Introduction

From their origin approximately 400 Mya1,2, insects have developed three basic types of metamorphic development: (i) ametaboly (Apterygota), the most primitive type that is characterized by the absence of a morphological transformation between the wingless immature individuals and the adults; (ii) hemimetaboly (Exopterygota), in which the juvenile wingless nymphs resemble miniature adults and metamorphose into winged adults during the last juvenile instar; and (iii) holometaboly (Endopterygota), in which the crawling juvenile larvae undergo a dramatic morphological transformation to form the winged adult through a two-stage metamorphic process bridged by the holometabolous-specific intermediate pupal stage3,4. Taking into consideration that holometaboly has facilitated the most spectacular radiation of animals, complete metamorphosis can be considered as one of the key evolutionary innovation in insect evolution5,6. And yet, in spite of the clear evolutionary advantages of holometaboly, the female of a number of holometabolan species have lost the typical complete metamorphic transformation.

The loss of complete metamorphosis in holometabolous insects is exclusively found in females and results from the disappearance of pupal and post-metamorphic stages, a process named paedomorphosis. One particular strategy leading to paedomorphosis is neoteny, where the larval morphology is retained and sexual maturity is attained without a pupal or an imaginal stage7,8. Interestingly, whereas holometaboly originated only once in insect evolution from ancestors exhibiting hemimetabolous development2, neoteny in holometabolous insects has occurred independently several occasions in the beetle superfamily Elateroidea9,10 and once in derived lineages of the endoparasitic order Strepsiptera11,12,13. Despite the interest in how insect neoteny has evolved, the molecular mechanisms underlying the loss of complete metamorphosis in holometabolous neotenic females, however, remain poorly understood.

Strepsiptera are a small, bizarre holometabolous order of obligate entomophagous endoparasitoids with extreme sexual dimorphism11,12,13,14,15. Males and females of the early branching order of Strepsiptera, the Mengenillidae, emerge to pupate externally from the host. In contrast, males of the families of the derived suborder Stylopidia pupate endoparasitically in the host and emerge as free-living adults, while females remain obligate endoparasites in the host and are neotenic. The neotenic female Stylopidia, such as Xenos vesparum Rossi, produces 1st instar larvae (planidia) which have rudimentary eyes, thoracic limbs, sclerotized cuticle, and are dedicated to host-seeking. Upon finding and entering a host, the planidium molts to a soft cuticle-apodous endoparasitic 2on instar larva. After three consecutive endoparasitic stages where no ecdysis occurs, the male larva extrudes the head (cephalotheca) through the host cuticle and continues the typical holometabolous development whereby it undergoes complete metamorphosis through pupation. At the end of pupation, the male emerges from the endoparasitic puparium as a free-living winged adult. In contrast, the female after the fourth larval instar extrudes the head, thoracic and first abdominal regions (cephalothorax) though the host cuticle and remains endoparasitic in the mobile host. A few days (3–4 days) after extrusion of the cephalothorax, the neotenic female assumes a calling position16. During this calling, the female produces a potent pheromone that attracts the male17,18,19. The male then inseminates the female through the brood canal opening in the cephalothorax16. After viviparous development in the endoparasitic female, the motile 1st instar planidia emerges through the brood canal opening in the cephalothorax11,13,15. Neotenic female Strepsiptera resembles a “bag of eggs” with an extruded cephalothorax without any external adult characteristics, such as antennae, legs and wings, and whose primary function is to be a repository for eggs13,19,20 (Fig. 1). Therefore, it could be considered that total endoparasitism in Stylopidia females has led to the simplification of their body plan and represents a state of extreme neoteny. However, it is worth noting that although the abdomen remains larviform, particular structures to aid mating, release of the planidia larvae and the pheromone glands are specifically found in the cephalothorax of neotenic females20 (Fig. 1).

