Group Title: Biotechnology for Biofuels 2009, 2:25
Title: Parallel metatranscriptome analyses of host and symbiont gene expression in the gut of the termite Reticulitermes flavipes
Full Citation
Permanent Link:
 Material Information
Title: Parallel metatranscriptome analyses of host and symbiont gene expression in the gut of the termite Reticulitermes flavipes
Series Title: Biotechnology for Biofuels 2009, 2:25
Physical Description: Archival
Creator: Tartar A
Wheeler MM
Zhou X
Coy MR
Boucias DG
Scharf ME
Publication Date: 40101
 Record Information
Bibliographic ID: UF00100252
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access:


This item has the following downloads:

parallel ( PDF )

Full Text

Biotechnology for Biofuels Bioed Central


Parallel metatranscriptome analyses of host and symbiont gene
expression in the gut of the termite Reticulitermes flavipes
Aurelien Tartar1'2, Marsha M Wheeler',3, Xuguo Zhou1'4, Monique R Coy',
Drion G Boucias' and Michael E Scharf*1

Address: 'Department of Entomology and Nematology, University of Florida, Gainesville, FL, USA, 2Division of Math, Science and Technology,
Nova Southeastern University, Fort Lauderdale, FL, USA, 3Current address: Department of Entomology, University of Illinois, Champaign-Urbana,
IL, USA and 4Current address : Department of Entomology, University of Kentucky, Lexington, KY, USA
Email: Aurelien Tartar; Marsha M Wheeler wheelel; Xuguo Zhou;
Monique R Coy; Drion G Boucias; Michael E Scharf*
* Corresponding author

Published: 15 October 2009 Received: 31 May 2009
Biotechnology for Biofuels 2009, 2:25 doi:10.1 186/1754-6834-2-25 Accepted: 15 October 2009
This article is available from:
2009 Tartar et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Background: Termite lignocellulose digestion is achieved through a collaboration of host plus prokaryotic and
eukaryotic symbionts. In the present work, we took a combined host and symbiont metatranscriptomic approach for
investigating the digestive contributions of host and symbiont in the lower termite Reticulitermes flavipes. Our approach
consisted of parallel high-throughput sequencing from (i) a host gut cDNA library and (ii) a hindgut symbiont cDNA
library. Subsequently, we undertook functional analyses of newly identified phenoloxidases with potential importance as
pretreatment enzymes in industrial lignocellulose processing.
Results: Over 10,000 expressed sequence tags (ESTs) were sequenced from the 2 libraries that aligned into 6,555
putative transcripts, including 171 putative lignocellulase genes. Sequence analyses provided insights in two areas. First,
a non-overlapping complement of host and symbiont prokaryoticc plus protist) glycohydrolase gene families known to
participate in cellulose, hemicellulose, alpha carbohydrate, and chitin degradation were identified. Of these, cellulases are
contributed by host plus symbiont genomes, whereas hemicellulases are contributed exclusively by symbiont genomes.
Second, a diverse complement of previously unknown genes that encode proteins with homology to lignase, antioxidant,
and detoxification enzymes were identified exclusively from the host library (laccase, catalase, peroxidase, superoxide
dismutase, carboxylesterase, cytochrome P450). Subsequently, functional analyses of phenoloxidase activity provided
results that were strongly consistent with patterns of laccase gene expression. In particular, phenoloxidase activity and
laccase gene expression are mostly restricted to symbiont-free foregut plus salivary gland tissues, and phenoloxidase
activity is inducible by lignin feeding.
Conclusion: To our knowledge, this is the first time that a dual host-symbiont transcriptome sequencing effort has been
conducted in a single termite species. This sequence database represents an important new genomic resource for use in
further studies of collaborative host-symbiont termite digestion, as well as development of coevolved host and symbiont-
derived biocatalysts for use in industrial biomass-to-bioethanol applications. Additionally, this study demonstrates that:
(i) phenoloxidase activities are prominent in the R. flavipes gut and are not symbiont derived, (ii) expands the known
number of host and symbiont glycosyl hydrolase families in Reticulitermes, and (iii) supports previous models of lignin
degradation and host-symbiont collaboration in cellulose/hemicellulose digestion in the termite gut. All sequences in this
paper are available publicly with the accession numbers FL634956-FL640828 (Termite Gut library) and FL641015-
FL645753 (Symbiont library).

Page 1 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

Lignocellulose, the principal constituent of plant bio-
mass, is currently targeted as a primary feedstock for the
production ofbiofuels [1]. However, the extant industrial
processes used for fuel conversion suffers from a lack of
efficiency that has been related in part to the extreme
recalcitrance of plant-produced lignocellulose [2,3]. Bio-
fuel production from plants requires the successful depo-
lymerization of the large carbohydrate polymers cellulose
and hemicellulose. In lignocellulose these polymers are
not directly accessible for chemical or enzymatic treat-
ments but are combined with lignins and other molecules
to form a complex, and highly resistant, three-dimen-
sional molecular structure. At present, the inability to effi-
ciently extract simple, utilizable sugars from
lignocellulose through depolymerization reactions is a
significant limiting factor for the bioethanol industry [ 1].
One strategy for identification of novel enzymes to
improve this process is through the mining of genomes of
lignocellulolytic organisms, both prokaryotic and eukary-
otic [4,5].

Termites (Insecta, Isoptera) are ubiquitous arthropods
that efficiently digest lignocellulose and flourish on this
seemingly nutritionally poor diet. The feeding activity of
termites, combined with their preference for wood, is
known to play a critical role in ecosystem nutrient recy-
cling, and to cause significant economic damage to
human-built structures. Recently, the ability of termites to
convert recalcitrant plant biomass into a useable energy
source has attracted much interest due to the numerous
potential applications in biofuel production [6-8]. Ligno-
cellulose digestion in termites is intimately correlated
with the presence of a highly specific flora of symbiotic
microbes [9,10]. Many of these microbes, essential for the
survival of the termite hosts, reside in a modified hindgut
or fermentation sac [11,12].

Current termite lignocellulose digestion models consider
both host and symbiont inputs (for example, [5,13,14]).
In this respect, recent studies conducted on the lower ter-
mites Coptotermes formosanus and Reticulitermes flavipes
have demonstrated that carbohydrate-active enzymes of
both termite and symbiont origins are produced simulta-
neously but in different regions of the gut [15-18]. For
example, termites and symbiotic protests produce glycosyl
hydrolases that belong to two separate families (GHFs)
(respectively): GHF9 endoglucanases and GHF7 exogluca-
nases/cellobiohydrolases [16,17]. Aside from semantic
disagreements [19], there is general agreement that
sequential depolymerization reactions involve insect and
symbiont enzymes with different and possibly comple-
mentary properties, allowing for efficient lignocellulose
digestion [6,16,17]. The respective contributions and
requirements of each enzyme, as well as the role played by

cellulases produced by prokaryotic symbionts, have yet to
be fully understood.

In efforts to define termite lignocellulolytic processes, sev-
eral previous studies have completed metagenome, tran-
scriptome, or metatranscriptome sequencing projects on
various termite digestive organs or symbiotic gut fauna.
Two studies have focused specifically on symbionts sam-
pled from the hindgut lumen. First, a metatranscriptomic
project to sequence cDNA clones representing expressed
genes of eukaryotic protistann) symbionts from the lower
Reticulitermes speratus was reported by Todaka et al. [20].
This study produced 910 expressed sequence tag (EST)
sequences that aligned into 580 tentative genes. The R.
speratus study revealed many diverse cellulase and hemi-
cellulase sequences, in particular, an unexpectedly large
complement of GHF7 cellulases putatively involved in
microcrystalline cellulose degradation. Second, a metage-
nomic project to sequence prokaryotic gut symbiont
genomes of a higher Nasutitermes termite (species
unknown) produced over 100 million bases of DNA
sequence [21] (see also [7,8,22]). This study revealed
genes containing more than 700 unique glycoside hydro-
lase catalytic domains from 45 different carbohydrate
active gene families, as well as other genes relating to
numerous aspects of microbial life in the gut microenvi-
ronment, such as nitrogen fixation and H2 production.
While both of these first two studies identified rich com-
plements of cellulases, hemicellulases, and other carbohy-
drolases, they revealed no evidence of lignases.

Still other large-scale sequencing projects have focused on
endosymbionts and ectosymbionts; specifically, bacterial
endosymbionts of protist gut symbionts and a termite-
cultivated symbiotic fungus. Two studies sequenced the
entire genomes of bacterial endosymbionts of hindgut
protests [23,24]. These bacteria included 'RS-D17', a bac-
terium from the protist gut symbiont Trichonympha agilis
of the termite R. speratus; and 'CfPtl-2', a Bacteroidales
endosymbiont of the protist gut symbiont Pseudotricho-
nympha grassii of the termite C. formosanus. Both of these
bacterial phylotypes are as yet unculturable and occur
only within cellulolytic protests of termite guts. Interest-
ingly, genome sequencing revealed that these bacteria are
highly adept at nitrogen fixation, monosaccharide metab-
olism and amino acid production, but encode no ligno-
cellulase genes and thus, likely play no roles in
lignocellulose degradation. The termite fungal cultivar
work was performed by sequencing cDNA ESTs from Ter-
mitomyces, a fungal ectosymbiont obtained from mounds
of the higher termite Macrotermes gilvus [25]. This work
identified 1,582 tentatively unique gene sequences
including well represented pectinases, hemicellulases and
cellulases, as well as 2 laccases putatively involved in
lignin degradation. Thus, while bacterial endosymbionts

Page 2 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

of protests confer no lignocellulose digestion capabilities,
cultivated fungal symbionts found in some higher termite
nests clearly do.

As exemplified by the above-summarized work, termite
gut symbionts play well established roles in cellulose and
hemicellulose digestion; however, there is no sequence-
based evidence to date supporting that gut symbionts are
capable of lignin degradation. Furthermore, it is also well
established that both lower and higher termites actively
secrete endogenous cellulases that play important roles in
cellulose digestion, such as endoglucanases and 3 glucosi-
dases [13-18,26-30]. In this respect, a recent study consid-
ered host digestive tissues of the lower termite
Hodotermopsis sjostedti [31]. This research employed five
cDNA libraries from various host termite tissues (salivary
gland, foregut, midgut, hindgut, and body). A total of
3,548 ESTs were obtained that aligned into 2,366 contig-
uous sequences. Many genes relating to cellulose and
hemicellulose digestion were identified (especially in
symbiont-free salivary gland tissue). While these various
host and symbiont sequencing efforts to date are both
important and provide fascinating information, to our
knowledge, no large-scale sequencing studies to date have
considered both host and symbiont concurrently from the
same termite species.

The termite 'digestome' is defined as the pool of host and
symbiont genes that collaborate to achieve high efficiency
lignocellulose digestion in the termite gut [5]. Industrial
lignocellulose 'pretreatment', alternatively, is a critical
step in bioethanol production; it is an early stage, and cur-
rently very costly industrial process whereby sugars con-
tained in cellulose and hemicellulose are separated from
lignin by non-enzymatic processes [1,32,33]. In this
study, our objectives were to: (1) develop a deeper under-
standing of the digestome of the lower termite R. flavipes,
and (2) functionally investigate genes with potential rele-
vance to industrial lignocellulose pretreatment. For this
purpose, we took a dual host plus symbiont metatran-
scriptome approach that included large-scale transcrip-
tome sequencing from two separate cDNA libraries; one
representing host gut cells and the other hindgut micro-
bial symbionts. Subsequent functional studies of candi-
date pretreatment enzymes involved efforts to correlate
phenoloxidase (laccase and catalase) enzyme activities
and gene expression in the termite gut. Our findings
reveal: (i) previously unidentified patterns of host and
symbiont production of cellulases, hemicellulases, a car-
bohydrolases, chitinases, and potentially lignases; and (ii)
novel host-derived laccase gene products that have poten-
tial relevance to industrial lignocellulose pretreatment.
The host-symbiont EST database described here (and pre-
liminarily by [5]) represents a significant new genomic
resource that will enable robust functional analyses of

coevolved host and symbiont digestive arsenals from the
same termite species.

