Temporal control of RNAi reveals both robust and labile feedback loops in the segmentation clock of the red flour beetle

Significance The generation of repetitive body parts such as ribs and vertebrae or insect segments has been a very successful evolutionary invention. How these repetitive units form during embryonic segmentation has been of key interest to developmental biologists. So far, they mostly used permanent knock-down of gene function for respective studies in insects. Using a unique tool to end a gene knock-down effect with temporal control, we find both robust and labile feedback loops within the insect segmentation machinery. The robust downstream loops probably buffer ongoing segmentation against external disturbance while the labile upstream loops may ensure that only one trunk is formed.


Supplementary Text
1) Establishment of CrPV1A as a tool for suppressing RNAi in T. castaneum Six VSRs were tested in order to identify a strong Viral Suppressor of RNAi (VSR) in T. castaneum.
First, we evaluated if any of those VSRs would be able to block RNAi targeting an exogenous gene expressed from a transgenic construct.For this purpose, we crossed a UAS-tGFP responder line (10)   In a second experiment, we tested the VSR effect with respect to RNAi targeting an endogenous gene.Larval RNAi against the pigmentation gene Tc-ebony was performed.This gene is an Nbeta-alanyl dopamine synthetase that is necessary for the synthesis of NBAD sclerotin.RNAi knocking down Tc-ebony leads to darkened or black body pigmentation.In this experiment, the UAS-VSR responder lines were crossed to a Gal4 driver, which showed ubiquitous Gal4deltaexpression in the epidermis except for early embryonic stages (until appr.2) The VSR CrPV1A Cricket Paralysis virus (CrPV) was initially identified and isolated from field crickets, Teleogryllus oceanicus and Teleogryllus commodus, and it is a highly potent virus of many species in the laboratory (4,11,11).CrPV is closely related to Drosophila C virus and likewise belongs to the positive-strand Dicistroviridae family.In contrast to DCV, CrPV leads to mortality upon infection of crickets and flies (4,12).This high pathogenesis of CrPV is partially based on its efficient RNAi suppressor protein, CrPV1A.It has been shown that adding CrPV1A to the Sindbis virus, which does not naturally encode an endogenous RNAi suppressor, resulted in increased virus production and fly lethality upon infection (4).The mode of action of CrPV1A relies on its interaction with the endonuclease Ago-2, a component of the RISC complex.This interaction blocks Ago-2 cleavage activity, resulting in inhibited RISC-mediated mRNA degradation and therewith RNAi disruption.Nevertheless, the suppressor protein CrPV1A did not interfere with the miRNA pathway or alter the physiology and development of the animals when expressed in flies (4).In order to check for lethal effects of VSR expression in T. castaneum, we heat-shocked animals of the vw strain wildtype and of the hsVSR-line at

3) Effect of heat-shock on segmentation gene expression
To test for the effect of heat-shock on patterning gene expression we performed the procedure given in Fig. 1F for wildtype animals (SB strain) with or without heat-shock and stained them for Tc-wg, Tc-eve and Tc-cad.Then, embryos were scored for wildype and aberrant patterns.Tc-eve has a very dynamic pattern such that minor modifications are difficult to detect.Therefore, we scored for embryos that showed 2 or more stripes (reflecting the expected pattern for ongoing segmentation) versus animals that had clearly aberrant patterns or only one posterior domain (reflecting either aberrant pattern in elongating embryos or a regular pattern observed at the end of segmentation).We found no difference between heat-shocked or non-heat-shocked.
Similarly, all embryos of both treatments had the expected single posterior domain.For Tcwingless, we found that 50% of the heat-shocked embryos had localized defects in some abdominal segments (Fig. D) while in non-heat-shocked controls this was not observed.
These observations would be in line with a temporally restricted effect of heat-shock on segmentation.The genes acting early in the hierarchy (Tc-cad, Tc-eve) had probably already recovered from potential earlier defects while the observed spatially restricted Tc-wg defects probably reflected patterning problems induced during the temporally limited heat-shock.This confirms our observation from cuticle analysis that a heat-shock leads to unspecific defects in a subset of embryos.

Non heat-shocked
Heat-shocked Tc-eve 2   Regular stripes posterior to the defect domain indicate that segmentation resumed after some time.

Strains and husbandry
Tribolium castaneum (HERBST) beetles were reared using standard conditions and methods (52).During experiments, beetles (embryos/larvae) were kept at 32°C and 40% RH while general stock keeping was done at 28°C and 40% RH.RNAi inhibition experiments were performed in the transgenic line containing the RNAi inhibitor CrPV1A from the Cricket Paralysis Virus under the control of the endogenous Tribolium heat shock promoter (40) and the 3xP3DsRed (eye marker) integrated into the genome using piggyback vector (57) in Tribolium line vermillion white .Nontransgenic vermillion white beetles were used for negative controls.

