SacPox from the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius is a proficient lactonase

Background SacPox, an enzyme from the extremophilic crenarchaeal Sulfolobus acidocaldarius (Sac), was isolated by virtue of its phosphotriesterase (or paraoxonase; Pox) activity, i.e. its ability to hydrolyze the neurotoxic organophosphorus insecticides. Later on, SacPox was shown to belong to the Phosphotriesterase-Like Lactonase family that comprises natural lactonases, possibly involved in quorum sensing, and endowed with promiscuous, phosphotriesterase activity. Results Here, we present a comprehensive and broad enzymatic characterization of the natural lactonase and promiscuous organophosphorus hydrolase activities of SacPox, as well as a structural analysis using a model. Conclusion Kinetic experiments show that SacPox is a proficient lactonase, including at room temperature. Moreover, we discuss the observed differences in substrate specificity between SacPox and its closest homologues SsoPox and SisLac together with the possible structural causes for these observations.

Both enzyme families exhibit the same (β/α) 8 -barrel topology [9,10] and belong to the amidohydrolase superfamily [11,12]. Their structure consists of 8 β-strands forming a central barrel surrounded by 8 α-helixes. The active site is constituted by a bimetallic center (two metal cations) localized at the C-terminus of the barrel. Metal cations are coordinated by four histidines, an aspartic acid and a carboxylated lysine residue [9]. While the nature of the bimetallic center can vary depending on the enzyme nature and the purification procedure [3,5,13,14], the catalytic mechanism is presumed to be identical. The bimetallic center activates a water molecule into a hydroxide ion which performs a nucleophilic attack onto the electrophilic center [9,15].
The difference in substrate specificities of PLLs and PTEs seems mainly governed by variation in the connecting loops of the barrel [2,16]. Major differences between PTEs and PLLs reside in the active site loop size and conformation [1,2]. Indeed, loop 7 is shorter in PLLs than in PTEs whereas the loop 8 is larger, forming a hydrophobic channel that accommodates lactones aliphatic chain [9]. Loop 7/8 length and sequence also differ within the PLL family and led to the identification of two different subfamilies: PLLs-A and PLLs-B [2]. Both subfamilies exhibit different substrate specificities: PLLs-B are exclusively oxo-lactonases ( Figure 1DE) whereas PLLs-A hydrolyze efficiently oxo-lactones and Acyl-Homoserine Lactones (AHLs, Figure 1C) [2]. AHLs are messenger molecules involved in a bacterial communication system dubbed quorum sensing (QS) [17]. QS regulates the expression of numerous genes, and enables bacterial population to adopt a "group" behavior, including the expression of virulence factors of some pathogens [18,19]. The involvement of PLLs-A in quorum sensing has not yet been demonstrated, and these enzymes are often found with no other AHL components, including in archaeal species [20]. However, the fact that they hydrolyze specifically the natural enantiomer of AHL indicates that it may be their native substrate [16].
PLLs are promiscuous enzymes that catalyze two chemical reactions of potential biotechnological interest. Indeed, the inhibition or "quenching" of the QS is seen as a possibly promising strategy to develop innovative therapies [21][22][23][24][25]. Indeed, lactonases such as PLLs can inhibit QS (known as quorum quenching, i.e. QQ) [26,27] and thereby annihilate the virulence of micro-organisms possessing an AHL-based QS system [28]. Moreover, PLLs are endowed with relatively low phosphotriesterase activity, but might be optimized against OPs and subsequently used for degrading organophosphorus pesticides [3,5,6,9,29] and nerve agents [30], for which no satisfactory remediation methods are currently available [31].
Here we focus on SacPox, the PLL from the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius (living conditions: 55-85°C, pH 2-3) [44]. SacPox was originally isolated and studied for its ability to hydrolyze OP compounds at high temperature [4]. The enzyme shares about 30% of sequence identity with BdPTE and about 70% with its closest homologues, i.e. SsoPox from Sulfolobus solfataricus [3] and SisLac from Sulfolobus islandicus [33,45]. Being an enzyme from a hyperthermophile, SacPox is however less stable than SsoPox (half-life of 5 min at 90°C [4] and of 4 h at 95°C [3,46], respectively). The kinetic characterizations performed on SacPox revealed that it hydrolyzes OP, ester and lactone molecules at high temperature [4,13]. However, only few substrates have been tested, and no natural lactones were assayed as substrate. In this study, we performed a broad kinetic characterization of SacPox at room temperature (25°C) for several OPs, esters ( Figure 1B) and lactone molecules including AHLs, γ-lactones and δ-lactones in the aim to evaluate the biotechnological potentialities of this enzyme.

