Received 11 April 2016; accepted 24 June 2016; published 28 June 2016

ABSTRACT

With the aim of identifying novel thermostable esterases, comprehensive sequence databases and cloned fosmid libraries of metagenomes derived from an offshore oil reservoir on the Norwegian Continental Shelf were screened for enzyme candidates using both sequence-and function-based screening. From several candidates identified in both approaches, one enzyme discovered by the functional approach was verified as a novel esterase and subjected to a deeper characterization. The enzyme was successfully over-produced in Escherichia coli and was shown to be thermostable up to 90˚C, with the highest esterase activity on short-chain ester substrates and with tolerance to solvents and metal ions. The fact that the thermostable enzyme was solely found by functional screening of the oil reservoir metagenomes illustrates the importance of this approach as a complement to purely sequence-based screening, in which the enzyme candidate was not detected. In addition, this example indicates the large potential of deep-sub-surface oil reservoir metagenomes as a source of novel, thermostable enzymes of potential relevance for industrial applications.

Keywords:

Metagenomics, Enzyme Discovery, Thermostable, Esterase

1. Introduction

The globe provides a large spectrum of environments with very diverse physical and chemical conditions for life to exist [1] . Some of these are considered extreme, providing for example high or low temperatures, pH or salt concentrations, or high pressure, radiation, as well as different combinations thereof. Deep sub-surface offshore oil reservoirs are poly-extreme environments due to a combination of high temperature, high pressure, high salinity, nutrient limitations, as well as the presence of toxic compounds like heavy metals and organic solvents [2] . Microorganisms populating these environments are generally considered (poly) extremophiles, attributing several unusual and interesting traits [3] . However, studies of these microbes are limited by difficulties in accessing relevant oil reservoirs, as well as technical challenges in representative oil reservoir sampling [4] [5] . Due to their nature, oil reservoir microorganisms with their special adaptations are potential high value targets for the discovery of useful biocatalysts for industrial processes [6] - [8] . By also covering the genetic information of the vast uncultivable fraction of microorganisms in nature [1] which is particularly pronounced in such special ecological niches, metagenomic libraries are an invaluable source for new discoveries by bioprospecting, using both metagenomic sequence data and cloned metagenomic DNA [9] [10] . Sequence-based bioprospecting using metagenomic DNA sequence databases has benefits in being fast and inexpensive to perform, and is independent of functional expression in a heterologous host and functional assays. However, its dependence on previously described sequence information limits its potential to discover entirely new enzymes. In contrast, function-based screening for activities of interest using cloned metagenomic libraries is capable of discovering novel enzymes based on the function, independent of sequence similarity to previously known genes. It represents therefore a valuable complement to sequence data mining, despite of its dependence on assay methods and vector-host systems for functional gene expression.

Esterases are enzymes with several applications in industrial processes, and novel variants of these providing special properties are still in demand [11] . They are today commercially used e.g. in the food industry, for animal feed and biofuel production, as well as detergents. New candidates with high tolerance to elevated temperature and pressure, stability and activity in organic solvents, and/or special substrate spectra may prove useful in these, or other applications [12] . In the current study, we have screened oil reservoir metagenomes [13] [14] for esterase enzyme candidates using both sequence-based and functional screening, and have subjected one candidate from the functional screening approach to further characterization. Our results document oil reservoir metagenomes as a valuable source of novel thermostable enzymes of potential industrial relevance.

