8/11/2019 Science 2014 Roy

1/12

DOI: 10.1126/science.1244624, (2014);343Scienceet al.Sougata Roy

Decapentaplegic Signaling ProteinDrosophilCytoneme-Mediated Contact-Dependent Transport of the

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by, you can order high-quality copies for yourIf you wish to distribute this article to others

here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

):August 1, 2014www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/343/6173/1244624.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2014/01/02/science.1244624.DC1.htmlcan be found at:Supporting Online Material

http://www.sciencemag.org/content/343/6173/1244624.full.html#relatedfound at:

can berelated to this articleA list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/343/6173/1244624.full.html#ref-list-1

, 16 of which can be accessed free:cites 52 articlesThis article

http://www.sciencemag.org/content/343/6173/1244624.full.html#related-urls2 articles hosted by HighWire Press; see:cited byThis article has been

http://www.sciencemag.org/cgi/collection/cell_biolCell Biology

subject collections:This article appears in the following

registered trademark of AAAS.is aScience2014 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrighAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by thScience

http://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/content/343/6173/1244624.full.htmlhttp://www.sciencemag.org/content/343/6173/1244624.full.htmlhttp://www.sciencemag.org/content/343/6173/1244624.full.htmlhttp://www.sciencemag.org/content/suppl/2014/01/02/science.1244624.DC1.htmlhttp://www.sciencemag.org/content/suppl/2014/01/02/science.1244624.DC1.htmlhttp://www.sciencemag.org/content/suppl/2014/01/02/science.1244624.DC1.htmlhttp://www.sciencemag.org/content/343/6173/1244624.full.html#relatedhttp://www.sciencemag.org/content/343/6173/1244624.full.html#relatedhttp://www.sciencemag.org/content/343/6173/1244624.full.html#ref-list-1http://www.sciencemag.org/content/343/6173/1244624.full.html#ref-list-1http://www.sciencemag.org/content/343/6173/1244624.full.html#ref-list-1http://www.sciencemag.org/content/343/6173/1244624.full.html#ref-list-1http://www.sciencemag.org/content/343/6173/1244624.full.html#related-urlshttp://www.sciencemag.org/content/343/6173/1244624.full.html#related-urlshttp://www.sciencemag.org/content/343/6173/1244624.full.html#related-urlshttp://www.sciencemag.org/content/343/6173/1244624.full.html#related-urlshttp://www.sciencemag.org/cgi/collection/cell_biolhttp://www.sciencemag.org/cgi/collection/cell_biolhttp://www.sciencemag.org/cgi/collection/cell_biolhttp://www.sciencemag.org/content/343/6173/1244624.full.html#related-urlshttp://www.sciencemag.org/content/343/6173/1244624.full.html#ref-list-1http://www.sciencemag.org/content/343/6173/1244624.full.html#relatedhttp://www.sciencemag.org/content/suppl/2014/01/02/science.1244624.DC1.htmlhttp://www.sciencemag.org/content/343/6173/1244624.full.htmlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://oascentral.sciencemag.org/RealMedia/ads/click_lx.ads/sciencemag/cgi/reprint/L22/530061541/Top1/AAAS/PDF-R-and-D-Systems-Science-1709891/SfN2014_TG_ScienceBanner.raw/1?x

8/11/2019 Science 2014 Roy

2/12

Cytoneme-Mediated Contact-Dependent Transport of the DrosophilaDecapentaplegic Signaling ProteinSougata Roy, Hai Huang, Songmei Liu, Thomas B. Kornberg*

Introduction:In multicellular organisms, morphogen signaling proteins move from signal-ing centers where they are produced to target cells whose growth and patterning they regulate.Whereas much progress has been made identifying and characterizing signaling proteins such asthe transforming growth factorfamily member Decapentaplegic (Dpp), which is produced in theDrosophilawing imaginal disc, the mechanisms that disperse signaling proteins remain controver-sial. We characterized Dpp signaling in a system in which cytonemes, a specialized type of filopodiaimplicated in long-distance signaling, could be imaged, and in which movement of signaling pro-teins and their receptors could be followed.

Methods:We expressed fluorescence-tagged forms of proteins that function in morphogen signal-ing to monitor Dpp in signal-producing cells, its receptor in signal-receiving cells, and proteins andcell structures that participate in trafficking of signaling proteins. Signaling was characterized in

live, unfixed tissue as well as by immunohistochemistry, and under conditions of both gain- andloss-of-function genetics.

Results:Cells that received Dpp and activated Dpp signal transduction extended cytonemes thatdirectly contacted Dpp-producing cells. The contacts were characterized by relative stability and mem-brane juxtaposition of less than 15 nm. Cytonemes that contained the Dpp receptor in motile punctaalso contained Dpp taken up from Dpp-producing cells. In contrast, a different set of cytonemes thatcontacted fibroblast growth factor (FGF)producing cells contained the FGF receptor but did not takeup Dpp. The cytonemes were reduced in number and length in genetic loss-of-function conditions fordiaphanous, which encodes a formin; for neuro-glian, which encodes an L1-type cell adhesionmolecule; and for shibire, which encodes a dyna-min. Cytonemes were present in loss-of-functionconditions for capricious, which encodes a leu-cine-rich repeat cell adhesion protein, but these

cytonemes failed to contact Dpp-producing cells.Signaling was abrogated in all these conditionsthat created defective cytonemes, although thesignal-producing cells were not compromised. Themutant conditions were not lethal to the affectedcells, and the mutant cells retained competence toautocrine signaling.

Discussion: This work describes cytonemesthat receive and transport signaling proteinsfrom producing cells to target cells, and showsthat cytoneme-mediated signal exchange isboth contact-dependent and essential for Dppsignaling and normal development. Contact-

mediated signal exchange and signaling arealso the hallmarks of neuronsan analogy thatextends to the functional requirements for thediaphanous, neuroglian,shibire, and capriciousgenes by both neurons and epithelial cells. Dis-coveries of cytonemes in many cell types and inmany organisms suggest that contact-mediatedsignaling may be a general mechanism that isnot unique to neurons.

FIGURES IN THE FULL ARTICLE

Fig. 1. Dpp produced in the wing disc signa

to disc-associated tracheal cells.

Fig. 2. The ASP takes up Dpp, and ASP

cytonemes contain activated Tkv receptor.

Fig. 3. Tkv-containing cytonemes transport

Dpp.

Fig. 4. Tracheal cytonemes contact Dpp-

and FGF-expressing disc cells.

Fig. 5. ASP cytonemes require diaand Shi

and nrg.

Fig. 6. ASP cytonemes require Caps.

SUPPLEMENTARY MATERIALS

Figs. S1 to S3Tables S1 to S5Movie S1

Cytonemes take up and transport morphogens. Micrograph showing a tracheal branch marked with mCheroverlying a wing disc expressing Dpp (tagged with green fluorescence). Cytonemes extend from the mediaregion of the branch to Dpp-expressing disc cells, and from the tip of the branch toward FGF-expressing dicells. Dpp has been taken up and transported by the cytonemes that contact Dpp-expressing cells.

READ THE FULL ARTICLE ONLINE

http://dx.doi.org/10.1126/science.1244624

Cite this article as S. Roy et al.,Science343, 1244624 (2014).DOI: 10.1126/science.1244624

Cardiovascular Research Institute, University of California, San Francisco, CA 94158, USA.*Corresponding author. E-mail: tkornberg@ucsf.edu

RESEARCH ARTICLE SUMMARY

21 FEBRUARY 2014 VOL 343 SCIENCE www.sciencemag.org52Published by AAAS

8/11/2019 Science 2014 Roy

3/12

Cytoneme-Mediated Contact-DependentTransport of the DrosophilaDecapentaplegic Signaling ProteinSougata Roy, Hai Huang, Songmei Liu, Thomas B. Kornberg*

Decapentaplegic (Dpp), a Drosophila morphogen signaling protein, transfers directly at synapsesmade at sites of contact between cells that produce Dpp and cytonemes that extend fromrecipient cells. The Dpp that cytonemes receive moves together with activated receptors towardthe recipient cell body in motile puncta. Genetic loss-of-function conditions for diaphanous,shibire, neuroglian, and capricious perturbed cytonemes by reducing their number or only thesynapses they make with cells they target, and reduced cytoneme-mediated transport of Dppand Dpp signaling. These experiments provide direct evidence that cells use cytonemes to exchangesignaling proteins, that cytoneme-based exchange is essential for signaling and normaldevelopment, and that morphogen distribution and signaling can be contact-dependent, requiringcytoneme synapses.

In many contexts during development, cell fateis determined by morphogen signaling pro-teins. TheDrosophilawing imaginal disc, for

instance, expresses the morphogen Decapentaplegic(Dpp), a transforming growth factorbfamily mem-ber that regulates the fate, proliferation, and pat-terning of its cells [reviewed in (1, 2)]. The discexpresses Dpp in a stripe of cells alongside theanterior/posterior (A/P) compartment border, andDpp disperses across the disc to form exponentialconcentration gradients to either side that regulatetarget genes in adjacent cells in a concentration-dependent manner. Whereas the dispersion of Dppacross the disc and the functional importance of

its concentration gradients are well established,the mechanism that moves Dpp from producingto target cells is not.

We tested the model that morphogens aretransported along specialized signaling filopodia(cytonemes) that receive protein released at siteswhere producing and receiving cells contact eachother (3). Cytonemes are on both the apical and thebasal surfaces of wing disc cells. Apical cytonemesthat orient toward Dpp-producing disc cells containthe Dpp receptor Thickveins (Tkv), and cytonemeshape, orientation, and distribution depend onthe expression of Dpp (35). There are basalcytonemes that contain Hedgehog (Hh) and theInterference Hedgehog (Ihog) proteins (6, 7).Hh is also present in short cytonemes that extendfrom Hh-producing cells in the female germlinestem cell niche (8). These correlations are sug-gestive, but they do not establish that cytonemesmediate transfers of signaling proteins from pro-ducing to target cells or that such transfers, ifthey occur, are required for signaling.

The wing disc has associated trachea whosedevelopment depends in part on signaling fromthe disc (9). Larval trachea form an intercon-nected network of oxygen-carrying tubes; one,the transverse connective (TC) of Tr2 is boundto the wing disc (Fig. 1A). During the third larvalinstar (L3), Branchless [the fly fibroblast growthfactor (FGF)] produced by a group of disc cellsinduces a new branch, the air sac primordium(ASP), to grow from the TC (9). The ASP is jux-taposed to the basal surface of the wing disccolumnar epithelium; it is a monolayered epi-thelial tube. At the late L3 stage, the ASP hasmany cytonemes that extend toward the disc (Fig.

1B). Cells at the ASP tip extend long (

30mm)cytonemes that contain the FGF receptor (FGFR)Breathless and appear to touch FGF-producingdisc cells. The presence and orientation of thesecytonemes are dependent on FGF (5,9). The lateL3 ASP also has shorter cytonemes that containTkv and that extend from its lateral flank towardDpp-expressing disc cells (5).

In the wing disc, Dpp induces several changesin responding cells: induction ofDaughters againstDpp(Dad) expression (10), increased phospho-rylation of the Mothers against dpp protein (pMad)(11), and decreasedtkvexpression (11). Dpp signaltransduction does not change expression of theother Dpp receptor subunit Punt (Put). ElevatedDad expression, increased pMad expression,and decreasedtkvexpression were observed inthe ASP, presumably due to Dpp signaling, andtheir abundance indicates that Dpp signal trans-duction is probably higher in the lower layercells that face the disc epithelium than in thecells that are further away in the upper layer(Fig. 1, C and D; fig. S1, A to D; and table S1).Put expression was uniform. Dpp expressionwas not detected in the TC or ASP (Fig. 1A andfig. S1E). These results show that Dpp signaltransduction in the ASP inversely correlates

with distance from Dpp-expressing cells in wing disc.

