Transcriptome of pleuropodia from locust embryos supports that these organs produce enzymes enabling the larva to hatch
Pleuropodia are limb-derived glandular organs that transiently appear on the first abdominal segment in embryos of insects from majority of “orders”. They are missing in the genetic model Drosophila and little is known about them. Experiments carried out on orthopteran insects 80 years ago indicated that the pleuropodia secrete a “hatching enzyme” that digests the serosal cuticle to enable the larva to hatch, but evidence by state-of-the-art molecular methods is missing.
We used high-throughput RNA-sequencing to identify the genes expressed in the pleuropodia of the locust Schistocerca gregaria (Orthoptera). First, using transmission electron microscopy we studied the development of the pleuropodia during 11 stages of the locust embryogenesis. We show that the glandular cells differentiate and start secreting just before the definitive dorsal closure of the embryo and the secretion granules outside the cells become more abundant prior to hatching. Next, we generated a comprehensive embryonic reference transcriptome for the locust and used it to study genome wide gene expression across ten morphologicaly defined stages of the pleuropodia. We show that when the pleuropodia have morphological markers of functional organs and produce secretion, they are primarily enriched in transcripts associated with transport functions. They express genes encoding enzymes capable of digesting cuticular protein and chitin. These include the potent cuticulo-lytic Chitinase 5, whose transcript rises just before hatching. Unexpected finding was the enrichment in transcripts for immunity-related enzymes. This indicates that the pleuropodia are equipped with epithelial immunity similarly as barrier epithelia in postembryonic stages.
These data provide transcriptomic support for the historic hypothesis that pleuropodia produce cuticle-degrading enzymes and function in hatching. They may also have other functions, such as facilitation of embryonic immune defense. By the genes that they express the pleuropodia are specialized embryonic organs and apparently an important though neglected part of insect physiology.
KeywordsAppendage Cuticle Ecdysone Embryo Gland Immunity Moulting fluid Orthoptera RNA-seq Schistocerca gregaria
First abdominal segment
Differentially expressed gene
- EC1, EC2, EC3
The first, the second, the third embryonic cuticle, respectively
Principal component analysis
Reads per kilobase of transcript per million reads mapped
Scanning electron microscopy
Third thoracic segment
Transmission electron microscopy
Eighty years ago Eleanor Slifer [30, 31] demonstrated that the pleuropodia of the grasshopper Melanoplus differentialis are necessary for the digestion of serosal cuticle (SC) before hatching. SC is a chitin and protein-containing sheet structurally similar to larval and adult cuticles and is produced by extraembryonic serosa in early embryogenesis [32, 33]. SC makes a layer just under the chorion and forms a sac-like structure around the embryo and yolk. Shortly before hatching the inner layer of SC (procuticle), which forms the major part of the cuticle, disappears. When Slifer  removed pleuropodia from embryos, SC remained thick and the hatching larva could not break through it to get out of the egg. She proposed that pleuropodia secrete the “hatching enzyme”. The exact molecular composition of this substance is unknown, but we may assume that it is similar to the cuticle degrading moulting fluid (MF) that is released by larval epidermis under the old cuticle when the insect is preparing to moult .
In a few insects the ultrastructure of the cells in the pleuropodia was examined by transmission electron microscopy (TEM). These studies showed that the organs are primarily formed by an epithelium with morphological features of transporting and secretory epithelia [20, 25, 35, 36, 37, 38, 39]. In some insects, including orthopterans, the cells produce some secretion, but it is not clear if this is equivalent to the “ecdysial droplets”  carrying the MF. Some of Slifer’s experiments  were successfully repeated on other orthopterans  and an extract from pleuropodia was capable of digesting pieces of SC , but validation by genetic methods, such as that the pleuropodia express genes for cuticle-degrading enzymes, is missing.
Endocrinologists Novak and Zambre  questioned Slifer’s conclusions by arguing that according to her hypothesis SC would be digested in an unusual way compared to a typical larval cuticle. During larval moulting epidermal cells deposit a cuticle and subsequently the same epidermal cells, not a special gland, secrete MF. Therefore, they  proposed that the SC degrading enzymes would most probably be secreted by the serosa itself. Based on their studies in the locust (swarming grasshopper) Schistocerca gregaria, they suggested that the pleuropodia reach the peak of their activity in young embryos during katatrepsis and stimulate the serosa, to secrete the “hatching enzyme”. At katatrepsis the serosa is still present, but from that on it starts to shrink and completely degenerates by the time of dorsal closure in mid-embryogenesis . Novak and Zambre  proposed that the stimulating substance released by the pleuropodia is likely a small molecule with the properties of the moulting hormone ecdysone or ecdysone itself and they carried out experiments in support of that. When a homogenate from pleuropodia was injected into the abdomen of the final instar larva of Drosophila that had been isolated (by a ligature) from the ecdysone producing prothoracic gland, the cuticle darkened like at pupariation. A similar effect was achieved by injection of ecdysone.
As a first step towards understanding the function of the pleuropodia we identified the genes that they express. We chose S. gregaria as a model, because first, it has large embryos (eggs > 7 mm long) and external pleuropodia that can easily be dissected out, second, previous experimental studies addressing the function of pleuropodia were carried out in orthopterans, and third, it is a model pest that is also used for physiological and developmental-genetic studies. Using TEM we were able to examine 11 developmental stages of the pleuropodia at high resolution. This helped us to identify when exactly the organs are fully developed and produce secretion. No detailed staging system using TEM exists for the pleuropodia of S. gregaria or any other orthopteran. Using high-throughput RNA sequencing (RNA-seq) we sequenced transcriptomes of ten morphologically defined stages of the pleuropodia and similarly aged hind legs and performed differential gene expression analysis between the two appendages. This enabled us to identify the transcripts enriched in the pleuropodia. The leg was chosen for comparison because the pleuropodium is a modified leg. The goal of this paper was to investigate whether the gene expression profile of the fully developed and secreting pleuropodia supports that these organs produce the “hatching enzyme”. We show that the pleuropodia of S. gregaria indeed express genes for some of the cuticle degrading enzymes previously identified in the MF. This brings a trancriptomic support for the Slifer’s historic hypothesis [30, 31]. As a part of our study we assembled a full embryonic transcriptome of S. gregaria, whose genome has not been sequenced yet.
