A transcription factor regulates development of a bacteria-housing eukaryotic cell through DNA damage mitigation

Significance

Intracellular microorganisms are often confined to specialized host cells called bacteriocytes in invertebrates. Transcription factors play key roles in regulating bacteriocyte development, but underlying molecular mechanisms remain unclear. Our previous work demonstrates that the transcription factor Adf-1 suppresses cell death of whitefly bacteriocytes. Here, we unravel that Adf-1 induces the expression of the ataxia telangiectasia and Rad3-related (ATR) gene in whitefly bacteriocytes. ATR negatively regulates DNA damage levels, apoptosis and autophagy of bacteriocytes, but positively regulates division and numbers of bacteriocytes. Thus, we identify a previously uncharacterized Adf-1-ATR regulatory pathway that safeguards whitefly bacteriocyte development and symbiosis by mitigating DNA damage. It highlights that DNA damage repair is indispensable for the functioning of evolutionarily novel cells like bacteriocytes.

Abstract

Intracellular microorganisms are widespread and housed within specialized host cells called bacteriocytes in many invertebrates. Although increasing attention has been devoted to understanding bacteriocyte development, differentiation, and proliferation, the molecular mechanisms that maintain these cells remain poorly understood. We previously showed that the transcription factor Adf-1 suppresses bacteriocyte cell death in whiteflies. Here, DNA affinity purification sequencing identified Adf-1 target genes involved in diverse cellular processes and revealed the DNA damage response kinase ataxia telangiectasia and Rad3-related (ATR) as a direct transcriptional target. Dual-luciferase reporter assays and RNA interference confirmed that Adf-1 binds the ATR promoter and positively regulates its expression in bacteriocytes, where ATR is highly expressed. Consistent with this regulatory relationship, silencing either Adf-1 or ATR increased DNA damage in bacteriocytes. ATR knockdown further promoted apoptosis and autophagy, reduced the proportion of dividing bacteriocytes, decreased bacteriocyte abundance, and lowered symbiont titers in whiteflies. In addition, yeast two-hybrid assays identified an interaction between Adf-1 and the ATP-dependent DNA helicase RecQ1, although RecQ1 silencing had no detectable effect on bacteriocyte abundance. Together, our findings demonstrate that Adf-1 preserves bacteriocyte homeostasis by promoting DNA damage repair through transcriptional activation of ATR. Given that bacteriocytes are exposed to persistent genotoxic stress associated with polyploidy and high metabolic activity, this work identifies DNA damage control as a key mechanism underlying bacteriocyte maintenance and symbiosis stability, and suggests that this pathway may represent a conserved feature of bacteriocyte biology.

Microbial symbionts play a pivotal role in the evolution of animals by forming with their hosts intimate associations that influence their physiology, development, and ecological success (1, 2). In the case of endosymbiotic partnerships, many invertebrates have evolved specialized structures, known as symbiotic organs, that are adapted to preserve and control microbial symbionts (3, 4). Symbiotic organs display a huge diversity of shapes, positions in the invertebrate body and structures. They include the light organ of squids (5), the midgut ceca of heteropteran insects (6) and the trophosomes and bacteriomes in many invertebrates, including insects, bivalves, sea slugs, sponges, leeches, shipworms, zombie worms, tubeworms, flatworms, snails, and filarial nematodes (7–15). Trophosomes and bacteriomes are composed of host cells, the bacteriocytes. Previous studies have demonstrated that homobox-containing transcription factors, including Ultrabithorax (Ubx), Abdominal-A (abd-A), Antennapedia (Antp), Distal-less (Dll), and Engrailed (En), are linked to the formation and development of bacteriocytes in aphids, seed bugs, and ants (16–18). It has also been demonstrated that the transcription factor Adf-1 impacts the sexual differentiation of development in whitefly bacteriocytes (19).

Transcription factors are primarily located to the cell nucleus where they modulate gene expression by recognizing and binding to specific promoter sequences of target genes (20, 21). Moreover, various other proteins (other transcription factors or cofactors) can facilitate or decrease the DNA-binding affinity of transcription factors, or influence translation, degradation, and subcellular localization of transcription factors via posttranslational modifications (22–24). Studies on plant systems have also shown that transcription factors play important roles in orchestrating plant-microbe symbioses. For instance, the rice phosphate starvation response transcription factor OsPHR2 is essential for mycorrhizal symbiosis, as it upregulates the expression of mycorrhiza-related genes under low-phosphate conditions, while SPX proteins (named after the Saccharomyces cerevisiae SYG1 and Pho81 proteins and the mammalian Xpr1) interact with OsPHR2 under high-phosphate conditions to repress symbiosis-related gene expression and inhibit mycorrhizal colonization, thereby dynamically fine-tuning the rice-arbuscular mycorrhizal fungal mutualism in response to phosphate availability (25, 26). Similarly, the legume transcription factor Nodule Inception governs rhizobial infection and nodule development by regulating targeted gene expression and through interactions with NLP1 protein (27–31). Despite these advances in plant-microbe symbioses, the mechanisms by which transcription factors orchestrate bacteriocyte development and differentiation in animals remain poorly delineated.

The whitefly Bemisia tabaci is a complex of more than 40 cryptic species, among which B. tabaci MEAM1 is a globally important agricultural pest (32–34). B. tabaci MEAM1 harbors symbionts “Candidatus Portiera aleyrodidarum” (hereafter Portiera) and “Candidatus Hamiltonella defensa” (hereafter Hamiltonella), which have the specificity to be localized within the same bacteriocyte (19, 35). Bacteriocytes are maternally inherited in B. tabaci (36, 37). Our previous work revealed that whitefly bacteriocytes proliferate in adult females but degenerate in adult males, and showed that the transcription factor Adf-1 is involved in bacteriocyte development by negatively regulating apoptosis and autophagy (19). However, the molecular and cellular mechanisms by which Adf-1 regulates whitefly bacteriocyte development are unclear. In this study, we employed yeast two-hybrid (Y2H) library screening and DNA affinity purification sequencing (DAP-seq) to identify interacting proteins and target genes of Adf-1, respectively. We found that, although one Adf-1-interacting protein does not impact bacteriocyte development, Adf-1 directly targets the ATR gene to maintain genomic integrity and support bacteriocyte development.