Overexpressing dominant negative formsTkv or Put, or Dad (which negatively regulaDpp signaling), in the trachea generated abnmally shaped ASPs and reduced Dpp signaliin the ASP (Fig. 1E; fig. S1, F to H; and tabS2 and S3). Expression ofdppRNAiin the widisc generated similar phenotypes and reduc

Dpp signaling (Fig. 1F and table S3), indicatthat the wing disc is the source of the Dpp thactivates signal transduction in the ASP, aestablishing that Dpp signaling from the discessential for normal ASP development.

ASP Cytonemes Receive Dpp from

the Wing Disc

To investigate the basis for disc-dependent Dsignaling in the ASP, we overexpressed an iform of Dpp coupled to green fluorescent prot(Dpp:GFP) (12, 13) in the disc dpp expressdomain (14). GFP fluorescence was detectboth in the Dpp-expressing disc cells and in

ASP. Amounts of Dpp:GFP in the ASP wehighest in the medial region of the ASP nearthe Dpp-expressing disc cells and in the lowlayer (Figs. 1D and 2A and table S1), showithat Dpp:GFP produced by the wing disc dtributed to the ASP in a manner that correlawith amounts of Dpp signal transduction (F1C and fig. S1, A to D). To examine the sucellular localization of marked Dpp in the ASwe expressed Dpp coupled to mCherry fluorcent protein (Dpp:Cherry) (5) in the disc dexpression domain, and Dpp signaling was motored in unfixed, livepreparations with a tragene that expresses nuclear-localized GFP (nGF

underDadcontrol. Dpp:Cherry puncta were oserved in multiple optical sections of ASP cewith strongly marked GFP-positive nuclei (F2, B and B); the presence of Dpp:Cherry punat apical positions (Fig. 2B) indicated that DpCherry had likely been taken up from the dby these ASP cells.

Whereas most tip cytonemes extended towthe region of the disc that expresses FGF (5, some TC and lateral cytonemes extended towaDpp-expressing disc cells (Fig. 1B). Expressiof Tkv:GFP marked puncta in these cytonem(Fig. 2C). To determine whether activated Twas present in cytonemes, we overexpressevariant of Tkv (TIPF) that fluoresces only in phosphorylated state and that has been usedmonitor receptor activation for Dpp or bone mphogenetic prote in (BMP) signaling (15). Acytonemes with bright fluorescent puncta wpresent under conditions of TIPF overexpress(Fig. 2D). Expression of Tkv:Cherry and TItogether in the TC and ASP generated puncta wboth green (TIPF) and red (Tkv:Cherry) fluorcence, indicating that Tkv in these puncta had beactivated (Fig. 2, E and E). We propose that presence of activated Tkv indicates that thecytonemes had received Dpp. The presence

RESEARCHARTICLE

Cardiovascular Research Institute, University of California,San Francisco, CA 94158, USA.

*Corresponding author. E-mail: tkornberg@ucsf.edu

www.sciencemag.org SCIENCE VOL 343 21 FEBRUARY 2014 124462

8/11/2019 Science 2014 Roy

4/12

cytonemes with only red fluorescence suggeststhat not all the cytonemes had received Dpp.

To further validate and characterize Dpp re-ception, ASPs were marked with either CD8:Cherry(mCherry fused to the extracellular and transmem-brane domains of the mouse lymphocyte proteinCD8), Tkv:Cherry, or FGFR:Cherry, and Dpp:GFPwas expressed in the discdppdomain in a pulseduring L3 (14). The ASP grows from the TC onthe anterior side of the disc and extends poste-

riorly across the stripe of Dpp-expressing cellsby late L3 (9)(Fig. 3A). At the midor latestages, animals that expressed CD8:Cherry andDpp:GFP had long ASP tip cytonemes markedwith Cherry fluorescence that oriented towardFGF-expressing disc cells. These cytonemes had noapparent GFP fluorescence (Fig. 3B). Lateral ASPcytonemes that projected toward Dpp-expressingdisc cells were also visible. These lateral cytonemeshad both Cherry and GFP fluorescence (Fig. 3,B and B), indicating that Dpp:GFP had beenreceived by these cytonemes. Dpp:GFP in punctafreefrom either cells or cytonemes was notdetected.

ASPs marked with Tkv:Cherry provided evi-dence that Dpp transport by cytonemes is asso-ciated with its receptor. Late-stage ASPs thatexpressed Tkv:Cherry had Dpp:GFP present intheir medial region and in lateral cytonemes thatextended from these cells, but there were fewTkv:Cherry-marked tip cytonemes, and Dpp:GFPwas present in much lower amounts in the distalASP cells (Fig. 3C). Some of the Dpp:GFPs presentin the medial ASP cells were associated withTkv puncta (Fig. 3C). These images show thatDpp:GFP appears to move from the disc and aretaken up by tracheal cells.

In mid-stage ASPs that expressed FGFR:Cherry

and whose tip had not grown beyond the Dpp-expressing zone of the disc, FGFR:Cherry-markedtip cytonemes extended over Dpp-expressing disccells toward the cells that expressed FGF (Fig. 3D).No Dpp:GFP puncta localized with the FGFR:Cherry-marked cytonemes. The absence of Dpp:GFP inthe FGFR:Cherry-containing tip cytonemes is con-sistent with the localization of the FGFR and Tkvreceptors to different cytonemes (5) and suggeststhat FGF and Dpp reception may be receptor-dependent and specific for cytonemes that con-tain FGFR or Tkv, respectively.

To better understand cytoneme-mediated move-ment of Dpp, we analyzed early- and mid-stagepreparations that had Tkv:Cherry expressed inthe trachea and Dpp:GFP expressed in the disc.Dpp source cells are distal to the ASP at thesestages. Long, Tkv:Cherry-marked cytonemes ex-tended toward Dpp-expressing disc cells (Fig. 3, Eand F). These cytonemes contained motile puncta(movie S1). Some cytonemes had both Tkv:Cherryand Dpp:GFP fluorescence and had brightly flu-orescent ends that localized with Dpp:GFP; theseimages suggest that these cytonemes contact Dpp-expressing disc cells. Not all cytonemes had bothTkv:Cherry and Dpp:GFP, suggesting that some,but not all, cytonemes had received Dpp:GFP.

These images are consistent with the patterns ofTIPF fluorescence (Fig. 2E). The presence ofDpp:GFP in tracheal cytonemes and the appar-ent contacts of cytonemes with Dpp-producingdisc cells suggest that the Dpp:GFP may movefrom the disc to the tracheal cells by direct trans-fer at sites of cytoneme contact.

Cytonemes Synapse with Wing Disc Cells

The cytoneme model of signaling protein dis-

persion posits that distant cells contact directlydespite their physical separation. To probe the ap-parent contacts at higher resolution, we adaptedthe GRASP (GFP Reconstitution Across Synap-tic Partner) technique, which was developed toimage membrane contacts at neuronal synapses(16,17). We expressed CD4:GFP110 (a fragment ofGFP that includes 10 strands of the GFP b-barrelphotocenter fused as an extracellular postscriptto the transmembrane domain of the mouse lym-phocyte protein CD4) and CD4:GFP11 (a frag-

ment that includes the 11th strand of the Gb-barrel). To image cytoneme contacts, the tparts of GFP were expressed separately in tcheal cells and in either FGF- or Dpp-expressidisc cells. These nonfluorescent GFP fragmegenerated fluorescence that localized specificly at the disc cells that expressed either FGFDpp (Fig. 4, A to C). Expression of mCherCAAX (CAAX is a plasma membranetargetmotif) in the discdppdomain revealed that GRA

fluorescence correlates withdpp-expressing ce(Fig. 4C). Fluorescence was separated from ASP cells by up to 40 mm (Fig. 4, A and B), approximate length of the longest cytonemthat projected from the ASP toward disc ceindicating that ASP and disc cells synapse evwhen separated. GFP fluorescence was not oserved in animals that expressed only one of fragments.

To show that the GRASP fluorescence wassociated with cytoneme contacts, cytonem

Fig. 1. Dpp produced in the wing disc signals to disc-associated tracheal cells. (A) Projectimage of a third instar wing disc (outlined with dashed red line) showing disc-associated trachea (markwith green; membrane-tethered GFP) and Dpp-expressing disc cells (red, marked by antibody against Lthat was expressed in thedppdomain). TC, DT (dorsal trunk), and ASP are labeled. Dotted circle indicaarea of disc that expresses FGF. (B) Expression of CD8:RFP marks dpp-expressing disc cells (red); exprsion of CD8:GFP in trachea (lexO-CD8:GFP, UAS-CD8:RFP/+; btl-LexA/+; dpp-Gal4/+) marks cytonemextending from ASP tip, from the lateral, medial region of the ASP, and from the TC (arrowhead), showithat some ASP and TC cytonemes orient toward Dpp-expressing cells. This plane of focus does not detect

dpp-expressing cells due to folds in the disc near the A/P organizer, but it did detect many dpp-expressanterior cells that are in the plane of focus as scattered in the A compartment region between the Aand TC. (C) Staining an ASP with antibody to pMAD (red) to show Dpp signaling in the medial regioAntibody to Dlg (discs large, green) marks cell outlines in ASP (bounded by white dashed line) adiscs. (D) Cartoon of a sagittal ASP section depicting the position of the disc epithelium and basal lamrelative to the ASP in the late L3; dashed lines represent approximate locations of the upper and lowoptical sections in all figures. (E) Overexpression of TkvDN in trachea (btl-Gal4) generated bifurcatabnormally shaped ASPs. (F) dppRNAiexpression in the dorsal compartment of the disc (ap-Gal4tuGal80ts) reducedDadexpression (Dad-nlsGFP, green) in disc-associated trachea (right panel) compato control (ap-Gal4tub-Gal80ts Dad-nlsGFP) ASP (left panel); abnormal ASP growths are indicatedwhite dotted lines; cells are marked with a-Dlg staining (red); and both panels show lower layer of AOrientation of discs in all figures: anterior, left; dorsal, down. Conditions of Gal80ts inactivation for and (F) are described in table S1 (14). Scale bars, 30 mm.

21 FEBRUARY 2014 VOL 343 SCIENCE www.sciencemag.org1244624-2

RESEARCH ARTICLE

8/11/2019 Science 2014 Roy

5/12

were marked independently of the GRASP GFPfragments by expression of mCherry-CAAX orTkv:Cherry. Fluorescence of reconstituted GFPwas mostly at or near cytoneme tips that contactedsource cells (Fig. 4, Aand B). Tkv:Cherry flu-orescence had a punctal distribution in thesecytonemes and was also present at contact sites(Fig. 4B). An estimate of the size of the CD4domains (diameter, ~65 ) (18, 19) and of thelinkers that join CD4 to the GFP fragments sug-

gests that the apposition of a cytoneme tip with atarget cell at a synapse is less than 20 nm. Thisdistance is comparable to neuronal and immunesynapses, and because GFP photocenter matura-tion is not instantaneous (20), the GRASP flu-orescence indicates that cytonemes can makerelatively stable contacts with target cells.

