The genomic and cellular basis of biosynthetic innovation in rove beetles

Abstract

How evolution at the cellular level potentiates macroevolutionary change is central to understanding biological diversification. The >66,000 rove beetle species (Staphylinidae) form the largest metazoan family. Combining genomic and cell type transcriptomic insights spanning the largest clade, Aleocharinae, we retrace evolution of two cell types comprising a defensive gland-a putative catalyst behind staphylinid megadiversity. We identify molecular evolutionary steps leading to benzoquinone production by one cell type via a mechanism convergent with plant toxin release systems, and synthesis by the second cell type of a solvent that weaponizes the total secretion. This cooperative system has been conserved since the Early Cretaceous as Aleocharinae radiated into tens of thousands of lineages. Reprogramming each cell type yielded biochemical novelties enabling ecological specialization-most dramatically in symbionts that infiltrate social insect colonies via host-manipulating secretions. Our findings uncover cell type evolutionary processes underlying the origin and evolvability of a beetle chemical innovation.

Keywords: Dalotia; Staphylinidae; beetles; biosynthesis; cell type evolution; gene expression programs; genomics; key innovations.

Figure 1.. Aleocharine rove beetles.

(A) Cladogram of tachyporine-group Staphylinidae^, showing major radiation of Higher Aleocharinae. Numbers in parentheses are described extant species. (B) Example of a free-living aleocharine (Atheta sp.) with confocal image of tergal gland showing position on dorsal body between tergites 6 and 7. The gland comprises two cell types: solvent cells (magenta) and BQ cells (green). (C) Cartoon of tergal gland showing solvent and BQ cells secreting into common reservoir that ejects between tergites. (D) Aleocharine symbionts of ants and termites displaying behavioral interactions with hosts (chemical manipulation of host ant by Lomechusa and grooming host ant by Sceptobius) and symbiotic morphologies (myrmecoid shape of myrmecophile Diploeciton and physogastric shape of termitophile Neodioxeuta).

Figure 2.. The Dalotia reference genome

(A) Genome assembly statistics of Dcor v3. (B) SPRITE assembled contact map reveals ten chromosomes. (C) Gene density (density plot, middle band) and synteny (links, inner band) between D. coriaria and T. castaneum chromosomes or linkage groups (outer band). Inner links colored according to originating D. coriaria chromosome. 6.6% of predicted protein-coding genes map to unplaced contigs (gray links).

Figure 3.. Chemical innovation across Aleocharinae.

(A) Dated ML phylogenomic tree inferred from 1520 orthologs, with key nodes and ages indicated. All nodes received maximal bootstrap support (See also Data S3). (B) Heat map of major and minor compounds from aleocharine tergal glands. Dashed box indicates deep conservation of benzoquinones across most aleocharines.

Figure 4.. Deep conservation of tergal gland gene repertoire in the Q clade.

(A) GC traces of Dalotia and Aleochara compounds and their cell type of origin. (B) Scheme for cell type-specific transcriptomes. (C) PCA of all expressed orthologues (n=9314) in Dalotia and Aleochara solvent cells, BQ cells and control tissue (tergite 6). (D) UpSet plot showing shared DEOs for each cell type by species and cell type. (E-F) Solvent pathways in Dalotia and Aleochara, with cases of paralogue co-expression in solvent cells. Transparency of purple boxes equates to maximum log₂ fold-change above control tissue for paralogues. (G) Example GC traces from wildtype Dalotia (top trace, n=14) and bgm-silenced animals (n=42). (H) Time-calibrated tree showing origins of key enzymes. (I-J) Schematic of abdominal cell types with gene expression programs (GEPs) for ventral fat body/oenocytes and cuticle cells (I), hybridization of which created the solvent cell type (J). (K, M) Aleochara solvent cell expression (red or grey) relative to Dalotia solvent cell expression (black) for orthologues of the highest z-score ranked genes in VFBO-GEP and CC-GEP. (L, N) Violin plots showing difference in Aleochara from Dalotia solvent cell expression for genes within each GEP.

Figure 5.. Evolution of BQ chemistry.

(A) Cartoon of BQ cell showing benzoquinone synthesis from tyrosine (Tyr). (B) Benzoquinone pathway in Q clade Aleocharinae, showing cellular locations of enzymatic steps. (C) In vitro conversion of 2-methyl-1,4-hydroquinone to benzoquinone by purified Aleochara or Dalotia Dmd. Asterisks denote p<0.0001 in Tukey post-hoc tests. (D–F) Tergal gland GC traces from wildtype Dalotia (D), ATP7-silenced animals (E) and BGLU-silenced animals (F). Dotted line indicates hexane contamination peaks (removed for clarity). Asterisks denote peaks of dimethyl-BQ spiked in as positive control. (G, H) BGLU expression in Dalotia BQ cells. Green, BGLU HCR, blue: WGA; In G, magenta is dmd HCR and red is Hoechst-labeled nucleus. Lu: Lumen. (I) ML tree of laccase gene family showing Higher Aleocharine Laccase (HAL) expansion in light blue; Dalotia HAL paralogues are indicated (substitution model LG+R10 with 1000 bootstrap replicates; support for larger clades is shown by the circle color: black = 95–100% and grey 94–90%). (J) Expanded Dmd clade from panel I reveals conservation across HA taxa. Node support values <90% are not displayed. (K) Genomic HAL clusters of selected aleocharine taxa. (L) Expression heat map of Dalotia laccases, including HALs, from RNAseq data obtained from tissues, life stages and sexes.