Results

RecQ1, A Protein Interacting with Transcription Factor Adf-1, Exerts Little Impact on Bacteriocyte Development.

The whitefly transcription factor Adf-1 contains both a MADF domain and a BESS domain, which is responsible for DNA binding and protein–protein interactions, respectively (19, 38–40). In Drosophila, the transcription factors Adf-1 and Dip3, both containing the MADF domain and BESS domain, can form homo- and heterodimers through the BESS domain to drive transcriptional activation (38, 41). To determine the mechanisms of action of B. tabaci Adf-1, we conducted yeast two-hybrid (Y2H) assays to assess whether Adf-1 employs a dimerization-dependent mechanism like that of its Drosophila homologs, and to identify its interaction partners. Y2H assays confirmed that Adf-1 exhibits neither autoactivation activity nor toxicity in yeast cells and is therefore suitable for use as a bait protein in subsequent Y2H screening (SI Appendix, Fig. S1). We cotransformed Y2HGold yeast cells with pGBKT7-Adf-1 (BD-Adf-1; full-length Adf-1 bait) paired with three prey constructs: full-length Adf-1 (AD-FAdf-1), the C-terminal region containing the BESS domain (AD-CAdf-1), and the N-terminal region containing the MADF domain (AD-NAdf-1). All yeast strains exhibited robust growth across serial dilutions on DDO (SD/-Trp/-Leu) plates, confirming successful cotransformation with both plasmids. Yeast cells coexpressing BD-Adf-1 with either AD-FAdf-1 or AD-CAdf-1 produced blue colonies at all dilutions on QDO/X-α-gal (SD/-Trp/-Leu/-His/-Ade/X-α-gal) plates. In contrast, yeast cells coexpressing BD-Adf-1 and AD-NAdf-1 showed no growth on QDO/X-α-gal plates. These results indicate that Adf-1 can form homodimers via its C-terminal BESS domain (Fig. 1A). We next performed Y2H library screening using Adf-1 as bait against the B. tabaci cDNA library generated from female whiteflies within 7 d after emergence. This screen identified 24 potential interacting proteins, among which only five contained in-frame sequences suitable for further validation (SI Appendix, Table S1). We then performed pairwise Y2H assays to validate these five candidate interactors. Y2HGold yeast cells were cotransformed with bait construct pGBKT7-Adf-1 and individual pGADT7 plasmids carrying the full-length sequence of each candidate protein. All yeast strains grew normally on DDO medium, confirming successful transformation. No growth was observed on TDO or QDO/X-α-gal plates for yeast cells coexpressing Adf-1 with Bta12552, Bta03301, or Bta00763. Yeast cells coexpressing Adf-1 with Bta13763 grew on TDO medium but showed no growth on QDO/X-α-gal plates. In contrast, yeast cells coexpressing Adf-1 and Bta01308 produced abundant colonies on both TDO and QDO/X-α-gal plates. Collectively, these pairwise assays revealed a weak interaction between Adf-1 and Bta13763 and a strong interaction between Adf-1 and Bta01308, namely the ATP-dependent DNA helicase Q1 (RecQ1), whereas the other three candidate proteins showed no detectable interaction with Adf-1 (Fig. 1B).

Fig. 1.

Identification of the protein interacting with Adf-1 and effect of RecQ1 silencing on bacteriocyte development. (A) Yeast two-hybrid (Y2H) assays to test Adf-1 homodimerization. BD represents the pGBKT7 bait vector, and AD represents the pGADT7 prey vector. FAdf-1, CAdf-1, and NAdf-1 correspond to full-length Adf-1 (1 to 912 bp), C-terminus containing the BESS domain (457 to 912 bp), and the N terminus containing MADF domain (1 to 456 bp), respectively. Y2HGold yeast cells were cotransformed with pGBKT7-Adf-1 (BD-Adf-1) combined with three prey constructs: AD-FAdf-1, AD-CAdf-1, and AD-NAdf-1. Transformed yeast suspensions were serially diluted from 10⁰ to 10⁻³ and spotted onto two selective media: DDO (SD/-Trp/-Leu) to confirm dual-plasmid transformation, and QDO/X-α-gal (SD/-Trp/-Leu/-His/-Ade/X-α-gal) to detect physical protein binding via blue colony formation. BD-53+AD-T served as the positive interaction control, and BD-Lam+AD-T as the negative noninteraction control. (B) Pairwise Y2H assays validating interactions between Adf-1 and candidate proteins. Y2HGold yeast cells were cotransformed with BD-Adf-1 bait and individual pGADT7 plasmids carrying the full-length sequence of each candidate protein (AD-Bta12552, AD-Bta03301, AD-Bta00763, AD-Bta13763, AD-Bta01308). Yeast suspensions were serially diluted from 10⁰ to 10⁻³ and spotted onto three selective media: DDO for cotransformation verification, TDO (SD/-Trp/-Leu/-His) for weak interaction screening, QDO/X-α-gal for detecting strong protein interactions via blue colony formation. (C) Expression of RecQ1 in adult females 1.5, 3, and 6 d after microinjection with dsRecQ1. n = 4 biological replicates. β-Actin was used as the reference gene. (D and E) Effect of RecQ1 silencing on the total number of bacteriocytes in adult females 3 (D) and 6 d (E) after dsRNA microinjections. The total number of bacteriocytes was calculated as the sum of the bacteriocytes present in whiteflies and the number of eggs laid by these whiteflies during 6 d following gene silencing, with n = 23 and 21 biological replicates, respectively. Data shown are mean ± SEM. Significant differences between treatments are indicated by asterisks in C (**P < 0.01, ***P < 0.001).

