Introduction

The epithelial apical extracellular matrix (ECM) controls development and structural maintenance of organisms, organ morphogenesis, wound healing, cell signaling and the local defense barrier against a hostile environment1,2. Numerous events, such as tissue metamorphosis and remodeling, require dynamic alterations and degradations of the ECM architecture. Epithelial cells secrete essential components, such as proteins, lipids and polysaccharides, into the extracellular space where they form a complex network of proteins and proteoglycans. These extracellular components, together with apical membrane bound proteins, organize the ECM as a first barrier at the epithelial cell surface2,3,4,5.

In mice and humans mutations of the extracellular collagen and the cornified envelope have been linked to skin disorder diseases, characterized by skin fragility, painful wounding and accompanied skin infections. The skin layers establish stability and an essential protective barrier by crosslinking of extracellular components5,6. In chitinous invertebrates, the ECM barrier is represented by the cuticle, an exoskeleton that covers the body and internal organs as an epithelial surface layer7,8. The cuticle is a highly organized ECM structure composed of the polysaccharide chitin and a variety of proteins and enzymes4. This exoskeleton is essential for body size control, epithelial barrier formation, epidermal wound healing and protects cells from direct contact with pathogens, toxins or pesticides1,9,10. It further provides barriers for maintaining homeostasis of body fluids11.

Genes involved in the biosynthesis and secretion of chitin-ECM components have been studied intensively in Drosophila4,12,13. Biosynthesis is assisted by conserved factors to control precise spatial and temporal fine-tuning of chitin deposition in the organs14. The chitin-polymers arise at the apical cell surface, where they establish a lamellate and tight chitin-ECM15. Recently, we showed that the chitin-ECM formation is coordinated within the cuticle assembly zone, which is a prominent part of the cuticle at the apical cell surface16. Several proteins and enzymes are located within the cuticle assembly zone in order to establish a tightly packed ECM. This involves the Obstructor-A (Obst-A) protein17 that binds chitin-fibrils to form a scaffold for localizing chitin deacetylases Serpentine (Serp) and Vermiform (Verm), to mature physical and chemical properties of the chitin-matrix16,18,19. The protection of the newly synthesized cuticle is mediated by the Knickkopf (Knk) protein in Tribolium and a similar role in Drosophila has been discussed recently20,21.

Chitin-polymers may spontaneously assemble into microfibrils of various lengths and diameter22. At the apical cell surface they need to be organized into a tight and compact ECM architecture. This process of chitin-ECM assembly and replacement is repeated several times due to cuticle molting and upon wounding and thus may involve a strict genetic control. Genes that control the developmental dynamics of chitin-ECM replacement are poorly known. The glycosylhydrolases were shown to be active in hydrolyzing chitin-polymers of different sizes and have been found in many species, including plants, insects and mammals23,24,25,26,27,28,29. In arthropods glycosylhydrolases are encoded by an evolutionary conserved multigene family, referred to as family18 due to the characteristic glyco18 domain. The family18 comprises numerous Chitinases (Chts) and Imaginal-disc-growth-factors (Idgfs)26,27, initially identified from Drosophila imaginal disc cells30. The glyco18 domains of Chts are catalytically active and may assist in the hydrolysis of simple or complex chitin-polymers, while glyco18 domains of Idgfs are inactive26,30,31,32. However, the roles of many family18 members in ECM organization and accompanied dynamics of the cuticle remained largely unknown.

We systematically investigated family18 members in ECM formation in Drosophila. By individually knocking down genes in the cuticle secreting tissues, we show that a large number of the Chts and idgfs are involved in cuticle molting during larval as well as in pupal stages. This is supported by gene specific spatial-temporal expression profiles and by developmental lethality profiles upon gene knockdowns. Moreover, we demonstrate that many Chts and idgfs genes are required also during the intermolt stages for forming the epidermal cuticle in order to organize the outermost protective barrier, since knockdown mutants are highly susceptible to mechanical stresses and bacterial infections. Investigating the molecular network in Cht2 and idgf6 RNAi induced mutants, which showed strongest lethality and most severe cuticle defects, we propose that certain Chts are involved in organizing the chitin-matrix scaffold, which is essential for further enzyme-mediated ECM maturation and cuticle thickness. In contrast to Chts, the non-enzymatic Idgfs play an essential role in the protection of the newly synthesized cuticle matrix against degradation in order to stabilize and expand the ECM size in the larvae.

Results

The Drosophila family18 comprises sixteen members consisting of ten Chitinase (Cht2, Cht4, Cht5, Cht6, Cht7, Cht8, Cht9, Cht11, Cht10, Cht12) and six idgf genes (idgf1, idgf2, idgf3, digf4, idgf5 and idgf6)25,27. Their protein domain arrangements and amino acid alignments are conserved across insects and have been studied elsewhere26,27,33. The Cht protein domains include N-terminal Signal-Peptides, a variable number of glyco-18 domains and protein specific transmembrane- and chitin-binding-domains. The Idgf proteins contain one N-terminal Signal-Peptide for protein secretion and one glyco-18 domain (Supplementary Fig. S1a). Previously, it was shown by an RNAi injection screen in Tribolium castaneum that Cht5, Cht10 and idgf4 genes are involved in the molting process34. However, the role of many other Chts and idgfs in cuticle molting remains elusive.

