Developmental Biology 402 (2015) 264–275

Contents lists available at ScienceDirect

Developmental Biology

http://d0012-16

n CorrE-m

journal homepage: www.elsevier.com/locate/developmentalbiology

Evolution of Developmental Control Mechanisms

The single fgf receptor gene in the beetle Tribolium castaneum codes fortwo isoforms that integrate FGF8- and Branchless-dependent signals

Rahul Sharma, Katharina Beer, Katharina Iwanov, Felix Schmöhl, Paula Indigo Beckmann,Reinhard Schröder n

University of Rostock, Biological Sciences, Department of Genetics, Albert-Einsteinstr. 3, 18059 Rostock, Germany

a r t i c l e i n f o

Article history:Received 3 December 2014Received in revised form31 March 2015Accepted 1 April 2015Available online 9 April 2015

Keywords:FGFRTriboliumSignallingEvolutionExon-specific RNAi

x.doi.org/10.1016/j.ydbio.2015.04.00106/& 2015 Elsevier Inc. All rights reserved.

esponding author.ail address: reinhard.schroeder@uni-rostock.d

a b s t r a c t

The precise regulation of cell–cell communication by numerous signal-transduction pathways is funda-mental for many different processes during embryonic development. One important signalling pathway isthe evolutionary conserved fibroblast-growth-factor (FGF)-pathway that controls processes like cellmigration, axis specification and mesoderm formation in vertebrate and invertebrate animals. In the modelinsect Drosophila, the FGF ligand / receptor combinations of FGF8 (Pyramus and Thisbe) / Heartless (Htl) andBranchless (Bnl) / Breathless (Btl) are required for the migration of mesodermal cells and for the formationof the tracheal network respectively with both the receptors functioning independently of each other.However, only a single fgf-receptor gene (Tc-fgfr) has been identified in the genome of the beetle Tribolium.We therefore asked whether both the ligands Fgf8 and Bnl could transduce their signal through a commonFGF-receptor in Tribolium. Indeed, we found that the function of the single Tc-fgfr gene is essential formesoderm differentiation as well as for the formation of the tracheal network during early development.Ligand specific RNAi for Tc-fgf8 and Tc-bnl resulted in two distinct non-overlapping phenotypes of impairedmesoderm differentiation and abnormal formation of the tracheal network in Tc-fgf8- and Tc-bnlRNAi

embryos respectively. We further show that the single Tc-fgfr gene encodes at least two different receptorisoforms that are generated through alternative splicing. We in addition demonstrate through exon-specificRNAi their distinct tissue-specific functions. Finally, we discuss the structure of the fgf-receptor gene from anevolutionary perspective.

& 2015 Elsevier Inc. All rights reserved.

Introduction

The fibroblast growth factor (FGF) signalling pathway plays arange of essential roles in various stages during animal develop-ment (Böttcher and Niehrs, 2005; Dorey and Amaya, 2010; Muhaand Müller, 2013) and is involved in many human diseasesincluding cancer (Beenken and Mohammadi, 2009; Turner andGrose, 2010). The complexity of this pathway in vertebrates isevident by the high number of extracellular ligands that impedesfunctional studies through possible redundant contributions(Böttcher and Niehrs, 2005; Tulin and Stathopoulos, 2010). Incomparison to 22 FGF ligands and 4 FGF receptors (FGFRs) inhumans, only three genes coding for the FGF ligands branchless(bnl), pyramus (pyr) and thisbe (ths) representing two Fgf sub-families and two FGF receptor (FGFRs) genes breathless (btl) andheartless (htl) have been identified in the Drosophila genome

e (R. Schröder).

(Gryzik and Mu, 2004; Itoh and Ornitz, 2011; Klämbt et al., 1992;Muha and Müller, 2013; Stathopoulos et al., 2004; Sutherlandet al., 1996).

The genome of the red flour beetle Tribolium castaneum con-tains four genes that code for the Tribolium Fgf ligands Tc-fgf1a,Tc-fgf1b, Tc-fgf8 and Tc-branchless and in contrast to Drosophilaonly one bona fide Fgf receptor coding gene (Tc-fgfr) (Beermannand Schröder, 2008; Richards et al., 2008).

The presence of a single FGFR is not unique to Tribolium.Likewise, only one fgfr gene exists also in the genomes of theplacozoan Trichoplax, the nematode C. elegans, the sea urchinStrongylocentrotus purpuratus and in that of the urochordate Ciona(Birnbaum et al., 2005; Itoh and Ornitz, 2011; Rebscher et al.,2009; Tulin and Stathopoulos, 2010). Thus, the multiplication offgfr-genes is not a prerequisite for the evolution of animal com-plexity since in Ciona the sole FGF receptor obviously is able tofunctionally integrate six different FGF-ligands. So far, multi-functionality of the fgf- receptor gene could only be demonstrated

R. Sharma et al. / Developmental Biology 402 (2015) 264–275 265

for C. elegans where two ligand-specific receptor-variants are theresult of alternative splicing (Goodman et al., 2003).

In Tribolium it is unknownwhether the sole Tc-fgf-receptor geneidentified in the genome mediates all FGF-dependent signallingevents. The role of FGF signalling during animal development isparticularly striking in processes that involve cell migration. Incontrast to vertebrates and hemichordates where FGF8 dependentsignalling induces mesoderm formation (Amaya et al., 1993; Greenet al., 2013), the fgf8-like genes pyramus (pyr) and thisbe (ths) inDrosophila function in mesoderm differentiation after its inductionvia Dorsal signalling (Jiang et al., 1991). The FGF ligands Pyr andThs activate the FGF receptor Htl to control the collapse of themesodermal tube, the flattening and the subsequent dorsalmigration of the mesoderm layer during gastrulation (Gryzik andMu, 2004; McMahon et al., 2010; Muha and Müller, 2013; Sta-thopoulos et al., 2004; Wilson et al., 2005).

During further development, various tissues derived from themesoderm such as muscle- and pericardial cells and the caudalvisceral mesoderm require the function of the FGF receptor Htl(Kadam et al., 2012; Kadam et al., 2009; Klingseisen et al., 2009;McMahon et al., 2010; Reim et al., 2012). In addition, the devel-opment of the neuroectoderm-derived glia in Drosophila alsodepends on Htl-mediated FGF signalling (Muha and Müller, 2013).

In Drosophila, the dorsal vessel or the heart is derived frommesodermal cells lined up laterally as two longitudinal bands atthe future dorsal side. The molecular network that regulates heartdevelopment involves different signalling pathways including theFGF-pathway and shares components between flies and verte-brates (Tao and Schulz, 2007). The FGF8-like ligands Pyr and Thsare required for the lateral spreading of the mesoderm to underlaythe dorsal ectoderm. The cardiogenic transcription factor Tinmanitself is induced and maintained by Dpp signals from the dorsalectoderm. Together with other transcription factors, Tinman isrequired during specification from the dorsal mesoderm to cardiacmesoderm and its differentiation into cardiac cells. The cardiactube builds when cardiac myocytes migrate dorsally and meet atthe dorsal midline (Seyres et al., 2012). Since tinman is expressedin cardiac cells from their specification until the differentiatedstate, this gene serves as an ideal marker for following cardiacdevelopment (Bodmer, 1993; Frasch, 1995).

The Branchless / Breathless ligand - receptor combinationregulates tracheal branching in the Drosophila embryo, the larvaand the adult (Ahmad and Baker, 2002; Ohshiro et al., 2002; Satoand Kornberg, 2002; Sutherland et al., 1996). Trachea developmentstarts with the invagination of ectodermal cell clusters, the tra-cheal precursor cells. Once the invaginated epithelial sac that isconnected to the surface has formed, no additional cell divisionsoccur and the subsequent formation of the tracheal systemdepends solely on cell migration and changes in cell shape, - sizeand the dynamic of cell–cell interactions. From the sac, a primarybranch buds off, that then gives rise to secondary branches. Finally,branch fusion leads to long, continuous tubes within the trachealnetwork (Ghabrial et al., 2003). In Drosophila, the development ofthe tracheal system depends primarily on the function of thebHLH-PAS domain transcription factor trachealess (trh) that isexpressed in tracheal cells throughout development (Isaac andAndrew, 1996; Wilk et al., 1996). Its expression starts in theectodermal primordia of the trachea before any morphologicaldifferentiation is visible. Together with another transcription fac-tor (ventral veins lacking (vvl) or drifter), trh can induce the for-mation of tracheal branches also ectopically (Boube et al., 2000).Among the numerous targets of the trh / vvl dimer is the FGFreceptor gene breathless (Anderson et al., 1996; Klämbt et al., 1992;Ohshiro and Saigo, 1997; Wilk et al., 1996) that is transcribed intracheal sac-cells just before the branching starts. Upon activationby the nearby produced Bnl-ligand, the FGF receptor then leads to

migration and the outgrowth of primary branches and their bud-ding into secondary ones (Klämbt et al., 1992; Sutherland et al.,1996).