EST sequencing overview
EST sequencing results from host and symbiont cDNA
libraries are summarized in Figure la,c. A total of 10,610
high-quality ESTs were generated from the 2 libraries. The
5,871 ESTs produced from the host library assembled into
875 contigs and 2,169 singlets to produce 3,044 puta-
tively unique transcripts. In the symbiont EST database, a
contig assembly performed using the same parameters
identified 358 contigs and 3,153 singlets (3,511 putative
transcripts). Thus, the total number of genes identified
from the 2 libraries is 6,555. Similarity searches showed
that a high proportion of transcripts do not produce any
significant match in the NCBI nr database (38% and 48%
of host and symbiont libraries, respectively), highlighting
the potential of both cDNA libraries for novel gene dis-
covery (Figure la,c). The symbiont library contains a
higher fraction of uncharacterized sequences (48%) than
the termite gut library (38%), reflecting the uniqueness of
the symbiont transcript pool and, likely, their unique
physiology and metabolism. The annotated transcripts
that exhibited significant similarities to known proteins

(A) Gut Tissue Library

ror 9

for ge
High Quality Reads: 5871 identi
Avg. Read Length: 863nt (n18
Redundancy: 52% gene
Contigs: 075
Singlets: 2169
Total genes: 3044
enes w significant identity: 62
Gens wi unknown identity: 38

(C) Symbiont Library (D)
Best BLASTx match
for genes w/sig.
High Quality Reads: 4739 identity
Avg. Read Length: 857nt (n-1826
Redundancy: 26% g"nes)
Contbgs: 318
Total genes: 3511
Genes wi significant identity: 52%
Genes wi unknown identity: 48%

ungi (1%)
!I\ Pkyote (1%
Vertebrate (7%)
-henomes (5%)
Nigig (8% )
~Plant (1%

ITermite bletaVomes (5%)

71_11 e~(7%)
V led (7.)
P0..,kyoe ()Is%)

Figure I
Expressed sequence tag (EST) sequencing results. (a)
and (c) present sequencing summaries for the host (termite
gut tissue) and symbiont cDNA libraries, respectively. (b)
and (d) are the taxonomic distributions of top BLASTX hits
for putative transcripts from the termite gut and symbiont
libraries (respectively). Hits producing an E-value < 105 were
considered significant. The charts illustrate that the majority
of transcripts sequenced from the termite gut library were
homologous to insect sequences, whereas the majority of
transcripts obtained from the symbiont library were homolo-
gous to genes previously sequenced from unicellular eukary-

Page 3 of 19
(page number not for citation purposes)

BLASTx match
nes Wsig.

were similar from each library, both in terms of numbers
and average sizes of putative genes (Figure la,c). Figure
lb,d depict the classification of annotated transcripts
based on the organism associated with the top BLASTX
hit. As expected, although a minority of microbial
sequences is present, >85% of the sequences generated
from the termite gut tissue are homologous to known
genes previously sequenced from insects or from other
invertebrates (Figure Ib). Likewise, while the symbiont
library contains a fraction of transcripts matching insect
and invertebrate sequences, it contains >85% microbe-
homologous sequences (Figure Id).

Phylogenetic analyses to investigate taxonomic origins
Because of apparent library cross contamination, we con-
sidered both phylogenetic signal and the library of origin
when making taxonomic classifications. In agreement
with BLAST analyses, phylogenetic analyses supported
that, independent of their cDNA library of origin, a
number of GHF family transcripts (for example, GHF1,
GHF9 and GHF16) are endogenous to the R. flavipes
genome (not shown). Likewise, symbiont sequences as
suggested by BLAST searching (for example, GHF7, GHF5
and GHF45) all cluster with genes previously attributed to
termite symbionts (not shown). As a representative exam-
ple, Figure 2 depicts the phylogenies of another apparent
symbiont-specific family, GHF11. In agreement with pre-
vious gut metagenomics research [21], the present analy-
sis shows that irrespective of library all GHF11 sequences
generated throughout our study cluster with termite sym-
biont GHF11 proteins with strong bootstrap support (Fig-
ure 2). Additionally, no GHF11 transcripts from either
library clustered with fungal or animal sequences, further
implying that all GHF11 sequences identified here origi-
nate from symbiotic genomes.

Identification of genes with carbohydrate-active functions
The termite and symbiont library EST databases were
screened for carbohydrate-active enzyme/protein coding
sequences. The results of these searches are summarized in
Figures 3 and 4, and are presented in more detail in Addi-
tional files 1 and 2. All accession numbers for carbohy-
drate-active enzyme sequences are provided in Additional
files 3 and 4. Five categories of carbohydrate-active
enzymes/proteins are summarized below.

Glycosyl hydrolase enzymes
Figure 3a presents GHFs. In all, 27 total families were rep-
resented in the host and symbiont sequence pools. The 27
families segregate into the 4 major functional categories of
cellulases (GHFs 1,3,7,9,16,30 and 45), hemicellulases
(GHFs 2,5,8,10,11,26,42,43 and 53), chitinases (GHFs
18,20,27 and 85), and a carbohydrolases (GHFs
37,48,47,76,77 and 92). Figure 3b summarizes putative
taxonomic origins of each family by functional group.

TG21 D5
Mooioo TS39-02
LJSP3 Redicufite~res speralus
USPi Refictuermes spevalus
USP2 Reticujjtermes speartus
USPI Hodowemoipsrs sestedh
USP3 Hodolerrmpsis sibstedhi
USP2 Hodclermopsis siobstedh
USP4 H oddermopsis sjiestedi
USP4 Relfculilermes speratus
Phaedon cochleadae
Aspergillus nidulans
Neurospora crassa
E Ve1)CllwM dahbiae
Trichoderma vinde
Thermomyces laniginross
Streptomyces coelio/or
Thermobdida fusca
Bacihls subbifs
- 0.05 changes

Termite Symbionts



Figure 2
Protein neighbor-joining phylogeny of glycoside
hydrolase family I I (GHF I I). The tree was rooted with
bacterial GHFI I enzyme sequences and showed that all
GHF I I transcripts generated during this study clustered with
GHFI I genes previously sequenced from termite symbionts
(uncultured symbiotic protist (USP)). The transcripts identi-
fied in the termite gut (TG) library are shown in blue,
whereas the transcript sequences that originated from the
termite symbiont (TS) library are shown in red. Numbers
(100/100) above the node represent bootstrap support
(1000 replicates). For clarity purposes only relevant boot-
strap values are indicated.

This analysis suggests that cellulose digestion is enabled
by a three-way collaboration of host plus protist plus
prokaryotic symbionts, hemicellulose digestion is accom-
plished completely by symbionts protistt plus prokaryo-
tic), and both chitin and a carbohydrate digestion are
achieved by host plus prokaryotic symbionts.

Glycosyl transferase enzymes
A summary of identified glycosyl transferase family (GTF)
coding sequences is provided in Figure 4a. In all, 10 total
subfamilies were represented in the host and symbiont
libraries, with 79% and 21% of total unigene sequences
being putatively of termite and prokaryotic origins,
respectively (Figure 4e). The represented prokaryotic GTFs
(families 2 and 8) both encode predicted functions in cell

Page 4 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

Taxonomic Predicted
rhinin ;_ laCth*f

No. Genes
I i r j

I ik.ra


n=1A n=lR



Cellulase Hemi- Chitinase Alpha-
cellulase carbohydrolase

I Prokaryote
Li Protist
H Termite

Figure 3
Distributions of expressed sequence tags (ESTs) encoding glycosyl hydrolase family (GHF) proteins generated
from the host gut tissue and hindgut symbiont libraries. (a) A summary of GHF family members with enzyme identities
and putative taxonomic origins that were determined based on library of origin and database homology. Shaded rows depict
GHF families with representatives from both termite and putative prokaryotic symbionts. CE = cellulase, H = hemicellulase,
CH = chitinase, A = a carbohydrolase. (b) Bar graph summarizing the taxonomic distributions of cellulase, hemicellulase, chiti-
nase and a carbohydrolase genes among prokaryotic, protist and host genomes. See Additional file I for gene by gene summa-
ries and Additional file 3 for Genbank accession numbers.

Page 5 of 19
(page number not for citation purposes)

1 beta-glucosidase termite CE 2
2 eta-galactosiase pro e H 1 1
3 beta-gluo~sidase prokaryote CE 2 10
5 endo-xylosidase protist H 2 11
7 exo-glucanase protest CE 10 36
8 lendo-/exo-xyosidase prokaryote H 1 1
9A endD-alucanmae larmite CE i
98 p rokaryote CE i
10 endo-xylanase H 1 2
11 endo-xvlanase protist H 5 10
13 alpha-amyase termite A 10
16 endo-1,3-glucanase termite CE 3 3
18A chlnase ftrme e CH 8
186 __________ prokaryote CH 5
20A 'eta-hexosaminidase lsrmite CH 4
20B _prokyde CH 3
26 beta-1.4-mannanase prokaryote H 1 6
27 alpha-Wacetylgalaclosaminidase termite CH 1
30A beta-alucosiyIeranidase bermite CE 3
30B prkaryoe CE 1
37 Irehalase termite A 1
38 alpha-mannosidase termite A 2
42 beta-gaaactosidase prokaryote H 1
43 beta-xylosidase prokaryote H 1
45 endo-glucanase Probst CE 1 4
47 alpha mannosidase termite A 1
53 endo-arabino-aalactosidase prokarvote H 1 1
76 alpha-1 6-mannanaso prokaryote A 1
77 alpha-amylorraltase prokaryo(e A I____
85 endo-beta N acetylglucosaminidase termite CH 1
92 alpha 1 2-mannosidase prokaryote A 1 1


n = 77 n = 45

Biotechnology for Biofuels 2009, 2:25

rUC Mn' Cnfioa IrlanIi

Biotechnology for Biofuels 2009, 2:25

Ia G Iyc Trwsfamess(GTF)

'r *

Bi :Carbohydrate Esteas (CE)

I -1


1.4tin Day4Eltted fnetrtrese

No. Genes No. Genes
Host Symbiont

No. Genes No. Genes
Host Symbiont
I tkhmu I Ivru

.1 s huie o,,lr~n.1.:aje, mi *~ahrtr o N a..4r b~ri I.v:as.1m,,,e
tJA o'id .rdan sutnst ee iLEi b r vie rkib~irla E

Carbohyd r te ining Modules (CBM)


No. Genes No. Genes
Host Symbiont

Icnlr Y. ruorein .ntiV Orfi Predicaraed Lnds Linrarv Lrarv
6 celluose bindng type 6, thio omdoreductsse-ike prokarvote celiluose 2
13 alpha-N-arabinofranosiase. arese ra e hemeildose 3 2
14 n termite chitin 12
20 starch,-brrdirgprotein, aiphe-elase prokaryoie starch 1
33 conserved hypothetical chitin-bindin prtin terrrete cistn 2
WDA 1okern proarrte elxrne 2


No. Genes No. Genes
Host Symbiont

rr i n e nuIuty u.m. rwiurnwu ljanu.u usmoaru Ljfary Ljary
xylose isomerase prokaryote 1erose, glucuronate fructose, mannose 4
UDP-N-acetyl glucosamine transporter termite IDP-N-acety glucosamine 1
sugar transporter lecin-lie termite carbohydrate 17
sugar Iransporter prokasyole hexose/ glucose 5
Fe-I-Hydtogernae prot H2 4
Fe-Hidnooapnesa roarvote 2H 1
ferreaoon-oxidareaoucase prlle Ferredoxni 10
Iferredoxin-exidoredudase prkaryoe Ferredon 2


n-29 n-24 n-8 n-5 n-12

Glycosyl Carbo-
transferase hydrate

Carbo- Fe- Ferredoxin
hydrate hydrogen- oxido-
euterase ase reductase

Figure 4
Distributions of expressed sequence tags (ESTs) encoding (a) glycosyl transferase family (GTF), (b) carbohy-
drate esterase (CE), (c) carbohydrate binding modules (CBM), and (d) miscellaneous carbohydrate-active and
hydrogen-active proteins generated from the host gut tissue and hindgut symbiont libraries. (a-d) Summaries of
GHF family members with protein identities and putative taxonomic origins that were determined based on library of origin
and database homology. Shaded rows depict gene families not included in the carbohydrate active enzyme (CAZy) database.
(e) Bar graph summarizing the distributions of glycosyl transferase, carbohydrate binding, carbohydrate esterase, Fe-hydroge-
nase, and ferredoxin oxidoreductase genes among prokaryotic, protist and host genomes. See Additional files 2 and 5 for gene
by gene summaries and Additional files 4 and 10 for Genbank accession numbers.

