5mC modification orchestrates choriogenesis and fertilization by preventing prolonged ftz-f1 expression

DNA methylation at the fifth position of cytosine (5-methylcytosine, 5mC) is a crucial epigenetic modification for regulating gene expression, but little is known about how it regulates gene expression in insects. Here, we pursue the detailed molecular mechanism by which DNMT1-mediated 5mC maintenance regulates female reproduction in the German cockroach, Blattella germanica. Our results show that Dnmt1 knockdown decreases the level of 5mC in the ovary, upregulating numerous genes during choriogenesis, especially the transcription factor ftz-f1. The hypomethylation at the ftz-f1 promoter region increases and prolongs ftz-f1 expression in ovarian follicle cells during choriogenesis, which consequently causes aberrantly high levels of 20-hydroxyecdysone and excessively upregulates the extracellular matrix remodeling gene Mmp1. These changes further impair choriogenesis and disrupt fertilization by causing anoikis of the follicle cells, a shortage of chorion proteins, and malformation of the sponge-like bodies. This study significantly advances our understanding of how DNA 5mC modification regulates female reproduction in insects.

the eggshell or chorion during choriogenesis 12 .During previtellogenesis, follicle cell growth, driven by insulin signaling and the target of rapamycin 13 .Vitellogenesis is mainly controlled by juvenile hormone (JH) and/or 20-hydroxyecdysone (20E, the active form of insect steroid hormone converted from ecdysone) in different insect orders [13][14][15] .In most, if not all, insects, 20E regulates the final oogenesis stage, choriogenesis [16][17][18] .Notably, the ovaries become the primary source of ecdysteroids in adult females of many species, as the larval/ nymphal-specific ecdysteroid synthesis (steroidogenesis) tissue, the prothoracic gland, degenerates post metamorphosis 19 .However, the regulatory mechanism of ovarian steroidogenesis is elusive.The 20E level in the ovaries of the German cockroach gradually increases and peaks before oviposition, and ovarian 20E triggers choriogenesis in an autocrine fashion 19 .In the silkworm, Bombyx mori, topical application of 20E during choriogenesis negatively impacts follicle cell development, and a particularly high level of 20E is able to induce follicle cell death 12 .
5mC modification has been reported to play a role in female reproduction in various insect species, such as the German cockroach, B. germanica; the silkworm, B. mori; the parasitic wasp, Nasonia vitripennis; the brown plant hopper, Nilaparvata lugens; and the milkweed bug, Oncopeltus fasciatus 7,9,[20][21][22][23] .However, the underlying regulatory mechanism remains largely unknown.Compared to those in other insect orders, 5mC levels are higher in cockroaches (e.g., the German cockroach) 6 , which are representative and classical hemimetabolous insects and suitable for studying panoistic ovary maturation and oogenesis.Previously, female reproduction and its regulation have been well studied in the German cockroach 18,[24][25][26][27][28] .In this study, using the German cockroach as a model insect, we elucidated how 5mC modification regulates female reproduction through controlling ftz-f1 medicated choriogenesis and fertilization.

DNMT1-mediated 5mC maintenance is indispensable for choriogenesis and fertilization
To investigate the relationship between 5mC modification and female reproduction in B. germanica, we first examined the spatiotemporal pattern of 5mC modification in female adults.The spatial pattern of 5mC was detected in six tissues/organs, and higher levels in the ovaries, thorax, and integuments compared to the heads, legs, and midguts (Fig. 1a).Subsequently, the temporal pattern of 5mC in the ovaries during the whole first reproductive period (8 days) was evaluated.A decreasing trend was observed from day 1 to day 4 post-adult emergence (PAE), followed by an increase until oviposition (Fig. 1a').Focusing on the DNMT1 protein, which has been shown to be essential for embryogenesis in this species 9 , the functional domains are highly conserved throughout evolution (Supplementary Fig. 1a).The fulllength DNMT1 of B. germanica shares 45% amino acid sequence identity with human DNMT1.Furthermore, the expression level of Dnmt1 in the ovaries during the first gonadal cycle was evaluated by using quantitative real-time PCR (qRT-PCR), and its developmental pattern was consistent with the genomic 5mC tendency (Fig. 1b).Together, these spatiotemporal patterns suggest that the DNMT1induced 5mC maintenance in ovaries may play a crucial role in female reproduction.
To test the hypothesis stated above, we performed RNA interference (RNAi) experiments in female adults.dsRNA was designed to target the Dnmt1, and the gene expression was significantly reduced upon dsDnmt1 treatment (Supplementary Fig. 1b).Dot blot analyses revealed that, compared to the negative control (dsCK) group, the 5mC levels were significantly reduced in the dsDnmt1 group (~40%) (Fig. 1c).To verify these results, a much more sensitive antibody-independent LC-MS/MS analysis was performed, and almost identical results were obtained (Fig. 1d).The impact of RNAi-mediated decreases in 5mC levels on female reproduction in B. germanica was then examined.
In the dsDnmt1 group, all oothecae were atrophied and wizened, and all the eggs inside the oothecae failed to hatch (Fig. 1e).Interestingly, neither mating nor ootheca-laying rates were affected (Supplementary Fig. 1c, d).We next examined the possible defects of dsDnmt1 treatment on eggs and ovaries.Firstly, DAPI staining of eggs was carried out to examine embryogenesis after ootheca-laying.Abundant energids in the yolk mass were observed in the dsCK-treated eggs, which tended to concentrate at the ventral surface at 60 h and developed into early embryos at 84 h 29,30 .Conversely, the dsDnmt1-treated eggs did not show cleavage energids in the yolk mass, only faint and diffuse staining was observed, and embryogenesis did not occur.The dsDnmt1induced phenotypic defects were similar to those in the unmated negative control group, as German cockroaches do not undergo parthenogenesis (Fig. 1f).Moreover, pronuclear staining of eggs that had just entered the oothecae revealed both female pronuclei from egg and male pronuclei from sperm in the dsCK-treated eggs, while no male pronuclei were detected in dsDnmt1-treated or unmated eggs (Fig. 1g).Importantly, the sponge-like body, an atretic structure appearing at the anterior pole of the basal oocyte through which sperms enter maturing egg for fertilization 31 , was severely malformed in the dsDnmt1-treated insects (Fig. 1h).Together, the experimental data demonstrate that dsDnmt1-induced malformation of sponge-like body disrupts fertilization and further terminates embryogenesis.
We further investigated other phenotypic defects in the dsDnmt1treated ovaries.There was no difference in weight or vitellogenin incorporation into oocytes between the dsDnmt1 and dsCK groups (Supplementary Fig. 1e, f).The ovarian follicle cells did not exhibit any noticeable effects on day 7 PAE (during the vitellogenesis stage) after dsDnmt1 treatment.However, on day 9 PAE (during the choriogenesis stage), the follicle cell nuclei in the dsDnmt1-treated ovaries displayed irregular and condensed shapes, and these follicle cells displayed signs of cytoskeletal disorganization (Fig. 1i).Additionally, the tunica propria, which covers the follicle cells 31 , exhibited a reticular distribution of hollows in the dsDnmt1 group (Supplementary Fig. 1g).These findings suggest that DNMT1-mediated 5mC maintenance is essential for timely choriogenesis and proper fertilization in the German cockroach.

