ISOLATION AND PARTIAL CHARACTERIZATION OF A NOVEL EIF-2-ALPHA KINASE FROM DROSOPHILA EMBRYONIC-CELLS

Page 1

ORIGINAL ARTICLE

Isolation and partial characterization of novel genesencoding acidic cellulases from metagenomes of buffalorumensC.-J. Duan, L. Xian, G.-C. Zhao, Y. Feng, H. Pang, X.-L. Bai, J.-L. Tang, Q.-S. Ma and J.-X. Feng

Guangxi Key Laboratory of Subtropical Bioresources Conservation and Utilization, The Key Laboratory of Ministry of Education for Microbial

and Plant Genetic Engineering, and College of Life Science and Technology, Guangxi University, Nanning, Guangxi, People’s Republic of China

Introduction

Cellulose is the most abundant biomass on earth and

can be degraded by cellulases, a group of enzymes

including endo-b-1,4-glucanases (EC3Æ2Æ1Æ4), cellobio-

hydrolases (EC3Æ2Æ1Æ91), cellodextrinases (EC3Æ2Æ1Æ74),

and b-glucosidases (EC3Æ2Æ1Æ21) (Lynd et al. 2002). The

product glucose can be easily fermented into useful

chemicals. Low efficiency of many cellulases in hydroly-

sing cellulose into glucose might have contributed to the

high cost of fuel ethanol production from lignocellulosic

biomass, therefore, studying the biology of cellulases are

necessary to overcome such technical obstacles. Cellulas-

es exist in many places including the rumens of rumi-

nants. Discovering novel genes encoding cellulases with

better properties from this natural source is the aim of

this study.

In ruminants, fibre digestion occurs mainly in the

rumen. Ruminal cellulolytic micro-organisms are able to

digest cellulosic materials of plants and produce fermen-

tation products and microbial proteins for the host

animals. The rumen microbial ecosystem mainly consists

of obligate anaerobic micro-organisms including bacteria,

archaea, fungi and protozoa. However, scientists generally

agree that cellulolysis in the rumen is primarily because

of the activities of the ruminal cellulolytic bacteria

because of their numerical predominance and metabolic

diversity (Cheng et al. 1991). Efficient breakdown of

cellulose in the rumen requires the synergistic actions of

cellulases with different properties. Many genes encoding

Keywords

cellulase, glycosyl hydrolase, metagenome,

properties, ruminal microorganisms.

Correspondence

Jia-Xun Feng, College of Life Science and

Technology, Guangxi University, the Eastern

Campus, 75 Xiuling Road, Nanning, Guangxi,

PO 530005, People’s Republic of China.

E-mail: [email protected];

[email protected]

2008 ⁄ 1189: received 10 July 2008,

revised 6 November 2008 and

accepted 4 December 2008

doi:10.1111/j.1365-2672.2009.04202.x

Abstract

Aims: To clone and characterize genes encoding novel cellulases from metage-

nomes of buffalo rumens.

Methods and Results: A ruminal metagenomic library was constructed and

functionally screened for cellulase activities and 61 independent clones express-

ing cellulase activities were isolated. Subcloning and sequencing of 13 positive

clones expressing endoglucanase and MUCase activities identified 14 cellulase

genes. Two clones carried two gene clusters that may be involved in the degra-

dation of polysaccharide nutrients. Thirteen recombinant cellulases were

partially characterized. They showed diverse optimal pH from 4 to 7. Seven

cellulases were most active under acidic conditions with optimal pH of 5Æ5 or

lower. Furthermore, one novel cellulase gene, C67-1, was overexpressed in

Escherichia coli, and the purified recombinant enzyme showed optimal activity

at pH 4Æ5 and stability in a broad pH range from pH 3Æ5 to 10Æ5. Its enzyme

activity was stimulated by dl-dithiothreitol.

Conclusions: The cellulases cloned in this work may play important roles in

the degradation of celluloses in the variable and low pH environment in

buffalo rumen.

Significance and Impact of the Study: This study provided evidence for the

diversity and function of cellulases in the rumen. The cloned cellulases may at

one point of time offer potential industrial applications.

Journal of Applied Microbiology ISSN 1364-5072

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256 245

Page 2

cellulases have been characterized from ruminal micro-

organisms (Krause et al. 2003), but most of them are

isolated from pure cultured micro-organisms. In recent

years, sequence analysis of the 16S rRNA genes has

revealed a highly diverse group of bacteria including

many previously uncultured ones in the ruminal samples

(Tajima et al. 1999; Koike et al. 2003; An et al. 2005).

The uncultured bacteria in the rumen may offer a better

and more natural source for the discovery of potential

novel cellulase genes.

Metagenomic approaches have been widely used to

isolate novel biocatalysts from environmental samples

(Daniel 2005; Lorenz and Eck 2005). Several meta-

genome-derived cellulase genes have been identified in

metagenomic libraries prepared from various environ-

mental samples (Healy et al. 1995; Rees et al. 2003; Voget

et al. 2003, 2006; Ferrer et al. 2005; Feng et al. 2007;

Palackal et al. 2007; Kim et al. 2008).

Ruminal pH is the central issue on keeping the rumen

healthy. Generally, the ruminal pH varies considerably

and can drop below 6 for a significant period of time

throughout the day (Yang et al. 2001; Michalet-Doreau

et al. 2002) because of the production and subsequent

removal of fermentation acid via absorption through the

rumen wall and neutralization by saliva (Allen 1997).

Many different factors can additionally influence or

reduce ruminal pH, such as decreases in the forage

particle size (Krause and Combs 2003) and the forage to

concentrate ratio (Nocek et al. 2002) in the diet. There-

fore, ruminal diurnal pH fluctuation can be more than

one unit (pH 5-7). In vitro and in vivo studies showed

that fibre degradation decreased when pH values in the

rumen are below 6 (Russell and Wilson 1996). In vitro

studies indicated that the ruminal ecosystem can adapt to

a temporary decrease or a small variation in pH under

in vivo conditions (Calsamiglia et al. 2002; Wales et al.

2004; Cerrato-Sanchez et al. 2008).

Many cellulase genes have been cloned from ruminal

micro-organisms. Generally, cellulases retrieved from cul-

tured ruminal bacteria require pH optima for activities

from 6 to 7, but some cellulases active at more acidic

conditions have also been identified. Among 14 cellulases

identified from Fibrobacter succinogenes, four required

optimal pH in the range of 5Æ3–5Æ8, four showed pH

optima of 5Æ9–6Æ2 and another six exhibited pH optima

of 6Æ3–7Æ0 (McGavin and Forsberg 1988; Bera et al. 1996;

Malburg et al. 1996, 1997). However, much is still to be

known about the cellulases functioning in the variable

and low pH environment that help to keep high rate of

cellulolysis in the rumen. In this study, we report the

isolation, sequence analysis of 14 cellulase genes from a

metagenomic library of buffalo ruminal contents, and

partial characterization of these gene products.

Materials and methods

DNA isolation from rumen sample and cosmid library

construction

Rumen contents were sampled from four slaughtered

buffalos that were fed with hay and leaves of sugarcane.

The samples were collected and stored at )80�C until the

DNA extraction was performed. To obtain the bacterial

fraction attached to cellulose particles, the rumen samples

were diluted in potassium-phosphate buffer (PPB;

0Æ18 mol l)1, pH 6Æ5), vortexed gently to allow cellulose

particles to settle. After removing the liquid fraction, the

settled pellet was washed twice with PPB and processed

for DNA extraction.

Total metagenomic DNA was extracted from rumen

samples and further purified by the method of Feng et al.

(2007). Cosmid library was prepared in cosmid vector

pWEB::TNC (pWEB::TNC Cosmid Cloning Kit; Epicentre,

Madison, WI) following the manufacturer’s instruction.