The proximity of the tubular ASP and thedisc varies along the ASP proximodistal axis (1.5

to 10mm), and the anatomies of the two epitheliaare complex (Fig. 4, D and D). The ASP cellsthat overlie Dpp-expressing disc cells are in closeapposition, yet in this region, cytonemes ema-nated from both the ASP (Fig. 4E) and the disc(Fig. 4F). The ASP cytonemes in this regionwere short (10 mm); the disc cytonemes wereas long as 30 mm, and many had bright bulboustips at apparent points of contact with ASP cells.GRASP marked the contacts between the lower

layer of the ASP and the disc (Fig. 4G), but didnot resolve the relative contribution of the ASPand disc cytonemes.

In the wing pouch primordium of the wingdisc, Dpp-dependent cytonemes on the apicalcell surfaces orient toward the stripe of Dpp-expressing cells at the A/P developmental or-ganizer and may ferry Dpp from the A/P organizerto cells as far away as the disc flanks (35). We

applied GRASP to image contacts between wing disc A/P organizer and flank cells by expreing the GFP fragments at the A/P organizer ain flank cells (Fig. 4H). In these discs, GFP forescence was observed in the region of the ganizer (Fig. 4H), in contrast to discs that expressonly one of the complementing fragments (F4H). This pattern of GFP reconstitution sugests that cytonemes may extend from the ceat the disc flanks to synapse with cells of the A

organizer.

Dpp Signaling in the ASP Requires

Cytoneme-Mediated Transport

We identified four genes that are required ASP morphogenesis and for cytoneme functidiaphanous (dia),shibire(shi), neuroglian(nrand capricious (caps). Mutant loss-of-functiconditions were induced selectively in trachduring the L3 stage (14), and mutant ASPs wabnormal or duplicated at variable expressivand penetrance (table S2 and fig. S1); we shand describe ASPs that were most normalappearance. Wing discs in these experimen

were not mutant, and wing disc developmeappeared normal.The formin Dia is an actin nucleation fac

(21) whose activated form localizes to the tof filopodia (22). When Dia:GFP and activaDia:GFP (23)were expressed in the ASP, Dia:Gwas mostly in the cell body and was presentlow levels in cytonemes, but activated Dia:Gwas prominent in most cytonemes and localizto cytonemes tips (Fig. 5A). The distributionactivated Dia indicates that cytoneme tips mbe sites of actin nucleation. To examine the roleDia, we expresseddiaRNAiin the ASP during L3 stage. In >85% of the animals (n= 26), grow

of the ASP was decreased and ASP morphgenesis was abnormal (for example, fig. S1The number of cytonemes was also decreasand many of the cytonemes that extended frmutant ASPs were abnormally short and hblunt tips (Fig. 5, B to E), and Dpp signal traduction (Dad-GFP expression and pMad abdance) was decreased (Fig. 5, F and G, and tabS3). We did not detect changes to cell shapnumber of dividing cells, or number of dyicells in mutant ASPs (fig. S2, A and B). ThuDia appears to be required by the ASP to macytonemes, and the defective cytonemes thare made in the absence of normal Dia functiare incapable of mediating Dpp signaling frothe disc.

We expressed a conditional mutant ofshib(fruit fly dynamin; shits1) (24) together wCD8:GFP in the trachea and compareda-pMstaining as well as the number and lengthcytonemes in ASPs that were isolated from lvae that had been incubated at either permiss(18C) or restrictive (30C) temperature (Fig. 5HDynamin is a multimer (25, 26), and under nopermissive conditions, the Shits1protein functioas a dominant negative (24). Control larvae sujected to 3 hours at 30C did not change

Fig. 2. The ASP takes up Dpp, and ASP cytonemes contain activated Tkv receptor. (A) Dpp:GFPexpressed in the disc dppdomain (dpp-LexA lexO-Dpp:GFP, dashed arrow) is present (arrows) in theupper and lower ASP layers in this unfixed preparation. ASP is outlined by white dotted lines. (BtoB)Dpp:Cherry expressed in the disc dppdomain (dpp-Gal4/UAS-Dpp:Cherry, Dad-nGFP/tub-Gal80ts) wasdetected as intracellular puncta (arrows) in ASP cells that also induce Dadexpression. ASP outline ismarked by white line [(B) and (B), sagittal sections; (B), transverse section]. (C) Expression of Tkv:GFP(btl-Gal4 UAS-Tkv:GFP) marks puncta (arrowheads) in ASP cytoneme. (D) Expression of TIPF (btl-Gal4UAS-TIPF) marks puncta in ASP cytoneme. (EandE) TIPF (green) and Tkv:Cherry fluorescence (btl-Gal4/UAS-Tkv:Cherry; tub-Gal80ts/UAS-TIPF) colocalizes (arrowheads) in puncta in some, but not all, ASPcytonemes. Arrow, cytoneme with Tkv:Cherry only; dashed arrow, cytoneme with both TIPF and Tkv:Cherry; left panel, merge; right panel, TIPF only. Gal80 ts inactivation for (B) and (E) was for 6 to 8 hoursin mid L3, followed by incubation at 25C for 6 to 12 hours. Scale bars, 10 mm, except for (A), 30mm.

www.sciencemag.org SCIENCE VOL 343 21 FEBRUARY 2014 124462

RESEARCH AR

8/11/2019 Science 2014 Roy

6/12

number of short(25 mm)ASP cytonemes (~4.4 and ~4.9% reduction, re-spectively) or reduce amounts of pMad (~7%).However,shi ts1 larvae subjected to 30C had de-creased cytoneme numbers and pMad abundance(Fig. 5I and tables S3 and S4). The number oflong cytonemes present after 30 min at 30Cwas less than 10% of that in control experiments;numbers of short cytonemes also declined after30 min at 30C. Reductions in numbers of short

cytonemes and amounts of pMad became moresevere over time intervals of up to 3 hours. A 2-hourheat pulse and a 1-hour incubation at 20C partiallyrestored both long and short cytonemes (14), butthe ASP morphology was not normal. Indeed, du-plicated, abnormally shaped ASPs were producedwhen a 24-hour incubation at 20C followed a1-hour heat pulse (fig. S1I). Adults that devel-oped at 20C after a 2-hour heat pulse appearedto have normal morphology, and we did not ex-amine the structure or function of their dorsal airsacs. Thus, Shi inactivation was not lethal in thecells of the ASP; the consequences of Shi inactivationon ASP development were partially reversible; and

the effects on cytonemes preceded the reduction insignaling (as revealed by amounts of pMad).To distinguish whether ASPs that are defi-

cient fordia orshiexpression failed to activateDpp signal transduction because they did notreceive Dpp from the wing disc or were incapableof initiating a response, we expressed Dpp:Cherrydirectly in ASPs with the btl-Gal4 driver (14).Ectopic Dpp induced pMad in ASPs with reduceddiaor Shi function (fig. S3, A and B). Thus, con-ditions that reduceddiaexpression or inactivatedShi did not abrogate the ability of ASP cells torespond to Dpp, and blocking cytoneme-mediateduptake of Dpp from the disc appears to be the

most likely cause of the signaling deficits.Cytonemes were also defective in loss-of-function conditions for nrgand caps, both ofwhich encode putative cell adhesion transmem-brane proteins. Nrg is an L1-type cell adhesionmolecule implicated in the development and sta-bility of neuronal synapses (27). Although flu-orescence of an in-frame protein trap Nrg:GFPfusion protein was detected in the ASP, ASPcytonemes could not be resolved because ofbackgroundexpression in the wing disc. How-ever, overexpression of Nrg:GFP in the ASP re-vealed that Nrg distributes in the ASP cytonemesand concentrates at the cytoneme tips (Fig. 5J).Expression ofnrgRNAi reduced the number ofboth tip and lateral cytonemes (Fig. 5K andtable S5), abrogated Dad-GFP expression anddpERK (diphosphoextracellular signalregulatedkinase) staining (Fig. 5, L and M), and causedgrowth of abnormal, duplicated ASP lobes (fig.S1K). Expression ofnrgRNAi had no apparenteffect on cell shape or the number or distributionof dividing or dying cells (fig. S2, C and D). Ex-pression of Dpp:Cherry together with nrgRNAiin the ASP restored Dpp signaling (fig. S3C),indicating thatnrg-deficient ASP cells can ac-tivate Dpp signal transduction.

We identifiedcapsin an enhancer trap screenfor genes that are expressed in the ASP (14) (fig.S2, E and F). Caps:GFP that was expressed inthe trachea was detected in ASP cytonemes and

concentrated at the tips (Fig. 6A). Caps and paralog Tartan (Trn) have extracellular domacontaining leucine-rich repeats (LRRs) and cotribute partially redundant functions to the f

Fig. 3. Tkv-containing cytonemes transport Dpp. (A) Drawings of three third instar stages depgrowth and development of the ASP (red) relative to wing disc cells expressing Dpp (green) and F(blue). (BandB) Expression of CD8:Cherry in the ASP and Dpp:GFP in thedppdomain of the disc (bGal4 UAS-CD8:Cherrydpp-LHG/lexO-Dpp:GFP) marks the ASP and ASP cytonemes (red) and dpp-expressdisc cells (green). GFP fluorescence is in the lateral ASP cytonemes (arrows) and in the lower medial regi

of ASP, but not in the tip of ASP cytonemes (arrowhead). Left panel, merge; right panel, GFP. ( CandExpression of Tkv:Cherry in the ASP and Dpp:GFP in the dppdomain of the disc (btl-Gal4/UAS-Tkv:Chedpp-LHG/lexO-Dpp:GFP) marks the ASP and lateral ASP cytonemes (red), but few tip cytonemes; lateTkv-containing ASP cytonemes and the medial region of the ASP have received Dpp:GFP (green) (Dpp:GFP and Tkv:Cherry colocalize in puncta in ASP cells (C, arrows). (D) FGFR:Cherry expressed in Aand Dpp:GFP in thedppdomain of the disc (btl-Gal4/UAS-Btl:Cherry dpp-LHG/lexO-Dpp:GFP) marks punin the ASP tip cytonemes (arrow) that project beyond Dpp-expressing disc cells (green); no localizationFGFR:Cherry with Dpp:GFP was apparent in tip cytonemes. (E, F, andF) Only cytonemes marked wTkv:Cherry that appear to contact Dpp:GFP-expressing disc cells (btl-Gal4UAS-Tkv:Cherry;dpp-LHG/lexO-DGFP) have GFP fluorescence in puncta and at their tips (arrows). Cytonemes that do not appear to macontact do not have GFP fluorescence at their tips or in their Tkv-containing puncta (F, arrowheads) lack Gfluorescence. (F) merge; (F) Dpp:GFP. Animals were raised at 18C to minimize transgene expression awere incubated at 22 to 25C for 12 to 16 hours before analysis. Scale bars, 10 mm.

21 FEBRUARY 2014 VOL 343 SCIENCE www.sciencemag.org1244624-4

RESEARCH ARTICLE

8/11/2019 Science 2014 Roy

7/12

mation of compartment boundaries of the wingdisc (28). caps mutants do not mediate selec-tion of synaptic partners normally (2932),and Caps protein localizes at filopodia tipsduring partner recognition and synapse stabi-lization (30). We observed similar types of ef-fects on ASP cytonemes.

Lack ofcapsfunction also led to abnormalASP development. Expression of capsRNAi,trnRNAi, or a dominant negative Caps mutant

(CapsDN

) that localizes similarly to wild-typeCaps in synapses and decreases synaptic contacts(30, 32) reduced Dpp signaling and yielded ab-normal ASPs (14) (Fig. 6B; fig. S1, L to N; andtables S2 and S3). Phenotypes were more ex-treme in a heterozygous caps trn double-mutantbackground. Expression of CapsDN did not causedetectable changes to cell polarity, cell morphol-ogy, mitotic activity, or cell viability (fig. S2, Gand H).

