The MicroRNA Ame-Bantam-3p Controls Larval Pupal Development by Targeting the Multiple Epidermal Growth Factor-like Domains 8 Gene (megf8) in the Honeybee, Apis mellifera

20-Hydroxyecdysone (20E) plays an essential role in coordinating developmental transitions in insects through responsive protein-coding genes and microRNAs (miRNAs). However, the interplay between 20E and miRNAs during insect metamorphosis is unknown. In this study, using small RNA sequencing, a comparative miRNA transcriptomic analysis in different development stages, and 20E treatment, we identified ame-bantam-3p as a key candidate miRNA involved in honeybee metamorphosis. Target prediction and in vitro dual-luciferase assays confirmed that ame-bantam-3p interacts with the coding region of the megf8 gene and promotes its expression. Meanwhile, temporal expression analysis revealed that the expression of ame-bantam-3p is higher in the larval stage than in prepupal and pupal stages, and that this expression pattern is similar to that of megf8. In vivo, we found that the mRNA level of megf8 was significantly increased after the injection of ame-bantam-3p agomir. A 20E feeding assay showed that 20E downregulated the expression of both ame-bantam-3p and its target gene megf8 on larval days five, six, and seven. Meanwhile, the injection of ame-bantam-3p agomir also reduced the 20E titer, as well as the transcript levels of essential ecdysteroid synthesis genes, including Dib, Phm, Sad, and Nvd. The transcript levels of 20E cascade genes, including EcRA, ECRB1, USP, E75, E93, and Br-c, were also significantly decreased after ame-bantam-3p agomir injection. However, ame-bantam-3p antagomir injection and dsmegf8 injection showed the opposite effect to ame-bantam-3p agomir injection. Ame-bantam-3p agomir treatment ultimately led to mortality and the failure of larval pupation by inhibiting ecdysteroid synthesis and the 20E signaling pathway. However, the expression of 20E signaling-related genes was significantly increased after megf8 knockdown, and larvae injected with dsmegf8 showed early pupation. Combined, our results indicate that ame-bantam-3p is involved in the 20E signaling pathway through positively regulating its target gene megf8 and is indispensable for larval–pupal development in the honeybee. These findings may enhance our understanding of the relationship between 20E signaling and small RNAs during honeybee development.


Introduction
MicroRNAs (miRNAs or miRs) constitute a class of endogenous, non-coding, singlestranded RNAs ranging from 21 to 24 nucleotides in length. They can regulate gene expression post-transcriptionally through binding to target sites in the 5 untranslated region (UTR), coding region, or 3 UTR of target mRNAs [1,2]. The role of miRNAs in the regulation of gene expression was initially thought to be repressive; however, it has been shown that miRNAs also play a positive regulatory role in gene expression [3,4]. For example, miR-466l binds to AU-rich elements in IL-10 mRNA, which in turn increases the expression of IL- 10 [5]. During amino acid starvation, miR-10a binding to the 5 end molecular mechanisms involved in the biogenesis and functions of 20E-regulated miRNAs and may also serve to expand our understanding of how miRNAs participate in insect metamorphosis.

Analysis of Small RNA Deep-Sequencing Data
To investigate the relationship between 20E signaling pathways and small RNAs during honeybee development and metamorphosis, small RNA libraries were constructed for sequencing from pooled samples of five-day-old larvae, six-day-old larvae, sevenday-old larvae, eight-day-old larvae, and Pw after 20E treatment (see Methods; Table S1). Between 13.9 and 35.9 million single-end raw reads were obtained from the 29 small RNA libraries generated from the developmental stages analyzed after 20E treatment. After removing low-quality reads, adaptor sequences, and sequences with lengths of <18 nt and >30 nt, approximately 9.7 to 16.0 million clean reads were obtained and were used to predict known and novel miRNAs. The Q30 percentage per individual ranged from 95.76% to 98.49% (Table S2). rRNAs, tRNAs, snRNAs, snoRNAs, other ncRNAs, and repeat sequences were filtered out, and 43.98% to 76.41% of genome-matched reads were screened out from the total reads (Table S3). The lengths of mature miRNAs ranged from 18 to 30 nt and one peak at approximately 22 nt was observed for all genome-matched reads, which is considered the classical length of miRNAs. Furthermore, 23.55% to 70.83% of the clean 22-nt reads were predicted to be miRNAs ( Figure S1, Table S4). For the 10 libraries, both known and novel miRNAs with longer sequences exhibited clear uridine (U) bias at the first nucleotide. Meanwhile, analysis of the frequency of bases at each position in all the miRNA reads showed that U and guanine (G) occurred more frequently than adenine (A) and cytosine (C) ( Figure S2, Tables S5 and S6).

