Isolation of bacterial strains able to metabolize lignin from screening of environmental samples

Authors


Correspondence

Timothy D.H. Bugg, Department of Chemistry, University of Warwick, Coventry CV4 7AL, UK. E-mail: t.d.bugg@warwick.ac.uk

Abstract

Aims

To develop a method to detect bacteria from environmental samples that are able to metabolize lignin.

Methods and Results

A previously developed UV–vis assay method for lignin degradation activity has been developed for use as a spray assay on agar plates. Nine mesophilic strains were isolated using this method from woodland soil incubated in enrichment cultures containing wheat straw lignocellulose: four Microbacterium isolates, two Micrococcus isolates, Rhodococcus erythropolis (all Actinobacteria) and two Ochrobactrum isolates (Alphaproteobacteria). Three thermotolerant isolates were isolated from the same screening method applied at 45°C to samples of composted wheat straw from solid-state fermentation: Thermobifida fusca and two isolates related to uncharacterized species of Rhizobiales and Sphingobacterium (Bacteroidetes), the latter strain showing tenfold higher lignin degradation activity than other isolates. The isolated strains were able to depolymerize samples of size-fractionated high molecular weight and low molecular weight Kraft lignin, and produced low molecular weight metabolites oxalic acid and protocatechuic acid from incubations containing wheat straw lignocellulose.

Conclusions

A new method for the isolation of bacteria able to metabolize lignin has been developed, which has been used to identify 12 bacterial isolates from environmental sources. The majority of isolates cluster into the Actinobacteria and the Alphaproteobacteria.

Significance and Impact of the Study

Lignin-degrading bacterial strains could be used to convert lignin-containing feedstocks into renewable chemicals and to identify new bacterial lignin-degrading enzymes.

Introduction

Lignin is a complex aromatic heteropolyme and found as a major component (20–35% dry weight) of plant lignocellulose (Lee 1997). Composed of highly cross-linked phenylpropanoid units, lignin is extremely resistant to microbial breakdown. However, there is considerable interest in lignin breakdown, because the removal of lignin is a key problem in the conversion of lignocellulose into second-generation biofuels and renewable chemicals (Lee 1997). The microbial degradation of lignin has been extensively studied in white-rot and brown-rot fungi, which use oxidative extracellular mechanisms to break down the lignin polymer (Wong 2009). The white-rot fungus Phanerochaete chrysosporium produces an extracellular lignin peroxidase (Tien and Kirk 1984) and manganese peroxidase enzymes (Glenn et al. 1986) to catalyse lignin depolymerization, and other fungi produce extracellular laccases (Kawai et al. 1988) that are involved in lignin breakdown. Bacteria have been much less well studied as degraders of lignin, but there are several reports of bacteria with this capacity (Zimmermann 1990; Bugg et al. 2011a,b), notably Streptomyces viridosporus (Ramachandra et al. 1988). If bacterial strains capable of efficient lignin breakdown could be found, then bacteria offer advantages over fungi for biotechnological applications, such as efficient large-scale growth, more convenient molecular genetics and protein expression, and the possibility of isolating thermophilic strains.

We have devised two novel spectrophotometric assays for lignin breakdown, using either fluorescently labelled lignin giving rise to a change in fluorescence upon lignin breakdown or chemically nitrated lignin (see Fig. 1) giving rise to a colour change at 430 nm owing to release of nitrated phenol breakdown products (Ahmad et al. 2010). Using this assay, in microtitre plate format, we identified a group of bacterial lignin degraders, including two known aromatic compound degraders Pseudomonas putida and Rhodococcus jostii RHA1 and compared their activity with that of white-rot and brown-rot fungi (Ahmad et al. 2010). We have also used a bioinformatic approach to identify a dyp-type peroxidase enzyme in R. jostii RHA1 as a bacterial lignin peroxidase (Ahmad et al. 2011).

Figure 1.

Chemical basis of nitrated lignin UV–vis assay (described further in Ahmad et al. 2010).

The aim of this study was to develop a colorimetric screening method, based on the nitrated lignin assay, which could be used on agar plates to detect bacteria capable of metabolizing lignin from soil or self-heating wheat straw compost. Here, we report the identification of a number of bacterial isolates using this method and evidence that they can metabolize different forms of lignin.