Screening for clones expressing cellulase activities

Cellulase activities, including activities of endoglucanase,

exoglucanase and b-glucosidase, were assayed on agar

plates containing appropriate substrates, as previously

described (Feng et al. 2007). Positive clones selected from

plates were growth expanded, and DNA were extracted,

and retransformed into Escherichia coli EPI100. Their

cellulase activities were confirmed by re-plating them

onto appropriate substrates. The DNA were also digested

with restriction enzymes and their digestion patterns were

analysed by agarose gel electrophoresis.

Annotation of cellulase genes and other open reading

frames (ORF) in positive clones

Plasmid DNA were extracted from positive transformants

of EPI100, and digested with the appropriate restriction

enzyme to yield smaller fragments by following the stan-

dard techniques (Sambrook and Russell 2001). They were

subcloned into plasmid vectors pGEM-3zf(+) or pBlue-

script M13(+), and selected for cellulase activity. The

shortest inserts encoding the corresponding cellulases

were selected for sequencing. The ORF were detected

using the online GeneMark software (http://opal.biology.

gatech.edu/GeneMark/gmhmm2_prok), and the ORF fin-

der (http://www.ncbi.nlm.nih.gov/gorf/gorf) provided by

the National Center for Biotechnology Information

(NCBI). The modular structures of the enzymes were pre-

dicted by SMART online (http://smart.embl-heidelberg.

de). Multiple alignments of catalytic domains of cellu-

lases were preformed with ClustalX 1.83 program or

Cellulases from rumen metagenomes C.-J. Duan et al.

246 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256

ª 2009 The Authors

Page 3

ClustalW online tool (http://www.ebi.ac.uk/Tools/

clustalw). Phylogenic tree was generated by MEGA2. The

amino acid sequences of family 5 glycosyl hydrolases most

similar with the cloned cellulases were retrieved from

GenBank.

Protein and enzyme assays

Selected clones were grown in 200-ml cultures over-

night. The cells were harvested, resuspended in 4 ml of

citrate-phosphate buffer (McIlvaine buffer, pH 6Æ5) and

lysed by sonication. Crude cell lysates were centrifuged

to remove cellular debris; the supernatants were

collected as crude protein extracts for the subsequent

assays.

The concentrations of proteins in the crude extracts

were measured by the method of Bradford (1976). The

crude protein extracts were characterized for cellulase

activities. Eleven cellulases (CMCases) and two MUCases,

the hydrolases for 4-methylumbelliferyl-b-d-cellobioside

(4-MUC), were assayed for their optimal pH values and

temperatures using carboxymethyl cellulose (CMC, low

viscosity; Sigma) and p-nitrophenyl-b-d-cellobioside

(p-NPC; Sigma) as substrates, respectively, as described

previously (Feng et al. 2007).

To analyse substrate specificities, the protein crude

extracts were measured for potential cellulase activities

10–120 min after incubation at the optimal temperature

in the optimal buffer containing 1% (w ⁄ v) polysaccha-

rides or 2Æ5 mmol l)1 p-nitrophenyl (p-NP) derivatives.

The tested polysaccharides were CMC, Avicel (Fluka),

lichenan (Sigma), barley glucan (Sigma), 2-hydroxyethyl

cellulose (Sigma), methyl cellulose (Sigma), oat spelt

xylan (Sigma), birch wood xylan (Sigma) and laminarin

(Sigma). The reducing sugars released from the sub-

strates were measured with 3,5-dinitrosalicylic acid

agents as described by Miller (1959). One unit (U) of

endoglucanase activity was defined as the amount of

enzyme releasing one micromole of reducing sugar per

minute from the substrate. p-NPC and its derivative

p-nitrophenyl-b-d-glucopyranoside (p-NPG; Sigma) were

used as substrates for exoglucanase and b-glucosidase,

respectively. The activity was determined by measuring

the amount of p-NP generated from p-NPC or pNPG

according to the method by Odoux et al. (2003). One

unit of the exoglucanase or b-glucosidase activity was

defined as the amount of enzyme releasing 1 lmol of

p-NP per minute. Highest enzymatic activity towards

one substrate for each crude enzyme extract of subclone

was taken as 100% activity, and the relative activity of

each enzyme towards other respective substrates were

calculated as the percentage of that of the highest

activity.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis

(SDS-PAGE), native-PAGE and cellulase zymography

SDS-PAGE was performed according to the method of

Laemmli (1970). Native PAGE was carried out similarly

with the exclusion of SDS from all solutions. Zymogram

of endoglucanases with native PAGE was performed

similar to zymogram analysis after SDS-PAGE as

described by Feng et al. (2007), except that the gels were

washed twice in 10 mmol l)1 PPB (pH 7Æ2) without pro-

tein refolding. For zymographic analyses of exoglucanases,

the treated native gel was incubated with 1 ml of

2Æ5 mmol l)1 p-NPC solution at 37�C for 20 min, and

then activities of exoglucanase were visualized as yellow

bands on the gel.

Expression and purification of recombinant enzyme C67-1

A new cellulase gene C67-1 was amplified from positive

clone C67 by polymerase chain reaction (PCR) with pri-

mer pairs C67BF1 (5¢-GACCATATGGATGCCGTCAAGA

ACATGGGTGT-3¢) and C67HR1 (5¢-AGCAAGCTTCTGT

TTCACGAACTTCTTTCCGTTCT-3¢). NdeI and HindIII

sites (underlined) were added to the forward and reverse

primers, respectively. PCR product was purified, digested

by NdeI and HindIII, and cloned into expression vector

pET-30a(+) (Novagen) for expressing recombinant pro-

teins with 6xHis tag at the C-termini. Recombinant

plasmid was transformed into host E. coli Rosetta (DE3)

pLysS (Novagen) and positive clones were selected by

kanamycin. Cells harbouring the recombinant plasmids

were first grown to OD 0Æ6 in Luria-Bertani (LB) broth

containing 25 lg ml)1 of kanamycin and 34 lg ml)1 of

chloramphenicol at 37�C and 200 rev min)1, and then

expression of the C67-1 gene was induced by adding

0Æ5 mmol l)1 IPTG in the medium. Following a further

incubation at 15�C and 100 rev min)1 overnight, the

recombinant protein was extracted from the cytoplasmic

fraction of the cell lysates, and purified by affinity

chromatography through two columns with nickel-

nitrilotriacetic acid agarose resin (Ni-NTA; Qiagen)

according to the product manual with some modifica-

tions. Imidazole concentrations in wash buffer were

increased to 40 mmol l)1 from the suggested 20 mmol l)1

for the first column, and 60 mmol l)1 for the second puri-

fication. The final purified protein solution was desalted

by ultrafiltration column Amicon Ultro-10 (Millipore)

and diluted into pH 4Æ5 citrate ⁄ phosphate buffer.

Characterization of C67-1 activities

To assay cellulase activities, a standard reaction contain-

ing 0Æ6 lg of recombinant C67-1 and 1% CMC in 0Æ5 ml

C.-J. Duan et al. Cellulases from rumen metagenomes

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256 247

Page 4

citrate ⁄ phosphate buffer, pH 4Æ5, was incubated at 45�C

for 10 min. Activities of C67-1 were assayed under a

range of pH values between pH 3 and 8 in the same buf-

fer, and a range of temperatures between 15 and 70�C.

pH and thermal stabilities were assayed as described pre-

viously (Eckert and Schneider 2003; Inoue et al. 2005).