Hypomethylation-induced excessive 20E impairs choriogenesis and fertilization
Given the distinct phenomena induced by dsDnmt1 on day 7 PAE and day 9 PAE, RNA-Seq was performed on ovaries at both time points.In the dsCK group, there was a significant difference in the transcriptomes between day 7 and day 9, as revealed by principal component analysis (PCA).However, the transcriptomes of the dsDnmt1 group on both day 7 and day 9 were closely resembled that of the dsCK group on day 7, suggesting that dsDnmt1 treatment did not have a significant effect on day 7 but delayed developmental progression on day 9 (Fig. 2a).Volcano plot analysis supported this finding, revealing approximately 300 differentially expressed genes (DEGs) between the dsDnmt1 and dsCK groups on day 7, and around 3000 DEGs on day 9. Notably, the number of upregulated genes exceeded the downregulated genes (1868/1306) on day 9 (Fig. 2b).These transcriptomic changes were consistent with the morphological changes observed at both time points, with insignificant effects on day 7 but significant effects on day 9 (Fig. 1i).
To investigate the effect of dsDnmt1 on choriogenesis, we analyzed the upregulated DEGs on day 9 PAE and examined their enrichment in KEGG pathways.Our analysis revealed that several nutrition-related pathways (e.g., cAMP signaling, PI3k-Akt signaling, insulin signaling, AMPK signaling, insulin secretion, biosynthesis of amino acids) and the insect hormone biosynthesis pathway were activated and enriched (Supplementary Fig. 2a), implying that nutrition or/and hormone might be involved in the DNMT1regulated choriogenesis.To test this hypothesis, exogenous glucose, amino acids (Aa) mixture, and insulin were injected into female adults, however, no significant decrease in embryonic development or hatching rate was observed for any of the treatments (Supplementary Fig. 2b, c).Thus, we focused on the insect hormone biosynthesis pathway, since phantom (phm) and shadow (sad), both of which are involved in steroidogenesis, were identified upregulated in dsDnmt1 group (Fig. 2b, c).As aforementioned, ovary is the main source of steroidogenesis and circulating 20E in Fig. 1 | DNMT1-mediated 5mC maintenance is indispensable for choriogenesis and fertilization.a, a' Spatial 5mC modification was detected in six different tissues, and temporal 5mC levels in ovaries during the first reproductive period were analyzed by dot blotting using a 5mC-specific antibody.The same amounts of gDNA were loaded as input in each panel.b Gene expression patterns of Dnmt1 in the ovaries during the first reproductive period were detected by qRT-PCR, n = 3 biologically independent samples.c 5mC levels were detected after RNAi of Dnmt1 by dot blotting, and the density for each sample was calculated, n = 4 biologically independent samples.d LC-MS/MS detection for the 5mC levels under dsRNA treatment, n = 3 biologically independent samples.e Calculation of hatching rates under RNAi of Dnmt1 and negative control.The atrophied and wizened phenotypes of oothecae in dsDnmt1 group compared with dsCK control group were showed, n = 3 biologically independent experiments.f DAPI staining was used to observe the development of embryos or unfertilized eggs, 36 h, 60 h, and 84 h post ootheca-laying were selected.The arrows point to the position of embryogenesis.More than three eggs were observed in each treatment group.g Pronucleus staining for the eggs using α-tubulin antibody.More than five eggs were observed in each treatment group.h SEM observation of sponge-like bodies on eggs, 9 of 13 sponge-like bodies were observed abnormal.i DAPI and PHA staining for nuclei and actin protein in ovarian follicle cells.Not less than three eggs were observed in each group.Of note, one more day is added and there are 9 days in the first reproductive period under dsRNA injection condition.Data are mean ± sd, the differences were analyzed by two-tailed Student's t test.Source data are provided as a Source Data file.female adults 11,19 .Notably, we observed a 3-fold increase in the level of 20E in the ovary of the dsDnmt1 group on day 9 PAE (Fig. 2d).To confirm the effect of 20E on choriogenesis, we injected a low dose of 20E into female adults.Similar to previous reports in B. mori 12 , most 20E-treated oothecae atrophied and wizened, and the majority of the eggs inside them failed to hatch (Fig. 2e).Moreover, the nuclei and cytoskeletal structures of some follicle cells became abnormally shaped (Fig. 2f), and most of the sponge-like bodies in the 20E-treated animals were malformed (Fig. 2g).In short, the dsDnmt1-induced phenotypic defects in choriogenesis and fertilization should be attributed in part to the aberrant elevation of 20E levels during choriogenesis.Hypomethylation-increased and -prolonged ftz-f1 expression causes excessive steroidogenesis To unravel the connection between differential gene expression and 5mC modification, we performed whole-genome bisulfite sequencing (WGBS) to analyze differences in 5mC levels in ovaries between the dsDnmt1 and dsCK groups on day 9 PAE (Supplementary Fig. 3a).In accordance with previous studies in B. germanica, methylation modifications were found predominantly at cytosines within the CpG dinucleotide 9 , while no noticeable cytosine methylation at CHG or CHH sites (Fig. 3a, b).Importantly, whole-genomic methylation levels at CpG sites decreased from 15% to 6% (P = 0.001) in the ovaries of dsDnmt1-treated insects (Fig. 3a).Significant differences in DNA methylation at CpG sites between the dsDnmt1 and dsCK groups were also revealed by PCA (Supplementary Fig. 3c).Compared to the dsCK group, 99.97% differentially methylated regions (DMRs) were observed as hypomethylated (hypo-) DMRs (205842) in the dsDnmt1 group, and only 0.03% hypermethylated (hyper-) DMRs (57) were found (Fig. 3b), showing that most DMRs were produced by dsDnmt1induced hypomethylation.Moreover, in the dsDnmt1 group, significant decreases in 5mC levels were observed in all gene regions of the genome, including the Up 2k region, the 5′-untranslated region (5′ UTR), the exons, the introns, the 3′-untranslated region (3′UTR), and the Down 2k region (Supplementary Fig. 3d).Then, we aimed to establish the relationship between alterations in steroidogenesis and changes in 5mC levels in the ovaries of dsDnmt1-treated insects on day 9 PAE.We first examined DEGs directly involved in steroidogenesis in the dsDnmt1 group.Although phm and sad were upregulated, no hypomethylation was observed in the Up 2k or promoter region of these genes (Fig. 3c), suggesting that the upregulation of phm and sad was not directly caused by 5mC level changes in their promoter regions.Then, genes involved in regulating steroidogenesis and 20E signaling, such as ecdysone receptor (EcR) 32 , Ecdysone-induced protein 75 (E75) 33 , Ecdysone-induced protein 93 (E93) 34 , and fushi tarazu factor 1 (ftz-f1) 35 were evaluated in the dsDnmt1 group.Hypomethylation occurred in the promoter and Up 2k regions of EcR, while EcR was only slightly upregulated (1.32-fold, P = 0.051); no significant differences were observed for these parameters of E75; E93 expression was slightly upregulated but no hypomethylation was found in its promoter or Up 2k region; notably, ftz-f1 was significantly upregulated and hypomethylation significantly occurred in its Up 2k and promoter regions (Fig. 3c).The composite analyses suggest that DNMT1 may inhibit steroidogenesis by maintaining 5mC levels in the promoter regions of ftz-f1.
Next, a correlation analysis was performed to determine the relationship between gene expression changes and 5mC decreases in the promoter region between the dsDnmt1 and dsCK groups.A stronger correlation was observed between gene upregulation and hypomethylation in the promoter regions than that of gene downregulation, and the number of upregulated genes was 1.5-fold higher than that of downregulated genes (413/275) (Fig. 3d).These data indicate that 5mC modification generally plays a more prominent role in repressing gene expression in the ovaries.Subsequently, we analyzed the 413 hypomethylated and upregulated genes in top-left quadrant, eight transcription factor (TF) genes, huckebein (hkb), ftz-f1, Mi-2, Hormone receptor-like in 39 (Hr39), bric a brac 2 (bab2), Cyclic-AMP response element binding protein A (CrebA), nejire (nej), and deformed wings (dwg), which might regulate the expression of steroidogenesis genes were screened out (Fig. 3e).Next, the top six TF genes (>2-fold upregulation) were knocked down individually (Supplementary Fig. 3e), and only dsftz-f1 treatment significantly decreased the expression of phm (Fig. 3f).As confirmed by qRT-PCR analysis, Dnmt1 RNAi caused approximately 4-fold upregulation of ftz-f1 in the ovaries on day 9 PAE (Supplementary Fig. 3f).Importantly, knocking down ftz-f1 also resulted in a significant decrease of EcR expression (Supplementary Fig. 3g); and moreover, knocking down of phm, ftz-f1, EcR, or taiman (tai, the co-activator of EcR) significantly reduced 20E levels (Fig. 3g and Supplementary Fig. 3h-S3j).These findings suggest that FTZ-F1 might not only act as a transcription factor promoting steroidogenesis but also as a competence factor working with Tai to enhance EcR activity for 20E signaling 36,37 .Altogether, we conclude that dsDnmt1-caused hypomethylation increases and prolongs ftz-f1 expression, leading to excessive steroidogenesis.

