Ew

HKa

b

c

d

e

f

g

h

i

a

ARRA

KSCLTBB

I

pdtc

tU

0h

Systematic and Applied Microbiology 37 (2014) 60– 67

Contents lists available at ScienceDirect

Systematic and Applied Microbiology

j ourna l ho mepage: www.elsev ier .de /syapm

nzyme activities of aerobic lignocellulolytic bacteria isolated fromet tropical forest soils

annah L. Wooa,b,d, Terry C. Hazena,c,d,e,f,g, Blake A. Simmonsa,h,risten M. DeAngelisa,c,i,∗

Microbial Communities Group, Deconstruction Division, Joint BioEnergy Institute, United StatesPhysical Biosciences Division, Lawrence Berkeley National Laboratory, United StatesEarth Sciences Division, Ecology Department, Lawrence Berkeley National Laboratory, United StatesDepartment of Civil & Environmental Engineering, The University of Tennessee, United StatesDepartment of Microbiology, The University of Tennessee, United StatesDepartment of Earth & Planetary Sciences, The University of Tennessee, United StatesBiosciences Division, Oak Ridge National Laboratory, United StatesBiomass Science and Conversion Technology Department, Sandia National Laboratories, United StatesMicrobiology Department, University of Massachusetts Amherst, United States

r t i c l e i n f o

rticle history:eceived 28 May 2013eceived in revised form 9 October 2013ccepted 15 October 2013

eywords:oil lignocellulolytic bacteriaelluloseigninropical forestiomass degradationiofuels

a b s t r a c t

Lignocellulolytic bacteria have promised to be a fruitful source of new enzymes for next-generation lig-nocellulosic biofuel production. Puerto Rican tropical forest soils were targeted because the residentmicrobes decompose biomass quickly and to near-completion. Isolates were initially screened basedon growth on cellulose or lignin in minimal media. 75 Isolates were further tested for the follow-ing lignocellulolytic enzyme activities: phenol oxidase, peroxidase, �-d-glucosidase, cellobiohydrolase,�-xylopyranosidase, chitinase, CMCase, and xylanase. Cellulose-derived isolates possessed elevated �-d-glucosidase, CMCase, and cellobiohydrolase activity but depressed phenol oxidase and peroxidaseactivity, while the contrary was true of lignin isolates, suggesting that these bacteria are specializedto subsist on cellulose or lignin. Cellobiohydrolase and phenol oxidase activity rates could classify ligninand cellulose isolates with 61% accuracy, which demonstrates the utility of model degradation assays.Based on 16S rRNA gene sequencing, all isolates belonged to phyla dominant in the Puerto Rican soils,Proteobacteria, Firmicutes, and Actinobacteria, suggesting that many dominant taxa are capable of therapid lignocellulose degradation characteristic of these soils. The isolated genera Aquitalea, Bacillus,

Burkholderia, Cupriavidus, Gordonia, and Paenibacillus represent rarely or never before studied lignolyticor cellulolytic species and were undetected by metagenomic analysis of the soils. The study revealeda relationship between phylogeny and lignocellulose-degrading potential, supported by Kruskal–Wallisstatistics which showed that enzyme activities of cultivated phyla and genera were different enoughto be considered representatives of distinct populations. This can better inform future experiments and

.

enzyme discovery efforts

ntroduction

Bioenergy from biomass is the leading form of renewable energyroduction in the United States. However, the main hurdle in pro-uction is the plant cell wall’s recalcitrance to saccharification due

o its tightly interwoven structural components – cellulose, hemi-elluloses, and lignin. Therefore, these structures and bonds must

∗ Corresponding author at: Department of Microbiology, 203 Morrill Science Cen-er IVN, University of Massachusetts, 639 North Pleasant Street, Amherst, MA 01003,nited States.

E-mail address: kristen@post.harvard.edu (K.M. DeAngelis).

723-2020/$ – see front matter © 2013 Elsevier GmbH. All rights reserved.ttp://dx.doi.org/10.1016/j.syapm.2013.10.001

© 2013 Elsevier GmbH. All rights reserved.

be more efficiently broken to liberate more fermentable sugars thatare inexpensive enough for biofuel production.

Although lignocellulolytic fungi such as Aspergillus, Penicil-lium, Schizophyllum, Trichoderma, Phanerochaete, and Sclerotium[1] can secrete industrial quantities of extracellular enzymes,bacterial enzyme production can be more cost-efficient. Thisis because they grow more rapidly, produce multi-enzymecomplexes with increased functionality and higher specificity,and can tolerate larger and more diverse environmental stress[2–6]. Lignocellulolytic bacteria could also potentially allow bet-

ter separation of lignin from cellulose and thereby increasethe value of both lignin, which is currently a waste prod-uct, and cellulose. The few bacterial species currently knownto degrade cellulose and lignin are within Pseudomonas (order

Applie

PmAc

rabapIbabmti[fbs

ttfasslpittpd

las

M

I

Eiacae2maamcaw

i8mao

H.L. Woo et al. / Systematic and

seudomonadales), Cellulomonas (order Actinomycetales), Strepto-yces (order Actinomycetales), and other genera within the orderctinomycetales [7,8] and are likely employing extracellular lac-ases and peroxidases.

The isolation and characterization of environmental strains areelatively simple strategies in the age of metagenomics, but theyre still crucial for understanding the broad range of natural micro-ial functions. For example, the isolation of bacteria subsisting onntibiotics by Dantas et al. [9] was a major discovery that uprootedaradigms about microbial metabolism and antibiotic resistance.

n our own lab, physiological study of a tropical soil isolate Entero-acter lignolyticus SCF1 was shown to use lignin as an assimilatorynd dissimilatory carbon source, which would have been impossi-le to discern based on sequence analysis alone [10]. Even whenetagenomes of an environment are available, the characteriza-

ion of pure strains can provide useful information for physiologicalnferences since metagenomes are only an “incomplete list of parts”11,12]. Though laboratory cultivations have only isolated a smallraction of bacterial organisms [13], the range of cultivable soilacteria is expanding with simple improvements to cultivationtrategies [14,15].

The diversity and functionality of lignocellulolytic Puerto Ricanropical forest soil bacteria have not been as thoroughly inves-igated as termite hindgut and compost, despite that tropicalorests have the fastest rates of terrestrial litter decompositionnd a vast unknown microbial diversity [16–22]. Fast decompo-ition combined with frequent low and fluctuating redox potentialuggests the presence of potentially novel and efficient lignocel-ulolytic bacteria. Small subunit ribosomal RNA (rRNA) ampliconyrosequencing indicates the presence of many uncultivated phyla

n these soils [17]. Furthermore, the metagenomic data con-ains numerous uncharacterized glycoside hydrolases and glycosylransferases that cannot be assigned to specific families [17], sup-orting the possibility of a vast, as-yet undiscovered functionaliversity.