A large number of distinct Chts and idgfs genes control larval and adult cuticle molting

In order to analyze Cht and idgf genes systematically in cuticle forming tissues, we performed individual gene specific transgenic RNA interference (RNAi)-mediated knockdown in Drosophila using lines of the very efficient (see below) Vienna stock collection (www.vdrc.at)35. For all experiments described below UAS-RNAi was driven in cuticle-forming organs with the help of the 69B-Gal4 driver line (Supplementary Fig. S1b-d). Wild-type (wt) and 69B-Gal4 animals were taken as control. Our knockdown studies identified a large subset of genes required in larval cuticle molting and a distinct set essential in pupal-adult molting (Figs 1,2). The Cht2,5,7,8,9,12 and the idgf1,3,4,6 knockdown mutants showed characteristic double cuticles of head skeleton and posterior spiracles in second and third instar larvae although with variances in the severity of phenotypes (Fig. 1). Individual depletion of Cht2,5,6,7 genes resulted in pupal molting defects and interfered with normal wing extension (Fig. 2). Pupal molting defects were not observed upon knockdown of idgf genes. Monitoring animals for twelve days, from egg laying until adulthood, demonstrated that Cht and idgf reduction caused larval and pupal lethality (Supplementary Fig. S1f,g). Relative gene expression profiles compared to first instar larvae showed up-regulation for Cht8,9,12 at second and third larval stages and for Cht2,5,6,7 at pupal stages (Supplementary Fig. S2) which partially reflects their time dependent knockdown phenotypes during molting (Figs 1,2). In contrast, idgf genes are expressed at rather constant levels throughout larval and pupal development (Supplementary Fig. S2). In summary, these studies show that numerous Cht and idgf genes are essential for cuticle molting throughout development. Importantly, among the family18 members, Cht2 and idgf6 knockdowns resulted in the strongest lethality during larval stages (Supplementary Fig. S1f, g).

Figure 1
figure 1

Chts and idgfs control larval cuticle molting.

Larval molting defects visualized by bright field microscopy of cuticle preparations. (a,b) In second (wt) and third instar stages (69B-Gal4 control animals) larvae contained single cuticles of mouth hooks, head skeleton (left panel) and posterior spiracles (psp) (right panel). (c–f) In contrast, individual idgf1, idgf3, idgf4 and idgf6 RNAi-mediated knockdown resulted in cuticle molting defects. Mutant larvae contained double cuticles (arrowheads), including old and new cuticles, of mouth hooks, head skeleton and posterior spiracles. (g–l) Similarly the individual Cht2, Cht5, Cht7, Cht9, Cht8 and Cht9 and Cht12 knockdown led to double cuticles (arrowheads). It is of note that Cht9 knockdown resulted in double cuticle only when larvae molt into third instar.

Full size image
Figure 2
figure 2

Chts coordinate pupal-adult molting and wing extension.

Pupal arrest and wing extension defects were documented by bright field microscopy. (a–a’) Early (a) and late (a’) 69B-Gal4 control pupae. The late pupa contained characteristic red colored eyes (arrowhead). The newly hatched 69B-Gal4 control fly completed wing extension ((a’), arrow). (b–b”) The RNAi-mediated Cht2 knockdown resulted in early (b) and late pupal (b’) lethality. Very few adult flies hatched but showed wing extension defects (b”) and died. (c–e’) The individual Cht5, Cht6 and Cht7 knockdown caused late pupal lethality (c–e) and eclosion defects (c’–e’) in which adult flies are trapped inside the pupal cuticle. Hatched adult Cht5 and Cht6 knockdown flies revealed wing extension defects (c”,d”).

Full size image

Chts and idgfs are essential for organizing the exoskeletal barrier

Next we asked whether Cht and idgf genes are involved in providing epithelial barriers, which was analyzed in wounded RNAi knockdown larvae of first instar stage. Gentle lateral pricking of wt and 69B-Gal4 control larvae with a microinjection needle resulted in short hemolymph bleeding16,19 but did not critically impair the cuticle barrier since most animals survived the wounding. In contrast, UAS-RNAi-mediated knockdown of individual Cht2, Cht5, Cht7, Cht9, Cht12 and of idgf1, idgf3, idgf4, idgf5 and idgf6 genes displayed large epidermal lesions at wounded sites of first instar. Epidermal burst was followed by a severe hemolymph bleeding, organ spill out and immediate lethality (Figs 3a–c and 4a–c). Quantitative real-time PCR analyses confirmed high knockdown efficiencies for these Cht and idgf genes in first instar larvae (Supplementary Fig. S1e). In summary, these findings provide evidence that a large number of Cht and idgf genes are required at the epidermis for exoskeletal cuticle barrier formation.

Figure 3
figure 3

Chts organize the cuticle barrier and ECM formation at the apical surface.

