University of Groningen Exploring the Biochemical and Biocatalytic

University of Groningen

Exploring the biochemical and biocatalytic properties of bacterial DyP-type peroxidases Colpa, Dana Irene

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2018

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Colpa, D. I. (2018). Exploring the biochemical and biocatalytic properties of bacterial DyP-type peroxidases. Rijksuniversiteit Groningen.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 03-10-2021 Exploring the biochemical and biocatalytic properties of bacterial DyP-type peroxidases

Dana I. Colpa The research described in this thesis was carried out in the department of Molecular Enzymology of the Groningen Biotechnology and Biomolecular Sciences Institute (GBB) of the University of Groningen, according to the requirements of the Graduate School of Science, Faculty of Mathematics and Natural Sciences, University of Groningen, The Netherlands.

This work was financially supported by The Netherlands Organisation for Scientific Research (NWO) via the graduate program: Synthetic Biology for Advanced Metabolic engineering, project number 022.004.006.

Cover: The cover shows an artistic representation of the decolorization of dyes by a DyP-type peroxidase.

Cover design: Dana I. Colpa and Alexander Ziel Printed by: Ipskamp Drukkers, Enschede (The Netherlands)

ISBN: 978-94-034-1070-8 (printed version) ISBN: 978-94-034-1069-2 (electronic version)

Copyright © D.I. Colpa 2018, Groningen, The Netherlands. All rights reserved. No part of this thesis may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without prior permission of the author. Exploring the biochemical and biocatalytic properties of bacterial DyP-type peroxidases

Proefschrift

ter verkrijging van de graad van doctor aan de Rijksuniversiteit Groningen op gezag van de rector magnificus prof. dr. E. Sterken en volgens besluit van het College voor Promoties.

De openbare verdediging zal plaatsvinden op

vrijdag 26 oktober 2018 om 11.00 uur

door

Dana Irene Colpa

geboren op 13 september 1988 te Assen Promotor Prof. dr. ir. M.W. Fraaije

Beoordelingscommissie Prof. dr. D.B. Janssen Prof. dr. G. Maglia Prof. dr. L.O. Martins Table of contents

Chapter 1 Introduction and outline of the thesis 7 DyP-type peroxidases: a promising and versatile class of enzymes

Chapter 2 Exploring the biocatalytic potential of a DyP-type 25 peroxidase by profiling the substrate acceptance of Thermobifida fusca DyP peroxidase

Chapter 3 Exploring the catalytic properties of DyP-type 43 peroxidase TfuDyP by site-directed mutagenesis

Chapter 4 High overexpression of dye decolorizing peroxidase 69 TfuDyP leads to the incorporation of heme precursor protoporphyrin IX

Chapter 5 Creating oxidase-peroxidase fusion enzymes as tool- 85 box for cascade reactions

Chapter 6 Conclusions and future perspectives 101

Chapter 7 Nederlandse samenvatting 111

Appendices List of publications 123

Acknowledgements/Dankwoord 124

Chapter 1:

Introduction and outline of the thesis

DyP-type peroxidases: a promising and versatile class of enzymes

Dana I. Colpa, Marco W. Fraaije and Edwin van Bloois

This chapter is based on: Journal of Industrial Microbiology and Biotechnology (2014) 41: 1–7. DOI: 10.1007/s10295-013-1371-6

DyP-type peroxidases

Introduction Peroxidases (EC 1.11.1.x) represent a large family of oxidoreductases that typically use hydrogen peroxide as an electron acceptor to catalyze the oxidation of substrate molecules. The vast majority of these enzymes contain heme as a cofactor1 and are ubiquitously present in prokaryotes and eukaryotes. Peroxidases take center stage in a variety of biochemical processes, ranging from the biosynthesis of cell wall material to immunological host-defense responses.2,3 Heme-containing peroxidases were originally classified into two superfamilies: the plant peroxidases and the animal peroxidases.4,a Remarkably, some members of the peroxidase superfamily have been studied for more than a century like horseradish peroxidase (HRP), a prototype plant peroxidase.5 In this respect, it was fascinating that the first member of a newly discovered peroxidase superfamily, the group of DyP-type peroxidases, was described in 1999.6 In this chapter, we discuss the biochemical and structural features of DyP- type peroxidases as well as their promising biotechnological potential.

Phylogenetic and structural comparison Dye-decolorizing (DyP-type) peroxidases were first discovered in fungi and named after their ability to degrade a wide range of dyes.6 Subsequently, additional members were found in the proteomes of other fungi as wellas in several bacteria.7 This indicates that these enzymes are widespread like other peroxidases. Interestingly, recent genome sequence analysis revealed that these enzymes are prominent in bacteria, whereas only a small number is found in fungi and higher eukaryotes. Their occurrence in archaea is even more limited. The most comprehensive overview of the DyP-type peroxidase superfamily is offered by the InterPro database.8 According to this database, the DyP superfamily currently (February 2018) comprises 12,670 enzymes of which 11,877 are found in bacteria, 741 in eukaryotes, and 52 in archaea. Additionally, DyP-type peroxidases are, according to PeroxiBase, further sub-classified into the phylogenetically distinct classes A, B, C, and D.9 Alternative classifications a More recently heme-containing peroxidases were reclassified in four independently evolved superfamilies: the peroxidase-catalase superfamily, the peroxidase-cyclooxygenase superfamily, the peroxidase-chlorite dismutase superfamily and the peroxidase-peroxygenase superfamily.78 The peroxidase-catalase superfamily is formed by members of the previously called ‘superfamily of bacterial, fungal and plant heme peroxidases’. The ‘superfamily of the animal heme-dependent peroxidases’ contains enzymes from all kingdoms of life and was renamed to the peroxidase- cyclooxygenase superfamily. The peroxidase-chlorite dismutase superfamily consists of three protein families that share a common fold: DyP-type peroxidases, chlorite dismutases and EfeB, previously called the CDE-superfamily79. In this chapter the original classification will be used.

9 Chapter 1

with three and five classes have also been proposed.10,11 Subfamilies C and D are grouped together in the classification with three classes. An overview of the DyP- type peroxidases characterized thus far is shown in Table 1 and a phylogenetic tree is shown in Figure 1. Many of the potential bacterial enzymes are putative cytoplasmic enzymes (class B and C), indicating that they are involved in intracellular metabolism. In contrast, enzymes belonging to class A contain a Tat-dependent signal sequence, which suggests that they function outside of the cytoplasm or extracellularly as previously confirmed by us and others.12–14 Class D contains primarily fungal variants. For some of these peroxidases, it has been shown that they are involved in dye decolorization.7 Nevertheless, the physiological function of the majority of DyP-type peroxidases is at present unclear, although evidence is accumulating that some bacterial variants are involved in the degradation of lignin.15–19 This suggests that these enzymes can be regarded as the bacterial equivalents of the fungal lignin degrading peroxidases.

38 DyPs DdDyP YfeX VcDyP ElDyP TyrA PflDyP1B DyPB PpDyP DyPPa PflDyP2B BtDyP MtDyP BlDyP

SaDyP B DyP2 AnaPX C A BsDyP GlDyP DyPA Ftr-DyP PsaDyP TceDyP_4GS1 D PoDyP TfuDyP TAP ScoDyP_4GT2 I. lacteus DyP DtpA DyP AauDyPI TcDyP PflDyPA EfeB MepDyP EglDyP SviDyP Msp1 Msp2

Figure 1. Phylogenetic tree of the DyP-type peroxidases characterized thus far (February 2018). The sequence alignment and phylogenetic tree were made by Geneious version 8.1.9 using ClustalW alignment with GONNET as cost matrix for the sequence alignment and neighbor-joining as tree build method.

10 DyP-type peroxidases

Table 1. DyP-type peroxidases characterized thus far (February 2018).

Class Protein name Organism PDB UniProtKB code Ref. A BsDyP/YwbN Bacillus subtilis P39597 [12,20] DtpA Streptomyces lividans TK24 5MJH A0A076MAJ9 [21,22] DypA Pseudomonas fluorescens Pf-5 Q4KBM1 [17] DypA Rhodococcus jostii RHA1 Q0S4I5 [16] EfeB/YcdB Escherichia coli O157 2Y4F P31545 [23] SviDyP Saccharomonospora viridis DSM 43017 C7MS11 [24] TcDyP Thermomonospora curvata 5JXU D1A807 [25] TfuDyP Thermobifida fuscaYX 5FW4 Q47KB1 [14,26] (ScoDyP) Streptomyces coelicolor ATCC BAA-471 4GT2 Q9ZBW9 - (TceDyP) Thermobifida cellulosilityica 4GS1 U3KRF5 - B BlDyP Brevibacterium linens M18 NCBI ref: [27] WP_101555111.1 BtDyP Bacteriodes thetaiotaomicron 2GVK Q8A8E8 [28] DdDyP Dictyostelium discoideum Q556V8 [29] DypB Rhodococcus jostii RHA1 3QNR Q0SE24 [16,30] Dyp1B Pseudomonas fluorescens Pf-5 Q4KAC6 [17] Dyp2B Pseudomonas fluorescens Pf-5 Q4KA97 [17] DyPPa Pseudomonas aeruginosa PKE117 D5LRR6 [31] ElDyP Enterobacter lignolyticus 5VJ0 E3G9I4 [32] MtDyP Mycobacterium tuberculosis H37Rv I6Y4U9 [33] PpDyP Pseudomonas putida Q88HV5 [34] TyrA Shewanella oneidensis 2HAG Q8EIU4 [28,35] VcDyP Vibrio cholerae 5DE0 Q9KQ59 [36] YfeX Escherichia coli O157:H7 str. Sakai 5GT2 P76536 [37] C AnaPX Anabaena sp. PCC 7120 5C2I Q8YWM0 [38] DyP2 Amycolatopsis sp. 75iv2 4G2C K7N5M8 [39] SaDyP2 Streptomyces avermitilis Q82HB1 [40] D AauDyPI/AjPI Auricularia auricula-judae 4AU9 I2DBY1 [41,42] DyP/BadDyP Bjerkandera adusta Dec 1 2D3Q Q8WZK8 [6,43] EglDyP Exidia glandulosa I2DBY2 [44] Ftr-DyP Funalia trogii (Coriolopsis trogii) GenBank: [45] AUW34346.1 GlDyP Ganoderma lucidum G0X8C9 [46] I. lacteus DyP Irpex lacteus A0A1R7T0P5 [47] MepDyP Mycena epipterygia I2DBY3 [44] MsP1/MscDyP1 Mycetinis scorodonius B0BK71 [48] MsP2/MscDyP2 Mycetinis scorodonius B0BK72 [48] PoDyP Pleurotus ostreatus Q0VTU1 [49] (r)PsaDyP Pleurotus sapidus A0A0F7VJ89 [50] TAP Termitomyces albuminosus Q8NKF3 [51]

11 Chapter 1

DyP-type peroxidases are unrelated at the primary sequence level to peroxidases of the plant and animal superfamilies. They also lack the typical heme-binding motif of plant peroxidases, comprising one proximal histidine, one distal histidine, and one crucial arginine (Fig. 2).3,5,7 Moreover, DyP-type peroxidases and plant peroxidases both bind the heme cofactor non-covalently, unlike animal peroxidases, which bind the heme cofactor covalently (Fig. 2).52 All DyP-type peroxidases contain the so-called GXXDG-motif in their primary sequence, which is part of the heme-binding region. This motif is important for peroxidase activity because replacement of the conserved aspartate by an alanine or asparagine inactivates the enzyme, while heme-binding is not affected.14,43 Based on these results, it was proposed that the conserved aspartate of the GXXDG-motif is functionally similar to the distal histidine of plant peroxidases.1,3 However, the catalytic role of this conserved aspartate was put into question by a recent study. It was shown that substitution of the aspartate of the GXXDG- motif ofEscherichia coli EfeB/YcdB by an asparagine only marginally affected the peroxidase activity of this enzyme.23 A limited number of fungal and bacterial DyP-type peroxidases have been characterized in some detail, including elucidation of their crystal structures, see Table 1. While DyP-type peroxidases from the different subclasses often exhibit a remarkable low sequence similarity, their overall structural topology is highly conserved. Structurally, DyP-type peroxidases comprise two domains that contain α‑helices and anti-parallel β-sheets, unlike plant and mammalian peroxidases that are primarily α-helical proteins (Fig. 2).5,52 Both domains in DyP-type peroxidases adopt a unique ferredoxin-like fold and form an active site crevice with the heme cofactor sandwiched in between. The heme-binding motif contains a highly conserved histidine in the C-terminal domain ofthe enzyme (Fig. 2), which seems to be an important heme ligand and is therefore functionally similar to the proximal histidine of plant peroxidases.23,28,35,43 To test the role of the proximal histidine of DyP-type peroxidases as a heme ligand, we replaced this residue by an alanine in TfuDyP from Thermobifida fusca. This resulted in a loss of heme, which demonstrates that this residue is indeed an important heme ligand of DyP-type peroxidases.14 In addition, fungal DyP-type peroxidases also contain a conserved histidine in the N‑terminal domain of the enzyme, which was previously assigned as heme ligand.53 However, this residue does not contribute to heme binding according to the available structures.43 Clearly, more structural studies are required to unveil the molecular details by which DyP-type peroxidases catalyze oxidations.

12 DyP-type peroxidases

Figure 2. Structural comparison of DyP from Bjerkandera adusta Dec1 (a), HRP from Armoracia rusticana (b) and human myeloperoxidase (hMPO) from Homo sapiens (c). α-helices are shown in green, β sheets are in blue, and the heme cofactor is in red. Close-up of key amino acids in the heme-surrounding region of DyP (d), HRP (e) and human myeloperoxidase (f). The proximal histidine of DyP (His308) and both the distal and proximal histidines of HRP (His42 and His170) and human myeloperoxidase (His95 and His336) are indicated, as well as catalytically important residues of DyP (e.g., Asp171 and Arg329) and HRP (e.g., Arg38 and Phe41). Heme is covalently bound by Asp94, Glu242, and Met243 in human myeloperoxidase. PDB files used: DyP, 2D3Q; HRP, 1ATJ; hMPO, 1CXP.

Biochemical properties The biochemical properties of about forty DyP-type peroxidases of fungal and bacterial origin have been analyzed thus far. These enzymes are typically 50- 60 kDa, while several bacterial variants are somewhat smaller (about 40 kDa). All characterized DyP-type peroxidases contain non-covalently bound heme (protoheme IX) as cofactor.13,14,28,31,35,43 In addition, several oligomeric states have been reported, ranging from monomers to hexamers.6,13,14,23,28,35,38 It has been well established that the catalytic mechanism of plant and animal peroxidases proceeds via formation of compound I. This is the first (high-oxidation) intermediate in the reaction cycle of peroxidases and is formed by a reaction between H2O2 and the Fe(III) resting state of the enzyme. It is therefore generally

13 Chapter 1 assumed that this is also the case for DyP-type peroxidases. Although the exact details about their catalytic cycle are still unclear, several recent studies point towards major differences between the catalytic mechanism of DyP-type peroxidases and other peroxidases. Based on four novel structures of a fungal DyP, it was proposed that the aspartate of the GXXDG-motif functions as acid- base catalyst and swings into a proper position that is optimal for interaction 54 with H2O2. The aspartate catalyzes compound I formation through the transfer of a proton from the proximal to the distal oxygen atom of H2O2, hence facilitating heterolytic cleavage of the O-O bond of2 H O2 (Fig. 3). Compound I, an oxoferryl iron with a porphyrin-based cation radical, reacts consecutively with two equivalents of substrate to return via compound II (Fe(IV)) back to the resting state of the heme (compound 0, Fe(III)). This crucial role of the conserved aspartate as a catalytic residue agrees well with the results of the mutagenesis studies on a fungal and a bacterial DyP as discussed above. However, it is in contrast to, for example, plant peroxidases where the distal histidine functions as an acid-base catalyst and compound I formation is assisted by an essential arginine (Fig. 2).5 Furthermore, analysis of the peroxidative cycle of DypB from Rhodococcus jostiiRHA1 established that its conserved aspartate is not required for peroxidase activity because replacement of this residue by alanine hada marginal effect on the reactivity towards H2O2 and the formation of compound I. Rather, a conserved arginine of DypB was found to be essential for peroxidase activity.55 It therefore appears that DyP-type peroxidases employ different residues as acid-base catalyst(s) during their catalytic cycle. Remarkably, DyP-type peroxidases are able to oxidize substrates that are too large to fit in the active site. DypB, for instance, shows saturation kinetics towards the large molecules of Kraft lignin.30 Long-range electron transfer (LRET) between the heme cofactor and the surface of DypB was suggested as a potential mechanism. More recently, an LRET pathway to the surface ofAau DyPI of Auricularia auricula-judae was identified.42,56 Residues Tyr337 and Leu357 facilitate electron transfer from the heme cofactor of AauDyPI to the surface of this fungal DyP, forming a surface-exposed oxidation site that might react with bulky substrates. Tyr337 is conserved in fungal DyPs. A comparable, but not identical, long-range electron transfer pathway is present in lignin peroxidases (LiP) from the plant superfamily of peroxidases (Fig. 4).42,57 For instance, in LiP from Phanerochaete chrysosporium, a surface-exposed tryptophan was shown to be the interaction site of veratryl alcohol.58 The most distinguishing feature of DyP-type peroxidases is their unparalleled catalytic properties. Firstly, these enzymes are active atlow pH, which is most likely dictated by the aspartate of the GXXDG-motif that

14 DyP-type peroxidases

Figure 3. Schematic representation of the proposed mechanism of DyP-type peroxidases in case substrate AH reacts with the heme directly. Aspartate is shown as acid-base catalyst for the formation of compound I, oxoferryl iron with a porphyrin-based cation radical. Compound I will be reduced in two one-electron reductions back to the resting state, thereby forming two substrate radicals.

Figure 4. Structural comparison of the residues involved in long-range electron transfer of AauDyPI from Auricularia auricula-judae (a) and LiP from Phanerochaete chrysosporium (b). Residues Tyr337 and Leu357 of AauDyP and Trp171 of LiP are involved in long-range electron transfer. PDB files used: AauDyP, 4AU9 and LiP, 1LLP.

15 Chapter 1 functions as an acid-base catalyst at low pH for at least a subset of DyP-type peroxidases.14,43 Secondly, DyP-type peroxidases exhibit a unique substrate acceptance profile. These enzymes are able to degrade various dyes efficiently and in particular anthraquinone dyes, which are poorly accepted by plant and animal peroxidases. Furthermore, DyP-type peroxidases display poor activity towards azo dyes and small non-phenolic compounds unlike plant and animal peroxidases.6,13,61–63,14,16,23,28,35,38,59,60 Moreover, we have relatively recently established that TfuDyP is able to oxidize aromatic sulfides enantioselectively, similar to plant peroxidases, thereby expanding their biocatalytic scope.5,14,64,65 Intriguingly, DyP-type peroxidases appear to be multifunctional enzymes displaying not only oxidative activity but also hydrolytic activity.7,48

Biotechnological potential Plant peroxidases are attractive biocatalysts because of their broad substrate range, neutral pH optimum, and ability to catalyze reactions such as halogenations, epoxidations, hydroxylations, and enantioselective oxidations, often accompanied with good yields.66 However, the exploitation of these enzymes is hampered by their notoriously difficult heterologous expression and limited stability. With regards to the latter, it is interesting to note that DyP- type peroxidases appear remarkable robust, as shown by us and others.14,67,68 Furthermore, our characterization ofTfu DyP showed that this enzyme is expressed well heterologously in E. coli.14 Combined, this shows that the bacterial enzymes are a promising alternative for known peroxidases of fungal origin because of the difficulties in fungal genetics and protein expression. The potential of DyP- type peroxidases as useful biocatalysts for industrial applications is further emphasized by their ability to degrade a variety of synthetic dyes, indicating that these enzymes can be used for the bioremediation of dye-contaminated waste water. Moreover, several recent studies showed that DyP-type peroxidases are involved in the biodegradation of lignocellulosic material, which is highly resistant to (bio)chemical degradation. For example, DypB from Rhodococcus jostii RHA 1 showed activity towards polymeric lignin as well as lignin model compounds.16 Additionally, the hydrolytic degradation of wheat straw was increased by external addition of DyP from Irpex lacteus.47 Together, these studies show that DyP-type peroxidases act as ligninolytic enzymes, thereby pointing towards a major role of these enzymes in the microbial degradation of lignin.19,69 Moreover, it was reported that two fungal DyP-type peroxidases are able to degrade β-carotene.48 The degradation of β-carotene is of interest for the food industry, enabling the enzymatic whitening of whey-containing foods and beverages. This specific application was patented and the respective fungal DyP-

16 DyP-type peroxidases type peroxidase is marketed under the name MaxiBright by DSM. The discovery of novel antimicrobial targets has become a pressing matter due to the vast increase of antibiotic-resistant, pathogenic bacteria.70 With regards to this issue, it is important to emphasize that, as noted earlier, DyP-type peroxidases are remarkably abundant in the proteomes of bacteria, including many pathogenic bacteria, while these enzymes are absent in mammals. This indicates that DyP- type peroxidases could be promising, novel anti-microbial (pro)drug targets. This notion is supported by a recent study that showed that a DyP-type peroxidase from Pseudomonas fluorescens GcM5-1A is toxic to cells of the Japanese black pine.71

Conclusions The group of DyP-type peroxidases comprises a newly identified superfamily of peroxidases, which are unrelated in sequence and structure to well-known peroxidases belonging to the plant or animal superfamilies. DyP-type peroxidases exhibit unique reaction features by displaying novel substrate specificities and reactivities. Additionally, DyP-type peroxidases can be remarkably robust and combined this unveils their potential use as biocatalysts in a variety of biotechnological applications. However, these enzymes are only active under acidic conditions, which severely restrict their number of applications. Itis therefore desirable to alter their pH optimum by enzyme redesign to broaden their applicability. Despite the promising biocatalytic potential of DyP-type peroxidases, much more work is needed to fully characterize the catalytic mechanism of DyP-type peroxidases, their heme biochemistry, as well as the exact role of the catalytic residues and in particular the function of the conserved aspartate. Additional high-resolution structures of DyP-type peroxidases from all the various subclasses are therefore required, preferably in combination with different ligands. The limited number of DyP-type peroxidases characterized so far has established that these enzymes exhibit a vastly different substrate scope than plant and animal peroxidases, using, however, a restricted set of diagnostic substrates. It is therefore desirable that more and diverse substrates should be tested in order to fully understand their biocatalytic scope. Lastly, future studies should be aimed at investigating the potential of DyP-type peroxidases in the biodegradation of lignocellulosic material and as novel microbial (pro)drug targets. In conclusion, it can be expected that the growing number of DyP-type peroxidases biochemically and structurally characterized will fully delineate their biotechnological potential. This will also provide new leads for the construction of improved variants suitable for biotechnological applications.

17 Chapter 1

Aim and outline of the thesis The research in this thesis was financed by the Netherlands Organization for Scientific Research (NWO) under the graduate program: synthetic biology for advanced metabolic engineering, project number 022.004.006. The aim of the research presented in this thesis was to broaden the knowledge on class A DyP-type peroxidases. Two DyPs were selected and used as model enzymes: TfuDyP from Thermobifida fusca YX and SviDyP from Saccharomonospora viridis DSM 43017.14,24 TfuDyP was isolated and characterized by van Bloois et al as a thermostable, Tat-dependently secreted peroxidase which was easily overexpressed in Escherichia coli. The substrate scope of this enzyme covers dyes and monophenolic compounds while it also shows peroxygenase activity in the enantioselective sulfoxidation of aromatic sulfides. SviDyP is homologous to TfuDyP (42% sequence identity) and has the advantage over TfuDyP that it is active at a slightly higher pH range. DyP-type peroxidases are named after their ability to degraded recalcitrant dyes.6 Previous studies focused predominantly on anthraquinone and azo dyes.6,20,38 To study the biotechnological potential of DyP-type peroxidases further, an extensive substrate profiling study was performed with TfuDyP as model enzyme. Chapter 2 presents the activity of TfuDyP on thirty dyes from seven distinct classes, three natural carotenoids and various lignin model compounds. For their activity DyP-type peroxidases rely on a tightly bound heme cofactor. On the proximal side of the heme a histidine functions as the fifth ligand of the heme iron, while on the distal side two catalytically important residues are found: an aspartate and an arginine. The oxidation site(s) of small and large substrates are presumably different and not fully understood. Some small compounds are known to enter the heme pocket and react with the heme cofactor directly. Large compounds, e.g. bulky dyes and lignin model compounds, are however too large to enter the active site.Chapter 3 describes a mutagenesis study on TfuDyP with the aim to identify more catalytically important residues or residues that determine the pH optimum of the enzyme. This chapter presents the effect of mutations in the heme pocket, in the predicted hydrogen peroxide tunnel and close to the surface exposed heme propionate. Some (DyP- type) peroxidases depend on long-range electron transfer for the activity of substrates that are too large to enter the active site. To explore whetherTfu DyP is dependent on such a mechanism, mutagenesis was performed on the surface exposed tyrosines and tryptophans. For the biotechnological applicability of DyP-type peroxidases, it would be beneficial to shift the pH optimum for activity

18 DyP-type peroxidases to a more neutral pH range. Mutations that were proven beneficial in shifting the pH optimum of PpDyP and BsDyP were studied in TfuDyP.72,73 For the industrial applicability of DyP-type peroxidases a high protein overexpression level is desired. Many DyP-type peroxidases are from bacterial origin and are therefore relatively easily heterologously overexpressed in a bacterial host. This is in stark contrast to peroxidases from eukaryotes. Even though TfuDyP originates from T. fusca and is easily overexpressed in E. coli, increasing the overexpression level to 200 mg per liter culture broth led to an almost inactive enzyme. Chapter 4 discusses the correlation between the activity of TfuDyP and the expression level in E. coli. Analysis of the protein by UV-vis absorbance spectroscopy and high-resolution mass spectroscopy on the extracted heme cofactor revealed the reason for the inactivity of the enzyme. Most enzymes in nature are involved in metabolic pathways in which the product of one enzyme is the substrate of another enzyme. In some cases this even led to the fusion of enzymes to bi/multifunctional protein complexes. Peroxidases and oxidases form catalytically logical combinations and are often coexpressed in nature: peroxidases require hydrogen peroxide for their activity, a by-product of the oxidases. Chapter 5 describes the first successful recombinant expression of artificial oxidase-peroxidase fusion enzymes. We fused SviDyP to four distinct oxidases: alditol oxidase (HotAldO), chitooligosaccharide oxidase (ChitO), eugenol oxidase (EugO) and 5-hydroxymethylfurfural oxidase (HMFO).74–77 Special attention was paid to exploring the potential applicability of the designed oxidase-peroxidase fusions. Two fusion enzymes were used in one-pot two-step cascade reactions while the other two fusion enzymes could be applied as biosensor for the detection of various sugars.

19 Chapter 1

References

1 Banci L (1997) Structural properties of peroxidases. J. Biotechnol. 53, 253–263. 2 Davies MJ, Hawkins CL, Pattison DI, and Rees MD (2008) Mammalian heme peroxidases: from molecular mechanisms to health implications. Antioxid. Redox Signal. 10, 1199–1234. 3 Passardi F, Cosio C, Penel C, and Dunand C (2005) Peroxidases have more functions than a Swiss army knife. Plant Cell Rep. 24, 255–265. 4 Welinder KG, Mauro JM, and Norskov-Lauritsen L (1992) Structure of plant and fungal peroxidases. Biochem. Soc. Trans. 20, 377–340. 5 Veitch NC (2004) Horseradish peroxidase: A modern view of a classic enzyme. Phytochemistry 65, 249–259. 6 Kim SJ and Shoda M (1999) Purification and characterization of a novel peroxidase from Geotrichum candidum Dec 1 involved in decolorization of dyes.Appl. Environ. Microbiol. 65, 1029–1035. 7 Sugano Y (2009) DyP-type peroxidases comprise a novel heme peroxidase family. Cell. Mol. Life Sci. 66, 1387–1403. 8 Finn RD, Attwood TK, Babbitt PC, Bateman A, Bork P, Bridge AJ, Chang H-Y, Dosztányi Z, El- Gebali S, Fraser M, Gough J, Haft D, Holliday GL, Huang H, Huang X, Letunic I, Lopez R, Lu S, Marchler-Bauer A, Mi H, Mistry J, Natale DA, Necci M, Nuka G, Orengo CA, Park Y, Pesseat S, Piovesan D, Potter SC, Rawlings ND, Redaschi N, Richardson L, Rivoire C, Sangrador-Vegas A, Sigrist C, Sillitoe I, Smithers B, Squizzato S, Sutton G, Thanki N, Thomas PD, Tosatto SCE, Wu CH, Xenarios I, Yeh L-S, Young S-Y, and Mitchell AL (2017) InterPro in 2017 - beyond protein family and domain annotations. Nucleic Acids Res. 45, D190–D199. 9 Fawal N, Li Q, Savelli B, Brette M, Passaia G, Fabre M, Mathé C, and Dunand C(2013) PeroxiBase: a database for large-scale evolutionary analysis of peroxidases. Nucleic Acids Res. 41, D441–D444. 10 Yoshida T and Sugano Y (2015) A structural and functional perspective of DyP-type peroxidase family. Arch. Biochem. Biophys. 574, 49–55. 11 Singh R and Eltis LD (2015) The multihued palette of dye-decolorizing peroxidases. Arch. Biochem. Biophys. 574, 56–65. 12 Jongbloed JDH, Grieger U, Antelmann H, Hecker M, Nijland R, Bron S, and van Dijl JM (2004) Two minimal Tat translocases in Bacillus. Mol. Microbiol. 54, 1319–1325. 13 Sturm A, Schierhorn A, Lindenstrauss U, Lilie H, and Brüser T (2006) YcdB from Escherichia coli reveals a novel class of Tat-dependently translocated hemoproteins. J. Biol. Chem. 281, 13972–13978. 14 van Bloois E, Torres Pazmiño DE, Winter RT, and Fraaije MW (2010) A robust and extracellular heme-containing peroxidase from Thermobifida fusca as prototype of a bacterial peroxidase superfamily. Appl. Microbiol. Biotechnol. 86, 1419–1430. 15 Adav SS, Ng CS, Arulmani M, and Sze SK (2010) Quantitative iTRAQ secretome analysis of cellulolytic Thermobifida fusca. J. Proteome Res. 9, 3016–3024. 16 Ahmad M, Roberts JN, Hardiman EM, Singh R, Eltis LD, and Bugg TDH (2011) Identification of DypB from Rhodococcus jostii RHA1 as a lignin peroxidase. Biochemistry 50, 5096–5107. 17 Rahmanpour R and Bugg TDH (2015) Characterisation of Dyp-type peroxidases from Pseudomonas fluorescens Pf-5: Oxidation of Mn(II) and polymeric lignin by Dyp1B. Arch. Biochem. Biophys. 574, 93–98. 18 Rahmanpour R, Rea D, Jamshidi S, Fülöp V, and Bugg TDH (2016) Structure of Thermobifida fusca DyP-type peroxidase and activity towards Kraft lignin and lignin model compounds. Arch. Biochem. Biophys. 594, 54–60. 19 de Gonzalo G, Colpa DI, Habib MHM, and Fraaije MW (2016) Bacterial enzymes involved in lignin degradation. J. Biotechnol. 236, 110–119. 20 Santos A, Mendes S, Brissos V, and Martins LO (2014) New dye-decolorizing peroxidases from

20 DyP-type peroxidases

Bacillus subtilis and Pseudomonas putida MET94: towards biotechnological applications. Appl. Microbiol. Biotechnol. 98, 2053–2065. 21 Kekilli D, Moreno-Chicano T, Chaplin AK, Horrell S, Dworkowski FSN, Worrall JAR, Strange RW, and Hough MA (2017) Photoreduction and validation of haem-ligand intermediate states in protein crystals by in situ single-crystal spectroscopy and diffraction. IUCrJ 4, 263– 270. 22 Petrus MLC, Vijgenboom E, Chaplin AK, Worrall JAR, van Wezel GP, and Claessen D (2016) The DyP-type peroxidase DtpA is a Tat-substrate required for GlxA maturation and morphogenesis in Streptomyces. Open Biol. 6, 150149. 23 Liu X, Du Q, Wang Z, Zhu D, Huang Y, Li N, Wei T, Xu S, and Gu L (2011) Crystal structure and biochemical features of EfeB/YcdB from Escherichia coli O157: Asp235 plays divergent roles in different enzyme-catalyzed processes. J. Biol. Chem. 286, 14922–14931. 24 Yu W, Liu W, Huang H, Zheng F, Wang X, Wu Y, Li K, Xie X, and Jin Y (2014) Application of a novel alkali-tolerant thermostable DyP-type peroxidase from Saccharomonospora viridis DSM 43017 in biobleaching of Eucalyptus Kraft pulp. PLoS One 9, e110319. 25 Shrestha R, Chen X, Ramyar KX, Hayati Z, Carlson EA, Bossmann SH, Song L, Geisbrecht B V, and Li P (2016) Identification of surface-exposed protein radicals and a substrate oxidation site in A-class dye-decolorizing peroxidase from Thermomonospora curvata. ACS Catal. 6, 8036–8047. 26 Rahmanpour R, Rea D, Jamshidi S, Fülöp V, and Bugg TDH (2016) Structure of Thermobifida fusca DyP-type peroxidase and activity towards Kraft lignin and lignin model compounds. Arch. Biochem. Biophys. 594, 54–60. 27 Sutter M, Boehringer D, Gutmann S, Günther S, Prangishvili D, Loessner MJ, Stetter KO, Weber-Ban E, and Ban N (2008) Structural basis of enzyme encapsulation into a bacterial nanocompartment. Nat. Struct. Mol. Biol. 15, 939–947. 28 Zubieta C, Krishna SS, Kapoor M, Kozbial P, McMullan D, Axelrod HL, Miller MD, Abdubek P, Ambing E, Astakhova T, Carlton D, Chiu H-J, Clayton T, Deller MC, Duan L, Elsliger M-A, Feuerhelm J, Grzechnik SK, Hale J, Hampton E, Han GW, Jaroszewski L, Jin KK, Klock HE, Knuth MW, Kumar A, Marciano D, Morse AT, Nigoghossian E, Okach L, Oommachen S, Reyes R, Rife CL, Schimmel P, van den Bedem H, Weekes D, White A, Xu Q, Hodgson KO, Wooley J, Deacon AM, Godzik A, Lesley SA, and Wilson IA (2007) Crystal structure of two novel dye-decolorizing peroxidases revel a β-barrel fold with a conserved heme-binding motif. Proteins 69, 223–233. 29 Rai A, Fedorov R, and Manstein DJ (2014) Expression, purification and crystallization of a dye-decolourizing peroxidase from Dictyostelium discoideum. Acta Crystallogr. Sect. FStructural Biol. Commun. 70, 252–255. 30 Roberts JN, Singh R, Grigg JC, Murphy MEP, Bugg TDH, and Eltis LD (2011) Characterization of dye-decolorizing peroxidases from Rhodococcus jostii RHA1. Biochemistry 50, 5108–5119. 31 Li J, Liu C, Li B, Yuan H, Yang J, and Zheng B (2012) Identification and molecular characterization of a novel DyP-type peroxidase from Pseudomonas aeruginosa PKE117. Appl. Biochem. Biotechnol. 166, 774–785. 32 Shrestha R, Huang G, Meekins DA, Geisbrecht B V, and Li P (2017) Mechanistic insights into dye-decolorizing peroxidase revealed by solvent isotope and viscosity effects. ACS Catal. 7, 6352–6364. 33 Contreras H, Joens MS, McMath LM, Le VP, Tullius M V, Kimmey JM, Bionghi N, Horwitz MA, Fitzpatrick JAJ, and Goulding CW (2014) Characterization of a Mycobacterium tuberculosis nanocompartment and its potential cargo proteins. J. Biol. Chem. 289, 18279–18289. 34 Sezer M, Santos A, Kielb P, Pinto T, Martins LO, and Todorovic S (2013) Distinct structural and redox properties of the heme active site in bacterial dye decolorizing peroxidase- type peroxidases from two subfamilies: Resonance raman and electrochemical study. Biochemistry 52, 3074–3084.