RecQ1 is involved in DNA damage repair and the maintenance of genomic stability in Homo sapiens. It is highly expressed in the nuclei of many human cancer cells, and silencing RecQ1 coding gene increases DNA damage levels and inhibits cancer cell proliferation (42–45). To investigate whether RecQ1 contributes to bacteriocyte development in B. tabaci, the female whitefly RecQ1 gene was silenced by microinjection with dsRecQ1. The expression of RecQ1 mRNAs in adult females was significantly reduced 1.5, 3, and 6 d after RNA Interference (RNAi) treatment (Fig. 1C, 1.5 d, F = 10.11, df = 3, P = 0.002; 3 d, F = 0.50, df = 6, P < 0.001; 6 d, F = 11.01, df = 5, P < 0.001). However, neither the number of bacteriocytes in adult females nor the total number of bacteriocytes was altered 3 or 6 d after RecQ1 silencing (Fig. 1D, F = 2.00, df = 44, P = 0.112; Fig. 1E, P = 0.314; SI Appendix, Fig. S2A, P = 0.912; SI Appendix, Fig. S2C, F= 0.02, df = 40, P = 0.865). The number of eggs laid by these females was not significantly altered at day 3, while it was significantly reduced at day 6 after RecQ1 silencing (SI Appendix, Fig. S2B, F= 1.39, df = 44, P = 0.109; SI Appendix, Fig. S2D, P = 0.037). These results demonstrate that, despite its strong interaction with Adf-1, RecQ1 is not required for bacteriocyte development in B. tabaci.

Adf-1 Targets Multiple Genes Involved in Cellular Processes Identified by DAP-Seq.

Transcription factors orchestrate mRNA transcription through sequence-specific binding to the promoter regions of target genes (20, 21). To elucidate the molecular mechanisms by which Adf-1 regulates bacteriocyte development in whiteflies, we performed DAP-seq experiments to identify genes directly targeted by Adf-1, following the previous protocol (46). This allowed us to obtain 47.4, 64.0, and 46.2 million clean reads, with genome mapping rates being 62.85%, 63.60%, and 89.14% for the two independent biological replicates and the negative control, respectively (SI Appendix, Table S2). Analyses of the sequencing data identified 12,710 putative Adf-1-binding sites in vitro (Fig. 2A and Dataset S1A). Among all detected binding peaks, 5.2% (664 peaks) were mapped to gene promoter regions, defined as the 2-kilobases (kb) upstream of the transcriptional start site (TSS) (Fig. 2B and Dataset S1B), with the consensus DNA binding motif of Adf-1 “AAAARAADAAAAAAWAAAAAA” (R represents A or G; D represents not C; W represents A or T) (Fig. 2C and Dataset S1C). Because genes with binding peaks localized in their promoter regions are generally considered putative transcription factor targets, we annotated a total of 650 target genes, each containing at least one Adf-1-binding site within its promoter region (Dataset S1B). KEGG enrichment analysis showed that 19 and 15 genes of these putative target genes were involved in the cell division and cell death processes, respectively (Fig. 2D and Dataset S1D). These results indicate that Adf-1 can target a diverse set of genes involved in key cellular processes, providing a molecular framework for further studying its role in the regulation of bacteriocyte development.

Fig. 2.

Genome-wide mapping of Adf-1 binding sites and motifs in B. tabaci using DNA affinity purification sequencing (DAP-seq). (A) Venn diagram showing the number of Adf-1-binding peaks identified from DAP-seq datasets. (B) Genomic distribution of Adf-1-binding peaks across the B. tabaci genome. (C) Enriched DNA-binding motif of Adf-1. (D) KEGG pathway analysis of the putative Adf-1 target genes involved in cellular processes.

Adf-1 Reduces DNA Damage Level of Bacteriocytes by Specifically Promoting ATR Gene Expression in Whitefly Bacteriocytes.

Among the putative Adf-1 target genes involved in cell division, the peaks associated with the ataxia telangiectasia and Rad3-related (ATR) gene exhibited the highest enrichment and statistical significance (Fig. 3A and Dataset S1D). ATR is a key regulator of DNA damage repair and genome integrity during DNA replication (47, 48). This suggests that ATR may support bacteriocyte development of whiteflies by mediating DNA damage repair. To determine whether Adf-1 directly modulates ATR transcription, we employed a dual-luciferase (LUC/REN) reporter assay. The results confirmed that the ATR promoter possesses basal transcriptional activity and that Adf-1 binding to the ATR promoter enhances the reporter gene expression (Fig. 3B, P_{pATR vs. CK or Adf-1} < 0.001, P_{pATR+Adf-1 vs. pATR} < 0.001), demonstrating that Adf-1 has the potential to exert a direct transcriptional regulatory control of the ATR gene. Consistent with the expression pattern of Adf-1 mRNA (19), ATR gene is highly expressed in female bacteriocytes compared to the female whole body of whiteflies (Fig. 3C, F = 7.71, df = 3, P < 0.001; SI Appendix, Fig. S3A, F = 0.55, df = 6, P < 0.001). Last, following Adf-1 silencing, ATR expression remained unchanged in the female whole body of whiteflies (Fig. 3D, Adf-1, F = 5.80, df = 6, P < 0.001; ATR, F = 1.24, df = 6, P = 0.136), but was significantly reduced in female bacteriocytes (Fig. 3E, Adf-1, F = 6.94, df = 4, P = 0.001; ATR, F = 2.16, df = 4, P = 0.002; SI Appendix, Fig. S3B, Adf-1, F = 0.54, df = 4, P < 0.001; ATR, F = 1.79, df = 4, P < 0.001). Together, these results demonstrate that Adf-1 binds to the ATR promoter and has the capability to specifically promote ATR expression in female whitefly bacteriocytes.

Fig. 3.