ftz-f1 promoter undergoes 5mC modification for preventing prolonged ftz-f1 expression
We next focused on the potential link between dsDnmt1-induced hypomethylation and prolonged ftz-f1 expression on day 9 PAE.To validate the reduction in 5mC levels in the ftz-f1 promoter region, the comprehensive 5mC levels around the ftz-f1 in the genome was checked using WGBS data.As expected, a sharp decrease of 5mC level in the ftz-f1 promoter region was observed in the dsDnmt1 group (Fig. 4a).Moreover, the high-throughput next-generation sequencingbased bisulfite sequencing PCR (NGS-BSP) was employed, and sharp reductions in 5mC levels were observed in six amplicons of the ftz-f1 promoter region (~7 kb) (Fig. 4b).We also evaluated the temporal pattern of 5mC levels in the promoter of ftz-f1 during the first gonadal cycle using NGS-BSP.The same 6 amplicons were analyzed, and there was an evident decreasing trend in amplicons 3 and 5 from day 1-3 PAE to day 4-6 PAE, followed by an increase until oviposition (Fig. 4c).Importantly, under our rearing conditions, ftz-f1 expression remained low until day 6 PAE and reached its peak at the end of vitellogenesis (day 7), and then declined during choriogenesis (Fig. 4d).In general, the 5mC levels in these two promoter regions of ftz-f1 are negatively correlated with ftz-f1 expression levels, indicating that the ftz-f1 promoter undergoes DNMT1-mediated 5mC modification to prevent prolonged ftz-f1 expression.