Page 6 of 19
(page number not for citation purposes)

V1-r. rl~lrvl E C uslouue Lra
1 JDP-glucu os tramferase glucunc asod 7
2A doichyl-pliosphate mannosy transferase termite mennoso 1
T2W polysaccharide biosynthesis protein i-roefyt exopolyaccharide monomers2
-3 "Wdine phosphoryse termite thyrmne + phshate 1
4 N-acetyIlucosaminw -phospho transferase termite W e lurosamine 2
8 lipopolysaacharide biosynthesis protein proakryote rtiomnose 4
10 alpha1,3-fucosyl transferase A termite tucose 4
13 N-actet-okuasaminyl transferase I termite UOP-N-etyl-O-ducosamine 1
22 GPI mannoei transferase 3 termite N-acetVkilucosamine, mannose I
2 UDP N-acetlucosamine traisferase termite UDP-N-acetyl-D-olucosamine 3
8 RTIposs a rra ferase terrade Iucose. mannose. Na2 uonsmnne 3

r%\ Miscellaneous




Biotechnology for Biofuels 2009, 2:25

wall biosynthesis. Alternatively, based on significant
homology to insect and animal GTFs, the majority of ter-
mite-derived GTFs have predicted functions in N-acetyl
glucosamine modification and/or chitin biosynthesis.

Carbohydrate esterases
A summary of carbohydrate esterase family members is
provided in Figure 4b. These esterases are involved in car-
bohydrate side chain modification and are considered dis-
tinct from the carboxylesterases described later. Predicted
functions of the carbohydrate esterases are as chitin
deacetylases and acetylxylan esterases. Despite being iden-
tified from both the host and symbiont libraries, all of the
eight carbohydrate esters identified are apparently
encoded by prokaryotic symbionts (Figure 4e).

Carbohydrate binding modules
A summary of genes encoding carbohydrate binding mod-
ule family (CBMF) members, as well as dockerins and
lectins is provided in Figure 4c. In all, five formal CBM
families were represented in the EST dataset (families 6,
13, 14, 20 and 33), of which two are apparently termite
derived (14, 33) and all others are apparently from
prokaryotic symbionts. Of the termite CBMF members, all
are involved in chitin binding; 1 group encodes 12 differ-
ent peritrophins that are putatively part of the peritrophic
membrane secreted around food in the termite midgut.
The remaining prokaryotic CBMF members encode puta-
tive functions in cellulose, hemicellulose and starch bind-
ing. Also included in this category are non-CBMF
members; specifically, prokaryotic dockerins that bind
cellulolytic enzymes in the secreted cellulosome (two
total; Additional file 5) and termite-derived carbohydrate
binding lectins that may have immune functions (two
total; Additional file 1). Of the 23 total CBMF members
identified, 58% are apparently derived from termites and
42% from prokaryotic symbionts (Figure 4e).

Miscellaneous carbohydrate active moieties
Four additional miscellaneous carbohydrate-active pro-
tein families were also identified (Figure 4d). These
include prokaryotic xylose isomerases active towards a
diversity of monosaccharides (four members), prokaryo-
tic sugar transporters (five members), and termite-pro-
duced sugar transporters (two members). Of the termite-
derived transporters, there is 1 N-acetyl glucosamine
transporter potentially involved in peritrophic membrane
or gut cuticle biosynthesis, and 17 lectin-like sugar trans-
porters of unknown significance.

Identification of genes encoding putative lignin
degradation, detoxification, and antioxidant functions
To examine potential members of lignin degradative path-
ways, we searched both host and symbiont cDNA library
datasets for laccases and peroxidases, which are two gene

families known to play roles in fungal lignin degradation
(Figure 5 and Additional file 6). In addition, because
lignin degradation produces toxic metabolites, both host
and symbiont sequence pools were also searched for
xenobiotic and antioxidant detoxification enzymes that
included alcohol dehydrogenases, catalases, superoxide
dismutases, cytochrome P450s, epoxide hydrolases, glu-
tathione S transferases, glutathione peroxidases and car-
boxylesterases. All accession numbers for the putative
lignase, detox and antioxidant sequences are provided in
Additional file 7. Whereas only 1 putative peroxidase gene
was identified in the symbiont EST pool (accession num-
bers FL643288, FL645697), the termite gut library
sequence pool was found to contain 52 enzyme coding
genes putatively associated with lignin degradation or
xenobiotic metabolism/detoxification (Figure 5 and Addi-
tional file 6). With regard to laccases, the presence of lac-
case transcripts in the termite library, but not the
symbiont library, supports that termite gut cells produce
laccases (six clones representing one putative gene; acces-
sion numbers FL639514, FL640712, FL635040.
FL635071, FL635132, FL635524). The termite laccase
contig also has highest homology to other insect laccases,
supporting that its origin is from the host genome (Addi-
tional file 8).

Carboxylesterases are another potentially important
group of termite-derived enzymes identified through the
current work. A total of 12 distinct carboxylesterase genes
were identified from the termite gut gene pool (Figure 5),

Laccase L
Catalase A.L
Epoxide Hydrolase D.L
Superoxide Dismutase A
Glutathione Peroxidase -I A.L
Glutathione-S-transferase D,A
Alcohol Dehydrogenase D
Carboxylesterase D,L
Cytochrome P450 D

0 2 4 6 8 10 12 14 16
Number of genes

Figure 5
Summary of potential lignase (L), detoxification (D),
and antioxidant (A) coding genes identified from the
termite gut library. No homologous or functionally similar
genes were identified from the symbiont library. Black bars
indicate gene families investigated in later functional studies
(see Figures 6-8). See Additional file 6 for gene by gene sum-
maries and Additional file 7 for Genbank accession numbers.

Page 7 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

all of which were absent from the symbiont library and
had highest homology to insect carboxylesterases. A
search for signal peptides at the N-termini of predicted
protein sequences suggests that termite gut cells secrete
laccases, catalases, cytochrome P450s, carboxylesterases,
superoxide dismutases, epoxide hydrolases, and glutath-
ione peroxidases (Additional file 6); however, because of
the known membrane-bound nature of P450 proteins,
they are not likely to be secreted.

Gut phenoloxidase activity, phenoloxidase induction by
lignin feeding, and laccase gene expression
The identification of a novel laccase (EC gene
from the host library prompted us to investigate phe-
noloxidase activity towards the model laccase substrate
pyrogallol in a colorimetric microplate assay (Figure 6).
All results shown are the average standard error deter-
mined from three colonies. pH-dependent pyrogallol oxi-
dation activity was highest at pH 7-8 (Figure 6a), which
partially encompasses the physiological pH range of 6-7
measured previously in the R. flavipes gut [34]. Next, the
tissue distribution of pyrogallol oxidation activity was
determined at the optimal pH of 7 (Figure 6b). In strong
agreement with laccase gene expression (see below), pyro-
gallol oxidation activity was approximately sixfold higher
in the foregut plus salivary gland than in the midgut and
hindgut. This activity profile did not correlate with cata-
lase, carboxylesterase, or reference gene expression (see
below). Finally, pyrogallol activity was investigated after
feeding live termites for 7 days on cellulose filter papers
treated with partially depolymerized lignin (called 'lignin
alkali'). Although no significant differences in feeding
occurred between controls and the different lignin bio-
assay concentrations (Figure 6c), there was a significant
induction of gut pyrogallol oxidation activity associated
with one lignin bioassay concentration (0.313%; Figure

Laccase (contig 659) and catalase (contig 230) gene
expression across the R. flavipes gut, relative to the RfEst3
carboxylesterase (contig 275) and P actin as a control
gene, is shown in Figure 7. Two laccase and two catalase
primer sets were tested; both primer sets for each gene
yielded identical results. Laccase expression was highest in
the foregut plus salivary gland, lower in the midgut, and
lowest in the hindgut. Catalase and P actin expression
were uniform across gut regions; in contrast, RfEst3
expression was highest in the midgut.

The broad goals of this research were to obtain robust
pools of host and symbiont lignocellulase and other car-
bohydrate active genes from the R. flavipes gut, and to
begin functional investigations of candidate genes with

potential roles in industrial lignocellulose pretreatment.
Two cDNA libraries were created for these purposes: (i) a
normalized host cDNA library prepared from gut tissues
cleared of symbionts and (ii) a non-normalized library
prepared from hindgut symbionts. The rationale for tak-
ing these approaches was that the normalization process
would maximize representation of host genes in the host
library (from our results it is clear that symbionts were not
entirely removed from guts upon dissection), while genes
represented in the non-normalized symbiont library
would correspond to the relative abundance (and pre-
sumably importance) of the hindgut symbiota. Prelimi-
nary sequencing summaries from this work are presented
in Scharf and Tartar [5]. The currently reported sequence
results presented in Figures 3, 4, 5 are the results of new
database identity searches in April 2009 after several new
large-scale sequencing projects were deposited in Gen-
bank [23,24,31].

In addition to identifying a robust complement of host
and symbiotic protist genes, numerous genes with strong
prokaryotic homology were also identified. Because
prokaryotic genes are known to be polyadenylated [35],
our cDNA synthesis strategy involving isolation of polyA
RNA with 'oligo-dT' priming resulted in the inclusion of
apparent prokaryotic cDNAs in our libraries. Thus,
although a large fraction of prokaryotic transcripts were
likely not captured, the prokaryotic genes identified here
are likely to be legitimate polyadenylated prokaryotic
genes. However, it is also possible that prokaryotic symbi-
ont genes have assimilated into host and/or protist
genomes [36-38], and/or that some microbial symbiont
sequences already deposited in Genbank may have been
misannotated. For example, with respect to the latter
point, previous metatranscriptome sequencing projects
on termite hindgut environmental cDNA did not consider
the possibility that prokaryotic transcripts can be polyade-
nylated and enriched via polyA RNA purification proto-
cols. Additional efforts beyond the scope of the current
work will be required to resolve these potential assimila-
tion and annotation issues.

With respect to functional analyses, we focused on previ-
ously unknown carboxylesterase and phenoloxidase cod-
ing genes ([39] and this work) with potential relevance to
industrial lignocellulose pretreatment [33]. Both of these
enzyme families are considered more relevant for pretreat-
ment than cellulases or hemicellulases, which are more
relevant in downstream carbohydrate depolymerization
processes that immediately precede simple sugar fermen-
tation [33]. Additionally, cellulase and hemicellulase
genes/activities in R. flavipes are already well documented
([17,18,30,40-43] and the present work). Fungal carboxy-
lesterases known as 'feruloyl esterases' solubilize hemicel-
lulose by cleaving ester bonds between hemicellulose and

Page 8 of 19
(page number not for citation purposes)








6- a a


0.00 0.156 0.313 0.625
Lignin Alkali Concentration (% w/v)



(D) 12

ZE 10

CL 8
T r"






a a

0 ---- ------- --- ---

0 .-

0.00 0.156 0,313 0,625
Lignin Alkali Concentration (% wlv)

Figure 6
Functional analysis of phenoloxidase activity towards the model substrate pyrogallol. (a) pH dependence of pyro-
gallol oxidation determined using whole guts of worker termites. (b) Distribution of pyrogallol oxidation activity across the
termite gut using 14,000 g supernatants from homogenized tissues of the foregut plus salivary gland (FG), midgut (MG), and
hindgut (HG). (c) Feeding by live termites for 7 days on papers treated with lignin alkali at concentrations ranging from 0% to
0.625% w/v. (d) Pyrogallol oxidation activity in whole termite guts after feeding on various concentrations as depicted in (c).
Bars with the same letters are not significantly different by least significant difference (LSD) t tests (P < 0.05).

monolignols and/or phenolic acids [44-47], and thus the
host carboxylesterases identified here represent poten-
tially important pretreatment enzymes. Based on the well
established importance of fungal laccases and related
enzymes in lignin degradation [48,49], phenoloxidases/
laccases are also relevant to industrial lignocellulose pre-

Finally, the host and symbiont sequence pools resulting
from this work constitute an important new genomic
resource. This resource has potential applications in
microarray, heterologous/recombinant protein expres-
sion, and biochemical characterization studies that can
provide more definitive insights into digestome function
and host-symbiont collaborative digestion than the cur-
rent sequencing work. In the sections that follow, details,
inferences and applied impacts of the current findings in

relation to host-symbiont collaborative digestion and bio-
fuel production are discussed.