Development of pleuropodia in the course of S. gregaria embryogenesis
We traced cell divisions in the pleuropodia by using Phospho-Histone H3 as a marker (Fig. 2c). The glandular cells were labeled only in the days 4 and 5. From day 6 onwards no cell divisions were detected and the nuclei started to enlarge as the cells became polyploid .
As development progresses the secretion granules (inside and outside the cells) become more abundant and are present also above the microvilli (Fig. 3k-p). On day 12 the apical side of the glandular cells changes: clusters of microvilli (usually at the borders between cells) elevate (Fig. 3n). Later the cells show signs of degeneration, the chromatin condenses and the cell content becomes disorganized (Fig. 3o,p). Large secretion granules are still abundant and probably released even on the last day before hatching, when the pleuropodia have shrunk and collapsed (Figs. 2b,3p).
At hatching, the larva enclosed in the apolysed second embryonic cuticle (EC2) leaves the eggshell and digs through the substrate up to the surface [46, 47]. Here the EC2 is shed and the degenerated pleuropodia are removed with it (; Fig. 2a).
Therefore, the timing of the high secretory activity (from ±69 % DT until just before hatching) corresponds to the stages when Slifer  demonstrated the presence of the “hatching enzyme” (Fig. 2a). By contrast, pleuropodia from embryos around katatrepsis (± 41–48 % DT) that Novak and Zambre  used for their experiments do not appear fully differentiated and secretory yet. Next we looked at the genes expressed in the pleuropodia.
Isolation of genes expressed in the pleuropodia of S. gregaria using RNA-seq
Prior to the gene expression analysis we prepared a comprehensive embryonic transcriptome that served as a reference (see Methods). It consists of 20,834 transcripts (Additional file 2: Table S1) and we aimed each transcript to represent one gene. The completeness of the transcriptome was assessed using the open-source software BUSCO (version 3) [48, 49]. 95.6, 96.3 and 94.6 % of the Metazoa, Arthropoda and Insecta orthologs, respectively, were found, a level comparable to published “complete” transcriptomes.
Top 10 % of the most abundant transcripts upregulated in the highly secreting pleuropodia
Uncharacterized, contains DUF4773 domain
Chitin binding Peritrophin-A
perotrophic matrix protein
Alpha-tocopherol transfer protein
intermembrane lipid transfer
Serine protease, Snake-like
proteolysis, Toll signaling
cuticular chitin degradation
Neutral endopeptidase 24.11
serine protease inhibitor
Glycosyl hydrolase, Myrosinase 1-like
Alpha-tocopherol transfer protein
intermembrane lipid transfer
Serine protease, H2-like
Pro-phenol oxidase subunit 2
Serine protease, Easter-like
Glutamate dehydrogenase mitochondrial
nitrogen and glutamate metabolism
proteolysis, lysosomal enzyme
Insect pheromone-binding protein A10/OS-D
Cytochrome P450 CYP4G102
synthesis of hydrocarbons, anti-dehydration
Sensory neuron membrane protein, 1-like
Lipopolysaccharide-induced tumor necrosis factor-alpha factor homolog
serine protease inhibitor, melanization
cuticular chitin degradation
Serine protease, Snake-like
Uncharacterized, contains Transcription activator MBF2 domain
Putative serine protease, K12H4.7-like / Serine carboxypeptidase
Uncharacterized, contains DUF3421 domain
antiproliferative protein, transcription corepressor
Lysosomal-associated membrane protein
lysosomal membrane protein
scaffolding protein in cell membrane
Leucine rich repeat
Fatty acyl-CoA reductase, waterproof-like
V-type proton ATPase proteolipid subunit
proton transporting ATPase
Bax inhibitor 1
negative regulation of apoptosis and autophagy
Sodium/potassium-transporting ATPase subunit alpha
sodium:potassium exchanging ATPase
Serine protease, Easter-like
RNA binding, sterol metabolism
iron ion transport, iron sequestration
Uncharacterized serine protease inhibitor
serine protease inhibitor
Phosphoenolpyruvate carboxykinase [GTP]
Alpha-tocopherol transfer protein
intermembrane lipid transfer
Gram-negative bacteria binding protein 3
The upregulated genes are primarily enriched in GO terms (Fig. 6, Additional file 2: Table S7) associated with active transport, which is consistent with that the pleuropodia have morphological characteristics of transporting organs. Genes for both V-ATPase and Na+, K+ ATPase are upregulated (Additional file 2: Table S10). We found enriched GO terms linked with lysosome organization, consistent with the observation that the pleuropodia contain numerous lysosomes (Fig. 3, ). We also found a cluster of GO terms associated with lipid metabolism, consistent with the presence of smooth endoplasmic reticulum in the cells (Fig. 3). Therefore, the genes differentially expressed between legs and pleuropodia are in agreement with the morphology of the organs.
Insect cuticle, such as SC, is digested by a cocktail of enzymes that degrade chitin, which is an aminoglycan polymer, and proteins [34, 55, 56]. In support of Slifer’s experiments demonstrating that pleuropodia produce cuticle degrading enzymes we found that these organs upregulate genes associated with carbohydrate derivative metabolism, aminoglycan catabolic process and proteolysis. A novel interesting finding was the upregulation of genes associated with immunity. Next we looked at particular genes in a detail.
Pleuropodia upregulate genes for cuticular chitin degrading enzymes
Cuticular chitin is hydrolyzed by a two-enzyme system composed of a β-N-acetyl-hexosaminidase (NAG) and a chitinase (CHT) . Both types of enzymes, a NAG and a CHT, have to be simultaneously present for efficient hydrolysis of chitin .
