HirFAAH homology and sequence analysis
Query of the H. verbana TSA using human FAAH amino acid sequence as the query sequence identified the transcribed RNA sequence GGIQ01042388.1. Further analysis of GGIQ01042388.1 resulted in the predicted HirFAAH with a nucleotide length of 1,578 bp, an amino acid length of 525 amino acids and estimated mass of 59.3 kDa. The predicted HirFAAH appears to be more similar to human FAAH-2 (HsaFAAH-2) isoform, based on a 44.1% sequence identity and 59.04% similarity (Fig. 1). The HirFAAH sequence compared to human FAAH-1 revealed 19.19% sequence identity and 33.39% similarity. Consistent with FAAH is other species the HirFAAH sequence possesses a GXSXG motif, a conserved feature in serine hydrolases (black box, Fig. 1). In addition, HirFAAH possesses an amidase signature sequence that exhibits 60.3% sequence identity and 71.1% similarity to HasFAAH2. This amidase sequence contained the serine-serine-lysine catalytic triad (grey highlight and bold/underlined residues, respectively, Fig. 1) and putative catalytic serine (S225) which aligns with the catalytic serine (S230) of HsaFAAH-2 (black arrow, Fig. 2) and are critical for FAAH serine hydrolase activity [33, 34]. Using these sequence data, subsequent qRT-PCR experiments confirmed the expression of HirFAAH in the Hirudo central nervous system (CNS; Fig. 2).
Sequence alignment of HirFAAH were made with FAAH sequences from other species including the mollusk Mizuhopectin yessoensis, the brachiopod Lingula anatina, arthropod Blattella germanica, and the mammal Pteropus alecto. Along and Homo sapiens FAAH (hsaFAAH) these show a high degree of homology in the region making up the catalytic triad and at the C-terminus (data not shown). A phylogenetic tree (Fig. 3) of confirmed or provisional FAAH2 orthologues from other species was constructed with HirFAAH using sequences from the Mediterranean mussel, Mytilus galloprovincialis (VDI01341.1), the Eastern oyster, Crassostrea virginica (XP_022339113.1), the scallop, Mizuhopectin yessoensis (OWF55125.1), the Korean mussel, Mytilus coruscus (CAC5390857.1), the Florida lancet, Branchiostoma floridae (XP_035686393.1), from the pacific oyster, Crassostrea gigas (XP_034331103.1), the termite, Cryptotermes secundus (XP_023715933.1), from the great scallop, Pectin maximus (XP_033740866.1), the snail, Pomacea canaliculate (XP_025099255.1), the German cockroach, Blattella germanica (PSN54917.1), the brachiopod, Lingula anatine (XP_013399812.1), Zebrafish, Danio rerio (NP_001002700.1), and Homo sapiens (NP_777572.2). Sequences ranged from 67.76–43.76% identity matches to the predicted HirFAAH sequence. The phylogenetic groupings of FAAH2 orthologues were largely as one might expect. The Hirudo HirFAAH sequenced grouped with other lophotrochozoan invertebrate species (e.g., mollusks and annelids) and these were distinct from ecdysozoans (e.g., cockroach and termites) and vertebrates (e.g., lancets, zebrafish, and humans). The FAAH2 sequence from the brachiopod Lingula appeared to be distinct from all the other phylogenetic groupings despite being a lophotrochozoan.
Features of HirFAAH expression in HEK-293 cells
The cloned HirFAAH was labeled in frame with a C-terminal eGFP tag resulting in a protein 768 amino acids in length and expressed in 293-HEK cells (Fig. 4). Immunofluorescence microscopy indicated the tagged HirFAAH was localized to hollow, ring-like cytoplasmic structures (Fig. 4a, b). This is consistent with immunofluorescence microscopy experiments performed with HsaFAAH2 that also localized to ring-like structures (Fig. 4a, b) [35]. The eGFP tagged HirFAAH protein was enriched in the membrane fractions of transfected 293-HEK cells (Fig. 4B, C), again consistent with the subcellular distribution of other FAAHs [36, 37]. We also tested a version of HirFAAH with a mutation of the active site serine (S225A). This mutation did not prevent expression of HirFAAH in the 293-HEK cells, nor did it disrupt localization to the membrane fraction as seen in wildtype FAAH (Fig. 4c, d).
