Fig. 1: Release of reduced-eye phenotype after HSP90 inhibition.

a Elucidating diagram showing Hsp83 knock-down using RNA interference (RNAi) in a male (♂) pupa of T. castaneum releasing a reduced-eye phenotype in F2 (shown in brown) followed by establishment of lines, denoted RNAi-lines henceforth. Normal-eye phenotype individuals are shown in blue. b Percentage of abnormal phenotypes showing the release of the reduced-eye trait after HSP90 impairment in the parental generation via RNAi or 17-DMAG (performed three times). RNAi was induced by injecting male pupae with 100 ng/μL dsRNA (“♂, high”) followed by crossing with naive beetles; as control we used AsnA-dsRNA. 17-DMAG was applied at 10 μg/mL (low) and 100 μg/mL (high), with treated males and females crossed (♂ & ♀); control indicates food without 17-DMAG. Total screened individuals per treatment are shown above each bar. Complete data on all used concentrations and crossing combinations are shown in Extended Data Fig. 1e. c Reduced-eye phenotype released after RNAi and 17-DMAG treatment in P (middle and bottom rows) compared to naive Cro1 beetles (top row). Three developmental stages are shown: larvae, pupae, and adults. Scale bar: 100 µm. Lateral views for pupae and adults are provided in Extended Data Fig. 1f. d Normalized eye area in normal- and reduced-eye individuals across the three developmental stages, showing significant differences in eye size within each stage. Eye areas were normalized by the stage mean. Lines connect individual beetles tracked from larval to adult stage. Phenotypes (normal- vs. reduced-eye) were assigned retrospectively based on adult morphology (n = 156 normal-eye and n = 44 reduced-eye beetles). Two-sided Mann–Whitney U tests were performed for larval (p = 5.18 × 10⁻⁹) and pupal (p = 4.87  × 10⁻²¹) stages, and a two-sided t-test for adults (t = 33.4, df = 198, p = 2.63 × 10⁻⁸³). e Average number of ommatidia in normal- and reduced-eye adult beetles (n = 8); each dot represents a single beetle. a two-sided t-test was performed (t = 21.55, df = 14, p = 3.91 × 10⁻¹²). f Number of adult offspring from parents kept under standard light/dark cycle or continuous light conditions, showing fitness consequences of reduced-eye phenotype. Numbers above bars indicate total families per treatment; each dot represents one family. Dark gray diamond = mean. Under continuous light, a generalized linear mixed model (negative binomial) with family as a random effect detected a significant difference (z = 3.35, SE = 0.08, p_adj. = 0.003). Box plots for (d–f) show the median (center line), 25th–75th percentile (box bounds), and whiskers extending to the minimum and maximum values within 1.5× the interquartile range. Significant differences indicated by asterisks, **p < 0.01, ***p < 0.001. Source data are provided via Zenodo and linked in the Data Availability section.