Article révisé par les pairs
Résumé : In termites with a true worker caste, the development pattern splits up from early developmental stages: primary reproductives develop through the nymphal line, whereas workers and soldiers follow the apterous developmental line. In some species, such as Cavitermes tuberosus Emerson (Isoptera: Termitidae, Termitinae), secondary reproductives (or neotenics) may also develop through the nymphal line from a transitional stage called aspirant. Aspirants originate mostly from automictic parthenogenetic reproduction. Therefore, C. tuberosus queens originate from sexual (primary queens) or parthenogenetic (neotenic queens) reproduction. A comparison of these two queen castes offers the possibility to better understand core molecular underpinnings of caste development and plasticity in termites. We investigated these molecular mechanisms by using high-throughput Illumina RNA sequencing of pooled individuals. We first assembled the de novo reference transcriptome of C. tuberosus, and then identified the transcripts consistently co-expressed across castes, sexes, and two alternative routes to female reproduction. Cavitermes tuberosus final transcriptome had 130 874 transcripts, N50 of 3398, and total length of 213 549 184 bp. We found that female reproductive maturation was characterized by gene expression down-regulation: primary queens expressed fewer transcripts overall and had the greatest number of down-regulated transcripts when compared to all other castes. In both secondary and primary queens, biological processes involved in muscle development and contraction, flight, and olfactory learning were enriched in the down-regulated gene cluster. In contrast, processes related to reproductive development, insulin receptor signaling pathway, isoprenoid biosynthesis, and multiple metabolic processes were enriched among up-regulated genes in primary queens. Finally, we found that 17% of all transcripts (21852) were differently co-expressed when females from sexual and parthenogenetic origins were compared, even though the expression profile of core reproductive-related gene clusters showed a similar trend in all reproductive females despite their origin. Our findings fit the genomic imprinting model predictions of a maternal effect that commonly regulates the expression of core reproductive genes in females from parthenogenetic and non-parthenogenetic origins, whereas the expression of non-reproductive genes varies.