The goal of this broad survey was to collect novel lignocellu-olytic strains and determine any associations between phylogenynd function that will better inform larger scale cultivation orequencing efforts.

aterials and methods

solate cultivation

Isolates were cultivated from soils collected from the Luquilloxperimental Forest, part of the NSF-sponsored Long-Term Ecolog-cal Research Program in Puerto Rico. The samples were collectednd transported under USDA permit number P526P-08-00634. Soilollected from two sites, the Bisley watershed ridge rain forest sitend Short Could Forest site, were used as inoculum. The rain for-st site is in a lower montane wet tropical forest at approximately70 m above sea level (18◦18′ N, 65◦50′ W), and receives approxi-ately 3.5 m of rainfall annually. The cloud forest site is located in

upper montane tropical cloud forest at approximately 1050 mbove sea level (18◦18′ N, 65◦50′ W) and experiences approxi-ately 4–5 m rainfall annually, and a high frequency of low redox

onditions. Soil cores were transported to the lab in plastic bagst ambient temperature, diluted, and used as inoculum for growthithin one week of collection.

About one gram of soil was added to 10 mL of one of the two min-mal salts bases with 0.1% sodium pyrophosphate and 0.03% Tween

0. The minimal salts bases were modified VL55 [15] or basal saltsinimal medium (BMM) [23]; the details of the media composition

re described below. The mixture was homogenized by 2 roundsf vortexing for 1 min and sonicating for another minute. A serial

d Microbiology 37 (2014) 60– 67 61

dilution of the slurry was created in the minimal salts base. Aliquotsof 100 �L of dilutions from 10−1 to 10−10 were spread onto isola-tion media agar plates using a sterile glass spreader, parafilmedto maintain moisture, and incubated at room temperature, 30 ◦C,37 ◦C, and 55 ◦C.

Isolation media agar consisted of one of the two differentdefined media (VL55 or BMM) and one of the three carbonsources (Sigma–Aldrich microgranular cellulose, carboxymethylcellulose, or alkali lignin). Modified VL55 defined medium con-tained 0.10 mM MgSO4, 0.30 mM CaCl2, and 0.20 mM (NH4)2HPO4and 2.50 mL L−1 trace minerals pH 6.0 (1.50 g L−1 of Nitrilo-triacetic acid disodium salt, 3.0 g L−1 MgSO4·7H2O, 0.50 g L−1

MnSO4·H2O, 1.0 g L−1 NaCl, 0.10 g L−1 FeSO4·7H2O, 0.10 g L−1

CaCl2·2H2O, 0.10 g L−1 CoCl2·6H2O, 0.13 g L−1 ZnCl, 0.01 g L−1

CuSO4·5H2O, 0.01 g L−1 AlK(SO4)2·12H2O, 0.01 g L−1 Boric Acid).The VL55 was buffered at pH 5.5 using 3.90 g L−1 of the buffer 2-[N-morpholino]ethanesulfonic acid (MES). Carbon sources were addedat 0.1% or 0.05% final concentration. Immediately before pouringinto petri dishes, 10 mL of vitamins solution (2 mg L−1 d-Biotin,2 mg L−1 Folic acid, 10 mg L−1 Pyridoxine HCl, 5 mg L−1 Riboflavin,5 mg L−1 Thiamine, 5 mg L−1 Nicotinic acid, 5 mg L−1 Pantothenicacid, 0.1 mg L−1 of Vitamin B12, 5 mg L−1 of P-amino benzoic acid,and 5 mg L−1 of d,l-6,8-thiotic acid) was added. Some plates wereamended with 1 mL of 1% antibiotic streptomycin or 1% antifungalcycloheximide per liter in attempt to prevent the overgrowth ofbacteria or fungi. A second defined media, BMM defined medium,was also used. BMM salts media contained 0.80 g L−1 NaCl, 1.0 g L−1

NH4Cl, 0.10 g L−1 KCl, 0.10 g L−1 KH2PO4, 0.80 g L−1 MgCl2·6H2O,and 4.0 g L−1 CaCl2·2H2O. BMM media was also buffered with10 g L−1 MES at pH 6, in addition to the trace minerals, vitaminsand one cellulose or lignin substrate at the previously listed con-centrations. Plates were poured with media and 15 g L−1 agar.Isolates were re-streaked to isolation before further testing. Iso-lated colonies were subcultured onto 10% Tryptic Soy Broth (TSB)agar to maximize biomass growth in preparation for enzymaticassays. Frozen stocks were prepared without any serial transfersof cells.

Enzyme assays

Enzyme activities were determined by measuringthe degradation of four different 4-methylumbelliferone(MUB) linked carbohydrate substrates and two phe-nolic L-3,4-dihydroxyphenylalanine (l-DOPA) solutions:MUB-�-d-glucopyranoside, MUB-�-d-cellobioside, MUB-�-d-xyloside, MUB-N-acetyl-�-d-glucosaminide dihydrate,l-3,4-dihydroxyphenylalanine (l-DOPA), l-DOPA with 0.3% H2O2.To prepare the cells for the enzymatic assays, cell cultures ofisolates were grown overnight in 10% TSB at 30 ◦C, shaking at200 rpm. The cultures were then diluted using 1× PhosphateBuffered Saline (PBS) to a normalized concentration of 0.2 OpticalDensity (OD) at 600 nm; cells were sometimes used at 0.10 OD ifthere was insufficient biomass to achieve 0.20 OD.

Cells were mixed at 1:1 volumetric ratio with each of the0.10 mM MUB-substrates or 10 mM l-DOPA in technical replicatesof three or more. Negative controls consisted of the substrate aloneand cells alone. Cells were allowed to digest substrates for 2–9.5 hduring which fluorescence or absorbance readings were taken. Thequantity of released MUB was calculated using a standard curveof 0–50 �M MUB in 50 mM sodium acetate buffer (pH 5.5). Ratesof MUB-substrate degradation are expressed as �moles of MUB-substrate degraded per OD600 nm cells loaded per hour. Rates of

l-DOPA degradation are expressed as l-DOPA absorbance at 460 nmper OD600 nm cells loaded per hour.

CMCase and xylanase activity were tested qualitatively bya Congo red based plate assay. Aliquots of 5 �L from 48-h-old

6 pplied Microbiology 37 (2014) 60– 67

loibada

I

s2Gcaop9f

otBPleiatoa(360ce(Dti

PfiB(btsBotwtfSngI

S

ec

Table 1Enzyme activities summary.