Comparisons of carbohydrate-active genes from host and
Despite apparent symbiont contamination in the host
library, >85% of carbohydrate active sequences from ter-
mite gut library had significant homology to translated
genes of animal origin. Likewise, despite apparent host
contamination, more than 85% of the sequences from the
symbiont library with carbohydrate-active database
matches were found to be most similar to predicted pro-
teins from microbes. However, interestingly, only 5% of
these transcripts were directly related to sequences from
known termite symbionts or metagenome/transcriptome
databases, highlighting the poor representation of close

Page 9 of 19
(page number not for citation purposes)

4.34 5 6 7 8
Buffer pH

Biotechnology for Biofuels 2009, 2:25

bp .. .. bp
~(Fontrol g7ne)

lows: FG, foregut plus salivary gland; MG, midgut; HG, hindgut; and (-), no template control. Two primer sets were tested for(-
400 ';'q" "C. 1 400
300 W i 300
200 Iu *1 TO m I I m 200
100a .

Figure 7
Laccase, catalase, RfEst3 esterase, and 3 actin (as a control) gene expression in Reticulitermes fiavipes worker
guts as determined from 35 cycles of conventional polymerase chain reaction (PCR) amplification. Shown is a
negative image of a representative agarose gel with equal loadings of PCR products for the four genes. Lane labels are as fol-
lows: FG, foregut plus salivary gland; MG, midgut; HG, hindgut; and (-), no template control. Two primer sets were tested for
the laccase and catalase genes, and one primer set for esterase and P actin. A 100 base pair (bp) low molecular weight ladder
(LML) is shown at the left and right. See Additional file 9 for PCR primer sequences.

homologues within the extant public protein database. It
is possible that with deeper sequencing, greater coverage,
and specific prokaryote targeting, more homologues to
existing metagenome sequences could be identified.

Contaminating sequences
Specific examples of contaminating sequences that we
observed are as follows. In the gut library such transcripts
included a and B tubulin homologues (data not shown)
similar to sequences recently produced from the paraba-
salian termite symbiont Spirotrichonympha leidyi [50], as
well as GHF7 glycosyl hydrolases (Figure 3; Additional file
1) previously shown to be associated with eukaryotic gut
symbionts [20,29]. Other potentially contaminating
sequences identified from the gut library were from uni-
cellular organisms such as fungi, protests and prokaryotes,
as well as sequences homologous to others previously
identified from termite metagenome projects [20,21].
However, by noting phylogenetic BLAST homologies of
specific genes and gene families from each library, we
have been able to deduce the likely origins of certain gene

Although hypotheses about lateral gene transfer and
potential misannotation of archived database sequences
cannot be dismissed, our conclusion from the present
sequence composition findings is that host and symbiont
are intimately connected and therefore cannot be com-
pletely excluded from each other and sampled/studied
independently. For example, a subset of prokaryotic sym-
bionts in Reticulitermes are attached to the gut wall
[51,52]. This phenomenon is concordant with our obser-
vation that the termite gut library contained a fraction of
sequences of prokaryotic origin. Also, protist gut symbi-
onts are now known to contain co-evolved bacterial endo-

symbionts [23,24,53-55], which, because of possible
assimilation, may limit the ability to distinguish host vs.
protist and bacterial endosymbiont genes.

Phylogenetic analyses
The tree presented in Figure 2 is representative of all GHF
phylogenetic analyses that were conducted. These analy-
ses are extremely concordant with the glycosyl hydrolase
family (GHF) gene listing available in the CAZy database The CAZy lists show that GHF1,
GHF9 and GHF16 genes have been widely sequenced
from insects (including termites) but have never been
reported in eukaryotic termite symbionts. In contrast,
GHF7, GHF11 and GHF45 have been sequenced in fungi
and termite symbionts but they are virtually absent from
metazoan genomes [21]. In addition to GHF enzymes
(Figure 3), significant differences in gene content of many
other carbohydrate-active and hydrogen metabolism
enzymes were also observed when comparing the two
libraries (Figures 3 and 4). Despite these trends, we took
the more conservative approach that considered both
phylogenetic signal and the library of origin when making
taxonomic classifications. If correct, our findings provide
seminal evidence of a tripartite collaboration of host plus
protist plus prokaryote in cellulose digestion and exclu-
sive dependence on symbiotic fauna (prokaryote plus
protist) for hemicellulose digestion.

Symbiont GHFs
Based on the hypothesis that the frequencies of ESTs in
non-normalized cDNA libraries are positively correlated
with gene expression levels, consistent with Todaka et al.
[20], the current data indicate that R. flavipes gut symbi-
onts overwhelmingly produce GHF7 enzymes. This GHF
family is mostly known for its cellobiohydrolase (exoglu-

Page 10 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

Biotechnology for Biofuels 2009, 2:25

canase) and endoglucanase activities against microcrystal-
line cellulose [16,20,56]. The symbiont library sequence
pool also suggests that gut symbionts abundantly produce
GHF3, GHF5, GHF10, GHF11, GHF26 and GHF45 hydro-
lases, whereas all other catalytic domains were observed
more sparingly (Figures 3 and 4). Although some of these
apparent symbiont sequences were also identified in the
host library, and assuming that horizontal gene transfer
and previous database misannotation have not occurred,
homology and phylogenetic analysis results confirm that
these GHF sequences are all symbiont derived. These sym-
biont library findings are concordant with another metat-
ranscriptome sequencing effort performed on the R.
speratus symbiotic protist community, which revealed that
members of GHF7 were the most extensively expressed
enzymes [20]. These similarities suggest that both R. fla-
vipes and R. speratus symbiont populations possess a sim-
ilar glycohydrolase arsenal, composed predominantly of
GHF7 hydrolases and a number of additionally important
cellulase and hemicellulase families (including GHF3,
GHF5, GHF11, GHF26 and GHF45). However, the cur-
rent study revealed at least thirteen GHF families from
both R. flavipes libraries that were not identified in R. sper-
atus, specifically GHF2, 9, 13, 16, 18, 20, 27, 30, 37, 38,
42, 47 and 53. Additionally, no laccases or other potential
lignase enzymes were identified in R. speratus hindgut
symbiont library as they were here from the R. flavipes host
library (see below). Another potentially important set of
symbiont genes identified in the current work encode
prokaryotic dockerin proteins. Dockerins are important
components of prokaryotic cellulosomes, which are
secreted complexes of enzymes and proteins that synergis-
tically collaborate in plant cell wall depolymerization
[57]. No dockerin genes have been reported previously
from lower termites.

Carbohydrate active host genes
The gene expression profile of the host termite gut cells is
not characterized by an overly abundant domain (Figures
3 and 4, Additional file 1); however, this may be related
to the fact that the host library was normalized. Important
carbohydrate-active transcripts from the termite host are
mainly associated with cellulose degradation (GHF1, 9,
16); chitin metabolism, which is an important antifungal
and cuticle melanization pathway in insects (GHF18, 20,
27, 85); depolymerization of a-linked carbohydrate poly-
mers (GHF13, 37, 38, 47); glycosyl transferases which are
likely to play roles in gut chitin/cuticle biosynthesis (GT1,
2, 3, 4, 10, 13, 22, 28, 66); carbohydrate binding proteins/
modules also involved in cuticle/peritrophic membrane
biosynthesis (CBM 14,33), a carbohydrate binding lectin
(2 genes); and finally, lectin-like sugar transporters (17
genes). No such genes were reported from the R. speratus
hindgut symbiont library [20]. In agreement with previ-
ous findings [13,16,17,19-21,27,31], the host library

dataset strongly suggests that termite gut cells produce
unique glycoside hydrolases that complement activities
derived from symbionts. Examples of such gene families
include GHF1, GHF9 and GHF16, which are predicted to
encode P glucosidase (GHF1) and endoglucanase (GHF9,
16) activities, and which accounted for 6.5%, 7.5% and
10.5 % of all GHF-domain-containing clones, respectively,
in the host library. Of the termite host-specific glycohy-
drolases, GHF9 endoglucanases and GHF1 P glucosidases
could play strong collaborative roles to liberate glucose
from cellulose, as they do in Trichoderma engineered for
maximal cellulase depolymerization in bioethanol pro-
duction from second-generation feedstocks [33,58].

Candidate lignin degradation, detoxification and
antioxidant factors
Lignin degradation is a critical first step in lignocellulose
digestion as it allows for the dissociation of lignin, cellu-
lose and hemicellulose, making carbohydrate polymers
available for enzymatic digestion [3,5,33]. This step
remains poorly understood in termites and other wood-
feeding insects, but it has been well characterized in
lignin-degrading fungi, where oxidative enzymes, includ-
ing laccases, manganese peroxidases and lignin peroxi-
dases have been shown to play major roles in
delignification [3,48,49]. A recent study demonstrated
indisputably that lignin is degraded on a scale of hours
within the guts of wood-feeding insects, including the
lower termite Zootermopsis angusticollis [59]. Many previ-
ous studies over several decades have also provided evi-
dence of lignin degradation capabilities in termite guts,
including R. flavipes [12,34,60-67]. No candidate lignase
genes from termites have been identified prior to the cur-
rent work.

Several genes identified in the present study may play
direct roles in either lignin degradation or protection from
reactive oxygen species and other toxic metabolites gener-
ated during lignin degradation. BLAST analyses demon-
strated that all genes listed in this category have
homologues in insects, and they were identified exclu-
sively in the host library, supporting that they are endog-
enous to the host genome. The search for signal peptides
at the N-termini of predicted protein sequences and sub-
sequent 5' RACE efforts [Coy NR, Salem TZ, Denton JS,
Kovaleva E, Liu Z, Campbell J, Davis D, Buchman G, Bou-
cias DG, Scharf ME, unpublished results] suggest that ter-
mite gut cells secrete laccases, catalases, esterases,
cytochrome P450s, superoxide dismutases, epoxide
hydrolases, and glutathione peroxidases (Additional file
6). However, as reviewed previously [68] cytochrome
P450 proteins are usually retained intracellularly in insect
cells where they remain anchored in the endoplasmic
reticulum or mitochondria. Despite their expected subcel-

Page 11 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

lular distributions, insect P450s are adept at catalyzing O-
demethylation reactions [68] such as the lignin side chain
oxidations reported by Geib et al. [59]. All other enzymes
(laccases, catalases, carboxylesterases, superoxide dis-
mutases, epoxide hydrolases and glutathione peroxi-
dases) are likely to be secreted in the gut lumen [39] and
therefore are promising candidates to test for activity
against lignin, its phenolic acid monomers, and/or its

Laccases and catalases
After obtaining sequence data indicating the presence of
laccase gene expression in the R. flavipes gut, we under-
took functional studies using the model substrate pyrogal-
lol. Pyrogallol is a proven substrate for fungal laccases that
are known to degrade lignin [49]. In addition to observing
an induction of pyrogallol activity from live termites after
lignin feeding (Figure 7), laccase activity has also now
been confirmed towards several other laccase model sub-
strates, including hydroquinone and dimethoxyphenol,
using both salivary gland homogenates and recombinant
laccase proteins [Coy NR, Salem TZ, Denton JS, Kovaleva
E, Liu Z, Campbell J, Davis D, Buchman G, Boucias DG,
Scharf ME, unpublished results]. Correlated laccase gene
expression and phenoloxidase activity in the foregut/sali-
vary gland region (Figures 6 and 7) and the presence of a
signal peptide in the laccase translation product [Coy NR,
Salem TZ, Denton JS, Kovaleva E, Liu Z, Campbell J, Davis
D, Buchman G, Boucias DG, Scharf ME, unpublished
results] are consistent with the hypothesis that laccase
protein is synthesized in the salivary gland and secreted
into the foregut. This expression pattern is in agreement
with the relatively high oxygen content occurring in the
foregut region [61], and is similar to that of host-secreted
salivary Cell-1 endoglucanase of R. flavipes [17,30]. Cata-
lases, while known to possess phenoloxidase activity
[69,70], were not considered further due to the lack of cor-
relation observed between catalase gene expression and
phenoloxidase activity (Figures 6 and 7).