Phenotypic data for Cht RNAi lines combined with the tissue specific 69B-Gal4 driver line. (a,b) Epidermal cuticle barrier on first instar larvae after puncture wounding visualized by Nomarski-optic. wt larvae (a) survived gentle lateral puncture wounding while Cht2 knockdown (b) resulted in epidermal burst, severe bleeding, organ spill out and immediate lethality (arrow). (c) The histogram indicates epidermal cuticle integrity defects in wounded Cht2, Cht5, Cht7, Cht9 and Cht12 knockdown larvae; n >50 individuals; error bars indicate standard deviations. (d) Confocal and bright field microscopy of cross-sections of third instar larvae show the epidermal chitin-matrix by Cbp (green) and the apical surfaces by WGA (red). In wt larvae the chitin-matrix detected by Cbp was the most prominent part of the cuticle and appeared in a stratified organization (red arrow) resembling procuticle chitin-lamellae. Cht2, Cht5, Cht7, Cht9 and Cht12 knockdown led to a deformed stratified chitin-matrix. In Cht2, Cht9 and Cht12 knockdown larvae the chitin-matrix was narrowed as indicated by the white bars. It is of note that Cht7 knockdown larvae showed unusual blisters (red arrow). For gene knockdown n ≥ 15 individuals were analyzed. White dashes indicate the apical cell surface. (e) Effects of Cht knockdown on the chitin-matrix organizer Obst-A in third instar larvae epidermis visualized by confocal microscopy. In wt larvae Obst-A was enriched at the apical cell surface (white arrowhead) marking the cuticle assembly zone16. Additional intracellular (red arrow) and extracellular (yellow arrow) Obst-A was detected. In contrast, Cht2, Cht7, Cht9 and Cht12 knockdown led to loss of Obst-A enrichment at the apical cell surface. Red arrow points to intracellular Obst-A expression. The extracellular Obst-A was either mislocalized at the entire chitin-matrix in Cht2 mutants or strongly reduced (Cht7, Cht9, Cht12). In Cht5 knockdown larvae Obst-A was only partially apically enriched (arrowhead). Note that the cuticle was detached from epidermal cells (asterisk) in Cht7 knockdown. Scale bars represent 10 μm.

Full size image
Figure 4
figure 4

Idgfs are required for apical cuticle barrier and ECM formation.

Phenotypic data for idgf RNAi lines combined with the 69B-Gal4 driver. (a,b) Cuticle barrier was analyzed in first instar larvae, as visualized by Nomarski-optic. wt larvae (a) survived gentle lateral puncture wounding while idgf6 knockdown resulted in epidermal rupture (arrow), organ spill out and lethality. (c) The histogram indicates that the epidermal cuticle barrier integrity of wounded larvae was severely affected as a result of individual idgf1, idgf3, idgf4, idgf5 and idgf6 knockdown; n > 50 individuals; error bars indicate standard deviations. (d) Confocal and bright-field microscopies of third instar larvae; the epidermal chitin-matrix was detected by Cbp (green) and the apical surfaces by WGA (red). For gene knockdown n ≥ 15 individuals were analyzed and representative images are shown. While wt larvae showed a stratified chitin-matrix (red arrow), the individual idgf1, idgf3, idgf4, idgf5 and idgf6 knockdown led to deformed chitin-matrix morphology and narrowed cuticle thickness, as indicated by the white bars. White dashes mark the apical cell surface. (e) Obst-A was enriched at the apical epidermal cell surface (arrowheads) marking the cuticle assembly zone16, in wt as well as in idgf1, idgf3, idgf4 knockdown larvae. Note that in RNAi-induced idgf5 and idgf6 mutants Obst-A was only partially enriched at the apical cell surface (arrowheads). Obst-A was additionally located in the lamellate procuticle of wt larvae (yellow arrow). Obst-A staining aggregated (blue arrows) in the procuticle in idgf3, idgf5 and idgf6 mutant larvae; Obst-A was reduced upon idgf1, idgf4 knockdown. Scale bars represent 10 μm.

Full size image

The roles of individual Cht2,5,7,9,12 and of idgf1,3,4,5,6 genes in establishing cuticle barriers were further investigated by studying epidermal morphology. Third instar larvae deposit massive amounts of cuticle material to expand their number of chitin-lamellae within the procuticle36. The chitin-rich procuticle was visualized by the chitin binding probe (Cbp) and the apical cell surface was detected by the lectin Wheat germ agglutinin (WGA)16, which recognizes sugar residues of glycoproteins and chitin37. Confocal analysis of wt third instar larvae showed WGA enrichment at the apical cell surface and Cbp accumulation in the chitin-lamellae16. In Cht and idgf knockdown mutants the morphology of the epidermal cuticle was defective (Figs 3d and 4d). Cht2 knockdown resulted in an extremely thin epidermal chitin-ECM. In Cht5 RNAi-induced mutant larvae the lamellate procuticle structure was deformed. In Cht7 knockdown larvae the chitin-ECM was disrupted and large blister-like cuticle extensions appeared. Cht9 knockdown larvae showed a deformed and thin chitin-ECM. Cht12 knockdown mutants revealed a thinned chitin-ECM and stratification defects. Our data of gene specific phenotypes provide first evidence that distinct Chts are required for epidermal cuticle morphogenesis. Gene expression profiles, individual domain variances (Supplementary Fig. S1) and enzyme specific activities against short or long polymeric chitin substrates32 suggest differences in Cht functions and thus further support our findings that several Chts are involved in organizing epidermal cuticles.

Analogous to Chts the knockdown of idgfs led to epidermal cuticle morphology defects. The histological analysis of third instar larvae showed unusually narrowed ECM thickness and deformed epidermal chitin-matrix in gene specific idgf1,3,4,5,6 knockdown larvae. Cbp signals appeared reduced and characteristic lamellate chitin-scaffolds were not observed in the mutant larvae. These data indicate that the epidermal chitin-ECM was severely deformed in idgf knockdown larvae (Fig. 4d).