Activity rates Unit Mean Standarddeviation

BG (umoles substrate/(O.D. cells × hour) 2.18 2.09CBH (umoles substrate/(O.D. cells × hour) 1.35 0.81NAG (umoles substrate/(O.D. cells × hour) 1.48 2.19XYL (umoles substrate/(O.D. cells × hour) 0.88 1.40PO (O.D. LDOPA/(O.D. cells × hour) 0.16 0.10HPO (O.D. LDOPA/(O.D. cells × hour) 0.21 0.26

Frequency(number of isolates)

2 H.L. Woo et al. / Systematic and A

iquid culture were spotted onto 10% TSB agar with either 0.5% CMCr xylan. Cells were incubated at 30 ◦C for 7–10 days before stain-ng with 0.1% Congo Red solution as described in Teather et al. [24]ut extended for an hour and followed by 2 washes in 1 M NaCl forn hour each. Lastly, plates were stained with 2% HCl for 5 min toevelop better contrast. Any indication of clearing was considered

positive result.

solate identification and characterization

Isolates were genotyped using the 16S rRNA geneequence. The 16S rRNA gene was PCR amplified using7F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GTTACCTTGTTACGACTT-3′) primers [25]. Each 50 �L PCR reactionontained 1× Takara ExTaq Buffer with MgCl2, 300 pM of both 27Fnd 1492R primer, 1 mg mL−1 of Bovine Serum Albumin, 200uMf each dNTP, and 2.5 U of Takara ExTaq. The PCR amplificationrotocol was a 95 ◦C denaturation step for 3 min, then 25 cycles of5 ◦C for 30 s, 50 ◦C for 30 s and 72 ◦C for 1 min, followed by a 72 ◦Cor 8 min to allow final extension.

A small amount of cell biomass from a single colony growingn 10% TSB agar plates was added to 50 �L of the PCR reaction asemplate (“colony PCR”). When colony PCR failed, Qiagen DNeasylood and Tissue kit was used to extract DNA from isolates, thenCR was performed using 10–50 ng of DNA template. Several iso-ates required 5 rounds of snap freezing and thawing followed byxtracting using a CTAB extraction method. The CTAB/NaCl buffers 41 g NaCl and 100 g CTAB per liter of solution. Cell biomass fromgar plates was resuspended in 740 �L of TE and the concentra-ion adjusted such that the OD600 nm equaled 1. At this point 20 �Lf lysozyme (100 mg mL−1) was added and incubated for 5 mint room temperature, then 40 �L 10% SDS and 8 �L proteinase K10 mg mL−1) were added and incubated for an additional 1 h at7 ◦C. Then, 100 �L 5 M NaCl and 100 �L CTAB/NaCl (pre-heated to5 ◦C) were added, then incubated at 65 ◦C for 10 min. After that,.5 mL of chloroform:isoamyl alcohol (24:1) was added mixed, andentrifuged for 10 min. The aqueous phase was transferred andxtracted again with 0.5 mL of phenol:chlorofom:isoamyl alcohol25:24:1), mixed and centrifuged, and the organic phase discarded.NA in the aqueous phase was precipitated by isopropanol precipi-

ation with one wash step, then resuspended in TE plus RNAse, andncubated at 37 ◦C for 20 min.

All PCR products were purified using the Qiagen Qiaquick PCRurification kit following the manufacturer’s instructions. Puri-ed PCR product was sent to either the University of Californiaerkeley Sequencing Facility (Berkeley, CA) or Quintara BiosciencesEmeryville, CA) for sequencing. Nucleotide sequences producedy the forward and reverse primers were concatenated into a con-iguous sequence read, NAST aligned against Greengenes referencetrains and checked for PCR chimeras using the Greengenes toolellerophon [26,27]. Strain identification at the genus level is basedn the Greengenes batch classification of sequences that provideshe best matching taxa from multiple taxonomies [28]. Strainsith less than 97% similarity were subsequently identified using

he EzTaxon-e server [29]. The 16S rRNA gene sequences can beound on Genbank, with accession numbers JQ917930–JQ918000.equences were clustered in 14 subgroups based on a nearesteighbor algorithm, then one representative isolate from each sub-roup was BLAST against Puerto Rican soil metagenomes within JGIMG [30]’s blastn module.

tatistical analyses

Activity rates from MUB-substrate degradation assays for allnzyme activities were normalized between 0 and 1 for prin-ipal component analysis and ordination using the package

CMCase Positive for activity 36Xylanase Positive for activity 33

vegan [31,32]. The neighbor joining tree was made by insertingGreengenes-aligned and chimera-free 16S rRNA gene FASTA filesinto program ARB using parsimony insertion [33]. The Greengenestree was used for genetic distances [28]. Four isolates that had 16SrRNA gene sequences with short reads or poorly aligned (<75%)with Greengenes reference strains were excluded from the tree –cmc19, cmc20, lig16, lig38.

All other statistical analyses were performed in JMP10 Pro. Con-tingency table analysis was used to compare nominal variables.Enzyme activities from MUB and l-DOPA tests were log trans-formed to more closely achieve normal distributions of valuesfor logistic regression analysis. Enzyme activities below detectablelimit assumed values of 0. Logistic regression’s response dichoto-mous variable set cultivation on cellulose as 1, and lignin growthas 0. Receiver operating characteristics (ROC) analysis was usedto evaluate the predictive value of the regression. Untransformeddata was used for all nonparametric analyses. Wilcoxon tests(non-parametric) were used to compare 2 groups’ distributions.Kruskal–Wallis tests (non-parametric) compared more than 2populations at once. Steel-Dwass (non-parametric) was used tocompare distributions pairwise following Kruskal–Wallis analysis.Statistical results were interpreted using techniques outlined byPeng et al. [34].

Results

Seventy-five bacterial strains were isolated from the PuertoRican Luquillo experimental forest soils based on their abilityto grow on a recalcitrant carbon source (lignin or cellulose) astheir sole carbon source. Each isolate was cultivated on car-boxymethyl cellulose (cmc), alkali lignin (lig), or microgranularcellulose (mgc) in minimal media agar. Isolates were picked fromplates with an inoculum dilution factor ranging from 10−1 to 10−10

with most colonies appearing on plates streaked from the 10−5

dilution (Table S3): 5 colonies of Gordonia, 4 of Bacillus, 12 ofBurkholderia were isolated at dilutions of 10−5 from at least 3 dif-ferent plates; 4 colonies of Aquitalea were isolated at dilutionsof 10−6. The high dilutions suggest that the most probable num-ber (MPNs) for Gordonia, Bacillus, Burkholderia and Aquitalea rangefrom 1 × 105 to 1 × 106 cells per milliliter [35]. To ensure isola-tion, single colonies were picked and re-streaked at least two times,checked for homogeneous morphology, and singular 16S ribosomalRNA sequence was determined by PCR. Isolated strains were testedfor phenol oxidase, peroxidases, �-d-glucosidase, cellobiohydro-lase, �-xylopyranosidases, chitinase, CMCase, and xylanase activity(Tables 1 and S1). Cellulose-derived isolates possessed elevated

�-d-glucosidase, CMCase, and cellobiohydrolase activity. Isolatedstrains were identified by assigning nearest neighbor based onBLAST results of 16S rRNA gene sequences. Forty of the 75 strainswere from lig minimal media agar, 31 strains on cmc, and 4 on

H.L. Woo et al. / Systematic and Applied Microbiology 37 (2014) 60– 67 63

Fig. 1. Neighbor-joining tree of all isolates with enzymatic activity rates. Organism labels in the phylogenetic tree are RDP taxonomy at the genus level. Ovals on the branchesrepresent branches with a bootstrap value above 80%. Isolates have red and/or blue dots for positive xylan and CMC degradation. Each isolate has 6 horizontally stacked bargraphs of the raw enzymatic activity rates: �-d-glucosidase (BG), cellobiohydrolase (CBH), N-acetyl-�-d-glucosaminidase (chitinase) (NAG), �-d-xylopyranosidases (XYL)i -dihydi terprv

mciaai

lA2sPMV

n �moles of substrate degraded per OD600 nm cells loaded per hour. Phenolic l-3,4n oxidized l-DOPA absorbance at 460 nm per OD600 nm cells loaded per hour. (For inersion of the article.)

gc (Table 2). On average, colonies appeared within 2 weeks. Toultivate strains of different temperature optimums, replicates ofnoculated agar plates were incubated at temperatures betweenmbient and 55 ◦C (Table 2). Only 10 of the 75 strains, all Gordoniand Paenibacillus species, were successfully cultured from platesncubated at or above 37 ◦C.