Identification of Differentially Expressed miRNAs
First, using miRBase (v22) and the reference honeybee genome (Apis_mellifera.Amel_4.5_ ncbi), we identified 297 miRNAs in the library from the five-day-old larvae, 227 in the library from the six-day-old larvae, 264 in the library from the seven-day-old larvae, 253 in the library from the eight-day-old larvae, and 265 in the library from Pw (Table S7). After the removal of redundant reads, a total of 424 miRNAs were obtained from all samples, 244 of which were predicted to be known miRNAs and 180 novel miRNAs (Tables S8 and S9). To understand the functional characteristics of these predicted miRNAs, we performed an analysis of miRNA families to identify which miRNAs were highly conserved. Some of the miRNAs were found to belong to families, such as the miR-9, miR-252, miR-279, and bantam families, among others (Table S10), suggesting that they may have conserved roles in the regulation of metamorphosis in honeybees.
The expression levels of miRNAs in each sample were calculated and normalized based on the TPM algorithm [43]. To reveal the functions of the identified miRNAs during honeybee metamorphosis, we compared miRNA expression levels between the 20E and control treatment groups using DESeq2 with |log2(FC)| ≥ 0.58 and p-value ≤ 0.05 as the significance cutoffs [44]. Pearson's correlation analysis showed that there was a high degree of correlation among the samples in each group, indicating that the sampling was reasonable ( Figure S3). Although their expression was detected in all libraries, the expression levels of several miRNAs were significantly altered at different developmental stages after 20E treatment. In five-day-old larvae, 39 miRNAs showed significant responses, including 19 that were upregulated and 20 that were downregulated. In six-day-old larvae, 26 miRNAs showed a significant response to 20E, 19 of which were significantly upregulated and 20 significantly downregulated. In seven-day-old larvae, 28 miRNAs displayed a significant response to 20E, with 13 exhibiting significant upregulation and 15 significant downregulation. However, 11 miRNAs in eight-day-old larvae showed a significant response (three upregulated, eight downregulated), four miRNAs showed a significant response in white eye pupae (two upregulated, two downregulated showed Table 1. miRNA count among different tissues.
involved in the larval to pupal transition of honeybees. Venn diagram analysis showed that the levels of ame-bantam-3p, miR-279b-3p, and miR-6001-5p were significantly decreased in five-day-old, six-day-old, and seven-day-old larvae after 20E treatment, indicating that these miRNAs were negatively regulated by 20E.Ame-bantam-3p was found to be the most abundant miRNA in all three developmental stages. Meanwhile, the expression of some miRNAs (e.g., ame-miR-7-5p, ame-miR-13b-3p, ame-miR-993-3p, and ame-miR-263a-5p) was upregulated in five-day-old and six-day-old larvae after 20E treatment ( Figure 2F). The expression patterns of these miRNAs suggested that they may play a significant role during larval-pupal transformation and metamorphosis in the honeybee and that 20E regulates the expression of metamorphosis-associated miRNAs. Based on all the above results, we subsequently focused on ame-bantam-3p, and we examined its function during honeybee metamorphosis.
To ensure the reliability of the small RNA deep-sequencing data and compare the expression profiles of some of the miRNAs among the different developmental stages after 20E treatment, we selected eight differential expression (DE) miRNAs that were highly expressed in one of the five developmental stages and confirmed their expression levels via RT-qPCR. As shown in the histogram in Figure 3, the expression trends of all the tested miRNAs were consistent with the small RNA deep sequencing data, suggesting that the latter were highly reliable.  Hierarchical clustering analysis based on the TPM method showed that larvae of the same day old treated with 20E were clustered in the same clade and separated from the respective controls (Figure 2A-E). Furthermore, some miRNAs, such as ame-bantam-3, miR-276b-3p, ame-let-7-5p, and ame-miR-282-5p, were differentially expressed at different developmental stages ( Figure S4), indicating that these miRNAs may be involved in the larval to pupal transition of honeybees. Venn diagram analysis showed that the levels of ame-bantam-3p, miR-279b-3p, and miR-6001-5p were significantly decreased in fiveday-old, six-day-old, and seven-day-old larvae after 20E treatment, indicating that these miRNAs were negatively regulated by 20E.Ame-bantam-3p was found to be the most abundant miRNA in all three developmental stages. Meanwhile, the expression of some miRNAs (e.g., ame-miR-7-5p, ame-miR-13b-3p, ame-miR-993-3p, and ame-miR-263a-5p) was upregulated in five-day-old and six-day-old larvae after 20E treatment ( Figure 2F). The expression patterns of these miRNAs suggested that they may play a significant role during larval-pupal transformation and metamorphosis in the honeybee and that 20E regulates the expression of metamorphosis-associated miRNAs. Based on all the above results, we subsequently focused on ame-bantam-3p, and we examined its function during honeybee metamorphosis.   To ensure the reliability of the small RNA deep-sequencing data and compare the expression profiles of some of the miRNAs among the different developmental stages after 20E treatment, we selected eight differential expression (DE) miRNAs that were highly expressed in one of the five developmental stages and confirmed their expression levels via RT-qPCR. As shown in the histogram in Figure 3, the expression trends of all the tested miRNAs were consistent with the small RNA deep sequencing data, suggesting that the latter were highly reliable.

Prediction and Functional Enrichment Analysis of the Target Genes of DE miRNAs
To characterize the regulatory roles of the identified miRNAs during larval-pupal

