1 Supplementary Information for 2 Genomic and transcriptomic analyses of the subterranean termite 3 Reticulitermes speratus: gene duplication facilitates social evolution 4 5 6 Shuji Shigenobu, Yoshinobu Hayashi, Dai Watanabe, Gaku Tokuda, Masaru Y Hojo, Kouhei 7 Toga, Ryota Saiki, Hajime Yaguchi, Yudai Masuoka, Ryutaro Suzuki, Shogo Suzuki, Moe Kimura, 8 Masatoshi Matsunami, Yasuhiro Sugime, Kohei Oguchi, Teruyuki Niimi, Hiroki Gotoh, Masaru K 9 Hojo, Satoshi Miyazaki, Atsushi Toyoda, Toru Miura, Kiyoto Maekawa 10 11 Corresponding authors: Shuji Shigenobu, Toru Miura, Kiyoto Maekawa 12 Email: [email protected], [email protected], [email protected] 13 14 This PDF file includes: 15 16 Supplementary text 17 Figures S1 to S36 18 Tables S1 to S28 19 Legends for Datasets S1 20 SI References 21 22 Other supplementary materials for this manuscript include the following: 23 24 Datasets S1 25 26 27 28 29 1 30 Supplementary Information Text 31 Supplementary Note: Genes involved in specific biological functions 32 Following the prediction of genes encoded in the Reticulitermes speratus genome, we manually 33 annotated and investigated the genes of some functional categories that characterize the 34 ecology, evolution, behavior, development and physiology of the subterranean termite. We 35 analyzed 15 categories, among which lipocalin, glycoside hydrolase family, lysozyme family, 36 geranylgeranyl diphosphate (GGPP) synthase and the novel secretion gene family TY are 37 described in the main text. Here, the other 10 categories (sex determination; epigenetics; 38 chemosensory genes; biogenic amines and neuropeptides; juvenile hormone-related genes; 39 ecdysone-related genes; insulin/insulin-like signaling pathway; toolkit genes involved in wing 40 formation; immunity; insecticide target and detoxification genes) are described. Finally, a report 41 on a caste-specific expression of microRNAs (miRNAs) is introduced. 42 Sex determination 43 In insects, sex determination is cell-autonomously controlled by a cascade of RNA splicing, in 44 which mRNAs are alternatively spliced in a sex-specific manner. Although the upstream genes in 45 this cascade differ among taxa, the most downstream key gene doublesex (dsx) is conserved 46 among insects, and also in some of crustaceans and chelicerates1–4. Since dsx encodes sex- 47 specific transcription factors, the cascade results in sex-specific transcriptions of downstream 48 genes that are responsible for sex differentiation. In contrast to holometabolous insects, the sex 49 determination cascades in hemimetabolous insects including termites are scarcely understood5,6. 50 Thus, our genome research in Reticulitermes speratus provides important information on gene 51 repertories related to the sex determination cascades in hemimetabolous insects. 52 The sex determination cascade might also be responsible for social organization in 53 termites, since some termite species show sex-specific or sex-biased caste ratio, suggesting that 54 some regulatory factors for caste differentiation sexually differ7. In R. speratus, sex ratio of 55 workers is known to be nearly equal, whereas those of nymph and soldier are female-biased8. For 56 example, differentiation into female secondary reproductives was inhibited by the pheromone 57 derived from female primary reproductives9, suggesting that they have the reaction mechanism to 58 sex-specific pheromones regulating the ergatoid differentiation. Therefore, the sex determination 59 cascade is one of the most likely candidate regulatory mechanisms for the sex-specific caste 60 differentiation. Thus, the annotation of sex-determination genes could also help to understand the 61 regulatory mechanisms underlying the eusociality in termites. Here, orthologs of the genes 62 reported as sex determination genes in other insect species were searched in the R. speratus 63 genome, and compared the expression levels of those genes between sexes and among castes 64 by transcriptomic analyses. 65 We selected the 27 candidate genes from the Drosophila melanogaster genes 66 categorized as “sex determination (BSID: 492283)” in BioSystems database at NCBI 67 (http://www.ncbi.nlm.nih.gov/biosystems/)10 and selected 4 candidate genes based on previous 68 studies on the sex determination in silkworm Bombyx mori11–13. Database searches were 69 performed using full length of amino acid sequences of 27 Drosophila genes or 3 Bombyx genes 70 against our Reticulitermes gene model database (RsGM8_pep) via BLASTP algorithm (See 71 Supplementary Table 12 for the accession numbers of each query sequence). Searching for an 72 ortholog of Bombyx Feminizer, which encodes microRNA, was performed via BLASTN algorithm. 