Proteolytic activity of the proteasome is required for female insect reproduction

Non-ATPase regulatory subunits (Rpns) are components of the 26S proteasome involved in polyubiquitinated substrate recognition and deubiquitination in eukaryotes. Here, we identified 15 homologues sequences of Rpn and associated genes by searching the genome and transcriptome databases of the brown planthopper, Nilaparvata lugens, a hemipteran rice pest. Temporospatial analysis showed that NlRpn genes were significantly highly expressed in eggs and ovaries but were less-highly expressed in males. RNA interference-mediated depletion of NlRpn genes decreased the proteolytic activity of proteasome and impeded the transcription of lipase and vitellogenin genes in the fat bodies and ovaries in adult females, and reduced the triglyceride content in the ovaries. Decrease of the proteolytic activity of the proteasome via knockdown of NlRpns also inhibited the transcription of halloween genes, including NlCYP307A2, NlCYP306A2 and NlCYP314A1, in the 20-hydroxyecdysone (20E) biosynthetic pathway in the ovaries, reduced 20E production in adult females, and impaired ovarian development and oocyte maturation, resulting in reduced fecundity. These novel findings indicate that the proteolytic activity of the proteasome is required for female reproductive processes in N. lugens, thus furthering our understanding of the reproductive and developmental strategies in insects.


Introduction
The brown planthopper, Nilaparvata lugens Stål (Hemiptera: Delphacidae), is one of the most threatening pests of rice crops in tropical Asia and southern China [1,2]. Nilaparvata lugens possess high fecundity, enabling intense population outbreaks in optimal environmental conditions [3]. Offspring production and fitness are directly influenced by oogenesis and ovulation in oviparous insects [4]. The mechanisms regulating fecundity in insects have attracted interest for decades, and many genes, including vitellogenin (Vg) and insulin associated with nutritional metabolism, and halloween genes involved in ecdysteroid biosynthesis, are known to play important roles in the regulation of oocyte maturation and embryonic development in most insect species [1,[5][6][7][8][9][10][11][12]. Progress has been achieved in understanding the molecular mechanisms governing female reproduction in N. lugens in recent years. The DICER1 gene, which is mostly responsible for microRNA precursor processing, was demonstrated to be important for oocyte maturation in the telotrophic ovary [13], while the female-specific gene for bicaudal-C was essential for oogenesis [14]. The Krüppel homologue 1 and Broad-Complex genes regulate ovarian development and jointly determine the number of ovarioles [15], while the mucin-like gene and eggshell-associated NlChP38 genes are specifically expressed in the follicular cells and are vital for ovulation [16,17]. We previously demonstrated the need for the pancreatic lipase-related protein 2 gene in oocyte maturation and development [18]. In addition to the characterized genes, other currently unknown genes may also be involved in fecundity. We previously sequenced the N. lugens genome and transcriptome in relation to developmental stage and tissue specificity [2,[19][20][21][22]. The subsequent annotated database enabled us to identify potential fecundity-related genes in this insect species. In the present study, we conducted a genome-and transcriptome-wide search and identified a group of genes encoding non-ATPase regulatory subunits and associated proteins of the 26S proteasome, which were highly expressed in the ovaries of adult females and in eggs laid in rice leaf sheaths.
The 26S proteasome is a large ATP-dependent multiprotein complex and the major proteolytic machine in eukaryotic cells. It functions primarily to maintain protein homeostasis by eliminating damaged or misfolded proteins via the ubiquitin-proteasome system [23]. Eukaryotic proteostasis plays a key role in most cellular processes, including DNA damage repair, transcription, signal transduction, cell cycle progression, apoptosis, ageing and disease [24,25]. The 26S proteasome is composed of the 19S regulatory particle (RP) and the 20S proteolytic core particle (CP). The CP has structurally and mechanistically well-characterized proteolytic activities [26], while the RP is responsible for the recognition, deubiquitylation, unfolding and translocation of substrates into the CP for proteolysis [27]. The RP can be separated into two sub-structures: a base complex that can associate with the CP, and a peripheral lid complex [24,28]. In budding yeast, the base comprises six ATPase regulatory subunits and three non-ATPase subunits (Rpns1, 2 and 13), and an additional cofactor Rpn10 [24,29], of which Rpn1 and Rpn2 are the two largest non-ATPase subunits that function as a scaffold, presumably stabilizing the RP [28,30]. Rpn10 and Rpn13 were reported to act as ubiquitin receptors to bind polyubiquitinated proteins [28]. The lid is composed of eight different Rpn subunits (Rpns3, 5-9, 11 and 12) and a small acidic protein, Sem1 [27]. The major function of the lid is to serve as a specialized isopeptidase coupling substrate deubiquitination [31]. Rpn11 is a MPN+/JAMM (JAB1/ MPN/Mov34 metalloenzyme)-domain-containing metalloprotease with deubiquitinase activity to remove ubiquitin chains from substrate proteins and to allow efficient protein degradation [27]. Rpn11 and its binding partner Rpn8 form a heterodimer in the lid, which is probably conducive to translocation-coupled substrate deubiquitination [32]. The Rpns3, 5-7, 9 and 12 subunits contain proteasome-COP9/signalosome-eIF3 (PCI) domains, and the Rpn15 subunit, a small acidic protein with neither PCI nor MPN domains, are thought to play scaffold roles in the lid [27,33,34]. Rpns have been extensively described in structural, biochemical and bioinformatics terms. The physiological functions for individual Rpns have been determined in yeast, Drosophila and Caenorhabditis elegans [31,[35][36][37][38]. For instance, C. elegans RPN-12 is required for reproduction [35]. Yeast Rpn11 has metalloisopeptidase activity [31]. Drosophila Rpn11 is a suppressor of progressive neurodegeneration [38]. However, systematic assessment of the in vivo physiological functions for Rpns has not been well studied in higher eukaryotes such as most insect species. We are interested in studying the biological significance of Rpns in the reproduction mechanisms in a Hemipteran model insect N. lugens because its whole genome sequence has been elucidated [22] and is susceptible to RNA interference (RNAi) [12,[39][40][41][42][43][44][45].
In this study, we identified 15 homologues genes of Rpn and associated chaperones in N. lugens by searching the genome and transcriptome databases. We investigated the functional roles of these genes in N. lugens using RNAi. Knockdown of the NlRpn genes notably reduced the proteolytic activity of the proteasome which then downregulated the transcript levels of several lipase and Vg genes in the fat bodies and the ovaries, decreased the triglyceride content of the ovaries, and resulted in remodelling of lipid droplets in the developing oocytes. In addition, decrease of the proteolytic activity via knockdown of NlRpns downregulated the transcript levels of halloween genes in the 20-hydroxyecdysone (20E) biosynthetic pathway in the ovary and reduced 20E synthesis throughout the body in adult females, leading to reproductive failure. Studies have established the 26S proteasome as an important regulator of transcription through proteolytic and non-proteolytic activities [46,47]. Genetic or chemical inhibition of proteasome function is known to result in significant changes in gene expression patterns [46][47][48][49]. Genes involved in mitochondrial function, stress response and protein degradation were upregulated, while ribosomal protein genes, mating genes and amino acid metabolism genes were downregulated in yeast [46]. In this study, our experiments revealed depletion of Rpns by RNAi significantly decreased the proteolytic activity of proteasome, which then affects the gene transcription during reproduction processes in N. lugens.

