Bacterial Enzymes for Lignin Depolymerisation: New Biocatalysts for Generation of Renewable Chemicals from Biomass

The conversion of polymeric lignin from plant biomass into renewable chemicals is an important unsolved problem in the biorefinery concept. This article summarises recent developments in the discovery of bacterial enzymes for lignin degradation, our current understanding of their molecular mechanism of action, and their use to convert lignin or lignocellulose into aromatic chemicals. The review also discusses recent developments in screening of metagenomic libraries for new biocatalysts, and the use of protein engineering to enhance lignin degradation activity.


Introduction
Lignin is an aromatic heteropolymer comprising 15-30% of the lignocellulose cell wall of plant biomass, and is the most abundant source of renewable aromatic carbon in the biosphere. Given the need to reduce greenhouse gas emissions in 21 st century society, there is considerable academic and commercial interest in finding new sustainable biocatalytic routes to fuels and chemicals from renewable sources of carbon such as plant biomass [1]. For aromatic chemicals, lignin is an obvious starting point, but is a very challenging polymer to deconstruct, due to the presence of nonhydrolysable ether C-O and C-C bond linkages, poor solubility in aqueous solution, and other technical challenges [2].
The search for microbial enzymes to deconstruct lignin has until recently focussed on white-rot basidomycete fungi such as Phanerochaete chrysosporium, that produce extracellular lignin peroxidases, manganese peroxidases, and multi-copper laccases that can attack lignin (see Figure 1) [3]. However, these fungal enzymes are often challenging to express in high yield, and their fungal hosts are not readily amenable to genetic modification for metabolic engineering, hence since 2010 there has been a resurgence of interest in lignin-oxidising enzymes from soil bacteria. A number of soil bacteria have been identified that can depolymerise lignin, mainly in the actinobacteria and and -proteobacteria phyla, albeit less rapidly than the most active basidiomycete fungi [4,5]. This article will describe recent developments in the enzymology of bacterial lignin-degrading enzymes, our current understanding of how they attack lignin, and applications for biotransformation. The first bacterial lignin-oxidising enzyme to be identified was peroxidase DypB from Rhodococcus jostii RHA1, a member of the dye-decolorising peroxidases, found in bacteria and fungi [6]. Dyp-type peroxidases have activity for dye decolorisation, but also oxidation of a range of phenolic substrates [7]. There are four sub-classes A-3 D, based upon sequence alignment, of which classes A-C are found in bacteria, and class D found in fungi [7]. Although many Dyp-type peroxidases have been identified, only some have been demonstrated to have activity for oxidation of polymeric lignin. Lignin-oxidising Dyp peroxidase enzymes have been identified in Amycolatopsis sp. 75iv2 (DypC) [8], Pseudomonas fluorescens Pf-5 [9], and Thermobifida fusca [10]. Pfl Dyp1B is able to release an oxidised lignin dimer from wheat straw lignocellulose in the presence of Mn 2+ [9]. The ability to oxidise polymeric lignin in vitro correlates in most cases with an ability to oxidise Mn 2+ by some B-type and C-type Dyp peroxidases [6,8,9], although some A-type Dyps can oxidise lignin model compounds [10].
One of the best-studied hosts for bacterial lignin degradation is Pseudomonas putida, a well characterised aromatic degrader that has been verified via different experimental approaches to break down lignin [4,11], and has been used as a host for The second class of bacterial lignin-degrading enzymes are multi-copper oxidase enzymes, or laccases, which are best characterised from fungal sources, but have also been found in bacteria. Bacterial multi-copper oxidase from Streptomyces coelicolor A3 has been implicated in lignin breakdown, from gene deletion studies in which decreases in acid-precipitable lignin (APPL) were observed [18]. Overexpression of a 4 related SLAC enzyme in Amycolatopsis sp 75iv3 has been shown to lead to 6-fold increases in APPL production, with enhanced syringyl (S) content in the APPL structure, and the release of monocyclic aromatic products such as vanillin, 4hydroxybenzoic acid, and 1,4-dihydroxy-3,5-dimethoxybenzene (via aryl-C cleavage) [19]. A blue multi-copper oxidase CueO from a lignin-degrading Ochrobactrum sp. strain has been characterised kinetically, and its crystal structure unit containing an  ketone group. These enzymes were first identified in Sphingobium SYK-6, a bacterium with the ability to degrade a range of lignin dimers that are likely to be lignin oxidation products, where dehydrogenases LigD and LigL catalyse oxidation of the benzylic -hydroxyl group, LigE, LigF are stereospecific etherase enzymes for ether cleavage, and LigG catalyses reductive elimination of glutathione to generate a benzylic ketone product [26][27][28]. Crystal structures of Sphingobium SYK-6 LigD, LigO, LigL, LigG have been solved, providing insight into mechanisms of catalysis by these enzymes [29]. Analogous dehydrogenases LigO and LigN and -etherase LigP have been identified in Novosphingobium, and have been shown to catalyse stereospecific -ether cleavage [30]. A Nu-class glutathione S-transferase was identified in Novosphingobium as a -etherase enzyme [30], which has been shown to act as a glutathione lyase, which can cleaves both enantiomers of the intermediate glutathione adduct [31]. A -etherase enzyme has also recently been identified in the white-rot fungus Dichomitus squalens [32].
Screening microbes isolated from a range of environments could yield novel enzymes for the degradation of lignin, however it is accepted that most environmental microbes are unculturable, and the methods used to isolate and characterise those that can be grown are time consuming and work intensive. Hence there is current interest in the use of biosensors for high-throughput screening of metagenomic DNA libraries, as shown in Figure 2. A biosensor comprises of a genetic regulatory unit, which is activated by a specific compound, coupled to a reporter gene, which gives a measurable output such as fluorescence or luminescence. Several biosensors have been developed for a range of compounds associated with the breakdown of lignin, including: vanillin [33,34], protocatechuate [35] and phenylpropenoic acids such as ferulic and p-coumaric acid [36]. These regulatory units have been discovered by screening regulator-reporter gene fusion libraries [33] or using known genetic regulation towards a target compound [36]. Rounds of mutagenesis and selection can be used to tighten the regulation towards a specific compound or improve the range of expression in the 'on' and 'off' states [35].
Genomic DNA is typically screened in ~40 kb fosmids in E. coli, incubated with lignin or lignin-like substrates, then the reporter strain containing the biosensor is added, and the output measurement taken to assess degradation of the substrate.
Positive clones are then investigated to discover the genes responsible for activity. 6 The success rate of these strategies demonstrates their worth compared to traditional methods of screening. Using the vanillin and syringaldehyde sensor strain, Ho et al. screened 42,520 clones and had 147 positive clones [33]. Metagenomic analysis from lignin-treated sugarcane soil has also recently identified unculturable microbial sequences with lignin-degrading activity [37]. For the bacterial Dyp-type peroxidases, the detection of vanillin from a lignin dimer substrate using R. jostii DypB indicated that C-C oxidative cleavage (route B, Figure 3) had occurred, although dimerization via coupling of phenoxy radicals also occurred [6]. P. fluorescens Dyp1B had been shown to release products from C-C cleavage (route B) or aryl-C cleavage (route C, Figure 3) from polymeric lignin substrates [38], and can release a lignin dimer product containing an oxidised ketodiol sidechain from treatment of wheat straw lignocellulose [9]. For Sphingobacterium MnSOD, the major site of reaction appears to be demethylation (route E, Figure 3) [25], but monomers arising from aryl-C cleavage and C-C cleavage are also formed [24]. In model studies, oxidative cleavage by Dyp-type peroxidases has only been observed using units containing a free phenolic 4-hydroxyl group [6], therefore it seems likely that for breakdown of polymeric lignin 7 breakdown, they cleave from the ends of a lignin chain (exo-cleavage), rather than in the middle of a chain (endo-cleavage). Expression of R. jostii DypB in tobacco plants has been shown to yield 200% more fermentable sugars, and reduced lignin content, demonstrating that Dyp-type peroxidases can be expressed heterologously to depolymerise lignin [39].
The reaction of multi-copper oxidases (laccases) with lignin often results in repolymerisation via phenoxy radical formation (route D, Figure 3) [18-21]. However, in the presence of mediators such as 1-hydroxybenzotriazole (HBT) or methyl syringate, detailed NMR studies have shown that oxidation of the -hydroxyl group of the -aryl ether unit (route A, Figure 3) is a major reaction [40,41], however, studies of product release from lignin by fungal laccases has shown differences in behaviour depending on the mediator used [42]. Recent studies on Amycolatopsis

