millepattes micropeptides are an ancient developmental switch required for embryonic patterning

Small open reading frames (smORFs) that code for “micropeptides” (10-100 amino acids) exhibit remarkable evolutionary complexity. Conserved micropeptides encoded by the millepattes (mlpt) gene are essential in Tribolium for embryogenesis but in Drosophila, function only in leg and cuticle differentiation. We find that a module identified in Drosophila trichome patterning, comprising Mlpt, UBR3, and Shaven-baby (Svb), coordinates early embryo patterning in several insect orders. Intriguingly, Mlpt segmentation function can be re-awakened in the Drosophila blastoderm, demonstrating the potency of an ancestral developmental switch retained despite evolving embryonic patterning modes. smORFs like millepattes thus illustrate plasticity of micropeptide functions despite constraints of essential genetic networks. One sentence summary A module comprising the small ORFs mlpt/pri/tal, the transcription factor Svb, and the ubiquitin ligase UBR3, possesses an ancestral function in insect embryo patterning which is lost in flies but reactivated when Svb expression is restored.

phenotypes share several similarities, including transformation of abdominal segments towards thoracic identity, shortened sensory bristles, shortened leg segments with a "bubble-like" terminus, and missing telson appendages. (Fig 1. B-G). Perturbed animals also exhibit subtle phenotypic differences in number of ectopic leg bearing segments and body length. Severe ubr3 RNAi phenotypes are stronger, often with severely shortened larvae missing most abdominal segments, while embryo RNAi even at very low concentrations is lethal (no cuticles develop), supporting the involvement of this ubiquitin E3 ligase in mlpt-independent functions. ubr3 RNAi leg segments are also more rounded and the pretarsi shorter. Knockdown svb larvae are characterized by the presence of legs on the first two abdominal segments only; one or more legs on segment A1 are often reduced. Presence of T2-like spiracles on A1 and the absence of spiracles on A2 in svb knockdowns suggest transformation of the two anterior abdominal segments into thoracic segments, T2 and T3, respectively. Leg segments are more severely shortened and rounded and pretarsi more strongly reduced in svb knockdowns than in mlpt knockdowns (Fig 1, Supplement).
CRISPR knock-out of svb highlights an additional phenotype not apparent from RNAi knockdown, which is considerable thinning of the epidermal cuticle similar to what has been observed in the fly (Andrew DJ and Baker BS, 2008; Figure 4; Figure 1, Supplement 3). These results strongly support similar functions of Mlpt, Ubr3 and Svb in Tribolium and flies. Moreover, similarities in transcript, and protein structure and disorder disposition profiles between Drosophila and Tribolium Svb suggest similar molecular interactions (Figure 1supplement 1). As in flies, we observe that Tc-mlpt knockdown embryos accumulate the repressor form of Svb protein, as detected by an antibody recognizing the N terminal repression domain of Drosophila Svb (Figure 1-supplement 1). Taken together, our data support a cooperation among Svb, Ubr3 and Mlpt peptides, defining a conserved module essential for Tribolium segmentation.
Notably, expression of the segmental marker wg confirms that only abdominal segments are disrupted in mlpt, svb, and ubr3 RNAi embryos while thoracic segments are formed normally ( Fig. 1 H). This is of interest since in the "short germ" embryo of Tribolium only head and thorax form in the "syncytial" or acellular environment of dividing nuclei in the blastoderm, while after cellularization, the embryo undergoes convergent extension, and abdominal segments continue to arise in a sequential and regular manner from the posterior "growth zone" (Rosenberg MI et al, 2009, Liu PZ andKaufman TC, 2005). In Drosophila, although both pri/mlpt and svb are expressed throughout embryogenesis, there is little overlap in their expression domains until after germband retraction, as pri/mlpt is expressed in stripes resembling pair rule stripes, while svb is expressed only in the head ( Figure 4); ubr3 expression is ubiquitous. Since functional conversion to an activator via cleavage of Svb protein relies on the presence of Pri/Mlpt peptides, the timing and co-incidence of their expression is important. The observation that loss of pri segmentation function in flies coincides with absence of svb expression in the early fly embryo led us to investigate the mRNA expression of Mlpt partners Svb and Ubr3 in Tribolium.
