BIOTECHNOLOGICALLY RELEVANT ENZYMES AND PROTEINS

Metagenome-derived haloalkane dehalogenases with novelcatalytic properties

Michael Kotik1& Pavel Vanacek2

& Antonin Kunka2 & Zbynek Prokop2& Jiri Damborsky2

Received: 15 April 2017 /Revised: 11 June 2017 /Accepted: 13 June 2017# Springer-Verlag GmbH Germany 2017

Abstract Haloalkane dehalogenases (HLDs) are environ-mentally relevant enzymes cleaving a carbon-halogen bondin a wide range of halogenated pollutants. PCR with degener-ate primers and genome-walking was used for the retrieval offour HLD-encoding genes from groundwater-derived envi-ronmental DNA. Using specific primers and the environmen-tal DNA as a template, we succeeded in generating additionalamplicons, resulting altogether in three clusters of sequenceswith each cluster comprising 8–13 closely related putativeHLD-encoding genes. A phylogenetic analysis of the translat-ed genes revealed that three HLDs are members of the HLD-Isubfamily, whereas one gene encodes an enzyme from thesubfamily HLD-II. Two metagenome-derived HLDs, eHLD-B and eHLD-C, each from a different subfamily, were heter-ologously produced in active form, purified and characterizedin terms of their thermostability, pH and temperature opti-mum, quaternary structure, substrate specificity towards 30halogenated compounds, and enantioselectivity. eHLD-Band eHLD-C showed striking differences in their activities,substrate preferences, and tolerance to temperature.Profound differences were also determined in the

enantiopreference and enantioselectivity of these enzymes to-wards selected substrates. Comparing our data with those ofknown HLDs revealed that eHLD-C exhibits a unique com-bination of high thermostability, high activity, and an unusu-ally broad pH optimum, which covers the entire range ofpH 5.5–8.9. Moreover, a so far unreported high thermostabil-ity for HLDs was determined for this enzyme at pH valueslower than 6.0. Thus, eHLD-C represents an attractive andnovel biocatalyst for biotechnological applications.

Keywords Haloalkane dehalogenase .Metagenomic DNA .

Heterologous production . Substrate specificity . Proteinstability

Introduction

Thousands of organohalogenated compounds exist in the bio-sphere. Many are man-made, but many are of natural originproduced through biotic or abiotic processes. Numerous stud-ies have documented the discovery of all kinds of simple andcomplex halogenated compounds in various habitats and or-ganisms. It appears that halogenated compounds have existedin nature long before their massive production by the chemicalindustry (Gribble 2000; Gribble 2003). For example, the nat-urally produced organohalogens chloromethane, chloroanisylalcohol, and pyrrolnitrin have been shown to be involved inveratryl alcohol biosynthesis (Harper et al. 1996), the degra-dation of lignin (de Jong and Field 1997), and the bacterialcontrol of plant pathogens (Hammer et al. 1997), respectively.Many more halogenated metabolites have been identified;however, their functions are often unknown (Anke andWeber 2006; Gribble 2015). There is evidence that in soilorganohalogens are formed through nonenzymatic processessuch as oxidation of organic matter by Fe(III) (Keppler et al.

Electronic supplementary material The online version of this article(doi:10.1007/s00253-017-8393-3) contains supplementary material,which is available to authorized users.

* Jiri Damborskyjiri@chemi.muni.cz

1 Laboratory of Biotransformation, Institute of Microbiology, CzechAcademy of Sciences, Videnska 1083, 14220 Prague, Czech Republic

2 Loschmidt Laboratories, Department of Experimental Biology andCentre for Toxic Compounds in the Environment RECETOX,Faculty of Science, Masaryk University, 625 00 Brno,Czech Republic, Brno, Czech Republic

Appl Microbiol BiotechnolDOI 10.1007/s00253-017-8393-3

2002) and enzymatic reactions as part of the metabolism ofmicrobial communities (Bastviken et al. 2009; Albers et al.2011). Given the considerable extent of natural halogenationand dehalogenation processes in soil and other environments(Ballschmiter 2003; Laturnus et al. 2005), it is not surprisingthat nature offers many enzyme systems which act on haloge-nated compounds (Smith et al. 2013; van Pée and Unversucht2003). One segment within the diverse collection ofdehalogenases (Weigold et al. 2016) is represented byhaloalkane dehalogenases (HLDs; EC 3.8.1.5), which cleavecarbon–halogen bonds in a wide range of organohalogenatedcompounds using water as a co-substrate. An important bio-logical role of HLDs is their involvement in the microbialcatabolism of halogenated compounds. As a result of hydro-lytic cleavage of the carbon–halogen bond, the toxicity andrecalcitrance of the generated compounds are usually reduced(Janssen et al. 2001). Many HLDs exhibit broad substratespecificities, accepting both primary and secondary carbon–halogen bonds (Daniel et al. 2015; Koudelakova et al. 2011);however, they cannot cleave carbon–halogen bonds at multi-ply halogenated or sp2-hybridized carbons (Swanson 1999).

HLDs hold the potential for various practical applications(Koudelakova et al. 2013) such as detoxification of warfareagents and bioremediation of environmental pollutants(Beschkov et al. 2008; Bosma et al. 2002; Erable et al.2005; Janssen 2007; Marchand et al. 2009; Prokop et al.2006; Stucki and Thüer 1995), and as tagging componentsin analytical and cell imaging methods (Bidmanova et al.2010; Los et al. 2008). Further, enantiomerically pure prod-ucts of HLD-catalyzed reactions are high-value buildingblocks in organic synthesis. Thus, enantioselective HLDs arehighly interesting catalysts for biocatalytic transformations(Pieters et al. 2001; Prokop et al. 2010; Westerbeek et al.2011, 2012). Generally, efficient biocatalytic processes re-quire enzymes of sufficient stability and high activity, prefer-ably with a broad pH optimum (Schoemaker et al. 2003). Onthe other hand, special applications call for enzymes adaptedfor low temperatures (Cavicchioli et al. 2002). As it is difficultto obtain an optimally performing enzyme for a target appli-cation, libraries of enzymes from various sources are created,which are then screened against a specific reaction. In thisrespect, metagenomic or environmental DNA is attractive be-cause it offers access to enzymes from presently uncultivableor as yet uncultivated microorganisms.

HLDs are members of the α/β-hydrolase fold superfamily.The active-site residues of the catalytic pentad are composedof a nucleophilic Asp residue, a basic His residue, Asp or Gluas the general acid, and a Trp–Trp or Trp–Asp pair of residueswhich stabilize the leaving halide. Based on phylogeneticanalyses, HLDs are divided into three subfamilies: HLD-I,HLD-II, and HLD-III (Chovancova et al. 2007). Currently,there are over 20 HLDs with experimentally verified and pub-lished dehalogenation activities. To meet the increasing

demand for new HLDs with interesting biotechnologicalproperties, microbial isolates have been screened for HLDactivities. Alternatively, HLD-encoding sequences have beenisolated from metagenomic DNA-based libraries using suc-cessive rounds of enrichment with gene-specific probes(Gray et al. 2003). Moreover, sequenced genomes can alsoserve as a source of novel HLDs using in silico screens(Chan et al. 2010).

In the present work, we retrieved three clusters ofmetagenome-derived putative HLD-encoding genes from amicrobial groundwater consortium, thereby predominantly fo-cusing on uncultivated bacteria as an HLD source. Wesucceeded in expressing two fully functional metagenomicHLDs, termed eHLD-B and eHLD-C, from two different clus-ters, which enabled us to perform a thorough biochemicalcharacterization. Interestingly, the two enzymes had pro-foundly different thermal stabilities with the highest activitiesat 25 and 55 °C. Moreover, eHLD-C with its high temperatureoptimum exhibited a distinct pH–activity profile and a so farunreported high thermostability at lower pH values, making ita promising candidate for biotechnological applications.

