whether NAAA was able to hydrolyze the RBM14 compounds. To this end, HEK293T cells were transiently transfected with an expression vector harboring the human NAAA gene, and the activity over RBM14 was determined in cell lysates at acidic pH. As shown in Fig. 7 , lysates from NAAA -transfected cells produced signifi cantly more fl uorescence from RBM14C14 and RBM14C16 than lysates from cells transfected with the empty vector. These increases were 1.6 and 2.7 for RBM14C14 and RBM14C16, respectively.

DISCUSSION
Ceramidases are amidohydrolases that catalyze the cleavage of ceramides and dihydroceramides to release fatty acids and sphingosine and sphinganine, respectively. Fluorescence spectroscopy-based methods to determine ceramidase activity have gained attention due to their high sensitivities and signal-to-noise ratios. In this context, different fl uorophores, such as NBD (21)(22)(23), BODIPY, ( 24 ), lissamine-rhodamine ( 24 ), and Nile Red ( 25 ) have been incorporated into either the fatty acyl or the sphingoid base moiety to produce fl uorescent (dihydro)ceramides as ceramidase substrates. Despite the advantage of fl uorescent over radioactive methods, the former are not amenable for high-throughput formats. In contrast, fl uorogenic substrates are susceptible for high-throughput confi gurations. Reported ceramidase fl uorogenic substrates include FRETbased dually labeled ceramides ( 12,26 ) in which donor and microsomes from tetracycline-treated ACER2-TET-ON cells had signifi cantly higher levels of free bases than controls ( Fig. 5D ).
The same tetracycline-inducible enzyme expression system was used to evaluate the ability of ACER1 to hydrolyze the fl uorogenic substrates. Addition of tetracycline to the culture medium provoked an increase in the levels of ACER1 mRNA in HeLa T-REX cells stably transfected with ACER1 (ACER1-TET-ON cells) over controls (cells incubated in tetracycline-free medium) ( Fig. 6A ). However, microsomes of ACER1-TET-ON cells cultured in the presence of tetracycline had no activity on the fl uorogenic substrates ( Fig. 6B ). Contrarily, tetracycline provoked a low, but signifi cant, decrease in fl uorescence from RBM14C14 and RBM14C16, as compared with tetracycline-free controls. Evidence that the overexpressed enzyme was active was obtained by LC/MS analysis of microsomes isolated from both cells treated with or without tetracycline. Thus, tetracycline induced a signifi cant decrease in the levels of 22:0, 22:1, Fig. 6. Hydrolysis of RBM14 compounds by ACER1. HeLa T-REx ACER1-TET-ON cells were grown in the absence ( Ϫ Tet) or presence (+Tet) of tetracycline (10 ng/ml) for 24 h. A: mRNA levels as quantifi ed by qPCR. B: Hydrolysis of the fl uorogenic substrates by microsomes isolated from HeLa T-REx ACER2-TET-ON cells grown in the absence ( Ϫ Tet) or presence (+Tet) of tetracycline (10 ng/ml) for 24 h. Incubations were carried out with 6 g of protein in 50 mM HEPES buffer with 1 mM CaCl 2 at pH 9.0. C: Amounts of ceramide and sphingoid bases in microsomes isolated from HeLa T-REx ACER2-TET-ON cells grown in the absence ( Ϫ Tet) or presence (+Tet) of tetracycline (10 ng/ml) for 24 h. Data correspond to the mean ± SD of two experiments with triplicates. In (B), data were analyzed by one-way ANOVA test ( P < 0.0001) followed by Bonferroni's multiple comparison test. Different letters denote a statistically signifi cant difference between means at P < 0.05 as found by this test. Lowercase letters correspond to the ACER2-TET-ON cells grown in the absence ( Ϫ Tet) of tetracycline group and capitalized letters correspond to the ACER2-TET-ON cells grown in the presence (+Tet) of tetracycline group. In (B) and (C), asterisks indicate signifi cant difference between samples from cells treated with and without tetracycline at P < 0.05 (unpaired two-tailed t -test). Mean (±SD) sphingoid base phosphate levels were: -Tet, 2.7 ± 0.3 pmol/mg protein; +Tet, 7.7 ± 0.5 pmol/mg protein. exclusive ACER, were able to hydrolyze the fl uorogenic substrates. As expected, lipid analysis of microsomes from ACER3 knockdown HCT116 cells contained lower amounts of free sphingoid bases than control cells. However, although they also had a higher content of C16:0, C24:1, and C24:2 ceramides, they were not as high (less than 30%) as we expected based on the decrease in ACER3 mRNA levels. Because accumulation of ceramide has been reported to induce cell death, further metabolism of ceramide to more complex lipids could explain the low levels found in these cells. The fact that ACER3 knockdown HCT116 cells contained signifi cantly higher levels of ceramide monohexosides and lactosylceramide supports this possibility.
As mentioned in the Results section, microsomes from ACER1 and ACER2 overexpressing cells produced less fl uorescence from the substrates than control microsomes, suggesting a lower capacity of the former to hydrolyze RBM14 compounds. These intriguing results suggest that overexpression of ACER1 and ACER2 provokes a downregulation of other ceramidases. Because the assay was conducted in microsomes at basic pH, ACER3, which hydrolyzes RBM14, is the most plausible candidate. Furthermore, ACER3 is highly expressed in HeLa T-REx cells ( 3 ). Experimental evidence has been reported that ACER3 knockdown upregulates the expression of ACER2 ( 4 ). However, no data have been reported that overexpression of one ACER affects the expression of other ceramidases. Nevertheless, transcript analysis showed that tetracycline did not affect ACER3 mRNA levels in either ACER1-TET-ON or ACER2-TET-ON cells (data not shown). Whether the enzyme activity is reduced by ACER1 and ACER2 overexpression is not known. In this regard, it is possible that activity is decreased by product (sphingoid bases) inhibition, because the total sphingosine/sphinganine levels in cells grown in the presence of tetracycline increased 2-and 3-fold over controls for ACER1 and ACER2 overexpressing cells, respectively. In any case, the overall results shown indicate that neither ACER1 nor ACER2 hydrolyze RBM14.