RNA-seq differential gene expression of cuticular chitin-degrading enzymes in the highly secreting pleuropodia
S. gregaria gene
124 (15.88 %)
46 (5.89 %)
592 (75.8 %)
306 (39.18 %)
I-Major "moulting" chitinases
15 (1.92 %)
400 (51.21 %)
III-Cuticle assembly chitinases
IV-Gut, fat body and other chitinases
188 (24.07 %)
V-Imaginal disc growth factors
391 (50.06 %)
The insect CHTs have been classified into several groups [56, 60], of which the major role in the digestion of cuticular chitin is played by Chitinase 5 and (perhaps with a secondary importance) by Chitinase 10 [61, 62] (Table 2; the classification of CHTs into five major groups that we use here is based on ). Some chitinases, for example, are expressed in the gut, trachea and fat body, where they are likely involved in digestion of dietary chitin, turnover of peritrophic matrix and immunity, some epidermal chitinases organize assembly of the new cuticle (e.g., [60, 63, 64, 65, 66]).
Our transcriptome contains 16 full or partial transcripts of CHTs representing all of the major CHT groups (Table 2, Additional file 1: Figure S5b, S6b). The pleuropodia specifically upregulate both of the genes for Chitinase 5, which show homology with cht5–1 and cht5–2 from the locust Locusta migratoria . One of the transcripts, Sg-cht5–1, was among the top 15 most abundant transcripts upregulated in the highly secreting pleuropodia (Table 1). The predicted amino acid sequence of Sg-cht5–1 contains a conserved catalytic domain and a signal peptide, and thus is likely to be active and secreted, respectively (Additional file 1: Figure S5b). The other upregulated CHTs were homologs of cht2 and idgf. By contrast, the S. gregaria homolog of cht10 that also has a role in cuticular chitin hydrolysis and required for larval moulting [62, 64] had low expression in both legs and pleuropodia.
We next focused on the transcript of the major chitinase, Sg-cht5–1. Unlike the NAGs, both RNA-seq and real-time RT-PCR have shown that the expression of this CHT is low in the early secreting stages, rises only later around day 12 and reaches highest levels when the pleuropodia are already degenerating (day 13 and 14) (Fig. 7 e,f,e'). Also real-time RT-PCR on cut embryos has shown that the pleuropodia are a major source of the Sg-cht5–1 mRNA on day 12 but not before (the high expression in the whole embryo on day 10 could be linked to the second embryonic moult and was also observed with Sg-cht7, although not with Sg-cht10, Additional file 1: Figure S7). These data show that the pleuropodia before hatching express a cuticle-degrading chitinase.
Pleuropodia upregulate transcripts for some proteases that could digest a cuticle
MF proteases that were upregulated in the highly secreting pluropodia
S. gregaria transcript IDc
Fold change UP
Neutral endopeptidase 24.11
Ecdysteroid-inducible angiotensin-converting enzyme
Digestive cysteine protease 1, cathepsin L
Serine protease HP21 precursor
Trypsin-like serine protease - fibroin heavy chain
Serine protease, Easter-like
Serine protease 1
Serine protease, Snake-like
Pleuropodia are enriched in transcripts for immunity-related proteins
In total we found upregulated 99 transcripts (13 % of the upregulated genes) for immunity-related proteins. These include proteins at all three levels, the pathogen recognition, signaling and response (Fig. 9, Additional file 2: Table S13). From the four signaling pathways, Toll was upregulated, but not IMD or JAK/STAT, and from the JNK signaling we found c-Jun. Genes for a range of immune responses were upregulated, including production of reactive nitrogen species (RNS), melanization, genes for lysozymes and one antimicrobial peptide (AMP) similar to Diptericin.
RNA-seq differential gene expression of S. gregaria lysozymes in the highly secreting pleuropodia
2 (0.26 %)
37 (4.74 %)
7 (0.90 %)
76 (9.73 %)
1251 (81.50 %)
150 (19.21 %)
Pleuropodia do not upregulate the pathway for ecdysone biosynthesis
Previous work has suggested that pleuropodia may be embryonic organs producing the moulting hormone ecdysone . During post-embryonic stages, ecdysone is synthesized in the prothoracic glands and several other tissues by a common set of enzymes [75, 76, 77], some of which have been characterized in the locusts [78, 79, 80, 81]. As shown in Drosophila, these genes are expressed in diverse cell types in embryos, and when the larval prothoracic glands are formed their expression co-localizes there [82, 83, 84, 85, 86].
RNA-seq differential gene expression of S. gregaria ecdysone biosynthesis enzymes in the highly secreting pleuropodia
shade (shd), Cyp314A1
shadow (sad), Cyp315A1
disembodied (dib), Cyp302A1
431 (55 %)
phantom (phm), Cyp306A1
spook (spo), Cyp307A1
1368 (89 %)
Pleuropodia of S. gregaria release secretion granules from shortly before dorsal closure until hatching
Using TEM we showed that the glandular epithelium in the pleuropodia of S. gregaria fully develops shortly before dorsal closure, on day 8 in our staging (55 % DT), when granular secretion outside the cells becomes visible. On day 6 and 7, which surround katatrepsis (45 % DT), the glandular cells only begin to differentiate and do not secrete. This would explain why no digestive effect on SC was detected by Novak and Zambre  when using a homogenate from S. gregaria pleuropodia isolated at around katatrepsis.