Laccases have long been associated with phenoloxidase
reactions required for insect cuticle tanning (for example,
[71,721); however, strong laccase gene expression and
activity in the midguts and salivary glands of herbivorous
insects [73,74] are suggestive of digestive roles. The
recombinant R. flavipes laccases noted above is essentially
inactive against tyrosine, dopamine and other melanin
precursors, which is consistent with the hypothesis that it
is not an insect cuticle laccase [Coy NR, Salem TZ, Denton
JS, Kovaleva E, Liu Z, Campbell J, Davis D, Buchman G,
Boucias DG, Scharf ME, unpublished results]. Addition-
ally, the R. flavipes laccase sequence has significant homol-
ogy to ESTs sequenced from a salivary gland cDNA library
of the termite Hodotermopsis sjostedti (Genbank accession
numbers DC232380, DC235488), and also, host-derived
laccase gene expression has now been putatively identi-

fled from the gut of a related lower termite, C. formosanus

Insect laccases are phylogenetically close to fungal lac-
cases [76], which are known for being prominently
responsible for lignin degradation in white rot fungi
[3,6,49] and possibly termite-cultivated fungi [77].
Sequence alignments of the translated laccase contig
assembled from the present sequence data with insect and
fungal laccases indicate that the termite laccase is very
similar to both insect and fungal laccases (Additional file
8). The emerging evidence of endogenous laccase gene
expression in termite guts is both interesting and worthy
of further pursuit given that (i) lignin degradation occurs
in termite guts [59], and (ii) no laccase genes have been
identified in any symbiont sequencing projects ([20-
24,59] and the present work).

Carboxylesterases are another potentially important
group of termite-derived digestive enzymes that were
identified through the current work. Although previous
studies have investigated termite esterase activity [78-83],
to our knowledge, this report and [39] provide seminal
evidence of esterase gene expression in termite guts. All
carboxylesterase genes were sequenced from only the host
library, they are expressed in symbiont-free tissues [39],
and they have significant homology to insect carboxy-
lesterases [39]; therefore they are believed at this time to
be encoded in the termite genome and produced by ter-
mite gut cells. In agreement with our identification of 12
host-derived carboxylesterase genes in the present study,
functional analysis of gut esterase activity revealed strong
levels of activity and diverse isoforms by native polyacry-
lamide gel electrophoresis (PAGE) analysis with an este-
rase model substrate [39]. Full-length gut esterase gene
sequences and more detailed functional biochemistry
results are presented in [39].

Many insect esterases have well defined biological func-
tions, such as xenobiotic, lipid, acetylcholine, and juve-
nile hormone metabolism [84]. However, while being
extremely efficient at metabolizing model substrates such
as naphthyl and p-nitrophenyl esters, the vast majority of
insect esterases have largely undefined functions. This lat-
ter category of esterases is referred to as the 'general este-
rases'. Our hypothesis [5,39] is that termite gut
carboxylesterases are potentially important in cleavage of
lignin and lignin monomers from hemicellulose [2,3,5],
as well as in the lignin side chain oxidations identified
recently in the termite gut [59].

The R. flavipes digestome sequence pool as a new genomic resource
The host-symbiont sequence dataset reported here repre-
sents an excellent genomic resource for addressing a
number of important applied research topics; most

Page 12 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

importantly, collaborative host-symbiont digestion and
development of novel biocatalysts for use in sustainable
bioethanol production. As a first step toward addressing
both topics, microarrays can now be developed that con-
tain the entire complement of host and symbiont genes.
Using digestome microarrays, termite colonies can be fed
different lignocellulosic materials or second-generation
bioethanol feedstocks [85] and then assessed for global
differences in host plus symbiont gene expression. It is
anticipated that this approach will reveal vastly different
complements of host and symbiont gene expression in
response to different lignocellulose diets, as well as which
complements of enzymes are most relevant to target for
use in digestion/depolymerization of specific feedstocks.
Another approach to resolve sequence origin will be to
probe microarrays that contain host and symbiont genes
together with enriched symbiont RNA fractions isolated
from the hindgut lumen.

Once relevant digestive genes are identified, enzyme
blends can be functionally expressed and tested as biocat-
alysts for collaborative digestion of natural materials and
agricultural/forestry feedstocks. Approaches to produce
recombinant enzymes can include heterologous expres-
sion in prokaryotic or eukaryotic systems. Examples of
prior successful efforts to functionally express termite and
symbiont digestive enzymes include GHF9 endogluca-
nases from C. formosanus and Coptotermes acinaciformis in
Escherichia coli [75,86,87]; GHF5, GHF9 and GHF45 cellu-
lases from Nasutitermes hindgut bacteria in E. coli [21],
and R. speratus protist symbiont GHF7 exoglucanases in
Aspergillus oryzae [88]. For industrial biorefinery applica-
tions, we hypothesize that coevolved host and symbiont
enzymes from the same system, that is, the R. flavipes
digestome, will be more efficient than a mix of enzymes
from different systems such as termite symbiont and Tri-
choderma. In this respect, our group has recently function-
ally expressed highly active forms of a host GHF9
endoglucanase and symbiont GHF7 exoglucanase identi-
fled in the current study (Figures 2 and 3) and previously
[17,18], as well as a host laccase and GHF1 P glucosidase
identified in the present study. For expression of these
proteins, we employed a baculovirus insect expression
system that enables large-scale production of recom-
binant proteins with the added benefit of eukaryotic tran-
scriptional and pre/post-translational processing [89].

Finally, the inability to completely physically separate
host from symbiont suggests that future efforts to investi-
gate host-symbiont lignocellulose digestive gene expres-
sion would be better approached by the parallel
sequencing of host and symbiont transcriptomes in a
comprehensive meta-analysis. In this respect, the host and
symbiont databases described here and previously [20-
25,31] will provide a powerful bioinformatic scaffold to

assist future efforts in this area, as well as proteomic

The research presented here took a unique host plus sym-
biont EST sequencing approach to identify host and sym-
biont contributions in collaborative lignocellulose
digestion by termites. EST sequencing from host and sym-
biont cDNA libraries provided >10,000 ESTs that aligned
into 6,555 putative genes with an immense diversity of
functions. Each library contained a significant proportion
(approximately 15%) of apparent cross contaminating
sequences from its opposite fraction, suggesting that
library composition alone cannot provide conclusive
information on host-symbiont collaboration. Because of
the apparent library cross contamination (and assuming
that assimilation and database misannotation did not
occur), we considered both phylogenetic signal and the
library of origin (host or symbiont) when making taxo-
nomic classifications based on sequence identity.

Most importantly, glycosyl hydrolase sequence analyses
revealed an apparent three-way collaboration between
host and protist plus prokaryotic symbionts in cellulase
production, a two-way collaboration between protist plus
prokaryotic symbionts for hemicellulase production, and
a two-way collaboration between host plus prokaryotic
symbionts in chitinase and a carbohydrolase production.
Searches for other carbohydrate-active moieties indicated
that glycosyl transferase, carbohydrate binding, and sugar
transport functions are shared between the termite host
plus prokaryotic symbionts, whereas acetyl side chain car-
bohydrate esterase activities are encoded by prokaryotic
symbionts. Other highly relevant prokaryotic genes such
as cellulosome dockerins that play known roles in facili-
tating cellulose digestion were also identified, supporting
that these activities are entirely symbiont-derived.

The gut (host) library exclusively revealed a number of
phenoloxidase and peroxidase genes that play potential
roles in lignin degradation/depolymerization, as well as
other detoxification and antioxidant genes that may pro-
tect termite cells and symbionts from damage by lignin
degradation products. The gut (host) library also exclu-
sively revealed a number of carboxylesterase genes with
potential roles as lignin depolymerizing esterases or feru-
loyl esterases that solubilize hemicellulose by cleaving
carboxyl ester bonds between hemicellulose (pentose)
sugars and lignin or monolignols [39]. Examination of
phenoloxidase activity also revealed, for the first time, sig-
nificant pyrogallol oxidation activity in the termite gut,
induction of pyrogallol activity by lignin feeding, and
strong agreement between laccase gene expression and
pyrogallol oxidation activity. More detailed results of car-
boxylesterase, laccase and various cellulase catalytic activ-

Page 13 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

ities, obtained using recombinant proteins and other
functional/genomic approaches will be the focus of forth-
coming reports.

It is well established that lignocellulose digestion in the
termite gut is a highly efficient process (for example,
[65]). Because lignocellulose is composed of 40% cellu-
lose, 25% hemicellulose, and 20% lignin [32], it would be
expected that a proportional diversity of digestive factors
specific to these three components should exist in the ter-
mite gut digestome. In this respect, the dual sequencing
approach presented here revealed a large and diverse com-
plement of host and symbiont genes with significant links
to lignocellulose digestion. The total number of candidate
lignocellulases identified here was 171 (45% cellulase,
26% hemicellulase, and 29% lignase/detox candidates).
In addition to identifying thousands of genes with both
well defined and undefined GO functions, including
many potential digestive gene families not revealed by
prior symbiont sequencing projects, the present research
has also revealed previously unseen patterns of host and
symbiont digestive gene expression. While the present
work clearly has limitations, these findings are unprece-
dented; not only do they offer the first concurrent, whole-
digestome glimpse into host-symbiont commensalism/
mutualism in termites, they also provide a new genomic
resource for developing coevolved host and symbiont bio-
catalysts for use in biomass-to-bioethanol applications.

Biological samples
All termites used in this study were verified as R. flavipes
using a combination of 16S rDNA sequence using
genomic DNA obtained from termite heads [90] and sol-
dier morphology. Termites were collected on the Univer-
sity of Florida campus (Alachua County, FL, USA) and
held in the laboratory on a pine wood and paper diet for
3-6 months before use.

Construction of the termite gut (host) cDNA library
Approximately 2,000 guts from Reticulitermes flavipes
worker termites from 5 colonies, with attached salivary
glands, were dissected and cleared of symbionts to yield
approximately 1000 mg of tissue. Dissected guts were
cleared of symbionts by placing individual guts into a
droplet of phosphate buffered saline (PBS; pH 7), open-
ing the paunch with dissecting scissors, rinsing in a second
droplet of PBS, and then by blotting the gut on a piece of
laboratory tissue paper with dissecting forceps. Guts were
stored for 1-2 months at -80 C until RNA isolation. Total
RNA was isolated using the SV Total RNA kit (Promega;
Madison, WI, USA) and mRNA purified using the Oligo-
tex mRNA mini-kit (Qiagen; Valencia, CA, USA). First
strand cDNA was synthesized from mRNA using SMART
technology [91] as part of the SMART cDNA Library Con-
struction kit (Clontech; Mountainview, CA, USA). Next,

the cDNA was normalized using proprietary technology
from Clontech that reduces abundant transcripts, enriches
for rare transcripts, and also enriches for full-length
mRNAs [92]. Normalization included cDNA denatura-
tion/reassociation, treatment by duplex-specific nuclease
[93] and amplification of the normalized fraction by
polymerase chain reaction (PCR). The normalized cDNA
pool was end digested and ligated into the pDNR-Lib vec-
tor (Clontech) then transformed into BH10B competent
cells (Clontech). To obtain enough high-quality plasmid
for library screening and for long-term storage, the library
was amplified and cultured to yield a final titer of 4.2 x
106 cfu/ml.