We find that, as in the fly, svb and mlpt mRNA are expressed during both blastoderm and germband stages of embryogenesis in Tribolium, and importantly, are co-expressed in the early embryo as Tribolium svb mRNA is expressed maternally (data not shown). Zygotic expression of Tribolium svb also overlaps with mlpt in the pre-growth zone at the onset of gastrulation (Fig. 1I,  1I'). In the extending embryo, this overlap resolves into complementary expression: the posterior svb domain evolves into a strong anterior band and a more diffuse posterior expression ( Figure  1J'') while mlpt has much stronger posterior expression ( Figure 1J). As the germ rudiment matures, svb forms two distinct expression domains flanking the mlpt expression domain ( Figure  1K, K'). High levels of mlpt and svb expression may be mutually repressive (Figure 1, supplement 5). Subsequently, svb and mlpt expression domains shift, wave-like, anteriorly while anterior svb expression fades and its posterior expression detaches from the posterior end ( Fig. 1M'). Concomitantly, mlpt trunk expression moves anteriorly and a second posterior domain arises that is again complementary to the now anteriorly shifted svb domain ( Fig. 1 M), with fading anterior mlpt overlapping with the fading anterior svb domain ( Fig. 1M and M'). The interaction at the interfaces of the complementary domains may be critical for patterning of the abdominal segments. Posterior trunk expression of mlpt eventually also detaches from the growth zone and gradually fades while svb persists as a strong narrow band of expression until segmentation is completed (Figure 1-supplement 2). Late svb is expressed in presumptive neurons during germ band extension, and later, in the head and appendages ( Figure 1M', Figure  1-supplement 2).
In summary, Tribolium mlpt and svb expression patterns are dynamic and often complementary, though at times, overlapping. Loss of function phenotypes of Tc mlpt, svb and ubr3 suggest that a functional module for mlpt discovered in Drosophila trichome patterning also works in concert in embryonic segmentation, leg patterning and cuticle formation in Tribolium.

Complementarity of expression of pri/mlpt and svb is conserved deeply in insects.
Our data revealed a surprising and essential role for this gene module in controlling posterior segment formation and identity in Tribolium. To determine whether these genes are similarly important for embryonic development of other insects, we investigated their expression patterns in additional, more basal insect species: the water strider, Gerris buenoi ("Gb"; Hemiptera, Gerridae) and the milkweed bug, Oncopeltus fasciatus ("Ofas"; Hemiptera, Lygaeidae). Figure 2 highlights the expression patterns of these genes at several timepoints in embryogenesis. The early development of water strider and milkweed bug are quite similar. Ubr3 expression is ubiquitous in both Tribolium and Gerris and was not examined further. mlpt and svb expression in the early hemipteran embryo is observed in adjacent domains at the anterior of the precellular blastoderm embryo (e.g. Oncopeltus, Figure 2A Ofas mlpt expression is also diffusely expressed through recently added segments anterior to the GZ ( Figure 2C'). Later expression in both species is seen in the limb buds and mouth parts ( Figure 2C-C', E-E'), consistent with a function in patterning the leg and head appendages, as well as in presumptive neurons of the head and along the midline. These data hint at a surprising role for this gene module in controlling segment formation and identity in representatives of the Coleoptera and Hemiptera, but not Diptera.

Conserved function of mlpt gene module in insect segmentation.
We next tested whether and how broadly mlpt, svb, and ubr3 may functionally cooperate during embryogenesis in these additional short germ insects. We used RNAi against each of these genes in water striders and milkweed bugs, and find that all cause severe patterning defects.
