Progress in the development of β-lactams as N-Acylethanolamine Acid Amidase (NAAA) inhibitors: Synthesis and SAR study of new, potent N-O-substituted derivatives.

The anti-in ﬂ ammatory effects resulting from raising the levels of palmitoylethanolamide (PEA), an endogenous bioactive lipid, led to envisage N -Acylethanolamine Acid Amidase (NAAA), the cysteine hydrolase mainly responsible for PEA degradation, as an attractive target for small molecule inhibitors. Previous work in our group identi ﬁ ed serine-derived b -lactams as potent and systemically active inhibitors of NAAA activity. Aiming to expand the SAR study around this class of compounds, we inves- tigated the effect of the substitution on the endocyclic nitrogen by designing and synthesizing a series of N -substituted b -lactams. The present work describes the synthesis of new N - O -alkyl and N - O -aryl substituted b -lactams and reports the results of the structure activity relationship ( SAR ) study leading to the discovery of a novel, single-digit nanomolar NAAA inhibitor ( 37 ). Compound 37 was shown in vitro to inhibit human NAAA via S -acylation of the catalytic cysteine, and to display very good selectivity vs. human Acid Ceramidase, a cysteine amidase structurally related to NAAA. Preliminary in vivo studies showed that compound 37 , administered topically, reduced paw edema and heat hyperalgesia in a carrageenan-induced in ﬂ ammation mouse model. The high in vitro potency of 37 as NAAA inhibitor, and its encouraging in vivo activity qualify this compound as a new tool for the study of the role of NAAA in in ﬂ ammatory and pain states.


Introduction
Palmitoylethanolamide (PEA), a member of the Fatty Acid Ethanolamides (FAEs) family, is an endogenous bioactive lipid able to inhibit peripheral inflammation and mast cell degranulation by activating the peroxisome proliferator-activated receptor-a (PPAR-a) [1,2] as well as to exert antinociceptive effects in rat and mouse models of acute and chronic pain [3e5]. The anti-inflammatory effects resulting from raising local levels of PEA led to envisage N-Acylethanolamine Acid Amidase (NAAA), the lysosomal cysteine hydrolase mainly involved in PEA degradation [6], as an attractive target for small molecule inhibitors for the treatment of inflammatory and pain conditions [7]. N-Acylethanolamine Acid Amidase (NAAA) is activated by autoproteolysis at acidic pH generating a catalytically competent subunit of the enzyme bearing a cysteine (Cys131 in mice, Cys126 in humans) as the nucleophile residue responsible for FAEs hydrolysis [8e10].
Although potent towards our target enzyme, b-lactone inhibitors showed a limited plasma stability (t ½ < 10 s), thus preventing their potential use as systemic drugs [16,18,19]. Aiming to further explore the therapeutic utility of NAAA inhibitors, the replacement of the b-lactone core with a structurally similar blactam moiety was envisioned, leading to a first series of serinederived b-lactam amides [20]. Among the different amide analogues, the N-[(S)-2-oxoazetidin-3-yl]nonanamide (2, Fig. 1) showed good inhibitory potency in vitro against human NAAA (h-NAAA IC 50 ¼ 0.34 mM) and an acceptable mouse plasma stability (t ½ ¼ 130 min). After intravenous and oral administration in rat, compound 2 exhibited a good oral bioavailability (F ¼ 67%), supporting the use of such class of compounds for systemic administration [20]. Additionally, as demonstrated by a recently published SAR study, the functionalization of the exocyclic amino group of the b-lactam scaffold as a carbamate afforded compounds with potent NAAA inhibitory activity [21]. Notably, compound 3 (ARN0726, Fig. 1) showed a good inhibitory activity in vitro, on both rat and human NAAA (r-NAAA IC 50 ¼ 0.028 mM, h-NAAA IC 50 ¼ 0.073 mM) and was recently reported to exert anti-inflammatory effects in mouse models of inflammation and suppress LPS-stimulated inflammatory reactions in human macrophages [22].