2. Materials and Methods

2.1. Sample Collection, DNA Isolation and Fosmid Library Generation

Oil reservoir samples were collected and processed as described earlier [13] [14] . Metagenomic DNA was isolated [14] from the sample material and used for direct high throughput sequencing (water phases of Wells I and II of [14] ), as well as whole genome amplification (WGA) using Qiagen REPLi-g kit according to the manufacturer’s protocol (water phase of Well II of [14] ) and both water and oil phase of a third, nearby production well (Well III) that, unlike Wells I and II, has experienced seawater breakthrough. Amplified DNA was cloned as fosmid libraries using the pCC1fos fosmid library kit from Epicentre according to the manufacturer’s instructions [15] . Library clones of each sample were selected for on square (215 ml volume) LB agar plates containing 12.5 µg/ml chloramphenicol, pooled and stored in 25% glycerol at −80˚C. Based on colony counts, the total library harbors approximately 77,000 independent fosmid clones. Prior to functional screening, library pools from the different samples were again plated onto the same type of selective agar plates (LB-agar + 12.5 µg/ml chloramphenicol), with 350 - 1500 CFU per plate. Proportional to the number of total clones represented in each library, the corresponding number of agar plates and 384 multi-well plates were used for the different samples. In total 20 agar plates were processed; 14 with Well II water phase library clones, 4 with Well III water phase library clones and 2 with Well III oil phase library clones. Correspondingly, 21 individual 384 well plates were inoculated with Well II water phase library clones, 6 plates with Well III water phase clones and 3 plates with Well III oil phase clones. Clones within the well plates were grown in 80 µl reduced Hi-YE medium per well containing 12.5 µg/ml chloramphenicol (37˚C, 900 rpm, 85% humidity, 20 hours), plates were duplicated and then stored at −80˚C with 25% final concentration glycerol.

2.2. Sequence Mining

Twenty Pfam [16] profiles for known lipases and esterases were used as input to tBLASTn searches [17] against ORFs of the metagenomic DNA libraries, using an E-value cutoff of 0.00001 and a minimal overlap of 50% on the Pfam motif. In addition, ORFs of the metagenomic libraries were classified according to EC numbers using tBLASTn with exemplar sequences derived from the BRENDA database [18] .

2.3. Functional Screening (HTS)

Approximately 11,520 fosmid clones carrying metagenomic DNA sequences originating from oil reservoir Well II (water phase sample) [14] and Well III (oil and water phase) were screened for lipase and/or esterase activities. Functional screening (schematically described in supplementary Figure S1) was performed using liquid culture extracts in 30 individual 384 well plates. Each of the 30 384 well plates of one stored replica set from library construction was used to inoculate 4 individual 96 well plates, in which the clones were grown for 20 hours under the same conditions as described above and subsequently the increase of fosmid copy number induced for 6 hours by addition of arabinose from a 10% (w/v) stock solution (0.01% final conc.). Cultivation plates were then stored at −80˚C until extraction. After thawing, crude cell extracts were prepared by adding 30 µl B-PER II (Thermo Scientific, prod# 89822) to each well, followed by an incubation at 25˚C, 700 rpm, 85% humidity for 1 hour. Cell debris was pelleted by centrifugation (4000 rpm, 30 minutes, 4˚C). Supernatants (extracts) were transferred in duplicates to 384 well plates and used in enzyme assays (Figure S1).

High throughput screening for short-chain esterase activity was performed in 384 well plates by addition of 5 µl crude cell extract to each well containing 40 µl reaction mixture (0.17 mM 4-nitrophenyl acetate substrate (Sigma-Aldrich, no. N8130) in 0.1 M phosphate buffer, pH 8, containing 4.4% EtOH), shaking of the plates (1600 rpm, 30 s) and absorption measurement at 410 nm immediately and after 100 min incubation at 37˚C. For screening for long-chain esterase activity, the same procedure was applied, with the exception of the reaction mixture containing 0.17 mM 4-nitrophenyl palmitate (Sigma-Aldrich, no. N2752), 4.4% EtOH and 1% Triton X-100 in 0.1 M phosphate buffer, pH 8, and performing absorption measurements at time points 0, 1 and 3 hours after 50˚C incubation.

2.4. Analysis of Positive Clones

Esterase positive clones identified in functional screening were subjected to fosmid isolation using the Promega Wizard Miniprep kit (arabinose induced cultures), and the genes of interest were identified by sequencing of respective fosmid clones and sub-cloning of candidate genes. Gene sequences were analyzed using BLASTx against the NCBI database. Amino acid sequences were analyzed on the Phylogeny.fr web server [19] , using T-Coffee [20] for multiple alignment, PhyML [21] for generating the phylogenetic tree (after removing all columns with gaps), and TreeDyn [22] for drawing the tree. Branch support in the phylogenetic tree was estimated by the approximate likelihood ratio test (aLRT) [23] . The multiple alignment was visualized using Jalview [24] . Pairwise comparison of protein sequences was done with pairwise BLAST.

