Bacterial versus fungal laccase: potential for micropollutant degradation

Relatively high concentrations of micropollutants in municipal wastewater treatment plant (WWTP) effluents underscore the necessity to develop additional treatment steps prior to discharge of treated wastewater. Microorganisms that produce unspecific oxidative enzymes such as laccases are a potential means to improve biodegradation of these compounds. Four strains of the bacterial genus Streptomyces (S. cyaneus, S. ipomoea, S. griseus and S. psammoticus) and the white-rot fungus Trametes versicolor were studied for their ability to produce active extracellular laccase in biologically treated wastewater with different carbon sources. Among the Streptomyces strains evaluated, only S. cyaneus produced extracellular laccase with sufficient activity to envisage its potential use in WWTPs. Laccase activity produced by T. versicolor was more than 20 times greater, the highest activity being observed with ash branches as the sole carbon source. The laccase preparation of S. cyaneus (abbreviated LSc) and commercial laccase from T. versicolor (LTv) were further compared in terms of their activity at different pH and temperatures, their stability, their substrate range, and their micropollutant oxidation efficiency. LSc and LTv showed highest activities under acidic conditions (around pH 3 to 5), but LTv was active over wider pH and temperature ranges than LSc, especially at near-neutral pH and between 10 and 25°C (typical conditions found in WWTPs). LTv was also less affected by pH inactivation. Both laccase preparations oxidized the three micropollutants tested, bisphenol A, diclofenac and mefenamic acid, with faster degradation kinetics observed for LTv. Overall, T. versicolor appeared to be the better candidate to remove micropollutants from wastewater in a dedicated post-treatment step.


Table of contents
1. Influence of the temperature on the pH -Correction of the activity to pH 4.5 (Fig. S1) 2 2. Fitting of a bi-exponential model to laccase activity stability (Table S1) 3 3. Fitting of a variable order reaction model to the oxidation of micropollutants (Table S2)  During the test to assess the influence of the temperature on laccase activity, the pH of the acetate buffer in the cuvettes decreased when the temperature increased, from pH 4.62 at 10°C to pH 4.05 at 70°C, following a linear relation (valid between T = 2 and 70°C, R 2 : 0.993): pH = -0.0099T (°C) + 4.715 (Fig. S1 a). The laccase activity with ABTS increased when the pH decreased from 5 to 4 ( Fig. S1 b and c). Therefore, to assess the temperature effect alone without the pH effect, the measured activity values (A pH ) were corrected to an equivalent activity at pH 4.5 (A 4.5 ) with the following relation:

Fitting of a bi-exponential model to laccase activity stability
The results of the laccase stability tests were fitted with a bi-exponential equation able to model various mechanisms of enzyme inactivation (Eq. 1) (Aymard and Belarbi 2000) by non-linear least squares regression using Matlab (MathWorks, USA), with A 0 and A t the activity at time 0 and at incubation time t respectively, a and b the pre-exponential factors, and k 1 and k 2 the apparent first order rate constants: The results of the fitting, the best-fit coefficients of the model and the estimated half-life of laccase at different pH are presented in Table S1. In pure water (both enzymes) and at pH 9 for L Sc , the inactivation followed a simple exponential decay, k 1 and k 2 being equal (Table S1). Except for pH 5, 6 and 7 for L Sc where the time series were too short to have confidence in the fitted model, a biexponential model was necessary to reproduce the behaviour observed.

Fitting of a variable order reaction model to the oxidation of micropollutants with laccase
The results (residual concentrations) of the micropollutant oxidation experiment were fitted by nonlinear least squares regression with a variable order reaction model (two coefficients, Eq. 2), as proposed by Margot et al. (2013), taking an initial concentration C 0 of 1 (arbitrary units as the initial concentration was always constant). C t is the residual concentration after a reaction time t, x the order of the reaction, and k the apparent variable order rate constant: The results of the fitting, the best coefficients of the model and the estimated half-life of the pollutants at different pH values are presented in Table S2. The order of reaction varied mainly between 1 and 3, as observed also by Margot et al. (2013).

Characterization of the commercial laccase preparation from Trametes versicolor from Sigma
The commercially available laccase preparation from Trametes versicolor obtained from Sigma (Ref. 38429) was analyzed by separating the proteins of 5 µl of concentrated laccase solutions (40 and 5 g l -1 ) by sodium dodecylsulfate polyacrylamide (12%) gel electrophoresis (SDS-PAGE), following Sambrook et al. (1989). The SDS-PAGE was done with and without 10 min boiling of the proteins. Prior to staining the proteins with Coomassie brilliant blue, one of the duplicate gels was incubated in acetate buffer 100 mM, pH 4.5, with 0.5 mM ABTS to detect the laccase activity.
As presented in Fig. S3, the commercially available laccase preparation contains a mixture of different proteins, from 17 to ~80 kDa, with a major band around 66 kDa, which corresponds approximately to the reported mass of the best-characterized T. versicolor laccase isoenzymes (Bourbonnais et al. 1995;Moldes et al. 2004). Similar protein bands from this laccase preparation were also observed by Wang et al. (2012). Despite the denaturing properties of the SDS gel, laccase activity was observed in at least two distinct bands in the gel with unboiled samples, around 40 kDa and 66-70 kDa, suggesting the presence of at least two enzymes with laccase activity in the preparation. The 40 kDa protein showed lower intensity with Coomassie staining but had high laccase activity, suggesting that this protein is thus either very active or more resistant to denaturation than the 66-70 kDa protein. These data show clearly that the commercially available laccase preparation contains a mixture of different proteins, several of which displaying laccase activity.