Figure 1
figure 1

Life cycle of Xenos vesparum. Free living 1st instar planidium enters a wasp nests to parasitize host larva. Upon entering the host, the planidium molts to an apodous endoparasitic 2nd instar larva, which successively molts two additional times. The male 4th instar larva undergoes metamorphosis through pupation (green arrow), and after, it extrudes through the abdomen of the wasp cuticle. The apolised cuticles of the larval instars form the puparium. The free-living male emerges from the puparium as a winged adult. In contrast, the 4th instar female larva does not undergo the pupal-based metamorphic transformation but develops a cephalothorax, which is extruded through the host. The cephalothorax contains adult-specific structures that facilitate the neotenic female in insemination and the release of the planidium, and also contains glands that release pheromones to attract the male. The abdomen (abd.) and the cephalothorax (ceph.) regions in the neotenic female are marked.

From an endocrine perspective, the metamorphic transition in holometabolous insects is controlled by the sesquiterpenoid juvenile hormone (JH) produced by the corpora allata glands (CA)21,22,23,24,25. The presence of JH during pre-ultimate larval stages prevents premature metamorphosis though the transcriptional induction of the anti-metamorphic transcription factor-encoding gene Krüppel-homolog 1 (Kr-h1)26,27. Conversely, the disappearance of JH and the down-regulation of Kr-h1 at the onset of the last larval instar trigger the metamorphic transformations to the pupal stage by allowing the up-regulation of the transcription factor Broad-complex (Br-C) by the steroid hormone 20-hydroxyecdysone (20E)28,29. During the final part of the last larval instar, Br-C controls the correct switch between larval and pupal forms, thus acting as the “pupal specifier”28,29,30,31,32. In pre-ultimate larval stages, Kr-h1 binds to the promoter region of Br-C and inhibits its expression, thus preventing larvae to undergo precocious larval-pupal transformation33. Finally, after pupation, a third critical transcription factor, E93, is strongly up-regulated by 20E and controls the transition from the pupa to the adult, thus acting as the “adult specifier”34. During the pupal stage, E93 also represses the expression of Kr-h1 and Br-C, thus ensuring the elimination of the two factors whose presence during this period is detrimental for proper adult differentiation1,34. Similar to the repressive activity on Br-C, Kr-h1 also binds to the promoter region of E93, suppressing its expression before the pupal stage and preventing larvae to undergo precocious larval-adult metamorphosis35. Due to their critical importance in the control of metamorphosis in holometabolous and hemimetabolous insects, Kr-h1, Br-C and E93 form what we have recently defined as the Metamorphic Gene Network (MGN)1.

In order to characterize the genetic control of neoteny in holometabolous insects, a recent pilot study proposed that the appearance of neotenic females in Strepsiptera is linked to the modification of regulatory pathways that underlie pupal determination36. The authors identified XvBr-C in X. vesparum and found upregulation in males at the larval-pupal transition but not in females. This work suggested that lack of XvBr-C upregulation might be one of the reasons underlying the loss of complete metamorphosis in neotenic females. However, it has been shown in T. castaneum that depletion of TcBr-C does not result in the maintenance of larval status but rather to a developmental arrest of the animal at the larval-pupal transition with knockdown animals showing a mix of larval, pupal and adult features1,28,29,30,31, thus suggesting that additional factors are required to induce complete metamorphosis.

Here, we further investigate the postulated link between the alteration of the pupal determination program controlled by the MGN and the occurrence of adult neotenic X. vesparum females. Due to the critical role of E93 in the formation of the adult form, we have identified X. vesparum E93 homolog (XvE93). Then, we have compared the expression dynamics of XvE93 in males and females during the metamorphic transition. Because the female is composed of a posterior abdomen that is larval-like and an anterior cephalothorax that has features for insemination and release of the 1st instars and the pheromone glands we analyzed the differential expression of XvE93 in these two regions of the female. Moreover, to complete the identification of the factors of the MGN, we have isolated a fragment of the anti-metamorphic Kr-h1 (XvKr-h1) factor and have measured its expression levels during the larva-to-adult transition in X. vesparum males and females. Finally, to complete the analysis of the MGN in the regulation of neoteny, we have measured the expression levels of XvBr-C in the cephalothorax and the abdomen of X. vesparum females. Our results strongly suggest that the low levels of XvE93 expression in X. vesparum females underlie the evolution of neotenic development.