Adf-1 impacts bacteriocyte DNA damage by regulating the expression of the ATR gene. (A) Adf-1-binding peaks at the ATR locus visualized via Integrative Genomics Viewer (IGV) software. Track height was set to 100, with peak height ranging from 0 to 140. Adf-1_1 and Adf-1_2 represent two replicates, and Adf-1_input represents the negative control. Binding peaks associated with the ATR gene were mapped to the genomic regions Scaffold754: 169,718 to 170,596 and Scaffold754: 170,106 to 170,676, respectively. The distance of the merged peaks to the transcription start site (TSS) of the ATR gene is −1613 bp. The minus sign (“−”) indicates a position upstream of the TSS. (B) Dual-luciferase (LUC/REN) reporter assay assessing the regulatory effect of Adf-1 on the ATR promoter. Control (CK): cotransfection of pGL3-Basic, pcDNA3.1, and pRL-TK; Adf-1: cotransfection of pGL3-Basic, pcDNA3.1-Adf-1, and pRL-TK; pATR: cotransfection of pGL3-pATR (−958 to +1), pcDNA3.1, and pRL-TK (to verify ATR promoter activity); pATR+Adf-1: cotransfection of pGL3-pATR (−958 to +1), pcDNA3.1-Adf-1, and pRL-TK (to assess Adf-1’s regulatory effect on the ATR promoter). Fluc/Rluc represents the ratio of firefly to renilla luciferase activity. n = 9 biological replicates. (C) Relative expression level of ATR in bacteriocytes compared with that in the whole body of adult female whiteflies. RPL7 was used as reference genes. n = 4 biological replicates. (D) Expression of Adf-1 and ATR in the whole body of adult females 6 d following microinjection with dsAdf-1. β-Actin was used as the reference gene. n = 4 biological replicates. (E) Expression of Adf-1 and ATR in the bacteriocytes of female whiteflies 6 d following microinjection with dsAdf-1. RPL7 was used as reference genes. n = 3 biological replicates. (F) Expression of ATR in female whiteflies 3 and 5 d following microinjection with dsATR. β-Actin was used as the reference gene. n = 4 biological replicates for whiteflies 3 and 5 d, respectively. (G) Localization of γH2AX (red) in the bacteriocytes of dsGFP- and dsATR-injected female whiteflies 5 d following microinjection. n = 21 biological replicates. (H) Localization of γH2AX (red) in the bacteriocytes of dsGFP- and dsAdf-1-injected female whiteflies 6 d following microinjection. n = 18 biological replicates. (I) Number of γH2AX foci in bacteriocytes of dsGFP- and dsATR-injected female whiteflies 5 d following microinjection, and in bacteriocytes of dsGFP- and dsAdf-1-injected female whiteflies 6 d following microinjection. n = 21 and 18 biological replicates, respectively. The data shown are mean ± SEM. Different letters above the bars indicate significant differences between treatments at P < 0.05 in B. Significant differences between treatments are indicated by asterisks in C–F and I (**P < 0.01, ***P < 0.001).

To investigate the role of ATR in DNA damage repair in whitefly bacteriocytes, we silenced the ATR gene of female whiteflies by microinjecting the corresponding dsRNAs. This resulted in a significant reduction of ATR expression in adult females at 3 and 5 d (Fig. 3F, 3 d, F = 0.68, df = 6, P < 0.001; 5 d, F = 0.24, df = 6, P < 0.001). DNA damage was then assessed by immunofluorescence staining to detect γH2AX, a well-established biomarker of DNA damage whose abundance positively correlates with DNA damage levels (49). γH2AX signals localized in the nuclei of female bacteriocytes, and the number of γH2AX foci in female bacteriocytes was significantly increased 5 d after ATR silencing (Fig. 3 G and I, dsGFP and dsATR, F = 20.63, df = 22, P < 0.001), which indicates that knock-down of this gene induced DNA damage accumulation in female whitefly bacteriocytes. We next evaluated whether Adf-1 silencing elicited a similar DNA damage phenotype. The number of γH2AX foci was significantly elevated in female whitefly bacteriocytes 6 d after Adf-1 silencing (Fig. 3 H and I, dsGFP and dsAdf-1, F = 20.30, df = 19, P < 0.001). The absence of signal in negative controls without antibody incubation confirmed the specificity of the γH2AX antibodies (SI Appendix, Fig. S4). These results indicate that transcription factor Adf-1 reduces the DNA damage level of bacteriocytes by specifically promoting ATR gene expression in female whitefly bacteriocytes.

ATR Silencing Induces Apoptosis and Autophagy in Bacteriocytes and Reduces the Bacteriocyte Numbers and Symbiont Titers in Whiteflies.

Given that our previous work demonstrated Adf-1 silencing induced apoptosis and autophagy in female whitefly bacteriocytes (19), we wanted to investigate whether ATR silencing elicits similar cellular phenotypes. Here we demonstrated that, 5 d after silencing ATR of female whiteflies, mRNA expression levels of Caspase 3, an apoptotic marker, and Atg8, an autophagy marker, were significantly elevated, and the corresponding proteins more abundant in whitefly bacteriocytes (Fig. 4A, Caspase 3, F = 1.99, df = 6, P = 0.012; Atg8, F = 4.61, df = 6, P = 0.001; Fig. 4 B and C and SI Appendix, Fig. S5A, Caspase 3, F = 0.03, df = 6, P = 0.002; Atg8, F = 0.01, df = 6, P < 0.001; SI Appendix, Fig. S5B, Caspase 3, F = 1.00, df = 46, P < 0.001; Atg8, F = 6.52, df = 33, P < 0.001). ATR silencing also impacted important physiological parameters relative to the B. tabaci symbiotic status 5 d after the gene knock-down: i) the proportion of bacteriocytes undergoing cell division decreased by 43.8%, ii) the number of bacteriocytes in the body of adult females decreased by 17.3%, iii) the number of eggs laid by these females decreased by 15.3%, and iv) the total number of bacteriocytes decreased by 16.6% (Fig. 4D, F = 22.82, df = 3, P = 0.015; Fig. 4E, F = 0.66, df = 42, P < 0.001; SI Appendix, Fig. S5C, F = 0.53, df = 42, P = 0.007; SI Appendix, Fig. S5D, F = 0.94, df = 42, P = 0.016; SI Appendix, Fig. S6). The reduction in the number of bacteriocytes corresponded to a decrease in the titers of the symbionts Portiera and Hamiltonella, as demonstrated by qPCR analysis (Fig. 4F, Portiera, F = 5.33, df = 28, P < 0.001; Hamiltonella, F = 15.43, df = 24, P = 0.014). These results indicate that ATR can affect bacteriocyte division and development by suppressing apoptosis and autophagy in bacteriocytes of female whiteflies.