Chts and idgfs coordinate chitin-scaffold and cuticle proteins for ECM formation at the apical cell surface

To understand the requirement of individual Cht and idgf genes in the molecular machinery of chitin-ECM formation, we studied if expression and localization of an essential cuticle component16,19 depends on the individual knockdowns. The chitin-binding protein Obstructor-A (Obst-A)17 coordinates dynamics of the growing chitin-scaffold19. Epidermal cells secrete Obst-A into the apical cuticle assembly zone, where it is required for the maturation of the chitin-ECM into a lamellar procuticle16. In addition, Obst-A is localized along chitin-lamellae of the procuticle (Figs 3e and 4e) where it is involved in chitin-matrix protection16. In contrast to wt epidermis, Obst-A enrichment in the cuticle assembly zone was disturbed in Cht2,5,7,9,12 knockdown larvae (Fig. 3e). These findings show that Chts are required for Obst-A localization within the cuticle assembly zone which is important for chitin-ECM maturation at the apical cell surface16.

In contrast to Chts, knockdown of idgf genes did not disturb apical enrichment of Obst-A within the cuticle assembly zone (Fig. 4e). However, Obst-A was strongly reduced within the lamellate chitin-ECM in idgf1,4,5,6 knockdown larvae and mislocalized in unusual plaque-like structures upon idgf3 knockdown. These findings suggest that the non-chitinolytic Idgfs26 are involved in Obst-A-mediated protection of the chitin-ECM16.

Cht2 is required for ECM growth at the cuticle assembly zone

Among family18 members the Cht2 and idgf6 knockdown resulted in the most severe epidermal cuticle defects (Figs 1, 2, 3, 4) and highest lethality of larvae (Supplementary Fig. S1f,g). Since our data suggest different functions in cuticle formation for Cht2 and Idgf6 we wanted to molecularly analyze their roles when third instar larvae increase epidermal chitin-ECM size. Our data suggest that the molecular Obst-A network can be affected in the cuticle assembly zone by Cht2 (Fig. 3e) and in the lamellate chitin-ECM by Idgf6 function (Fig. 4e). It was shown that Obst-A forms a core unit-complex with the deacetylases Serp, Verm18,38 and the GPI-linked Knk39 to promote chitin-matrix stability in trachea19. Epidermal cells secrete Obst-A into the cuticle assembly zone, where it is required for the localization of Serp and Verm in order to mature the chitin-matrix into a lamellar stratified ECM16. A second critical aspect for exoskeletal barrier formation is the Knk-mediated chitin-ECM stability20,21 and its putative role against degradation in Drosophila which involves Obst-A at the chitin-lamellae16. Defects either in expression levels or extracellular localization of these proteins lead to severe larval cuticle defects16 which we found in Cht2 and idgf6 knockdown larvae (Figs 3 and 4). We therefore addressed whether Cht2 or idgf6 have any influence in expression and localization of the Obst-A cuticle partners Serp, Verm and Knk. In wt third instar larvae the Serp and Verm deacetylases were enriched at the epidermal apical cell surface. In contrast, apical enrichment of both appeared disturbed in the Cht2 knockdown mutants. Confocal studies revealed reduced extracellular Serp levels and mislocalization of Verm (Fig. 5a,a’). These findings suggest that Cht2 is involved in Serp and Verm-mediated chitin-matrix maturation, providing essential chitin-ECM stability18,38. We then asked whether Cht2 could additionally play a role in protection of the established ECM. It was shown that Knk is involved in defending the newly formed chitin-matrix against chitinolytic degradation in the Tribolium body wall cuticle20 which could be also the case in the growing epidermal procuticle of Drosophila third instar larvae16. However, wt and Cht2 knockdown larvae showed similar Knk localization pattern indicating that Knk-mediated cuticle function appeared unaffected (Fig. 5c,c’). Our localization data propose that Cht2 activity affects the cuticle assembly zone machinery to organize and mature the chitin-matrix in third instar larvae at intermolt stage.

Figure 5
figure 5

Cht2 and idgf6 are required for epidermal procuticle formation.