Pleuropodia at stages before and after katatrepsis were previously examined by TEM in other orthopterans, Eyprepocnemis plorans, Ailopus thalassinus and Ailopus strepens . Unlike our observations on S. gregaria, this study detected release of secretion granules already in stages preceding katatrepsis. Their images (presence of secretion granules outside cells that have long microvilli, thick layer of fibres released from the tips of microvilli forming the EC1 cuticle) corresponds to developmentally more advanced pleuropodia, of S. gregaria, around day 8 or 8–9 (compare e.g., Fig. 3a in , before katatrepsis with our Fig. 3h,i, before and just after katatrepsis). This might be because in different species the glandular cells differentiate at different speed or the samples studied by Viscuso and Sottile  were taken from embryos that were slightly older than described. The fine staging and sampling at precise developmental age in our study on S. gregaria supports that pleuropodia of orthopterans (Orthoptera: Caelifera) are not yet functional at katatrepsis. All pleuropodia that we examined from several 6- and 7-day embryos (before and after katatrepsis) gave similar results per stage (see Methods). Intense secretion from pleuropodia of S. gregaria that we observed after dorsal closure and close to hatching is in agreement with other TEM studies on orthopterans E. plorans, A. thalassinus, A. strepens  and Locusta migratoria .
Pleuropodia of S. gregaria express genes for the “hatching enzyme”
Our RNA-seq analysis revealed that the secreting pleuropodia highly express genes encoding enzymes that are capable of digesting a typical chitin-protein insect cuticle. These include genes for proteolytic enzymes similar to those present in the MF and cuticular chitin-degrading NAGs and Chitinase 5. The pleuropodia also express genes for Chitinase 2 and Idgf, which have low effect on cuticular chitin digestion, but were shown to organize proteins and chitin fibres during cuticle deposition and for Idgf to have an immune function [63, 64]. These CHTs may organize the fibres in the cuticle secreted by the pleuropodia (Fig. 3).
We showed that, while the expression of the Sg-nag1 and Sg-nag2 started to rise in parallel with the differentiation of the glandular cells, the Sg-cht5–1 and Sg-cht5–2 transcripts raised shortly before hatching. Chitinase 5 is a critical chitin-degrading chitinase in insects: it is highly abundant in the moulting fluid and its silencing in diverse insects including the locust L. migratoria leads to failure in larval moulting [55, 62, 64, 67, 88] and as shown in L. migratoria also in moulting of the embryonic cuticle . Our data indicate that the sudden rise in the expression of Sg-cht5 in the pleuropodia at the end of embryogenesis and presumably secretion of this CHT into the extraembryonic space is an important component of the “hatching enzyme” effect [30, 31] in locusts and grasshoppers. Since silencing of this single gene in embryos of L. migratoria did not prevent hatching , we conclude that the other chitin and protein degrading enzymes produced in the pleuropodia, and perhaps the serosa (see below), are essential as well.
Pleuropodia in some other insects could secrete the “hatching enzyme” and their function may also vary among species
There is evidence to suggest that the process occurs similarly in some insects. As in orthopterans, the pleuropodia of the rhagophthalmid beetle Rhagophthalmus ohbai release secretion after katatrepsis and SC rapidly degrades just shortly before hatching . In the large water true bugs from the family Belostomatidae, the male carries a batch of eggs on his back. It is believed that the detachment of the eggs just before hatching is also caused by the secretion from the pleuropodia .
The molecular mechanism of SC degradation may also vary between insects and as previously hypothesized  the serosa may contribute to the SC degradation. The serosa of the beetle Tribolium castaneum, expresses cht10 and cht7 , of which the former CHT is important for cuticular chitin digestion. Silencing of cht10, but not cht5 prevented the beetle larvae from hatching . Transcripts for cht10 were not upregulated in the pleuropodia of S. gregaria. This suggests that the SC is degraded by enzymes produced by both, the serosa and the pleuropodia and that the indispensable roles in cuticle digestion are played by different enzymes in different insects. Likely, the serosa, before it degenerates in mid-embryogenesis, releases some of the SC degrading enzymes, but these do not make up the complete cocktail that would be able to digest the cuticle efficiently. The secretion from the pleuropodia than adds the missing enzymes.
In some insects the pleuropodia may not be involved in hatching but have another function. In the viviparous cockroach Diploptera punctata , the secretion from the pleuropodia is very low and the large pleuropodia of the melolonthid beetle Rhizotrogus majalis have not been observed to produce any secretion granules at all . In dragonflies, one of the more basal lineages of insects, the secretion likely has a different function than in orthopterans, because their SC is not digested before hatching . The special epithelium in the pleuropodia shares features with transporting epithelia [36, 38] that function in water transport and ion balance . Our data do not exclude this function, but it is yet to be tested. Our data do not support that the pleuropodia specialize in producing the moulting hormone  because they do not upregulate the whole set of the moulting hormone ecdysone biosynthetic enzymes. However, it cannot be excluded, that the pleuropodia produce some ecdysone intermediates; particularly products of biosynthetic steps catalyzed by the dib and spo enzymes that we found upregulated at some stages.
The pleuropodia of S. gregaria are enriched in transcripts for enzymes functioning in immunity
We found that many of the genes expressed in the pleuropodia encode proteins involved in immunity . This indicates that the pleuropodia are also organs of epithelial immunity, similar to other barrier epithelia in postembryonic stages (such as the gut) , which are in a constant contact with microorganisms. The pleuropodia differ from such tissues in that they are not directly exposed to the environment, but enclosed in the eggshell, seemingly limiting their contact with microorganisms. Proteins associated with immune defense are also found in the MF , where they prevent invasion of pathogens through a “naked” epidermis after the separation of the old cuticle from the epidermis in the process of apolysis. As found in the beetle T. castaneum, during the early embryonic stages the frontier epithelium providing the egg with an immune defense is the extraembryonic serosa . The serosa starts to degenerate after katatrepsis and disappears at dorsal closure . The pleuropodia of S. gregaria differentiate just before dorsal closure, suggesting that they take over this defense function in late embryogenesis. It will be interesting to clarify in the upcoming research whether apart from their role in hatching the pleuropodia are important organs for fighting against potential pathogens that have gained access to the space between the embryo and the eggshell.