Substrate specificity of C67-1 was tested against a panel

of different substrates in standard reactions with CMC

replaced by lichenan, barley-glucan, laminarin, oat spelt

xylan, avicel or acid-swollen cellulose that was prepared

by the method of Wood (1971) or using filter paper

(Whatman). Km and Vmax values for C67-1 were deter-

mined under optimal condition at CMC concentrations

ranging from 8 to 50 mg ml)1. A series of chemical

reagents, including several metal chloride salts (exception

of 2 mmol l)1 MnCl2) at 5 mmol l)1, chelating agent

EDTA at 5 mmol l)1, surfactants SDS and Triton X-100,

0Æ25% each, and reducing agent dithiothreitol (DTT) at

2 mmol l)1 were investigated for their effects on activities

of C67-1 under the standard assay conditions.

An endo or exo mode of action of C67-1 was determined

by incubating 0Æ3 lg of the purified enzyme with one of the

following cellooligosaccharides: 20 lg cellobiose (Sigma),

60 lg cellotriose (Sigma), 80 lg cellotetraose (Sigma), 80 lg

cellopentaose (Sigma) and 80 lg cellohexaose (Seikagaku,

Tokyo, Japan). After 2 h at 45�C, the hydrolysis products

from each reaction was detected by thin layer chromatogra-

phy (TLC) as described previously (Feng et al. 2007).

Nucleotide sequence accession numbers

All DNA sequences reported in this paper were deposited

in the GenBank database under accession numbers

EU449481–EU449493.

Results

Construction and screening of the metagenomic library

from bacteria in the buffalo rumen

To clone cellulase genes from natural micro-organisms in

buffalo rumens, a cosmid library was constructed with

metagenomic DNA isolated directly from uncultured

micro-organisms of rumen samples. The library yielded

c. 15 000 clones, and its quality and insert sizes were

analysed by restriction digestion of purified cosmids from

14 randomly chosen clones. The results showed that

inserted DNA fragments of these selected clones ranged

from 20 to 50 kb with distinct restriction patterns, indi-

cating that they each represented a unique sequence.

Therefore, the library most likely harboured randomly

cloned DNA. The average insert size for the 14 clones was

estimated to be 35 kb and the full library size was about

525 Mb.

Colonies of the library were transferred to LB agar

plates containing different substrates to screen for cellu-

lase activities. Eleven independent clones expressing

CMCase activities, two expressing MUCase activities and

forty-eight expressing b-glucosidase activities were

isolated. Cosmids isolated from all these clones were

transformed into E. coli EPI100, and all transformants

retained the corresponding enzyme activities, thus their

cellulase activities were encoded by the respective cosmid

inserts. The clones expressing endoglucanase and MUCase

activities were selected for further study.

Subcloning and sequencing of the positive clones for

cellulase genes

In order to locate the cellulase genes from positive clones

and facilitate sequencing, we further subcloned them with

shorter insert sizes of 1Æ7–9 kb from each clone. Sub-

clones that retained the corresponding cellulase activities

were characterized by partial or complete sequencing for

cellulase genes and their neighbouring ORF. As shown in

Table 1, each subclone harboured a complete ORF encod-

ing for either an endo-b-1,4-glucanase or MUCase except

subclone DC2B that only contained a partial ORF for

DC2-4. Both upstream and downstream of this partial

ORF of DC2B were sequenced with the source clone DC2

as template. The sequencing of this clone found another

cellulase gene (DC2-3) in clone DC2 as shown in Table 1.

Altogether, 12 endoglucanase genes from the 11 positive

clones and 2 MUCase genes from another two clones

have been identified (Table 1).

Analyses of predicted cellulase sequences and domain

structures

The predicted products of the 14 cellulase genes consisted

of 332–553 amino acids and their molecular masses

ranged from 38 208 to 62 308 Da. Six of them showed

less than 80% similarities to other cellulases in the NCBI

databases. They shared less than 55% identities and 70%

similarities at amino acid levels with cellulases from

cultured ruminal bacteria from Prevotella, Clostridium or

Ruminococcus except M40-2 (Table 1).

SMART analysis of deduced amino acid sequences of

the 14 cellulase genes showed that they each contained a

glycosyl hydrolase (GH) family 5 catalytic domain and a

signal peptide except that DM1-1 and M8-2 lacked signal

peptides. Multi-alignment with known three-dimensional

(3D) structure of GHF5 cellulases in GenBank by

ClustalW online tools revealed that two important

Cellulases from rumen metagenomes C.-J. Duan et al.

248 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256

ª 2009 The Authors

Page 5

residues in the catalytic active sites, the proton donor Glu

and the nucleophile amino acid Glu, were well conserved

among catalytic domains of these newly identified

cellulases (Table 1 and Fig. S1).

Other ORF identified from the positive clones

Apart from the cellulase genes, other putative GH

family genes were also localized on some of these posi-

tive cosmid clones. Cosmid DC2 harboured three other

GH genes encoding putative polygalacturonase, man-

nanase and glycosylase. As shown in Table 1, they dis-

played 41%, 69% and 88% similarity with their

respective closest homologues in databases. Another

predicted mannanase gene DC23-2 was found in cosmid

DC23 upstream of cellulase gene DC23-3. The deduced

protein of DC23-2 shared 38% identity and 51% simi-

larity with an endoglycosidase precursor from an

unidentified rumen bacterium. DC23-2 and the endo-

glycosidase precursor were only similar at the N-termi-

nal GH family 26 module that is responsible for

mannanase activity (Palackal et al. 2007). There were

other ORF neighbouring the GH on these positive

clones that encoded proteins with similarities to

proteins from a wide range of organisms, mainly from

Bacteriodes (Table S1).

Table 1 Putative glycosyl hydrolases obtained from subclones or source clones expressing cellulase activities from the buffalo rumen

metagenomes

Cosmid

(subclone)

Predicted

ORF

Protein

Length (aa) Most homologous protein

Identity ⁄similarity (%)

GH ⁄ conserved domain

(range) ⁄ catalytic residues

DC2 (DC2B) DC2-2 681 Prevotella bryantii putative

polygalacturonase (AAC97595)

22 ⁄ 41 Putative polygalacturonase

DC2-3 512 Unidentified micro-organism

endoglucanase (ABX76048)

50 ⁄ 63 Cellulase ⁄ GHF5(173–491) ⁄ E323 and E446

DC2-4 517 Unidentified micro-organism

cellulase (CAJ19139)

68 ⁄ 82 Cellulase ⁄ GHF5(152–486) ⁄ E305 and E438

DC2-5 355 Unidentified micro-organism

glycosyl hydrolase (CAJ19136)

54 ⁄ 69 Mannanase ⁄ GHF26(22–326) ⁄ E185

and E289

DC2-6 391 Unidentified micro-organism

conserved hypothetical protein

(CAJ19137)

78 ⁄ 88 Predicted glycosylase ⁄ DUF377(31–362)

DC3 (DC3B) DC3-1 520 Unidentified micro-organism

endoglucanase (ABX76048)

72 ⁄ 81 Cellulase ⁄ GHF5(161–502) ⁄ E328 and E459

M8 (M8-5) M8-2 344 Clostridium thermocellum

endoglucanase (BAA00793)

44 ⁄ 61 Cellulase ⁄ GHF5(15–328) ⁄ E147 and E284

DC9 (DC9BE) DC9-2 346 Uncultured rumen bacterium

beta-glucanase (CAP07661)

81 ⁄ 89 Cellulase ⁄ GHF5(45–321) ⁄ E177 and E273

M11 (M11-15) M11-2 335 Uncultured rumen bacterium

beta-glucanase (CAP07661)

80 ⁄ 90 Cellulase ⁄ GHF5(34–309) ⁄ E165 and E261

DC20 (DC20HS) DC20-2 332 Uncultured rumen bacterium

beta-glucanase (CAP07661)

83 ⁄ 91 Cellulase ⁄ GHF5(31–307) ⁄ E163 and E259

DC23 (DC23B) DC23-2 496 Uncultured bacterium endoglycosidase

precursor protein (ABB46200)

