The Occurrence of the Holometabolous Pupal Stage Requires the Interaction between E93, Krüppel-Homolog 1 and Broad-Complex

Abstract

Complete metamorphosis (Holometaboly) is a key innovation that underlies the spectacular success of holometabolous insects. Phylogenetic analyses indicate that Holometabola form a monophyletic group that evolved from ancestors exhibiting hemimetabolous development (Hemimetaboly). However, the nature of the changes underlying this crucial transition, including the occurrence of the holometabolan-specific pupal stage, is poorly understood. Using the holometabolous beetle Tribolium castaneum as a model insect, here we show that the transient up-regulation of the anti-metamorphic Krüppel-homolog 1 (TcKr-h1) gene at the end of the last larval instar is critical in the formation of the pupa. We find that depletion of this specific TcKr-h1 peak leads to the precocious up-regulation of the adult-specifier factor TcE93 and, hence, to a direct transformation of the larva into the adult form, bypassing the pupal stage. Moreover, we also find that the TcKr-h1-dependent repression of TcE93 is critical to allow the strong up-regulation of Broad-complex (TcBr-C), a key transcription factor that regulates the correct formation of the pupa in holometabolous insects. Notably, we show that the genetic interaction between Kr-h1 and E93 is also present in the penultimate nymphal instar of the hemimetabolous insect Blattella germanica, suggesting that the evolution of the pupa has been facilitated by the co-option of regulatory mechanisms present in hemimetabolan metamorphosis. Our findings, therefore, contribute to the molecular understanding of insect metamorphosis, and indicate the evolutionary conservation of the genetic circuitry that controls hemimetabolan and holometabolan metamorphosis, thereby shedding light on the evolution of complete metamorphosis.

Fig 1. Depletion of TcKr-h1 in T. castaneum last instar larvae causes direct transformation to the adult form bypassing the pupal stage.

(A-D) Newly molted L7 larvae were injected with dsMock (Control) or with dsTcKr-h1 (TcKr-h1i). Ventral views of (A) a Control pupa, and (B) a Control adult. Ventral and lateral views of (C) 6- and (D) 10-day-old TcKr-h1i animals. (E-X) Comparison of the external morphology of appendages between Control pupae and adults, and TcKr-h1i animals. Scanning electron microscopy photographs of (E, I, M, Q and U) a newly molted Control pupa, (F, J, N, R and V) a 1-day-old Control adult, and (G, K, O, S and W) a 6-day-old TcKr-h1i animal, showing strong direct adult differentiation in TcKr-h1i animals. (G) TcKr-h1i legs are clearly segmented, including the tarsus (arrow), and presents double claws typical for the adult leg (arrowhead). (K) TcKr-h1i antennae show adult sensillae (asterisks) with well-shaped and segmented funicle and club (arrow). (O) All TcKr-h1i mouthparts are strongly segmented (arrows), including the (ln) labial palps, (p) maxillary palps and (ga) galea. (S) TcKr-h1i compound eyes present several rows of well-developed ommatidia. (W) Elytra of TcKr-h1i animals are highly sclerotized with the typical adult microsculpture. (H, L, P, T and X) External morphology of the appendages of TcKr-h1i animals showing that they are highly sclerotized and present uniformly dark brown pigmentation, characteristic of adult appendages.

Fig 2. The cuticle of the appendages of TcKr-h1-depleted animals shows adult-specific microsculpture.

Comparison of scanning electron microscopy photographs of the cuticle surface in (A-C) legs, (D-F) antennae, and (G-I) maxillae of (A, D and G) Control pupae, (B, E and H) Control adults, and (C, F and I) TcKr-h1i animals. Instead of the pupal-like surface, the cuticle of TcKr-h1i animals shows the characteristic adult-specific microsculpture, including rounded pits with sensillae (arrows in C and F). Scale bars represent 50 μm in (C), (F) and (I).

Fig 3. Loss of TcKr-h1 in last instar larvae induces precocious up-regulation of the adult specifier TcE93 and the repression of the pupal specifier TcBr-C.

(A-C) Transcript levels of (A) the adult-specific gene TcCPR27, (B) TcE93, and (C) TcBr-C were measured by qRT-PCR in Control and TcKr-h1i animals. Transcript abundance values are normalized against the TcRpL32 transcript. Fold changes are relative to the expression of each gene in 3-day-old Control larvae, arbitrarily set to 1. In the abscissa axis, P0/L7d7 represents equivalent developmental time points between Control (P0, pupal day 0) and TcKr-h1i animals (L7d7, day 7 in arrested TcKr-h1i animals). Error bars indicate the SEM (n = 5). Asterisks indicate differences statistically significant at p≤0.05 (*); p≤0.005 (**), and p≤0.0005 (***) (t-test).

Fig 4. Precocious high levels of TcE93 mediate the direct adult transformation of TcKr-h1-depleted larvae and prevent the up-regulation of TcBr-C.

