Molecular evolution of juvenile hormone esterase-like proteins in a socially exchanged fluid

Abstract

Socially exchanged fluids are a direct means by which an organism can influence conspecifics. It was recently shown that when workers of the carpenter ant Camponotus floridanus feed larval offspring via trophallaxis, they transfer Juvenile Hormone III (JH), a key developmental regulator, as well as paralogs of JH esterase (JHE), an enzyme that catalyzes the hydrolysis of JH. Here we combine proteomic, phylogenetic and selection analyses to investigate the evolution of this esterase subfamily. We show that Camponotus JHE-like proteins have undergone multiple duplications, experienced positive selection, and changed tissue localization to become abundantly and selectively present in trophallactic fluid. The Camponotus trophallactic esterases have maintained their catalytic triads and contain a number of positively-selected amino acid changes distributed throughout the protein, which possibly reflect an adaptation to the highly acidic trophallactic fluid of formicine ants. To determine whether these esterases might regulate larval development, we fed workers with a JHE-specific pharmacological inhibitor to introduce it into the trophallactic network. This inhibitor increased the likelihood of pupation of the larvae reared by these workers, similar to the influence of food supplementation with JH. Together, these findings suggest that JHE-like proteins have evolved a new role in the inter-individual regulation of larval development in the Camponotus genus.

Figure 1

The esterase repertoire in C. floridanus. Maximum likelihood tree of re-annotated C. floridanus carboxylesterases. Bootstrap (BS) values greater than 90 and equal to 100 are indicated by empty circle nodes and black nodes, respectively. The four tandem arrays of esterases observed in the genome are represented by symbols. Cflo.Est14 is on a single-gene scaffold; circle and square arrays are found on the edge of contigs, indicating that these seven esterases may be a continuous DNA segment. Protein abundance is also shown in Supplementary Table S2 (percent of total NSAF). Catalytic triad sequence motifs in each carboxylesterase gene are shown in black if all motifs are present, and are shown in bold if the specific amino acids are consistent with known JHEs. Sequences are shown in grey if one or more residues of the catalytic triad are missing; ellipses indicate missing sequence regions. Sequence length of each esterase is shown in grey circles, where filled circles indicate a complete or nearly complete sequence.

Figure 2

Camponotus trophallactic esterases have duplicated, experienced positive selection and increased in trophallactic fluid abundance. Protein trees of esterases in the clade containing the Camponotus trophallactic esterases (corresponding to the highlighted subtree in Supplementary Fig. S3), either for sequences longer than 100 amino acids (a) or 400 amino acids (b). For (b), red branches indicate positive selection (FDR < 0.1). Branch numbers and significance values can be found in Supplementary Fig. S6 and Table S5. Nodes with bootstrap values greater than 85 and equal to 100 are shown in empty circles and filled circles, respectively. The portion of the tree corresponding to formicine ants is indicated in light blue, and the sequences from the genus Camponotus in yellow. For the four species where trophallactic fluid protein abundance has been measured, sequence names are in bold and color-coded by protein abundance (specific percentages can be found in Supplementary Table S2). Branch length is based on amino acid changes. (a) Sequence length of each of the 101 esterase sequences from 31 species of ants, A. mellifera and N. vitripennis is shown in grey circles, where filled circles indicate a complete or nearly complete sequence. (b) Gold diamonds mark the three positively selected branches within which specific positively-selected sites are highlighted in Fig. 3.

Figure 3

Positively selected amino acid changes in the abundant trophallactic esterases. The protein structure on the left is the lepidopteran M. sexta JHE (2JF0) crystallized with the JHE inhibitor OTFP (blue) covalently bound in the binding pocket. The Camponotus protein, on the right, is the same structure with the amino acids replaced at sites under positive selection (posterior probability >0.95, shown in red) in the three significant branches that differentiate the derived Camponotus sequences from the more ancestral sequences in Camponotus and Formica. At the mouth of the binding pocket, M. sexta has glycine at 2FJ0 position 312, Formica and Cflo.Est13 have proline, and Cflo.Est16 has phenylalanine (see Supplementary File 3). The catalytic triad is shown in yellow, with the exception of the D/E (M. sexta numbering: amino acid 357) shown with backbone in yellow and sidechain in red to indicate positive selection.

Figure 4

The trophallactic esterases influence development in vivo. (a) Head width of pupae raised by workers who were fed food supplemented with JH, OTFP, both JH and OTFP, or solvent alone. Median values and interquartile ranges are shown. GLMM testing for the effect of the two treatments on head width with colony as a random factor, JH, p < 0.001; OTFP, not significant. (b) Proportion of total larvae that pupated and the proportion of total larvae that metamorphosed when reared by workers that were fed food supplemented with JH, OTFP, both, or solvent only. Binomial GLMM testing effect of JH and OTFP on survival past metamorphosis with colony as a random factor, p < 0.018 for each of JH and OTFP. Two-way ANOVA yielded p < 0.017 for each treatment and there was no interaction between treatments. Source data available in Supplementary File 4. (a,b) contain information from the same experiment which contained five replicates per treatment where a replicate is 20–30 workers rearing five larvae over seven weeks.