Material and Methods

Strepsiptera collection

X. vesparum specimens used for cloning and gene expression analysis were collected from the following locations:

(i) X. vesparum parasitic in Polistes dominulus were collected in Italy in 2017, in Tuscany (Central Italy) Via Madonna del Piano, 6. Sesto Fiorentino, It-50019, Italy; N43°49.064′ E11°12.263′ by Helmut Kovac; (ii) X. vesparum parasitic in P. dominulus were collected in 2014, Tuscany (Central Italy) from University of Pisa farm, San Piero a Grado, Pisa (43°40′01.6″N 10°20′21.0″E) and in a private winery, “La Sughera” farm, Loc. Tonini, Spianate, Altopascio, Lucca (43°48′00.2″N 10°42′02.6″E) by Giovanni Benelli and J. Kathirithamby; (iii) X. vesparum parasitic in P. dominulus were collected in 2017 in Austria, Gschwendterstraße 76, A-8062 Kumberg, Austria (N47 10.721 E15 34.373) by H. Kovac; (iv) X. vesparum parasitic in P. dominulus were collected from nests in plastic bottles in 2016 in Fauglia, Pisa, by Consolato Latella and J. Kathirithamby. In all cases, specimens were dissected immediately in saline and fast frozen at −80 °C, or fast frozen and dissected for identification of the endoparasitic stages.

X. vesparum transcriptome assembly for sequence analysis

Raw reads were downloaded from NCBI SRA repository with accession numbers SRR1784898 and SRR1784897 which belong to RNA-seq studies on a X. vesparum female 4th instar larvae and neotenic females, respectively. Two independent transcriptomes assemblies were carried out with rnaSPAdes37 v.0.0.1. Before the assembly, raw reads were trimmed with Trimmomatic38 v3.059 using the following options ILLUMINACLIP:/adapters/TruSeq. 3-PE.fa:2:30:10 SLIDINGWINDOW:4:28 MINLEN:50 in order to remove Truseq adapters and low quality reads. Next, by using known E93, Kr-h1, HR3 and E75 protein sequences from Blattella germanica and Drosophila melanogaster, we performed tBLASTn39 searches (BLAST v.2.2.31) in order to identify the putative transcripts in X. vesparum and design primers for further experimental confirmation of gene sequences.

Phylogenetic analysis of E93

To understand the phylogenetic relationship of E93 proteins, amino acid sequences from E93 proteins were collected from different insect taxa, including that of X. vesparum as well as from two arachnid species as an outgroup (Supplementary Table 1), and aligned using MAFFT40 v7.299b L-INS-I with 1000 iterations. Ambiguously aligned positions were trimmed using trimAl41 v14, with the automated 1 algorithm. The best substitution model for phylogenetic inference was selected using IQ-TREE42 with the TESTNEW model selection procedure and following the BIC criterion. The LG substitution matrix with a 4-categories discrete Γ distribution and allowing for invariant sites was selected as the best-fitting model. Maximum likelihood inferences were performed with IQ-TREE, and statistical supports were drawn from 1,000 ultrafast bootstrap values with a 0.99 minimum correlation as convergence criterion43, and 1,000 replicates of the SHlike approximate likelihood ratio test44.

Pipsqueak Domain Similarity Analysis

A semi-automated BLAST and HMM-based search was conducted with the online Pfam database45 and the Pfam HMM profile for pipsqueak domain. E93 sequences of D. melanogaster (NP_652002.2), Tribolium castaneum (KYB25179.1), B. germanica (CCM97102.1), Bombyx mori (AIL29268.1), Apis mellifera (BAB64310.1) and Zootermopsis nevadensis (KDR22086.1), Caenorhabditis elegans (AB236333.1) and Homo sapiens (AAH53359.1) were obtained from GenBank database, were used for the analysis.