(A-C) Newly molted L7 larvae were injected with dsMock (Control) or with dsTcKr-h1 and dsTcE93 simultaneously (TcKr-h1i + TcE93i). (A) Ventral view of a Control pupa. (B) Ventral and (C) lateral views of TcKr-h1i + TcE93i knockdown animals. In C, the TcKr-h1i + TcE93i animal arrested development and could not undergo pupal ecdysis, so the larval cuticle has been removed. The inset is a high magnification of the abdomen of the arrested TcKr-h1i + TcE93i pupa (rectangle in C) showing the pupal gin traps (arrows). (D-K) Comparison of the external morphology of appendages between (D-G) Control, and (H-K) TcKr-h1i + TcE93i pupae. TcKr-h1i + TcE93i pupae undergo normal pupation as no differences in appendage morphology are observed when compared to Control animals. (D and H) Hindlegs, (E and I) antennae, (F and J) mandible, and (G and K) maxilla. (L) Simultaneous depletion of TcKr-h1 and TcE93 leads to normal expression of TcBr-C and TcCPR27 in T. castaneum prepupae. TcBr-C (upper), and TcCPR27 (lower) mRNA levels were measured by qRT-PCR in Control and TcKr-h1i + TcE93i animals. Transcript abundance values are normalized against the TcRpL32 transcript. Fold changes are relative to the expression of each gene in 3-day-old Control larvae, arbitrarily set to 1. Error bars indicate the SEM (n = 5). Scale bars: 0.5 mm in (A); inset in (C), 0.1 mm; 200 μm in (D); 100 μm in (E) and (G); 50 μm in (F).

Fig 5. Loss of TcBr-C induces premature adult differentiation without affecting TcE93 and TcKr-h1 expression.

(A-C) Newly molted L7 larvae were injected with dsMock (Control) or with dsTcBr-C (TcBr-Ci). (A) Ventral view of a Control pupa. (B and C) Ventral views of (B) a TcBr-Ci animal arrested at the end of the L7 stage, and (C) a TcBr-Ci animal just after the pupal molt, showing abnormal shape with short and blister wings (arrowheads), imperfect gin traps (arrows) and short legs (red arrows). (D and E) Transcript levels of (D) TcE93, and (E) TcKr-h1 in Control and TcBr-Ci larvae at the beginning of the last L7 instar (d0) and during the quiescent stage (d4-d6), measured by qRT-PCR. Transcript abundance values are normalized against the TcRpL32 transcript. Fold changes are relative to the expression of each gene in d0 Control larvae, arbitrarily set to 1. Error bars indicate the SEM (n = 5–10). Scale bar: 0.5 mm.

Fig 6. The crosstalk between the metamorphic network genes is conserved in the derived holometabolous insect D. melanogaster.

(A) Expression levels of DmE93A (left), DmE93B (middle), and DmBr-C (right) relative to DmRpL32 in the whole body of ActGal4 (in black) and ActGal4;UASDmKr-h1^RNAi (in red) 0 h after puparium formation (APF) animals, measured by qRT-PCR. Fold changes are relative to the expression of each gene in ActGal4 larvae, arbitrarily set to 1. (B) Expression levels of DmE93A (left) and DmE93B (right) relative to DmRpL32 specifically in the wing pouch of rnGal4 (in black) and rnGal4;UASDmKr-h1^RNAi (in red) 0 h after puparium formation (APF) animals, measured by qRT-PCR. Fold changes are relative to the expression of each gene in rnGal4 larvae, arbitrarily set to 1. Error bars in A and B indicate the SEM (n = 5–10). Asterisks indicate differences statistically significant at p≤0.05 (*), p≤0.005 (**), and p≤0.0005 (***) (t-test). (C) DmBr-C protein levels revealed by immunocytochemistry in wings from 0 h after puparium formation (APF) larvae expressing a UASDmKr-h1^RNAi constructs under the control of rnGal4 driver, which is expressed in the wing pouch (silhouetted). DmBr-C protein is absent in DmKr-h1-depleted cells. (D) Dorsal views of rnGAL4 (left panel) and rnGal4;UASDmKr-h1^RNAi (right panel) pupae. In the absence of DmKr-h1 in the wing pouch, wings degenerate after eversion during the pupal stage (arrowheads). Scale bars: 200 μm in (D).

Fig 7. DmBr-C expression is repressed by DmE93 in the prepupal stage of D. melanogaster.