21 Chapter 1

35 Zubieta C, Joseph R, Krishna SS, McMullan D, Kapoor M, Axelrod HL, Miller MD, Abdubek P, Acosta C, Astakhova T, Carlton D, Chiu H-J, Clayton T, Deller MC, Duan L, Elias Y, Elsliger M-A, Feuerhelm J, Grzechnik SK, Hale J, Han GW, Jaroszewski L, Jin KK, Klock HE, Knuth MW, Kozbial P, Kumar A, Marciano D, Morse AT, Murphy KD, Nigoghossian E, Okach L, Oommachen S, Reyes R, Rife CL, Schimmel P, Trout C V., van den Bedem H, Weekes D, White A, Xu Q, Hodgson KO, Wooley J, Deacon AM, Godzik A, Lesley SA, and Wilson IA (2007) Identification and structural characterization of heme binding in a novel dye-decolorizing peroxidase, TyrA. Proteins 69, 234–243. 36 Uchida T, Sasaki M, Tanaka Y, and Ishimori K (2015) A dye-decolorizing peroxidase from Vibrio cholerae. Biochemistry 54, 6610–6621. 37 Liu X, Yuan Z, Wang J, Cui Y, Liu S, Ma Y, Gu L, and Xu S (2017) Crystal structure and biochemical features of dye-decolorizing peroxidase YfeX from Escherichia coli O157 Asp143 and Arg232 play divergent roles toward different substrates.Biochem. Biophys. Res. Commun. 484, 40– 44. 38 Ogola HJO, Kamiike T, Hashimoto N, Ashida H, Ishikawa T, Shibata H, and Sawa Y (2009) Molecular characterization of a novel peroxidase from the cyanobacterium Anabaena sp. strain PCC 7120. Appl. Environ. Microbiol. 75, 7509–7518. 39 Brown ME, Barros T, and Chang MC (2012) Identification and characterization ofa multifunctional dye peroxidase from a lignin-reactive bacterium. ACS Chem. Biol. 7, 2074– 2081. 40 Sugawara K, Nishihashi Y, Narioka T, Yoshida T, Morita M, and Sugano Y (2017) Characterization of a novel DyP-type peroxidase from Streptomyces avermitilis. J. Biosci. Bioeng. 123, 425– 430. 41 Liers C, Bobeth C, Pecyna M, Ullrich R, and Hofrichter M (2010) DyP-like peroxidases of the jelly fungus Auricularia auricula-judae oxidize nonphenolic lignin model compounds and high-redox potential dyes. Appl. Microbiol. Biotechnol. 85, 1869–1879. 42 Strittmatter E, Liers C, Ullrich R, Wachter S, Hofrichter M, Plattner DA, and Piontek K (2013) First crystal structure of a fungal high-redox potential dye-decolorizing peroxidase substrate interaction sites and long-range electron transfer. J. Biol. Chem. 288, 4095–4102. 43 Sugano Y, Muramatsu R, Ichiyanagi A, Sato T, and Shoda M (2007) DyP, a unique dye- decolorizing peroxidase, represents a novel heme peroxidase family: ASP171 replaces the distal histidine of classical peroxidases. J. Biol. Chem. 282, 36652–36658. 44 Liers C, Pecyna MJ, Kellner H, Worrich A, Zorn H, Steffen KT, Hofrichter M, and Ullrich R (2013) Substrate oxidation by dye-decolorizing peroxidases (DyPs) from wood- and litter- degrading agaricomycetes compared to other fungal and plant heme-peroxidases. Appl. Microbiol. Biotechnol. 97, 5839–5849. 45 Kolwek J, Behrens C, Linke D, Krings U, and Berger RG (2018) Cell-free one-pot conversion of (+)-valencene to (+)-nootkatone by a unique dye-decolorizing peroxidase combined with a laccase from Funalia trogii. J. Ind. Microbiol. Biotechnol. 45, 89–101. 46 Kung CP, Wu YR, and Chuang HW (2014) Expression of a dye-decolorizing peroxidase results in hypersensitive response to cadmium stress through reducing the ROS signal in Arabidopsis. Environ. Exp. Bot. 101, 47–55. 47 Salvachúa D, Prieto A, Martínez ÁT, and Martínez MJ (2013) Characterization of a novel dye-decolorizing peroxidase (DyP)-type enzyme from Irpex lacteus and its application in enzymatic hydrolysis of wheat straw. Appl. Environ. Microbiol. 79, 4316–4324. 48 Scheibner M, Hülsdau B, Zelena K, Nimtz M, de Boer L, Berger RG, and Zorn H (2008) Novel peroxidases of Marasmius scorodonius degrade β-carotene. Appl. Microbiol. Biotechnol. 77, 1241–1250. 49 Faraco V, Piscitelli A, Sannia G, and Giardina P (2007) Identification of a new member of the dye-decolorizing peroxidase family from Pleurotus ostreatus. World J. Microbiol. Biotechnol. 23, 889–893.

22 DyP-type peroxidases

50 Lauber C, Schwarz T, Nguyen QK, Lorenz P, and Lochnit G (2017) Identification, heterologous expression and characterization of a dye-decolorizing peroxidase ofPleurotus sapidus. AMB Express 7, 164. 51 Johjima T, Ohkuma M, and Kudo T (2003) Isolation and cDNA cloning of novel hydrogen peroxide-dependent phenol oxidase from the basidiomycete Termitomyces albuminosus. Appl. Microbiol. Biotechnol. 61, 220–225. 52 Zamocky M, Jakopitsch C, Furtmüller PG, Dunand C, and Obinger C (2008) The peroxidase- cyclooxygenase superfamily: reconstructed evolution of critical enzymes of the innate immune system. Proteins 72, 589–605. 53 Sugano Y, Ishii Y, and Shoda M (2004) Role of H164 in a unique dye-decolorizing heme peroxidase DyP. Biochem. Biophys. Res. Commun. 322, 126–132. 54 Yoshida T, Tsuge H, Konno H, Hisabori T, and Sugano Y (2011) The catalytic mechanism of dye-decolorizing peroxidase DyP may require the swinging movement of an aspartic acid residue. FEBS J. 278, 2387–2394. 55 Singh R, Grigg JC, Armstrong Z, Murphy MEP, and Eltis LD (2012) Distal heme pocket residues of B-type dye-decolorizing peroxidase: Arginine but not aspartate is essential for peroxidase activity. J. Biol. Chem. 287, 10623–10630. 56 Strittmatter E, Wachter S, Liers C, Ullrich R, Hofrichter M, Plattner DA, and Piontek K (2013) Radical formation on a conserved tyrosine residue is crucial for DyP activity.Arch. Biochem. Biophys. 537, 161–167. 57 Choinowski T, Blodig W, Winterhalter KH, and Piontek K (1999) The crystal structure of lignin peroxidase at 1.70 A resolution reveals a hydroxy group on the Cβ of tryptophan 171: a novel radical site formed during the redox cycle. J. Mol. Biol. 286, 809–827. 58 Doyle WA, Blodig W, Veitch NC, Piontek K, and Smith AT (1998) Two substrate interaction sites in lignin peroxidase revealed by site-directed mutagenesis. Biochemistry 37, 15097– 15105. 59 Burner U, Krapfenbauer G, Furtmüller PG, Regelsberger G, and Obinger C (2000) Oxidation of hydroquinone, 2,3-dimethylhydroquinone and 2,3,5-trimethylhydroquinone by human myeloperoxidase. Redox Rep. 5, 185–190. 60 Reszka KJ, McCormick ML, and Britigan BE (2001) Peroxidase- and nitrite-dependent metabolism of the anthracycline anticancer agents daunorubicin and doxorubicin. Biochemistry 40, 15349–15361. 61 Reszka KJ, Wagner BA, Burns CP, and Britigan BE (2005) Effects of peroxidase substrates on the Amplex red/peroxidase assay: Antioxidant properties of anthracyclines. Anal. Biochem. 342, 327–337. 62 Stolz A (2001) Basic and applied aspects in the microbial degradation of azo dyes. Appl. Microbiol. Biotechnol. 56, 69–80. 63 Chen H (2006) Recent advances in azo dye degrading enzyme research. Curr. Protein Pept. Sci. 7, 101–111. 64 Klibanov AM (2003) Asymmetric enzymatic oxidoreductions in organic solvents. Curr. Opin. Biotechnol. 14, 427–431. 65 van Rantwijk F and Sheldon RA (2000) Selective oxygen transfer catalysed by heme peroxidases: synthetic and mechanistic aspects. Curr. Opin. Biotechnol. 11, 554–564. 66 Regalado C, García-Almendárez BE, and Duarte-Vázquez MA (2004) Biotechnological applications of peroxidases. Phytochem. Rev. 3, 243–256. 67 Liers C, Ullrich R, Hofrichter M, Minibayeva F V, and Beckett RP (2011) A heme peroxidase of the ascomyceteous lichen Leptogium saturninum oxidizes high-redox potential substrates. Fungal Genet. Biol. 48, 1139–1145. 68 Pühse M, Szweda RT, Ma Y, Jeworrek C, Winter R, and Zorn H (2009) Marasmius scorodonius extracellular dimeric peroxidase - Exploring its temperature and pressure stability. Biochim. Biophys. Acta 1794, 1091–1098.

23 Chapter 1

69 Bugg TDH, Ahmad M, Hardiman EM, and Singh R (2011) The emerging role for bacteria in lignin degradation and bio-product formation. Curr. Opin. Biotechnol. 22, 394–400. 70 Amini S and Tavazoie S (2011) Antibiotics and the post-genome revolution. Curr. Opin. Microbiol. 14, 513–518. 71 Kong L, Guo D, Zhou S, Yu X, Hou G, Li R, and Zhao B (2010) Cloning and expression of a toxin gene from Pseudomonas fluorescens GcM5-1A. Arch. Microbiol. 192, 585–593. 72 Mendes S, Catarino T, Silveira C, Todorovic S, and Martins LO (2015) The catalytic mechanism of A-type dye-decolourising peroxidase BsDyP: neither aspartate nor arginine is individually essential for peroxidase activity. Catal. Sci. Technol. 5, 5196–5207. 73 Mendes S, Brissos V, Gabriel A, Catarino T, Turner DL, Todorovic S, and Martins LO (2015) An integrated view of redox and catalytic properties of B-type PpDyP from Pseudomonas putida MET94 and its distal variants. Arch. Biochem. Biophys. 574, 99–107. 74 Winter RT, Heuts DPHM, Rijpkema EMA, van Bloois E, Wijma HJ, and Fraaije MW (2012) Hot or not? Discovery and characterization of a thermostable alditol oxidase fromAcidothermus cellulolyticus 11B. Appl. Microbiol. Biotechnol. 95, 389–403. 75 Ferrari AR, Lee M, and Fraaije MW (2015) Expanding the substrate scope of chitooligosaccharide oxidase from Fusarium graminearum by structure-inspired mutagenesis. Biotechnol. Bioeng. 112, 1074–1080. 76 Dijkman WP and Fraaije MW (2014) Discovery and characterization of a 5-hydroxymethylfurfural oxidase from Methylovorus sp. strain MP688. Appl. Environ. Microbiol. 80, 1082–1090. 77 Jin J, Mazon H, van den Heuvel RHH, Janssen DB, and Fraaije MW (2007) Discovery of a eugenol oxidase from Rhodococcus sp. strain RHA1. FEBS J. 274, 2311–2321. 78 Zámocký M, Hofbauer S, Schaffner I, Gasselhuber B, Nicolussi A, Soudi M, Pirker KF, Furtmüller PG, and Obinger C (2015) Independent evolution of four heme peroxidase superfamilies. Arch. Biochem. Biophys. 574, 108–119. 79 Goblirsch B, Kurker RC, Streit BR, Wilmot CM, and Dubois JL (2011) Chlorite dismutases, DyPs and EfeB: 3 microbial heme enzyme families comprise the CDE structural superfamily. J. Mol. Biol. 408, 379–398.

24 Chapter 2:

Exploring the biocatalytic potential of a DyP-type peroxidase by profiling the substrate acceptance of Thermobifida fusca DyP peroxidase

Nikola Lončar*, Dana I. Colpa* and Marco W. Fraaije * these authors contributed equally to this work

This chapter is based on: Tetrahedron (2016) 72: 7276-7281 DOI: 10.1016/j.tet.2015.12.078 Abstract Dye-decolorizing peroxidases (DyPs) represent a new class of oxidative enzymes for which the natural substrates are largely unknown. To explore the biocatalytic potential of a DyP, we have studied the substrate acceptance profile of a robust DyP peroxidase, a DyP from Thermobifida fusca (TfuDyP). While previous work established that this bacterial peroxidase is able to act on a few reactive dyes and aromatic sulfides, this work significantly expands its substrate scope towards lignin related compounds, flavors, and various dyes. Exploring the substrate scope of TfuDyP

Introduction Substrate promiscuity has often been attributed to different oxidoreductases and in particular to peroxidases. These enzymes perform hydrogen peroxide driven one electron oxidations of a wide range of phenolic and nonphenolic substrates.1 Substrate promiscuity of enzymes is an interesting biocatalytic property as it broadens up the applicability of an enzyme as a biocatalyst. Plant and animal peroxidases are notorious for their high activity and wide range of substrates. These eukaryotic biocatalysts are known already for several decades, have been thoroughly studied, and are currently used in numerous processes.2,3 Despite being powerful catalysts, application of these enzymes is often hampered by their low temperature stability and sensitivity to salt and organic solvents. Furthermore, it has been proven to be difficult and often impossible to produce these peroxidases in recombinant form. For example, it has been shown that it is extremely difficult to produce horseradish peroxidase in a heterologous host.4 As a result, horseradish peroxidase is still mainly produced by isolating it from plant roots which results in a mixture of various peroxidase isoforms. As alternatives for the plant and animal peroxidases, the newly discovered DyP-type peroxidases (DyPs) may offer advantages. One advantage is the possibility to produce such peroxidases using bacterial expression hosts as most DyPs are of bacterial origin.5 Except for facilitating the production of peroxidases and eliminating the existence of isoforms, the ability to produce DyPs in a recombinant form also allows engineering of these biocatalysts. The first DyPs were identified less than two decades ago.6 DyPs are unrelated in sequence and structure to peroxidases belonging to the plant or animal peroxidase superfamilies.7 While numerous putative DyP-encoding genes can be identified in sequenced bacterial genomes, only a small number of DyPs have been characterized. Originally, their activity was established based on the decolorization of dyes, and hence their name (DyP stands for dye decolorizing peroxidase). DyPs are typically identified by their activity on anthraquinone dyes. While DyPs are efficient in oxidizing these synthetic dyes, the physiological substrates for DyPs are unclear and therefore their role in nature is enigmatic. Interestingly, recent studies suggest that bacterial DyPs may play an important role in the degradation of lignin which suggests that DyPs represent the bacterial counterparts of the fungal lignin peroxidases. Except for establishing their activity on synthetic dyes and possible role in lignin degradation, little data is available concerning their biocatalytic potential. Therefore, we set out a study aimed at profiling the potential of a newly identified DyP which can be easily produced as recombinant enzyme and is thermostable: DyP from Thermobifida fusca (TfuDyP).

27 Chapter 2

T. fusca is a moderate thermophilic soil actinomycete with a growth temperature of approximately 50 ºC. It is a major degrader of plant cell walls in heated organic materials.4 It produces many extracellular enzymes, including cellulases. A number of these secreted enzymes has been studied because of their thermostability, broad pH range and high activity.8 TfuDyP is a robust and secreted peroxidase described previously as a member of the DyP family.9 Activity of TfuDyP towards several reactive dyes was described in addition to enantioselective sulfoxidation.9 In this paper we present an exhaustive substrate profiling study which provides a better view on the biocatalytic repertoire of this newly discovered robust peroxidase.

Results and Discussion Establishing optimal conditions To investigate the experimental boundaries at whichTfu DyP can be applied, the apparent melting temperature of TfuDyP was measured at different pH values.

In the pH range of 5-8 the enzyme shows a Tm, app of ~56 ºC (Fig. 1). This is in line with temperatures at which Thermobifida fusca thrives and it shows that TfuDyP is a rather thermostable peroxidase. However, its thermostability decreases dramatically at a pH below 5(Tm, app = 35 ºC at pH 3). This contrasts the pH optimum for optimal TfuDyP activity which is in the range of pH 3-4 (vide infra). Such a low pH optimum for activity has also been observed for other DyPs.10 These data indicate that there is a delicate balance in pH optima for activity and stability. Related to this, one should realize that TfuDyP and many other DyPs are secreted and may have to operate at a pH different from neutral pH that is normal for intracellular enzymes. The broad pH optimum for stability is in line with the pH optima observed for other secreted enzymes of T. fusca that typically display a pH optimum of 4-10.8 Another noteworthy observation is the fact that the pH optimum for activity seems to depend on the type of substrate (vide infra). For peroxidases different from DyPs, it has been established that hydrogen peroxide can be replaced by organic peroxides such a tert-butyl peroxide.11 To our knowledge, DyPs had not been tested before with these peroxide alternatives. However, when testing TfuDyP activity with 0.10, 1.0 or 10 mM tert-butyl peroxide as electron acceptor and Reactive Blue 19 as substrate, no activity was observed. Moreover, when monitoring the UV/Vis absorbance spectrum of TfuDyP upon the addition of 0.10 mM tert-butyl peroxide, no spectral changes were observed in the Soret band. This indicates that TfuDyP is very selective for hydrogen peroxide. For the substrate profiling experiments performed in this study, 0.10 mM of hydrogen peroxide was used as cosubstrate.

28 Exploring the substrate scope of TfuDyP

Figure 1. Effect of the pH on the apparent melting temperature of TfuDyP.

Degradation of synthetic and natural dyes DyP-type peroxidases are named after their ability to convert dyes. In previous studies, DyP activity has been mainly probed using a restricted number of synthetic anthraquinone and azo dyes for each reported DyP.6 Activity towards triarylmethane dyes and natural pigment β-carotene has also been reported.12,13 This study aimed at an extensive exploration of the substrate scope of a DyP- type peroxidase, TfuDyP. The activity of TfuDyP towards hemin, three natural carotenoids and thirty members of seven different classes of dyes was determined.

For every dye, the initial activity (kobs) and the amount of dye degraded in one hour were determined at pH 3, pH 4, and pH 5. The amount of dye degraded in one hour was defined as the observed decrease in absorbance at λmax. One should note that the degree of dye degradation is an underestimation in case the product has a comparable absorption spectrum. The absorbance maxima of carminic acid and the copper phthalocyanine tetrasulfonic acid dye were pH dependent. For these compounds the isosbestic point of the spectra at pH 3, 4 and 5 was used to analyze the activities. A few dyes were found to be poorly soluble in buffer and were prepared in DMSO and used in the reaction mixture with a final concentration of 2.5% DMSO (resorufin) or 10% DMSO (Disperse Blue 1, curcumin, and β‑carotene). Only a small number of tested compounds did not show any activity with TfuDyP: hemin, β‑carotene, the azo dyes Direct Yellow 27 and Acid Yellow 23, and the heterocyclic dyes methylene blue, neutral red, and resorufin. The highest activities and conversions were observed for the anthraquinone dyes (Tables 1 and 2, and Table S1). Most of the representatives of the other dye classes displayed

29 Chapter 2

Table 1. Structural formulas of the tested dyes. Compound names see Table 2.

19 27

Phthalocyanine dye Phthalocyanine 14

18 5

26 Indigoid dyes 11

Dyes

13 25

1 Anthraquinone dyes dyes Azo dyes Arylmethane Xanthene dyes Carotenoids

30 Exploring the substrate scope of TfuDyP

19 27

Phthalocyanine dye Phthalocyanine 14

18 5

26 Indigoid dyes 11

Dyes

13 25

1 Anthraquinone dyes dyes Azo dyes Arylmethane Xanthene dyes Carotenoids

31 Chapter 2

Table 2. Activity of TfuDyP on representatives of different classes of dyes.

λmax Dye kobs Dye degraded Nr. Dye at pH 4 concentration at pH 4 in 1h (nm) (µM) (s-1)(1) (%)(2) Anthraquinone dyes 1 Disperse Blue 1 588 50#, ## 10 48 (13)(3) 2 Carminic acid 503* 50 2.4 · 10-2 41 3 Acid Blue 129 629 50# 22 82 4 Acid Blue 80 629 50# 0.11 34 5 Reactive Blue 19 595 50# 1.7 22(4) 6 Reactive Blue 4 598 50# 1.4 12 7 Cibacron Blue 3G-A 615 50# 1.5 4.2 Azo dyes 8 Acid Orange 7 484 25 1.4 · 10-2 16(5) 9 Reactive Red 2 512 25 3.5 · 10-3 2.1(4) 10 Acid Red 18 507 25 2.7 · 10-2 15 11 Acid Red 14 516 25 4.7 · 10-2 19(5) 12 Reactive Black 5 597 25 1.4 · 10-2 7.7 13 Reactive Red 120 510 10 8.4 · 10-3 3.2 14 Direct Red 80 543 10 4.1 · 10-3 4.1 Di/Tri-arylmethane dyes 15 Basic Yellow 2 432 25 5.0 · 10-3 16 16 Crystal Violet 590 25 5.2 · 10-3 15 17 Acid Green 50 635 10 3.2 · 10-2 48 18 Acid Blue 9 630 25 - 35 19 Acid Blue 93 592 50 6.1 · 10-2 5.4(5) Xanthene dyes 20 Rhodamine B 554 25 1.6 · 10-2 12 21 Fluorescein 474** 25 9.4 · 10-3 12(6) 22 Eosin Y 517 25 0.93 92 Indigoid dyes 23 Indigo carmine 611** 50 2.2 · 10-2 31 (1.9)(7) 24 Indigotetrasulfonate 590 50 2.3 · 10-2 8.2(5) Carotenoids 25 Crocin 441** 50 8.8 · 10-3 72 (1.2)(7) 26 Curcumin 431 50## 7.2 · 10-2 37 (4.1)(3) Phthalocyanine dye 27 Copper phthalocyanine- 612* 25 0.85 64(5) 3,4’,4’’,4’’’-tetrasulfonic acid

32 Exploring the substrate scope of TfuDyP

(1) If necessary kobs was corrected for the background activity. (2) Percentage of dye degraded in one hour is based on the observed decrease in absorbance at λmax. The actual amount of degraded dye is higher in case the product absorbs in the same range. High background activities of dye degradation with2 H O2 but without enzyme are given in parenthesis.

(3) Higher kobs and more degradation after 1h at pH 5. (4) Higherk obs and more degradation after 1h at pH 3. (5) Lower kobs but more degradation after 1h at pH 5. (6) Measured at pH 5 as only activity at pH 5 could be observed. (7) Measured at pH 5, dye is not stable at pH 4. * In this case the isosbestic point at pH 3, 4 and 5 was taken as wavelength to monitor activity. ** At pH 5. # 30 nM enzyme was used instead of 300 nM. ## Containing 10% DMSO.

2-3 orders of magnitude lower initial activities. There were two exceptions: the xanthene dye Eosin Y and the copper phthalocyanine tetrasulfonic acid dye were -1 good substrates with kobs values of around 1 s (Table 2). Anthraquinone dye -1 Acid Blue 129 was the best substrate with a kobs of 22 s and a 82% decrease in absorbance at λmax in one hour. While the initial rates of decolorization among the different dyes varied significantly, significant decolorization of most of the dyes was observed after 1 hour. This can be caused by various factors, for example the affinity of TfuDyP for dyes can be different and/or the formed dye degradation products may inhibit decolorization by the peroxidase. For most dyes only one oxidation/decolorization step was observed as no other color developed during the decolorization. Only for the phthalocyanine dye two clear oxidation steps were visible. First, the color changed from light blue to dark blue, after which the solution decolored fully. The products of the decolorization reactions were not characterized. In fact, it is worth noting that, although DyPs can effectively decolor various dyes, they do not fully degrade dyes into regular metabolites. Still, such enzymes may develop as valuable biocatalysts for dye degradation, for example, for textile wastewater treatment, as the degradation products may be less toxic and/or easily degraded by follow-up microbial catabolic routes.14,15 In general, higher initial activities were observed at pH 3 (Table S1). However, the enzyme is poorly stable at this pH and rapidly inactivates, resulting in lower dye degradation in the first hour. The measured initial rates of decolorization also revealed that the pH optimum for activity is clearly substrate dependent. Such phenomenon was also observed for other DyP-type 6,16 peroxidases in previous studies. In most cases, when taking both kobs and the degree of degradation in one hour into account, TfuDyP was most effective at pH 4 (Table S1). Some substrates were however an exception. TfuDyP showed a higher activity for the anthraquinone dye Reactive Blue 19 and azo dye Reactive Red 2 at pH 3 while it performed better with the anthraquinone dye Disperse Blue 1 and curcumin at pH 5.

33 Chapter 2

When inspecting the reactivity of TfuDyP towards the tested dyes, it is not obvious why some substrates were degraded fast while others were not. Except for a general preference for anthraquinone dyes, neither the size nor the charge of the substrates seemed to have a large influence on the activity. To shed light on this, the redox potentials of the dyes were determined using cyclic voltammetry. Interestingly, the lowest redox potentials were observed for the anthraquinone dyes, E½ = 0.3 ‑ 0.65 V (Table S1). This may explain why TfuDyP is most active towards this class of dyes. Unfortunately no redox potentials could be obtained for Eosin Y or the copper phthalocyanine tetrasulfonic acid dye, the other two compounds to which TfuDyP displayed a high activity. The oxidation of the azo dyes was found to be irreversible when measuring the redox potentials and high oxidation potentials were obtained, Ep = 0.7 - 1.1 V. The observed peak potentials of the azo dyes Direct Yellow 27 and Acid Yellow 23 were both above 0.95 V, which might explain why TfuDyP could not degrade these dyes. The arylmethane dyes and Rhodamine B, showed a high observed oxidation peak potential as well, with values between 0.6 - 1.1 V.

Oxidation of lignin-related compounds Our initial study on TfuDyP already revealed that the substrate scope is not restricted to dyes. Also activity on phenolic compounds was observed by identifying guaiacol, 2,6-dimethoxyphenol and veratryl alcohol as substrates.9 As part of the current study we explored some more simple and complex phenolic compounds. Activity on several other monophenols could be confirmed: catechol, acetosyringone, syringaldehyde, vanillin, vanillyl alcohol and vanillyl acetone. Activities towards these small phenolic substrates were rather low with observed rates of 0.1 – 0.7 s-1. Yet, activity on these compounds may hint to a role of TfuDyP in delignification of plant biomass as such phenols are often described as natural mediators that are used by laccases and peroxidases.17 For vanillin and vanillin-related compounds, vanillyl alcohol and vanillyl acetone, product analysis was performed. LC-MS analysis revealed the appearance of one dominant product upon TfuDyP-catalyzed oxidation of vanillin. The formed compound could be identified as divanillin with a mass of 301.46 Da (negative mode, see Supporting Information Fig. S1, S4 and S5). Formation of divanillin from vanillin has been proposed to result from oxidative phenolic coupling and keto-enol tautomerisation to give the final product.18 For vanillyl alcohol and vanillyl acetone, also dimerization products were observed (Supporting Information Fig. S2 and S3). The selective oxidation of vanillin into divanillin may open up new avenues for the application of DyP peroxidases. Divanillin is valued as taste enhancer and efficient methods to prepare this food

34 Exploring the substrate scope of TfuDyP flavor are in demand.18,19 In addition to flavor production, vanillin oxidation may be used for polymer production because renewables-based monomers, such as furfural, 2,5-furandicarboxylic acid, and vanillin, are currently considered as polymer precursors.20 As several recent papers hint at a role of DyPs in oxidizing lignin or lignin- derived complex molecules21,22, we also investigated the activity of DyP on more complex aromatic molecules. Analysis of lignin degradation can be extremely complex. Therefore, lignin model compounds are often used to identify targets of enzyme action. In this work two model lignin dimers were tested: guaiacylglycerol-β-guaiacyl ether and veratrylglycerol-β-guaiacyl ether (Fig. 2).

Figure 2. Lignin model dimers: guaiacylglycerol-β-guaiacyl ether (A) and veratrylglycerol-β-guaiacyl ether (B).