Evolution of the pleuropodia
Although only the functions of pleuropodia in digestion of SC and production of the moulting hormone in orthopterans were supported experimentally, current little data indicate that the function of these organs has been changing during evolution. In the diverse insect lineages that have them, the ultrastructure of the cells in pleuropodia appear similar, but the organs have different shapes. Hypothetically then, the glandular and/or water and ion transporting cells [20, 25, 35, 36, 37, 38, 39] were adapted to suit the needs of the developing embryo, such as digestion of the tough SC in orthopterans  or exchanging fluids and ions by particularly elongated pleuropodia in viviparous cockroaches . In advanced insect lineages this type of cells is likely not needed and the pleuropodia do not develop. Future research by modern techniques is needed to uncover the diverse functions of these ancient insect organs and their role in insect evolution, physiology and development.
Transcriptomic profiling of pleuropodia from S. gregaria supports that the conclusions that Eleanor Slifer drew from her experiments over 80 years ago that these organs secrete cuticle degrading enzymes, were correct. The pleuropodia likely have other functions, such as in immune defense. The pleuropodia appear to be true live embryonic organs and likely an important but neglected part of insect physiology. The sequencing data that we generated will in future studies enable to dissect the functions of these enigmatic organs in a detail.
Schistocerca gregaria (gregarious phase) originated from a long-term colony at the Department of Zoology, University of Cambridge. Eggs were collected into pots with damp sand in two- or four-hour intervals. The pots with eggs were stored in an incubator at 30 °C, in constant darkness and the sand was kept moist.
Imaging of embryonic stages
Embryos and appendages were dissected in phosphate buffer saline (PBS). Whole eggs were treated with 50 % household bleach to dissolve the chorion. All were photographed using the Leica M125 stereomicroscope equipped with DFC495 camera and associated software. Photos were processed using Adobe Photoshop CC 2017.1.1. Photos of eggs and embryos (Fig. 2a and Additional file 1: Figure S1) had the background cleaned using the software.
Immunohistochemistry on paraffin sections
Embryos were dissected in PBS and pieces including A1 were fixed in PEMFA (4 % formaldehyde in PEM buffer: 100 mM PIPES, 2.0 mM EGTA, 1.0 mM MgSO4) at room temperature (RT) for 15–30 min, washed in PBT (PBS with 0.1 % Triton-X 100) and stored in ethanol at − 20 °C. Samples were cleared 3 × 10 min in Histosol (National Diagnostics) at RT, infiltrated with paraffin at 60 °C for 2–3 days and hardened in moulds at RT. Sections 6–8 μm thick were prepared on a Leica RM2125RTF microtome. Slides with sections were washed with Histosol, ethanol and re-hydrated to PBT. Slides were placed in a humidified chamber, blocked with 10 % sheep serum (Sigma-Aldrich) in PBT for 30 min at RT and incubated with Phospho-Histone H3 antibody (Invitrogen) diluted 1:130 at 4 °C overnight, Alexa Fluor 568 anti-rabbit secondary antibody (Invitrogen) diluted 1:300 at RT for 2 h and DAPI (Invitrogen) diluted 1:1000. Sections were imaged with a Leica TCS SP5 confocal microscope and photos processed using Fiji (https://fiji.sc).
For TEM embryos were dissected in PBS and pieces including A1 were fixed in 2.5–3.0 % glutaraldehyde in 0.1 M phosphate buffer (PB) pH 7.2 for a few hours at RT and at 4 °C for several days. Pleuropodia and legs were separated from other tissues and embedded into 2 % agar as previously . Appendages in agar were incubated in solution of 3 % OsFeCN in cacodylate buffer with 4 mM CaCl2 for 1–2 days at 4 °C, 0.01 mg/ml thiocarbohydrazide (Sigma-Aldrich) for 20–30 min at RT in dark, 2 % OsO4 30–45 min at RT and 1 % uranyl acetate (pH 5.5) at 4 °C overnight. Washing steps were done with deionized water. Samples were dehydrated in ethanol, washed with dry acetone, dry acetonitrile, infiltrated with Quetol 651 resin (Agar Scientific) for 4–6 days and hardened in moulds at 60 °C for 2–3 days. Ultrathin sections were examined in the Tecnai G280 microscope. From each stage at least three pleuropodia and three legs were examined; all replications showed similar morphology.
For SEM whole embryos were fixed in 3 % glutaraldehyde in PB, post-fixed with OsO4, dehydrated in ethanol, critical point dried, gold coated, and observed in a FEI/Philips XL30 FEGSEM microscope. Photos from TEM and SEM were processed using Adobe Photoshop CC 2017.1.1.
Preparation of the reference transcriptome
The reference transcriptome includes transcripts that were assembled from (a) RNAs isolated from whole eggs and (b) RNAs isolated from legs and pleuropodia at a stage shortly before dorsal closure (sample “day 8–9”). (a) Whole egg transcriptome: Eggs from each 24-h egg collections incubated for the desired time were briefly washed with 50 % household bleach, washed with water and frozen in liquid nitrogen. Total RNA was isolated using TRIzol (Invitrogen), treated with TURBO DNase (Invitrogen) and purified on a column supplied with the RNAeasy Kit (Quiagen). The purified RNA from each of the 14 one-day samples was pooled into four: day 1–4, 5–7, 8–10 and 11–14. Ten μg of RNA from each was sent to BGI (Hong Kong). The total RNA was enriched in mRNA using the oligo (dT) magnetic beads and cDNA library was prepared using a standard protocol. 100 bp paired-end (PE) reads were sequenced on Illumina HiSeq 2000; the numbers of reads are in Additional file 2: Table S2. Non-clean reads were filtered using filter_fq. Transcripts from the samples were assembled separately using Trinity (release 20,130,225) ; parameters: --seqType fq --min_contig_length 100; −-min_glue 4 --group_pairs_distance 250; −-path_reinforcement_distance 95 --min_kmer_cov 4. The four assemblies were merged together to form a single set of non-redundant transcripts using TGICL software (version 2.1) ; parameters: -l 40 -c 10 -v 20. (b) Legs and pleuropodia transcriptome (age about 8.5–8.75 days): Appendages were dissected in RNase-free PBS and total RNA was isolated and cleaned as above. Ten μg of RNA from each leg and pleuropodium samples were transported to the Eastern Sequence and Informatics Hub, Cambridge (UK). cDNA libraries were prepared including mRNA enrichment. 75 bp PE reads were sequenced on Illumina GAIIX; the numbers of reads are in Additional file 2: Table S2. Reads were trimmed to the longest contiguous read segment for which the Phred quality score at each base Q > 13 (or 0.05 probability of error) using DynamicTrim (version 1.7) from the SolexQA package  and filtered to remove sequence adapter using Cutadapt (version 0.9) (http://code.google.com/p/cutadapt/). Sequences < 40 bp were discarded. The transcriptome was assembled using combination of Velvet (version 1.1.07)  (parameters: -shortPaired –fastq; −short2 –fastq; −read_trkg yes) and Oases (version 0.2.01)  (parameterss: -ins_length 350): the contigs that were output by Velvet were used by Oases to build likely transcripts from the RNA-seq dataset. K-mer sizes of 21, 25 and 31 were attempted for the two separate samples as well as the combined samples and optimal K-mer sizes of 21 were found for both samples.