Materials and methods

Sources of cultures

Soil samples were collected from six woodland sites in Warwickshire and Hampshire, United Kingdom. Each soil sample (25 mg) was used to establish enrichment cultures (5 ml) in M9 salts (Na2HPO4, 12·8 g l−1; KH2PO4, 3·0 g l−1; NaCl, 0·5 g l−1; NH4Cl, 1·0 g l−1) containing milled wheat straw lignocellulose (2·0 g l−1) at 30°C for 21 days. Lignocellulose enrichment cultures were established from composted wheat straw (25 mg), obtained from a solid-state fermentation of wheat straw amended with 1% (w/w) chicken manure, operating at 60–65°C after 14 days of fermentation (Dr K. Burton, Warwick HRI).

Isolate identification

The relatedness of isolated strains to known bacteria was assessed by analysis of 16S rRNA gene sequences. Gene fragments of isolated cultures were amplified by colony PCR using forward (5′-AGAGTTTGATCMTGGCTCAG-3′) and reverse (5′-TACGGYTACCTTGTTACGACTT-3′) primers. PCR products (approximately 1·5 kb fragments) were sequenced and analysed using the Blast algorithm on the EBI web-server (www.ebi.ac.uk).

Transmission electron microscopy

Liquid cultures of Sphingobacterium sp. strain T2 were grown in Luria-Bertani broth at 45°C. Cultures were coated onto formvar–carbon grids and negatively stained with 2% uranium acetate. TEM images were obtained using a Jeol 2011 LaB6 microscope with a Gatan ultrascan 1000 camera.

Screening of environmental samples

Aliquots (200 μl) of mesophilic, lignocellulose enrichment cultures were diluted 100-fold and spread onto nutrient agar plates of the same media. Each plate was sprayed with a solution of nitrated lignin and incubated at 30°C for 2 days. The nitrated lignin solution was prepared as previously described (Ahmad et al. 2010), via treatment of milled wood lignin (MWL, prepared from wheat straw, 5 mg) in acetic acid (1 ml) with concentrated nitric acid (0·2 ml). After neutralization to pH 7, the solution was diluted 25-fold in sterile H2O for use in the spray assay and microtitre plate assay. Yellow colonies (indicative of nitrated lignin breakdown) were picked and streaked onto agar plates containing M9 salts, 0·04% (w/v) glucose and 0·5% (w/v) yeast extract and restreaked until pure cultures were obtained. Activity was confirmed by repeating the nitrated lignin spray assay. Screening for thermophilic degraders used the same procedure, but following growth on liquid and solid media at 45°C.

Nitrated lignin UV–vis assays were carried out using a similar procedure to that previously described (Ahmad et al. 2010). A stock solution of nitrated MWL from wheat straw was prepared as described previously. Assays (200 μl total volume) were carried out in 96-well Nunclon Surface clear plates, using a TECAN GENios plate reader. To each well was added 30 μl of bacterial supernatant, 160 μl of nitrated lignin, 10 μl of 2 mmol l−1 H2O2. Absorbance change at 430 nm was measured after a 20 min assay. The whole plate was repeated without addition of H2O2, where 10 μl of 750 mmol l−1 Tris pH 7·4 containing 50 mmol l−1 NaCl was added instead. Each assay was carried out in duplicate, with controls in which nitrated lignin or bacterial culture supernatant was replaced with 750 mmol l−1 Tris pH 7·4 containing 50 mmol l−1 NaCl. All readings were taken in duplicate.

Growth on selected carbon sources

Selected strains were streaked onto agar plates containing M9 salts and the following carbon sources, at the final concentrations indicated: veratryl alcohol (0·84 g l−1); m-cresol (0·22 g l−1) and p-cresol (0·22 g l−1); biphenyl (1·0 g l−1), ferulic acid (1·0 g l−1); vanillic acid (1·0 g l−1); and microgranular cellulose (0·5 g l−1). Cultures were grown at 30°C (mesophiles) or 45°C (thermophiles) for 7–10 days, and growth scored as good (103–104 colonies), moderate (102–103 colonies), weak (<50 colonies) or no growth.