Up to the present, no crystallographic structure of the target protein (NAAA) is available. Therefore, although potent enzyme inhibitors have been identified, the precise interactions between those molecules and the amino acidic residues of the catalytic binding site cannot be described in detail. In order to collect additional information on the NAAA binding pocket, novel b-lactam derivatives substituted on the endocyclic nitrogen (N1) were designed. The chemical growth of those compounds on their lefthand side is expected to shed additional light on the stereoelectronic requirements of the NAAA catalytic site. To this purpose, the extensively characterized ARN0726 was selected as starting point for our SAR study. With the aim to maintain fixed the chemical functions essential for NAAA inhibition (i.e. the b-lactam core and the exocyclic nitrogen derivatization as carbamate) as well as to further explore the available chemical space on the b-lactam scaffold, the functionalization of the endocyclic nitrogen (N1) with different substituents was envisaged. To this purpose, a few synthetically accessible N1-substituted derivatives were designed, synthesized and tested against h-NAAA. Notably, among the possible chemical modifications on N1, the corresponding N-Oalkylation was clearly envisaged as the most beneficial to retain significant h-NAAA inhibition ( Table 1).
The aim of the present work was to investigate the effect of the N-O-substitution on h-NAAA inhibitory activity as well as to evaluate the available space in the enzyme active site by designing and synthesizing a novel class of N-O-substituted b-lactams. A conventional SAR study was performed to uncover the chemical features beneficial for NAAA inhibition, leading to the discovery of new potent b-lactam inhibitors that were further characterized also in term of stability, selectivity and in vivo activity.

Chemistry
A slight modification of the reported synthetic pathway to unprotected serine-derived b-lactams [21] provided an efficient procedure to easily prepare the new N-substituted analogues. Basedcatalyzed nucleophilic substitutions using 3 (ARN0726) and the corresponding methyl iodide, acetyl chloride and methylchloroformate afforded compounds 4, 5, and 6, respectively, in  moderate yield (Scheme 1). Concerning the N-O-methyl substituted serine-and threoninederived stereoisomers, the final compounds were obtained starting from the corresponding enantiomerically pure amino acid in a five-step synthetic sequence, as already described in a previous work (see Supporting Information, Schemes S1-3) [21].
The novel N-O-substituted a-amino b-methyl b-lactams were prepared by the enantioselective synthetic pathway reported in Scheme 2. Except for the N-O-tert-butyl derivative 13 (see Supporting Information, Scheme S4), all the N-O-alkyl and aryl derivatives were obtained by installation of the corresponding substituent onto the endocyclic nitrogen of the common building block 18. Two different series of N-O-substituted analogues were synthesized.
Starting from the enantiomerically pure Boc-L-threonine (14), the first step consisted in the conversion of the free carboxylic acid into the benzyloxy amide derivative, followed by transformation of the secondary alcohol into the corresponding mesylate (15). An intramolecular base-catalyzed (potassium carbonate) cyclization reaction in acetone afforded the corresponding N-O-benzyl aamino b-methyl b-lactam core 16 with inversion of stereochemistry at C3. The insertion of the desired carbamic acid side chain on the exocyclic amino group was accomplished as previously reported for the synthesis of other b-lactam derivatives [21]. First, compound 16 was Boc-deprotected and the obtained amine transformed into the tosylate salt. The carbamoylation reaction was carried out with a mixture (1.6:1 isomeric ratio) of 2-pyridyl carbonate (S18 1 ) and 2oxopyridine-1-carboxylate (S18 2 ) derivatives, previously obtained by activation of 4-cyclohexylbutan-1-ol with 2-DPC in the presence of catalytic DMAP (see Supporting Information, Scheme S1) [21,23], to afford the 4-cyclohexylbutyl derivative 17. As last step, the hydrogenolytic removal of the benzyl group of 17 afforded the Nhydroxy substituted b-lactam key intermediate 18 in high yield. The yields of the single steps were generally good, varying from 50% in the carbamoylation reaction to quantitative for the salt preparation.