Fig. 4.

Effect of ATR silencing on bacteriocyte cell death, division, numbers, and symbiont titers in whiteflies. (A) Expression of Caspase 3 and Atg8 in the bacteriocytes of dsGFP- and dsATR-injected adult females 5 d following microinjection. RPL7 was used as the reference genes. n = 4 biological replicates. (B and C) Localization of Caspase 3 (green) (B) and Atg8 (green) (C) in the bacteriocytes of dsGFP-injected and dsATR-injected female whiteflies 5 d following microinjection. n = 24 biological replicates. (D) The proportion of bacteriocytes undergoing division in dsGFP- and dsATR-injected female whiteflies 5 d following microinjection. n = 4 biological replicates. (E) Effect of ATR silencing on the total number of bacteriocytes in female whiteflies 5 d following microinjection. n = 22 biological replicates. (F) Effect of ATR silencing on symbiont titers in female whiteflies 5 d after following microinjection. n = 19 biological replicates. β-Actin was used as the reference gene. Data are presented as mean ± SEM. Significant differences between treatments are indicated by asterisks in A and D–F (*P < 0.05; **P < 0.01; ***P < 0.001).

Discussion

Symbiosis with microbes drives the ecological and evolutionary diversification of eukaryotic organisms (1, 2). In many invertebrates, symbiotic partners are housed in evolutionarily novel cells, the bacteriocytes. Despite the importance of these cells, the molecular mechanisms governing the bacteriocyte development remain poorly resolved (14). This question is particularly important in insects having evolved mutualistic interactions with symbiotic bacteria necessary for their development and reproduction. In this study we focus on the Adf-1 transcription factor, that we previously showed as being an important regulator of bacteriocyte cell death in whiteflies (19). We show that Adf-1 specifically activates ATR expression in bacteriocytes to mitigate DNA damage, thereby ensuring genomic integrity and normal bacteriocyte proliferation (Fig. 5). These findings provide insights into how transcription factors regulate bacteriocyte development and highlight that proteins essential for DNA damage repair, conserved across evolution, are also important for the functioning of evolutionarily novel cells like bacteriocytes.

Fig. 5.

Schematic model of transcription factor Adf-1 mediated regulation of bacteriocyte development in the whitefly B. tabaci. Adf-1 is highly expressed in bacteriocytes of female whiteflies, where it modulates bacteriocyte development by suppressing apoptosis and autophagy. Adf-1 contains a MADF domain, which mediates binding to DNA transcription factor binding sites (TFBS), and a BESS domain, which facilitates protein–protein interactions. Our results indicate that Adf-1 regulates bacteriocyte development by directly targeting the ATR gene, ultimately suppressing apoptosis and autophagy in these cells.

Concerning its mechanisms of action, we showed here that, consistent with its homologs in Drosophila (38, 41), whitefly Adf-1 forms homodimers via its BESS domain, suggesting evolutionary conservation of this structural feature. Y2H screening identified five proteins potentially capable of interacting with Adf-1, with only RecQ1 showing robust interaction in pairwise validation assays. RecQ1 is known to mediate DNA damage repair and promote cancer cell proliferation in humans (43–45). However, RecQ1 silencing had no significant impact on the number of whitefly bacteriocytes. This discrepancy may stem from two nonmutually exclusive mechanisms. First, partial RecQ1 silencing may leave sufficient protein to maintain its function, even if the transcript expression levels were substantially reduced. Second, the genome of B. tabaci contains five ATP-dependent DNA helicase Q1 genes, and functional redundancy within this gene family may compensate for RecQ1 depletion. Moreover, the limited number of Adf-1 interactors identified in the present study may reflect technical constraints of the Y2H assay. Complementary approaches, such as coimmunoprecipitation coupled with mass spectrometry (Co-IP/MS) would enable identification of a more comprehensive repertoire of Adf-1-interacting proteins.

Bacteriocytes are polyploid, and undergo division and differentiation during development of aphids (50), weevils (51), ants (52), and whiteflies (19, 53). In aphids, weevils, and male whiteflies, bacteriocytes stop their division at the adult stage. On the contrary, in female whiteflies, bacteriocytes continue to undergo division at the adult stage (19, 53). In general, polyploidization combined with sustained division and differentiation processes can increase cell stress and induction of genomic DNA damage (54, 55). Moreover, intense metabolic activity and flux can cause the production of reactive oxygen species further contributing to elevated DNA damage accumulation (56). Consistently with this activity and dynamics of the bacteriocyte cell, our DAP-seq analysis identified 650 putative Adf-1 target genes including those involved in cell cycle progression and cell death, indicating that Adf-1 may coordinate a broad transcriptional network. Notably, Adf-1 binding peaks at the ATR were mostly enriched at the promoter of the ATR gene. As a core kinase in the DNA damage response pathway, ATR is activated by replication stress, free radicals, hypoxia, ultraviolet light, radiation, and other genomic insults, triggering cell cycle checkpoint signals and facilitating DNA repair (57, 58). Our results demonstrate that Adf-1 positively regulates ATR expression to mitigate DNA damage. Adf-1 silencing specifically reduced ATR expression in bacteriocytes but not in whole bodies. This suggests that bacteriocytes may express a unique combination of transcription factors and cofactors that cooperate with Adf-1 to drive cell type–specific regulation of ATR transcription. This finding points to a specialized regulatory program tailored to bacteriocyte biology. Consistently, silencing either Adf-1 or ATR increased γH2AX foci in bacteriocytes, supporting a conserved role for ATR in maintaining genomic stability and identifying Adf-1 as an upstream regulator of this pathway in whiteflies. Unlike most somatic cells, which are ultimately discarded at the end of an organism’s lifespan, bacteriocytes are vertically transmitted in B. tabaci, and must preserve functional integrity to sustain symbiosis across generations (36, 37). This physiological and evolutionary constraint may explain the elevated expression of Adf-1 and ATR in bacteriocytes relative to the whole body of female whiteflies, serving as an adaptive mechanism to control replication-associated DNA damage in these specialized cells. To our knowledge, such a tissue-specific DNA repair regulatory pattern within insect symbiotic organs has not been previously reported.