In Tribolium, the fgf8 ortholog Tc-fgf8 has been discussed to beinvolved in mesoderm differentiation similar to the fgf8 genes pyrand ths in Drosophila. This assumption was based on its sharedexpression domain with the mesoderm marker Tc-twist in thesegments and in the growth zone (Beermann and Schröder, 2008).Furthermore, the expression of Tc-fgfr in close vicinity to Tc-fgf8expression in the dorsal region suggests its involvement in otherkey processes including heart development and in cooperationwith Tc-bnl in trachea formation (Beermann and Schröder, 2008).While the fgf1-like genes, Tc-fgf1a and Tc-fgf1b, in Tribolium haverecently been characterized (Sharma et al., 2013), the functionalroles of both the ligands Tc-fgf8 and Tc-bnl in Tribolium are yet tobe determined. Moreover, it is also not clear, whether a single FGFreceptor represents the common acceptor for both of these ligandsin Tribolium. In Drosophila both the FGF receptors Heartless andBreathless strictly function independent of each other and so far,no combinatorial and overlapping function in any developmentalprocess has been observed (Muha and Müller, 2013). We thereforeasked whether and how in Tribolium a single FGF receptor is ableto function as a multi-specific FGF receptor.

In this study, we show through the analyses of tissue-specificmarker genes that the mesoderm and its derivatives, the heart andthe muscles, are severely affected by Fgf8-downregulation inTribolium. We further show that the differentiation of the trachealsystem but not the mesoderm is impaired through Tc-bnl RNAi andprovide the evidence that both these tissues are severely affectedin Tc-fgfrRNAi embryos. We identified two different Tc-fgf-receptorsplice variants and demonstrate their tissue-specific functionthrough exon-specific RNAi.

Materials and Methods

Template preparation and parental RNAi

To amplify the gene specific product, cDNA template was gen-erated from the total RNA (RNeasy/Qiagen) using random hexamerprimers provided with the Transcriptor First Strand cDNA Synth-esis Kit (Roche). PCR-amplified fragments were then subclonedinto a pCR4 plasmid vector of the TOPO-TA Cloning Kit (Invitrogen)and verified by sequencing (LGC Genomics, Berlin). Double-stranded RNA (dsRNA) was produced from the linearised plasmidDNA using T3 and T7 MEGAscript in-vitro transcription kit(Ambion). DIG-labelled-RNA probes were synthesised with T3/T7RNA polymerase (Roche) along with the DIG RNA Labelling Mix(Roche). For RNA interference (RNAi), dsRNA was injected intofreshly hatched adults (parental RNAi) at several different con-centrations in multiple independent experiments (see Table S1 inthe supplementary material) under standard conditions (Bucheret al., 2002; Schröder, 2003). In situ hybridisation to RNA tran-scripts was performed as published (Tautz and Pfeifle, 1989).Conditions were strictly parallel for each in-situ hybridisationreaction that allowed precise comparative analysis betweenwildtype and the affected RNAi embryos. Some of the markergenes used for expression studies were already published likeTc-dof (Beermann and Schröder, 2008), Tc-twist (Handel et al.,2005) and Tc-tinman (Cande et al., 2009; Janssen and Damen,2008). The Tribolium trachealess ortholog (Tc-trh, TC001448) wascloned using the primer pair: FW: TCAGAC CTCTTAGATTACACGGC;RW: AATGCTTCGACATTGCGTTCTC.

R. Sharma et al. / Developmental Biology 402 (2015) 264–275266

dsRNA fragments used for RNAi

To rule out the possibility of off-target effects, different non-over-lapping / partially overlapping / nested fragments (NOFs) were injectedto confirm the function of targeted genes. For Tc-fgf8 (TC000278) threefragments: Tc-fgf8_fragment1 (658 bp; binding position: CDS 17-674;primers: FW_CAAAATGTGGGGCTAAAGGTA; RV_GTCGAT AAAG-TAAACGTGTGA); Tc-fgf8_fragment2 (420 bp; binding position: CDS175-594; primers: FW_CATGTAACGTCGGTGTGAGAAA; RV_ GTCCGA-CAAGAAAGAACTA TCAC) and Tc-fgf8_fragment3 (605 bp; bindingposition: 257 bp 5′UTRþCDS 1-376; primers: FW_GTCGCTTAT-CCGCTCTCCATGT GC; RV_TCCAAGTGCGAAACCAGCTCCTG) were used.For Tc-fgfr (TC004713) two fragments: Tc-fgfr_fragment1 (556 bp;binding position: CDS 166-719; primers: FW_CAAAATGTGGGGC-TAAAGGTA; RV_GTCGATAAAGTAAAC GTGTGA) and Tc-fgfr_fragment2(846 bp; binding position: CDS 918-1781; primers: FW_AGAGGAGGAGAACACTGTGCCTG; RV_ GGTTCGGAGACATCCTCGTTGGT G) wereused. In addition an iBeetle dsRNA fragment (iB_03821) for Tc-fgfr genewas obtained from a company (Eupheria) and injected with 1 mg/ml(see Table S1). For Tc-branchless gene (TC001760/TC001705) a singlefragment of 401 bp corresponding to positions: CDS 1-264 (TC001760)and 557-658 (TC001705), amplified with the primers: FW_CATCTG-GATCTTTCTGGCGGTGCCGG; RV_CTTCGTACACCCGCGTGCTGA CG wasused (Beermann et al 2008). The Tc-distalless (Tc-dll) clone used in thisstudy as a positive control experiment was obtained from A. Beermann(Beermann et al., 2001).

The following primer pairs were used for amplifying exon-specific DNA fragments: exon7: the primer pair fw7 ACG-TTTCCCGTCAAAACCATATATTA/rv7CTTTAAAACGCTTTTGTTATAATT-AGTGAA resulted in a 187 bp long fragment; exon8: the primerpair fw8 AACGCTATCACAACAAACCTGTTCTTACACATCC, rv8 CAACC-GCACTCTTATTTGGGTATT resulted in a 161 bp long fragment; andexon 7/8: the primer pair fw7 / rv8 resulted in a 432 bp longfragment that includes the 34 bp long intron between exon7 andexon8.

Genomic analysis of the Tc-fgfr gene & PCR-amplification of splice-variants

The analysis of the Tc-fgfr gene-organisation is based on thenew annotation of the Tribolium genome version Tcas4.0 that isaccessible via: http://bioinf.uni-greifswald.de/tcas/genes/annotation/. Fig. 5C shows a screenshot of the Genome-Browser: http://bioinf.uni-greifswald.de/gb2/gbrowse/tcas4/?name¼TC004713.The RNAseq profile has been obtained using GSNAP (Wu and Nacu,2010) and is the result of a community effort: the RNA-Seq datawas supplied by the iBeetle consortium and 9 external con-tributors; RNA was sampled from several developmental stagesfrom eggs to adults.

Using gene specific primers located in exon 5 (Forward primer:AGAGGAGGAGAACACTGTGCCTG) and 11 (reverse primer GGTTCGGA-GACATCCT-CGTTGGTG), PCR-fragments from cDNA were amplified,cloned and sequenced. The sequences, the original sequencing reactionand assignment of exon numbers of two representative clones areshown in Suppl. Mat. 1 and 2.

Automatic annotation of the fgfr-gene structure by NCBI led tothe following gene predictions: XM_ 008192524 & XM_008192525that both contain Exon7 and XM_ 008192526 that contains Exon8(see also Suppl. Mat. Table S1).

Analyses of first instar larvae and embryos

All first instar larvae were embedded in Hoyer's medium andincubated overnight at 60 °C. Analysis of the larva was possiblethrough the autofluorescence of all cuticularized structures(including trachea and gut) under the illumination of UV light.

Embryos were mounted in 100% glycerol and photographed on anAxioImager Z1 Zeiss microscope using AxioCam MRc (for colourimages) and AxioCam MRm (for black and white images) camerasand AxioVision 4.8.1 software. For precise interpretation of resultsstage-matched embryos were compared.

Sequence analysis

The domain structure of the predicted proteins from the variousfgfr genes and the Tribolium splice variants was identified using theSMART tool (http://smart.embl-heidelberg.de/). fgfr genes withinthe arthropod genomes were identified using EnsemblMetazoa(http://metazoa.ensembl.org/index.html) or NCBI.

Animal husbandry

For the current study all functional experiments were per-formed onwildtype “San Bernardino” strain of the beetle Triboliumcastaneum. Beetles were reared on a regular changing diet ofwhole wheat and instant flour supplemented with 5% yeast extractin a 30 °C incubator. Collection of eggs was performed as pre-viously described (Beermann et al., 2004).

Analysis of the FGFR-protein structure

The domain-structure of the predicted FGFR proteins havebeen analysed using the SMART-tool, EMBL-Heidelberg (Letunicet al., 2015).

Results

Mesoderm morphogenesis, hindgut morphology, Malpighian tubulesand heart development are strongly affected in Tc-fgf8RNAi- and inTc-fgfrRNAi embryos

The functional analysis of both the ligands Tc-fgf8 and Tc-bnl aswell as the receptor Tc-fgfr during embryonic development of thebeetle Tribolium were achieved through parental RNAi by injectingrespective dsRNAs at various concentrations into freshly hatchedadult females (Bucher et al., 2002).

The single knockdown of Tc-fgf8, Tc-bnl or Tc-fgfr resulted innon-hatching larvae that did not display significant deviationsfrom the wildtype cuticle pattern in respect to the basic mor-phology of the head, thorax or the abdomen.

To further analyse the development of crucial tissues that couldrequire FGF input during embryogenesis we analysed the expres-sion of tissue-specific marker genes in the respective RNA-inter-ference knockdown embryos.

