Here we present 7-hydroxyresorufinyl octanoate (1c) as a new fluorogenic substrate to evaluate MAGL activity by HTS. Although numerous assay methodologies have been published in the past decades for measurement of MAGL activity, few of them allow the screening of a large number of compounds showing high sensibility, high throughput, and low assay cost. In order to overcome these issues, we investigated the use of an alternative substrate for measuring MAGL activity. The native substrates for MAGL include 2-AG (C20:4) and 2-oleoyl glycerol (2-OG, C18:1) both containing long hydrocarbon fatty acid chains with relatively unstable cis olefins. Nevertheless, MAGL is able to hydrolyze a vast array of esters structurally very different, such as p-nitrophenyl esters [9] or 7-hydroxycoumarinyl esters [22]. We chose resorufine as fluorescent probe because its derivatives have proven to be an excellent fluorescence marker and have the advantages to be hardly susceptible to background signals. Often pharmacological and biochemical research need to measure enzymatic activity in crude cell lysate or turbid solution [23], the shift of emission wavelengths at higher values give less interference [24]. In the current approach, 10 fluorogenic alkyl esters of resorufin with different acyl chains are synthesized and evaluated as MAGL assay substrates (Figure 1). Selected compound 1c ensure lower background signals, thanks to resorufine in comparison with 7-methylumbelliferone, higher stability and lower costs if compared to arachidonoyl or oleyl derivatives.

MAGL fluorescence assays were carried out in the presence of 5 μM substrate, 10 % DMSO and at pH 7.4, a value that attempts to mimic the native intracellular conditions. Diverging too far from the physiological pH could affect the interactions between the enzyme and the substrates, thus giving an altered view of the potential biochemical and cellular activity of these compounds. Moreover, this pH value can be maintained by various buffers, such as Tris-HCl, PBS, sodium phosphate, Mops and Hepes, and allowed us a prompt comparison with previous data [12]. Among the synthesized compounds, linear acylic chains of up to 12 carbon atoms are substrates for MAGL, as well as the derivative of oleic acid, as expected, since 2-OG, an oleic acid derivative, is one of the native substrates of the protein. The addition of just two more carbons to the side chain makes the hydrolysis of this compound almost non-existent. The same result is obtained when the branching becomes as long as four carbon atoms. When in aqueous buffer, 1a and 1b show the highest rate of spontaneous hydrolysis among our compounds (Figure 2), to the point that fluorometric determination of their kinetic constants proved impossible (Figure 3). For resorufin acetate 1a, this behavior was already reported by Maeda and colleagues [25], leading to the impossibility of using that compound as a probe. Even if 1b showed an inferior hydrolysis rate than 1a (Figure 1), this compound is unsuitable for the achievement of Michaelis-Menten conditions in an aqueous environment (Figure 3) [26]. On the other hand, compounds with a long linear chain or with a branched side chain, showed low to negligible hydrolysis. These differences may be explained with the susceptibility of acyl resorufins towards nucleophilic attack [27]. Short linear side chains (1a, 1b) leave the α carbon exposed to the attack of water, while side chains with branches next to the carboxyl group (1h, 1i, 1j) exert a steric hindrance on water molecules, protecting the ester group from hydrolysis. A similar phenomenon reasonably happens with medium (1c, 1d) and long (1f, 1g) alkyl chains. When in an aqueous environment, the alkyl chain folds on itself via van der Waals interactions next to the carboxyl group, thus mimicking the side branch protecting effect. Moreover, the formation of aggregates, as highlighted by DLS analyses (Supplementary Figure S2), could be a reasonable explanation for the unusual behavior observed in spontaneous hydrolysis experiments of compounds 1d–h where to a five-fold increase of ester concentration corresponds a less increase in the rate of the hydrolysis. In Figure 6, hydrolysis of different substrates by MAGL, normalized against compound 1c, is reported to better underline the high difference between the hydrolysis rate displayed by 1c and the other proposed MAGL substrates.

The presence of aggregates in the substrates’ solutions at 25 μM concentration prompted us to consider only solutions until 10 μM concentration for calculations of kinetic parameters. Observing obtained results (Figure 3), 1f gives a good fit with Michaelis-Menten type kinetics indeed, its side chain, the same of natural substrate 2-OG, seems to allow a better accommodation in the enzyme active site. It is possible that the enhanced rigidity of the alkyl chain, conferred by the presence of the double bond, guides the substrate to a conformation favorable to enzymatic hydrolysis, hindering disadvantageous folding in the enzyme active site.

The scale of Y axis (Figure 3), and the V_max and K_m values (Table 1), highlight how steric hindrance, due to the different length of acyl side chain on C2, reduce the rate of substrate hydrolysis. Interestingly, 1j shows a Michaelis-Menten kinetic behavior even if with very low hydrolysis rate. In this case, the bulkier side chain likely exerts an even stronger effect of steric hindrance, which does not allow reaching the concentration needed to manifest the inhibitory effect. Compound 1c proved to be a very interesting substrate for MAGL. Indeed, with respect to the previously described arachidonate 1g [12] and oleate 1f, the two resorufine derivatives of the MAGL natural substrates, 1c shows slightly higher XP Glide docking score, MM-GBSA binding energy and K_m values. In the molecular docking poses, 1c resorufin perfectly overlaps the 1g one, while the linear side chain of 1c mimics the 2-AG side chain. Focussing our attention specifically on the difference in activity pointed out for compound 1c versus 1b, as already reported in the Results section and in Figure S4, a significant relative difference in the extension of the hydrophobic interactions can be one of the factors contributing to the different stability of the complexes, as shown by the different of the solvation/desolvation energy reported in the Prime MM-GBSA energy column in Table 1. The V_max for 1c, however, is much higher than that of 1g, as indicated also by the direct comparison of the catalytic speed of the two substrates. In addition, the smaller substrate 1c has superior aqueous solubility properties and contains no potential labile cis olefins. Therefore, the signal to background ratio of 1c (seven to one) is suitable for the screening for MAGL inhibitors in high-throughput screening. For all these reasons, 1c was selected and validated as new red fluorogenic probe for the HTS method.

MAFP and URB602, two well-known serine hydrolase and MAGL inhibitors, were chosen for validation [7,19]. After a 60 min pre-incubation of the enzyme with the inhibitor, the IC₅₀ values for both MAGL inhibitors using hMAGL and 1c (5 μM) were obtained and compared with those reported in literature [12,20]. In order to evaluate and validate the assay for its potential in high-throughput screening applications, we further determined the Z-factor [28], a dimensionless statistical parameter that reflects both the assay signal dynamic range and the data variation associated with the signal measurements. The assay was highly reproducible with an overall average Z′ value of 0.86. The objective of this study was to identify a substrate readily accessible, stable in solution and with low rate of spontaneous hydrolysis for realizing an HTS method for MAGL activity using human recombinant MAGL. It should be considered that MAGL belongs to the group of serine hydrolases (SHs), a superfamily of enzymes able to cleavage ester, amide, or thioester bonds of protein, peptide, and small molecule substrates [29] including esterases, lipases, peptidases, and amidases [30]. Within this family, there are the SHs that regulate the biosynthesis and degradation of two major endocannabinoids, AEA and 2-AG, as well as other lipid-metabolizing serine hydrolases that are ubiquitously distributed [31]. Therefore, in the potentiality to apply 1c as a probe for experiments performed in complex biological systems, as living cells or cell lysates, the hydrolysis of the probe 1c by other SHs enzymes should be considered. These results demonstrated that our fluorometric method could be successfully applied to identify compounds that modulate MAGL activity.