Materials and methods

Sample collection, extraction and purificationof metagenomic DNA

Groundwater samples were taken in May 2010 at a depth of10.0 m below ground level, ~7 m below the water table, froma site located in a former industrial area in the Czech Republic(N 50° 37′ 20″, E 15° 37′ 00″). The 2-l samples weretransported on ice to the laboratory and immediately filteredthrough a 0.2-μm polycarbonate membrane filter (Nuclepore,Whatman) using a vacuum pump (Waters). The collected bio-mass was frozen in liquid nitrogen and stored at −80 °C. Lysisof the thawed biomass and extraction and purification of themetagenomic DNA were performed as described previously(Kotik and Faměrová 2012).

Amplification of 16S rRNA genes, pyrosequencing,and sequence processing

The V4–V6 region of the eubacterial 16S ribosomal RNA(rRNA) genes was amplified by PCR using the metagenomicDNA as a template, the PfuUltra II Fusion Hot Start DNAPolymerase (Agilent Technologies), and the primers eub530fand eub1100br (Kotik et al. 2013). All preparatory steps forpyrosequencing on a GS Junior System (Roche) and the se-quence data processing were performed as described else-where (Kotik et al. 2013). Diversity estimates were calculatedusing Mothur (Schloss et al. 2009).

Appl Microbiol Biotechnol

Amplification of gene fragments

The candidate HLD gene fragments were PCR-amplifiedusing the GoTaq Hot Start Polymerase (Promega), andthe consensus primers Hld95f and Hld320r (Kotik andFaměrová 2012) in the presence of metagenomic DNA,3.5 mM MgCl2, and 200 μg ml−1 of bovine serumalbumin (New England Biolabs). Forty cycles were per-formed using an annealing temperature of 60.5 °C(60 s) and an extension temperature of 70 °C (55 s).The major PCR product (~700 bp) was purified from anagarose gel using a High Pure PCR Product PurificationKit (Roche) and cloned into the pGEM-T Easy vector(Promega). TOP10 cells (Invitrogen) were transformedwith the ligation mixes, and the plasmid DNA of whitecolonies was isolated using a High Pure PlasmidIsolation Kit (Roche). Insert sequences were obtainedusing the sequencing primers T7 and SP6.

Retrieval of full-length genes

The purified total metagenomic DNA was partially digestedwith Sau3A I, resulting in a smear (0.3–10 kbp) when loadedonto an agarose gel. An unphosphorylated double-strandedSau3A I linker (Takara Bio) was ligated to the restrictedDNA. After purification using a High Pure PCR ProductPurification Kit (Roche), the modified DNA was usedfor genome-walking PCR as described in Kotik et al. (2009,2010). Briefly, the missing upstream and downstream DNAsegments of the genes were retrieved in two consecutive PCRswith one primer being linker-specific (C1 or C2, Takara Bio)and the other being specific for the known HLD-encodingDNA sequence (Supplementary Table S1). The hot startproofreading Herculase Enhanced DNA polymerase (AgilentTechnologies) was used with or without dimethyl sulfoxidefor all genome-walking PCRs. The PCR products werepurified by agarose gel electrophoresis and cloned into thepGEM-T Easy vector for sequencing. Specific primers(Supplementary Table S2) which were derived from thedetermined upstream and downstream DNA sequencesenabled the amplification of the full-length genes using theHerculase Enhanced DNA polymerase in the presence ofdimethyl sulfoxide (1–4%) with the originally purifiedmetagenomic DNA as a template. The resulting PCR productswere cloned into the pGEM-T Easy vector for sequencing.The possibility of having retrieved chimeric HLD-encodinggenes was assessed by repeatedly analyzing the data set ofeach cluster with Mallard 1.02 (Ashelford et al. 2006),selecting in each run a different sequence as a reference.Alternatively, the following GenBank sequences, which wereidentified in BlastN searches, were used as references:CP000927.1 (for cluster A sequences), CP006936.2 (for

cluster B sequences), and CP000774.1 (for cluster C se-quences). Outliers were identified using a 99.9% cutoff value.

Overexpression and purification

The HLD-encoding genes were PCR-amplified using thePfuUltra II Fusion HS DNA polymerase (AgilentTechnologies), the TOPO-cloning primers (SupplementaryTable S3), and the pGEM-T-based plasmids as templates.The PCR products were gel-purified and re-purified using aHigh Pure PCR Cleanup Micro Kit (Roche). Cloning into theexpression vectors pET100/D-TOPO and pET101/D-TOPOwas performed according to the manufacturer’s instructions(Invitrogen). The pET100- and pET101-based constructswere transformed into Escherichia coli BL21 Star (DE3) cells(Life Technologies), which harbored the plasmid pGro7(Takara Bio) for co-expression of the chaperones GroES andGroEL. A transformant was grown for 6–8 h in 10 ml of LBmedium at 35 °C in the presence of ampicillin (100 μg ml−1)and chloramphenicol (34 μg ml−1). Four milliliters of this pre-culture was used for the inoculation of 1 l of LB mediumcontaining both antibiotics. The cells were cultivated over-night at 30 °C; the next day in the morning, the cultivationtemperature was reduced to 20 °C, and L-arabinose was addedto a final concentration of 11 mM. After 0.5 h, the expressionof the HLD was induced with IPTG (final concentration of0.45mM). The cultivation continued for 6–7 h before harvest-ing the biomass by centrifugation. The washed biomass wasre-suspended in binding buffer (10 ml per gram of wet weight;see below) containing DNase I (1.3 mg per ml of cell suspen-sion). The cells were subsequently broken by sonication (ul-trasonic cell disruptor Hielscher UP200S) using 0.3-s pulsesand an amplitude of 85%. The soluble fraction of the lysate,which was obtained by centrifugation (21,000×g at 4 °C, 1 h),was loaded on a Ni-nitrilotriacetic acid (Ni-NTA) Superflowcolumn (Qiagen) equilibrated with binding buffer (20 mMpotassium phosphate, 500 mM NaCl, 10 mM imidazole,pH 7.5). The column was washed with 10% elution buffer,and the bound target enzyme was eluted with a 60% stepgradient of elution buffer (20 mM potassium phosphate,500 mM NaCl, 300 mM imidazole, pH 7.5). The eluted pro-tein was dialyzed overnight at 4 °C against 50 mM phosphatebuffer (pH 7.5).

The HLD DbeA was overexpressed and purified aspreviously described (Chaloupkova et al. 2014).Protein concentrations were determined using theBradford method (Sigma-Aldrich). The purified proteinswere stored at 4 °C.

Circular dichroism

The circular dichroism (CD) spectra were recorded at roomtemperature using a Chirascan CD spectrometer equipped

Appl Microbiol Biotechnol

with a Peltier thermostat (Applied Photophysics, UK). Thepurified enzyme was diluted with 50 mM phosphate buffer(pH7.5) to reach a protein concentration of0.18–0.25mgml−1.The data were collected from 187 to 260 nm using a 1-mmquartz cuvette with a 100 nm min−1 scan rate, a response timeof 1 s, and a 1-nm bandwidth. Five scans were averaged togive a spectrum, which was subsequently corrected for thebuffer absorbance and smoothed using a Chirascan softwaretool.

The unfolding of eHLD-B and eHLD-C in 50 mM phos-phate buffer (pH 7.5) was followed by monitoring the ellip-ticity at 222 nm over a temperature range of 20 to 90 °C usinga heating rate of 1 °Cmin−1. The recorded denaturation curveswere fitted to a Boltzmann function (generating a sigmoidalcurve) using the OriginPro8 software (OriginLab). Themelting temperatures (Tm) were defined as the midpoint (x0)of the normalized thermal transitions.