For the functional analysis of FGF signalling pathway compo-nents on mesoderm development and its derivatives in Triboliumwe compared the expression of the mesoderm marker Tc-twist(Handel et al., 2005) in Tc-bnlRNAi-, Tc-fgf8RNAi- and Tc-fgfrRNAi

embryos with that of wildtype embryos.During blastoderm formation in wildtype embryos, Tc-twist is

expressed in a mid-ventral stripe marking the prospective meso-derm (Fig. 1A) and during germband formation and –elongation insegmental packages along the anterior–posterior axis as well as ina patch of mesodermal cells between the head lobes (arrowhead)and at the posterior end of growth zone (Fig. 1B and C). Duringthese early stages of embryogenesis, no critical deviations from thewildtype Tc-twist pattern were seen in Tc-bnlRNAi (Fig. 1E–G),Tc-fgf8RNAi (Fig. 1I–K) and Tc-fgfrRNAi (Fig. 1M–O) embryos.

In the fully elongated wildtype germband embryo, Tc-twistexpression is seen at the base of the gnathal and thoracic appen-dages, in the lateral mesoderm of the abdominal segments and in

Fig. 1. Analysis of mesoderm development in wildtype and affected embryos. (A–D,Q–S′) wildtype; (E–H,T–V′) Tc-bnlRNAi; (I–L,W–Y′) Tc-fgf8RNAi; (M–P,AA-AC′) Tc-fgfrRNAi

embryos stained for Tc-twist mRNA. All images are DIC images and show ventral views with anterior to the left. In the blastoderm and during germband elongation, Tc-twistexpression is indistinguishable in WT (A–C) or RNAi treated embryos (E–G,I–K,M–O). In fully segmented embryos, Tc-twist expression in lateral positions and in theappendages is similar in WT (D,Q–S) and Tc-bnlRNAi embryos (H,T–V) but drastically reduced in Tc-fgf8RNAi (L,W–Y) and Tc-fgfrRNAi (P,AA-AC) embryos. In these RNAi-affectedembryos, Tc-twist expression is seen at the base of the appendages but missing from the interior (Y′,AC′,red asteriks) which like in the WT (S′) seems unaffected in Tc-bnlRNAi

embryos (V′). T: thoracic segment; A: abdominal segment; hg: hindgut; MT: malpighian tubules; black asterisks mark foregut anlage in older staged embryos.

R. Sharma et al. / Developmental Biology 402 (2015) 264–275 267

the posterior hindgut precursor (Fig. 1D). While upon Tc-bnl RNAino obvious differences to the wildtype pattern were observed(Fig. 1H), Tc-twist expression in those sites was drastically reducedin Tc-fgf8RNAi (Fig. 1L) and in Tc-fgfrRNAi (Fig. 1P) embryos butremained unchanged in the hindgut region.

During germband retraction of the WT embryo, Tc-twistexpression marks a variety of tissues including the precursors of thefore- and the hindgut, the gnathal appendages, growing musclefibres in the legs, the area surrounding the tracheal openings andlaterally, in rows of muscle precursor cells (Fig. 1Q–S). While instage-matched Tc-bnlRNAi embryos (Fig. 1T–V) Tc-twist expressiondomains were indistinguishable from the wildtype pattern, Tc-twistexpression was significantly altered in both, Tc-fgf8RNAi- (Fig. 1W–Y)and Tc-fgfrRNAi embryos (Fig. 1AA-AC). Particularly, the Tc-twistexpression domain that runs along the proximo–distal axis withinthe WT larval legs (Fig. 1S′), was severely reduced (red asterisks) inTc-fgf8RNAi- (Fig. 1Y′) and Tc-fgfrRNAi (Fig. 1AC′) embryos, althoughfew Tc-twist-positive cells can still be recognised at the proximalbase of the appendage. This domain was unchanged in Tc-bnl RNAi

embryos (Fig. V′).In addition to the above-described phenotypes, Malpighian

tubules were abnormally orientated in Tc-fgf8RNAi- (Fig. 1Y) and inTc-fgfrRNAi embryos. In the hindgut assembly of a wildtype embryo,

the Malpighian tubules (MT, black arrowheads) insert into theproctodeum at the midgut–hindgut boundary and adhere closelyto the outer hindgut wall (Fig. 2A). Tc-twist-staining in the hindgutmarks the visceral musculature (white arrows) and is absent fromthe midgut precursor (Fig. 2A). The 2–3 cell-wide sheet of cellsaround the hindgut tube is called the caudal visceral mesoderm(CVM). This tissue is irregularly thin (1-cell-wide) in weaklyaffected embryos or missing (black asterisks) in strongly affectedTc-fgf8RNAi and Tc-fgfrRNAi embryos (Fig. 2B,C).

Still, due to the dense packaging, the single MTs are hardlyvisible in the wildtype embryo. Tissue organisation of the hindgutand the extent of Tc-twist expression in this organ is markedlydisturbed and reduced in both Tc-fg8RNAi or Tc-fgfrRNAi embryoswhere the MTs were shortened and irregularly arranged aroundthe hindgut tube and twist expression is confined to small pos-terior / lateral patches of the hindgut (Fig. 2B and C).

The expression pattern of Tc-tinman (Tc-tin) during Triboliumwildtype embryogenesis has been shown previously (Cande et al.,2009; Janssen and Damen, 2008). In brief: in the fully segmentedwildtype embryo, Tc-tin is strongly expressed as a pair of longitudinaldorsal packages in each segment from the labial to abdominal seg-ment 10 and in two head spots near the antennal segment (Fig. 3A).In older embryos during germband retraction, a second Tc-tin domain

WT RNAiTc-fgf8

RNAiTc-fgfr

WTRNAi

Tc-fgf8RNAi

Tc-fgfr

Tc-twist Tc-twist Tc-twist

MTsCVM

* ** *

**

*CVM

MTs

CVM

MTs*

CVM

CVM

MG

Fig. 2. Hindgut development in wildtype and affected embryos. (A,D) wildtype; (B,B′) Tc-fgf8RNAi; (C,C′) Tc-fgfrRNAi embryos. (A–C) DIC images; (D,B′,C′) Hoechst stainednuclei (fluorescence). Hindgut morphology in embryos of the retracted germband stage. For better visibility, the hindgut structures were flipped by 180° during mounting tobetter demonstrate the morphology of this organ. (A,D) Tc-twist expression marks mesodermal precursor cells around the hindgut (hg) tube including the CVM cells (whitearrows). Malpighian tubules (MTs) strictly adhere to the hindgut-tube (black arrowheads in A) and therefore are hardly visible as single structures. (B,B′,C,C′) In Tc-fgf8- andTc-fgfrRNAi embryos MTs are detached from the tube-surface and can be recognised individually. Some of the MTs might be ruptured from the hindgut during mounting andtherefore are missing. Black asterisks (*) mark the missing CVM. CVM: caudal visceral mesoderm, MG: midgut membrane.

Fig. 3. Analysis of heart development in wildtype and affected embryos. (A–C) wildtype; (D–F) Tc-bnlRNAi; (G–I) Tc-fgf8RNAi; (J–L) Tc-fgfrRNAi embryos stained for Tc-tinmanmRNA. All images are DIC images with anterior to the left. (A) Tc-tinman marks the heart precursor cells in the dorsal region from the labial (Lb) to abdominal segment 10(A10). (B,C) In the retracting WT embryo, from A2–A7 the Tc-tin domains are doubled indicating the region of the future heart while in the more anterior segments (Lb-A1)where the aorta develops the Tc-tin domains remain as single ones. While Tc-tin expression is wildtype-like in Tc-bnlRNAi embryos (D–F), Tc-tin expression in the anteriorsegments Lb-T3 is markedly reduced or absent in fgf8- (G–I) and in fgfrRNAi (J-L) embryos. Also in the abdominal segments, only remnants of Tc-tin expression in patches areseen. (I) Weakly affected embryo. Lr: labrum; Ant: antenna; Md: mandible; Mx: maxilla; T/A: thoracic/abdominal segment; hg: hindgut; MTs: Malpighian tubules.

R. Sharma et al. / Developmental Biology 402 (2015) 264–275268

in the abdominal segments 2–7 has formed in parallel to the firstdomain, while in more anterior or posterior segments, Tc-tinremained expressed in a single domain (Fig. 3B). In principle, thisexpression pattern is stable until germband compaction only that the

cephalic expression domains of Tc-tin have expanded to thin stripesalong the antenna perpendicular to the AP axis (Fig. 3C). While awildtype-like expression pattern for Tc-tin was observed throughoutdevelopment in Tc-bnlRNAi embryos (Fig. 3D–F), significant deviations

R. Sharma et al. / Developmental Biology 402 (2015) 264–275 269

from the wildtype Tc-tin expression pattern were noticed inTc-fgf8RNAi- (Fig. 3G–I) as well as in Tc-fgfrRNAi embryos (Fig. 3J–L). Inthe fully segmented germband stage, Tc-tin expression in the thoracicregion was markedly reduced or absent in both, Tc-fgf8RNAi- andTc-fgfrRNAi embryos. In the abdomen however, the reduction ofexpression strength was visible in Tc-fgf8RNAi- (Fig. 3G) and Tc-fgfrRNAi-(Fig. 3J) embryos when compared to the wildtype although lesssevere in the former. Some Tc-tin positive cells were present butdisordered in older embryos. Again, Tc-fgfrRNAi embryos seemed moreseverely affected than embryos after Tc-fgf8 RNAi and the anterior ofthe embryo including the thoracic segments are more severely dis-turbed than the abdomen. The severity of the RNAi effect can bejudged by the degree of malformation in the hindgut and the Mal-pighian tubules (Fig. 3H,I; K,L). According to that criterion and therelatively normal expression pattern of Tc-tin in the abdomen, theembryo in Fig. 3I clearly represents a weaker phenotype.