CapsDN reduced amounts of dpERK (Fig. 6C),indicating thatcaps function was also requiredfor FGF signaling. Evidence that signal transduc-tion per se was not abrogated in ASP cells thatlackcapsfunction was obtained by overexpress-ing FGF ubiquitously. Heat shockinduced expres-sion of FGF or expression of Dpp:Cherry in theASP increased amounts of dpERK or pMad, re-spectively, throughout the ASP, attenuating theeffects of CapsDN (fig. S3D). These experiments

show that Dpp and FGF proteins that are producedby the disc (Fig. 1) (9)require caps function inthe ASP to activate signal transduction in ASPcells, and show that mutant ASP cells that cannotreceive FGF and Dpp from the disc are com-petent for signal transduction.

The presence of Caps:GFP in the tips ofcytonemes (Fig. 6A), the role of Caps at neu-ronal synapses (30), the fact that cytonemesmake contact with Dpp-producing cells (Fig. 4,

B, C, and G) and receive Dpp at apparent poiof contact (Fig. 3, E and F), and the essentrole ofcapsfor Dpp signaling suggest that Camay be required for cytonemes to establish funtional contacts for Dpp exchange. However, number and distribution of ASP cytonemes not change under caps loss-of-function contions (fig. S2I), indicating that the ASP cells not require Caps to make cytonemes. In contrathe contacts that ASP cytonemes made with Dp

expressing disc cells requiredcaps. We monitothese contacts with GRASP fluorescence: Gfluorescence at the interface of Dpp-expressdisc cells and the lower layer of the ASP, andcytoneme contacts of the lateral ASP and Twas reduced when CapsDN was expressed in trachea (Figs. 1D; 4, C and D; and 6, D and In addition, CapsDN reduced uptake of Dpp:Gfrom the disc (Fig. 6F), suggesting that althouASP cells make cytonemes in the absence

Fig.4. Tracheal cytonemes con-tact Dpp- andFGF-expressingdisccells.(A, A, B,and B) Green

fluorescence (arrowheads) fromreconstituted GFP (GRASP) dueto contact between ASPcytonemesand disc shown in projection im-ages composed of several upperto mid optical sections. ASP (dashedwhite line), disc, and TC lumenwere imaged at 405 nm for back-ground fluorescence (gray). Nor-maldppexpression includes cellsanterior to the stripe at the A/Pcompartmentborder (see Fig. 1, Aand B). Marking cytonemes withCherry-CAAX (A) or Tkv:Cherry(B) showed that GRASP fluores-

cence was cytoneme-associated(arrowheads).(C) Left panel: draw-ing of third instar wing disc de-pictingDpp-expressing cells (red)and ASP and TC (gray). Right pan-el:region outlined by dashedlinesin left panel for GRASP fluores-cence (green) at the basal surfaceofdpp-expressing disc cells (red).(DandD) Sagittal (D) and trans-verse (D) sections in the mid-region of ASP show the spatialrelationship of the ASP(red) lowerlayer todpp-expressing disc cells(green, dpp-CD8:GFP:red,btl-Cherry-

CAAX). (E) CD8:GFP expressed intheASPmarks cytonemesemanat-ing from the lower aspect of theASP; they orient toward the disc.(F) CD8:GFP expressed in the discmarks cytonemes that extendtoward and appear to contact(arrowheads) ASP cells markedwith Cherry:CAAX (btl-Cherry:CAAXdpp-CD8:GFP). (G) GFP reconstitution in four optical sections of (B) from the upperlayer,from thetwomiddlelayers, andfrom theinterfacebetween lower layer anddisc.(H) Drawing of the wing pouch region of a wing disc showing the stripe of dppexpression at theorganizer(purple) and theflanking regions that express brinker(brk,

orange). Box with dashed line indicates region imaged in (H) and (H). (H) constitutedGFP(arrowheads)in theorganizer regionin disc with expression of theGfragments in thebrkanddppdomains. (H) Control with CD4:GFP

110 expressionthedppdomain only. Scale bars, 30mm, except for (A), (B), (E), and (F), 10mm

www.sciencemag.org SCIENCE VOL 343 21 FEBRUARY 2014 124462

RESEARCH AR

8/11/2019 Science 2014 Roy

8/12

Fig. 5. ASP cytonemes requirediaand Shi and nrg. (A) In the ASP, lo-calization of Dia:GFP is predominantly in the cell bodies; activated Dia (DiaDDad:

GFP) localizes to cytoneme tips (btl-Gal4,UAS-CD8:Cherry/+;tub-Gal80ts

/UAS-Dia:GFP, orUAS-DiaDDad:GFP). (BtoG) Expression ofdiaRNAishortened lateral andtip of ASP cytonemes (btl-Gal4,UAS-CD8:Cherry/tub-Gal80ts;Dad-GFP/UAS-diaRNAi),and reduced expression ofDad-GFP. Control genotype:btl-Gal4, UAS-CD8:Cherry/tub-Gal80ts; Dad-GFP/+. (HandI) Late third instar larvae that coexpressed shts1 andCD8:GFP (btl-Gal4, UAS-CD8:GFP, UAS-shits1) were incubated at 30C for theindicated times and after dissection; GFP fluorescence and a-pMad staining(red) were imaged in the lower layer of the ASPs (see Fig. 1C). The perimeterof each of the five ASPs was measured, cytonemes were counted (I) around theperimeter in about 35 to 40 optical sections, and the length of each cytonemewas measured. Graph (I) shows the average percentage change to the numberof ASP cytonemes per micrometer perimeter in the length ranges of 25 mm (red). Amounts of pMad were determined by measuring

the mean fluorescence intensity (555 nm) in four ASPs for each time pofor a region of the lower ASP level that contained about 11 cells. ( J) N

GFP (btl-Gal4, UAS-CD8:Cherry/UAS-Nrg:GFP; tub-Gal80ts

) localizes to aconcentrates at the tips (arrowheads) of ASP cytonemes. (K) Late thinstar larvae that coexpressednrgRNAiand CD8:GFP (btl-Gal4, UAS-CDGFP/UAS-nrgRNAi;tub-Gal80ts/+). Lateral and tip cytonemes were stuntand reduced in number. (L) Expression of Dad-GFP was reduced in a lowASP layer that expresses nrgRNAi (lower panel; btl-Gal4, UAS-CD8:CherUAS-nrgRNAi; Dad-GFP/tub-Gal80ts) compared to control (upper panel; bGal4,UAS-CD8:Cherry/+; Dad-GFP/tub-Gal80ts). (M) dpERK staining (arrored) is partially reduced in ASP that expresses nrgRNAi(lower panel,btl-GaUAS-nrgRNAi; tub-Gal80ts/+; upper panel, control btl-Gal4/UAS-nrgRNAi; tGal80ts/+); outline of ASP marked with dashed line anda-Dlg (green) outlicells. Conditions for conditional inactivation are described in table S2 (1Scale bars, 25 mm.

21 FEBRUARY 2014 VOL 343 SCIENCE www.sciencemag.org1244624-6

RESEARCH ARTICLE

8/11/2019 Science 2014 Roy

9/12

Caps function, Caps-deficient cytonemes thatdo not make stable synapses do not transfer Dppfrom producing to recipient cells.

Discussion

This study revealed an essential role for cytoneme-based transport of signaling proteins in long-distance paracrine signaling. This mechanisminvolves contact-dependent transfer of signalingproteins from producing to responding cells, and

although we studied signaling between two epi-thelial tissues in a Drosophila larva, evidencefrom other systems supports a general role forcytonemes in paracrine signaling.

Studies of cells in culture indicate that filopo-dia receive and transport signaling proteins thatare taken up from culture medium. In experi-ments with human adenocarcinoma cells, uptakeof epidermal growth factor (EGF) protein fromthe culture medium led to retrograde transportby filopodia along with activated EGF receptor(EGFR) and was sensitive to cytochalasin D, adisruptor of F-actin (33). Actin-based cytonemes

that carry FGFR-rich puncta and that are depen-dent on the small GTPase (guanosine triphos-phatase) RhoD are present in cultured mousemesenchymal cells (34).

Some characteristics of Dpp signaling in theASP are consistent with these cell culture exper-iments. Dpp that was taken up by an ASP cellwas present in motile puncta that translocatedalong the ASP cells cytoneme, and some punctain the ASP cytonemes contained both Dpp and

its receptor (Figs. 2, C and D; 3, C, E, and F;and 4B).Drosophilacytonemes are actin-based(3). However, in contrast to cultured cells, sig-naling in the ASP did not appear to involveuptake of signaling proteins from the extracellularmilieu, but was dependent on synaptic contactbetween the tip of a cytoneme that extended froma responding ASP cell and the cell body of a Dpp-expressing disc cell. This signaling mechanismappears to involve specific dynamic interactionsbetween signaling and responding cells.

ASP cells express both the Tkv Dpp receptorand FGFR, and segregate these receptors to

puncta in distinct cytonemes (5). At the early stage, the ASP is small and does not extend acrothe disc, and both the Dpp- and FGF-expressidisc cells are distal to its tip. Both Tkv- and FGFcontaining cytonemes extended distally from tip (5). The FGFR-containing cytonemes extendbeyond the Dpp-expressing cells and did not tup Dpp (Fig. 3D). At later L3 stages, the Ahas grown across the disc, and although the FGexpressing disc cells are distal to it, the Dp

expressing disc cells lie under its medial regionthese ASPs, the Tkv-containing cytonemes eanated from the medial region and reachedmuch as 40 mm to pick up Dpp from disc ce(Fig. 3, B and C). Thus, in the ASP, spatiarestricted Dpp signal transduction (Fig. 1, C aF) and uptake (Figs. 2A and 3, B and C) wassociated with cytonemes whose orientation acomposition appeared to be specific for Dpp.

The dynamism of this signaling system mbe inferred from steady-state images. The dtribution and orientation of cytonemes changeexpression of signaling protein is compromis

Fig. 6. ASP cytonemes requireCaps.(A) Caps:GFP (btl-Gal4 UAS-Caps:GFP, tub-Gal80ts) localizesto and concentrates at the tips(arrowheads) of cytonemes. Con-ditions for conditional inactiva-tion are described in (14). (B) ASPexpression (btl-Gal4) of CapsDN

(middle panel) and capsRNAi(right panel) reduced Dad ex-pression in the ASP (Dad-nlsGFP;green); left panel, control. (C) ASPexpression (btl-Gal4) of CapsDN

reduced dpERK staining (red) atthe tip of ASP. Cells are marked

witha-Dlg staining [red, (B); green,(C)]. (D) Sagittal optical sectionsat lower level of ASP (left andmiddle panels) and in coronalsections (right panels) showingthat GRASP fluorescence is re-duced by expression of CapsDN

(at 29C); CapsDN genotype in-cludes two copies ofUAS-CapsDN.TC indicates the lumen of thetransverse connective. (E) CapsDN

expression in the TC reducesGRASP fluorescence (arrows) as-sociated with Dpp-expressing disccells. Genotypes: same as (D). (F)

Dpp:GFP uptake in ASP (arrow) inthe presence (bottom panel) andabsence (top panel) of CapsDN.Genotypes: same as (D). In (D) to(F), ASP, disc, and TC are imagedfor autofluorescence at 405 nm(gray). Scale bars, 30mm, exceptfor (A), 10mm.

www.sciencemag.org SCIENCE VOL 343 21 FEBRUARY 2014 124462

RESEARCH AR

8/11/2019 Science 2014 Roy

10/12

and if signaling protein is overexpressed in ec-topic locations (4, 5, 9). These properties sug-gest that cytonemes are changeable and thattheir distributions reflect the relative positions ofsignal-producing and signal-receiving cells. Thedifferent distributions of Tkv-containing cytonemesin the temporal sequence described above areconsistent with this idea and with a model ofcytoneme impermanence. The observation thatsome ASP cytonemes contain Tkv, make contact

with Dpp-producing cells, and take up Dpp,whereas other cytonemes contain Tkv but do notmake contact with Dpp-producing cells or takeup Dpp (Figs. 2E and 3F), may also suggest thatcytoneme contacts may be transient.