Bioinformatics analysis of Rpn and associated genes in Nilaparvata lugens
We identified 15 genes encoding Rpns and associated proteins by searching the N. lugens genomic and transcriptomic databases (electronic supplementary material, table S1). The deduced N. lugens amino acid sequences had high identities with their homologues in human, yeast and most insect species. The predicted domains were well conserved. Here, we follow the nomenclature strategy of human Rpn subunits to name the N. lugens homologues as NlRpn1-4, NlRpn6-8, NlRpn11-14 and NlADRM1 (table 1). To distinguish Rpn subunits and assembly chaperones, we name three N. lugens assembly chaperones as NlRac1, NlRac2 and NlRac5. The two largest proteins, NlRpn1 and NlRpn2, were composed of 1037 and 889 amino acid residues containing three and two characteristic proteasome/cyclosome repeat (PC) domains, respectively (figure 1a). NlRpn3, NlRpn6, NlRpn11, NlRpn12 and NlRpn13 consisted of 383-498 amino acids and shared a conserved PCI domain. NlRpn3 and NlRpn6 also contained a PCI/PINT-associated module (PAM) and an Rpn7 domain (26S proteasome subunit Rpn7, known as the Rpn6 in higher eukaryotes) (pfam10602), respectively. NlRpn4 contained a von Willebrand factor type A (vWFA) and three ubiquitininteracting motif (UIM) domains. NlRpn8 had only a COP9 signalosome/proteasome 26S non-ATPase subunit 8 Phylogenetic analysis of insect RPNs and assembly chaperones. The phylogenetic tree was constructed based on the deduced amino acid sequences of 165 RPNs and assembly chaperones from 12 insect species by the maximum-likelihood method using MEGA X (http://www.megasoftware.net/). Phylogenetic relationships were determined using the Jones- Taylor Rpn11 (MitMem region) at the C-terminus. NlRac1 was predicted to be a small protein, containing 226 amino acid residues and composed of five ankyrin repeat domains. NlRac2 was also a small protein, with 201 amino acid residues containing a PSD95-DLG-ZO1/2 (PDZ) domain. NlRac5 had a proteasome subunit beta type domain (PSMB: pfam10508). We constructed a phylogenetic tree based on the complete amino acid sequences available in the GenBank database in order to understand the evolutionary relationships among NlRpns and associated assembly chaperones and their homologues in other insect species. Phylogenetic analysis indicated that the insect Rpns and assembly chaperones were separated into individual branches (figure 1b). Rpn1 and Rpn2 had the characteristic PC domains, were clustered together and were closely related to Rpn4 and Rpn10/Rac1, and formed an independent cluster divergent from all other Rpns. The Rpn5/Rac5, Rpn9/Rac2 and Rpn8 branches were closely related, clustered together, and diverged from Rpn3 and Rpn11 branches with a conserved PCI domain. The Rpn7 and Rpn14 branches that shared JAB/MPN and MitMem domains were closely related. The Rpn6, Rpn12, and Rpn13 branches had a PCI domain and were clustered together. Most branches of N. lugens RPNs and assembly chaperones showed the closest phylogenetic relationships with homologues in Hemiptera such as Halyomorpha halys and Bemisia tabaci, but were distant from their homologues in Diptera, such as Aedes aegypti and Drosophila hydei. The nomenclature of the Rpn and assembly chaperone genes varies among organisms. The names of the Rpn genes in most insect species, except Drosophila melanogaster, are consistent with the standardized names of the human Rpns, which differ from the gene names in the yeast Saccharomyces cerevisiae and D. melanogaster (table 1). ScRpn13 in S. cerevisiae and DmRpn13 in D. melanogaster are homologous to ADRM1 in humans, which is a 19S RP subunit that does not follow the human PSMD naming strategy. In N. lugens, we name the homologue of ADRM1/ScRpn13/DmRpn13 as NlADRM1. NlRac1 and NlRac2, the RP assembly chaperones, are shown to be the homologues of PSMD10/ScNas6p/DmAnk2/ Rpn10 and PSMD9/ScNas2p/Deml_CG9588/Rpn9 in humans, yeast, Drosophila and most insect species. NlRac5 is the homologue of PSMD5/Deml_CG12096/Rpn5 in humans, Drosophila and most insect species, but is absent in yeast. Intriguingly, homologues of human and yeast Sem1/ HOD1 are lacking in most insects.

Temporospatial expression patterns of NlRpn and NlRac genes
We investigated the functional roles of the NlRpn and NlRac genes by examining their expression patterns throughout development and in different tissues using quantitative realtime polymerase chain reaction (qRT-PCR). These genes displayed similar developmental stage specificities, with the highest transcript levels in eggs and/or adult females and the lowest levels in adult males (figure 2a  . 1st, 2nd, 3rd, 4th and 5th refer to 1st-5th instar nymphs. One microgram of starting RNA was used for reverse transcription in a 20 µl reaction. Two microlitres of the first-strand cDNA (diluted 10 times) was analysed in each 20 µl reaction by qRT-PCR. The relative transcript levels of each gene in each developmental stage were normalized using the N. lugens 18S rRNA or β-actin threshold cycle (C t ) values that were obtained from reactions run on the same plate. In each assay, the transcript level was normalized to the lowest transcript level, which was arbitrarily set at one. (b) Tissue-specific expression analysis (in fold) of NlRpns and NlRacs. Total RNAs were extracted from the ovary, fat body, gut, salivary gland (SG), brain and cuticle of adult female N. lugens (n = 50-100). Relative transcript levels of the target genes in each tissue were determined by qRT-PCR as described above. Statistical analysis was performed using GRAPHPAD PRISM 8 software (San Diego, CA, USA). Three independent biological replicates (mean ± standard deviation) were conducted and relative transcript levels in each sample were measured using the ΔΔC t method. **p < 0.01 between indicated developmental stages or tissues (Student's t-test).
royalsocietypublishing.org/journal/rsob Open Biol. 11: 200251 levels in eggs than in adult females. The other Rpn transcripts, including NlRpn2, NlRpn3, NlRpn4, NlRpn8, NlRpn11 and NlRpn12 were expressed at similarly high levels in both eggs and adult females. These observations suggest that NlRpn and NlRac genes may have physiological functions in female reproduction and egg development. We subsequently investigated their tissue specificity in adult females. These genes, except NlRac2, displayed the highest transcript levels in ovaries and the lowest levels in fat bodies and cuticles among all the tested tissues ( figure 2b). Notably, NlRac2 showed the highest transcript levels in the brain followed by ovaries. These results further suggest that NlRpn and NlRac genes play vital roles in ovarian tissues in adult females, and may have additional functions in N. lugens.

Investigation of physiological functions of NlRpn and NlRac genes by RNA interference
We investigated the physiological functions of NlRpn and NlRac genes by silencing their expression in N. lugens in vivo using an RNAi approach. Newly emerged females were individually microinjected with each target double-stranded RNA (dsRNA) and phenotypic variations were observed at 24 h intervals following RNAi. Knockdown of NlRpn and NlRac1 transcription generated apparently lethal phenotypes. Following RNAi, greater than 70% of individuals in dsNlRpn1-, dsNlRpn2-, dsNlRpn3-, dsNlRpn4-, dsNlRpn6-, dsNlRpn7-, dsNlRpn8-, dsNlRac1-, dsNlRpn11-, dsNlRpn12-, dsNlRpn13or dsNlRpn14-injected group survived at 3 days post-injection (dpi), but these rates decreased at 4 dpi and declined dramatically to less than 40% at 11 dpi and to less than 20% at 14 dpi (figure 3a). By contrast, knockdown of NlRac5 and NlRac2 had no significant effect on the survival of N. lugens throughout the test period, similar to controls treated with double strand green fluorescent protein (dsGFP). Interestingly, adult females treated with dsNlRpns and dsNlRacs, except NlRpn4, NlRac5 and NlRac2, showed an abnormal phenotype from 7 dpi, with an obviously inflated abdomen and stretched intersegmental membranes in the tergum (figure 3b). Knockdown of NlRpns and NlRacs, except NlRpn4, NlRac5 and NlRac2, also significantly increased female body weight by approximately 29-48% compared with dsGFP-injected controls at 7 dpi (figure 3c). qRT-PCR analysis confirmed that transcript levels of each NlRpn or NlRac gene were notably reduced in RNAiinjected insects compared with those in dsGFP-injected controls at 3 dpi (figure 3d). Because NlRpn and NlRac genes were highly expressed in N. lugens ovaries, we investigated the ovarian phenotypes resulting from the knockdown of each NlRpn or NlRac in newly emerged adult females. The ovaries were dissected and observed on the 5th day after emergence. Adult females injected with dsNlRpn4, dsNlRac5 or dsNlRac2 showed normal ovaries with regular, banana-shaped mature oocytes, as in dsGFP-injected females (figure 5). By contrast, females injected with other dsNlRpns and dsRac1 displayed apparently abnormal ovaries, with no or malformed oocytes with irregular shapes. DsNlRac1-and dsNlRpn13-injected females had chubby, globose oocytes in the ovarioles, while dsNlRpn6-and dsNlRpn8-injected females had segmented ovarioles with malformed, round-shaped oocytes. Females injected with dsNlRpn1, dsNlRpn2, dsNlRpn3, dsNlRpn7, dsNlRpn11, dsNlRpn12 or dsNlRpn14 had no obviously segmented ovarioles in the ovaries, but had abnormal-looking immature oocytes containing large lipid droplets, loosely distributed in the ovarioles. The lateral oviducts of the ovaries also appeared milky white, compared with brown in dsGFP-injected females.