Sequence alignment
The sequence alignment was performed based on the previously published PLL sequence alignment [2], using the T-coffee server (expresso) [47,48] and manually improved with the seaview software [49]. It contains 29 different sequences (Additional file 1: Table S1). The sequence alignment was represented using the BioEdit 7.1.3 software [50]. Protein sequence identities were computed using ClustalW server [51]. The phylogenetic tree was performed using PhyML [49] and default parameters.

Protein production and purification
The protein production and subsequent purification steps were performed analogously to previously described [16,33,34,45,[52][53][54]. In brief, the protein was heterologously produced in Escherichia coli strain BL21 (DE 3 )-pGro7/GroEL (TaKaRa) at 37°C in ZYP medium [55]. When OD 600nm reaches 0.8, protein production was induced with addition of arabinose (0.2%, w/v) and CoCl 2 (2 mM) and temperature transition to 25°C for 20 hours. Cells were harvested by centrifugation, and pelleted cells were suspended in lysis buffer (50 mM HEPES pH 8, 150 mM NaCl, 0.2 mM CoCl 2 , lysozyme 25 mg/ml, PMSF 0.1 mM, DNase I 10 mg/ml), stored at −80°C during 2 hours; then sonicated 3 times during 30 seconds (Branson Sonifier 450, 80% intensity and microtype limit of 8) and centrifuged. Taking advantage of the high stability of SacPox, the supernatant was heated at 70°C during 30 minutes and centrifuged before proceeding a STREP-TRAP affinity chromatography step (GE Healthcare, Uppsala, Sweden). The sample was then cleaved by the Tobacco Etch Virus protease (TEV, ratio 1:20, w/w [56]) during 20 hours at 30°C prior to be loaded a second time on STREP-TRAP affinity chromatography. The flow through containing the cleaved protein was then concentrated and loaded on a size exclusion column (S75-16-60; GE Healthcare, Uppsala, Sweden). The protein purity and identity were checked by SDS-PAGE and mass spectrometry analysis (MS platform Timone, Marseille, France). The protein concentration was determined using a nanospectrophotometer (Nanodrop, Thermofisher Scientific, France) using its molar extinction coefficient (SacPox ε 280 nm = 35 307.7 M −1 cm −1 ) calculated by the PROT-PARAM server [57].

Kinetic characterization General procedures
Catalytic parameters were evaluated at 25°C and recorded with a microplate reader (Synergy HT, BioTek, USA) and the Gen5.1 software as previously explained [16,33,52,54]. The reaction was performed in a 200 μL volume using a 96-well plate with a 6.2 mm path length as previously described [33]. The collected data were subsequently fitted to the Michaelis-Menten (MM) equation [58] using Graph-Pad Prism 5.00 (GraphPad Software, San Diego California USA, www.graphpad.com). In cases where V max could not be reached, the catalytic efficiency was obtained by fitting the linear part of MM plot to a linear regression using Graph-Pad Prism 5.00 software.

Structural modeling and structural analysis
The SacPox structure was modelled using the ESyPred3D server using SacPox protein sequence as query and SsoPox structure (2VC5) as template [59]. Structures were analyzed and figure made using PyMol [60].

Results
First classified within the bacterial PTEs, SacPox shares in fact only 33.8% sequence identity with BdPTE (Additional file 1: Table S2). SacPox indeed belongs to the PLLs-A (Figure 2A) [2]: it shares 76.1% of sequence identity with its closest homologues SsoPox and SisLac, and only 30.6% identity with the PLL-B DrOPH. Together with SisLac and SsoPox, SacPox comprises the creanarcheal clade of the PLLs-A (Figure 2A). The sequence alignment highlights the strict conservation of essential active site residues between the different clades ( Figure 2B).