Analysis of metabolites from lignocellulose breakdown

Wheat straw lignocellulose (2·5 g, milled to a coarse powder) was inoculated with a 2 ml culture of log-phase bacteria in 100 ml Luria-Bertani broth (10 g tryptone, 5 g yeast extract, 10 g NaCl per l). Samples (2 ml) were taken from cultures, which were shaken (180 rev·min−1) at 30°C or 45°C for up to 7 days. The samples were centrifuged (5 min, 10 000 rev min−1) and supernatants treated with 50 μl of cooled trichloroacetic acid (100% w/v) and left on ice for 3 min. Precipitates were removed by centrifugation again, and final supernatants were analysed by reverse phase HPLC on a Phenomenex HyperClone C18 reverse phase column. The gradient was 20–30% MeOH/H2O over 5 min; 30–50% MeOH/H2O over 5–12 min; 50–80% MeOH/H2O over 12–25 min at a flow rate of 0·5 ml min−1, detection at 310 nm. Samples (1 ml) were taken for GC/MS and extracted into dichloromethane (1 ml), then dried (MgSO4) and filtered using a 20-μm syringe filter (Whatman). The samples were derivatized with 200 μl of N,O-bis(trimethylsily)acetamide and 10 μl of chlorotrimethylsilane, left overnight and diluted by a factor of 100 before analysis by GC-MS on a Varian 3800 apparatus, using a Varian Factor Four column, 30 m × 0·25 mm × 0·25 μm, with electron ionization and ion trap. The temperature gradient was 75°C for 0–1 min and increased by 25°C min−1 to 300°C.

Data for oxalic acid (ethanedioic acid) GC-MS (disilylated) RT = 4·4 min, m/z 234 M+, 233 (M-H)+, 147 (M-SiMe3-CH2)+, 131 (M-SiMe3-2CH3)+, 132 (M-SiMe3-2CH2)+, 116 (M-SiMe3-3CH3)+. Data for protocatechuic acid (3,4-dihydroxybenzoic acid) GC-MS (trisilylated) RT = 12·2 min, m/z 371 M+, 370 (M-H)+, 355 (M-CH3)+, 281 (M-SiMe3-O)+, 193 (M-2SiMe3-O-CH3)+. Control incubations containing either no bacterial culture or no lignocellulose were also analysed and gave no observable small molecule products.

Size fractionation of Kraft lignin

A solution of Kraft alkali lignin (Sigma-Aldrich 471003, average Mw ~10 000; Gillingham, Dorset, UK) in water (20 mg/ml, 5 ml) was loaded onto an Ultrogel AcA44 gel filtration column (25 ml) and eluted with water, collecting 16 × 12 ml fractions. Fractions were analysed by UV–vis spectroscopy at 280 nm, and divided into three fractions, eluted sequentially from the column, corresponding to high-, medium- and low molecular weight Kraft lignin.

Incubation of Kraft lignin fractions with bacterial isolates

Solutions of high-, medium- and low molecular weight fractionated Kraft lignin were incubated at 30 or 45°C (for Sphingobacterium sp. T2) in 2 ml M9 minimal media (Na2HPO4: 12·8 g l−1, KH2PO4: 3·0 g l−1, NaCl: 0·5 g l−1, NH4Cl: 1·0 g l−1) with each bacterial strain. Aliquots of each sample (200 μl) were collected after 72 and 192 h, which were centrifuged (10 000 rev min−1, 5 min). The supernatant from each sample (200 μl) was analysed using size exclusion HPLC, using a BioSep SEC S-3000 column, eluting with water at a flow rate of 0·5 ml min−1, with UV detection at 280 nm.

Results

Development of colorimetric spray assay

A yellow colour developed when agar plate cultures of Pseudomonas fluorescens or R. jostii RHA1 were sprayed with nitrated lignin at 25 μg ml−1 concentration, tenfold higher than that used for microtitre plate assays, and incubated for 24–48 h at 30°C (see Supporting Information, Fig. S1). In contrast, there was no colour change with agar plates containing nondegraders Bacillus subtilis or Leuconostoc mesenteroides. This spray assay was suitable for selecting isolates from soil samples for further purification, because sprayed colonies remained viable for re-streaking.