The final step to achieve the N-O-substituted a-amino b-methyl b-lactam analogues started from 18 which was either submitted to a classic Mitsunobu reaction [24,25] with various alcohols to obtain the N-O-alkyl derivatives (19e29), or to a copper catalyzed coupling reaction [26] with arylboronic acids to afford the corresponding N-O-aryl b-lactams (30e39) (Scheme 2).

Results and discussion
As the first step, in our study we investigated the possible substitutions on the endocyclic nitrogen (N1) of b-lactam derivative 3 (ARN0726), compatible with the NAAA inhibitory activity of this chemical class. We synthesized a small set of derivatives (Table 1) and tested their ability to inhibit the hydrolysis of N-(4-methyl-2oxo-chromen-7-yl)-hexadecanamide (PAMCA) by recombinant h-NAAA heterologously expressed in HEK293 cells (See Supporting Info) [21].
Previous works on b-lactams carried out by our research group reported compound 3 as a potent, double-digit nanomolar NAAA inhibitor with pronounced in vivo anti-inflammatory properties in mouse models [21,22]. In this context, a strong stereorecognition by the target enzyme was highlighted, being the (R)-enantiomer around 40-fold less active than the (S)-enantiomer in vitro and failing to show anti-inflammatory activity in vivo.
Data reported in Table 1 showed that, while the insertion of a methyl (4), an acetyl (5) or a methylcarbamoyl (6) moiety resulted in decrease or loss of activity compared to ARN0726, the substitution of the hydrogen on the endocyclic nitrogen with a methoxy group, as in compound 7, produced a significant improvement in the h-NAAA inhibitory potency (IC 50 ¼ 0.016 mM). Surprisingly, the (R)-serine-derived compound 8, the enantiomer of 7, showed inhibitory potency in the low nanomolar range (IC 50 ¼ 0.056 mM), leading to a loss of the stereorecognition previously observed in similar derivatives [21]. Furthermore, based on previous findings, we asked whether the introduction of a b-methyl substitution on the b-lactam ring would recover such stereorecognition. This modification was earlier reported to restore the effect of stereochemistry on the activity of isomeric threonine-derived analogues [21], and also to beneficially affect the selectivity towards human acid ceramidase (h-AC), a cysteine amidase that exhibits 33% amino acid identity with NAAA [27,28]. To answer this question, the corresponding four stereoisomeric N-O-methyl substituted threoninederived analogues (9e12) were synthesized. Interestingly, while stereoisomers 9e11 resulted to be poorly active against h-NAAA (Table 1), the b-lactam 12, bearing the (2S,3S) configuration, retained a good potency on the target enzyme (IC 50 ¼ 0.069 mM).