Testing these substrates allows to discriminate between two possible degradation pathways: (1) oxidation of the phenoxy group, or (2) oxidative cleavage of the β-ether linkage, which constitutes up to 50% of the bonds in lignin. Interestingly, no peroxidase activity was detected for veratrylglycerol-β- guaiacyl ether. In contrast, for guaiacylglycerol-β-guaiacyl ether 50% substrate depletion was measured. Only the latter lignin model compound contains a phenolic moiety which suggests that TfuDyP acts on this part of the lignin dimer, using the phenolic group as electron donor. This is in contrast to the observation that a DyP from Rhodococcus jostii (DyPB) degrades lignin dimers by acting on the β-ether bond.22,23 An explanation of the observed difference in reactivity may be the fact that the two respective DyPs are representatives from two different DyP subgroups, with TfuDyP being an A-type DyP and DyPB a B-type DyP. The sequence clustering of these DyP subgroups may reflect differences in the type of reaction they catalyze. As the DyPB-catalyzed conversion of lignin results in

35 Chapter 2 the liberation of vanillin22,23, we set out to determine which products are formed upon TfuDyP-catalyzed conversion of guaiacylglycerol-β-guaiacyl. As vanillin itself is also a substrate for TfuDyP, we did not anticipate formation of this monophenol.18 Indeed, upon incubation of 0.5 mM guaiacylglycerol-β-guaiacyl ether and 1.0 mM H2O2 with TfuDyP, neither vanillin nor divanillin was observed as product. This again confirms that TfuDyP does not cleave the ether bond in lignin-derived compounds but oxidizes the phenolic moiety. Our LC-MS analyses reveal that, similar to vanillin, TfuDyP catalyzed the oxidative coupling of the guaiacylglycerol-β-guaiacyl ether resulting in several products with higher mass. Mainly dimers and trimers of guaiacylglycerol-β-guaiacyl ether were formed (masses of 661.44 and 979.62 Da in positive mode, 637.58 Da and 955.76 Da in negative mode, see Supporting Information Fig. S6-S11). As our study reveals that TfuDyP has a preference for acting on phenolic moieties, we also tested several proteins are substrates. Other oxidative enzymes have been shown to oxidize tyrosyl residues of proteins which trigger formation of cross-linked proteins. Such biocatalytic protein cross-linking methods are highly valuable in the food industry as it introduces new properties in protein- based food.24 Two model proteins were tested: β-lactoglobulin (18 kDa) and lactalbumin (14 kDa). However, incubation of these proteins with hydrogen peroxide and TfuDyP did not result in any protein cross-linking. While a small fraction of naturally occurring dimers were observed in both β-lactoglobulin and lactalbumin preparations, TfuDyP did not promote formation of additional cross-linked proteins as judged by SDS-PAGE and gel permeation analysis.

Conclusions DyPs represent a relatively poorly explored group of biocatalysts that have only been recently identified. While DyPs have been shown to act on a variety of dyes, this study provides a better view on the biocatalytic potential of a DyP. The studied TfuDyP was found to be active on a wide variety of dyes. Furthermore, it shows activity on phenolic compounds ranging from monophenols to lignin model compounds. This shows that, except for the degradation of dyes, DyPs can be used for selective polymerization reactions (e.g. production of divanillin) and modification of lignin derived compounds.

Materials and methods Chemicals and reagents Lignin model dimers were supplied by TCI Europe. Hydrogen peroxide was obtained from Merck. Reactive Blue 19, Neutral Red and indigo carmine were

36 Exploring the substrate scope of TfuDyP supplied by Acros Organics and resorufin by TCI Europe.Tert -butyl hydroperoxide (Luperox), copper phthalocyanine-3,4’,4’’,4’’’-tetrasulfonic acid tetrasodium salt and all other dyes were obtained from Sigma-Aldrich.

Enzyme purification and growth conditions TfuDyP was expressed and purified as described before5, with some modifications. TfuDyP was expressed in E. coli TOP10 (Invitrogen) using Terrific broth medium for cultivation, in order to achieve a higher cell density. An initial culture in Luria- Bertani medium was grown to saturation at 37 °C overnight. This preculture was

1:100 diluted in 1.6 L Terrific broth medium and grown at 37 °C. At OD600 nm= 0.6 the culture was induced with 0.02% L-arabinose and grown to saturation at 30 °C overnight. All cultures were supplemented with 50 µg/mL ampicillin. Purification was performed using Ni-Sepharose obtained from GE. The enzyme was eluted using a sodium-acetate buffer of pH 4.5 to avoid possible inhibition of the peroxidase by imidazole.

Thermal stability assays The ThermoFluor method was used to determine the apparent melting points of TfuDyP using an enzyme concentration of 0.5-1.0 mg/ml. This method is based on the fluorescence increase upon binding of SYPRO Orange to hydrophobic protein surfaces that become exposed upon thermal protein unfolding or multimer dissociation. The fluorescence of the SYPRO Orange dye was monitored using a RT-PCR machine (CFX-Touch, Bio-Rad). The temperature was increased with 0.5 °C per step, starting at 25 °C and ending at 99 °C, using a 10 s holding time at each step. The temperature at the maximum of the first derivative of the observed fluorescence was taken as the apparent melting temperature. Stability was assayed in 100 mM sodium acetate buffers pH 3.0-5.0, 100 mM MES buffer pH 6.0 and 100 mM Tris-HCl buffers pH 7.0-8.0.

Kinetic analyses of the oxidation of dyes The activity of TfuDyP towards seven different classes of dyes (anthraquinone, azo, arylmethane, heterocyclic, xanthene, indigoid and phthalocyanine dyes), three carotenoid pigments and hemin was measured spectrophotometrically (Jasco V-660). Reaction mixtures containing 50 mM citrate buffer pH 3.0, 4.0 or 5.0, supplemented with 50 µM dye, 30-300 nM purified enzyme and 100 µM

H2O2 were used. In case the absorbance of a dye at λmax was too high, a lower concentration of 25 or 10 µM was used. The enzyme was added to start the reaction. First the initial rate of oxidation was measured at the corresponding wavelength maximum for each dye (Table 1). Reactions were subsequently incubated for 1h

37 Chapter 2 at ambient temperature after which spectra between 350-750 nm were taken to estimate the level of degradation, calculated as percentage compared to the starting solution. Control reactions were included without enzyme. Thetwo dyes Disperse Blue 1 and resorufin, the two pigments β-carotene and curcumin are poorly water soluble. Therefore, stock solutions were prepared in DMSO and added to the reaction mixtures to a final concentration of 2.5% or 10% DMSO. The initial activity of TfuDyP in the presence of 10% DMSO was tested using Reactive Blue 19 as substrate and revealed that the enzyme remained 40% of its activity. The stock solution of porphyrin hemin was obtained by dissolving hemin to a concentration of 5.0 mM in 0.2 M NaOH and used at a final concentration of 50 µM in buffer.

Cyclic voltammetry Redox potentials of the dyes (when soluble without DMSO) were measured using cyclic voltammetry (CH-Instruments, Electrochemical analyzer CHI630B, USA). A Saturated Calomel Electrode (SCE) (BAS Inc, Japan) was used as reference electrode, glassy carbon as working electrode and a platinum wire as counter electrode. Scans were taken between -0.4 and 1.1 V at ambient temperature. Dyes were dissolved in a 0.1-0.2 M citric acid buffer of pH 3-5 to a concentration of 0.25-1.0 mM. The redox potential of a dye was analyzed at the pH at which TfuDyP was most active towards that dye.

Oxidation of small phenolics Catechol, sinapic acid, syringaldehyde and acetosyringone were tested in different concentrations and oxidation was followed spectrophotometrically. Oxidation of vanillin, vanillyl alcohol and vanillyl acetone was performed in 1 mL reactions consisting of 90 nM TfuDyP, 0.5 mg/mL substrate and 1.0 mM H2O2 in 100 mM citric acid buffer pH 4.0. Reactions were incubated at 30º C for 150 min and analyzed by HPLC and LC-MS.

By TfuDyP catalyzed degradation of lignin and lignin model compounds Lignin model compounds, guaiacylglycerol-β-guaiacyl ether and veratrylglycerol- β-guaiacyl ether, were dissolved in DMSO at a concentration of 10 mM. The reaction mixtures contained 2.0 µMTfu DyP, 0.20 mM lignin dimer and 0.10 mM

H2O2 in 100 mM Na-acetate buffer pH 3.5. Reactions were incubated at 30 ºC for 1h. Control reactions were made with H2O2 and buffer mixed with substrates and also with enzyme and substrate without H2O2. Reactions was incubated for 150 min at 30 ºC, 500 rpm and samples were then centrifuged for 10 min at 13,000 rpm. Products of the reaction of TfuDyP with lignin model compounds

38 Exploring the substrate scope of TfuDyP and other small phenolics were analyzed by reverse phase HPLC using Jasco HPLC system. 10 µL samples were injected on a Grace Altima HP C18 column (3 µm, 4.6x100 mm, with precolumn of same material). Solvents used: A water with 0.1% formic acid and B acetonitrile. HPLC method: 2 min 15% B, 2-16 min 80% B, 14-16 min 80% B, 16 min 15% B followed by re-equilibration. Detection with UV detector at 254 nm and flow rate of 0.5 mL/min. LC-MS analysis was performed on Surveyor HPLC-DAD coupled to LCQ Fleet detector.

Cross linking of proteins Beta-lactoglobulin and lactalbumin (final concentration 1 mg/mL) were tested as model proteins for TfuDyP mediated cross-linking in the presence and absence of syringaldehyde (0.2 mM). Reactions were carried out in 100 mM Na-acetate buffer pH 3.5 with 0.10 mM 2H O2 at 30 ºC with shaking for 4h. Samples were analyzed by gel permeation on a Superdex 200 column and by SDS-PAGE.

Acknowledgements We thank Prof. Dr. W.R. Browne, Molecular Inorganic Chemistry, University of Groningen, The Netherlands, for his help with the cyclic voltammetry experiments. Nikola Lončar was financially supported by the Serbian Ministry of Education, Science and Technological development through the project ON172048 and by COST action Systems biocatalysis CM1303. Dana I. Colpa was financially supported by the NWO graduate program: synthetic biology for advanced metabolic engineering, project number 022.004.006, The Netherlands.

Supporting information The supporting information includes the activity measurements forTfu DyP on the twenty seven dyes from Table 2 at pH 3, 4 and 5 (Table S1). Figures S1-S11 show the HPLC and LC-MS results of the oxidation of vanillin, vanillyl alcohol, vanillyl acetone and guaiacylglycerol-β-guaiacyl ether by TfuDyP. This information can be found online, connected to the publication: DOI: 10.1016/j.tet.2015.12.078.

39 Chapter 2

References

1 Battistuzzi G, Bellei M, Bortolotti CA, and Sola M (2010) Redox properties ofheme peroxidases. Arch. Biochem. Biophys. 500, 21–36. 2 Vigneswaran C, Ananthasubramanian M, and Kandhavadivu P (2014) 5 - Enzymes in textile effluents. In Bioprocessing of Textiles pp. 251–298. Woodhead publishing India PVT LTD, New Delhi. 3 Azevedo AM, Martins VC, Prazeres DMF, Vojinovic V, Cabral JMS, and Fonseca LP (2003) Horseradish peroxidase: a valuable tool in biotechnology. Biotechnol. Annu. Rev. 9, 199– 247. 4 Krainer FW and Glieder A (2015) An updated view on horseradish peroxidases: recombinant production and biotechnological applications. Appl. Microbiol. Biotechnol. 99, 1611–1625. 5 Yoshida T and Sugano Y (2015) A structural and functional perspective of DyP-type peroxidase family. Arch. Biochem. Biophys. 574, 49–55. 6 Kim SJ and Shoda M (1999) Purification and characterization of a novel peroxidase from Geotrichum candidum Dec 1 involved in decolorization of dyes.Appl. Environ. Microbiol. 65, 1029–1035. 7 Colpa DI, Fraaije MW, and van Bloois E (2014) DyP-type peroxidases: a promising and versatile class of enzymes. J. Ind. Microbiol. Biotechnol. 41, 1–7. 8 Lykidis A, Mavromatis K, Ivanova N, Anderson I, Land M, DiBartolo G, Martinez M, Lapidus A, Lucas S, Copeland A, Richardson P, Wilson DB, and Kyrpides N (2007) Genome sequence and analysis of the soil cellulolytic actinomycete Thermobifida fusca YX. J. Bacteriol. 189, 2477–2486. 9 van Bloois E, Torres Pazmiño DE, Winter RT, and Fraaije MW (2010) A robust and extracellular heme-containing peroxidase from Thermobifida fusca as prototype of a bacterial peroxidase superfamily. Appl. Microbiol. Biotechnol. 86, 1419–1430. 10 Sugano Y, Muramatsu R, Ichiyanagi A, Sato T, and Shoda M (2007) DyP, a unique dye- decolorizing peroxidase, represents a novel heme peroxidase family: ASP171 replaces the distal histidine of classical peroxidases. J. Biol. Chem. 282, 36652–36658. 11 Pesic M, López C, López-Santín J, and Álvaro G (2013) From amino alcohol to aminopolyol: one-pot multienzyme oxidation and aldol addition. Appl. Microbiol. Biotechnol. 97, 7173– 7183. 12 Yu W, Liu W, Huang H, Zheng F, Wang X, Wu Y, Li K, Xie X, and Jin Y (2014) Application of a novel alkali-tolerant thermostable DyP-type peroxidase from Saccharomonospora viridis DSM 43017 in biobleaching of Eucalyptus Kraft pulp. PLoS One 9, e110319. 13 Scheibner M, Hülsdau B, Zelena K, Nimtz M, de Boer L, Berger RG, and Zorn H (2008) Novel peroxidases of Marasmius scorodonius degrade β-carotene. Appl. Microbiol. Biotechnol. 77, 1241–1250. 14 Ulson de Souza SMAG, Forgiarini E, and Ulson de Souza AA (2007) Toxicity of textile dyes and their degradation by the enzyme horseradish peroxidase (HRP). J. Hazard. Mater. 147, 1073–1078. 15 Bilal M and Asgher M (2015) Sandal reactive dyes decolorization and cytotoxicity reduction using manganese peroxidase immobilized onto polyvinyl alcohol-alginate beads. Chem. Cent. J. 9, 1–14. 16 Ogola HJO, Kamiike T, Hashimoto N, Ashida H, Ishikawa T, Shibata H, and Sawa Y (2009) Molecular characterization of a novel peroxidase from the cyanobacterium Anabaena sp. strain PCC 7120. Appl. Environ. Microbiol. 75, 7509–7518. 17 Nousiainen P, Kontro J, Manner H, Hatakka A, and Sipilä J (2014) Phenolic mediators enhance the manganese peroxidase catalyzed oxidation of recalcitrant lignin model compounds and synthetic lignin. Fungal Genet. Biol. 72, 137–149. 18 Nishimura RT, Giammanco CH, and Vosburg DA (2010) Green, enzymatic syntheses of

40 Exploring the substrate scope of TfuDyP

divanillin and diapocynin for the organic, biochemistry, or advanced general chemistry laboratory. J. Chem. Educ. 87, 526–527. 19 Krings U, Esparan V, and Berger RG (2015) The taste enhancer divanillin: a review on sources and enzymatic generation. Flavour Fragr. J. 30, 362–365. 20 Amarasekara AS and Razzaq A (2012) Vanillin-based polymers - part II: Synthesis of Schiff base polymers of divanillin and their chelation with metal ions. ISRN Polym. Sci. 2012, 1–5. 21 Salvachúa D, Prieto A, Martínez ÁT, and Martínez MJ (2013) Characterization of a novel dye-decolorizing peroxidase (DyP)-type enzyme from Irpex lacteus and its application in enzymatic hydrolysis of wheat straw. Appl. Environ. Microbiol. 79, 4316–4324. 22 Ahmad M, Roberts JN, Hardiman EM, Singh R, Eltis LD, and Bugg TD (2011) Identification of DypB from Rhodococcus jostii RHA1 as a Lignin Peroxidase. Biochemistry 50, 5096–5107. 23 Sainsbury PD, Hardiman EM, Ahmad M, Otani H, Seghezzi N, Eltis LD, and Bugg TDH (2013) Breaking down lignin to high-value chemicals: The conversion of lignocellulose to vanillin in a gene deletion mutant of Rhodococcus jostii RHA1. ACS Chem. Biol. 8, 2151–2156. 24 Heijnis WH, Dekker HL, de Koning LJ, Wieringa PA, Westphal AH, de Koster CG, Gruppen H, and van Berkel WJH (2011) Identification of the peroxidase-generated intermolecular dityrosine cross-link in bovine α-lactalbumin. J. Agric. Food Chem. 59, 444–449.

41 42 Chapter 3:

Exploring the catalytic properties of DyP- type peroxidase TfuDyP by site-directed mutagenesis

Dana I. Colpa, Thomas Hilberath, Bastiaan de Wit and Marco W. Fraaije Abstract In this work we used site-directed mutagenesis to establish the role of various residues in catalysis by TfuDyP. Different point mutations in the heme pocket, next to heme propionate and in the predicted hydrogen peroxide tunnel were studied, together with mutations that were designed to shift the pH optimum of TfuDyP. Also mutations that might give information on whether the activity of TfuDyP is based on long-range electron transfer were analyzed. Mutagenesis of the fifth heme ligand (H338C), the catalytically active aspartate (D242N) and the residue next to this aspartate (N236K) underlined the importance of these residues in catalysis. These variants yielded soluble proteins that bound the heme cofactor but were not active towards Reactive Blue 19. Two other mutations highlighted the importance of the area around the heme propionate and catalytically active arginine on the pH range/optimum for activityTfu of DyP. A245R and N246L broadened the pH range for activity towards Reactive Blue 19 by one pH unit. The dependence of TfuDyP on long-range electron transfer for the oxidation of bulky substrates was studied by individually mutating the surface exposed tyrosines and tryptophans to phenylalanine. None of the tyrosine or tryptophan mutants lost its activity, indicating that TfuDyP is either flexible to use different tyrosines/tryptophans or that TfuDyP is not depended on long- range electron transfer for the oxidation of bulky substrates. This site-directed mutagenesis study has provided a better insight in how a DyP-type peroxidase performs catalysis and provides leads for future engineering of this family of biotechnologically interesting peroxidases. Exploring the catalytic properties of TfuDyP by site-directed mutagenesis

Introduction The family of DyP-type peroxidases was discovered relatively recently.1 The first member of this family was discovered two decades ago and named after its ability to degrade dyes: a dye decolorizing peroxidase (DyP).2 Members of the DyP-type peroxidase family have the advantage over plant/fungal peroxidases that they are mainly found in bacteria and are easily heterologously overexpressed in a bacterial host, e.g. Escherichia coli.3–8 The family of DyP-type peroxidases is divided in four subclasses by PeroxiBase based on sequence homology.9 Classes A-C are predominantly formed by bacterial enzymes while members of class D are mainly found in fungi. Class A is formed by Tat-dependently exported DyPs, while classes B and C are intracellular enzymes of which the smallest family members are found in class B. Classifications with three or five subclasses were also reported.10,11 Several structurally and catalytically important residues were identified in DyP-type peroxidases.3,10–12 A histidine is found as the fifth ligand of the non- covalently bound heme cofactor, similar to peroxidases from the peroxidase- catalase superfamily.13 However, DyP-type peroxidases, which belong to the peroxidase-chlorite dismutase superfamily, differ from the peroxidase-catalase superfamily on the distal side of the heme. For the large family of peroxidase- catalases a histidine is found as acid-base catalyst together with an essential arginine13–15, while for DyP-type peroxidases an aspartate and an arginine are found. The aspartate is part of the highly conserved GXXDG-motif and functions (in most DyPs) as acid-base catalyst.1,12 The pH optimum for activity of DyPs is in the acidic range and is (partially) defined by this residue.6,8,12 For most DyP-type peroxidases both the aspartate and arginine are important for the activity of the enzyme.6–8,16–19 However, for DyPB from Rhodococcus jostii RHA1 (class B) only the arginine was reported to be essential.5 The role of these two residues on the activity of DyP-type peroxidases with respect to different substrates is still unclear. For EfeB/YcdB from Escherichia coli O157 (class A) it was shown that D235N had a strongly reduced activity towards catechol but was fully active towards guaiacol, while mutants of R347 showed a strongly reduced activity towards both substrates.17 Mutants of D220 in TcDyP from Thermomonospora curvata (class A) retained up to 85% of the wild-type activity towards guaiacol, while activity towards ABTS was lost.7 D142 mutants of YfeX from Escherichia coli O157 (class B) showed almost no activity, while mutant R232L lost its activity only towards Reactive Blue 19 and ABTS but was still for 50% or more active towards guaiacol and catechol.19 This suggests that DyP-type peroxidases oxidize different substrates via different pathways. The substrate scope of DyP-type peroxidases is diverse and covers

45 Chapter 3 monophenolic compounds, bulky dyes and lignin model compounds.2,20–22 Substrate access to the active site is limited in DyP-type peroxidases. Small molecules, such as hydrogen peroxide, may enter the active site via a narrow channel on the distal side of the heme or potentially via the heme propionate pocket (in case the heme propionate is surface exposed).19,23,24 For two DyPs, TfuDyP and engineered AauDyP from Auricularia auricula-judae (class D), peroxygenase activity in the form of enantioselective sulfoxidation ofsmall aromatic sulfides was reported.4,25 For enantioselective sulfoxidation substrates need to enter the active site and react with the oxoferryl heme species directly. However, bulky compounds such as dyes and lignin model compounds are too big to enter the active site of a DyP-type peroxidase. These compounds are proposed to react at an oxidation site on the protein surface. Substrate binding sites on the protein surface were observed for various peroxidases. Crystal structures of multiple plant and mammalian peroxidases showed substrate binding at the δ-edge of the heme.26 For other enzymes, e.g. manganese and ascorbate peroxidases, enzyme-substrate complexes with substrates bound at the surface exposed heme propionate (γ-edge of the heme) were revealed.26–28 And lastly, bulky substrates could be oxidized by a surface exposed radical oxidation site formed via long-range electron transfer (LRET). Tyrosines and tryptophans are redox-active amino acids that are able to propagate and stabilize a radical.29,30 LRET-pathways and surface exposed radical sites were observed in lignin and versatile peroxidases.31–33 In DyP-type peroxidases two potential surface exposed oxidation sites were noted. DyP from Bjerkandera adusta Dec 1 (class D) was crystalized with ascorbic acid and 2,6-dimethoxyphenol bound to the heme propionate pocket (the γ-edge of the heme).23 Yoshida et al and Roberts et al proposed that the heme propionate pocket might be the oxidation site of larger substrates, in line with ascorbate and manganese peroxidases.23,24,27,28 Furthermore, LRET-pathways were reported for three DyP-type peroxidases: AauDyP (class D), TcDyP (class A) and VcDyP from Vibrio cholerae (class B).6,29,34 The aim of this study is to gain a better understanding of the catalytic machinery of DyP-type peroxidase TfuDyP. TfuDyP, a thermostable dimeric class A DyP-type peroxidase from Thermobifida fusca, is active on small phenolic compounds, bulky dyes and displays peroxygenase activity in the enantioselective sulfoxidation of aromatic sulfides.4,20 In earlier work we showed the importance of the fifth heme ligand H338 and catalytically active D242 of TfuDyP.4 H338 is required for proper heme binding. When this histidine was mutated to an alanine no Soret-band was observed in UV-vis absorbance spectra and the activity towards Reactive Blue 19 was reduced to 3%.4 D242 is part of the conserved GXXDG-motif of DyP-type peroxidases and is crucial for the activity of

46 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis the enzyme. Mutant D242A was able to bind the heme cofactor but the activity towards Reactive Blue 19 was almost fully lost (0.7% activity).4 Besides the catalytic aspartate also an arginine was shown to be essential for the activity of various DyP-type peroxidases.6–8,16–19 Rahmanpour et al solved the crystal structure of TfuDyP which confirmed the close proximity of the Asp and Arg to the heme cofactor.16 In agreement with the work of van Bloois4, Rahmanpour observed a strongly reduced enzyme activity for mutant D242A (D203A in truncatedTfu DyP used by Rahmanpour). D242A was not active towards phenolic compounds and showed reduced activity towards ABTS compared to the wild-type enzyme. Mutant R354Q (R315Q in truncated TfuDyP) was not active towards ABTS nor towards phenolic substrates. To gain a better understanding ofTfu DyP the effect of a variety of rationally designed point mutations were studied in this work. First, mutations were made around the heme propionate, in the H2O2-tunnel and in the heme pocket. In the heme pocket the fifth heme ligand, a histidine, was mutated to a cysteine as is found in heme-thiolate enzymes such as peroxygenases and cytochrome P450 monooxygenases.35,36 With this mutation we aimed to enhance the peroxygenase activity of TfuDyP. Furthermore, mutations were made in the heme pocket with the aim to shift the pH optimum of TfuDyP from an acidic to a more neutral pH range. Mutations that were proven beneficial in shifting the pH optimum of PpDyP from Pseudomonas putida MET94 (class B) were applied on TfuDyP.8 Lastly we studied the dependence of TfuDyP on long-range electron transfer (LRET) for the oxidation of bulky substrates. In order to study this, we mutated the surface exposed tyrosines and tryptophans individually to phenylalanine. Substitution to phenylalanine is a conservative mutation, but Phe is unlike Trp/ Tyr not able to stabilize a radical.29 We also made a knock-out mutant in which all five tyrosines and seven tryptophans ofTfu DyP were mutated to phenylalanine.

Results and discussion Enzyme expression and purification Wild-type TfuDyP and mutants thereof were heterologously overexpressed in Escherichia coli TOP10. All enzymes were soluble and purified by His-tag affinity chromatography with yields of 20 to 35 mg enzyme per liter culture broth and a Reinheitszahl (Rz-value) of 1.0-1.5. Most enzymes were brown-red due to the heme cofactor and showed an UV-vis absorbance spectroscopy profile comparable to the wild-type protein, with a protein peak at 280 nm, a Soret band around 410 nm and one or two Q-bands between 450-600 nm depending on the oxidation state of the enzyme (Fig. 1).4 However, different features were

47 Chapter 3 observed for mutants H338C and N236K. Purified TfuDyP N236K was green while mutant H338C was yellow. These differences were also visible in UV-vis absorbance spectroscopy (Fig. 1).

0.3

0.2

WT N236K H338C

Absorbance 0.1

0 250 350 450 550 650 750 Wavelength (nm)

Figure 1. UV-vis absorbance spectroscopy profile of purified enzyme TfuDyP wild-type (solid line) and mutants N236K (dashed line) and H338C (dotted line). The inset shows the color of the enzymes, from left to right: wild-type (brown-red), N236K (green) and H338C (yellow).

Mutation of heme ligand H338 to cysteine TfuDyP is the only native DyP reported to date to perform enantioselective sulfoxidations.4 Even though the overall conversion was low, TfuDyP was shown to enantioselectively sulfoxidize four aromatic sulfides of which thioanisole (methyl phenyl sulfide) was the best substrate with an enantiomeric excess (ee‑value) of 61% for the R-sulfoxide. This peroxygenase activity clearly showed that some substrates access the active site ofTfu DyP and react with the oxoferryl heme directly, an activity which is typical for peroxygenases and cytochrome P450 monooxygenases.35,36 Recently, AauDyP from Auricularia auricula-judae was successfully engineered to perform enantioselective sulfoxidations. The F359G variant of AauDyP gave a 95-99% conversion of thioanisole and methyl- p-tolyl sulfide with up to 99% excess of the S-products.25 Peroxygenases and cytochrome P450 monooxygenases are heme-thiolate enzymes, with a proximal cysteine functions as fifth ligand to the heme iron instead of a histidine as is found in DyP-type peroxidases.1,35,36 To enhance the sulfoxidation activity ofTfu DyP we

48 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis mutated the proximal heme ligand H338 to a cysteine. TfuDyP variant H338C was purified as a soluble enzyme that bound, in contrast to an earlier studied variant H338A, a heme cofactor.4 The enzyme was however yellow instead of red and showed an UV-vis absorbance spectrum which significantly differed from the wild-type enzyme (Fig. 1). For TfuDyP H338C a broader Soret band, a lower Rz-value (0.4) and no Q-bands were observed. The observed UV-vis spectrum is comparable to the results observed for TcDyP variant H312C, which also showed a broader Soret band and a lower Rz-value when compared to the wild-type enzyme.7 TcDyP variant H312C not only bound the heme cofactor, it showed some residual activity towards ABTS (0.5%), hydroquinone (HQ, 1.1%) and guaiacol (5.4%).7 TfuDyP H338C did not show any significant residual activity towards Reactive Blue 19 or in the (enantioselective) sulfoxidation of thioanisole.

Effect of mutations in the predicted hydrogen peroxide tunnel TfuDyP and other DyP-type peroxidases have a narrow tunnel from the surface of the enzyme to the heme iron.19 We made two mutations in this tunnel to make the tunnel narrower and more hydrophobic. G243 from the highly conserved GXXDG-motif and G356 were mutated to alanine. The modification of G243 to alanine reduced the initial activity towards Reactive Blue 19 at pH 3.5 to 15% of the wild-type activity (Fig. 2). Mutation G356A had a smaller effect on the enzyme: the activity decreased to 53% of the wild-type activity. Remarkably, the sulfoxidation of thioanisole was unaffected by both mutations: boththe conversion in two hours and the enantioselectivity were comparable to the wild- type enzyme (Table 1).

Table 1: Conversion of thioanisole by different mutants of TfuDyP.

Mutant ee R-product (%)a Conversion (%)b WT 77 23 G243A 79 23 A245R 79 37 N246L 77 20 H338C - - G356A 81 20

[a] The enantiomeric excess (ee) of the R-sulfoxide was calculated by ee = (AR – AS)/(AR + AS) x 100%, in which the peak areas (A) of the R- and S-sulfoxide products were determined by chiral GC. [b] Conversion of thioanisole (S) after incubation at 30 ºC, 100 rpm for two hours. Conversion

= (1 - ([S]f/[S]i)) x 100%.

49 Chapter 3

9.0 pH 8.0 2.5 7.0 3.0 3.5

6.0 4.0 ) 1

- 4.5 5.0 (s 5.0

obs 4.0 k 3.0 2.0 1.0 0.0 WT G243A G356A A245R N246L Figure 2. Initial activity ofTfu DyP wild-type and mutants G243A, G356A, A245R and N246L towards Reactive Blue 19 at pH 2.5-5.0.

D242N and N246L, designed to shift the pH optimum of TfuDyP, are also located in the predicted hydrogen peroxide tunnel. Mutant N246L, which forms a hydrogen bond with the backbone of R354 and makes the entrance of the hydrogen peroxide tunnel more hydrophobic, shifted the pH range for activity slightly (Fig. 2). Mutant D242N was inactive.

Effect of mutations in the heme propionate pocket One possible site for substrate binding and oxidation in TfuDyP is the surface exposed heme propionate pocket (Fig. 3). Substrate binding in such a pocket is well known for various heme peroxidases. Manganese and ascorbate peroxidases were for instance crystallized in complex with their substrates bound to the heme propionate pocket.26–28 Even though DyP-type peroxidases are structurally unrelated to these peroxidases, a surface exposed heme propionate is found in many DyP-type peroxidases.16,23,24,29 To study the importance of this pocket to TfuDyP we mutated an alanine directly next to the surface exposed heme propionate to positively charged arginine, negatively charged glutamate and hydrophobic valine. The initial activity of mutants A245E and A245V towards Reactive Blue 19 at pH 2.5-3.5 decreased slightly to 50-80% of the wild-type activity. Interestingly, variant A245R was active at a broader pH range. The wild- type enzyme is active towards Reactive Blue 19 at pH 2.5-4.0, while mutant A245R is active up to pH 5.0 (Fig. 2). The activity at pH 4.0 increased fourfold and

50 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis No side chains initial activities of respectively 0.31 and 0.11 s-1 were measured at pH 4.5 and 5.0. The conversion of thioanisole by A245R was increased significantly while the mutant enzyme retained its enantioselectivity (Table 1).

D242 R354 A245

H338

Figure 3. Surface exposed heme propionate pocket of TfuDyP. Heme and residues D242, A245, H338 and R354 are shown in sticks. The heme cofactor is shown in orange. Heme ligand H338, residue A245 and catalytically active D242 and R354 are shown in green. The active site and the position of A245 are shown on the left while the protein surface is shown on the right. The picture is based on a model created by Yasara, the crystal structure (PDB: 5fw4) is missing a loop in this area.