dsDnmt1-induced defects are partially rescued by ftz-f1 knockdown
To confirm the negative regulatory effect of Dnmt1 on ftz-f1 expression, dsDnmt1 treatment and dsftz-f1 treatment were conducted simultaneously.Co-treatment with dsftz-f1 attenuated the dsDnmt1-induced upregulation of ftz-f1 (Fig. 5a).Notably, the excessive 20E levels induced by dsDnmt1 were also partially attenuated by the co-treatment of dsftz-f1 (Fig. 5b).Since dsftz-f1 treatment can completely disrupt oviposition process, the co-treatment with dsDnmt1 increased the oviposition rate back to a relatively high level (~80%) (Fig. 5c).Additionally, we paid more attentions on the obvious phenotypic defects on follicle cells and sponge-like bodies.The co-treatment with dsftz-f1 partially rescued the dsDnmt1-Fig.2 | Hypomethylation-induced excessive 20E impairs choriogenesis and fertilization.a PCA of gene expression among four groups: the dsCK day 7 (d7), the dsCK day 9 (d9), the dsDnmt1 d7, and the dsDnmt1 d9.b Volcano plot analysis of the DEGs under dsDnmt1 treatment conditions on d7 and d9.Green dots indicate downregulated genes, and red dots indicate upregulated genes.c Heat map of genes enriched in the pathways of insect hormone biosynthesis.d 20E levels detection under dsDnmt1 treatment on day 9, n = 8 biologically independent samples.e Calculation of hatching rates under exogenous 20E treatment.The atrophied and wizened phenotypes of oothecae in 20E group compared with ethanol control group were showed, n = 3 biologically independent experiments.f DAPI and PHA staining for nuclei and actin protein in ovarian follicle cells.g SEM observation of sponge-like bodies on eggs under exogenous 20E treatment.8 of 10 sponge-like bodies were observed abnormal.Data are mean ± sd, the differences were analyzed by two-tailed Student's t test.Source data are provided as a Source Data file.

5mC modification prevents ftz-f1-induced Mmp1 expression and anoikis
To investigate how hypomethylation-induced ftz-f1 upregulation regulates choriogenesis and fertilization, we conducted additional investigations since exogenous 20E treatment only partially phenocopy dsDnmt1-induced defects.Focusing on the abnormal disorganization  of follicle cells, excessive follicle cells dissociation was found after dsDnmt1 treatment (Fig. 6a).Moreover, the activity of cell death marker Caspase-3 was dramatically increased in follicle cells of dsDnmt1-treated insects, and this dsDnmt1-induced Caspase-3 activity could be blocked by the Pan Caspase inhibitor Z-VAD-FMK (Fig. 6b).These results suggest that the dsDnmt1-treated follicle cells were likely undergoing anoikis, a cell death process related to detachment 38 .It is well-documented that FTZ-F1 induces the expression of matrix metalloproteinase (Mmp) genes, thus regulate the maturation and rupture of follicle cells in the ovaries of Drosophila melanogaster 39,40 .
Consistent with this regulatory mechanism, Mmp1 was upregulated in the dsDnmt1-treated ovaries and rescued by dsDnmt1 and dsftz-f1 double RNAi (Fig. 6c).When ftz-f1 was knocked down, the expression of Mmp1 in the ovary was also significantly downregulated (Supplementary Fig. 4a).Additionally, no hypomethylation was observed in the Up 2k or promoter region of Mmp1 (Supplementary Fig. 4b), indicating that Mmp1 expression is not directly regulated by DNMT1mediated 5mC modification.Moreover, the expression of five wellknown chorion genes, cathepsin-L, yellow-g, citrus, Fcp3c, and brownie, in the transcriptome were explored 27,31,41 .There was no significant difference in their expression on day 7 PAE, and three of them (yellow-g, citrus, and Fcp3c) were significantly downregulated in the dsDnmt1 group on day 9 PAE (Supplementary Fig. 4c).The downregulation of these three genes was validated by qRT-PCR, while cathepsin-L was also found downregulated significantly when detected by qRT-PCR (Fig. 6d).Collectively, the hypomethylation-increased and prolonged ftz-f1 expression results in the dissociation and anoikis of follicle cells, which might lead to the failure to synthesize sufficient chorionic proteins and further disrupt fertilization (Fig. 1h, i).