Quantitative real-time reverse transcriptase polymerase chain reaction (qRT-PCR)

Total RNA from individual X. vesparum specimens was extracted using the GenEluteTM Mammalian Total RNA kit (Sigma). cDNA synthesis was carried out as previously described46,47. Relative transcript levels were determined by quantitative real-time PCR (qPCR), using Power SYBR Green PCR Mastermix (Applied Biosystems). To standardize the quantitative real-time RT-PCR (qPCR) inputs, a master mix that contained Power SYBR Green PCR Mastermix and forward and reverse primers was prepared to a final concentration of 100 µM for each primer. The qPCR experiments were conducted with the same quantity of tissue equivalent input for all treatments, and each sample was run in duplicate using 2 µl of cDNA per reaction. As a reference, same cDNAs were subjected to qRT-PCR with a primer pair specific for X. vesparum Ribosomal 18 S36. All the samples were analyzed on the iCycler iQReal Time PCR Detection System (Bio-Rad). Primer sequences used for qPCR for X vesparum are:

XvE93-F: 5′-GGTACAACGCGGTGAAATGTC-3′XvE93-R: 5′-GTTTTCGTGGCCGCATTAAATGC-3′XvKrh1-F: 5′-TATGCGACGATGTACGCTTA-3′XvKrh1-R: 5′-CTTGCACGTTTAACACGTCAT-3′XvHR3-F: 5′-CTACGAGCAAACACCATCGA-3′XvHR3-R: 5′-GGATTGTAATAAGTCGTATACG-3′XvE75-F: 5′-GAAAGAGGAACCAACAAGTTC-3′XvE75-R: 5′-CTTACAACCTTCACATGAATGA-3′XvBr-C-F: 5′-GCAGCATTACCTCTGCTT-3′XvBr-C-R: 5′-CGAAAATATGGGCTGCAG-3′Xv18S-F: 5′-TGCGGCGTATCTTTCAATTGT-3′Xv18S-R: 5-CTGCCTTCCTTAGATGTGGT-3′.

Results and Discussion

Identification and phylogenetic analysis of XvE93

X. vesparum E93 (XvE93) full-length sequence was obtained by the assembly of X. vesparum transcriptomes from 4th instar female larvae and neotenic females (GenBank accession number MH220841). A single cDNA that corresponds to a transcription factor of 1083 amino acid was identified (Fig. 2A and Supplementary Fig. 1). The comparison with other E93 sequences (Fig. 2B,C), revealed that XvE93 possess two helix-turn-helix (HTH) DNA binding motifs of the pipsqueak family, termed RHF1 and RHF2 (Fig. 2A,B). RHF-1 presents 91–98% similarity compared to other insects and a 67% compared to the Caenorhabditis elegans homolog MBR-1, while RHF-2 presents 93–100% similarity compared to other insects, 71% compared to MBR-1 and 53% compared to the human homolog Ligand Co-Repressor (LCoR). Although D. melanogaster DmE93 and human LCoR have only the RHF-2 domain, the two RHF DNA-binding motifs have been shown to be critical for proper induction of 20E dependent genes in the lepidepteran B. mori48, suggesting that both domains, RHF-1 and RHF-2, are required for binding to DNA.

Figure 2
figure 2

The structure of Xenos vesparum XvE93. (A) Predicted amino acid sequence of XvE93. The two HTH-DNA binding motifs, RHF1 (grey box) and RHF2 (stripped box), are indicated. The two NR-box are boxed in black. The CtBP-interaction motif (CtBP-im) is single-line stripped boxed. (B) Comparison of the domain structure of E93 homologs. Tribolium castaneum (TmE93), Apis mellifera (AmE93), Bombyx mori (BmE93), Drosophila melanogaster (DmE93), Zootermopsis nevadensis (ZnE93), Blattella germanica (BgE93), Caenorhabditis elegans (CeE93) and Homo sapiens (HsLCoR). The percentages indicate the sequence identities within each domain. (C) Comparison of the RHF2 domain of XvE93 with those from other E93 homologs. Amino acid residues highly conserved in E93 insects, MBR-1 from C. elegans and human LCoR are indicated by a reverse background, whereas moderate conserved residues are shaded in grey. The positions of the three helices are indicated by horizontal bars.