The 16S rRNA gene sequencing revealed that the 75 iso-ates were from three phyla: Proteobacteria (65% of isolates),ctinobacteria (21%), and Firmicutes (13%) with 14 genera and

potentially novel species represented (Fig. 1). More than one

train of Aquitalea, Bacillus, Burkholderia, Cupriavidus, Gordonia, andaenibacillus were isolated. Other isolates were Mycobacterium,ethylobacterium, Novosphingobium, Pseudomonas, Roseateles, and

ariovorax. The two potentially novel strains were most closely

roxyphenylalanine (l-DOPA) with hydrogen peroxide (HPO) and without (PO) areetation of the references to color in figure legend, the reader is referred to the web

related to Betaproteobacteria and the family Neisseriaceae: one ofthe genus Gulbenkiania (lig3) and the other Pseudogulbenkiania(lig4). Both 16S rRNA gene sequences were only 94–95% simi-lar to strains in the RDP database; similarly low matches werefound within the Greengenes and NCBI databases (Table S2). Bothhave 93–94% similarity to the nearest bacteria with a full genomesequence, Chromobacterium violacearum ATCC 12472. However, afairly new database, EzTaxon-e, identified both strains as Vogesellaperlucida DS-28 (family Neisseriaceae) with 97% similarity. Becauseof the discrepancy between these classification databases, addi-

tional genetic information would be needed to confirm the noveltyof these strains [36,37].

Seventy percent of the cellulose isolates demonstrated CMCaseand xylanase activity by forming zones of clearing on CMC or xylan

64 H.L. Woo et al. / Systematic and Applied Microbiology 37 (2014) 60– 67

Table 2Descriptive statistics of Puerto Rican isolate collection.

Isolates N = 75

Frequency Percentage

Isolation carbon sourceAlkali lignin 40 53Carboxymethyl cellulose 31 41Microgranular cellulose 4 5

Genus (Phyla)Aquitalea (Proteobacteria) 11 15Bacillus (Firmicutes) 6 8Burkholderia (Proteobacteria) 23 31Cupriavidus (Proteobacteria) 6 8Gordonia (Actinobacteria) 15 20Paenibacillus (Firmicutes) 3 4Other 11 15

Isolation temperature30◦C 12 1637◦C 9 12

◦

ptlssc(a

hrdeeoushlamOrnclA�opclpBads

acl(ip(

Fig. 2. Ordination of 6 enzyme activities on MUB-linked carbohydrates and l-DOPAphenolics. Results were plotted and clusters were grouped based on similar activityprofile. Points without isolate labels had relatively lower activity and do not clusterwith any activity profiles. Labeled points possess one of the following activity pro-

(P < 0.01) and peroxidase (P < 0.01). On the contrary, Gordonia

55 C 1 1Ambient 53 71

lates stained with Congo red. Paenibacillus sp. cmc12 producedhe most notable zones of clearing on both substrates that were ateast twice the diameter of biomass growth, suggestive of aggres-ively dispersing enzymes. As expected, contingency table analysishowed that strains initially isolated on cellulose had a statisti-ally higher likelihood of having positive results of CMCase activityOdds ratio: 3.14, 95% Confidence interval: 1.22–8.08) and xylanasectivity (Odds ratio: 2.76, 95% Confidence interval: 1.08–7.09).

Principal component analysis based on the level of carbo-ydrate or phenolic compound degradation for each isolateevealed four functional groups. This analysis was based onegradation of four carbohydrates and two phenolics at differ-nt levels; these compounds are frequently used to study soilnzymes [38,39]. The lignin-degrading enzyme activities phenolxidase (EC 1.10.3.2) and peroxidase (EC 1.11.1.7) were testedsing colorimetric assay based on degradation of the model sub-trate l-dihydroxyphenylalanine (l-DOPA) with and without addedydrogen peroxide, respectively. �-d-glucosidase (EC 3.2.1.21), cel-

obiohydrolase (EC 3.2.1.91), �-xylopyranosidases (EC 3.2.1.37),nd chitinase (EC 3.2.1.30) were tested using their respectiveonomers linked to fluorogenic 4-methylumbelliferone (MUB).rdination of these activities using principal component analysis

evealed 4 major groupings among the isolates (Fig. 2) which doot include the 32 isolates with the lowest total enzyme activityalculated by the sum of their 6 enzymatic rates. The other 43 iso-ates separated into clusters that represent four different profiles:, B, C, or D. Most cellulose isolates possessed profile A (elevated-d-glucosidase and cellobiohydrolase). Strains initially isolatedn lignin typically had profile B (elevated cellobiohydrolase andhenol oxidase), C (elevated phenolics degradation), or D (elevatedellobiohydrolase and �-d-xylopyranosidases). Profile B and D iso-ates could not be differentiated because D profile isolates alsoossess a small amount of phenol oxidase activity. However, Profile

and D isolates are very different in terms of �-d-xylopyranosidasectivity. Lig8 and lig17 were included with profile D since they allemonstrated elevated phenol degradation activities in compari-on to the rest of the isolates.

Reducing the enzymatic activity data to just cellobiohydrolasend phenol oxidase in a logistic regression model can correctlylassify over half of the collection whether they were initially iso-ated on cellulose or lignin isolates. Addition of cellobiohydrolaseCBH) and phenol oxidase (PO) enzymatic activity data significantly

mproved the fit of the model to the experimental data when com-ared to a null model with all regression parameters equal to 0�2(2) = 12.65, P = 0.0018); each parameter was also statistically

files: profile A (elevated �-d-glucosidase and cellobiohydrolase), profile B (elevatedcellobiohydrolase and phenol oxidase), profile C (elevated phenolics degradation)or profile D (elevated cellobiohydrolase and �-d-xylopyranosidases).

significant (P < 0.05). The prediction formula for the probability ofcolony formation on cellulose given values of CBH and PO is asfollows:

Probability (isolate growing on cellulose minimal media agar |Log (CBH), Log (PO)) = −3.37 + 1.04 × Log (CBH) + -1.57 × Log (PO).

