Group Title: BMC Molecular Biology 2010, 11:28
Title: Socio-environmental and endocrine influences on developmental and caste-regulatory gene expression in the eusocial termite Reticulitermes flavipes
Full Citation
Permanent Link:
 Material Information
Title: Socio-environmental and endocrine influences on developmental and caste-regulatory gene expression in the eusocial termite Reticulitermes flavipes
Series Title: BMC Molecular Biology 2010, 11:28
Physical Description: Archival
Creator: Tarver MR
Zhou X
Scharf ME
Publication Date: 40291
 Record Information
Bibliographic ID: UF00100277
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access:


This item has the following downloads:

01471-2199-11-28 ( PDF )

Full Text

BMC Molecular Biology

This Provisional PDF corresponds to the article as it appeared upon acceptance. Fully formatted
PDF and full text (HTML) versions will be made available soon.

Socio-environmental and endocrine influences on developmental and
caste-regulatory gene expression in the eusocial termite Reticulitermes flavipes

BMC Molecular Biology 2010, 11:28 doi:10.1186/1471-2199-11-28

Matthew R Tarver (Matt.Tarver@ARS.USDA.GOV)
Xuguo Zhou (
Michael E Scharf (


Article type

Submission date

Acceptance date

Publication date

Article URL


Research article

28 October 2009

23 April 2010

23 April 2010

Like all articles in BMC journals, this peer-reviewed article was published immediately upon
acceptance. It can be downloaded, printed and distributed freely for any purposes (see copyright
notice below).

Articles in BMC journals are listed in PubMed and archived at PubMed Central.

For information about publishing your research in BMC journals or any BioMed Central journal, go to

2010 Tarver et al., licensee BioMed Central Ltd.
This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.ora/licenses/bv/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Socio-environmental and endocrine influences on
developmental and caste-regulatory gene expression
in the eusocial termite Reticulitermes flavipes

Matthew R. Tarver1'2, Xuguo Zhou'3 and Michael E. Scharf'

Department of Entomology and Nematology, University of Florida, Gainesville FL,


2 Present Address: Formosan Subterranean Termite Research Unit, USDA-ARS-SRRC,

New Orleans LA, USA

3 Present Address: Department of Entomology, University of Kentucky, Lexington,


Corresponding author

Email addresses:






Strict regulation of caste differentiation, at the molecular level, is thought to be important to
maintain social structure in insect societies. Previously, a number of extrinsic and intrinsic
factors have been shown to influence caste composition in termite colonies. One important
factor is the influence of nestmates; in particular, soldier termites are known to inhibit
hormone-dependent worker-to-soldier differentiation. However, soldier influences on nestmates
at the molecular level are virtually unknown. Here, to test the hypothesis that soldiers can
influence nestmate gene expression, we investigated the impact of four treatments on whole-
body gene expression in totipotent Reticulitermesflavipes workers: (i) juvenile hormone III
(JHIII; a morphogenetic hormone), (ii) soldier head extracts (SHE), (iii) JHIII+SHE, and (iv)
live soldiers.


Using quantitative-real-time PCR we determined the expression patterns of 49 previously
identified candidate genes in response to the four treatments at assay days 1, 5, and 10. Thirty-
eight total genes from three categories (chemical production / degradation, hemolymph protein,
and developmental) showed significant differential expression among treatments. Most
importantly, SHE and live soldier treatments had a significant impact on a number of genes
from families known to play roles in insect development, supporting previous findings and
hypotheses that soldiers regulate nestmate caste differentiation via terpene primer pheromones
contained in their heads.


This research provides new insights into the impacts that socio-environmental factors (JH,
soldiers, primer pheromones) can have on termite gene expression and caste differentiation, and
reveals a number of socially-relevant genes for investigation in subsequent caste differentiation

Phenotypic plasticity can be described as the production of variable phenotypes

from a single genotype based on conditions encountered throughout an organism's

development [1]. Phenotypic plasticity can be divided into gradual or discrete

polyphenisms. Reaction norms are phenotypically graded responses to environmental

factors. Polyphenisms, occur when two or more discrete alternative phenotypes occur

without intermediate forms [2].

Social insects have evolved to produce and use multiple alternate phenotypes

(i.e., polyphenism) to accomplish a wide range of tasks within their colonies. Castes are

phenotypically and behaviorally discrete individuals that cooperate to perform colony

tasks [3]. Termites are hemimetabolous social insects that utilize castes to meet various

needs within the colony. Most termite colonies are made up of three distinct castes:

workers, soldiers, and reproductive [4]. All termite eggs, except when a rare genetic

component might be involved [5], are considered totipotent, and most evidence

supports that castes differentiate based on gene expression responses to intrinsic and

extrinsic factors. The research presented here examined gene expression responses in

worker termites to both intrinsic and extrinsic factors.

All castes except soldiers and reproductive retain the ability to molt, while

soldiers and reproductive are considered terminally developed [6]. Caste

differentiation can proceed along two routes; imaginal (winged) or apterous (wingless).

The first developmental branch is the point at which larvae differentiate into apterous

workers or imaginal nymphs. Nymphs can either (i) regress into worker-like

"pseudergates", (ii) differentiate into fully winged and eyed adult alates that disperse

and found new colonies, or (iii) differentiate into winged and eyed non-dispersive

brachypterous reproductive that serve as supplemental reproductive within the

colony. Workers are totipotent in that they can (i) undergo status quo worker-to-worker

molts, (ii) differentiate into soldiers (after passing through an intermediate presoldier

stage), or (iii) differentiate into apterous and eyeless neotenic reproductive that serve

supplementary reproductive roles [6-8].

The entire complement of intrinsic and extrinsic factors that dictate each of the

developmental switches in termites, and how they interact, are yet to be fully

understood. Examples of intrinsic factors include juvenile hormone (JH), storage

proteins, and nutrition; whereas examples of extrinsic factors include primer

pheromones, temperature, food quality, nestmates (soldiers and reproductives, and

season [9-19].

Phenotypic divergence from the worker to the soldier caste can be mediated by

multiple JH-related factors. For example, elevated JH titers in workers are correlated

with presoldier differentiation [20-22]. Additionally, the presence of soldiers has been

shown to inhibit the formation of new soldiers, implying that soldier termites produce

inhibitory factors that cause reduced responsiveness to JH or reduced JH biosynthesis

in nestmates [15,16,23,24]. This inhibition is presumed to be caused by soldier-derived

primer pheromones [10,25,26]. Primer pheromones are defined as chemical messengers

that are passed among individuals and trigger physiological responses in recipients

[27]. Recently, R. flavipes soldier head extracts (SHE), when applied in combination

with juvenile hormone III (JH III), were found to enhance presoldier production

compared to JH III alone [19]. Two major components of R. flavipes SHE are y-

cadinene and its aldehyde y-cadinenal; they represent the first candidate primer

pheromones to be identified from termites [19]. Interestingly, SHE alone does not

impact presoldier formation [19]. Also, while the SHE blend is active at influencing

JH-dependent presoldier differentiation, the individual impacts of its constituents and

whether they are being actively released or absorbed has yet to be determined.

Functional genomics is a powerful approach for elucidating the functions of

genes, including genes that mediate pheromone and hormone action [28]. Transcript

levels generally correlate with the physiological demand for the product they produce;

thus, changes in transcript abundance can reveal genes that are most important in

relation to a stimulus [28]. Such an approach has been used to elucidate the chemical

ecology of the bark beetle (Ips pini) [28-31] and the honeybee (Apis mellifera) [32].