Investigation of lipid accumulation in ovaries
We monitored the morphological changes and distribution of lipid droplets in the ovaries of dsNlRpn-and dsNlRac-injected females using Nile red staining, which revealed abnormal sizes and uneven distribution of lipid droplets in oocytes after knockdown of NlRpn or NlRac transcript levels. dsGFPinjected ovaries included large numbers of small, round, evenly distributed lipid droplets in the oocytes (figure 6a). Uniformly distributed round lipid droplets with similar sizes were observed in the oocytes of females injected with dsNlRpn4, dsNlRac5 and dsNlRac2, while the lipid droplets in the immature oocytes were remodelled and obviously larger in the ovaries of females injected with dsNlRpn1, dsNlRpn2, dsNlRpn3, dsNlRpn6, dsNlRpn7, dsNlRpn8, dsNlRac1, dsNlRpn11, dsNlRpn12, dsNlRpn13 and dsNlRpn14. These large lipid droplets were surrounded by some mediumsized or small droplets. Interestingly, some immature oocytes in dsNlRpn1-and dsNlRpn2-injected females had lipid droplets distributed at the periphery of the oocytes. These observations implied that knockdown of most NlRpns and NlRac1 led to the remodelling of lipid droplets in oocytes. In addition, as shown in figure 6a, each oocyte is surrounded by a single layer of follicular epithelial cells with two nuclei in ovarioles in dsGFP-injected females, but the oocyte is surrounded by the disordered follicular cells that did not possess the typical two nuclei after knockdown of NlRpns and NlRacs, except for NlRpn4, NlRac5 and NlRac2, suggesting that most NlRpns royalsocietypublishing.org/journal/rsob Open Biol. 11: 200251 and NlRac1 were involved in the accumulation of lipid droplets in oocytes and the development of follicular cells.
Based on the observation of abnormal lipid droplets in oocytes in dsNlRpn-and dsNlRac-injected females, we measured the triglyceride content, as a major component of lipid droplets, in female ovaries and fat bodies. The triglyceride contents in dsGFP-injected ovaries on the 5th day after emergence were 1.2 nmol µg −1 total protein (figure 6b). RNAi of NlRpns and NlRacs, except NlRac5, significantly reduced the triglyceride contents of the ovaries by 24-57% in the case of dsNlRpn4, dsNlRac2, dsNlRac1, and dsNlRpn13-injected females, 67-74% in dsNlRpn6, dsNlRpn7 and dsNlRpn8-injected females, and 82-89% in dsNlRpn1, dsNlRpn2, dsNlRpn3, dsNlRpn11, dsNlRpn12 and

NlRpn1
NlRpn2 Figure 3. Phenotypic observation of RNAi targeting NlRpns and NlRacs. (a) Dynamic analysis of survival rates of N. lugens. Newly emerged females were injected with each dsNlRpn or dsNlRac, respectively, and observed for phenotypic variations at 24 h intervals for 14 days. dsGFP was injected as a negative control to determine the non-specific effects of dsRNA. Three independent biological replicates were conducted for each treatment (mean ± standard deviation; n = 60 females). (b) Morphology was observed on the 7th day after RNAi with each dsNlRpn or dsNlRac in the newly emerged females. Arrowheads indicate intersegmental membranes. Scale bar, 1.0 mm. (c) Body weights of females were measured on the 7th day after RNAi with each dsNlRpn or dsNlRac in the newly emerged females. Three independent biological replicates were conducted for each treatment (mean ± standard deviation; n = 20 females). **p < 0.01 between dsGFP and dsNlRpn or dsNlRac treatments (Student's t-test); ns: no significant difference. (d ) Determination of RNAi efficiency. Total RNAs were extracted from N. lugens on the 3rd day after RNAi and transcript levels of each NlRpn and NlRac were analysed by qRT-PCR, as described in figure 2. One microgram of each RNA sample was used for reverse transcription in a 20 µl reaction. Two microlitres of the first-strand cDNA (diluted 10 times) was analysed in each 20 µl reaction by qRT-PCR. The relative transcript levels of each NlRpn and NlRac gene in dsNlRpn, dsNlRac and dsGFP treatments were normalized using the N. lugens 18S rRNA or β-actin C t values. In each assay, the transcript level was normalized to the lowest transcript level, which was arbitrarily set at one. Results of triplicate experiments are shown with standard deviations. **p < 0.01 compared with dsGFP-injected controls (Student's t-test).  dsNlRpn14-injected females, compared with dsGFP-injected controls. The triglyceride content in the fat bodies in dsGFP-injected females was 0.9 nmol µg −1 total protein (figure 6c), and was not significantly affected by RNAi with any NlRpn or NlRac. These results suggest that knockdown of NlRpn and NlRac affected the accumulation of triglycerides in oocytes in the ovaries, but not in fat body cells.

Changes in lipase gene expression in ovaries and fat bodies
To understand whether the NlRpn and NlRac genes are involved in lipid metabolism in female ovaries and fat bodies, we investigated the effects of NlRpn and NlRac knockdown on lipase gene expression using qRT-PCR. Transcript levels of adipose triglyceride lipase (NlATGL), pancreatic triacylglycerol lipase 1 (NlPNLIP1), lipoprotein lipase-like 2 (NlLPL2) and gastric triacylglycerol lipase-like 1 (NlLIPF1) genes in the ovaries were markedly decreased by knockdown of NlRpn1-3, NlRpn6-8, NlRpn11-12 and NlRpn14 compared with dsGFPinjected controls (figure 7a). Knockdown with dsNlRpn4, dsNlRac5, dsNlRac2, dsNlRac1 and dsNlRpn13 notably reduced the transcript levels of NlATGL, but had no effect on transcript levels of NlPNLIP1. DsNlRpn4, dsNlRac1 and dsNlRpn13 significantly decreased the transcript levels of NlLPL2 and NlLIPF1, whereas dsNlRac5 and dsNlRac2 had no effect on the transcript levels of these two genes. These data suggest that knockdown of NlRpn and NlRac genes affects lipid-metabolizing gene transcription in female ovaries. In fat bodies, NlATGL, pancreatic lipase-related protein 2 (NlPLRP2), and pancreatic triacylglycerol lipase 2 (NlPNLIP2)  RNAi was performed by microinjecting each specific dsNlRpn or dsNlRac into the newly emerged female adults. Total RNAs were extracted from the whole bodies of the insects on the 5th day after RNAi (n = 10 adults). One microgram of starting RNA was used for reverse transcription in a 20 µl reaction. Two microlitres of the first-strand cDNA (diluted 10 times) was analysed in each 20 µl reaction by qRT-PCR. The relative transcript levels of each gene in each RNAi treatment with specific dsRNA were normalized using the N. lugens 18S rRNA or β-actin C t values as described in figure 2. In each assay, the transcript level was normalized to the lowest transcript level, which was arbitrarily set at one. Statistical analysis was performed using GRAPHPAD PRISM 8 software (San Diego, CA, USA). Three independent biological replicates (mean ± standard deviation) were conducted and relative transcript levels were measured using the ΔΔC t method. **p < 0.01, *p < 0.05 between dsGFP and a specific dsRNA treatments (Student's t-test). NlRpn and NlRac in the upper-left of each graph refer to the gene expression; dsNlRpn and dsNlRac of x-axis refer to specific dsRNA treatment. (b) Heatmap of transcript level variations of NlRpn and NlRac genes. Heatmap was plotted using GRAPHPAD PRISM 8 software based on the above qRT-PCR data from three biological replicates. NlRpn and NlRac of y-axis refer to the genes; dsNlRpn and dsNlRac of x-axis refer to specific dsRNA treatment. Red and blue refer to significantly downregulated and upregulated transcript levels of the genes after depleting specific NlRpn or NlRac gene expression. White refers to no significant change at the transcript levels.

Effect of RNA interference on expression of NlVg and NlVg-like genes in ovaries and fat bodies
Vg is the key factor in vitellogenesis in insects. We therefore investigated the effects of NlRpn and NlRac knockdown on expression levels of NlVg and NlVg-like gene transcription in female ovaries and fat bodies. qRT-PCR analysis indicated that dsNlRpns (except dsNlRpn4) and dsNlRac1 significantly downregulated NlVg and NlVg-like2 transcripts in both ovaries and fat bodies (figure 8a,b). Western blotting assay using an antibody against N. lugens Vg antigen prepared in our laboratory further confirmed these results at the protein level. Faint-specific protein bands were detected in the ovaries and fat bodies of dsNlRpn1-3, dsNlRpn6-8, dsNlRpn11-12 and dsNlRpn14-injected females, and stronger bands were detected in the ovaries of dsNlRpn4, dsNlRac5, dsNlRac2, dsNlRac1 and dsNlRpn13-injected females (figure 8c). These findings indicate that the changes in NlRpn-mediated Vg expression patterns were consistent with that of lipase genes. We subsequently explored the physiological function of the Vg  figure S1). These data suggest that NlRpns affect the expression of the NlVg gene but did not affect the expression of the NlVgR gene in female ovaries.