Tracheal development is impaired in Tc-bnlRNAi and in Tc-fgfrRNAi

embryos

To visualise the development of the tracheal network in wild-type and in RNAi treated Tribolium embryos, we made use of thetracheal marker gene trachealess that in the fly is expressed in alltrachea cells from their induction to the differentiated stage (Wilket al., 1996). Like in Drosophila, the development of the trachealnetwork starts as pairs of groups of cells in ten segments fromthoracic segment 2 (T2) to abdominal segment eight (A8) asindicated by the expression of the Tribolium trachealess orthologTc-trh (Fig. 4A,E). The groups of Tc-trh positive cells stretch laterallyin the retracted Tribolium germband (Fig. 4I,M). The initiation andmaintenance of Tc-trh expression are not affected during germ-band formation in either Tc-bnlRNAi-, Tc-fgf8RNAi- or in Tc-fgfrRNAi-embryos (Fig. 4B–D,F–H). The subsequent distribution of the Tc-trhexpressing cells within each hemisegment was observed in a

Fig. 4. Tracheal development in wildtype and affected embryos. (A,E,I,M,Q) wildtype; (B,FTc-trh (A–P) and Tc-dof (Q–T) mRNA. (A–T) DIC images with anterior to the left (A,E,I,M) Tround shaped group and during trachea formation stretch and distribute within the anteTc-trh and thus the allocation of the placode are not affected in any of the dsRNA treatembryos (L,P), also illustrated by the expression of Tc-dof (Q–T). Lr: labrum; Ant: antennaMalpighian tubules.

wildtype-like manner in Tc-fgf8RNAi embryos (Fig. 4K,O). However,this crucial step during trachea formation does not take place inTc-bnlRNAi- and in Tc-fgfrRNAi embryos where the expressiondomains of Tc-trh remained unchanged as a round-shaped domain(Figs. 4J,N; 4L,P). The loss of the dynamic behaviour of the trachealprecursor cells was also demonstrated in the respective RNAitreated embryos for the Tc-dof (downstream-of-FGF-receptor)expression pattern (Fig. 4R–T) that among other tissues also labelsthis cell type in the wildtype (Fig. 4Q) (Beermann and Schröder,2008).

Differential splicing results in two FGFR isoforms

The simple modular architecture search tool (SMART/EMBL)predicts that the derived Tribolium FGFR protein contains fiveextracellular Ig-like domains (Fig. 5A). This is one more thanpreviously described where the analysis was based on the BLAST-Palgorithm (Beermann and Schröder, 2008) and two more than thatof a prototypical vertebrate FGFR (Kelleher et al., 2013). With that,the Tribolium-FGFR matches the Drosophila FGF receptor Breath-less with five domains (Klämbt et al., 1992) while the Heartlessprotein contains only two such domains (Muha and Müller, 2013).We then asked whether in Tribolium differential splicing of theTc-fgfr gene could result in different fgfr gene products. Using PCRprimers directed against exon 5 and exon 10 we could amplify twocDNA variants that differ in exon content. We obtained threeindependent clones for the splice variant including exon 7 andmore than 10 clones for the splice-variant including the alternativeexon 8. Both these exons, 7 and 8, each code for the N-terminalpart of IG-domain V of which the second, C-terminal half is con-tributed to by the common exon 9 (Fig. 5A and B). The sequencesof representative clones can be found in Supplementary data.

,J,N,R) Tc-bnlRNAi; (C, G, K, O, S) Tc-fgf8RNAi; (D,H,L,P,T) Tc-fgfrRNAi embryos stained forc-trachealess marks the cells of the tracheal placode that initially are organised as arior of segment T2–A8 perpendicular to the AP axis. While the initial expression ofed embryos, stretching of the placode does not occur in bnl (J,N) and in fgfr-RNAi; Md: mandible; Mx: maxilla; T/A: thoracic/abdominal segment; hg: hindgut; MTs:

R. Sharma et al. / Developmental Biology 402 (2015) 264–275270

Exon-specific RNAi results in distinct, receptor-specific phenotypesthat mirror Tc-fgf8RNAi- and Tc-bnlRNAi- effects

At the cuticle level, the shape of the hindgut as well as thewiring of the tracheal network is visible through the auto-fluorescence of cuticularized structures under UV illumination. Inthe wildtype, the hindgut tube shows a stereotypic double curvedoutline (Fig. 6A). The main trunk of the tracheal system runs in astraight line along the AP axis (Fig. 6D) and links to the segmen-tally formed short primary branches that are connected to thetracheal openings.

The analysis of Tc-fgf8RNAi cuticles revealed abnormal formationof hindgut structures. In those, the hindgut lost its S-shape andformed either condensed within the posterior abdomen (Fig. 6B)

1 2 3 87654

IG I IG IIIIG II IG IV I

Fig. 5. Differential splicing generates two Tribolium fgfr isoforms. (A) Structure of the(encoded by exon 7 or 8) and the constant second half (encoded by exon 9). SequencesTribolium fgfr gene (TC004713). Exon 7 has been added to the original gene model takrepresent the 5′ and the 3′ untranslated region. Bars in dark blue, green and light blabundance of the Tribolium fgfr-gene indicates that exon 7 is less frequently transcribed w(see methods). Screenshot of Tribolium genome database Tcas4: http://bioinf.uni-greifs

* **

hg

hg

100

90

80

70

60

50

40

30

20

10

0 fgfr

n=132fgfr-exon 8

n=51fgf 8n=86

RNAi

%

G

to

to to

to

to

to

A

D

E

F

B

Fig. 6. RNAi cuticle phenotypes of the FGF signalling components compared to exon-spwildtype (A,D); Tc-fgf8RNAi (B,C,E); Tc-bnlRNAi (F). Anterior points to the left; (G) Statisticswith two turns (A) is abnormal in fgf8- and fgfrRNAi embryos where it either can take a contracheal system (arrows in D) forms in curves in an fgf8RNAi background (E; stippled) andbranch normally forms) while the primary tubes (arrowheads) and the tracheal openinsingle knockdown of Tc-fgf8, Tc-fgfr and Tc-bnl as well as after exon-specific fgfr-RNAi. Wadriven artefacts. “n” represents the total number of cuticles analysed specifically for thgenerally analysed (see Supplementary Table S2).

or extends without any bending towards the anterior (straightphenotype, Fig. 6C). In addition, the main tracheal branch waspresent but showed a meandering pattern and abnormal curvingnot seen in the wildtype (Fig. 6E).

While in Tc-bnlRNAi-larvae, the hindgut, tracheal openings andthe ingrowing primary tracheal branches were formed essentiallynormal, the main tracheal trunk was found absent (Fig. 6F).

Larval Tc-fgfrRNAi -cuticles showed a combined tracheal/hindgutphenotype where the main tracheal branch was missing and thehindgut formed was either condensed or straight. In segmentswhere the main trachea was present, it showed the meanderingphenotype.

By injecting exon-specific double stranded RNA we revealedtissue specific function of the two FGF-receptor splice variants.

14131211109

RTKG V

TM

deduced FGFR-protein. Composite Ig-domain V with two alternative first halvesof the splice variants are provided in Suppl. Data. (B) Genomic organisation of theen from the genome assembly 4.0 (ChLGX, position 786,486 to 790,713). grey barsue symbolise the exons: black line indicates the position of the introns. (C) RNAhen compared to the other exons. RNA-seq data come from the iBeetle consortium

wald.de/gb2/gbrowse/tcas4/?name¼TC004713.

trachea WT / hg WT

trachea missing / hg WT

trachea missing / hg cond.

trachea winding / hg straight

trachea winding / hg cond.

trachea winding / hg WT

trachea missing / hg straight

hg

fgfr-exon7/8n=84

fgfr-exon7n=98

bnln=97

controln=129

C

ecific FGFR-RNAi experiments. (A–F) First instar larval cuticles (autofluorescence):of observed RNAi phenotypes. The typical shape of the hindgut tube in the WT larvadensed (cond.) (B) or an elongated shape without turns (C). The main branch of theis absent in bnl- and fgfrRNAi embryos (stars in (F) indicating the position where thisgs (to) are present. (G) The frequency of the observed cuticle phenotypes after theter injections served as a negative control as well as a reference to exclude injectioneir specific gut- and tracheal phenotypes. A larger set of cuticles have been more

R. Sharma et al. / Developmental Biology 402 (2015) 264–275 271

While RNAi for the version containing Tc-fgfr-exon8 led to cuticlephenotypes seen in Tc-fgf8-knock-down experiments, the Tc-fgfr-exon7-knockdown resulted in cuticle phenotypes seen in Tc-bnl-RNAi experiments (summarised in Fig. 6G). Injection of Tc fgfr-dsRNA that includes both exon 7 and 8 specific sequences resultedlike Tc-fgfr-RNAi in a combined phenotype (compare column 3 and4 in Fig. 6G).

Discussion

In this work we provide the first evidences that the Tribolium FGFreceptor, which is encoded by the single Tc-fgfr gene, is able totransmit the signals from both the ligands Tc-FGF8 and Tc-Branchless.This is made possible through different receptor isoforms of Tc-Fgfrgenerated by differential splicing.