Embryos of hemimetabolous insects, including water striders and milkweed bugs, complete embryogenesis and undergo a series of molts (which does not include a pupal stage) through which these hatchlings eventually achieve adult size and morphology. These intermediate nymph stages exhibit the full complexity of adult structures, allowing for detailed phenotypic analysis. In water strider and milkweed bugs, the wildtype hatchling possesses three long pairs of legs, which extend along the ventral side, curling around the posterior, as well as a long pair of antennae that extend posteriorly along the ventral midline ( Figure 3A-B; 3I-J). mlpt RNAi in both Gb and Ofas results in loss of posterior abdominal segments and fusion of thoracic segments, with shortened rounded legs that terminate proximal to the trunk; reduction and fusion of head appendages is also apparent (Fig 3C-D, 3K-L). In Ofas, severely affected embryos fail to gastrulate, resulting in an everted gut. Gb and Ofas svb RNAi also results in loss of abdominal segments and rounding of more distally truncated legs ( Figure 3E-F, 3M-N). ubr3 RNAi in both species gives the most severe phenotype, reflecting its presumed additional functions independent of svb and mlpt. Gb ubr3 RNAi embryos exhibit loss of posterior abdomen, as well as fusion or loss of most leg-bearing segments and reduced head appendages and eyes; remaining legs are fused and shortened with bubble-like termini ( Figure 3G-H). In Ofas, severe ubr3 RNAi embryos are almost completely ablated, leaving unidentifiable ectodermal tissue connected to everted presumptive visceral tissue. More mildly affected embryos possess some apparent segment identity, with head and eyes, but no appendages and limited evidence for correct axial polarity ( Figure 3O-P).

Conservation in alternative long germ insects.
Since all basal insect species examined showed evidence of conserved function of this module in segmentation, we tested whether other more derived species may retain this functionality. Like Drosophila, Nasonia, Apis, and other Hymenoptera exhibit "long germ" embryogenesis, in which the embryo is mostly patterned in the context of the syncytical blastoderm, and which has evolved several times in the insect phylum ( . Late expression of svb in Nasonia appears in stripes, prefiguring the larval trichomes. Late expression of mlpt is also in stripes, which are reminiscent of segmental genes en and wg, and coincident with svb expression, as well as in spots in the presumptive nervous system ( Figure 2-supplement 2). Thus, in a wide range of insects, complementary and sporadic overlapping expression of svb and mlpt in the embryo correlates with an essential role in embryonic segmentation.
In holometabolous insects, like Drosophila and Nasonia, the embryo hatches into a larva, whose cuticle is ornamented with dense bristles called trichomes. Since the pattern of trichomes, or denticles, is predictable and distinctive along the anterior-posterior (A-P) and dorsoventral axes and within each segment, it can serve as a readout for correct segmentation. The fly larva has a characteristic trichome pattern in WT flies ( Figure 4I), which is severely reduced (hence, "shaven") in the thin cuticles of larvae mutant for svb, pri, or ubr3 ( Figure 4J-L). However, all segments are still formed in these mutants. The larval cuticle pattern of Nasonia differs from the fly; three thoracic and 10 abdominal segmental bristle bands can be discerned, as well as 4 spiracles, located on thoracic segment T2, and abdominal segments A1, A2 and A3, which provide essential landmarks for segment identification (Pultz MA et al 2000; Figure 4M). In Nasonia, mlpt RNAi causes posterior truncation and segment fusions, evident as a severely shortened larval cuticle, with two remaining bands of denticles that are likely thoracic and anterior abdomen ( Figure 4N). Similarly, svb RNAi causes severe posterior truncation and loss of most abdominal segments. Two clear bands of denticles and two pairs of spiracles remain ( Figure 4O). Larvae from ubr3 RNAi were almost uniformly too fragile to recover (not shown), likely owing to the observed absence/thinning of cuticle; mildly affected ubr3 RNAi larvae exhibit thin cuticle, devoid of denticles on the posterior, with anterior fusion of remaining denticles ( Figure 4P). Altogether, our data support conserved functions for mlpt, svb and ubr3 in embryonic segmentation of a long germ insect, leaving only long germ Drosophila from among species tested without such an early patterning function.