An interesting feature of all E93 homologs is the presence of at least one nuclear receptor interaction motif (NR-box; also named as LXXLL motif). Interestingly, XvE93 contains two of such motifs (Fig. 2A,B). Through these motifs LCoR interacts with different hormone-bound nuclear receptors attenuating their transcriptional activity49,50. Likewise, B. mori BmE93 interacts through the NR-box with ultraspiracle (USP), a nuclear receptor that dimerizes with the ecdysone receptor (EcR) to act as the functional 20E receptor, to impair the transcriptional activity of the heterodimer48. Finally, a third conserved domain present in XvE93 is the co-repressor C-terminal-binding protein interaction motif (CtBP-im) (Fig. 2A,B). Overall, the highly conserved protein structure observed in all E93 homologs, including that of X. vesparum, suggests that XvE93 functions through similar molecular mechanisms.

On the other hand, maximum-likelihood analysis of E93 sequences, using the sequences of arachnids E93 as outgroup, showed that XvE93 sequence grouped with those of coleopteran species, which is consistent with phylogenomic inferences2,51 (Fig. 3). A remarkable feature of the tree was the different length of the branches. Diptera, and specially flies, had the longest lengths, clearly indicating a faster rate of divergence of these sequences with respect to other insect sequences, whereas non-dipteran branches were shorter, indicating a higher conservation of these sequences.

Figure 3
figure 3

Phylogenetic analysis of E93 proteins. Phylogenetic tree based in protein sequences of E93 from 24 different insect taxa including the E93 sequence of Xenos vesparum described in this study. Two Arachnida E93 sequences are used as outgroup. Branch colors at the tip of the tree indicated the order of the different species used in this analysis. X. vesparum is highlighted in bold.

Differential expression of XvE93 in males and females of X. vesparum

As a first step towards the characterization of E93 in neoteny, we measured the expression levels of XvE93 in the larva-to-adult transition of X. vesparum. Males of X. vesparum progress through four larval stages (one free-living planidium stage and three endoparasitic stages) before undergoing endoparasitic pupation, when the metamorphic transition results in the formation of free-living winged adults (Fig. 1). In contrast, X. vesparum females pass through four larval molts (one free-living planidium stage and three endoparasitic stages) which then transform directly into endoparasitic larviform neotenic animals without transiting through the metamorphic-pupal stage (Fig. 1). As shown in Fig. 4A, mRNA expression levels of XvE93 in males were low in 4th larval instar but significantly increased in the pupal and adult stages. In contrast, XvE93 levels in females were very low in 4th instar larvae and did not show any detectable increase in neotenic adults (Fig. 4A).

Figure 4
figure 4

Developmental expression profiles of Xenos vesparum XvE93 in the transition from larvae to adults. (A) XvE93 mRNA levels were measured by qRT-PCR in the larva-to-adult transition in male (4th, pupa, adult) and female (4th and adult) whole bodies. (B) XvE93 mRNA levels (qRT-PCR) in the whole body of 3th, early and late 4th and neotenic females. Transcript abundance values in panels A and B are normalized against the XvRibosomal 18S transcript. The range of expression is shown by Boxplot: boxes demarcate the upper and lower quartiles, while the heavy bar indicates the median value of normalized expression. Whiskers extend to the most extreme values. Points represent each individual measurement. Different letters in panel B represent groups with significant differences according to ANOVA test (Tukey, p ≤ 0.005).