2.5. Enzyme Production

The esterase candidate genes were amplified using PCR from the relevant fosmid (primer sequences GGGGCC- ATGGGAGATAAGGAGGAGGG and GGGGATCCAAAGATAGAGGAGCAGATC for His-tagged enzyme, and sequences ATAAGGAGGAGGGCATATGGCTGA and GGGGATCCAAAGATAGAGGAGCAGATC for non-tagged enzyme), or synthesized as E. coli codon optimized versions (GenScript, Piscataway, NJ, USA) and sub-cloned in expression vector pET16b (with and without hexahistidyl-tag; however, only untagged protein was analyzed as enzyme isolation by heat incubation proved to be feasible). Enzymes were produced in E. coli BL21 (DE3) REPL (codon+) grown in shake flasks in 50 ml LB containing 100 µg/ml ampicillin. Gene expression was induced at OD₆₀₀ = 1.0 using IPTG (0.5 mM final conc.). After continued incubation at 37˚C until observed OD₆₀₀ decrease after approx. 7 h, the cultures were harvested by centrifugation (4000 rpm, 15 minutes, 4 ˚C) and cell pellets stored at −20˚C until extraction. Crude cell extracts were prepared by sonication in 2 ml 0.1 M Tris-HCl buffer, pH 8, per 1 g wet weight biomass (flat tip, duty cycle 50%, output control 4; 7 × 1 min). Cell debris was pelleted by centrifugation (20 000 x g, 20 min, 4˚C) and supernatant stored at −20˚C.

Batch cultivation for the production of esterase FS02 was performed in two 3 l Applicon fermentor vessels using 1.8 l 3xLB medium (Tryptone 30 g/l, Yeast extract 15 g/l, NaCl 10 g/l) each, containing 100 µg/ml ampicillin, with constant airation at 0.25 VVM (0.45 l/min) and 1000 rpm fixed stirrer speed. The cultivation medium was inoculated with 1.5% (27 ml) overnight culture (LB, 100 µg/ml ampicillin). Gene expression was induced at OD₆₀₀ = 4 using (3 hours cultivation) 0.5 mM IPTG. Cultures were thereafter harvested after 6 hours of induction and thereafter treated as described above to obtain crude cell extracts. Crude cell extracts were heat treated at 65˚C for 30 minutes, centrifuged at top speed (13 000 x g) in a table top centrifuge for 5 minutes to precipitate denatured host proteins, aliquoted and stored at −20˚C.

2.6. Enzyme Characterization

Verification of lipase activity. Crude cell extracts were used to assess lipase enzyme function using the same assay procedure as applied for esterase screening in multi-well plates, however, up-scaled to 1 ml volume for cuvette measurements. To 880 µl master mix (4.4% EtOH and 1% Triton X-100 in 0.1 M phosphate buffer, pH 8), 20 µl substrate solution (20 mM 4-nitrophenyl palmitate in acetonitrile) was added, incubated at 37˚C for 30 s prior to addition of 100 µl cell extract and absorbance measurement at 420 nm every 2 seconds for 4 minutes. End point measurements were performed using the same setup but with measurements after 30, 60 and 120 minutes, with assay and incubations done at 37˚C.

Heat stability. Temperature stability of enzymes was determined by incubating crude extracts at elevated temperatures (65˚C - 100˚C) for different time periods (15 min - overnight), followed by SDS-PAGE and/or activity assay as described above. Thermal unfolding transition midpoint Tm of the enzyme of interest was analyzed using the Thermal Shift assay from Applied Biosystems on a 7500 rtPCR machine according to the manufacturer’s instructions.

pH optimum. The pH optimum of the enzyme of interest was analyzed by performing the esterase activity assay at 37˚C as described above but using buffer with different pH (0.1 M glycine buffer pH 2 - 3 and pH 9 - 12; 0.05 M citrate buffer pH 5, 0.1 M phosphate buffer pH 6 - 8).

Substrate spectrum. Enzyme activity on a spectrum of commercial ester substrates of different chain lengths (C2-C18; p-nitrophenyl acetate, p-nitrophenyl-butyrate, p-nitrophenyl-decanoate, p-nitrophenyl-myristate, p-nit- rophenyl-palmitate and p-nitrophenyl-stearate) was analyzed using the assay procedure describe above.