ATR silencing not only increased DNA damage but also triggered bacteriocyte apoptosis and autophagy and resulted in reduced bacteriocyte number. These findings are consistent with established links between unresolved DNA damage and programmed cell death (59–61). Bacteriocytes have evolved independently across multiple insect orders, arising from distinct cell lineages (17, 62). Bacteriocytes degenerate in old parthenogenetic aphids (50, 63, 64), in both female and male adult weevils 7 d after molting (51), and in male adult whiteflies soon after emergence (14, 19). These processes are activated in response to host physiology through the activation of complex cell death mechanisms that may be facilitated by accumulation of extensive DNA damage. Notably, a homolog of Adf-1 containing well conserved MADF and BESS domains has been recently annotated in the genome pea aphid, Acyrthosiphon pisum, a symbiotic insect containing bacteriocytes and belonging to the order Hemiptera, like B. tabaci (65). This raises a key evolutionary question as to whether other insects than whiteflies employ analogous DNA damage repair pathways to safeguard bacteriocyte genomic integrity during development, differentiation, or aging. Comparative studies across diverse insect symbioses will help determine whether Adf-1-ATR regulatory network is evolutionarily conserved or lineage-specific, shedding light on mechanisms underlying bacteriocyte evolution.

Silencing the ATR gene significantly reduced the titers of Portiera and Hamiltonella in bacteriocytes and decreased egg production in whiteflies. The dependence of symbiont persistence on ATR-mediated bacteriocyte homeostasis underscores the importance of the ATR gene or of DNA damage repair pathway in maintaining the whitefly–symbiont mutualism, as Portiera and Hamiltonella symbionts are both essential for whitefly fitness (66, 67). Given that ATR is being explored as a therapeutic target in cancer (68), it may also represent a potential target for whitefly control strategies. Pharmacological inhibition of ATR using specific inhibitors could disrupt bacteriocyte development by inducing unresolved DNA damage, apoptosis, and symbiont loss, ultimately reducing whitefly fitness.

In conclusion, our study identifies a previously uncharacterized Adf-1-ATR regulatory pathway that safeguards whitefly bacteriocyte development and symbiosis by mitigating DNA damage. These findings advance our understanding of the molecular basis of bacteriocyte biology, highlight the central role of DNA repair in endosymbiosis, and suggest a potential avenue for pest management. More broadly, we provide a framework for investigating how specialized animal cells evolve mechanisms to preserve genomic stability in the context of long-term symbiotic associations.

Materials and Methods

Insect Rearing and Plant Materials.

A B. tabaci colony from the MEAM1 species (mtCO1 GenBank accession no. GQ332577) was maintained on cotton plants (Gossypium hirsutum, cv. Shiyuan 321) as previously described (19). The B. tabaci MEAM1 population harbors the symbionts Portiera and Hamiltonella within the same bacteriocyte (69, 70). Cotton plants were grown individually in 1.5-L pots containing potting substrate [Pingshi Tuopu Horticulture (Shanghai) Co., Ltd, Shanghai, China] until they reached the 6 to 7 true-leaf stages for experiments. Whiteflies and plants were maintained in separate climate-controlled chambers at 26 ± 2 °C, with a 14:10 h (L:D) photoperiod and 60 to 80% relative humidity (RH).

Construction of a Yeast Two-Hybrid cDNA Library of B. tabaci.

Total RNA was extracted from female whiteflies within 7 d after emergence using TRIzol reagent (Sigma-Aldrich, St. Louis, MO). mRNA was further purified using the Oligotex mRNA Kit (Qiagen, Hilden, Germany) and used for cDNA library construction following the manufacturer’s instructions. First-strand cDNA was synthesized using PowerScript™ Reverse Transcriptase (Takara Bio, Beijing, China), followed by long-distance PCR amplification using Advantage polymerase mix (Takara Bio, Beijing, China) to generate double-stranded cDNA. The resulting cDNA was then purified using the QIAquick PCR Purification Kit (Qiagen, Hilden, Germany). To reduce the abundance of highly expressed transcripts and equalize transcript concentrations in the sample, thereby improving sequencing efficiency and facilitating rare gene discovery, double-stranded cDNA was normalized using the Trimmer-2 cDNA normalization kit (Evrogen, Moscow, Russia). The normalized double-stranded cDNA was cloned into the linearized three-frame pGADT7 vector, and the resulting construct was used to transform DH5α cells by electroporation. The transformation mix was plated onto the LB/ampicillin agar and incubated at 37 °C for 24 h. The following day, bacteria were eluted from the agar plates with liquid medium, and library plasmids were extracted from the eluted bacteria using the QIAGEN Plasmid Maxi Kit (Qiagen, Hilden, Germany).

Yeast Two-Hybrid Library Screening.