The odds ratios indicate a reverse relationship between CBH andPO activity when predicting cellulose and lignin isolates. For everyunit increase in CBH activity, the chance that the isolate was cul-tivated on cellulose has increased 2.8-fold (Odds ratio: 2.78, 95%Confidence interval: 1.24–7.32). Meanwhile, for every unit increasein PO value, the chance that the isolate grew on lignin insteadhas increased 4.8-fold (Odds ratio: 0.20, 95% Confidence interval:0.06–0.61). Interaction between PO and CBH was not significant(P > 0.05). Out of 75 isolates, this prediction equation using a proba-bility cutoff of 0.50 correctly predicted 19 of the 35 cellulose isolatesand 27 of the 40 lignin isolates. Isolates of profile B were the mostdifficult to accurately predict with this model due to their elevatedCBH and PO activities. The predictive value of the logistic regressionformula was supported by a sensitivity and specificity plot, com-monly known as a Receiver operating characteristics (ROC) (Fig.S2). The Area under curve (AUC) value of 0.71 was higher thanthe expected AUC value of 0.5 from random guessing [40]. Thusthe logistic regression model based only on CBH and PO performedsignificantly better than pure chance at predicting initial isolationconditions.

Aquitalea and Gordonia isolates were the most active cel-lulose and lignin degraders, respectively. When comparing theenzyme activity distribution of Aquitalea strains against allother isolates using Wilcoxon rank sums, Aquitalea strains hadsignificantly higher distributions of �-d-glucosidase (P < 0.01), cel-lobiohydrolase (P < 0.01), and �-d-xylopyranosidases (P < 0.01),but meanwhile possessing depressed levels of phenol oxidase

strains possessed higher phenol oxidase (P < 0.001) and peroxi-dase (P < 0.0001) activities but depressed levels of �-d-glucosidase(P < 0.05).

Applie

eFicq[mpt(cBhvtg

vsTmts9tpahTtltrtmat

D

RabbpeRoot

BuSlqtB1tdlm

H.L. Woo et al. / Systematic and

There is indication of a relationship between phylogeny andnzyme function. The enzyme activities of Actinobacteria andirmicutes isolates differed significantly in lignin-associated activ-ty phenol oxidase (P < 0.001) and peroxidase (P < 0.001) whenompared using Kruskal–Wallis rank sums, the most robust and fre-uently used non-parametric method to compare multiple groups41]. The rejection of the most conservative null hypothesis of this

ethod suggests that the sample groups originate from differentopulations or simply stochastic heterogeneity [42,43]. Actinobac-eria and Proteobacteria isolates differed in �-d-glucosidase activityP < 0.05) in addition to the phenolics degradation rates. Whenomparing the 6 dominant taxa (Gordonia, Bacillus, Paenibacillus,urkholderia, Cupriavidus, and Aquitalea), �-d-glucosidase, cellobio-ydrolase, �-d-xylopyranosidases, phenol oxidase, and peroxidasealues were all different (P < 0.05). The small P-values indicate thathe probability of observing such differences between phylogeneticroupings is very unlikely to happen by pure chance.

A majority of the isolates could not be detected in several pre-iously reported metagenomes generated from Puerto Rican forestoil, based on a high level of similarity of 16S rRNA gene sequence.he BLAST alignment lengths between isolates’ 16S rRNA genes andetagenomic scaffolds within Puerto Rican Luquillo forest soil and

wo soil microbial consortia enriched on switchgrass as a carbonource [44] were not long enough to conclude a confident match.2% of isolates matched metagenomic scaffolds at E-values lowerhan 10−5, with percent identities higher than 90%, but all matchesroduced alignment shorter than 200 bp long (Fig. S1). The shortlignment lengths led to the same set of 37 scaffolds being BLASTits for all isolates, even when the strains were from different phyla.his does not seem plausible when scaffolds are too short to con-ain multiple 16S rRNA genes. A BLAST search of the two relativelyonger scaffolds against the NCBI database yielded uncultured bac-erium clones from leaf-litter degrading studies, subtropical forest,hizosphere soil and crop residue digestion studies at 99% iden-ity. The overall low similarity between the isolates’ sequences and

etagenomics data is expected due to the large bacterial diversitynd limitations with sequencing depth, all of which demonstrateshe importance of bacterial isolations in novel strain discovery.

iscussion

A targeted cultivation strategy was successfully used on Puertoican rain forest soil to broaden the known diversity and function-lity of bacteria in lignocellulose degradation that would not haveeen found with metagenomics alone. Puerto Rican soil has noteen sufficiently sequenced to determine all the species of bacteriaresent. Less than 80 16S rRNA gene sequences were annotated inach of the three metagenomes [44]. Pyrotags SSU rRNA of Puertoican soil microbial communities trapped on lignin beads couldnly provide resolution at the phylum and genus level [45]. With-ut sufficient sequencing to identify dominants, isolation may behe easiest method to discover strain-level diversity.

The multiple strains of Gordonia, Aquitalea, Bacillus, andurkholderia could be dominant taxa in Puerto Rican soil or contrib-tors to rapid plant litter degradation. This agrees with previousSU rRNA pyrosequencing and Phylochip work where the preva-ence of phyla Proteobacteria, Firmicutes, and Actinobacteria mostuickly increased in response to lignin enrichments [45]. In addi-ion, the most probable number of the present study’s Gordonia,acillus, Burkholderia and Aquitalea strains being at least 105 and06 organisms per milliliter suggest a large contribution of these

o the overall soil population. However, it is still possible thatominant strains eluded cultivation. The most well characterized

ignin degrading bacteria like Sphingomonas paucimobilis, Strepto-yces viridosporus, and Rhodococcus were not found [46,47,48].

d Microbiology 37 (2014) 60– 67 65

Acidobacteria and Chloroflexi constituted a significant amount ofsome clone libraries or pyrotags from tropical soil but were not cul-tivated either [49,45]. Adjusting the media composition and lengthof incubation may reveal additional diversity of isolates.

One of the few existing culture collections from tropical for-est soils also captured similar diversity on lignocellulose fromthe genera Mycobacterium (genus of Actinobacteria), Bacillus andPaenibacillus (genera of Firmicutes), Burkholderia (genus of Betapro-teobacteria), Variovorax, Cupriavidus, and Pseudomonas (genera ofGammaproteobacteria) [49]. However, no prior reports exist onAquitalea, Gulbenkiania, and Pseudogulbenkiania as lignocellulosedegraders. The latter two strains were potential novel species basedon their 16S rRNA sequences, but were also functionally differentfrom type strains of the genera, which were either negative for �-d-glucosidase activities or failed to grow on cellobiose in previouslypublished characterizations [50–53].