Broadening the pH range for activity DyP-type peroxidases show a pH optimum for activity in the acidic range.2–4,6–8 For the applicability of these enzymes in for instance cascade reactions with other enzymes it would be desirable to shift the pH optimum to a more neutral pH range.37 Shifting the pH optimum through site-directed or random mutagenesis was shown before to be successful for various enzymes.8,38–40 Mendes et al shifted the pH optimum of PpDyP (class B) by four pH units by mutagenesis of the catalytically active aspartate to an asparagine (D132N, GXXDG).8 The pH optimum of PpDyP D132N was shifted from 4.3 to 7.4 for ABTS as substrate. Mutagenesis of N136L, a residue that forms a hydrogen bond with the backbone of the catalytically active arginine, had a smaller effect and shifted the pH optimum to 5.6. Mutagenesis of the arginine (R214L) itself had an opposite effect; it shifted the pH optimum to 3.6. These results are in line with

the pKa of the side chains of these residues: aspartate and arginine have a pKa of respectively 3.9 and 12.5 in water.6,8 Mutagenesis of the same residues in BsDyP (class A) had a smaller effect: D240N shifted the pH optimum for activity on ABTS

51 Chapter 3 from 3.8 to 4.4.41 Brissos et al shifted the pH optimum for activity Ppof DyP even further by random mutagenesis to a pH of 8.5.40 We aimed to shift the pH optimum for activity of TfuDyP to a higher pH range by applying the mutations that were proven beneficial for PpDyP and BsDyP on TfuDyP. The catalytically active aspartate was mutated to an asparagine (D242N) and N246 was mutated to a leucine. N246 forms a hydrogen bond with the catalytically active arginine (R354). The mutations discovered by Brissos et al could not be applied on TfuDyP: the sequence identity between TfuDyP and PpDyP is only 29%. Besides these two mutations we mutated the residue next to the catalytically active aspartate to a lysine to potentially shift the pKa of this aspartate to a higher range (N236K, pKa of lysine is 10.5 in water). In contrast to the results obtained for PpDyP and BsDyP, mutagenesis of the catalytically active aspartate of TfuDyP (D242N) resulted in an inactive enzyme.8,41 TfuDyP N246L broadened the pH range for activity towards Reactive Blue 19 slightly (Fig 2). It showed 51% of the wild-type activity towards Reactive Blue 19 at pH 3.5. The initial activity at pH 4.0 increased threefold and at pH 4.5 and 5.0 initial activities of 0.93 and 0.11-1 s were measured, respectively. TfuDyP N236K showed different features, the enzyme was green, showed an Rz-value of 0.6 and a plateau between 450-700 nm in the UV-vis absorbance spectrum (Fig. 1). The color of mutant N236K might be shifted to green due to a change in the heme environment by the positively charged lysine directly above the heme, or due to a heme modification.42–44 TfuDyP N236K was not activity towards Reactive Blue 19.

Surface exposed tyrosines and tryptophans Small (phenolic) substrates can enter the active site of DyP-type peroxidases and react with the oxoferryl heme directly. Direct evidence for such a mechanism is the enantioselective sulfoxidation of thioanisole and related small aromatic sulfides by TfuDyP and engineered AauDyP.4,25 Large substrates, e.g. dyes and lignin model compounds, are however too large to enter the active site. For AauDyP (class D), TcDyP (class A) and VcDyP (class B) long-range electron transfer pathways between the heme and surface exposed tyrosines and tryptophans were demonstrated.6,29,34,45 Tyrosines and tryptophans are the only amino acids that can stabilize and propagate a radical.29,30 In AauDyP and TcDyP mainly a tryptophan is used while in VcDyP mainly a tyrosine is used. After mutagenesis of these residues, radicals were formed by another surface exposed Tyr/Trp. We studied the effect of mutagenesis of all surface exposed tyrosines and tryptophans of TfuDyP. TfuDyP contains in total five tyrosines and seven tryptophans of which four tyrosines and four tryptophans are partially surface

52 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis exposed and within 20 Å of the heme cofactor (Fig. 4 and Supporting Information Table S1). These eight residues were individually mutated to phenylalanine. Phenylalanine was chosen because of its resemblance to tyrosine and tryptophan by being an aromatic residue, it is however unable to form a stable radical.29 Individually mutating the eight surface exposed tyrosines and tryptophans that are within 20 Å from the heme had no or only a small effect on the expression level, the heme content and the activity of TfuDyP towards Reactive Blue 19 at pH 3.0 (Fig. 5). From these results it was concluded that individual tyrosines or tryptophans are not solely responsible for the oxidation of large substrates. In case TfuDyP uses long-range electron transfer for the oxidation of bulky substrates, multiple tyrosines/tryptophans are involved and TfuDyP is capable of using a different pathway if one of the Tyr/Trp is knocked-out. These results are in line with the results recorded for AauDyP, VcDyP and TcDyP: knocking out one Tyr/Trp did not completely abolish the activity of these enzymes.6,29,34,45 A knock-out mutant of TfuDyP in which all twelve tyrosines and tryptophans were substituted with a phenylalanine could not be expressed. Further research is required to study whether the oxidation of bulky substrates byTfu DyP is based on LRET.

Figure 4. Protein structure of TfuDyP visualizing the heme cofactor in red and the two monomeric units in gray and green. The five tyrosines and seven tryptophans of TfuDyP are shown in orange and blue sticks respectively. The picture is based on the crystal structure with pdb code: 5fw4.16 The numbering in this figure is based on the full protein sequence.

53 Chapter 3

140 pH 3.0 120 7.0

6.0 100 5.0 80 4.0 60 3.0 40 2.0

20 1.0 (U/mg) activity Specific

Residual activity relative to WT (%) WT to relative activity Residual 0 0.0 WT Y276F Y357F Y359F Y403F W179F W215F W289F W363F

Figure 5. Initial specific activity of TfuDyP mutants towards Reactive Blue 19 at pH 3.0.

Linde et al compared the tyrosine and tryptophan content of basidiomycete DyPs to other ligninolytic peroxidases (versatile peroxidases (VP) and lignin peroxidases (LiP)) from the same organisms.46 VPs and LiPs show a low number or even the absence of tyrosines, in contrast to DyPs. Low abundance of tyrosines protects the enzyme from oxidation damage due to the formation and the subsequent (coupling) reactions of tyrosyl radicals.46 From these results, Linde et al presumed that DyPs function in a less oxidizing environment than VPs or LiPs. LRET-pathways and surface exposed radical sites have however been recorded for lignin and versatile peroxidases.31–33 The lower number of Tyr/Trp in VPs and LiPs presumably promotes one specific LRET- pathway on the one hand and protects the enzyme from oxidation damage on the other hand. We compared the tyrosine and tryptophan content of thirty-eight well-studied DyP-type peroxidases from all four subclasses to the average amino acid composition of proteins.47–49 The thirty-eight DyPs contained on average 1.2 ± 0.5% tryptophans and 1.9 ± 0.9% tyrosines (Supporting Information Table S2). When compared to the average values recorded for prokaryotic/eukaryotic intra/extracellular proteins (1.01-1.44% tryptophans and 2.60-4.03% tyrosines), DyP-type peroxidases show an average tryptophan content and a slightly lower percentage of tyrosines (Supporting Information Table S3). These results show no indication for a relatively high or low tyrosine or tryptophan content in DyPs which would be expected when LRET-pathways are a main contributor to the catalysis of DyP-type peroxidases.

Conclusions In this work we studied the effect of seventeen rationally designed point mutations and one 12‑fold mutant in TfuDyP (Fig. 6). Mutations were made

54 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis

in the heme pocket, in the H2O2-tunnel, next to the heme propionate, and with the aim to shift the pH optimum of TfuDyP. Furthermore, mutations were made to study the dependence of TfuDyP on LRET for the oxidation of bulky substrates. All designed TfuDyP variants were overexpressed as soluble enzymes and bound the heme cofactor. Mutagenesis of the fifth heme ligand (H338C), of the catalytically active aspartate (D242N) and of the residue next to this aspartate (N236K) had a great impact on enzyme activity. Although these mutant enzymes still bound the heme cofactor, the heme environment was significantly changed and these mutants were inactive towards Reactive Blue 19. The usually brown-red color of TfuDyP was shifted to yellow and green for H338C and N236K, respectively. Variant H338C was designed to increase the peroxygenase activity of TfuDyP as it was inspired on peroxygenases and cytochrome P450 monooxygenases in which a cysteine is found as the fifth heme ligand.35,36 TfuDyP H338C did not show any significant activity towards Reactive Blue 19 or thioanisole. To increase the peroxygenase activity ofTfu DyP in future work, mutations that increased the peroxygenase activity of AauDyP could be introduced in TfuDyP25, or the predicted hydrogen peroxide tunnel could be broadened to give the aromatic substrates a better access to the heme iron.

Figure 6. Protein structure of TfuDyP in which the heme cofactor is shown in sticks and the Cα atoms of the mutated residues are shown in spheres. Heme ligand H338A is shown as a red sphere, mutations N236K, D242N, G243A, A245(E/V/R), N246L, G356A are shown in cyan. All five tyrosines (yellow spheres) and seven tryptophans (blue spheres) were mutated to phenylalanine in the 12-fold mutant; eight of these mutations were also studied individually. The figure is based on the crystal structure with pdb code: 5fw4, chain A.16

55 Chapter 3

In order to shift the pH-optimum of TfuDyP to a more neutral pH range, mutations that were proven beneficial in shifting the pH-optimumPp of DyP and BsDyP were introduced in TfuDyP.8,41 However, the mutation that shifted the pH- optimum ofPp DyP by four pH units (D132N, in TfuDyP D242N) was not beneficial for TfuDyP. TfuDyP D242N was not active towards Reactive Blue 19. Variant N246L, which forms a hydrogen bond with the backbone of the catalytically active arginine, had a small and similar effect on the pH range for activity as observed for BsDyP and PpDyP.8,41 In PpDyP and BsDyP this mutation shifted the pH optimum slightly, in TfuDyP it broadened the pH range for activity towards Reactive Blue 19 from pH 2.5-4.0 to 2.5-5.0.8,41 A similar broadening of the pH range for activity was observed for a mutant of the residue next to N246, A245R which is located in the heme propionate pocket. This area around the heme propionate and close to the catalytically active arginine is important for the pH optimum for activity of TfuDyP. In future work, the pH optimum for activity of TfuDyP might be shifted to a more neutral pH range by exploring a library of TfuDyP mutants based on random mutagenesis around the heme cofactor, including the area around the surface exposed heme propionate. Two mutations were made in the predicted hydrogen peroxide tunnel to analyze the effect of making the tunnel narrower: G243, which is part of the highly conserved GXXDG-motif of DyP-type peroxidases, and G356 were mutated to alanine. The activity of variants G243A and G356A towards Reactive Blue 19 was reduced to respectively 15 and 53% of the wild-type activity at pH 3.5. This result underlines the importance of the conserved motif and/or the presence of a glycine at position 243 in the narrow hydrogen peroxide tunnel of TfuDyP. Mutation N246L, which broadened the pH range for activity towards Reactive Blue 19 by one unit, is located in the hydrogen peroxide tunnel and makes this tunnel more hydrophobic. Besides broadening the pH range for activity it reduced the activity at the optimal pH, pH 3.5, to 51%. The enantioselectivity and the total conversion of thioanisole in two hours were not influenced by the mutations in the hydrogen peroxide tunnel. This is remarkable since thioanisole is expected to access the heme pocket via the same tunnel to react with the oxoferryl heme. Another possibility might be that this substrate enters the heme pocket via the heme propionate pocket. Mutations that influence/reduce the activity towards one type of substrates, but not towards another type, were observed before for various DyP-type peroxidases.7,17,19 These results support the hypothesis that different types of substrates get oxidized at different locations in or on the surface of DyP-type peroxidases. Long-range electron transfer from the protein surface to the heme has been reported for AauDyP, TcDyP and VcDyP.6,29,34 These enzymes were observed

56 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis to be flexible in the used pathway and were able to form radicals on more than one tyrosine/tryptophan. Mutating the surface exposed tyrosines and tryptophans of TfuDyP to phenylalanine individually yielded proteins with a similar activity as the wild-type enzyme towards Reactive Blue 19 at pH 3.0. From this it was concluded that, if the activity of TfuDyP towards bulky substrates is based on LRET and surface exposed Tyr/Trp radicals, more than one Tyr/Trp could be used by TfuDyP as was observed for AauDyP, TcDyP and VcDyP.6,29,34 Unfortunately, a “knock-out” TfuDyP in which all five tyrosines and seven tryptophans were exchanged for a phenylalanine did not express. Another possibility for the activity towards bulky substrates might be that the activity of TfuDyP is not based on surface exposed radicals formed by LRET, but that bulky substrates get oxidized in the heme propionate pocket. Binding of ascorbic acid and 2,6-dimethoxyphenol to this pocket was also observed in the crystal structure of DyP from Bjerkandera adusta Dec 1.23 To study the oxidation site(s) for bulky substrates further, more Tyr/Trp knock-outs could be studied. Furthermore it would be very interesting to co-crystalize DyP-type peroxidases from all four subclasses with a variety of substrates.

Materials and methods Chemicals, reagents and enzymes Oligonucleotide primers were obtained from Sigma-Aldrich. Restriction enzyme DpnI was obtained from New England Biolabs. Media components, reagents and salts were obtained from Merck, Sigma, Aldrich, BD and Fisher Scientific. Reaction Blue 19 was obtained from Acros Organics.

Site directed mutagenesis Plasmid pBAD/Myc-His A-TfuDyP was used for the overexpression of TfuDyP.4 Point mutations (A245E, A245R, A245V, D242N, G243A, G356A, H338C, N236K, N246L, Y276F, Y357F, Y359F, Y403F, W179F, W215F, W289F, W363F) were made by site directed mutagenesis using PfuUltra (I/II) hotstart PCR master mix (Agilent technologies). The twelve fold mutant of TfuDyP, in which all five tyrosines and seven tryptophans were mutated to a phenylalanine, was ordered as gene fragment from GenScript. The nativeTfu DyP gene in plasmid pBAD/Myc- His A-TfuDyP was exchanged with the twelve fold mutant by In-Fusion cloning (ClonTech). The obtained plasmids were used to transform Escherichia coli strain TOP10.

Protein expression and purification TfuDyP (mutants) were expressed and purified as described before, with minor

57 Chapter 3 changes.50 In short, 100 mL cultures were grown in Terrific Broth (TB) medium and induced with 0.02% L-arabinose at an OD600 ~ 0.6. All cultures were supplemented with 50 μg/mL ampicillin. Cells were harvested by centrifugation 3000 x g, for 30 min at 4 ºC (Eppendorf centrifuge 5810R). Pellets were resuspended in buffer A (50 mM potassium phosphate, 500 mM NaCl, 5% glycerol, pH 8.0) supplemented with cOmplete EDTA-free protease inhibitor cocktail (Roche). After breaking the cells by sonication, the enzymes were purified from the cell- free extract using Ni-NTA agarose (Qiagen) or a 1-mL His-Trap HP column (GE Healthcare). Columns were washed with Buffer A and Buffer A supplemented with 20 mM imidazole. Enzymes were eluted with Buffer A containing 300 mM imidazole. The buffer was exchanged to Buffer B (50 mM potassium phosphate, 150 mM NaCl, 10% glycerol, pH 7.5) using an Econo-Pac 10DG desalting column (BioRad). Purified enzymes were concentrated, flash frozen in liquid nitrogen and stored at -20 ºC.

UV-vis absorption spectroscopy The purified enzymes were analyzed by UV-vis absorption spectroscopy (V‑660 spectrophotometer, Jasco) as described before.50 Absorbance spectra were recorded between 250-800 nm at ambient temperature. The predicted molecular extinction coefficients at 280 nm were used to calculate the protein 51 ‑1 -1 concentrations (ProtParam tool, Expasy ): TfuDyP, ε280 nm = 45,950 M cm ; -1 -1 TfuDyP with mutation Tyr to Phe,280 ε nm = 44,460 M cm and TfuDyP with -1 -1 mutation Trp to Phe, ε280 nm = 40,450 M cm . Absorbance spectra of the oxidized and reduced enzymes were taken after the addition of 100 μM H2O2 or a few grains of sodium dithionite, respectively.

Activity towards Reactive Blue 19 Activity towards Reactive Blue 19 was measured as described before.4,50 In short, oxidation of Reactive Blue 19 was followed at 595 nm at ambient temperature (V-660 spectrophotometer, Jasco). Reaction mixtures contained 25 μM Reactive

Blue 19, 100 μM H2O2 and 20 nM enzyme in a citric acid/Na2HPO4 buffer of pH 2.5-5.5.52

Peroxygenase activity: sulfoxidation of thioanisole Sulfoxidation of thioanisole was studied as described before with minor changes.4 A 250 mM stock of thioanisole was prepared in ethanol. Reaction mixtures of

1 mL contained 2.5 mM thioanisole, 1.0 mM H2O2 and 2.0 μM enzyme in a 25 mM citrate buffer of pH 3.5. Reactions were incubated in 4 mL glass vials at 30 ºC, 100 rpm for two hours (Innova 44 incubator shaker, New Brunswick). Samples

58 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis of 0.5 mL were taken after two hours. Substrate and products were extracted and analyzed by chiral GC as described before.53 In short, the substrates and products were extracted by 0.5 mL tert-butyl methyl ether containing 0.1% mesitylene as internal standard. Samples were vortexed for one minute with the extracting solvent and centrifuged at 16,000 xg for two minutes to improve phase separation. The organic fractions were collected and the reaction mixtures were extracted for a second time by the same procedure. The organic fractions were combined and dried over anhydrous magnesium sulfate, followed by centrifugation at 16,000 x g for two minutes. The supernatant was analyzed by chiral GC (6890 GC system, Hewlett Packard) and a Chiraldex G-TA column (30 m x 0.25 mm x 0.125 mm, Grace Alltech). Program: (1) 35 ºC for 1 min, (2) gradient of 35 to 170 ºC with increment of 10 ºC/min, (3) 170 ºC for 8 minutes and (4) 170 to 35 ºC at 10 ºC/min. With an inlet temperature of 180 ºC, inlet volume of 2 μL and a split ratio of 1:20.

Acknowledgement This work was supported by the NWO graduate program: synthetic biology for advanced metabolic engineering, project number 022.004.006, The Netherlands.

59 Chapter 3

References

1 Sugano Y (2009) DyP-type peroxidases comprise a novel heme peroxidase family. Cell. Mol. Life Sci. 66, 1387–1403. 2 Kim SJ and Shoda M (1999) Purification and characterization of a novel peroxidase from Geotrichum candidum Dec 1 involved in decolorization of dyes.Appl. Environ. Microbiol. 65, 1029–1035. 3 Colpa DI, Fraaije MW, and van Bloois E (2014) DyP-type peroxidases: a promising and versatile class of enzymes. J. Ind. Microbiol. Biotechnol. 41, 1–7. 4 van Bloois E, Torres Pazmiño DE, Winter RT, and Fraaije MW (2010) A robust and extracellular heme-containing peroxidase from Thermobifida fusca as prototype of a bacterial peroxidase superfamily. Appl. Microbiol. Biotechnol. 86, 1419–1430. 5 Singh R, Grigg JC, Armstrong Z, Murphy MEP, and Eltis LD (2012) Distal heme pocket residues of B-type dye-decolorizing peroxidase: Arginine but not aspartate is essential for peroxidase activity. J. Biol. Chem. 287, 10623–10630. 6 Uchida T, Sasaki M, Tanaka Y, and Ishimori K (2015) A dye-decolorizing peroxidase from Vibrio cholerae. Biochemistry 54, 6610–6621. 7 Chen C, Shrestha R, Jia K, Gao PF, Geisbrecht B V, Bossmann SH, Shi J, and Li P (2015) Characterization of Dye-decolorizing peroxidase (DyP) from Thermomonospora curvata reveals unique catalytic properties of A-Type DyPs. J. Biol. Chem. 290, 23447–23463. 8 Mendes S, Brissos V, Gabriel A, Catarino T, Turner DL, Todorovic S, and Martins LO (2015) An integrated view of redox and catalytic properties of B-type PpDyP from Pseudomonas putida MET94 and its distal variants. Arch. Biochem. Biophys. 574, 99–107. 9 Fawal N, Li Q, Savelli B, Brette M, Passaia G, Fabre M, Mathé C, and Dunand C(2013) PeroxiBase: a database for large-scale evolutionary analysis of peroxidases. Nucleic Acids Res. 41, D441–D444. 10 Yoshida T and Sugano Y (2015) A structural and functional perspective of DyP-type peroxidase family. Arch. Biochem. Biophys. 574, 49–55. 11 Singh R and Eltis LD (2015) The multihued palette of dye-decolorizing peroxidases. Arch. Biochem. Biophys. 574, 56–65. 12 Sugano Y, Muramatsu R, Ichiyanagi A, Sato T, and Shoda M (2007) DyP, a unique dye- decolorizing peroxidase, represents a novel heme peroxidase family: ASP171 replaces the distal histidine of classical peroxidases. J. Biol. Chem. 282, 36652–36658. 13 Zámocký M, Hofbauer S, Schaffner I, Gasselhuber B, Nicolussi A, Soudi M, Pirker KF, Furtmüller PG and Obinger C (2015) Independent evolution of four heme peroxidase superfamilies. Arch. Biochem. Biophys. 574, 108–119. 14 Veitch NC (2004) Structural determinants of plant peroxidase function. Phytochem. Rev. 3, 3–18. 15 Welinder KG (1992) Superfamily of plant, fungal and bacterial peroxidases. Curr. Opin. Struct. Biol. 2, 388–393. 16 Rahmanpour R, Rea D, Jamshidi S, Fülöp V, and Bugg TDH (2016) Structure of Thermobifida fusca DyP-type peroxidase and activity towards Kraft lignin and lignin model compounds. Arch. Biochem. Biophys. 594, 54–60. 17 Liu X, Du Q, Wang Z, Zhu D, Huang Y, Li N, Wei T, Xu S, and Gu L (2011) Crystal structure and biochemical features of EfeB/YcdB from Escherichia coli O157: Asp235 plays divergent roles in different enzyme-catalyzed processes. J. Biol. Chem. 286, 14922–14931. 18 Shrestha R, Huang G, Meekins DA, Geisbrecht B V, and Li P (2017) Mechanistic insights into dye-decolorizing peroxidase revealed by solvent isotope and viscosity effects. ACS Catal. 7, 6352–6364. 19 Liu X, Yuan Z, Wang J, Cui Y, Liu S, Ma Y, Gu L, and Xu S (2017) Crystal structure and biochemical features of dye-decolorizing peroxidase YfeX from Escherichia coli O157 Asp143 and Arg232

60 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis

play divergent roles toward different substrates.Biochem. Biophys. Res. Commun. 484, 40– 44. 20 Lončar N, Colpa DI, and Fraaije MW (2016) Exploring the biocatalytic potential of a DyP-type peroxidase by profiling the substrate acceptance of Thermobifida fusca DyP peroxidase. Tetrahedron 72, 7276–7281. 21 Ogola HJO, Kamiike T, Hashimoto N, Ashida H, Ishikawa T, Shibata H, and Sawa Y (2009) Molecular characterization of a novel peroxidase from the cyanobacterium Anabaena sp. strain PCC 7120. Appl. Environ. Microbiol. 75, 7509–7518. 22 Ahmad M, Roberts JN, Hardiman EM, Singh R, Eltis LD, and Bugg TD (2011) Identification of DypB from Rhodococcus jostii RHA1 as a Lignin Peroxidase. Biochemistry 50, 5096–5107. 23 Yoshida T, Tsuge H, Hisabori T, and Sugano Y (2012) Crystal structures of dye-decolorizing peroxidase with ascorbic acid and 2,6-dimethoxyphenol. FEBS Lett. 586, 4351–4356. 24 Roberts JN, Singh R, Grigg JC, Murphy MEP, Bugg TDH, and Eltis LD (2011) Characterization of dye-decolorizing peroxidases from Rhodococcus jostii RHA1. Biochemistry 50, 5108–5119. 25 Linde D, Cañellas M, Coscolín C, Davó-Siguero I, Romero A, Lucas F, Ruiz-Dueñas FJ, Guallar V, and Martínez AT (2016) Asymmetric sulfoxidation by engineering the heme pocket of a dye-decolorizing peroxidase. Catal. Sci. Technol. 6, 6277–6285. 26 Gumiero A, Murphy EJ, Metcalfe CL, Moody PCE, and Lloyd E (2010) An analysis of substrate binding interactions in the heme peroxidase enzymes : A structural perspective. Arch. Biochem. Biophys. 500, 13–20. 27 Macdonald IK, Badyal SK, Ghamsari L, Moody PC, and Raven EL (2006) Interaction of ascorbate peroxidase with substrates: A mechanistic and structural analysis. Biochemistry 45, 7808–7817. 28 Sundaramoorthy M, Kishi K, Gold MH, and Poulos TL (1994) The crystal structure of manganese peroxidase from Phanerochaete chrysosporium at 2.06 A resolution. J. Biol. Chem. 269, 32759–32767. 29 Shrestha R, Chen X, Ramyar KX, Hayati Z, Carlson EA, Bossmann SH, Song L, Geisbrecht B V, and Li P (2016) Identification of surface-exposed protein radicals and a substrate oxidation site in A-class dye-decolorizing peroxidase from Thermomonospora curvata. ACS Catal. 6, 8036–8047. 30 Gray HB and Winkler JR (2015) Hole hopping through tyrosine/tryptophan chains protects proteins from oxidative damage. Proc. Natl. Acad. Sci. U. S. A. 112, 10920–10925. 31 Choinowski T, Blodig W, Winterhalter KH, and Piontek K (1999) The crystal structure of lignin peroxidase at 1.70 A resolution reveals a hydroxy group on the Cβ of tryptophan 171: a novel radical site formed during the redox cycle. J. Mol. Biol. 286, 809–827. 32 Blodig W, Smith AT, Winterhalter K, and Piontek K (1999) Evidence from spin-trapping for a transient radical on tryptophan residue 171 of lignin peroxidase. Arch. Biochem. Biophys. 370, 86–92. 33 Pérez-Boada M, Ruiz-Dueñas FJ, Pogni R, Basosi R, Choinowski T, Martínez MJ, Piontek K, and Martínez AT (2005) Versatile peroxidase oxidation of high redox potential aromatic compounds: Site-directed mutagenesis, spectroscopic and crystallographic investigation of three long-range electron transfer pathways. J. Mol. Biol. 354, 385–402. 34 Strittmatter E, Liers C, Ullrich R, Wachter S, Hofrichter M, Plattner DA, and Piontek K (2013) First crystal structure of a fungal high-redox potential dye-decolorizing peroxidase substrate interaction sites and long-range electron transfer. J. Biol. Chem. 288, 4095–4102. 35 Yoshioka S, Takahashi S, Hori H, Ishimori K, and Morishima I (2001) Proximal cysteine

residue is essential for the enzymatic activities of cytochrome camP450 . Eur. J. Biochem. 268, 252–259. 36 Bormann S, Gomez Baraibar A, Ni Y, Holtmann D, and Hollmann F (2015) Specific oxyfunctionalisations catalysed by peroxygenases: opportunities, challenges and solutions. Catal. Sci. Technol. 5, 2038–2052.

61 Chapter 3

37 Colpa DI, Lončar N, Schmidt M, and Fraaije MW (2017) Creating oxidase-peroxidase fusion enzymes as a toolbox for cascade reactions. ChemBioChem 18, 2226–2230. 38 Tomschy A, Brugger R, Lehmann M, Svendsen A, Vogel K, Kostrewa D, Lassen SF, Burger D, Kronenberger A, van Loon APGM, Pasamontes L, and Wyss M (2002) Engineering of phytase for improved activity at low pH. Appl. Environ. Microbiol. 68, 1907–1913. 39 Pokhrel S, Joo JC, and Yoo YJ (2013) Shifting the optimum pH of Bacillus circulans xylanase towards acidic side by introducing arginine. Biotechnol. Bioprocess Eng. 18, 35–42. 40 Brissos V, Tavares D, Sousa AC, Robalo MP, and Martins LO (2017) Engineering a bacterial DyP-type peroxidase for enhanced oxidation of lignin-related phenolics at alkaline pH. ACS Catal. 7, 3454–3465. 41 Mendes S, Catarino T, Silveira C, Todorovic S, and Martins LO (2015) The catalytic mechanism of A-type dye-decolourising peroxidase BsDyP: neither aspartate nor arginine is individually essential for peroxidase activity. Catal. Sci. Technol. 5, 5196–5207. 42 Metcalfe CL, Ott M, Patel N, Singh K, Mistry SC, Goff HM, and Lloyd Raven E (2004)

Autocatalytic formation of green heme: evidence2 forH O2-dependent formation of a covalent methionine-heme linkage in ascorbate peroxidase. J. Am. Chem. Soc. 126, 16242– 16248. 43 Fenna R, Zeng J, and Davey C (1995) Structure of the green heme in myeloperoxidase. Arch. Biochem. Biophys. 316, 653–656. 44 Fraaije MW, Roubroeks HP, Hagen WR, and van Berkel WJH (1996) Purification and characterization of an intracellular catalase-peroxidase from Penicillium simplicissimum. Eur. J. Biochem. 235, 192–198. 45 Baratto MC, Sinicropi A, Linde D, Sáez-Jiménez V, Sorace L, Ruiz-Duenas FJ, Martinez AT, Basosi R, and Pogni R (2015) Redox-active sites inAuricularia auricula-judae dye-decolorizing peroxidase and several directed variants: A multifrequency EPR study.J. Phys. Chem. B 119, 13583–13592. 46 Linde D, Ruiz-Dueñas FJ, Fernández-Fueyo E, Guallar V, Hammel KE, Pogni R, and Martínez AT (2015) Basidiomycete DyPs: Genomic diversity, structural-functional aspects, reaction mechanism and environmental significance. Arch. Biochem. Biophys. 574, 1–9. 47 Nakashima H and Nishikawa K (1994) Discrimination of intracellular and extracellular proteins using amino acid composition and residue-pair frequencies. J. Mol. Biol. 238, 54– 61. 48 Cedano J, Aloy P, Pérez-Pons JA, and Querol E (1997) Relation between amino acid composition and cellular location of proteins. J. Mol. Biol. 266, 594–600. 49 Gaur RK (2009) Prokaryotic and eukaryotic non-membrane proteins have biased amino acid distribution. J. Comput. Sci. Syst. Biol. 2, 298–299. 50 Colpa DI and Fraaije MW (2016) High overexpression of dye decolorizing peroxidase TfuDyP leads to the incorporation of heme precursor protoporphyrin IX. J. Mol. Catal. B. Enzym. 134, 372–377. 51 Gasteiger E, Hoogland C, Gattiker A, Duvaud S, Wilkins MR, Appel RD, and Bairoch A (2005) Protein identification and analysis tools on the ExPASy server. In The Proteomics Protocols Handbook (Walker JM, ed), pp. 571–607. Humana Press Inc., Totowa, NJ. 52 Dawson RMC, Elliot DC, Elliot WH, and Jones KM (1986) Data for Biochemical Research, 3rd ed. Oxford Science Publications, Oxford. 53 Romero E, Castellanos JRG, Mattevi A, and Fraaije MW (2016) Characterization and crystal structure of a robust cyclohexanone monooxygenase. Angew. Chemie - Int. Ed. 55, 15852– 15855. 54 Bendtsen JD, Nielsen H, Widdick D, Palmer T, and Brunak S (2005) Prediction of twin-arginine signal peptides. BMC Bioinformatics 6, 167. 55 Petersen TN, Brunak S, von Heijne G, and Nielsen H (2011) SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat. Methods 8, 785–786.