Transcripts from the egg and legs plus pleuropodia transcriptomes were first merged with evigene (version 2013.03.11) using default parameters. Because this selection of transcripts (Selection 1) eliminated some genes (represented by zero transcripts, although the transcripts were present in the original transcriptomes), we repeated the step with less strict parameters (cd-hit-est - version 4.6, with -c 0.80 -n 5). This Selection 2 contained several genes represented by multiple transcripts, therefore we aligned Selection 1 and 2 to each other. Selection 1 was then completed by adding the missing transcripts from Selection 2. The resulting selection was edited as follows. Several redundant transcripts were removed manually: these were found by blasting diverse insect sequences against the transcriptome using the local ViroBLAST interface . Some transcripts were edited manually: e.g., when we found that two transcripts were combined into one (S. gregaria transcript blasted against sequences in GenBank resulted in different parts in high scoring alignments against different protein sequences) we split the transcripts, or when the transcript had a frameshift mutation that was not in the other transcripts from the mRNA, we corrected this. These manually changed transcripts were found during a random inspection of the selection. The resulting selection was blasted against itself (Blast+ suite, version 2.6.0) and if there was an alignment spanning ≥300 bp with a sequence identity of ≥98 % the redundant shorter transcript was removed. Transcripts < 200 bp were discarded. These steps were carried out in R  using the Biostrings package .
Basic transcript analysis was done using CLC Sequence Viewer7 (QIAGEN). Signal peptide and transmembrane regions were predicted by Phobius . Conserved domains were identified using SMART (http://smart.embl-heidelberg.de/). The reference transcriptome was annotated using Trinotate (version 3.1.1) . The longest candidate ORF of each transcript was identified using the inbuilt software TransDecoder . The transcriptome was blasted against Uniprot sequences of Schistocerca gregaria, Locusta migratoria, Apis melifera, Tribolium castaneum, Bombyx mori and Drosophila melanogaster (blastx with default parameters and -max_target_seqs 1).
RNA-seq expression analysis
Pleuropodia and legs from embryos at the same age (day 4, 5, 6, 7, 8, 10, 11, 12 and 13) were dissected in RNase-free PBS and total RNA was isolated as described above, but cleaned with RNA Clean & Concentrator (Zymo Research). One μg of RNA from each sample was sent to BGI (Hong Kong). cDNA was prepared including mRNA enrichment as above. Over 45 millions of 50 bp single-end (SE) reads were sequenced from each sample (Additional file 2: Table S2) on Illumina HiSeq 2000. A pair of samples from embryos that was used for the preparation of the reference transcriptome, part (b) above (“day 8–9”), was included in the expression analysis, but prior to mapping, the 75 bp PE reads were trimmed to 50 bp, using Trimmomatic in the paired-end mode (version 0.36) and using the CROP function (CROP:50). Each sample for expression analysis contained tens to hundreds of appendages. A single sample from each pleuropodia and legs was sequenced per stage.
The quality of the sequenced reads was assessed using the FastQC. All samples showed a Per base sequence quality > 30. Reads were mapped to the reference transcriptome using Bowtie2 (version 2.2.5) with default parameters and –local alignment mode . Trimmed pairs of reads were concatenated for each stage and treated as single reads. A PCA plot was prepared using the plotPCA() function in the DESeq2 R package ; the count matrix was transformed with the rlog() function. The plot showed that differences in sequencing type and processing of SE and PE samples had no noticeable effect on the results. The R package HTSFilter  was used with default parameters to remove transcripts with constantly low expression; 12,988 transcripts were left.
The differential expression analysis was performed using the NOISeq R package (version 2.22.1 ;). Reads were normalized using the RPKM method . DEGs between legs and pleuropodia for each stage were searched using NOISeq-sim; parameters for simulation of “technical replicates” prior to differential expression analysis without replicates: k = NULL, norm = “n”, pnr =0.2, nss =5, v = 0.02, lc = 1, replicates = “no”. DEGs between highly secreting pleuropodia and equally aged legs (samples from day 10, 11 and 12 treated as replicates) were searched using the NOISeq-real algorithm; parameters: k = 0.5, norm = “n”, factor = “type”, nss = 0, lc = 1, replicates = “technical”. Thresholds for significant differential expression were probability (“prob”) > 0.7 for single stage comparisons and > 0.8 for the triplicated comparison, RPKM > 10 and fold change > 2 for all comparisons (thresholds were set arbitrarily based on the values for the genes whose expression dynamics in the pleuropodia are known, Additional file 2: Table S4).
GO enrichment analysis
The transcriptome was blasted against the UniProt/Swiss-Prot database. GO enrichment with blast hits of an e-value ≤1e− 5 was performed using the R package GOSeq (version 1:30.0 ;) implemented in the Trinotate pipeline (see above). Enriched GO terms were summarized and visualized using REVIGO . Dot plots were prepared from DEGs having RPKM > 50, fold change > 3.