The yeast-two-hybrid library screening was conducted following the protocol of the Matchmaker gold yeast two-hybrid system (Takara Bio, Beijing, China). Briefly, the full-length cDNA of Adf-1 was cloned into the pGBKT7 vector and transformed into the Y2HGold yeast strains. The transformation mix was plated on SD/-Trp and SD/-Trp/-His/-Ade/X-α-gal agar and incubated at 30 °C for 3 to 5 d to verify the autoactivation and toxicity of the bait protein Adf-1 in yeast cells. The empty pGBKT7 and the pGBKT7-GAL4 vectors served as negative and positive controls, respectively. The cDNA library plasmids were cotransformed, together with the pGBKT7-Adf-1 vector, into Y2HGold cells and plated on the SD/-Leu/-Trp/-His/-Ade medium to identify interacting proteins. Plasmids were rescued from positive colonies using a yeast plasmid extraction kit (Pyeast, Wuhan, China), and sequenced to identify prey inserts.

Pairwise Yeast Two-Hybrid Assay.

The full-length coding sequence (1 to 912 bp), the N-terminal region containing the MADF domain (1 to 456 bp), and the C-terminal region containing the BESS domain (457 to 912 bp) of Adf-1 were amplified using specific primers (SI Appendix, Table S3), cloned into the pGADT7 vector by homologous recombination, and then cotransformed with the pGBKT7-Adf-1 into Y2HGold yeast strains. The transformation mixture was plated on the SD/-Leu/-Trp and SD/-Trp/-Leu/-His/-Ade/X-α-gal agar plates and incubated at 30 °C for 3 to 5 d to determine whether Adf-1 forms homodimers through the BESS domain. The CDs of genes encoding candidate interacting proteins were separately cloned into the pGADT7 vector using specific primers (SI Appendix, Table S3), and cotransformed with pGBKT7-Adf-1 into Y2HGold yeast cells. The transformation mixture was plated on SD/-Leu/-Trp, SD/-Trp/-Leu/-His, and SD/-Trp/-Leu/-His/-Ade/X-α-gal agar plates to further confirm the specific interactions between Adf-1 and candidate proteins. Yeast strains containing pGBKT7-Lam+pGADT7-T and pGBKT7-P53+pGADT7-T vectors served as negative and positive controls, respectively.

DNA Affinity Purification Sequencing (DAP-Seq) and Data Analysis.

DAP-seq was performed as previously described (46). The Adf-1 coding sequence was cloned into the pFN19K (HaloTag®7) T7 SP6 Flexi® vector to express Halo-Adf-1 fusion proteins using the TnT® Coupled Wheat Germ Extract System (Promega, Madison, USA) following the manufacturer’s specifications. Genomic DNA was extracted from abdomens of female whiteflies 7 d after emergence using DNAiso Reagent (Takara Bio, Beijing, China) and sheared to 200 to 800 bp fragments by sonication. Halo-Adf-1 proteins were bound to anti-Halo monoclonal antibody agarose beads (Promega, Madison, USA) and incubated with 200 ng fragmented genomic DNA. Adf-1 bound DNA fragments were subsequently eluted and amplified using indexed primers. DNA libraries were sequenced on the Illumina NovaSeq platform for 150 bp paired-end sequencing (Bluescape Biotechnology Co, Beijing, China). Mock DAP-seq libraries without protein added to the beads were used as negative controls. Clean reads were aligned to the B. tabaci genome using the BWA-MEM alignment tool (71). Peak calling was performed with MACS2 (72), and peaks from two biological replicates were merged using Homer software (73) with a significance threshold (Q-value) < 0.05. The genomic distribution of identified peaks was annotated using ChIPseeker software (74). Motif discovery was carried out using the MEME-ChIP software (75). All peak-related target genes were searched against the Kyoto Encyclopedia of Genes and Genome database (r94) using BLAST v2.6.0+ with a maximum E value of 1.0E⁻⁵ (76).

Dual-Luciferase Reporter Gene Assay.

The dual-luciferase reporter gene assay was conducted as previously described (77). Briefly, the 958 bp promoter sequence of ATR identified by DAP-seq was amplified from B. tabaci genomic DNA extracted from abdomens of female whiteflies 7 d after emergence by PCR and subcloned into the firefly luciferase reporter vector pGL3-Basic (Promega, Madison, USA). The Adf-1 coding sequence was inserted into the pcDNA3.1 (−) expression vector. Human HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37 °C in a humidified incubator with 5% CO₂. Cells were counted and seeded into 96-well plates. When the cells had reached 70 to 80% confluence, they were cotransfected using GenXPIII (Probegene, Xuzhou, China) with different reporter plasmids and the Renilla luciferase reporter vector pRL-TK across experimental groups. Luciferase activity was measured 48 h posttransfection using the dual-luciferase reporter assay system following the manufacturer’s instructions. Firefly luciferase activity was normalized to Renilla luciferase activity. Nine biological replicates were conducted. All primers are listed in SI Appendix, Table S3.

RNAi Experiments.

Double-strand RNAs (dsRNAs) targeting the whitefly RecQ1, Adf-1, ATR, and the control gene GFP were synthesized using a T7 RiboMAX Express RNAi System kit (Promega, Madison, WI) with primers listed in SI Appendix, Table S3, following the manufacturer’s instructions as previously described (19, 78). Adult female whiteflies at day 1 after emergence were injected with 1.5 μg/μL dsRNA (dsRecQ1, dsAdf-1, or dsATR) in injection buffer (5 mM KCl, 0.1 mM NaH₂PO₄) using an Eppendorf microinjection system (Hamburg, Germany) as previously described (66). Control whiteflies were injected with dsGFP. The average injection volume for each whitefly was 10 nL. After injection, whiteflies were transferred to cotton leaf disks placed on 1.5% agar plates in an incubator at 26 ± 2 °C, 14:10 h (L:D) photoperiod and 60 to 80% RH.

qRT-PCR and qPCR Analysis.