Similarly, the use of functional genomics in studies of termite caste regulation can help

to better understand potential primer pheromone function as well as the influences of

intrinsic and extrinsic factors on caste differentiation.

The central goal of this research was to use a functional genomics approach to

identify candidate caste-regulatory genes from R. flavipes workers that potentially

mediate hormonal and soldier primer pheromone signaling. Four treatments were

tested on isolated groups of worker termites: (i) JH III alone, (ii) soldier head extract

(SHE) alone, (iii) JH III + SHE, and (iv) live soldiers. These treatments represent key

intrinsic and extrinsic / socio-environmental factors that are thought to impact soldier

development in totipotent workers. Our central hypothesis was that these four

treatments will be associated with the differential expression of key genes through time,

and that key responsive genes will play significant roles in meditating caste

differentiation and/or caste-regulatory signaling. Our approach involved determining

the impacts of the four treatments on both phenotypic caste differentiation and the

expression of forty-nine candidate and reference genes during the first 10 days of

differentiation from worker to presoldier, with subsequent validation of reference genes

and post-hoc analyses to identify genes with significant differential expression among

treatments. Here, we identify and discuss a number of responsive genes from three

categories (chemical production / degradation, hemolymph protein, and developmental)

with significant links to caste differentiation.


Phenotypic responses

Phenotypic bioassays (Fig. 1A) showed that the combination of JH III + SHE

significantly increased presoldier development when compared to JH III alone. A two-

way ANOVA and adjusted LS means were used for analysis (whole model F=24.092,

df=14, P<0.0001; treatment F= 54.32, df=4, P<0.0001; colony F=24.140, df=2,

P<0.0001; treatment*colony F=11.513, df=8, P<0.0001). Variation was observed

between the different colonies tested, with Colony 1 showing the greatest presoldier

induction response to JH III (40%) and JH III+SHE (80%). But, as seen in previous

research [19], the overall trend was the same in that JH III+SHE increased presoldier

differentiation compared to JH III alone and no presoldiers formed in the acetone-

treated controls, SHE-alone treatments, or live soldier treatments (Fig. 1A).

Because no phenotypic effects were observed with live soldier treatments in the

small-format dish assays noted above, we conducted post-hoc presoldier induction

assays using larger groups of workers (Fig. 1B). Our objective was to determine if

soldiers could inhibit natural presoldier formation in the absence of ectopic JH, using

greater numbers of workers (100 termites) over a longer period of time (30 days). Two

treatments were tested: (1) 100 workers + 0 soldiers, and (2) 90 workers + 10 soldiers.

Interestingly, no presoldiers formed after 30 d. in treatments that included 10% soldiers

at the beginning of assays; and conversely, presoldiers appeared only in the treatments

that included 100% workers at the beginning of assays. This finding verifies that R.

flavipes soldiers are capable of inhibiting worker-to-presoldier differentiation, and

provides evidence that is directly supportive of the soldier and SHE impacts on gene

expression presented below.

Reference gene selection

To accurately determine relative gene expression in totipotent workers, we chose

three reference genes that had stable expression across all treatments and colonies

(Stero-1, LIM, and Mev-1). These reference genes were selected by comparing the

standard deviation of the raw Ct values for all 49 genes across treatments (Additional

file 1: Table Si). This determination is important because it allows normalization of the

expression of target genes (n=46) to reference genes (n=3) that have stable expression

across all treatments and colonies.

Gene expression overview

All target and reference genes investigated in this study have been annotated

based on significant translated identity to insect sequences deposited in the Genbank nr

and EST databases. Full-length gene names are provided in Additional file 2: Table S2.

All reported gene expression data represent the average of three independently sampled

and replicated R. flavipes colonies. Gene expression changes in response to all

treatments were determined via qRT-PCR. To identify genes with significant

differential expression across treatments, two-way ANOVAs were used with adjusted

LS means and FDR correction on normalized CT (ACT) values (Additional files

3,4,5,6: Tables S3, S4, S5, S6). Additionally, gene expression at three days (1, 5 and

10) was analyzed separately using the ANOVA procedure noted above. For a large

proportion of the genes tested there was a significant colony effect. This was to be

expected because (i) there was also a significant colony effect in the phenotypic

bioassay and (ii) the colonies tested each have different mitochondrial haplotypes (see

later). Colony effects were compensated for by using adjusted LS means in the


To easily visualize gene expression responses, genes showing significant

expression changes across treatments were organized by day into heat maps (Fig.

2,3,4). Genes with similar expression profiles are horizontally clustered together. By

clustering genes in this manner we are able to identify groups of genes that respond

similarly and putatively belong to the same gene networks (also see Additional file 7:

Table S7).

Gene expression: day 1

As shown in the Day 1 heat map (Fig. 2), 17 out of the 46 genes that were tested

showed significant differences in their expression across treatments (Additional file 4:

Table S4). Day 1 receives focus here because we presume Day 1 responsive genes to

be important immediate-early responders. Three main clusters of genes were identified,

with sub groupings of genes in some clusters. Genes in group IIB overall were affected

by SHE and live solder treatments, with IIB2ii genes Carbx-1, 3!v, ,Win. B-actin, fl-tube,

R-Pro, ATPase, and HMG all being down-regulated with live soldiers. Genes in group

IIB1, NADH and nanos, were up-regulated in live soldier treatments. Group IIB2i

genes Hex-2 and 18s were down-regulated in SHE treatments. The P450 protein coding

genes in group IIA, CYP15FI, CYP4C48, CYP6G?, and CYP4C47 were down

regulated with JH III and JH III+SHE treatments, while group I genes, CYP4C46 and

CYP4U3, were up-regulated with JH III. These Day 1 results reveal a number of early

response genes in totipotent workers that are both up and down-regulated in response to

the different treatments. Perhaps most importantly, a number of P450 genes that may

play roles in semiochemical or hormone processing were differentially expressed

among treatments at this early time point.

Gene expression: day 5

Five days into assays, 23 genes showed significant differential expression among

the five treatments (Fig. 3, Additional file 5: Table S5). A larger number of genes

showed significant variation in expression at this point compared with Days 1 and 10,

with the majority of the genes showing down-regulated responses to most treatments.

Genes in group IIB2iib3, CYP4C44vl, broad, and APO had a slight expression

increase with JH, while being down-regulated with SHE and live soldier treatments.

Group IIB2iib2 genes, CoxII, HSP, and Shp displayed an up-regulation with live

soldier treatments. Genes SH3, NADH and CYP15F1, in group IIB2iibl, were down-

regulated with JH+SHE and SHE treatments. Group IIB2iia genes, Famet-2, Carbx-1,

CYP4U3, Carbx-2, and To-F were all down-regulated with live soldier treatments. Bic

and nanos, in group IIB2i, were down-regulated with JH III, JH III+SHE and SHE

treatments. Genes that clustered into group IIB1, Hex-2, Hex-1, and CYP4C46 were

up-regulated with JH III and JH+SHE treatments. Finally, two hemolymph protein

coding genes, Vit-1 (IIA) and Vit-2 (I) were up-regulated with JH III and JH+SHE and

down-regulated with SHE and live soldier treatments. Five days into assays represents

the middle of the worker-to-solder differentiation process [8]. Therefore, genes

identified at this time point could be playing mid-level signaling roles in the caste

differentiation cascade. The hemolymph protein coding genes Vit-1, Vit-2, Hex-1 and

Hex-2, have been linked to caste differentiation in past research in termites and honey

bees [17,18,33-38]. Thus, their differential expression during the worker-to-presoldier

differentiation process was expected, and serves to validate our approach for

determining gene expression during differentiation.