Effect of NlRpn and NlRac knockdown on ovulation and egg hatching
To further understand the functional roles of NlRpn and NlRac genes in reproduction, we investigated ovulation in adult females and hatchability of the eggs following NlRpn and NlRac knockdown in newly emerged adult females. Following mating with non-injected males, females that were treated with dsNlRpn1, dsNlRpn2, dsNlRpn3, dsNlRpn6, dsNlRpn7, dsNlRpn8, dsNlRpn11, dsNlRpn12 or dsNlRpn14 failed to lay eggs, suggesting that these genes were required for oocyte maturation and ovulation ( figure 9a)  and fat bodies (c) dissected from dsNlRpnor dsNlRac-injected females on the 5th day after emergence. dsGFP-injected samples were used as controls. Triglyceride content (nmol triglyceride μg −1 total protein) was determined from three independent biological replicates (mean ± standard deviation; n = 5). **p < 0.01 between dsNlRpn or dsNlRac and dsGFP treatments (Student's t-tests); ns: no significant difference.
royalsocietypublishing.org/journal/rsob Open Biol. 11: 200251 dsNlRpn13-injected females were significantly lower than those laid by dsGFP-injected controls (figure 9a). DsNlRpn4and dsNlRac1-injected females laid an average of 28 and 43 eggs, respectively, while dsNlRpn13-injected females laid an average of 0.1 eggs. By contrast, dsNlRac5-and dsNlRac2injected females laid 95 and 90 eggs, respectively, similar to dsGFP-injected females (83 eggs). Furthermore, dsNlRac1and dsNlRpn13-injected females laid eggs without an eye spot (eye pigmentation) (figure 9b), dsNlRpn4-injected females laid 89% of eggs without and 11% with an eye spot, and dsNlRac5-and dsNlRac2-injected females laid 93% and 86% of eggs with an eye spot, similar to dsGFP-injected controls (90%). In N. lugens, an eye spot usually appears 4-5 days after ovulation, and eggs with an eye spot are considered viable, whereas those without an eyespot are considered to be dead [50]. None of the eggs laid by dsNlRpn4-, dsNlRac1-or dsNlRpn13-injected females hatched into nymphs (figure 9c), compared with around 66% of eggs laid by dsNlRac5-injected females, which was similar to the 72% hatching rate in dsGFPinjected controls. Approximately 56% of the eggs laid by dsNlRac2-injected females developed into nymphs, which was significantly lower than in dsGFP-injected controls (figure 9c). Eggs with and without eye spots laid on rice leaf sheaths by dsNlRpn4-, dsNlRac5-, dsNlRac2-, dsNlRac1-, dsNlRpn13-and dsGFP-injected females are shown in figure 9d.    Figure 8. Effects of RNAi on NlVg and NlVg-like expression in ovaries and fat bodies. Transcript levels (in fold) of NlVg and NlVg-like genes after knockdown of NlRpns and NlRacs in ovaries (a) and fat bodies (b) from dsNlRpn-or dsNlRac-injected females on the 5th day after emergence. dsGFP-injected samples were used as controls. Relative transcript levels of NlVg and NlVg-like genes in each treatment were determined by qRT-PCR. Three independent biological replicates (mean ± standard deviation; n = 10) were conducted, as described in figure 2. (c) Determination of NlVg protein levels in female ovaries and fat bodies from dsNlRpn-or dsNlRac-injected females on the 5th day after emergence, determined by western blot analysis. β-Actin was used as a loading control. (d ) Ovarian and whole-body morphologies were observed under a stereomicroscope in dsNlVg-injected adult females on the 5th and 7th days after emergence. Scale bars, 1.0 mm. (e) Observation of lipid droplets in oocytes after knockdown of NlVg. Ovaries were isolated from dsNlVg-injected adult females on the 5th day after emergence and lipid droplets were observed as described in figure 5. ( f ) Lipid droplets were measured using IMAGEJ 1.52a software (National Institutes of Health, Maryland, USA). Statistical analysis was performed using GRAPHPAD PRISM 8. Lipid droplets in oocytes from dsNlVg-and dsGFP-injected females exhibited different diameters (mean ± standard deviation; n = 100).**p < 0.01 between dsNlVg and dsGFP treatments (Student's t-test). (g) NlVg gene transcript levels were analysed by qRT-PCR, as described in figure 2. (h) NlVg protein levels in ovaries from dsNlVg-and dsGFP-injected females on the 5th day after emergence, detected by western blot analysis (n = 10). Protein samples were diluted 400 times. β-actin was used as a loading control. Western blotting data represent one of three experiments. and NlCYP314A1. Transcript levels of NlCYP307A2, NlCYP306A2 and NlCYP314A1 were significantly decreased after knockdown of NlRpns and NlRac1, but were not affected by knockdown of NlRac5 and NlRac2 (figure 10a), indicating that NlRpns and NlRac1 mediated 20E biosynthetic gene expression. We previously quantified 20E levels in 4th-5th instar nymphs and adults at 0-72 h after emergence [45]. In the present study, we quantified 20E levels in females at 120 h after emergence, as the oocytes matured in the ovaries. royalsocietypublishing.org/journal/rsob Open Biol.

11: 200251
A typical chromatogram of 20E was produced from dsGFPinjected adult females at 120 h using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The MS/ MS spectra of 20E extracted from dsNlRac5-and dsNlRac2injected females were identical to the typical 20E spectra (figure 10b), but spectra from other dsNlRpn-and dsNlRac1injected females produced very small 20E peaks (figure 10b), suggesting that dsNlRpn and dsNlRac1 RNAi significantly reduced 20E levels in adult female N. lugens. Furthermore, we quantified the amounts of 20E in adult females at 120 h after emergence (figure 10c). 20E levels were very low in dsNlRpn-and dsNlRac1-injected adult females (0.0001-0.002 ng mg −1 ) compared with dsGFP-, dsNlRac5-and dsNlRac2-injected females (0.0067-0.0072 ng mg −1 ). These data suggest that knockdown of NlRpns and NlRac1 affects transcription of 20E biosynthetic enzymes and production of 20E in female adults. We conducted rescue experiments by injecting 0.4 µg of 20E in each dsNlRpn-injected insect. The phenotype of female ovaries was observed at 120 h after emergence. 20E could not rescue the abnormal ovary phenotypes resulting from dsNlRpn treatments (data not shown).