(A) Representative wing disc of CiGal4 UASGFP; UASKr-h1^RNAi 0 h APF prepupae labelled to visualize the Ci anterior domain (GFP in green), the nuclei (DAPI) and DmBr-C protein (in red). In the Ci domain, where the cells express the transgene and are depleted of DmKr-h1, DmBr-C is strongly repressed. (B) Representative wing disc of CiGal4 UASGFP; UASKr-h1^RNAi 0 h APF prepupae labelled to visualize the Ci anterior domain (GFP in green), DmBr-C protein (in red), and the transcription factor Spalt (in white). Whereas DmBr-C protein levels are strongly reduced in the Ci domain, normal levels of Spalt protein are present in the anterior and posterior domains of the wing disc indicating that the viability of the DmKr-h1-depleted cells is not compromised. (C) Examples of clones of cells in TubUASGFP; UASDmKr-h1^RNAi wing discs. Wings discs are labeled to visualize the clones (GFP in green), the nuclei (DAPI) and DmBr-C protein (in red). DmBr-C protein in cells within the clones is absent. (D) Representative wing disc of CiGal4 UASGFP; UASKr-h1^RNAi UASDmE93^RNAi 0 h APF prepupae labelled to visualize the Ci anterior domain (GFP in green), the nuclei (DAPI) and DmBr-C protein (in red). In the Ci domain, where the cells are depleted of DmKr-h1 and DmE93, DmBr-C protein levels are not reduced. (E) Cuticle preparations of adult wings expressing the indicated transgenes under the control of the rnGAL4 driver.

Fig 8. The Kr-h1-dependent repression of adult differentiation is relayed through E93 repression also in hemimetabolous insects.

(A-C) Newly molted N5 nymphs of B. germanica were injected with dsMock (Control), with dsBgKr-h1 (BgKr-h1i) or with dsBgK-h1 and dsBgE93 simultaneously (BgKr-h1i + BgE93i) and left until the next molts. (A) Dorsal (upper panels) and ventral (lower panels) views of a N6 Control nymph and a winged adult. (B) Dorsal and ventral views of a BgKr-h1i animal after the next molt showing a premature adult morphology, including functional hindwings and forewings, adult-specific pigmentation of the cuticle and the pronotum, and adult cerci. In the dorsal view, the left forewing has been removed to allow the observation of the membranous hindwing (red arrow). (C) Dorsal and ventral views of a BgKr-h1i + BgE93i animal after the next molt showing a perfect N6 morphology that includes black cuticle, two thick stripes of black melanin in the pronotum, nymphal cerci and external wing pads. (D and E) Loss of BgKr-h1 during N5 induces (D) precocious up-regulation of BgE93, and (E) repression of BgBr-C in N5 nymphs. The repression of BgBr-C is averted when BgKr-h1 and BgE93 are simultaneously depleted in N5 nymphs. Transcript levels of (D) BgE93, and (E) BgBr-C were measured by qRT-PCR in wings from 5-day-old Control, BgKr-h1i and BgKr-h1i+BgE93i N5 nymphs. Transcript abundance values are normalized against the BgActin5C transcript. Fold changes are relative to the expression of each gene in Control nymphs, arbitrarily set to 1. Error bars indicate the SEM (n = 5–10). Asterisks indicate differences statistically significant at p≤0.0005 (***) (t-test). Scale bar represents 2 mm.

Fig 9. The anti-metamorphic activity of BgKr-h1 is stage-specific.

(A and B) Newly molted N4 nymphs of B. germanica were injected with dsMock (Control), or with dsBgKr-h1 (BgKr-h1i) and left until the next molts. (A) Dorsal views of Control animals after molting into normal N5 and N6 nymphs, and finally to a winged adult. (B) Dorsal views of BgKr-h1i animals after molting into normal N5 nymphs and then to a premature adult with functional hindwings and forewings, adult-specific pigmentation of the cuticle and the pronotum, and adult cerci. (C and D) Loss of BgKr-h1 in N4 does not induce precocious up-regulation of BgE93, but represses BgBr-C expression. Transcript levels of (C) BgE93, and (D) BgBr-C were measured by qRT-PCR in wings from 4-day-old Control and BgKr-h1i N4 nymphs. Transcript abundance values are normalized against the BgActin5C transcript. Fold changes are relative to the expression of each gene in Control nymphs, arbitrarily set to 1. Error bars indicate the SEM (n = 5–10). Asterisks indicate differences statistically significant at p≤0.0005 (***) (t-test). Scale bar represents 2 mm.

Fig 10. Regulation of holometabolan and hemimetabolan metamorphosis.

(A) Expression profiles of Kr-h1, E93 and Br-C during the last larval and pupal stages (upper part) are from T. castaneum. Model depicting the regulatory interactions between the metamorphic toolkit genes in the prepupal and pupal stages that underlie the formation of the pupa and the adult using T. castaneum as a holometabolan model (lower part). Black arrows represent inductive effects, and red lines represent repressive effects. Gray colors denote genes and transcriptional regulatory events that are absent during each particular period. (B) Equivalent model depicting the regulation of hemimetabolan metamorphosis during the penultimate and last nymphal instar of B. germanica as a model. The expression profiles of Kr-h1, E93 and Br-C are from B. germanica [19,25,33]. (C) In the absence of the prepupal TcKr-h1 peak, the expression dynamics and regulatory interactions of the metamorphic toolkit genes during the last larval instar of T. castaneum closely resemble those in the last nymphal instar of the hemimetabolous B. germanica. Question marks denote unknown identities or functional relations. Thick black arrows above the expression profiles represent metamorphic periods.