Figure 6.. Glandular biology of the earliest-branching HA lineage.

(A-C) GC traces of hypocyphtine glandular compounds: Holobus (A), Oligota (B, B’) and Cypha (C). B’ shows headspace volatiles from 20 Oligota beetles detected via SPME. (D) Drosophila larval survival following immersion in synthetic hypocyphtine or Dalotia secretions. Outcome of Tukey post-hoc test between treatments is shown (n.s., not significant; *** denotes p<0.0001 in all individual comparisons between “all Dalotia” gland compounds and the other treatments). (E, F) HCR of MFASN (E, E’, magenta) and O-CYP4G (F, F’, green) in Oligota solvent reservoir (E, F: labelling within plane of solvent cell epithelium; E’, F’: cross section through reservoir). (G, H) HCR of MFASN (magenta) and O-CYP4G (green) in Oligota fat body and oenocytes (blue: Hoechst-stained nuclei). (I) Synteny reveals origin of TG-CYP4G in Q clade (Aleochara sp. 3, Geostiba and Dalotia) via duplication of O-CYP4G, present as a single copy in Hypocyphtini (Holobus, Oligota and Cypha), the glandless gymnusine Adinopsis, and outgroup silphine Nicrophorus. The upstream gene asense (as) is a conserved syntenic feature of all species except Dalotia. For further details of synteny see Data S5A. (J, K) TEM of Dalotia (top) and Holobus (bottom) BQ cells. Lu: lumen. Insets in J and K show differing mitochondrial densities between the two species (electron-dense structures). J’ and K’ show differing microvillar organization and density within BQ cell lumens. J” and K” show differing shield thickness within internal lumen (Lu) of ducts. L: Topology of deepest divergences in Aleocharinae. Alternative scenarios posit benzoquinones were gained in Q clade or lost in Hypocyphtini. (M, N) HCR of dmd (M, magenta) and meos (N, green) in Oligota BQ cells. (O) In vitro conversion of 2-methyl-1,4-hydroquinone to benzoquinone by purified Holobus or Dalotia Dmd. Asterisks denote p<0.0001 in Tukey post-hoc tests.

Figure 7.. Tergal gland evolution in myrmecophiles.

(A) Liometoxenus newtonarum with Liometopum occidentale host (photo: David Miller). (B) GC trace of Liometoxenus gland compounds. Magnification of geranial peak (compound 3) in grey. Asterisks: contaminants. (C) Volcano plot of Liometoxenus solvent cells (positive log₂ fold-change) and BQ cells (negative log₂ fold-change). DEOs encoding key enzymes are colored (solvent cell=purple, BQ cell=green) along with novel enzymes including inferred monoterpene pathway (blue). IDI: isopentenyl-diphosphate delta-isomerase 1; FPPS: farnesyl pyrophosphate synthase; HMG-CoA: 3-hydroxy-3-methylglutaryl-coenzyme A reductase; PMVK: phosphomevalonate kinase; FNTA: farnesyltransferase/geranylgeranyltransferase type-1 subunit alpha; GGPS1_1/GGPS1_2: geranylgeranyl pyrophosphate synthase; PDSS2: decaprenyl-diphosphate synthase subunit 2; SCD5.2: stearoyl-CoA desaturase 5; CES1: Carboxylesterase 4A; ACBP: Acyl-CoA-binding protein homolog; CES3: type-B carboxylesterase. (D) Cartoon showing hypothesized pathway for Liometoxenus aliphatic esters. SCD: Acyl-CoA Delta(11) desaturase. “CES” denotes hypothesized function of either or both Carboxylesterase 4A (CES1) or type-B carboxylesterase (CES3) in solvent cells. (E) Inferred terpene pathway leading to geranial. (F, G) Mass spectra of molecular ion regions of compounds from Liometoxenus fed with dead ants infused with ¹³C₆-Tyr. Spectra were recorded in single-ion mode. 2-methyl-1,4-BQ (MW 122) and 2-methoxy-3-methyl-1,4,-BQ (MW=152) exhibit strong [M+6]⁺ enrichment (green bars) (F), as does 2-hydroxy-6-methyl-benzoate (red bar). H: Ecitophya simulans beetle. I, J: TG-CYP4G and methoxyless gene models from Ecitochara-group species showing inactivating mutations. Negative/positive numbers are frameshift base pair deletions/insertions against the reference genome (Dalotia). Premature stop codons are shown; splice junction mutations are shown at intron-exon boundaries.