Differential scanning calorimetry

Prior to differential scanning calorimetry (DSC), the samplecontaining 1 mgml−1 of the purified enzymewas degassed forat least 20 min under vacuum. For DSC, a MicroCalVP-Capillary calorimeter with a cell volume of 0.3 ml (GEHealthcare Life Sciences) was used at a constant pressure of50–55 psi. The scans were recorded using a scan rate of60 °C h−1 with a filtering period of 10 s. The melting temper-ature (Tm) was determined after baseline subtraction and con-centration normalization using the software Microcal Origin8.0 (OriginLab). Each denaturation profile is the average ofthree individual scans. All DSC measurements were per-formed in 50 mM phosphate buffer (pH 7.5).

Differential scanning fluorimetry

The thermal unfolding of proteins was also followed by re-cording the ratio of fluorescence intensities at 335 and 350 nmupon excitation at 295 nm using a heating rate of 1 °Cmin−1 inthe range of 20 to 90 °C with a Prometheus NT.48 scanningfluorometer (NanoTemper Technologies). The samplescontained 6 mg ml−1 of purified enzyme. The intensity ratioswere plotted as a function of the temperature, and the inflec-tion point of the resulting curve was defined as the meltingtemperature.

Determination of enzyme activity

Enzyme activities towards a set of 30 halogenated sub-strates (Sigma-Aldrich) were determined using aMicrolab STARlet liquid handling workstation (HamiltonRobotics). The substrate specificities were analyzed at25 °C for eHLD-B or 37 °C for eHLD-C using 2-ml vialscontaining 1 ml of 100 mM glycine–NaOH buffer (pH 8.6)

and 1 μl of the substrate. The reaction was initiated byadding 50 μl of purified enzyme (0.01–8 mg ml−1). Theprogress of the reaction was monitored by taking 75-μlsamples and mixing them with 10 μl of 35% HNO3 toterminate the reaction. A colorimetric end point assaybased on the detection of released halide ions usingHg(SCN)2 (9.5 mM in ethanol) and NH4Fe(SO4)2(250 mM in 9 M HNO3) (Iwasaki et al. 1952) was usedfor determining the temperature and pH dependence of theenzyme activities. The absorbance was detected at 460 nm.

Determination of enantioselectivity

The enzymatic conversions were performed at 20 °C in 2-mlvials. The reaction mixtures contained 1 ml of glycin buffer(100 mM, pH 8.6) and 1 μl of the racemic substrate. Thereaction was initiated by adding 50 μl of the enzyme (0.01–2.5 mg ml−1). Samples of 1 μl were withdrawn and subse-quently analyzed by chiral gas chromatography using anA s t e c CH IRALDEX B -DM cap i l l a r y c o l umn(50 m × 0.25 mm; Sigma-Aldrich) at an isothermal tempera-ture of 60 °C for 8 min and at 70 °C for 12 min for 2-bromopentane and ethyl 2-bromopropionate, respectively.The enantioselectivity was expressed as the enantiomeric ratio(E) defined as the ratio between the specificity constants (kcat/Km) for the two enantiomers:

E ¼kRcat

.KR

m

kScat.KS

m

The progress curves were fitted to the competitiveMichaelis-Menten equation (Lutje Spelberg et al. 1998) usingthe nonlinear least-squares regression software DynaFit(Kuzmič 2009).

Size exclusion chromatography

Themolecular masses of eHLD-B and eHLD-Cwere assessedby SEC using a Superdex 200 10/300 GL column (GEHealthcare Life Sciences) with a flow rate of 0.4 ml min−1.The mobile phase was composed of 50 mM Tris–HCl and150 mM NaCl (pH 7.2). The calibration proteins were cyto-chrome c, carbonic anhydrase, albumin, alcohol dehydroge-nase, and β-amylase.

Statistical analyses

The specific activity data of eHLD-B, eHLD-C, and 13 pre-viously examined wild-type HLDs towards 30 substrates(Supplementary Table S4) were analyzed by two principalcomponent analyses (PCAs) as described in Koudelakovaet al. (2011) using Statistica 10.0 (StatSoft, USA).

Appl Microbiol Biotechnol

Dendrograms based on similarities of individual HLDs in theirsubstrate specificity profiles were constructed using theneighbor-joining method based on the PCA results as de-scribed in Koudelakova et al. (2011).

Construction of phylogenetic trees and homology models

Phylogenetic trees were generated in MEGA using theneighbor-joining method with 1000 bootstrap re-samplings (Tamura et al. 2007). The evolutionary dis-tances were computed using the Poisson correction meth-od. Structural homology models were constructed usingthe automated modeling server SWISS-MODEL with de-fault settings (Arnold et al. 2006). The distances of res-idues to the solvent-accessible surface of the modelswere calculated using the SurfDis command in the pro-gram YASARA (Krieger and Vriend 2014).

Nucleotide sequence accession numbers

The nucleotide sequences of the retrieved full-length HLD-encoding genes are available in GenBank under the accessionnumbers KF483136–KF483146.

Results

Site characteristics

Groundwater and soil of the investigated area were continu-ously contaminated with various products of the oil industryfor 50 to 70 years until 1995. The groundwater was sampledfrom a slightly contaminated monitoring well with cis-1,2-dichloroethene, trichloroethylene, and tetrachloroethylene be-ing the primary contaminants at this site, reaching concentra-tions of 15.9, 3.9, and 1.8 μg l−1, respectively. 1,2-Dichloroethane was not detected at the site using standardanalytical procedures (<0.5 μg l−1). A temperature of 8.5 °Cand pH of 6.9 were determined at the time of sampling.

Microbial diversity

The pyrosequencing generated 2000 sequence reads with se-quence lengths ≥400 bases, resulting in 627 operational taxo-nomic units (OTUs). The statistical description of the 16SrRNA gene library indicates a high level of species richness(Supplementary Table S5; Kotik et al. 2013). One should notethat the diversity estimates are always dependent on the sam-pling depth (Hughes et al. 2002). While recognizing that, oursole intension was to estimate the microbial diversity to showthat a complex community had been analyzed. All OTUs withseqmatch scores of ≥0.950 (Cole et al. 2014) or abundances of≥2‰ can be found in the Supplementary Table S6. The

microbial community was dominated by Betaproteobacteria,Gammaproteobacteria, and Alphaproteobacteria, with mem-bers of the orders Burkholderiales, Rhodocyclales,Xanthomonadales, Chromatiales, and Rhizobiales being themost abundant (Supplementary Figs. S1 and S2).

Retrieved full-length HLD-encoding sequences

The fragments of the putative HLD-encoding genes were di-rectly amplified from the metagenomic DNA using degener-ate hld-specific consensus primers (Kotik and Faměrová2012). The obtained PCR products enabled us to amplify theregions adjacent to these core sequences using a PCR-basedgenome-walking technique with linker-coupled DNA as atemplate in combination with linker-specific and nestedprimers. With the exception of hldC_gw, two upstream walk-ing rounds were performed to reach the start codons of the hldgenes. One downstream walking round was sufficient to am-plify the missing fragments which encoded the C-termini ofthe HLDs (Supplementary Table S1). The sequences of thefull-length genes were obtained via the assembly of overlap-ping fragment sequences (Supplementary Figs. S3 to S6). Inaddition, three longer stretches of DNA adjacent to the assem-bled hld genes were obtained; a BlastX analysis of these se-quences revealed the presence of ORFs encoding hypotheticalproteins or an arylsulfatase (Supplementary Table S7).