To characterize in more detail the expression pattern of XvE93 in X. vesparum females, we next measured XvE93 mRNA levels in staged 3rd, early and late 4th instar larvae as well as in neotenic female adults. As expected, mRNA expression levels of XvE93 were very low in 3rd and at the beginning of the 4th larval instars, but, surprisingly, the levels of XvE93 were significantly up-regulated at the final part of the 4th larval instar and the adult stage (Fig. 4B). This stage-specific up-regulation of XvE93 is surprising in that, as previously stated, neotenic X. vesparum adult females are characterized for the lack of complete metamorphosis and the retention of an overall juvenile morphology. However, the female cephalothorax presents some adult-specific features (Fig. 5A), such as the pheromone glands and structures that support insemination and the release of the 1st instar planidia. This observation raised the possibility that the late increase in XvE93 levels would be related with the development of such structures. To test this hypothesis, we measured the transcript levels of XvE93 separately in cephalothorax and abdomen of the late last juvenile instar, when XvE93 was up-regulated (Fig. 4B), and the neotenic-adult stage. Remarkably, as shown in Fig. 5B, XvE93 was significantly up-regulated in the cephalothorax of 4th instar larvae and neotenic females but not in the abdominal part. This confirms the correlation between the specific increase in XvE93 expression in the cephalothorax with the occurrence of adult-specific structures in this part of the neotenic female.

Figure 5
figure 5

Expression levels of XvE93, XvHR3, XvE75, and XvBr-C in the cephalothorax and abdomen of Xenos vesparum females. (A) Dorsal views of a 4th instar larva and a neotenic adult female. The abdomen (abd.) and the cephalothorax (cep.) regions are marked. (BE) Transcript levels of (B) XvE93, (C) XvHR3, (D) XvE75, and (E) XvBr-C (without discriminating isoforms) were measured by qRT-PCR in the cephalothorax and abdomen of 4th instar larvae and neotenic adult females. Transcript abundance values in panels B-E are normalized against the XvRibosomal 18S transcript. The range of expression is shown by Boxplot: boxes demarcate the upper and lower quartiles, while the heavy bar indicates the median value of normalized expression. Whiskers extend to the most extreme values. Points represent each individual measurement. Asterisks indicate differences statistically significant at *p ≤ 0.05; **p ≤ 0.01; and ***p ≤ 0.001 (t-test).

Next, we investigated how the expression of XvE93 is differentially regulated in neotenic females. Since E93 expression is induced by 20E in holometabolous insects35,48,52,53, the differential expression could be explained by differences in the signaling, or response to 20E in different parts of X. vesparum female body. To test this, we analyzed the differential expression of XvHR3 and XvE75 (GenBank accession numbers MH220843 and MH220844, respectively), two 20E-dependent nuclear receptors of the stereotypic genetic cascade that responds to the 20E signal in insects47,54. Remarkably, the mRNA levels of both factors were higher in the cephalothorax of female 4th instar larvae compared to the abdominal region (Fig. 5C,D). Higher levels of XvHR3 and XvE75 remained in the cephalothorax of neotenic females, although the levels of both factors were very low (Fig. 5C,D). Overall, these results suggest that higher levels of 20E signaling in the cephalothorax of last instar X. vesparum female larvae could be responsible for the particular increase in the expression of XvE93 in this region, which, in turn, results in the “metamorphic” transformation of the anterior part of the body.

