A novel family VIII carboxylesterase hydrolysing third- and fourth-generation cephalosporins

A metagenomic library was constructed from a soil sample of spindle tree-rhizosphere. From this library, one clone with esterase activity was selected. The sequence analysis revealed an open reading frame (EstSTR1) encoded protein of 390 amino acids. EstSTR1 is a family VIII carboxylesterase and retains the S-X-X-K motif conserved in both family VIII carboxylesterases and class C β-lactamases. The estSTR1 gene was overexpressed in E. coli and the recombinant protein was purified by purified by metal chelating affinity chromatography and size-exclusion chromatography. EstSTR1 hydrolysed p-nitrophenyl esters, exhibited the highest activity toward p-nitrophenyl butyrate. Furthermore, EstSTR1 could hydrolyse third- and fourth-generation cephalosporins (cefotaxime and cefepime) as well as first-generation cephalosporin (cephalothin). Site-directed mutagenesis studies revealed that a catalytic residue, Ser71, of EstSTR1 plays an essential role in hydrolysing both antibiotics and p-nitrophenyl esters. We demonstrate that a metagenome-derived carboxylesterase displays β-lactam-hydrolysing activities toward third- and fourth-generation cephalosporins. Electronic supplementary material The online version of this article (doi:10.1186/s40064-016-2172-y) contains supplementary material, which is available to authorized users.


Background
Carboxylesterases (EC3.1.1.1) hydrolyse emulsified esters of short-chain carboxylic acids, retaining a characteristic α/β hydrolase fold with the G-X-S-X-G motif in the active site which contains a conserved serine residue (Bornscheuer 2002;Nardini and Dijkstra 1999). Based on conserved sequence motifs and biological properties, microbial carboxylesterases have been classified into eight families (Arpigny and Jaeger 1999). Particularly, the primary sequences of family VIII carboxylesterases with a conserved S-X-X-K motif instead of the G-X-S-X-G motif of canonical carboxylesterases are similar to those of class C β-lactamases (Arpigny and Jaeger 1999).
A previous report (Wagner et al. 2002) revealed that the overall structure of EstB from Burkholderia gladioli, composed of a mixed α/β domain and a small helical domain, was similar to that of class C β-lactamases.
Functional metagenomic screening, in which shotguncloned DNA fragments are selected for survival to antibiotic exposure, have been increasingly applied to the characterisation of many antibiotic resistance reservoirs. In recent study, a new class A β-lactamase derived from a polluted river metagenome library has been identified by functional screening (Vercammen et al. 2013). These experiments have demonstrated that antibiotic resistance genes are highly diverse and widely distributed, frequently bearing little to no similarity to known sequences (Pehrsson et al. 2013). The relationship between serineβ-lactamases and family VIII carboxylesterases has been surprisingly evidenced by the metagenomic approach of screening ester-hydrolysing activities that perform unrelated functions to previously identified resistance genes. New family VIII carboxylesterases that originated from metagenomic libraries of various environmental samples (Elend et al. 2006;Rashamuse et al. 2009;Kim et al. 2010;Jeon et al. 2011;Yu et al. 2011;Mokoena et al. 2013), including EstC (Rashamuse et al. 2009), EstM-N1 (Yu et al. 2011), EstM-N2 (Yu et al. 2011), and Est22 (Mokoena et al. 2013), exhibited hydrolysing activity for nitrocefin, but did not display β-lactamase activity for β-lactam antibiotics. Recently, the family VIII carboxylesterase, EstU1, which was selected from a soil metagenomic library by screening for esterase activity showed a bona fide hydrolysing activity toward both the ester bond of p-nitrophenyl esters and the amide bond of β-lactams, apparently utilising the same active site residues for both reactions, as supported by site-directed mutagenesis, confirming the functional relationship between family VIII carboxylesterases and β-lactamases (Jeon et al. 2011).
In this study, we report a novel family VIII carboxylesterase identified from a soil metagenomic library. Est-STR1 display an esterase activity for p-nitrophenyl esters. In particular, EstSTR1 exhibits a broad range of β-lactam hydrolytic activities for first-generation cephalosporin (cephalothin), third-generation cephalosporin (cefotaxime), and fourth-generation cephalosporin (cefepime). Here, we demonstrate the functional characterisation of EstSTR1, as well as site-directed mutagenesis and its structural discussion to understand the mechanism of EstSTR1.