Prediction and Functional Enrichment Analysis of the Target Genes of DE miRNAs
To characterize the regulatory roles of the identified miRNAs during larval-pupal transformation and metamorphosis in the honeybee, the potential targets of the miRNAs were predicted using miRanda [45] and targetScan [46] software. A total of 161 target genes were predicted for 43 DE miRNAs between control and 20E in five-day-old larvae, and these were successfully annotated against the COG, NR, Swiss-Prot, GO, KEGG, Pfam, and eggNOG databases (Table S11). KEGG pathway enrichment analysis showed that the target genes of the DE miRNAs were mostly associated with ECM-receptor interaction, Hippo signaling pathway, MAPK signaling pathway, Dorso-ventral axis formation, lysine degradation, and FoxO signaling pathway, suggesting that these genes may play important roles in honeybee metamorphosis ( Figure 4). The most abundant terms relating to biological process, cellular component, and molecular function in the GO term enrichment analysis of the predicted target genes are shown in Figure 5. The most significantly enriched terms in the biological process category included homophilic cell adhesion via plasma membrane adhesion molecules, instar larval or pupal development, muscle attachment, PCD involved in cell development, and skeletal muscle organ development. For the cellular component category, the most significantly enriched terms included axoneme, collagen trimer, dynein complex, mitochondrial matrix, and plasma membrane-bounded cell projection cytoplasm. Meanwhile, the most significantly enriched terms in the molecular function class included act in binding, calcium ion binding, extracellular matrix structural constituent, molecular function, and transcription regulator activity. and these were successfully annotated against the COG, NR, Swiss-Prot, GO, KEGG, Pfam, and eggNOG databases (Table S11). KEGG pathway enrichment analysis showed that the target genes of the DE miRNAs were mostly associated with ECM-receptor interaction, Hippo signaling pathway, MAPK signaling pathway, Dorso-ventral axis formation, lysine degradation, and FoxO signaling pathway, suggesting that these genes may play important roles in honeybee metamorphosis ( Figure 4). The most abundant terms relating to biological process, cellular component, and molecular function in the GO term enrichment analysis of the predicted target genes are shown in Figure 5. The most significantly enriched terms in the biological process category included homophilic cell adhesion via plasma membrane adhesion molecules, instar larval or pupal development, muscle attachment, PCD involved in cell development, and skeletal muscle organ development. For the cellular component category, the most significantly enriched terms included axoneme, collagen trimer, dynein complex, mitochondrial matrix, and plasma membrane-bounded cell projection cytoplasm. Meanwhile, the most significantly enriched terms in the molecular function class included act in binding, calcium ion binding, extracellular matrix structural constituent, molecular function, and transcription regulator activity.

In Vitro Validation of Ame-Bantam-3p Targets
To identify the target genes of ame-bantam-3p, two miRNA target-prediction software programs, miRanda and TargetScan, were simultaneously employed, leading to the identification of only one putative target gene, megf8. Based on the predicted results, miRNA/mRNA combinations with low energy and high score were selected. The putative target sites in the megf8 coding sequence and the free energy value between ame-bantam-3p and megf8 mRNA is shown in Figure 6A. In addition, ame-bantam-3p is highly conserved in insects and the honeybee ( Figure 6B). To further assess the relationship between ame-bantam-3p and its target gene, we conducted an in vitro dual-luciferase reporter assay. The approximately 400 bp fragment, which covered the bantam-3p binding site in megf8, was cloned into a luciferase reporter plasmid and co-transfected into HEK-293 cells with the ame-bantam-3p mimic. A null target sequence upstream of the luciferase coding sequence served as a NC. The results showed that luciferase reporter

In Vitro Validation of Ame-Bantam-3p Targets
To identify the target genes of ame-bantam-3p, two miRNA target-prediction software programs, miRanda and TargetScan, were simultaneously employed, leading to the identification of only one putative target gene, megf8. Based on the predicted results, miRNA/mRNA combinations with low energy and high score were selected. The putative target sites in the megf8 coding sequence and the free energy value between ame-bantam-3p and megf8 mRNA is shown in Figure 6A. In addition, ame-bantam-3p is highly conserved in insects and the honeybee ( Figure 6B). To further assess the relationship between amebantam-3p and its target gene, we conducted an in vitro dual-luciferase reporter assay. The approximately 400 bp fragment, which covered the bantam-3p binding site in megf8, was cloned into a luciferase reporter plasmid and co-transfected into HEK-293 cells with the ame-bantam-3p mimic. A null target sequence upstream of the luciferase coding sequence served as a NC. The results showed that luciferase reporter activity was increased by approximately 41% when the pGL3-CMV-LUC-megf8 (LOC725895)-CDS was co-transfected with the ame-bantam-3p mimic compared with when megf8 was co-transfected with the NC ( Figure 6C). Together, these results confirmed that there is a functional target of amebantam-3p in the coding region of megf8, and megf8 is a target gene of am-bantam-3p.
(knockdown) into six-day-old larvae. Since neither NC-agomir nor NC-antagomir had a significant effect on the larvae after injection in all measured indices, we used NC-agomir as a control for further analysis in the following experiments. As shown in Figure 6D, the ame-bantam-3p expression level was significantly upregulated in worker bee larvae injected with agomir, whereas antagomir treatment resulted in a significant downregulation of ame-bantam-3p expression. The results showed that, compared with the control group, the mRNA level of megf8 was increased by approximately 86% following ame-bantam-3p agomir injection, whereas antagomir injection resulted in a 50% decrease in megf8 expression levels ( Figure 6E), which was consistent with the results of the dual-luciferase assay. These findings further indicated that ame-bantam-3p positively regulates megf8 expression in vivo by binding to the megf8 coding region. UTR NC represents a reporter without the target sequence and mimics NC represents mimic controls. The mean ± SEM of the relative luciferase expression ratio (firefly luciferase/Renilla luciferase, Luc/R-luc) was calculated from three biological replicates (Student's t-test, n = 3) and compared with the negative control (NC). The expression level of ame-bantam-3p (D) and megf8 (E) in larvae injected with ame-bantam-3p agomir, ame-bantam-3p antagomir or NC. Reverse transcription quantitative polymerase chain reaction (RT-qPCR) data are presented as means ±SEM of triplicate samples. * p < 0.05, ** p < 0.01 (Student's t-test). (C) Relative luciferase activity in the dual-luciferase assay. UTR NC represents a reporter without the target sequence and mimics NC represents mimic controls. The mean ± SEM of the relative luciferase expression ratio (firefly luciferase/Renilla luciferase, Luc/R-luc) was calculated from three biological replicates (Student's t-test, n = 3) and compared with the negative control (NC). The expression level of ame-bantam-3p (D) and megf8 (E) in larvae injected with ame-bantam-3p agomir, ame-bantam-3p antagomir or NC. Reverse transcription quantitative polymerase chain reaction (RT-qPCR) data are presented as means ±SEM of triplicate samples. * p < 0.05, ** p < 0.01 (Student's t-test).
To further determine whether ame-bantam-3p regulates Megf8 expression in vivo, we injected either ame-bantam-3p agomir (overexpression) or ame-bantam-3p antagomir (knockdown) into six-day-old larvae. Since neither NC-agomir nor NC-antagomir had a significant effect on the larvae after injection in all measured indices, we used NC-agomir as a control for further analysis in the following experiments. As shown in Figure 6D, the ame-bantam-3p expression level was significantly upregulated in worker bee larvae injected with agomir, whereas antagomir treatment resulted in a significant downregulation of ame-bantam-3p expression. The results showed that, compared with the control group, the mRNA level of megf8 was increased by approximately 86% following ame-bantam-3p agomir injection, whereas antagomir injection resulted in a 50% decrease in megf8 expression levels ( Figure 6E), which was consistent with the results of the dual-luciferase assay. These findings further indicated that ame-bantam-3p positively regulates megf8 expression in vivo by binding to the megf8 coding region.