Screening of soil samples

Aliquots of lignocellulose enrichment cultures (see Methods) were diluted and plated onto lignocellulose-containing agar. After 3 days incubation at 30°C, colonies were clearly visible on plates from 100-fold diluted samples. Each plate was sprayed with a solution of nitrated lignin and incubated at 30°C for 2 days. Yellow colonies were picked from each plate and restreaked onto minimal medium containing glucose, yeast extract and nutrient agar (see Materials and Methods). Each colony was restreaked several times to obtain a pure culture and checked again using the nitrated lignin spray assay (see Supporting Information, Fig. S2). This protocol resulted in 20 positive isolates, from 51 colonies originally picked.

Each isolate was grown in Luria-Bertani broth at 30°C, and culture supernatants were assayed in the nitrated lignin microtitre plate assay, in which active strains typically show a response of ≥1·0 mAU over a 20 min assay, and inactive strains show a zero or negative response, compared with controls (Ahmad et al. 2010). The results (see Fig. 2) indicated that most of the isolates showed higher activity in the assay than the previously studied Ps. putida and R. jostii RHA1 (Ahmad et al. 2010). Several isolates showed higher activity in the absence of added hydrogen peroxide, notably strains A4.3 and A5.2, in contrast to previously studied bacterial strains, which were more active in the presence of 2 mmol l−1 hydrogen peroxide (Ahmad et al. 2010).

Figure 2.

UV–vis assay data for mesophilic lignin-degrading strains shown in Table 1, compared with lignin degraders Rhodococcus jostii RHA1 and Pseudomonas putida previously studied (Ahmad et al. 2010), and nondegrader Bacillus subtilis. Assays were carried out using culture supernatant, as described in the Methods section, in the presence and absence of 2 mmol l−1 hydrogen peroxide. The data shown for each strain are averaged over all readings taken for duplicate strains isolated from screening. (□) Activity without hydrogen peroxide and (image) activity in the presence of hydrogen peroxide.

To validate the spray assay, colonies originally picked but found not to give a yellow colouration in the assay after restreaking were also grown in Luria-Bertani broth at 30°C and culture supernatants were assayed in the nitrated lignin microtitre plate assay. The spray positives segregated well from the spray negatives, with only three false positives (see Fig. 3), indicating that the spray assay had a fairly good degree of reliability as a primary screen for activity.

Figure 3.

Validation of screen for lignin degradation activity, via activity assay of ‘positives’ and ‘negatives’ from agar plate screen, using the nitrated lignin UV–vis assay (Ahmad et al. 2010). Activity in presence of 2 mmol l−1 hydrogen peroxide (y-axis) is plotted against activity in the absence of hydrogen peroxide (x-axis), and the ‘positives’ can be seen to segregate from the ‘negatives’. The activity of Rhodococcus jostii RHA1 and Pseudomonas putida mt-2 previously studied (Ahmad et al. 2010) is also shown. (△) Positive in spray assay; (●) negative in spray assay; (□) Rhodococcus RHAI and (×) Pseudomonas putida.

Strain identification of mesophilic isolates and growth on aromatic carbon sources

Nine different isolates were identified from 16S rRNA gene sequence analysis (see Table 1). Seven actinobacteria (four Microbacterium isolates, two Micrococcus isolates and Rhodococcus erythropolis) and two alphaproteobacteria (Ochrobactrum species) were found. Seven of the isolates were found in enrichment cultures from Warwickshire woodland soil (strains A1.1, A1.2, A2.1, A4.3, A5.1, A5.2 and B5.3), and two strains (C4.1 and E1.1) were isolated from enrichment culture from Hampshire heathland soil. High levels of sequence identity (>99·5% identity over 500–1400 nucleotides) were found to isolates in the Genbank database (see Table 1), which were obtained from various phyllospheres, rhizospheres or sediment and sludge samples. The majority shared high levels of 16S rRNA gene sequence identities to type strains of the species to which they most likely belong (Table 1). In addition, the mesophile strain A.4.3 shared over 98% sequence identity with the type strain of Ochrobactrum pseudogrignonense (GenBank accession no. NR-042589) and strains B.5.3 and E1.1 shared 99·9% sequence identity with the type strain of Micrococcus luteus (NR-037113).