Reawakening functionality mlpt/svb/ubr3 module in segmentation by restoring Svb expression in the early Drosophila embryo. Since we find this functional module to be ancestral and deeply conserved in both short and long germ insects, we sought to determine how the module lost its segmentation role in Drosophila. ubr3 is ubiquitous and mlpt is expressed in pairrule like stripes, but svb expression is absent at these early Drosophila embryonic stages. We therefore hypothesized that the loss of the segmentation function of this gene module occurred through the loss of svb expression during embryogenesis in the Drosophila lineage. To test this hypothesis, we added back svb expression to the early embryo context using the Gal4/UAS system (Brand AH & Perrimon N, 1993), as it is observed in the blastoderm embryos of Tribolium, Oncopeltus, Gerris, and Nasonia. Strikingly, over-expression of svb in the early embryo resulted in segmentation defects ( Figure 5B, middle panels) while loss of mlpt function did not (Galindo MI et al, 2007). This result suggests that the function of the mlpt/svb/ubr3 module in segmentation is contingent upon expression of all three partners and that the loss of this function in flies evolved through the loss of svb expression and the dissociation of the trio during early embryogenesis. We next tested whether the segmentation defects observed were due to the processing of the ectopically expressed svb into an activator by mlpt and ubr3. To do this, we over-expressed OvoB, a transcriptional activator originating from the same locus as Svb that behaves like the cleaved form of Svb but that does not require cleavage by Mlpt and Ubr3. Maternal (ectopic) expression of OvoB activator causes segmentation defects that are reminiscent of those obtained by Svb over-expression ( Figure 5B, right panels). This result indicates that the segmentation defects observed upon over-expression of either Svb or OvoB are the result of the accumulation of the activator form of Svb/Ovo protein, and that, in this context, ectopic Svb can be regulated and converted to activator form by Mlpt peptides. We therefore conclude that the cooperativity of this gene module has remained intact throughout evolution, and the inactivation of its function in Drosophila segmentation likely occurred through the promoter-mediated abrogation of early expression of Svb, an essential component of the module.

Discussion.
In summary, our experiments reveal how a cooperative trio of genes, initially discovered in Drosophila where it has a more restricted functional capacity, possesses important ancestral functions in insect embryonic segmentation, representing a significant addition to the anteriorposterior patterning network in insects. Segmentation function of this module has receded in the Drosophila lineage through the dissipation of temporal coexpression, while retaining cooperative potency, a solution that preserves connectivity for use in other processes in the fly, such as leg patterning, trichome patterning, and metamorphosis. We surmise that the loss of the essential function of Svb/Mlpt/Ubr3 in segmentation of flies was facilitated by the shift to complete segment patterning before cellularization, which leaves no segments aside for delayed patterning, thus obviating the need for a switch between modes of segmentation. Our current study also illustrates how the use of non-Drosophilid insects with varied modes of embryogenesis can uncover a deeply conserved function, which would have been missed using only the well-studied Drosophila model.
We observe several conserved functions of this module of mlpt, svb and ubr3. All insects from the most basal studied (Gerris) to the most derived (Drosophila) exhibit strong epidermal patterning defects, evident both in the patterns of apical projections on the larval cuticles of flies and wasps, and in the thinning of the cuticle (Figures 1, 4; Figure 1, Supplement 3) as well as defects in leg patterning in all species examined ( Figure 1G, Figure 3). The conserved outputs of this module highlight transcriptional networks downstream of Svb whose connectivity are largely intact over large evolutionary distance, as are sequence features within Svb protein (conserved Nterminus, proteolytic cleavage sites) required for mlpt/ubr3 mediated cleavage ( Figure 1supplement 1). Several Svb epidermal targets identified in flies are similarly regulated in Tribolium (Li C et al, 2016). Expression patterns of several Svb epidermal target genes in Nasonia support this conclusion (Figure 1, Supplement 4). Together, these observations underscore the ancestral conservation of both Svb transcriptional networks and phenotypes dating to early in the radiation of arthropods.
Outside of Drosophila, we observe conserved function in insect embryonic segmentation, affecting formation of posterior segments in all insects tested, as well as additional defects in head and thoracic segment derivatives, delineating a new module for insect embryonic segmentation. Our experimental data show that promoter control of Svb expression timing may be sufficient to bring segment patterning potency on-or off-line in the embryo ( Figure 5). The phylogenetic distribution within insects of short/long germ modes of development, whose hallmark is delayed abdominal segment formation and patterning, suggests that evolution has likely repeatedly sampled these modes (Misof, B. et al, 2014). Our data suggest one mechanism by which delayed posterior segment formation may be switched on/off via Svb/Mlpt/Ubr3.