Gene expression: day 10

On the last day investigated (Day 10) nineteen genes showed significant variation

in expression across treatments (Fig. 4, Additional file 6: Table S6). Live soldier effects

were not investigated at this time point due to limitations imposed by the 96-well PCR

plate format and an inability to include all treatments for individual genes on a single

plate. The group II genes Epox-1 and Vit-2 were up-regulated with JH III and JH+SHE

treatments. Genes in group IB3iib, CYP15F1, Shp, and Tro-1 were down-regulated

with JH+SHE treatment, while Hex-1 and To-F (IB3iia) were down regulated with JH

III and JH+SHE treatments. The putative ribosomal RNA coding 18s gene was down-

regulated in live soldier treatments (IB3i). Group IB2ii genes CYP4U3, 28s, and

CYP4C46 were up-regulated with JH III but down-regulated with JH+SHE treatment,

while genes in group IB2i, Lprs, Famet-1, and NADH were down-regulated with JH III.

Genes that clustered in group IB1, Myosin, APO, and broad were up-regulated with

JH+SHE treatment. Finally group IA genes, Carbx-1 and SH3, were down-regulated

with JH+SHE treatment. These Day 10 results reveal a number of potential late

responding genes that are both up- and down-regulated in response to the different

treatments. Thus, these late responding genes likely are part of multiple pathways that

are involved in the later stages of the worker-to-presoldier differentiation process.

Uniformly responsive genes and hierarchical clustering

Across days 1, 5 and 10, four genes showed consistent, significant differential

expression: CYP15FI, CYP4C46, CYP4U3, and NADH. This finding suggests that


these four genes are of broad general importance in worker-to-soldier caste

differentiation and / or caste regulation / homeostasis.

Finally, gene expression results were hierarchically (vertically) clustered by

treatment across days based on the expression patterns of all genes (Fig. 5a,b,c).

Results for Day 1 and 5 are similar with control and live soldier treatments clustering

together, and JH III and JH+SHE treatments clustering together. Day 10 results show a

different clustering pattern in which control and SHE treatments cluster together, and

the JH and JH+SHE treatments show a more distant relationship. These results suggest

that effects of the different treatments on genes and gene networks are not temporally

fixed, but change through time.


Social organisms, including hemimetabolous lower termites like R. flavipes,

utilize phenotypic plasticity to achieve caste polyphenism and division of labor.

Because all termite colony members share essentially the same genetic background,

they rely on differential gene expression for caste differentiation [3]. The development

of termites along alternate caste pathways is regulated by a number of interacting

intrinsic and extrinsic factors (e.g., [18]); however, detailed global gene expression

responses through the differentiation process have been lacking, and no prior studies

have investigated nestmate or primer-pheromone-responsive gene expression in


This study correlates clear phenotypic effects of R. flavipes hormones,

semiochemicals, and social treatments with patterns of gene expression and reveals

potentially important candidate caste-regulatory genes. Changes in expression of

several genes having homology to other well-characterized developmental and

-11 -

hormone / semiochemical biotransformation genes were detected in association with

the different treatments. Several gene networks apparently important in caste

differentiation and social interactions were also identified.

The model bioassay system used here induces changes in phenotype, and gene /

protein expression, and has been used repeatedly to monitor and elucidate mechanisms

of caste differentiation, specifically the worker-to-soldier transition [17-19,34,39,40].

Here, we investigated the effects of specific hormone / semiochemical (JH III, JH

III+SHE, SHE) and socio-environmental conditions (live soldiers) on soldier caste

differentiation and gene expression by totipotent termite workers. Although there are

certainly other semiochemical and socio-environmental conditions that could play a

role in worker-to-soldier differentiation, we focused on the components listed above

because they build concisely on preceding work.

Phenotypic assay results were similar to past findings in that JH III induced

presoldier formation, JH III+SHE synergistically increased presoldier formation, and

SHE alone had no effect on presoldier development [19]. The addition of live soldiers

to the bioassay did not impact soldier formation. Because presoldier differentiation

only occurs naturally in larger groups of workers over longer periods of time (see Fig.

1B and [21]), our small-scale model bioassay cannot allow for determination of any

inhibitory effects by soldiers at the whole organism level. Nonetheless, our results

provide good support to the hypothesis that SHE, or a component of it, acts with JH as

a primer pheromone to help regulate caste proportions within termite colonies.

This research also monitored phenotypic effects in concert with the expression

patterns of multiple genes. This was accomplished with destructive sampling of some

assay replicates for RNA isolation, while allowing others to proceed without

disturbance. The typical worker-to-presoldier differentiation process takes


approximately 15 days. To capture potential expression changes up to apolysis, gene

expression levels were monitored at 1, 5, and 10 days post treatment, which are

considered early, middle and late time points, respectively, in the presoldier

developmental transition. A total of forty-nine genes were investigated across three

replicate colonies. Statistically significant genes that passed the FDR cutoff were

clustered together based on expression pattern (Fig. 2,3,4). As discussed below, three

main groups of responsive genes were identified: (i) chemical production / degradation,

(i) hemolymph protein coding, and (iii) developmental.

Chemical production / degradation genes

Chemical production and degradation genes code for enzymes that are potentially

responsible for the production and / or degradation of many types of semiochemicals in

termites, including hormones such as JH and ecdysone, as well as the soldier head

terpenes y-cadinene and y-cadinenal. The three groups of genes included in this

category are cytochrome P450, hydrolytic, and mevalonate pathway protein-coding


Cytochrome P450s are known for their role in the oxidation of endogenous and

xenobiotic substrates including hormones, pheromones, insecticides, and secondary

plant compounds [41,42]. Specifically, P450s have been shown to play a role in the

biosynthesis and metabolism of morphogenic hormones (JH, ecdysone) and terpenoids

[41]. On Day 1, two groups of P450s were differentially expressed. In the first group

(IIA), CYP15FI, CYP4C48, CYP6G? and CYP4C47 were down-regulated with JH III

and JH III+SHE treatments, while in the second group (I), CYP4C46 and CYP4U3 were

up-regulated with JH III and JH III+SHE treatments. This opposite expression profile


of the two P450 groups suggests they have different functions, likely acting on multiple


Past research has identified P450s that play significant roles in JH biosynthesis

and degradation in insects. In the cockroach, Diploptera punctata, CYP15A] epoxidizes

methyl farnesoate to form JH III [43]. In the present study, those P450s that were

down-regulated with JH treatment (CYP15FI, CYP4C48, CYP6G? and CYP4C47)

could have a similar function. Insect P450s have also been shown to play a role in the

degradation of JH III, as is the case with CYP4C7, which converts JH III to 12-trans-

hydroxy JH III in Diploptera punctata [44,45]. The group I P450s (CYP4C46 and

CYP4U3) that were up-regulated in the present study could be playing this role and / or

the group of genes that were down-regulated could be inactivated, potentially blocking

the worker-to-soldier transition.