The specific primers which were derived from the noncod-ing regions upstream and downstream of the four assembledhld genes enabled us to retrieve a set of putative full-lengthHLD-encoding genes from the metagenomic DNA sample(Supplementary Figs. S7 to S9). The conceptual translationof the sequences resulted in protein sequences, which afteralignment generated the clusters A, B, and C in the dendro-gram (Supplementary Fig. S10). The phylogenetic analysisrevealed that the cluster A and B sequences belong to theHLD-I clade with DpcA (Drienovska et al. 2012) as the clos-est functionally tested HLD representative (sequence identi-ties of 60%). The cluster C sequences were found to be mem-bers of the HLD-II clade with DadB (Li and Shao 2014) beingthe closest characterized HLD (sequence identity of 66%).The protein sequences within each cluster exhibited high se-quence identities: 95–99% for cluster A sequences, 98–99%for cluster B sequences, and 97–99% for cluster C sequences(Supplementary Figs. S11 to S13). Two nucleotide sequencesencoding two HLDs within cluster Awere identified as poten-tially chimeric (Supplementary Fig. S11) with a breakpoint atposition ~600 bp.

The residues which are essential for catalysis were predict-ed through sequence comparison with HLDs with known ter-tiary structures—DmrA (Fung et al. 2015) for cluster A and Bsequences and HanR (Novak et al. 2014) for cluster C se-quences. The positions of the deduced catalytic residues inthe cluster sequences were found to represent classical

Appl Microbiol Biotechnol

catalytic pentad arrangements of subfamily I and II HLDswith Asp as the nucleophile, His as the catalytic base, Asnand/or Trp as the halide-stabilizing residues, and Asp or Gluas the catalytic acid (Supplementary Figs. S11 to S13).

Model structures

In an attempt to locate the various amino acid exchanges of themetagenome-derived HLD variants in the protein structures,homology models were constructed using the known crystalstructures of DmrA (Fung et al. 2015) and HanR (Novak et al.2014) as the templates (Supplementary Figs. S14 to S16). Itturned out that the vast majority of these exchange sites is notburied in the protein, as the distances between the residues ofthese sites and the protein surface were determined to be 2.9–3.5 Å, which are values of highly accessible residues(Chakravarty and Varadarajan 1999). However, two locationswere found to be rather distant from the surface of the modelstructures: 5.8 Å for Val-77 in eHLD-A and 4.5 Å for Tyr-60in eHLD-B (Supplementary Figs. S17 to S19). Interestingly,the detected exchanges at these buried positions can be con-sidered conservative with Val-77 and Tyr-60 being replacedby Ile and Phe, respectively. On the other hand, at sites close tothe surface, we observed all kinds of exchanges, conservativeones, such as Lys↔Arg, Ile↔Val, or Asp↔Glu, but alsothose that fundamentally alter the physicochemical propertiesof a residue, e.g., Pro↔Ser, Ala↔Thr, Gly↔Ser, orGly↔Arg (Supplementary Table S8). The sites of the muta-tions appear to be unevenly distributed over the HLD se-quences with stretches of higher variability and stretches ofvirtually no exchange. Our data suggest that a nearly invari-able segment of ~50–70 amino acids may stretch fromβ-sheet5 up to α-helix 4, which includes the catalytic nucleophileAsp and both halide-stabilizing Trp residues in the HLD-Isubfamily.

Purification and preliminary activity tests of recombinantHLDs

Four metagenome-derived hld genes were selected for theheterologous expression and purification of the correspondingenzymes containing C-terminal His-tags: eHLD-A andeHLD-D from cluster A, eHLD-B from cluster B, andeHLD-C from cluster C (Supplementary Figs. S10, S11 toS13). Only eHLD-C exhibited dehalogenase activity towardsthe representative and universal HLD substrates 1-iodobutane,1,2-dibromoethane, and 4-bromobutyronitrile. Elution duringprotein purification did not result in any protein fractions con-taining eHLD-A or eHLD-D. In a second attempt, eHLD-Aand eHLD-B with N-terminal His-tags were tested fordehalogenase activity with 1,3-dibromopropane and 1-chlorohexane. In this case, eHLD-B was found to be active.Active eHLD-B and eHLD-C were purified to a high degree

of homogeneity using Ni-NTA affinity chromatography(Supplementary Fig. S20).

Biophysical characterization of eHLD-B and eHLD-C

The purified His-tagged enzymes eHLD-B and eHLD-C elut-ed during size exclusion chromatography as monomers withdeduced molecular masses of 37.9 and 37.4 kDa, respectively(Supplementary Table S9 and Fig. S21). No indication of di-mer formation was observed. The far-UV CD spectra of bothenzymes revealed the typical structural features of foldedHLD proteins with a characteristic positive peak at ~195 nmand two negative peaks at 208 and 222 nm, which is indicativeof a large fraction of α-helices (Supplementary Fig. S22). Thethermal unfolding of both enzymes was compared at pH 7.5using (i) DSC and (ii) the temperature-dependent CD signal at222 nm. The determined melting temperatures using eithertechnique were in good agreement: 34.0–35.1 and 55.5–58.7 °C for eHLD-B and eHLD-C, respect ively(Supplementary Table S10 and Fig. S23).

Biochemical characterization of eHLD-B and eHLD-C

The temperature and pH optima of the purified HLDs were de-termined using 1,3-dibromopropane as a substrate. For eHLD-B,the optimal pH was 8.5 with enzyme activities exceeding half-maximum values between pH 6.0 and ~9.7 (Fig. 1a). On theother hand, the pH optimumof eHLD-Cwas found to be broaderwith a plateau region of maximum activity between pH 5.5 and8.9. Activities of eHLD-C that exceeded half-maximum valueswere determined between pH 4.7 and 10.0 (Fig. 1a). eHLD-Bexhibited a low optimal temperature of 25 °C at pH 8.5, whereasa significantly higher optimal temperature of 55 °C (pH 6.5) wasdetermined for eHLD-C (Fig. 1b).

The substrate specificity profiles of eHLD-B and eHLD-Cwere investigated using a panel of 30 halogenated hydrocarbons,which were previously selected on grounds of large differencesin their physicochemical properties (Koudelakova et al. 2011).Generally, eHLD-C exhibited substantially higher specific activ-ities than eHLD-B, the difference in the activity being two tothree orders of magnitude (Table 1). The highest specific activityof eHLD-B was determined for 1,3-dibromopropane (substrateno. 48) and 1-bromo-3-chloropropane (no. 52). On the otherhand, eHLD-C was most active towards 1,2-dibromoethane(no. 47) and 1-bromo-2-chloroethane (no. 137). Moreover, thesubstrate range of this enzyme was broader with detected activ-ities towards 28 out of 30 tested substrates. Both enzymes did notreact with 1,2,3-trichloropropane (no. 80) and chlorocyclohexane(no. 115). The substrate specificity profile of eHLD-C, with theexception of 1-bromobutane (no. 18), is characterized by a pref-erence for multisubstituted compounds with chain lengths of ≤3carbon atoms (substrates nos. 47, 48, 52, 72, 76, 137, 154, 155).Furthermore, eHLD-C exhibited a marked preference for

Appl Microbiol Biotechnol

brominated compounds over chlorinated and iodinated ones. Bycontrast, eHLD-B showed less preference for haloalkanes withshort chain lengths, i.e., relatively high activities were determinedfor 1,5-dichloropentane and 1-bromohexane (Fig. 2).

The enantioselectivities of eHLD-B and eHLD-C wereassessed using 2-bromopentane, represent ing β-bromoalkanes, and ethyl-2-bromopropionate, which repre-sents α-brominated esters (Supplementary Fig. S24).Interestingly, the kinetic resolution experiments with the for-mer compound revealed an S-enantiopreference for eHLD-B,though with a very low E-value of 3. eHLD-C also exhibited alow discrimination between the enantiomers of this compoundwith a determined E-value of 6; however, it preferentiallyhydrolyzed the R-enantiomer. A medium enantioselectivitywith an E-value of 40 was observed for ethyl-2-b romoprop iona t e us ing eHLD-B, whi l e a h ighenantioselectivity with an E-value of 101 was determined foreHLD-C; both enzymes exhibited a preference for the R-en-antiomer (Table 2). Generally low enantioselectivities can berelated to the environmental origin of the host organisms,which could be exposed to the toxicities of both enantiomers.