Esterase activity
The ability of SacPox to hydrolyze phenyl-acetate, pNPacetate and pNP-decanoate (Additional file 1: Figure  S1VII-IX) has been evaluated ( Table 2). While no activity could be detected against pNP-decanoate, SacPox exhibits low catalytic efficiencies against both phenyl-acetate and pNP-acetate (k cat /K M ≈ 50 M −1 .s −1 ). This weak activity against classical esters differs from previous studies on the close homologues SsoPox and SisLac, for which activity has only been recorded on pNP-acetate [33].

Lactonase activity
The catalytic parameters of SacPox for various lactone substrates have been measured, including against oxolactones (lipophilic aroma), AHLs and dihydrocoumarin (Table 3). Our results indicate a preference of SacPox for oxo-lactone substrates; i.e. γ-heptanolide and nonanoic-γlactone (k cat /K M ≈ 2.5×10 4 M −1 .s −1 ), while AHLs are about 10 times worse substrates (i.e.; C8 AHLs, k cat /K M ≈ 5×10 3 M −1 .s −1 ). Furthermore, it seems that SacPox prefers AHLs vs 3-oxo-AHLs since the K M for C8 aliphatic chains is 5fold lower than that for 3-oxo-C8 AHLs. Overall, long aliphatic chain substrates AHLs are better substrates for the enzyme. Indeed, short aliphatic chain AHLs are not hydrolyzed by SacPox. Interestingly, this preference is not retained for oxo-lactones, for which molecules with short (See figure on previous page.) Figure 2 Phylogenetic analysis of the PLL family. A. Phylogenetic tree of PLLs, PTEs, and close homologues. Members of PLL-B are colored in green while within the PLL-As, mesophilic and archaeal PLLs are respectively colored in red and orange. The clades of PHPs, PTEs and RTXs were collapsed for clarity. All the sequences used for this tree are listed in Additional file 1: Table S1. B. Sequence alignment of BdPTE from B. diminuta, SsoPox from S. solfataricus, SacPox from S. acidocaldarius and DrOPH from D. radiodurans. Conserved amino acid residues are highlighted in black and similar residues in grey. Conserved active site residues involved in metals coordination are highlighted by red stars. Secondary structures are represented according to SsoPox structure (with pink arrows depicting β-sheets and red cylinders depicting α-helixes). or without aliphatic chain are efficiently hydrolyzed (k cat / K M ≈ 10 4 M −1 .s −1 ). As previously observed for SsoPox and SisLac [16,33], this feature may reveal a potential alternative binding mode of these compounds in SacPox active site. Finally, contrary to SsoPox and SisLac [16,33], SacPox does not hydrolyze dihydrocoumarin.

Structural analysis
Numerous attempts to crystallize SacPox were made, with no success (Elias, Hiblot, Gotthard & Chabriere, unpublished). A previous structural model was generated by homology modeling based on BdPTE structure [4] (~33.8% sequence identity with SacPox), but yielded little insights given the moderate sequence identity with the template and the very significant differences in the active site loops between these two representatives of distinct enzyme families [1,9,16]. Here we generated a homology-based model using the structure of SsoPox as template (76.1% of sequence identity; Additional file 1: Table S2). As expected, the SacPox model structure almost perfectly superimposes to the SsoPox crystal structure ( Figure 3A). Residues forming the active site are all conserved and residues involved in loops 7 and 8 occupy nearly identical conformation in SacPox and SsoPox but also in SisLac structures ( Figure 3B). Noteworthy, loop 8 is partially structured into an α-helix, as seen in X-ray structures of SsoPox and SisLac. A substitution (I266 in SacPox; T265 in SsoPox and SisLac) in loop 8 may slightly alter the shape of the aliphatic channel. But overall, the active site of SacPox and SsoPox are nearly identical ( Figure 2B). Furthermore, four other substitutions between SacPox and its close homologues can be seen in loop 8: SacPox exhibits a K at position 268, instead of an R residue (R267 in SisLac), Y271 instead of L (L270 in SisLac), K278 instead of R (R277 in both SisLac and SsoPox), and M281 instead of I (I280 in SsoPox) (Additional file 1: Figure S2). While the structural model suggests that these substitutions are not affecting directly the binding cleft of SacPox, they might modulate loop 8 conformation and its dynamics. Indeed, it was shown in the close homologue SsoPox that a single substitution in loop 8 (W263 in SsoPox, equivalent to W264 in SacPox) increases the conformational flexibility of loop 8, thereby conferring higher promiscuity to the enzyme [16]. The effect is in fact so dramatic that the substitution in SsoPox of W263 by any of the 19 other natural amino acids yields a variant with improved phosphotriesterase activity [16]. Additionally, loop 8 being involved in the accommodation of the aliphatic substituent of lactones substrates [9], mutations in this loop can also affect the lactonase activity [16].