Discussion
The composite data demonstrate that dsDnmt1 induces reproductive disability primarily through two mechanisms associated with ftz-f1 overexpression: by causing aberrantly high levels of 20E, anoikis of follicle cells, and a shortage of chorion proteins.This impairment of choriogenesis eventually causes sponge-like body malformation and thus disrupts fertilization.In conclusion, 5mC modification prevents prolonged ftz-f1 expression, ensuring timely choriogenesis and proper fertilization in the German cockroach (Fig. 7).
Identifying the key gene that tightly links DNA 5mC modification with insect reproduction in females is the primary challenge in the field.In a previous study on O. fasciatus, 5mC levels in the ovaries decreased after Dnmt1 RNAi, but no quantitative changes in gene transcript expression were observed 7 .This negative result was likely due to inappropriate selection of the time point for RNA-Seq.In our study, we obtained a similar result with only a few DEGs on day 7 PAE, but numerous DEGs were found on day 9 PAE.In other words, the reduction in 5mC levels caused delayed developmental progression on day 9 PAE (Fig. 2a, b).Moreover, in both vertebrates and insects, there's an ongoing debate about the facilitative or preclusive effects of DNA methylation on gene expression 42,43 .In this study, a large number of both upregulated genes and downregulated genes were observed after Dnmt1 depletion, while the number of upregulated genes was much higher than downregulated genes (Fig. 2b), indicating that DNMT1-mediated DNA methylation mainly functions as a gene repressor in the German cockroach.Notably, an interesting question that arises now is how the expression of Dnmt1 is precisely regulated during choriogenesis.
It has been reported that FTZ-F1 acts as a key factor involved in choriogenesis regulation 17 , while this is the first instance where the gene has been shown to be controlled by 5mC modification regarding this reproductive process.ftz-f1 expression was likely regulated by both positive and negative factors during choriogenesis: the 20E pulse should be responsible for inducing ftz-f1 expression, consequently, DNMT1-mediated 5mC maintenance prevents prolonged ftz-f1 expression.Interestingly, hypomethylation-induced aberrant elevations in FTZ-F1 levels overactivated the expression of steroidogenic genes and thus enhanced 20E levels in ovarian follicle cells, which disturbed the formation of sponge-like bodies and impaired fertilization.These results show that FTZ-F1 is involved in steroidogenesis regulation in ovarian follicle cells of B. germanica, and a similar relationship has been observed in the prothoracic glands of D. melanogaster larvae 35,[44][45][46] .Additionally, exogenous 20E treatment only partially mimicked dsDnmt1-induced phenotypic defects, which suggests more genes or factors are directly affected by dsDnmt1 treatment (Fig. 3d).Moreover, the regulation of female reproduction among insects might be regulated by multiple factors (e.g., JH, neuropeptides, and nutrition), and the epigenetic regulation might be involved in different physiological adaptations.We found that food deprivation downregulated the expression of Dnmt1 (Supplementary Fig. 5), making it interesting to address these mechanisms further.
Previous studies have shown that FTZ-F1 can regulate follicle cell dissociation, choriogenesis, and fertilization by affecting the expression of downstream genes, including Mmps 39,40,47 .Cell dissociation is a classic marker for detachment-induced apoptosis, anoikis 38 .Here, we found that FTZ-F1 significantly promoted the expression of Mmp1, which might further lead to the cleavage of E-cadherin and result in excessive follicle cell dissociation eventually.The resulting follicle cell anoikis might be the reason for the abnormalities in the nuclei and cytoskeleton (Fig. 1I), leading to the failure to synthesize sufficient chorionic proteins (Fig. 6d).Moreover, FTZ-F1 has been found controlling the degeneration and death of the prothoracic gland by stimulating the expression of E93 during the nymphal-adult transition in B. germanica 48,49 .ftz-f1 was also reported playing a dual function in oviposition in the B. germanica recently; it promotes the expression of cytoskeleton and muscle-related genes at the early stage, while at the late stage, it maintains the expression levels of these genes 50 .In short, DNA 5mC modification plays key roles in follicle cell dissociation and choriogenesis.Of note, the effects of DNMT1 are likely genome-wide, and other DEGs or pathways except ftz-f1 mentioned in this study Fig. 3 | Hypomethylation-induced ftz-f1 overexpression regulates steroidogenesis in follicle cells.a Calculated cytosine methylation rates for CG, CHG, and CHH motifs, where H equals A, T, or C. Methylation rate=methylated reads/ (methylated reads + unmethylated reads), n = 2 biologically independent samples, data are mean ± sd.b Violin plots of the distributions of DMRs in hypo-and hypermethylated groups.The median, maximum, and minimum values are shown.c Analysis of steroidogenesis-related genes.Fold change statistics for the FPKM value and methylation level differences between the dsDnmt1 and dsCK groups (dsDnmt1/dsCK).Bar charts with padding means the methylation level was not detectable ("nd") at Up 2k or promoter regions, there are three biologically independent samples for gene transcription assays and two biologically independent samples for gene methylation assays.d Correlation analyses of 5mC changes and gene expression differences.The horizontal axis shows the difference in methylation (dsDnmt1/dsCK), and the vertical axis shows the difference in gene expression (log2 dsDnmt1/dsCK).Red dots indicate upregulated genes, green dots indicate downregulated genes, and the number in each quadrant indicates the number of genes with the corresponding methylation and expression change patterns.e Eight TF genes were found by doing the intersecting between 413 hypo-methy and up DEGs and 259 trusted TF genes in D. melanogaster (Dmel).Fold change statistics for the FPKM value and methylation level differences of these eight TF genes were shown with bar chart, there are three biologically independent samples for gene transcription assays and two biologically independent samples for gene methylation assays.f phm relative expression levels were detected when top six TF genes with >2-fold upregulation were knocked down separately, n = 3 biologically independent samples.g 20E levels detected under dsRNAs treatment on day 9, n = 4 biologically independent samples.Data are mean ± sd, the differences were analyzed by two-tailed Student's t-test.Source data are provided as a Source Data file.might also be involved in the choriogenesis, because dsftz-f1 co-treatment only partially mimicked dsDnmt1-induced phenotypic defects.For example, other unknown DEGs are likely involved in the reticular distribution of hollows on the tunica propria (Supplementary Fig. 1g).Altogether, we propose a detailed regulatory mechanism by which DNMT1-mediated 5mC maintenance regulates female reproduction partially by repressing ftz-f1 expression in B. germanica.Acting as a brake during normal choriogenesis, 5mC in the ftz-f1 promoter regions prevents prolonged ftz-f1 expression (Fig. 7a).Reductions in 5mC levels relieve the 5mC-medidated inhibition of ftz-f1 expression during choriogenesis, and the aberrant elevations in ftz-f1 expression disturb choriogenesis and disrupt fertilization by inducing an aberrantly high 20E level and causing excessive Mmp1 expression, which in turn trigger anoikis of follicle cells and disrupt the formation of sponge-like bodies (Fig. 7b).This study provides a detailed regulatory mechanism by which 5mC modification precisely regulates insect reproduction.