The highly dynamic expression of mlpt and svb and the complementarity of their expression patterns, seen in every non-Drosophilid insect examined in this study, are striking. Mlpt/svb boundaries occur at varying positions in different insect embryos. Their expression domains also overlap at times, as in the head in the early Drosophila embryo, and later, in the same denticle prepattern. We often observe a sharp boundary in non-Drosophilid insects between two domains of strong expression, suggesting mutual exclusivity between high svb and mlpt expression, or a negative feedback loop. Indeed, we observe derepression of svb when mlpt is reduced by RNAi (Supplementary Figure 7). The position of the complementary Svb/Mlpt expression boundaries at the interface between patterned segments and unpatterned cells in the growth zone is intriguing. A strong domain of svb expression in the growth zone is observed in all short germ species examined, often adjacent to a strong mlpt expression domain. In cells distant from the boundary, high levels of Svb repressor protein that excludes Mlpt expression would silence target promoters, while Svb targets in the region of overlap at the junctions between the svb-mlpt domains would be activated due to more moderate Svb repressor levels that are no longer inhibitory to Mlpt expression. Alternatively, potential non-cell autonomous activity of Mlpt peptides could activate Svb and its target promoters close to the boundary even in the absence of an overlap between the mRNA expression domains. The potential of this interface to produce a narrow adjacent domain of Svb target activation is indeed interesting. The short-range gradient of Svb activator could constitute a retracting wavefront (regulated by or acting in conjunction with a speed/frequency regulator; Zhu, X et al, 2017) that determines spatial patterning. The position of the complementary Svb/Mlpt expression boundaries at the interface between blastoderm and oscillation-driven growth zone invites further study.
Beyond insects, Svb/Ovo is conserved in all animals, and predates bilateria ( In conclusion, our experiments build on the strength of deep evolutionary understanding of a micropeptide switch and its genetic context to revive ancestral functionality in an essential developmental network. This process demonstrates potential for the rational design of novel biological switches, using evolutionarily conserved components in contexts in which they no longer possess native function, but in which they are poised to do so.

Materials and Methods.
Experimental Design. All insect cultures were reared at ambient temperature of 25ºC, unless otherwise noted. Wild type Nasonia embryos were collected from virgin AsymCx (34) females host fed on Sarcophaga bullata pupae (Carolina Biological), aged as needed at 25ºC, and fixed for 25 minutes in 4% heptane-saturated formaldehyde/1X Phosphate Buffered Saline (PBS), with vigorous shaking. Embryos were transferred to 100% methanol for storage, or further processed. Embryos were then rehydrated to 1X PBS with 0.1% Tween (PBT) through a methanol/PBT series, and hand peeled under 1X PBT using 1mL insulin needles (Becton-Dickinson).
Wild-type Oncopeltus embryos were collected on cotton from mated females, and aged, as needed, in a 25ºC incubator. Embryos were first boiled for 1 to 3 minutes in a microfuge tube in water, followed by a 1 minute incubation on ice, before further processing. Embryos were fixed in 12% heptane-saturated formaldehyde/1X PBS for 20 minutes with shaking. The heptane was replaced by methanol, and the embryos either stored under methanol at -20ºC or processed immediately. Embryos were then rehydrated to 1X PBT through a methanol/PBT series, and dechorionated, before further fixation for 90-120 minutes in 4% formaldehyde/1X PBT. Embryos were then transferred to and stored in 100% methanol.
Wild type Gerris buenoi were collected from a pond in Toronto, Ontario, Canada and established in the lab. Stocks were maintained in aquaria at 25 °C with a 14-h light/10-h dark cycle, and fed with fresh crickets. Styrofoam float pads were provided to females as substrate for egg lays. Embryos were collected and incubated at 20-25°C until desired developmental timepoints, at which time they were dissected in 1x PBS with 0.05% tween-20 ("PTW"). Once dissected, embryos were fixed in 4% paraformaldehyde and stored under 100% Methanol at -20ºC until use.
Wild type Drosophila lines were reared at ambient temperature under standard conditions. Wild type Tribolium embryos were fixed and single whole-mount in situ hybridizations were performed as previously described (35,36,37). Digoxigenin-labeled RNA probes were detected using alkaline phosphatase-conjugated anti-DIG antibodies (1:2000; Roche) and NBT/BCIP substrates (Roche) as per manufacturer instructions.