Plasticity may be an important attribute ofcytonemes that function in a developmental sys-tem such as the ASP, in which relations betweenproducing and receiving cells change as the larvadevelops. Cytonemes may have the capacity toregulate release and uptake of signals and to di-rect signals to a preselected target. Regulatedrelease may be implied by the absence of Dppuptake and Dpp signal transduction in ASP mu-

tant conditions that abolish synaptic contacts byASP cytonemes. In these experiments, the signal-producing cells were not mutant, and the wingdiscs, which depend on Dpp signaling, developednormally, indicating that the signaling defectwas specific to the ASP cells that made defectivecytonemes. Because filopodia of cultured cellstake up signaling proteins from culture mediumand activate signal transduction (33), we may as-sume that ASP cytonemes are similarly capableof taking up signaling protein that their receptorsencounter and that the inability of cytoneme-defective cells to take up Dpp or activate signaltransduction may indicate that Dpp was not re-

leased in the absence of cytoneme contact.There may be a functional analogy to neuronalsignaling. Neurons make asymmetric extensionsthat send and receive signals, make specific con-tacts where signal release and uptake are reg-ulated, and require the diaphanous, neuroglian,shibire, andcapriciousgenes for contact-mediatedsignal exchange and signaling. In the developingDrosophila retina, Hh moves to the termini ofretinal axons, where it signals to postsynapticlaminal neurons in the brain (35). Perhaps thestrongest precedent is Wingless delivery at de-veloping neuromuscular junctions in the Dro-sophila larva; in this case, Wingless moves tothe postsynaptic cell after release in a vesicularform from the presynaptic neuron (36). Our studieshave been limited to the cytonemes that are madeby receiving cells, but in other contexts, cytonemesextend from cells that deliver signaling proteins,such as the Hh-containing cytonemes of the wingdisc (7) and the cytonemes that extend from capcells in the female germline stem cell niche (8)and that are associated with Notch and EGF sig-naling (3740). Cytonemes that transport Hhacross the chick limb bud from Hh-producingcells have also been described (41). The wide-spread presence of cytonemes in many cell types

and in many contexts suggests that they mayprovide a general mechanism to move signalingproteins between nonneuronal cells at sites ofdirect contact.

Materials and Methods

Drosophila Stocks

Transgenes: btl-Gal4 (9),ap-Gal4 [BloomingtonStock Center (BSC)]; dpp-Gal4/CyO; HS-Bnl

(9);UAS-Tkv:GFP (4);UAS-Dpp:Cherry, UAS-CD8:Cherry,UAS-CD8:GFP (5),Dad-nEGFP (III)(42), UAS-FGFRDN (43); dpp-LHG/TM6(LexA-Gal4 activation domain fusion; III) (44),dpp-LHG(II; this study), lexO-Dpp:GFP/TM6andbrkBM14-LHG (44),btl-LHG (II and III) (this study), lexO-CD4:GFP11 (II),UAS-CD4:GFP110 (III) (17),UAS-dppRNAi(BSC),UAS-putRNAi(BSC);tub-Gal80ts (II and III; BSC), UAS-Dad (II) (BSC),UAS-TkvQD/TM6B [activated Tkv (45)], UAS-TkvDN, UAS-PutDN [DGSK, dominant negativeforms of Tkv and Put lacking GS and kinase do-main (46)],UAS-TIPF (15),UAS-capsRNAi[BSC,Vienna Drosophila RNAi Center (VDRC)], UAS-

trnRNAi [National Institute of Genetics (NIG),BSC], UAS-CapsDN (30); UAS-CD4:GFP10 (II;this study), UAS-diaRNAi (BSC, NIG), UAS-shits1 (24),lexO-CherryCAAX [II and III (44)];UAS-Dia:GFP (47) and UAS-DiaDDad:GFP (23);UAS-nrgRNAi (II and III) (BSC); 10XUAS-IVS-mCD8::RFP;13XlexO-mCD8::GFP (BSC);UAS-Nrg:GFP (II) (27).

Insertions and mutations:Dadj1E4-LacZ/TM3,tkvk16713-LacZ/CyO,dpp10638-LacZ/CyO,put10460-LacZ/TM3(BSC), and Nrg:GFP protein trap line(flytrap line G00305). Conditional inactivation ofDpp was indpp4/dpp56L3 larvae for 18 hours at29C as described (4).

Overexpression

tub-Gal80ts was present to limit expression tothe L3 stage. Expression drivers were as fol-lows: ap-Gal4 fordppRNAi; btl-Gal4 for Dad,TkvDN, TkvQD, PutDN,putRNAi, CapsDN, capsRNAi,trnRNAi, diaRNAi, Dia:GFP, DiaDDad:GFP,nrgRNAi, and Nrg:GFP. Animals were reared at18C until L3 and were incubated at 29C, as in-dicated in table S2, before dissection. For knock-down under heterozygous mutant background(table S2 and fig. S1M), CapsDN andcapsRNAiex-pression was driven bybtl-Gal4 or bydpp>Gal4incapsC28fs trnD17andcaps65.2 trnS064117 double mu-tants. At 25C, CapsDN andcapsRNAi overex-pression is embryonic lethal in the Caps mutantbackground; animals were therefore reared at20C to the L3 stage and were incubated at 25Cfor 1 day before dissection.

Ectopic Expression

For fig. S3, A to C, crosses were, fordia,shi, andnrg:btl-Gal4,UAS-CD8:GFP/+;tub-Gal80ts/UAS-dpp:Cherry to eitherUAS-diaRNAi, UAS-shits, orUAS-nrgRNAi. Control larvae expressed eithershits, diaRNAi, ornrgRNAi, but lackeddpp:Cherry;experimental larvae had UAS-dpp:Cherry. Ani-

mals were reared at 18C to minimize the effeof Dpp overexpression. To express diaRNAi, larvae were incubated at 25C for 5 to 6 houShits larvae were treated similarly and were thshifted to 29C for 1 hour. ASPs in the Shits larvdid not grow normally because of temperatsensitivity ofshits at 25C.nrgRNAiinduction wfor 14 to 18 hours at 29C. CapsDN larvae (btl-GUAS-CD8:GFP/UAS-CapsDN; UAS-CapsDN/HS-Bwere reared at 20C until L3; heat shock indu

tion of Bnl was for 30 min at 37C followed 3 hours of incubation at 20C.

Dual Expression

LexA and Gal4:10XUAS-IVS-mCD8::RFP,13XlexmCD8::GFP flies (BSC) were crossed to dpp-GaSM5;btl-LHG flies to mark Dpp-producing cein wing disc with RFP (red fluorescent proteand trachea with GFP. To express either Tkv:Cheor FGFR:Cherry in trachea simultaneously wDpp:GFP in the wing disc,UAS-Tkv:Cherry/CyWeep;dpp-LHG/TM6orUAS-FGFR:Cherry/CyWeep;dpp-LHG flies were crossed to btl-GalexO-Dpp:GFP/TM6flies. To minimize toxic

fects, btl-Gal4/UAS-Tkv:Cherry (or FGFR:CherrlexO-Dpp:GFP/dpp-LHG animals were grown18C until the L2 stage and were shifted to 20

Enhancer Trap Screening

About 500 lines with randomly inserted enhantrap transposons (gift from E. Heberlein) wescreened for tracheal expression (UAS-GFP)line with elevated expression in the ASP was idtified; its Gal4 transposon was mapped by enout sequencing to the first exon ofcaps. Wing dGFP expression was similar to the expressioncapsas indicated by in situ hybridization (28

GFP Reconstitution

Genotype for reconstitution between Dpp signalpartners: dpp-Gal4/lexO-CD4:GFP11; btl-LHUAS-CD4:GFP110. Genotype for reconstition between FGF signaling partners: btl-LHlexO-CD4:GFP11; bnl-Gal4/UAS-CD4:GFP1

For reconstitution in the wing disc: dpp-GalexO-CD4:GFP11;brk-LexA/UAS-CD4:GFP110. Freconstitution with marked cytonemes:btl-LHlexO-CherryCAAX/lexO-CD4:GFP11; bnl-Gabtl-LHG/UAS-CD4:GFP10. For reconstitution win the presence of marked Tkv:btl-Gal4,dpp-LHUAS-Tkv:Cherry;lexO-CD4:GFP11,UAS-CDGFP110. For reconstitution with marked Dsource: btl-Gal4, dpp-LHG/+; lexO-CherCAAX/UAS-CD4:GFP110, lexO-CD4:GFPFor reconstitution together with CapsDN overpression, btl-Gal4, dpp-lexA; UAS-CD4:GFP1

lexO-CD4:GFP11 flies were crossed with eithw-(control) orUAS-CapsDN. Larvae were rearin room temperature and shifted to 25 or 29for 1 day before assay.

shibire Inactivation

Larvae [btl>Gal4, UAS>CD8:GFP, UAS>sh(24)] were raised at 18C before a single heat shofor 0.5, 1, 2, or 3 hours at 30C. Larvae were d

21 FEBRUARY 2014 VOL 343 SCIENCE www.sciencemag.org1244624-8

RESEARCH ARTICLE

8/11/2019 Science 2014 Roy

11/12

sected and imaged for ASP cytonemes or werefixed for pMad staining. Rescue after heat shockwas by returning larvae to 18C before dissectionand imaging. Control heat shock was with larvaeexpressing CD8:GFP in trachea (btl-Gal4 UAS-CD8:GFP) at 30C for 0 and 3 hours. No sig-nificant change in numbers of cytonemes [either25 mm (4.7 T7.6% increase)] was detected. Rescue after 30Cat 2 hours was at 20C for 1 hour, followed by

dissection and imaging. Increases in numbers ofcytonemes [25 mm (11 T 2.9%, P = 0.0196)] wereevaluated by the unpaired ttest.

Quantitation and Statistical Analyses

Cytonemes were counted and measured in z-sectionstacks of confocal images from five ASPs for eachdata point and were binned as 25 mm.Lengths represent measures from each tip alongthe connecting shaft to the point of its widen-ing base either at the plasma membrane or at thelamellipodia-like protrusion. The size variationbetween ASPs was normalized by measuring

the perimeter of each ASP and then calculatingthe number of cytonemes per unit length. Valuesin Fig. 5I are plotted as percentage of the 0-hourtime point. pMad levels were quantified by mea-suring the mean intensity of 555-nm fluorescencein the cells of the lower layer of ASP, subtractingbackground fluorescence, and normalizing withrespect to pMad fluorescence at the A/P borderof the same wing disc. Values were plotted aspercentage of the 0-hour time point. Statisticalsignificance values were calculated withttest oranalysis of variance (ANOVA) followed by Tukeyhonestly significant difference (HSD) test.