Ame-Bantam-3p and megf8 Are Co-Expressed in the Larval-Pupal Stages
To investigate the potential function of ame-bantam-3pduringhoneybee metamorphosis, the temporal expression levels of ame-bantam-3p and megf8 in the larval-pupal stages were assessed by RT-qPCR. Our results showed that ame-bantam-3p expression rapidly increased during the larval stages, peaking at larval day 5, and then it gradually declined. At the prepupal stage (Pp2 to Pp3), ame-bantam-3p expression was almost undetectable ( Figure 7A). In the larval stage, the mRNA level of megf8 gradually increased, followed by a decrease after the prepupal stage, which was consistent with the ame-bantam-3p expression pattern. Focusing on five-day-old larvae, we measured ame-bantam-3p and megf8 transcript levels in the midgut, fat body, head, epidermis, and silk gland. The results showed that ame-bantam-3p was ubiquitously expressed in all investigated tissues, with the highest relative expression levels being observed in the head, epidermis, and fat body. Similar results were seen for megf8 expression ( Figure 7B,C). The consistency in expression mode between ame-bantam-3p and megf8, both in the larval-pupal stages and isolated tissues, showed that ame-bantam-3p might be involved in regulating the development of worker bee larvae by positively regulating its target gene megf8.

Ame-Bantam-3p and megf8 Are Co-Expressed in the Larval-Pupal Stages
To investigate the potential function of ame-bantam-3pduringhoneybee metamorphosis, the temporal expression levels of ame-bantam-3p and megf8in the larvalpupal stages were assessed by RT-qPCR. Our results showed that ame-bantam-3p expression rapidly increased during the larval stages, peaking at larval day 5, and then it gradually declined. At the prepupal stage (Pp2 to Pp3), ame-bantam-3p expression was almost undetectable ( Figure 7A). In the larval stage, the mRNA level of megf8gradually increased, followed by a decrease after the prepupal stage, which was consistent with the ame-bantam-3p expression pattern. Focusing on five-day-old larvae, we measured amebantam-3p and megf8 transcript levels in the midgut, fat body, head, epidermis, and silk gland. The results showed that ame-bantam-3p was ubiquitously expressed in all investigated tissues, with the highest relative expression levels being observed in the head, epidermis, and fat body. Similar results were seen for megf8 expression ( Figure  7B,C). The consistency in expression mode between ame-bantam-3p and megf8, both in the larval-pupal stages and isolated tissues, showed that ame-bantam-3p might be involved in regulating the development of worker bee larvae by positively regulating its target gene megf8.
Next, the levels of 20E were measured in the larval-pupal developmental stages. We found that the 20E titer was negatively correlated with the expression of ame-bantam-3 and megf8, remaining at a very low level in the larval stages and at a high level in the prepupal stage ( Figure 7D). These observations further confirmed that ame-bantam-3p might have a negative impact on 20E production in the honeybee by targeting the megf8 gene.

Ame-Bantam-3p Inhibits Ecdysteroid Production through Upregulating Target Gene megf8
To elucidate the effect of 20E oname-bantam-3p and its target gene megf8, we measured the expression levels of ame-bantam-3p and megf8 in five-day-old, six-day-old, Next, the levels of 20E were measured in the larval-pupal developmental stages. We found that the 20E titer was negatively correlated with the expression of ame-bantam-3 and megf8, remaining at a very low level in the larval stages and at a high level in the prepupal stage ( Figure 7D). These observations further confirmed that ame-bantam-3p might have a negative impact on 20E production in the honeybee by targeting the megf8 gene.