In all holometabolous insects studied to date, E93 is highly expressed specifically during the pupal and adult stages1,34,35,48,53. Functional studies have shown that E93 is required during the pupal period for proper adult differentiation, as RNA interference (RNAi)-mediated depletion of E93 prevented pupal-adult transition in D. melanogaster or even produced a supernumerary second pupa in T. castaneum1,34. Similar results were also observed in the lepidopteran B. mori48. Remarkably, the function of E93 is conserved in hemimetabolous insects34,55. In the hemimetabolous cockroach B. germanica, BgE93 is highly expressed in metamorphic tissues during the last nymphal instar, and RNAi-mediated knockdown of BgE93 in the nymphal stage prevented the nymphal–adult transition, inducing endless reiteration of nymphal stages34. Due to the evolutionary conservation of its expression and function, E93 is considered the “master factor” of adult metamorphosis in winged insects34. Importantly, the expression dynamic of XvE93 in X. vesparum males reported in our work is consistent with the expression of E93 in all holometabolous insects analyzed to date. In addition to the expression pattern of XvE93 in males, it has been previously suggested that the expression of the pupal-specifier XvBr-C presents a consistent increase during 4th larval instar and early pupa stages in X. vesparum males36, an expression pattern that is consistent with that observed in other holometabolous insects28,29,32,56,57. Altogether, these data show that male X. vesparum displays the typical holometabolous metamorphic development characterized by the sequential occurrence of Br-C during the larval-pupal transition and of E93 during the pupal-adult transition. In contrast, in females we observed very low levels of XvE93 in the 4th instar larvae and in neotenic females. Likewise, the expression of XvBr-C in females does not show a detectable induction in the same developmental period36. The absence in neotenic females of the typical holometabolous expression pulses of XvE93 (this paper) and XvBr-C36 during the larval-to-adult transition correlates with the lack of the pupal stage, and clearly suggests that the suppression of the pupal determination program has contributed to the occurrence of female neoteny in X. vesparum.

It is interesting to note, however, that although X. vesparum neotenic females retain an overall larviform appearance at the posterior part of the body, the anterior extruded part of the body, the cephalothorax, undergoes a “metamorphic-like” transformation between the 4th larval instar and the adult stages without an intermediate pupal stage. This transformation includes the development of structures to aid insemination, the release of the planidia as well as the pheromone glands20. Remarkably, we have found that the occurrence of such “metamorphic-like” process in the cephalothorax correlates with the significant increase of XvE93, XvHR3 and XvE75 expression in this part of the body (Fig. 5). Interestingly, we observed that XvBr-C was also up-regulated in the cephalothorax of 4th instar larvae and neotenic females compared to the abdominal part (Fig. 5E), suggesting that a differential increase in 20E signaling in the cephalothorax results in higher expression levels of XvE93 and XvBr-C, which controls the development of adult features. Altogether, our results show that the suppression of XvE93 and XvBr-C expression in X. vesparum females, especially in the abdominal part of the body, may underlie the development of female neoteny in Strepsiptera.

A change in the expression of key developmental genes seems to be also underlying the evolution of progenesis, a second developmental strategy of paedogenesis that consists in the early growth and maturation of the ovaries during larval development. In the dipterans Heterozepa pygmaea and Mycophila speyeri, for example, a precocious up-regulation of the heterodimer EcR/USP in the ovaries of first instar larvae correlates with the early maturation of the ovary within the larval body in paedogeneic females58. Therefore, critical variations in the expression levels of genes that are important for the transduction of JH and 20E hormonal signaling may be responsible for the occurrence of paedogenesis in insects.

Expression of XvKr-h1 during metamorphic and neotenic development in X. vesparum

Finally, to characterize the expression of the genes that form the MGN, we identified the transcription factor XvKr-h1 in this species. To this aim, we isolated a fragment that encompasses seven zing-fingers of the DNA-binding domain through the assembling of X. vesparum transcriptomes (GenBank accession number MH220842) (Supplementary Fig. 2). We, then, analyzed the expression of XvKr-h1 during the larval-adult transition in X. vesparum. In males, XvKr-h1 mRNA levels were low in 4th instar larvae, almost undetectable in the pupal period, and significantly high in the adult (Fig. 6A). In contrast, consistent with the absence of pupal stage in the neotenic female, XvKr-h1 mRNA levels did not show differences between 4th instar larva and the neotenic female (Fig. 6A). As before, we next analyzed in detail the expression levels of XvKr-h1 in females and found that XvKr-h1 was clearly detected in 3rd instar larvae, but its levels significantly dropped at the onset of the 4th larval instar, only to reappear at the final part of the instar (Fig. 6B). In the neotenic adult female, low levels of XvKr-h1 were detected (Fig. 6B). Interestingly, although not significant, the increase in XvKr-h1 expression in late 4th larval instar was higher in the cephalothorax compared to the abdominal region, while in neotenic animals very low levels of XvKr-h1 were detected in both parts of the body (Fig. 6C).