Molecular Cloning

btl>LHG: The P[B123] fragment upstream of thebtlgene (48)was amplified from a genomic cloneobtained from (49), with 5primer GGCTCGA-GATAATCGCATTCTGACCTCGGTAAAC and 3primer GGTCTAGAGGATCGTACCCGTAATCCG,and the product was cloned inpCASPER4. TheLexA:Gal4H-GAD portion was isolated from thepDppattB-LHGplasmid (44) and was inserted atthepCASPER4Not I site.

Tkv:Cherry: The Not IHind III fragment froma Tkv:GFP construct (4) was ligated to a mCherryfragment with 5Hind III and 3Kpn I sites in thepresence of pUAST that had been digested withNot I and Kpn I. Primers for mCherry amplification:5 primer, GCAAGCTTATGGTGAGCAAGGGC-GAGGAGG; 3primer, AGGTACCTTACTTG-TACAGCTCGTCCATGCCGC. Tkv:Cherry andTkv:GFP are similar in phenotype, activity, andlocalization in cytonemes.

In Situ Hybridization

RNA in situ hybridization was performed accord-ing to(50). Digoxigenin (DIG)labeled antisenseprobe was generated by transcription from a T7promoter joined to a 600base pair fragment ofdppcomplementary DNA (cDNA) amplified with

polymerase chain reaction primers: CAAGGAGGC-GCTCATCAAG andTTGTAATACGACTCACT-ATAGGGAGACACCAGCAGTCCGTAGTTGC.Alkaline phosphataseconjugated a-DIG antibody(Roche) was used to detect the DIG-labeled probe.

Immunohistochemistry

The following antisera were used:a-pMad [fromE. Laufer and P. ten Dijke; at 1:2000 ( 51)];a-dpERK (Sigma; 1:250) and a-apontic [from

R. Schuh (52)]; and a-discs large (4F3; 1:50),a-DE-cadherin (DECAD2; 1:20), and antib-galactosidase (Developmental Studies HybridomaBank). dpERK staining was carried out as de-scribed (9) with antibody obtained from CellSignaling Technology. Secondary antibodies wereconjugated to Alexa Fluor 488, 555, or 647. Toassay for cell lethality,a-cleaved caspase-3 (Asp175;Cell Signaling Technology) was used as de-scribed (53). Cell proliferation was monitoredwith a-phosphohistone H3 antibody (Ser10; CellSignaling Technology).

Imaging Techniques

Wing discs were dissected and mounted as de-scribed (5), except that the second small cover-slip was omitted. Images were made with a LeicaTCS SPE or TCS SP2 confocal microscope witheither 405, 488, 561, or 635 wavelength lasers andwith LAS-AF software; or with a custom-built Zeissspinning disc confocal with electron-multiplyingcharged-coupled device (EM-CCD) Hamamatsucamera (9100-13) and Volocity 5.5 software; orwith a standard Zeiss AxioPlan 2 fluorescencemicroscope with sensicam CCD camera (CookeCorporation) and SlideBook 4 acquisition soft-ware (Intelligent Imaging Innovations). Patternsof cytonemes were consistent in all three types

of systems. Brightfield images were made ona Leica DMR microscope equipped with SPOTCCD camera (Diagnostics Instruments) and SPOTacquisition software. Final images were analyzedand processed with National Institutes of Health(NIH) ImageJ.

References and Notes1. W. M. Gelbart, Thedecapentaplegicgene: A TGF-b

homologue controlling pattern formation in Drosophila.

Development107 (suppl.), 6574 (1989). pmid: 2699859

2. O. Wartlicket al., Dynamics of Dpp signaling and

proliferation control. Science 331, 11541159 (2011).

doi: 10.1126/science.1200037;pmid: 21385708

3. F. A. Ramrez-Weber, T. B. Kornberg, Cytonemes:

Cellular processes that project to the principal signaling

center in Drosophila imaginal discs. Cell 97,599607 (1999). doi: 10.1016/S0092-8674(00)80771-0;

pmid: 10367889

4. F. Hsiung, F. A. Ramirez-Weber, D. D. Iwaki, T. B. Kornberg,

Dependence of Drosophila wing imaginal disc cytonemes

on Decapentaplegic. Nature437, 560563 (2005).

doi: 10.1038/nature03951; pmid: 16177792

5. S. Roy, F. Hsiung, T. B. Kornberg, Specificity ofDrosophila

cytonemes for distinct signaling pathways. Science 332,

354358 (2011). doi: 10.1126/science.1198949;

pmid: 21493861

6. A. Bilioniet al., Balancing Hedgehog, a retention

and release equilibrium given by Dally, Ihog, Boi and

shifted/DmWif. Dev. Biol. 376, 198212 (2013).

doi: 10.1016/j.ydbio.2012.12.013;pmid: 23276604

7. A. Callejoet al., Dispatched mediates Hedgehog

basolateral release to form the long-range morphogene

gradient in theDrosophilawing disk epithelium. Proc. N

Acad. Sci. U.S.A.108, 1259112598 (2011). doi:10.10

pnas.1106881108; pmid: 21690386

8. P. Rojas-Ros, I. Guerrero, A. Gonzlez-Reyes,

Cytoneme-mediated delivery of Hedgehog regulates

the expression of bone morphogenetic proteins to

maintain germline stem cells in Drosophila.PLOS Biol.

e1001298 (2012). doi: 10.1371/journal.pbio.1001298;

pmid: 22509132

9. M. Sato, T. B. Kornberg, FGF is an essential mitogen

and chemoattractant for the air sacs of the Drosophi

tracheal system. Dev. Cell 3, 195207 (2002).

doi: 10.1016/S1534-5807(02)00202-2; pmid:121948

10. K. Tsuneizumiet al., Daughters against dpp modulat

dpp organizing activity in Drosophila wing developm

Nature389, 627631 (1997). doi: 10.1038/39362;

pmid:9335506

11. H. Tanimoto, S. Itoh, P. ten Dijke, T. Tabata, Hedgeh

creates a gradient of DPP activity in Drosophila wing

imaginal discs.Mol. Cell 5, 5 971 (2000). doi: 10.10

S1097-2765(00)80403-7 ;pmid:10678169

12. E. V. Entchev, A. Schwabedissen, M. Gonzlez-Gaitn

Gradient formation of the TGF-bhomolog Dpp. Cell

103, 981992 (2000). doi: 10.1016/S0092-

8674(00)00200-2 ;pmid: 11136982

13. A. A. Teleman, S. M. Cohen, Dpp gradient formation

the Drosophila wing imaginal disc. Cell 103, 9719

(2000). doi: 10.1016/S0092-8674(00)00199-9;

pmid:1113698114. See supplementary materials on Science Online.

15. M. Michel, I. Raabe, A. P. Kupinski, R. Prez-Palencia

C. Bkel, Local BMP receptor activation at adherens

junctions in theDrosophila germline stem cell niche

Nat. Commun. 2, 415 (2011). doi: 10.1038/

ncomms1426;pmid: 21811244

16. E. H. Feinberget al., GFP Reconstitution Across Syna

Partners (GRASP) defines cell contacts and synapses

in living nervous systems. Neuron 57, 353363 (200

doi: 10.1016/j.neuron.2007.11.030; pmid: 1825502

17. M. D. Gordon, K. Scott, Motor control in a Drosophil

taste circuit. Neuron 61, 373384 (2009).

doi: 10.1016/j.neuron.2008.12.033; pmid: 1921737

18. S. E. Ryu et al., Crystal structure of an HIV-binding

recombinant fragment of human CD4. Nature 348,

419426 (1990). doi: 10.1038/348419a0; pmid:2247

19. J. H. Wanget al., Atomic structure of a fragment

of human CD4 containing two immunoglobulin-like

domains. Nature 348 , 411418 (1990).

doi: 10.1038/348411a0;pmid:1701030

20. S. Cabantous, T. C. Terwilliger, G. S. Waldo, Protein

tagging and detection with engineered self-assemblin

fragments of green fluorescent protein. Nat. Biotech

23, 102107 (2005). doi: 10.1038/nbt1044;

pmid:15580262

21. D. H. Castrillon, S. A. Wasserman,diaphanousis requ

for cytokinesis in Drosophila and shares domains of

similarity with the products of the limb deformity ge

Development120 , 33673377 (1994). pmid: 78212

22. C. C. Homem, M. Peifer, Exploring the roles of

diaphanous and enabled activity in shaping the bala

between filopodia and lamellipodia. Mol. Biol. Cell 2

51385155 (2009). doi: 10.1091/mbc.E09-02-0144

pmid:19846663

23. T. Rousso, A. M. Shewan, K. E. Mostov, E. D. SchejteB.-Z. Shilo, Apical targeting of the formin Diaphano

inDrosophila tubular epithelia.eLife2, e00666 (2013

doi: 10.7554/eLife.00666; pmid: 23853710

24. T. Kitamoto, Conditional modification of behavior in

Drosophilaby targeted expression of a temperature-sens

shibire allele in defined neurons. J. Neurobiol. 47, 81(2001). doi: 10.1002/neu.1018; pmid:11291099