Table 1. Identity and activity of bacterial isolates from screening
StrainGrowth temp (°C)Highest identity sequence match from 16S rRNA sequence (with GenBank accession number)Sequence identitya(%)Bacterial familyActivity in nitrated lignin assayGrowth on aromatic carbon sourcesb
+H2O2 (mAU)−H2O2 (mAU)Biphenylm-cresolp-cresolVanillic acidVeratryl alcohol
  1. a

    % identity over 500–1400 aligned nucleotides.

  2. b

    Growth on aromatic carbon sources: +++, good growth; ++, moderate growth; +, weak growth; −, no growth.

  3. c

    Type strain.

A1.130 Microbacterium phyllosphaerae c NR025405 100Actinobacteria1·92·7++++
A1.230 Microbacterium marinilacus c AB286020 99·9Actinobacteria1·70·6++++++
A2.130 Microbacterium marinilacus c AB286020 99·9Actinobacteria0·91·4+++
A4.330 Ochrobactrum pseudogrignonense GQ203110 100α-Proteobacteria1·24·7+++++
A5.130 Rhodococcus erythropolis c EU729738 99·8Nocardioform1·61·9+++++++++++
A5.230 Microbacterium oxydans c NR-04493199·7Actinobacteria1·15·4
B5.330 Micrococcus luteus JQ795852 100Actinobacteria1·91·1+++++++++
C4.130 Ochrobactrum rhizosphaerae c AM490632 100α-Proteobacteria2·01·9++++
E1.130 Micrococcus luteus GU085223 99·9Actinobacteria0·91·1++++++++++
T145 Rhizobiales bacterium AB563785 100α-Proteobacteria<0·22·0+++++
T245 Sphingobacterium AB563783 99·6Bacteroides9·030+++++++++

Their ability to degrade a selection of aromatic compounds (biphenyl, m-cresol, p-cresol, vanillic acid, and veratryl alcohol) was examined. Most of the strains showed some growth on one or more of the aromatic carbon sources (see Table 1). Good growth on vanillic acid was observed with two of the Microbacterium strains (A1.1, A1.2). The R. erythropolis strain A5.1 showed good growth on biphenyl and was able to grow on each of the carbon sources tested. Only a few of the isolated strains were able to grow on m-cresol or p-cresol as sole carbon source. The ability to grow on minimal M9 salts containing 0·5 g l−1 cellulose was also assessed. Moderate levels of growth were observed with strains A1.2, A2.1 (both Microbacterium marinilacus), B5.3 and E1.1 (both M. luteus), and R. erythropolis (strain A5.1); poor levels of growth were observed for A1.1 (Microbacterium phyllosphaerae) and A5.2 (Microbacterium oxydans).

Screening for thermophilic isolates

Samples from 45°C enrichment cultures from composted wheat straw were plated on the agar-solidified medium followed by incubation at 45°C for 3 days. The well-studied lignocellulose-degrading thermophilic actinobacterium Thermobifida fusca was prominent on the plates (with 100% rRNA gene sequence identity to the type strain) and in samples of the wheat straw compost examined by microscopy. Two less readily identifiable, nonsporing bacteria produced colonies with yellow colouration after use of the nitrated lignin spray. These were restreaked as before to obtain pure cultures and confirmed as positive isolates using the nitrated lignin spray, with one (strain T2) producing an intense colour change (see Supporting Information, Fig. S3). Analysis of 16S rRNA gene sequences (Table 1) indicated that the thermotolerant strain T1 has an identical sequence to that of unnamed species of Rhizobiales bacteria (Table 1) and was less related (no more than 96% sequence identity) to named species of Nitratireductor, Mesorhizobium and Phyllobacterium, for example, 96% sequence identity to the type strain of Phyllobacterium trifolii (NR-043193). The thermotolerant strain T2 also shared very high identities to unnamed species, in this case species of Sphingobacterium, but only 92% sequence identity to the most closely related type strain, that of Sphingobacterium composti (NR-041363).