Strains, metagenomics library construction, and screening lipolytic activity
E. coli DH5α and BL21(DE3) were used for all cloning and expression experiments, respectively. A soil sample (35°52′N, 127°3′E) was collected from the Korea Expressway Corporation Arboretum in Jeonju City, South Korea. Soil DNA was prepared by directed DNA extraction and purification as previously described (Kim et al. 2007). The metagenomic cosmid library was constructed according to the method of Yun et al. (2005). Two-step DNA purification was applied to remove humic compounds present in the soil DNAs using pulsed-field gel electrophoresis (PFGE) (CHEF, BioRad). To remove humic compounds in DNA extracted from soil sample, 1 % low melting point agarose was prepared and crude DNA was fractionated by PFGE under a 4 V cm −1 electrical field at 14 °C for 12 h. A gel containing 100-190 kb of DNA was purified by agarase (1 U per 100 mg slice, Takara, Japan). The isolated DNA was partially digested by Sau3AI (0.05 U μl −1 of DNA, 37 °C for 1 h), and then the digested DNA was fractionated by PFGE. A gen containing approximately 40-kb lengths DNA was again purified by agarase. The 40-kb DNA was ligated into a pSuperCosI (Stratagene, La Jolla, CA) and packaged using MaxPlax Lambda Packaging Extracts (Epicentre, Madison, WI). For screening esterase activity, the transformants were plated on Luria-Bertani (LB) agar plates with chloramphenicol (12.5 μg ml −1 ) and tributyrin (1 %). After incubation at 37 °C for 1 day, the plates were subsequently incubated at 4 °C for a week. Candidate colonies were selected based on the presence of clear zone on the plate.

Subcloning and sequence analysis
Positive colony (pCosSTR1) with a clear halo was selected on the plate. To identify the gene encoding esterase active, the subclone library was constructed by method as described by Jeon et al. (2011). The transformants with esterase activity were selected by the presence of the clear halo zone on LB agar plates containing 100 μg ml −1 ampicillin and 1 % tributyrin. DNA sequencing of the subclone with esterase activity was performed with an ABI3100 (PE Applied Biosystems, Foster City, CA, USA) and assembled using the Vector NTI suite 7 software package (InforMax, North Bethesda, MD, USA). The open reading frames (ORFs) and sequence homology searches in the complete assembled sequence were analysed by the ORF finder and BlastX search provided by the National Center for Biotechnology Information (NCBI) (Altschul et al. 1997).

Multiple alignment and phylogenetic analysis
Multiple sequence alignments were carried out by the ClustalW program (Thompson et al. 1994) for the protein sequences. To compare the amino acid sequences of EstSTR1 and EstU1, we performed sequence alignment using the Expresso program on the T-Coffee Server (Notredame et al. 2000), found at the ExPASy web site (http://tcoffee.vital-it.ch/apps/tcoffee/index.html). Expresso is able to combine sequence information with protein structural information. Molecular Evolutionary Genetics Analysis 4.1 software (MEGA, version 4.1) (Tamura et al. 2007) was used to make the phylogenetic tree using a neighbour-joining method (Saitou and Nei 1987).

β-Lactamase assay of EstSTR1
Antibiotics (cephalothin, cefoxitin, cefotaxime, and cefepime) were obtained from Sigma-Aldrich (St. Louis, MO, USA). The chemical structures of cephalosporins (cephalothin, cefoxitin, cefotaxime, and cefepime) are shown in the Additional file 1: Figure S1. The hydrolysing activity of EstSTR1 for cephalosporins was measured by the paper disc method as previously described (Jeon et al. 2011). The enzyme (330 μM) was incubated with antibiotic substrates (3 mM cephalothin, 1 mM cefoxitin, 1 mM cefotaxime, and 1 mM cefepime) in 50 mM Tris-HCl (pH 8.0) for 2 h at 35 °C, and then reaction mixtures were loaded onto small paper discs. After 8 h incubation at 37 °C, the diameters of the inhibition zones around the small paper discs were recorded. For comparison of β-lactam hydrolysing activity, a negative control containing antibiotics without enzyme and a positive control containing antibiotics and the CMY-10, a plasmid-encoded class C extended-spectrum β-lactamase (ESBL) from Enterobacter aerogenes (Kim et al. 2006), were used.