38 ⁄ 51 Mannanase ⁄ GHF26(150–476) ⁄ E319

and E439

DC23-3 518 Unidentified micro-organism

endoglucanase (ABX76048)

51 ⁄ 65 Cellulase ⁄ GHF5(178–495) ⁄ E328 and E451

C29 (C29-4X) C29-2 553 Unidentified micro-organism

endo-1,4-beta-D-glucanase (ABX76045)

76 ⁄ 86 Cellulase ⁄ GHF5(39–340) ⁄ E178 and E288

C35 (C35S8) C35-2 552 Unidentified micro-organism

endo-1,4-beta-d-glucanase (ABX76045)

74 ⁄ 84 Cellulase ⁄ GHF5(39–336) ⁄ E174 and E284

C5614 (C5614E7) C5614-1 537 Unidentified micro-organism

endo-1,4-beta-d-glucanase (ABX76045)

54 ⁄ 66 Cellulase ⁄ GHF5(40–334) ⁄ E172 and E282

C67 (C67E4) C67-1 546 Unidentified micro-organism

endo-1,4-beta-d-glucanase (ABX76045)

67 ⁄ 79 Cellulase ⁄ GHF5(40–334) ⁄ E172 and E282

M40 (M40ES) M40-2 386 Prevotella ruminicola cellulase (BAA74515) 83 ⁄ 90 Cellobiosidase ⁄ GHF5 (46–363) ⁄ E178

and E320

DM1 (DM1P17) DM1-1 332 Ruminococcus flavefaciens

cellodextrinase A (P16169)

52 ⁄ 70 Cellodextrinase ⁄ GHF5 (36–317) ⁄ E157

and E274

ORF, open reading frame.

C.-J. Duan et al. Cellulases from rumen metagenomes

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256 249

Page 6

Phylogenic relationship for the cloned cellulases

The cloned cellulases shared less than 80% similarity with

each other except DC9-2, M11-2 and DC20-2. The pro-

tein sequences of the 14 cellulases were compared with

published cellulases for phylogenic analysis. The phylo-

genic tree showed that the cloned cellulases clustered into

three main groups (Fig. 1). At least nine of them were

closely related to cellulases of Prevotella. Three cellulases,

DC20-2, M11-2 and DC9-2, were closely related to cellu-

lases of an uncultured rumen bacterium. But based on

the low bootstrap confidence, they were affiliated to Buty-

rivibrio fibrisolvens, a typical rumen bacterium. Cellulases

M8-2 and DM1-1 were most closely related to cellulases of

Clostridium thermocellum and Ruminococcus flavefaciens,

respectively.

Characterization of enzyme activities for the cloned

cellulases

Because the inserts of the 13 positive subclones ranged

from 1Æ7 kb to 9Æ0 kb, it is highly possible that some

of them may represent more than one cellulase activity.

Cellulase zymogram with native PAGE showed a single

50 C29-2

Piromyces equi endoglucanse 5A CAB92326

Ruminoccus albus beta-1, 4-endoglucanase V BAA92146Ruminoccus flavefaciens cellulase AAB19708

Prevotella ruminicola carboxymethylcellulase AAA22909Prevotella bryantii beta-1,4-endoglucanase AAC97596

Unidentified microorganism endoglucanase ABX76048

Uncultured bacterium endogluycosidase precursor protein ABB46200

Unidentified microorganism endo-1,4-beta-D-glucanase ABX76045C35-2C5614-1

C67-1

DC2-3

DC3-1

DC2-4Unidentified microorganism cellulase CAJ19139

Unidentified microorganism cellulase CAJ19135

Prevotella ruminicola endoglucanase 1814455A

Neocallma stix patriciarum endoglucanase B CAA83238

M11-2

Uncultured rumen bacterium beta-glucanse CAPP07661

DC9-2

M40-2

M8-2Clostridium thermocellum glucoside hydrolase, family 5 ZP_00504672

Unidentfied bacterium cellulase AAA91966

DM1-1Ruminococuus flavefaciens Cellodextrinase A P16169

DC20-2

Cellvibrio mixtus cellulase AAAB61461Uncultured bacterium cellulase ABA02176

Butyrivibrio fibrisolvens endoglucanase A precursor P22541

Prevotella ruminicola cellulase BAA74515Fibrobacter succinogenes cellodextrinase AAA50210

DC23-3

87

100

9299

70

71

100

98

99

92

96

93

45

100

100100 Orpinomyces sp.cellulase AAD04193

100100

681899

41100

100

100

78

0·20

9342

72

84

84

93

Figure 1 Phylogenic tree of catalytic domains of GHF 5 cellulases generated by the neighbour-joining method. The amino acid sequences of

family 5 glycosyl hydrolases were retrieved from GenBank by homology searching against protein database with the cloned cellulases, and submit-

ted to SMART for determining the catalytic domain of each enzyme. Multiple sequence alignments were performed in CLUSTALX 1Æ83. The obtained

alignments were then used in MEGA 2.1 to establish the phylogenetic tree. Bootstrap values for 1000 replications and Poisson correction were

performed. Database accession numbers are shown in bold after each enzyme. The bar represents 0Æ2 changes per amino acid.

Cellulases from rumen metagenomes C.-J. Duan et al.

250 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256

ª 2009 The Authors

Page 7

positive band for each protein crude extract from all the

subclones (Fig. 2), indicating that each of them produced

only a single cellulase protein responsible for the observed

activity.

The 13 cellulases were assayed for their activities under

a range of pH values and temperatures with CMC or

p-NPC as substrates. They showed diverse pH optima

ranging from 4 to 7; seven were most active at pH 5Æ5 or

lower, and C5614-1 exhibited optimal activity at pH 4Æ0.

To our knowledge, C5614-1 is the most acidophilic cellu-

lase from ruminal bacteria that require the lowest pH

value to perform its best (Table 2). The optimal tempera-

tures of all are 35–50�C.

All the cloned cellulases were also tested for their sub-

strate specificities (Table 3), and 11 CMCases showed

highest activities towards barley glucan (91–23 730 U g)1

protein and taken as 100%), and decreased their activities

in order towards lichenan (14–51%), CMC (2Æ9–53Æ4%)

and 2-hydroxyethyl cellulose (0Æ31–26Æ8%). Activities of

these enzymes towards methyl cellulose and Avicel were

generally low except DC3-1 that exhibited higher activities

towards methyl cellulose (6Æ4%). The 11 CMCases

could also hydrolyse xylan from birchwood and oat spelt.

Interestingly, C67-1 showed highest activities towards all

these substrates except Avicel among these CMCases. The

M8-2, DC9-2, M11-2 and DC20-2 could hydrolyse

p-NPC with higher activities (12Æ2–89Æ4%). The two

MUCases, DM1-1 and M40-2, had the highest activities

towards p-NPC (26 993 and 685 U g)1, respectively and

taken as 100%), followed by those towards barley glucan

(0Æ35% and 38Æ8%, respectively) and lichenan (0Æ06%

and 5Æ7%, respectively). They also hydrolysed CMC,

2-hydroxyethyl cellulose, methyl cellulose, Avicel, xylan

and p-NPG to lesser extents. The abilities of these

enzymes to hydrolyse laminarin, a b-1,3 ⁄ 1,6-glucan, was

also tested but no such activities had been detected. These

data indicated that it is probably more appropriate to

consider CMCases and MUCases cloned in this study as

endo-b-1,4-glucanases and cellodextrinases, respectively.