Lignin degradation accessory enzymes
One limitation of using recombinant lignin-oxidising enzymes for in vitro biotransformation of lignin substrates is that dimerization or repolymerisation is often observed, due to the formation of phenoxy radicals that spontaneously recombine.
Therefore it is likely that there are accessory enzymes in vivo that can trap phenoxy radicals via one-electron reduction. One candidate enzyme for this activity has been recently identified, a highly expressed extracellular dihydrolipoamide dehydrogenase from Thermobifida fusca, that has been shown to prevent dimerization of a lignin model compound in vitro, and change the profile of low molecular weight products formed [46]. Reductase enzymes such as this could be valuable accessory enzymes for lignin biotransformation.
Another accessory enzyme activity needed for lignin degradation in vivo is the generation of hydrogen peroxide co-substrate for lignin-oxidising peroxidase enzymes, probably generated from dioxygen by oxidase enzymes. Two such oxidase enzymes have been identified recently that appear to be linked to lignin degradation.
A copper-dependent oxidase enzyme has been identified in Thermobifida fusca, that has been shown to result in reduced lignin content in sugarcane bagasse, and generates dilignol products [47]. A new pathway for metabolism of aryl-C2 lignin fragments in Rhodococcus jostii RHA1 has also been shown to involve an FMNdependent oxidase enzyme that can oxidise aldehyde intermediates and simultaneously generate hydrogen peroxide [48].

Protein engineering studies
Two recent reports describe the application of directed evolution methods to bacterial Dyp-type peroxidase enzymes. Brissos et al report the engineering of Pseudomonas putida DyP using error-prone polymerase chain reaction, giving a mutant enzyme containing three mutations (E188K, A142V, H12V), each on the surface of the enzyme (see Figure 4), that enhance kcat/KM for 2,6-dimethoxyphenol by 100-fold, and shift the optimum pH to 8.5 [49]. Rahmanpour et al report the use of focused libraries around the active site of Pseudomonas fluorescens Dyp1B, enhancing the kcat/KM for 2,4-dichlorophenol by 7-8 fold, and mutation H169L was found to enhance product release from polymeric lignin [50]. This paper provides a more detailed mechanistic insight into the catalytic cycle of a bacterial dye-decolorizing peroxidase enzyme. This study demonstrates that bacterial multi-copper oxidase enzymes can be used to generate enhanced yields of acid-precipitable lignin and specific monocyclic aromatic products from lignin breakdown. This paper elucidates a novel mechanism of action of an unusual manganese superoxide dismutase enzyme that can generate hydroxyl radical to attack lignin, and explores the molecular basis for this unusual reactivity.