Derivative 12 not only showed promising activity on h-NAAA, but also a considerably high selectivity versus h-AC, displaying ca. 50%  Table 2 and Table 3 (Table 2). First, the optimal length of a linear alkyl chain was evaluated (12, 19e21). Increasing the length from a simple methyl group (12, IC 50 ¼ 0.069 mM, Table 2) up to longer alkyl chains led to a progressive decrease in potency. Whereas a good inhibitory activity was retained up to a C5 linear alkyl moiety (20, IC 50 ¼ 0.387 mM), a remarkable drop was observed for compound 21 (IC 50 ¼ 2.23 mM) bearing a longer alkyl chain (C7). Notably, both the cyclohexyl (22, IC 50 ¼ 3.01 mM) and the tert-butyl substitutions (13, IC 50 > 50 mM) were detrimental for enzyme inhibitory activity, probably due to steric hindrance in close proximity of the catalytic active site. Based on these findings, aiming to get additional structural information around the catalytic site, compound 20 was selected as a starting point for the synthesis of a small set of derivatives (23e29) bearing different substitutions on the alkyl chain. As shown in Table 2, the insertion of a terminal cyclohexyl (28) as well as of a phenyl (29) moiety on the C5 linear chain led to a decrease in potency (h-NAAA IC 50 ¼ 3.47 mM and 1.24 mM, respectively), suggesting a limited space in this portion of enzyme catalytic site. Interestingly, the derivatives 23e27, bearing an ether (23,24), a thioether (25), a methylsulfonyl (26), and a ketone (27) moiety respectively, displayed good inhibitory potencies with IC 50 values ranging between 0.049 and 0.187 mM.  Reagents and conditions: a) O-Benzyl-hydroxylamine, EDC HCl, HOBt, THF, room temperature, 5 h; b) Mesyl chloride, pyridine, 0 C to room temperature, 15 h; c) K 2 CO 3 , acetone, reflux, 3 h; d) p-Toluensulfonic acid, TFA, room temperature, 15 min; e) Mixture (1.6:1 ratio) of 4-cyclohexylbutyl pyridin-2-yl carbonate (S18 1 ) and 4-cyclohexylbutyl 2oxopyridine-1(2H)-carboxylate (S18 2 ), DIPEA, dry DCM, room temperature, 15 h; f) Cyclohexene, 10% palladium on activated carbon (Pd/C), EtOH, room temperature, 2 h; g) R 1 OH, PPh 3 , DEAD, dry THF, 0 C to room temperature, 5e12 h; h) R 2 B(OH) 2 , CuCl, pyridine, dry DCM, room temperature, 2e10 h. of the h-NAAA inhibitory activity. The introduction of a methylsulfonyl group (26: IC 50 ¼ 0.049 mM) on the left hand side of the molecule led to an increase in potency, suggesting the beneficial effect of a di-oxygenated substituent on NAAA inhibition. We next turned our attention to the N-O-aryl bÀlactams and investigated the role of the substituted phenyl rings on NAAA inhibition (Table 3). In general, all N-O-aryl derivatives showed a very good inhibitory potency, higher than that of the alkyl analogues, with IC 50 s in the low nanomolar range. We speculated that the higher potency of these derivatives could be explained by an extended conjugation of the phenoxy residue with the b-lactam amidic bond, making the carbonyl more prone to nucleophilic attack by the catalytic cysteine of NAAA.
The unsubstituted phenyl derivative 30 showed a three-fold improvement in IC 50 compared to the methyl derivative 12 (IC 50 ¼ 0.023 mM). A similar inhibitory potency (IC 50 ¼ 0.017e0.031 mM) was also displayed by compounds 31e34, bearing electron-donating (31) or electron-withdrawing (32e34) substituents on the phenyl ring. The apparent lack of correlation between the stereoelectronic properties of the phenyl substitution and the inhibitory activity on the enzyme remains to be explained. Despite the increased bulkiness, the para-phenyloxy (35) and paraphenyl (36) substitutions on the phenyl ring demonstrated to be well tolerated by the enzyme (IC 50 ¼ 0.048 mM and 0.039 mM, respectively). At this point, considering the beneficial effect on activity of a methylsulfonyl substitution on the N-O-alkyl substituted analogues, we explored such modification also on the aryl substituted b-lactam derivatives. Noteworthy, a significant increase in potency with respect to phenyl compound (30) was observed with the para-methylsulfonyl derivative 37 (IC 50 ¼ 0.006 mM), suggesting a possible electrostatic interaction of the two oxygen atoms with the amino acid residues of NAAA active site. To better investigate the effect of this substitution on the phenyl ring, the meta-(38) and ortho-(39) methylsulfonyl derivatives were also synthesized. As shown by the IC 50 values (meta-: IC 50 ¼ 0.010 mM and ortho-: IC 50 ¼ 0.006 mM) both substituted compounds resulted to be equally active to the para analogue 37, suggesting this type of substitution to be beneficial for a potent NAAA inhibition regardless of the substituent position on the phenyl ring.