Ame-Bantam-3p Inhibits Ecdysteroid Production through Upregulating Target Gene megf8
To elucidate the effect of 20E oname-bantam-3p and its target gene megf8, we measured the expression levels of ame-bantam-3p and megf8 in five-day-old, six-day-old, and sevenday-old larvae fed exogenous 20E. Compared with the control group, the expression of ame-bantam-3p was downregulated by approximately 60%, 50%, and 34% in, respectively, five-day-old, six-day-old, and seven-day-old larvae fed 20E ( Figure 8A). Meanwhile, there was a significant decrease in megf8 expression levels between the 20E treatment group and the control group ( Figure 8B). This 20E feeding assay demonstrated that 20E can suppress the transcription of ame-bantam-3p and that of its target gene megf8. honeybee larvae. These observations constitute additional evidence that ame-bantam-3p and megf8 may be involved in larval-pupal development in the honeybee through the regulation of 20E levels.  That ame-bantam-3p and 20E have opposing effects on systemic growth suggests that the function of ame-bantam-3p during honeybee development is, at least in part, to inhibit ecdysteroid production. To test this possibility, we injected ame-bantam-3pagomir and ame-bantam-3p antagomir into the hemolymph of six-day-old larvae, respectively, and measured the circulating levels of 20E. We found that agomir-injected larvae failed to molt to white pupae ( Figure 8C), and their mortality rate was approximately 12% higher than that of NC-agomir-injected larvae ( Figure 8D). However, antagomir injection had no significant effect on larval mortality, but significantly reduced the pupation time of larvae ( Figure 8E). This may have been due to interference with ecdysteroid biosynthesis and 20E signaling, ultimately affecting larval pupation. Compared with following NC-agomir injection, the level of 20E was significantly increased in larvae 24 h after agomir injection, while 20E levels were significantly lower in antagomir-injected larvae ( Figure 8F). We further evaluated the expression levels of the Dib, Phm, Sad, Nvd, and Shd genes, which encode enzymes required for ecdysteroid biosynthesis, by RT-qPCR and found that the expression levels of Dib, Phm, Sad, and Nvd were lower in agomir-treated larvae than in larvae injected with NC-agomir, in contrast to that seen with antagomir treatment ( Figure 8G). This suggested that amebantam-3p could negatively affect ecdysteroid biosynthesis. 20E signaling was also lower in the agomir treatment group than in the NC-agomir treatment group, as determined by the expression levels of genes encoding the nuclear receptors ECRA, ECRB1, and USP and those of 20E primary response genes (i.e., the transcription factor-encoding genes Br-C, E75, and E93). In contrast, the expression levels of 20E signaling genes, including ECRA, ECRB1, USP, E93, E75, and Br-C were significantly increased in antagomir-injected larvae compared to NC-injected larvae ( Figure 8H).
We further assayed the proliferation of fat body cells in larvae injected with amebantam-3p agomir or am-bantam-3p antagomir. A greater number of proliferating fat body cells was observed in ame-bantam-3p agomir-injected larvae than in larvae treated with NC agomir, as detected by EdU staining, whereas there was no significant effect in the larvae injected with ame-bantam-3p antagomir ( Figure 8I). The levels of programmed cell death (PCD) were also detected using TUNEL staining in the fat body. Compared with NC agomir, the apoptosis was not observed at 24 h after ame-bantam-3p agomir injection ( Figure 8I), whereas apoptotic signals were greater in larval fat body cells after ame-bantam-3p antagomir injection. In summary, our findings suggested that high amebantam-3p levels in young larvae contribute to the maintenance of low 20E titers and the promotion of systemic growth, whereas reduced ame-bantam-3p activity in the prepupal stage contributes to the generation of the 20E peak and the cessation of growth.
To investigate the role of the target gene megf8 in the regulation of ecdysteroid synthesis by bantam-3p, we injected dsmegf8 into the hemocoel of six-day-old honeybee larvae and examined the expression levels of genes related to ecdysteroid synthesis. Compared with dsGFP injected larvae, mortality of larvae after dsmegf8 injection was not significantly affected, but it caused early pupation of larvae ( Figure 9A-C). The 20E titer and the expression levels of Dib, Phm, Sad, Nvd and Shd were significantly increased in dsmegf8injected larvae compared to dsGFP-injected larvae (Figures 9D and 8E). The expression of several key genes in the 20E-triggered transcriptional cascade was measured in dsmegf8injected larvae. Of the six genes tested, ECRA, ECRB1, USP, and E75 showed an 80-90% increase in mRNA levels compared to control larvae ( Figure 9F). These data suggest that bantam-3p inhibits ecdysteroid synthesis and interferes with 20E signaling by upregulating the expression of the target gene megf8 during metamorphosis in honeybee larvae. These observations constitute additional evidence that ame-bantam-3p and megf8 may be involved in larval-pupal development in the honeybee through the regulation of 20E levels.

Discussion
The steroid hormone 20E triggers significant developmental changes in insects, including molting and metamorphosis. The functions of hormones might be achieved through the interaction of genes involved in hormone signaling with miRNAs that control complex gene networks directing developmental decisions [47,48]. In this study, we investigated the complex relationship between 20E and miRNAs during honeybee metamorphosis by systematically examining differential miRNA expression after 20E treatment and at different developmental stages (from the larval to the pupal stage) and screening for conserved miRNAs. We identified the conserved ame-bantam-3p as a candidate miRNA potentially involved in honeybee metamorphosis, and we explored its role in the regulation of ecdysteroid synthesis and the 20E signaling pathway during larval-pupal development in the honeybee. In recent years, numerous studies have demonstrated that miRNAs can activate the expression of the target genes, such as that shown for miR-373, which can bind to the promoters of E-calmodulin and C2, a cold excitation domain-containing protein, thereby upregulating their expression [49]. In specific conditions, micoRNA-10a binds to ribosomal protein 5'-UTR to activate the translation of ribosomal proteins and ultimately promote total protein synthesis [6]. In the present study, the complete megf8 gene sequence was subjected to binding site analysis using target-prediction software, leading to the identification of ame-bantam-3p-binding sequences in the coding region of megf8. Using a luciferase assay, we further found that ame-bantam-3p positively regulates megf8 expression through its target site in the coding