73 These analyses revealed that the R. speratus genome possessed orthologs of the 24 Drosophila 74 genes and those of the 3 Bombyx genes (Supplementary Table 12). For the orthologs of major 75 components of the sex determination cascade in Drosophila, orthologs of Sex-lethal (Sxl), 76 transformer (tra), transformer-2 (tra2) and fruitless (fru) were found, whereas, surprisingly, the dsx 77 ortholog was not found. The BLAST search for Drosophila dsx as a query hit three orthologs of 78 the doublesex-mab3 related transcription factors genes (Dmrt11B, Dmrt93B, Dmrt99B), but not 79 dsx ortholog. The insect dsx is a member of the Dmrt gene family that is conserved among a wide 80 array of animal phyla14. Both dsx and Dmrt paralogs share a well conserved DNA-binding domain 81 (DM domain), and some Dmrt genes also plays essential roles in gonad development and sexual 82 differentiation outside Insecta14. The dsx ortholog would have been lost in R. speratus genome, 83 and any other factors, e.g., Dmrt genes, might be substituted as the most downstream genes in 2 84 their sex determination cascade. Alternatively, because the German cockroach Blattella 85 germanica has the conserved dsx ortholog6, the domain sequences may have diverged during 86 the course of termite evolution. 87 The R. speratus orthologs of 3 regulatory genes (deadpan, groucho, scute) for splicing of 88 the most upstream gene (Sxl) in the Drosophila cascade were duplicated, whereas no ortholog of 89 two (sisterless-A and degringolade) of them were found. An ortholog of stand still, required for the 90 Drosophila germline sex determination, was neither found. Additionally, the ortholog of B. mori 91 Feminizer was not found, while the others coding proteins were found in the R. speratus genome. 92 Although a primary signal for sex determination in R. speratus was unknown, the signal should be 93 different from those in D. melanogaster (the dose of X-linked signal element15) and B. mori 94 (Feminizer piRNA on the W chromosome11). 95 RNA-seq analysis revealed the expression patterns for 25 out of 30 candidate orthologs 96 (Supplementary Table 12). These data were compared the expression levels between sexes and 97 among three castes (primary reproductives, soldiers and workers) in two body parts [heads and 98 the remaining parts (thorax + abdomen)] (biological triplicates; NCBI BioProject Accession No. 99 PRJDB5589). Statistical analysis revealed that 14 out of 25 orthologs showed caste-biased 100 expression patterns while none showed sex-biased (Supplementary Table 12). For example, the 101 expressions of Dmrt11 orthologs (RS007930) were higher in reproductive heads, moderate in 102 worker heads, and lower in soldier heads (FDR = 7.85E-11, GLM, Supplementary Fig. 9), the 103 orthologs of outstreched (os RS015475, FDR = 1.87E-08) and ovarian tumor (out, RS009292, 104 8.00E-07) were highly expressed in female reproductive bodies (Supplementary Fig. 9). It was 105 suggested that these orthologs were involved in their sex determination or the downstream 106 pathways of sex differentiation. 107 Our analyses revealed that most orthologs of sex determination genes identified in other 108 insects, were conserved in the R. speratus genome, and that some of them were expressed in a 109 cased-biased manner. However, it remains unknown which of them play roles in their sex 110 determination cascade. In order to test their roles for sex determination, it should be examined 111 whether these genes were spliced in a sex-specific manner, and whether these genes actually 112 regulate their sex-specific trait expressions. 113 Epigenetics: Histone modifying enzymes 114 Histone posttranslational modifications (PTMs), one of the epigenetic mechanisms, play important 115 roles in gene-expression regulations without genomic changes, resulting in alteration of 116 development and behavior in various organisms16–18. Since such PTMs can be changed in 117 response to environmental stimuli16,19, polyphenic developments in insects, including caste 118 differentiation in social insects, are considered to be affected by PTMs20–23. For example, when 119 honeybee larvae were fed with royal jelly, which contain (E)-10-hydroxy-2-decenoic acid 120 possessing histone deacetylase inhibitor activity, such larvae differentiate into queen bees24. 121 Furthermore, an association between PTM (especially, histone acetylation and methylation) 122 patterns and caste identities was shown in the carpenter ant Camponotus floridanus25. It 123 suggests that histone acetylation and methylation have important roles in the regulation of caste 124 differentiation in social hymenopterans. However, in termites, such roles of PTMs remain 125 unknown. Here, as a first step of understanding the PTM roles in the caste differentiation in R. 126 speratus, we searched four