Tissues were dissected, total RNA was isolated and DNase treated as described for sequencing and cleaned with RNA Clean & Concentrator (Zymo Research). cDNA was prepared from 0.5 μg (legs, pleuropodia) or 1 μg (cut embryos) of the RNA using oligo-dT primer (Invitrogen) and ThermoScript RT-PCR System (Invitrogen) at 55 °C (lower amount of RNA from legs and pleuropodia, compared to the amount of RNA from whole cut embryos, was used because this RNA was in a short supply and difficult to obtain since these appendages are small and had to be dissected). PCR reactions (20 μl) contained 5 μl of cDNA diluted to 40 ng/μl, 10 μl of SYBR Green PCR Master Mix (Applied Biosystems) and 5 μl of a 1:1 mix of forward and reverse primers (each 20 nM in this mix). Reactions were run in the LightCycler480 (Roche) and analyzed using associated software (release 1.5.0 SP1) according to comparative Ct method and normalized to the eEF1α gene. Amplification was 40 cycles of 95 °C for 10 s, 60 °C for 15 s, 72 °C for 12 s. Primers (Additional file 2: Table S18) were designed using Primer3PLUS . To check for the presence of a single PCR product, the melting curve was examined after each run and for each pair of primers at least two finished runs were visualized on a 2 % agarose gel.
Majority of the work was carried out in the lab of Michael Akam (University of Cambridge) and the data analysis was finished in the lab of Gregor Bucher (University of Göttingen); BK thanks both for hosting and financial support. Electron microscopy was done at the Cambridge Advanced Imaging Centre (University of Cambridge). Immunolabeling was done in the lab and with help of Andrew Gillis. Stereomicroscopic pictures were taken in the lab of Paul Brakefield. For RNA-seq we used the services of BGI (Hong Kong) and the Eastern Sequence and Informatics Hub, Cambridge (UK). We also thank for help and advice to Ken Siggens, Jenny Barna, Jeremy Skepper and lab, Steven Van Belleghem, Barry Denholm, Jan Šobotník, and Gareth Griffiths, for scripts to Erik Clark and Simon Martin. We thank to Michael Akam, Siegfried Roth, Stuart Reynolds, Nico Posnien and Maurijn van der Zee for comments on the manuscript.
BK initiated the study, designed research, carried out all experimental work, supervised the bioinformatics analysis, interpreted the data and wrote the paper; EB performed majority of the bioinformatics analysis and edited the draft; AC carried out the initial steps in the selections of transcripts for the reference transcriptome and did a preliminary expression analysis. All authors read and approved the manuscript.
This work was supported by Human Frontier Science Program (Long-Term postdoctoral fellowship LT000733/2009-L), Biotechnology and Biological Sciences Research Council (grant number grant BB/ K009133/1), Isaac Newton Trust (University of Cambridge) and Balfour-Browne Fund (University of Cambridge).
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
- 1.Rathke H. Zur Entwickelungsgeschichte der Maulwurfsgrille (Gryllotalpa vulgaris). Arch Anat Physiol wiss Med. 1844:27–37.Google Scholar
- 2.Wheeler WMM. On the appendages of the first abdominal segment of embryo insects. Trans Wis Acad Sci Arts Lett. 1889;8:87–140 pls 1–3.Google Scholar
- 6.Hussey PB. Studies on the Pleuropodia of Belostoma flumineum Say and Ranatra fusca Palisot de Beauvois, with a discussion of these organs in other insects. Entomol Am. 1926;7:1–82.Google Scholar
- 8.Ando H. The comparative embryology of Odonata with special reference to a relic dragonfly Epiophlebia superstes Selys: Sugadaira Biological Laboratory of Tokyo Kyoiku University; 1962.Google Scholar
- 9.Ando H, Haga K. Studies on the Pleuropodia of Embioptera, Thysanoptera, and Mecoptera. Tokyo U. Educ. Sugadaira Biol. Lab. Bull. 1974;6:1–8.Google Scholar
- 12.Graber V. Ueber den Bau und die phylogenetische Bedeutung der embryonalen Bauchänhange der Insekten. Biol Zent Bl. 1889;9:355–63.Google Scholar
- 14.Heming BS. Origin and fate of pleuropodia in embryos of Neoheegeria verbasci (Osborn) (Thysanoptera: Phlaeothripidae). In: Bhatti JS, editor. Advances in Thysanopterology. New Delhi: Sciantia Publishing. Journal of Pure and Applied Zoology 4; 1993. p. 205–223.Google Scholar
- 15.Kamiya A, Ando H. 1985. External morphogenesis of the embryo of Ascalaphus ramburi (Neuroptera, Ascalaphidae). In: Ando H, Miya K, editors. Tsukuba: ISEBU Co Ltd.; 1985. p. 203–213.Google Scholar
- 17.Kobayashi Y, Suzuki H, Ohba N. Development of the pleuropodia in the embryo of the glowworm Rhagophthalmus ohbai (Rhagophthalmidae, Coleoptera, Insecta), with comments on their probable function. Proc Arthropod Embryol Soc Jpn. 2003;38:19–26.Google Scholar
- 19.Larink O. Embryonic and postembryonic development of Machilidae and Lepismatidae (Insecta: Archaeognatha). Entomol Gen. 1983;8:119–33.Google Scholar
- 21.Machida R, Tojo K, Tsutsumi T, Uchifune T, Klass K-D, Picker MD, Pretorius L. Embryonic development of heel-walkers: reference to some prerevolutionary stages (Insecta: Mantophasmatodea). Proc Arthropod Embryol Soc Jpn. 2004;39:31–9.Google Scholar