Confocal analysis of cross sections of third instar larvae. (a–b”) Serp (a) and Verm (b) were enriched in the cuticle assembly zone of wt larval epidermis16. The Cht2 knockdown larvae showed no apical Serp (a’) and Verm (b’) enrichment. Upon idgf6 knockdown Serp (a”) and Verm (b”) were enriched at the apical cell surface (arrowheads). (c–c”) In wt larvae Knk was distributed in the procuticle at chitin-lamellae. In Cht2 knockdown larvae Knk was normally distributed. Upon idgf6 knockdown, Knk staining was affected when compared to wt. Scale bars represent 10 μm. (d–d”) Representative images show ultrastructure analysis of the larval body wall cuticle. For gene knockdown phenotypes of different larvae (Cht2 RNAi n = 6 second instar; idgf6 RNAi n = 5 second instar) were similar though variable in severity. The wt second instar (d) epidermal ECM contains the cuticle assembly zone (az), which was located at the apical cell surface, the lamellate chitinous procuticle (pro), the envelope and the outermost epicuticle (epi). In Cht2 knockdown larvae (d’) the cuticle assembly zone was absent and the procuticle contained irregularly spaced chitin-lamellae. The idgf6 knockdown larvae (d”) contained the cuticle assembly zone, but irregular chitin-lamellae architecture (arrow) and defective separation of the epicuticle. Red dashes mark the border between az and pro; blue dashes indicate the apical cell membrane. (e–e”) In wt third instar (e) ECM the number of chitin-lamellae increases (overview in inset). Upon Cht2 knockdown (n = 6; third instar) (e’) the number of chitin-lamellae was small. In idgf6 RNAi mutants (n = 5; third instar) (e”) chitin-lamellae degraded. They show numerous electron dense pore canal-like structures (yellow arrows) which were distributed across the deformed procuticle. In wt third instar pore canal-like structures appeared restricted to the epicuticle and outermost procuticle lamellae (Fig S7). (f–f”) Immunogold-labeled WGA recognizes chitin16,37. In wt second instar WGA was enriched in the chitin-dense procuticle but was less often found in the cuticle assembly zone (f). In Cht2 RNAi mutants (f’) the cuticle assembly zone was absent and WGA was distributed towards the apical cell surface. In idgf6 knockdown larvae, WGA was enriched in the procuticle. WGA enrichment was partially found adjacent to the apical cell surface (arrow) which may reflect the wavy pattern of the procuticle (d”). Supplementary Fig. S3e shows the control. TEM studies n = 3; Scale bars represent 1 μm, in D–F” and 2 μm in the inset.

Full size image

In order to verify whether the cuticle assembly zone is defective in the Cht2 knockdown we addressed epidermal ECM histology by ultrastructure analysis (Fig. 5d,d’). In wt larvae the chitin-ECM of the cuticle is represented by the cuticle assembly zone, next to the the apical cell surface and the outer lamellate procuticle (Fig. 5d). In third instar larvae procuticle size starts to increase dramatically by producing one chitin-lamella per hour36,40. This leads to a multiple number of chitin-lamellae at the end of larval development (Fig. 5e). In Cht2 knockdown larvae the assembly zone was barely detectable or even absent. Procuticle lamellae appeared irregularly stratified and, moreover, the number of procuticle lamellae did not increase during intermolt of third instar larvae (Fig. 5d’), likely causing the narrowed ECM size in these larvae (Fig. 3d). The absence of the cuticle assembly zone was further verified by immunogold-labelled WGA, which recognizes tightly packed chitin. In the wt epidermis numerous WGA-gold conjugates accumulated in the chitin-rich lamellate procuticle but not in the assembly zone. In contrast, in Cht2 knockdown larvae such differences in WGA distribution were not detected (Fig. 5f–f’). The unusual WGA accumulation towards the apical cell surface indicates strong reduction or loss of the cuticle assembly zone in Cht2 mutants.

idgf6 is essential for cuticle maintenance

In sharp contrast to Cht2 mutants, knockdown of the idgf6 gene did not disturb enrichment of Obst-A, Serp and Verm at the apical cell surface within the cuticle assembly zone (Figs 4e and 5a–b”). Upon idgf6 knockdown Knk was localized within in the cuticle but normal extracellular accumulation appeared affected when compared to wt (Fig. 5c,c”). This suggests that potentially chitin protection could be altered and this may cause a premature degradation of the lamellate procuticle16,20,21.

Analogous to the Cht2 mutant ultrastructure analysis we investigated the epidermal cuticle in idgf6 mutant larvae. In second instar larvae idgf6 knockdown led to unusually wrinkled epidermal procuticle morphology which partially affected the cuticle assembly zone (Fig. 5d”). In third instar, the epidermal procuticle appeared degraded (Fig. 5e”) and its thickening by adding numerous chitin-lamellae was disturbed, finally causing the narrowed cuticle size. The premature procuticle degradation may have further consequences. In insects pore canals transport material across the procuticle41,42,43,44. In Drosophila pore canals are found in the procuticle where they expand to form a lumen15. In wt second instar we identified pore canal-like structures running across the procuticle. The wt third instar showed pore canal-like structures only in the epicuticle and the outermost two chitin-lamellae of the procuticle (Supplementary Fig. S3). In contrast, upon depletion of idgf6 in third instar larvae, numerous unusually electron dense pore canal-like structures were found in the deformed procuticle (Fig. 5e”; Supplementary Fig. S3). Degradation of the procuticle and defective structure of pore canals may interfere with outer cuticle layers since the epicuticle was irregularly and not clearly separated from the underlying procuticle in idgf6 knockdown larvae. Altogether, this implies that the chitinolytically inactive Idgf632 may contribute to the stability and protection of the growing procuticle.

Cht2 and idgf6 function is essential for epithelial defense against pathogens

Cuticles continuously withstand numerous environmental physical and chemical traumas, such as infections by invading microorganisms or harmful and toxic reagents. We addressed the in vivo consequence of defective cuticle barriers in protecting RNAi knockdown animals against invasive microorganisms. GFP expressing Pseudomonas entomophila normally do not penetrate the epidermal cuticle but infect Drosophila larvae orally leading to damage of the gut epithelia which causes melanization spots in the wounds45. Pseudomonas successfully penetrated the defective cuticles of Cht2 and idgf6 knockdown larvae. With the exception of one animal all wt larvae (n = 45) were not penetrated by Pseudomonas at the epidermal cuticle, as GFP spots and the accompanied melanization darkening were not observed. In contrast, most of the Cht2 and idgf6 RNAi induced mutants (n = 45 for each genotype) showed melanization-like darkening spots at the epidermal wounding sites where GFP-expressing bacteria accumulated (Fig. 6a–c). These results indicate the fragility of the outermost physical defense barrier in the mutants and show that Cht2 and Idgf6 are critical for the susceptibility of the body wall cuticle to invading microorganisms.