Antibody staining of Tribolium embryos: Freshly fixed embryos were washed with PBT (PBS and 0.1% Triton X) and exposed to rabbit anti-Svb 1S at a concentration of 1:2000 for 4h at room temperature on a rotating wheel. Following washes, the embryos were incubated with preabsorbed alkaline phosphatase conjugated goat anti-rabbit secondary antibody at a concentration of 1:1000 for 2h at room temperature on the wheel. Following washes the embryos were stained with NBT-BCIP substrate and visualized under a stereomicroscope.

Parental RNAi (pRNAi) and larval/hatchling cuticle preparation:
Tribolium pRNAi: dsRNA synthesis and parental RNAi were performed as described previously (38; 39). dsRNAs were injected into female pupae or virgin adult females at a concentration of 1-3µg/µl. RNAi phenotypes were confirmed by injection of non-overlapping dsRNA fragments.
First instar larval cuticles were cleared in Hoyer's medium/lactic acid (1:1) overnight at 60 °C. Cuticular autofluorescence was detected on a Zeiss Axiophot microscope. Z stacks and projections were created with a Zeiss ApoTome microscope using the Axiovision 4.6.3.SP1 Software. Colour images were taken by (ProgResC14) using the ProgResC141.7.3 software and maximum projection images were created from z stacks using the Analysis D software (Olympus).
Oncopeltus pRNAi: dsRNA template was amplified from target gene fragments which had been cloned into either pCR-Topo (Qiagen) or pGEM (invitrogen), using T7 promoter-containing oligos, as described previously (40). Purified PCR product was used for dsRNA transcription using Megascript RNAi (Ambion) according to manufacturer's instructions. dsRNA was injected into newly eclosed virgin female milkweed bugs, at a concentration of 1-3µg/µl. After injection, females were mated to uninjected males, and embryos were collected for the duration of egglaying. Embryos for phenotypic evalution were incubated at 28ºC for 8 days, and unhatched embryos were dissected from their membranes and imaged for hatchling phenotypes.
Gerris buenoi parental RNAi. Knock down of pri, svb and ubr3 genes was performed by injection of double-stranded RNAi into adult females abdomen (41). dsRNA was synthesized by in vitro transcription and purified following the protocol described in (40) Nature Protocols. dsRNA was injected at a concentration of 1ug/uL in 1X injection buffer prepared following the protocol in (43). dsRNA directed against the yellow fluorescent protein gene was used as an injection control. Nasonia embryo RNAi: dsRNA template was amplified from target gene fragments that had been previously cloned into pCR-Topo (Qiagen) or directly from embryo cDNA, using standard T7 promoter-containing oligos, as described previously (41). Purified PCR product was used for dsRNA transcription using Megascript RNAi (Ambion) according to manufacturer's instructions, and purified product diluted to 1-3µg/µl for injection. pRNAi for mlpt and svb in Nasonia resulted in sterility. Therefore, haploid embryos laid by unmated host-fed virgin Nasonia females were microinjected with dsRNA using a Femto-Jet microinjector (Eppendorf), and transferred to a slide to develop in a humid chamber at 28ºC for 36H. Unhatched larvae were dissected from extraembryonic membranes and cleared in freshly prepared Lacto:Hoyer's medium overnight at 65ºC, and imaged for cuticle autofluorescence the following day.
Immunostaining on Gerris embryos: Embryos were cleaned with dilute bleach solution and washed PTW 0.05%. After dissection, embryos were fixed for 20min in 4% Formaldehyde/1X PTW 0.05%. Embryos were then permeabilized with PBT 0.3% for 30 minutes and incubated in antibody blocking solution (1X PBS; 0.1% Triton X100; 0.1% BSA; 10% NGS) at room temperature for 1 hour. Embryos were transferred to blocking solution containing primary antibody (mouse anti-Ubx-AbdA Hybridoma bank 1:5 dilution; rabbit anti-Dll (42), catalog number; 1:200 dilution) and incubated overnight at 4°C. The next day embryos were washed in PTW 0.05% (two quick washes, then two longer washes 10 min each) and incubated for 30 minutes in blocking solution at room temperature with shaking, before adding the secondary antibody (Rabbit anti-mouse-HRP [1:1000] from Promega or Donkey anti-Rabbit-HRP [1:500] from Jackson Immuno research) diluted in PTW. All secondary antibodies were incubated with embryos for 2 hours at room temperature with shaking. Following secondary antibody, embryos were rinsed in PBT 0.3% and PTW 0.05% three times each for 10 minutes at room temperature. Before enzymatic developing with DAB with color enhancer (Diaminobenzidine tetrahydro-Chloride from Sigma), embryos were briefly incubated with DAB solution for 5 minutes at room temperature. Upon completion, staining was stopped by washing the embryos briefly in PBT 0.3%, followed by 5 times, five minute washes in PBT 0.3%. Five more washes of 5 minutes in PTW 0.1% followed. Embryos were transferred to 30% glycerol/1X PBS for 5 min, and then 50% Glycerol/1X PBS for 5 min, before sinking in 80% glycerol/1X PBS at 4°C until mounting in 80% glycerol under coverslips for imaging.