Culture supernatants from thermophilic strains T1 and T2 were subjected to the nitrated lignin assay in a microtitre plate reader. Strain T2 showed approximately tenfold higher activity than the mesophilic strains isolated previously, with highest activity in the absence of hydrogen peroxide. Strain T1 showed detectable levels of activity in the absence of hydrogen peroxide. Strains T1 and T2 grew optimally at 45–50°C with growth inhibited at higher temperatures. Their ability to utilize aromatic carbon sources for growth was assessed on minimal medium (see Table 1). Strain T2 showed good growth with biphenyl or vanillic acid as sole carbon sources. Neither T1 nor T2 showed any growth on cellulose as sole carbon source. Analysis of cells of the Sphingobacterium isolate by transmission electron microscopy showed that the strain formed filaments containing small branches (see Fig. 4), unlike most sphingobacteria, which are generally rod-shaped bacteria.

Figure 4.

Transmission electron micrograph of Sphingobacterium sp. strain T2, grown on Luria-Bertani broth at 45°C. Scale bar (bottom left) 1 μmol l−1.

Biotransformation of lignocellulose to low molecular weight products

Microbacterium strain A1.1 and Sphingobacterium strain T2 were incubated with milled wheat straw lignocellulose at 30 and 45°C, respectively, and samples removed for analysis by LC-MS and GC-MS. A series of UV-absorbent lignin-derived peaks were observed by reverse phase HPLC, which were found to decrease in intensity over the 7 days of incubation, indicating lignin breakdown, whereas the peaks showed no change with nondegrader B. subtilis.

A new peak at time 4·4 min present in the sample taken from strain T2 at day 6 had a retention time and fragmentation pattern consistent with an authentic standard of oxalic acid (see Supporting Information, Fig. S4). Oxalic acid was observed as a major product in earlier treatments of lignocellulose by lignin degraders R. jostii RHA1 and Ps. putida (Ahmad et al. 2010). We have recently shown that R. jostii RHA1 DypB is able to cleave the Cα-Cβ bond of a β-aryl ether lignin model compound, thereby generating a two-carbon glycolaldehyde product (Ahmad et al. 2011); therefore, our hypothesis is that the observed oxalic acid product is formed via oxidation of glycolaldehyde generated via this route (see Fig. 5).

Figure 5.

Scheme showing a hypothesis for production of oxalic acid and protocatechuic acid from the β-aryl ether component of lignin.

A peak at retention time 12·2 min and m/z 371 present in the Microbacterium strain A1.1 sample at day 5 was further identified as protocatechuic acid by comparison with an authentic sample (see Supporting Information, Fig. S5). Protocatechuic acid is a well-known intermediate in aromatic degradation pathways (Bugg and Winfield 1998), which could also result from oxidative cleavage of β-aryl ether component of lignin (see Fig. 5).

Biotransformation of Kraft lignin

To examine whether the bacterial isolates were able to depolymerize high molecular weight vs low molecular weight forms of lignin, a sample of commercially available Kraft alkali lignin was fractionated by Ultrogel AcA44 gel filtration chromatography and divided into three fractions, high-, medium- and low molecular weight lignin based on their order of elution.

Incubations of five of the most active bacterial isolates (Microbacterium strains A1.1, A1.2, A5.2, R. erythropolis strain A5.1 and Sphingobacterium strain T2) with each of the fractionated Kraft lignins in minimal M9 media were then set up at 30 or 45°C, and aliquots were removed after 72 and 192 h. Growth was observed in all cases after 72 and 192 h, with somewhat faster growth on low molecular weight Kraft lignin, compared with high molecular weight and medium molecular weight Kraft lignin. Following centrifugation of bacterial cells, the supernatants were analysed by gel filtration HPLC, using a BioSep SEC S-3000 column. The observed data for Sphingobacterium sp. strain T2 are shown in Fig. 6, and the data for other strains are shown in Supporting Information (Fig. S6).

Figure 6.

