The Lolal-dpp axis mediates the regulation of host reproduction by gut symbionts in insects

Abstract

Gut commensal microbiota can play an integral role in shaping insect reproduction, but the underlying mechanisms remain largely unexplored. Here, we report that gut bacteria promote host reproduction in the oriental fruit fly Bactrocera dorsalis by inducing ubiquitin–proteasome system (UPS)-mediated degradation of the key transcription factor Longitudinals lacking-like (Lolal). Antibiotic-induced gut bacterial depletion impairs ovarian development and fertility. These reproductive defects can be reversed by nicotinic acid (NA) supplementation or recolonization with a potent NA provider, Enterobacter hormaechei. Gut bacteria-derived NA enhances coenzyme nicotinamide adenine dinucleotide (NAD) biosynthesis and mitochondrial energy production, thereby activating the UPS. Ubiquitinome analysis reveals that gut bacteria enhance Lolal ubiquitination and promote its degradation. Lolal overabundance in gut bacteria-depleted females leads to decapentaplegic (dpp) overexpression and impaired reproduction. Conversely, Lolal knockdown suppresses dpp expression, resulting in disrupted mature egg formation. Our results reveal a link between gut bacteria-derived metabolites and host protein homeostasis which determines host reproductive success.

Here, the authors show that gut bacteria promote reproduction in the oriental fruit fly by producing nicotinic acid that regulates energy metabolism and Lolal-dpp homeostasis.

Introduction

The insect gut can harbor a complex and diverse symbiotic microbial system, including bacteria, fungi, archaea, protozoa, and viruses. These resident microbes have been implicated in the regulation of a wide range of host functions, including immunity, nutrition, behavior, and development^1–3. During the long co-evolution of insects and microbiota, beneficial resident microbes that exert an important influence on host health were selected. Insect genomes often lack genes for the biosynthesis of certain metabolites, and this gap can be filled by gut symbiotic microbiota⁴. Transcriptomic studies have revealed that gut bacteria are enriched in biosynthetic pathways for essential amino acids⁵. Gut bacteria carrying genes for B vitamin synthesis are particularly important for the development of insects that feed on nutrient-poor diets^6,7. Some gut bacteria encode enzymes that can convert tryptophan into indole derivatives, which promote the learning and memory behaviors of the honeybee⁸. Although some symbionts have undergone extensive reduction of genome size and gene content over long-standing host–microbe associations⁹, they still retain key genes that function in the host and help reduce host metabolic costs to allow the allocation of energy for other physiological processes^10,11. These host-complementary genetic features expand the functional diversity of microbiota and establish the inseparable host-microbiota relationship.

The reproductive process in female insects is crucial for species propagation. Juvenile hormone, ecdysteroids, and insulin signaling contribute to oocyte maturation and nutrient sensing during ovarian development¹². Growing evidence shows that gut bacteria can maintain these signaling pathways^13,14. In Drosophila melanogaster, gut microbiome changes can reshape host gene expression^15,16; specifically, the downregulation of aldehyde dehydrogenase caused by the absence of gut Acetobacter species is sufficient to suppress oogenesis¹⁷. Germ-free D. melanogaster also exhibits ovarian and even systemic metabolism disorders¹⁸. Under amino acid-deficient diets, the contribution of gut bacteria to fecundity becomes particularly pronounced in Bactrocera oleae¹⁹. Importantly, reproductive defects resulting from gut bacteria depletion can be rescued through specific gut bacteria or dietary supplementation^17,18,20.

Although the metabolic and nutritional roles of gut bacteria are closely linked to the reproductive health of insect hosts, the underlying molecular mechanisms remain largely unknown. Among the potential regulatory pathways, ubiquitination may also play a role. Ubiquitination is a post-translational modification (PTM) implemented by a three-step enzymatic cascade involving ubiquitin-activating enzymes (E1), ubiquitin-conjugating enzymes (E2), and ubiquitin ligases (E3), which covalently link ubiquitin to proteins²¹. Ubiquitination has been shown to function in insect reproduction, including gametogenesis²². Notably, some microbes have developed a range of strategies to manipulate the host ubiquitination system. Pathogenic bacteria can directly alter host E1 enzyme expression to manipulate global ubiquitination²³ and secrete effectors that diminish host ubiquitin binding activity, thereby counteracting the host immune response²⁴. Some viruses manipulate host E3 ligases to alter the ubiquitin-mediated protein degradation profile that facilitates their infection²⁵. However, whether and how gut bacteria influence ubiquitination in insect hosts remains unknown.

The oriental fruit fly Bactrocera dorsalis (Hendel) is one of the world’s most invasive pests, with females ovipositing in a wide variety of fruits²⁶. B. dorsalis has a stable and complex gut bacterial community, with Enterobacteriaceae as the predominant family. Enterobacter and Klebsiella are the most abundant genera in both laboratory-reared and wild populations²⁷. Several culturable species, such as Enterobacter cloacae, Klebsiella michiganensis, Klebsiella oxytoca, and Citrobacter freundii, have been isolated from the gut and shown to facilitate larval development and contribute to the adaptation of B. dorsalis to environmental stress^6,28,29. In this study, we show that gut bacteria-derived nicotinic acid (NA) contributes to B. dorsalis reproduction by promoting ubiquitination of the Lolal transcription factor. B. dorsalis relies on gut bacteria to supply NA, especially from Enterobacter hormaechei. Gut bacteria-derived NA promotes NAD synthesis and enhances ATP production, which in turn enhances host UPS activity. Moreover, gut bacteria-activated UPS promotes ubiquitinated Lolal degradation, and in this way maintains the requisite Lolal level for female reproduction.

Results

Gut bacteria-derived NA is essential for B. dorsalis reproduction

First, we characterized the gut microbiota composition of B. dorsalis females by 16S ribosomal DNA (rDNA) sequencing. In conventionally reared females (Ctrl), gut bacteria were largely composed of Enterobacteriaceae, with Enterobacter and Klebsiella as the predominant genera (Supplementary Fig. 1a, b). To deplete gut bacteria, females were administered an oral antibiotic cocktail (ABX). After three days of treatment, gut bacteria were decreased by 95% compared with Ctrl (unpaired t-test, t(4) = 79.0, p < 0.0001, mean difference = 0.94, 95% CI [0.90, 0.97]; Fig. 1a), accompanied by a significant reduction in cultivable bacterial load (Supplementary Fig. 2a). We reintroduced the total cultivable bacteria into ABX females (Supplementary Fig. 2b) and found that gut bacteria depletion or recolonization did not affect female survival (Supplementary Fig. 2c). However, depletion of gut bacteria inhibited ovarian development, which was partially restored after cultivable bacteria reintroduction, with 70% of the flies reaching ovarian development grades IV–V (Chi-square test, χ²(8) = 126.8, p < 0.0001, Cramér’s V = 0.72, 95% CI [0.58, 0.83]; Supplementary Fig. 2d, e).

Fig. 1. Gut bacteria supply NA that promotes female reproduction of B. dorsalis.

a qPCR verification of gut bacterial depletion in antibiotic-treated (ABX) and conventional-reared (Ctrl) females (n = 3). KEGG pathways enrichment (two-sided Fisher’s exact test) (b), volcano plot (c) of differential metabolites (two-tailed unpaired t-test). Representative ovary images (d) and development grades (e) of Ctrl (n = 41), ABX (n = 42), and ABX supplemented with nicotinic acid (ABX + NA) (n = 32). Total egg production (f) and egg hatching rates (g) of females in d (n = 8). h Relative NA levels in hemolymph and ovaries of Ctrl, ABX, and ABX females recolonized with gut microbiota (n = 4 (hemolymph), 5 (ovary)). i NA content secreted by five gut bacteria (n = 4). j Relative NA levels in hemolymph of Ctrl, ABX, and ABX mono-colonized with gut bacteria (n = 4). k Relative NA levels in hemolymph and ovaries of Ctrl, ABX, and ABX recolonized with E. hormaechei (ABX + EH) or ∆PncA E. hormaechei (ABX + ∆PncA EH) (n = 4). Representative ovary images (l) and development grades (m), shown for Ctrl (n = 50), ABX (n = 40), ABX + EH (n = 58), and ABX + ∆PncA EH (n = 47). Total egg production (n) and egg hatching rates (o) of females treated as l (n = 8). a, f, g, i, n, o Data are shown as mean ± SD. h, j, k Box plots show the median, 25th and 75th percentiles, minimum and maximum values. a Two-tailed unpaired t-test, p = 1.54e−07. One-way ANOVA followed by Tukey’s test, p = 3.34e−08 (f), 8.11e-08 (g), 4.08e−18 (i), 2.62e−08 (j), 4.76e−11 (n), 3.13e−07 (o). h p = 0.0022 (hemolymph), 0.0002 (ovary). k One-way ANOVA followed by Tukey’s test, p = 1.29e−07 (hemolymph); Kruskal–Wallis test followed by Dunn’s test, p = 0.0007 (ovary). Two-sided Pearson’s χ² test with Bonferroni-corrected post-hoc comparisons, p = 1.13e−13 (e), 1.99e−27 (m). Different letters indicate significant differences. ****p < 0.0001. Scale bar: 1 mm. Source data are provided as a Source Data file.

The insect gut microbiota is involved in shaping host metabolism¹, and hemolymph serves as a key medium for the systemic transport of metabolites³⁰. To investigate how gut bacteria influence reproduction in B. dorsalis, we performed untargeted metabolomics of hemolymph from ABX and Ctrl females (Supplementary Data 1). KEGG pathway analysis revealed that the elimination of gut bacteria caused widespread disruptions in host metabolic pathways, particularly those involved in vitamin and cofactor biosynthesis, amino acid metabolism, and energy metabolism (Fig. 1b). Within these pathways, 9 metabolites had decreased and 26 increased in ABX females (Supplementary Fig. 3a). Given the role of gut bacteria in host nutrition support, we focused on the downregulated metabolites and identified six candidates that are potentially derived from gut bacteria, including nicotinic acid (NA), pantothenic acid (PA), mannitol, xanthurenic acid (XA), melibiose, and neopterin (Fig. 1c). To assess their potential roles in ovarian development, each metabolites was individually supplemented in the diet of ABX females. Among them, dietary NA supplementation partly rescued ovarian development (Supplementary Fig. 3b, c), with 41% of ovaries reaching development grades Ⅳ–Ⅴ (Chi-square test, χ²(8) = 78.2, p < 0.0001, Cramér’s V = 0.58, 95% CI [0.42, 0.69]; Fig. 1d, e) and increased vitellogenin deposition (Supplementary Fig. 3d). Consistently, ABX females exhibited a 75% decrease in fecundity compared with Ctrl females, whereas NA supplementation significantly increased egg production to 3.0-fold of ABX females (one-way ANOVA, F(2, 21) = 43.6, p < 0.0001, η² = 0.81, 95% CI [0.58, 0.87]; Fig. 1f). Depletion of gut bacteria or dietary NA had no effect on the timing of first reproduction, mature egg size or larval viability (Supplementary Fig. 3e–g). In addition, the egg hatching rate was significantly reduced in ABX females, and dietary NA supplementation increased it to 1.7-fold of ABX females, although it did not reach the level observed in the Ctrl (one-way ANOVA, F(2, 21) = 39.2, p < 0.0001, η² = 0.79, 95% CI [0.55, 0.86]; Fig. 1g). These results indicate that NA is required for female reproduction in B. dorsalis.

NA is a form of vitamin B3 that cannot be autonomously synthesized by insects³¹. LC–MS/MS analysis revealed that gut bacteria removal led to a reduction in host NA levels, with decreases of 69% in the hemolymph and 68% in the ovaries (Fig. 1h). Interestingly, reintroduction of total culturable gut bacteria into ABX females restored NA levels (one-way ANOVA, F(2, 9) = 12.9, p = 0.0022, η² = 0.74, 95% CI [0.21, 0.84] for hemolymph; F(2, 12) = 18.6, p = 0.0002, η² = 0.76, 95% CI [0.34, 0.85] for ovary; Fig. 1h). Our previous study identified Klebsiella aerogenes, Enterobacter hormaechei, Providencia vermicola, Providencia alcalifaciens, and Klebsiella quasipneumoniae as the dominant culturable bacterial species in the female gut²⁰. To determine potential NA-producing bacterial providers, we assessed the NA production capacity of these isolates using LC–MS/MS. Among these species, E. hormaechei produces substantially higher levels of NA (one-way ANOVA, F(4, 15) = 1034.8, p < 0.0001, η² = 1.00, 95% CI [0.99, 1.00]; Fig. 1i). Mono-colonization experiments revealed that P. alcalifaciens, P. vermicola, K. aerogenes, and K. quasipneumoniae had no effect on NA levels in the hemolymph of ABX females (one-way ANOVA, F(6, 21) = 23.6, p < 0.0001, η² = 0.87, 95% CI [0.67, 0.90]; Fig. 1j). Notably, E. hormaechei significantly increased NA levels in both hemolymph and ovaries compared with ABX females (one-way ANOVA, F(3, 12) = 63.0, p < 0.0001, η² = 0.94, 95% CI [0.79, 0.96] for hemolymph; Kruskal–Wallis test, H(3) = 11.6, p = 0.0007, ε² = 0.58, 95% CI [0.49, 0.89] for ovary; Fig. 1k). Consistently, E. hormaechei recolonization promoted ovarian development in ABX females, with 55% ovaries reaching grades Ⅳ–Ⅴ (Chi-square test, χ²(12) = 157.2, p < 0.0001, Cramér’s V = 0.53, 95% CI [0.42, 0.58]; Fig. 1l, m) and increased egg production by 1.9-fold (one-way ANOVA, F(3, 28) = 46.9, p < 0.0001, η² = 0.83, 95% CI [0.67, 0.88]; Fig. 1n). These results imply that E. hormaechei-derived NA contributes to female fertility. Genomic analysis revealed that E. hormaechei possesses the complete set of genes for NA synthesis (Supplementary Fig. 4a). To further investigate the role of NA, we knocked out the nicotinamidase/pyrazinamidase gene (∆PncA), a key gene in the NA synthesis pathway (Supplementary Fig. 4b). NA secretion became undetectable in the mutant bacterial culture (Supplementary Fig. 4c), whereas PncA depletion did not affect the growth of E. hormaechei in either nutrient-rich or minimal media, nor its gut colonization (Supplementary Fig. 4d, e). Recolonization of ∆PncA E. hormaechei did not significantly increase NA levels in hemolymph or ovaries of ABX females (Fig. 1k), nor did it improve ovarian development as observed with wild-type E. hormaechei (Fig. 1l, m). Consequently, fecundity of ABX females recolonized with the ∆PncA strain remained markedly low (Fig. 1n), and the restored egg hatching rate observed in the wild strain was not reproduced in ∆PncA strain-recolonized females (one-way ANOVA, F(3, 28) = 20.5, p < 0.0001, η² = 0.69, 95% CI [0.42, 0.78]; Fig. 1o). Collectively, these results indicate that gut bacteria promote reproductive success by supplying NA to B. dorsalis females.

NA supports host NAD biosynthesis and ATP production

NA is metabolized into the bioactive NAD pool (NAD⁺ and NADH)³², and we found that the ovarian NAD⁺ and NADH content of ABX females were significantly reduced compared with Ctrl females (one-way ANOVA, F(3, 12) = 79.6, p < 0.0001, η² = 0.95, 95% CI [0.83, 0.97]; Fig. 2a). Recolonization with E. hormaechei, but not with ∆PncA E. hormaechei, effectively improved NAD⁺ and NADH content (Fig. 2a). Moreover, dietary NA supplementation increased ovarian NAD⁺ and NADH content in ABX females (one-way ANOVA, F(2, 9) = 79.2, p < 0.0001, η² = 0.95, 95% CI [0.78, 0.97]; Fig. 2b). To investigate the impact of E. hormaechei-derived NA on NAD biosynthesis, we analyzed the expression of key genes involved in the NAD biosynthesis pathway and found that ABX females exhibited decreased expression of the Naprt (nicotinate phosphoribosyltransferase), Nmnat (nicotinamide mononucleotide adenylyltransferase), and Nadsyn (NAD synthetase) genes (Fig. 2c). Notably, NA supplementation upregulated the expression of these genes, particularly Nmnat and Nadsyn, irrespective of whether the NA originated from gut bacteria or dietary sources (Fig. 2c). These results suggest that NA activates the NAD metabolic pathway and enhances its biosynthesis.

Fig. 2. NA boosts host NAD biosynthesis and mitochondrial function.

a Ovarian NAD(H) content of Ctrl, ABX, ABX + EH and ABX + ∆PncA EH females. Values were normalized to tissue protein (n = 4). b Ovarian NAD(H) content of Ctrl, ABX and ABX + NA females (n = 4). c Heatmap representing Naprt, Nmnat and Nadsyn transcriptional levels, genes involved in NAD biosynthesis. Each column represents an independent sample (n  =  3). d Ovarian ATP levels of females treated as in (a). Values were normalized with tissue protein (n = 5). e Ovarian ATP levels of Ctrl, ABX, and ABX + NA females (n = 5). f Transmission electron micrographs showing mitochondrial morphology of Ctrl, ABX, and ABX + NA female ovaries. N nucleus, M mitochondria. The white arrow denotes disruptive mitochondrion membrane, red arrow denotes swollen mitochondrion, and blue arrow denotes broken cristae. Quantitative analysis of the swollen mitochondria (g) and mitochondria with impaired cristae (h) in f (n = 3). i COX activity of Ctrl, ABX, and ABX + NA fly ovaries visualized by COX single staining. j Relative COX activity in i (n = 4). Data are shown as mean ± SD. One-way ANOVA followed by Tukey’s test, p = 3.46e−08 (a), p = 1.93e−06 (b), 3.37e−08 (d), 4.66e−09 (e), 0.0009 (g), 0.0058 (h), 0.0045 (j). Different letters indicate significant differences. Source data are provided as a Source Data file.

NAD serves as a crucial mitochondrial coenzyme that is involved in cellular energy metabolism³³. To evaluate the role of gut bacteria-derived NA on mitochondrial function, we assessed mitochondrial energy production in ABX female ovaries, which exhibited clear signs of energy depletion (one-way ANOVA, F(3, 16) = 47.6, p < 0.0001, η² = 0.90, 95% CI [0.72, 0.93]; Fig. 2d). E. hormaechei partially restored ATP content, while the ∆PncA E. hormaechei failed to do so (Fig. 2d). Similarly, dietary NA enhanced ATP production in ABX ovaries, although the reversal remained incomplete (one-way ANOVA, F(2, 12) = 140.8, p < 0.0001, η² = 0.96, 95% CI [0.87, 0.97]; Fig. 2e). Transmission electron microscopy (TEM) of ABX ovaries further revealed an increased number of damaged mitochondria characterized by distorted cristae, disrupted membranes, and swollen morphology (Fig. 2f). These defects were reversed in ovaries of NA-supplemented ABX females (Fig. 2g, h), suggesting an improvement in mitochondrial oxidative metabolism. Thus, we investigated the impact of NA on mitochondrial oxidative phosphorylation (OXPHOS), the main pathway for cellular energy production³⁴. NA supplementation increased the expression of OXPHOS subunits including ND1, ND3, SDHA, SDHC, and COX1 (Supplementary Fig. 5a–d), as well as ATPase subunits ATP5A, ATP6 and ATP8 compared with ABX females (Supplementary Fig. 5e). Further assessment of OXPHOS complex IV (COX) activity by histochemical staining of frozen ovary sections revealed a marked reduction of COX staining intensity in ABX ovaries and that was improved by NA supplementation (one-way ANOVA, F(2, 9) = 10.47, p = 0.0045, η² = 0.70, 95% CI [0.14, 0.82]; Fig. 2i, j). Overall, these results suggest that gut bacteria-derived NA contributes to host mitochondrial energy production.

NA reduction suppresses female reproduction by inhibiting UPS activity

To further investigate the underlying mechanism by which NA-mediated ATP production affects ovarian development, we focused on the differential expression of ovarian proteins between Ctrl and ABX females (Supplementary Data 2). Protein turnover-related functions, including protein folding, degradation, ubiquitin binding, and proteasome activity, were found to be downregulated in ABX females (Supplementary Fig. 6a). KEGG pathway analysis further revealed significant downregulation of the proteasome pathway (Supplementary Fig. 6b). Mechanistically, the proteasome recognizes ubiquitin chains and degrades polyubiquitinated proteins, which play a central role in protein homeostasis²¹. These findings suggested that the ubiquitin–proteasome system (UPS) may be disrupted in ABX females. Indeed, global ubiquitination reduction was observed in ABX ovaries (Fig. 3a). ATP is involved in the conjugation of ubiquitin to substrate proteins as part of proteasomal degradation³⁵. We found that feeding females a diet containing rotenone (Rot), an inhibitor of mitochondrial complex I (NADH dehydrogenase), significantly reduced ovarian ATP content (Mann–Whitney test, U = 0, p = 0.0022, r = 0.83, 95% CI [0.56, 0.85]; Supplementary Fig. 7) as well as protein ubiquitination (Fig. 3b). We hypothesize that ATP-dependent ubiquitin modification is regulated by gut bacteria via NA-mediated ATP production. Supporting this hypothesis, protein ubiquitination level in ABX ovary was restored by wild-type E. hormaechei recolonization, but not by ∆PncA E. hormaechei (Fig. 3c). This is in line with the observation that E. hormaechei rescue is thwarted by Rot administration (one-way ANOVA, F(4, 10) = 79.0, p < 0.0001, η² =  0.97, 95% CI [0.86, 0.98]; Fig. 3c). Notably, dietary NA supplementation increased ubiquitination in ABX females (Fig. 3d). These results indicate that gut bacteria systematically maintain ovarian ubiquitin modification by promoting NA-induced energy production.

Fig. 3. Reduced NA suppresses female reproduction by inhibiting ubiquitination.

a Western blot analysis for ubiquitinated-protein of Ctrl and ABX fly ovaries. β-Tubulin was used as the loading control. b Western blot analysis for ovarian ubiquitinated-protein following rotenone (Rot) exposure. c Changes in ubiquitinated-protein abundance in ovaries of Ctrl, ABX, ABX + EH, ABX + ∆PncA EH, and ABX flies supplemented with EH and Rot (ABX + EH + Rot). The right panel displays band intensity (n = 3). d Western blotting analysis for ubiquitinated-protein of Ctrl, ABX, and ABX + NA female ovaries. e Western blot analysis for ubiquitinated-protein in ovaries after dsGFP and dsUba1 injection. Representative ovary images (f), the development grades (g), shown for females injected with dsGFP (n = 62) and dsUba1 (n = 66). Total egg production (h) and egg hatching rates (i) of females treated as f (n = 8). Representative ovary images (j) and development grades (k), shown for Ctrl (n = 46), ABX (n = 46), and ABX + NA females injected with dsGFP (n = 42) or dsUba1 (n = 46). Total egg production (l) and egg hatching rates (m) of females treated as j (n = 8). Data are shown as mean ± SD. a, b, d, e Three independent biological samples for each group. One-way ANOVA followed by Tukey’s test, p = 1.59e−07 (c), 5.38e−14 (l), 4.76e−06 (m). h Two-tailed unpaired t-test, p = 9.55e−10. i Two-tailed Man–Whitney test, p = 0.0047. Two-sided Pearson’s χ² test with Bonferroni-corrected post-hoc comparisons, p = 4.79e−11 (g), 8.91e−25 (k). Different letters indicate significant differences. **p < 0.01; ****p < 0.0001. Scale bar: 1 mm. Source data are provided as a Source Data file.

To investigate whether reduced ubiquitination is associated with the negative impact on reproduction in ABX females, we silenced Uba1, the major E1 enzyme responsible for initiating ubiquitination, by dsRNA-mediated RNA interference (RNAi) (Supplementary Fig. 8a). Uba1 knockdown led to an overall reduction in ovarian protein ubiquitination (Fig. 3e) and phenocopied the suppressed ovarian development observed in bacterial-depleted females, as 80% of the ovaries remained at the early Ⅰ–Ⅲ grades (Chi-square test, χ²(4) = 54.2, p < 0.0001, Cramér’s V = 0.65, 95% CI [0.46, 0.81]; Fig. 3f, g), consistent with the reduced vitellogenin deposition (Supplementary Fig. 8b, c). Moreover, Uba1 knockdown caused a 68% reduction in egg production (unpaired t-test, t(14) = 14.3, p < 0.0001, Cohen’s d = 7.15, 95% CI [4.34, 9.93]; Fig. 3h) and a 31% decrease in egg hatching rate (Mann–Whitney test, U = 6, p = 0.0047, r = 0.67, 95% CI [0.26, 0.84]; Fig. 3i). To determine whether restoration of ubiquitination activity contributes to improved ovarian development and fecundity in NA-supplemented ABX females, we supplemented NA for ABX females injected with dsUba1. When ubiquitin modification was blocked, dietary NA only promoted development in 9% of the ovaries to grades Ⅳ–Ⅴ, as compared with 48% in dsGFP-injected ABX females (Chi-square test, χ²(12) = 144.1, p < 0.0001, Cramér’s V = 0.52, 95% CI [0.41, 0.58]; Fig. 3j, k). Similarly, NA supplementation failed to improve the fecundity (one-way ANOVA, F(3, 28) = 82.2, p < 0.0001, η² = 0.90, 95% CI [0.79, 0.93]; Fig. 3l) or egg hatching rate (one-way ANOVA, F(3, 28) = 15.2, p < 0.0001, η² = 0.62, 95% CI [0.32, 0.73]; Fig. 3m) in dsUba1-injected ABX females, as ubiquitination lies downstream of the NA pathway. Taken together, these results suggest that gut bacteria modulate female fertility via ubiquitination upregulation by NA.

Gut bacteria target Lolal ubiquitination and proteasomal degradation

We further explored changes in host ubiquitination profiles using quantitative mass spectrometry-based ubiquitinome analysis. Antibodies that recognize diglycine-modified lysine residues were used to enrich ubiquitin-modified peptides (K-GG), followed by mass spectrometry (Fig. 4a). A total of 3331 ubiquitinated proteins with 12,514 K-GG sites were identified in ovaries from the Ctrl, ABX, and E. hormaechei recolonized (EH) females (Supplementary Data 3). Among them, 66% of the proteins contained 1–3 K-GG sites and 8% exhibited 10 or more K-GG sites (Supplementary Fig. 9a). Integrated analysis of the proteome and ubiquitinome revealed that 819 ubiquitinated proteins showed a negative correlation between changes in protein abundance and ubiquitination level in ABX vs. Ctrl (Fig. 4b), and 665 ubiquitinated proteins showed similar opposing changes in EH vs. ABX (Fig. 4c). KEGG pathway analysis of these datasets highlighted pathways related to protein processing and ubiquitin-mediated proteolysis (Fig. 4d). We identified 27 proteins that exhibited decreased ubiquitination and increased protein abundance in the ABX females compared with those in the Ctrl females; ubiquitination of these proteins was rescued in the EH females (Fig. 4e and Supplementary Table 1). GO analysis revealed that 19 of these proteins were enriched in developmental processes (Supplementary Fig. 9b). Among this set, the transcription factor Lolal (also known as batman) showed the highest expression level in the ovary (Supplementary Fig. 9c, d). Lolal is a small protein of 127 amino acids containing a BTB/POZ domain, which interacts with genes involved in retinal morphogenesis, epithelial tube elongation, and ovariole number limitation in Drosophila^36–38. These findings imply that Lolal may mediate gut bacteria-induced regulation of ovarian development.

Fig. 4. Lolal ubiquitination and degradation are regulated by gut bacteria.

a Schematic of the enrichment procedure for ubiquitinated proteins using a diglycine antibody that recognizes the GG remnants of ubiquitinated regions, followed by mass spectrometry analysis. Created in BioRender. Qiao, J. (2026) https://BioRender.com/g1axqba. b, c Scatter plot of 9-quadrant associate analyses of ubiquitination and protein level from log₂ FC (ABX/Ctrl) and (EH/ABX). Each point represents a protein or K-GG site. The plot is divided into nine quadrants (numbered 1–9), with the number of points in each quadrant indicated in parentheses. d Bubble charts displaying KEGG pathway enrichment based on ubiquitinated proteins from quadrant NO. 9 in b and quadrant NO. 1 in c, analyzed by two-sided Fisher’s exact test. e Venn diagram of proteins from quadrant NO. 9 in b and quadrant No. 1 in (c). f Analysis of Lolal ubiquitination in the ovaries by immunoprecipitation. g Half-life analysis of Lolal in B. dorsalis embryonic cells. Cells were incubated with the protein synthesis inhibitor cycloheximide (CHX) (80 µg/mL). h Lolal degradation pathway analysis in B. dorsalis embryonic cells with treatment of CHX + DMSO and CHX + MG132 (20 µM). The panel below shows Lolal band intensity (n = 3). i Detection of Lolal protein in the ovaries of Ctrl, ABX, ABX + EH, ABX + ∆PncA EH, and ABX + NA females. The panel below shows Lolal band intensity (n = 3). h, i Data are shown as mean ± SD. f, g Similar results were obtained from two biological repeats. h two-way ANOVA, p _interaction = 5.83e−14. i One-way ANOVA followed by Tukey’s test, p = 0.0025. Different letters indicate significant differences. ****p < 0.0001. Source data are provided as a Source Data file.

Therefore, we next examined how gut microbes regulate Lolal. The Co-immunoprecipitation (Co-IP) assay confirmed that Lolal was highly ubiquitinated in Ctrl and EH females, while its ubiquitination was reduced in ABX and ∆PncA E. hormaechei-recolonized females (Fig. 4f). In addition, NA supplementation in ABX females increased Lolal ubiquitination (Fig. 4f). To investigate Lolal protein stability, we treated a B. dorsalis embryonic cell line with the protein synthesis inhibitor cycloheximide (CHX) and collected cells at different time points. The results revealed that Lolal is an unstable protein with a relatively short half-life, as its abundance progressively declined over time (Fig. 4g). Treatment with the proteasome inhibitor MG132 (carbobenzoxy-Leu-Leu-leucine) delayed degradation of Lolal protein (two-way ANOVA, F(5, 24) = 76.7, p _interaction < 0.0001, η² = 0.94, 95% CI [0.86, 0.96]; Fig. 4h), while treatment with the lysosome inhibitor chloroquine (CQ) did not prevent degradation (two-way ANOVA, F(5, 24) = 1.7, p_interaction = 0.174, η² = 0.26, 95% CI [0.00, 0.45]; Supplementary Fig. 10a). These results indicated that Lolal protein degradation occurs primarily via the proteasome. Consistently, we found that hypo-ubiquitinated Lolal in ABX and ∆PncA E. hormaechei-recolonized ABX females resulted in the accumulation of ovarian Lolal protein compared with Ctrl females (Fig. 4i). In contrast, higher ubiquitination levels facilitated Lolal degradation, leading to reduced Lolal abundance in E. hormaechei-recolonized and NA-supplemented ABX females compared with ABX females (one-way ANOVA, F(4, 10) = 8.9, p = 0.0025, η² = 0.78, 95% CI [0.24, 0.84]; Fig. 4i). Notably, the detection of Lolal mRNA levels excluded the possibility that changes in Lolal protein level are due to transcriptional regulation (Supplementary Fig. 10b). Overall, these results indicate that gut bacteria facilitate the ubiquitination of Lolal and regulate its protein abundance via the ubiquitin–proteasome pathway.

Lolal-dpp pathway homeostasis is required for female reproduction

In Drosophila, Lolal is generated during oogenesis and is necessary for proper maintenance of decapentaplegic (dpp) transcription³⁹. To determine whether gut bacteria regulate B. dorsalis reproductive development via the Lolal-dpp pathway, we first investigated how Lolal regulates dpp expression. Lolal RNAi led to reduced dpp expression in ovaries (Fig. 5a). We then performed chromatin immunoprecipitation (ChIP) assays on B. dorsalis ovaries with anti-Lolal antibody. Both PCR product and quantitative PCR analyses revealed significant enrichment of dpp promoter sequence in samples immunoprecipitated with the anti-Lolal antibody compared with the IgG control (Mann–Whitney test, U = 0, p = 0.0286, r = −0.82, 95% CI [−0.56, −0.29]; Fig. 5b, c). In B. dorsalis embryonic cells, siRNA-mediated Lolal knockdown decreased dpp mRNA expression (Supplementary Fig. 11a), while Lolal overexpression increased dpp expression (Supplementary Fig. 11b, c). These results suggest that Lolal binds to the dpp promoter to modulate its expression.

Fig. 5. Accumulation of Lolal inhibits ovarian development by upregulating dpp in ABX females.

a Relative expression level of Lolal and dpp in ovaries after Lolal knockdown (n = 4). b, c PCR amplification (b) and ChIP results of qPCR (c) (n = 4). Ovary chromatin was immunoprecipitated with anti-Lolal antibody, IgG; Ec, empty control. d SDS–PAGE and western blot analysis for recombinant GST-His and Lolal-His. Similar results were obtained from two biological repeats. Representative ovary images (e) and development grades (f) of females injected with Lolal-His (n = 61) or GST-His (n = 76). g Relative expression of dpp in ovaries of females treated as e (n = 3). Total egg production (h) and egg hatching rates (i) of females treated as e (n = 8). Representative ovary images (j) and development grades (k) of Ctrl, ABX flies injected with dsGFP, dsLolal, or dsdpp (n = 60). Total egg production (l) and egg hatching rates (m) of females treated as j (n = 8). n The egg length of females treated as j (n = 15). o Relative expression of dpp mRNA in ovaries of Ctrl, ABX, and ABX + NA females (n = 3). Data are shown as mean ± SD. Two-tailed unpaired t-test, (a) p = 0.0007 (Lolal), 0.0021 (dpp), p = 0.0231 (g), 0.0003 (h), 0.0118 (i). c Two-tailed Mann–Whitney test, p = 0.0286. One-way ANOVA followed by Tukey’s test, p = 7.46e−24 (l), 3.38e−07 (m), 0.1624 (n), 0.0094 (o). Two-sided Pearson’s χ² test with Bonferroni-corrected post-hoc comparisons, p = 2.35e−08 (f), 8.81e−33 (k). Different letters indicate significant differences. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns non-significant. Scale bar: 1 mm. Source data are provided as a Source Data file.

Since gut bacteria affect Lolal protein homeostasis, we hypothesized that elevated Lolal contributes to impaired female reproduction in ABX females. To test this, we expressed and purified recombinant Lolal-His in a prokaryotic system. SDS–PAGE and western blot analyses confirmed expression of Lolal-His with a predicted molecular weight of 20 kDa (GST-His as a control) (Fig. 5d). Lolal-His injection inhibited ovarian development (Chi-square test, χ²(4) = 41.3, p < 0.0001, Cramér’s V = 0.55, 95% CI [0.36, 0.70]; Fig. 5e, f) and increased dpp expression compared with the control group (unpaired t-test, t(4) = 3.6, p = 0.0231, Cohen’s d = −2.93, 95% CI [−5.39, −0.35]; Fig. 5g). In addition, egg production and hatching rate were reduced by 33% and 21%, respectively (unpaired t-test, t(14) = 4.7, p = 0.0003, Cohen’s d = 2.37, 95% CI [1.04, 3.66]; t(14) = 2.9, p = 0.0118, Cohen’s d = 1.45, 95% CI [0.31, 2.54]; Fig. 5h, i). Moreover, we performed rescue experiments in ABX females by knocking down Lolal and dpp, respectively. Lolal knockdown markedly reduced Lolal protein levels and dpp transcription in ABX females (Supplementary Fig. 12a–c). This resulted in a partial rescue of ovarian development, with only 30% of ovaries remaining at grade Ⅰ compared to 60% of ovaries from dsGFP-injected ABX females (Chi-square test, χ²(12) = 183.3, p < 0.0001, Cramér’s V = 0.55, 95% CI [0.42, 0.56]; Fig. 5j, k). Similarly, the reduction of dpp expression in ABX females led to a higher proportion of ovaries reaching grades Ⅳand Ⅴ (28%) than dsGFP-injected ABX females (7%) (Fig. 5j, k). Moreover, egg production and hatching rates increased in dsLolal-injected and dsdpp-injected ABX females, though they did not fully recover to the levels observed in Ctrl females (one-way ANOVA, F(3, 28) = 455.4, p < 0.0001, η² = 0.98, 95% CI [0.96, 0.97]; F(3, 28) = 20.3, p < 0.0001, η² = 0.69, 95% CI [0.42, 0.78]; Fig. 5l,m). Lolal and dpp knockdown had no effect on egg length in ABX females (Fig. 5n). We indeed observed that dpp mRNA expression was upregulated in ABX ovaries as Lolal abundance increased, whereas NA supplementation recovered dpp mRNA expression (Figs. 4i and  5o). These results suggest that excessive Lolal induces dpp overexpression and inhibits female reproduction in ABX females.

Interestingly, upon knockdown of Lolal or dpp in normal females (Supplementary Fig. 13a), most ovaries from dsLolal-injected and dsdpp-injected females at 8 days post-eclosion were at grades Ⅳ and Ⅴ, whereas dsGFP-injected ovaries predominantly reached grade Ⅴ filled with mature eggs (Chi-square test, χ²(4) = 11.8, p = 0.019, Cramér’s V = 0.22, 95% CI [0.03, 0.32]; Supplementary Fig. 13b, c). The average egg length of dsLolal-injected females was 1.09 ± 0.05 mm, which was shorter than that of dsGFP females (Supplementary Fig. 13d, e). We surmised that this phenomenon was due to decreased dpp signaling, the blockage of which has been associated with egg chamber degeneration⁴⁰. Indeed, dpp knockdown resulted in a similar phenotype, with average egg length reduced to 1.07 ± 0.12 mm (one-way ANOVA, F(2, 42) = 22.6, p < 0.0001, η² = 0.52, 95% CI [0.28, 0.65]; Supplementary Fig. 13d, e). In addition, knockdown of Lolal and dpp impaired female reproductive performance, resulting in reduced egg production (one-way ANOVA, F(2, 21) = 114.9, p < 0.0001, η² = 0.92, 95% CI [0.81, 0.94]; Supplementary Fig. 13f) and a lower hatching rate (one-way ANOVA, F(2, 21) = 7.8, p = 0.003, η² = 0.43, 95% CI [0.07, 0.61]; Supplementary Fig. 13g), but had no effect on larval viability (Supplementary Fig. 13h). These findings indicate that the Lolal-dpp pathway acts as an intermediary in the formation of mature eggs and its homeostasis is required for female reproduction. Taken together, our results show that the transcription factor Lolal is a key mediator between gut symbionts and host insect reproduction.

Discussion

In this study, we show that the metabolite of the gut bacterium E. hormaechei regulates host reproduction via ubiquitination of the transcription factor Lolal, a critical hub for host-microbiota interactions during reproductive processes. This regulation is driven by gut bacteria-derived nicotinic acid (NA), which activates the ubiquitin-proteasome system (UPS) to prevent Lolal overexpression from suppressing ovarian development and to maintain Lolal levels that allow normal female reproduction.

The gut microbiota exerts widespread effects on host gene expression beyond the gut, modulating pathways involving immunity, metabolism, and growth^13,15,41. In Drosophila, Acetobacter pomorum activates the insulin pathway to promote growth¹³. Further molecular investigations have revealed that some microbiota reshape gene expression via histone modifications and DNA methylation. For instance, Hamiltonella alters the histone H3 methylation profile to regulate the expression of mitochondria-related genes in whiteflies⁴², while viral infections in Bombyx mori can induce DNA methylation changes⁴³. Short-chain fatty acids produced by microbiota may inhibit host histone deacetylase activity, leading to altered acetylation proximal to promoter regions⁴⁴.

Beyond transcriptional control, microbiota can regulate the post-translational modification levels. Microbiota can regulate host ubiquitin and ubiquitin-like modifications by secreting effectors or modulating the activity of regulatory factors⁴⁵. For example, the SdeA effector secreted by Legionella has been reported to control host conventional ubiquitination⁴⁶. In this study, we found that gut bacteria-derived NA influences global ubiquitination and proteasome activity in an ATP availability-dependent manner. Indeed, a reduced ATP level could inhibit the ubiquitination pathway under energy stress^35,47. Ubiquitination ensures protein homeostasis by working with the proteasome, enabling rapid adaptation to environmental changes. Specifically, ubiquitination facilitates the timely degradation of transcription factors. UPS-mediated degradation of clock-associated transcription factors enables dynamic circadian gene expression⁴⁸. In the Drosophila immune response, ubiquitination activation is required for the cleavage of the transcription factor Relish, which initiates downstream signaling⁴⁹. Additionally, Relish has been identified as a central regulator of microbiota-mediated transcriptional responses in the Drosophila gut¹⁵. Regulation of transcription factors facilitates broader microbial regulation of host gene networks.

The performance of insects, particularly fertility, can be shaped by their microbial communities. In natural environments, insects often face fluctuating or limited nutrient availability. They have evolved nutritional symbiosis with microbial partners that provide essential nutrients, particularly for amino acids and vitamins that insects cannot synthesize de novo or obtain sufficiently from their diets. The gut symbiont Candidatus Erwinia dacicola recycles urea derived from bird droppings (a natural food source) into available nitrogen that supports the reproductive processes of B. oleae⁵⁰. In aphids, Buchnera aphidicola supplies the essential amino acid tryptophan, which is absent from their plant sap diet^51,52. Insects lack kynureninase and quinolinate synthase, which are required for NA synthesis from tryptophan³¹, making them dependent on external NA or NAM sources. Here, we show that the gut bacterium E. hormaechei contributes to female reproductive success by supplying vitamin B3 (NA), a key NAD precursor. This bacterium possesses a complete NA biosynthesis pathway and secretes abundant NA, as is the case with other NA-producing symbionts like Coriobacterium glomerans, A. pomorum, and Lactobacillus plantarum^7,53. Metabolite analysis of E. hormaechei detected the presence of NA rather than NAM, highlighting the central role of gut bacteria-derived NA in insect NAD metabolism. During coevolution, some endosymbionts have also developed cooperative interactions with their hosts to synthesize essential metabolites (e.g., vitamins B5 and B9) that support host reproduction^42,54. Their reproductive manipulations further enhance vertical transmission and facilitate the spread of symbionts in natural populations⁵⁵. Endosymbionts and their hosts have an ancient and stable symbiotic relationship, while gut microbiota represent a more flexible interface. Polyphagous insects usually undergo shifts in gut microbiota to adapt to different host plants⁵⁶. In Riptortus pedestris, gut microbiota are selectively acquired from the soil to enhance egg production under natural selection pressure⁵⁷. Although most gut bacteria are primarily acquired through horizontal transmission, host-mediated selection and functional coadaptation can ensure the long-term stability of specific microbial partners within host populations.

Some microbiota-derived metabolites function not only as nutritional sources but also as precursors of active coenzymes that drive metabolic reactions. Certain bacterial metabolites have been shown to influence epigenetic modifications in the reproductive system^20,42. During oogenesis, numerous tightly clustered mitochondrial networks are present in the egg chamber, and the activity of maternally derived mitochondria is crucial for early embryogenesis^58,59. Deficiency of gut microbiota and B vitamins leads to mitochondrial dysfunction in ovarian follicle cells¹⁸, whereas adequate provision of mitochondrial cofactor precursors can enhance intracellular energy efficiency⁶⁰. Our study showed that gut bacteria supply the developing ovary with NA, which enables long-distance regulation of reproductive health. The observed reduction in fecundity resulted from suppressed ovarian development, which was caused by NA insufficiency-induced impairment of ATP production and UPS activity. This phenotype would be expected to persist if the gut microbiota or dietary NA remained absent over an extended period. Notably, bacterial supplementation restored ovarian energy and development more effectively than NA supplementation alone, suggesting synergistic effects of other microbial factors, such as folic acid and riboflavin^18,42.

Dpp, a member of the bone morphogenetic protein ligand family, is a key signaling molecule that maintains the balance between germline stem cell (GSC) self-renewal and differentiation by regulating Mad (Mothers against decapentaplegic) phosphorylation in Drosophila⁶¹. In addition, dpp signaling is also required for the formation of mature eggs by establishing proper follicle cell patterning for egg chamber formation and oocyte polarity^12,62. Notably, mutations in the dpp receptor saxophone lead to partial disruption of posterior follicle cell fate and egg chamber degeneration⁴⁰. We found that gut bacteria regulate female reproduction by modulating dpp expression in the ovary. Upregulated dpp expression was detected in the gut bacteria-depleted ovaries, and this increase was associated with suppressed ovarian development and reduced egg production. Indeed, dpp overexpression can induce GSC-like tumors and repress germline differentiation⁶³. This repression is characterized by increased phosphorylation of Mad and transcriptional repression of bag-of-marbles in the germarium⁶¹. We further found that the transcription factor Lolal binds to the dpp promoter region and regulates its transcription. Lolal has been found to localize at more than 300 sites on Drosophila polytene chromosomes, with Trithorax-like as a potential partner of Lolal for recruiting DNA-binding proteins and regulating gene expression⁶⁴. Similar to our results, the BTB transcription factors bric-à-brac paralogs (bab1/bab2) have been reported to activate dpp expression in adult ovaries⁶⁵. In particular, Lolal has been shown to genetically interact with the bab locus⁶⁴. These findings suggest that Lolal likely functions cooperatively with other BTB transcription factors to regulate dpp transcription during oogenesis. We found that Lolal downregulation reduces dpp expression, leading to decreased mature egg production and shortened egg length. These phenotypes may be a consequence of impaired transfer of nurse cell contents to the oocyte, due to an irregular F-actin network^40,66. However, Lolal RNAi and dpp RNAi did not fully inhibit oogenesis, suggesting that the residual dpp activity is sufficient to support partial ovarian development.

In summary, our study shows that a gut microbiota metabolite enhances host reproduction by upregulating ubiquitination and proteasomal activity (Fig. 6). Gut commensal bacterium E. hormaechei-derived nicotinic acid improves host mitochondrial function and energy production, which reshapes the ubiquitination profile. In addition, we determined that ubiquitin-mediated Lolal homeostasis is required for ovarian development in B. dorsalis. Our study underscores the interplay between microbial products and ubiquitin-mediated proteostasis for maintaining reproduction. Thus, manipulation of gut microbiota may offer novel strategies for insect pest control by inhibiting female reproduction.

Fig. 6. Model of how gut bacteria affect B. dorsalis reproductive development.

Gut bacteria, particularly E. hormaechei, provide nicotinic acid (NA) to the host. The removal of gut bacteria by antibiotic treatment (ABX) reduces NA content and suppresses ovarian development. NA is essential for mitochondrial coenzyme NAD(H) biosynthesis. In ABX females, reduced NA metabolism diminishes mitochondrial ATP production, subsequently disrupting UPS activity. Hypo-ubiquitinated Lolal exhibits enhanced stability, leading to its accumulation in ABX female ovaries. Lolal induces the transcription of dpp and is required for female reproduction. However, excessive accumulation of Lolal induces dpp overexpression, disrupting ovarian development. Gut bacteria-derived NA promotes ATP production, which supports ubiquitination and proteasomal degradation, a process that modulates Lolal homeostasis and promotes female reproduction (Thicker lines emphasize greater regulatory impact). Created in BioRender. Qiao, J. (2026) https://BioRender.com/g1axqba.

Methods

Insect rearing

Oriental fruit flies were raised at the Institute of Horticulture and Urban Entomology, Huazhong Agricultural University (Wuhan, China), under conditions of 27 ± 1 °C, 70–80% relative humidity with a 12:12 h (L:D) photoperiod. Larvae were fed with larval food (banana, 200 g; corn flour, 200 g; sucrose, 40 g; yeast powder, 40 g; water, 200 mL). After eclosion, the adult flies were transferred to a mesh cage and provided with an artificial diet composed of a sucrose-to-yeast extract ratio of 3:1. We have complied with all relevant ethical regulations for B. dorsalis testing and research.

Bacterial removal and re-colonization

To eliminate gut microbiota, about 60 newly emerged females were released into a box and fed with the aforementioned sterile liquid diet supplemented with 3 mg/mL penicillin and 5 mg/mL streptomycin for 3 days. After antibiotic treatment, the flies were fed with a sterile diet, and the obtained females were denoted as ABX females. Control insects (Ctrl) were provided with diets containing no antibiotics. Both the antibiotic-infused and standard diets were replaced every 24 h.

In the bacterial recolonization experiment, ABX females were provided with a bacteria-containing diet on the fourth day. For the re-introduction of all culturable gut bacteria, 10 female guts were dissected and homogenized in 500 μL sterile PBS. A 100-μL aliquot of the gut homogenate was inoculated into 100 mL LB medium and cultured at 37  °C for 8 h. The culture was then centrifuged at 5000×g for 10 min, and the bacterial cells were resuspended in sterile diet to achieve a final concentration of OD₆₀₀ = 5. For single-bacterium reintroduction, equivalent levels of each bacterial strain were introduced to ABX females: individual bacterial strains were cultured in LB medium at 37 °C, and cultures were collected at different time points to measure optical density. Then, serial dilutions of the bacterial cultures were plated on LB agar to determine colony-forming units (CFU) and establish the OD–CFU relationship for each strain. The resuspension volume of each strain was determined based on the CFU mL⁻¹ at an OD of 5 to standardize the supplemented diet to 1 × 10⁹ CFU/mL²⁹. The bacteria were cultured freshly, and the diet was renewed daily to ensure continuous feeding over a period of 5 days. The above diets were added to cotton pads and placed within ultraviolet-sterilized plastic dishes.

Gut bacterial load assessment

To evaluate the effects of antibiotic treatment and bacterial recolonization, the flies were surface-sterilized with 75% ethanol, and their guts were dissected under sterile conditions using PBS. Experiments were performed as previously described²⁹: Guts from 10 flies were pooled as one sample and homogenized in 1 mL PBS. CFUs were determined by spread plating diluted gut suspension. Gut microbial DNA was extracted from 10 fly guts using the E.Z.N.A.® soil DNA Kit (Omega Bio-tek, D5625). The total bacterial loads were quantified by real-time PCR on gut bacterial DNA using the universal bacterial 16S rRNA primers (forward: 5’-ACTCCTACGGGAGGCAGCAG-3’; reverse: 5’-TACCGCGGCTGCTGG-3’)⁶⁷, with the host β-actin gene used as an internal reference.

Food preparation

Liquid diet was prepared using a sucrose: yeast extract: water ratio of 6:1:60, and sterilized at 121 °C for 15 min. To ensure sterile conditions, the cooled diet was packed into sterile plastic tubes. For experiments with nutrient metabolite replenishment or rotenone, the diet was mixed with the following metabolites to achieve a final concentration of 100 μg/mL NA (Solarbio, N8060), 100 μg/mL PA⁶⁸ (Solarbio, IPA10490), 200 mM melibiose (Solarbio M8420), 200 mM mannitol (Solarbio M8141), 9 μM XA⁶⁹ (MedChemExpress, HY-W014666), 9 μM neopterin (MedChemExpress, HY-W040055), and rotenone (APExBIO, B5462) at 1 mM.

Cells and reagents

The B. dorsalis embryonic cell line was established from insect eggs and has been stably passaged more than 120 times (RRID: CVCL_F0G0). Cells were cultured at 28 °C in TNM-FH medium (Pricella, PM152010) supplemented with 10% fetal bovine serum. Cycloheximide (CHX, Sigma, C7698), MG132 (Santa Cruz Biotechnology, sc-201270), and Chloroquine (CQ, MCE, HY-17589) were dissolved in 100% dimethyl sulfoxide (DMSO) and then diluted in medium for in vitro studies. Plasmids and siRNAs were transfected into cells using TransIT®-Insect Transfection Reagent (Mirus Bio, MIR 6105) following the manufacturer’s instructions.

Construction of PncA knockout bacteria

To disrupt the PncA gene (Gene ID: 93199804) of E. hormaechei bacteria, we employed the SacB-based gene knockout system⁷⁰. The suicide vector for deleting the target gene was constructed by amplifying the flanking region using primers with homologous integration sequences, allowing directed insertion into the vector pH73sacB. Briefly, the upstream region (827 bp) and downstream region (750 bp) PCR products were seamlessly connected to the vector pH73sacB digested with SmaⅠ. The constructs were confirmed by sequencing and then transformed into E. coli S17 λpir, which served as the donor for plasmid conjugational transfer to wild-type E. hormaechei. E. coli S17 λpir containing the plasmid was transferred to LB medium containing 34 µg/mL chloramphenicol (CM), while E. hormaechei was cultured in LB medium without antibiotics. When the OD₆₀₀ reached 0.7, 700 µL of E. hormaechei and 700 µL of S17 λpir were mixed and centrifuged at 4000×g for 3 min. The bacterial mixture was transferred to an LB agar plate and incubated overnight at 37 °C. Conjugational transferred colonies were diluted and inoculated onto LB plates containing both ampicillin and chloramphenicol, and then cultured at 37 °C. Single colonies were cultured in LB medium containing 10% sucrose. After counter-selection on CM-resistant LB agar plate, the obtained clones were subjected to PCR amplification and sequencing to confirm gene knockout.

RNAi experiments

This study employed dsRNA for RNA interference. The target sequence fragments were amplified by template-specific primers conjugated with the T7 RNA polymerase promoter (5’-GGATCCTAATACGACTCACTATAGG-3’)⁶⁷. The primer sequences are detailed in Supplementary Data 4. The dsRNA was transcribed by T7 Ribomax Express RNAi System (Promega, USA) according to the manufacturer’s instructions. Microinjection of dsRNA was performed using the Eppendorf micromanipulation system. First, a 2000 ng/mL dsRNA solution was injected into the abdomen of 1-day-old females and antibiotic-treated flies before providing an E. hormaechei or NA-containing diet. The control flies were injected with dsGFP. The average injection volume was 1 µL.

Reproductive phenotypic analysis

Following the experimental treatments, ovaries were dissected from females at 8 days post-eclosion to evaluate development grade. Ovarian development grades are as follow⁷¹: Grade Ⅰ: Ovary is small and ovarioles are not distinguishable; Grade Ⅱ: Oocytes begin to differentiate and ovarioles become visible; Grade Ⅲ: Yolk deposition occurs and oocytes rapidly increase in size; Grade Ⅳ: Mature oocytes are visible in some ovarioles; Grade Ⅴ: Ovarioles are fully packed with mature oocytes, and lateral oviducts contain mature eggs.

One 10-day-old female and two 10-day-old males were placed in individual 200 mL bottles equipped with a yellow egg collection cup containing banana medium. After mating, eggs were collected five times at 2-day intervals, with the banana replaced each time. Flies were allowed to lay eggs for 3 h during each collection period. Then, the total number of deposited eggs, egg length, and egg hatching rates were counted.

LC–MS/MS analysis of NA

For the targeted analysis of NA, sample preparation and detection methods were developed based on reports from Igor et al. and Feng et al.^72,73. To test the ability of culturable gut bacteria to produce NA, equal amounts of E. hormaechei, PA, PV, KA, and KQ were inoculated into M9 medium supplemented with 3 mM ASP and incubated in the dark for 36 h. The ∆PncA E. hormaechei strain and E. hormaechei strain were inoculated into M9 medium containing 60 µM NAM and incubated in the dark for 20 h. The bacterial culture was centrifuged at 5000×g for 10 min. The supernatant was then filtered through a 0.22 µm membrane and subjected to LC–MS/MS analysis. For hemolymph samples, the samples were diluted 20–40 times with 80% methanol/water (v/v) and extracted at 4 °C in the dark for 1 h. The mixture was then centrifuged at 12,000×g for 15 min at 4 °C, and the supernatant was transferred to amber glass vials for analysis. For ovarian tissue, the samples were homogenized in PBS and sonicated for 25 s. A portion of the supernatant was used for protein quantification. Lysates with equal protein concentrations were extracted in 80% methanol/water (v/v) at 4 °C in the dark for 1 h. After centrifugation at 12,000×g for 15 min, the supernatant was used for NA analysis.

The calibration standard of NA was prepared by diluting 100 µg/mL NA standard stock solution (LGC Standards, CAS Rn: 59-67-6) with methanol, resulting in working concentrations of 5, 25, 50, 100, 300 ng/mL for NA. Standard solutions were stored in the dark at 4 °C. Analyses were performed using an HPLC system (Shimadzu LC-30A) coupled with a Qtrap 4000 mass spectrometer (SCIEX). The LC separation was performed on a reversed-phase C18 column (150 mm × 2.1 mm, 5 µm particle size, GL Sciences). Solvent A was 0.1% (V/V) formic acid in water, and solvent B was methanol with 2 nM ammonium acetate. The gradient was 0–2 min, 99%–99% A; 3–8 min, 5%–5% A; 8–15 min, 99%–99% A. Other LC parameters: flow rate 300 µL/min, column temperature 30 °C, auto-sampler temperature 4 °C, and injection volume 10 µL. Mass spectrometry was performed with MRM mode, monitoring the transition of the precursor ion m/z 124 to the product ion m/z 80 for NA. Ion spray voltage was set to 4000 V, with a source temperature of 350 °C. The calibration curve was generated based on the peak area and the concentration of the standard. The sample concentration was calculated based on the calibration curve.

NAD⁺/NADH and ATP assay

Cellular NAD and ATP levels were quantified using the Coenzyme I NAD (H) Content Assay Kit (Boxbio, AKCO001M) and ATP assay kit (Beyotime, S0026B), respectively. Samples were homogenized individually in extraction buffers at low temperatures, and subsequent steps were carried out according to the manufacturer’s instructions. The NAD and ATP contents were normalized to the concentration of tissue protein.

COX activity staining and TEM

Ovaries from Ctrl, ABX, and NA-fed flies were snap frozen in liquid nitrogen. Then, 10 μm-thick frozen sections were assayed for in situ histochemical COX activity. The buffer composition used for the assay was 50 mM phosphate buffer (pH 7.4) with 4 mM 3,3-diaminobenzidine (DAB), 2 μg/mL catalase, 200 μM cytochrome c, and sucrose. Sections were incubated with COX staining solution for 30 min, washed twice in 50 mM phosphate (pH 7.4) for 5 min each, and scanned under white light. Relative COX activity in the ovaries was quantified using ImageJ.

For transmission electron microscopy (TEM) analysis, ovarian samples were fixed in 2.5% glutaraldehyde, dehydrated in ethanol containing 1% osmium tetroxide, and embedded in epoxy resin. Semi-thin sections (1 μm) were stained with 0.5% (w/v) toluidine blue and 1% (w/v) boric acid to locate regions of interest, which were analyzed by light microscopy. Ultra-thin sections (60–90 nm) were then cut, stained with uranyl acetate and lead citrate, and observed using a JEOL 1400 transmission electron microscope.

Quantitative RT-PCR

RNA was extracted from dissected ovaries using the TRIzol-chloroform method. Each pool contained 10 ovaries. cDNA was synthesized from 1 µg RNA using PrimeScript^TM RT reagent kit (Takara) with gDNA eraser to remove DNA contamination. qPCR was performed using QuantStudio™ 7 Flex Real-Time PCR system (Thermo Fisher Scientific) with SYBR Green qPCR mix in 384-well plates under the following cycling conditions: 95 °C for 5 min, 40 cycles of 95 °C for 15 s, 60 °C for 30 s and 72 °C for 30 s; dissociation curves: 95 °C for 15 s, 60 °C for 1 min, 95 °C for 15 s to confirm single amplification. At least three independent biological replicates and three technical replicates were calculated by the 2^−ΔΔCt method. All values were normalized to internal control α-tubulin. Primers are listed in Supplementary Data 4.

Plasmids and siRNA transfection

The cDNA encoding Lolal was cloned into the pBac-IE1 plasmid stored in the laboratory. For transfection, cells were incubated with 2 µg plasmid and 1 µL Insect Transfection Reagent for 12 h. Gene knockdown experiments were carried out by siRNA. siRNA oligos were obtained from Beijing Tsingke (The sequences of siRNAs are shown in Supplementary Data 4). Cells were transfected with the individual siRNA pools at 50 nM in 1 µL Insect Transfection Reagent for 12 h.

Polyclonal antibody generation

Whole ORFs (open reading frames) of Bd-Lolal were amplified and cloned into the pET28b vector. The recombinant proteins were expressed in BL21 induced by 1 mM IPTG at 37 °C for 4 h. The bacterial cells were ultrasonically lysed in the lysis buffer (PBS, pH 7.5, 10% Glycerol, 1 mM PMSF) after centrifugation. The inclusion bodies, which contained recombinant protein, were re-suspended in denaturing solution (8 M urea in PBS, pH 7.5) and subjected to ultrasonic crusher. The recombinant protein was purified with Ni NTA Beads 6FF (SMART Life Sciences, China) and used by AtaGenix (Wuhan, China) to immunize rabbits for polyclonal antibody production in accordance with its institutional animal care and ethical guidelines. The resulting antisera were affinity-purified before being applied in immunoblotting.

Prokaryotic expression and protein injection

We constructed 6×His-Lolal-6ⅹHis (Lolal-His) and GST-His protein expression plasmids using the pET-28a vector. Plasmids were transformed into Escherichia coli BL21 competent cells, and clones were selected on kanamycin LB medium followed by DNA sequencing. A 2 mL culture of positive clones was inoculated into 200 mL of LB medium supplemented with kanamycin and incubated at 37 °C, 220 rpm, until the OD₆₀₀ reached 0.7. Protein expression was induced with 0.6 mM IPTG at 20 °C, 180 rpm for 16 h. Cells were collected at 6000×g for 15 min followed by resuspension in NETN lysis buffer (0.5% NP-40, 1 mM EDTA, 20 mM Tris–HCl, and 100 mM NaCl). After sonication for 5 min, the supernatant was collected and purified using a Ni-NTA column. Proteins were concentrated using ultrafiltration tubes (Millipore) with imidazole replaced by PBS. The purified proteins were analyzed by SDS–PAGE and Western blotting.

The concentrations of Lolal-His and GST-His recombinant proteins were determined using a BCA protein assay (Beyotime, P0010S). A dose of 3 ng Lolal-His was injected into 2-day-old females by Eppendorf micromanipulation system, followed by a second injection after 4 days. GST-His was injected as a control. Ovarian development grades, number of eggs laid, and egg hatching rates were recorded as described above.

Protein extraction and western blot analysis

Ovaries and cells were lysed in RIPA lysis buffer (150 mM NaCl, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris) and protease inhibitor cocktail. The lysate was incubated on ice for 30 min, and then centrifuged at 13,000×g at 4 °C for 10 min. Boiled samples were separated by SDS–PAGE and then transferred onto polyvinylidene fluoride membranes (Bio-Rad, 1620177). The membrane was blocked with 5% skim milk at room temperature for 30 min, and then incubated with primary antibody (mouse anti-β tubulin (Frdbio, RAB0102), 1:5000; rabbit anti-Lolal, 1:5000; mouse anti-ubiquitin (Santa, sc-271289), 1:3000 in 3% BSA at 4 °C overnight. Secondary antibody (Goat anti-Mouse IgG and Goat anti-Rabbit IgG (Biosharp, BL001A, BL003A), 1:5000) was incubated at room temperature for 1 h. ECL highly sensitive substrate (Monad, PW30601S) was used for immunological blot detection.

Ubiquitination assays

For the ubiquitination assay, cells transfected with indicated plasmids or siRNAs and insects were treated with 10 µM MG132 for 8 h. Cell pellets or tissue samples were lysed in RIPA lysis buffer as described above and ultrasonicated for 5 min (10 pulses, 20 s each). The lysates were incubated with anti-Lolal antibody or rabbit IgG (GenScript, A01008) coupled with rProtein A/G MagPoly Beads suspensions (SMART Life Sciences, SM015005) at 4 °C overnight. The subsequent steps were the same as those for immunoprecipitation described above.

Chromatin immunoprecipitation (ChIP)

The ChIP assay was performed using the Sonication ChIP Kit (ABclonal, RK20258) following the manufacturer’s protocol. Briefly, 120 ovaries (4-day-old) were cross-linked in PBS (1% formaldehyde) for 20 min at room temperature. The samples were homogenized in the lysate (ChIP sonication buffer, 1 mM DTT (dithiothreitol), 1% SDS) and sonicated to shear chromatin into 200–1000 bp fragments. After the chromatin concentration was equalized, 5% input was used as the control. Lysate containing 15 μg of chromatin was further incubated with anti-IgG antibodies (GenScript, A01008, 8 μg per ChIP assay) or anti-Lolal (8 μg per ChIP assay) and rProtein A/G MagPoly Beads (SMART Life Sciences, SM015005) at 4 °C overnight. BSA served as an empty control (EC). After incubation, the beads were washed, and the DNA was eluted. Finally, DNA was de-cross-linked and purified for PCR and qPCR analysis.

16S rDNA sequencing and analysis

Total gut bacterial DNA was extracted from Ctrl and ABX females using the Soil DNA Kit (Omega Bio-tek, D5625). The hypervariable region V3–V4 of the bacterial 16S rDNA gene was amplified with primer pairs 338 F: 5’-ACTCCTACGGGAGGCAGCA-3’ and 806R: 5’-GGACTACHVGGGTWTCTAAT-3’. The amplicon library was paired-end sequenced (2 × 250) on the Illumina MiSeq sequencing platform (Biomarker Technologies Co., Ltd., Beijing, China). After quality filtering, reads were assembled and clustered into operational taxonomic units (OTUs) at 97% similarity using USEARCH software. The complexity of species diversity of each sample was analyzed using QIIME2 software⁶. The bioinformatics analysis of this study was performed with the aid of the BMKCloud (www.biocloud.net).

Nontargeted metabolomics analysis

Hemolymph samples (100 µL) were collected from Ctrl and ABX females and mixed with acetonitrile (1:4, v/v) to precipitate proteins²⁹. Six biological replicates were included for each group. Samples were sent to Novogene Bioinformatics Technology Co. Ltd. (Beijing, China) for metabolite detection using Vanquish UHPLC system coupled with Q Exactive HF mass spectrometer (Thermo Fisher Scientific, USA). Briefly, samples were separated on a Hypersil Gold C18 column (100 × 2.1 mm, 1.9 µm, Thermo Fisher Scientific, USA) at a flow rate of 0.2 mL/min using solvent A (0.1% formic acid (FA) in water) and solvent B (methanol). The mass spectrometer was operated in positive/negative polarity mode with spray voltage of 3.5 kV, capillary temperature of 320 °C, sheath gas flow rate of 35 psi, and aux gas flow rate of 10 L/min, S-lens RF level of 60, Aux gas heater temperature of 350 °C.

The raw data files were processed using Compound Discoverer 3.3 (Thermo Fisher Scientific, USA) for peak alignment, peak detection, and metabolite quantification. Molecular formula prediction was performed based on additive ions, molecular ion peaks, and fragment ions. Metabolite identification was achieved by matching the mzCloud and mzVault databases. Differential metabolites were identified based on VIP > 1, p < 0.05, and a fold change of ≥1.5. Metabolomics data analysis was then performed using MetaboAnalyst 6.0.

Sample preparation for proteome and ubiquitinated peptides enrichment

Ovaries of Ctrl, ABX, and E. hormaechei-recolonized ABX (EH) females were collected, with each sample weighing 200 mg. Two biological replicates were used for the assessment. Ovaries were frozen in liquid nitrogen and finely ground, and protein was extracted using lysis buffer (8 M urea, 1% protease inhibitor cocktail, and 50 µM PR-619). The protein concentration in the supernatant was measured using the BCA assay, and the protein quantity was determined by Coomassie staining following SDS-PAGE. Proteins were reduced in 5 mM DTT at 30 °C for 45 min and alkylated in 15 mM iodoacetamide (IAA) at room temperature in the dark for 30 min. Next, 100 mM tetraethylammonium bromide was added to dilute the urea concentration to 2 mM. All samples were digested overnight at 37 °C with trypsin being added at a 1:50 trypsin-to-protein mass ratio. Finally, the peptides were desalted by a C18 SPE column, which was then concentrated by vacuum centrifugation for proteomic and ubiquitinome mass spectrometry identification, respectively.

The lyophilized peptides were dissolved in NETN buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris–HCl, 0.5% NP-40, pH 8.0), and the supernatant was incubated with the pre-washed di-Gly-Lys antibody beads (PTM1104, PTM Bio, China) overnight at 4 °C with gentle rotation. Beads were washed four times with ice-cold NETN buffer and twice with ice-cold mass spectrum water. The enriched diGly-modified peptides were eluted from the beads using 0.1% trifluoroacetic acid. Finally, the eluted fractions were vacuum-dried for the next step LC–MS/MS analysis.

Mass spectrometry identification

Peptide samples were dissolved in solvent A (0.1% FA, 2% acetonitrile (ACN)) and loaded onto a reverse-phase trap column. Peptides were separated by the nanoElute UHPLC system (Bruker Daltonics) and analyzed by the timsTOF Pro (Bruker Daltonics) mass spectrometer. Peptides were separated with a gradient from 6% to 24% solvent B (0.1% FA, 98% ACN) over 70 min, a 24–35% gradient in 14 min, and increasing to 80% in 3 min, then held at 80% for the last 3 min, all at a constant flow rate of 450 nL/min. MS spectra were acquired with an MS/MS scan range from 100 to 1700m/z. The timsTOF Pro was operated in parallel accumulation serial fragmentation (PASEF) mode. Precursors with charge states of 0–5 were selected for fragmentation, and 10 PASEFMS/MS scans were acquired per cycle. A protein was confidently identified only if it contained at least one unique peptide, applying a threshold of false discovery rate (FDR) ≤ 1.0%.

Proteins showing significant differences between groups were identified based on a fold change of ≥1.3 and a coefficient of variation (CV) < 0.1. Functional annotation was performed using the Bactrocera dorsalis protein database from UniProt. Gene Ontology (GO) terms, including biological processes (BPs) and molecular function (MF), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using the OmicsBean platform (www.omicsbean.cn).

Statistical analysis

Statistical analyses were performed using GraphPad Prism 8 and SPSS 30. Data normality and homogeneity of variance were assessed to determine appropriate parametric or nonparametric tests. Gut bacterial load after antibiotic treatment, bacterial growth, and gene expression following Uba1, Lolal, or dpp knockdown were compared using the two-tailed unpaired t-test. Egg production, NAD(H) and ATP contents, egg length, gene expression, and protein abundance among ABX, bacteria-recolonized, or NA-supplemented females, egg production and hatching rates across three or more groups were analyzed by one-way ANOVA with Tukey’s multiple comparisons test. Analysis of NA levels was performed using one-way ANOVA, except for ovarian NA levels among Ctrl, ABX, and bacteria recolonized females (Kruskal–Wallis with Dunn’s test). ATP content after Rot treatment, egg hatching rates after dsUba1 injection, and ChIP-qPCR were compared using the two-tailed Mann–Whitney test. Lolal degradation rates were analyzed by two-way ANOVA. Differences in ovarian development grades were evaluated using a two-tailed Pearson’s χ² test with Bonferroni-corrected post-hoc comparisons. Survival curves were analyzed by the Log-rank test. p < 0.05 was defined as statistically significant.

Statistics and reproducibility

No statistical method was used to predetermine sample size. Sample sizes were chosen based on previous studies. No data were excluded from the analyses. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment. All experiments included at least three independent biological replicates, unless otherwise stated.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

H.Z. and X.L. supervised the study. H.Z., J.Q., and X.L. designed the study. J.Q. performed the experiments, collected samples, and analyzed data. Z.L. contributed to the cell experiments. Q.Z. assisted with the isolation and identification of gut bacteria. C.Z. maintained the insect colonies. H.Z., X.L., J.Q., and W.Z. prepared the manuscript. All authors read and approved the final manuscript.

Peer review

Peer review information

Nature Communications thanks Yoshitomo Kikuchi and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The 16S rDNA sequencing data reported in this study have been deposited in the Genome Sequence Archive under accession code CRA033033. Non-targeted and targeted metabolomic data have been deposited in MetaboLights under accession numbers MTBLS13264 and MTBLS12798. The proteomic data have been deposited in the Proteome Xchange Consortium under the dataset identifier PXD066837. Additional data are available through Figshare (10.6084/m9.figshare.30672845). Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-69021-y.