Juvenile hormone metabolism is also potentially mediated by hydrolytic

enzymes, including JH esterases and epoxide hydrolases [46]. Three genes having

homology to JH esterases and epoxide hydrolases displayed significant expression

differences among treatments. Carbx-1 has highest homology to a JH esterase of the

wood-feeding beetle P'a, ,'ih a hilaris (BAE94685) [47]. The Carbx-2 gene has

highest homology to a JH esterase of the sawfly Athalia rosae (BAD91555). Both the

Carbx-1 and Carbx-2 genes also have significant homology to honey bee JH esterases

[48] as described by Mackert et al. [49]. Both genes are expressed in the gut, and thus

could be acting on JH acquired via trophallaxis, but also could play digestive roles by

hydrolyzing lignin or hemicellulose carboxyl esters [48].

Epoxide hydrolases are known to degrade JH by hydrolyzing the epoxide bond

that is formed by CYP15 action as described above. The epoxide hydrolase studied

here, Epox-1, has significant homology to an Aedes Cat ,lvli epoxide hydrolase


(XP_001651935), among others. If Epox-1 is acting as a JH epoxide hydrolase, its

observed up regulation could contribute to the degradation or inactivation of any

endogenous remaining JH prior to apolysis or ecdysis, which is expected to occur at

around day 10 in our model presoldier induction assays [18].

The production of JH and other sesquiterpenes derived from the mevalonate

pathway is important to termite colony success, not only for development and caste

differentiation, but also for production of defensive chemicals and pheromones that

possess a sesquiterpene backbone [29,50]. Both up- and down-regulation of genes in

the mevalonate pathway can significantly impact the production of JH and pheromones

[30,51]. In the present study, five mevalonate pathway genes were investigated: Famet-

1, Famet-2, Famet-3, Mev-1, and HMG. Two genes homologous to famesoic acid

methyl transferases (Famet-1, Famet-2) showed differential expression. Farnesoic acid

methyl transferase methylates farnesoic acid, producing the immediate JH precursor

methyl farnesoate [50]. RNAi-mediated knockdown of this gene in Tribolium

castaneum has led to reduced JH levels and precocious molting [52]. R. flavipes Famet-

1 shares strongest homology to a FAMet protein from the hymenopteran Melipona

scutellaris (AM493719) [53]. Our results revealed that JH causes increased Famet-1

expression. Increased expression of this gene could theoretically increase JH

biosynthesis rates and enable soldier formation. Our results also revealed that the

presence of live soldiers down-regulates Famet-2 gene expression, which theoretically

could lead to reduced JH production and decreased worker-to-soldier differentiation.

In general, these results suggest that JH III causes up-regulation of mevalonate

pathway genes, while live soldiers are suppressive. Consistent with our phenotypic

bioassay results, suppression of the mevalonate pathway by live soldiers would likely


result in reduced pathway products, such as JH, resulting in reduced JH titers and

subsequent reductions in soldier caste differentiation.

Hemolymph protein coding genes

Four hemolymph protein coding genes, Hex-1, Hex-2, Vit-1, and Vit-2 showed

significant differential expression through all assay days. These four genes are

important in caste differentiation and sociobiology for a number of social insects;

therefore, it was not surprising that they showed responsiveness in our experiments.

The termite hexamerin genes have been shown to act as part of an environmentally

responsive socio-regulatory mechanism that affects the activity of JH, possibly limiting

its availability [18,33,34,54].

Two other hemolymph protein genes, Vit-1 and Vit-2, were up-regulated with

JH and JH +SHE treatments at Day 5, but only Vit-2 was differentially expressed at

Day 10. Throughout the experiment, both Vit-1 and Vit-2 genes displayed a high

amount of variability among treatments and replicates. One explanation for such

variance is the inclusion of both sexes of worker termites in assays. In most insects,

vitellogenin (Vg) serves as a female-specific yolk precursor protein that functions in

oocyte provisioning. However, Vg has also been shown to play a role in social insect

caste regulation; for example, Vg in female honeybee workers, has been shown to

interact with JH. Specifically, higher JH levels and lower Vg levels increased the

transition from nursing to foraging behavior by worker bees [35], while a reduction of

JH delayed the onset of foraging [55]. Honeybee workers with RNAi-suppressed Vg

levels performed foraging behaviors earlier than untreated workers [36,37]. Nutrition

has also been shown to affect Vg and JH by regulating the transition from nursing to

foraging [56]. Finally, Vg has been shown to affect queen honeybee longevity by


interacting with insulin signaling [57]. Together, these findings suggest that honeybee

Vg has been co-opted away from reproduction to serve as a regulator of caste

behavioral polyethism [38]. Results of the current study, showing that Vit-1 and Vit-2

are up-regulated with JH and JH +SHE treatments, but down-regulated with SHE and

live soldier treatments, suggest interesting possibilities with respect termite vitellogenin

and caste polyphenism.

Developmental genes

The dramatic morphological change that occurs as worker termites become

soldiers requires significant body plan rearrangement [3]. The soldier termite's large

mandibles and their associated muscles represent a large change from the smaller head

and reduced muscle mass present in worker termites [58,59]. Thus, it is likely that

multiple genes are required to coordinate and achieve this transition [60]. Six

developmental genes from two groups, cytoskeletal/ structural and body-plan, showed

significant differential expression in the current study.

The cytoskeletal/ structural protein coding gene "fl-tube" was significantly

differentially expressed at Day 1 among treatments. f-tubulins are also hormone-

responsive and have been linked to the production of ecdysteroids in Manduca sexta

[61,62]. a- and f-tubulin genes were also identified in Bombyx mori from several EST

libraries linked to imaginal wing disk metamorphosis and 20-hydroxyecdysone [63,64],

suggesting roles in restructuring during adult wing formation. Our findings suggest

potential roles for R. flavipes fl-tube in either soldier head muscle function or possibly

ecdysone-linked developmental-regulatory processes.

A number of developmental/ body plan genes also showed significant differential

expression. One body plan gene, broad (BTB/POZ) [8], which is homologous to broad


(br) transcription factor genes of the hemimetabolous and holometabolous insects

(Oncopeltusfasciatus and T. castaneum), was up-regulated at Day 5 with JH and

JH+SHE treatment and at Day 10 with JH+SHE treatment. Erezylimaz et al. [65] used

RNAi to silence the br gene in 0. fasciatus, causing an additional immature molt.

Erezylimaz et al. suggested that br is required for morphogenesis, and that its

expression is regulated by JH. RNAi silencing of br in T. castaneum caused similar

results [66]. If br is acting in the same manner in termites, up-regulation of the gene by

JH+SHE would promote the worker-to-soldier transition, which is in agreement with

phenotypic bioassay results showing increased presoldier formation in the JH+SHE



The research presented here demonstrates for the first time the influence that the

SHE blend, live soldier caste members, and JH together have on phenotype and gene

expression of totipotent termite workers (Fig. 6). To summarize phenotypic assay

results (Fig. 6A): (i) JH III induced significant presoldier differentiation, (ii) JH + SHE

induced significantly higher levels of presoldier differentiation, (iii) the crude SHE

blend by itself did not have any observable phenotypic effects, and (iv) live soldiers

inhibited presoldier formation in the absence of ectopic JH. In support of primer

pheromone hypotheses initially proposed by Liischer [67] and further developed by

Henderson [68], our results provide the first evidence that the soldier caste has direct

impacts on caste-regulatory gene networks, and subsequently, worker caste

differentiation. Significant responsive gene categories identified here include chemical

production / degradation genes, hemolymph protein coding genes, and developmental

genes (Fig. 6b). Past reports (e.g., [16]) and the present research (Fig. 1B) have


demonstrated that live soldiers do indeed inhibit natural presoldier formation. These

results, in addition to the current gene expression findings, support earlier hypotheses

that live soldiers act as part of a negative feedback loop, inhibiting new soldier

formation by regulating the expression of genes important for caste differentiation (Fig.