Tc-fgf8 and Tc-fgfr in Tribolium are essential for mesoderm differ-entiation at late embryonic stages

The involvement of FGF signalling in mesoderm formation anddevelopment is one of the key roles that has been identified in avariety of both vertebrates and non-vertebrates species. Especially,FGF8-dependent signalling was found instrumental for meso-dermal morphogenesis in metazoans (Birnbaum et al., 2005;Dorey and Amaya, 2010; Fletcher et al., 2006; Gryzik and Mu,2004; Muha and Müller, 2013; Röttinger et al., 2008; Stathopouloset al., 2004; Wilson et al., 2005).

In Tribolium, the expression patterns of the single Fgf-receptorgene Tc-fgfr and the adaptor molecule gene Tc-dof (Downstream-of-Fgfr) have been found in close vicinity to expression sites of theligands Tc-fgf8 and Tc-bnl. On this basis we previously hypothe-sised that in Tribolium both these ligands signal through onereceptor during mesoderm and trachea formation, which is dif-ferent to Drosophila (Beermann and Schröder, 2008).

The expression analysis of the mesoderm marker Tc-twist inTc-fgf8- and Tc-fgfr- RNAi embryos showed a significant reductionor the loss of mesodermal cells from various tissues in affectedmatured embryos when compared to the wildtype (Fig. 1). Thiseffect was more prominently visible at later embryonic stageswhere tissue specification was already apparent which clearlyshows that FGF signalling has a specific role in mesoderm mor-phogenesis and -differentiation after gastrulation. During earlystages of development from the blastoderm to the elongatinggermbands, the initiation and the maintenance of Tc-twistexpression was normal in Tc-fgf8RNAi and Tc-fgfrRNAi embryos. Likein the wildtype (Fig. 1A), a ventral stripe of Tc-twist expressionmarking the mesodermal primordium was also visible in theTc-fgf8RNAi (Fig. 1I) and Tc-fgfrRNAi embryos (Fig. 1M). Likewiseduring the stages of germband extension, the segmental packagesof mesodermal cells marked by Tc-twist expression were indis-tinguishable in WT- (Fig. 1B and C) and Tc-fgf8RNAi embryos (Fig. 1Jand K) although somehow weaker in more anterior segments inTc-fgfrRNAi embryos (Fig. 1N and O).

Thus, FGF signalling in Tribolium has no role in mesoderminduction, but is essential for mesoderm migration in post-gas-trulating embryos as this has been shown previously for theecdysozoan model systems C. elegans and D. melanogaster (DeVoreet al., 1995; Kadam et al., 2009; Lo et al., 2008; Stathopoulos et al.,2004).

In contrast, FGF signalling in vertebrates (Xenopus, zebrafish,mouse), the cephalochordate Amphioxus and in the ascidian Cionahas been found crucial for mesoderm induction suggesting anancestral role of FGFs in deuterostome mesoderm specification(Dorey and Amaya, 2010; Fletcher and Harland, 2008; Green et al.,2013).

The analysis of Tc-twist expression reveals that in fully retractedTc-fgf8- and in Tc-fgfr knockdown embryos twist-expressing cellsare located at the proximal bases of growing legs but were absentfrom their interior (Fig. 1Y′, AC′). This is in strong contrast to stage-matched wildtype embryos where mesodermal cells that at thesame time also express Tc-fgfr (Beermann and Schröder, 2008) arealigned along the proximal-distal axis (Fig. 1S′) (Handel et al.,2005). Obviously, in embryos with compromised FGF8- and FGFR-dependent signalling, the allocation of twist-expressing cells to thebasal position of an appendage still occurs but the subsequentinvading of the leg requires FGF8 signalling and agrees with thegeneral function of FGF signalling in organising directional cellmovements (Muha and Müller, 2013).

The role of FGF signalling for development of the caudal visceralmesoderm (CVM) and the Malpighian tubules

The combined phenotype of disturbed development of thehindgut and the Malpighian tubules observed in Tc-fg8RNAi orTc-fgfrRNAi embryos is striking and sheds light on the role of FGFsignalling during morphogenesis of these organs. In Drosophila,Htl-dependent signalling in a subset of the visceral mesoderm, thecaudal visceral mesoderm (CVM) significantly contributes to theformation of the founder cells of the longitudinal musculature ofthe gut. Htl is expressed in the CVM cells that actively migratealong the trunk visceral mesoderm under the guidance of the fgf8-like genes pyr and ths (Beiman et al., 1996; Kadam et al., 2012;Reim et al., 2012).

In Tribolium, the prominent expression of Tc-fgf8 and Tc-dof inthe hindgut anlagen and in the MTs of wildtype embryos descri-bed previously (Beermann and Schröder, 2008) suggests aninvolvement of the FGF pathway during development of thesetissues. Our results demonstrate that as in Drosophila, FGF sig-nalling also plays a critical role in hindgut morphogenesis in Tri-bolium. More specifically, the CVM cells that normally ensheaththe hindgut tube (Fig. 2D) are reduced in number or missing at theposterior of the hindgut likely because they are not able to spreadalong the hindgut tube. If true than this finding would underlinethe general importance of the FGF signals for the migratory cap-ability of mesodermal derived cells during development (Kadamet al., 2012; McMahon et al., 2010; Muha and Müller, 2013; Reimet al., 2012; Röttinger et al., 2008). The phenotype of a malformedhindgut – either condensed or stretched – seen in the larval cuticlecan be explained by the inability to get into final shape due to thelack of a functional musculature or because the gut tube is dis-continued through the general instability of gut structures.

The development of the Malpighian- or renal tubules inDrosophila requires the input of different signalling pathways(Jung et al., 2005; Liu et al., 1999; Soderberg et al., 2011) but so far,no evidences point to the contribution of FGF signalling to itsmorphogenesis in the fly. This is in contrast to Tribolium, where weshowed previously that Tc-fgf8 is expressed along the MTsexcluding the most distal part while Tc-dof marks the tip-cells ofthe tubes (Beermann and Schröder, 2008).

The abnormal upright shape and position of the MTs at theproctodeum seen in Tc-fg8RNAi or Tc-fgfrRNAi embryos (Fig. 2B,C),show that in Tribolium FGF signalling has an important role in thedevelopment of the MTs, the organ mainly responsible for tightregulation of fluid homoeostasis in insects (Huang and Stern,2005).

Likely, FGF signalling in the tip-cell of a MT controls its motilityand adhesive property to be able to contact and to move andelongate along on the gut outer surface. Without this signal, theMTs are not able to adhere to the gut and show the observed tree-like appearance. It remains open, whether the MTs still are able to

R. Sharma et al. / Developmental Biology 402 (2015) 264–275272

elongate properly since older staged Tc-fg8RNAi or Tc-fgfrRNAi

embryos have not been analysed.The parallel roles of FGF signalling in the regulation of fluid

homoeostasis in C. elegans (Huang and Stern, 2005) and in renaldevelopment in vertebrates (Bates, 2011) is striking and points toan evolutionary early origin of FGF-function in this excretion-organs. Still, it is remarkable that the important process of MTdevelopment in Drosophila manages without the input of FGFsignalling.

Differential requirement of FGF signalling along the AP axis?

Interestingly, anterior segments seem to be more vulnerable forthe reduction of FGF8-dependent FGF-signalling along the AP axisat the level of Tc-twist (Fig. 1) and Tc-tin expression (Fig. 3). Whilein anterior head appendages of Tc-fgf8- and in Tc-fgfr knockdownembryos a complete loss of twist-positive cells was clearly appar-ent (Fig. 1L,P,W,AA), faint expression of Tc-twist was visible in acluster of cells at the proximal bases of more posteriorly posi-tioned gnathal and thoracic appendages (Fig. 1L,P,W,AA). Likewise,Tc-tin-positive cells were missing in the head- and gnathal seg-ments, strongly reduced in the thorax and fairly normal in theabdomen of both, Tc-fgf8- and in Tc-fgfrRNAi knockdown embryos(Figs. 2G–I; 2J–L).

Since the head and the gnathal appendages are derived fromsegments that develop early during embryogenesis, a loss of themesodermal domain in these appendages of affected embryoscould represent a combined result of impaired cell differentiationand some influences of the AP patterning system and the inabilityto maintain the differentiated state of mesodermal cells.

Tracheal development depends on FGF signalling

The expression of the FGF-ligand branchless in the developingtracheal system of the Tribolium embryo (Beermann and Schröder,2008) suggested the involvement of FGF signalling in trachealdevelopment in the beetle as this has been intensively shownalready for Drosophila (Sutherland et al., 1996). Due to the specificexpression of Tc-fgfr near the tracheal openings we speculated arole in trachea formation also for this receptor that has beenthoroughly studied for breathless in Drosophila (Ghabrial et al.,2003; Klämbt et al., 1992; Muha and Müller, 2013). The expressionof the tracheal marker gene Tc-trh in the tracheal placodes hasfully developed in the elongated embryo (Fig. 4A, E). The initialround shaped domain elongates laterally as cells spread in theanterior part of each segment. Spreading occurs normally inTc-fgf8RNAi embryos but does not happen in Tc-bnl- or in Tc-fgfrRNAi

embryos where the tracheal placodes stay as round-shaped groupsof cells. Clearly, the motility of these cells requires the function ofboth, the Bnl ligand and the FGF receptor in Tribolium. At the levelof the cuticle the main longitudinal secondary branch of the tra-cheal network has not formed. However, trachea formation is notcompletely omitted since tracheal openings and the primary tra-cheal tubes still develop, implying that the formation of thesestructures is FGF-independent.

Does the mesoderm serve as a substrate for the hindgut and elon-gating trachea?