Effect of selected additives on enzyme activity. Esterase activity in the presence of different additives was assayed as described above using p-nitrophenyl-butyrate as the substrate. Chemicals added were; metal ions (1 mM final concentrations from CaCl₂∙2H₂O, CoSO₄∙7H₂O, CuCl₂∙2H₂O, FeCl₃, MgCl₂∙6H₂O, MnCl₂∙4H₂O, RbCl, ZnCl₂, KCl, NaCl and Li-acetate*2H₂O salt solutions, respectively), isopropanol, ethanol, methanol, acetone, acetonitrile, DMF and DMSO (1%, 10% and 30% each), Triton X-100, Tween 80 and SDS 5 (1% and 5% each) as well as DTT and EDTA (1, 5 and 10 mM each).

3. Results and Discussion

3.1. Homology-Based Sequence Data Mining for Lipase/Esterase Enzyme Candidates

The established metagenomic sequence databases [13] [14] were screened for lipase enzyme candidates using 1) BLAST searches with Pfam profiles for lipases and esterases and 2) screening by a local ORF database classified by EC-number. Using Pfam profiles against the Well I and Well II metagenomics databases, 72 and 55 candidates, respectively, were identified and in total 24 candidates were found associated with E.C. number 3.1.1.3 after matching ORFs against BRENDA exemplar sequences and requiring >60% sequence identity and >80% coverage (Supplementary Table S1). The list of candidates from sequence mining using the Pfam strategy contained candidates with very low sequence identity to Pfam motif (annotated lipases/esterases/hydrolases). In total only 19 candidates had sequence identity higher than 30% (37.5% as highest). increasing the probability of false positive predictions. However, the top five candidates (37.5% - 35.4% id) were used in a BLASTx search, and were indeed matching esterases, hydrolases and lipolytic proteins with fairly high sequence identity (>80%) for most, but with some exceptions having sequence identity as low as 40% - 50%. As examples, two gene candidates from sequence mining had sequence id of 78% against alpha/beta hydrolase fold (Campylobacterales bacterium GD) and 62% against biotin biosynthesis protein (Pelobacter carbinolicus DSM 2380), respectively (in addition verified using latest version of Pfam, resulting in abhydrolase 6, alpha/beta hydrolase family (e-value 2e-19), and abhydrolase 1, alpha/beta hydrolase fold (e-value 7e-25), respectively). These candidates were synthesized as codon optimized versions (GenScript) and produced in E. coli, however showing non-con- clusive data in activity analysis and only moderate thermostability (data not shown). These results illustrate a key challenge of sequence mining to reliably identify candidates with the predicted enzyme activity. In the light of the obtained results, these candidates were not analyzed further. Therefore, with the purpose of finding esterase enzymes with high novelty and functionality, focus was turned to functional screening of the constructed metagenome libraries.

3.2. Functional Screening for Lipases and Esterases

Functional screening for long-and short-chain esterases encoded in the 11,520 arrayed fosmid clones resulted in 13 positive hits from the initial screen. For identification of the genes of interest within the respective fosmids, fosmid DNA was isolated and subjected to a combination of end-sequencing and primer walking, supplemented by 454 pyrosequencing of pooled fosmid clones. For the end-sequencing and primer walking approach, technical difficulties appeared, and sequencing results could not be correlated with fosmid size determinations by gel electrophoresis (data not shown). After thorough evaluation of the fosmid library construction procedure, it was concluded that the encountered problems likely were associated with the amplification of the metagenomic DNA. Due to the low amount of DNA that could be isolated from the oil reservoir samples, target DNA was amplified using WGA based on Multiple Strand Displacement Amplification (MDA) [25] prior to fosmid cloning. MDA has been reported earlier to result in chimeric sequences [26] , and for long insert cloning (such as fosmids of around 35 kilobases in size), repetitive and chimeric DNA fragments can potentially be expected. The obtained sequencing results suggest frequent occurrence of chimeric repeats with multiple primer binding sites within the cloned inserts during end-sequencing/primer walking, hampering elucidation of the complete sequence information contained in the cloned DNA. Nevertheless, some genes potentially associated with the detected esterase phenotype could still be identified. One of these genes was found to encode a previously described lipase of Archaeoglobus fulgidus [27] which was not chosen for further analysis due to lack of novelty, as well as two other candidate genes (FS01 and FS02, with annotations as hypothetical proteins). These could be associated with strong esterase and lipase phenotypes and were therefore chosen for sub-cloning, expression from pET16b in E. coli and then subjected to further characterization. Respective crude cell extracts were produced and analyzed by esterase activity assays. Sub-cloned candidate FS01 did not show any esterase activity, whereas the FS02 clone, despite of low enzyme production levels, clearly exhibited such an activity (Figure S2). FS02 was in addition found to be highly thermostable, as incubation at 85˚C for 30 minutes led only to a small decrease in activity compared to the not heat treated control (Figure S2).