Figure 6
figure 6

Developmental expression profiles of Xenos vesparum XvKr-h1 in the larva to adult transition. (A) XvKr-h1 mRNA levels were measured by qRT-PCR in the larva-to-adult transition in male (4th, pupa, adult) and female (4th and adult) whole bodies. The inset panel shows a magnification of expression levels in 4th larval instar and pupa. (B) XvKr-h1 mRNA levels (qRT-PCR) in the whole body of 3th, early and late 4th and neotenic females. (C) XvKr-h1 mRNA levels (qRT-PCR) in the cephalothorax (cep.) and abdomen (abd.) of 4th instar larvae and neotenic adult females. Transcript abundance values in all panels are normalized against the XvRibosomal 18S transcript. The range of expression is shown by Boxplot: boxes demarcate the upper and lower quartiles, while the heavy bar indicates the median value of normalized expression. Whiskers extend to the most extreme values. Points represent each individual measurement. Different letters in panel B represent groups with significant differences according to ANOVA test (Tukey, p ≤ 0.005). Asterisks in panel C indicate differences statistically significant at **p ≤ 0.01 (t-test).

Similar to XvE93, the expression pattern of XvKr-h1 in the male is comparable to those observed in holometabolous insects, in which Kr-h1 is expressed through larval development to become strongly repressed by E93 during the pupal stage, and to reappear during the adult stage1,26,27,59. On the other hand, the levels of XvKr-h1 in females display a particular dynamic: at the early last larval stage, the decline of XvKr-h1 is similar to that observed at the onset of the last larval instar of holometabolous insects; however, at the final part of the 4th instar, XvKr-h1 is up-regulated in the cephalothorax region in parallel to the increase in the expression of XvE93 and XvBr-C. This result suggests that in the neotenic female, XvKr-h1 has lost the strong repressive activity upon XvE93 and XvBr-C expression that is observed in holometabolous insects studied1,35. Alternatively, it is plausible that the up-regulation of XvKr-h1 in 4th instar larva is necessary to prevent a stronger transcriptional induction of XvE93 that would result in the activation of the adult genetic program in the body of female X. vesparum.

Concluding Remarks

Loss of complete metamorphosis in the strepsipteran suborder Stylopidia has evolved exclusively in females. Males, however, display the typical holometabolous larva-pupa-adult transition. Correlating with the presence of the metamorphic pupal stage in X. vesparum males, high levels of expression of the adult specifier XvE93, along with the disappearance of the anti-metamorphic XvKr-h1, are detected in the pupal stage. In contrast, very low levels of XvE93 are observed throughout development of X. vesparum females. However, a specific up-regulation of XvE93 and XvBr-C in the cephalothorax of late female 4th instar larva correlates with the occurrence of adult-specific features in the anterior part of the neotenic female body. Overall, our results, together with previous work36, suggest that neoteny in strepsipteran females arose by the suppression of the genetic program that controls the formation of the pupa, that is the sequential expression of XvBr-C and XvE93 factors at the end of larval development. The loss of expression of XvE93 and XvBr-C in X. vesparum females could be the result of changes in the endocrine milleu, with alterations in the titer of JH and/or 20E in particular stages of development. There might also be changes in the regulation of the expression of these genes, as suggest by the fact that XvKr-h1 does not seem to act as a potent repressor of XvE93 and XvBr-C in females. In summary, our work highlights the importance of the MGN in the regulation of insect metamorphosis and in the evolution of insect neoteny in the order Strepsiptera.