6b) [16,24,68]. Recent findings have further revealed that y-cadinene and y-cadinenal

levels increase in workers that are held with soldiers (MR Tarver, unpublished results),

which lends significant strength to the results presented here that show live soldier and

SHE impacts on gene expression. The next steps in this research will follow up on

these observations by investigating the impacts of pure y-cadinene and y-cadinenal on

phenotypic caste differentiation and on the expression of responsive genes identified in

the current study.

This research provides important new evidence of impacts on nestmate gene

expression by live termite soldiers and crude soldier head extracts. While further

research is needed to resolve the roles of soldiers and SHE blend components in termite

caste regulation (via RNA interference, gene expression localization, further analysis of

SHE constituents, investigating impacts of SHE constituents on gene expression, or

using whole-genome micro-arrays or next-generation transcriptome sequencing) the

findings of this study provide a solid foundation on which to conduct further

translational studies.


R. flavipes colonies were collected from different locations near Gainesville,

Florida USA. Termites were held in the laboratory for at least two months before use in

bioassays. Colonies were maintained in darkness within sealed plastic boxes, at 220C.

All colonies contained male and female neotenic reproductive. Termites were


considered true workers if they did not possess any sign of wing buds or distended

abdomens. Termites were identified as R. flavipes by a combination of soldier

morphology [69], and 16S mitochondrial-ribosomal RNA gene sequencing [70]. The

partial mitochondrial 16S sequences of the four colonies used were deposited,

respectively, in Genbank under accession numbers: FJ265704 (colony-1 "GB 1"),

FJ627943 (colony-2 "K2"), FJ265705 (colony-3 "A8") and GQ403073 (colony-4

"K5"). Using the 16S mitochondrial sequences, colony 1 was 99% identical to

mitochondrial haplotypes F22 and Fl (EU259755, EU259734), colony 2 was 98%

identical to haplotype F20 (EU259753), colony 3 was 96% identical to haplotypes F34,

28, and 21 (EU259767, EU259761, EU259754) and colony 4 was 98% identical to

haplotype F20 (EU259753).

Phenotypic bioassays

Small-scale dish bioassays were conducted at 270C as described previously

[19,39]. Paired paper towel sandwiches were treated with acetone (controls), JH III, or

SHE treatments delivered in acetone. JH III (75% purity; Sigma; St. Louis, MO) was

provided at a rate of 112.5 pg per dish in a volume of 200 p1 acetone. The JH III rate

was chosen based on maximal efficacy with minimal mortality observed in previous

concentration range studies [39]. After solvent evaporation, paper towel sandwiches

were placed in 5 cm plastic Petri dishes and moistened with 150 1p of reverse osmosis

water. Fifteen worker termites were placed in each assay dish. Live solider treatments

consisted of holding two live soldiers with 15 workers from the same colony. Every

five days, termites were counted, presoldier formation was noted, and water was added

if needed. Each treatment was monitored for 25 days.


For larger-scale soldier inhibition assays, two treatments were examined to assess

the influence of live solders on presoldier formation in large group format of 100

termites per dish. Control treatments included 100 workers only; whereas, soldier

treatments included 90 workers + 10 soldiers. Termites from a single colony (K5) were

used, and assays were run in large Petri dishes (9 cm diam.). Caste composition and

survival were monitored every ten days for a total of 30 days. Each treatment was

replicated five times and results pooled (avg SEM) for reporting.

Preparation of solider head extracts

Soldier head extracts were prepared as described in Tarver et al. [19]. Soldiers

(80-150 total) were isolated from lab colonies, and their heads removed and

homogenized in 5 mL acetone using a Tenbroeck glass homogenizer. To remove

particulate matter, the homogenate was fractionated by passing it through a glass

Pasteur pipette filled with approximately 250 mg of silica gel (60-200 mesh) on top of a

glass wool plug. The SHE was eluted with 10 column volumes of acetone and brought

to 50 ml with acetone in a volumetric flask.

Gene expression bioassays

A total of five different treatments were tested including acetone controls (300

a1), JH III (200 11 acetone containing 112.5 pg JH III), JH III+SHE (112.5 pg JH III in

acetone + 1.5 soldier head equivalents in acetone), SHE (1.5 head equivalents extracted

in acetone), and live soldiers (two per assay replicate). Each treatment was replicated

five times for colony-1 and six times for colonies-2 and 3 (GB1, K2, and A8

respectively). Three biological replicates were used per treatment for colony 1 and four

for colonies 2 and 3. Additional replicates for colonies 2 and 3 were added to improve

-21 -

statistical power. Samples of 15 termites were collected for destructive sampling at

days 0, 1, 5, and 10. Collected samples were immediately frozen at -800C.

RNA isolation and cDNA synthesis

Total RNA was isolated from frozen samples using the SV total RNA Isolation

System (Promega; Madison, WI) according to the manufacturer's protocol. Whole

body RNA extracts were isolated from all 15 worker termites included in each bioassay

dish. The amount of RNA was quantified by spectrophotometry and equal amounts of

RNA were used in cDNA synthesis reactions. First-strand cDNA was synthesized using

the iScript cDNA synthesis Kit (Bio-Rad; Hercules, CA) according to the

manufacturer's protocol.

Gene expression

The 49 candidate and reference genes were identified in recent R. flavipes

sequencing projects and were chosen based on their homology to developmental or JH

biosynthesis / metabolism genes [8,17,34,40,54,71]. The identity of all 49 PCR

products was verified by direct sequencing. Quantitative real-time PCR (qRT-PCR)

was performed using the iCycler iQ real-time PCR detection system (Bio-Rad) with

SYBR-green product tagging (similar to [8,34]). cDNA, obtained as described above,

was used as the qRT-PCR template. Gene specific primers are listed in Additional file

2: Table S2. Eleven total biological replicates were conducted for qRT-PCR (three

from colony-1, and four each from colonies 2 and 3). Average Ct values of three

technical replicates were pooled for analysis to represent each biological replicate.


Reference gene selection

To select appropriate reference genes, all of the Ct values across all colonies,

treatments, biological replicates, and technical replicates for each gene were analyzed

to identify genes with the least amount of variation in expression (see [17,34]). Three

genes with the lowest standard deviation were chosen for use as reference genes: Stero-

1, LIM, and Mev-1 (Additional file 1: Table Si).

Data and statistical analyses

Relative expression of target genes was calculated by comparing the average of

the three technical replications first normalized to the reference genes and then

normalized to the control treatment using the 2-ActAct method [72]. Normalized

expression values (2-ActAct) from all colony replicates were initially analyzed using the

microarray visualization software ArrayStar (DNASTAR, Inc, Madison, Wisconsin,

USA). To identify potential gene networks, genes with significant differential

expression were clustered hierarchically using Euclidean distance metrics and centroid

linkage for each day (1, 5 and 10) using ArrayStar.

To identify similarities between treatments, all genes were clustered

hierarchically using Euclidean distance metrics and centroid linkage for each day (1, 5

and 10) using Array StarTM software.