3.3. Analysis of the FS02 Gene

The nucleotide sequence of the FS02 gene was used in BLASTx searches against the NCBI protein database, resulting in hits against hypothetical proteins or hydrolases of the alpha/beta superfamily (Table S2). The closest match was against a hypothetical protein encoded in the Archaeoglobus fulgidus genome [27] , showing 99% sequence identity (1 amino acid mismatch) by 100% coverage. Further close homologues were a hypothetical protein from Ferroglobus placidus (60% identity, 97% coverage), a putative hydrolase of the alpha/beta superfamily from Archaeoglobus sulfaticallidus (55% identity, 96% coverage) and a hypothetical protein from Archaeoglobus veneficus SNP6 (52% identity and 99% coverage). For further evaluation of novelty, the FS02 amino acid sequence was combined with known lipase/esterase sequences of several Archaea and Bacteria (Table S3) in a multiple sequence alignment (Figure S3) in order to build a phylogenetic tree (Figure 1). This tree confirms the initial picture that the FS02 enzyme is closely related to predicted proteins from Archaea, but similar proteins are also found in Bacteria. Based on the functional confirmation of FS02 as an esterase, it is now possible to assign putative esterase function to these hypothetical proteins, demonstrating the strength of functional screening of metagenomes in elucidating function of the high number of genes of yet unknown function.

3.4. Production of FS02 Enzyme and Enzyme Characterization

Due to the apparently low production levels of FS02 enzyme in E. coli, the FS02 gene was synthesized in an E. coli codon usage optimized version (GenScript), cloned in the T7 expression vector pET16b and expressed in E. coli. Enzyme production based on the new gene version was found to be significantly higher compared to the

native gene sequence (data not shown). FS02 thermostability (Figure S4) and activity were verified, and a larger batch of enzyme was produced to allow multiple analysis of the enzyme from the same batch (3 - 4 g wet cell mass/fermentor).

Batch-produced FS02 was successfully tested for heat stability by incubation at 65˚C for 30 minutes, thus leading to an FS02 enriched crude extract for subsequent studies. The enzyme’s temperature stability was analyzed in detail by incubation at different temperatures for various periods of time, demonstrating that FS02 is thermostable up to 90˚C for at least 1 hour without any notable decrease in activity, whereas a 100˚C incubation resulted in rapid inactivation of the enzyme (Figure 2(a)). Longer incubation at 90˚C resulted in lowered residual activity, however, high enzyme activity could still be detected after incubations at up to 80˚C for as long as 5 hours (Figure 2(a)), indicating that FS02 is a highly thermostable enzyme. Tm values were determined by Thermal Shift, resulting in values of Boltzman Tm (TmB; Figure 2(b)) that were consistent with observed residual enzyme activities at high temperature. When comparing Tm values determined under different pH conditions, it could be concluded that the chosen pH range did not affect heat stability at notable levels (Figure 2(b)).

In attempts to analyze the pH optimum of FS02, it was found that enzyme activity was high at pH 11, however, assay background was high as well at this pH (data not shown), and therefore pH 9 was chosen as standard pH in all subsequent additional enzyme characterizations. The substrate spectrum analysis of FS02 revealed a higher enzymatic activity on short-chain substrates compared to long-chain esters (Figure 3). Enzyme activity was highest using p-nitrophenyl acetate as substrate.