Oncopeltus in situ hybridizations were carried out (as described for Nasonia in 14) on embryos peeled and stored under 100% methanol, and rehydrated through an methanol/1x PBS, 0.1%Tween (1xPBT) series. Briefly, rehydrated embryos were washed several times in 1x PBT before a 5 minute post-fix in 5% formadelhyde/1X PBT, followed by 3 five minute washes in 1X PBT. Embryos were briefly treated with Proteinase K (4µg/mL final concentration) in 1X PBT for 5 min, followed by 3 five minute washes in 1X PBT, and an additional 5 minute post-fix in 5% formaldehyde/1X PBT. Following 3 x three minute washes in 1X PBT, embryos were incubated in hybridization buffer for 5 minutes at room temperature, followed by incubation in fresh hybe buffer for a 1 hour pre-hybridization step at 65ºC. RNA probes were prepared and added to a fresh portion of hybridization buffer and incubated at 85ºC for 5 minutes, then one minute on ice, before replacing pre-hybe with hybe buffer containing denatured RNA probe. Tubes were incubated overnight at 65ºC. After washes in formamide wash buffer, embryos were washed in several changes of 1X MABT buffer, before incubation in 1X MABT+ 2% Blocking Reagent (BBR; Roche Diagnostics) for 1 hour, and then 1X MABT/2%BBR/20% sheep serum for an additional hour, before addition of fresh 1X MABT/2%BBR/20% sheep serum containing anti-DIG AP Fab fragments (1:2000; Roche) for overnight incubation at 4ºC. In the morning, extenstive 1X MABT washes were carried out before equilibration of embryos with AP staining buffer and then staining with AP staining buffer containing NBT/BCIP (Roche; as per manufacturers instructions). After staining, three 1X PBT washes were carried out before a final post-fixation step (5% formaldehyde/1xPBT), and then one PBT wash before sinking in 50% glycerol/1xPBS, and then 70% glycerol/1xPBS, which was also used for mounting before imaging.
Nasonia in situ hybridizations. Nasonia embryos were collected and fixed in 4% formaldehyde/1X PBS/Heptane for 28 minutes with moderate shaking. The embryos were handpeeled under 1x PBS+ 0.1% tween, after being affixed in room air to double sided tape, and hand-peeled as described previously, except that embryos were collected from host-fed, mated females. Fixed embryos were stored under methanol at -20ºC until use.
In situ hybridizations were carried out as described previously (44). Briefly, fixed embryos that had been stored under methanol were gradually brought up to 1X PBT in a PBT/MeOH series, and washed three times in 1x PBS +0.1% tween-20 (PBT) before a 30 minute post-fixation in 5% formaldehyde/1XPBT. The embryos were then washed three times in 1X PBT, and digested in Proteinase K for five minutes (final concentration of 4mg/mL) before three PBT washes. Embryos were blocked for 1 hour in hybridization buffer before probe preparation (85ºC, 5'; ice 1') and addition for overnight incubation at 65ºC. The next day, embryos were washed in formamide wash buffer three times, and and then 1X MABT buffer three times, before blocking in 2% Blocking Reagent (BBR; Roche Diagnostics) in 1X MABT for 1 hour, then in 10% horse serum/2% BBR/1XMABT for 2 hours. Embryos were incubated overnight at 4ºC with primary antibody (anti-DIG-AP Fab fragments; Roche, 1:2000). The third day, embryos were washed in 1X MABT for ten x 20 minutes washes before equilibrating the embryos in AP staining buffer and developing in AP buffer with NBT/BCIP solution (Roche Diagnostics). After staining, embryos were washed in 1x PBT three times for five minutes each before a single 25 minutes post-fix step in 5% formaldehyde/1XPBT. Embryos were then washed several times with 1X PBT, and allowed to sink in 50% glycerol/1XPBS and then 70% glycerol/1XPBS, which was subsequently used for mounting.