Functional classification of eHLD-B and eHLD-C

PCA was performed to compare the substrate specificities ofboth novel HLDs with 13 previously characterized HLDs.First, the data containing the untransformed specific activitiestowards the 30 abovementioned substrates was analyzed byPCA, establishing three significant principal components. Theenzymes were ranked by the first principal component accord-ing to their overall activities (Fig. 3a), which accounted for39% of the variance in this dataset. eHLD-B positionedamong the less active enzymes, whereas eHLD-C was identi-fied as an enzyme with a high overall activity. In order to gaininsight into the differences of the individual HLD activityprofiles, the activity data were log-transformed and weightedprior to PCA. Again, three statistically significant principalcomponents were identified. In the resulting three-dimensional t1–t2–t3 score plot, which explained 56% of thevariance in the dataset, the objects (HLDs) were found to beclustered in four substrate specificity groups (SSG) accordingto the similarities in their specificity profiles (Fig. 3b). Thedistribution of the individual HLDs into four SSGs is also seenin a two-dimensional t1–t3 score plot defined by the first andthird principal components (Fig. 3d). Comparison of the t1–t3score plot with the corresponding p1–p3 loading plot (Fig. 3c)enabled the identification of the relative activities with partic-ular substrates that were primarily responsible for the cluster-ing of the enzymes (Koudelakova et al. 2011). eHLD-B andeHLD-C were positioned in two different clusters, SSG-IVand SSG-I, respectively (Fig. 3d). The constructed substratespecificity dendrogram also supports the functional classifica-tion of the HLDs into four SSGs (Supplementary Fig. S25).

Thermostabilities of eHLD-C and DbeA

The thermostability experiments revealed an unusually highrobustness of eHLD-C, which was explored by a comparativeanalysis with DbeA. DbeA was selected for the comparisonbecause its thermostability was reported to be the highestamong HLDs (Chaloupkova et al. 2014; Fung et al. 2015; Liand Shao 2014). The melting temperatures were determinedfor both enzymes in the pH range of 4.0 to 11.0 by monitoringthe tryptophan fluorescence signal. The data confirmed a veryhigh thermal stability of both enzymes with eHLD-C beingsignificantly more stable in acidic conditions with pH valuesof <6.0 (Fig. 4).

Discussion

Recent literature data suggest that HLD-encoding genes arenot rare in bacteria (Chovancova et al. 2007; Kotik andFaměrová 2012; Koudelakova et al. 2011). Moreover, theiroccurrence is not limited to bacteria which exist at sites

0.0

20.0

40.0

60.0

80.0

100.0

2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0

Rel

ativ

e ac

tivity

(%)

pH

eHLD-CeHLD-B

a

0.0

20.0

40.0

60.0

80.0

100.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0

Rel

ativ

e ac

tivity

(%)

Temperature (°C)

eHLD-CeHLD-B

b

Fig. 1 Temperature and pH optima of eHLD-B and eHLD-C activitiesdetermined with 1,3-dibromopropane. a Dependence of HLD activity onpH at 25 °C. bTemperature dependence of activities determined at pH 8.5for eHLD-B and pH 6.5 for eHLD-C

Appl Microbiol Biotechnol

contaminated with haloalkanes (Fung et al. 2015; Jesenskáet al. 2002), which may reflect the presence of natural haloge-nation processes described in various environments (Alberset al. 2011; Bastviken et al. 2009; Clarke et al. 2009). Asshown in this report, HLD-encoding genes can also be re-trieved from metagenomes with no obvious relation tohaloalkane contamination, since the major contaminants atthe investigated site were chlorinated ethenes, which are notsubstrates for HLDs. The investigated site exhibited a highdegree of microbial diversity and consequently was thoughtto be a good metagenomic source of diverse HLD-encodinggenes. For obvious reasons, our PCR-based approach did notenable us to identify the hosts of the HLDs.

The biochemical characterization revealed marked differ-ences between eHLD-B and eHLD-C in their optimal

temperatures, thermostabilities, and substrate specificities.With a denaturation temperature of 34–35 °C, eHLD-B isclearly less stable than most of the characterized HLDs, withthe notable exception of DpcA from Psychrobactercryohalolentis K5T (Drienovska et al. 2012). Thus, the origi-nal host of eHLD-B appears to be a cold-adapted organism.Although retrieved from the same metagenome, eHLD-C ex-hibited a high denaturation temperature of 55–58 °C, which isat the upper end of currently characterized HLDs (Drienovskaet al. 2012).

eHLD-C showed exceptionally high activities towards sixsubstrates (nos. 47, 48, 52, 72, 137, 154, 155), considering thereported specific activities of other biochemically character-ized HLDs (Koudelakova et al. 2011). eHLD-B generally ex-hibited much lower specific activities, which is, however, not

Table 1 Specific activities ofeHLD-B and eHLD-C towards apanel of 30 selected halogenatedsubstrates

Numbera Substrate Specific activity (nmol s−1 mg−1)

eHLD-B eHLD-C

4 1-Chlorobutane –b 1.1

6 1-Chlorohexane 1.7 1.2

18 1-Bromobutane 5.0 74.1

20 1-Bromohexane 3.8 7.9

28 1-Iodopropane 1.3 9.4

29 1-Iodobutane 1.9 8.9

31 1-Iodohexane 0.2 0.5

37 1,2-Dichloroethane – 1.9

38 1,3-Dichloropropane – 20.0

40 1,5-Dichloropentane 3.7 5.5

47 1,2-Dibromoethane 2.2 203.8

48 1,3-Dibromopropane 8.1 80.5

52 1-Bromo-3-chloropropane 6.7 148.1

54 1,3-Diiodopropane 2.7 7.2

64 2-Iodobutane 0.6 6.3

67 1,2-Dichloropropane – 1.8

72 1,2-Dibromopropane – 132.3

76 2-Bromo-1-chloropropane 0.4 126.7

80 1,2,3-Trichloropropane – –

111 Bis(2-chloroethyl)ether – 8.3

115 Chlorocyclohexane – –

117 Bromocyclohexane – 9.1

119 (1-Bromomethyl)cyclohexane 0.2 1.0

137 1-Bromo-2-chloroethane 0.9 172.1

138 Chlorocyclopentane – 17.6

141 4-Bromobutyronitrile 6.0 51.0

154 1,2,3-Tribromopropane – 105.8

155 1.2-Dibromo-3-chloropropane 0.4 79.6

209 3-Chloro-2-methylpropene 1.4 31.0

225 2,3-Dichloropropene 0.5 32.5

a Substrate numbering is based on the database of halogenated compounds described in Koudelakova et al. (2011)b No activity was detected under the tested conditions

Appl Microbiol Biotechnol

unusual in view of the available activity data for other HLDssuch as DmbC, DmrB, DrbA, and DspA (Fortova et al. 2013;Fung et al. 2015; Jesenská et al. 2009). The poor catalyticefficiency of eHLD-B, which appears to be a typical represen-tative of the subfamily I of HLDs (Supplementary Fig. S10),raises the question of its physiological role. One should keepin mind that the halogenated compounds commonly used forHLD characterization may be structurally different from thecognate substrates, which may explain the low enzymatic ac-tivities. Comparing substrate specificity groupings (Fig. 3)with evolutionary distance data (Supplementary Fig. S10)suggests that there is no correlation between the substratespecificity profile of a particular HLD and its sequence-based phylogeny. In other words, HLDs within a given SSGare not necessarily evolutionary related. For instance, the sev-en SSG-IV members DatA, DbeA, DmbC, DpcA, DspA,DsvA, and eHLD-B fall into the three existing phylogeneticsubgroups, not just into one. A lack of correlation betweenphylogenic and substrate specificity data of HLDs had beenpreviously reported (Koudelakova et al. 2011). eHLD-C wasplaced in SSG-I, whose members are characterized by theircatalytic robustness, accepting most of the selected halogenat-ed compounds as substrates (Koudelakova et al. 2011). In

addition, these enzymes share similar substrate preferences,which include the following substrates: 1,2-dibromoethane(no. 47), 1,3-dibromopropane (no. 48), 1-bromo-3-chloropropane (no. 52), and 1-bromo-2-chloroethane (no.137). HLDs within the SSG-IV cluster represent enzymeswith a more restricted substrate range. The representatives ofthis cluster have a preference for terminally substituted bromi-nated butanes and propanes (Koudelakova et al. 2011), as wasshown for eHLD-B.