Construction of the symbiont fauna cDNA library
The digestive tracts from 300 mature workers were dis-
sected and released hindgut luminal contents were col-
lected into a small drop of phosphate buffered saline. The
gut contents, pooled in a microcentrifuge tube on ice,
were centrifuged at 13,000 rpm at 40C for 15 min. Total
RNA was extracted from cell pellets using the SV Total
RNA Isolation System (Promega). The ability of lysis
buffer to disrupt the microbiota in the pellets was con-
firmed microscopically. Approximately 100 gg fresh
(never frozen) total RNA was precipitated with ethyl alco-
hol/ammonium acetate/glycogen and then processed
through the PolyA Purist kit (Ambion; Austin, TX, USA).
cDNA was synthesized immediately from the polyA RNA
(approximately 1 Gg) fraction and ligated into the
pDONR 222 plasmid utilizing the CloneMiner cDNA
Library construction kit (Invitrogen; Carlsbad, CA, USA).
Plasmid preparations were electroporated into competent
ElectroMAX DH10B T1 phage-resistant cells (Invitrogen).
The library was amplified and cultured to yield a final titer
of 1.5 x 107 cfu/ml.

Isolation of clones for sequencing
For sequencing, frozen library aliquots were scraped and
placed into 1 ml ice cold 1 x PBS. The resulting aliquots
(1-10 gl) were diluted further into 1 ml SOC media [94]
and plated on LB-agar plates containing 41 gg/ml chlo-
ramphenicol (gut library) or LB-agar containing 60 lg/ml
kanamycin (symbiont library). Plates were inoculated
with variable volumes of the diluted library mixture to
achieve optimal spacing. Streaked plates were incubated
at 37 C for 16-18 h. Well spaced colonies were selected at
random and placed into 500 gl LB media with 8% glyc-
erol containing either 34 lg/ml chloramphenicol (gut
library) or 50 lg/ml kanamycin (symbiont library). After
inoculation, liquid cultures were shaken (250 rpm) at
37C for 16-18 h. Cultures were frozen at -800C until
plasmid isolation and sequencing as described below.

Sequencing and sequence analysis
Both libraries were considered independently throughout
the sequencing and analysis processes. Sequencing was

Page 14 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

performed by the Interdisciplinary Center for Biotechno-
logical Research (ICBR) at the University of Florida, and it
consisted of single pass reactions from the 5' end of the
cDNA, using the M13 forward primer. Trace files were
screened to identify and eliminate vector sequences and
low-quality reads, using the ICBR software package 'Finch-
Suite' (Geospiza Inc., Seattle, WA, USA). Sequences
shorter than 150 base pairs (bp) were also removed from
the final sequence pools. The remaining high-quality
sequences were used as inputs for contig assemblies per-
formed using CAP3 [95] with default parameters (includ-
ing overlap cut-off values fixed at 30 bp and 75%
identity). The clustered sequences were then annotated
based on similarity searches performed by batch BLAST
analyses using the Greengene interface http://green BLAST analyses included searches against
the non-redundant (nr) NCBI protein database (BLASTX),
as well as against the Conserved Domain Database. An E-
value cut-off of 10-5 was used for all BLAST searches, and
in >95% of instances, both identities and taxonomic ori-
gins were supported by at least the top 10 BLAST hits.
Sequences associated with carbohydrate catabolism were
classified according to the CAZy nomenclature (http_:/; [96]). Identification of signal peptides at
the N-terminus of putative enzymes was performed using
the SignalP program [97].

Phylogenetic analyses
Sequences predicted to encode glycoside hydrolases were
translated in silicon and aligned with homologous
sequences accessed from the CAZy databases. Alignments
were performed using CLUSTALX [98]. Phylogenetic rela-
tionships were reconstructed in PAUP* (Sinauer Associ-
ates, Sunderland MA, USA) using maximum parsimony
and distance (neighbor joining) models, with default
parameters. Support for relationships was assessed by
bootstrap analyses (1000 replicates).

Phenoloxidase functional analyses
Protein isolations
All manipulations were performed on ice. For tissue local-
ization studies, 25 termite worker guts from 3 separate
colonies were removed and dissected into the 3 regions of
foregut plus salivary gland, midgut and hindgut. Each gut
region preparation was homogenized using a Tenbroeck
glass homogenizer in potassium phosphate buffer (0.1 M,
pH 7.6), and then centrifuged for 15 min at 14,000 g and
4 C. The supernatant was saved for assays and the pellet
discarded. Whole gut preparations were prepared identi-
cally, except that 25 guts were dissected and homoge-
nized. Protein content of protein preparations was
estimated by a microplate Bradford assay (Bio-Rad; Her-
cules, CA, USA) using bovine serum albumin as a stand-
ard and corresponding buffers as blanks.

Phenoloxidase assay
Oxidation of the model substrate pyrogallol to the prod-
uct purpurogallin was investigated using the method of
Chance and Maehly [99], modified for a 96-well micro-
plate format. Each assay reaction (250 gl) contained 168
gl nanopure water, 25.6 gl potassium phosphate (100
mM, pH 4-8), 12.8 gtl of 0.5% hydrogen peroxide (not
necessary for activity; Aldrich, Milwaukee, WI, USA; pre-
pared fresh in water), 25.6 gl of 5% pyrogallol solution
(Sigma, St Louis, MO, USA; prepared fresh in water), and
8 gl protein preparation. For assays, protein preparations
were first added to microplates on ice. All reagents were
then combined as described above into a master mix.
Assays were started by adding 242 pl master mix solution
to reaction wells using a multichannel pipette. Assays
were conducted at room temperature and read kinetically
at 420 nm for a total of 30 min. Blank reactions contained
an equivalent volume of potassium phosphate in place of
protein preparation. Specific activity in nmol/min/mg
was determined from the linear portions of reaction
curves using an extinction coefficient of 24.7 mM/cm [99]
and by correcting for protein content. The equation used
to calculate specific activity was as follows: (extinction
coefficient) x (velocity per min) x ((protein concentration
in mg/ml)/(protein dilution factor in assay)). Feeding
assays were also conducted that investigated phenoloxi-
dase activity after termite feeding for 7 days on filter
papers treated with three serial concentrations of lignin
alkali extract (0.625, 0.313, 0.156% w/v; Sigma). All
results are summarized from three protein preparations,
each assayed in triplicate. Statistical analyses consisted of
analysis of variance (ANOVA) with least significant differ-
ence (LSD) t tests for mean separation.

PCR to verify gene expression
All PCR primer sequences are provided in Additional file
9. The stable expression of the reference gene 3 actin across
gut regions was validated previously [17]. PCR primers
were designed with specificity to unique sequence regions
of the target genes, and to produce products in the 100-
300 bp range. PCR primers were designed using Primer3
[100]. cDNA from the foregut plus salivary gland, midgut
and hindgut regions served as the template for PCR. Gut
dissections and homogenizations were performed as
described above, using RNA lysis buffer (Promega) in
place of potassium phosphate. cDNA was synthesized
from the total RNA of 25 individual gut regions per exper-
imental replicate. Total RNA and cDNA were obtained
using the SV total RNA isolation kit (Promega) and the
iScript cDNA Synthesis Kit (Bio-Rad), respectively, follow-
ing manufacturer protocols. PCR reactions contained
equal template loadings and proceeded for 35 cycles.
Control reactions were conducted in the absence of cDNA
template. PCR products were viewed on 1.5% agarose gels

Page 15 of 19
(page number not for citation purposes)

and imaged using a Gel-Doc 2000 imaging system (Bio-

Sequence accession numbers
The sequences were submitted to the Genbank dbEST
database using the ESTin program [101] and are available
publicly with the accession numbers FL634956-FL640828
(host gut tissue library) and FL641015-FL645753 (hind-
gut symbiont library). Annotated accession numbers are
provided in Additional files 3, 4, 7 and 10.

Competing interests
Pending US patent 61/168,275.

Authors' contributions
DGB and MES designed the research; AT, MMW, MRC,
DGB and MES performed the research; AT, DGB and MES
analyzed the data;AT, DGB and MES wrote the paper; XZ
provided technical assistance.

Additional material

Additional file 1
Table S1 Carbohydrate active genes, gut (host) library. Summary of
glycoside hydrolase (GH), glycosyl transferase (GT), carbohydrate este-
rase (CE), carbohydrate binding modules (CBM) and other miscellane-
ous (Misc.) carbohydrate active domain protein coding genes identified
from the termite gut (host) library sequencing. Accession Numbers are
provided in Additional file 3.
Click here for file

Additional file 2
Table S2 Carbohydrate active genes, SYMBIONT library. Summary
of glycoside hydrolase (GH), glycosyl transferase (GT), carbohydrate este-
rase (CE), carbohydrate binding modules (CBM) and other miscellane-
ous (Misc.) carbohydrate active domain protein coding genes identified
from the symbiont library sequencing. Accession Numbers are provided in
Additional file 4.
Click here for file

Additional file 3
Table S3. Genbank accession numbers for carbohydrate active enzymes,
gut (host) library.
Click here for file

Additional file 4
Table S4. Genbank accession numbers for carbohydrate active enzymes,
SYMBIONT library.
Click here for file

Additional file 5
Table S5 Dockerin, Fe-hydrogenase, ferredoxin oxidoreductase and
nitroreductase genes, SYMBIONT library. Summary of dockerin, Fe-
hydrogenase, ferredoxin oxidoreductase and nitroreductase genes identi-
fied from the symbiont library sequencing. Accession Numbers are pro-
vided in Additional file 10.
Click here for file

Additional file 6
Table S6 Candidate lignase, detoxification and antioxidant genes,
gut (host) library. Summary of candidate lignin degradation, detoxifica-
tion and antioxidant enzyme coding genes identified from the termite gut
(host) library sequencing. Accession Numbers are provided in Additional
file 7.
Click here for file

Additional file 7
Table S7. Genbank accession Nos. for candidate lignase, detoxification
and antioxidant genes, gut (HOST) library.
Click here for file

Additional file 8
Figure Sl. Deduced amino .. i .. ...... .. ..f the R. flavipes laccase con-
tig obtained in the present study (Contig 659; indicated by arrows) with
homologous insect and fungal laccases. The fungal laccases shown play
known roles in lignin degradation. Shaded amino acids are those that
match the R. flavipes sequence; insect sequences are above the arrows and
fungal sequences are below. The ESTs assembling into the R. flavipes con-
tig are as follows: FL639514, FL640712, FL635040, FL635071,
FL635132, and FL635524. Sequence accession numbers for all homologs
are shown in parentheses: Ms (Manduca sexta), Tribolium (T. casta-
neum), Ag (Anopheles gambiae), Termitomyces (Termitomyces sp.
NS/Mg.), C. cinera (Coprinus cinereus), A. bisporus (Agaricus bis-
Click here for file

Additional file 9
Table S8. PCR primer sequences used for validating laccase and catalase
gene expression relative to the control gene -actin.
Click here for file

Additional file 10
Table S9. Genbank accession Nos. for Dockerin, Fe-hydrogenase, ferre-
doxin oxidoreductase and nitroreductase genes, SYMBIONT library.
Click here for file

Page 16 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

Biotechnology for Biofuels 2009, 2:25

We thankV Lietze, M Tarver, T Conklin, K Simms and M Schwinghammer
for assistance with gut dissections. We also thank W Farmerie (University
of Florida), and E Kovaleva, G Buchman, and R Balcerzak (Chesapeake-
PERL) for helpful discussions and advice. This research was supported by a
University of Florida IFAS Innovation Grant to the authors, CSREES-USDA-
NRI grant no. 2007-35607-17777 to M ES and XZ, and through the Consor-
tium for Plant Biotechnology Research, Inc. by DOE Prime Agreement no.
DE-FG36-02GO 12026 to MES and DGB (this support does not constitute
an endorsement by DOE or by the Consortium for Plant Biotechnology
Research, Inc. of the views expressed in this publication).