Gel filtration HPLC analysis of the breakdown of size-fractionated high molecular weight (a) and low molecular weight (b) Kraft lignin by Sphingobacterium sp. strain T2, in M9 minimal media at 30°C. Experimental procedure is described in the Materials and Methods section. The retention time for the high molecular weight lignin peak at 11 min is indicated with a dashed line. New peaks 1 (25 min), 2 (17 min) and 3 (9 min) are discussed in the text. Data for other bacterial strains are given in Supporting Information.

For the high molecular weight Kraft lignin (see Fig. 6A), a peak at retention time 11 min corresponding to polymeric lignin was found to reduce in intensity vs time, to the extent of approximately 75% reduction after 72 h. In each case, after 3 and 7 days, a new peak (Peak 1) was observed at 24 min, corresponding to the formation of low molecular weight products, and in some cases, a second new peak at 17 min (Peak 2), corresponding to an intermediate molecular weight species. A third new peak at 9 min (Peak 3) was also formed after 72 h, indicating that a higher molecular weight species was also formed to some extent, suggesting that some repolymerization to form higher molecular weight material was taking place. In most cases, Peak 3 then decreased in intensity after 192 h. The most rapid increase in the 24 min peak was by the Sphingobacterium T2 strain, consistent with the high levels of activity observed in the nitrated lignin assay.

For the low molecular weight Kraft lignin (see Fig. 6B), a broader molecular weight envelope was observed at RT 11–20 min, consistent with a lower molecular weight distribution. Upon incubation with each bacterial strain, again a reduction in absorbance of this envelope was observed vs time, and a new peak at RT 24 min corresponding to low molecular weight products observed. For Microbacterium strains A1.1, A1.2 and A5.2 and Sphingobacterium strain T2, the residual lignin envelope after 192 h was greater for low molecular weight lignin than for high molecular weight lignin, suggesting that these strains show some preference for degradation of high molecular weight lignin; whereas in the case of R. erythropolis strain A5.1, the residual lignin envelopes were small in both cases.

Discussion

In this paper, we have reported a novel screening method for the detection and isolation of bacteria able to metabolize lignin from environmental samples. Previous screening methods for identification of lignin degraders have involved radiochemical assays using 14C-labelled lignin (6), a method not well suited to high-throughput analysis. The nitrated lignin assay in spray format selected for the small population of lignin degraders, with a low level of false positives, among the variety of organisms in enrichment cultures (including cellulose degraders). Some strains showed higher activity in the presence of hydrogen peroxide, which would be consistent with the involvement of extracellular peroxidase enzymes in lignin breakdown. However, the majority of these environmental strains showed higher activity in the absence of hydrogen peroxide, suggesting either that they utilize lignin-oxidizing enzymes that accept dioxygen as a substrate or that they are able to generate hydrogen peroxide in situ for use by peroxidase enzymes. Evidence that these bacterial strains are able to metabolize lignin is based upon: (i) the ability to grow on minimal media containing either wheat straw lignocellulose or size-fractionated Kraft lignin as sole carbon source; (ii) the depolymerization of high molecular weight Kraft lignin into low molecular weight products, observed by gel permeation chromatography; (iii) the detection of specific low molecular weight metabolites whose structures are consistent with breakdown pathways for lignin fragments; (iv) activity using three different lignin preparations, namely wheat straw lignocellulose (for growth and metabolite production), nitrated MWL (for screening and assay), and Kraft lignin (for depolymerization studies). Biotransformation of size-fractionated Kraft lignin by the isolated strains indicates that they are able to metabolize both high molecular weight and low molecular weight forms. An initial repolymerization to form higher molecular weight species (Peak 3, Fig. 6) was also observed, as has been observed using P. chrysosporium lignin peroxidase (Hammel et al. 1993).