The lack of hydrolysis of RBM14 compounds by ACER1 and ACER2 could be explained by their reported preference for very long chain ceramides ( 1,2,17,27 ), while ACER3 hydrolyzes long chain, but not very long chain, unsaturated ceramides ( 4 ). Considering that RBM14 bears a modifi cation in the sphingoid base moiety, another plausible explanation is that ACER3 tolerates this modifi cation, while ACER1 and ACER2 do not. In this regard, ACER3 can hydrolyze ceramides, dihydroceramides, and phytoceramides, while ACER1 and ACER2 are more restrictive toward the ceramide sphingoid base portion ( 28 ).
Bioactive N -acylethanolamines, including the endocannabinoid anandamide and the anti-infl ammatory and neuroprotective N -palmitoylethanolamine, are hydrolyzed to fatty acids and ethanolamine by fatty acid amide hydrolase and NAAA. The latter, discovered by Ueda and collaborators ( 14 ), was shown to work only at acidic pH. NAAA has no homology to fatty acid amide hydrolase, but shares homology with AC and was reported to hydrolyze N -lauroylsphingosine at acidic pH ( 15 ). In agreement with the AC activity of NAAA, in this article, we show that NAAA hydrolyzes acceptor FRET pairs are located as part of the acyl chain and/or the sphingoid base of the substrate, and fl uorescence is displayed by the ceramidase hydrolysis-promoted fl uorescence quencher release. On the other hand, a series of fl uorogenic coumarin-containing ceramidase substrates has also been described ( 10,11 ). These compounds (RBM14 compounds, Fig. 1 ) were fi rst reported as substrates of the AC ( 10, 11 ), but the studies described herein show that they are also hydrolyzed by other amidohydrolases such as NC, ACER3, and, to a lesser extent, NAAA.
In the fi rst case, NC includes bacterial, human, and mouse enzymes. Comparison of the K m / V max ratios indicate that the bacterial enzyme exhibits a preference for RBM14C12 and RBM14C14 ( K m / V max = 0.032 and 0.048, respectively), while the human enzyme hydrolyzes RBM14C14 and RBM14C16 preferentially ( K m / V max = 0.0083 and 0.0069, respectively). In both cases, the substrate affi nity increases directly with increasing the N -acyl chain length, as concluded from the decrease of the K m values. This correlation is particularly evident with the human enzyme, which has, in general, less affi nity for the coumarinic substrates than the bacterial enzyme. In contrast, the latter hydrolyzes RBM14 with lower reaction rates than the human enzyme. As compared with reported substrates, hydrolysis of N -hexadecanoylsphingosine by pure pCDase occurs with a K m of 139 M and a V max of 5,300 nmol/min/mg of protein ( K m / V max = 0.026) ( 23 ) ( Table 1 ). Thus, the enzyme exhibits a lower affi nity for N -hexadecanoylsphingosine than for the best fl uorogenic substrate, RBM14C12, but the reaction rate is ten times faster with the former than with the later. On the other hand, pure hNC hydrolyzes N -[12-[(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]dodecanoyl]D-erythro -sphingosine (CerC12NBD) with a K m of 33.41 M and a V max of 826 nmol/min/mg of protein ( K m / V max = 0.040) ( 29 ) ( Table 1 ). Therefore, RBM14C16 is a better substrate of hNC than CerC12NBD, as its K m value is one-half lower (16 M) and its V max is almost three times higher, which affords a K m / V max ratio 5.8 times lower (0.0069) ( Table 1 ).
Regarding the mNC, studies were carried out using acid ceramidase-defi cient Farber cells transiently transfected with the mouse ASAH2 gene. In this case, a preferential trend for the short acyl chain substrates was observed both in intact and lysed cells. The similarity between both sets of results excludes the possibility that the substrate preference observed in intact cells is due to putative transacylation of the administered substrate. Intriguingly, the mNC also hydrolyzes RBM14C8, which is not accepted as a substrate by any other ceramidase so far studied. The different substrate preferences between bacterial NC, hNC, and mNC are likely due to differences in their amino acid sequences that may impart different 3D structures to the substrate binding pockets.
Among the three ACERs, only ACER3 has been shown to hydrolyze RBM14, with a preferential activity over the C14 derivative. This conclusion was drawn from results obtained in ASAH2 ( Ϫ / Ϫ ) MEFs, which exhibited an expected lower capacity to hydrolyze RBMM14C12 than wild-type MEFs at neutral pH. These cells, containing ACER3 as the almost RBM14C14 and RBM14C16, although both appear to be poor substrates.
In summary, besides AC, NC, ACER3, and NAAA also hydrolyze fl uorogenic RBM14 compounds. Although the selectivity of the substrates toward ceramidases can be modulated by the length of the N -acyl chain, none of them is specifi c for a particular enzyme, except for RBM14C8, which seems to be accepted only by the ectopically expressed mNC. Despite the lack of specifi city, the excellent Z factors obtained for the assays with both NC and ACER3 warrant the suitability of these substrates in high-throughput library screening programs aimed at identifying potent and selective inhibitors of NC and ACER3, which are currently unknown.