RNAi and heat-shock treatment
Parental RNAi was performed according to established methods (5,58).Templates were prepared by PCR with T7 overhanging-primers from plasmid templates containing varying lengths of coding and/or regulatory mRNA sequence (Tc-even-skipped ~1400 bp, Tc-odd-skipped ~380 bp, Tc-paired ~540 bp, Tc-arrow ~1800 bp, Tc-Wnt8/D ~500 bp, Tc-wntless ~600 bp).DsRNA was produced using MEGAscript T7 Transcription Kit (Life Technologies).The concentration of injected dsRNA for parental RNAi was 1000 ng/μl (Tc-even-skipped), 500 ng/μl (Tc-odd-skipped, Tc-paired), or 100 ng/μl (Tc-arrow, Tc-Wnt8/D, Tc-wntless).Add NCBI accession number?Tc-eve: NM_001039449.1;Tc-odd: XM_008198532.2;Tc-run: XM_964184.3;Tc-wg: NM_001114350 Heat-shock conditions for T. castaneum had been extensively optimized in a previous work (40).Starting from these values, we optimized the conditions using the number of surviving embryos and the number of rescued segments as criterion.Embryos from dsRNA injected animals of either the hsVSR or wild type control (vermillion white ) were collected and kept without flour at 32°C in small plastic fly culturing vials before the treatment.For the heat-shock, the eggs were transferred to a small (40ml) glass beaker with a flat bottom making sure that all embryos had direct contact with the bottom.Then, the beaker was put into a pre-heated 48°C waterbath.The beaker was covered with perforated aluminum foil.The bottom of the glass beaker was kept submerged for 10 min.To ensure a controlled and quick termination of the heatshock, the beaker was put into a room temperature waterbath.After transferring the embryos back into the plastic vials, they were allowed to recover for two hours at 32°C, until they were heatshocked a second time for 10 min at 48°C, following the same procedure.Based on (40), the VSR mRNA is predicted to be expressed for about 5 hours by that treatment.In the absence of an VSR antibody we were not able to measure the protein dynamics.Thereafter, the embryos were kept at 32°C until fixation or cuticle preparation.We note that the results were sensitive even to minor changes of the procedure and each step needed to be optimized carefully.

Staining, and microscopy
Embryo fixation and in-situ hybridisation were performed as described previously (59).Digoxigenin (DIG)-labeled riboprobes targeting Tc-wg (DIG RNA Labeling Kit, Roche), was detected by anti-DIG-AP antibodies (Roche) and visualized by NBT/BCIP staining.HCR staining was performed as published with small modification (kindly provided by Eric Clark and Olivia Tidswell prior publication) (60, 61).HCR probes for Tc-eve were purchased from Molecular Technologies while HCR probes for all other genes were purchased from Molecular Instruments.Binding sequences are available from vendors at request due to intellectual properties restrictions.Cuticles of L1 larvae were analyzed and documented using either a Zeiss AxioPlan 2 (10x air objective) with ImagePro 6 or a Leica SP5 inverted cLSM (10x air objective) with Leica LAS-X software, utilizing the cuticle's autofluorescence.HCR stainings were documented using a Leica SP8 confocal laser-scanning microscope (20x objectives with 100% glycerol as immersion medium) and the Leica LAS-X software (v 3.5.2).In-situ hybridization was documented using Zeiss AxioPlan 2 or Zeiss AxioScope.

qPCR
RNA from whole embryos was extracted using the Quick-RNA Tissue/Insect Kit (Zymo Research) with DNase on-column digest (DNaseI Set, Zymo Research).cDNA was synthesized using the MAXIMA First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Fisher Scientific) according to manufacturer's instructions.qPCRs were performed using the CFX96 Real-Time PCR System (Bio-Rad Laboratories) with 5x HOT FIREPol® EvaGreen® qPCR Mix Plus (ROX) Master mix (Solis Biodyne).Reference genes were identified using RefFinder (62).qPCR data analysis was done in the CFX Manager 3.1 (Bio-Rad Laboratories) and pyQPCR with the delta-delta-Ct method (63).