The clustering of the mesophilic isolates, collected from two geographical areas, into the Actinobacteria and the Alphaproteobacteria is consistent with previous observations that the majority of lignin-degrading bacteria cluster into the Actinobacteria, Alphaproteobacteria and Gammaproteobacteria (Bugg et al. 2011a,b), suggesting that there are metabolic capabilities in these groups that support lignin breakdown. Bioinformatic analysis of bacterial pathways for β-aryl ether and biphenyl degradation, and the presence of homologues of R. jostii RHA1 DypB (Ahmad et al. 2011) suggests that genes for these pathways are found most commonly in these groups of bacteria (Bugg et al. 2011b). We have previously detected lignin degradation activity in several Rhodococci (Ahmad et al. 2010), and we have identified R. jostii RHA1 DypB as a lignin peroxidase enzyme (Ahmad et al. 2011), so it is interesting that one of the isolates is a R. erythropolis strain. Bioconversion of lignin model compounds has also recently been reported using Rhodococcus opacus strains (Kosa and Ragauskas 2012).

There are no previous reports of Microbacterium strains with lignin degradation ability, but Microbacterium and Ochrobactrum species were both found to be present in the gut of the wood-infesting termite Zootermopsis angusticollis (Wenzel et al. 2002). There are also reports of strains of Microbacterium that are able to degrade polychlorinated biphenyls (Rybkina et al. 2003) and polycyclic aromatic hydrocarbons (Zhang et al. 2004), consistent with the ability of the Microbacterium strains found here to grow on aromatic carbon sources and consistent with a possible link between lignin degradation and aromatic degradation. Biphenyl is a component of lignin, and the R. jostii RHA1 strain studied previously is a biphenyl degrader (Seto et al. 1995); therefore, it is interesting that some strains were able to grow on biphenyl as sole carbon source. Vanillic acid is a known intermediate on the catabolic pathway for the β-aryl ether component of lignin in Sphingomonas paucimobilis (Masai et al. 2007), and nearly all of the strains isolated were able to grow on vanillic acid as sole carbon source. There are also previous reports of strains of Ochrobactrum, found in this study, that are able to degrade polychlorinated biphenyls (Zermeno-Equia et al. 2009) and polycyclic aromatic hydrocarbons (Ghosal et al. 2010).

There are no previous reports of lignin-degrading sphingobacteria, but S. composti has been found among bacteria in paper mill pulps containing recycled fibres, as well as strains of Microbacterium barkeri (Suihko and Skytta 2009). Several strains of Sphingobacterium have been isolated from compost (Ten et al. 2006), but Sphingobacterium strain T2 is a novel species and not closely related to characterized species such as S. composti (Yoo et al. 2007). Sphingobacterium species have also been identified in a consortium able to degrade benzpyrene (Kanaly et al. 2000), and Sphingobacterium sp. ATM was reported to degrade a textile dye Direct Blue GLL, using an oxidative laccase enzyme activity (Tamboli et al. 2010). In the wider Bacteroides class, Spirosoma-like strains of Flexibacter have been detected in the gut of the termite Z. angusticollis (Wenzel et al. 2002), and anaerobic Bacteroides have been detected in wetwoods of living trees (Schink et al. 1981). Sphingobacterium T2 is unusual in growing aerobically at 45–50°C and forming filaments (see Fig. 4), as the majority of sphingobacteria are rod-shaped anaerobic organisms, however, there are reports of pleomorphism in the Bacteroides class, with filamentous forms reported for Bacteroides funduliformis (Smith et al. 1948) and Bacteroides forsythus (Sabet et al. 2003). Thermobifida fusca, which was observed on screening plates, is known as a major degrader of plant cell walls in self-heating compost (Lykidis et al. 2007).

In summary, the method reported herein will prove amenable for the identification of bacteria from environmental samples that are able to metabolize lignin, which may have application to the bioconversion of lignocellulose into biofuels and renewable chemicals. Studies to isolate and identify the extracellular enzymes responsible for lignin oxidation by these bacteria are in progress and will be reported in due course.

Acknowledgements

This work was supported by studentships from the Tuck Foundation Enerbio (C.R.T.), the IMRC ‘Wealth out of waste’ Project (M.A.), and BBSRC (P.S.) and a research grant from the BBSRC IBTI Refinery Club (grant BB/H004270/1). The authors would also like to thank Dr. Kerry Burton (Warwick HRI) and Dr Dan Eastwood (Warwick HRI) for samples of composted wheat straw, Ian Hands-Portman (Department of Life Sciences) for assistance with electron microscopy and Ludovic Laigle (Department of Life Sciences) for valuable technical assistance.

Ancillary