62 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis

56 Jongbloed JDH, Grieger U, Antelmann H, Hecker M, Nijland R, Bron S, and van Dijl JM (2004) Two minimal Tat translocases in Bacillus. Mol. Microbiol. 54, 1319–1325. 57 Santos A, Mendes S, Brissos V, and Martins LO (2014) New dye-decolorizing peroxidases from Bacillus subtilis and Pseudomonas putida MET94: towards biotechnological applications. Appl. Microbiol. Biotechnol. 98, 2053–2065. 58 Kekilli D, Moreno-Chicano T, Chaplin AK, Horrell S, Dworkowski FSN, Worrall JAR, Strange RW, and Hough MA (2017) Photoreduction and validation of haem-ligand intermediate states in protein crystals by in situ single-crystal spectroscopy and diffraction. IUCrJ 4, 263– 270. 59 Petrus MLC, Vijgenboom E, Chaplin AK, Worrall JAR, van Wezel GP, and Claessen D (2016) The DyP-type peroxidase DtpA is a Tat-substrate required for GlxA maturation and morphogenesis in Streptomyces. Open Biol. 6, 150149. 60 Rahmanpour R and Bugg TDH (2015) Characterisation of Dyp-type peroxidases from Pseudomonas fluorescens Pf-5: Oxidation of Mn(II) and polymeric lignin by Dyp1B. Arch. Biochem. Biophys. 574, 93–98. 61 Yu W, Liu W, Huang H, Zheng F, Wang X, Wu Y, Li K, Xie X, and Jin Y (2014) Application of a novel alkali-tolerant thermostable DyP-type peroxidase from Saccharomonospora viridis DSM 43017 in biobleaching of Eucalyptus Kraft pulp. PLoS One 9, e110319. 62 Sutter M, Boehringer D, Gutmann S, Günther S, Prangishvili D, Loessner MJ, Stetter KO, Weber-Ban E, and Ban N (2008) Structural basis of enzyme encapsulation into a bacterial nanocompartment. Nat. Struct. Mol. Biol. 15, 939–947. 63 Zubieta C, Krishna SS, Kapoor M, Kozbial P, McMullan D, Axelrod HL, Miller MD, Abdubek P, Ambing E, Astakhova T, Carlton D, Chiu H-J, Clayton T, Deller MC, Duan L, Elsliger M-A, Feuerhelm J, Grzechnik SK, Hale J, Hampton E, Han GW, Jaroszewski L, Jin KK, Klock HE, Knuth MW, Kumar A, Marciano D, Morse AT, Nigoghossian E, Okach L, Oommachen S, Reyes R, Rife CL, Schimmel P, van den Bedem H, Weekes D, White A, Xu Q, Hodgson KO, Wooley J, Deacon AM, Godzik A, Lesley SA, and Wilson IA (2007) Crystal structure of two novel dye-decolorizing peroxidases revel a β-barrel fold with a conserved heme-binding motif. Proteins 69, 223–233. 64 Rai A, Fedorov R, and Manstein DJ (2014) Expression, purification and crystallization of a dye-decolourizing peroxidase from Dictyostelium discoideum. Acta Crystallogr. Sect. FStructural Biol. Commun. 70, 252–255. 65 Li J, Liu C, Li B, Yuan H, Yang J, and Zheng B (2012) Identification and molecular characterization of a novel DyP-type peroxidase from Pseudomonas aeruginosa PKE117. Appl. Biochem. Biotechnol. 166, 774–785. 66 Contreras H, Joens MS, McMath LM, Le VP, Tullius M V, Kimmey JM, Bionghi N, Horwitz MA, Fitzpatrick JAJ, and Goulding CW (2014) Characterization of a Mycobacterium tuberculosis nanocompartment and its potential cargo proteins. J. Biol. Chem. 289, 18279–18289. 67 Sezer M, Santos A, Kielb P, Pinto T, Martins LO, and Todorovic S (2013) Distinct structural and redox properties of the heme active site in bacterial dye decolorizing peroxidase- type peroxidases from two subfamilies: Resonance raman and electrochemical study. Biochemistry 52, 3074–3084. 68 Zubieta C, Joseph R, Krishna SS, McMullan D, Kapoor M, Axelrod HL, Miller MD, Abdubek P, Acosta C, Astakhova T, Carlton D, Chiu H-J, Clayton T, Deller MC, Duan L, Elias Y, Elsliger M-A, Feuerhelm J, Grzechnik SK, Hale J, Han GW, Jaroszewski L, Jin KK, Klock HE, Knuth MW, Kozbial P, Kumar A, Marciano D, Morse AT, Murphy KD, Nigoghossian E, Okach L, Oommachen S, Reyes R, Rife CL, Schimmel P, Trout C V., van den Bedem H, Weekes D, White A, Xu Q, Hodgson KO, Wooley J, Deacon AM, Godzik A, Lesley SA, and Wilson IA (2007) Identification and structural characterization of heme binding in a novel dye-decolorizing peroxidase, TyrA. Proteins 69, 234–243. 69 Brown ME, Barros T, and Chang MC (2012) Identification and characterization ofa

63 Chapter 3

multifunctional dye peroxidase from a lignin-reactive bacterium. ACS Chem. Biol. 7, 2074– 2081. 70 Sugawara K, Nishihashi Y, Narioka T, Yoshida T, Morita M, and Sugano Y (2017) Characterization of a novel DyP-type peroxidase from Streptomyces avermitilis. J. Biosci. Bioeng. 123, 425– 430. 71 Liers C, Bobeth C, Pecyna M, Ullrich R, and Hofrichter M (2010) DyP-like peroxidases of the jelly fungus Auricularia auricula-judae oxidize nonphenolic lignin model compounds and high-redox potential dyes. Appl. Microbiol. Biotechnol. 85, 1869–1879. 72 Liers C, Pecyna MJ, Kellner H, Worrich A, Zorn H, Steffen KT, Hofrichter M, and Ullrich R (2013) Substrate oxidation by dye-decolorizing peroxidases (DyPs) from wood- and litter- degrading agaricomycetes compared to other fungal and plant heme-peroxidases. Appl. Microbiol. Biotechnol. 97, 5839–5849. 73 Kolwek J, Behrens C, Linke D, Krings U, and Berger RG (2018) Cell-free one-pot conversion of (+)-valencene to (+)-nootkatone by a unique dye-decolorizing peroxidase combined with a laccase from Funalia trogii. J. Ind. Microbiol. Biotechnol. 45, 89–101. 74 Kung CP, Wu YR, and Chuang HW (2014) Expression of a dye-decolorizing peroxidase results in hypersensitive response to cadmium stress through reducing the ROS signal in Arabidopsis. Environ. Exp. Bot. 101, 47–55. 75 Salvachúa D, Prieto A, Martínez ÁT, and Martínez MJ (2013) Characterization of a novel dye-decolorizing peroxidase (DyP)-type enzyme from Irpex lacteus and its application in enzymatic hydrolysis of wheat straw. Appl. Environ. Microbiol. 79, 4316–4324. 76 Scheibner M, Hülsdau B, Zelena K, Nimtz M, de Boer L, Berger RG, and Zorn H (2008) Novel peroxidases of Marasmius scorodonius degrade β-carotene. Appl. Microbiol. Biotechnol. 77, 1241–1250. 77 Faraco V, Piscitelli A, Sannia G, and Giardina P (2007) Identification of a new member of the dye-decolorizing peroxidase family from Pleurotus ostreatus. World J. Microbiol. Biotechnol. 23, 889–893. 78 Lauber C, Schwarz T, Nguyen QK, Lorenz P, and Lochnit G (2017) Identification, heterologous expression and characterization of a dye-decolorizing peroxidase ofPleurotus sapidus. AMB Express 7, 164. 79 Johjima T, Ohkuma M, and Kudo T (2003) Isolation and cDNA cloning of novel hydrogen peroxide-dependent phenol oxidase from the basidiomycete Termitomyces albuminosus. Appl. Microbiol. Biotechnol. 61, 220–225.

64 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis

Supporting information

Table S1. Location and orientation of tyrosine and tryptophan residues in TfuDyP.

Residue Distance to heme (Å) Location and orientation Y276 14.4 Surface exposed -OH Y357 13.0 Backbone surface exposed, in a tunnel in dimer interface Y359 9.7 Backbone surface exposed, in a tunnel in dimer interface Y403 9.2 Surface exposed -OH W179 16.3 Surface exposed -NH W215 15.1 Backbone and part of the indole are surface exposed, NH-group faces inside W289 8.3 Surface exposed -NH W363 17.2 Surface exposed -NH Tyr/Trp that were not mutated in this study Y422 30.6 Surface exposed -OH W111 26.2 Internal W270 22.0 Backbone and part of the indole are surface exposed, NH-group faces inside W378 12.0 Internal * Lower the number by 39 to obtain the amino acid number of the same residues in the truncated version of TfuDyP used by Rahmanpour et al, PDB entry 5FW4.16.

Table S2. Tyrosine and tryptophan content of DyP-type peroxidases. Sequences of members from class A and D are in the mature form, without the predicted (TAT) signal sequences predicted by the TatP 1.0 and the SignalP 4.1 servers.54,55

65 Chapter 3 Ref. [56,57] [58,59] [60] [22] [17] [61] [29] [4,16] - - [62] [63] [64] [22,24] [60] [60] [65] [18] [66] [67] [63,68] [6] [19] UniProtKB UniProtKB accession code P39597 A0A076MAJ9 Q4KBM1 Q0S4I5 P31545 C7MS11 D1A807 Q47KB1 Q9ZBW9 U3KRF5 NCBI ref: WP_101555111.1 Q8A8E8 Q556V8 Q0SE24 Q4KAC6 Q4KA97 D5LRR6 E3G9I4 I6Y4U9 Q88HV5 Q8EIU4 Q9KQ59 P76536 3.8 2.6 2.5 3.2 3.6 3.0 2.7 3.0 2.6 3.1 2.2 3.8 3.9 2.9 3.7 2.8 4.3 4.3 2.1 3.5 4.5 3.6 3.7 (%) W + Y 3.00 ± 0.43 3.49 ± 0.79 Y 2.7 1.0 1.5 1.6 2.6 0.8 0.8 1.3 1.0 1.3 1.6 3.2 3.3 2.3 2.0 1.9 2.3 3.0 1.5 1.7 3.9 3.0 2.3 (%) 1.46 ± 0.67 2.46 ± 0.73 W 1.1 1.6 1.0 1.6 1.0 2.2 1.9 1.8 1.6 1.8 0.5 0.6 0.7 0.6 1.7 0.9 2.0 1.3 0.6 1.7 0.6 0.7 1.3 (%) 1.54 ± 0.40 1.03 ± 0.53 4 6 6 3 3 5 4 5 6 8 6 6 7 9 5 5 9 7 Y 10 10 10 10 12 (#) 4 6 4 6 4 8 7 7 6 7 2 2 2 2 5 3 6 4 2 5 2 2 4 W (#) (#) 372 386 396 379 388 367 369 396 386 393 367 316 306 350 295 324 299 299 335 287 311 302 299 Residue Residue H37Rv DSM 43017 PKE117 ATCC BAA-471 ATCC M18 TK24 RHA1 RHA1 O157 Sakai O157:H7 str. Organism Bacillus subtilis lividans Streptomyces Pseudomonas fluorescens Pf-5 jostii Rhodococcus Escherichia coli viridis Saccharomonospora Thermomonospora curvata Thermobifida fusca coelicolor Streptomyces cellulosilityica Thermobifida linens Brevibacterium thetaiotaomicron Bacteriodes discoideum Dictyostelium jostii Rhodococcus Pseudomonas fluorescens Pf-5 Pseudomonas fluorescens Pf-5 Pseudomonas aeruginosa lignolyticus Enterobacter tuberculosis Mycobacterium Pseudomonas putida oneidensis Shewanella Vibrio cholerae Escherichia coli PDB 5MJH 2Y4F 5JXU 5FW4 4GT2 4GS1 2GVK 3QNR 5VJ0 2HAG 5DE0 5GT2 DyP) DyP) Pa DyP DyP DyP DyP DyP DyP Tc Tfu ( Sco ( Tce Bl DyP Bt DdDyP DypB Dyp1B Dyp2B DyP El Mt DyP Pp DyP TyrA Vc YfeX DyP Bs DyP DtpA DypA DypA EfeB Svi Class A B Average class A Average class B Average

66 Exploring the catalytic properties of TfuDyP by site-directed mutagenesis Ref. [21] [69] [70] [34,71] [2,12] [72] [73] [74] [75] [72] [76] [76] [77] [78] [79] UniProtKB UniProtKB accession code Q8YWM0 K7N5M8 Q82HB1 I2DBY1 Q8WZK8 I2DBY2 GenBank: AUW34346.1 G0X8C9 A0A1R7T0P5 I2DBY3 B0BK71 B0BK72 Q0VTU1 A0A0F7VJ89 Q8NKF3 4.3 3.0 3.7 2.3 2.3 2.1 2.1 2.0 2.9 1.8 1.4 1.6 4.7 4.8 3.3 (%) W + Y 3.67 ± 0.63 2.62 ± 1.14 3.10 ± 1.02 Y 2.8 1.9 2.2 1.4 1.3 1.3 1.0 1.0 1.6 0.8 0.6 0.8 3.0 3.0 2.3 (%) 2.30 ± 0.43 1.50 ± 0.82 1.88 ± 0.89 W 1.5 1.1 1.5 0.8 1.1 0.8 1.0 1.0 1.3 1.0 0.8 0.8 1.8 1.8 1.0 (%) 1.37 ± 0.25 1.11 ± 0.35 1.22 ± 0.50 9 7 6 6 5 5 7 4 3 4 Y 13 10 15 15 11 (#) 7 5 7 4 5 4 5 5 6 5 4 4 9 9 5 W (#) (#) 469 464 456 487 476 480 484 488 447 505 493 491 506 497 482 Residue Residue 75iv2 PCC 7120 Streptomyces avermitilis Streptomyces Auricularia auricula-judae Dec 1 adusta Bjerkandera Exidia glandulosa trogii (Coriolopsis trogii) Funalia Ganoderma lucidum lacteus Irpex epipterygia Mycena scorodonius Mycetinis scorodonius Mycetinis Pleurotus ostreatus Pleurotus sapidus albuminosus Termitomyces Organism Anabaena sp. sp. Amycolatopsis PDB 5C2I 4G2C 4AU9 2D3Q DyP DyP DyP AnaPX DyP2 Sa DyP2 Aau DyP DyP Egl DyP Ftr- Gl DyP I. lacteus DyP Mep DyP Msp1 MsP2 Po Psa DyP TAP Class C class C Average D class D Average class A-D Average

67 Chapter 3

Table S3. Tyrosine and tryptophan content of prokaryotic and eukaryotic intra/extracellular proteins.

Protein origin/location W (%) Y (%) Literature Prokaryote 1.35 2.60 [49] Eukaryote 1.14 2.91 Prokaryote, intracellular 1.11 2.80 [47] Prokaryote, extracellular 1.44 4.03 Eukaryote, intracellular 1.01 3.16 Eukaryote, extracellular 1.43 3.50 Intracellular 1.2 ± 0.82 3.1 ± 1.49 [48] Extracellular 1.4 ± 1.3 3.6 ± 2.02

68 Chapter 4:

High overexpression of dye decolorizing peroxidase TfuDyP leads to the incorporation of heme precursor protoporphyrin IX

Dana I. Colpa and Marco W. Fraaije

This chapter is based on: Journal of Molecular Catalysis B: Enzymatic (2016) 134: 372-377 DOI: 10.1016/j.molcatb.2016.08.017 Abstract The heterologous overexpression level of the bacterial dye decolorizing peroxidase TfuDyP in Escherichia coli was increased sixty fold to approximately 200 mg of purified enzyme per liter culture broth by fusing the enzyme to the small ubiquitin-related modifier protein (SUMO). The highly overexpressed SUMO-TfuDyP was, however, almost inactive. Analysis of the enzyme by UV-vis absorption spectroscopy and high-resolution mass spectrometry showed that a large fraction of the highly overexpressed enzyme contained the iron deficient heme precursor protoporphyrin IX (PPIX) instead of heme. Here we show that the activity of the enzyme was dependent on the expression level of the protein. High overexpression of TfuDyP leads to the incorporation of PPIX

Introduction Dye decolorizing peroxidases (DyPs) comprise a family of peroxidases which were discovered only two decades ago.1 The physiological role of these enzymes is unknown, but they display a strikingly broad substrate scope including anthraquinone and azo-dyes1–4, lignin(-model) compounds5, β-carotene6 and aromatic sulfides7. Their enigmatic physiological role and broad substrate scope make DyPs interesting targets for biochemical and biotechnological studies. One member of this family is already applied in the food industry to degrade β-carotene in whey-containing beverages.6 Another application might be the bioremediation of waste water polluted with synthetic dyes. Many DyPs are active on various synthetic dyes, including TfuDyP from Thermobifida fusca, BsDyP from Bacillus subtilis, PpDyP from Pseudomonas putida MET94 and AnaPX from Anabaena sp. strain PCC 7120.2–4 Furthermore, DyPs might become valuable biocatalytic tools in the enzymatic degradation of lignin and lignocellulose containing biomass. It has been shown that the DyP from Irpex lacteus promotes the enzyme-mediated degradation of wheat straw and DyPB from Rhodococcus jostii RHA1 has been found to be active on lignin and lignin model compounds.5,8 DyP-type peroxidases have a ferredoxin-like fold and rely on a non- covalently, but tightly, bound heme b cofactor for their activity.9 Heme b is the most common heme cofactor, and is found in a wide variety of proteins, e.g. cytochromes P450, hemoglobin, and most peroxidases.10,11 Heme cofactors are produced through a highly conserved heme synthesis pathway in which heme b is synthesized from δ-aminolevulinic acid (δ-ALA).10 In the last step ferrous iron is inserted in protoporphyrin IX (PPIX) to form heme b. Our group characterized a thermostable member of the DyP-type peroxidase family, TfuDyP. TfuDyP from Thermobifida fusca is a dimeric, Tat- dependently exported DyP which degrades several types of dyes, oxidizes small aromatic compounds and performs enantioselective sulfoxidations.2,7 We have recently shown that the enzyme can be exploited as biocatalyst for the production of valuable flavor compound divanillin, using vanillin as substrate. The crystal structure of TfuDyP was solved recently.12 A high expression level in a heterologous host, like Escherichia coli, would make DyPs more attractive for industrial applications. In this work we boosted the heterologous overexpression of TfuDyP in E. coli through a fusion to SUMO (small ubiquitin-related modifier protein). The expression level of TfuDyP was efficiently raised from 3 mg TfuDyP7 to approximately 200 mg SUMO-TfuDyP from 1 L culture. However, a large fraction of the highly overexpressed enzyme contained the iron deficient heme precursor protoporphyrin IX instead of heme.

71 Chapter 4

Results and discussion TfuDyP versus SUMO-TfuDyP In earlier work, TfuDyP was expressed in E. coli MC1061 with a yield of 3 mg purified enzyme per liter LB medium.7 This construct includes the Tat-signal sequence and a C-terminal myc-His tag. The protein yield per liter culture broth could be increased to approximately 25 mg purified enzyme from a culture of E. coli TOP10 grown on TB medium. To increase the expression level of TfuDyP further, TfuDyP was fused to SUMO. Fusing an enzyme to SUMO often increases the expression level and solubility of an enzyme.13 By fusing it to SUMO the overexpression level of TfuDyP in TB medium was increased to approximately 200 mg enzyme per liter culture broth. Clear differences were, however, observed between TfuDyP and SUMO- TfuDyP. PurifiedTfu DyP was brown-red while a more purple color was observed for highly overexpressed SUMO-TfuDyP. This difference could also be visualized by UV-vis absorption spectroscopy (Fig. 1). The two spectra are on the one hand very similar, as both spectra show a protein peak at 280 nm and a Soret band at around 406-416 nm. Clear differences were on the other hand observed in the range of 450-700 nm; in this range the so-called Q-bands are observed. TfuDyP showed a broad peak between 450 and 650 nm, while for SUMO-TfuDyP five peaks were observed at 509, 546, 568, 621 and 663 nm. In previous work, we showed that TfuDyP could be reduced using sodium dithionite, shifting the Soret band to 4317 nm. When sodium dithionite was added to SUMO-TfuDyP only minor spectral changes were observed, suggesting that a major part of the cofactors could not be reduced (data not shown). This observation is in agreement with the activity measurements. TfuDyP had an observed reaction rate towards Reactive Blue 19 of 5.8 s-1, while for SUMO- TfuDyP a relatively low rate of 1.3 s-1 was observed.

Influence of the expression conditions onTfu DyP In an attempt to increase the specific activity of recombinantTfu DyP the effects of different expression conditions and protein fusion partners were studied. Hemin and δ-aminolevulinic acid (δ‑ALA) are often added to the growth medium when overexpressing heme-containing enzymes.4,14 Ferrous sulfate and ferric citrate are added to the culture media when overexpressing iron-containing proteins.15 Addition of δ-ALA, ferrous sulfate or ferric citrate to the culture media had only a minor effect on the specific activity of the purified SUMO-fused enzyme. The protein yields were increased by 35-45% while the activities were increased by only 15-20%. The observed specific activities were still approximately 3.7-fold

72 High overexpression of TfuDyP leads to the incorporation of PPIX

Figure 1. UV-vis absorption spectra of TfuDyP (top) and SUMO-TfuDyP (bottom). The insets show the range between 450 nm and 700 nm enlarged.

73 Chapter 4 lower compared to TfuDyP. A bigger effect was observed for the addition of hemin to the growth medium, the protein yield was increased by 50% and the activity by a factor of 1.8 to an observed rate of catalysis of 2.4 s-1. It was hypothesized that SUMO might have an effect on the proper folding, dimerization and/or the accessibility of the active site of TfuDyP. To determine whether the fusion of SUMO to TfuDyP had an influence on the activity, SUMO was cleaved off using SUMO protease. However, the activity did not increase. To analyze the effect of SUMO further, SUMO was exchanged by another often- used fusion protein, maltose binding protein (MBP). The overexpression of MBP- TfuDyP by E. coli TOP10 gave comparable results; a highly overexpressed but almost inactive protein was obtained (Fig. 2). To analyze the effect of any fusion protein (SUMO or MBP) on TfuDyP, SUMO was deleted from the plasmid pBAD His-SUMO TfuDyP yielding pBAD His-TfuDyP (not to be confuse with the original plasmid pBAD TfuDyP Myc-His A, which includes the Tat-sequence). Expression of His-TfuDyP by E. coli TOP10 yielded approximately 210 mg purified enzyme per liter culture broth. The highly overexpressed enzyme had again a relatively low specific activity towards Reactive Blue 19 of 1.5-1 s (Fig. 2). From these results it could be concluded that the lower specific activity of SUMO/MBP-TfuDyP was not caused by the protein fusion partner.

7.0 TfuDyPTfuDyP 6.0 SUMO-TfuDyPSUMO-TfuDyP MBP-TfuDyPMBP-TfuDyP 5.0

His-TfuDyPHis-TfuDyP ) 1 - 4.0 (s

obs 3.0 k

2.0

1.0

0.0 0 100 200 300 Expression yield (mg/L culture broth medium)

Figure 2. Comparison between the protein expression yield per liter culture broth medium and the observed initial activity towards Reactive Blue 19 for variants of TfuDyP. The reported expression yields and activities were determined upon purification of the enzyme.

74 High overexpression of TfuDyP leads to the incorporation of PPIX

Expression of SUMO-TfuDyP by different E. coli strains In an attempt to increase the activity of the highly overexpressed enzyme, several expression strains were screened. Different E. coli strains can be used for the heterologous overexpression of enzymes. Here we compared the expression level and the activity of SUMO-TfuDyP when being overexpressed by seven different E. coli strains (Fig. 3). Four E. coli K-strains (TOP10, MC1061, DH5α and SHuffle) and three E. coli B-strains (BL21 (DE3), C41 (DE3) and C43 (DE3)) were used. Each strain was grown in duplicate in 50-100 mL TB medium, and induced with the same amount of L-arabinose (0.02%). Interestingly, a correlation between the expression level and the enzyme activity was observed. E. coli TOP10 and MC1061 produced 190 and 140 mg of purified enzyme per liter culture broth respectively. The enzyme purified from these cultures had a relatively low activity for Reactive Blue 19 of respectively 1.3 s-1 and 0.8 s‑1. The three E. coli B-strains, on the other hand, produced only approximately 20 mg of enzyme per liter culture broth displaying activities between 5.0-6.7 s-1. E. coli SHuffle and DH5α yielded low levels of expression and intermediate specific activities.

The activity of TfuDyP is influenced by the expression level To prove that the differences between active TfuDyP and highly overexpressed but poorly active SUMO-TfuDyP were caused by the expression level of the protein, different overexpression conditions were studied. The expression level of TfuDyP was boosted through the induction of the E. coli TOP10 cultures with 0.2% instead of 0.02% L-arabinose. The expression level increased from 26 to 39 mg per liter culture broth, while the specific activity decreased by 12%. To probe the opposite effect, the relatively high expression level of SUMO-TfuDyP was reduced by inducing the respective E. coli TOP10 cultures with 0.002% L-arabinose instead of 0.02%. These cultures were grown at 17 ºC for 3 days, instead of at 37 ºC overnight. The expression level was indeed efficiently decreased from 194 mg/L to 70 mg/L culture broth, while the observed rate of catalysis improved by 35%. These results show that the activity of the enzyme was dependent on the expression level of the protein.

Analysis of the cofactor by mass spectrometry To analyze the differences between active TfuDyP and poorly active SUMO- TfuDyP further, the heme cofactors were extracted from the enzymes. The red ethyl acetate samples were analyzed by HR-MS. The isolated cofactor of TfuDyP had a mass of 616.1758 for the [M + H]+ ion, which is in accordance with the calculated mass of heme (616.1768) (Fig. 4A). On the other hand, revealed the

75 Chapter 4 cofactor isolated from the highly overexpressed but almost inactive SUMO- TfuDyP three peaks. Intriguingly, the main peak, with a mass of 563.2648 for the [M + H]+ ion, corresponds to the mass of the heme precursor protoporphyrin IX (PPIX, heme deficient in iron) (Fig. 4B). The two smaller peaks correspond to heme (616.1758) and oxidized PPIX (595.2540). PPIX is known to be sensitive to photo-oxidation in the presence of oxygen, which may explain the presence of oxidized PPIX.16 These results are in agreement with the observed specific activities of the two proteins: fully active TfuDyP contained heme, while highly overexpressed and poorly active SUMO-TfuDyP was mainly loaded with heme precursor PPIX which cannot support catalysis.

7.0

6.0

5.0

) 1 -

(s 4.0

obs

k 3.0

2.0

1.0

0.0 250 TfuDyP, BL21 C41 C43 DH5a SHuffle TOP10 MC1061 TOP10

200

yield yield 150

100 Expression Expression 50 (mg/L culture broth medium) broth culture (mg/L 0 TfuDyP, TOP10 BL21BL21 C41 C43 DH5aDH5a SHuffle TOP10 MC1061 TfuTOP10DyP

Figure 3. Comparison of the expression level and the observed initial reaction rate of SUMO- TfuDyP towards Reactive Blue 19, when being overexpressed by seven different E. coli strains. TfuDyP from E. coli TOP10 is added for comparison. The reported expression yields and activities were determined upon purification of the enzyme.

76 High overexpression of TfuDyP leads to the incorporation of PPIX

Figure 4A. HR-MS analysis of the extracted cofactor of TfuDyP. The cofactor had an expected mass of 616.1758 for the [M + H]+ ion.

Figure 4B. HR-MS analysis of the extracted cofactor of SUMO-TfuDyP. The enzyme contained three different compounds. The main peak corresponds to the [M + H]+ ion of PPIX (563.26 m/z), and the smaller two peaks to heme (616.18 m/z) and oxidized PPIX (592.25 m/z).

77 Chapter 4

Highly overexpressed SviDyP binds PPIX in vivo To explore whether the in vivo binding of PPIX to the recombinantly overexpressed DyP-type peroxidase is specific for TfuDyP or a recurrent complicating factor for producing DyPs, we also expressed a DyP from another bacterium. For this we choose the recently described DyP-type peroxidase SviDyP from Saccharomonospora viridis DSM 4301717 which exhibits a different substrate scope and pH profile when compared toTfu DyP. Like TfuDyP, SviDyP also belongs to the subclass of A-type DyPs but only shares 40% sequence identity and displays different enzymatic properties including higher pH and temperature optima.17 An overexpression level of approximately 100 mg of His-tagged SviDyP per liter culture broth was obtained. The UV-vis absorption spectrum of SviDyP revealed that it also contains a significant amount of PPIX as it displays the identical Q-bands when compared with the recombinantly produced SUMO- TfuDyP (Fig. 5). This shows that the typical PPIX spectral fingerprint in the Q-bands region can be used to verify incorporation of heme or PPIX in DyPs or other hemoproteins.

Figure 5. UV-vis absorption spectrum of SviDyP. The inset show the range between 450 nm and 700 nm enlarged.

78 High overexpression of TfuDyP leads to the incorporation of PPIX

Conclusions In this study we showed that the specific activity of DyP-type peroxidaseTfu DyP was dependent on the overexpression level of the enzyme. UV-vis absorption spectroscopy and high-resolution mass spectrometry showed that the inactive enzyme contained, instead of a heme cofactor, the heme precursor PPIX. A heme enzyme lacking the redox active iron is unable to catalyze oxidations. In previous studies, it was shown that apo forms of DyP-type peroxidases Mt-DyP and YfeX could bind PPIX in in vitro titration experiments.18,19 The UV-vis absorption spectra of Mt-DyP and YfeX reconstituted with PPIX are comparable to the spectrum of SUMO-TfuDyP loaded with PPIX (Fig. 1). Moreover, hemoproteins nitric oxide synthase from Geobacillus stearothermophilus and dye decolorizing peroxidase EfeB from E. coli O157, had been described to bind PPIX in vivo when being overexpressed in E. coli.14,20 These data indicate that several DyPs have a high affinity towards PPIX, and that the expression host is not always able to provide enough heme to outcompete PPIX binding. It may also (partly) explain the observed amounts of PPIX in the studies on EfeB and YfeX20 leading to the erroneously conclusion that these DyPs are dechelatases.21 In this context it is worth noting that TfuDyP is not active on hemin.3 The low affinity of these enzymes towards heme may also offer interesting possibilities. For example, it could potentially be exploited to exchange the heme or PPIX for an artificial porphyrin. Key et al demonstrated that apo heme enzymes reconstituted with abiological porphyrins show unique catalysis.22 Laboratory E. coli strains are often unable to import or synthesize enough heme for highly overexpressed heme enzymes.14,23,24 In this work we tried to circumvent this problem by lowering the expression level of the enzyme or through the expression of TfuDyP in alternative E. coli strains. To improve heme incorporation one can also opt for engineering the host. For example, overexpression of the heme receptor ChuA, for hemin uptake, or heme ferrochelatase, for iron incorporation in heme, have been shown to increase the available amount of heme in the cell.14,23 This study shows that it is important to carefully tune the conditions for recombinant expression of DyPs in order to obtain enzyme that is fully loaded with the heme cofactor. Members of this newly discovered family of DyP- type peroxidases are currently of great interest concerning their biochemical properties and biotechnological potential.3,5,6,8,9 Our findings may assist in the development of effective recombinant production methods for DyPs.

79 Chapter 4

Materials and methods Reagents and chemicals Salts, media components, reagents and oligonucleotide primers were obtained from Merck, Fisher chemicals, Sigma-Aldrich and BD (USA). Reactive Blue 19 was from Acros Organics. The In-Fusion HD cloning kit was obtained from Clontech, Phusion DNA polymerase was from Thermo Scientific and the PfuUltra Hotstart PCR master mix was from Agilent Technologies. Restriction enzymes NdeI and HindIII were from New England Biolabs and T4 DNA ligase was from Promega.

Cloning To improve the expression level of TfuDyP (or any other protein) a new pBAD- based expression vector was designed, called pBAD SUMO. Fusing a protein to the C-terminus of SUMO (small ubiquitin-related modifier protein) often improves the expression level and solubility of a protein.13 The SUMO gene (including an N-terminal 6xHis-tag) was amplified from pET SUMO STMO25 and inserted in front of the multiple cloning site of an empty pBAD NK plasmid. The pBAD NK plasmid is a pBAD/Myc-His A-derived expression vector (Invitrogen) in which the NdeI site was removed and the NcoI site replaced by NdeI.26 The SUMO gene was inserted in the first restriction site (NdeI) of the multiple cloning site of pBAD NK using In-Fusion cloning (In-fusion HD cloning kit, Clontech). The NdeI restriction site was retained on both sides of the gene. To finalize the pBAD SUMO vector, the NdeI restriction site in front of SUMO was exchanged to an NcoI site, while a stop codon (TAA) was created after the multiple cloning site using QuikChange mutagenesis. In a second step, the TfuDyP gene was cloned into the obtained pBAD SUMO vector using restriction enzymes NdeI and HindIII. The first 35 residues predicted to be the Tat-signal sequence of TfuDyP were excluded. The lengths of the signal sequences were predicted using the program TatP.27 Plasmid pBAD MBP TfuDyP was obtained by exchanging SUMO of pBAD SUMO TfuDyP with MBP from pBAD MBP AldO using restriction free cloning.28,29 Plasmid pBAD His-TfuDyP was obtained by deleting SUMO, and pBAD His-SviDyP was obtained by replacing the TfuDyP gene in plasmid pBAD His-TfuDyP with the gene of SviDyP17 using restriction free cloning.29 The native gene of SviDyP from Saccharomonospora viridis DSM 43017 was synthesized by GenScript. The first 37 residues predicted to be the Tat-signal sequence were excluded. Plasmid pBAD SUMO TfuDyP was used to transform E. coli strains TOP10, MC1061, DH5α, SHuffle, BL21 (DE3), C41 (DE3) and C43 (DE3). pBAD TfuDyP Myc-His A, pBAD MBP TfuDyP, pBAD His-SviDyP and pBAD His-TfuDyP were used to transform E. coli strain TOP10.