Discussion
Here we show that SacPox is a proficient lactonase (~10 4 M −1 .s −1 ) and can hydrolyze both oxo-lactones and AHLs. Nevertheless, SacPox have a slightly different substrate specificity than its close homologues [16,33]. Indeed, SacPox exhibits slightly lower catalytic efficiencies, prefers AHLs over 3-oxo-AHLs and does not show any activity against dihydrocoumarin. Interestingly, as noted for SisLac and SsoPox [16,33], SacPox clearly prefers long chain AHLs, but can efficiently hydrolyze short chain or oxo-lactones without aliphatic substituents. This feature could reflect a putatively different binding mode of AHLs and oxo-lactones into PLLs active sites. We note that the biological role of lactonases such as PLLs is yet unclear, especially in extremophilic archaea where no AHL-based quorum sensing systems have been identified so far.
SacPox also exhibits promiscuous esterase and phosphotriesterase activities, a common feature of PLLs. Similarly to SsoPox and SisLac [33,52], SacPox prefers OPs with small substituents. Moreover, SacPox also shows a clear preference for oxono-phosphotriesters, rather than thionophosphotriesters, a feature previously dubbed thiono-effect [52]. Interestingly, SsoPox, SisLac and SacPox exhibit similar catalytic efficiencies against OPs (10 2-3 M −1 .s −1 ) at 25°C, efficiencies that are close to those measured at much higher temperatures [4].  The structural model shows that SacPox structure is very close to that of SsoPox (Figure 2A). Most critically, the active sites of both enzymes are essentially identical ( Figure 2B), with the exception of position 266 (I in SacPox, T in SsoPox and SisLac). This substitution might partly account for the observed differences in substrates specificity between these enzymes, and would thereby represent an interesting target for future mutagenesis studies. But four other substitutions in loop 8 between these close homologues might be involved as well, and comprise also interesting options for mutagenesis studies (K268R, Y27IL, K278R and M281I). A recent study on SsoPox highlighted how profound the effect on catalysis of a single substitution on loop 8 (W263) can be [16]. Therefore, substitution T266I, and/or the four others on loop 8, might contribute to the observed differences between SacPox and SsoPox in substrate specificity, in combination with other factors that cannot be assessed by a structural model such as subtle changes in active site loops conformation and dynamics [16,33]. Indeed, the observed differences in the detergent stimulation between both enzymes (SacPox is only weakly stimulated by SDS, as compared to SsoPox) could well be a manifestation of different dynamics of their respective active site loops.

Conclusions
To conclude, we here demonstrate that albeit being initially isolated, characterized, and named after its ability to degrade the insecticide paraoxon (pox; [4]), SacPox is putatively a native lactonase, capable of hydrolyzing these compounds with significant catalytic efficiencies at 25°C (up to 10 4 M −1 .s −1 ). The extensive kinetic characterization reveals some substrate specificity differences between Sac-Pox and its close homologues SisLac and SsoPox, and the proposed structural model of SacPox suggests putative candidates (e.g. I266) that could account for these observations. Such positions might constitute interesting targets for future engineering studies, with the aim of improving or altering the catalytic properties of SacPox.