The most active isolates were the cellulose-degrading Aquitaleaand the lignin-degrading Gordonia. The five isolates Aquitalea spp.cmc2, cmc3, cmc4, cmc7, and cmc9, (family Neisseriaceae) demon-strated the ability to produce all three glycoside hydrolase (GH)activities involved in complete enzymatic hydrolysis of cellulose.These general classes of enzymes are �-1,4-endoglucanase (EC3.2.1.4) which cleave internal 1-4-glycosidic linkages and are activeagainst CMC; cellobiohydrolase (EC 3.2.1.91) which is an exocellu-lase that cleaves cellobiose from the non-reducing end and is mostactive on crystalline cellulose; and �-d-glucosidase (EC 3.2.1.21)which finally hydrolyses cellobiose into glucose monomers [16,54].Most known cellulolytic strains have rate-limiting steps caused bydeficiencies in one or more of these glycoside hydrolases [6,18].While some bacteria such as Bacillus polymyxa have been foundto have all three types of activities [55], the Aquitalea isolates hadsignificantly higher �-d-glucosidase and cellobiohydrolase activ-ity than other isolates of the collection, which include the knowncellulose degraders Paenibacillus, Methylobacterium, and Bacillus[55–57]. The most active lignin isolates were Gordonia species,which supports prior observation of a Gordonia strain growing onlignin [58] and its ability to degrade xenobiotics and alkanes includ-ing lignin-related alkyl ethers [59,60]. It is closely related to otherlignocellulolytic organisms such as Mycobacterium, Norcardia, andRhodococcus within the suborder Corynebacterineae [59,61–63].

Nutrient contamination is a concern with cultivation stud-ies, but the general association between cellulose isolates andhigh carbohydrate degradation (�-d-glucosidase, cellobiohydro-lase, CMCase) and conversely, lignin isolates and high phenolicsdegradation (phenol oxidase, peroxidase) makes it likely that theisolates were growing on cellulose or lignin provided as sub-strates in the isolation media. Current enzyme assays, like CongoRed, are limited in resolution but improvements in sensitivi-ties could strengthen the link between carbon source utilizationand enzyme activities so that low CMCase production can bedetected in all isolates grown on CMC [64]. However, the gen-eral association between isolation substrate and enzyme activitywas evident through three different analyses: contingency analysisusing CMCase and xylanase rates, principal component analysis ofthe carbohydrate and phenolics degradation activities, and logisticregression using enzyme activities as a predictor for cellulose andlignin isolates.

This is the first experimental evidence of a possible relation-ship between lignocellulose degradation activity and phylogeny.Few lignocellulose degrading culture collections have been pub-lished within the past decade because of the increasing prevalenceof solely culture-independent studies. This has consequently left

questions regarding lignocellulolytic function in relationship tophylogeny unaddressed, though at a broad phylogenetic scalethere is evidence for phylogenetic coherence based on nichedifferentiation or the distribution of cellulase genes [65,66]. A

6 pplied

pl#cptsbctb

C

hpqadacodntspcf

F

(Otlfd

A

smN

A

f2

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

6 H.L. Woo et al. / Systematic and A

revious isolation study reported no associations between phy-ogeny and function in terms of cellulose utilization {Ulrich, 2008112}[67,68], though there could be niche differentiation betweenellulose and lignin based on time of degradation. It has beenosited that cellulose degradation emerged via convergent evolu-ion, where traits arise irrespective to lineage [7]. As shown in thistudy, it is extremely unlikely that the observed activity differencesetween phylogenetic clades, both at the phyla and genus level,ould be attributed to random chance and thereby supports a rela-ionship between phylogeny and function among lignocellulolyticacteria.

onclusion

This collection of tropical forest soil isolates are members of aighly lignocellulolytic microbial community and should arguablyossess more robust degradation capabilities than bacteria fre-uently studied only due to literature precedents. Despite the biasssociated with cultivation, culture collections can demonstrate theiversity of a functional group of bacteria in a way that molecularnalyses alone cannot, and in fact, culture collections overcomeultivation bias if they discover novel or rarely studied speciesf a selected function as this study has done with lignocelluloseegradation. With this approach, we were able to observe that lig-ocellulolytic function may not be randomly distributed in respecto phylogeny. This investigation of tropical forest soils has providedome of the much-needed insight into the microbial communityhylogeny and enzyme capabilities that account for the rapid ligno-ellulose degradation and whose enzymes could be used for biofueleedstock deconstruction.

unding

This work was performed at the DOE Joint BioEnergy Institutehttp://www.jbei.org) supported by the U.S. Department of Energy,ffice of Science, Office of Biological and Environmental Research,

hrough contract DE-AC02-05CH11231 between Lawrence Berke-ey National Laboratory and the U.S. Department of Energy. Theunders had no role in study design, data collection and analysis,ecision to publish, or preparation of the manuscript.

cknowledgments

This work conducted by the Joint BioEnergy Institute wasupported by the Office of Science, Office of Biological and Environ-ental Research, of the U.S. Department of Energy under Contracto. DE-AC02-05CH11231.

ppendix A. Supplementary data

Supplementary data associated with this article can beound, in the online version, at http://dx.doi.org/10.1016/j.syapm.013.10.001.

eferences

[1] Goyal, A., Ghosh, B., Eveleigh, D. (1991) Characteristics of fungal cellulases.Bioresour. Technol. 36, 37–50.

[2] Gilkes, N.R., Kilburn, D.G., Miller, R.C., Warren, R.A.J., Bacterial cellulases, (1991)Bioresour. Technol. 36, 21–35.

[3] Bhat, M.K., Cellulases, (2000) Related enzymes in biotechnology. Biotechnol.Adv. 18, 355–383.

[4] Coughlan, M.P. (1992) Enzymatic hydrolysis of cellulose: an overview. Biore-

sour. Technol. 39, 107–115.

[5] Sun, Y., Cheng, J. (2003) Hydrolysis of lignocellulosic materials for ethanol pro-duction: a review. ChemInform 34, 1–11.

[6] Maki, M., Leung, K.T., Qin, W. (2009) The prospects of cellulase-producing bacte-ria for the bioconversion of lignocellulosic biomass. Int. J. Biol. Sci. 5, 500–516.

[

Microbiology 37 (2014) 60– 67

[7] Lynd, L.R., Weimer, P.J., Van Zyl, W.H., Pretorius, I.S. (2002) Microbial cellu-lose utilization: fundamentals and biotechnology. Microbiol. Mol. Biol. Rev. 66,506–577.

[8] Pérez, J., Munoz-Dorado, J., de la Rubia, T., Martínez, J. (2002) Biodegradationand biological treatments of cellulose, hemicellulose and lignin: an overview.Int. Microbiol. 5, 53–63.

[9] Dantas, G., Sommer, M.O.A., Oluwasegun, R.D., Church, G.M. (2008) Bacteriasubsisting on antibiotics. Science 320, 100–103.

10] DeAngelis, K.M., Sharma, D., Varney, R., Simmons, B.A., Isern, N.G., Markil-lie, L.M., Nicora, C.D., Norbeck, A.D., Taylor, R.C., Aldrich, J.T., Robinson, E.W.(2013) Evidence supporting dissimilatory and assimilatory lignin degradationin Enterobacter lignolyticus SCF1. Front. Microbiol. 4, 1–14.

11] DeLong, E.F. (2009) The microbial ocean from genomes to biomes. Nature 459,200–206.

12] Giovannoni, S., Stingl, U. (2007) The importance of culturing bacterioplanktonin the ‘omics’ age. Nat. Rev. Microbiol. 5, 820–826.

13] Lorenz, P., Schleper, C. (2002) Metagenome – a challenging source of enzymediscovery. J. Mol. Catal. B: Enzym. 19–20, 13–19.