Effects of proteasome activity on ovarian morphology and gene transcription in females
As described above, knockdown of NlRpns resulted in various female reproduction phenotypes. To explore the possibility that these phenotypes resulted from decreases or changes in proteolysis within N. lugens, we determined the chymotrypsin-like proteasome activity through in vitro proteasome activity assay after knockdown of transcript levels of all    Figure 11. RNAi effects of a mixture of dsNlRpns and dsNlRacs on ovarian development. Newly emerged adult females were injected with a mixture of dsNlRpns and dsNlRacs including dsNlRpn1-4, dsNlRac5, dsNlRpn6-8, dsNlRac2, dsNlRac1 and dsNlRpn11-14 at an equal quality ratio. (a) Measurement of chymotrypsin-like proteasome activity in the whole-body of the adult females using the fluorogenic peptide Suc-Leu-Leu-Val-Tyr-AMC. y-axis refers to the fluorescence intensities measured at 10 min intervals in dsNlRpn + NlRac mix-and dsGFP-injected samples. x-axis refers to the reaction time over a period of test time (60 min). Each sample represents 20 mg of the whole-body lysates of the adult females. For each treatment, five independent biological replicates were conducted. **p < 0.01, *p < 0.05 between dsNlRpn + NlRac mix-and dsGFP treatments (Student's t-test). Data are means ± standard deviation. Statistical analysis was performed using GRAPHPAD PRISM 8. (b) Morphology of adult females was observed on the 7th day after RNAi. Scale bar, 1 mm. Ovaries were dissected from the females on the 5th day after RNAi. Insects injected with dsGFP were used as controls. Scale bar, 500 µm. (c) RNAi effects depleting NlRpns and NlRacs together on individual NlRpn and NlRac gene expression. RNAi was performed by microinjecting dsNlRpn + NlRac mix into the newly emerged female adults. Total RNAs were extracted from the whole bodies of the insects on the 5th day after RNAi (n = 10 adults). The relative transcript levels of individual NlRpn and NlRac gene were determined by qRT-PCR as described in figure 2. Statistical analysis was performed using GRAPHPAD PRISM 8 software. Three independent biological replicates (mean ± standard deviation) were conducted and the relative transcript levels were measured using the ΔΔC t method. **p < 0.01 between dsGFP and dsNlRpn + NlRac mix treatments (Student's t-test). dsNlRpns refers to dsNlRpn + NlRac mix.
royalsocietypublishing.org/journal/rsob Open Biol. 11: 200251 Subsequently, we investigated the effects of decreasing proteolysis through knockdown of a subunit in the 20S CP by RNAi. Subunits β2, β5 and β1 of the 20S CP have proteolytic activity, displaying trypsin-like, chymotrypsin-like and caspase-like peptidase activity, respectively. Here, we focus on a N. lugens proteasome subunit β5 (NlPSMB5) gene to examine RNAi effects on ovarian morphology and lipid accumulation in oocytes. Knockdown of NlPSMB5 generated an abnormally inflated abdomen and stretched intersegmental membranes in the tergum of adult females (figure 12a, upper panel). These females displayed apparently abnormal ovaries with no or malformed immature oocytes in the ovarioles ( figure 12a, lower panel). Globular lipid droplets with different sizes were observed in the immature oocytes (figure 12b). By contrast, dsGFP-injected females had normal ovaries containing the mature oocytes with the uniform sizes of lipid droplets ( figure 12a,b). We conducted a kinetic assay to monitor the chymotrypsin-like proteasome activity in adult females by knockdown NlPSMB5 gene. Our experiments revealed that NlPSMB5 knockdown significantly decreased the chymotrypsin-like proteasome activity in adult females over a period of 60 min. In dsGFP-injected female samples, the fluorescence intensities were 2.0-2.3 fold higher than that in dsNlPSMB5-injected sample within 60 min (figure 12c). A negative control did not show the change of the fluorescence intensities over the test period. We investigated the effects of NlPSMB5 knockdown on the transcript level variations of Vg, VgR and lipase genes in fat bodies and ovaries of adult females. qRT-PCR analysis confirmed that transcript levels of NlPSMB5 gene were notably reduced in dsNlPSMB5-injected fat bodies and ovaries compared with those in dsGFP-injected controls (figure 12d). The transcript levels of the NlATGL, NlPNLIP1, NlLPL2, NlLIPF1, NlVg, NlVg-like2, NlCYP307A2 and NlCYP314A1 genes were significantly decreased in the ovaries (figure 12d); and the transcript levels of NlATGL, NlPNLIP2, NlPLRP2, NlVg and NlVg-like2 genes were significantly decreased in the fat bodies by knockdown of NlPSMB5 compared with dsGFP-injected controls (figure 12e). By contrast, NlPSMB5 knockdown did not change the NlVgR transcript levels in the ovaries (figure 12d ). These observations suggest that decreasing chymotrypsin-like proteasome activity through

Effects of oocyte-specific NlBic-C knockdown on gene expression in female ovaries
As NlRpn knockdown decreased the levels of lipidmetabolizing, 20E biosynthetic enzymes, Vg and Vg-like, and halted oogenesis at late mid-stage, we investigated whether the levels of these genes could be altered after knockdown of genes encoding non-proteasomal subunits that are also known to affect oogenesis. We used femalespecific N. lugens Bicaudal-C (NlBic-C) as a control because this gene is expressed only in developing oocytes and required for oogenesis and oocyte maturation. NlBic-C knockdown resulted in abnormal ovaries with severely inhibited oocyte growth in ovarioles (figure 13a), whereas fully developed oocytes were observed in the ovarioles of dsGFPinjected female ovaries ( figure 13a). Subsequently, we investigated the effects of NlBic-C knockdown on Vg and Vg-like levels. qRT-PCR analysis showed that NlBic-C expression was significantly decreased on the 5th day after emergence in dsNlBic-C-injected females ( figure 13b). NlBic-C knockdown did not significantly alter the expression of lipid-metabolizing genes (NlATGL, NlPNLIP1, NlLPL2 and NlLIPF1), 20E biosynthetic enzymes (NlCYP307A2 and NlCYP314A1) and Vg genes (NlVg and NlVg-like2) ( figure 13b).
These results indicate that knockdown of NlBic-C, which is . RNAi effects of dsNlBic-C on ovarian development. Newly emerged adult females were injected with dsNlBic-C. (a) Ovaries were dissected from the females on the 5th day after RNAi. Insects injected with dsGFP were used as controls. Scale bar, 500 µm. (b) RNAi effects depleting NlBic-C on the transcript levels of lipid-metabolizing (NlATGL, NlPNLIP1, NlLPL2 and NlLIPF1), 20E biosynthetic enzymes (NlCYP307A2, NlCYP306A2 and NlCYP314A1) and Vg (NlVg and NlVg-like2) genes. RNAi was performed by microinjecting dsNlBic-C into the newly emerged female adults. Total RNAs were extracted from the ovaries of the females on the 5th day after RNAi (n = 10 adults). The relative transcript levels of the target genes were determined by qRT-PCR as described in figure 2. Statistical analysis was performed using GRAPHPAD PRISM 8 software. Three independent biological replicates (mean ± standard deviation) were conducted and the relative transcript levels were measured using the ΔΔC t method. **p < 0.01 between dsGFP and dsNlBic-C treatments (Student's t-test). ns: no significant difference.
necessary for the late stages of oogenesis, did not affect the levels of lipid-metabolizing genes, 20E biosynthetic enzymes and the Vg gene, whereas knockdown of NlRpns affected the expression of these genes in adult females.