Site-directed mutagenesis of EstSTR1
A site-directed change to alanine (S71A) was made using the Stratagene Quik Change kit (La Jolla, CA, USA). The primers (S71A-F and S71A-R) designed to introduce the S71A substitution were as follows: S71A-F (5′-CGCT CATCAATACCTATGCGACCACCAAGGGCATGG-3′) and S71A-R (5′-CCATGCCCTTGGTGGTCGCATAGG TATTGATGAGCG-3′). The sequence corresponding to the mutated codons are underlined. The catalytic activity of the variant was tested and compared with that of the wild-type enzyme.

Nucleotide sequence accession number
The nucleotide sequence of EstSTR1 has been deposited in the GenBank database under the accession number KJ530984.

Metagenomic library screening and sequence analysis
A cosmid metagenomic library consisting of 7968 clones was constructed using high molecular weight DNA extracted from a soil sample of spindle treerhizosphere, taken near the Korea Expressway Corporation Arboretum in South Korea. The insert sizes for the cosmid clones ranged from 35 to 40 kb, with nonredundant patterns. The entire library was screened for lipolytic activity on 1 % tributyrin plates. One cosmid clone (36 kb insert size) forming a clear halo zone on the plate was selected for further analysis. To characterise the gene exhibiting esterase activity, the subcloning experiment with a pUC118/HincII/BAP plasmid was performed and subclone (pUCSTR1) with a short insert of 1825 bp was found. The sequence analysis of the pUCSTR1 insert DNA showed the presence of one open reading frame (ORF) of 1173 bp (estSTR1), encoding a polypeptide of 390 amino acids. Primary sequence analysis of EstSTR1 indicated that it was similar to a putative esterase (WP_007223494) from Marine gamma proteobacterium HTCC2143 (47 % identity), class C β-lactamase (WP_004621811) from Caulobacter vibrioides (47 % identity), and β-lactamase (WP_005321278) from Streptomyces pristinaespiralis (46 % identity). Multiple sequence alignment of EstSTR1 and its homologs showed that the S-X-X-K motif is well conserved in family VIII carboxylesterases (Jeon et al. 2011;Kim et al. 2010;Rashamuse et al. 2009;Elend et al. 2006), class C β-lactamases (Knox et al. 1996), and penicillin-binding proteins (PBPs) (Joris et al. 1988) (Fig. 1). Based on the esterase/lipase classification system (families I-VIII) proposed by Arpigny and Jaeger (1999), the phylogenetic relationship was analysed. The phylogenetic analysis showed that EstSTR1 was grouped with family VIII carboxylesterases (Fig. 2).

Purification and characterisation of EstSTR1
To invest the biochemical properties of EstSTR1, the est-STR1 gene ligated into pET-24a(+) was overexpressed in E. coli. SDS-PAGE analysis of purified EstSTR1 showed a single protein band which correlated well with the theoretical mass (42 kDa) of EstSTR1 (Additional file 1: Figure S2). The hydrolytic activity of purified recombinant EstSTR1 for p-nitrophenyl esters (C2-C18) was investigated. EstSTR1 showed substrate preference for a wide range of substrates (C2 to C10), and p-nitrophenyl butyrate (C4) was most rapidly hydrolysed (Fig. 3). However, no enzyme activity was detected for p-nitrophenyl esters (C12-C18) with a longer chain (Fig. 3). The optimum activity of EstSTR1 was measured at a temperature range of 5-70 °C and pH range of 6.0 and 10.0. The enzyme showed a temperature optimum at 40 °C (Additional file 1: Figure S3A). The optimum activity occurred at an alkaline pH in the range of pH 9-10, threefold higher than the activity at pH 8.0, indicating that Est-STR1 is an alkaline esterase (Additional file 1: Figure  S3B).