In this context, we also synthesized and tested few N-O-aryl blactams bearing a substitution in meta-or ortho position. As seen for methylsulfonyl substituted compounds (37e39), the position of  the substituent on the phenyl ring did not affect the compounds' NAAA inhibitory activity (data not reported).
The activity data on h-AC inhibition showed in the case of N-Oalkyl derivatives 12 and 26 a very high selectivity (h-NAAA/h-AC > 1000 fold), which drastically decreased when the endocyclic nitrogen was substituted with a phenyloxy moiety (30), showing only a 13-fold difference between the activity of the two cysteine hydrolases. The selectivity on h-NAAA was recovered when a methyl sulfonyl group was inserted in para position (37, h-AC/h-NAAA ¼ ca. 340 fold).
At the tested concentration (10 mM), none of the selected N-Osubstituted b-lactam carbamates inhibited h-FAAH in a significant manner (Table 4).
To investigate the physicochemical properties of this new blactam class, the same subset of N-O-substituted a-amino b-methyl b-lactams was also evaluated for the chemical stability in buffer (PBS) at pH 7.4 and for the mouse and rat plasma stability ( Table 4).
The tested compounds showed an overall high chemical stability to hydrolytic cleavage, with half-life of over 24 h (12, 26, 30: t ½ > 1440 min). A lower but still acceptable half-life (t ½ ) value was observed for the para-methylsulfonyl derivative 37 (t ½ ¼ 727 min).
With respect to the plasma stability, all selected compounds displayed very low half-life (t ½ < 5 min) either in rat and mouse plasma, due to a fast hydrolytic cleavage.
In contrast with previously described b-lactam derivatives [21], the low plasma stability displayed by N-O-substituted analogues clearly prevents the potential use of this new class of NAAA inhibitors for systemic administration. However, this limitation (i.e. low plasma stability) may be well turned into an advantage when administering these compounds by topical applications, thus restricting NAAA inhibition to the site of administration and limiting their potential side effects. To assess the in vivo activity of a representative compound of this series, we tested compound 37 by topical administration in a mouse model of carrageenan-induced inflammation (see Supporting Information). After induction of paw edema and heat hyperalgesia, a fresh suspension of 37 was given by intraplantar administration.
A single dosing of 37 (0.5e50 mg/paw) was able to significantly attenuate both responses ( Fig. 2A and 2B), and when 37 was administered at the highest dose (50 mg/paw), the effect was still statistically significant 24 h after treatment ( Fig. 2A, P < 0.01).
The same compound 37 was also used as a tool to evaluate the mechanism of NAAA inhibition by this new class of b-lactams. As previously shown for other b-lactam NAAA inhibitors [16,21], high resolution liquid chromatography mass spectrometry (HRLC/MS-MS) studies proved the formation of a covalent thioester adduct with the N-terminal cysteine of the active form of h-NAAA (see Supporting Information). The covalent nature of NAAA inhibition by this class of compound was also confirmed by a competitive Activity Based Protein Profiling (ABPP) study using ARN14686, a recently published probe for NAAA [30]. We used the lysosomal enriched fraction from HEK-293 cells stably overexpressing NAAA (NAAA HEK-293) as source of the target enzyme. This fraction was pre-incubated with compound 37 (20 mM) for one hour and then a five-fold higher concentration of ARN14686 was added after 15 min or 4 h. Subsequent click chemistry reaction with ARN14686 allowed the introduction of a fluorophore (i.e., rhodamine) to detect the catalytically active portion of the enzyme. The SDS-PAGE analysis (Fig. 3) clearly showed that b-lactam 37 was able to rapidly and fully inhibit NAAA, as demonstrated by the absence of any fluorescent signals after 15 min incubation with ARN14686. The same result was obtained after 4 h incubation with ARN14686, thus indicating that NAAA inhibition by compound 37 is irreversible, at least under this experimental conditions.