To compare ATR expression in bacteriocytes and whole bodies of female whiteflies, 2,000 bacteriocytes were dissected from 100 female adult whiteflies 7 d after emergence for each of four biological replicates. To examine the effect of Adf-1 silencing on Adf-1 and ATR expression, bacteriocytes (2,000 per replicate, isolated from 100 female whiteflies for each of three biological replicates) and whole bodies (10 female whiteflies per replicate, four biological replicates) were collected 6 d after microinjection with dsAdf-1. To examine the effects of ATR silencing on the expression of Caspase 3 and Atg8 genes, 2,000 bacteriocytes were dissected from 100 female whiteflies 5 d after microinjection with dsATR for each of four biological replicates. Total RNA was extracted from dissected bacteriocytes and whole bodies of female whiteflies using TRIzol (Sigma-Aldrich, St. Louis, MO) and quantified with a NanoDrop 2000 spectrophotometer (Thermal Scientific, Wilmington). cDNA was synthesized using a reverse transcription kit from Bimake (Houston, TX). Primers for qRT-PCR were designed with Primer Premier 5.0 software (Premier Biosoft International, Palo Alto, CA) and the assay was performed as previously described (70, 79). The β-actin gene was used as the reference gene to normalize Adf-1 and ATR expression in whole bodies of dsAdf-1- and dsGFP-injected female whiteflies. RPL7 and RPL13 genes were used as reference genes to analyze ATR gene expression in bacteriocytes versus whole-body samples, Adf-1 and ATR gene expression in bacteriocytes from dsAdf-1- and dsGFP-injected female whiteflies, and Caspase 3 and Atg8 expression in bacteriocytes from dsATR- and dsGFP-injected female whiteflies, as described in a previous study (80).

For quantification of Portiera and Hamiltonella, DNA was extracted from 19 adult female whiteflies 5 d after microinjection with dsATR. qPCR was performed by determining the copy number of the 16S rRNA gene, with the β-actin gene used as an internal reference for normalization, as previously described (66, 70). Primers are shown in SI Appendix, Table S3. Two technical replicates were included for each biological replicate. Relative gene expression and symbiont density were calculated using the 2^−ΔCt method (81).

Effect of Silencing RecQ1 or ATR Genes on Bacteriocyte Numbers.

RNA was extracted from eight adult female whiteflies 1.5, 3, and 6 d following microinjection with dsRecQ1, or 3 and 5 d following microinjection with dsATR, to examine the expression of RecQ1 or ATR for each of the four biological replicates, respectively. The β-actin gene was used as the reference gene. In parallel, individual dsRNA-injected whitefly was transferred onto cotton leaf disks placed on 1.5% agar plates and maintained at 26 ± 2 °C, with 14:10 h (L:D) photoperiod and 60 to 80% RH. Egg numbers were recorded, and the number of bacteriocytes in adult female bodies was counted in 23 and 21 individuals, respectively, 3 and 6 d after microinjection with the dsRecQ1, and in 22 whitefly individuals 5 d after microinjection with the dsATR. The total number of bacteriocytes was calculated as the sum of bacteriocytes present in the female bodies plus the number of eggs deposited following dsRNA microinjection as previously described (19).

Immunofluorescence Microscopy.

To investigate whether ATR silencing affects DNA damage, apoptosis, autophagy, and cell division, bacteriocytes were dissected from adult female whiteflies 5 d after microinjection. Bacteriocytes were subsequently fixed, permeabilized, and incubated with rabbit γH2AX antibody (Cell Signaling Technology, Boston), rabbit cleaved Caspase 3 antibody (Cell Signaling Technology, Boston), rabbit Atg8 antibody (ProbeGene Life Sciences, Xuzhou, China), or mouse α-tubulin antibody (Sigma, St. Louis), following a previously described protocol (19, 53). Samples were then washed in PBS and incubated with goat anti-rabbit secondary antibody conjugated to Alexa Fluor 594 for the γH2AX antibody and to Alexa Fluor 488 for the Caspase 3 and Atg8 antibodies, or with a goat anti-mouse secondary antibody conjugated to Alexa Fluor 488 for α-tubulin antibody, following previously described protocols (19, 53). After the removal of the secondary antibody and washing, samples were incubated with DAPI (0.1 µg/mL in 3% BSA in PBS) for 15 min at room temperature. Caspase 3, Atg8, and α-tubulin antibody had been validated in a previous work (19). Samples incubated without the γH2AX primary antibody served as negative controls. Images were acquired using a FV3000 confocal microscope (Olympus, Tokyo, Japan). γH2AX protein levels are positively correlated with DNA damage (49). γH2AX foci numbers were analyzed in 21 bacteriocytes, and fluorescence intensity of Caspase 3 and Atg8 were analyzed in 24 bacteriocytes from dsATR-injected and dsGFP-injected whiteflies, respectively, using Image J software. Spindle formation, visualized by α-tubulin staining, indicated bacteriocytes undergoing division. At least 100 bacteriocytes were dissected from five to eight dsRNA-injected adult females for each of four biological replicates to determine the proportion of dividing bacteriocytes.

Quantification and Statistical Analysis.

All datasets were tested for normality using the Shapiro–Wilk test and for homogeneity of variance using Levene’s test. For comparisons between two groups, the Mann–Whitney U test was used when the data did not meet the assumption of normality. When the data met the assumptions of normality and homogeneity of variance, an independent-samples Student’s t test was used. When the data met the assumption of normality but showed unequal variances, Welch’s t test was adopted. For comparisons involving more than two groups of data, the Kruskal–Wallis test was used when the data did not conform to the normal distribution. When the data simultaneously conformed to the normal distribution and homogeneity of variance, one-way ANOVA followed by the least significant difference (LSD) post hoc test was conducted. If the data met the assumption of normality but showed unequal variances, one-way ANOVA followed by Tamhane’s T2 post hoc test was used. The significance threshold was set at P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001) for all tests. Proportion data were arcsine square-root transformed before analysis. All statistical analyses were performed using IBM SPSS Statistics software version 26 (SPSS Inc., Chicago, IL).

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Data, Materials, and Software Availability

The raw reads for DAP-seq are available at the National Center for Biotechnology Information (NCBI) Short Read Archive (SRA), with accession number PRJNA1256209 (82). Data used in the study are included in the manuscript and/or supporting information.