Discussion
Rpns are constituents of the 26S proteasome that have been reported to be involved in polyubiquitinated substrate protein recognition and deubiquitination in eukaryotes. Using RNAi, we report the functions of Rpn and Rac in N. lugens, a model hemipteran insect, based on the whole genome sequence. Fifteen Rpns and Racs were identified in the N. lugens genome and transcriptome databases, and their deduced amino acid sequences showed close phylogenetic relationships with their homologues in many insect species and humans. In D. melanogaster, however, the nomenclatures of Rpns were different from those of other insect species and humans but similar to their homologues in S. cerevisiae. Bioinformatics analysis indicated that each NlRpn and NlRac contained a characteristic domain(s), including PC, PAM, PCI, vWFA, UIM, JAB/MPN, MitMem, PDZ and ANK, implying their functional diversity. qRT-PCR analysis detected NlRpn and NlRac expression throughout development, with the highest levels in females and/or laid eggs and the lowest levels in males, suggesting their functional importance in adult females. Tissue-specific expression analysis revealed that NlRpn and NlRac levels were highest in the ovaries, further supporting their physiological roles in female reproduction. Based on these observations, we silenced NlRpn and NlRac expression in N. lugens by RNAi, which resulted in the generation of lethal phenotypes. The survival rates of insects injected with dsNlRpns and dsNlRac1 were less than 20% at 14 dpi compared with greater than 80% in dsGFPinjected controls, suggesting that these genes are important for survival. However, NlRac5 and NlRac2 were not required for survival. Except for NlRpn4, NlRac5 and NlRac2, adult female N. lugens injected with dsNlRpns and dsNlRacs showed an abnormally expanded abdomen and a significantly increased body weight at 7 dpi.
To determine whether knockdown of NlRpn or NlRac had any effect on other NlRpns and NlRacs, we examined the levels of other NlRpns and NlRacs in the absence of a different NlRpn or NlRac. Our results revealed that knockdown of NlRpn or NlRac either upregulated or failed to alter the levels of most other NlRpns and NlRacs. In C. elegans, knockdown of rpn-6.1 or rpn-10 increased the expression of the other 26S proteasome subunits, probably to compensate for the reduction in proteasomal activity induced by the decreased levels of these subunits [36,37]. In this study, we report that knockdown of NlRpn or NlRac did not affect the levels of most other NlRpns and NlRacs.
To understand the reasons for the abnormally expanded abdomen phenotype, we dissected adult females and observed fat bodies in the abdomens of individuals injected with most dsNlRpns and dsNlRac1. However, we observed fewer fat bodies in dsNlRpn4-, dsNlRac5-, dsNlRac2-and dsGFP-injected females (data not shown). DsNlRpn1-, dsNlRpn2-, dsNlRpn3-, dsNlRpn6-, dsNlRpn7-, dsNlRpn8-, dsNlRac1-, dsNlRpn11-, dsNlRpn12-, dsNlRpn13-and dsNlRpn14-injected females also showed defects in the ovaries, which contained no or irregularly shaped immature oocytes in the ovarioles. By contrast, dsNlRpn4-, dsNlRac5-, dsNlRac2-and dsGFP-injected females showed normal ovaries with mature banana-shaped oocytes. These findings indicate that most NlRpns and NlRac1 might be involved in lipid metabolism in the fat bodies, that is, they might affect the absorption and utilization of lipids in the developing oocytes. We therefore investigated lipid droplet accumulation in the ovary. Knockdown of NlRpns resulted in the accumulation of large lipid droplets in the oocytes, whereas dsNlRpn4-, dsNlRac5-, dsNlRac2-and dsGFP-injected females showed small round lipid droplets that were distributed uniformly in the oocytes.
In Hemipteran insects, such as Cixius nervosus (Cixiidae) and Javesella pellucida, Conomelus anceps, N. lugens (Delphacidae), the developing oocytes are surrounded by a single layer of follicular cells. At the start of vitellogenesis, each follicular cell generates two nuclei, with the two nuclei positioned one on top of the other in the long axis of each cell [16,51]. Knockdown of NlRpns disrupted the arrangement of follicular cells surrounding the oocytes. Furthermore, these follicular cells did not possess the typical two nuclei, whereas dsNlRpn4-, dsNlRac5-, dsNlRac2-and dsGFP-injected females had normally arranged follicular cells with two nuclei surrounding the oocytes. These observations support the hypothesis that NlRpns are associated with the accumulation of lipid droplets in oocytes and the development of follicular cells during vitellogenesis in female adults.
Lipids, mostly triglycerides, comprise approximately 30-40% of the oocyte dry weight and greater than 50% of the fat body dry weight in insects [52]. Triglycerides are stored in lipid droplets in oocytes and fat body cells and serve as the main energy source for oocyte maturation and embryonic development [53][54][55]. We measured the triglyceride content of the ovaries and fat bodies in N. lugens. Except for NlRac5, knockdown of NlRpns and NlRacs significantly reduced triglyceride levels in the ovaries, but not in fat bodies. Combined with the abnormally expanded abdomens of NlRpn-injected females, these results indicate that lipid metabolism was adversely affected in the fat bodies, thereby disrupting the provision of triglycerides to the ovaries as well as the uptake and utilization of triglycerides in the oocytes.
We further investigated the reasons for the triglyceride metabolic disorder by examining lipase expression in the ovaries and fat bodies following RNAi. The levels of NlATGL, NlPNLIP1 and NlLIPF1, which encode triglyceride metabolism-related lipases, and NlLPL2, which encodes a lipoprotein lipase, were significantly decreased in the ovaries, and the levels of NlATGL and NlPNLIP2, which encode triglyceride metabolism-related lipases, and NlPLRP2, which encodes a pancreatic lipase-related protein, were significantly decreased in the fat bodies after the knockdown of NlRpns and NlRacs, but not NlRpn4, NlRac5, NlRac2, NlRac1 and NlRpn13. These results indicate that knockdown of most NlRpns can affect lipase expression in the ovary and fat bodies.
ATGL is a vital enzyme in the first step of triglyceride hydrolysis [55,56], and the suppression of ATGL expression significantly reduced the glycerin content and the number of eggs laid in N. lugens [57]. Another study has reported that a lipoprotein lipase is important in the uptake of lipids by developing oocytes in Manduca sexta [58]. PLRP2 has various lipolytic activities in mammals, but its role in female reproduction remains unknown. We previously reported that NlPLRP2 was required for oocyte maturation and development in insects [18]. The findings of the present study reveal the regulatory mechanisms of six genes encoding various lipases in N. lugens. The suppression of lipase expression by knockdown of NlRpns resulted in triglyceride metabolism disorders, leading to lipid accumulation in the ovaries and fat bodies, abnormally expanded abdomens, and large lipid droplet accumulation in developing oocytes in affected females. The oocytes were probably unable to use the products of the lipase reaction, such as glycerol and free fatty acids, thereby impairing lipid hydrolysis metabolism and oocyte maturation.
Developing oocytes in all oviparous insects accumulate massive amounts of yolk to ensure embryo development [59]. Vg, as the major yolk protein precursor, is synthesized in fat bodies outside of the ovaries, released into the haemolymph, and transported into the growing oocytes via membrane receptor-mediated endocytosis [59,60]. The uptake of Vg by developing oocytes during egg maturation is essential for successful female reproduction [60]. Vg knockdown resulted in ovarian atrophy and reduced egg production in the hemipteran bedbug Cimes lectularius [61]. One Vg and two Vg-like (Vg-like1 and Vg-like2) genes were identified and investigated in oocyte development in N. lugens [3]. In the present study, we examined the changes in the levels of these genes following NlRpn and NlRac knockdown. Except for NlRpn4, NlRac5, NlRac2, NlRac1 and NlRpn13, knockdown of NlRpns and NlRacs significantly decreased the levels of Vg and Vg-like2, but not Vg-like1, in the ovaries and fat bodies, as confirmed at the protein level by western blotting. These data suggest that Vg and Vg-like2 expression is affected by knockdown of most NlRpns. Furthermore, the suppression of Vg expression in adult females resulted in the accumulation of hypertrophied fat bodies in the abnormally expanded abdomen and large lipid droplets in the developing oocytes, consistent with the observations of NlRpn gene suppression. However, Vg-like2 knockdown did not produce an abnormal phenotype in the oocytes and ovaries. Vg and Vg-like2 shared a close phylogenetic relationship that was distant from Vg-like1 in N. lugens [3]. In addition, Vg and Vg-like2 shared the same developmental stage-specific expression patterns, with transcripts primarily detected in adults, whereas Vg-like1 transcripts were detected throughout all stages, implying different regulatory mechanisms and physiological functions in N. lugens development and reproduction. In the present study, NlRpns affected Vg and Vg-like2 expression, and Vg accumulation in the ovaries and fat bodies. Vg is a primary yolk protein that provides nutrients for the oocytes, whereas Vg-like2 is not essential for oocyte maturation. Vg is taken up by developing oocytes via VgR-mediated endocytosis. qRT-PCR analysis indicated that knockdown of NlRpns or NlRacs did not alter the levels of NlVgR in the ovaries, suggesting that the depletion of NlRpns or NlRacs did not affect the uptake of Vg into developing oocytes. Instead, it impeded Vg production in ovaries and fat bodies, thereby disrupting oocyte maturation.
Except for NlRac5 and NlRac2, knockdown of NlRpns and NlRacs resulted in the complete failure of ovulation or a significant reduction in the number of eggs laid, which was owing to the severely deformed ovaries. The eggs laid by NlRpn4-, NlRac1-and NlRpn13-injected females failed to hatch into nymphs, and all of the eggs laid by dsNlRac1and dsNlRpn13-injected females, and nearly 90% of those laid by dsNlRpn4-injected females, lacked an eye spot. These observations indicate that knockdown of NlRpns can downregulate lipase and Vg expression, which might impede the digestion and absorption of nutrients and energy in the developing oocytes, thereby inhibiting oocyte development.
Ecdysteroids are steroid hormones in insects that are required for ovarian development and egg production. Ecdysteroids are produced in the ovaries of adult female insects [62,63]. We previously reported that 20E was an active steroid hormone in N. lugens, and knockdown of 20E biosynthetic enzymes in newly emerged females resulted in failed egg production, fewer vitellogenic mature oocytes, fewer eggs laid, and abnormal embryonic development of the laid eggs, demonstrating that 20E biosynthesis in ovaries is critical for ovarian development in N. lugens [45]. In the present study, we investigated the changes in expression of halloween genes in the 20E biosynthetic pathway in N. lugens ovaries following knockdown of NlRpns and NlRacs. Except for NlRac5 and NlRac2, knockdown of NlRpns and NlRacs significantly decreased the levels of the halloween genes NlCYP306A2, NlCYP307A2 and NlCYP314A1 in female ovaries and reduced 20E production throughout the whole body in adult females. Among the halloween genes in N. lugens, NlCYP306A2, NlCYP307A2 and NlCYP314A1 were highly expressed in the ovaries, whereas NlCYP307B1 and NlCYP315A1 were expressed in various tissues [45], indicating that NlRpns are involved in the expression of ovarian-specific halloween genes. NlRpn-mediated halloween gene knockdown resulted in reproductive failure. We conducted rescue experiments in dsNlRpn-injected insects and observed that 20E could not rescue the abnormal ovarian phenotype. This might have been owing to the fact that knockdown of NlRpns decreased the expression of many genes, including Vg, lipases, and 20E biosynthetic genes, which are vital for oocyte maturation and development. The biosynthesis of 20E alone was not enough to rescue the reproductive defects.
The roles of NlRpns and NlRacs in N. lugens survival and reproduction are summarized in table 2. Among the NlRpns, NlRpn1-3, NlRpn6-8, NlRpn11-12 and NlRpn14 are associated with the expression of functional genes encoding a Vg and six lipases, including four triglyceride metabolism-related lipases and three halloween genes in the ovaries and/or fat bodies, which are necessary for oocyte formation and maturation. NlRpn4, NlRpn13, and an associated chaperone, NlRac1, are also involved in Vg, lipase, and halloween expression, and affect oocyte maturation and egg hatching. By contrast, two associated chaperones, NlRac5 and NlRac2, had no effect on Vg, lipase and halloween expression. In addition, they had no effect on female development and reproduction, including oocyte maturation, ovulation, and egg hatching.
It is well established that the proteasome initiates gene transcription through proteolytic and non-proteolytic activities, and an inhibition of 26S proteasome function can affect transcription [46,47]. To determine whether the transcriptional changes reported in the present study were owing to a decrease in proteolytic activity or the functions of NlRpns and NlRacs, we examined the effects of decreasing proteolysis by silencing the NlPSMB5 subunit in the 20S CP. NlPSMB5 knockdown notably decreased chymotrypsin-like proteasome activity in N. lugens, significantly reduced the levels of NlATGL, NlPNLIP1, NlLPL2, NlLIPF1, NlVg, NlVg-like2, NlCYP307A2 and NlCYP314A1 but not that of the ovarian-specific VgR gene, and generated deficient phenotypes in the ovaries of adult females, as observed in dsNlRpn-injected females, suggesting that the changes were owing to a decrease in proteolytic activity. We also investigated the effects on proteolytic activity following simultaneous dsNlRpns and dsNlRacs knockdown. Knockdown significantly decreased chymotrypsin-like proteasome activity in adult females and resulted in a deficient phenotype in the ovaries. These observations indicate that depleting NlRpn decreased proteasome activity, which downregulated the transcription of genes involved in ecdysteroidogenesis, vitellogenesis and lipogenesis, and resulted in ovaries lacking the late stages of oogenesis.
To further confirm the results, we included a developing oocyte-specific gene, NlBic-C, whose knockdown halted oogenesis as a control to determine whether the same transcripts could be affected under non-Rpn knockdown conditions. qRT-PCR analysis showed that NlBic-C knockdown did not alter the levels of lipid-metabolizing genes (NlATGL, NlPNLIP1, NlLPL2 and NlLIPF1), 20E biosynthetic enzymes (NlCYP307A2 and NlCYP314A1), and Vg genes (NlVg and NlVg-like2) in female ovaries. By contrast, the levels of these genes were significantly decreased in NlRpns or NlPSMB5depleted ovaries compared with dsGFP-injected controls. These results indicate that NlRpn or NlPSMB5 knockdown decreases proteasome activity and affects the same genes involved in ecdysteroidogenesis, vitellogenesis and lipogenesis.
There is crosstalk between ecdysteroidogenesis, vitellogenesis and lipogenesis in reproduction in female insects. The regulatory pathways differ depending on the reproductive strategies adopted by various insects [60]. In D. melanogaster, 20E signalling regulates vitellogenesis and oogenesis in the presence of sufficient nutritional resources. In addition, 20E has been linked to lipid accumulation in oocytes during oogenesis [60,64]. In the present study, we reported that Rpns are involved in female reproduction in a model hemipteran insect. We proposed a regulatory mechanism for the roles of Rpns in ovarian development in N. lugens.
Furthermore, crosstalk between 20E, Vg, lipids and Rpns was hypothesized ( figure 14). In summary, Vg and lipid (i.e. triglycerides) production is associated with the proteolytic activity of the proteasome in the ovaries and fat bodies. Knockdown of NlRpns decreased proteasome activity, repressed Vg and lipase expression, reduced triglyceride content, decreased the Vg protein level and disrupted lipid metabolism in the ovaries, thereby leading to the defective absorption and utilization of nutrients in developing oocytes. Decrease of the proteolytic activity of the proteasome also decreased halloween expression and reduced 20E synthesis in the ovaries. Furthermore, insufficient levels of lipids, Vg and 20E in the ovaries disrupted oocyte maturation and development. Therefore, the proteolytic activity of the proteasome is required for insect survival, ovarian development and oocyte maturation, and our results point to the importance of the proteolytic activity of the proteasome in reproduction.

