Comparative analysis of expressed sequence tags from three castes and two life stages of the termite Reticulitermes flavipes

Abstract

Background: Termites (Isoptera) are eusocial insects whose colonies consist of morphologically and behaviorally specialized castes of sterile workers and soldiers, and reproductive alates. Previous studies on eusocial insects have indicated that caste differentiation and behavior are underlain by differential gene expression. Although much is known about gene expression in the honey bee, Apis mellifera, termites remain relatively understudied in this regard. Therefore, our objective was to assemble an expressed sequence tag (EST) data base for the eastern subterranean termite, Reticulitermes flavipes, for future gene expression studies.

Results: Soldier, worker, and alate caste and two larval cDNA libraries were constructed, and approximately 15,000 randomly chosen clones were sequenced to compile an EST data base. Putative gene functions were assigned based on a BLASTX Swissprot search. Categorical in silico expression patterns for each library were compared using the R-statistic. A significant proportion of the ESTs of each caste and life stages had no significant similarity to those in existing data bases. All cDNA libraries, including those of non-reproductive worker and soldier castes, contained sequences with putative reproductive functions. Genes that showed a potential expression bias among castes included a putative antibacterial humoral response and translation elongation protein in soldiers and a chemosensory protein in alates.

Conclusions: We have expanded upon the available sequences for R. flavipes and utilized an in silico method to compare gene expression in different castes of an eusocial insect. The in silico analysis allowed us to identify several genes which may be differentially expressed and involved in caste differences. These include a gene overrepresented in the alate cDNA library with a predicted function of neurotransmitter secretion or cholesterol absorption and a gene predicted to be involved in protein biosynthesis and ligase activity that was overrepresented in the late larval stage cDNA library. The EST data base and analyses reported here will be a valuable resource for future studies on the genomics of R. flavipes and other termites.

Figure 1

Proportions of different GO cellular component terms (level two) in each R. flavipes cDNA library. The caste/life stage for each cDNA library is indicated above the pie diagram. The numbers next to each GO term are the percent and number of genes that are in that category.

Figure 2

Proportions of different GO biological process terms (level two) in each R. flavipes cDNA library. The caste/life stage for each cDNA library is indicated above the pie diagram. The numbers next to each GO term are the percent and number of genes that are in that category.

Figure 3

Proportions of different GO molecular function terms (level two) in each R. flavipes cDNA library. The caste/life stage for each cDNA library is indicated above the pie diagram. The numbers next to each GO term are the percent and number of genes that are in that category.

Figure 4

Frequency of genes putatively matching reproductive GO terms in each R. flavipes EST library as a fraction to total number of unigenes. No genes with putative reproductive functions were detected in the late larval cDNA library.

Figure 5

A. R-statistic comparison of contigs composition and alate singleton bias. (A) contig 794 of predicted putative ejaculatory bulb protein in soldiers; (B) contig 827 of predicted putative ejaculatory bulb protein in alates; (C) contig 545 of predicted neurotransmitter secretion or cholesterol absorption functioning in alates; (D) contig 366 of predicted protein biosynthesis or ligase activity functionality in late larval stage.