All previously investigated HLDs have been found to havepH optima of ≥8.0 with the notable exception of DmbB. Thisenzyme, however, possesses a very low activity (Jesenskáet al. 2005), which hampers its use in practical applications.

Fig. 2 Comparison of substratespecificity profiles of eHLD-Band eHLD-C. Relative activitiesare shown towards a set of 30chlorinated (white),chlorobrominated (dark gray),brominated (light gray), andiodinated substrates (black)

Table 2 Enantioselectivities of eHLD-B and eHLD-C towards 2-bromopentane and ethyl-2-bromopropionatea

Racemic substrate eHLD-B eHLD-C

2-Bromopentane 3 (S) 6 (R)

Ethyl-2-bromopropionate 40 (R) 101 (R)

a The numbers indicate E values with corresponding enantiopreferencesin parentheses

Appl Microbiol Biotechnol

In contrast, eHLD-C shows an optimal activity betweenpH 5.5 and 8.9; this characteristic and the high thermostabilitymake this enzyme a promising candidate for various practicalapplications. In particular, the unusual pH tolerance makesthis enzyme interesting for processes which require operationat lower pH: (i) bioremedation or biosensing of environmentalpollutants in neutral to acidic environments or (ii) industrialbiocatalysis of halogenated hydrocarbons showing generallylower stability at neutral to alkaline conditions.

The metagenome represents a huge trove of genes fromthousands of microorganisms which are present at a specificsampling site (Handelsman et al. 1998). Many of these micro-organisms have not yet been cultivated or are at present un-cultivable. Thus, metagenomic DNA can fundamentally en-large our collection of biotechnologically interesting enzymes,providing access to an immense number of genes (Steele et al.2008; Uchiyama and Miyazaki 2009). As shown in this andprevious works, PCR with degenerate primers in conjunctionwith genome walking is one strategy to prospect for novelenzyme-encoding genes with potential applications in

biotransformations (Bell et al. 2002; Eschenfeldt et al. 2001;Hayashi et al. 2005; Jiang et al. 2006; Kotik 2009; Kotik et al.2010; Labes et al. 2008; Sunna and Bergquist 2003; Tang et al.

Fig. 3 Quantitative comparisonof activity and specificity ofeHLD-B and eHLD-C with 13previously characterized HLDsusing principle componentanalysis (PCA). a PCA ofuntransformed specific activitydata, resulting in a t1 score plot.Ranking of the HLDs by the firstprincipal component according totheir overall enzyme activities isshown. The dataset was based on30 selected substrates. b PCAwith the transformed dataset. Thet1–t2–t3 score plot shows thedistribution of the objects (HLDs)based on their substratespecificity profiles. t1, t2, and t3are the principal componentscores of the individual HLDs.Using PCA, the HLDs weredistributed into four substratespecificity groups (SSGs)according to their substratespecificity profiles. c, d PCAusing the transformed dataset: thep1–p3 loading plot (c) and thecorresponding t1–t3 score plot (d)are shown. The t1–t3 score plotaccounts for 43% of the variancein the dataset. The p1–p3 loadingplot quantifies the contribution ofeach variable (relative activitytowards a particular substrate) to aparticular principal component

Fig. 4 Comparison of melting temperatures for eHLD-C (white bars)and DbeA (black bars) determined in the pH range 4.0–11.0. Thetryptophan fluorescence signal was used as a probe of the unfolding ofthe tertiary structures

Appl Microbiol Biotechnol

2006; Uchiyama and Watanabe 2006; Weerachavangkul et al.2012). Here, we have shown that metagenomic DNA is avaluable source of novel HLD-encoding genes. Moreover,the pool of evolutionary closely related genes present at thesampling site represents a resource of HLD variants with po-tentially altered substrate specificities or other enzyme char-acteristics. These natural HLD variants can be of biotechno-logical value because replacing a few residues at key sites cansignificantly alter the substrate specificities in HLDs (Pavlovaet al. 2009; Pries et al. 1994). This study illustrates the highpotential of PCR-based gene mining directly frommetagenomic DNA to enlarge the existing collection of prom-ising HLDs. We have shown that the technique is suitable forthe isolation of genes encoding novel HLDs with markedlydifferent properties such as thermostabilities, pH–activity pro-files, substrate specificities, and enantioselectivities.

Acknowledgements We would like to thank Envisan—GEM a.s.(Prague, Czech Republic) for the groundwater samples and their analyses.This work was supported by the Czech Science Foundation (grantsP504/10/0137 and 16-07965S). We also acknowledge support for thelong-term conceptual advancement of the Institute of Microbiology(RVO61388971). The research infrastructure for this project was support-ed by the following grants from the Czech Ministry of Education:LO1214, LM2015051, LM2015047, and LM2015055. Antonin Kunkaacknowledges the Brno Municipality for funding a Ph.D. TalentScholarship.

Compliance with ethical standards

Conflict of interest The authors declare no conflict of interest.

Ethical statement All laboratory experiments were carried in line withthe ethical guidelines.

References

Albers CN, Jacobsen OS, Flores ÉMM, Pereira JSF, Laier T (2011)Spatial variation in natural formation of chloroform in the soils offour coniferous forests. Biogeochemistry 103:317–334

Anke H, Weber RWS (2006) White-rots, chlorine and the environment—a tale of many twists. Mycologist 20:83–89

Arnold K, Bordoli L, Kopp J, Schwede T (2006) The SWISS-MODELworkspace: a web-based environment for protein structure homolo-gy modeling. Bioinformatics 22:195–201

Ashelford KE, Chuzhanova NA, Fry JC, Jones AJ,WeightmanAJ (2006)New screening software shows that most recent large 16S rRNAgene clone libraries contain chimeras. Appl Environ Microbiol 72:5734–5741

Ballschmiter K (2003) Pattern and sources of naturally producedorganohalogens in the marine environment: biogenic formation oforganohalogens. Chemosphere 52:313–324

Bastviken D, Svensson T, Karlsson S, Sandén P, Oberg G (2009)Temperature sensitivity indicates that chlorination of organic matterin forest soil is primarily biotic. Environ Sci Technol 43:3569–3573

Bell PL, Sunna A, Gibbs MD, Curach NC, Nevalainen H, Bergquist PL(2002) Prospecting for novel lipase genes using PCR.Microbiology-(UK) 148:2283–2291

Beschkov V, Sapundzhiev T, Torz M, Wietzes P, Janssen DB (2008)Modeling of 1,2-dichloroethane biodegradation by Xanthobacterautotrophicus GJ10 under shock loading of other halogenated com-pounds in continuous stirred tank bioreactor. Chem Biochem Eng Q22:339–348

Bidmanova S, Chaloupkova R, Damborsky J, Prokop Z (2010)Development of an enzymatic fiber-optic biosensor for detectionof halogenated hydrocarbons. Anal Bioanal Chem 398:1891–1898