80 High overexpression of TfuDyP leads to the incorporation of PPIX

Enzyme expression and purification All E. coli strains were grown on 5 mL Luria Bertani (LB) medium at 37 ºC, 200 rpm, overnight to saturation. The cultures were diluted 1:100 in Terrific

Broth (TB) medium and grown at 37 ºC, 135 rpm till OD600 ~ 0.6. The cultures were subsequently induced with 0.02% L-arabinose and grown at 30 ºC, 130 rpm overnight. TfuDyP was expressed in E. coli strain TOP10 using a culture volume of 400 mL. SUMO-TfuDyP was expressed in E. coli DH5α, SHuffle, BL21 (DE3), C41 (DE3) and C43 (DE3) on a 100-mL scale. All other cultures were grown on 50‑mL scale. If required, media were supplemented with 500 μM δ-aminolevulinic acid (δ-ALA), 10 μM hemin, 100 μM ferrous sulfate or 100 μM ferric citrate. For comparison, cultures of E. coli TOP10 expressing TfuDyP and SUMO-TfuDyP were also induced with 0.2% (100‑mL culture, expression at 30 ºC) and 0.002% (50-mL culture, expression at 17 ºC) L-arabinose, respectively. All cultures were supplemented with 50 μg/mL ampicillin and grown in duplicate. Cells were harvested by centrifugation at 6000xg (Beckman Coulter, Avanti JE centrifuge, JLA 14 rotor) for 15 minutes at 4º C or at 3000xg (Eppendorf centrifuge 5810R) for 30 minutes at 4 ºC. The pellets were washed in Buffer A (50 mM potassium phosphate, 0.5 M NaCl, 5% glycerol, pH 8) after which the cells were disrupted by sonication (5 minutes total on time with cycles of 10 sec on and 10 sec off at 70% amplitude). The cell-free extract was obtained by centrifugation at 16,000xg for 15 min at 4 ºC. The enzymes were purified from the cell-free extract using a 1-mL His- Trap HP column (GE Healthcare). After loading the cell-free extract onto a pre- equilibrated column, the column was washed with buffer A (see above) followed by Buffer A supplemented with 20 mM imidazole. The enzymes were eluted by a gradient from 20 mM to 300 mM imidazole in Buffer A. After the purification, the buffer was exchanged to buffer B (50 mM potassium phosphate, 150 mM NaCl, 10% glycerol, pH 7.5) using an Econo-Pac 10DG desalting column (BioRad). The enzymes were concentrated using Amicon centrifugal filters with a 30 kDa cut-off (Merck Millipore), flash frozen with liquid nitrogen and stored at ‑20 ºC.

UV-vis spectral analysis All purified enzymes were analyzed by UV-vis absorption spectroscopy (V-660 spectrophotometer, Jasco). Absorbance spectra were recorded (250-800 nm) at ambient temperature. Enzyme concentrations were determined using a predicted 30 molecular extinction coefficient at 280 nm (ProtParam tool, Expasy ), ε280 nm = -1 -1 ‑1 -1 45,950 M cm for TfuDyP and His-TfuDyP, ε280 nm = 47,440 M cm for SUMO- ‑1 -1 ‑1 -1 TfuDyP, ε280 nm = 112,300 M cm for MBP-TfuDyP and ε280 nm = 48,470 M cm for His-SviDyP.

81 Chapter 4

Activity assay Activity towards Reactive Blue 19 was measured as described before, with minor changes.7 The activity was analyzed spectrophotometrically (Jasco V-660) in a

50 mM citrate buffer pH 3.5, containing 50 μM Reactive Blue 19, 100 μM H2O2 and 100 nM enzyme at ambient temperature. Oxidation of Reactive Blue 19 was followed at 595 nm (ε = 10 mM-1 cm-1).

Cofactor extraction and mass spectrometry The cofactors of purified (SUMO)-TfuDyP were extracted using an adaptation of the method of Adamczack et al.31 Before extraction, the buffer was exchanged to double-distilled H2O using an Econo-Pac 10DG desalting column (BioRad). Enzyme concentrations of 25-150 μM were used to yield a (calculated) heme concentration of 5-40 μg/mL in the final extract, assuming that the enzyme was fully loaded with heme and all cofactors were extracted. To release the cofactor from the enzyme, hydrochloric acid was added to a concentration of 0.25 M. Two volumes of ethyl acetate were added and samples were vortexed for one minute. To improve phase separation, the samples were centrifuged for 2 minutes at 16,000xg. The red ethyl acetate layer was collected, and the mass of the cofactor was determined by high-resolution mass spectrometry (HR-MS) on an LTQ Orbitrap XL (Thermo Fisher Scientific), with electrospray in positive mode.

Acknowledgements We thank Ing. Theodora Tiemersma-Wegman (Stratingh Institute for Chemistry, University of Groningen, The Netherlands) for technical assistance with mass spectrometry, and Dr. Nikola Lončar for meaningful discussions. This work was supported by the NWO graduate program: synthetic biology for advanced metabolic engineering, project number 022.004.006, The Netherlands.

82 High overexpression of TfuDyP leads to the incorporation of PPIX

References

1 Kim SJ and Shoda M (1999) Purification and characterization of a novel peroxidase from Geotrichum candidum Dec 1 involved in decolorization of dyes.Appl. Environ. Microbiol. 65, 1029–1035. 2 Lončar N, Colpa DI, and Fraaije MW (2016) Exploring the biocatalytic potential of a DyP-type peroxidase by profiling the substrate acceptance of Thermobifida fusca DyP peroxidase. Tetrahedron 72, 7276–7281. 3 Santos A, Mendes S, Brissos V, and Martins LO (2014) New dye-decolorizing peroxidases from Bacillus subtilis and Pseudomonas putida MET94: towards biotechnological applications. Appl. Microbiol. Biotechnol. 98, 2053–2065. 4 Ogola HJO, Kamiike T, Hashimoto N, Ashida H, Ishikawa T, Shibata H, and Sawa Y (2009) Molecular characterization of a novel peroxidase from the cyanobacterium Anabaena sp. strain PCC 7120. Appl. Environ. Microbiol. 75, 7509–7518. 5 Ahmad M, Roberts JN, Hardiman EM, Singh R, Eltis LD, and Bugg TDH (2011) Identification of DypB from Rhodococcus jostii RHA1 as a lignin peroxidase. Biochemistry 50, 5096–5107. 6 Scheibner M, Hülsdau B, Zelena K, Nimtz M, de Boer L, Berger RG, and Zorn H (2008) Novel peroxidases of Marasmius scorodonius degrade β-carotene. Appl. Microbiol. Biotechnol. 77, 1241–1250. 7 van Bloois E, Torres Pazmiño DE, Winter RT, and Fraaije MW (2010) A robust and extracellular heme-containing peroxidase from Thermobifida fusca as prototype of a bacterial peroxidase superfamily. Appl. Microbiol. Biotechnol. 86, 1419–1430. 8 Salvachúa D, Prieto A, Martínez ÁT, and Martínez MJ (2013) Characterization of a novel dye-decolorizing peroxidase (DyP)-type enzyme from Irpex lacteus and its application in enzymatic hydrolysis of wheat straw. Appl. Environ. Microbiol. 79, 4316–4324. 9 Colpa DI, Fraaije MW, and van Bloois E (2014) DyP-type peroxidases: a promising and versatile class of enzymes. J. Ind. Microbiol. Biotechnol. 41, 1–7. 10 Munro AW, Girvan HM, Mclean KJ, Cheesman MR, and Leys D (2009) Chapter 10: Heme and Hemoproteins. In Tetrapyrroles: Birth, Life and Death (Warren MJ, and Smith AG, eds), pp. 160–183. Landes Bioscience and Springer Science+Business Media. 11 Banci L (1997) Structural properties of peroxidases. J. Biotechnol. 53, 253–263. 12 Rahmanpour R, Rea D, Jamshidi S, Fülöp V, and Bugg TDH (2016) Structure of Thermobifida fusca DyP-type peroxidase and activity towards Kraft lignin and lignin model compounds. Arch. Biochem. Biophys. 594, 54–60. 13 Marblestone JG, Edavettal SC, Lim Y, Lim P, Zuo X, and Butt TR (2006) Comparison of SUMO fusion technology with traditional gene fusion systems: enhanced expression and solubility with SUMO. Protein Sci. 15, 182–189. 14 Sudhamsu J, Kabir M, Airola M V, Patel BA, Yeh S-R, Rousseau DL, and Crane BR (2010) Co-expression of ferrochelatase allows for complete heme incorporation into recombinant proteins produced in E. coli. Protein Expr. Purif. 73, 78–82. 15 Jaganaman S, Pinto A, Tarasev M, and Ballou DP (2007) High levels of expression of the iron- sulfur proteins phthalate dioxygenase and phthalate dioxygenase reductase in Escherichia coli. Protein Expr. Purif. 52, 273–279. 16 Krieg M and Whitten DG (1984) Self-sensitized photo-oxidation of protoporphyrin IX and related porphyrins in erythrocyte ghosts and microemulsions: a novel photo-oxidation pathway involving singlet oxygen. J. Photochem. 25, 235–252. 17 Yu W, Liu W, Huang H, Zheng F, Wang X, Wu Y, Li K, Xie X, and Jin Y (2014) Application of a novel alkali-tolerant thermostable DyP-type peroxidase from Saccharomonospora viridis DSM 43017 in biobleaching of Eucalyptus Kraft pulp. PLoS One 9, e110319. 18 Contreras H, Joens MS, McMath LM, Le VP, Tullius M V, Kimmey JM, Bionghi N, Horwitz MA, Fitzpatrick JAJ, and Goulding CW (2014) Characterization of a Mycobacterium tuberculosis

83 Chapter 4

nanocompartment and its potential cargo proteins. J. Biol. Chem. 289, 18279–18289. 19 Létoffé S, Heuck G, Delepelaire P, Lange N, and Wandersman C (2009) Bacteria capture iron from heme by keeping tetrapyrrol skeleton intact. Proc. Natl. Acad. Sci. U. S. A. 106, 11719–11724. 20 Liu X, Du Q, Wang Z, Zhu D, Huang Y, Li N, Wei T, Xu S, and Gu L (2011) Crystal structure and biochemical features of EfeB/YcdB from Escherichia coli O157: Asp235 plays divergent roles in different enzyme-catalyzed processes. J. Biol. Chem. 286, 14922–14931. 21 Dailey HA, Septer AN, Daugherty L, Thames D, Gerdes S, Stabb E V, Dunn AK, Dailey TA, and Phillips JD (2011) The Escherichia coli protein YfeX functions as a porphyrinogen oxidase, not a heme dechelatase. MBio 2, e00248-11. 22 Key HM, Dydio P, Clark DS, and Hartwig JF (2016) Abiological catalysis by artificial haem proteins containing noble metals in place of iron. Nature 534, 534–537. 23 Varnado CL and Goodwin DC (2004) System for the expression of recombinant hemoproteins in Escherichia coli. Protein Expr. Purif. 35, 76–83. 24 Ramzi AB, Hyeon JE, and Han SO (2015) Improved catalytic activities of a dye-decolorizing peroxidase (DyP) by overexpression of ALA and heme biosynthesis genes in Escherichia coli. Process Biochem. 50, 1272–1276. 25 Franceschini S, van Beek HL, Pennetta A, Martinoli C, Fraaije MW, and Mattevi A (2012) Exploring the structural basis of substrate preferences in Baeyer-Villiger monooxygenases: insight from steroid monooxygenase. J. Biol. Chem. 287, 22626–22634. 26 Fraaije MW, Wu J, Heuts DPHM, van Hellemond EW, Lutje Spelberg JH, and Janssen DB (2005) Discovery of a thermostable Baeyer-Villiger monooxygenase by genome mining. Appl. Microbiol. Biotechnol. 66, 393–400. 27 Bendtsen JD, Nielsen H, Widdick D, Palmer T, and Brunak S (2005) Prediction of twin-arginine signal peptides. BMC Bioinformatics 6, 167. 28 Heuts DPHM, van Hellemond EW, Janssen DB, and Fraaije MW (2007) Discovery, characterization, and kinetic analysis of an alditol oxidase from Streptomyces coelicolor. J Biol Chem 282, 20283–20291. 29 van den Ent F and Löwe J (2006) RF cloning: a restriction-free method for inserting target genes into plasmids. J. Biochem. Biophys. Methods 67, 67–74. 30 Gasteiger E, Hoogland C, Gattiker A, Duvaud S, Wilkins MR, Appel RD, and Bairoch A (2005) Protein identification and analysis tools on the ExPASy server. In The Proteomics Protocols Handbook (Walker JM, ed), pp. 571–607. Humana Press Inc., Totowa, NJ. 31 Adamczack J, Hoffmann M, Papke U, Haufschildt K, Nicke T, Bröring M, Sezer M, Weimar R, Kuhlmann U, Hildebrandt P, and Layer G (2014) NirN protein from Pseudomonas aeruginosa is a novel electron-bifurcating dehydrogenase catalyzing the last step of heme

d1 biosynthesis. J. Biol. Chem. 289, 30753–30762.

84 Chapter 5:

Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions

Dana I. Colpa, Nikola Lončar, Mareike Schmidt and Marco W. Fraaije

This chapter is based on: ChemBioChem (2017) 18: 2226-2230 DOI: 10.1002/cbic.201700478 Abstract In this work we prepared a set of bifunctional oxidase-peroxidases by fusing four distinct oxidases to a peroxidase. While such fusion enzymes have not been observed in nature, they could be expressed and purified with good yields. Characterization revealed that the artificial enzymes retained the capability to bind the two required cofactors and were catalytically active as oxidase and peroxidase. The peroxidase fusions of alditol oxidase and chitooligosaccharide oxidase could be used for selective detection of xylitol and cellobiose with a detection limit in the low µM range. The peroxidase fusions of eugenol oxidase and 5-hydroxymethylfurfural oxidase could be used for dioxygen-driven one-pot two-step cascade reactions to convert vanillyl alcohol into divanillin and eugenol into lignin oligomers, respectively. The designed oxidase-peroxidase fusions represent attractive biocatalysts that allow efficient biocatalytic cascade oxidations that only require molecular oxygen as oxidant. Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions

Introduction In nature most enzymes take part in metabolic pathways in which each formed product is a substrate for the next enzymatic reaction. For optimizing the efficiency of such intricate biocatalytic cascades the enzymes are often brought together to form enzyme complexes, e.g. the pyruvate dehydrogenase complex, the microbial type I fatty-acid synthase complex and the cellulosome.1–3 The cellulosome is found in anaerobic microorganisms and consists of a scaffolding protein which brings together the required hydrolytic enzymes to degrade cellulosic biomass.3 In some cases, this has even led to the fusion of two or more enzymes creating a bi/multifunctional protein4, e. g. pyrroline-5- carboxylate synthase which features both glutamate kinase and γ-glutamyl phosphate reductase activities or the pentafunctional AROM complex from Aspergillus nidulans that is involved in aromatic amino acid biosynthesis.5,6 Inspired by the latter observation, various artificial enzyme fusions have been created in recent years in order to engineer efficient multifunctional biocatalysts. The first artificial bifunctional fusion enzyme, a histidinol dehydrogenase/ aminotransferase, was published in 1970.7 Several fusion enzymes have been made since.4,8,9 For example, a fusion between a fatty acid decarboxylase cytochrome P450 (OleTJE) and alditol oxidase (AldO) was made to fuel the reactions of OleTJE with hydrogen peroxide produced by the oxidase.10 For enabling efficient cofactor regeneration we have shown that various NAD(P)H-dependent monooxygenases can be produced fused to phosphite dehydrogenases which efficiently regenerates NAD(P)H.11,12 Fusion enzymes have several advantages over separate enzymes. They are cheaper and less labor intensive concerning their production, since only one enzyme needs to be expressed and purified. Another advantage is the close proximity of the catalytic sites enabling substrate channeling.4,9,13,14 Substrate channeling circumvents diffusion of the intermediate product in the solution, and hence increases the combined reaction rate. We were particularly inspired by the interplay betweendases oxi and peroxidases as is found in nature. Oxidases and peroxidases are often coexpressed as oxidases produce hydrogen peroxide which again is a substrate for peroxidases. Well-known examples of such interplay between oxidases and peroxidases are found in fungi.15,16 Many fungi secrete specialized peroxidases (e.g. lignin peroxidase and manganese peroxidase) that aid in biomass degradation.15 Except for secreting these heme-containing enzymes, these fungi also secrete various oxidases (e.g. pyranose oxidase and aryl alcohol oxidase) to serve as hydrogen peroxide producing enzymes to fuel the peroxidases.15,16 The consecutive reactions of oxidases and peroxidases are also applied in enzyme

87 Chapter 5 activity screening approaches and biosensors. Numerous assays and biosensors are based on the combination of an oxidase and a peroxidase, for instance for the detection of glucose or uric acid levels in blood serum.17–19 The activities of various oxidases were studied in a coupled assays in which typically horseradish peroxidase (HRP) is employed.20–23 HRP, however, is still extracted from horseradish because of the difficulties in the heterologous expression of this plant peroxidase.24 Therefore, for this study we selected a recently discovered bacterial peroxidase, SviDyP, which is easily produced using Escherichia coli.25,26 Fusion enzymes between oxidases and peroxidases were not made before while they form catalytically logical combinations as the oxidase-formed hydrogen peroxide will drive the fused peroxidase (Fig. 1).

O substrate oxidase 2

product peroxidase H2O2 substrate

product H2O

Figure 1. Fused oxidase-peroxidases (P-oxidases) enable O2-driven oxidative cascade reactions. The cascade reaction starting from vanillyl alcohol to divanillin is shown as example.

In this work we fused a bacterial peroxidase (SviDyP from Saccharomono- spora viridis DSM 43017, EC 1.11.1.19) to four different bacterial oxidases (EC 1.1.3.x).25,26 SviDyP belongs to the family of DyP-type peroxidases which are known for their activity on dyes and phenolic compounds.27–30 The peroxidase has been shown to be easily expressed in a bacterial host while it is also a very robust enzyme. SviDyP was fused to two FAD-containing oxidases that are active towards sugars: alditol oxidase (HotAldO) from Acidothermus cellulolyticus 11B and chitooligosaccharide oxidase (ChitO) from Fusarium graminearum.20,23 In this study, a ChitO triple mutant Q268R/G270E/S410R (ChitO*) was used because of its increased catalytic efficiency towards glucose, lactose, cellobiose and maltose. Besides these oxidases, SviDyP was fused to two other flavoprotein

88 Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions oxidases that feature a partially overlapping substrate/product scope to SviDyP: eugenol oxidase (EugO) from Rhodococcus sp. strain RHA1 and 5-hydroxy methylfurfural oxidase (HMFO) from Methylovorus sp. strain MP688.21,22 This overlap in substrate/product scope, with both fusion partners active on phenolic compounds, would allow one-pot cascade reactions. In such a way we were able to produce four novel bifunctional fusion biocatalysts that can either serve a role in biosensing or act as a catalyst for one-pot two-step cascade reactions.

Results and discussion Expression and UV-vis analysis of the fusion enzymes The four DyP-type peroxidase/oxidase fusion enzymes (which we termed P-oxidases) were made by cloning the genes of the individual oxidases ChitO*, EugO, HMFO and HotAldO C-terminally to the gene encoding for His-tagged SviDyP. The resulting fusion enzymes were overexpressed and subsequently purified by affinity chromatography yielding 26-60 mg of enzyme per liter culture broth medium. The fusion enzymes displayed an intense red-brown color indicative for binding of the heme and flavin cofactors. Analysis by UV- vis absorbance spectroscopy revealed absorbance maxima at 280 nm (protein) and 406 nm (heme) for all enzymes. The Reinheitszahl (Rz-value) of the fusion enzymes varied between 0.61 and 0.97 and suggests effective incorporation of the heme cofactor. The typical absorbance maxima of FAD, around 350-385 and 440-460 nm20–22,31, could not be observed due to the high absorbance of the heme cofactor. To confirm binding of the FAD cofactor, the purified ChitO*, EugO and HotAldO fusion enzymes were analyzed for in-gel fluorescence after SDS- PAGE. This revealed that all three fusion enzymes contained a covalently bound flavin cofactor. Such analysis was not feasible for the HMFO fusion enzyme as this flavoprotein oxidase contains a dissociable FAD. Yet, activity measurements (vide infra) confirmed that also this fusion enzyme was functional as oxidase confirming the presence of the flavin cofactor.

Oxidase and peroxidase activities of the fusion enzymes In order to verify that the prepared fusion enzymes are fully functional, activities of both fusion partners were measured (Table 1). The observed peroxidase activities for all fusion enzymes were in good agreement with thek cat determined for the isolated peroxidase. From this it can be concluded that the activity of the peroxidase was unaffected by fusing it to the oxidases. Also, the oxidases displayed activity when fused to SviDyP, although the activities were somewhat lower than the activities of the non-fused enzymes. Oxidase activities of 15-43% were observed for the fused oxidases ChitO*, EugO, HMFO and HotAldO. This

89 Chapter 5 can be partly explained by that the activity was measured at a fixed substrate concentration (kobs) which will yield lower rates when compared with kcat values taken from literature. Another explanation of the lower observed rates may lay in incomplete flavin cofactor incorporation. Yet, prolonged incubation of the fusion enzymes did not result in higher activities. The somewhat lowered oxidase activities may also be caused by structural effects of bringing the enzymes together. Nonetheless, it can be concluded that both fusion partners of the created fusion enzymes show significant activities. Therefore we started to explore their use as bifunctional biocatalysts.

Table 1. Peroxidase and oxidase activities of the fusion enzymes.a Fusion enzyme Peroxidase Oxidase -1 -1 kobs (s ) kobs (s ) P-ChitO* 7.7 (6.6) 1.0 (6.5 [23]) P-EugO 5.0 (6.6) 2.5 (12 [22]) P-HMFO 7.1 (6.6) 9.0 (21 [21]) P-HotAldO 8.6 (6.6) 0.43 (1.9 [20])

[a] The peroxidase activity was measured using Reactive Blue 19 as substrate at pH 4.0. Oxidase activity of P-EugO and P-HMFO towards vanillyl alcohol was measured at pH 7.5 and8.0 respectively. Activity of P-HotAldO towards xylitol was measured at pH 7.5 and activity of P-ChitO* towards cellobiose was measured at pH 7.6. The value in brackets indicates the kcat values of the separate enzymes as determined for SviDyP (see SI) or as reported in literature.

Biosensors: sugar detection by P-ChitO and P-HotAldO There are numerous applications in which the combined use of a peroxidase and oxidase is exploited for detection purposes. One known application that uses such an oxidase-peroxidase couple is the combined use of glucose oxidase and HRP in biosensors to determine the glucose level in blood.17 Glucose oxidase oxidizes glucose to gluconic acid in the presence of molecular oxygen, and the formed hydrogen peroxide is subsequently used to translate the oxidase activity into a readout. Here we explored SviDyP-oxidase fusion enzymes for their use in detecting sugars. SviDyP is a representative of a newly discovered class of peroxidases, the DyP-type peroxidases, which have the advantage over HRP that they are typically easily overexpressed and purified from a heterologous host, such as E. coli.24,27 First, we produced and probed native SviDyP for its performance, and found it to be mainly active at pH 3-7 at ambient temperature, with an optimum for activity towards Reactive Blue 19 at pH 4.0.Svi DyP showed to

90 Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions be active towards 4-aminoantipyrine (AAP) and 3,5-dichloro-2-hydroxybenzene sulfonic acid (DCHBS) which are commonly used as chromogenic substrates in peroxidase assays (AAP/DCHBS assay). The P-ChitO* and P-HotAldO fusion enzymes were tested with the AAP/DCHBS assay for their use in detecting sugars. ChitO is active towards mono-, di- and oligosaccharides and is the only oxidase known to be able to oxidize N-acetylated carbohydrates.23,31 Various ChitO mutants have been engineered that display distinct preferences for different carbohydrates. This would allow generating dedicated oxidase-peroxidase fusions for detection of specific mono- and oligosaccharides. HotAldO is mainly active on alditols, such as xylitol and sorbitol, which would allow its use for xylitol or sorbitol sensing.20 For testing the fusion enzymes a pH of 6 was used as this is where the pH optima of the oxidases and the peroxidase overlap. With saturating concentrations of test sugars (24 mM cellobiose for P-ChitO* and 1.4 mM xylitol for P‑HotAldO) a clear and rapid color developed with a rate of 0.3 s-1 for both sugars. In fact, the rate of color formation was close to the observed rate when native SviDyP was tested in the AAP/DCHBS assay (0.4 s-1). This indicates that at the employed conditions the peroxidase is rate-determining in the assay. A more detailed analysis of the sensitivity of P-ChitO* and P-HotAldO revealed that the fusion enzymes are able to detect low levels of cellobiose (25 µM) and xylitol (10 µM) respectively (Supporting Information Fig. S3). When using Amplex Red as fluorogenic peroxidase substrate, we could even lower the detection limit by one order of magnitude (Supporting Information Fig. S4). This shows that such peroxidase-oxidase fusion enzymes are perfectly suited for sensing purposes by harboring the full catalytic arsenal for an oxygen-driven biosensor.

One-pot two-step cascade reactions by P-HMFO and P-EugO For the generated P-HMFO and P-EugO fusion enzymes, we explored their use in fully linked cascade reactions. We imagined that, except for the use of the oxidase- generated hydrogen peroxide, also the aromatic product formed by the oxidases could be used as substrate for the fused peroxidase. DyP-type peroxidases have been shown to act on various aromatic compounds while HMFO and EugO are, among other substrates, active on monophenolic compounds.21,22,27–30 This overlap in substrate/product scope is perfect for one-pot cascade reactions. In earlier work we showed that another DyP-type peroxidase, TfuDyP, dimerizes vanillyl alcohol, vanillin and vanillyl acetone.28 Dimerization of phenolic compounds is a known reaction for peroxidases and laccases and involves oxidative phenolic coupling and keto-enol tautomerization.32,33 Divanillin is a desired taste/flavor enhancer and was reported to give an impression of creaminess to food and to mask the sense of bitterness.33 In this work we examined whether P-EugO

91 Chapter 5 and P-HMFO could produce divanillin from vanillyl alcohol, a cascade reaction in which vanillyl alcohol is oxidized to vanillin by an oxidase and subsequently dimerized to divanillin by SviDyP (Fig. 1). P-EugO and P-HMFO were incubated with vanillyl alcohol at pH 5.5, and the reaction mixtures were subsequently analyzed by LC-MS (Supporting Information Fig. S5-S14). Both P-EugO and P-HMFO were found to convert vanillyl alcohol. After 21 hours P-HMFO had oxidized 90% of vanillyl alcohol into vanillin (69%) and divanillin and related oligomers (21%). Using the same conditions, P-EugO converted 92% of vanillyl alcohol into vanillin (53%) and a higher amount of oligomers (39%) of which the most dominant product was divanillin (see Supporting Information). These results demonstrate that the fusion enzymes are suitable for the production of taste enhancer divanillin.33 Besides being recognized as flavors, vanillin and divanillin are also considered as renewable building blocks for the production of biobased plastics.34–36 Furthermore, divanillin and related phenolic dimers were reported to have an antimetastatic potential.37 Recently, we developed a one-pot two-step cascade reaction in which EugO and HRP or SviDyP were combined to produce low molecular weight lignin-like oligomers from eugenol.38 The created fusion enzyme P-EugO simplifies this newly developed approach to synthesize lignin oligomer starting from eugenol. HPLC analysis revealed that incubation of eugenol with P-EugO gave the same lignin products: phenyl coumaran, pinoresinol, coniferyl alcohol, dieugenol and a lignin tetramer (Supporting Information Fig. S15).

Conclusions In this work we made four active fusion enzymes of DyP-type peroxidaseSvi DyP and four different oxidases which we termed P-oxidases. All designed fusion enzymes could be overexpressed by E. coli as a soluble protein. SviDyP proved to be a good substitute for HRP in the horseradish peroxidase coupled assay and could be applied at an acidic pH. This SviDyP-assay could be applied to explore the substrate scope of oxidases or as biosensor for the detection of for instance sugars. SviDyP has an overlapping substrate/product scope with multiple oxidases, which is perfect for cascade reactions. Fusion enzymes P-HMFO and P-EugO were used in one-pot two-step cascade reactions. P-HMFO could be used to prepare divanillin as main product while P-EugO could be used for the synthesis of lignin oligomers. For future work it would be interesting to shift the pH optima of the oxidase and peroxidase closer together. The pH-optima of several enzymes were shifted through site directed mutagenesis in the past.39–41 By optimizing these artificial fusions of redox enzymes, novel effective bifunctional biocatalysts can be developed.

92 Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions

Materials and methods Chemicals, reagents and enzymes Chemicals, media components and reagents were obtained from Sigma(-Aldrich), Merck, BD, Acros Organics, TCI, Alfa Aesar, Thermo Fisher and Fisher Scientific. Amplex Red (Amplisyn Red) was obtained from SynChem. Oligonucleotides and horseradish peroxidase were obtained from Sigma. Restriction enzyme HindIII was obtained from New England Biolabs. The PfuUltra Hotstart PCR master mix was from Agilent Technologies, and the In-Fusion HD EcoDry cloning kit was obtained from Clontech.

Cloning The genes of oxidases ChitO* (ChitO triple mutant, Q268R/G270E/S410R), EugO, HMFO and HotAldO were amplified and cloned C-terminal to the SviDyP gene in vector pBAD His-SviDyP26, original and new plasmids (Supporting Information Table S1). pBAD His-SviDyP contains C-terminal to the SviDyP gene a stop codon, a HindIII restriction site and another stop codon. The vector was linearized using restriction enzyme HindIII. The gene of HotAldO (including a C-terminal 6xHis- tag, without the first codon for methionine) was cloned into pBAD His-SviDyP using restriction free cloning.42 The obtained plasmid contained a stop codon between the genes of SviDyP and HotAldO. The stop codon was mutated to serine by QuikChange PCR yielding vector pBAD His-SviDyP-HotAldO-His. The above mentioned stop codon was mutated to serine before the cloning of the other oxidase genes. The oxidase genes were subsequently amplified and cloned into the obtained plasmid by In-Fusion cloning (In-Fusion HD EcoDry cloning kit, Clontech). The HindIII restriction site was retained on both sides of the oxidase genes. E. coli strain TOP10 (Invitrogen) was transformed by the obtained plasmids.

Culture growth and enzyme purification Precultures were grown on 5 mL Luria-Bertani (LB) medium at 37 ºC, 135 rpm, overnight. To inoculate 400 mL Terrific Broth (TB) medium, 1:100 preculture was added. These cultures were grown at 37 ºC, 135 rpm until OD600 ~0.4-0.6, after which they were induced by 0.02% L-arabinose and grown at 17 ºC, 135 rpm for 70 hours. All cultures were supplemented with 50 μg/mL ampicillin. Cells were harvested by centrifugation at 6700xg and 4 ºC for 20 minutes (Beckman Coulter, Avanti JE centrifuge, JLA 10.500 rotor). Pellets were washed with buffer A (50 mM potassium phosphate, 0.5 M NaCl, pH 8.0), harvested by centrifugation (3000xg, 4 ºC, 40 minutes, Eppendorf centrifuge 5810R) and stored at -20 ºC till use. Prior to enzyme purification the pellets were thawed and resuspended in

93 Chapter 5 buffer A supplemented with 0.1 mM PMSF. Cells were disrupted by sonication (70% amplitude, 5 min total on time with cycles of 5 sec on and 10 sec off) and the cell-free extract was obtained by centrifugation at 16,000xg, 4 ºC for 15 minutes (VWR, Micro Star 17R centrifuge). The enzymes were purified from the cell-free extract using a 5-mL His-Trap HP column (GE Health care). The columns were washed with buffer A and buffer A supplemented with 6, 12 and 24 mM imidazole. The enzymes were eluted with 300 mM imidazole in buffer A. Subsequently the buffer was exchanged to buffer B (20 mM potassium phosphate, 150 mM NaCl, pH 7.5) using a 10‑mL Econo-Pac 10 DG desalting column (BioRad). The purified enzymes were flash frozen with liquid nitrogen and stored at ‑20 ºC. UV-vis absorbance spectra of the enzymes were recorded between 250-800 nm at ambient temperature (V-660 spectrophotometer, Jasco). The protein concentrations were determined using Lambert Beer’s law and the predicted molecular extinction coefficients (ExPASy ProtParam tool43): -1 -1 -1 -1 ε280 nm = 48,470 M cm for SviDyP, ε280 nm = 124,915 M cm for P-ChitO* (SviDyP- -1 -1 ChitO*, in case it contains one disulfide bond),280 ε nm = 127,770 M cm for ‑1 -1 P-EugO (SviDyP-EugO), ε280 nm = 126,850 M cm for P-HMFO (SviDyP-HMFO), -1 -1 ε280 nm = 116,880 M cm for P-HotAldO (SviDyP-HotAldO).

Steady-state kinetic analysis of SviDyP The steady-state kinetic parameters of SviDyP were determined for Reactive ‑1 -1, 26 Blue 19 (ε595 nm = 10 mM cm ) in a 50 mM sodium citrate buffer at pH 4.0 with 100 μM H2O2 and 20 nM enzyme. SviDyP was added to start the reaction. Oxidation of Reactive Blue 19 was followed spectrophotometrically (JASCO V-660) at ambient temperature.