Insects
Nilaparvata lugens were originally collected from a rice field in the Huajiachi Campus of Zhejiang University, Hangzhou, China in 2008 and subsequently maintained in our laboratory on a diet of fresh rice seedlings (Oryza sativa strain Xiushui 110) at 26 ± 0.5°C and 50 ± 5% humidity under a 16 L : 8 D photoperiod, as described previously [44].

Bioinformatics analysis
NlRpn and NlRac sequences were searched against the N.

Sample collection for developmental stage-and tissue-specific expression analyses
To investigate the expression levels of NlRpn and NlRac genes at different developmental stages, we collected samples from eggs, 1st, 2nd, 3rd, 4th and 5th instar nymphs, and adult males and females. For tissue-specific analysis, we dissected fat body, ovary, brain, cuticle, salivary gland and gut tissues from adult females under an S8AP0 stereomicroscope (Leica Microsystems GmbH, Wetzlar, Germany), and washed them quickly in diethylpyrocarbonate-treated NaCl/Pi solution (pH 7.4), as described previously [65].   Figure 14. Schematic diagram of NlRpn affecting female reproduction in N. lugens. Knockdown of NlRpns decreased the proteolytic activity of proteasome in N. lugens and impeded the transcription of Vg, Vg-like and lipid-metabolizing genes, including adipose triglyceride lipase (NlATGL), pancreatic triacylglycerol lipase 1 (NlPNLIP1), lipoprotein lipase-like 2 (NlLPL2), gastric triacylglycerol lipase-like1 (NlLIPF1), pancreatic lipase-related protein 2 (NlPLRP2), and pancreatic triacylglycerol lipase 2 (NlPNLIP2) in fat bodies and/or ovaries. NlRpns are also related to the expression of the 20E biosynthetic enzyme genes NlCYP307A2, NlCYP306A2, and NlCYP314A1 in female ovaries. Knockdown of NlRpns affects Vg synthesis in the fat body and utilization in the developing oocytes, disrupts lipid metabolism in the fat body and oocyte, and also inhibits 20E synthesis in the ovaries, leading to failure of oocyte maturation and development. denaturation at 95°C for 3 min followed by 40 cycles at 95°C for 15 s and 55°C for 30 s. The gene-specific primers were designed using the PRIMER PREMIER 6.0 program based on the N. lugens transcriptomic sequences (electronic supplementary material, table S2). The use of reference genes as internal controls is the most appropriate normalization strategy for achieving the reliable qRT-PCR assay [66]. In our previous study, the use of N. lugens housekeeping 18S ribosomal RNA (18S rRNA) and β-actin genes (GenBank accession nos. JN662398 and XP_022202043) has been validated for their stable expressions in N. lugens tissues and developmental stages [42,44,45,67]. In this study, the N. lugens 18S rRNA and β-actin genes were used as internal controls and the results were normalized to the expression levels of the two internal genes, respectively. The relative quantitative method (ΔΔC t method, C t is the threshold cycle) was used to evaluate the relative differences in the transcript levels as described previously [42,44,45,67]. Namely, the following equation was used: ΔC t = the C t of specific gene-the C t of an internal gene (18S rRNA or β-actin). Three biological replicates were performed.