Conclusions
The present study expanded the previous SAR on serine-and threonine-derived b-lactams as NAAA inhibitors by investigating the influence of the endocyclic nitrogen O-substitution on the activity. First, we explored the effect of a b-methyl substitution on the b-lactam ring: as already seen with other analogues, a strong recognition towards the (2S,3S) configuration (compound 12) was confirmed also for this class. Then, N-O-alkyl and N-O-aryl substituted compounds were investigated, disclosing a significant preference of the enzyme towards aryl and polar substituents. The expansion of the aryl series, starting from the phenyl derivative 30, led us to compound 37, a single-digit nanomolar NAAA inhibitor (h-NAAA IC 50 ¼ 0.006 mM). Surprisingly, similar IC 50 values were observed for all para-substituted phenyl compounds, indicating that the stereo-electronic properties of the substituent did not affect NAAA inhibitory potency. Furthermore, the four-fold increased potency displayed by the methylsulfonyl derivative 37 compared to compound 30, suggested the importance of oxygencontaining substituents for the interaction with the enzyme. The results of our SAR study showed that b-lactam NAAA inhibitors could be substituted at the endocyclic nitrogen with O-alkyl and Oaryl groups, besides the exocyclic nitrogen of the a-amino group.
This results in linearly extended compounds where the cysteine warhead is no longer at one end of the molecule, as is the case in blactones or b-lactams, like compounds 2 and 3. The enzyme appears therefore able to accommodate relatively extended small molecules at the active site. The introduction of properly substituted O-aryl moieties at the endocyclic nitrogen was particularly beneficial in terms of NAAA inhibitory activity, and offered the possibility of modulating the substituent at the exocyclic nitrogen in order to Table 4 Inhibitory potency (IC 50 ) or % inhibition of selected N-O-substituted b-lactams (12, 26, 30,  access further b-lactams with improved NAAA inhibitory activity and desired chemical functionalities. Based on the interesting outcomes reported for the first NAAA activity-based probe [29], N-O-substituted b-lactams could also be exploited for the design of novel ABPs. This would allow the development of additional pharmacological tools for the in vitro and in vivo detection of the catalytically active form of NAAA, thus contributing to improving our understanding of the enzyme's physiological roles.
A set of this new b-lactams was also evaluated for hydrolytic and plasma stability. While all selected compounds showed a good chemical stability at pH 7.4 in PBS, upon incubation with rat and mouse plasma they were rapidly hydrolyzed, suggesting their possible use as soft drugs for skin diseases. In line with these data, a preliminary in vivo study with compound 37 in a carragenaaninduced inflammation mouse model demonstrated the promising activity of such compound when administered topically. This encouraging result lends support to its possible use as soft drug for skin diseases.

Chemicals, materials and methods
All the commercial available reagents and solvents were used as purchased from vendors without further purification. Dry solvents (THF, Et 2 O, CH 2 Cl 2 , DMF, DMSO, MeOH) were purchased from Sigma-Aldrich. Automated column chromatography purifications were performed using a Teledyne ISCO apparatus (CombiFlash ® Rf) with pre-packed silica gel columns of different sizes (from 4 g up to 120 g). Mixtures of increasing polarity of cyclohexane (Cy) and ethyl acetate (EtOAc) or cyclohexane and methyl tert-butyl ether (MTBE) were used as eluents. NMR experiments were run on a Bruker Avance III 400 system (400.13 MHz for 1 H, and 100.62 MHz for 13 C), equipped with a BBI probe and Z-gradients. Spectra were acquired at 300 K, using deuterated