Bosma T, Damborsky J, Stucki G, Janssen DB (2002) Biodegradation of1,2,3-trichloropropane through directed evolution and heterologousexpression of a haloalkane dehalogenase gene. Appl EnvironMicrobiol 68:3582–3587

Cavicchioli R, Siddiqui KS, Andrews D, Sowers KR (2002) Low-temperature extremophiles and their applications. Curr OpinBiotechnol 13:253–261

Chakravarty S, Varadarajan R (1999) Residue depth: a novel parameterfor the analysis of protein structure and stability. Structure 7:723–732

Chaloupkova R, Prudnikova T, Rezacova P, Prokop Z, Koudelakova T,Daniel L, Brezovsky J, Ikeda-Ohtsubo W, Kuty M, Sato Y, NagataY, Smatanova K, Damborsky J (2014) Structural and functionalanalysis of a novel haloalkane dehalogenase with two halide-binding sites. Acta Cryst D70:1884–1897

Chan WY, Wong M, Guthrie J, Savchenko AV, Yakunin AF, Pai EF,Edwards EA (2010) Sequence- and activity-based screening of mi-crobial genomes for novel dehalogenases. Microb Biotechnol 3:107–120

Chovancova E, Kosinski J, Bujnicki JM, Damborsky J (2007)Phylogenetic analysis of haloalkane dehalogenases. Proteins 67:305–316

Cole JR, Wang Q, Fish JA, Chai B, McGarrell DM, Sun Y, Brown CT,Porras-Alfaro A, Kuske CR, Tiedje JM (2014) Ribosomal databaseproject: data and tools for high throughput rRNA analysis. NuclAcids Res 42:633–642

Clarke N, Fuksova K, Gryndler M, Lachmanova Z, Liste H-H,Rohlenova J, Schroll R, Schröder P, Matucha M (2009) The forma-tion and fate of chlorinated organic substances in temperate andboreal forest soils. Environ Sci Pollut Res 16:127-143

Daniel L, Buryska T, Prokop Z, Damborsky J, Brezovsky J (2015)Mechanism-based discovery of novel substrates of haloalkanedehalogenases using in silico screening. J Chem Inf Model 55:54–62

Drienovska I, Chovancova E, Koudelakova T, Damborsky J,Chaloupkova R (2012) Biochemical characterization of a novelhaloalkane dehalogenase from a cold-adapted bacterium. ApplEnviron Microbiol 78:4995–4998

Erable B, Maugard T, Goubet I, Lamare S, Legoy MD (2005)Biotransformation of halogenated compounds by lyophilized cellsof Rhodococcus erythropolis in a continous solid–gas biofilter.Process Biochem 40:45–51

Eschenfeldt WH, Stols L, Rosenbaum H, Khambatta ZS, Quaite-RandallE, Wu S, Kilgore DC, Trent JD, Donnelly MI (2001) DNA fromuncultured organisms as a source of 2,5-diketo-D-gluconic acid re-ductases. Appl Environ Microbiol 67:4206–4214

Fortova A, Sebestova E, Stepankova V, Koudelakova T, Palkova L,Dambo r sky J , Cha l oupkova R (2013 ) DspA f r omStrongylocentrotus purpuratus: the first biochemically characterizedhaloalkane dehalogenase of non-microbial origin. Biochimie 95:2091–2096

Fung HKH, Gadd MS, Drury TA, Cheung S, Guss JM, Coleman NV,Matthews JM (2015) Biochemical and biophysical characterisationof aloalkane dehalogenases DmrA and DmrB in Mycobacteriumstrain JS60 and their role in growth on haloalkanes. Mol Microbiol97:439–453

Gray KA, Richardson TH, Robertson DE, Swanson PE, SubramanianMV (2003) Soil-based gene discovery: a new technology to

Appl Microbiol Biotechnol

accelerate and broaden biocatalytic applications. Adv ApplMicrobiol 52:1–27

Gribble GW (2000) The natural production of organobromine com-pounds. Environ Sci Pollut Res 7:37–49

Gribble GW (2003) The diversity of naturally produced organohalogens.Chemosphere 52:289–297

Gribble GW (2015) A recent survey of naturally occurring organohalogencompounds. Environ Chem 12:396–405

Hammer PE, Hill DS, Lam ST, Van Pée K-H, Ligon JM (1997) Fourgenes from Pseudomonas fluorescens that encode the biosynthesisof pyrrolnitrin. Appl Environ Microbiol 74:2147–2154

Handelsman J, Rondon MR, Brady SF, Clardy J, Goodman RM (1998)Molecular biological access to the chemistry of unknown soil mi-crobes: a new frontier for natural products. Chem Biol 5:245–249

Harper DB, McRoberts WC, Kennedy JT (1996) Comparison of theefficacies of chloromethane, methionine, and S-adenosylmethionineas methyl precursors in the biosynthesis of veratryl alcohol andrelated compounds in Phanerochaete chrysosporium. ApplEnviron Microbiol 62:3366–3370

Hayashi H, Abe T, Sakamoto M, Ohara H, Ikemura T, Sakka K, Benno Y(2005) Direct cloning of genes encoding novel xylanases from thegut. Can J Microbiol 51:251–259

Hughes JB, Hellmann JJ, Ricketts TH, Bohannan BJM (2002) Countingthe uncountable: statistical approaches to estimating microbial di-versity. Appl Environ Microbiol 68:448

Iwasaki I, Utsumi S, Ozawa T (1952) New colorimetric determination ofchloride using mercuric thiocyanate and ferric ion. Bull Chem SocJpn 25:226

Janssen DB (2007) Biocatalysis by dehalogenating enzymes. Adv ApplMicrobiol 61:233–252

Janssen DB, Oppentocht JE, Poelarends GJ (2001) Microbialdehalogenation. Curr Opin Biotechnol 12:254–258

Jesenská A, Bartoš M, Czerneková V, Rychlík I, Pavlík I, Damborský J(2002) Cloning and expression of the haloalkane dehalogenase genedhmA from Mycobycterium avium N85 and preliminary characteri-zation of DhmA. Appl Environ Microbiol 68:3724–3730

Jesenská A, Pavlová M, Strouhal M, Chaloupková R, Těšínská I,Monincová M, Prokop Z, Bartoš M, Pavlík I, Rychlík I, Möbius P,Nagata Y, Damborský J (2005) Cloning, biochemical properties, anddistribution of mycobacterial haloalkane dehalogenases. ApplEnviron Microbiol 71:6736–6745

Jesenská A, Monincová M, Koudeláková T, Hasan K, Chaloupková R,Prokop Z, Geerlof A, Damborský J (2009) Biochemical characteri-zation of haloalkane dehalogenases DrbA and DmbC, representa-tives of a novel subfamily. Appl Environ Microbiol 75:5157–5160

Jiang Z, Wang H, Ma Y, Wei D (2006) Characterization of two novellipase genes isolated directly from environmental sample. ApplMicrobiol Biotechnol 70:327–332

de Jong E, Field JA (1997) Sulfur tuft and turkey tail: biosynthesis andbiodegradation of organohalogens by Basidiomycetes. Annu RevMicrobiol 51:375–414

Keppler F, Borchers R, Pracht J, Rheinberger S, Scholer HF (2002)Natural formation of vinyl chloride in the terrestrial environment.Environ Sci Technol 36:2479–2483

Kotik M (2009) Novel genes retrieved from environmental DNA bypolymerase chain reaction: current genome-walking techniques forfuture metagenome applications. J Biotechnol 144:75–82

Kotik M, Faměrová V (2012) Sequence diversity in haloalkanedehalogenases, as revealed by PCR using family-specific primers.J Microbiol Methods 88:212–217

Kotik M, Štěpánek V, Marešová H, Kyslík P, Archelas A (2009)Environmental DNA as a source of a novel epoxide hydrolasereacting with aliphatic terminal epoxides. J Mol Catal B Enzym56:288–293