Figure 6
figure 6

The developmental roles of the Cht- and idgf family.

(a–c) GFP expressing Pseudomonas entomophila usually cannot penetrate through the cuticle barrier of the wt instar larva (n = 45). For most Cht2 and idgf6 knockdown (each n = 45) larvae Pseudomonas penetrate the defective cuticle barrier for invasion resulting in immune triggered melanization-like darkening spots at the wound surface (arrowheads). Normarski-optic reveals darkening of cuticle material which was attacked by invading bacteria. (d) The drawing scheme summarizes phenotypic data resulting from knockdown of Chts and idgf genes. (e) The schematic model summarizes our findings in size regulation of the epidermal cuticle in third instar larvae, which requires core cuticle components such as Obst-A, Serp and Verm enriched in the cuticle assembly zone to mature the chitin-lamellae16. It further requires the Knk protein, which has an important role in protecting the newly synthesized cuticle in Tribolium and potentially also in Drosophila16,20. Our data provide evidence that in Cht2 knockdown larvae the cuticle assembly zone is absent and core cuticle components required for maturation are mislocalized, finally resulting in reduced ECM thickness. In idgf6 RNAi mutants, chitin-lamellae are degraded and localization of procuticle components is disturbed causing a defective cuticle.

Full size image

Discussion

We investigated the role of Drosophila family18 genes in organizing cuticle molting and barrier protection (Fig. 6d). Our data are in line with previous observations26,46. It was recently shown in the brown plant hopper Nilaparvata lugens that double stranded RNA injection directed against Cht1, Cht5, Cht7, Cht9 and Cht10 causes mortality46. We observed larval and pupal lethality upon knockdown of the individual of family18 members. The Cht5 gene knockdown caused pupal molting defects in Tribolium similar to our study. Findings that the Cht10 (encoded by the Drosophila Cht1/3/10 gene) RNAi-induced knockdown showed lethality and severe cuticle molting defects in larvae and pupa of the red flour beetles47 supplements our knockdown screen for Chts in Drosophila. The observation that Drosophila Cht11 did not cause any molting or barrier defects is consistent with its recently proposed function in the mitochondria48. The Cht4 knockdown did not show any cuticle phenotypes in Drosophila and other insects, but Cht4 is rather involved in reproduction46. However, previous data indicated functional specialization among Cht genes for insect molting26. Our systematic study in Drosophila extends initial observations and shows that a large number of Cht and idgf genes are required to control the repetitive reorganization of the chitin-ECM at the epidermis.

Moreover, our screen characterizes which Cht and idgf genes are essential for forming the epidermal barrier-ECM. We show that the Cht2,5,7,9,12 and idgf1,3,4,5,6 genes are required for chitin-ECM formation and barrier function throughout larval development. These morphological data and in vivo barrier analysis were supported by analysis of the cuticle structure and the involved molecular network. The cuticle assembly zone is associated with the apical cell surface49 where chitin-fibrils and associated proteins arise. We recently showed that apically secreted components are organized within the assembly zone into the functional ECM matrix16. Here we provide evidence that Cht2 is required for cuticle assembly zone function. In the Cht2 knockdown the assembly zone was absent and consequently Obst-A, Serp and Verm were not enriched at the apical cell surface but mislocalized or reduced. This finally led to reduced number and defective morphology of chitin-lamellae when normally the number of lamellae should increase in third instar and the cuticle becomes thickened. There are different molecular scenarios how the glycosylhydrolase Cht2 could determine procuticle size. It could act in the cuticle assembly zone in order to modify for example length of chitin-fibrils, which supports Obst-A-mediated chitin-scaffold formation and maturation by the deacetylases Serp and Verm. A distinct scenario could be that Cht2 inhibits premature assembly, or acts in later stages for elongation and ordering the filaments. Defects in each of these processes could lead to pushing the cuticle proteins towards the membrane and the disappearance of cuticle assembly zone. Indeed, preliminary data for Cht2 protein localization within the epidermal cuticle support both scenarios (Supplementary Fig. S4).

Recently, it has been shown in Tribolium that downregulation of Tc-Cht5 and Tc-Cht10 along with Tc-knk rescued the knk mutant phenotype and restored chitin levels but not laminar architecture. The single and even double knockdown of Tc-Cht5 and Tc-Cht10 did not disturb laminar organization of the cuticle20. Based on our finding, we suggest that Chts may have different roles in organizing cuticle molting and cuticle formation. For example, the Drosophila Cht5 knockdown showed partially enriched Obst-A at the assembly zone and only minor reduction of cuticle thickening. Whether it plays a role in Knk turnover at the old cuticle as found for Tc-Cht5 remains elusive20. In contrast, Cht2 knockdown did not affect Knk localization or levels, but led to the absence of assembly zone formation and mis-localization of associated molecular modifiers resulting in a lamellar architecture but cuticle thickening was disturbed. Furthermore, in contrast to Cht2, in idgf6 mutants the procuticle was degraded, showing rudimentary chitin-lamellae and an irregularly and not clearly separated epicuticle. Reduced idgf6 did not disturb the assembly zone and its modifiers Obst-A, Serp and Verm at the apical cell surface but interfered with normal Knk levels. Altogether, this implies that the chitinolytically inactive Idgf6 (synonymous to Cht13 and DmDS47)32 may contribute to the stability and preservation of the growing procuticle. Similar phenotypes have been found upon Tc-Knk downregulation at the body wall and elytra cuticle20. Whether Idgf6 binds chitin for rather direct protection or it may act indirectly via potential Knk function is not known.