I. Ragauskas AJ, Williams CK, Davison BH, Britovsek G, Cairneyj, Eck-
ert CA, Frederick WJ Jr, Hallett JP, Leak DJ, Liotta CL, Mielenz JR,
Murphy R, Templer R, Tschaplinski T: The path forward for bio-
fuels and biomaterials. Science 2006, 3 11:484-489.
2. Saha BC: Hemicellulose bioconversion. j Ind Microbiol Biotechnol
2003, 30:279-291.
3. Anderson WF, Akin DE: Structural and chemical properties of
grass lignocelluloses related to conversion for biofuels. j Ind
Microbiol Biotechnol 2008, 35:355-366.
4. Rubin EM: Genomics of cellulosic biofuels. Nature 2008,
5. Scharf ME, Tartar A: Termite digestomes as sources for novel
lignocellulases. Biofuels Bioprod Bioref 2008, 2:540-552.
6. Ohkuma M: Termite symbiotic systems: efficient bio-recycling
of lignocellulose. Apple Microbiol Biotechnol 2003, 61:1-9.
7. Brune A: Microbiology: woodworker's digest. Nature 2007,
8. Chaffron S, von Mering C: Termites in the woodwork. Genome
Biol 2007, 8:229.
9. Brune A: Symbiotic associations between termites and
prokaryotes. Prokaryotes 2006, 1:439-474.
10. Ohkuma M, Hongoh Y, Noda S: Symbiotic complex in the ter-
mite gut microbial community. Tanpakushitsu Kakusan Koso
2008, 53:1841-1849.
I I. Wood TG, Johnson RA: The biology, physiology, and ecology of
termites. In Economic impact and control of social insects Edited by:
Vinson SB, Johnson RA. Santa Barbara, CA, USA: Praeger; 1986:1-68.
12. Breznak JA, Brune A: Role of microorganisms in the digestion
of lignocellulose by termites. Annu Rev Entomol 1994,
13. Watanabe H, Nakamura M, Tokuda G, Yamaoka I, Scrivener AM,
Noda H: Site of secretion and properties of endogenous endo-
P-1,4-glucanase components from Reticulitermes speratus
(Kolbe), a Japanese subterranean termite. Insect Biochem Mol
Biol 1997, 27:305-313.
14. Watanabe H, Noda H, Tokuda G, Lo N: A cellulase gene of ter-
mite origin. Nature 1998, 394:330-331.
15. Nakashima K, Azumajl: Distribution and properties of endo-p-
1,4-glucanase from a lower termite Coptotermes formosanus.
Biotechnol Biochem 2000, 64:1500-1506.
16. Nakashima K, Watanabe H, Saitoh H, Tokuda G, AzumaJ-l: Dual cel-
lulose-digesting system of the wood-feeding termite, Copto-
termes formosanus. Insect Biochem Mol Biol 2002, 32:777-784.
17. Zhou X, Smith JA, Oi FM, Koehler PG, Bennett GW, Scharf ME: Cor-
relation of cellulase gene expression and cellulolytic activity
throughout the gut of the termite Reticulitermes flavipes.
Gene 2007, 395:29-39.
18. Wheeler MM, Zhou X, Scharf ME, Oi FM: Molecular and biochem-
ical markers for monitoring dynamic shifts in cellulolytic
protozoa in Reticulitermes flavipes. Insect Biochem Mol Biol 2007,
19. Tokuda G, Watanabe H, Lo N: Does correlation of cellulase
gene expression and cellulolytic activity in the gut of termite
suggest synergistic collaboration of cellulases? Gene 2007,
20. Todaka N, Moriya S, Saita K, Hondo T, Kiuchi I, Takasu H, Ohkuma
M, Piero C, Hayashizaki Y, Kudo T: Environmental cDNA analy-
sis of the genes involved in lignocellulose digestion in the
symbiotic protist community of Reticulitermes speratus. FEMS
Microbiol Ecol 2007, 59:592-599.

21. Warnecke F, Luginbuhl P, Ivanova N, Ghassemian M, Richardson TH,
Stege JT, Cayouette M, McHardy AC, Djordjevic G, Aboushadi N,
Sorek R, Tringe SG, Podar M, Martin HG, Kunin V, Dalevi D, Madejska
J, Kirton E, Platt D, Szeto E, Salamov A, Barry K, Mikhailova N, Kyrpi-
des NC, Matson EG, Ottesen EA, Zhang X, Hernindez M, Murillo C,
Acosta LG, Rigoutsos I, Tamayo G, Green BD, et al.: Metagenomic
and functional analysis of hindgut microbiota of a wood-feed-
ing higher termite. Nature 2007, 450:560-565.
22. Margonelli L: Gut reactions. [
23. Hongoh Y, Sharma VK, Prakash T, Noda S, Taylor TD, Kudo T, Sakaki
Y, Toyoda A, Hattori M, Ohkuma M: Complete genome of the
uncultured Termite Group I bacteria in a single host protist
cell. Proc NatlAcad Sci USA 2008, 105:5555-5560.
24. Hongoh Y, Sharma VK, Prakash T, Noda S, Toh H, Taylor TD, Kudo
T, Sakaki Y, Toyoda A, Hattori M, Ohkuma M: Genome of an endo-
symbiont coupling N2 fixation to cellulolysis within protist
cells in termite gut. Science 2008, 322:1108-1109.
25. Johjima T, Taprab Y, Noparatnaraporn N, Kudo T, Ohkuma M:
Large-scale identification of transcripts expressed in a sym-
biotic fungus (Termitomyces) during plant biomass degrada-
tion. Apple Microbiol Biotechnol 2006, 73:195-203.
26. Slaytor M, Sugimoto A, Azuma J, Murashima K, Inoue T: Cellulose
and xylan utilisation in the lower termite Reticulitermes sper-
atus. j Insect Physiol 1997, 43:235-242.
27. Tokuda G, Watanabe H, Matsumoto T, Noda H: Cellulose diges-
tion in the wood-eating higher termite, Nasutitermes
takasagoensis (Shiraki): distribution of cellulases and proper-
ties of endo-beta-1,4-glucanase. Zoolog Sci 1997, 14:83-93.
28. Tokuda G, Saito H, Watanabe H: A digestive beta-glucosidase
from the salivary glands of the termite, Neotermes koshunen-
sis (Shiraki): distribution, characterization and isolation of its
precursor cDNA by 5'- and 3'-RACE amplifications with
degenerate primers. Insect Biochem Mol Biol 2002, 32:1681-1689.
29. Watanabe H, Nakashima K, Saito H, Slaytor M: New endo-beta-
1,4-glucanases from the parabasalian symbionts, Pseudotri-
chonympha grassii and Holomastigotoides mirabile of Coptoter-
mes termites. Cell Mol Life Sci 2002, 59:1983-1992.
30. Zhou X, Wheeler MM, Oi FM, Scharf ME: Inhibition of termite
cellulases by carbohydrate-based cellulose inhibitors: evi-
dence from in vitro biochemistry and in vivo feeding studies.
Pestic Biochem Physiol 2008, 90:3 1-4 1.
31. Yuki M, Moriya S, Inoue T, Kudo T: Transcriptome analysis of the
digestive organs of Hodotermopsis sjostedti, a lower termite
that hosts mutualistic microorganisms in its hindgut. Zoolog
Sci 2008, 25:401-406.
32. Lange JP: Lignocellulose conversion: an introduction to chem-
istry, process and economics. Biofuels Bioprod Bioref 2007,
33. Yang B, Wyman CE: Pretreatment: the key to unlocking low-
cost cellulosic ethanol. Biofuels Bioprod Bioref 2008, 2:26-40.
34. Brune A, Miambi E, BreznakJA: Roles of oxygen and the intesti-
nal microflora in the metabolism of lignin-derived phenyl-
propanoids and other monoaromatic compounds by
termites. Apple Env Microbiol 1995, 61:2688-2695.
35. Sarkar N: Polyadenylation ofmRNA in prokaryotes. Annu Rev
Biochem 1997, 66:173-197.
36. Genereux DP, Logsdon JM Jr: Much ado about bacteria-to-verte-
brate lateral gene transfer. Trends Genet 2003, 19:191-195.
37. Hotopp JC, Clark ME, Oliveira DC, Foster JM, Fischer P, Torres MC,
Giebel JD, Kumar N, Ishmael N, Wang S, Ingram J, Nene RV, Shepard
J, Tomkins J, Richards S, Spiro DJ, Ghedin E, Slatko BE, Tettelin H,
Werren JH: Widespread lateral gene transfer from intracellu-
lar bacteria to multicellular eukaryotes. Science 2007,
38. Lozupone CA, Hamady M, Cantarel BL, Coutinho PM, Henrissat B,
Gordon JI, Knight R: The convergence of carbohydrate active
gene repertoires in human gut microbes. Proc NatlAcad Sci USA
2008, 105:15076-15081.
39. Wheeler MM, Tarver MR, Coy MR, Scharf ME: Characterization of
four esterase genes and esterase activity from the gut of the
termite Reticulitermes flavipes. Arch Insect Biochem Physiol in
40. Zhou X, Wheeler MM, Oi FM, Scharf ME: RNA interference in the
termite Reticulitermes flavipes through ingestion of double-
stranded RNA. Insect Biochem Mol Biol 2008, 38:805-815.