Overexpression of C67-1, purification and

characterization of the translated product C67-1

Because C67-1 showed apparently highest activity towards

a wide range of tested substrates among those CMCases,

it was selected for further study. The recombinant C67-1

Table 2 Optimal pH values and temperatures of the cellulases cloned from metagenomic library of buffalo rumen

Cellulase DC2-4 DC3-1 M8-2 DC9-2 M11-2 DC20-2 DC23-3 C29-2 C35-2 C5614-1 C67-1 DM1-1 M40-2

Optimal pH 6Æ5 5 5Æ5 7 6Æ5 5Æ5 6Æ5 5 4Æ5 4Æ0 4Æ5 6Æ0 6Æ5

Optimal temperature 50�C 50�C 50�C 45�C 45�C 50�C 50�C 50�C 45�C 45�C 45�C 45�C 35�C

Protein crude extracts were prepared from cultures of Escherichia coli EPI100 positive subclones, and activities of the cloned endo-b-1,4-glucanases and the MUCases were

measured at various pH values and temperatures to determine the optimal values for substrates carboxymethyl cellulose or p-nitrophenyl-b-D-cellobioside. Details for

protein crude extracts and the assay procedures are described in ‘Materials and methods’.

1 2 3 4 5 6 7(a) (b) (c)

8 9 10 111213 1 2 3 4 5 6 7 8 9 1011 12 13

Figure 2 Zymograph of the 13 cloned cellulases in native polyacrylamide gel electrophoresis. (a) Staining of proteins in the crude extracts

separated on the polyacrylamide gel with Coomassie brilliant blue G-250. (b) Carboxymethyl cellulose (CMCase) activities detected on CMC plates,

the agar replica of the polyacrylamide gel. (c) Exoglucanase activities detected with 2Æ5 mmol l)1 p-nitrophenyl-b-D-cellobioside solution spreading

on the polyacrylamide gel. Lane 1, DC2-4; lane 2, DC3-1; lane 3, DC8-2; lane 4, DC9-2; lane 5, M11-2; lane 6, DC20-2; lane 7, DC23-3; lane 8,

C29-2; lane 9, C35-2; lane 10, C5614-1; lane 11, C67-1; lane 12, DM1-1; lane 13, M40-2.

C.-J. Duan et al. Cellulases from rumen metagenomes

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256 251

Page 8

Tab

le3

Subst

rate

spec

ifici

tyof

the

pro

tein

crude

extr

acts

from

culture

sof

subcl

ones

expre

ssin

gce

llula

seac

tivi

ties

Test

subst

rate

Rel

ativ

esp

ecifi

cac

tivi

ty(%

)

CM

Cas

esM

UC

ases

DC

2-4

DC

3-1

M8-2

DC

9-2

M11-2

DC

20-2

DC

23-3

C29-2

C35-2

C5614-1

C67-1

DM

1-1

M40-2

Bar

ley

glu

can

(b-1

,3⁄4

-glu

can)

100

±4Æ5

100

±5Æ9

100

±1Æ6

100

±4Æ4

100

±6Æ2

100

±1Æ1

100

±0Æ5

100

±2Æ5

100

±6Æ8

100

±4Æ2

100

±0Æ3

0Æ3

5±

0Æ0

26

38Æ8

±0Æ2

2

Lich

enan

(b-1

,3⁄

4-g

luca

n)

14Æ5

±0Æ5

49Æ3

±3Æ8

51Æ4

±3Æ7

35Æ3

±4Æ3

231Æ9

±0Æ7

49

±1Æ5

29Æ5

±3

32Æ9

±2Æ1

26Æ8

±2Æ1

24Æ3

±3Æ7

23Æ7

±0Æ2

0Æ0

6±

0Æ0

08

5Æ7

±0Æ4

7

Car

boxy

met

hyl

cellu

lose

(b-1

,4-g

luca

n)

17Æ5

±0Æ8

64Æ1

±3Æ5

2Æ9

±0Æ1

53Æ4

±1Æ4

49Æ5

±0Æ6

329Æ1

±3Æ0

12Æ3

±0Æ2

512Æ6

±0Æ7

86Æ6

±0Æ0

736Æ9

±1Æ3

12Æ4

±0Æ2

20Æ0

02

±0Æ0

005

–

2-h

ydro

xyet

hyl

cellu

lose

(b-1

,4-g

luca

n)

15Æ0

±0Æ6

26Æ8

±2Æ8

1Æ0

±0Æ0

316Æ5

±0Æ9

13Æ2

±1Æ1

4Æ0

±0Æ0

911Æ3

±0Æ4

6Æ9

±0Æ3

12Æ1

±0Æ0

21Æ1

±0Æ0

50Æ3

1±

0Æ0

35

0Æ0

15

±0Æ0

04

–

Met

hyl

cellu

lose

(b-1

,4-g

luca

n)

0Æ0

5±

0Æ0

26Æ4

±0Æ8

0Æ0

5±

0Æ0

08

0Æ3

±0Æ0

80Æ8

8±

0Æ1

90Æ8

7±

0Æ0

87

1Æ8

±0Æ6

31Æ2

±0Æ2

01Æ6

±0Æ2

00Æ7

2±

0Æ0

60Æ3

2±

0Æ0

15

––

Avi

cel(b

-1,4

-glu

can)

0Æ4

±0Æ0

80Æ6

±0Æ1

30Æ0

4±

0Æ0

03

0Æ6

8±

0Æ1

70Æ6

6±

0Æ0

66

0Æ1

7±

0Æ0

05

0Æ1

8±

0Æ0

14

0Æ1

8±

0Æ0

45

––

–0Æ0

3±

0Æ0

08

0Æ8

3±

0Æ1

8

Xyl

anfr

om

birch

wood

(b-1

,4-x

ylan

)

2Æ9

±0Æ3

11Æ0

±0Æ7

0Æ3

4±

0Æ0

1–

0Æ8

8±

0Æ2

40Æ1

2±

0Æ0

19

9Æ3

±0Æ2

36Æ7

±0Æ3

84Æ0

±0Æ1

46Æ6

±0Æ7

76Æ5

±0Æ2

4–

–

Xyl

anfr

om

oat

spel

t(b

-1,4

-xyl

an)

8Æ9

±0Æ2

18Æ3

±2Æ0

0Æ9

1±

0Æ0

5–

0Æ5

5±

0Æ0

50Æ0

35

±0Æ0

07

15Æ8

±0Æ1

511Æ7

±0Æ3

95Æ8

±0Æ3

18Æ5

±0Æ3

09Æ9

±0Æ0

80Æ0

2±

0Æ0

02

0Æ4

7±

0Æ0

9

p-n

itro

phen

yl-D

-

cello

bio

side

––

89Æ4

±1Æ6

16Æ5

±0Æ2

215Æ1

±0Æ3

812Æ2

±0Æ1

8–

––

––

100

±2Æ0

100

±0Æ1

4

p-n

itro

phen

yl-D

-

glu

copyr

anosi

de

––

––

––

––

––

–0Æ0

01

±0Æ0

0002

–

Prote

incr

ude

extr

acts

wer

epre

par

edas

des

crib

edin

‘Mat

eria

lsan

dm

ethods’

from

culture

sof

Esch

eric

hia

coli

EPI1

00

posi

tive

subcl

ones

,an

das

saye

dfo

rth

eir

spec

ific

hyd

roly

tic

activi

ties

tow

ards

diffe

rent

subst

rate

s.Sp

ecifi

c

activi

ties

wer

em

easu

red

asen

zym

eunits

inea

chgra

mof

pro

tein

,w

hic

his

defi

ned

in‘M

ater

ials

and

met

hods’

.Th

e100%

activi

tyw

aseq

uiv

alen

tto

each

cellu

lase

hav

ing

hig

hes

tac

tivi

tyto

war

ds

one

subst

rate

,th

ere

lative

activi

tyis

defi

ned

in‘M

ater

ials

and

met

hods’

.D

ata

wer

eex

pre

ssed

asm

ean

±SD

from

valu

esobta

ined

inth

ree

exper

imen

ts,

each

with

trip

licat

em

easu

rem

ents

.N

ohyd

roly

tic

activi

ties

wer

edet

ecte

dag

ainst

any

of

the

test

ed

subst

rate

sw

ith

pro

tein

crude

extr

act

from

culture

sof

E.co

lihar

bouring

the

empty

cosm

idve

ctor

pW

EB::

TNC

alone.