To determine significantly differentially expressed genes, CT expression values

for target genes were normalized to the CT values from the reference genes (ACT). A

two-way ANOVA, with adjusted least squares (LS) means and false discovery rate

(FDR) correction was used to distinguish significantly differentially expressed genes

among treatments using JMP statistical software (SAS Institute, Cary, NC, USA)


(Additional file 3: Table S3). Tukey's HSD test was used for separating means by

treatment for each gene.

Authors' contributions

MRT conceived the study design, performed all the experimental procedures and was
the primary author of the manuscript. XZ participated in the design of the study. MES
conceived the study design, analyzed data, and critically revised the manuscript. All the
authors read and approved the final manuscript.

We thank Aur6lien Tartar for assistance with gut gene bioinformatics, Daniel Hahn for
valuable discussions on expression analysis, and Alan Lax and Dunhua Zhang for
critical reading of manuscript drafts. This work was supported by CSREES-USDA-NRI
grant No. 2007-35607-17777 to MES and XZ, by The Consortium for Plant
Biotechnology Research, Inc. and DOE Prime Agreement No. DE-FG36-02GO12026
to MES, and a University of Florida IFAS Innovation Grant to MES.

1. Nijhout FH: Control mechanisms of polyphenic development in insects.
Bioscience 1999, 49:181-192.

2. Nijhout FH: Development and evolution of adaptive polyphenisms. Evol Dev 2003,

3. Miura T: Proximate mechanisms and evolution of caste polyphenism in social
insects: From sociality to genes. Ecol Res 2004, 19:141-148.

4. Wilson EO: The insect societies. Cambridge: Belknap Press; 1971.

5. Hayashi Y, Lo N, Miyata H, Kitade O: Sex-linked genetic influence on caste
determination in a termite. Science 2007, 318:985-987.

6. Laine LV, Wright DJ: The life cycle of Reticulitermes spp. (Isoptera:
Rhinotermitidae): what do we know? Bull Entomol Res 2003, 93:267-378.

7. Buchli H: L'origine des castes et les poetntialities ontogeniques des termites
europeens du genre Reticulitermes holmgren. Ann Sci Nat Zool 1958, 20:261-429.

8. Scharf ME, Wu-Scharf D, Pittendrigh BR, Bennett GW: Caste- and development-
associated gene expression in a lower termite. Genome Biol 2003, 4(R62).

9. Howard RW, Haverty MI: Production of soldiers and maintenance of soldier
proportion by laboratory experimental groups of Reticulitermes flavipes (Kollar)
and Reticulitermes virginicus (Banks) (Isoptera: Rhinotermitidae). Insect Soc

10. Lefeuve P, Christian Bordereau Soldier formation regulated by a primer


pheromone from the soldier frontal gland in a higher termite, Nasutitermes lujae.
Proc Natl Acad Sci 1984, 81:7665-7668.

11. Waller DA, La Fage JP: Environmental influence on soldier differentiation in
Coptotermes formosanus. Insectes Soc 1988, 35:144-152.

12. Horiuchi S, Yamamura N, Abe T: Soldier production strategy in lower termites:
from young instars or old instars? J Theo Biol2002, 218(195-205).

13. Fei H, Henderson GR: Formosan subterranean termite wood consumption and
worker survival as affected by temperature and soldier proportion. Environ
Entomo12002, 31:509-514.

14. Liu Y, Henderson G, Mao L, Roger A. Laine Seasonal variation of juvenile hormone
titers of the formosan subterranean termite, Coptotermes formosanus
(Rhinotermitidae). Environ Entomol2005, 34(3):557-562.

15. Park YI, Raina AK: Juvenile hormone III titers and regulation of soldier caste in
Coptotermes formosanus. J Insect Physiol 2004, 50:561-566.

16. Mao L, Henderson G, Liu Y, Laine RA: Formosan subterranean termite (Isoptera:
Rhinotermitidae) soldiers regulate juvenile hormone levels and caste
differentiation in workers. Ann Entomol Soc AM 2005, 98(3):340-345.

17. Zhou X, Song C, Grzymala TL, Oi FM, Scharf ME: Juvenile hormone and colony
conditions differentially influence cytochrome P450 gene expression in the
termite Reticulitermes flavipes. Insect Mol Biol 2006, 15(6):749-761.

18. Scharf ME, Buckspan CE, Grzymala TF, Zhou X: Regulation of polyphenic
differentiation in the termite Reticulitermes flavipes by interaction of intrinsic
and extrinsic factors. J Exp Biol 2007, 24:4390-4398.

19. Tarver MR, Schmelz EA, Rocca JR, Scharf ME: Effects of soldier-derived terpenes
on soldier caste differentiation in the termite Reticulitermes flavipes. J Chem Ecol
2009, 35(2):256-264.

20. Okot-Kotber BM, Prestwich GD, Strambi A, Strambi C: Changes in morphogenetic
hormone titers in isolated workers of the termite Reticulitermes flavipes (Kollar).
Gen Comp Endocrinol 1993, 90:290-295.

21. Elliott KL, Stay B: Changes in juvenile hormone synthesis in the termite
Reticulitermes flavipes during development of soldiers and neotenic
reproductive from groups of isolated workers. J Insect Physiol2008, 54:492-500.

22. Korb J, Hoffmann K, Hartfelder K: Endocrine signatures underlying plasticity in
postembryonic development of a lower termite, Cryptotermes secundus
(Kalotermitidae). Evol Dev 2009, 11(3):269-277.

23. Park YI, Raina AK: Factors regulating caste differentiation in the formosan
subterranean termite with emphasis on soldier formation. Sociobiology 2003, 41:1-

24. Park YI, Raina AK: Regulation of juvenile hormone titers by soldiers in the
Formosan subterranean termite, Coptotermes formosanus. J Insect Physiol2005,

25. Okot-Kotber BM, Ujvary I, Mollaaghababa R, Szurdoki F, Matolcsy G, Prestwich GD:
Physiological influence on fenoxycarb pro-insecticides and soldier head extracts
of various termite species on soldier differentiation in Reticulitermes flavipes
(Isoptera). Sociobiology 1991, 19(1):77-89.


26. Korb J, Roux EA, M. Lenz Proximate factors influencing soldier development in
the basal termite Cryptotermes secundus (Hill). Insect Soc 2003, 50:299-303.

27. Wilson EO, Bossert WH: Chemical communication among animals. Recent Prog
Horm Res 1963, 19:673-716.

28. Tittiger C: Functional genomics and insect chemical ecology. J Chem Ecol2004,

29. Seybold SJ, Tittiger C: Biochemistry and molecular biology of de novo isoprenoid
pheromone production in the scolytidae. Annu Rev Entomol2003, 48:425-453.

30. Keeling CI, Blomquist GJ, Tittiger C: Coordinated gene expression for pheromone
biosynthesis in the pine engraver beetle, Ips pini(Coleoptera: Scolytidae).
Naturwissenschaften 2004, 91:324-328.

31. Keeling CI, Bearfield JC, Young S, Blomquist GJ, Tittiger C: Effect of juvenile
hormone on gene expression in the pheromone-producing midgut of the pine
engraver beetle, Ips pini. Insect Mol Biol2006, 15(2):207-216.