Statistical analysis
Comparisons of abdominal segment numbers in cuticles and comparisons of number of expression stripes in germbands were tested using unpaired, two-sided Mann-Whitney U tests for independent samples.All measured data points were included in the calculations and were not checked for outliers beforehand.Outliers were determined for the plots using the R package ggplot2, considering data above 1.5 *IQR of the 75th percentile or below 1.5 *IQR of the 25th percentile as outliers, which are indicated in in the respective plots in red.Comparisons of the number of stripes in germbands were done using the Pearson's Chi-squared Test for Count Data with simulated p-values by Monte Carlo simulations (B=1000).All graphs and statistical calculations were performed using R (v3.5.2;R Core Team, 2018) and RStudio (v1.1.x;RStudio Team, 2015) with the following packages: dplyr, ggplot2, ggpubr, ggsigni, patchwork, readxl, reshape2.B) The early germ band can be separated in head and thoracic segments (red and purple), which are specified during the blastoderm stage and the SAZ (green).C,D) During elongation, abdominal segments are added in a sequential way in the SAZ.The depicted stages correspond approximately to the stages in which heat-shocks were performed.E) Heat-shocks alone did not much affect anterior segments.F) Importantly, the earlier heat-shocks did affect the number of abdominal segments indicating that this treatment disturbs segmentation to some degree.G and H) Repetition of the Tc-prd experiment (see text for details).In this repetition, the negative influence of the heat-shock is more pronounced than in the experiment shown in Fig. 2.   A) The number of pair rule gene and Tc-wg stripes increased significantly when comparing the heat-shocked batches from the hsVSR line and wildtype.B) In an alternative analysis, the resulting embryos were assigned to three classes: no stripes (-), intermediate (+/-) and close to wildtype (WT).The p-value for Tc-eve reached significance levels while for the other genes the p-value was low but not significant.qPCR was performed in the hsVSR line (reddish colors) and vw wildtype (blue colors).Rps3 and gapdh were used for normalization.As expected, heat-shock treatment increased the amount of VSR transcript (VSR -compare bars with heatshock ("pos") with those without heatshock ("neg").Surprisingly, Tc-eve expression was not much reduced after Tc-eve RNAi.However, nuclear transcripts did increase dramatically in eve RNAi embryos that were not rescued (eve intron, intermediate red bar).Tc-odd expression was not much altered in line with ongoing expression in the knock-down embryos.Tc-runt may be downregulated in Tc-eve RNAi (middle bar) but the expression was rescued in Tc-eve RNAi combined with hsVSR rescue (light red bar).
(red eye marker) to a driver line that expressed Gal4delta in a thoracic region during late larval, pupal and adult stages (name: Bauchbinde-Gal4; black eye marker).The double heterozygous offspring of this mating showed a strong thoracic fluorescence (Fig. A -leftmost columns).In parallel, we generated animals where in addition to UAS-tGFP also the UAS-VSR construct was present (right columns).We performed RNAi targeting tGFP both, in the cross without UAS-VSR (middle column) and with UAS-VSR (right column; see table for overview).The RNAi was predicted to deplete tGFP expression unless the UAS-VSR was able to block RNAi.In this experiment, only CrPV1A blocked RNAi to a degree that tGFP expression remained visible (Fig. A top left panel, red box).

Figure A
Figure A Testing six VSRs for inhibition of RNAi.See text for details 24h at 32°C; we did not check internal organs; name: Boje-Gal4).In the presence of an active VSRs a reduction of the portion of black animals and an increase of animals with wildtype color was expected.Again, CrVP1A showed a complete rescue for one insertion (Fig. B, panel D) and partial rescue for another insertion (Fig. B, panel C).The FHV B2 VSR was active to a lesser degree (Fig. B panels E,F -two different insertions led to an increase of wt coloured animals).Taken together, these tests revealed CrPV1A as an efficient suppressor of RNAi in T. castaneum transgenically expressed.Based on these results, we chose CrPV1A as tool for temporal control of RNAi in T. castaneum.

Figure B
Figure B Testing six VSRs for inhibition of RNAi See text for details different stages of development (adult, pupae and larvae).The survival rate was reduced significantly in heat-shock-VSR line animals compared to wildtype (Figure C) indicating adverse effects.

Figure C
Figure C Survival rate after heat-shock activation of CrPV1A Light grey columns indicate the survival rate of wildtype animals after heat-shock whereas dark filled columns indicate the survival rate of hsVSR animals after heatshock treatment.After heat-shock activation of viral suppressor of RNAi, the survival rate was reduced significantly in heatshock-VSR line animals compared to wildtype animals indicating an adverse effect of CrPV1A expression.

Figure D
Figure D Developmental defects induced by heatshock A) Embryo from the non-heat-shock control showing the typical pattern of segmental Tc-wg staining in addition to the domain in the growth zone.B-D) Embryos from the heat-shocked sample showing deviations of segment formation in the abdomen.Regular stripes posterior to the defect domain indicate that segmentation resumed after some time.

Fig. S1
Fig. S1 Embryonic schemes and replicate of Tc-paired experiments Fig. S1 Embryonic schemes and replicate of Tc-paired experiments

Fig. S5
Fig. S5 Definition of developmental stages based on head expression of Tc-wg

Fig. S7
Fig. S7 Quantification of gain of striped pPRG expression after hsVSR treatment

Fig
Fig. S8 Rescued Tc-eve embryos and classification See text for details