Although gene expression patterns are common among essential Cht genes (Supplementary Fig. S2c–j), individual domain arrangements (Supplementary Fig. S1), such as chitin-binding and transmembrane domains and in case of Cht5 and Cht9 enzymatic specificity for substrates and for pH profile25,31, support our studies and suggest that a large variety of different glyco-18 domain hydrolases could be required for modulating the ECM in the growing cuticle at the same time. Essentially, also idgf genes show a common expression pattern in the cuticle producing organs, but their gene-products are most likely enzymatically inactive, as shown for the Drosophila Idgf6 and the Tribolium IDGF2 and IDGF432. This group of proteins likely binds to chitin25 and may thereby contribute to the protection of the ECM.

Our study demonstrates that Cht2 and idgf6 are very critical for larval cuticle barrier formation (Fig. 6e) and protection against invasive microorganisms and mechanical stresses. Cht2 and Idgf6 genes offer new targets for a gene specific search of insecticides, since the epithelial barrier was impaired and resulted in a premature lethality of animals and a higher susceptibility to pathogenic infections. Homologues of the family18 genes exist in parasites and pest harmful to mankind and agriculture, such as mosquitos, lice and the spotted wing Drosophila50,51,52. Blast searches showed that parts of the C-terminus of Cht2 and the N-terminus of Idgf6 were not conserved among the insects providing a specific target for insecticides (Supplementary Fig. S5). This opens the possibility to search for small molecules that may inhibit function of insect species specific Cht2 or Idgf6 without affecting beneficial organisms. Chitin-ECM is the most important insect barrier against any environmental stresses. Affecting function of hydrolase enzymes and the ECM protection system provide new strategies for eliminating problematic animals already during larval stages and preventing the growth of next generations of upcoming pest.

Methods

Fly stocks

Vienna (Austria) /Bloomington (USA) stock centers (http://stockcenter.vdrc.at/control/main and http://flystocks.bio.indiana.edu/): w1118 (referred to as wild-type (wt)), 69-BGal4 line, UAS-RNAi-Cht and UAS-RNAi-idgf lines (Vienna stock center). For knockdown analysis, 69B-Gal4 flies were mated with UAS-RNAi flies. For negative controls we used Wild-type (wt) animals and animals from the 69B-Gal4 and the different UAS-RNAi lines but without inducing the RNAi system. In general, RNAi induced phenotypes can be variable to some extent. We therefore show representative images of observed phenotypes and indicate the number of analyzed animals.

Lethality and cuticle assays

Embryos from wt, 69B-Gal4 stocks and from individual 69B-Gal4 driven UAS-RNAi cht/idgf crosses were grown on apple juice agar plates at 29 °C for until they reached end of embryogenesis. After dechorionation with 2.5% sodium hypochlorite for 5 minutes, living stage (st) 17 embryos (n ≥ 100 for Chts n ≥ 75 for idgfs) and were transferred onto new agar plates and incubated at 29 °C. The number of dead and living animals were counted every 24 hours with Stemi stereomicroscopes (Zeiss, Germany) and characterized from first instar larvae to adult flies. Every genetic experiment was repeated at least three times. For cuticle preparation first, second and third instar larvae were washed with water and incubated in a 1:1 lactate and Hoyer’s reagent mixture overnight at 65 °C. Larval stages were determined by the morphology of mouth hooks. Larvae were analyzed with 60x and 100x objectives in Olympus immersion oil and by Normarski microscopy (DIC; Olympus AX70). Pupae and adult flies were analyzed using an Olympus Binocular microscope (SZX12). All images were taken with Cell’F software. In general for all assays (see below) images were cropped in Adobe Photoshop CS6 and figures designed with Adobe Illustrator CS6. For statistical analysis Excel 2010 was used.

Epithelial cuticle/epidermis integrity assay

First instar larvae (n > 50) were gently laterally wounded with a standard glass needle for micro injections as recently described16,19. The integrity test was repeated at least three times each with a minimum of thirty animals and immediately monitored by Normarski microscopy (DIC; Olympus AX70) and Cell’F software.

Larval fixation and immunofluorescence studies

Third instar larvae were fixed overnight in 4% paraformaldehyde and embedded in JB-4 resin (Polysciences, USA, Warrington, USA) as described by the manufacturer. Slices of 7 μm were cut using an Ultracut E (Reichert-Jung, Solms, Germany). For each genotype n ≥ 15 individuals were analyzed. The immunofluorescent labeling was performed as recently described16. For labeling chitin we used Alexa488-conjugated chitin-binding-probe (Cbp, 1:50, New England BioLabs, USA). For detecting the apical cell surfaces we used Alexa633-conjugated wheat germ agglutinin (WGA, 1:250; Molecular Probes, Carlsbad, USA). The WGA selectively detects N-acetylneuraminacid and N-acetylglucosamine sugar residues and can be used as a marker for plasma membrane surfaces and cuticle layers. Used antibodies: Knk (1:333; rabbit)39, Obst-A (1:300; rabbit)19, Serp (1:175; rabbit) and Verm (1:175; rabbit)18. Secondary antibodies were linked with fluorescent dyes (Dianova, Hamburg, Germany and JacksonImmuno, Baltimore, USA). The samples were mounted in Vectashield (H1500, DAPI, Vector Laboratories, Burlingame, USA).