Kotik M, Štěpánek V, Grulich G, Kyslík P, Archelas A (2010) Access toenantiopure aromatic epoxides and diols using epoxide hydrolasesderived from total biofilter DNA. J Mol Catal B Enzym 65:41–48

Kotik M, Davidová A, Voříšková J, Baldrian P (2013) Bacterial commu-nities in tetrachloroethene-polluted groundwaters: a case study. SciTotal Environ 454-455:517–527

Koudelakova T, Chovancova E, Brezovsky J, Monincova M, Fortova A,Jarkovsky J, Damborský J (2011) Substrate specificity of haloalkanedehalogenases. Biochem J 435:345–354

Koudelakova T, Bidmanova S, Dvorak P, Pavelka A, Chaloupkova R,Prokop Z, Damborsky J (2013) Haloalkane dehalogenases: biotech-nological applications. Biotechnol J 8:32–45

Krieger E, Vriend G (2014) YASARA view—molecular graphics for alldevices—from smartphones to workstations. Bioinformatics 30:2981–2982

Kuzmič P (2009) DynaFit—a software package for enzymology. MethEnzymol 467:247–280

Labes A, Karlsson EN, Fridjonsson OH, Turner P, Hreggvidson GO,Kristjansson JK, Holst O, Schönheit P (2008) Novel members ofglycoside hydrolase family 13 derived from environmental DNA.Appl Environ Microbiol 74:1914–1921

Laturnus F, Fahimi I, Gryndler M, Hartmann A, Heal MR, Matucha M,Schöler HF, Schroll R, Svensson T (2005) Natural formation anddegradation of chloroacetic acids and volatile organochlorines inforest soil. Environ Sci Pollut Res 12:233–244

Li A, Shao Z (2014) Biochemical characterization of a haloalkanedehalogenase DadB from Alcanivorax dieselolei B-5. PLoS One9:e89144

Los GV, Encell LP, McDougall MG, Hartzell DD, Karassina N, ZimprichC, Wood MG, Learish R, Friedman Ohana R, Urh M, Simpson D,Mendez J, Zimmerman K, Otto P, Vidugiris G, Zhu J, Darzins A,Klaubert DH, Bulleit RF,WoodKV (2008) HaloTag: a novel proteinlabeling technology for cell imaging and protein analysis. ACSChem Biol 3:373–382

Lutje Spelberg JH, Rink R, Kellogg RM, Janssen DB (1998)Enantioselectivity of a recombinant epoxide hydrolase fromAgrobacterium radiobacter. Tetrahedron-Asymmetry 9:459–466

Marchand P, Lamare S, Legoy M-D, Goubet I (2009) Dehalogenation ofgaseous 1-chlorobutane by dehydrated whole cells: influence of themicroenvironment of the halidohydrolase on the stability of the bio-catalyst. Biotechnol Bioeng 103:687–695

Novak HR, Sayer C, Isupov MN, Gotz D, Spragg AM, Littlechild JA(2014) Biochemical and structural characterisation of a haloalkanedehalogenase from marine Rhodobacteraceae. FEBS Lett 588:1616–1622

Pavlova M, Klvana M, Prokop Z, Chaloupkova R, Banas P, Otyepka M,Wade RC, Tsuda M, Nagata Y, Damborsky J (2009) Redesigningdehalogenase access tunnels as a strategy for degrading an anthro-pogenic substrate. Nat Chem Biol 5:727–733

van Pée K-H, Unversucht S (2003) Biological dehalogenation and halo-genation reactions. Chemosphere 52:299–312

Pieters RJ, Lutje Spelberg JH, Kellog RM, Janssen D (2001) Theenantioselectivity of haloalkane dehalogenases. Tetrahedron Lett42:469–471

Pries F, van den Wijngaard AJ, Bos R, Pentenga M, Janssen DB (1994)The role of spontaneous cap domain mutations in haloalkanedehalogenase specificity and evolution. J Biol Chem 269:17490–17494

Prokop Z, Opluštil F, DeFrank J, Damborský J (2006) Enzymes fightchemical weapons. Biotechnol J 1:1370–1380

Prokop Z, Sato Y, Brezovsky J, Mozga T, Chaloupkova R, KoudelakovaT, Jerabek P, Stepankova V, Natsume R, van Leeuwen JGE, JanssenDB, Florian J, Nagata Y, Senda T, Damborsky J (2010)Enantioselectivity of haloalkane dehalogenases and its modulationby surface loop engineering. Angew Chem Int Ed 49:6111–6115

Appl Microbiol Biotechnol

Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB,Lesniewski RA, Oakley BB, ParksDH, RobinsonCJ, Sahl JW, StresB, Thallinger GG, Van Horn DJ, Weber CF (2009) Introducingmothur: open-source, platform-independent, community-supportedsoftware for describing and comparing microbial communities.Appl Environ Microbiol 75:7537–7541

Schoemaker HE, Mink D, Wubbolts MG (2003) Dispelling the myths—biocatalysis in industrial synthesis. Science 299:1694–1697

Smith DRM, Grüschow S, Goss JM (2013) Scope and potential ofhalogenases in biosynthetic applications. Curr Opin Chem Biol 17:276–283

Steele HL, Jaeger K-E, Daniel R, Streit WR (2008) Advances in recoveryof novel biocatalysts from metagenomes. J Mol MicrobiolBiotechnol 16:25–37

Stucki G, Thüer M (1995) Experiences of a large-scale application of 1,2-dichloroethane degrading micro-organisms for groundwater treat-ment. Environ Sci Technol 29:2339–2345

Sunna A, Bergquist PL (2003) A gene encoding a novel extremely ther-mostable 1,4-β-xylanase isolated directly from an environmentalDNA sample. Extremophiles 7:63–70

Swanson PE (1999) Dehalogenases applied to industrial-scale biocataly-sis. Curr Opin Biotechnol 10:365–369

Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: molecular evo-lutionary genetics analysis (MEGA) software version 4.0. Mol BiolEvol 24:1596–1599

Tang K, Utairungsee T, Kanokratana P, Sriprang R, Champreda V,Eurwilaichitr L, Tanapongpipat S (2006) Characterization of a novel

cyclomaltodextrinase expressed from environmental DNA isolatedfrom Bor Khleung hot spring in Thailand. FEMS Microbiol Lett260:91–99

Uchiyama T, Miyazaki K (2009) Functional metagenomics for enzymediscovery: challenges to efficient screening. Curr Opin Biotechnol20:616–622

Uchiyama T, Watanabe K (2006) Improved inverse PCR scheme formetagenome walking. BioTechniques 41:183–188

Weerachavangkul C, Laothanachareon T, Boonyapakron K,Wongwilaiwalin S, Nimchua T, Eurwilaichitr L, Pootanakit K,Igarashi Y, Champreda V (2012) Alkaliphilic endoxylanase fromlignocellulolytic microbial consortium metagenome forbiobleaching of eucalyptus pulp. J Microbiol Biotechnol 22:1636–1643

Weigold P, El-Hadidi M, Ruecker A, Huson DH, Scholten T, JochmannM, Kappler A, Behrens S (2016) A metagenomic-based survey ofmicrobial (de)halogenation potential in a German forest soil. SciRep 6:28958

Westerbeek A, Szymański W, Wijma HJ, Marrink SJ, Feringa BL,Janssen DB (2011) Kinetic resolution of α-bromoamides: experi-mental and theoretical investigation of highly enantioselective reac-tions catalyzed by haloalkane dehalogenases. Adv Synth Catal 353:931–944

Westerbeek A, van Leeuwen JGE, Szymański W, Feringa BL, JanssenDB (2012) Haloalkane dehalogenase catalysed desymmetrisationand tandem kinetic resolution for the preparation of chiralhaloalcohols. Tetrahedron 68:7645–7650

Appl Microbiol Biotechnol