Animal culture
The strain of B. germanica used in this study was originally collected from downtown Shanghai in the 1970s and is a well-established laboratory strain bred for nearly 50 years without exposure to insecticides.To maintain the colony, cockroaches were reared in plastic jars at 28 °C and 70% relative humidity in the dark.They were provided with rat chow and water 51 .Freshly emerged adults were separated from the colony after molting.The animals were randomly assigned to the test groups in all experiments.All animal protocols were reviewed and approved by the University Animal Care and Use Committee of South China Normal University.

5mC detection using dot blotting
For spatial comparative analysis of 5mC modification, six tissues, including the integument, head, thorax, legs, midgut, and ovaries, were dissected on day 5 PAE.For temporal analysis, the ovaries were collected daily throughout the first reproductive cycle (days 1-8 PAE).For 5mC detection after RNAi treatments, the ovaries were collected on day 5 PAE.Genomic DNA (gDNA) was extracted from the tissues using a gDNA extraction kit (Aidlab, DN14) and then treated with RNase to digest the RNA.gDNA on nitrocellulose membranes was incubated with a 1:5000 dilution of a 5mC-specific antibody (Abcam, 33D3) and then incubated with a 1:5000 dilution of an HRP-conjugated anti-rabbit IgG secondary antibody (Beyotime, A0239).The membranes were then treated with enhanced chemiluminescence reagent.The same amount of gDNA was loaded as input and stained with ethidium bromide dye at a 1:500 dilution (Thermo Fisher, 15585011).

qRT-PCR for gene expression detection
For temporal analyses of Dnmt1 gene expression in ovaries, tissues from three individuals were collected and mixed to form one sample at For other gene expression analyses, ovaries from three individuals were collected after the corresponding treatments and were mixed to form one sample.Three replicates were tested for each gene.qRT-PCR was performed in triplicate using Hieff qPCR SYBR Green Master Mix (Yeasen, 11202ES03).The relative gene expression was calculated relative to Actin-5c expression using the ΔΔCt method following the manufacturer's instructions 14 .The primers used for qRT-PCR are shown in Supplementary Table 1, and all the specificities of the primers were confirmed by Sanger sequencing of the PCR products.

Double-stranded RNA treatment
For the negative control, dsCK targeting a clone vector sequence was used 15 .Sense and antisense RNA was synthesized in a single reaction using a T7 RiboMAX Express RNAi System (Promega, P1700).After purification, the double-stranded RNA (dsRNA) concentration was adjusted to 3 μg/μl 15 .Under normal conditions, female adults were injected with 6 μg of dsDnmt1 between the abdominal sternites using an insulin syringe for each injection.To thoroughly disrupt gene function, three injections were performed (on days 2, 4, and 6 PAE).The Dnmt1 RNAi efficiency was detected on day 5 PAE after twice injection on day 2 and day 4 PAE.For transcript factor genes hkb, ftz-f1, Hr39, Mi-2, bab2, and CrebA, RNAi efficiency was detected on day 5 PAE after twice injection on day 2 and day 4 PAE.For phm, tai, and EcR, 6 μg of double-strand RNAs were injected on day 7 PAE, and relative expression levels were detected on day 9 PAE.In the double RNAi experiments, 6 μg of dsCK or dsDnmt1 were injected on day 2, 4, 6 PAE, and dsftz-f1 were injected on day 7 PAE with 10 ng.All dsRNA-injected female adults were paired with wild-type male adults for mating.The primer sequences for dsRNA synthesis can be found in Supplementary Table 2.

LC-MS/MS detection of DNA 5mC modification
DNA 5mC modification was detected by MetWare Company (Wuhan, China) with an AB Sciex QTRAP 6500 LC-MS/MS platform.The ovaries were dissected from female adults treated with dsCK or dsDnmt1, and at least three replicates were tested for each group.gDNA was digested by DpnI (NEB, R0176S) to remove possible bacterial DNA.gDNA completely digested with deoxyribonuclease was analyzed via LC-MS/MS.The ESI-MS/MS instrument used was an AB 6500 QTRAP LC-MS/MS system equipped with an ESI Turbo Ion-Spray interface operating in both positive and negative ion modes and controlled by Analyst 1.6.3software (AB Sciex).