For immunofluorescent analysis LSM710 and LSM780 (Zeiss, Jena, Germany) confocal microscopes and 63x LCI Plan Neofluar objectives were used. For scanning procedure pinhole “airy unit 1” ZEN software standard settings were used and cross-talk was avoided by using according filter settings and lasers. Images were taken with the ZEN 2010 and ZEN 2011, cropped with Adobe Photoshop CS6 and figures were designed with Adobe Illustrator CS6.

Ultrastructure analysis

Drosophila second and third instar larvae were placed on a 150 μm flat embedding specimen holder (Engineering Office M. Wohlwend GmbH, Sennwald, Switzerland) and frozen using 1-hexadecene in a Leica HBM 100 high pressure freezer (Leica Microsystems, Germany) according to the described method16. Wheat Germ Agglutinin (5 μg/ml, Vector Labs Burlingame, USA) was detected by a rabbit-anti-biotin antibody (3.9 μg/ml, Rockland, USA) and Protein A gold (10 nm, 1:100, G. Posthuma, Utrecht, Netherlands) as described recently16. Images were taken by a Philips CM120 electron microscope (Philips Inc.) using a TemCam 224A slow scan CCD camera (TVIPS, Gauting, Germany).

Infection assay

Pseudomonas entomophila (PE) was cultured in LB-medium with 50 μg/ml rifampicin at 29 °C. For infection experiments, second instar larvae were placed on filter papers moistened with PBS. For each genotype and condition, three dishes were prepared. Uninfected control larvae were fed with 500 μl saccharose solution. For infection of larvae, 500 μl of 10% saccharose solution containing 0.5 g resuspended bacteria was added to the filter paper. Larvae were incubated at 29 °C for 24 h and afterwards imaged with a Zeiss Apotome Axiovert 200 M (25x Plan Neofluar objective; Zeiss, Jena, Germany). Pseudomonas positive larvae reveal the GFP signal. Multidimensional photographs were taken with AxioVision software (Zeiss). For each genotype n = 45 animals were analyzed and the infection experiment was repeated three times. To prove infection by Pseudomonas and to rule out contamination of the control, homogenized larval material was streaked out on Rifampicin-agar plates with milk, on which pathogenic Pseudomonas grows colonies. After overnight culturing at 29 °C only plates with infected animals showed Pseudomonas colonies, while non-infected did not.

Quantitative real-time PCR (qPCR)53

To test RNAi kockdown efficiency L1 larvae were used. Samples were collected into lysis buffer with ß-Mercapoethanol (Roth, Karlsruhe, Germany) and lysed in a Precellys homogenizer using a 1.25–1.65 diameter glass bead mix (Roth, Karlsruhe, Germany). The lysed samples were frozen in liquid nitrogen and stored at −80 °C. RNA was extracted within two weeks. The NucleoSpin RNA II Kit (Macherey-Nagel, Düren, Germany) was used following the manual instructions including DNase treatment. RNA quantification was performed by the use of NanoDrop (ThermoScientific, Waltham, USA). A second genomic DNA digestion and the first strand cDNA synthesis was carried out with 500 ng of total RNA using the QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany) following the manual instructions and using the provided oligo-dT and random primer mix. All RNA reactions were performed on ice, cDNA was stored at −20 °C. qPCR experiments were done as recently described54 with the iQ5 Real-Time PCR detection system (Bio-Rad, München, Germany). qPCR primers have been designed with the “primer3plus” software using standard adjustments for qPCR primers leading to an amplicon length between 75 and 150 nucleotides (manufactured by Invitrogen, Karlsruhe, Germany). Desalted qPCR primers have been manufactured by Invitrogen (Karlsruhe, Germany). For PCR reaction we used 1 μl cDNA, primers in a final concentration of 200 nM and the iQ SYBR Green Supermix (Bio-Rad, Munich, Germany) in a total volume of 25 μl. Reactions were performed as technical duplicates or triplicates in clear 96-well qPCR plates (Bioplastics, Landgraaf, Nederlands). For normalization of transcript levels we used rp49 (Ribosomal protein L32) as a reference. The specificity of all primers was tested by melt curve analysis. Standard control PCR reactions (no-template control; negative control) were performed and showed no signal. Expression data were calculated according to the delta-delta-CT method and analyzed using Bio-Rad iQ5 Optical System Software and Microsoft Excel 2010. All primer sequences and efficiencies are provided for Chts (Supplementary Table S1), for idgfs (Supplementary Table S2) and the FlyBase gene card web-links are given in (Supplementary Table S3).

Additional Information

How to cite this article: Pesch, Y.-Y. et al. Chitinases and Imaginal disc growth factors organize the extracellular matrix formation at barrier tissues in insects. Sci. Rep. 6, 18340; doi: 10.1038/srep18340 (2016).