Discussion
The steroid hormone 20E triggers significant developmental changes in insects, including molting and metamorphosis. The functions of hormones might be achieved through the interaction of genes involved in hormone signaling with miRNAs that control complex gene networks directing developmental decisions [47,48]. In this study, we investigated the complex relationship between 20E and miRNAs during honeybee metamorphosis by systematically examining differential miRNA expression after 20E treatment and at different developmental stages (from the larval to the pupal stage) and screening for conserved miRNAs. We identified the conserved ame-bantam-3p as a candidate miRNA potentially involved in honeybee metamorphosis, and we explored its role in the regulation of ecdysteroid synthesis and the 20E signaling pathway during larval-pupal development in the honeybee. In recent years, numerous studies have demonstrated that miRNAs can activate the expression of the target genes, such as that shown for miR-373, which can bind to the promoters of E-calmodulin and C2, a cold excitation domain-containing protein, thereby upregulating their expression [49]. In specific conditions, micoRNA-10a binds to ribosomal protein 5'-UTR to activate the translation of ribosomal proteins and ultimately promote total protein synthesis [6]. In the present study, the complete megf8 gene sequence was subjected to binding site analysis using target-prediction software, leading to the identification of ame-bantam-3p-binding sequences in the coding region of megf8. Using a luciferase assay, we further found that ame-bantam-3p positively regulates megf8 expression through its target site in the coding region of megf8 ( Figure 6B), while bioinformatic analysis and a dual-luciferase reporter assay confirmed the functionality of the target site. Studies have shown that antagomir treatment can markedly reduce miRNA expression levels in insects [50][51][52][53][54]. In this study, ame-bantam-3p agomir treatment upregulated ame-bantam-3p expression, whereas ame-bantam-3p antagomir exerted a significant inhibitory effect on the expression of bantam-3p. The expression level of megf8 was significantly upregulated in larvae 24 h after ame-bantam-3p agomir injection compared to NC-agomir-injected larvae, while ame-bantam-3pantagomir injection elicited the opposite effect. These results suggest that bantam-3p promotes the expression of megf8 by binding to its coding region.
In our study, temporal expression analysis indicated that ame-bantam-3p expression rapidly increased in the larval stage, peaked at five-day-old larval, and then gradually decreased, which was consistent with the megf8 expression pattern. Additionally, we measuredame-bantam-3p and megf8 transcript levels in the midgut, fat body, head, epidermis, and silk gland, and found that ame-bantam-3p was ubiquitously expressed in all investigated tissues, with the highest relative expression levels being observed in the head, epidermis, and fat body. Importantly, similar results were seen for megf8 expression (Figure 7). Combined, these findings suggested that ame-bantam-3p exerts an important function in larvaldevelopment. 20-Hydroxyecdysone (20E), a key steroid hormone, also regulates the expression of many miRNAs in insects. The transcriptional levels of miR-8-5p and miR-2a-3p in the planthopper Nilaparvatalugens [55]; miR-281 in B. mori [8]; and miR-8 and miR-14 in Drosophila melanogaster [56,57] were all reported to be inhibited in response to 20E treatment. Similarly, we found that the expression level of ame-bantam-3p was reduced in five-, six-, and seven-day-old larvae fed exogenous 20E. Meanwhile, we also measured the levels of 20E during larval-pupal developmental stages, and we found that the 20E titer was strongly and negatively correlated with ame-bantam-3p and megf8 expression, remaining at a very low level in the larval stages and at a high level in the prepupal stage ( Figure 7D). Using a 20E injection assay, we demonstrated that this hormone can repress the transcription of ame-bantam-3p and that of its target gene, megf8. Together, these results shed light on the molecular mechanism underlying the regulatory effect of 20E on ame-bantam-3p, as well as its pivotal role in insect development.
Some miRNAs are components of additional systemic signaling pathways in insects, including the 20E, insulin, stress, immune, and circadian pathways [41]. Bantam is involved in the regulation of insect metamorphosis by targeting spook and spookier in the ecdysteroid synthesis pathway [13,42]. Moreover, bantam promotes systemic growth by inhibiting the synthesis of the steroid hormone ecdysteroid [13]. In summary, we present evidence that both conserved and lineage-specific miRNAs participate in the regulation of metamorphosis in insects by directly controlling ecdysteroid biosynthesis. In our study, we found that the disruption of ecdysteroid biosynthesis through the use of ame-bantam-3p agomir led to a decrease in the expression levels of the dib, phm, sad, nvd, and shd genes, which encode enzymes required for ecdysteroid biosynthesis, whereas ame-bantam-3pantagomir injection increased the expression of these genes. Furthermore, 20E signaling was also reduced following ame-bantam-3pagomir treatment, as evidenced by the expression levels of genes coding for nuclear receptors (ECRA, ECRB1, and USP) and those of 20E primary response genes (i.e., the transcription factor-encoding genes Br-C, E75, and E93) (Figure 8). In addition, by employing RNA interference, we found that megf8 is also involved in the regulation of ecdysteroid synthesis and 20E signaling pathways. These results showed that the expression of these 20E response genes is dependent on the effects of ame-bantam-3p on megf8 expression. Meanwhile, ame-bantam-3p agomir treatment upregulated the expression of megf8 and resulted in pupation failure and the promotion of cell proliferation and, consequently, body growth, which blocked larval metamorphosis. Bantam represses apoptosis directly by inducing the expression of DIAP1 and indirectly by inducing the expression of Yki, which, in turn, suppresses the expression of the proapoptotic gene hid [16]. Additionally, we observed a greater number of proliferating fat body cells in ame-bantam-3p agomir-injected larvae than in larvae injected with NC agomir, as detected by EdU staining. The dMegf8 gene is essential for Drosophila viability and the loss of its function results in lethality during larval stages, and the phenotype of dMegf8 larval mutants is similar to that of some BMP signaling mutants, such as impaired formation of denticles and bristle [30]. In the present study, megf8 knockout caused early pupation of larvae and significantly increased the expression levels of genes involved in ecdysteroid synthesis and 20E signaling-related genes in honeybee larvae (Figure 9). Accordingly, our data suggest that, as in Drosophila, ame-bantam-3p is involved in larval development by inhibiting 20E release. Ame-bantam-3p plays an important role in positively regulating the levels of the megf8 gene, thus contributing to the homeostasis of 20E signaling pathway-associated genes during larval-pupal development in the honeybee. This study may provide novel insights into miRNA/gene interactions during metamorphosis in the honeybee.