The

‘—’

sign

isuse

dto

indic

ate

valu

esnot

det

ecta

ble

.

Cellulases from rumen metagenomes C.-J. Duan et al.

252 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256

ª 2009 The Authors

Page 9

was purified by nickel-nitrilotriacetic acid agarose resin,

and its high purity was confirmed by SDS-PAGE (data

not shown). Activity of C67-1 towards CMC was the best

at approximately pH 4Æ5 and 45�C, which was also the

optimal condition for the crude protein extract from the

subclone containing the C67-1 gene (Table 2). Low activi-

ties were detected above pH 7Æ5 and below 3Æ5. The pH

stability study demonstrated that C67-1 retained more

than 75% activity after it was stored at 4�C for 24 h at

broad pH values ranging from 3Æ5 to 10Æ5 (Fig. S2a). The

enzyme was stable for 1 h incubation at temperatures

below 45�C with over 80% of the activity remaining and

lost complete activity at temperatures above 50�C

(Fig. S2b). Km and Vmax of the recombinant C67-1

towards CMC were determined to be 36Æ83 mg ml)1 and

367Æ7 U mg)1, respectively.

Substrate specificity of the recombinant C67-1 was

determined under the optimal condition with 1% poly-

saccharides and 2Æ5 mmol l)1 p-NP derivatives. The

enzyme had the highest activities towards barley glucan

(464 U mg)1), then decreased activities in descending

order towards CMC (73 U mg)1), lichenan (57 U mg)1),

2-hydroxyethyl cellulose (20Æ5 U mg)1) and methyl cellu-

lose (19Æ6 U mg)1). The enzyme also hydrolysed xylan

(14Æ4 U mg)1), and showed low activity to acid-swollen

cellulose (2Æ96 U mg)1), but did not hydrolyse insoluble

cellulose (Avicel and filter paper), b-1,3 ⁄ 6-glucan (lami-

narin) or p-NPC and p-NPG. The hydrolysis of cello-

oligosaccharides was analysed by TLC, and C67-1 showed

no activity against cellobiose and cellotriose. It degraded

cellotetraose, cellopentaose and cellohexaose into cellotri-

ose, cellobiose and glucose (Fig. 3). Based on these data,

C67-1 is a typical endo-b-1,4-glucanase.

The influence of different metal ions, chelating agent

(EDTA), surfactants (SDS and Triton X-100) and reduc-

ing agent (DTT) on the activity of C67-1 was tested using

CMC as a substrate (Table S2). In these tests, the enzyme

activity was stimulated strongly only in the presence of

the reducing agent DTT (2 mmol l)1), up to 135% of its

original activity. Divalent metal ions Cr2+, Zn2+, and

Cu2+ caused almost complete losses of activities of C67-1,

while Mn2+, Fe2+ , Co2+ only partially inhibited the

enzyme. Chelating agent EDTA and surfactant TritonX-100

showed no significant effects on the enzyme activity,

whereas SDS was an effective inhibitor, completely abol-

ishing activity of C67-1.

Discussion

Metagenomics that was initially used to identify micro-

bial diversity from uncultured sources has now been

widely employed to discover novel and potentially

important biocatalysts in different environments (Daniel

2005; Lorenz and Eck 2005). DNA can be extracted from

micro-organisms recovered from environmental samples

followed by cell lysis, or directly from sample lysates

without cell recovery (Miller et al. 1999). In this study,

the metagenomes of ruminal bacteria associated with

feed particles were isolated by direct lysis of rumen sam-

ples. Because ruminal bacteria associated with feed parti-

cles are considered to be the most important group for

fibre degradation (Michalet-Doreau et al. 2001), the

direct lysis of their metagenomes has a high possibility

of obtaining cellulase-producing clones. Using this

approach, we constructed a metagenomic library of

c. 525 Mb, and identified 14 cellulase genes (excluding

b-glucosidase genes).

Although the hit rate obtained from this library was

lower than that reported by Ferrer et al. (2005), it was

higher than most screenings for industrial relevant

enzymes from other metagenomic libraries (Lorenz and

Eck 2005), probably resulting from relatively higher num-

bers of cellulase genes in the rumen bacteria. On the

other hand, previous genome sequencing reported 24

G1

G2

G3

G4

1 2 3 4 5 6

Figure 3 Hydrolysis products of cello-oligosaccharides by the recom-

binant C67-1. Lane 1, mixed standard sugars: glucose (G1), cellobiose

(G2), cellotriose (G3) and cellotetraose (G4); lanes 2–6, G2–G4, cello-

pentaose (G5) and cellohexaose (G6) treated with the recombinant

C67-1.

C.-J. Duan et al. Cellulases from rumen metagenomes

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256 253

Page 10

cellulase-encoding genes from ruminal bacterium

F. succinogenes, much more than what had been cloned

with recombinant DNA techniques in E. coli (Krause

et al. 2003), also more than what we found by the activity

screening from the present study. This may be that the

cellulase genes present in the uncultured ruminal bacteria

failed to express in the heterogenous host E. coli, or their

expression level was too low to be detected.

Analyses for protein structures showed that all 14

cellulases belonged to GH family 5 (GHF5). GHF5 activities

have been identified as the most abundant cellulases from

both the cultured ruminal bacterial species, such as F. suc-

cinogenes containing more than 50% of GHF5 in the total

cellulases identified from its genome (Krause et al. 2003),

and the uncultured ruminal bacteria (Ferrer et al. 2005;

Palackal et al. 2007). The majority of metagenomic cellulas-

es cloned and characterized from other environments also

belonged to GHF5 (Healy et al. 1995; Voget et al. 2006;

Feng et al. 2007). Large-scale metagenomic sequencing of

hindgut bacteria of a wood-feeding higher termite revealed

that GHF5 was predominant in all identified GH except

GH family 3 (Warnecke et al. 2007). These data indicate

that the GHF5 is the largest member of GH in nature;

therefore, GHF5 cellulases were the most cloned celluases

in this study.

In addition to genes encoding glycosyl hydrolases, other

genomic properties were also identified from ruminal bac-

teria, reflecting their adaptation to the rumen ecosystem.

Cosmid DC2 harboured a gene cluster or operon involved

in utilization of polysaccharides nutrients including genes

encoding proteins for nutrient binding and transportation

(Table S1), and different family GH responsible for poly-

saccharide degradation. Another GH gene cluster was

found in clone DC23. Similar GH gene operon was also

reported in cultured ruminal bacteria in Prevotella bryantii

(Gardner et al. 1997), which encoded putative polygalactu-

ronase, b-1,4-endoglucanase, and mannanase. As different

family GH genes were found in those gene clusters, they

may be responsible for the degradation of different polysac-

charides in the complex nutrients in the rumen. The plant

materials indigested by animals were very complex, with

cell walls containing polysaccharides such as cellulose,

hemicellulose and pectin, and hemicellulose consists of

xylan and glucomannan (de Vries and Visser 2001). Hydro-

lysis of the combination of polysaccharides requires many

enzymes expressed from different family GH genes. These

genes organize in clusters in the genomes of the ruminal

bacteria and their proteins function synergistically, so the

energy expenditure can be minimized and optimal rates for

degradation of the plant components can be ensured. This

may have resulted from the natural selection for bacterial

species with such genetic advantages to offer the best abili-

ties to digest complex nutrients in the rumens.