14] Janssen, P.H., Yates, P.S., Grinton, B.E., Taylor, P.M., Sait, M. (2002) Improved cul-turability of soil bacteria and isolation in pure culture of novel members of thedivisions Acidobacteria, Actinobacteria, Proteobacteria, and Verrucomicrobia.Appl. Environ. Microbiol. 68, 2391–2396.

15] Sait, M., Hugenholtz, P., Janssen, P.H. (2002) Cultivation of globally dis-tributed soil bacteria from phylogenetic lineages previously only detected incultivation-independent surveys. Environ. Microbiol. 4, 654–666.

16] Breznak, J.A., Brune, A. (1994) Role of microorganisms in the digestion of lig-nocellulose by termites. Annu. Rev. Entomol. 39, 453–487.

17] DeAngelis, K., Gladden, J., Allgaier, M., D’haeseleer, P., Fortney, J., Reddy, A.,Hugenholtz, P., Singer, S., Vander Gheynst, J., Silver, W., Simmons, B., Hazen,T. (2010) Strategies for enhancing the effectiveness of metagenomic-basedenzyme discovery in lignocellulolytic microbial communities. BioEnergy Res.3, 146–158.

18] Lu, W.-J., Wang, H.-T., Yang, S.-J., Wang, Z.-C., Nie, Y.-F. (2005) Isolation andcharacterization of mesophilic cellulose-degrading bacteria from flower stalks-vegetable waste co-composting system. J. Gen. Appl. Microbiol. 51, 353–360.

19] Tuomela, M., Vikman, M., Hatakka, A., Itavaara, M. (2000) Biodegradation oflignin in a compost environment: a review. Bioresour. Technol. 72, 169–183.

20] Kato, K., Kozaki, S., Sakuranaga, M. (1998) Degradation of lignin compounds bybacteria from termite guts. Biotechnol. Lett. 20, 459–462.

21] Pasti, M.B., Pometto, A.L., 3rd, Nuti, M.P., Crawford, D.L. (1990) Lignin-solubilizing ability of actinomycetes isolated from termite (Termitidae) gut.Appl. Environ. Microbiol. 56, 2213–2218.

22] Parton, W., Silver, W.L., Burke, I.C., Grassens, L., Harmon, M.E., Currie, W.S.,King, J.Y., Adair, E.C., Brandt, L.A., Hart, S.C., Fasth, B. (2007) Global-scale simi-larities in nitrogen release patterns during long-term decomposition. Science315, 361–364.

23] Tanner, R. (2007) Cultivation of Bacteria and Fungi. In: Hurst, C.J., Crawford,R.L., Garland, J.L., Lipson, D.A., Mills, A.L., Stetzenbach, L.D. (Eds.), Manual ofenvironmental microbiology, ASM Press, Washington D.C., pp. 69–78.

24] Teather, R.M., Wood, P.J. (1982) Use of Congo red-polysaccharide interactionsin enumeration and characterization of cellulolytic bacteria from the bovinerumen. Appl. Environ. Microbiol. 43, 777–780.

25] Lane, D. (1991) 16S/23S rRNA sequencing. In: Goodfellow, E.S.a.M. (Ed.), Nucleicacid techniques in bacterial systematics, Wiley, Chichester, pp. 115–175.

26] DeSantis, T.Z., Hugenholtz, P., Keller, K., Brodie, E.L., Larsen, N., Piceno, Y.M.,Phan, R., Andersen, G.L. (2006) NAST: a multiple sequence alignment server forcomparative analysis of 16S rRNA genes. Nucleic Acids Res. 34, W394–W399.

27] Huber, T., Faulkner, G., Hugenholtz, P. (2004) Bellerophon: a program todetect chimeric sequences in multiple sequence alignments. Bioinformatics20, 2317–2319.

28] DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K., Huber,T., Dalevi, D., Hu, P., Andersen, G.L. (2006) Greengenes, a chimera-checked16S rRNA gene database and workbench compatible with ARB. Appl. Environ.Microbiol. 72, 5069–5072.

29] Kim, O.S., Cho, Y.J., Lee, K., Yoon, S.H., Kim, M., Na, H., Park, S.C., Jeon, Y.S., Lee, J.H.,Yi, H., Won, S., Chun, J. (2012) Introducing, EzTaxon-e: a prokaryotic 16S rRNAgene sequence database with phylotypes that represent uncultured species.Int. J. Syst. Evol. Microbiol. 62, 716–721.

30] Markowitz, V.M., Chen, I.-M.A., Palaniappan, K., Chu, K., Szeto, E., Grechkin,Y., Ratner, A., Jacob, B., Huang, J., Williams, P., Huntemann, M., Anderson, I.,Mavromatis, K., Ivanova, N.N., Kyrpides, N.C. (2012) IMG: the integrated micro-bial genomes database and comparative analysis system. Nucleic Acids Res. 40,D115–D122.

31] Dixon, P. (2003) VEGAN, a package of R functions for community ecology. J.Veg. Sci. 14, 927–930.

32] Team, R.D.C.R. 2008 A language and environment for statistical computing. In:R Foundation for Statistical Computing, Vienna, Austria.

33] Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar,Buchner, A., Lai, T., Steppi, S., Jobb, G., F

örster, W., Brettske, I., Gerber, S., Gin-

hart, A.W., Gross, O., Grumann, S., Hermann, S., Jost, R., K√

∂nig, A., Liss, T.,L√◦√ümann, R., May, M., Nonhoff, B.r., Reichel, B., Strehlow, R., Stamatakis,

A., Stuckmann, N., Vilbig, A., Lenke, M., Ludwig, T., Bode, A., Schleifer, K.Ä.(2004) ARB: a software environment for sequence data. Nucleic Acids Res. 32,1363–1371.

34] Peng, C.J., So, T.J. (2002) Logistic regression analysis and reporting: a primer.Underst. Stat. 1, 31–70.

Applie

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[classification of carboxymethyl-cellulose decomposing bacteria isolated from

H.L. Woo et al. / Systematic and

35] Oblinger, J.L., Koburger, J.A. (1975) Understanding and teaching most probablenumber technique. J. Milk Food Technol. 38, 540–545.

36] Gest, H., Favinger, J. (2001) Letter to the editor. Taxonomic ambiguities: a casehistory. Int. J. Syst. Evol. Microbiol. 51, 707–710.

37] Ciccarelli, F.D., Doerks, T., von Mering, C., Creevey, C.J., Snel, B., Bork, P. (2006)Toward automatic reconstruction of a highly resolved tree of life. Science 311,1283–1287.

38] Marx, M.C., Wood, M., Jarvis, S.C. (2001) A microplate fluorimetric assay for thestudy of enzyme diversity in soils. Soil Biol. Biochem. 33, 1633–1640.

39] Hendel, B., Sinsabaugh, R., Marxsen, J. (2005) Lignin-degrading enzymes: phe-noloxidase peroxidase. In: Grac a, M., Bärlocher, F., Gessner, M. (Eds.), Methodsto study litter decomposition, Springer, Netherlands, pp. 273–277.