Despite relatively high standard deviations in replicate measurements (Figure 4) it was found that enzyme activity on this substrate was little affected by a number of different additives, such as Ca²⁺, Cu²⁺, Na⁺, Li⁺, Co²⁺, Fe³⁺ and Mg²⁺ ions, as well as by addition of 5 to 10 mM DTT (Figure 4). Ions like Mn²⁺, Rb²⁺, Zn²⁺ and K⁺ however, led to a notable increase in activity up to 5-fold. Stimulation and inhibitory effects of different metal ions on esterase activity have been observed before [28] , including an increased activity in the presence of Zn²⁺ and a decreased activity in the presence of e.g. Ca²⁺, Mn²⁺ and Na⁺. Addition of low concentrations of detergent (1% Triton X-100 or SDS) was found to increase activity, whereas a higher concentration of the same compounds (5%) resulted in loss of activity, which is consistent with observations with other esterases (e.g. [29] ). Presence of solvents such as ethanol, methanol, DMF, acetonitrile, acetone and isopropanol apparently led to significant increases in activity (Figure 4). Lipase activity in the presence of different solvents has been analyzed in several previous studies, mainly showing no or negative effects; e.g. unaffected activity with methanol, ethanol and acetone, and a decreased lipase activity in the presence of isopropanol [30] , or reduced lipase activity in the presence of isopropanol, methanol, acetone, acetonitrile and/or ethanol [27] [29] . On the other hand,

similar to the observed activity increase of FS02 in the current work, enzymatic activity of an alkaline lipase was shown to be increased in the presence of e.g. methanol [31] , and high solvent tolerance of a Pseudomonas lipase has been shown [30] . Hence enzyme activity increase in the presence of solvents observed for FS02 is not unique; however, despite of some uncertainties concerning the absolute activity levels from high standard deviations in the analyses, the increases in activity observed for enzymes in previous studies are rarely as high as observed for FS02, rendering the newly discovered enzyme with distinct features compared to other similar enzymes.

4. Conclusions

Esterases are used in several industrial processes [1] [32] , and many of these enzymes are well characterized and applied in different processes [33] [34] . However, the discovery of novel esterases with beneficial properties, e.g. high thermo-and solvent-stability, might still serve current and potential new uses in e.g. the food, polymer, pharmaceutical and fine chemical industries [32] [34] . Metagenome-based bioprospecting is a powerful tool for biodiscovery of novel enzymes for various applications, and metagenomic libraries from (poly) extreme environments might be a particularly valuable source of enzymes with unique features. Several lipase/esterase enzyme candidates were identified from the oil reservoir-originating metagenomes, both using homology-based sequence mining and functional screening. The latter helps assigning function to the vast pool of hypothetical proteins encoded and annotated in genomes and metagenomes.

One lipase/esterase candidate from functional screening, denoted FS02, was found to be active, highly thermostable, as well as showed high activity in the presence of various compounds, generating noteworthy results (Figure 4) on the enzyme’s tolerance to e.g. solvents, potentially relevant for some industrial applications. Solvent tolerance is often observed for lipases (e.g. [31] ), but the large stimulation of enzyme activity appears rare. However, dependent on a potentially envisioned industrial application, a more profound investigation of these properties will be necessary.

The discovery of an active and thermostable esterase in a metagenomic library originating from a deep sub- surface oil reservoir illustrates the powerful approach of functional metagenomics in search for novel enzymes with desired properties from poly-extreme environments. It also demonstrates the potential for biodiscovery from such environments as deep subsurface oil reservoirs. The finding of the described thermostable esterase represents only the top of the ice berg of what potentially can be found by harvesting and screening metagenomes from this special ecological niche.

Acknowledgements

This work is supported by the Research Council of Norway (grant numbers 187317/S30, and 208541/O10), Statoil ASA and SINTEF.

Cite this paper

Anna Lewin,Trine Aakvik Strand,Tone Haugen,Geir Klinkenberg,Hans Kristian Kotlar,Svein Valla,Finn Drabløs,Alexander Wentzel, (2016) Discovery and Characterization of a Thermostable Esterase from an Oil Reservoir Metagenome. Advances in Enzyme Research,04,68-86. doi: 10.4236/aer.2016.42008

References

Appendix

Submit your manuscript at: http://papersubmission.scirp.org/

NOTES

^*Corresponding author.