25. A. B. Muhlberg, D. E. Warnock, S. L. Schmid, Domai

structure and intramolecular regulation of dynamin

GTPase. EMBO J.16, 66766683 (1997). doi: 10.10

emboj/16.22.6676 ;pmid:9362482

26. J. E. Hinshaw, Dynamin spirals. Curr. Opin. Struct. Bi

9, 260267 (1999). doi: 10.1016/S0959-

440X(99)80036-0 ; pmid: 10322220

www.sciencemag.org SCIENCE VOL 343 21 FEBRUARY 2014 124462

RESEARCH AR

http://www.ncbi.nlm.nih.gov/pubmed/2699859http://dx.doi.org/10.1126/science.1200037http://dx.doi.org/10.1126/science.1200037http://www.ncbi.nlm.nih.gov/pubmed/21385708http://dx.doi.org/10.1016/S0092-8674(00)80771-0http://dx.doi.org/10.1016/S0092-8674(00)80771-0http://www.ncbi.nlm.nih.gov/pubmed/10367889http://dx.doi.org/10.1038/nature03951http://www.ncbi.nlm.nih.gov/pubmed/16177792http://dx.doi.org/10.1126/science.1198949http://dx.doi.org/10.1126/science.1198949http://www.ncbi.nlm.nih.gov/pubmed/21493861http://dx.doi.org/10.1016/j.ydbio.2012.12.013http://dx.doi.org/10.1016/j.ydbio.2012.12.013http://www.ncbi.nlm.nih.gov/pubmed/23276604http://dx.doi.org/10.1073/pnas.1106881108http://dx.doi.org/10.1073/pnas.1106881108http://www.ncbi.nlm.nih.gov/pubmed/21690386http://dx.doi.org/10.1371/journal.pbio.1001298http://dx.doi.org/10.1371/journal.pbio.1001298http://www.ncbi.nlm.nih.gov/pubmed/22509132http://dx.doi.org/10.1016/S1534-5807(02)00202-2http://www.ncbi.nlm.nih.gov/pubmed/12194851http://dx.doi.org/10.1038/39362http://dx.doi.org/10.1038/39362http://www.ncbi.nlm.nih.gov/pubmed/9335506http://dx.doi.org/10.1016/S1097-2765(00)80403-7http://dx.doi.org/10.1016/S1097-2765(00)80403-7http://dx.doi.org/10.1016/S1097-2765(00)80403-7http://www.ncbi.nlm.nih.gov/pubmed/10678169http://dx.doi.org/10.1016/S0092-8674(00)00200-2http://dx.doi.org/10.1016/S0092-8674(00)00200-2http://dx.doi.org/10.1016/S0092-8674(00)00200-2http://www.ncbi.nlm.nih.gov/pubmed/11136982http://dx.doi.org/10.1016/S0092-8674(00)00199-9http://dx.doi.org/10.1016/S0092-8674(00)00199-9http://www.ncbi.nlm.nih.gov/pubmed/11136981http://dx.doi.org/10.1038/ncomms1426http://dx.doi.org/10.1038/ncomms1426http://dx.doi.org/10.1038/ncomms1426http://www.ncbi.nlm.nih.gov/pubmed/21811244http://dx.doi.org/10.1016/j.neuron.2007.11.030http://dx.doi.org/10.1016/j.neuron.2007.11.030http://www.ncbi.nlm.nih.gov/pubmed/18255029http://dx.doi.org/10.1016/j.neuron.2008.12.033http://dx.doi.org/10.1016/j.neuron.2008.12.033http://www.ncbi.nlm.nih.gov/pubmed/19217375http://dx.doi.org/10.1038/348419a0http://www.ncbi.nlm.nih.gov/pubmed/2247146http://dx.doi.org/10.1038/348411a0http://dx.doi.org/10.1038/348411a0http://www.ncbi.nlm.nih.gov/pubmed/1701030http://dx.doi.org/10.1038/nbt1044http://dx.doi.org/10.1038/nbt1044http://www.ncbi.nlm.nih.gov/pubmed/15580262http://www.ncbi.nlm.nih.gov/pubmed/7821209http://dx.doi.org/10.1091/mbc.E09-02-0144http://www.ncbi.nlm.nih.gov/pubmed/19846663http://dx.doi.org/10.7554/eLife.00666http://www.ncbi.nlm.nih.gov/pubmed/23853710http://dx.doi.org/10.1002/neu.1018http://www.ncbi.nlm.nih.gov/pubmed/11291099http://dx.doi.org/10.1093/emboj/16.22.6676http://dx.doi.org/10.1093/emboj/16.22.6676http://dx.doi.org/10.1093/emboj/16.22.6676http://www.ncbi.nlm.nih.gov/pubmed/9362482http://dx.doi.org/10.1016/S0959-440X(99)80036-0http://dx.doi.org/10.1016/S0959-440X(99)80036-0http://www.ncbi.nlm.nih.gov/pubmed/10322220http://www.ncbi.nlm.nih.gov/pubmed/10322220http://dx.doi.org/10.1016/S0959-440X(99)80036-0http://dx.doi.org/10.1016/S0959-440X(99)80036-0http://www.ncbi.nlm.nih.gov/pubmed/9362482http://dx.doi.org/10.1093/emboj/16.22.6676http://dx.doi.org/10.1093/emboj/16.22.6676http://www.ncbi.nlm.nih.gov/pubmed/11291099http://dx.doi.org/10.1002/neu.1018http://www.ncbi.nlm.nih.gov/pubmed/23853710http://dx.doi.org/10.7554/eLife.00666http://www.ncbi.nlm.nih.gov/pubmed/19846663http://dx.doi.org/10.1091/mbc.E09-02-0144http://www.ncbi.nlm.nih.gov/pubmed/7821209http://www.ncbi.nlm.nih.gov/pubmed/15580262http://dx.doi.org/10.1038/nbt1044http://www.ncbi.nlm.nih.gov/pubmed/1701030http://dx.doi.org/10.1038/348411a0http://www.ncbi.nlm.nih.gov/pubmed/2247146http://dx.doi.org/10.1038/348419a0http://www.ncbi.nlm.nih.gov/pubmed/19217375http://dx.doi.org/10.1016/j.neuron.2008.12.033http://www.ncbi.nlm.nih.gov/pubmed/18255029http://dx.doi.org/10.1016/j.neuron.2007.11.030http://www.ncbi.nlm.nih.gov/pubmed/21811244http://dx.doi.org/10.1038/ncomms1426http://dx.doi.org/10.1038/ncomms1426http://www.ncbi.nlm.nih.gov/pubmed/11136981http://dx.doi.org/10.1016/S0092-8674(00)00199-9http://www.ncbi.nlm.nih.gov/pubmed/11136982http://dx.doi.org/10.1016/S0092-8674(00)00200-2http://dx.doi.org/10.1016/S0092-8674(00)00200-2http://www.ncbi.nlm.nih.gov/pubmed/10678169http://dx.doi.org/10.1016/S1097-2765(00)80403-7http://dx.doi.org/10.1016/S1097-2765(00)80403-7http://www.ncbi.nlm.nih.gov/pubmed/9335506http://dx.doi.org/10.1038/39362http://www.ncbi.nlm.nih.gov/pubmed/12194851http://dx.doi.org/10.1016/S1534-5807(02)00202-2http://www.ncbi.nlm.nih.gov/pubmed/22509132http://dx.doi.org/10.1371/journal.pbio.1001298http://www.ncbi.nlm.nih.gov/pubmed/21690386http://dx.doi.org/10.1073/pnas.1106881108http://dx.doi.org/10.1073/pnas.1106881108http://www.ncbi.nlm.nih.gov/pubmed/23276604http://dx.doi.org/10.1016/j.ydbio.2012.12.013http://www.ncbi.nlm.nih.gov/pubmed/21493861http://dx.doi.org/10.1126/science.1198949http://www.ncbi.nlm.nih.gov/pubmed/16177792http://dx.doi.org/10.1038/nature03951http://www.ncbi.nlm.nih.gov/pubmed/10367889http://dx.doi.org/10.1016/S0092-8674(00)80771-0http://www.ncbi.nlm.nih.gov/pubmed/21385708http://dx.doi.org/10.1126/science.1200037http://www.ncbi.nlm.nih.gov/pubmed/2699859

8/11/2019 Science 2014 Roy

12/12

27. E. M. Ennekinget al., Transsynaptic coordination of

synaptic growth, function, and stability by the L1-type

CAM Neuroglian. PLOS Biol. 11, e1001537 (2013).

doi: 10.1371/journal.pbio.1001537; pmid:23610557

28. M. Miln, U. Weihe, L. Prez, S. M. Cohen, The LRR

proteins Capricious and Tartan mediate cell interactions

during DV boundary formation in the Drosophila wing.

Cell 106, 785794 (2001). doi: 10.1016/S0092-

8674(01)00489-5 ;pmid: 11572783

29. W. Hong et al., Leucine-rich repeat transmembrane

proteins instruct discrete dendrite targeting in an

olfactory map. Nat. Neurosci. 12, 15421550 (2009).

doi: 10.1038/nn.2442; pmid: 19915565

30. H. Kohsaka, A. Nose, Target recognition at the tips of

postsynaptic filopodia: Accumulation and function of

Capricious. Development136 , 11271135 (2009).

doi: 10.1242/dev.027920;pmid: 19270171

31. E. Shishido, M. Takeichi, A. Nose,Drosophilasynapse

formation: Regulation by transmembrane protein with

Leu-rich repeats, CAPRICIOUS. Science 280, 21182121

(1998). doi: 10.1126/science.280.5372.2118;

pmid:9641918

32. H. Taniguchi, E. Shishido, M. Takeichi, A. Nose, Functional

dissection of Drosophilacapricious: Its novel roles in

neuronal pathfinding and selective synapse formation.

J. Neurobiol.42, 104116 (2000). doi: 10.1002/

(SICI)1097-4695(200001)42:13.0.

CO;2-V; pmid:10623905

33. D. S . Lidke, K. A. Lidke, B. Rieger, T. M. Jovin, D. J. Arndt-Jovin,

Reaching out for signals: Filopodia sense EGF and respond

by directed retrograde transport of activated receptors.J. Cell Biol.170, 619626 (2005). doi: 10.1083/

jcb.200503140; pmid: 16103229

34. K. Koizumi et al., RhoD activated by fibroblast growth

factor induces cytoneme-like cellular protrusions

through mDia3C. Mol. Biol. Cell 23, 46474661 (2012).

doi: 10.1091/mbc.E12-04-0315; pmid: 23034183

35. Z. Huang, S. Kunes, Hedgehog, transmitted along retinal

axons, triggers neurogenesis in the developing visual

centers of the Drosophila brain.Cell 86, 411422

(1996). doi: 10.1016/S0092-8674(00)80114-2;

pmid: 8756723

36. C. Korkut et al., Trans-synaptic transmission of vesicular

Wnt signals through Evi/Wntless. Cell 139 , 393404

(2009). doi: 10.1016/j.cell.2009.07.051; pmid:19837038

37. M. Cohen, M. Georgiou, N. L. Stevenson, M. Miodownik,

B. Baum, Dynamic filopodia transmit intermittent

Delta-Notch signaling to drive pattern refinement

during lateral inhibition. Dev. Cell 19, 7889 (2010).

doi: 10.1016/j.devcel.2010.06.006;pmid: 20643352

38. C. de Joussineauet al., Delta-promoted filopodia mediate

long-range lateral inhibition in Drosophila. Nature

426, 555559 (2003). doi: 10.1038/nature02157;

pmid: 14654840

39. O. Renaud, P. Simpson,scabrous modifies epithelial

cell adhesion and extends the range of lateral signalling

during development of the spaced bristle pattern

in Drosophila. Dev. Biol. 240, 361376 (2001).

doi: 10.1006/dbio.2001.0482;pmid: 11784069

40. Y. Peng, C. Han, J. D. Axelrod, Planar polarized

protrusions break the symmetry of EGFR signaling during

Drosophila bract cell fate induction. Dev. Cell 23 ,

507518 (2012). doi: 10.1016/j.devcel.2012.07.016;

pmid: 22921201

41. T. A. Sanders, E. Llagostera, M. Barna, Specialized

filopodia direct long-range transport of SHH during

vertebrate tissue patterning. Nature 497, 628632

(2013). doi: 10.1038/nature12157; pmid:23624372

42. N. Ninov et al., Dpp signaling directs cell motility and

invasiveness during epithelial morphogenesis. Curr. Biol.

20, 513520 (2010). doi: 10.1016/j.cub.2010.01.063;

pmid: 20226662

43. M. Reichman-Fried, B.-Z. Shilo, Breathless, a Drosophila

FGF receptor homolog, is required for the onset of

tracheal cell migration and tracheole formation.

Mech. Dev. 52, 265273 (1995). doi: 10.1016/

0925-4773(95)00407-R ;pmid:8541215

44. R. Yagi, F. Mayer, K. Basler, Refined LexA transactivatorsand their use in combination with the Drosophila

Gal4 system. Proc. Natl. Acad. Sci. U.S.A. 107,

1616616171 (2010). doi: 10.1073/pnas.1005957107;

pmid: 20805468

45. D. Nellen, R. Burke, G. Struhl, K. Basler, Direct and

long-range action of a DPP morphogen gradient. Cell

85, 357368 (1996). doi: 10.1016/S0092-

8674(00)81114-9 ;pmid:8616891

46. T. E. Haerry, O. Khalsa, M. B. OConnor, K. A. Wharton,

Synergistic signaling by two BMP ligands through the SAX

and TKV receptors controls wing growth and patterning in

Drosophila. Development125, 39773987 (1998).

pmid: 9735359

47. R. Massarwa, E. D. Schejter, B.-Z. Shilo, Apical secretion

in epithelial tubes of the Drosophila embryo is directed

by the Formin-family protein Diaphanous. Dev. Cell 16,

877888 (2009). doi: 10.1016/j.devcel.2009.04.010

pmid:19531358

48. T. Ohshiro, Y. Emori, K. Saigo, Ligand-dependent

activation of breathless FGF receptor gene in

Drosophila developing trachea. Mech. Dev. 114 , 31

(2002). doi: 10.1016/S0925-4773(02)00042-4;

pmid:12175485

49. C. Ribeiro, A. Ebner, M. Affolter, In vivo imaging rev

different cellular functions for FGF and Dpp signalin

in tracheal branching morphogenesis. Dev. Cell 2,

677683 (2002). doi: 10.1016/S1534-5807(02)00171

pmid: 12015974

50. A. Klebes, B. Biehs, F. Cifuentes, T. B. Kornberg, Expres

profiling of Drosophila imaginal discs. Genome Biol. 3

RESEARCH0038 (2002). doi: 10.1186/gb-2002-3-8-

research0038; pmid: 12186645

51. U. Persson et al., The L45 loop in type I receptors

for TGF-bfamily members is a critical determinant in

specifying Smad isoform activation. FEBS Lett. 434,

8387 (1998). doi: 10.1016/S0014-5793(98)00954-

pmid:9738456

52. K. G. Eulenberg, R. Schuh, The tracheae defective ge

encodes a bZIP protein that controls tracheal cell

movement during Drosophila embryogenesis. EMBO J.