40] Fawcett, T. (2006) An introduction to ROC analysis. Pattern Recognit. Lett. 27,861–874.

41] Moore, D.S. 2010 The Basic Practice of Statistics, W.H. Freeman and Company,New York, NY.

42] Kruskal, W.H., Wallis, W.A. (1952) Use of ranks in one-criterion variance anal-ysis. J. Am. Stat. Assoc. 47, 583–621.

43] Vargha, A., Delaney, H.D. (1998) The Kruskal–Wallis test and stochastic homo-geneity. J. Educ. Behav. Stat. 23, 170–192.

44] DeAngelis, K.M., D’Haeseleer, P., Chivian, D., Simmons, B., Arkin, A.P., Mavro-matis, K., Malfatti, S., Tringe, S., Hazen, T.C. (2013) Metagenomes of tropicalsoil-derived anaerobic switchgrass-adapted consortia with and without iron.Standards in Genomic Sciences 7, 382–398.

45] DeAngelis, K.M., Allgaier, M., Chavarria, Y., Fortney, J.L., Hugenholtz, P., Sim-mons, B., Sublette, K., Silver, W.L., Hazen, T.C. (2011) Characterization of trappedlignin-degrading microbes in tropical forest soil. PLoS ONE 6, e19306.

46] Masai, E., Shinohara, S., Hara, H., Nishikawa, S., Katayama, Y., Fukuda, M. (1999)Genetic and biochemical characterization of a 2-pyrone-4,6-dicarboxylic acidhydrolase involved in the protocatechuate 4,5-cleavage pathway of Sphin-gomonas paucimobilis SYK-6. J. Bacteriol. 181, 55–62.

47] Ramachandra, M., Crawford, D.L., Hertel, G. (1988) Characterization of an extra-cellular lignin peroxidase of the lignocellulolytic actinomycete Streptomycesviridosporus. Appl. Environ. Microbiol. 54, 3057–3063.

48] Ahmad, M., Taylor, C.R., Pink, D., Burton, K., Eastwood, D., Bending, G.D., Bugg,T.D. (2010) Development of novel assays for lignin degradation: comparativeanalysis of bacterial and fungal lignin degraders. Mol. BioSyst. 6, 815–821.

49] Bruce, T., Martinez, I., Maia Neto, O., Vicente, A., Kruger, R., Thompson, F. (2010)Bacterial community diversity in the Brazilian Atlantic forest soils. Microb. Ecol.60, 840–849.

50] Jyoti, V., Narayan, K.D., Das, S.K. (2010) Gulbenkiania indica sp. nov., isolatedfrom a sulfur spring. Int. J. Syst. Evol. Microbiol. 60, 1052–1055.

51] Lin, M.-C., Chou, J.-H., Arun, A.B., Young, C.-C., Chen, W.-M. (2008) Pseudogul-

benkiania subflava gen. nov., sp. nov., isolated from a cold spring. Int. J. Syst.Evol. Microbiol. 58, 2384–2388.

52] Vaz-Moreira, I., Nobre, M.F., Nunes, O.C., Manaia, C.I.M. (2007) Gulbenkianiamobilis gen. nov., sp. nov., isolated from treated municipal wastewater. Int. J.Syst. Evol. Microbiol. 57, 1108–1112.

[

d Microbiology 37 (2014) 60– 67 67

53] Weber, K.A., Hedrick, D.B., Peacock, A.D., Thrash, J.C., White, D.C., Achenbach,L.A., Coates, J.D. (2009) Physiological and taxonomic description of the novelautotrophic, metal oxidizing bacterium, Pseudogulbenkiania sp. strain 2002.Appl. Microbiol. Biotechnol. 83, 555–565.

54] Gilbert, H.J., Hazlewood, G.P. (1993) Bacterial cellulases and xylanases. J. Gen.Microbiol. 139, 187–194.

55] Ivanen, D.R., Rongjina, N.L., Shishlyannikov, S.M., Litviakova, G.I., Isaeva-Ivanova, L.S., Shabalin, K.A., Kulminskaya, A.A. (2009) Novel precipitatedfluorescent substrates for the screening of cellulolytic microorganisms. J.Microbiol. Methods 76, 295–300.

56] Wang, C.M., Shyu, C.L., Ho, S.P., Chiou, S.H. (2008) Characterization of a novelthermophilic, cellulose-degrading bacterium Paenibacillus sp. strain B39. Lett.Appl. Microbiol. 47, 46–53.

57] Jayashree, S., Lalitha, R., Vadivukkarasi, P., Kato, Y., Seshadri, S. (2011) Cellu-lase production by Pink Pigmented Facultative Methylotrophic Strains (PPFMs).Appl. Biochem. Biotechnol. 164, 666–680.

58] Kummer, C., Schumann, P., Stackebrandt, E. (1999) Gordonia alkanivorans sp.nov., isolated from tar-contaminated soil. Int. J. Syst. Evol. Microbiol. 49,1513–1522.

59] Arenskotter, M., Broker, D., Steinbuchel, A. (2004) Biology of the metabolicallydiverse genus Gordonia. Appl. Environ. Microbiol. 70, 3195–3204.

60] Kim, Y.-H., Cha, C.-J., Engesser, K.-H., Kim, S.-J. (2008) Degradation of var-ious alkyl ethers by alkyl ether-degrading Actinobacteria isolated fromactivated sludge of a mixed wastewater treatment. Chemosphere 73,1442–1447.

61] Sundman, V., Kuusi, T., Kuhanen, S., Kilpi, S., Sederholm, H. (1968) Observationson bacterial utilization of the lignin from brown-rotted spruce wood and ofBrauns’ native lignin. Finska Kemists. Medd. 77, 70–86.

62] Haider, K., Trojanowski, J., Sundman, V. (1978) Screening for lignin degrad-ing bacteria by means of 14C-labelled lignins. Arch. Microbiol. 119,103–106.

63] Trojanowski, J., Haider, K., Sundman, V. (1977) Decomposition of 14C-labelled lignin and phenols by a Nocardia sp. Arch. Microbiol. 114,149–153.

64] Carder, J.H. (1986) Detection and quantitation of cellulase by Congo red stainingof substrates in a cup-plate diffusion assay. Anal. Biochem. 153, 75–79.

65] Philippot, L., Andersson, S.G.E., Battin, T.J., Prosser, J.I., Schimel, J.P., Whitman,W.B., Hallin, S. (2010) The ecological coherence of high bacterial taxonomicranks. Nat. Rev. Microbiol. 8, 523–529.

66] Berlemont, R., Martiny, A.C. (2013) Phylogenetic distribution of potential cel-lulases in bacteria. Appl. Environ. Microbiol. 79, 1545–1554.

67] Wirth, S., Ulrich, A. (2002) Cellulose-degrading potentials and phylogenetic

soil. Syst. Appl. Microbiol. 25, 584–591.68] Ulrich, A., Klimke, G., Wirth, S., Diversity, (2008) Activity of cellulose-

decomposing bacteria, isolated from a sandy and a loamy soil after long-termmanure application. Microb. Ecol. 55, 512–522.