Determination of β-lactamase activity
Previously, we reported that a metagenome derived esterase, a family VIII carboxylesterase (EstU1) could hydrolyse first-generation cephalosporins (cephaloridine, cephalothin, and cefazolin) (Jeon et al. 2011). The crystal structure of cephalothin complex of EstU1 revealed an acyl-enzyme intermediate in which Ser100 (the nucleophile) in first motif of EstU1 is covalently linked to the carbonyl carbon of the hydrolysed β-lactam ring of cephalothin (Cha et al. 2013). Though the amino acid sequence of EstSTR1 showed low similarity to EstU1 from uncultured bacterium (15 % identity), three active site residues (S100, K103, and Y218) essential for the β-lactam hydrolytic activity in EstU1 could be found in EstSTR1 (Fig. 1). Fig. 1 Conserved sequence blocks from a multiple sequence alignment of EstSTR1 and related family VIII carboxylesterases, class C β-lactamases, and penicillin-binding proteins (PBPs). Family VIII carboxylesterases are represented by EstU1 (uncultured bacterium, JF791800), EstC (uncultured bacterium, ACH88047), and EstB (Burkholderia gladioli, AAF59826). Class C β-lactamases are represented by Lac-1 (Escherichia coli, AAA23441) and Lac-2 (Enterobacter cloacae, P05364) and penicillin-binding proteins are represented by PBP-1 (Streptomyces sp. R61, P15555) and PBP-2 (Bacillus cereus, CAA09676). Asterisks indicate conserved active site residues. Identical residues are shown as white letters on a dark background For these reasons, we examined whether EstSTR1 could hydrolyse various β-lactam antibiotics, cephalothin (firstgeneration cephalosporins), cefoxitin (second-generation cephalosporin), cefotaxime (third-generation cephalosporins), and cefepime (fourth-generation cephalosporin). Compared with that of the negative control, the diameters of the inhibition zones around the discs containing cephalothin, cefotaxime, and cefepime incubated with EstSTR1 were decreased from 27 to 22 mm in cefotaxime and from 26 to 23 mm in cefepime (Fig. 4). The result implies that the antibiotic efficacies of cephalothin, cefotaxime, and cefepime are affected by EstSTR1 activity. However, EstSTR1 did not appear to change the antibiotic efficacy of cefoxitin because the size alteration of the clear zone around paper disc did not show compared with that of the negative control. Thus, EstSTR1 displays β-lactam hydrolytic activity for cephalothin, cefotaxime, and cefepime except for cefoxitin.

Determination of kinetic parameters
The kinetic parameters of EstSTR1 for cephalosporins (cephalothin, cefotaxime, and cefepime) and p-nitrophenyl butyrate were investigated. The K m and k cat values for p-nitrophenyl butyrate were approximately 10-and 20-fold higher than those of EstU1 (6.03 μM and 15.72 s −1 ) (Jeon et al. 2011), respectively (Table 1). Kinetic analysis of EstSTR1 revealed that this enzyme has a hydrolysing activity for cephalothin, cefotaxime, and cefepime (Table 1). The catalytic efficiencies of EstSRT1 for cephalothin, cefotaxime, and cefepime were similar to that of EstU1 for cefazolin (3.035 M −1 s −1 ) (Jeon et al. 2011) (Table 1). Although catalytic efficiencies of EstSTR1 for β-lactams were very low level, catalytic efficiency of β-lactamase I (4 M −1 s −1 ) from Bacillus cereus for a β-lactam (Martin Villacorta et al. 1991) is similar to the level of the EstSTR1.

Site-directed mutagenesis
Because EstSTR1 retains the catalytic (or nucleophilic) serine (Ser71) correlated with the Ser100 of EstU1, we wondered whether a single nucleophile (Ser71) was involved or whether the enzyme harbored two active sites. We performed site-directed mutagenesis to confirm the effect of substitution of a single serine on both activities of EstSTR1. The nucleophilic serine (Ser71) in EstSTR1 was mutated into an alanine residue (Ala71) and the mutant protein (S71A) was purified and characterised, compared with the wild-type EstSTR1. The mutant protein (S71A) was inactive for cephalosporins and p-nitrophenyl butyrate (Table 1). This result suggests that the serine residue is essential for both activities.