The general integrity of the tracheal system in Tc-fgf8RNAi

embryos is not impaired since the secondary, main branch wasseen connected to the tracheal openings via the primary branches.However, the main branch was not following the shortest straightway but developed as a loopy shaped, meandering tube (Fig. 6E).Since in Tc-fgf8RNAi embryos the dorsal part of the mesoderm isstrongly reduced (Fig. 1W–Y) it is likely, that in the wildtype, the

underlying mesoderm could provide guidance cues for themigrating tracheal tubes. Such a scenario already has been shownfor Drosophila, where migration of the dorsal tracheal trunkdepends on a specific cell type within the mesoderm (Wolf andSchuh, 2000) and migration was misguided in embryos with amissing or malformed mesoderm (Franch-Marro and Casanova,2000).

Alternative splicing results in different FGF-receptor isoforms inTribolium

The functions of FGF-receptors are required in a variety of tis-sues and in different developmental contexts. In principle, multi-ple receptor isoforms with different affinities for the high numberof ligands or intracellular adaptor proteins can be generated bygene duplication and through differential splicing. Both thesestrategies have been noted for vertebrates that contain four dif-ferent Fgfr genes in their genomes and at the same time generatedifferent receptor variants from one Fgfr gene by the differentialusage of exons (Burgar et al., 2002; Gong, 2014; Itoh and Ornitz,2011; Zhang et al., 2006). Alternative splicing that results infunctionally exclusive receptor isoforms has been shown for themouse Ig-domain-III (Haugsten et al., 2010) that is crucial forligand specificity and located N-terminal adjacent to the trans-membrane domain. It is composed of two parts where the first iscommon to both receptor isoforms and the second half comesfrom two alternatively used exons (Kelleher et al., 2013). In theTribolium FGFR, it is the first half of the IG domain N-terminal tothe transmembrane domain that is variable while the second halfremains constant (Fig. 5A).

Differently to vertebrates and Tribolium, the two C. elegans FGFRisoforms differ in an insertion between IG-domain 1 and Ig-2 andfunction tissue-specific (Goodman et al., 2003). All these examplesillustrate the variability of gaining functional diversity of a singlefgf-receptor gene during evolution.

Varying number of fgfr genes and their introns in arthropods

The number of fgfr-gene copies within genomes varies sub-stantially (Coulier et al., 1997; Rebscher et al., 2009). Only a single-fgfr-gene is found e.g. in the genomes of Daphnia, Apis and Nasonia,Bombyx, Tribolium and in that of the nematoceran fly Anopheles(Fig. 7). Multiple copies of the fgfr-gene in other genomes could beeither the result of lineage-specific gene duplications like inmyriapods (Chipman et al., 2014), in the nematoceran (Aedes andCulex) and brachyceran flies or of whole genome duplication thatlikely occurred in the spider lineage (Schwager et al., 2007).

The integration of tissue-specific signals from different ligandsrequires several receptor isoforms that can be provided either byduplicated fgfr-genes or by differential splicing of a single-copygene.

In the case of a single fgfr gene present in a genome, variousreceptor isoforms can only be generated through differentialsplicing which requires multiple introns within the coding region.This has recently been demonstrated for the single fgfr-gene ofC. elegans (Goodman et al., 2003) and we show here for Tribolium(Fig. 5B).

In Drosophila, the two fgfr-genes heartless and breathless do notcontain introns within the region coding for the ligand bindingpart of the protein. Thus, none of these genes are able to generatedifferent protein-isoforms. Consequently, Drosophila and otherspecies that contain only intron-less receptor genes in their gen-omes need at least two paralogs to be able to specifically andfaithfully transduce the signal of more than one ligand.

Various evolutionary scenarios can arrive with intron-less fgfr-paralogs. Either, the two Drosophila-fgfr genes are the result of a

Strigamia(Myriapoda)

Daphnia(Crustacea)

Apis(Hymenoptera)

Tribolium(Coleoptera)

Bombyx(Lepidoptera)

Anopheles(Nematocera)

Aedes(Nematocera)

Culex(Nematocera)

Musca(Brachycera)

Ceratitis(Brachycera)

Drosophila(Brachycera)

# FGFR-genes introns # IG domains

2Heartless -

Breathless -

2 --

2--

2 ++

1 +

1 +

2 +

1 +

1 +

1 +

6 + / - -

Diptera

Parasteatoda(Chelicerates) 2 +

Fig. 7. Structural properties of the FGFR genes and -proteins in different arthropods. Within the different arthropod classes, the number of fgfr genes within a genome,the presence or absence of introns in the coding region and the number of IG-like domains involved in ligand binding vary substantially. Stars mark putative gene duplicationevents during evolution. Accession numbers of the respective genes are listed in Supplementary data, Table 2.

R. Sharma et al. / Developmental Biology 402 (2015) 264–275 273

gene duplication event and became intron-free secondarily or oneintron-less gene originated via a retrotransposition event, a gen-eral important mechanism responsible for gene amplification inevolution (Schrider et al., 2011) and duplicated subsequently.

The reduction of the length and number of introns can be seenas an adaptation to extremely short early mitotic cycles in Droso-phila. In this way, developmental disorders as a result of intron-delay are avoided (Artieri and Fraser, 2014; Guilgur et al., 2014;Rothe et al., 1992) and thus can be assured that enough functionalheartless transcripts are present at the beginning of gastrulation. Inthe case of retrotransposition, one retrogene functionally replacedthe intron-containing original gene, duplicated and eventuallyresulted in functionally diversified paralogs. One such example hasrecently been discussed for a vertebrate gene (Macqueen et al.,2014). For the fgfr-genes, such a retrotransposition-event shouldhave occurred in the last common ancestor of the brachyceran fliesand independently in the ancestor of the nematoceran fly Culex(Fig. 6). Since the Drosophila-fgfr genes still contain one intron, theretrotransposition event started with an incompletely splicedgene.

Conclusions

The combined mesoderm / trachea phenotype upon RNAiknockdown of the sole Tribolium FGF receptor gene demonstratesthe multifunctionality of this receptor for both, FGF8 and Bnlinputs. We have shown that the specific functions of the two FGFR-isoforms generated in Tribolium are transmitted by receptor var-iants that originate through differential splicing. As seen throughexon-specific RNAi, the isoform that includes exon 8 functionssimilarly as the Drosophila-Heartless during mesoderm migrationand the isoform that includes exon 7 transmits like Drosophila-Breathless the Branchless signal during trachea formation.

Acknowledgements

We thank Manfred Frasch (Univ. of Erlangen) for sending theTc-tinman clone, Andrei Lupas (MPI Tübingen) for discussions onthe IG-domain numbers, Nico Posnien (Univ. Göttingen) forpointing out the spider genome resources, Lizzy Gerischer and

R. Sharma et al. / Developmental Biology 402 (2015) 264–275274

Mario Stanke (Univ. Greifswald) for sharing bioinformaticsknowledge and the DFG for funding.

Appendix A. Suplementary information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.ydbio.2015.04.001.

References

Ahmad, S.M., Baker, B.S., 2002. Sex-specific deployment of FGF signaling in Dro-sophila recruits mesodermal cells into the male genital imaginal disc. Cell 109,651–661.

Amaya, E., Stein, P.A., Musci, T.J., Kirschner, M.W., 1993. FGF signalling in the earlyspecification of mesoderm in Xenopus. Development 118, 477–487.

Anderson, M.G., Certel, S.J., Certel, K., Lee, T., Montell, D.J., Johnson, W.A., 1996.Function of the Drosophila POU domain transcription factor drifter as anupstream regulator of breathless receptor tyrosine kinase expression in devel-oping trachea. Development 122, 4169–4178.

Artieri, C.G., Fraser, H.B., 2014. Transcript length mediates developmental timing ofgene expression across Drosophila. Mol. Biol. Evol. 31, 2879–2889.

Bates, C.M., 2011. Role of fibroblast growth factor receptor signaling in kidneydevelopment. Am. J. Physiol. Ren. Physiol. 301, F245–F251.

Beenken, A., Mohammadi, M., 2009. The FGF family: biology, pathophysiology andtherapy. Nat. Rev. Drug Discov. 8, 235–253.

Beermann, A., Aranda, M., Schröder, R., 2004. The Sp8 zinc-finger transcriptionfactor is involved in allometric growth of the limbs in the beetle Triboliumcastaneum. Development 131, 733–742.

Beermann, A., Jay, D.G., Beeman, R.W., Hülskamp, M., Tautz, D., Jürgens, G., 2001.The Short antennae gene of Tribolium is required for limb development andencodes the orthologue of the Drosophila Distal-less protein. Development 128,287–297.

Beermann, A., Schröder, R., 2008. Sites of Fgf signalling and perception duringembryogenesis of the beetle Tribolium castaneum. Dev. Genes Evol. 218,153–167.

Beiman, M., Shilo, B.Z., Volk, T., 1996. Heartless, a Drosophila FGF receptor homolog,is essential for cell migration and establishment of several mesodermal linea-ges. Genes Development 10, 2993–3002.

Birnbaum, D., Popovici, C., Roubin, R., 2005. A pair as a minimum: the two fibroblastgrowth factors of the nematode Caenorhabditis elegans. Dev. Dyn. 232, 247–255.

Bodmer, R., 1993. The gene tinman is required for specification of the heart andvisceral muscles in Drosophila. Development 118, 719–729.

Böttcher, R.T., Niehrs, C., 2005. Fibroblast growth factor signaling during earlyvertebrate development. Endocr. Rev. 26, 63–77.