Double-stranded RNA synthesis and RNA interference
The open reading frame of each NlRpn or NlRac around 500 base pairs was amplified and cloned into the pMD-19 T vector (TaKaRa). dsRNA were synthesized in vitro using PCR-generated DNA templates using a T7 High Yield RNA Transcription Kit (Vazyme). The specific primers used to generate the DNA templates are shown in the electronic supplementary material, table S3. Newly emerged adult females were anaesthetized with carbon dioxide and microinjected with approximately 250 ng of dsNlRpn or dsNlRac using a FemtoJet microinjection system (Eppendorf-Netheler-Hinz, Hamburg, Germany). Insects injected with GFP dsRNA from Aequorea victoria were used as controls. The treated insects were reared on fresh rice seedlings at 26 ± 0.5°C and 50% ± 5% humidity under a 16 L : 8 D photoperiod. Phenotypes were observed following RNAi.
For examining the effects of depleting specific NlRpns and NlRacs on other NlRpn and NlRac gene expression, RNAi was performed by injecting each dsNlRpn or dsNlRac into newly emerged adult females. Total RNAs were extracted from the whole bodies of the female adults on the 5th day after RNAi and used as starting RNAs for qRT-PCR analysis as described in Materials and methods, §4.4. dsGFP-injected insects were used as controls.

Determination of lipid droplet accumulation in oocytes of Nilaparvata lugens ovaries
Ovaries from adult females on the 5th day after emergence were used to determine lipid droplet accumulation. Intracellular lipid droplets were visualized by staining ovaries with Nile red (Sangon Biotech, Shanghai, China) [68]. Briefly, ovaries were dissected from adult females and washed three times in 1 × phosphate-buffered saline (PBS), and then fixed in 4% paraformaldehyde for 30 min at room temperature. After washing three times with PBS containing 0.1% Triton X-100 (PBST), the ovaries were incubated in 0.1% PBST buffer with 1 µg ml −1 Nile red solution for 90 min at room temperature to stain the lipid droplets. For nuclear staining, the ovaries were incubated for 15 min in 1 µg ml −1 4 0 ,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific) at room temperature followed by washing with 0.1% PBST three times. Finally, the ovaries were washed twice with 1 × PBS and transferred to microslides. Lipid droplets were observed under an LSM 800 confocal microscope (Carl Zeiss MicroImaging, Göttingen, Germany). The absorption and emission wavelengths for the lipid droplets and nuclei were 550/570 nm and 358/461 nm, respectively. The confocal images were analysed with ZEN 2.3 software (Carl Zeiss MicroImaging) and the diameters of the lipid droplets were measured using IMAGEJ 1.52a software (National Institutes of Health, Maryland, USA).

Detection of Vg in ovaries by western blotting assay
RNAi was conducted by injecting newly emerged females with each dsNlRpn or dsNlRac. Ovaries were dissected at the 5th dpi. dsGFP-injected insects were used as controls.
The ovaries were washed three times with ice-cold PBS solution (137 mM NaCl, 2.68 mM KCl, 8.1 mM Na 2 HPO 4 , 1.47 mM KH 2 PO 4 , pH 7.4) and homogenized, followed by adding 6 × protein loading buffer and boiling for 15 min. The proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis at 80 V for 2 h and transferred to polyvinylidene fluoride membranes (Merck KGaA, Darmstadt, Germany) at 60 V for 2 h. Blots were probed with mouse primary antibody (1 : 5000 dilution) against N. lugens Vg ( prepared in our laboratory) and detected using goat anti-mouse IgG-conjugated horseradish peroxidase antibody (Genscript, Nanjing, China) at a dilution of 1 : 5000. Western blot signals were visualized using a Chemiluminescence Detection Kit (Bio-Rad) and photographed with the Molecular Imager ® ChemiDoc™ XRS + System (Bio-Rad). The β-actin polyclonal rabbit primary antibody was used as a loading control [69].

Measurement of triglycerides in ovaries and fat bodies
Ovaries and fat bodies were dissected from dsNlRpn-or dsNlRac-and dsGFP-injected females at the 5th day after emergence. Following washing three times in 1× PBS, the ovaries and fat bodies were homogenized separately in lysis buffer and kept at room temperature for 10 min, and the homogenates were then heated for 10 min at 70°C and centrifuged for 5 min at 2000g. The triglyceride contents of the supernatants were measured using a Tissue Triglyceride Assay Kit (Applygen Technologies, Beijing, China) and protein concentrations were analysed using a Bicinchoninic Acid (BCA) Protein Assay Kit (Thermo Fisher Scientific), according to the manufacturer's instructions.

Functional analysis of NlRpn and NlRac genes during ovulation and egg hatching
Ovulation and egg hatching experiments were conducted according to Lou et al. [16]. Briefly, newly emerged females were microinjected with each dsNlRpn or dsNlRac and royalsocietypublishing.org/journal/rsob Open Biol. 11: 200251 dsGFP, respectively. Brachypterous females have higher fecundity than macropterous females, and we therefore used brachypterous female N. lugens for ovulation experiments in this study. Newly emerged brachypterous females were injected with the target dsRNAs and reared on fresh rice seedlings for 1 day. To ensure successful mating, a single dsNlRpn-or dsNlRac-injected or dsGFP-injected control adult female was mated with two non-injected adult males in a long glass tube containing three-leaf stage fresh rice seedlings (9 ± 0.5 cm long) at 26 ± 0.5°C and 50% ± 5% relative humidity under a 16 L : 8 D photoperiod for 6 days. The adult insects were then removed and hatched nymphs were observed and counted at 24 h intervals, and were removed from the rice seedlings 11 days later. Unhatched eggs were dissected and counted. Biological replicates were carried out for each mating (n = 17-20 females × males).

Measurement of 20E in Nilaparvata lugens adult females
We previously determined the 20E contents of N. lugens nymphs and adults [45]. In this study, 20E contents were measured in adult females after NlRpn and NlRac RNAi. Briefly, newly emerged adult females were injected with each dsNlRpn or dsNlRac, with dsGFP-injected insects as controls. The 20E content was measured in the whole body at 120 h after emergence, according to Nakaoka et al. [70] with some modifications, as described in Zhou et al. [45].

In vitro proteasome activity assay
Each newly emerged adult female was injected with 250 ng of dsNlPSMB5 or 250 ng of a mixture of dsNlRpns and dsNlRacs at equal quality ratio (each dsRNA 100 µg in a mixture of dsRNAs). The adult females were collected at the 5th day after RNAi. dsGFP-injected insects were used as controls.
The in vitro proteasome activity assay using the fluorogenic peptide N-Succinyl-Leu-Leu-Val-Tyr-7-amido-4-methylcoumarin (Suc-LLVY-AMC; Sigma, Missouri, USA) was performed with some modifications to measure the chymotrypsin-like proteasome activity, which has been reported to be the most specific substrate to measure proteasome activity as described previously [71,72]. Briefly, 20 mg of the whole bodies of adult females were homogenized in 200 µl TSDG buffer (10 mmol l −1 Tris/HCl, 1.1 mmol l −1 MgCl 2 , 10 mmol l −1 NaCl, 0.1 mmol l −1 EDTA, 1 mmol l −1 DTT, 2 mmol l −1 ATP, 10% (v/v) glycerol, pH 7.0). The homogenates were centrifuged at 4°C at 20 000g for 10 min and the supernatants were used for an in vitro proteasome activity assay. The fluorogenic peptide substrate Suc-LLVY-AMC was dissolved in 100% DMSO as a stock solution of 10 mmol l −1 . Ten microlitres of the supernatants were added to 20 µl of PBS and 30 µl of 200 µmol l −1 Suc-LLVY-AMC solution that was diluted 50 times with TEAD buffer (20 mmol l −1 Tris/ HCl, 1 mmol l −1 EDTA, 1 mmol l −1 DTT, pH 7.2) from the Suc-LLVY-AMC stock solution and incubated for 1 h at 25°C. The fluorescence intensities were measured at 10 min intervals for 60 min under a multifunctional microplate reader Varioskan LUX (Thermo Fisher Scientific). The absorption and emission wavelengths for the Suc-LLVY-AMC were 380/ 440 nm. 10 µl TSDG buffer was added to 20 µl of PBS and 30 µl of 200 µmol l −1 Suc-LLVY-AMC solution and used as a negative control.

Statistical analysis
Ovulation experiment results were calculated from at least 17 biological replicates; the other assays were from at least three biological replicates. The significance values were determined using Student's t-test (*p < 0.05; **p < 0.01) with GRAPHPAD PRISM 8 software (San Diego, CA, USA).
Data accessibility. This article has no additional data. Authors' contributions. W.W. carried out the molecular laboratory work, participated in data analysis, carried out sequence alignments, statistical analyses and drafted the manuscript; Y.B. participated in the design of the study and critically revised the manuscript; R.Y., Y.P., L.Z. and Y.Y. carried out the molecular laboratory work. All authors gave final approval for publication and agree to be held accountable for the work performed therein.
Competing interests. The authors declare no competing or financial interests.
Funding. This work was supported by the National Natural Science