- 24.Miyakawa K. Embryology of the dobsonfly, Protohermes grandis Thunberg (Megaloptera: Corydalidae), I. Changes in External form of the embryo during development. Kontyû. 1979;47:367–75.Google Scholar
- 27.Tanaka M, Kobayashi Y, Ando H. Embryonic development of the nervous system and other ectodermal derivatives in the primitive moth, Endoclita sinensis (Lepidoptera, Hepialidae). In: Ando H, Miya K, editors. Recent Advances in Insect Embryology in Japan. Tsukuba: Isebu Co Ltd.; 1985. p. 215–29.Google Scholar
- 28.Tsutsumi K, Machida R. Embryonic development of a snakefly, Inocellia japonica Okamoto: an outline (Insecta: Neuroptera, Raphidiodea). Proc Arthropod Embryol Soc Jpn. 2006;41:37–45.Google Scholar
- 30.Slifer EH. The origin and fate of the membranes surrounding the grasshopper egg; together with some experiments on the source of the hatching enzyme. Q J Micr Sci. 1937;79:493–506.Google Scholar
- 35.Bullière F. L’évolution des pleuropodes au cours du développement embryonnaire de la Blabera craniifer (Insecte Dictyoptère). Arch Anat Microsc. 1970;59:201–20.Google Scholar
- 36.Louvet JP. L’ultrastructure du pleuropode et son ontogenèse, chez l’embryon du phasme Carausius morosus Br. I. – Étude du pleuropode de l’embryon agé. Ann Sci Nat Zool. 1973;12:525–94.Google Scholar
- 37.Louvet JP. Premières observations sur l’ultrastructure du pleuropode chez le Criquet migrateur. C R Acad Sci Paris D. 1975;280:1301–4.Google Scholar
- 41.Jones B. Endocrine activity during insect embryogenesis. Control of events in development following the embryonic moult (Locusta migratoria and Locustana pardalina, Orthoptera). J Exp Biol. 1956;33:685–96.Google Scholar
- 42.Shutts JH. Some characteristics of the hatching enzyme in the eggs of Melanoplus differentialis (Thomas). Proc S Dak Acad Sci. 1952;31:158–63.Google Scholar
- 43.Novak VJA, Zambre SK. To the problem of structure and function of pleuropodia in Schistocerca gregaria FORSKÅL embryos. Zool Jb Physiol. 1974;78:344–55.Google Scholar
- 46.Bernays EA. The vermiform larva of Schistocerca gregaria (Forskål): form and activity (Insecta, Orthoptera). Z Morph Tiere. 1971;70:183–200.Google Scholar
- 82.Chávez VM, Marqués G, Delbecque JP, Kobayashi K, Hollingsworth M, Burr J, Natzle JE, O’Connor MB. The Drosophila disembodied gene controls late embryonic morphogenesis and codes for a cytochrome P450 enzyme that regulates embryonic ecdysone levels. Development. 2000;127:4115–26.PubMedPubMedCentralGoogle Scholar
- 84.Petryk A, Warren JT, Marqués G, Jarcho MP, Gilbert LI, Kahler J, Parvy JP, Li Y, Dauphin-Villemant C, O’Connor MB. Shade is the Drosophila P450 enzyme that mediates the hydroxylation of ecdysone to the steroid insect molting hormone 20-hydroxyecdysone. Proc Natl Acad Sci USA. 2003;100:13773–13,778.PubMedCrossRefPubMedCentralGoogle Scholar
- 85.Warren JT, Petryk A, Marques G, Jarcho M, Parvy JP, Dauphin-Villemant C, O’Connor MB, Gilbert LI. Molecular and biochemical characterization of two P450 enzymes in the ecdysteroidogenic pathway of Drosophila melanogaster. Proc Natl Acad Sci USA. 2002;99:11043–8.PubMedCrossRefPubMedCentralGoogle Scholar
- 86.Warren JT, Petryk A, Marqués G, Parvy JP, Shinoda T, Itoyama K, Kobayashi J, Jarcho M, Li Y, O’Connor MB, Dauphin-Villemant C, Gilbert LI. Phantom encodes the 25-hydroxylase of Drosophila melanogaster and Bombyx mori: a P450 enzyme critical in ecdysone biosynthesis. Insect Biochem Mol Biol. 2004;34:991–1010.PubMedCrossRefPubMedCentralGoogle Scholar
- 90.Tanizawa T, Ando H, Tojo K. Notes on the pleuropodia in the giant water bug Appasus japonicus (Heteroptera, Belostomatidae). Proc Arthropod Embryol Soc Jpn. 2007;42:9–11.Google Scholar
- 93.Grabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson DA, Amit I, Adiconis X, Fan L, Raychowdhury R, Zeng Q, Chen Z, Mauceli E, Hacohen N, Gnirke A, Rhind N, di Palma F, Birren BW, Nusbaum C, Lindblad-Toh K, Friedman N, Regev A. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat Biotechnol. 2011;29:644–52.PubMedPubMedCentralCrossRefGoogle Scholar
- 94.Pertea G, Huang X, Liang F, Antonescu V, Sultana R, Karamycheva S, Lee Y, White J, Cheung F, Parvizi B, Tsai J, Quackenbush J. TIGR Gene Indices clustering tools (TGICL): a software system for fast clustering of large EST datasets. Bioinformatics. 2003;19:651–2.PubMedCrossRefPubMedCentralGoogle Scholar
- 99.R Development Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3–900,051–07-0; 2008.Google Scholar
- 100.Pagès H, Aboyoun P, Gentleman R and DebRoy S. Biostrings: Efficient manipulation of biological strings. R package version 2.46.0.; 2017.Google Scholar
- 102.Haas BJ, Papanicolaou A, Yassour M, Grabherr M, Blood PD, Bowden J, Couger MB, Eccles D, Li B, Lieber M, MacManes MD, Ott M, Orvis J, Pochet N, Strozzi F, Weeks N, Westerman R, William T, Dewey CN, Henschel R, LeDuc RD, Friedman N, Regev A. De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nat Protoc. 2013;8:1494–512.PubMedCrossRefPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.