In Vitrorearing of Honeybee Worker Larvae
One-day-old honeybee larvae were collected from colonies of the experimental apiary at Shandong Agricultural University (Tai'an, China). The queen bee was confined to a defined area on a comb containing empty cells for 12 h to control the age of the brood. Subsequently, the combs with newly laid eggs were moved to a separate place in the same colony using a queen excluder. After four days, the combs containing the synchronized two-day-old larvae (no more than 12 h old) were quickly transferred from the brood comb to 48-well culture plates and maintained in an incubator (temperature: 33 ± 1 • C; relative humidity: 55 ± 10%; in the dark). At the prepupal and pupal stages, the relative humidity was decreased to 80% and 70%, respectively. Daily diet volumes provided to the larvae and the approximate composition of the diet are detailed in Table 2. For feeding, 20E (Sigma, St. Louis, MO, USA) was dissolved in 100% ethanol (50 mg/mL) to prevent crystallization of the steroid and then diluted in Ringer solution containing 130 mM NaCl, 6 mM KCl, 4 mM MgCl 2 , 5 mM CaCl 2 , 160 mM sucrose, 25 mM glucose, and 10 mM HEPES (pH 6.7, 500 mOsmol) to give a final concentration of 10 mg/mL 20E. For the treatment group, larvae were fed a diet containing 20E (0.01 mg/mL). Control larvae were fed the normal diet with the corresponding doses of ethanol diluted in Ringer solution as the solvent control.

MicroRNA Sequencing
For small RNA sequencing, samples from five-day-old larvae, six-day-old (last-instar) larvae, seven-day-old larvae (prepupae), eight-day-old larvae (prepupae), and 10-day-old larvae (white-eye pupae, Pw) from different treatments were collected and, respectively, pooled. The samples were collected at least three times during different rearing cycles at each developmental stage and from each treatment group, thus serving as biological replicates. The mixed samples were immediately frozen in liquid nitrogen and stored at −80 • C for RNA extraction and small RNA library construction. Total RNA was extracted from each sample using TRIzol reagent (TaKaRa, Dalian, China) according to the manufacturer's instruction. RNA concentrations were measured using a Qubit RNA Assay Kit and a Qubit 2.0 Fluorometer (Life Technologies, Santa Clara, CA, USA). RNA integrity was assessed using an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA). Total RNA was stored at −80 • C for subsequent sequencing. RNA libraries were constructed and generated using a NEBNext Ultra II small RNA Sample Library Prep Kit for Illumina (NEB, San Diego, CA, USA) following the manufacturer's recommendations, then they were sequenced using the Illumina HiSeq.2500 Platform. Sequencing was undertaken by Biomarker Technologies Company (Beijing, China). Clean data (clean reads) were obtained by removing reads containing adapter, reads containing ploy-N, and low-quality reads from raw data. The clean reads were, respectively, mapped to the SILVA, GtRNAdb, Rfam, and Repbase databases. Sequence alignment was undertaken using the Bowtie tool [58] to filter out ribosomal RNAs (rRNAs), transfer RNAs (tRNAs), small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNAs), other ncRNAs, and repeats. The remaining reads were used to detect novel miRNAs predicted via a comparison with the honeybee genome (Apis_mellifera. Amel_4.5_ncbi) and known miRNAs from miRBase (v22) [59]. Randfold was used for the prediction of novel miRNA secondary structure. Differential expression analysis between two conditions/groups was performed using the "DESeq2" R package (1.10.1) [44], which provides statistical routines for determining differential expression in digital miRNA expression data using a model based on the negative binomial distribution. The resulting p-values were adjusted using the Benjamini and Hochberg approach for controlling the false discovery rate. MiRNAs with |log2(FC)| ≥ 0.58 and a p-value ≤ 0.05 were assigned as differentially expressed.

MiRNA Target Gene Prediction and Functional Analysis
Based on the sequences of the miRNAs (known and novel), miRanda [45] and tar-getScan [46] were used to identify the targets of miRNAs and run by default parameters. The final targets were intersected with the gene sets identified by two softwares. The sequences of the predicted target genes were compared with those of the non-redundant (NR) [60], Swiss-Prot [60], Gene Ontology (GO) [61], Clusters of Orthologous Groups of Proteins (COG) [62], Kyoto Encyclopedia of Genes and Genomes (KEGG) [63], Eukaryotic Orthologous Groups (KOG) [64], and Pfam [65] databases using BLASTX searches with an E-value cutoff of <1.00 × 10 −5 to obtain annotation information.

Quantitative PCR for miRNA and mRNA
Total RNA was extracted using TRIzol reagent (TaKaRa) following the manufacturer's instructions. First-strand cDNA for mRNA and miRNA was synthesized using, respectively, Evo M-MLV Premix and a miRNA 1st strand cDNA Synthesis Kit (Accurate Biotechnology Co., Ltd, ChangSha, China) according to the manufacturer's protocols. All the primers for RT-qPCR were designed and synthesized by Sangon Biotechnological Co. (Shanghai, China). RT-qPCR was conducted on a 7500 Real-time PCR System (ABI, USA) using TransStart® Top Green qPCR SuperMix (TransGen Biotech, Beijing, China) according to the manufacturer's instructions. The sequences of all the primers used for PCR are listed in Table 3. Each treatment had at least three biological repeats with three technical repeats. β-actin (GenBank accession no. XM_017065464) and U6 were used as the reference genes for mRNA and miRNA, respectively. Relative gene expression levels were calculated using the 2 −∆∆Ct method [66]. Table 3. The primer sequences used for qPCR and construction of luciferase reporter vector.