The ruminal pH fluctuates diurnally between pH 5 and

7, and decreases to pH 6 or lower during the production

of fermented acid after feeding (Yang et al. 2001; Micha-

let-Doreau et al. 2002). Under such pH condition, the

growth of cellulolytic bacteria is inhibited and fibre degra-

dation is decreased (Russell and Wilson 1996). Acidic

cellulases are then needed to guarantee the degradation of

fibre particles in the low pH condition at an appropriate

rate. Ten of the thirteen cellulases characterized in this

study showed optimal pH range of 5–7 and seven of them

were most active under acidic conditions (pH 4–5Æ5).

These cloned enzymes may play roles in the cellulose

degradation at low pH condition in the buffalo rumen.

In the detailed analyses with recombinant enzyme C67-1

expressed from one highly active clone, the metal ions

could not enhance the enzyme activity and EDTA showed

no significant effect on the enzyme activity, indicating that

the activities of C67-1 were not dependent on metal ions.

This property was distinguished from most cellulases that

activities can be induced by one or more metal ions (Voget

et al. 2006; Feng et al. 2007). However, interestingly, activ-

ity of C67-1 was highly stimulated by DTT, a disulfide

reducing agent that may reduce disulfide bonds within

enzyme structures. DTT has been shown to be required for

activity of cellulosome from the anaerobic bacterium Clos-

tridium thermocellum, but had no effect on the Trichoderma

enzyme (Johnson et al. 1982; Johnson and Demain 1984).

These results indicated that C67-1, like the anaerobic clos-

tridial cellulase, unlike the enzyme from aerobic fungi, con-

tains essential sulfhydryl groups that are important for its

activity. As the rumen is an anaerobic environment, adding

DTT or other reducing agent in the assay may help attain

higher enzymatic activities of those hydrolases derived from

the rumen. Although C67-1 was an acidic cellulase with

optimum pH at 4Æ5, it was very stable both at acidic

(pH ‡ 3Æ5) and alkaline conditions up to pH 10Æ5. This

property was also different from the most acidic enzymes

that were not stable in alkaline condition (Parry et al. 2002;

Gao et al. 2008). The predicted isoelectric point of C67-1 is

8Æ535 determined by using software Editseq (DNASTAR).

This is consistent with the result of the native PAGE of the

crude enzyme extract of subclone C67E4, the active band of

C67-1 located between the stacking gel and the separating

gel (Fig. 2), indicating that C67-1 had a higher isoelectric

point. The stability of C67-1 at alkaline conditions may be

because of its uniquely high isoelectric point; thus, it can

maintain its correct structure in the alkaline environment.

In conclusion, our study identified novel cellulases

with distinct properties, and provided evidence for the

diversity and function of cellulase and mechanisms of cel-

lulose hydrolysis in the rumen. We believe that these

findings may at some point of time offer potential appli-

cations in the industry.

Cellulases from rumen metagenomes C.-J. Duan et al.

254 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256

ª 2009 The Authors

Page 11

Acknowledgements

This work was supported by the National Natural Science

Foundation of China (30560003), Programme for

New Century Excellent Talents in University of China

(NCET-05-0752), Hi-tech Research and Development

Programme of China (863 Programme 2007AA021307,

2004AA214140) and Director’s Fund of Guangxi Key

Laboratory of Subtropical Bioresources Conservation and

Utilization (Zhuji-06-08).

References

Allen, M.S. (1997) Relationship between fermentation acid

production in the rumen and the requirement for physi-

cally effective fiber. J Dairy Sci 80, 1447–1462.

An, D., Dong, X. and Dong, Z. (2005) Prokaryote diversity in

the rumen of yak (Bos grunniens) and Jinnan cattle (Bos

taurus) estimated by 16S rDNA homology analyses. Anaer-

obe 11, 207–215.

Bera, C., Broussolle, V., Forano, E. and Gaudet, G. (1996)

Gene sequence analysis and properties of EGC, a family

E(9) endoglucanase from Fibrobacter succinogenes BL2.

FEMS Microbiol Lett 136, 79–84.

Bradford, M.M. (1976) A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing

the principle of protein-dye binding. Anal Biochem 72,

248–254.

Calsamiglia, S., Ferret, A. and Devant, M. (2002) Effects of pH

and pH fluctuations on microbial fermentation and nutri-

ent flow from a dual-flow continuous culture system.

J Dairy Sci 85, 574–579.

Cerrato-Sanchez, M., Calsamiglia, S. and Ferret, A. (2008)

Effect of the magnitude of the decrease of rumen pH on

rumen fermentation in a dual-flow continuous culture

system. J Anim Sci 86, 378–383.

Cheng, K.J., Forsberg, C.W., Minato, H. and Costerton, J.W.

(1991) Microbial ecology and physiology of feed degrada-

tion within the rumen. In Physiological Aspects of Digestion

and Metabolism in Ruminants ed. Tsuda, T., Sasaki, Y. and

Kawashima, R. pp. 595–624. Ontario, Canada: Academic

Press.

Daniel, R. (2005) The metagenomics of soil. Nat Rev Microbiol

3, 470–478.

de Vries, R.P. and Visser, J. (2001) Aspergillus enzymes

involved in degradation of plant cell wall polysaccharides.

Microbiol Mol Biol Rev 65, 497–522.

Eckert, K. and Schneider, E. (2003) A thermoacidophilic endo-

glucanase (CelB) from Alicyclobacillus acidocaldarius dis-

plays high sequence similarity to arabinofuranosidases

belonging to family 51 of glycoside hydrolases. Eur J Bio-

chem 270, 3593–3602.

Feng, Y., Duan, C.J., Pang, H., Mo, X.C., Wu, C.F., Yu, Y.,

Hu, Y.L., Wei, J., et al. (2007) Cloning and identification

of novel cellulase genes from uncultured microorganisms

in rabbit cecum and characterization of the expressed

cellulases. Appl Microbiol Biotechnol 75, 319–328.

Ferrer, M., Golyshina, O.V., Chernikova, T.N., Khachane,

A.N., Reyes-Duarte, D., Santos, V.A., Strompl, C., Elbor-

ough, K., et al. (2005) Novel hydrolase diversity retrieved

from a metagenome library of bovine rumen microflora.

Environ Microbiol 7, 1996–2010.

Gao, J., Weng, H., Zhu, D., Yuan, M., Guan, F. and Xi, Y.

(2008) Production and characterization of cellulolytic

enzymes from the thermoacidophilic fungal Aspergillus

terreus M11 under solid-state cultivation of corn stover.

Bioresour Technol 99, 7623–7629.

Gardner, R.G., Wells, J.E., Fields, M.W., Wilson, D.B. and

Russell, J.B. (1997) A Prevotella ruminicola B(1)4 operon

encoding extracellular polysaccharide hydrolases. Curr

Microbiol 35, 274–277.

Healy, F.G., Ray, R.M., Aldrich, H.C., Wilkie, A.C., Ingram,

L.O. and Shanmugam, K.T. (1995) Direct isolation of

functional genes encoding cellulases from the microbial

consortia in a thermophilic, anaerobic digester maintained

on lignocellulose. Appl Microbiol Biotechnol 43, 667–674.

Inoue, T., Moriya, S., Ohkuma, M. and Kudo, T. (2005)

Molecular cloning and characterization of a cellulase gene

from a symbiotic protist of the lower termite, Coptotermes

formosanus. Gene 349, 67–75.

Johnson, E.A. and Demain, A.L. (1984) Probable involvement

of sulfhydryl groups and a metal as essential components

of the cellulase of Clostridium thermocellum. Arch Microbiol

137, 135–138.

Johnson, E.A., Sakajoh, M., Halliwell, G., Madia, A. and

Demain, A.L. (1982) Saccharification of complex cellulosic

substrates by the cellulase system from Clostridium thermo-

cellum. Appl Environ Microbiol 43, 1125–1132.