71567165 (1997). doi: 10.1093/emboj/16.23.7156;

pmid:9384592

53. A. Guha, L. Lin, T. B. Kornberg, Organ renewal and

cell divisions by differentiated cells in Drosophila. Pr

Natl. Acad. Sci. U.S.A. 105, 1083210836 (2008).

doi: 10.1073/pnas.0805111105; pmid: 18664581

Acknowledgments:We thank K. Basler, M. Affolter, K. Sc

E. Laufer, P. ten Dijke, T. Kitamoto, J. Pielage, C. Bkel,

K. Wharton, A. Nose, L. Luo, M. Milan, B. Shilo, the BSC,

and the Developmental Studies Hybridoma Bank for reage

L. Lin and P. Rao for discussions; and M. Barna for access

the spinning disc confocal microscope. Funding: NIH

K99HL114867 to S.R. and GM030637 to T.B.K.

Supplementary Materialswww.sciencemag.org/content/343/6173/1244624/suppl/DC1

Figs. S1 to S3

Tables S1 to S5

Movie S1

13 August 2013; accepted 13 December 2013

Published online 2 January 2014;

10.1126/science.1244624

RESEARCH ARTICLE

http://dx.doi.org/10.1371/journal.pbio.1001537http://www.ncbi.nlm.nih.gov/pubmed/23610557http://dx.doi.org/10.1016/S0092-8674(01)00489-5http://dx.doi.org/10.1016/S0092-8674(01)00489-5http://dx.doi.org/10.1016/S0092-8674(01)00489-5http://www.ncbi.nlm.nih.gov/pubmed/11572783http://dx.doi.org/10.1038/nn.2442http://dx.doi.org/10.1038/nn.2442http://www.ncbi.nlm.nih.gov/pubmed/19915565http://dx.doi.org/10.1242/dev.027920http://dx.doi.org/10.1242/dev.027920http://www.ncbi.nlm.nih.gov/pubmed/19270171http://dx.doi.org/10.1126/science.280.5372.2118http://dx.doi.org/10.1126/science.280.5372.2118http://www.ncbi.nlm.nih.gov/pubmed/9641918http://dx.doi.org/10.1002/(SICI)1097-4695(200001)42:1%3C104::AID-NEU10%3E3.0.CO;2-Vhttp://dx.doi.org/10.1002/(SICI)1097-4695(200001)42:1%3C104::AID-NEU10%3E3.0.CO;2-Vhttp://dx.doi.org/10.1002/(SICI)1097-4695(200001)42:1%3C104::AID-NEU10%3E3.0.CO;2-Vhttp://www.ncbi.nlm.nih.gov/pubmed/10623905http://dx.doi.org/10.1083/jcb.200503140http://dx.doi.org/10.1083/jcb.200503140http://www.ncbi.nlm.nih.gov/pubmed/16103229http://dx.doi.org/10.1091/mbc.E12-04-0315http://www.ncbi.nlm.nih.gov/pubmed/23034183http://dx.doi.org/10.1016/S0092-8674(00)80114-2http://dx.doi.org/10.1016/S0092-8674(00)80114-2http://www.ncbi.nlm.nih.gov/pubmed/8756723http://dx.doi.org/10.1016/j.cell.2009.07.051http://www.ncbi.nlm.nih.gov/pubmed/19837038http://dx.doi.org/10.1016/j.devcel.2010.06.006http://dx.doi.org/10.1016/j.devcel.2010.06.006http://www.ncbi.nlm.nih.gov/pubmed/20643352http://dx.doi.org/10.1038/nature02157http://dx.doi.org/10.1038/nature02157http://www.ncbi.nlm.nih.gov/pubmed/14654840http://dx.doi.org/10.1006/dbio.2001.0482http://dx.doi.org/10.1006/dbio.2001.0482http://www.ncbi.nlm.nih.gov/pubmed/11784069http://dx.doi.org/10.1016/j.devcel.2012.07.016http://dx.doi.org/10.1016/j.devcel.2012.07.016http://www.ncbi.nlm.nih.gov/pubmed/22921201http://dx.doi.org/10.1038/nature12157http://dx.doi.org/10.1038/nature12157http://www.ncbi.nlm.nih.gov/pubmed/23624372http://dx.doi.org/10.1016/j.cub.2010.01.063http://dx.doi.org/10.1016/j.cub.2010.01.063http://www.ncbi.nlm.nih.gov/pubmed/20226662http://dx.doi.org/10.1016/0925-4773(95)00407-Rhttp://dx.doi.org/10.1016/0925-4773(95)00407-Rhttp://dx.doi.org/10.1016/0925-4773(95)00407-Rhttp://www.ncbi.nlm.nih.gov/pubmed/8541215http://dx.doi.org/10.1073/pnas.1005957107http://dx.doi.org/10.1073/pnas.1005957107http://www.ncbi.nlm.nih.gov/pubmed/20805468http://dx.doi.org/10.1016/S0092-8674(00)81114-9http://dx.doi.org/10.1016/S0092-8674(00)81114-9http://dx.doi.org/10.1016/S0092-8674(00)81114-9http://www.ncbi.nlm.nih.gov/pubmed/8616891http://www.ncbi.nlm.nih.gov/pubmed/9735359http://dx.doi.org/10.1016/j.devcel.2009.04.010http://www.ncbi.nlm.nih.gov/pubmed/19531358http://dx.doi.org/10.1016/S0925-4773(02)00042-4http://dx.doi.org/10.1016/S0925-4773(02)00042-4http://www.ncbi.nlm.nih.gov/pubmed/12175485http://dx.doi.org/10.1016/S1534-5807(02)00171-5http://www.ncbi.nlm.nih.gov/pubmed/12015974http://dx.doi.org/10.1186/gb-2002-3-8-research0038http://dx.doi.org/10.1186/gb-2002-3-8-research0038http://www.ncbi.nlm.nih.gov/pubmed/12186645http://dx.doi.org/10.1016/S0014-5793(98)00954-5http://www.ncbi.nlm.nih.gov/pubmed/9738456http://dx.doi.org/10.1093/emboj/16.23.7156http://www.ncbi.nlm.nih.gov/pubmed/9384592http://dx.doi.org/10.1073/pnas.0805111105http://dx.doi.org/10.1073/pnas.0805111105http://www.ncbi.nlm.nih.gov/pubmed/18664581http://www.sciencemag.org/content/343/6173/1244624/suppl/DC1http://www.sciencemag.org/content/343/6173/1244624/suppl/DC1http://www.ncbi.nlm.nih.gov/pubmed/18664581http://dx.doi.org/10.1073/pnas.0805111105http://www.ncbi.nlm.nih.gov/pubmed/9384592http://dx.doi.org/10.1093/emboj/16.23.7156http://www.ncbi.nlm.nih.gov/pubmed/9738456http://dx.doi.org/10.1016/S0014-5793(98)00954-5http://www.ncbi.nlm.nih.gov/pubmed/12186645http://dx.doi.org/10.1186/gb-2002-3-8-research0038http://dx.doi.org/10.1186/gb-2002-3-8-research0038http://www.ncbi.nlm.nih.gov/pubmed/12015974http://dx.doi.org/10.1016/S1534-5807(02)00171-5http://www.ncbi.nlm.nih.gov/pubmed/12175485http://dx.doi.org/10.1016/S0925-4773(02)00042-4http://www.ncbi.nlm.nih.gov/pubmed/19531358http://dx.doi.org/10.1016/j.devcel.2009.04.010http://www.ncbi.nlm.nih.gov/pubmed/9735359http://www.ncbi.nlm.nih.gov/pubmed/8616891http://dx.doi.org/10.1016/S0092-8674(00)81114-9http://dx.doi.org/10.1016/S0092-8674(00)81114-9http://www.ncbi.nlm.nih.gov/pubmed/20805468http://dx.doi.org/10.1073/pnas.1005957107http://www.ncbi.nlm.nih.gov/pubmed/8541215http://dx.doi.org/10.1016/0925-4773(95)00407-Rhttp://dx.doi.org/10.1016/0925-4773(95)00407-Rhttp://www.ncbi.nlm.nih.gov/pubmed/20226662http://dx.doi.org/10.1016/j.cub.2010.01.063http://www.ncbi.nlm.nih.gov/pubmed/23624372http://dx.doi.org/10.1038/nature12157http://www.ncbi.nlm.nih.gov/pubmed/22921201http://dx.doi.org/10.1016/j.devcel.2012.07.016http://www.ncbi.nlm.nih.gov/pubmed/11784069http://dx.doi.org/10.1006/dbio.2001.0482http://www.ncbi.nlm.nih.gov/pubmed/14654840http://dx.doi.org/10.1038/nature02157http://www.ncbi.nlm.nih.gov/pubmed/20643352http://dx.doi.org/10.1016/j.devcel.2010.06.006http://www.ncbi.nlm.nih.gov/pubmed/19837038http://dx.doi.org/10.1016/j.cell.2009.07.051http://www.ncbi.nlm.nih.gov/pubmed/8756723http://dx.doi.org/10.1016/S0092-8674(00)80114-2http://www.ncbi.nlm.nih.gov/pubmed/23034183http://dx.doi.org/10.1091/mbc.E12-04-0315http://www.ncbi.nlm.nih.gov/pubmed/16103229http://dx.doi.org/10.1083/jcb.200503140http://dx.doi.org/10.1083/jcb.200503140http://www.ncbi.nlm.nih.gov/pubmed/10623905http://dx.doi.org/10.1002/(SICI)1097-4695(200001)42:1%3C104::AID-NEU10%3E3.0.CO;2-Vhttp://dx.doi.org/10.1002/(SICI)1097-4695(200001)42:1%3C104::AID-NEU10%3E3.0.CO;2-Vhttp://dx.doi.org/10.1002/(SICI)1097-4695(200001)42:1%3C104::AID-NEU10%3E3.0.CO;2-Vhttp://www.ncbi.nlm.nih.gov/pubmed/9641918http://dx.doi.org/10.1126/science.280.5372.2118http://www.ncbi.nlm.nih.gov/pubmed/19270171http://dx.doi.org/10.1242/dev.027920http://www.ncbi.nlm.nih.gov/pubmed/19915565http://dx.doi.org/10.1038/nn.2442http://www.ncbi.nlm.nih.gov/pubmed/11572783http://dx.doi.org/10.1016/S0092-8674(01)00489-5http://dx.doi.org/10.1016/S0092-8674(01)00489-5http://www.ncbi.nlm.nih.gov/pubmed/23610557http://dx.doi.org/10.1371/journal.pbio.1001537