Whole-genome bisulfite sequencing analysis of cytosine methylation
For detection of 5mC levels in B. germanica after dsCK or dsDnmt1 treatment, ovaries at day 9 PAE that had been treated with dsRNA were dissected; two biological replicates were tested for each treatment.dsRNA injections were performed on days 2, 4, and 6 PAE (three times), and ovaries from six individuals were dissected and mixed to form one sample.The gDNA was extracted for WGBS library construction using an AceGen Bisulfite-Seq Library Prep Kit (AceGen, AG0311) according to the manufacturer's protocol.The gDNA from ovaries was extracted for WGBS library construction.After bisulfite treatment, the DNA was amplified in 10 cycles of PCR with Illumina 8 bp dual index primers.The constructed WGBS libraries were then analyzed with an Agilent 2100 Bioanalyzer and finally sequenced on Illumina platforms using a 150×2 paired-end sequencing protocol.The FastQC tool was used to perform basic statistical analysis on quality of the raw reads.BSMAP software was used to map the bisulfite sequences to the reference genome.The methylation levels of individual cytosines were calculated with the formula ML = mC/(mC+umC), where ML is the methylation level and mC and umC represent the numbers of reads supporting methylated C and unmethylated C, respectively.Differentially methylated regions (DMRs) were identified using metilene (ver.0.2-7) 52 , and the regions with Kolmogorov-Smirnov (KS)-test P value < 0.05 and Benjamini & Hochberg (BH)-corrected P values < 0.05 were considered DMRs.Based on the results of DMR annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotation 53 , functional enrichment analysis was performed on genes whose gene bodies or 2 kb upstream or downstream regions overlapped with a DMR.

Tissue staining
For ovary staining, fresh ovaries were fixed in 4% paraformaldehyde in PBS for 2 h, and then washed in PBS supplemented with 0.3% Triton (1×PBT).For DAPI (Beyotime, C1002) and phalloidin (PHA) (Yeasen, 40734ES75) staining, the ovaries were incubated with a 1:1000 dilution of phalloidin in 1× PBT for 2 h, and 1 µg/ml DAPI was added for the last 5 min.Three washes with 1× PBT were performed before observation.For egg pronucleus staining, the processes were performed based on modified CUBIC protocol 54 .The pronucleus was stained by primary αtubulin antibody (Sigma, MAB1864-I), and Cy3-labeled Goat Anti-Rat IgG(H + L) (Beyotime, A0507) was used as the secondary antibody.For Caspase-3 activity detection, the ovaries were incubated with a 1:2000 dilution of a Caspase-3 (Beyotime, AC033) rabbit monoclonal antibody overnight at 4 °C.After three washes with 1× PBT, the ovaries were incubated with a 1:1000 dilution of an Alexa Fluor 488-labeled goat anti-rabbit IgG (H + L) (Beyotime, A0423) secondary antibody.All the signals were observed via confocal microscopy.

Scanning electron microscopy
Selected eggs from dsDnmt1-and dsCK-treated female adults were fixed in 2.5% glutaraldehyde and 2% paraformaldehyde in cacodylate buffer (0.2 M) for at least 2 h.After rinsing twice with the same buffer, the samples were treated with 1% osmium tetroxide (Ted Pella) at 4 °C for 1 h.The tissues were dehydrated with increasing concentrations of ethanol at 15 min intervals.Finally, the samples were subjected to critical-point drying to complete the dehydration process.The samples were attached to stubs with double-stick tape, coated with gold-palladium in a sputter coating apparatus, and then observed via scanning electron microscopy (SEM) at 5 kV (JEOL JSM-6360LV).

Transmission electron microscopy
Ovarian follicles were isolated from both dsDnmt1-and dsCK-treated female adults at the end of the choriogenic period and fixed in fixative for transmission electron microscopy (TEM; Servicebio, G1102) for at least 2 h.After rinsing three times with PBS, the samples were treated with 1% osmium tetroxide (Ted Pella) in 0.1 M PBS (pH 7.4) at room temperature for 2 h.The osmium tetroxide was removed by rinsing in 0.1 M PBS three times.The tissues were dehydrated with increasing concentrations of ethanol at 15 min intervals and infiltrated with increasing concentrations of EMbed 812.Ultrathin sections (60-80 nm) were cut with an ultramicrotome.The sections were stained with uranyl acetate and lead citrate in a stepwise manner.Finally, the sections were air-dried overnight at room temperature and then observed with a transmission electron microscope (Hitachi, HT7700) 55 .

RNA-Seq and data analysis
Ovaries at day 7 (d7) and day 9 (d9) PAE that had been treated with dsRNAs were dissected.Three biological replicates were tested for each treatment.The dsRNA injections were performed on days 2, 4, and 6 PAE (three times), and ovaries from six individuals were dissected and mixed to form one sample.A total of 1 μg of RNA per sample was used as input material for RNA preparation.Sequencing libraries were generated using an NEBNext Ultra TM RNA Library Prep Kit for Illumina (NEB, E7770S) following the manufacturer's recommendations.The clustered libraries were sequenced on an Illumina platform, and paired-end reads were generated.Raw FASTQ-format data were first processed through in-house Perl scripts.HISAT2 software was used to map the reads to the reference genome (B.germanica Bger_1.1).Gene expression levels were estimated as fragments per kilobase of transcript per million fragments mapped (FPKM) values, and differential expression analysis between the two groups was performed using DESeq.Genes with adjusted P values < 0.05 in DESeq were considered as differentially expressed.KEGG pathway enrichment analysis was performed for the differentially expressed genes using the R package clusterProfiler v4.8.3.