Discussion
In this study, a novel family VIII carboxylesterase, designated as EstSTR1, was identified by active screening of a metagenomic library constructed from a soil sample of tree-rhizosphere collected near in the Korea Expressway Corporation Arboretum in South Korea. EstU1 and EstSTR1 were found within two soil functional metagenomes of really a few clones (<10,000), which means that the retrieved genes are likely to be highly prevalent in soil. The amino acid sequence of EstSTR1 showed that the S-X-X-K motif encompassing the nucleophilic serine residue of serine-β-lactamases was conserved and EstSTR1 was grouped together with the family VIII carboxylesterases. The β-lactam hydrolytic activities of Est-STR1 for first-generation cephalosporin (cephalothin), third-generation cephalosporin (cefotaxime), and fourthgeneration cephalosporin (cefepime) were clearly demonstrated by the disc diffusion assay and kinetics as well.
A previous report (Cha et al. 2013) revealed that the structure of EstU1 had a β-lactamase-like modular architecture and that the active site residues (Ser100, Lys103, and Tyr218) of EstU1 played an important role in hydrolysing both ester and amide bonds (Cha et al. 2013;Jeon et al. 2011). The corresponding active site residues, Ser71, Lys74, and Tyr160, could be predicted in EstSTR1 and site-directed mutagenesis of the EstSTR1 serine residue demonstrated that this residue was crucial for both the esterase and β-lactamase activities. In addition, the active

Fig. 4
Disc diffusion assay for confirming the hydrolysis of antibiotics by EstSTR1. This assay was performed by incubating purified EstSTR1 with antibiotics, including 3 mM cephalothin, 1 mM cefoxitin, 1 mM cefotaxime, and 1 mM cefepime in 50 mM Tris-HCl (pH 8.0) for 2 h at 35 °C. The reaction mixtures were adsorbed onto a paper disk and placed on agar seeded with E. coli BL21(DE3). The diameters of the inhibition zones around the discs appeared after an 8 h at 37 °C. The negative controls were samples containing antibiotics only and the positive controls were reaction mixtures containing CMY-10 and antibiotics. Antibiotics: LOT cephalothin, FOX cefoxitin, TAX cefotaxime, FEP cefepime site in EstU1 was divided into R1 and R2 subsites. The R1 subsite is defined by the R1 segment, the β8-β9 turn, and the Ω-loop. The R2 subsite is surrounded by the R2' segment, the R2 segment, and β8, with the nucleophilic serine. The R1 subsite accommodates the R1 side-chain of cephalosporins and the R2 subsite represents the opposite region interacting with the R2 side-chain of cephalosporins (Kim et al. 2006). EstU1 only hydrolysed first-generation cephalosporins (e.g., cephalothin, cephaloridine, and cefazolin); however, EstSTR1 hydrolysed third-and fourth-generation cephalosporins as well as first-generation cephalosporin. Cefotaxime and cefepime have the R1 side chain larger than that of cephalothin (Additional file 1: Figure S1). This result implies that the R1 subsite of EstSTR1 may be wider than that of EstU1. We confirmed variations in the Ω-loop and R1 segment in the R1 subsite via sequence alignment of EstSTR1 and EstU1 (Additional file 1: Figure S4). Therefore, these results suggest that variation of the R1 subsite in EstSTR1 allows the active site of EstSTR1 to access substrates (cefotaxime and cefepime), resulting in the efficient acylation.
The kinetic study of EstSTR1 also indicated that Est-STR1 showed weak β-lactamase activity for the antibiotic substrates, cephalothin, cefotaxime, and cefepime. A hydrolase mechanism for antibiotics has recently been demonstrated from the crystal structure of cephalothin complex of EstU1. A two-strep hydrolysis process composed of acylation and deacylation steps, mediated by the nucleophilic serine has been proposed in family VIII carboxylesterases and serine-β-lactamases. The weak hydrolytic activity of EstU1 for β-lactam antibiotics likely indicates that the active site of EstU1 cannot show the efficient deacylation reaction. The EstSTR1 deacylation efficiency on β-lactam antibiotics may therefore be lower than that of typical extended-spectrum β-lactamase, lowering the overall hydrolysis efficiency.

Conclusion
One gene screened from a soil metagenomic library was identified as having an S-X-X-S motif found in family VIII carboxylesterases and serine-β-lactamases. The purified EstSTR1 protein showed hydrolysing activity for p-nitrophenyl esters and antibiotic substrates such as cephalothin, cefotaxime, and cefepime. Similar to EstU1, EstSTR1 contains only the active site residues (Ser100, Lys103, and Tyr218) responsible for the hydrolysing activity toward both β-lactam antibiotics and p-nitrophenyl esters. EstSTR1 is the first esterase that hydrolyses third-generation cephalosporin (cefotaxime) and fourthgeneration cephalosporin (cefepime). To elucidate the detailed catalytic mechanism of EstSTR1, further studies will be required, including the crystallographic determination of EstSTR1's structure.