Kim, S.J., Lee, C.M., Han, B.R., Kim, M.Y., Yeo, Y.S., Yoon,

S.H., Koo, B.S. and Jun, H.K. (2008) Characterization of a

gene encoding cellulase from uncultured soil bacteria.

FEMS Microbiol Lett 282, 44–51.

Koike, S., Yoshitani, S., Kobayashi, Y. and Tanaka, K. (2003)

Phylogenetic analysis of fiber-associated rumen bacterial

community and PCR detection of uncultured bacteria.

FEMS Microbiol Lett 229, 23–30.

Krause, K.M. and Combs, D.K. (2003) Effects of forage particle

size, forage source, and grain fermentability on perfor-

mance and ruminal pH in midlactation cows. J Dairy Sci

86, 1382–1397.

Krause, D.O., Denman, S.E., Mackie, R.I., Morrison, M., Rae,

A.L., Attwood, G.T. and McSweeney, C.S. (2003)

Opportunities to improve fiber degradation in the rumen:

microbiology, ecology, and genomics. FEMS Microbiol Rev

27, 663–693.

Laemmli, U.K. (1970) Cleavage of structural proteins during

the assembly of the head of bacteriophage T4. Nature 227,

680–685.

Lorenz, P. and Eck, J. (2005) Metagenomics and industrial

applications. Nat Rev Microbiol 3, 510–516.

C.-J. Duan et al. Cellulases from rumen metagenomes

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256 255

Page 12

Lynd, L.R., Weimer, P.J., van Zyl, W.H. and Pretorius, I.S.

(2002) Microbial cellulose utilization: fundamentals and

biotechnology. Microbiol Mol Biol Rev 66, 506–577.

Malburg, L.M. Jr., Iyo, A.H. and Forsberg, C.W. (1996) A

novel family 9 endoglucanase gene (celD), whose product

cleaves substrates mainly to glucose, and its adjacent

upstream homolog (celE) from Fibrobacter succinogenes

S85. Appl Environ Microbiol 62, 898–906.

Malburg, S.R., Malburg, L.M., Jr., Liu, T., Iyo, A.H. and

Forsberg, C.W. (1997) Catalytic properties of the cellulose-

binding endoglucanase F from Fibrobacter succinogenes

S85. Appl Environ Microbiol 63, 2449–2453.

McGavin, M. and Forsberg, C.W. (1988) Isolation and charac-

terization of endoglucanases 1 and 2 from Bacteroides

succinogenes S85. J Bacteriol 170, 2914–2922.

Michalet-Doreau, B., Fernandez, I. and Fonty, G. (2002) A

comparison of enzymatic and molecular approaches to

characterize the cellulolytic microbial ecosystems of the

rumen and the cecum. J Anim Sci 80, 790–796.

Michalet-Doreau, B., Fernandez, I., Peyron, C., Millet, L. and

Fonty, G. (2001) Fibrolytic activities and cellulolytic bacte-

rial community structure in the solid and liquid phases of

rumen contents. Reprod Nutr Dev 41, 187–194.

Miller, G.L. (1959) Use of dinitrosalicyclic acid reagent

for determination of reducing sugar. Anal Chem 31,

426–428.

Miller, D.N., Bryant, J.E., Madsen, E.L. and Ghiorse, W.C.

(1999) Evaluation and optimization of DNA extraction

and purification procedures for soil and sediment samples.

Appl Environ Microbiol 65, 4715–4724.

Nocek, J.E., Allman, J.G. and Kautz, W.P. (2002) Evaluation of

an indwelling ruminal probe methodology and effect of

grain level on diurnal pH variation in dairy cattle. J Dairy

Sci 85, 422–428.

Odoux, E., Escoute, J., Verdeil, J.L. and Brillouet, J.M. (2003)

Localization of beta-d-glucosidase activity and

glucovanillin in vanilla bean (Vanilla planifolia Andrews).

Ann Bot (Lond) 92, 437–444.

Palackal, N., Lyon, C.S., Zaidi, S., Luginbuhl, P., Dupree, P.,

Goubet, F., Macomber, J.L. and Short, J.M. (2007) A

multifunctional hybrid glycosyl hydrolase discovered in an

uncultured microbial consortium from ruminant gut. Appl

Microbiol Biotechnol 74, 113–124.

Parry, N.J., Beever, D.E., Owen, E., Nerinckx, W., Claeyssens,

M., Van Beeumen, J. and Bhat, M.K. (2002) Biochemical

characterization and mode of action of a thermostable

endoglucanase purified from Thermoascus aurantiacus.

Arch Biochem Biophys 404, 243–253.

Rees, H.C., Grant, S., Jones, B., Grant, W.D. and Heaphy, S.

(2003) Detecting cellulase and esterase enzyme activities

encoded by novel genes present in environmental DNA

libraries. Extremophiles 7, 415–421.

Russell, J.B. and Wilson, D.B. (1996) Why are ruminal cellulo-

lytic bacteria unable to digest cellulose at low pH? J Dairy

Sci 79, 1503–1509.

Sambrook, J. and Russell, D.W. (2001) Molecular Cloning: A

Laboratory Manual, 3rd edn. Cold Spring Harbor, New

York: Cold Spring Harbor Laboratory Press.

Tajima, K., Aminov, R.I., Nagamine, T., Ogata, K., Nakamura,

M., Matsui, H. and Benno, Y. (1999) Rumen bacterial

diversity as determined by sequence analysis of 16S rDNA

libraries. FEMS Microbiol Ecol 29, 159–169.

Voget, S., Steele, H.L. and Streit, W.R. (2006) Characterization

of a metagenome-derived halotolerant cellulase. J Biotech-

nol 126, 26–36.

Voget, S., Leggewie, C., Uesbeck, A., Raasch, C., Jaeger, K.E.

and Streit, W.R. (2003) Prospecting for novel biocatalysts

in a soil metagenome. Appl Environ Microbiol 69, 6235–

6242.

Wales, W.J., Kolver, E.S., Thorne, P.L. and Egan, A.R. (2004)

Diurnal variation in ruminal pH on the digestibility of

highly digestible perennial ryegrass during continuous

culture fermentation. J Dairy Sci 87, 1864–1871.

Warnecke, F., Luginbuhl, P., Ivanova, N., Ghassemian, M.,

Richardson, T.H., Stege, J.T., Cayouette, M., McHardy,

A.C., et al. (2007) Metagenomic and functional analysis of

hindgut microbiota of a wood-feeding higher termite.

Nature 450, 560–565.

Wood, T.M. (1971) The cellulase of Fusarium solani. Purifica-

tion and specificity of the b-(1-4)-glucanase and the

b-d-glucosidase components. Biochem J 121, 353–362.

Yang, W.Z., Beauchemin, K.A. and Rode, L.M. (2001) Effects

of grain processing, forage to concentrate ratio, and forage

particle size on rumen pH and digestion by dairy cows.

J Dairy Sci 84, 2203–2216.

Supporting Information

Additional Supporting Information may be found in the

online version of this article:

Figure S1 Multiple sequence alignment of the catalytic

domains of the cloned cellulases in this study with other

members of family 5 glycoside hydrolases with known

three-dimensional structures collected from CAZy data-

base.

Figure S2 Effects of pH and temperatures on the

stability of the recombinant C67-1.

Table S1 Annotation of predicted open reading frames

based on DNA sequences obtained from subclones or

source clones expressing cellulase activities from the

buffalo rumen metagenomes.

Table S2 Effects of metal ions, chelating agent, surfac-

tants and reducing reagent on the enzyme activity of C67-1.

Please note: Wiley-Blackwell are not responsible for the

content or functionality of any supporting materials sup-

plied by the authors. Any queries (other than missing

material) should be directed to the corresponding author

for the article.

Cellulases from rumen metagenomes C.-J. Duan et al.

256 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 107 (2009) 245–256

ª 2009 The Authors