dimethylsulfoxide (DMSO-d 6 ) as solvent. UPLC/MS analyses were run on a Waters ACQUITY UPLC/ MS system consisting of a SQD (single quadrupole detector) mass spectrometer equipped with an electrospray ionization interface and a photodiode array detector. The PDA range was 210e400 nm. Analyses were performed on an ACQUITY UPLC BEH C 18 column (100 Â 2.1mmID, particle size 1.7 mm) with a VanGuard BEH C 18 precolumn (5 Â 2.1 mmID, particle size 1.7 mm). Mobile phase was 10 mM NH 4 OAc in H 2 O at pH 5 adjusted with CH 3 COOH (A) and 10 mM NH 4 OAc in CH 3 CN-H 2 O (95:5) at pH 5.0. Electrospray ionization in positive and negative mode was applied. Purifications by preparative HPLC/MS were run on a Waters Autopurification system consisting of a 3100 single quadrupole mass spectrometer equipped with an Electrospray Ionization interface and a 2998 Photodiode Array Detector. HPLC system included a 2747 sample manager, 2545 binary gradient module, system fluidic organizer and 515 HPLC pump. PDA range was 210e400 nm. Purifications were performed on a XBridge™ Prep C 18 OBD column (100 Â 19 mmID, particle size 5 mm) with a XBridge™Prep C 18 (10 Â 19 mmID, particle size 5 mm) guard cartridge. Mobile phase was 10 mM NH 4 OAc in MeCN-H 2 O (95:5) at pH 5. Electrospray ionization in positive and negative mode was used. Optical rotations were measured on a Rudolf Research Analytical Autopol II automatic polarimeter using a sodium lamp (589 nm) as the light source; concentrations are expressed in g/100 mL using CHCl 3 as a solvent and a 1 dm cell. All tested compounds (4e13, 18e39) showed ! 95% purity by NMR and UPLC/MS analysis.

Synthesis of the N-substituted serine-derived b-lactams (4e6)
Experimental procedure and 1 H NMR relative to 3 (ARN0726) are according to previously reported literature [21].   filtered over a celite pad, diluted with DCM and evaporated to dryness. The crude product was absorbed over silica gel and purified by column chromatography using a Teledyne ISCO apparatus, eluting with Cy/EtOAc (from 100:0 to 20:80) to afford a crude product, which was submitted to a preparative HPLC/MS using a Waters Autopurification system, affording the title compound   (6). Under nitrogen atmosphere, to a stirred solution of 3 (0.05 g, 0.18 mmol) in dry DCM (3.0 mL), triethylamine (0.056 mL, 0.409 mmol) and methyl-carbonyl chloride (0.017 mL, 0.223 mmol) were slowly added at 0 C. After stirring 15 h at room temperature (still starting material left) the solution evaporated to dryness. The crude product was absorbed over silica gel and purified by column chromatography using a Teledyne ISCO apparatus, eluting with Cy/EtOAc (from 100:0 to 40:60) to afford a crude product, which was submitted to a preparative HPLC/MS using a Waters Autopurification system, affording the title compound Step 1 to 4. See Supporting Information.
Step 5. Under nitrogen atmosphere, a solution of S13 (0.082 g, 0.463 mmol) in dry CH 2 Cl 2 (7.0 mL) was treated with DIPEA (0.092 mL, 0.55 mmol). A second solution containing a mixture (ratio 1.6:1) of 4-cyclohexylbutyl 2-pyridyl carbonate (S18 1 ) and 4cyclohexylbutyl 2-oxopyridine-1-carboxylate (S18 2 ) (0.352 g, 1.27 mmol) in dry CH 2 Cl 2 (2.0 mL) was then added. The reaction was left to stir under nitrogen atmosphere at room temperature for 15 h. The solvent was evaporated and the crude product was dissolved in EtOAc (10 mL), washed first with a saturated NH 4 Cl aqueous solution (20.0 mL), subsequently with a saturated NaHCO 3 (3 Â 20 mL) and NaCl (20 mL) aqueous solutions. The organic layer was dried over Na 2 SO 4 , filtered and evaporated to dryness. The crude was absorbed over silica gel and purified by column chromatography using a Teledyne ISCO apparatus, eluting with Cy/ TBME (from 100:0 to 30:70) to afford the title compound Step 1 to 3. See Supporting Information.