Boube, M., Llimargas, M., Casanova, J., 2000. Cross-regulatory interactions amongtracheal genes support a co-operative model for the induction of tracheal fatesin the Drosophila embryo. Mech. Development 91, 271–278.

Bucher, G., Scholten, J., Klingler, M., 2002. Parental RNAi in Tribolium (Coleoptera).Curr. Biol. 12, R85–R86.

Burgar, H.R., Burns, H.D., Elsden, J.L., Lalioti, M.D., Heath, J.K., 2002. Association ofthe signaling adaptor FRS2 with fibroblast growth factor receptor 1 (Fgfr1) ismediated by alternative splicing of the juxtamembrane domain. J. Biol. Chem.277, 4018–4023.

Cande, J.D., Chopra, V.S., Levine, M., 2009. Evolving enhancer-promoter interactionswithin the tinman complex of the flour beetle, Tribolium castaneum. Develop-ment 136, 3153–3160.

Chipman, A.D., Ferrier, D.E., Brena, C., Qu, J., Hughes, D.S., Schröder, R., Torres-Oliva, M.,Znassi, N., Jiang, H., Almeida, F.C., Alonso, C.R., Apostolou, Z., Aqrawi, P., Arthur, W.,Barna, J.C., Blankenburg, K.P., Brites, D., Capella-Gutierrez, S., Coyle, M., Dearden, P.K.,Du Pasquier, L., Duncan, E.J., Ebert, D., Eibner, C., Erikson, G., Evans, P.D.,Extavour, C.G., Francisco, L., Gabaldon, T., Gillis, W.J., Goodwin-Horn, E.A., Green, J.E.,Griffiths-Jones, S., Grimmelikhuijzen, C.J., Gubbala, S., Guigo, R., Han, Y., Hauser, F.,Havlak, P., Hayden, L., Helbing, S., Holder, M., Hui, J.H., Hunn, J.P., Hunnekuhl, V.S.,Jackson, L., Javaid, M., Jhangiani, S.N., Jiggins, F.M., Jones, T.E., Kaiser, T.S., Kalra, D.,Kenny, N.J., Korchina, V., Kovar, C.L., Kraus, F.B., Lapraz, F., Lee, S.L., Lv, J.,Mandapat, C., Manning, G., Mariotti, M., Mata, R., Mathew, T., Neumann, T., News-ham, I., Ngo, D.N., Ninova, M., Okwuonu, G., Ongeri, F., Palmer, W.J., Patil, S.,Patraquim, P., Pham, C., Pu, L.L., Putman, N.H., Rabouille, C., Ramos, O.M.,Rhodes, A.C., Robertson, H.E., Robertson, H.M., Ronshaugen, M., Rozas, J., Saada, N.,Sanchez-Gracia, A., Scherer, S.E., Schurko, A.M., Siggens, K.W., Simmons, D., Stief, A.,Stolle, E., Telford, M.J., Tessmar-Raible, K., Thornton, R., van der Zee, M., von Hae-seler, A., Williams, J.M., Willis, J.H., Wu, Y., Zou, X., Lawson, D., Muzny, D.M.,Worley, K.C., Gibbs, R.A., Akam, M., Richards, S., 2014. The first myriapod genomesequence reveals conservative arthropod gene content and genome organisation inthe centipede Strigamia maritima. PLoS Biol. 12, e1002005.

Coulier, F., Pontarotti, P., Roubin, R., Hartung, H., Goldfarb, M., Birnbaum, D., 1997.Of worms and men: an evolutionary perspective on the fibroblast growth factor(FGF) and FGF receptor families. J. Mol. Evol. 44, 43–56.

DeVore, D.L., Horvitz, H.R., Stern, M.J., 1995. An FGF receptor signaling pathway isrequired for the normal cell migrations of the sex myoblasts in C. eleganshermaphrodites. Cell 83, 611–620.

Dorey, K., Amaya, E., 2010. FGF signalling: diverse roles during early vertebrateembryogenesis. Development 137, 3731–3742.

Fletcher, R.B., Baker, J.C., Harland, R.M., 2006. FGF8 spliceforms mediate earlymesoderm and posterior neural tissue formation in Xenopus. Development 133,1703–1714.

Fletcher, R.B., Harland, R.M., 2008. The role of FGF signaling in the establishmentand maintenance of mesodermal gene expression in Xenopus. Dev. Dyn. 237,1243–1254.

Franch-Marro, X., Casanova, J., 2000. The alternative migratory pathways of theDrosophila tracheal cells are associated with distinct subsets of mesodermalcells. Dev. Biol. 227, 80–90.

Frasch, M., 1995. Induction of visceral and cardiac mesoderm by ectodermal Dpp inthe early Drosophila embryo. Nature 374, 464–467.

Ghabrial, A., Luschnig, S., Metzstein, M.M., Krasnow, M.A., 2003. Branching mor-phogenesis of the Drosophila tracheal system. Annu. Rev. Cell Dev. Biol. 19,623–647.

Gong, S.G., 2014. Isoforms of receptors of fibroblast growth factors. J. Cell. Physiol.Goodman, S.J., Branda, C.S., Robinson, M.K., Burdine, R.D., Stern, M.J., 2003. Alter-

native splicing affecting a novel domain in the C. elegans EGL-15 FGF receptorconfers functional specificity. Development 130, 3757–3766.

Green, Sa, Norris, R.P., Terasaki, M., Lowe, C.J., 2013. FGF signaling induces meso-derm in the hemichordate Saccoglossus kowalevskii. Development 140,1024–1033.

Gryzik, T., Müller, H.A., 2004. FGF8-like1 and FGF8-like2 encode putative ligands ofthe FGF receptor Htl and are required for mesoderm migration in the Drosophilagastrula. Curr. Biol. 14, 659–667.

Guilgur, L.G., Prudencio, P., Sobral, D., Liszekova, D., Rosa, A., Martinho, R.G., 2014.Requirement for highly efficient pre-mRNA splicing during Drosophila earlyembryonic development. eLife 3, e02181.

Handel, K., Basal, A., Fan, X., Roth, S., 2005. Tribolium castaneum twist: gastrulationand mesoderm formation in a short-germ beetle. Dev. Genes Evol. 215, 13–31.

Haugsten, E.M., Wiedlocha, A., Olsnes, S., Wesche, J., 2010. Roles of fibroblastgrowth factor receptors in carcinogenesis. Mol. Cancer Res.: MCR 8, 1439–1452.

Huang, P., Stern, M.J., 2005. FGF signaling in flies and worms: more and morerelevant to vertebrate biology. Cytokine Growth Factor Rev. 16, 151–158.

Isaac, D.D., Andrew, D.J., 1996. Tubulogenesis in Drosophila: a requirement for thetrachealess gene product. Genes Dev. 10, 103–117.

Itoh, N., Ornitz, D.M., 2011. Fibroblast growth factors: from molecular evolution toroles in development, metabolism and disease. J. Biochem. 149, 121–130.

Janssen, R., Damen, W.G., 2008. Diverged and conserved aspects of heart formationin a spider. Evol. Dev. 10, 155–165.

Jiang, J., Kosman, D., Ip, Y.T., Levine, M., 1991. The dorsal morphogen gradientregulates the mesoderm determinant twist in early Drosophila embryos. GenesDev. 5, 1881–1891.

Jung, A.C., Denholm, B., Skaer, H., Affolter, M., 2005. Renal tubule development inDrosophila: a closer look at the cellular level. J. Am. Soc. Nephrol.: JASN 16,322–328.

Kadam, S., Ghosh, S., Stathopoulos, A., 2012. Synchronous and symmetric migrationof Drosophila caudal visceral mesoderm cells requires dual input by two FGFligands. Development 139, 699–708.

Kadam, S., McMahon, A., Tzou, P., Stathopoulos, A., 2009. FGF ligands in Drosophilahave distinct activities required to support cell migration and differentiation.Development 136, 739–747.

Kelleher, F.C., O'Sullivan, H., Smyth, E., McDermott, R., Viterbo, A., 2013. Fibroblastgrowth factor receptors, developmental corruption and malignant disease.Carcinogenesis 34, 2198–2205.

Klämbt, C., Glazer, L., Shilo, B.Z., 1992. Breathless, a Drosophila FGF receptorhomolog, is essential for migration of tracheal and specific midline glial cells.Genes Dev. 6, 1668–1678.

Klingseisen, A., Clark, I.B.N., Gryzik, T., Müller, H.-A.J., 2009. Differential and over-lapping functions of two closely related FGF8-like growth factors in mesodermdevelopment. Development 136, 2393–2402.

Letunic, I., Doerks, T., Bork, P., 2015. SMART: recent updates, new developments andstatus in 2015. Nucleic Acids Res. 43, D257–D260.

Liu, X., Kiss, I., Lengyel, J.A., 1999. Identification of genes controlling malpighiantubule and other epithelial morphogenesis in Drosophila melanogaster. Genetics151, 685–695.

Lo, T.W., Branda, C.S., Huang, P., Sasson, I.E., Goodman, S.J., Stern, M.J., 2008. Dif-ferent isoforms of the C. elegans FGF receptor are required for attraction andrepulsion of the migrating sex myoblasts. Dev. Biol. 318, 268–275.

Macqueen, D.J., Fuentes, E.N., Valdes, J.A., Molina, A., Martin, S.A., 2014. The ver-tebrate muscle-specific RING finger protein family includes MuRF4 – a novel,conserved E3-ubiquitin ligase. FEBS Lett. 588, 4390–4397.