Cloning of Tribolium cDNA sequences.
For Tribolium svb, all primer pairs shown were used to generate template for dsRNA synthesis. Amplicons generated by the last four pairs were also used for antisense RNA probe synthesis. dsRNA fragments corresponding to different regions of the svb transcript were used for gene knockdown by RNAi. All dsRNA fragments resulted in similar knockdown phenotypes with high penetrance. Primers were designed based on the Next-RNAi software, Primer3 or MacVector. The nucleotides shown in red indicate tags of parts of T7 (3' primer) and SP6 (5' primer) promoter sequences attached to gene-specific sequences in the manner described by Schmitt-Engel et al. 2015. The products were used for a second PCR using T7 and SP6-T7 primers for generating a double stranded template for in-vitro transcription by T7 polymerase. For in-situ RNA probes, the second PCR was done using the complete T7 and SP6 promoter sequences and subsequently in-vitro transcription was performed to generate a digoxigenin-labeled antisense RNA probe with the appropriate polymerase. Amplicons that were cloned into pBluescript vector were amplified with T7 and T3-T7 primers for subsequent dsRNA synthesis or T7 and T3 primers for subsequent antisense RNA probe synthesis using either T3 or T7 RNA polymerase. The primer design was based on the RNAseq data (Tcas au5 prediction) for svb available on iBeetle-Base. For mlpt dsRNA and probe synthesis, a full-length mlpt cDNA cloned into pBluescript was obtained from Dr. Michael Schoppmeier. For ubr3, all primer pairs shown were used for dsRNA synthesis. All dsRNA fragments resulted in similarly strong knockdown phenotypes with very high penetrance. The fragments generated with the primers containing iBeetle numbers were also used as probes.

Generating a Tcas svb mutant using CRISPR/Cas9:
The svb gRNAs were directed to the putative transactivation domain in exon 2 of Tc-svb. The G (required by the U6 promoter for transcription initiation) is marked in green. The PAM sequence is shown in blue. The sequences in orange represent the complementary overhangs generated by Bsa1 digestion. A fourth gRNA was directed to the Tribolium single-minded gene (Tc-sim, Rode and Klingler, unpublished). Embryonic injection mix consisted of 125 ng each of the four gRNA expression vectors, 500ng of the donor eGFP vector containing the sim target sequence, and 500ng of the Cas9 expression vector. Non-homologous end joining (NHEJ) method was employed for directed knock-in of an eGFP containing donor marker plasmid into the exon2 of the endogenous svb gene. The sim gRNA was used to target the sim sequence in the marker plasmid leading to its Cas9-induced linearization. This was followed by insertion of the linearized plasmid into one or more target sites in the svb genome. A successful knock-in of the marker plasmid was obtained only at gRNA target site 3. This insertion site was present in all svb transcripts and was also downstream from a putative second start codon, thus increasing the chances of obtaining a svb null phenotype.    RNAi results in larvae with little or no cuticle that are impossible to recover. Here, the more mild larval phenotype includes a shortened larva with a thin, naked cuticle at the posterior, and few denticles in the anterior. and Svb in wild type embryos (left embryos) and in embryos expressing D.mel Svb maternally from a strong ubiquitous promoter (right embryos). (B) Phenotype of maternal expression of Svb in D. melanogaster. Top panels are larval cuticle preps showing the pattern of ventral denticles that are used as a proxy for proper patterning. Left most panels show denticle pattern and engrailed protein staining in wild type Drosophila melanogaster. Middle panels show denticle pattern and engrailed protein staining in D. mel overexpressing Svb from a ubiquitous maternal promoter (a-tubulin). Right panels show denticle pattern and engrailed protein staining in D.mel overexpressing OvoB activator from the maternal a-tubulin promoter.