Cell Transfection and Dual-Luciferase Reporter Assay
The pGL3-CMV-LUC vector (Genomeditech, Shanghai, China) was used as a firefly luciferase reporter vector and the fragment tested covering the binding site of the miRNA to the target gene of roughly 400 bp were cloned downstream of the firefly luciferase gene. HEK-293 cells were seeded into 24-well plates with 500 µL of DMEM (Gibco, Carlsbad, CA, USA). At approximately 70% confluency, the cells were transfected as follows. First, 100 ng of plasmid DNA and 30 nM ame-bantam-3p mimics (5 -TGAGATCATTGTGAAAGCTGATT-3 ) (GeneBiogist, Shanghai, China) were dissolved in 50 µL of Opti-MEM (Invitrogen, Carlsbad, CA, USA), 3 µL of Lipofectamine (Invitrogen) was dissolved in 50 µL of Opti-MEM, and the two solutions were mixed together and incubated at room temperature for 30 min. At 48 h post-transfection, cell lysates were prepared and assayed for firefly luciferase activity using a Dual-Luciferase Reporter Assay Kit (Promega, Madison, WI, USA) according to the manufacturer's protocol. Luciferase activity was assessed using an Infinite M1000 microplate reader (Tecan, Switzerland). Each experiment was performed in triplicate. The mean of the relative firefly luciferase/Renilla luciferase expression ratio of the control was set to 1. The data were analyzed with twotailed t-tests.

Injection of Ame-Bantam-3p Agomir and Antagomir
Ame-bantam-3p and negative control (NC) agomir and antagomir were synthesized by Sangon Biotechnological Co. After dilution to 1 nmol/µL, 3 µL of miRNA agomir and antagomir or negative control (NC) agomir and antagomir were injected into the hemolymph of six-day-old larvae, respectively, using a microsyringe. A total of 24 larvae were injected with each miRNA agonist and antagomir. After 24 h, some of the larvae (four biological replicates) were dissected and stored at −80 • C for further analysis, and the remaining larvae were used for phenotypic analysis and determination of survival.

20Etiter Determination
To estimate 20E titers, 100 µL of hemolymph was collected from no less than 10 agomirinjected or antagomir-injected larvae 24 and 36 h post-injection. A 20E enzyme-linked immunosorbent assay (ELISA) kit (Meimian Industrial Co., Ltd., Jiangsu, China) was used to determine the 20E titer following the manufacturer's instructions. The experiment was performed with three biological replicates.

5-Ethynyl-2 -Deoxyuridine (EdU) Incorporation Assay
Cell proliferation assays were undertaken using the BeyoClick EdU Cell Proliferation Kit with Alexa Fluor 488 (Beyotime Biotechnology, ShangHai, China) according to the manufacturer's instructions. The solution-injected larvae were injected with 10 ngof EdU and the larval fat bodies were dissected out 12 h after injection. The samples were subsequently fixed in 4% paraformaldehyde for 12 h and were transversely sliced into 7-µm-thick sections on a freezing microtome. After washing three times in PBS, the sections were permeabilized with 0.5% Triton X-100 at room temperature for 15 min. The fixative was then removed, and the sections were washed first with PBS containing 3% BSA and then three times with PBS. Subsequently, the sections were incubated in Click Additive Solution in the dark and stained with DAPI. Fluorescence images of EdU incorporation were then obtained under Leica TCS SPE confocal microscope (Carl Zeiss, Oberkochen, Germany).

Terminal Deoxynucleotidyl Transferase-Mediated dUTP-Biotin Nick End Labeling (TUNEL) Assay
The fat bodies of injected worker bee larvae were harvested, fixed in 4% paraformaldehyde for 12 h, and then transversely sliced into 7-µm-thick sections on a freezing microtome. The One Step TUNEL Apoptosis Assay Kit from Beyotime was used for the detection of programmed cell death (PCD) in the tissues based on the manufacturer's instructions. The sections were subsequently incubated with 0.5% Triton X-100 on a shaker at room temperature for 15 min, washed three times in PBST, incubated with 50 µL of TUNEL inspection fluid for 1 h at 37 • C, and counterstained with DAPI at room temperature for 10 min. After washing three times with PBST, the stained sections were observed and imaged using Leica TCS SPE confocal microscope (Carl Zeiss, Oberkochen, Germany).

Statistical Analysis
All data were expressed as means ± standard error of the mean (SEM) derived from at least three biological repeats. Student's t-tests (two groups) or one-way ANOVA followed by Dunnett's post hoc test (three or more groups) was employed for comparisons using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA). Differences were considered significant at p < 0.05.

Conclusions
In conclusion, high ame-bantam-3p activity in young larvae was found to promote the expression of its target gene megf8, thus contributing to the maintenance of low 20E titers and the promotion of systemic growth. In contrast, reduced ame-bantam-3p activity contributed to the generation of the 20E peak, the cessation of growth, and entry into metamorphosis ( Figure 10).

Statistical Analysis
All data were expressed as means ± standard error of the mean (SEM) derived from at least three biological repeats. Student's t-tests (two groups) or one-way ANOVA followed by Dunnett's post hoc test (three or more groups) was employed for comparisons using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA). Differences were considered significant at p < 0.05.

Conclusions
In conclusion, high ame-bantam-3p activity in young larvae was found to promote the expression of its target gene megf8, thus contributing to the maintenance of low 20E titers and the promotion of systemic growth. In contrast, reduced ame-bantam-3p activity contributed to the generation of the 20E peak, the cessation of growth, and entry into metamorphosis ( Figure 10).