Comparison of "commercial" versus "in house-produced" Trametes versicolor laccases
Trametes versicolor is known to produce two main laccase isoenzymes with slightly different kinetic properties (Moldes and Sanromán 2006). The proportion of these two isoenzymes is reported to change depending on the growth substrate, especially in case of addition of lignocellulosic material (Moldes et al. 2004). Thus, the commercially available laccase preparation from Trametes versicolor obtained from Sigma (Ref. 38429) may not be fully representative of the laccase produced in a biofilter system with wood chips as the substrate/support. To assess if there was significant difference on micropollutant oxidation kinetics by both laccase preparations, we compared the oxidation kinetics of three micropollutants, bisphenol A (BPA), diclofenac (DFC) and mefenamic acid (MFA), by either the commercial laccase (from Sigma) or laccase produced on wood substrate.
Laccase production on wood substrate T. versicolor was grown in a glass column (used as a trickling filter) on oak wood by addition of mycelium inoculum on moistened autoclaved wood chips. Once the wood was completely colonized by the mycelium, a synthetic wastewater containing micro and macro nutrients (Borràs et al. 2008), 4 g l -1 of glucose and 10 mM MOPS buffer (pH 7), was filtered through the colonized wood chips as in a trickling filter. The water was continuously recirculated and laccase activity was regularly monitored. After 3 d of recirculation, when the activity reached 2000 U l -1 , the solution was filtrated at 0.22 µm and used as "produced on-site" laccase preparation.

Micropollutants oxidation essay
Oxidation of a mixture of three micropollutants, BPA, DFC and MFA, at 20 mg l -1 , was conducted as described in the main manuscript, in 20 mM citrate-phosphate buffer at two different pH values: 5.8 and 6.8. "Produced" or "commercial" laccase preparations were added to the reaction mixture at the same initial activity of 570 to 580 U l -1 . To have similar reaction mixture compositions between both experiments, the same amount of "produced" laccase preparation was also added, after heat inactivation, to the solution containing commercial laccase. Indeed, the "produced" preparation contained some organic substances leached from the wood substrate that may have an effect on the oxidation kinetics. Micropollutant concentration was then followed during 10 h as described in the main manuscript. Duplicate experiments were conducted at 25°C.

Results
As presented in Fig. S4, for both pH values tested, both laccase preparations had very similar oxidation kinetics for BPA and MFA, with no significant difference in the degradation rates. For DFC, the commercial laccase preparation was slightly less efficient at both pH values than the "produced" one, but with less than 10% difference in the removal rates. These very similar oxidation kinetics observed at two different pH values on three different micropollutants show that the commercial laccase preparation is representative, for micropollutant oxidation, of the laccase produced on wood substrate in a trickling filter.

Identification of one laccase candidate from S. cyaneus culture supernatant
Extracellular crude enzyme preparation of S. cyaneus culture supernatant was concentrated 80 times by ultrafiltration as described in the manuscript, and then separated by sodium dodecylsulfate polyacrylamide (12%) gel electrophoresis (SDS-PAGE) following Sambrook et al. (1989). A protein band around 75 kDa corresponding to the predicted S. cyaneus laccase molecular mass (Arias et al. 2003) was analysed by mass spectrometry (MS) after trypsin digestion and compared to profiles of peptides generated from available Streptomyces genomes and from the deposited S. cyaneus laccase sequence (GenBank HQ857207). This analysis was performed by the PCF laboratory (EPFL, Switzerland).
MS analysis of the excised 75 kDa protein band obtained after concentrating S. cyaneus culture supernatant showed a profile matching with nine unique peptides (20% coverage) of the deposited laccase sequence (GenBank HQ857207). This latter protein sequence shows 84% amino acid sequence identity with the phenoxazinone synthase (PHS) of S. antibioticus (Hsieh and Jones 1995) (see sequence alignment below). This laccase, along with several other Streptomyces proteins, form a distinct multi-copper oxidase family either classified as laccase (EC 1.10.3.2) or phenoxazinone synthase (EC 1.10.3.4). Functional differentiation between these two classes is unclear (Le Roes-Hill et al. 2009). The reported S. cyaneus laccase shows 33% sequence identity with the wellcharacterized CotA laccase of Bacillus subtilis (GenBank AAB62305) (Martins et al. 2002), and only very limited sequence identity with the EpoA laccase of S. griseus (GenBank BAB64332)  or the laccase of T. versicolor (GenBank CAA77015) (Fig. S7). Despite its relatively low sequence homology with other well-characterized laccases, the structure and active site configuration of S. antibioticus PHS, a close parent of S. cyaneus laccase, is reported to be very similar to other laccases, with three conserved cupredoxin-like domains, T1 (type 1 Cu centre) where the substrate oxidation takes place, and a trinuclear Cu cluster T2 and T3 where the electrons are transferred and where the reduction of oxygen to water take place (Enguita et al. 2003;Smith et al. 2006). Thus, similar catalytic mechanisms for these enzymes are expected. It is, however, important to mention that the S. cyaneus laccase activity was measured in the culture supernatant, which might also contain several other laccases not yet identified or reported in databases.

Comparison of sequences of S. cyaneus laccase and S. antibioticus phenoxazinone synthase
The amino acids sequences of S. cyaneus laccase (Scy-laccase, GenBank HQ857207) and S. antibioticus phenoxazinone synthase (San-PhsA, GenBank AAA86668) are presented below. A high sequence identity (84%, in black) exists between both enzymes. The residues that bind the different copper atoms (active sites T1 and T2, mononuclear, and T3, binuclear) are presented in color, following Smith et al.(2006). A fifth copper centre, not present in other laccases, was identified. This copper is thought to participate in the stability of the structure but not in the oxidation mechanisms (Smith et al. 2006). The proteins are separated into three main domains, presented with the dashed lines with different colours.