20E level detection
Fresh ovaries were dissected on day 9 PAE and extracted with 200 μl of 100% methanol.The tissues were homogenized and centrifuged at 13000 × g for 5 min, and the supernatant was collected in a new tube.The remaining precipitate was resuspended with another 200 μl of methanol and centrifuged again to collect the supernatant 56 .The supernatants were mixed together and air-dried with nitrogen to collect the 20E.The 20E concentration was detected with 20E ELISA kit according to the manufacturer's instructions (Mlbio, ml062795).
For the nutrition-related components (insulin, Aa mixture, and glucose) treatment, bovine insulin (Yuanye Bio-Technology, 11070-73-8) was solubilized in DMSO at a concentration of 25 μg/μl; the Aa mixture (type H) (Wako, 013-08391) was diluted with PBS solution into 12.5 μg/μl; the glucose (Beyotime, ST1024) was diluted with PBS into 25 μg/μl.2 μl of different kinds of components and their corresponding solvents were injected into the abdomen of female adults on day 5 PAE.All injected females were paired with wild-type male adults for mating.The eggs were collected at 84 h post oviposition.
In brief, six pairs of BSP primers were designed using the online MethPrimer software and were listed in Table S3.One microgram of genomic DNA was converted using a ZYMO EZ DNA Methylation-Gold Kit (Zymo, D5005), and one-twentieth of the elution product was used as the template for PCR amplification with 35 cycles using a KAPA HiFi HotStart Uracil + ReadyMix PCR Kit (Kapa, KK2801).For each sample, the BSP products of multiple DMRs were pooled equally, 5′-phosphorylated, 3′-dA-tailed, and ligated to a barcoded adapter using T4 DNA ligase (NEB, M0202S).The barcoded libraries from all samples were sequenced on the Illumina platform.

Statistics and reproducibility
The hatchability experiment of B. germanica was repeated three times, with 15 individuals per group treated.The mass spectrometry and quantitative experiments had three biological replicates involving 12 individuals.The data statistics in the paper were analyzed by the software SPSS 25, and the specific analysis method is shown in figure legends.

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

Fig. 4 |
Fig.4| ftz-f1 promoter undergoes 5mC modification thus preventing prolonged expression.a 5mC level changes of the ftz-f1 in the genome detected by WGBS.The blue and red lines indicate the 5mC tendencies for the dsCK and dsDnmt1 groups, respectively.b 5mC level decreasing at the promoter region of ftz-f1 was detected by NGS-BSP.Six amplicons were used and three replicates for each group were performed.c Boxplot results of temporal 5mC levels in the promoter region of ftz-f1 during the first reproductive period (D1-D8) detected by NGS-BSP.Six amplicons were selected and three replicates for each group were performed.The median, maximum, and minimum values are shown.d Gene expression patterns of ftz-f1 in the ovaries during the first reproductive period were detected by qRT-PCR, data are mean + sd, n = 3 biologically independent samples.Source data are provided as a Source Data file.

Fig. 5 |
Fig. 5 | dsDnmt1-induced phenomena can be partially rescued by ftz-f1 knockdown during choriogenesis.a ftz-f1 relative expression level under separate or combined injection of dsDnmt1 and dsftz-f1.b 20E level detection under separate or combined injection of dsDnmt1 and dsftz-f1.c Statistics of oviposit of female adults under separate or combined injection of dsDnmt1 and dsftz-f1.d DAPI and PHA staining for nuclei and actin protein in ovarian follicle cells under separate or combined injection of dsDnmt1 and dsftz-f1.e SEM observation of sponge-like bodies on eggs under separate or combined injection of dsDnmt1 and dsftz-f1, 9 of 17 sponge-like bodies were observed normal.Data are mean ± sd, different letters indicate statistically significant differences between groups using one-way ANOVA and LSD multiple comparisons test, P < 0.05, n = 3 biologically independent samples.Source data are provided as a Source Data file.

Fig. 6 |
Fig. 6 | 5mC modification prevents ftz-f1 induced anoikis during cockroach choriogenesis.a TEM observation of ovarian sections.TP: tunica propria; Oo: oocyte; FC: follicle cell.Two different magnifications, ×700 and ×1500, were selected.The asterisks indicate the cytomembrane detachment between follicle cells.b Apoptosis-associated Caspase-3 staining.Z-VAD means Pan Caspase inhibitor Z-VAD-FMK.c Mmp1 relative expression level under separate or combined injection of dsDnmt1 and dsftz-f1.The difference was analyzed by one-way ANOVAand LSD multiple comparisons test.Data are mean ± sd, different letters indicate statistically significant differences between groups, P < 0.05, n = 3 biologically independent samples.d The gene expression levels of five chorion genes detected by qRT-PCR at both d7 and d9 PAE.Data are mean ± sd, the significant difference was analyzed by two-tailed Student's t test, n = 3 biologically independent samples.Source data are provided as a Source Data file.

Fig. 7 |
Fig. 7 | Schematic view of the regulatory effect of DNMT1 on reproduction in B. germanica.a In wild-type females (left panel), DNMT1-mediated DNA methylation in the promoter region of the ftz-f1 was able to repress ftz-f1 expression at suitable levels during choriogenesis.At its normal expression level, ftz-f1 plays important roles in ovarian development, including the maturity of sponge-like bodies.b In dsDnmt1-treated females (right panel), hypomethylation-induced ftz-f1 upregulation impairs choriogenesis and disrupts fertilization by causing aberrantly high levels of 20E titer and 20E signaling, excessive follicle cell dissociation, and a shortage of chorion proteins.