McMahon, A., Reeves, G.T., Supatto, W., Stathopoulos, A., 2010. Mesoderm migra-tion in Drosophila is a multi-step process requiring FGF signaling and integrinactivity. Development 137, 2167–2175.

Muha, V., Müller, H.-A.J., 2013. Functions and mechanisms of fibroblast growthfactor (FGF) signalling in Drosophila melanogaster. Int. J. Mol. Sci. 14,5920–5937.

Ohshiro, T., Emori, Y., Saigo, K., 2002. Ligand-dependent activation of breathless FGFreceptor gene in Drosophila developing trachea. Mech. Dev. 114, 3–11.

R. Sharma et al. / Developmental Biology 402 (2015) 264–275 275

Ohshiro, T., Saigo, K., 1997. Transcriptional regulation of breathless FGF receptorgene by binding of TRACHEALESS/dARNT heterodimers to three central midlineelements in Drosophila developing trachea. Development 124, 3975–3986.

Rebscher, N., Deichmann, C., Sudhop, S., Fritzenwanker, J.H., Green, S., Hassel, M.,2009. Conserved intron positions in FGFR genes reflect the modular structure ofFGFR and reveal stepwise addition of domains to an already complex ancestralFGFR. Dev. Genes Evol. 219, 455–468.

Reim, I., Hollfelder, D., Ismat, A., Frasch, M., 2012. The FGF8-related signals Pyramusand Thisbe promote pathfinding, substrate adhesion, and survival of migratinglongitudinal gut muscle founder cells. Dev. Biol. 368, 28–43.

Richards, S., Gibbs, R.A., Weinstock, G.M., Brown, S.J., Denell, R., Beeman, R.W.,Gibbs, R., Bucher, G., Friedrich, M., Grimmelikhuijzen, C.J.P., Klingler, M., Lor-enzen, M., Roth, S., Schröder, R., Tautz, D., Zdobnov, E.M., Muzny, D., Attaway, T.,Bell, S., Buhay, C.J., Chandrabose, M.N., Chavez, D., Clerk-Blankenburg, K.P.,Cree, A., Dao, M., Davis, C., Chacko, J., Dinh, H., Dugan-Rocha, S., Fowler, G.,Garner, T.T., Garnes, J., Gnirke, A., Hawes, A., Hernandez, J., Hines, S., Holder, M.,Hume, J., Jhangiani, S.N., Joshi, V., Khan, Z.M., Jackson, L., Kovar, C., Kowis, A.,Lee, S., Lewis, L.R., Margolis, J., Morgan, M., Nazareth, L.V., Nguyen, N.,Okwuonu, G., Parker, D., Ruiz, S.-J., Santibanez, J., Savard, J., Scherer, S.E.,Schneider, B., Sodergren, E., Vattahil, S., Villasana, D., White, C.S., Wright, R.,Park, Y., Lord, J., Oppert, B., Brown, S., Wang, L., Weinstock, G., Liu, Y., Worley, K.,Elsik, C.G., Reese, J.T., Elhaik, E., Landan, G., Graur, D., Arensburger, P.,Atkinson, P., Beidler, J., Demuth, J.P., Drury, D.W., Du, Y.-Z., Fujiwara, H.,Maselli, V., Osanai, M., Robertson, H.M., Tu, Z., Wang, J.-j, Wang, S., Song, H.,Zhang, L., Werner, D., Stanke, M., Morgenstern, B., Solovyev, V., Kosarev, P.,Brown, G., Chen, H.-C., Ermolaeva, O., Hlavina, W., Kapustin, Y., Kiryutin, B.,Kitts, P., Maglott, D., Pruitt, K., Sapojnikov, V., Souvorov, A., Mackey, A.J.,Waterhouse, R.M., Wyder, S., Kriventseva, E.V., Kadowaki, T., Bork, P.,Aranda, M., Bao, R., Beermann, A., Berns, N., Bolognesi, R., Bonneton, F., Bopp, D.,Butts, T., Chaumot, A., Denell, R.E., Ferrier, D.E.K., Gordon, C.M., Jindra, M.,Lan, Q., Lattorff, H.M.G., Laudet, V., von Levetsow, C., Liu, Z., Lutz, R., Lynch, Ja, daFonseca, R.N., Posnien, N., Reuter, R., Schinko, J.B., Schmitt, C., Schoppmeier, M.,Shippy, T.D., Simonnet, F., Marques-Souza, H., Tomoyasu, Y., Trauner, J.,Van der Zee, M., Vervoort, M., Wittkopp, N., Wimmer, Ea, Yang, X., Jones, A.K.,Sattelle, D.B., Ebert, P.R., Nelson, D., Scott, J.G., Muthukrishnan, S., Kramer, K.J.,Arakane, Y., Zhu, Q., Hogenkamp, D., Dixit, R., Jiang, H., Zou, Z., Marshall, J.,Elpidina, E., Vinokurov, K., Oppert, C., Evans, J., Lu, Z., Zhao, P., Sumathipala, N.,Altincicek, B., Vilcinskas, A., Williams, M., Hultmark, D., Hetru, C., Hauser, F.,Cazzamali, G., Williamson, M., Li, B., Tanaka, Y., Predel, R., Neupert, S.,Schachtner, J., Verleyen, P., Raible, F., Walden, K.K.O., Angeli, S., Forêt, S.,Schuetz, S., Maleszka, R., Miller, S.C., Grossmann, D., 2008. The genome of themodel beetle and pest Tribolium castaneum. Nature 452, 949–955.

Rothe, M., Pehl, M., Taubert, H., Jäckle, H., 1992. Loss of gene function through rapidmitotic cycles in the Drosophila embryo. Nature 359, 156–159.

Röttinger, E., Saudemont, A., Duboc, V., Besnardeau, L., McClay, D., Lepage, T., 2008.FGF signals guide migration of mesenchymal cells, control skeletal morpho-genesis and regulate gastrulation during sea urchin development. Development135, 353–365.

Sato, M., Kornberg, T.B., 2002. FGF is an essential mitogen and chemoattractant forthe air sacs of the Drosophila tracheal system. Dev. Cell 3, 195–207.

Schrider, D.R., Stevens, K., Cardeno, C.M., Langley, C.H., Hahn, M.W., 2011. Genome-wide analysis of retrogene polymorphisms in Drosophila melanogaster. GenomeRes. 21, 2087–2095.

Schröder, R., 2003. The genes orthodenticle and hunchback substitute for bicoid inthe beetle Tribolium. Nature 422, 621–625.

Schwager, E.E., Schoppmeier, M., Pechmann, M., Damen, W.G., 2007. DuplicatedHox genes in the spider Cupiennius salei. Front. Zool. 4, 10.

Seyres, D., Roder, L., Perrin, L., 2012. Genes and networks regulating cardiacdevelopment and function in flies: genetic and functional genomic approaches.Brief. Funct. Genomics 11, 366–374.

Sharma, R., Beermann, A., Schröder, R., 2013. FGF signalling controls anteriorextraembryonic and embryonic fate in the beetle Tribolium. Dev. Biol. 381,121–133.

Soderberg, J.A., Birse, R.T., Nassel, D.R., 2011. Insulin production and signaling inrenal tubules of Drosophila is under control of tachykinin-related peptide andregulates stress resistance. PloS One 6, e19866.

Stathopoulos, A., Tam, B., Ronshaugen, M., Frasch, M., Levine, M., 2004. Drosophilaembryos pyramus and thisbe : FGF genes that pattern the mesoderm of Dro-sophila embryos. Genes Dev. (18)6, 687–699.

Sutherland, D., Samakovlis, C., Krasnow, Ma, 1996. Branchless encodes a DrosophilaFGF homolog that controls tracheal cell migration and the pattern of branching.Cell 87, 1091–1101.

Tao, Y., Schulz, R.A., 2007. Heart development in Drosophila. Semin. Cell Dev. Biol.18, 3–15.

Tautz, D., Pfeifle, C., 1989. A non-radioactive in situ hybridization method for thelocalization of specific RNAs in Drosophila embryos reveals translational controlof the segmentation gene hunchback. Chromosoma 98, 81–85.

Tulin, S., Stathopoulos, A., 2010. Extending the family table: insights from beyondvertebrates into the regulation of embryonic development by FGFs. BirthDefects Res. Part C : Embryo Today : Rev. 90, 214–227.

Turner, N., Grose, R., 2010. Fibroblast growth factor signalling: from development tocancer. Nature reviews. Cancer 10, 116–129.

Wilk, R., Weizman, I., Shilo, B.Z., 1996. trachealess encodes a bHLH-PAS protein thatis an inducer of tracheal cell fates in Drosophila. Genes Dev. 10, 93–102.

Wilson, R., Vogelsang, E., Leptin, M., 2005. FGF signalling and the mechanism ofmesoderm spreading in Drosophila embryos. Development 132, 491–501.

Wolf, C., Schuh, R., 2000. Single mesodermal cells guide outgrowth of ectodermaltubular structures in Drosophila. Genes Dev. 14, 2140–2145.

Wu, T.D., Nacu, S., 2010. Fast and SNP-tolerant detection of complex variants andsplicing in short reads. Bioinformatics 26, 873–881.

Zhang, X., Ibrahimi, O.A., Olsen, S.K., Umemori, H., Mohammadi, M., Ornitz, D.M.,2006. Receptor specificity of the fibroblast growth factor family. The completemammalian FGF family. J. Biol. Chem. 281, 15694–15700.