32. Grozinger CM, Sharabsh NM, Whitfield CW, Robinson GE: Pheromone-mediated
gene expression in the honey bee brain. Proc NatlAcad Sci2003, 100(2):14519-

33. Zhou X, Oi FM, Scharf ME: Social exploitation of hexamerin: RNAi reveals a major
caste-regulatory factor in termites. Proc Natl Acad Sci2006, 103(12):4499-4504.

34. Zhou X, Tarver MR, Scharf ME: Hexamerin-based regulation of juvenile hormone-
dependent gene expression underlies phenotypic plasticity in a social insect.
Development 2007, 134(3):601-610.

35. Amdam GV, Norberg K, Hagen A, Omholt SW: Social exploitation of vitellogenin.
Proc Natl Acad Sci 2003, 100(4):1799-1802.

36. Nelson CM, Ihle KE, Fondrk MK, Page Jr. RE, Amdam GV: The gene vitellogenin has
multiple coordinating effects on social organization. PLOS Biol2007, 5(3):0673-

37. Antonio DSM, Guidugli-Lazzarini KR, Nascimento AMD, Simoes ZLP, Hartfelder K:
RNAi-mediated silencing of vitellogenin gene function turns honeybee (Apis
mellifera) workers into extremely precocious foragers. Naturwissenschaften 2008.

38. Denison R, Raymond-Delpech V: Insights into the molecular basis of social
behaviour from studies on the honeybee, Apis mellifera. Invert Neurosci2008, 8:1 -

39. Scharf ME, Ratliff CR, Hoteling JT, Pittendrigh BR, Bennett GW: Caste differentiation
responses of two sympatric Reticulitermes termite species to juvenile hormone
homologs and synthetic juvenoids in two laboratory assays. Insect Soc 2003,

40. Scharf ME, Wu-Scharf D, Pittendrigh BR, Bennett GW: Gene expression profiles
among immature and adult reproductive castes of the termite Reticulitermes
flavipes. Insect Mol Biol 2005, 14(1):31-44.

41. Andersen JF, Walding JK, Evans PH, Bowers WS, Feyereisen R: Substrate
specificity for the epoxidation of terpenoids and active site topology of house fly
cytochrome P450 6A1. Chem Res Toxicol 1997,10:156-164.


42. Feyereisen R: Insect cytochrome P450. In: Comprehensive Molecular Insect Science
[vol 4]: Biochemistry and Molecular Biology. Edited by latrou GK, Gill SS, vol. 4.
Amsterdam: Elsevier; 2005: 1-77.

43. Helvig C, Koener JK, Unnithan GC, Feyereisen R: CYP15A1, the cytochrome P450
that catalyzes epoxidation of methyl farnesoate to juvenile hormone III in
cockroach corpora allata. Proc Natl Acad Sci2004, 101(12):4024-4029.

44. Sutherland TD, Unnithan GC, Andersen JF, Evans PH, Murataliev MB, Szabo LZ,
Mash EA, Bowers WS, Feyereisen R: A cytochrome P450 terpenoid hydroxylase
linked to the suppression of insect juvenile hormone synthesis. Proc NatlAcad
Sci1998, 95:12884-12889.

45. Sutherland TD, Unnithan GC, Feyereisen R: Terpenoid w-hydroxylase (CYP4C7)
messenger RNA levels in the corpora allata: a marker for ovarian control of
juvenile hormone synthesis in Diploptera punctata J Insect Physiol 2000, 46:1219-

46. Roe RM, Kallapur V, Linderman RJ, Viviani F, Harris SV, Walker EA, Thompson DM:
Mechanism of action and cloning of epoxide hydrolase from the cabbage looper
Tricoplusia ni. Arch. Insect Biochem. Phys. 1996, 32:527-535.

47. Munyiri FN, Ishikawa Y: Molecular cloning and developmental expression of the
gene encoding juvenile hormone esterase in the yellow-spotted longicorn beetle,
Psacothea hilaris. Insect Biochem Molec Biol 2007, 37:497-505.

48. 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. Phys. 2010. 73: 30-48.

49. Mackert A, Nascimento AMd, Bitondi MMG, Hartfelder K, Simoes ZLP: Identification
of a juvenile hormone esterase-like gene in the honey bee, Apis mellifera L.-
expression analysis and functional assays. Comp Biochem Physiol B 2008, 150:33-

50. Belles X, Martin D, Piulachs MD: The mevalonate pathway and the synthesis of
juvenile hormone in insects. Annu Rev Entomol2005, 50:181-199.

51. Tillman JA, Lu F, Goddard LM, Donaldson ZR, Dwinell SC, Tittiger C, Hall GM, Storer
AJ, Blomquist GJ, Seybold SJ: Juvenile hormone regulates de novo isoprenoid
aggregation pheromone biosynthesis in pine bark beetles, Ips, SPP., through
transcriptional control of HMG-CoA reductase. J Chem Ecol2004, 30(12):2459-

52. Minakuchi C, Namiki T, Yoshiyama M, Shinoda T: RNAi-mediated knockdown of
juvenile hormone acid O-methyltransferase gene causes precocious
metamorphosis in the red flour beetle Tribolium castaneum. FEBS Journal 2008,

53. Vieira CU, Bonetti AM, Simoes ZLP, Maranhao AQ, Costa CS, Costa MCR, Siquieroli
ACS, Nunes FMF: Farnesoic acid O-methyl transferase (FAMeT) isoforms:
Conserved traits and gene expression patterns related to caste differentiation in
the stingless bee, Melipona scutellaris. Arch. Insect Biochem. Phys. 2008, 67(2):97-

54. Zhou X, Tarver MR, Bennett GW, Oi FM, Scharf ME: Two hexamerin genes from the
termite Reticulitermes flavipes: Sequence, expression, and proposed functions
in caste regulation. Gene 2006, 376(1):47-58.


55. Schulz DJ, Barron AB, Robinson GE: A role for octopamine in honey bee division
of labor. Brain Behav Evol2002, 60(6).

56. Toth AL, Robinson GE: Worker nutrition and division of labour in honeybees. Anim
Behav 2005, 69(2):427-435.

57. Corona M, Velarde R, Remolina S, Moran-Lauter A, Wang Y, Hughes KA, Robinson
GE: Vitellogenin, juvenile hormone, insulin signalling, and queen honey bee
longevity. Proc Natl Acad Sci 2007, 107(17):7128-7133.

58. Miura T, Matsumoto T: Soldier morphogenesis in a nasute termite: discovery of a
disc-like structure forming a soldier nasus. Proc Roy Soc Lon B Bio 2000,

59. Koshikawa S, Matsumoto T, Miura T: Mandibular morphogenesis during soldier
differentiation in the damp-wood termite Hodotermopsis sjoestedti (Isoptera:
Termopsidae). Naturwissenschaften 2003, 90:180-184.

60. Koshikawa S, Cornette R, Hojo M, Maekawa K, Matsumoto T, Miura T: Screening of
genes expressed in developing mandibles during soldier differentiation in the
termite Hodotermopsis sjostedti. FEBS Lett 2005, 579:1365-1370.

61. Rybczynski R, Gilbert LI: Prothoracicotropic hormone elicits a rapid,
developmentally specific synthesis of beta tubulin in an insect endocrine gland.
Dev Biol 1995, 169:15-28.

62. Rybczynski R, Gilbert LI: Cloning of a beta 1 tubulin cDNA from an insect
endocrine gland: Development and hormone-induced changes in mRNA
expression. Mol Cell Endocrinol 1998, 141:141-151.