Step 4. To a refluxing (90 C) slurry of powdered K 2 CO 3 (0.14 g, 1.02 mmol) in acetone (5.0 mL), a solution of S20 (0.1 g, 0.254 mmol) in acetone (3.0 mL) was added. The resulting suspension was stirred at 100 C for 3 h. Upon cooling, the thick slurry was filtered through celite and the collected solid was extracted with EtOAc (20.0 mL). The organic layers were washed sequentially with a 1 N HCl aqueous solution (10.0 mL), a saturated NaHCO 3 solution (20.0 mL) and a saturated NaCl solution (10.0 mL), dried over Na 2 SO 4 , filtered and concentrated to dryness. The crude product was absorbed over silica gel and purified by column chromatography using a Teledyne ISCO apparatus, eluting with Cy/ TBME (from 100:0 to 70:30) to afford the title compound
In a heart shaped flask purged with nitrogen, 16 (2.72 g, 8.87 mmol) was mixed with p-toluensulfonic acid (1.60 g, 9.30 mmol) and the solid mixture was cooled to 0 C. Subsequently trifluoroacetic acid (36 mL) was added dropwise over 20 min and the reaction was left to stir at 0 C for 15 min. The solution was rotary evaporated maintaining the bath below 30 C and the obtained oil was left under high vacuum for 1 h. Et 2 O (20.0 mL) was added to the obtained crude and a white precipitate was observed. After washing with Et 2 O (5 Â 10 mL) the title compound Step 5. 4-Cyclohexylbutyl 2-pyridyl carbonate (S18 1 ) and 4cyclohexylbutyl 2-oxopyridine-1-carboxylate (S18 2 ).

Analysis of covalent adducts by LC-MS/MS
The presence of possible inhibitor covalent adducts on h-NAAA was investigated by high-resolution nano LC-MS/MS analysis. A solution of purified h-NAAA (2.5 mM) was incubated with compound 37 (25 mM in 5% DMSO, final concentration). A no-inhibitor control (DMSO 5%) was also included. After the reaction time, samples were precipitated with 10 vol of cold acetone and centrifuged 10 min at 12000 Â g. The pellets were re-suspended in 50 ml of 50 mM NH 4 HCO 3 (pH 8) and trypsin (1:50 w/w) was added for 16 h at 37 C. The pellets were resuspended in 50 ml of NH 4 HCO 3 (pH 8) and proteomic grade trypsin (1:50 w/w) was added for 16 h at 37 C. The resulting peptides were analyzed on a UPLC chromatographic system equipped with a BEH C18 reversed phase column (1 Â 100 mm). Peptides were eluted with a linear gradient of CH 3 CN in water (both added with 0.1% formic acid) from 3 to 50% in 8 min. Flow rate was set to 0.09 mL/min. Eluted peptides were analyzed in positive ion mode by high-resolution tandem mass spectrometry on a Synapt G2 qTOF mass spectrometer (UPLC, column and qTOF instrument were purchased from Waters, Milford MA, USA). A linear ramp of the collision energy from 15 to 45 eV was used to induce backbone fragmentation of the eluting peptides. 500 nM gluco-fibrino peptide infused at 500 nL/min was used as lock spray mass. MS/MS data were analyzed using the BioLynx software embedded in the MassLynx software suite. MassLynx and ProteinLynx software (Waters, USA) were used for the interpretation of LC-MS data.

Competitive Activity Based Protein Profiling (ABPP)
For competitive ABPP, 50 ml of lysosomal enrichment (0.5 mg/ mL) from of h-NAAA-overexpressing HEK293 cell line were incubated 1 h at 37 C with compound 37 at a final concentration of 10 mM (DMSO 2%). Following the incubation time, the activity based probe ARN14686 [30] was added at 10 mM for 15 min or for 3.5 h at 37 C. Next, click chemistry reaction was performed by adding the