Oxidase and peroxidase activity of the fusion enzymes The activities of both fusion partners were determined separately. Forall reactions, a saturating substrate concentration of twenty times the KM-value was used. For SviDyP the same reaction conditions were used as described above, with a substrate concentration of 100 μM Reactive Blue 19 (KM = 4.6 μM). For the oxidases the same reaction mixtures and pH’s were used as described before.20–23 Prior to the reactions the oxidases were incubated with 100 μM FAD for one hour at ambient temperature. Vanillyl alcohol was used as substrate for EugO22 and HMFO21, D-(+)-cellobiose for ChitO*,23 and xylitol for HotAldO20. The oxidation of vanillyl alcohol was followed spectrophotometrically at 340 nm -1 ‑1 22 21 (vanillin, ε340 nm = 14 mM cm at pH 7.5 and 8.0 ). The oxidation of D-(+)- cellobiose and xylitol were followed via a horseradish peroxidase-coupled assay. In this assay the hydrogen peroxide is produced by the oxidases and used

94 Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions by horseradish peroxidase (HRP) to couple 3,5-dichloro-2-hydroxybenzene- sulfonic acid (DCHBS) and 4-aminoantipyrine (AAP) to a pink product (ε515 nm = 26 mM-1 cm-1).20,23

SviDyP-coupled assay for the detection of oxidase substrates This assay is a variant of the horseradish peroxidase-coupled assay mentioned above and makes use of the peroxidase activity of dye decolorizing peroxidase SviDyP instead of HRP. The coupled activity of fusion enzymes P-ChitO* and P-HotAldO were determined at pH 5 (50 mM sodium citrate buffer) and pH 6 (50 mM potassium phosphate buffer). The reaction mixtures contained 0.1 mM

AAP, 1.0 mM DCHBS and 23.8 mM (20x KM) D-(+)-cellobiose for ChitO* or 1.4 mM xylitol (20x KM) for HotAldO. The formation of the pink product was followed -1 ‑1 spectrophotometrically at ambient temperature (ε515 nm = 26 mM cm ). To determine whether the oxidase or the peroxidase was the limiting factor in these reactions, the reactions were repeated in the presence of 100 μM2 H O2 to determine the optimal reaction rate of SviDyP.

Analysis of the sensitivity of the SviDyP-assay The sensitivity of the coupled-assay was studied by determining the lower concentration limit for substrate detection, as described 44before. Reaction mixtures of 200 μL contained 0.1 mM AAP, 1.0 mM DCHBS, 150 nM fusion enzyme and varying substrate concentrations (0.5 μM - 1 mM) in50 mM potassium phosphate buffer pH 6.0. D-(+)-cellobiose and xylitol were used as substrates for P‑ChitO* and P-HotAldO, respectively. The enzymes were added to start the reaction. Reactions were performed in triplicate and absorbance -1 ‑1 at 515 nm (pink product, ε515 nm = 26 mM cm ) was followed at ambient temperature for 15 minutes by a SynergyMX (BioTek) plate reader. The obtained values after 15 minutes were corrected for both the path length and the blank. For comparison, the sensitivity of the coupled-assay was also determined using 60 μM Amplex Red (10-acetyl-3,7-dihydroxyphenoxazine, Amplisyn Red) instead of AAP/DCHBS. A stock of 6.0 mM Amplex Red was prepared in DMSO. The oxidation of Amplex Red was followed by measuring the fluorescence ofthe product resorufin (excitation 530 nm, emission 590 nm) for 15 min at ambient temperature.

One-pot cascade reaction for synthesis of divanillin and related dimers and oligomers Vanillyl alcohol was dissolved in water at a concentration of 50 mM. Reaction mixtures of 2.0 mL contained 2 mM vanillyl alcohol and 1.0 µM SviDyP, P-HMFO

95 Chapter 5 or P-EugO in 50 mM sodium citrate buffer pH 5.5. In case of SviDyP 500 μM

H2O2 was added. Reactions were incubated in 15 mL closed tubes at30 °C, 100 rpm for 21 hours. Control reactions were prepared without enzyme. After 2, 3 and 21 hours samples were taken. Enzymes were heat inactivated at 95 °C for 10 min after which the samples were centrifuged for 5 min at 13,200 rpm. Reaction products were analyzed by reverse phase HPLC using a Jasco HPLC system. Samples of 10 µL were injected on a Grace Altima HP C18 column (5 μm, 2.1x150 mm, with a 1.0 cm precolumn of the same material). Solvents used: A, water with 0.1% formic acid and B, acetonitrile. HPLC method: 2 min 10% B, 2-20 min gradient to 70% B, 20-23 min 70% B, 23 min 10% B followed by 7 min re-equilibration. Detection by a UV-detector at 280 nm and flow rate of 0.5 mL/ min. LC-MS analysis was performed on Surveyor HPLC-DAD coupled to LCQ Fleet detector using scanning for both positive and negative mode. Samples were injected on a Grace Altima HP C18 column (3 μm, 2.1x100 mm, with 1.0 cm precolumn of the same material), flow rate 0.3 mL/min. Solvents used: A, water with 0.1% formic acid and B, acetonitrile with 0.08% formic acid. LC-MS method: 2 min 100% A, 2-32 min gradient to 80% B, 32-37 min 80% B, 37-38 min 100% A, 38-48 min 100% A re-equilibration.

One-pot cascade reaction for the synthesis of lignin-like oligomers from eugenol The activity of P-EugO towards eugenol was assayed as described 38before. Reaction mixtures of 2.0 mL contained 1.0 μM ‑P EugO, 10 mM eugenol and 5% DMSO (v/v) in 20 mM potassium phosphate buffer pH 6.0. A stock solution of 300 mM eugenol was prepared in DMSO. For comparison a reaction mixture containing 1.0 μM SviDyP and 1.0 μM EugO was assayed. All reactions were performed in duplicate and compared to a reaction without enzyme. Reaction mixtures were incubated at 30 ºC and 50 rpm in 20-mL pyrex tubes with a headspace to volume ratio of 10:1. Samples of 200 μL were taken after 24 and 96 hours. These samples were heat treated and analyzed by reverse phase HPLC as described above for the production of divanillin and related oligomers.

96 Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions

Acknowledgements This work was supported by the NWO graduate program: synthetic biology for advanced metabolic engineering, project number 022.004.006, The Netherlands.

Supporting information The supporting information includes Tables S1-S4 and Figures S1-S15. This information can be found online, connected to the publication: DOI: 10.1002/ cbic.201700478

97 Chapter 5

References

1 Schweizer E and Hofmann J (2004) Microbial type I fatty acid synthases ( FAS ): Major players in a network of cellular FAS systems. Microbiol. Mol. Biol. Rev. 68, 501–517. 2 Patel MS, Nemeria NS, Furey W, and Jordan F (2014) The pyruvate dehydrogenase complexes: Structure-based function and regulation. J. Biol. Chem. 289, 16615–16623. 3 Doi RH and Kosugi A (2004) Cellulosomes: Plant-cell-wall-degrading enzyme complexes. Nat. Rev. Microbiol. 2, 541–551. 4 Elleuche S (2014) Bringing functions together with fusion enzymes - from nature’s inventions to biotechnological applications. Appl. Microbiol. Biotechnol. 99, 1545–1556. 5 Pérez-Arellano I, Carmona-Álvarez F, Martínez AI, Rodríguez-Díaz J, and Cervera J (2010) Pyrroline-5-carboxylate synthase and proline biosynthesis: From osmotolerance to rare metabolic disease. Protein Sci. 19, 372–382. 6 Charles IG, Keyte JW, Brammar WJ, Smith M, and Hawkins AR (1986) The isolation and nucleotide sequence of the complex AROM locus of Aspergillus nidulans.Nucleic Acids Res. 14, 2201–2213. 7 Yourno J, Kohno T, and Roth JR (1970) Enzyme evolution: generation of a bifunctional enzyme by fusion of adjacent genes. Nature 228, 820–824. 8 Yu K, Liu C, Kim B-G, and Lee D-Y (2015) Synthetic fusion protein design and applications. Biotechnol. Adv. 33, 155–164. 9 Conrado RJ, Varner JD, and Delisa MP (2008) Engineering the spatial organization of metabolic enzymes: mimicking nature’s synergy. Curr. Opin. Biotechnol. 19, 492–499. 10 Matthews S, Tee KL, Rattray NJ, Mclean KJ, Leys D, Parker DA, Blankley RT, and Munro AW (2017) Production of alkenes and novel secondary products by P450 OleT JE using novel H2O2-generating fusion protein systems. FEBS Lett. 591, 737–750. 11 Torres Pazmiño DE, Snajdrova R, Baas B-J, Ghobrial M, Mihovilovic MD, and Fraaije MW (2008) Self-sufficient Baeyer-Villiger monooxygenases: Effective coenzyme regeneration for biooxygenation by fusion engineering. Angew. Chemie - Int. Ed. 47, 2275–2278. 12 Beyer N, Kulig JK, Bartsch A, Hayes MA, Janssen DB, and Fraaije MW (2016) P450BM3 fused to phosphite dehydrogenase allows phosphite-driven selective oxidations. Appl. Microbiol. Biotechnol., Ahead of print. 13 Meynial Salles I, Forchhammer N, Croux C, Girbal L, and Soucaille P (2007) Evolution of a Saccharomyces cerevisiae metabolic pathway in Escherichia coli. Metab. Eng. 9, 152–159. 14 Seo HS, Koo YJ, Lim JY, Song JT, Kim CH, Kim JK, Lee JS, and Choi YD (2000) Characterization of a bifunctional enzyme fusion of trehalose-6-phosphate synthetase and trehalose-6- phosphate phosphatase of Escherichia coli. Appl. Environ. Microbiol. 66, 2484–2490. 15 Abdel-Hamid AM, Solbiati JO, and Cann IKO (2013) Insights into Lignin Degradation and its Potential Industrial Applications. Adv. Appl. Microbiol. 82, 1–28. 16 Ander P and Marzullo L (1997) Sugar oxidoreductases and veratryl alcohol oxidase as related to lignin degradation. J. Biotechnol. 53, 115–131. 17 Barham D, and Trinder P (1972) An improved colour reagent for the determination of blood glucose by the oxidase system. Analyst 97, 142–145. 18 Chun HJ, Park YM, Han YD, Jang YH, and Yoon HC (2014) Paper-based glucose biosensing system utilizing a smartphone as a signal reader. Biochip J. 8, 218–226. 19 Mundaca-Uribe R, Bustos-Ramírez F, Zaror-Zaror C, Aranda-Bustos M, Neira-Hinojosa J, and Peña-Farfal C (2014) Development of a bienzymatic amperometric biosensor to determine uric acid in human serum, based on mesoporous silica (MCM-41) for enzyme immobilization. Sensors Actuators, B Chem. B Chem. 195, 58–62. 20 Winter RT, Heuts DPHM, Rijpkema EMA, van Bloois E, Wijma HJ, and Fraaije MW (2012) Hot or not? Discovery and characterization of a thermostable alditol oxidase fromAcidothermus cellulolyticus 11B. Appl. Microbiol. Biotechnol. 95, 389–403.

98 Creating oxidase-peroxidase fusion enzymes as toolbox for cascade reactions

21 Dijkman WP and Fraaije MW (2014) Discovery and characterization of a 5-hydroxymethylfurfural oxidase from Methylovorus sp. strain MP688. Appl. Environ. Microbiol. 80, 1082–1090. 22 Jin J, Mazon H, van den Heuvel RHH, Janssen DB, and Fraaije MW (2007) Discovery of a eugenol oxidase from Rhodococcus sp. strain RHA1. FEBS J. 274, 2311–2321. 23 Ferrari AR, Lee M, and Fraaije MW (2015) Expanding the substrate scope of chitooligosaccharide oxidase from Fusarium graminearum by structure-inspired mutagenesis. Biotechnol. Bioeng. 112, 1074–1080. 24 Krainer FW and Glieder A (2015) An updated view on horseradish peroxidases: recombinant production and biotechnological applications. Appl. Microbiol. Biotechnol. 99, 1611–1625. 25 Yu W, Liu W, Huang H, Zheng F, Wang X, Wu Y, Li K, Xie X, and Jin Y (2014) Application of a novel alkali-tolerant thermostable DyP-type peroxidase from Saccharomonospora viridis DSM 43017 in biobleaching of Eucalyptus Kraft pulp. PLoS One 9, e110319. 26 Colpa DI and Fraaije MW (2016) High overexpression of dye decolorizing peroxidase TfuDyP leads to the incorporation of heme precursor protoporphyrin IX. J. Mol. Catal. B. Enzym. 134, 372–377. 27 Colpa DI, Fraaije MW, and van Bloois E (2014) DyP-type peroxidases: a promising and versatile class of enzymes. J. Ind. Microbiol. Biotechnol. 41, 1–7. 28 Lončar N, Colpa DI, and Fraaije MW (2016) Exploring the biocatalytic potential of a DyP-type peroxidase by profiling the substrate acceptance of Thermobifida fusca DyP peroxidase. Tetrahedron 72, 7276–7281. 29 Ogola HJO, Kamiike T, Hashimoto N, Ashida H, Ishikawa T, Shibata H, and Sawa Y (2009) Molecular characterization of a novel peroxidase from the cyanobacterium Anabaena sp. strain PCC 7120. Appl. Environ. Microbiol. 75, 7509–7518. 30 Kim SJ and Shoda M (1999) Purification and characterization of a novel peroxidase from Geotrichum candidum Dec 1 involved in decolorization of dyes.Appl. Environ. Microbiol. 65, 1029–1035. 31 Heuts DPHM, Janssen DB, and Fraaije MW (2007) Changing the substrate specificity of a chitooligosaccharide oxidase from Fusarium graminearum by model-inspired site-directed mutagenesis. FEBS Lett. 581, 4905–4909. 32 Nishimura RT, Giammanco CH, and Vosburg DA (2010) Green, enzymatic syntheses of divanillin and diapocynin for the organic, biochemistry, or advanced general chemistry laboratory. J. Chem. Educ. 87, 526–527. 33 Krings U, Esparan V, and Berger RG (2015) The taste enhancer divanillin: a review on sources and enzymatic generation. Flavour Fragr. J. 30, 362–365. 34 Fache M, Boutevin B, and Caillol S (2015) Vanillin, a key-intermediate of biobased polymers. Eur. Polym. J. 68, 488–502. 35 Llevot A, Grau E, Carlotti S, Grelier S, and Cramail H (2015) ADMET polymerization of biobased biphenyl compounds. Polym. Chem. 6, 7693–7700. 36 Amarasekara AS and Razzaq A (2012) Vanillin-based polymers - part II: Synthesis of Schiff base polymers of divanillin and their chelation with metal ions. ISRN Polym. Sci. 2012, 1–5. 37 Jantaree P, Lirdprapamongkol K, Kaewsri W, Thongsornkleeb C, Choowongkomon K, Atjanasuppat K, Ruchirawat S, and Svasti J (2017) Homodimers of vanillin and apocynin decrease the metastatic potential of human cancer cells by inhibiting the FAK/PI3K/Akt signaling pathway. J. Agric. Food Chem. 65, 2299–2306. 38 Habib MHM, Deuss PJ, Lončar N, Trajkovic M, and Fraaije MW (2017) A biocatalytic one- pot approach for the preparation of lignin oligomers using an oxidase/peroxidase cascade enzyme system. Adv. Synth. Catal. 359, 3354–3361. 39 Pokhrel S, Joo JC, and Yoo YJ (2013) Shifting the optimum pH of Bacillus circulans xylanase towards acidic side by introducing arginine. Biotechnol. Bioprocess Eng. 18, 35–42. 40 Mendes S, Brissos V, Gabriel A, Catarino T, Turner DL, Todorovic S, and Martins LO (2015)

99 Chapter 5

An integrated view of redox and catalytic properties of B-type PpDyP from Pseudomonas putida MET94 and its distal variants. Arch. Biochem. Biophys. 574, 99–107. 41 Tomschy A, Brugger R, Lehmann M, Svendsen A, Vogel K, Kostrewa D, Lassen SF, Burger D, Kronenberger A, van Loon APGM, Pasamontes L, and Wyss M (2002) Engineering of phytase for improved activity at low pH. Appl. Environ. Microbiol. 68, 1907–1913. 42 van den Ent F and Löwe J (2006) RF cloning: a restriction-free method for inserting target genes into plasmids. J. Biochem. Biophys. Methods 67, 67–74. 43 Gasteiger E, Hoogland C, Gattiker A, Duvaud S, Wilkins MR, Appel RD, and Bairoch A (2005) Protein identification and analysis tools on the ExPASy server. In The Proteomics Protocols Handbook (Walker JM, ed), pp. 571–607. Humana Press Inc., Totowa, NJ. 44 Ferrari AR, Gaber Y, and Fraaije MW (2014) A fast, sensitive and easy colorimetric assay for chitinase and cellulase activity detection. Biotechnol. Biofuels 7, 37. 45 Habib MHM, Deuss PJ, Lončar N, Trajkovic M, and Fraaije MW (2017) A facile biocatalytic one-pot approach for the preparation of lignin oligomers. Adv. Synth. Catal. 359, 3354– 3361.

100 Chapter 6:

Conclusions and future perspectives

Dana I. Colpa and Marco W. Fraaije

Conclusions and future perspectives

DyP-type peroxidases are a relatively recently discovered family of heme- containing peroxidases. About forty DyPs were characterized to some extent in the last two decades, rapidly increasing the knowledge of this enzyme family. DyP-type peroxidases are different in fold and function from other peroxidases; they show a ferredoxin-like fold, are active at an acidic pH and oxidize/decolorize recalcitrant synthetic dyes. Based on sequence homology the family of DyP-type peroxidases can be divided in four subclasses (A-D). The crystal structures of sixteen DyPs have been elucidated, including members from every subclass. The aim of this thesis was to broaden the knowledge on class A DyP- type peroxidases. Enzymes from this subclass are from bacterial origin and Tat- dependently secreted to the periplasm. TfuDyP from Thermobifida fusca YX and SviDyP from Saccharomonospora viridis DSM 43017 were selected as model enzymes. TfuDyP is a thermostable dimeric DyP-type peroxidase which is active on various dyes, monophenolic compounds and performs the enantioselective sulfoxidation of small aromatic sulfides. SviDyP, another class A DyP-type peroxidase, shares 42% sequence identity with TfuDyP and is active at a slightly higher pH range.

Dye decolorization and divanillin production The family of dye decolorizing peroxidases (DyP) got its name due to the common activity of its members on synthetic (anthraquinone and azo) dyes. To investigate the dye decolorizing potential of this enzyme family further, we studied the ability of TfuDyP to degrade a more diverse palette of synthetic dyes and natural pigments in chapter 2. Thirty compounds from seven classes of dyes were studied together with three natural carotenoids. TfuDyP showed to be a true dye decolorizing enzyme, as it showed activity towards most studied dyes. The highest initial activities and overall conversions were observed for a xanthene dye, a copper phthalocyanine dye and dyes from the most frequently studied class of dyes, the anthraquinone dyes. Many synthetic dyes are not easily biodegraded. Although DyP-type peroxidases do not degrade these dyes to primary metabolites, they could form a good first step in the full biodegradation as the formed products might be more easily degraded in microbial catabolic routes. DyP-type peroxidases are not only active on dyes but show a much broader substrate scope, including monophenolic and lignin model compounds, Kraft lignin, β-carotene and aromatic sulfides. To explore the biocatalytic potential of these enzymes further, the research in chapter 2 was extended to a broader range of substrates. TfuDyP was found to show activity towards the six tested phenolic

103 Chapter 6 compounds, hinting at a potential role of the bacterial peroxidase in biomass (lignin) degradation. Phenolic compounds are known to be natural mediators used by other peroxidases in the degradation of lignin. Additionally, TfuDyP showed activity towards the phenolic lignin model compound guaiacylglycerol- β-guaiacyl ether, but not towards the nonphenolic veratrylglycerol-β-guaiacyl ether. The products formed upon conversion of phenolic compounds (vanillin, vanillyl alcohol, vanillylaceton and the phenolic lignin model compound) were analyzed by HPLC and LC-MS and consisted of dimers and higher oligomers. Interestingly, the formed dimer of vanillin, divanillin, is in demand as taste enhancer.

Exploring the catalytic properties of TfuDyP To gain a better understanding of the catalytic machinery ofTfu DyP, various point mutations were studied in chapter 3. Mutations were made around the heme cofactor, in the proposed hydrogen peroxide tunnel, to shift the pH optimum for activity and to study the dependence of TfuDyP activity on long-range electron transfer (LRET). The importance of the catalytically active arginine and the aspartate from the GXXDG-motif, which is the proposed acid-base catalyst, were reported in earlier work. In this work, the heme ligand, a histidine, was mutated to a cysteine to convert TfuDyP into a heme-thiolate enzyme with potentially improved peroxygenase activity. This histidine is however crucial for TfuDyP: mutating this residue yielded an inactive yellow-colored enzyme. DyP-type peroxidases contain a highly conserved GXXDG-motif. Mutant G243A (GXXDG), designed to narrow the proposed hydrogen peroxide tunnel, reduced the activity towards Reactive Blue 19 to 15%. Furthermore, the area around the heme propionate and the catalytically active arginine was found to be important for the pH optimum for activity. Mutations A245R and N246L in this area broadened the pH range for activity towards Reactive Blue 19 by one pH unit. For the oxidation of substrates that are too large to enter the active site, long- range electron transfer pathways from the protein surface to the heme cofactor were reported for multiple (DyP-type) peroxidases. To test whether TfuDyP is dependent on such a mechanism all eight surface exposed tyrosines and tryptophans within 20 Å from the heme cofactor were mutated to phenylalanine individually. The obtained mutants showed however a comparable activity as the wild-type enzyme, indicating thatTfu DyP is either not dependent on LRET or able to use multiple pathways.

104 Conclusions and future perspectives

High overexpression of TfuDyP leads to the incorporation of heme precursor PPIX For the industrial applicability of enzymes a high overexpression level is desired. In contrast to eukaryotic peroxidases are bacterial DyPs relatively easily expressed in a heterologous host, such as Escherichia coli. The heterologous expression of TfuDyP in E. coli was successfully boosted through a SUMO-fusion, increasing the enzyme expression from 25 mg to 200 mg per liter culture broth (chapter 4). The highly overexpressed SUMO-fused enzyme was more purple in color and while it displayed the typical Soret band, it showed five instead of the expected one or two Q-bands in UV-vis absorbance spectroscopy. Moreover, the specific activity decreased to 22%. In an attempt to understand the reason behind the lower activity, different expression conditions, expression hosts and enzyme fusions were studied. The decrease in activity was not due to the fusion to SUMO; highly overexpressed SUMO-TfuDyP, MBP-TfuDyP and TfuDyP (with only an N-terminal His-tag) gave comparable results. Addition of hemin, iron salts or a heme precursor (δ-aminolevulinic acid) to the cultures increased the enzyme specific activity only slightly. Important information was obtained when SUMO-TfuDyP was expressed in different E. coli strains and under varying arabinose concentrations for induction. These experiments clearly showed that the activity of the purified enzyme was dependent on the expression level. Strains which yielded less than 50 mg enzyme per liter culture broth produced fully active enzyme, while strains with an expression level above 150 mg yielded enzyme with a reduced specific activity. Further analysis of the isolated heme cofactors by mass spectrometry revealed a crucial difference between the active and less active enzymes: the active enzyme samples contained the expected heme cofactor while the less active enzyme samples were found to contain a significant amount of an iron deficient heme precursor protoporphyrin IX (PPIX). As a heme enzyme deficient in iron is unable to perform catalysis, this solved the riddle of the relatively low specific activity when the enzyme was highly overexpressed. To study whether the in vivo binding of heme precursor PPIX is specific to TfuDyP or a more general feature of DyP-type peroxidases, we boosted the expression level of SviDyP from S. viridis to approximately 100 mg per liter culture broth. UV-vis absorbance spectroscopy revealed that SviDyP too binds a significant amount of PPIX, showing the same five PPIX specific Q-bands in UV- vis absorbance spectroscopy. This shows that E. coli has to be tuned when it is used for the overexpression of DyP-type peroxidases. The insufficient conversion of PPIX into iron-containing heme b may be solved by the coexpression of heme ferrochelatase.

105 Chapter 6

Oxidase-peroxidase fusion enzymes: biosensors and cascade reactions Many enzymes take part in metabolic routes in nature. During evolution some enzymes formed complexes or fused together to reach a higher efficiency. Oxidases and peroxidases form catalytically logical combinations and are often coexpressed, for instance in biomass degradation. The oxidases use molecular oxygen as oxidant and produce hydrogen peroxide as by-product, which in turn fuels the peroxidases. Even though various artificial fusion enzymes were made before, no fusion enzyme between an oxidase and a peroxidase had been reported. Chapter 5 presents the first artificial oxidase-peroxidase fusion enzymes produced as recombinant proteins. In chapter 5, bacterial DyP-type peroxidase SviDyP was fused to four different flavoprotein oxidases: chitooligosaccharide oxidase (ChitO), eugenol oxidase (EugO), HMF oxidase (HMFO) and alditol oxidase (HotAldO). Fusion of these enzymes yielded four well expressed bi-functional enzymes. Two of the artificial fusion enzymes were applied as biosensor for the detection of sugars; P-ChitO and‑ P HotAldO (in which P stands for peroxidase SviDyP). With such fusion enzymes, the model substrates cellobiose and xylitol could be detected with a detection limit in the low μM range. The coupled assay fundamental to these biosensors is closely related to the well-established HRP-assay in which the activity of an oxidase is coupled to the activity of plant peroxidase horseradish peroxidase (HRP).Svi DyP proved to be a good substitution for HRP in this assay; it is well expressed in a bacterial host, performs the same reaction, and can be applied at a more acidic pH range. The other two oxidases, EugO and HMFO, show a partially overlapping substrate scope with peroxidase SviDyP. These three enzymes are all active on phenolic compounds. This overlap in substrate scope is perfectly suited for molecular oxygen driven one-pot two-step cascade reactions. ‑ P HMFO and P-EugO produced the taste enhancer divanillin in a two-step reaction from vanillyl alcohol. P-EugO converted eugenol in lignin-like dimers and oligomers.

Future perspectives DyP-type peroxidases are named after their activity on synthetic dyes. The physiological substrate is however still unknown. Until now, most evidence points towards a role in biomass degradation. DyP-type peroxidases from class A and D are (Tat-dependently) secreted, which supports this hypothesis. Identification of the physiological substrates of (intra/extracellular) DyPs would be very informative. Besides the unknown physiological substrates is the catalytic mechanism of DyP-type peroxidases not well understood. DyP-type peroxidases rely on a heme cofactor for their activity. Different oxidation sites were observed

106 Conclusions and future perspectives or proposed for different members of this family. AauDyP, TcDyP and VcDyP showed long-range electron transfer (LRET) from the heme cofactor to a Tyr/ Trp on the protein surface, implying that this is the oxidation site for (bulky) substrates.1–3 Whether all DyP-type peroxidases use LRET is unknown. DyP from Bjerkandera adusta Dec 1 was co-crystallized with two substrates, each bound to the surface exposed heme propionate, suggesting that this is the oxidation site of DyP.4 In future work, oxidation sites and reaction pathways could be studied further through site-directed mutagenesis and enzyme crystallization in the presence and absence of small and bulky substrates. Small substrates might react at the heme-iron directly while large substrates get oxidized at the surface exposed heme propionate or via LRET. For the industrial applicability of enzymes a high overexpression level in a heterologous host is desired. Both bacterial and fungal DyPs are easily heterologously overexpressed in a bacterial host which is in stark contrast to peroxidases from other enzyme families. Enzymes from class D seem to be more effective catalysts, as they show a higher catalytic efficiency towards 6 7 -1 ‑1 anthraquinone dyes (kcat/KM = 10 -10 s M ) when compared to enzymes from classes A and B (104-105 s-1 M-1).5 The sequence identity between DyPs from bacterial classes A and B and fungal class D is low, <20%. The higher catalytic efficiency and the possibility to express fungal DyPs from class D heterologously in a bacterial host makes them interesting candidates for future research. D-type DyPs could potentially substitute plant and fungal peroxidases, which are not easily heterologously produced.6 Most characterized DyPs were heterologously overexpressed in E. coli. Chapter 4 showed however that the high overexpression of TfuDyP and SviDyP resulted in partially inactive enzymes due to the incorporation of the iron deficient heme precursor protoporphyrin IX. To circumvent this problem, the expression level was lowered to approximately 50 mg enzyme per liter culture broth, at which the enzyme was still fully loaded with heme. In future work this problem could be solved by increasing the heme amount in the cell by co-expression of heme ferrochelatase, an enzyme which incorporates iron in heme.7 Furthermore, the low selectivity of some DyP-type peroxidases between heme and heme precursor protoporphyrin IX, and the expression of other DyP- type peroxidases as apo-enzyme offers some interesting possibilities. Key et al. (2016) showed that the reconstitution of apo heme b (Fe-porphyrin IX, Fe- PIX) containing enzymes with unnatural metalloporphyrins (M-PIX) resulted in abiological catalysis.8 It would be very interesting to load DyP-type peroxidases with e.g. Mn, Co, Cu, Zn or Ru-based porphyrins and probe such engineered enzymes for novel activities.

107 Chapter 6

DyP-type peroxidases show a pH optimum for activity in the acidic range, while oxidases show a pH optimum in the neutral to mild alkaline range. For the applicability of DyPs in cascade reactions with oxidases, it would be beneficial to shift the pH optimum towards a more neutral pH range. The pH optimum of other enzymes, including PpDyP, were shifted by site-directed mutagenesis in the past.9–11 Yet, as shown in chapter 3, analogous mutations in TfuDyP did not result in a major change of the pH optimum, which suggests that changing the pH optimum of a DyP is not always straightforward. Alternatively, DyPs with a higher pH optimum for activity could be isolated from alkaliphilic or alkalitolerant microorganisms.12 The Gram-positive alkaliphilic bacteria Cellulomonas bogoriensis and Dietzia natronolimnaea contain one and two DyP-type peroxidases respectively. These DyPs are most related to the Tat- dependently secreted DyPs from class A. The DyP from C. bogoriensis shows an interesting feature; the aspartate of the conserved GXXDG-motif is substituted for a glutamate. Another highly promising DyP was found in the genome of the biomass degrading thermophilic and alkalitolerant Aspergillus fumigatus Z5. This enzyme is predicted to belong to class D but no secretion sequence could be identified. In the genome of fungus Aspergillus versicolor five genes were found that resemble to DyPs; three DyPs D, a DyP C and one DyP that surprisingly resembled to the bacterial DyPs from class B. Small modifications in the GXXDG- motif were also observed for some of these DyPs; the last glycine of the GXXDG- motif is substituted for a Leu, Thr or His in the DyP from A. fumigatus, and two DyPs D and the DyP C from A. versicolor. Besides shifting the pH optimum, DyP-type peroxidases could be engineered towards an improved thermostability and/or an improved tolerance towards hydrogen peroxide or organic solvents. To improve the biomass degrading capacity of DyPs, it would be desirable to increase the specific activity towards natural mediators such as phenols and Mn2+. Next to improving the activity or the robustness of these enzymes, DyPs could be engineered towards non-natural activities; for instance towards (an improved) peroxygenase activity or to become an oxygen carrier as hemoglobin and myoglobin. Exchanging the activities of heme b containing peroxidases, peroxygenases and oxygen carriers was done before.13–15 The research presented in this thesis showed that the dye decolorizing peroxidase TfuDyP is a true dye decolorizing enzyme. In future work, DyP- type peroxidases could be applied in the bioremediation of synthetic dye contaminated waste water. Even though DyP-type peroxidases are not capable of the full biodegradation of these dyes to regular metabolites, they form a good first step in this process. To go one step further, a whole cell based system could

108 Conclusions and future perspectives be created for this purpose through the extracellular coexpression of a DyP-type peroxidase with other oxidative enzymes. Upon coexpression of other required catabolic enzymes the dyes could be fully mineralized and detoxified. Besides waste water treatment, in-textile bleaching for the creation of a pattern or for the production of a stonewashed-look could be an interesting application area for DyPs. Bleaching is also of interest to the food industry. In fact, a fungal DyP was patented and marketed under the name MaxiBright for the enzymatic bleaching of carotenoids in whey containing food and beverages. Next to bleaching it is worth noting that DyPs can be used for the dimerization and polymerization of phenolic compounds. Both DyPs that were studied as part of this thesis were found to catalyze the dimerization of the phenolic compound vanillin to the taste enhancer divanillin and higher oligomers. Divanillin gives an impression of creaminess to food and masks the sense of bitterness, and is therefore attractive for the food industry. Polymerization of vanillin is considered for the production of bio-based plastics.16 Lastly, biosensors form an interesting field of application for DyP-type peroxidases. Oxidase substrates could be detected through a coupled reaction between an oxidase and a peroxidase, during which a chromogenic compound is formed. In this work, sugars cellobiose and xylitol were detected with a μM detection limit by P-ChitO and P-HotAldO, respectively. Future enzyme discovery and enzyme engineering efforts will reveal the full potential of DyP-type peroxidases.

109 Chapter 6

References

1 Strittmatter E, Liers C, Ullrich R, Wachter S, Hofrichter M, Plattner DA, and Piontek K (2013) First crystal structure of a fungal high-redox potential dye-decolorizing peroxidase substrate interaction sites and long-range electron transfer. J. Biol. Chem. 288, 4095–4102. 2 Uchida T, Sasaki M, Tanaka Y, and Ishimori K (2015) A dye-decolorizing peroxidase from Vibrio cholerae. Biochemistry 54, 6610–6621. 3 Shrestha R, Chen X, Ramyar KX, Hayati Z, Carlson EA, Bossmann SH, Song L, Geisbrecht B V, and Li P (2016) Identification of surface-exposed protein radicals and a substrate oxidation site in A-class dye-decolorizing peroxidase from Thermomonospora curvata. ACS Catal. 6, 8036–8047. 4 Yoshida T, Tsuge H, Hisabori T, and Sugano Y (2012) Crystal structures of dye-decolorizing peroxidase with ascorbic acid and 2,6-dimethoxyphenol. FEBS Lett. 586, 4351–4356. 5 Yoshida T and Sugano Y (2015) A structural and functional perspective of DyP-type peroxidase family. Arch. Biochem. Biophys. 574, 49–55. 6 Veitch NC (2004) Horseradish peroxidase: A modern view of a classic enzyme. Phytochemistry 65, 249–259. 7 Sudhamsu J, Kabir M, Airola M V, Patel BA, Yeh S-R, Rousseau DL, and Crane BR (2010) Co-expression of ferrochelatase allows for complete heme incorporation into recombinant proteins produced in E. coli. Protein Expr. Purif. 73, 78–82. 8 Key HM, Dydio P, Clark DS, and Hartwig JF (2016) Abiological catalysis by artificial haem proteins containing noble metals in place of iron. Nature 534, 534–537. 9 Mendes S, Brissos V, Gabriel A, Catarino T, Turner DL, Todorovic S, and Martins LO (2015) An integrated view of redox and catalytic properties of B-type PpDyP from Pseudomonas putida MET94 and its distal variants. Arch. Biochem. Biophys. 574, 99–107. 10 Tomschy A, Brugger R, Lehmann M, Svendsen A, Vogel K, Kostrewa D, Lassen SF, Burger D, Kronenberger A, van Loon APGM, Pasamontes L, and Wyss M (2002) Engineering of phytase for improved activity at low pH. Appl. Environ. Microbiol. 68, 1907–1913. 11 Pokhrel S, Joo JC, and Yoo YJ (2013) Shifting the optimum pH of Bacillus circulans xylanase towards acidic side by introducing arginine. Biotechnol. Bioprocess Eng. 18, 35–42. 12 Grant WD (2006) Cultivation of aerobic alkaliphiles. Meth. Microbiol. 35, 439–449. 13 Ozaki S, Matsui T, and Watanabe Y (1996) Conversion of myoglobin into a highly stereo- specific peroxygenase by the L29H/H64L mutation. J. Am. Chem. Soc. 118, 9784–9785. 14 Du J, Sono M, and Dawson JH (2010) Functional switching of Amphitrite ornata

dehaloperoxidase from O2-binding globin to peroxidase enzyme facilitated by halophenol

substrate and H2O2. Biochemistry 49, 6064–6069. 15 Li L-L, Yuan H, Liao F, He B, Gao S-Q, Wen G-B, Tan X, and Lin Y-W (2017) Rational design of artificial dye-decolorizing peroxidases using myoglobin by engineering Tyr/Trp in the heme centre. Dalt. Trans. 46, 11230–11238. 16 Fache M, Boutevin B, and Caillol S (2015) Vanillin, a key-intermediate of biobased polymers. Eur. Polym. J. 68, 488–502.

110 Chapter 7:

Nederlandse samenvatting Voor de geïnteresseerde leek

Dana I. Colpa

Nederlandse samenvatting

Biochemie en moleculaire enzymologie Biochemie ligt op het grensvlak van biologie en scheikunde en probeert het leven op het niveau van moleculen te begrijpen. Moleculen zijn de kleinste deeltjes van een stof, die nog de chemische eigenschappen van die stof bezitten. In de biotechnologie wordt gebruik gemaakt van deze kennis voor praktische doeleinden. Eeuwen geleden werden micro-organismen al gebruikt voor de productie van yoghurt, brood, wijn en bier, hoewel toen nog niet bekend was dat bacteriën en gisten verantwoordelijk zijn voor de fermentatie van deze producten. Tegenwoordig wordt deze kennis toegepast bij het vervaardigen van allerlei producten, zoals voedingsmiddelen, geneesmiddelen en materialen. Hierbij worden niet alleen micro-organismen zoals bacteriën en gisten gebruikt, maar ook specifieke eiwitten (enzymen) uit deze micro-organismen. Enzymen worden bijvoorbeeld toegepast in glucosemeters voor het bepalen van het glucosegehalte in een druppel bloed, in de productie van kaas voor het stremmen van melk en in wasmiddelen voor de afbraak van vetten, eiwitten en zetmeel. Een moleculaire enzymoloog bestudeert hoe enzymen werken op moleculair niveau en welke reacties ze versnellen (katalyseren). Binnen de moleculaire enzymologie worden natuurlijke enzymen niet alleen onderzocht, ze worden ook aangepast en geoptimaliseerd. Daarbij kunnen verschillende eigenschappen van het enzym worden aangepast of verbeterd, zoals de activiteit, de stabiliteit of de optimale zuurgraad waarbij ze werken (het pH-optimum).

Enzymen Enzymen zijn eiwitten die chemische reacties versnellen oftewel katalyseren; ze doen dat zonder daarbij zelf verbruikt te worden. Eiwitten bestaan uit ketens van kleine schakels, aminozuren genaamd. In de natuur bestaan twintig verschillende aminozuren. Door deze aminozuren steeds in een andere volgorde aan elkaar te koppelen, ontstaan zeer veel verschillende aminozuurketens. Deze ketens vouwen zich op tot verschillende eiwitten. Veel van de eiwitten, die in de natuur voorkomen, functioneren als enzymen. De in dit proefschrift onderzochte enzymen, TfuDyP en SviDyP, bestaan elk uit een keten van ongeveer 400 aminozuren, die opgevouwen min of meer eivormig zijn. Veel enzymen hebben een extra molecuul, een cofactor, nodig om reacties te katalyseren. Een deel van de cofactoren wordt gemaakt uit vitaminen, zoals FAD en NAD(P)H. Andere cofactoren zijn niet op vitaminen gebaseerd, zoals heem in hemoglobine. TfuDyP en SviDyP bevatten ook een heem cofactor en zijn hierdoor net als hemoglobine rood van kleur. TfuDyP en SviDyP vervoeren in tegenstelling tot hemoglobine geen zuurstof, maar voeren oxidatiereacties uit met behulp van waterstofperoxide. In hoofdstuk 5 worden naast peroxidasen ook oxidasen bestudeerd. Oxidasen

113 Chapter 7 bevatten een flavine cofactor (afgeleid van riboflavine, vitamine B2) en zijn hierdoor geel van kleur. Deze enzymen katalyseren oxidatiereacties met behulp van zuurstof en produceren daarbij waterstofperoxide als bijproduct. De geproduceerde waterstofperoxide kan vervolgens door de peroxidasen worden gebruikt. Dit proefschrift beschrijft het onderzoek naar twee kleurstof afbrekende peroxidasen (Engels: dye decolorizing peroxidases, DyPs). Hiervoor zijn TfuDyP van de bacterie Thermobifida fusca en SviDyP van de bacterie Saccharomonospora viridis gebruikt als modelenzymen. Zoals de naam van deze enzymen aangeeft, zijn ze betrokken bij de ontkleuring en afbraak van oplosbare kleurstoffen. Het onderzoek beschreven in dit proefschrift is vooral gericht op het vergroten van de kennis over deze enzymfamilie. We hebben hierbij onder andere onderzocht hoe het reactiemechanisme van het enzym in elkaar zit en op welke kleurstoffen en andere substraten het enzym actief is. Tot slot hebben we een van de enzymen gekoppeld aan een oxidase om tweestapsreacties uit te voeren. Hiermee konden we in twee stappen de smaakversterker divanilline maken of verschillende suikers detecteren.

Hoofdstuk 2: Exploring the biocatalytic potential of a DyP-type peroxidase by profiling the substrate acceptance of Thermobifida fusca DyP peroxidase. Vertaling: Het verkennen van het biokatalytisch potentieel van een DyP-type peroxidase door te analyseren welke (kleur)stoffen omgezet worden door de DyP-type peroxidase uit Thermobifida fusca.

DyP-type peroxidasen (of dye decolorizing peroxidases) ontkleuren kleurstoffen. Deze familie van peroxidasen is zelfs vernoemd naar deze eigenschap. De meeste synthetische kleurstoffen, kleurstoffen die bijvoorbeeld in de textielindustrie worden toegepast, zijn moeilijk biologisch afbreekbaar. De twintig jaar geleden ontdekte DyP-type peroxidasen breken deze kleurstoffen gedeeltelijk af en zouden kunnen worden toegepast als eerste stap in dit afbraakproces. Andere enzymen kunnen de gevormde producten verder afbreken tot nieuwe grondstoffen. In hoofdstuk 2 is onderzocht welke kleurstofgroepen door een DyP- type peroxidase kunnen worden afgebroken. In dit hoofdstuk is TfuDyP gebruikt als modelenzym. De activiteit van TfuDyP werd getest op dertig kleurstoffen uit zeven verschillende kleurstofgroepen en drie natuurlijke pigmenten. TfuDyP toonde aan actief te zijn op een groot aantal kleurstoffen van zes verschillende kleurstofgroepen, waaronder kleurstofgroepen die nog niet eerder waren onderzocht met DyPs.

114 Nederlandse samenvatting

In de scheikunde behoren kleurstoffen tot de groep van aromatische verbindingen (gebaseerd op een aromatische groep in het molecuul, niet op de geur). Behalve naar de activiteit op kleurstoffen is in hoofdstuk 2 ook gekeken naar de activiteit van TfuDyP op andere aromatische stoffen, waaronder fenolen en bouwstenen van lignine. Lignine is naast cellulose het polymeer dat structuur en stevigheid geeft aan planten. TfuDyP is actief op fenolen en zorgt voor een koppeling tussen twee fenolen (dimerisatie) of zelfs tussen drie, vier of meer fenolen (oligomerisatie, polymerisatie). Eén van de geteste fenolen was vanilline, de smaak- en geurstof van vanille. Vanilline wordt door TfuDyP omgezet (gedimeriseerd) in een andere smaakversterker, divanilline. Divanilline versterkt de smaak van romigheid en maskeert bitterheid in voedingsmiddelen.

Hoofdstuk 3: Exploring the catalytic properties of DyP-type peroxidase TfuDyP by site-directed mutagenesis. Vertaling: Het verkennen van de katalytische eigenschappen van DyP-type peroxidase TfuDyP door middel van gerichte mutagenese.

Het doel van het onderzoek in hoofdstuk 3 is het vergroten van de kennis over de precieze werking van het enzym TfuDyP. Het onderzoek was er op gericht om zowel de locatie als het reactiemechanisme van de substraatoxidatie vanTfu DyP te bepalen. Zoals in hoofdstuk 2 is aangetoond oxideert TfuDyP veel verschillende substraten, waaronder fenolen en grotere kleurstofmoleculen. Uit eerder onderzoek is gebleken dat TfuDyP andere relatief kleine moleculen (aromatische sulfiden) direct bij het katalytische centrum van het enzym oxideert, bijhet ijzeratoom van de heem. De heem van TfuDyP bevindt zich echter in het enzym en is alleen bereikbaar via een paar nauwe tunnels. Daardoor kunnen grote moleculen zoals kleurstoffen de heem niet bereiken; zij zullen op een andere plek aan het oppervlak van het enzym moeten reageren. Om de locatie hiervan te bepalen zijn bepaalde gebieden rondom de heem en aan het oppervlak van het enzym onderzocht met behulp van gerichte mutagenese. Dit is, een techniek waarbij één aminozuur van het enzym gemuteerd wordt naar een van de andere 19 aminozuren. Vaak is er een vermoeden welke aminozuren betrokken zijn bij de reactiviteit van het enzym. Door deze aminozuren te veranderen kan dit vermoeden bevestigd of ontkracht worden. Voor TfuDyP werden aminozuren direct naast de heem cofactor en tyrosines en tryptofanen aan het oppervlak onderzocht. De tyrosines en tryptofanen werden onderzocht omdat deze twee aminozuren radicalen kunnen vormen en peroxidasen radicaalreactie uitvoeren. TfuDyP is voornamelijk actief bij een zure pH van 3 tot 4. Naast mutaties om de locatie van de substraatoxidatie te bepalen zijn er in hoofdstuk 3 ook mutaties

115 Chapter 7 gemaakt met als doel het pH-optimum vanTfu DyP te verschuiven naar een meer neutrale pH. Uit het onderzoek volgde dat het gebied rondom de heem belangrijk is voor het pH-optimum van TfuDyP en dan voornamelijk rondom de heempropionaat (staart van de heem) en in de buurt van de katalytische arginine naast de heem. Twee mutaties in dit gebied verruimden het pH-bereik voor de activiteit met één pH-unit richting een neutrale pH. Verder onderzoek naar dit gebied zou het pH-optimum van TfuDyP mogelijk verder kunnen verschuiven. Daarnaast was er eerder voorspeld dat grote substraten bij tyrosines en tryptofanen aan het enzymoppervlak reageren. Mutagenese van deze aminozuren had geen effect op de kleurstofafbraak. Hieruit concluderen we dat de oxidatie van grote moleculen via een ander mechanisme verloopt of dat TfuDyP na het muteren van één van de tyrosines of tryptofanen een andere tyrosine of tryptofaan kan gebruiken.

Hoofdstuk 4: High overexpression of dye decolorizing peroxidase TfuDyP leads to the incorporation of heme precursor protoporphyrin IX. Vertaling: Hoge overexpressie (productie) van kleurstof afbrekende peroxidase TfuDyP leidt tot de binding van heemprecursor protoporphyrine IX.

Voor de industriële toepasbaarheid van enzymen is een eenvoudige en groot- schalige productie essentieel. Enzymen worden bij voorkeur in ongevaarlijke laboratoriumstammen geproduceerd, zoals Saccharomyces cerevisiae (bakkers- gist), Lactococcus lactis (melkzuurbacterie) of een onschuldige Escherichia coli stam. In het onderzoek dat beschreven is in dit proefschrift is E. coli gebruikt. Om E. coli enzymen te laten produceren, wordt een plasmide (een stuk DNA met daarop de informatie voor het maken van deze enzymen) inde E. coli cellen geïntroduceerd. In eerder werk werd hiermee ongeveer 3 mg TfuDyP per liter E. coli cultuur geproduceerd. Om het enzym aantrekkelijker te maken voor industriële toepassingen is in hoofdstuk 4 weergegeven hoe de productie van TfuDyP verhoogd kon worden tot ongeveer 200 mg per liter E. coli cultuur. Het verkregen eiwit was rood, maar voor het grootste deel inactief. In de rest van hoofdstuk 4 is het onderzoek beschreven waarbij werd nagegaan wat de daling in activiteit van TfuDyP veroorzaakte. Het actieve en inactieve enzym zijn met verschillende methoden bestudeerd. DyP-type peroxidasen bevatten net als hemoglobine een heem cofactor. Met behulp van massaspectrometrie is de massa van de heem cofactor bepaald; dit liet een duidelijk verschil zien tussen het actieve en het inactieve enzym. Het actieve enzym bevatte zoals verwacht heem, het inactieve enzym bevatte echter de heem precursor protoporphyrine IX (PPIX). PPIX is het voorlaatste molecuul in de heemproductie

116 Nederlandse samenvatting en bevat nog geen ijzer. Het ijzeratoom is essentieel voor de activiteit van het enzym. In dit hoofdstuk laten we zien dat dit probleem zich alleen voordoet bij een enzymproductieniveau boven de 100 mg per liter. In de toekomst zou dit probleem verholpen kunnen worden door de E. coli cellen naast TfuDyP ook een enzym te laten produceren dat betrokken is bij de laatste stap van de productie van heem.

Hoofdstuk 5: Creating oxidase-peroxidase fusion enzymes as a toolbox for cascade reactions. Vertaling: Creëren van oxidase-peroxidase fusie-enzymen voor kettingreacties.

In de natuur zijn de meeste enzymen betrokken bij de stofwisseling (het metabolisme) van de cel. Voor de vertering van voedsel of de productie van complexere stoffen zoals hormonen en vitaminen zijn reactiepaden ontwikkeld waarbij meerdere enzymen betrokken zijn. Om deze cascadereacties (kettingreacties) efficiënt te laten verlopen vormen sommige enzymen inde natuur een complex. In sommige gevallen is dit complex zelfs gefuseerd tot één eiwit met twee of meer functies. Peroxidasen en oxidasen vormen katalytisch gezien een logische combinatie. Oxidasen produceren waterstofperoxide als bijproduct, wat vervolgens door de peroxidasen gebruikt wordt als ‘brandstof’ voor hun reacties. In de literatuur zijn veel voorbeelden van natuurlijke en artificiële fusie-enzymen te vinden, maar geen fusies tussen oxidasen en peroxidasen. Hoofdstuk 5 beschrijft de ontwikkeling van vier artificiële oxidase- peroxidase fusie-enzymen. In dit onderzoek is DyP-type peroxidase SviDyP gefuseerd met vier oxidasen; chitooligosaccharide oxidase (ChitO), eugenol oxidase (EugO), HMF oxidase (HMFO) en alditol oxidase (HotAldO). De fusie- enzymen SviDyP-ChitO en SviDyP-HotAldO zijn gebruikt als biosensor voor de detectie van de suikers cellobiose en xylitol. De detectielimiet van deze suikers was ongeveer een milligram per liter. Met deze fusie-enzymen kunnen naast xylitol en cellobiose ook andere suikers gedetecteerd worden, namelijk suikers waarop ChitO en HotAldO actief zijn. De andere twee oxidasen, EugO en HMFO, zijn gedeeltelijk actief op dezelfde substraten als SviDyP. Deze overlap in substraten is ideaal voor cascadereacties. Hoofdstuk 2 liet zien dat TfuDyP vanilline omzet in de smaakversterker divanilline. SviDyP katalyseert dezelfde reactie. In hoofdstuk 5 gaan we een stap verder en produceren divanilline met behulp van de fusie-enzymen SviDyP-EugO en SviDyP-HMFO in een tweestapsreactie uit vanillyl alcohol. Daarnaast is SviDyP-EugO toegepast in de cascadereactie voor de productie van lignine-achtige dimeren en oligomeren uit eugenol (kruidnagelgeur).

117 Chapter 7

Vooruitblik In dit proefschrift is een aantal vragen over DyP-type peroxidasen (kleurstof afbrekende peroxidasen, DyPs) beantwoord. In hoofdstuk 2 is aangetoond dat TfuDyP actief is op fenolen en een groot aantal kleurstoffen uitzes kleurstofgroepen, waaronder kleurstofgroepen die nog niet eerder onderzocht waren met DyP-type peroxidasen. In hoofdstuk 4 is de productie van TfuDyP verbeterd en is geleerd wat er fout kan gaan met de heem cofactor bij een hoge productie van DyPs. Andere vragen blijven echter nog onbeantwoord. Hoewel er verschillende aanwijzingen zijn dat DyPs betrokken zijn bij de afbraak van biomassa, is het fysiologische substraat onbekend. Daarnaast wordt ook het reactiemechanisme van deze DyP-type peroxidasen nog niet goed begrepen. Voor een aantal DyPs is er meer onderzoek gedaan naar het reactiemechanisme en is het opgehelderd. De familie van DyP-type peroxidasen bestaat uit verschillende klassen; de resultaten uit deze klassen komen echter niet altijd overeen. Voor een beter begrip van deze familie van peroxidasen is het daarom belangrijk om in de toekomst het fysiologische substraat en het reactiemechanisme van de verschillende klassen DyPs te onderzoeken. Dit kan een basis vormen voor het maken van verbeterde varianten van de enzymen. Voor de industriële toepassing van enzymen zijn activiteit en stabiliteit essentieel. Hoewel DyP-type peroxidasen stabieler zijn bij een neutrale pH, zijn de meeste DyPs actief in een zuur milieu (lage pH). Ook voor een kettingreactie met andere enzymen, zoals oxidasen, zou het gunstig zijn als DyPs bij een hogere pH actief zouden zijn. Het is daarom interessant om te zoeken naar een DyP die van nature actief is bij een hogere pH. Hiervoor zou gezocht kunnen worden in bijvoorbeeld alkalifiele/alkalitolerante organismen. Een andere optie is het aanpassen van het pH-optimum voor de activiteit met behulp van (gerichte) mutagenese. Ook zouden de thermostabiliteit of de tolerantie van deze enzymen voor organische oplosmiddelen of het cosubstraat waterstofperoxide, verbeterd kunnen worden. Naast de stabiliteit van DyPs zou de activiteit op verschillende gewenste substraten verbeterd kunnen worden, bijvoorbeeld voor een aantal fenolen die als mediator betrokken zijn bij de afbraak van biomassa. De mogelijke toepassingsgebieden voor DyP-type peroxidasen zijn divers. DyPs kunnen bijvoorbeeld worden toegepast in de textielindustrie voor de zuivering van met kleurstof verontreinigd afvalwater, of voor het creëren van patronen of een stone-washed-look door middel van enzymatische bleking van textiel. Ook in de voedingsmiddelenindustrie worden DyPs toegepast voor het afbreken van kleurstoffen. Een DyP uit de schimmel Mycetinis scorodonius is op de markt gebracht onder de naam MaxiBright. Bij de productie van kaas wordt in sommige gevallen carotenoiden toegevoegd voor een gelere kleur.

118 Nederlandse samenvatting

Als bijproduct wordt wei opgevangen dat vervolgens gebruikt wordt in zuivel- en frisdranken. In wei is deze gele kleur niet gewenst en wordt MaxiBright toegevoegd om deze carotenoiden af te breken. Daarnaast zouden DyPs in de voedingsmiddelenindustrie ook kunnen worden toegepast voor de productie van de smaakversterker divanilline, zoals beschreven is in hoofdstuk 2 en 5. Naast de voedsel- en de textielindustrie zijn bioplastics en biosensoren interessante toepassingsgebieden. Hoofdstuk 2 en 5 lieten zien dat TfuDyP en SviDyP vanilline omzet in dimeren en oligomeren, als deze polymerisatie doorgaat dan worden plastics gevormd. Met biosensoren zouden (in combinatie met een oxidase) verschillende stoffen gedetecteerd kunnen worden, waaronder suiker, zoals beschreven is in hoofdstuk 5. In de toekomst kunnen we nog meer te weten komen over deze veelbelovende en veelzijdige enzymfamilie door het katalytische mechanisme verder te onderzoeken en de toepassingsmogelijkheden uit te breiden.

119 120 Appendices

List of publications

Colpa DI, Lončar N, Schmidt M, and Fraaije MW (2017) Creating oxidase- peroxidase fusion enzymes as a toolbox for cascade reactions. ChemBioChem 18, p. 2226–2230.

Colpa DI and Fraaije MW (2016) High overexpression of dye decolorizing peroxidase TfuDyP leads to the incorporation of heme precursor protoporphyrin IX. J. Mol. Catal. B. Enzym. 134, p. 372–377.

Lončar N*, Colpa DI* and Fraaije MW (2016) Exploring the biocatalytic potential of a DyP-type peroxidase by profiling the substrate acceptance of Thermobifida fusca DyP peroxidase. Tetrahedron 72, p. 7276–7281. de Gonzalo G, Colpa DI, Habib MHM, and Fraaije MW (2016) Bacterial enzymes involved in lignin degradation. J. Biotechnol. 236, p. 110-119.

Floor RJ, Wijma HJ, Colpa DI, Ramos-Silva A, Jekel PA, Szymański W, Feringa BL, Marrink SJ, and Janssen DB (2014) Computational library design for increasing haloalkane dehalogenase stability. ChemBioChem 15, p. 1660-1672.

Colpa DI, Fraaije MW and Bloois E (2014) DyP-type peroxidases: a promising and versatile class of enzymes. J. Ind. Microbiol. Biotechnol. 41, p. 1-7.

Winter RT, van den Berg TE, Colpa DI, van Bloois E, and Fraaije M.W. (2012) Functionalization of oxidases with peroxidase activity creates oxiperoxidases: a new breed of hydrid enzyme capable of cascade chemistry. ChemBioChem 13, p. 252-258.

* Shared first authorship.

123 Appendices

Acknowledgements/Dankwoord

This thesis was not accomplished without the help of other people. Even though I cannot mention everyone here, and if I try I will probably forget someone, I am very grateful for all the help and advice I have received during my PhD. I would like to thank the (former) members of the Molecular Enzymology and the Biotransformation & Biocatalysis groups for the open, friendly and stimulating environment and for all the great conversations about both scientific and non- scientific topics. I would like to thank the following people in particular.

Allereerst wil ik graag mijn promotor prof. dr. ir. Marco Fraaije bedanken. Aan het eind van mijn Master kreeg ik de gelegenheid om met een eigen onderzoeksvoorstel te solliciteren voor een PhD positie bij het GBB. Marco, bedankt dat je samen met mij deze uitdaging aan wilde gaan! Daarnaast wil ik je bedanken voor jouw begeleiding en advies. Ik waardeer het zeer dat jouw deur altijd open staat voor vragen en dat je mijn manuscripten bijzonder snel van commentaar en suggesties voorzag. Ook ben ik dankbaar dat je me bij hoofdstuk 4 langdurig in de gelegenheid hebt gesteld om uit te zoeken waarom de activiteit van het enzym zo laag was. Je had dit project ook kunnen stoppen. Ik ben blij dat we er uiteindelijk toch achter zijn gekomen wat er met het enzym gebeurd.

Prof. dr. Dick Janssen, naast jouw deelname aan de leescommissie wil ik je bedanken voor jouw bijdrage en de vele vragen tijdens mijn presentaties. Nogmaals bedankt voor het motiverende en inspirerende Masteronderzoek onder begeleiding van Hein en Robert. Ik stel het ook heel erg op prijs dat ik het afgelopen half jaar aan een paar projecten in jouw groep mocht werken. Verder kijk ik altijd erg uit naar de wetenschappelijke quizen tijdens de labuitjes!

Furthermore, I would like to thank the other two members of the reading committee: Prof. dr. G. Maglia and Prof. dr. L.O. Martins.

Daarnaast wil ik graag mijn paranimfen Christiaan en Nina bedanken. Christiaan, bedankt voor de vele goede gesprekken die varieerden van het voor de zoveelste keer misdragen van een autoclaaf of HPLC tot de Grijze Jagers, scouting of de meest hilarische vogel op aarde, de lyrebird. Nina, bedankt voor de gezelligheid en ik vond het erg fijn dat we konden overleggen over onze projecten en het begeleiden van onze studenten. Ik wil je vooral ook bedanken voor het plezier dat we samen hadden, bijvoorbeeld tijdens het maken van de filmpjes voor Gosia en Hugo. Hier wil ik de rest van de crew uiteraard ook voor bedanken! Hoewel

124 Acknowledgements/Dankwoord ik bijzonder goede herinneringen heb aan het maken van het filmpje voor Gosia, zal iedereen die het filmpje heeft gezien het begrijpen als ik zeg: Ik wil revanche! Wie gooide die emmer water ook alweer en mocht ik die dag niet nat maken?! Christiaan, Nina, Hilde? Hilde, nogmaals bedankt voor de uitnodiging en de heerlijke quiche!

It is not always easy to get a project started. Luckily, I could work on an enzyme which was discovered and partially characterized before I started my PhD.Edwin , thanks for your previous work on TfuDyP and for your advice in the beginning of my PhD. Approximately when Edwin left the group, Nikola arrived. Nikola, thanks for all your help during my PhD, for answering all my questions, for your motivating words and for your usually very easy and practical solutions. Thanks for staying in contact when you were in Belgrade. I am very happy with our collaboration and publications.

Mohamed, thanks for joining the DyP-team. It was a pleasure to work with you and to discuss our projects together. I am happy to hear that one of the new DyPs forms crystals. I wish you all the best with the remaining of your PhD! Please, keep me updated.

This work could not have been performed without the help of Master and Bachelor students Mareike, Thomas and Bastiaan. Thanks for your contribution in the projects, for finishing your reports on time and for the nice scientific and non-scientific conversations.Bastiaan , unfortunately we did not know about the problems with the heme cofactor during your project, which made the results very difficult to interpret. I hope you still had a great time in our lab. Thomas, you continued the work of Bastiaan on a more active version of the enzyme. It is great to see that you enjoy doing research and like to know exactly what is going on with the enzyme. Thanks for your very accurate work and for the conversations we had, which were not only about your project (part of chapter 3), but also about the troubles with the heme cofactor (chapter 4). I am very happy to hear that you decided to do a PhD and I wish you all the best with the last part of it! Mareike, it was nice to have you in the lab and to talk with you about our shared interests. On paper your project looked risky, so we decided to spread our chances and to make four fusion enzymes. Surprisingly, all four fusion enzymes were expressed and active. Thanks for your detailed Master thesis; it was very helpful when writing the paper. Mareike, Bastiaan and Thomas, I wish you all the best in life and your future careers.

125 Appendices

Piet en Christiaan, bedankt voor het draaiend houden van het lab, voor het bestellen en repareren van de apparaten, voor het waarborgen van de veiligheid en voor alles wat jullie verder regelen, inclusief de dingen die we soms over het hoofd zien en onterecht als vanzelfsprekend beschouwen. Piet, bedankt voor jouw technische ondersteuning. Het leek soms wel alsof de apparaten wisten wanneer je op vakantie ging en dan zelf ook ‘vakantie’ namen. Henriëtte en Lotteke, bedankt voor de pogingen om de DyPs te kristalliseren. Hein, bedankt voor de computationele hulp en Sandra, voor de secretariële ondersteuning. Johan en Theodora, bedankt voor de hulp met massaspectrometrie en prof. dr. Wesley Browne voor het gebruik maken van de faciliteiten en de hulp bij het bepalen van het redoxpotentiaal van de kleurstoffen.

Remko, Robert, Hein, Marysia en Josy, de begeleiders van mijn Bachelor- en Masterprojecten. Bedankt voor het overdragen van jullie kennis, praktische tips en vooral voor jullie enthousiasme voor het veld. Dit heeft er zekeraan bijgedragen dat ook ik voor enzymologie heb gekozen. Remko, alles wat je me tijdens mijn Bacheloronderzoek geleerd hebt over het transformeren van een oxidase in een oxiperoxidase heeft me erg geholpen bij het maken en karakteriseren van de oxidase-peroxidase fusie-enzymen in hoofdstuk 5.

All the current and former members of lab .123, office .106 and the rest of the two groups: thanks for all the good coffee breaks, lunches, Friday evening dinners (even though I did not join that often), the lab outings, ice skating and the relaxed times at the lake. Caterina, Elvira, Max, Sam, Suzan, Hugo, Marzena, Gosia and Yasser, thanks for the friendly (and quiet) environment in the office, which made it easy to concentrate. Hugo, thanks for: ‘what is the minimal amount of experiments necessary to get the answer?’. This approach kept me more focused on ‘why am I doing this experiment?’ instead of ‘blindly’ producing data. Thank you for checking the Dutch summary. Suzan, thanks for motivating me by showing interest in dye degradation. Marleen, je bent al een paar jaar verder, bedankt voor alle tips en adviezen. Antonija, thank you for all the fun and good talks. I have always appreciated your different perspective on topics and your honest opinions.

Naast een PhD project/werk is ontspanning erg belangrijk. Hiervoor wil ik mijn vrienden van binnen en buiten het lab bedanken, bedankt voor de afleiding en ontspanning in de weekenden en tijdens de pauzes. Ik heb goede herinneringen aan de tijd die we samen doorbrachten; aan de wandelingen, de (telefoon) gesprekken, het zwemmen en de spelletjes die we speelden. Sommigen van

126 Acknowledgements/Dankwoord jullie doen/deden zelf ook een PhD waardoor we veel herkenden in elkaars verhalen. Lotteke, Laura, Sonja, Maaike, Erik en Arjan bedankt voor de (bio) chemische gesprekken. Erik, bedankt voor het beantwoorden van al mijn organisch chemische vragen. En Alexander bedankt voor je hulp bij het maken van de cover.

Tot slot wil ik mijn ouders en zusje, Opa en Oma en tantes en ooms bedanken voor alle gezelligheid en steun. Fijn dat jullie altijd voor me klaar staan. Mam en Pap, de interesse voor de natuur, nieuwsgierigheid en doorzettingsvermogen heb ik van jullie. Bedankt voor alle mogelijkheden die jullie me gegeven hebben.

127