63. Kawasaki H, Sugaya K, Quan G-X, Nohata J, Mita K: Analysis of alpha- and beta-
tubulin genes of Bombyx mori using an EST database. Insect Biochem Molec Biol
2003, 33:131-137.

64. Kawasaki H, Ote M, Okano K, Shimada T, Quan G-X, Mita K: Change in the
expressed gene patterns of the wing disc during the metamorphosis of Bombyx
mori. Gene 2004, 343:133-142.

65. Erezyilmaz D, Riddiford LM, Truman JW: The pupal specifier broad directs
progressive morphogenesis in a direct-developing insect. Proc NatlAcad Sci
2006, 103(18):6925-6930.

66. Parthasarathy R, Tan A, Bai H, Palli SR: Transcription factor broad suppresses
precocious development of adults structures during larval-pupal metamorphosis
in the red flour beetle, Tribolium castaneum. Mech Deve/2008, 125(3-4):299-313.

67. LOscher M: Environmental control of juvenile hormone (JH) secretion and caste
differentiation in termites. Gen Comp Endocrinol 1972, 3:509-514.

68. Henderson GR: Primer pheromones and possible soldier caste influences on the
evolution of sociality in lower termites. In: Pheromone Communication in Social
Insects. Edited by Vander Meer RK, Breed MD, Espelie KE, Winston ML. Boulder:
Westview Press; 1998.

69. Nutting WL: Insect: Isoptera. New York: Wiley and Sons; 1990.

70. Szalanski AL, Austin JW, Owens CB: Identification of Reticulitermes spp. (Isoptera:
Reticulitermatidae) from South Central United States by PCR-RFLP. J Econ
Entomol2003, 96(5):1514-1519.


71. Tartar A, Wheeler MM, Zhou X, Coy MR, Boucias DG, Scharf ME: Parallel meta-
transcriptome analyses of host and symbiont digestive factors in the gut of the
termite Reticulitermes flavipes. Biotech. Biofuels 2009, 2(5).

72. Livak KJ, Schmittgen TD: Analysis of relative gene expression data using real-time
quantitative PCR and the 2-delta delta Ct method. Methods 2001, 25:402-408.


Figure captions
Fig. 1. Impact of semiochemical and socio-environmental treatments on
soldier caste differentiation. (A) Cumulative presoldier formation through Day 25
of assays that compared five different treatments: untreated controls, JH III, JH
III+SHE, SHE, and live soldiers for three different colonies. Each replicate dish (n=5
per colony) contained 15 workers. Results for the three colonies were pooled for
analysis. Adjusted LS means are shown; bars with the same letter are not significantly
different (P<0.05). (B) Post-hoc presoldier induction assays conducted using 100
termites per replicate dish (n=5). The two treatments that were examined included
starting compositions of 100 workers + 0 soldiers, or 90 workers + 10 soldiers. No
presoldiers formed after 30 d. in treatments that included soldiers at the beginning of
assays. This finding reveals that R. flavipes soldiers indeed are capable of inhibiting
worker differentiation to presoldiers.

Fig. 2. Expression changes for significant genes in termite workers in
response to hormonal, semiochemical and socio-environmental
treatments after 1 day. Results shown represent the relative expression values of
significant differentially expressed genes under five different treatments: control, JH
III, JH III+SHE, SHE, and live soldiers after one day; Blue boxes represent genes that
are down-regulated while red boxes represent genes that are up-regulated. Boxes with
the same letter within a row are not significantly different (FDR). Dendrograms at the
left group genes by similar expression pattern.

Fig. 3. Expression changes for significant genes in termite workers in
response to hormonal, semiochemical and socio-environmental
treatments after 5 days. See Fig. 2 caption for details.

Fig. 4. Expression changes for significant genes in termite workers in
response to hormonal, semiochemical and socio-environmental
treatments after 10 days. See Fig. 2 caption for details.

Fig. 5. Hierarchical clustering results of gene expression pattern
clustered by treatment. All three days (A- Day 1, B- Day 5, and C- Day 10) were
analyzed separately using the relative expression of each gene by the Euclidean
distance metric, with centroid squared linkage.

Fig. 6. Diagrams summarizing the influence of socio-environmental and
semiochemical factors on caste differentiation. A) Semiochemical and socio-
environmental factors tested and their effects on worker-to-soldier differentiation. JH
III and JH+SHE caused an increase in soldier formation, while SHE had no effect on
presoldier/soldier formation. Past research [15,16, personal observations] indicates that
soldiers inhibit worker differentiation. B) Diagram representing how socio-
environmental and semiochemicals factors might modulate the expression patterns of
multiple genes and caste differentiation. Networks including the following gene
categories showed significant changes among treatments: chemical
production/degradation, hemolymph protein coding, and developmental genes. Dotted
lines represent the possible feedback loop when colony worker termites molt into
soldiers, the increase in the soldier number consequently inhibits the formation of
additional soldiers.


Additional files

Additional file 1
Title: Table S1
Description: Meta-analysis of all genes used to identify reference genes
having the most stable expression. Genes highlighted in yellow are those
with the most stable expression across time and treatments (i.e., smallest
standard deviation; SD).

Additional file 2
Title: Table S2
Description: Sequence accession numbers and quantitative real-time
PCR primer details

Additional file 3
Title: Table S3
Description: Summary of ANOVAs for each gene (down) and day (across).

Additional file 4
Title: Table S4
Description: Day 1 relative expression values and summarized ANOVA
results with FDR q-values

Additional file 5
Title: Table S5
Description: Day 5 relative expression values and summarized ANOVA
results with FDR q-values

Additional file 6
Title: Table S6
Description: Day 10 relative expression values and summarized ANOVA
results with FDR q-values

Additional file 7
Title: Table S7
Description: Summary of horizontal gene clustering for Figures 2, 3 and

-31 -

Colonies 1-3

a a


- Worker %
1 Soldier %
SPresoldier %

day 0 day 30
100 Workers
0 Soldiers

Figure 1

day 0
90 Work
10 Soldi

day 30 Assay Days
ers Treatments






Day 1
Control JH-llI JH+SHE SHE Live Soldiers

b b b b Carbx-1

b b ab b Myosin

b b b ab Bactin

ab b a ab ab Btube

ab b ab ab a R-Pro

2 ab b ab ab a ATPase
ab b b ab HMG

ab b ab a Hex-2

ab b b Vab 18s

a a a a NADH

II ab a a ab nanos

b I b b CYP15F1

b b b CYP4C48
Sbc c bc CYP6G?

b b b b CYP4C47

a a a CYP4C46

Figure 2 a a ab CYP4U3

Day 5



I' te

SHE Live Soldiers
U nanos

Day 10 Control

















Live Soldiers SHE JHIII


Day 5


Live Soldiers SHE JHIII


Day 10


Figure 5




I Social environment I I



JH+SHE -- Soldiers






Gene Networks

uctionidegradation pr(
coding genes

molymph protein codi

developmentall genes



Worker Soldier





- *

Figure 6




Additional files provided with this submission:

Additional file 1: Additional File 1.doc, 259K .doc
Additional file 2: Additional File 2.doc, 277K
Additional file 3: Additional File 3.doc, 1325K
Additional file 4: Additional File 4.doc, 313K
Additional file 5: Additional File 5.doc, 313K
Additional file 6: Additional File 6.doc, 306K
Additional file 7: Additional File 7.doc, 238K

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