Page 17 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

41. Smith JA, Koehler PG: Changes in Reticulitermes flavipes gut
xylanolytic activities in response to dietary xylan content.
Ann Entomol Soc Am 2007, 100:568-573.
42. Smith JA, Scharf ME, Pereira RM, Koehler PG: Comparison of gut
carbohydrolase activity patterns in Reticulitermes flavipes
and Coptotermes formosanus workers and soldiers. Sociobiology
2009, 53:13-22.
43. Smith JA, Scharf ME, Pereira RM, Koehler PG: pH optimization of
gut cellulase and xylanase activities from the Eastern subter-
ranean termite, Reticulitermes flavipes. Sociobiology 2009,
44. de Vries RP, Michelsen B, Poulsen CH, Kroon PA, Heuvel RH van den,
Faulds CB, Williamson G, Hombergh JP van den, Visser J: The faeA
genes from Aspergillus niger and Aspergillus tubingensis encode
ferulic acid esterases involved in degradation of complex cell
wall polysaccharides. Apple Environ Microbiol 1997, 63:4638-4644.
45. de Vries RP, vanKuyk PA, Kester HC, VisserJ: The Aspergillus niger
faeB gene encodes a second feruloyl esterase involved in pec-
tin and xylan degradation and is specifically induced in the
presence of aromatic compounds. Biochem 2002, 363:377-386.
46. Crepin VF, Faulds CB, Connerton IF: Functional classification of
the microbial feruloyl esterases. Apple Microbiol Biotechnol 2004,
47. Benoit I, Danchin EGJ, Bleichrodt RJ, de Vries RP: Biotechnological
applications and potential of fungal feruloyl esterases based
on prevalence, classification and biochemical diversity. Bio-
technol Letters 2008, 30:387-396.
48. Smith M, Thurston CF, Wood DA: Fungal laccases role in delig-
nification and possible industrial applications. In Multi-copper
oxidases Edited by: Messerschmidt A. Hackensack, NJ, USA: World
Scientific; 1997:201-224.
49. Baldrian P: Fungal laccases occurrence and properties. FEMS
Microbiol Rev 2006, 30:215-242.
50. Ohkuma M, Saita K, Inoue T, Kudo T: Comparison of four protein
phylogeny of parabasalian symbionts in termite guts. Mol Phy-
logen Evol 2007, 42:847-853.
51. LeadbetterJR, BreznakJA: Physiological ecology ofMethanobre-
vibacter cuticularis sp. nov. and Methanobrevibacter curvatus
sp. nov., isolated from the hindgut of the termite Reticuli-
termes flavipes. Apple Environ Microbiol 1996, 62:3620-3631.
52. Yang H, Schmitt-Wagner D, Stingl U, Brune A: Niche heterogene-
ity determines bacterial community structure in the termite
gut (Reticulitermes santonensis). Environ Microbiol 2005,
53. Herlemann DP, Geissinger 0, Ikeda-Ohtsubo W, Kunin V, Sun H,
Lapidus A, Hugenholtz P, Brune A: Genomic analysis of Elusimi-
crobium minutum, the first cultivated representative of the
phylum Elusimicrobia (formerly termite group I). Apple Environ
Microbiol 2009, 75:2841-2849.
54. Ikeda-Ohtsubo W, Brune A: Cospeciation of termite gut flagel-
lates and their bacterial endosymbionts: Trichonympha spe-
cies and Candidatus Endomicrobium trichonymphae. Mol Ecol
2009, 18:332-342.
55. Sato T, Hongoh Y, Noda S, Hattori S, Ui S, Ohkuma M: Candidatus
Desulfovibrio trichonymphae, a novel intracellular symbiont
of the flagellate Trichonympha agilis in termite gut. Environ
Microbiol 2009, I 1:1007-1015.
56. Todaka N, Lopez CM, Inoue T, Saita K, Maruyama JI, Arioka M, Kita-
moto K, Kudo T, Moriya S: Heterologous expression and char-
acterization of an endoglucanase from a symbiotic protist of
the lower termite, Reticulitermes speratus. Apple Biochem Biotech-
nol in press.
57. Bayer EA, Belaich JP, Shoham Y, Lamed R: The cellulosomes: mul-
tienzyme machines for degradation of plant cell wall polysac-
charides. Annu Rev Microbiol 2004, 58:521-554.
58. Wyman CE: Twenty years of trials, tribulations, and research
progress in bioethanol technology. Apple Biochem Biotechnol 200 1,
59. Geib SM, Filley TR, Hatcher PG, Hoover K, Carlson JE, del Mar
Jimenez-Gasco M, Nakagawa-lzumi A, Sleighter RL, Tien M: Lignin
degradation in wood-feeding insects. Proc Natl Acad Sci USA
2008, 105:12932-12937.
60. Esenther GR, Kirk TK: Catabolism of aspen sapwood in Reticu-
litermes flavipes. Ann Entomol Soc Am 1974, 67:989-991.
61. Brune A, Emerson D, Breznak JA: Termite gut microflora as an
oxygen sink: microelectrode determinations of oxygen and

pH gradients in guts of lower and higher termites. Apple Env
Microbiol 1995, 61:2681-2687.
62. Butler JHA, Buckerfield JC: Digestion of lignin by termites. Soil
Biol Biochem 1979, 11:507-511.
63. Cookson LJ: 14C-lignin degradation by three Australian ter-
mite species. Wood Sci Technol 1987, 21:11-25.
64. Cookson LJ: The site and mechanism of 14C-lignin degrada-
tion by Nasutitermes exitosus. j Insect Physiol 1988, 34:409-414.
65. Shuji I, Ueshima K, Tanaka H, Enoki A: Degradation of wood com-
ponents by subterranean termite, Coptotermes formosanus.
Mokuzai Gakkaishi 1995, 4 1:580-586.
66. Kuhnigk T, Konig H: Degradation of dimeric lignin model com-
pounds by aerobic bacteria isolated from the hindgut of xylo-
phagous termites. j Basic Microbiol 1997, 37:205-211 I.
67. Katsumata KS,Jin Z, Hori K, liyama K: Structural changes in lignin
of tropical woods during digestion by termite, Cryptotermes
brevis. j Wood Sci 2007, 53:419-426.
68. Feyereisen R: Insect cytochrome P450. In Comprehensive molecu-
lar insect science Volume 4. Edited by: Gilbert LI, latrou K, Gill SS. New
York, USA: Elsevier; 2005:1-77.
69. Vetrano AM, Heck DE, Mariano TM, Mishin V, Laskin DL, Laskin JD:
Characterization of the oxidase activity in mammalian cata-
lase. j Biol Chem 2005, 280:35372-35381.
70. Sutay-Kocabas D, Bakir U, Phillips SE, McPherson MJ, Ogel ZB: Puri-
fication, characterization, and identification of a novel
bifunctional catalase-phenol oxidase from Scytalidium ther-
mophilum. Apple Microbiol Biotechnol 2008, 79:407-415.
71. Arakane Y, Muthukrishnan S, Beeman RW, Kanost MR, Kramer KJ:
Laccase 2 is the phenoloxidase gene required for beetle cuti-
cle tanning. Proc Natl Acad Sci USA 2005, 103:1 1337- 1342.
72. Yatsu J, Asano T: Cuticle laccase of the silkworm, Bombyx mori:
purification, gene identification and presence of its inactive
precursor in the cuticle. Insect Biochem Mol Biol 2009, 39:254-262.
73. Dittmer NT, Suderman RJ,Jiang H, Zhu Y-C, Gorman MJ, Kramer KJ,
Kanost MR: Characterization of cDNAs encoding putative lac-
case-like multicopper oxidases and developmental expres-
sion in the tobacco hornworm, Manduca sexta, and the
malaria mosquito, Anopheles gambiae. Insect Biochem Mol Biol
2003, 34:29-41.
74. Hattori M, Konishi H, Tamura Y, Konno K, Sogawa K: Laccase-type
phenoloxidase in salivary glands and watery saliva of the
green rice leafhopper, Nephotettix cincticeps. j Insect Physiol
2005, 51:1359-1365.
75. Zhang D, Lax AR, Raina AK, Bland JM: Differential cellulolytic
activity of native-form and C-terminal tagged-form cellulase
derived from Coptotermes formosanus and expressed in E.
coli. Insect Biochem Mol Biol 2009, 39:5 16-522.
76. Hoegger PJ, Kilaru S, James TY, Thacker JR, Kues U: Phylogenetic
comparison and classification of laccase and related multi-
copper oxidase protein sequences. FEBSJ 2006, 273:2308-2326.
77. Taprab Y, Johjima T, Maeda Y, Moriya S, Trakulnaleamsai S, Noparat-
narapron N, Okhuma M, Kudo T: Symbiotic fungi produce lac-
cases potentially involved in phenol degradation in fungus
combs of fungus-growing termites in Thailand. Apple Env Micro-
biol 2005, 71:7696-7704.
78. Sreerama L, Veerabhadrappa PS: Purification and properties of
carboxylesterases from the mid-gut of the termite Odontot-
ermes horni. Insect Biochem 1991, 21:833-844.
79. Sreerama L, Veerabhadrappa PS: Isolation and properties of car-
boxylesterases of the termite gut-associated fungus, Xylaria
nigripes, and their identity from the host termite, Odontoter-
mes horni. Intj Biochem 1993, 25:1637-165 1.
80. Wyss-Huber M: Caste differences in haemolymph proteins in
two species of termites. Insectes Soc 1981, 28:71-86.
81. Ruvolo-Takasusuki MCR, Collet T: Characterization of Nasuti-
termes globiceps esterases. Biochem Genet 2000, 38:367-375.
82. Davis RW, Kamble ST, Prabhakaran SK: Characterization of gen-
eral esterases in workers of the Eastern subterranean ter-
mite. J Econ Entomol 1995, 88:574-578.
83. Valles SM, Oi FM, Strong CA: Purification and characterization
of trans-permethrin metabolizing microsomal esterases
from workers of the eastern subterranean termite, Reticuli-
termes flavipes. Insect Biochem Mol Biol 2001, 3 1:715-725.
84. Oakeshott JG, Claudianos C, Campbell PM, Newcombe RD, Russell
RJ: Biochemical genetics and genomics of insect esterases. In

Page 18 of 19
(page number not for citation purposes)

Comprehensive molecular insect science Volume 5. Edited by: Gilbert LI,
latrou K, Gill SS. New York, USA: Elsevier-Pergamon; 2005:309-382.
85. Simmons BA, Loque D, Blanch HW: Next-generation biomass
feedstocks for biofuel production. Genome Biol 2008, 9:242.
86. Ni J, Takehara M, Watanabe H: Heterologous overexpression of
a mutant termite cellulase gene in Escherichia coli by DNA
shuffling of four orthologous parental cDNAs. Biosci Biotechnol
Biochem 2005, 69:171 1-1720.
87. Ni J, Takehara M, Miyazawa M, Watanabe H: Random exchanges
of non-conserved amino acid residues among four parental
termite cellulases by family shuffling improve thermostabil-
ity. Protein Eng Des Sel 2007, 20:535-542.
88. Sasaguri S, Maruyama J, Moriya S, Kudo T, Kitamoto K, Arioka M:
Codon optimization prevents premature polyadenylation of
heterologously-expressed cellulases from termite-gut sym-
bionts in Aspergillus oryzae. j Gen AppI Microbiol 2008, 54:343-35 I.
89. O'Connell KP, Kovaleva E, Campbell JH, Anderson PE, Brown SG,
Davis DC, Valdes JJ, Welch RW, Bentley WE, van Beek NA: Produc-
tion of a recombinant antibody fragment in whole insect lar-
vae. Mol Biotechnol 2007, 36:44-5 I.
90. Szalanski AL, Austin JW, Owens CB: Identification of Reticuli-
termes spp. (Isoptera: Rhinotermitidae) by polymerase chain
reaction restriction fragment length polymorphism. j Econ
Entomol 2003, 96:1514-1519.
91. Zhu YY, Machleder EM, Chenchik A, Li R, Siebert PD: Reverse tran-
scriptase template switching: a SMART approach for full-
length cDNA library construction. Biotechniques 2001,
92. Zhulidov PA, Bogdanova EA, Shcheglov AS, Vagner LL, Khaspekov
GL, Kozhemyako VB, Matz MB, Meleshkevitch A, Moroz LL, Lukyanov
SA, Shagin DA: Simple cDNA normalization using Kamchatka
crab duplex-specific nuclease. Nucleic Acids Res 2004, 32:e37.
93. Shagin DA, Rebrikov DV, Kozhemyako VB, Altshuler IM, Schlegov AS,
Zhulidov PA, Bogdanova EA, Staroverov DB, Rasskazov VA, Lukyanov
S: A novel method for SNP detection using a new duplex-spe-
cific nuclease from crab hepatopancreas. Genome Res 2002,
94. Sambrook J, Fritsch EF, Maniatis T: Molecular cloning, a labora-
tory manual. >Cold Spring Harbor Laboratory Press; 1989.
95. Huang X, Madan A: CAP3: A DNA sequence assembly pro-
gram. Genome Res 1999, 9:868-877.
96. Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard V, Hen-
rissat B: The Carbohydrate-Active EnZymes database
(CAZy): an expert resource for glycogenomics. Nucleic Acids
Res 2009, 37:D233-238.
97. Bendtsen JD, Nielsen H, von Heijne G, Brunak S: Improved predic-
tion of signal peptides: SignalP 3.0. j Mol Biol 2004, 340:783-795.
98. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG: The
ClustalX windows interface: flexible strategies for multiple
sequence alignment aided by quality analysis tools. Nucleic
Acid Res 1997, 24:4876-4882.
99. Chance B, Maehly AC: Assays of catalases and peroxidases.
Meth Enzymol 1955, 2:773-775.
100. Rozen S, Skaletsky HJ: Primer3 on the WWW for general users
and for biologist programmers. In Bioinformatics methods and pro-
tocols, methods in molecular biology Edited by: Krawetz S, Misener S.
Totowa, NJ, USA: Humana Press; 2000:365-386.
101. Lee B: ESTin: a program for building a dbEST submission file
from massive EST sequences. Genom Info 2006, 4:170-172.

Publish with BioMed Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours you keep the copyright

Submit your manuscript here: BioMedcentral -

Page 19 of 19
(page number not for citation purposes)

Biotechnology for Biofuels 2009, 2:25

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs