Genetic variation in gut microbe as a key regulator of host social behavior in C. elegans

ABSTRACT

Gut microbiota have been shown to influence the social behaviors of their hosts, while variations in host genetics can affect the composition of the microbiome. Nonetheless, the degree to which genetic variations in microbial populations impact host behavior, as well as any potential transgenerational effects, remains inadequately understood. Utilizing C. elegans as a model organism, we identified 77 strains of E. coli from a total of 3,983 mutants that significantly enhanced aggregation behavior through various neurobehavioral pathways. This discovery underscores a collaborative regulatory mechanism between microbial genetics and host behavior. Notably, we observed that some mutant bacteria might affect social behavior via the mitochondrial pathway. Additionally, the modulation of social behavior has been identified as a heritable trait in offspring. Our results provide a novel perspective on the regulatory role of microbial genetic variation in host behavior, which may have significant implications for human studies and the development of genetically engineered probiotics aimed at enhancing well-being across generations.

Introduction

Innate social behavior is regulated by neuronal circuits that interpret sensory inputs, influenced by an individual’s genetic makeup, gender, and life experiences.¹ While genetic and environmental factors have long been known to shape behavior, decision-making, and emotions through central nervous system pathways, innate behaviors also demonstrate a remarkable capacity for rapid adaption to environmental changes.² Recent studies have highlighted gut microbes’ role in bidirectional communication with the brain, influencing host social behavior via multiple pathways.^3,4 For example, germ-free mice display deficits in social behavior that can be significantly restored by reintroducing gut microbiota through fecal transplantation.³ This microbial-gut-brain axis provides a new framework for understanding the mechanisms regulating host social behavior. However, the therapeutic potential of gut microbiota is currently limited by challenges in standardizing co-probiotic transplantation techniques and insufficient characterization of specific probiotic strains.^5,6

The co-evolutionary relationship between intestinal microbes and their hosts is shaped by genetic variation within the microbes, enabling them to adapt to changes in the host environment, modify metabolic functions, and fine-tune their interactions with the host.^7–9 These microbial genetic variations can, in turn, impact host social behavior through the production of metabolites and modulation of both nervous and endocrine systems. However, given the complexity and slow mutation rates of many microbial genes, significant genetic shifts within short timeframes remain challenging for most microbial gene regions. Understanding how microbial genetic variation affects innate social behavior is crucial for addressing the mechanisms behind behavior and social disorders. Currently, knowledge of the influence of microbial genetic variation on inherent social behavior is limited, leaving a critical gap in understanding the microbial factors that may drive social traits and behaviors in hosts. Moreover, the mechanisms by which microbial genetic variations regulate host behavior remain unclear, particularly regarding potential transgenerational effects.

C. elegans exhibits both solitary and gregarious social behaviors,¹⁰ with key roles attributed to various signaling pathways, including the C. elegans neuropeptide Y receptor (TAX-2/TAX-4-GCY-35/GCY-36),¹¹ TGF-β, TRPV,¹² and dopaminergic pathways,¹³ in regulating these behaviors. The simplicity of C. elegans’ gut microbiota makes it an excellent model for studying the relationship between microbial genetic variation and host social behavior. In controlled laboratory environments, C. elegans feeds exclusively on E. coli strains that colonize its gut, forming a defined gut microbiome.¹⁴ However, the mechanisms by which bacterial genetic factors influence host social behavior remain incompletely understood, and it is still unclear whether these microbial genetic variants produce transient or long-lasting effects on host social behavior.

In this study, we utilized the Keio collection library, a comprehensive set of 3,983 single-gene knockout mutants of nonessential genes in E. coli K-12 ¹⁵, to perform an unbiased high-throughput screening.¹⁵ Our research uncovers particular microbial genetic variants that influence the social behavior of hosts, and for the first time, we are elucidating the mechanisms involved through signaling pathways that are known to affect social behavior. Additionally, we observed that extended exposure to certain mutant microorganisms can affect the social behavior of C. elegans offspring, and we identified a link between mitochondrial genetic variation and social behavior in C. elegans. These discoveries establish an essential connection between microbiome genetics and host social behavior, offering valuable insights for the potential design of genetically engineered probiotics aimed at modifying social behavior. Our findings also shed light on the chemical communication network between bacterial metabolites and host neuronal pathways, broadening our understanding of host–microbe interactions in behavior regulation.

Results

High-throughput screening identifies E. coli mutants that increase C. elegans social behavior

Under natural conditions, N2 C. elegans typically display solitary behavior.¹¹ In this study, we used high-throughput screening to systematically identify E. coli gene knockout strains derived from the Keio library,¹⁵ which are capable of modulating the aggregation behavior of N2 C. elegans (Figure 1(a)). When exposed to wild-type control bacteria BW25113, N2 C. elegans displayed an average of 13% aggregation behavior, with 95% of the 35 replications showing social behavior values below 18% (Figure 1(b)). Based on these results, we established a threshold of 20% as a baseline for initial screening of mutant bacteria that could potentially enhance social behavior. Our primary screen of 3,985 E. coli mutants from the Keio library, which consists of single-gene deletions of all nonessential genes in E. coli K-12 that were previously developed, identified 407 bacterial mutants that increased social behavior in C. elegans (Figure 1(c)). These mutants were selected based on their ability to enhance social aggregation beyond a threshold of 20% compared to wild-type E. coli (BW25113). Subsequent re-screening of these 407 E. coli mutants in triplicate revealed that 77 mutants significantly enhanced the social behavior of C. elegans compared to the wild-type (WT) E. coli control strain (p < 0.05). From these, the top ten mutants were selected for further mechanistic analysis (Figure 1(d)) (Table S1).

Figure 1.

Identification of E. coli mutants that enhanced C. elegans social behavior. (a) Concept of high-throughput screening of C. elegans raised on E. coli mutants from the Keio library. (b) The average social behavior levels of 35 replicates in N2 C. elegans fed with wild-type E. coli bw25113. (c) Detailed flowchart for schematic in (a). (d) 10 E. coli mutants significantly increase the aggregating rate of the C. elegans, which exhibit early death due to age-related progression of germline tumors (*p < 0.05, **p < 0.01,***p < 0.001, T.Test) (n > 3).

Interaction of beneficial microbial factors with host social behavior mechanisms

Oxygen perception is a key modulator of social aggregation behavior in C. elegans.¹⁶ The N2 strain of C. elegans harbors a gain-of-function mutation in the neuropeptide receptor NPR-1, which reduces both aggregation and oxygen avoidance responses.^1,16 Under natural conditions, C. elegans tend to aggregate in lower-oxygen environments, such as the edges of dense colonies or near moss. Our study found a weak correlation between the influence of mutant bacteria on the social behavior of wild-type N2 C. elegans and their tendency to aggregate near the moss edge (Figure S1). To minimize potential confounding effects due to oxygen variability, we ensured consistent population sizes across each C. elegans group, adjusted by food moss concentration, thereby standardizing environmental oxygen levels.

To explore the regulatory mechanisms by which mutant bacteria enhance social behavior in C. elegans, we examined the aggregation behavior in mutants of gcy-35 and tax-4 (the GCY-35/TAX-4 pathway), daf-3 and daf-5 (the TGF-β pathway), ocr-2 (the TRPV signaling pathway), and dop-2 (the dopaminergic signaling pathway) (Figure 2a-f, red column). We also assessed gene expression levels of gcy-35, tax-4, daf-3, daf-5, ocr-2, and dop-2 in N2 C. elegans after exposure to the ten bacterial mutants (Figure 2a-f, green column).

Figure 2.

The relationship between the ten E. coli mutants and the signaling pathways of GCY-35/TAX-4, TRPV, TGF-β, and dopamine. (a-f: red bar column) E. coli mutants were examined for their effects on the aggregating rate of the gcy-35 (ok769), tax-4 (p678), ocr-2 (ak47), daf-3 (ok3610), daf-5 (e1386), and dop-2 (vs105) C. elegans mutants. (a-f: green bar column) the gene expression of gcy-35, tax-4, ocr-2, daf-3, daf-5, and dop-2 in N2 C. elegans was examined when fed with the ten E. coli mutants. An independent BW25113 control experiment was established for each mutant bacterium to analyze aggregating rate and associated gene expression in nematodes following feeding. The relative expression levels of relevant genes in N2 C. elegans, as well as the aggregating rate of C. elegans with different gene mutations, were quantified relative to their respective independent control experiments after feeding on different E. coli mutants * p < 0.05, ** p < 0.01, *** p < 0.001 *** p < 0.0001, Student’s t -test), error bars represent SEM.

Among the ten bacterial mutants examined, crL, iaaA, yccU, and yjgW significantly upregulated gene expression of gcy-35. Furthermore, the gcy-35 mutant demonstrated a pronounced impairment in the modulation of social behavior among these bacterial mutants, suggesting that their influence may operate via the gcy-35 signaling pathway. Similarly, ycdC functioned through tax-4 and dop-2. crL and yjgW mutant bacteria enhanced C. elegans social behavior through daf-3. Additionally, yjgW mutants regulate social behavior through daf-5, ocr-2, and dop-2. The results indicate that specific genetic variants of bacteria have an impact on the social behavior of the host through various signaling pathways.

Mutant bacteria exert influence on C. elegans social behavior through intricate regulatory networks

The analysis of the protein interaction network revealed that GCY-35, TAX-4, DAF-3, DAF-5, OCR-2, and DOP-2 play significant roles in social foraging pathways (Figure S2). These findings highlight a complex regulatory network governing C. elegans’ social behaviors and underscore the intricate pathways involved in modulating these behaviors through interactions with mutant bacteria.

This study also identified 18,800 genes in the tissues of C. elegans fed with mutant bacteria. A significant difference in the expression of 2615 genes was observed in C. elegans fed by ten mutant bacteria compared to control strain BW25113, with the majority of differentially expressed genes being upregulated (Figure 3(a)) (Table S2). Specifically, C04G2.5, nlp-25, C34B2.3, and M7.7 showed high expression levels in over 90% of C. elegans exhibiting enhanced social behavior (Figure 3(a)). Remarkably, 323 genes demonstrated significant upregulation or downregulation relative to the control group in more than half of the C. elegans with high aggregation behavior (Figure 3(b-c)) (Table S3). The majority of upregulated genes were notably found among the differential genes in c. elegans fed with ten mutant bacteria (Figure 3(b-c)) (316 vs 7). The 323 genes that exhibited differential expression in over half (≥50%) of the enhanced C. elegans aggregation behavior demonstrated significant alterations in gene composition and relative abundance across diverse cellular compartments, encompassing the cytoplasm, extracellular space, mitochondria, and nucleus (Figure 3(d)). GO function enrichment analyses indicated that these 323 genes were significantly enriched in cuticle development, protein phosphorylation, and mitochondrial phosphate ion transmembrane transport (Figure S3). The results suggest that the enhancement of social behavior in C. elegans necessitates intricate gene regulatory networks.

Figure 3.

The gene expression of C. elegans fed with ten mutant bacteria differs from that of the control strain BW25113. (a) Genes differentially expressed in C. elegans fed mutant and control bacteria. The red dots represent genes that are significantly up-regulated compared to the control group, while the blue dots represent genes that are significantly down-regulated relative to the control group. Differential gene expression analysis was conducted using the R package edgeR, applying a Fold change threshold of 2 and a significance level of 0.05 for the p value cutoff. The four labeled genes, C04G2.5, nlp25, C34B2.3, and M7.7, exhibited significantly elevated expression levels in over 90% of highly social behavior C. elegans. (b-c) He upset plot illustrates the intersection patterns of differentially expressed genes (DEGs) across ten experimental groups of C. elegans fed mutant bacterial strains compared to the control group (|FC| > 2, p value < 0.05). The horizontal histogram (left) displays the total number of DEGs per treatment group, while the vertical histogram (top) indicates the magnitude of intersecting genes across combinations of experimental groups. The connected dots matrix (central panel) visualizes specific group combinations, where filled dots represent participating groups in each intersection. (d) Subcellular localization of genes related to social aggregation behavior in C. elegans. Brown represents up-regulated genes in 5 or more C. elegans groups fed by mutant bacteria, while blue represents down-regulated genes in 5 or more C. elegans groups fed by mutant bacteria, (p < 0.05, Fold change > 2).

Genetic interaction analyses highlighted that ypdF, purH, torY, and ycgJ mutants were unrelated to specific neural genes. However, a comparative gene expression analysis identified 2442 differentially expressed genes, which were found to be enriched in mitochondrial functions, such as phosphorylation and ATP binding (Figure 4(a)). Furthermore, KEGG pathway analysis linked them to mitochondrial pathways and C. elegans social behavior. Additionally, a total of 323 genes exhibited significant alterations in response to treatment with multiple bacterial mutants in C. elegans, which displayed pronounced aggregation behavior. These genes were notably enriched in functions associated with mitochondrial activity (Figure 4(b)). Weighted gene co-expression network analysis (WGCNA) identified five distinct phenotypic modules associated with social behavior among the ten mutant bacteria (Figure 4(c)). Notably, torY and yccU mutants showed significant association with mitochondria function-related modules (MEgreen) (Figure 4(d)). Collectively, these findings indicate a significant role of mitochondria in the interplay between microbial genetic variations and the social behavior of the host.

Figure 4.

Analysis of gene expression profiles of C. elegans fed with ypdF, purH, torY, and ycgJ mutant bacteria and wildtype control strains. (a) GO gene functional enrichment analysis of the genes fed with mutant bacteria of ypdF, purH, torY, and ycgJ. (b) KEGG and Reactome pathways analysis of the genes fed with mutant bacteria of ypdF, purH, torY, and ycgJ. (c) Clustering dendrogram of genes, with dissimilarity measured by topological overlap, accompanied by assigned module colors. (d) Module-trait associations, where each row represents a gene module and each column corresponds to mutant bacteria. The table is color-coded by correlation, according to the color legend. The numerical values in the squares of the table represent the Pearson correlation coefficient (R value), while the numerical values in brackets indicate the p-value for correlation statistical analysis. The green module represents mitochondrial function.

Mitochondria play a role in regulating the social behavior of C. elegans

NPR-1 is a key gene governing social behavior in C. elegans, known to play a significant role in the behavioral differences observed between solitary C. elegans N2 and social C. elegans CB4856¹¹. Furthermore, compared to the wild-type C. elegans N2, the wild isolates of CB4856 exhibit a unique p.A12S amino acid substitution in the COX1 core catalytic subunit, encoded by mitochondrial DNA (mtDNA) of mitochondrial complex IV (CIV). The acquisition of a MAPK-1 binding site, along with the subsequent phosphorylation of the newly introduced serine in the CIV COX1 subunit, has been shown to enhance CIV activity.¹⁷

To eliminate the confounding effects of the nuclear genome, we generated C. elegans strains with swapped nuclear and mitochondrial backgrounds: CN30, with a CB4856 nuclear genome and N2 mitochondria. The social aggregation behavior of CB4856 C. elegans was significantly higher than N2 C. elegans，and CN30 exhibited a significant reduction in social aggregation behavior compared to CB4856 (Figure 5), suggesting mitochondria play a role in social behavior.

Figure 5.

The correlation between mitochondrial genotypes and C. elegans social behaviors. Wild isolates CB4856 shows a unique p.A12S substitution in COX1 (mtDNA), enhancing mitochondrial complex IV activity compared to CN30 C. elegans. (*** p < 0.0001, Student’s t -test), error bars represent SD.

Previous studies indicate that C. elegans behavior can be modulated by a bacterial biofilm in its oral cavity.¹⁸ However, the SEM analysis revealed no significant differences in biofilm formation between ten mutant bacterial strains and wild-type E. coli (Figure S4).

The mutant bacteria can influence the social behavior of C. elegans offspring

To assess the duration of behavioral changes induced by specific mutant bacteria, we fed L1-stage N2 C. elegans with three selected mutant bacterial strains (ΔyghZ, ΔyjgW, and ΔYPDF) until they reached the L4 stage. Afterward, the C. elegans were transferred to an NGM medium containing the control bacterium BW25113. The aggregation rate of N2 C. elegans was observed on Day 1, Day 3, Day 5, and Day 7. The results showed that these three mutant bacteria significantly increased the social behavior of C. elegans only on Day 1. By Day 3, the aggregation behavior returned to pre-treatment levels and remained stable through Days 5 and 7 (Figure 6(a-c)).

Figure 6.

The mutant bacteria influence the social behavior of C. elegans offspring. (a-c) N2 C. elegans fed with ΔypdF, ΔyjgW, and ΔyghZ strains from L1 stage were replaced with control bacteria BW25113 at L4 stage. The social aggregation behavior of N2 was measured on day 1, 3, 5, and 7 after replacement. (d) The F2 generation N2 C. elegans, which have been fed mutant strains for two generations, were examined for social aggregation behavior on plates with wild type E. coli as food. Statistical analysis was performed using independent Student’s t -test) (mut vs. ctrl), (*p < 0.05, * *p < 0.01, ***p < 0.001). Error bars represent SEM.

To investigate the transgenerational persistence of behavioral changes induced by mutant bacteria, we exposed C. elegans to ten distinct E. coli mutants for two consecutive generations. Young adult first generation (F1) offspring derived from these cultures was then transferred to plates seeded with the wild-type control strain BW25113. The social aggregation behavior of the F1 was assessed during adulthood. Our findings indicated that F1 individuals whose parents had been exposed to the ypdF, crL, torY, ycdC, iaaA, and purH mutants exhibited a significantly elevated aggregation rate (Figure 6(d)). This suggests that certain mutant bacterial strains can have a lasting impact on the social behavior of their progeny.

Discussion

Our study highlights the crucial role of microbial genetic diversity in modulating host social behavior. The microbiome encompasses a diverse array of microbial species, each exhibiting substantial genetic variation.^19,20 These microbial organisms frequently undergo genetic changes, and their gene expression is dynamically regulated by the host’s intestinal environment.^21,22 Using C. elegans as a model organism, we systematically demonstrate that the deletion of specific bacterial genes can impact the social behavior of the host. Through this approach, we identified 77 bacterial mutants that enhance the social behavior of C. elegans.

gGut microbes have the ability to influence the metabolism and neural signaling of their host. For instance, the microbiome of Drosophila affects the host’s developmental and metabolic balance through insulin signaling.²³ Gut microbiota interacts with the vagus nerve and various brain areas, including the hypothalamus, mesolimbic system, and prefrontal cortex, all of which are crucial for controlling feeding behavior.²⁴

Our study demonstrates that gut microbes regulate host social behavior through three evolutionarily conserved mechanisms: (1) environmental sensing via the GCY-35/TAX-4 oxygen-responsive pathway, (2) neuronal integration through TGF-β signaling and dopaminergic/TRPV circuits, and (3) bioenergetic regulation mediated by mitochondrial functional modules. These pathways are essential for regulating feeding behavior and social interactions.

In recent years, the role of mitochondria in mediating nervous system diseases has garnered growing attention.^25,26 Our previous studies also demonstrated the pivotal role of mitochondria in mediating the regulation of host lifespan and neural function by gut microbes.^27,28 Lynn Margulis first proposed the origin of mitochondria from bacteria in her work, the Origin of Eukaryotic Cells.²⁹ Franco-Obregon et al. introduced the concept of a bacteria-mitochondrial axis, emphasizing the interplay between bacteria and mitochondria and their collaborative involvement in host metabolism.³⁰ Furthermore, this investigation revealed that both bacteria and mitochondria possess autonomous genetic systems, similar division mechanisms, and genomic characteristics that facilitate information exchange between them.³⁰ Notably, we have discovered for the first time that mitochondrial genetic variation can influence the social behavior of C. elegans. Additionally, several mutant bacteria were found to regulate host social behavior through the mitochondrial pathway. Transcriptome analysis and protein interaction network analysis further support the involvement of the mitochondrial pathway in classical neurobehavioral regulatory pathways. These findings provide a novel perspective for studying the intricacies of social behavior and highlight the importance of microbial–mitochondrial interactions in shaping host behavior.

In this study, we have made the novel discovery that the effects of mutant bacteria on C. elegans social behavior persist for 1 d but return to control levels by the third day. This observation suggests that as the mutant bacteria are gradually replaced by the wild-type BW25113 strain within the C. elegans gut, the behavioral changes revert to baseline levels. Furthermore, our findings demonstrate that these acquired alterations in social behavior can be inherited by subsequent generations after long-term exposure to mutant bacteria. Additionally, it is noteworthy that not all behavioral changes induced by mutant bacteria in the host can be consistently transmitted to offspring, suggesting the existence of diverse regulatory mechanisms, independent of oxygen sensing, through which host behavior is modulated by mutant bacteria.

Summary

Our results offer a novel perspective on the regulatory role of microbial genetic variation in host behavior and suggest a potential involvement of mitochondria in mediating the effects of bacterial mutants on host social behavior. However, it remains unclear whether this phenomenon is highly conserved across species. The precise molecular mechanisms underlying microbial regulation of host behavior via the mitochondrial pathway require further investigation. For instance, whether these mutant bacteria exert their effects via specific metabolites or directly interact with mitochondrial components requires further investigation. Additionally, identifying the exact mitochondrial molecules involved in this regulation will be crucial for future studies.

Methods

Experimental model

In this study, we employed various strains of C. elegans, including N2 (Bristol wild-type), PR678 (tax-4 gene knockout), A×1295 (gcy-35 gene knockout), C×454 (ocr-2 gene knockout), L×702 (dop-2 gene knockout), CB1386 (daf-5 gene knockout), and RB2589 (daf-3 gene knockout), to investigate specific behavioral pathways. These strains were obtained from the Caenorhabditis Genetics Center (CGC). The C. elegans were cultured and subjected to experimental analysis at a temperature of 20°C on standard agar plates containing C. elegans growth medium (NGM). Unless otherwise specified, these plates were inoculated with pre-cultured bacterial strains. To ensure age synchronization, hermaphrodites of C. elegans were isolated from eggs using the method previously described by Sulston and Brenner.³¹

E. coli deletion collection screen for changes in C. elegans aggregation behavior bacterial strain selection and cultivation

The 3,983 E. coli mutants were obtained from the Keio collection library, which contains single-gene deletions of all nonessential genes in E. coli K-12¹⁵. Each mutant strain was individually inoculated into an LB liquid medium supplemented with kanamycin sulfate (25 μmol/L) and incubated overnight for 12–14 h at 37°C. The bacterial density was adjusted to an optical density at 600 nm (OD600) value of approximately 0.6.

Bacterial harvesting and resuspension

Following the cultivation process, the bacterial suspensions were subjected to centrifugation at 6,000 rpm for 5 min to separate the cells from the growth medium. The resulting liquid supernatant was carefully discarded, and the bacterial pellet was subsequently resuspended in 1 mL of LB liquid medium that did not contain any antibiotics. This resuspension aimed to create a homogeneous bacterial suspension suitable for subsequent applications.

Plate preparation

Using a pipette, 150 μL of the reconstituted bacterial culture was dispensed onto the central region of a sterile Petri dish with a diameter of 60 mm, resulting in the formation of a typical circular bacterial lawn. Subsequently, the plates were air-dried within a laminar flow cabinet for a period of 12–24 hours to ensure complete sterility prior to utilization.

Aggregation behavior assay

1. The Worm Synchronization Procedure: To synchronize the wild-type C. elegans strain BW25113 at the L1 larval stage, suitable nutrition was provided. A small amount of M9 buffer solution was added, and the worms were incubated overnight for 10–12 h.

2. Worm Translocation and Cultivation: Approximately 300 synchronized L1 larvae were transferred using a pipette into Petri dishes filled with various mutant bacterial strains. The worms were subsequently cultivated for 48 h, allowing sufficient time for interaction with the bacteria.

3. Worm Transfer to Fresh Dishes: To ensure consistent worm density of 80–120 worms per dish across all replicates, a technique involving agar blocks was employed for individual transfer of worms that had fed on different mutant strains to three new dishes, each containing the same mutant bacterial strain.

4. Further Cultivation and Observation: The C. elegans were cultivated for an additional 10 h until they reached adulthood. To assess their aggregation patterns, images were captured using a standard optical microscope equipped with an industrial-grade camera (Motic, SMZ-171).

5. Screening and Analysis: The initial screening procedure was performed thrice. For mutant strains exhibiting an average aggregation rate surpassing 20% in the initial screening, additional replications (at least six times) were conducted to mitigate the risk of false positives.

Calculation of aggregation rate

1. Aggregation occurred when the body contact area between two worms was 50% or more, or when their bodies intersected.^10,11 Worms were considered non-aggregated if they had no body contact or a contact area below 50%. The Aggregation Rate was calculated as follows: M=$\sum _{\mathit{i}=1}^n\mathit{ai}(\frac{\mathit{Nij}}{\mathit{Nim}})/\mathit{n}$. The number of aggregated C. elegans on the i-th plate was denoted as N_ij, while the total C. elegans count on the same plate was represented as N_im.

2. Edge aggregation described a scenario where over half of a C. elegans’s body was positioned at the edge of the bacterial lawn. The edge aggregation rate indicated the proportion of C. elegans at the edge relative to the total number. For each plate, both the total number of worms and those aggregated at the edge were counted. The Edge Aggregation Rate was determined by the formula: M=$\sum _{\mathit{i}=1}^n\mathit{ai}(\frac{\mathit{Nij}}{\mathit{Nim}})/\mathit{n}$. The count of C. elegans at the edge of bacteria on the i-th plate was denoted as N_ij, and the total C. elegans count on that plate was denoted as N_im.

RNA-seq data analysis

In this study, C. elegans samples were subjected to sequencing on the DNBSEQ platform after being fed with E. coli mutants (ypdF, yghZ, crl, torY, yjgW, yccU, ycdC, iaaA, purH, and ycgJ). Three control samples of BW25113 were also included in the analysis. On average, each sample generated approximately 1.19 GB of data (BioProject ID: F21FTSECWLJ1283_NEMyjzwN). RNA-seq data were analyzed previously.³² The raw sequencing reads underwent a series of filtering steps to eliminate rRNA sequences, low-quality reads, and adapter contamination. Reads containing a high proportion of unknown bases were also discarded. The cleaned reads were then aligned to the reference genome (NCBI: GCF_000002985.6_WBcel235) using HISAT and assembled using StringTie software. Additionally, Bowtie 2 software was employed to align against the reference sequence for comparative analysis.

Gene expression measurement

Trizol (Invitrogen, Carlsbad, USA) was used to extract total RNA from synchronized worms according to the manufacturer’s instructions. The quality and quantity of RNA were assessed using an Agilent 2100 and NanoDrop bioanalyzer (Thermo Fisher Scientific, USA). The gcy-35, tax-4, daf-3, daf-5, ocr-2, and dop-2 gene expressions were determined by qRT-PCR. The cDNA was synthesized using the ABScript II cDNA First-Strand Synthesis Kit (RK20400) reverse transcription kit. Subsequently, real-time fluorescence quantitative PCR was conducted on the target genes employing the SYBR Green method. The expression levels of the target genes were assessed. The PCR primers are as follows:

ACT-1-F: ACGACGAGTCCGGCCCATCC

ACT-1-R: GAAAGCTGGTGGTGACGATGGTT

DAF −3-F: GGCGGATTCATTTGCTCTG

DAF −3-R: CGGGTACTTCATGCGGTTT

DAF −5-F: GGGGAGATTCTTGTGGATTT

DAF-5-R: GCGTGACGACTTTATGTGTG

OCR-2-F: GAAAGCTGGTGGTGACGATGGTT

OCR-2-R: GTTCAAATACCATTTATCACAGGGA

GCY-35-F: ACTCGTTTCAAACTTACACCC

GCY-35-R: GCCGAACATATTCTACTCTCC

TAX-4-F: GTGCATACGACTACGGCTC

TAX-4-R: CTCTTTTGGGGGCGGTACTT

DOP-2F: ACCTCCAAAGGAATACCGACG

DOP-2 R: GTAGTTCGCGCTATTTGCCG

Scanning electron microscopy (SEM) analysis

1. The adult C. elegans were washed from petri dishes fed with different mutant strains using 3 ml of M9 buffer into 15 ml aseptic centrifuge tubes. They were then washed three times, and the supernatant was discarded after natural precipitation of C. elegans. Next, they were immersed in a 2.5% glutaraldehyde solution overnight for 12 h.

2. The glutaraldehyde solution was discarded, and the samples were cleaned three times with 1 ml of PBS buffer solution. Then, they were fixed with a 1% osmic acid solution for 2 h.

3. The samples were dehydrated using ethanol solutions at concentrations of 50%, 70%, 90%, and finally,100% in increments of 10 min each time.

4. The samples underwent critical point drying by placing them into a vacuum plating machine to dry for 4 h.

5. For sample mounting, dried C. elegans were individually attached to conductive glue using a wormpicker coated with nail polish.

6. Finally, the samples were observed using an emission scanning electron microscope.

E. coli growth detection

The E. coli were cultured in 200 μl of LB medium supplemented with kanamycin. The initial culture density was adjusted to OD600 = 0.1, and the bacteria were incubated at 37°C for 24 h under static conditions (without shaking). After incubation, the cultures were vortexed thoroughly to ensure homogeneity, and the absorbance was measured at 600 nm using a spectrophotometer.

Statistical analysis

DEGs were obtained by comparing the gene expression level in C. elegans fed with BW25113 and ten mutant E. coli bacteria separately using DESeq2. Differentially expressed genes analysis was performed with the R package edgeR. Fold change (cutoff, 2) and p value (cutoff, 0.05). To clarify the cellular localization of these DEGs, we used the org.Ce.eg.db R package for annotation and statistics. The Weighted Gene Co-expression Network Analysis (WGCNA) package³³ was employed to screen co-expressed gene networks significantly associated with ten mutant E. coli bacteria. Enrichment analysis for metabolites was carried out via the MetaboAnalyst platform (https://www.metaboanalyst.ca/), with visualizations generated using the R language. The PPI network was performed on the STRING database with an interaction score >0.4. The selected target genes and PPI network were subjected to gene function evaluation with the GO database and pathway enrichment analysis with the KEGG pathway database.

Funding Statement

This work was supported by the Natural Science Foundation of Jiangsu Province [BK20241953], Xuzhou Science and Technology Innovation Project [KC23009], Xuzhou Key Research and Development Program [KC22096], National Natural Science Foundation of China [81701390], Innovative Group Cultivation Project for Basic Medicine [CX25XT03] and Laboratory for Clinical Medicine, Capital Medical University.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

Z.Z., X.W., and Y.L. designed the experiments, conducted data analysis, and drafted the manuscript. R.B., Y.Z., P.S., T.W., R.B., Y.Z., X.D., H.F., and Y.L. performed the majority of the experiments. D.Z., J.Z., T.Z., X.Z., D.Z., J.Z., Y.L., X.W., and Z.Z. discussed the data and its interpretation. D.Z. and J.Z. carried out transcriptomics detection as well as data analysis. Y.L. provided technical support throughout the study. The final manuscript was reviewed and approved by all authors.

Data availability statement

RNA-seq data have been deposited at GEO at GEO: GSE283454.

Highlights

• The microbial-mitochondrial pathway emerges as a novel mechanism through which E. coli modulates C. elegans social behavior.

• High-throughput screening identifies E. coli mutants that regulate the social behavior of C. elegans.

• Parental exposure to specific E. coli mutants influences social behaviors in offspring.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19490976.2025.2490828