Activity-Based Protein Profiling
The catalytic activity of HirFAAH was assessed by ABPP, which detects functionally active serine hydrolases [34, 38]. Membrane fractions were prepared from 293HEK cells transfected with HirFAAH, HirFAAH(S225A), or GFP vector and the relative expression of each protein was determined via western blot (Fig. 5a). These findings indicated successful transfection and expression for all three proteins. The ActiveX™ TAMRA-FP probe labeled an ~ 83 kDa protein in the HirFAAH expressing membrane fractions that corresponded with the expected molecular weight of GFP tagged HirFAAH (Figs. 5b; N = 9). Hydrolase activity at this molecular weight was not present in any of the GFP lanes (Fig. 5b; N = 7). Next, we tested the effects of URB597, a selective, irreversible inhibitor of mammalian FAAH [39], on the serine hydrolase activity of HirFAAH at concentrations 0.01, 0.1, 1 and 10 µM plus a vehicle control (2% DMSO). A 2-way ANOVA was used to assess serine hydrolase activity in samples from HirFAAH-transfected cells vs. those transfected with the GFP-containing vector and the effects of increasing concentrations of URB97 on hydrolase activity. This analysis detected a significant effect of transfection construct with serine hydrolase activity with HirFAAH samples have much greater activity than GFP-only samples (F1,46 = 16.84, p < 0.0001). Preincubation of sample with URB597 decreased hydrolase activity in the HirFAAH in a concentration-dependent manner that was not observed in the GFP samples (Fig. 5b, c). This was confirmed by a significant concentration effect (F4,46 = 23.16, p < 0.0001) and a significant interaction effect (F4,46 = 22.88, p < 0.0001). Post-hoc analysis indicated significant inhibition of HirFAAH activity by 0.1, 1, and 10 µM URB597 (p < 0.0001 for all). No other proteins labeled by the AcivX™ TAMRA-FP probe exhibited any obvious sensitivity to URB597.
AAMCA-based Fluorescence Assay
While the ABPP assay confirms serine hydrolase activity, it does not confirm that HirFAAH actually metabolizes AEA. To address this issue, we used a high-throughput fluorescent screening assay developed by Ramarao et. al. that specifically measures FAAH activity [40]. In this assay, FAAH catalyzes the hydrolysis of a nonfluorescent AAMCA, a substrate that is specific to FAAH, to produce arachidonic acid and the highly fluorescent AMC (excitation 355nm/emission 460nm).
In order to maximize the amount of HirFAAH used in the AAMCA assay and increase the assay’s sensitivity, the membrane fraction of HirFAAH was further purified into a microsomal fraction. Previous work has shown that FAAH activity is highest in the microsomal fraction [40]. As a validation step to assess our microsomal isolation process, the ABPP assays were repeated using microsomal fractions from the transfected 293HEK cells. In these experiments, serine hydrolase activity was compared from microsome samples prepared from cells transfected with HirFAAH (N = 5), the GFP-only vector (N = 2), and HirFAAH(S225A) mutant (N = 4). Consistent with our previous experiments using membrane fraction samples (Fig. 5), the ActiveX™ TAMRA-FP probe labeled a protein (~ 83kDa) in the HirFAAH-expressing microsomal fractions that was not present in the GFP samples (Fig. 6a, b). In addition, no serine hydrolase activity was observed samples transfected with the HirFAAH(S225A) mutant, indicating that the mutation at the proposed active site did disrupt enzyme function. 100 nM URB597 significantly inhibited HirFAAH activity with no effect on samples from the GFP-only and HirFAAH(S225A) groups. This was confirmed by 2-way ANOVA which showed a significant effect of gene product (F4,46 = 9.68, p < 0.005), URB597 treatment (F4,46 = 5.36, p < 0.05), and interaction effect (F4,46 = 7.38, p < 0.01). To summarize, these TAMRA-FP experiments demonstrated that HirFAAH exhibited the expected serine hydrolase activity, and this activity was reduced on a concentration-dependent manner by the FAAH inhibitor URB597. Furthermore, mutation of HirFAAH at the putative active site did prevent serine hydrolase activity.
Next, microsomal fractions from 293HEK cells expressing GFP (N = 3), HirFAAH (N = 4) or HirFAAH(S225A) (N = 4) were incubated with AAMCA plus either 2% DMSO or increasing concentrations of URB597. Figure 6c shows background levels of AMC production in samples expressing GFP alone incubated in 2% DMSO. No statistically significant changes in AMC levels was observed in GFP-only sample incubated in any of the URB597 concentrations (1nM, 10nM, 100nM, 1µM and 10µM). Samples containing wildtype HirFAAH and incubated in DMSO showed substantial AMC production over background levels indicating HirFAAH specific AAMCA hydrolysis (Fig. 6c), consistent with other FAAHs that metabolize AEA. Additionally, there was a statistically significant decrease in the AMC production when the HirFAAH microsomal fraction was incubated with URB597 (Fig. 6c). Concentrations ranging from 1 nM to 10 µM URB597 produced significant inhibition of enzymatic activity (post hoc p < 0.001). Interestingly, samples in this analysis appeared more sensitive to inhibition by URB597 compared to samples in the ABPP assay. In the AAMCA assay, 1 nM URB597 produced a significant decrease in enzymatic activity, whereas a minimum of 100 nM was required to produce a significant decrease relative to the vehicle control in the ABPP assays (see Fig. 5c vs. 6c). This may be a consequence of different sensitivities between the AAMCA and ABPP assays. In the case of HirFAAH(S225A) samples incubated in DMSO, AMC production were at background levels (Fig. 6c) indicating that no hydrolysis activity in this mutant FAAH, consistent with the site mutation interfering with enzymatic activity. HirFAAH(S225A)-containing samples exhibited no change in AMC production when pre-incubated with increasing concentrations of URB597 (Fig. 6c). Two-way ANOVA showed a significant effect of gene product indicating that only HirFAAH-containing samples exhibited enzymatic activity (F2,48 = 12.47, p < 0.001). Analysis also showed a significant effect of URB597 concentration (F5,48 = 13.53, p < 0.001) and a significant interaction effect (F10,48 = 13.10, p < 0.001), indicating that the FAAH inhibitor did reduce enzymatic activity of HirFAAH with no effect on the HirFAAH(S225A) mutant or the GFP control.
Effects of FAAH inhibition on synaptic transmission
Next, we wanted to examine whether URB597 has functional effects on the Hirudo CNS consistent with inhibition of FAAH. Therefore, we examined the effects of URB597 on synaptic transmission by pressure sensory neurons (P cells) in acutely isolated Hirudo ganglia. In previous studies, a 15 min AEA application was found to produce long-lasting (1 hour) in P cell synapses [24, 41]. If URB597 is increasing AEA levels by inhibiting HirFAAH, then one would expect URB597 to mimic the effect of exogenously applied AEA. EPSP amplitude was recorded prior to and then 1 hour following a 15 min application of 1 µM URB597 (N = 7) since this was a concentration that appeared to effectively inhibit HirFAAH activity (see Figs. 5c and 6c). Experiments using 1 µM AEA (N = 5) were also conducted in parallel to compare to the URB597 results. The FAAH inhibitor did significantly potentiate the P cell EPSP compared to the 0.001% DMSO control group (N = 5) and did so at a level similar to AEA (Fig. 7a, b). One-way ANOVA showed a significant effect of treatment group (F4,25 = 6.55, p < 0.0001), with the AEA- and URB597-treated ganglia exhibiting significant potentiation compared to the DMSO control groups (p < 0.05 and < 0.01, respectively). In past experiments, synaptic potentiation by AEA was blocked by the TRPV inhibitor SB366791, indicating that AEA acted on a Hirudo TRPV-like channel [24]. Here the ability of 10µM SB366791 to block synaptic potentiation by AEA was confirmed (Fig. 6B; AEA + SB group, N = 6; p < 0.05). Furthermore, the co-application of the TRPV inhibitor with URB597 also blocked the FAAH inhibitor’s capacity to produce synaptic potentiation (URB + SB, N = 7; p < 0.05).
Consistent with prior studies, AEA-induced synaptic potentiation produced no change in paired-pulse facilitation ratio (PPR) compared to the DMSO controls (Fig. 7c). This suggests that synaptic potentiation occurs at the post-synaptic level. As with AEA, URB597 also did not significantly change PPR. One-way ANOVA confirmed no significant difference in the percent change in PPR across an of the treatment groups (F4,25 = 1.93, p > 0.05). For all the treatment groups, no change was observed in the input resistance of the postsynaptic cell (Fig. 7d), indicating that the observed increases in EPSP amplitude were not due to changes in the intrinsic excitability of the postsynaptic neurons, at least as measured in the soma (F4,25 = 1.42, p < 0.05).