RESEARCH ARTICLE

Gain of Affinity Point Mutation in theSerotonin Receptor Gene 5-HT2Dro AcceleratesGermband Extension Movements DuringDrosophila GastrulationB. Schaerlinger,1,2 J.M. Launay,3,4 J.l. Vonesch,2 and L. Maroteaux2,5,6*

Serotonin (5-HT) not only works as a neurotransmitter in the nervous system, but also as a morphogeneticfactor during early embryogenesis. In Drosophila, a previous report showed that embryos that lack the5-HT2Dro receptor locus, display abnormal gastrulation movements. In this work, we screened for pointmutations in the 5-HT2Dro receptor gene. We identified one point mutation that generates a gain of serotoninaffinity for the receptor and affects germband extension: 5-HT2Dro

C1644. Embryos homozygous for this pointmutation display a fourfold increase in the maximal speed of ectodermal cell movements during the rapidphase of germband extension. Homozygous 5-HT2Dro

C1644 embryos present a cuticular phenotype, includinga total lack of denticle belt. Identification of this gain of function mutation shows the participation ofserotonin in the regulation of the cell speed movements during the germband extension and suggests a roleof serotonin in the regulation of cuticular formation during early embryogenesis. Developmental Dynamics236:991–999, 2007. © 2007 Wiley-Liss, Inc.

Key words: serotonin; receptor; mutation; ectoderm; gastrulation; cuticle; affinity

Accepted 4 February 2007

INTRODUCTIONIn vertebrates, the biogenic amine se-rotonin (5-hydroxytryptamine, 5-HT)affects a wide variety of central andperipheral functions involved in be-havioral and physiological activitiesin addition to embryonic development.Numerous receptor subtypes mediatethe many functions of 5-HT. At least14 different receptors have been mo-

lecularly characterized in mammals.All receptors except 5-HT3 belong tothe G protein coupled receptor familyand activate different signaling path-ways. In insects, biogenic aminesare similarly involved as neuromodu-lators and neurotransmitters, in addi-tion to their role in cross-linkingproteins and chitin during the sclero-tization of the cuticle (Wright, 1987).

In Drosophila, four 5-HT receptorshave been described (Witz et al., 1990;Saudou et al., 1992; Colas et al., 1995)in addition to the transporter (Park etal., 2006).

We previously reported that theG-protein coupled receptor 5-HT2Dro isan ortholog of the mammalian 5-HT2A,

B, C receptor subfamily. The expres-sion of 5-HT2Dro receptor mRNA and

The Supplementary Material referred to in this article can be found at http://www.interscience.wiley.com/jpages/1058-8388/suppmat1Univ Nancy, Faculte des sciences et techniques, Vandoeuvre-les-Nancy, France2CNRS UMR7104, Illkirch, France3Service de Biochimie, Hopital Lariboisiere, AP-HP, Paris, France4IFR139, EA3621, Paris, France5Inserm, U839, Paris, France6Universite Pierre et Marie Curie-Paris 6, Institut du Fer a Moulin, UMR-S0839, Paris, FranceGrant sponsors: Centre National de la Recherche Scientifique; Institut National de la Sante et de la Recherche Medicale; Universite LouisPasteur; Fondation pour la Recherche Medicale; Fondation de France; French Ministry of Research ACI and Agence Nationale pour leRecherche; Association pour la Recherche contre le Cancer.*Correspondence to: L. Maroteaux, INSERM, U616-U839; Hopital Pitie-Salpetriere, Universite Pierre et Marie Curie, BatPediatrie, 47 Bd de l�Hopital 75013 Paris, France. E-mail: luc.maroteaux@chups.jussieu.fr

DOI 10.1002/dvdy.21110Published online 20 March 2007 in Wiley InterScience (www.interscience.wiley.com).

DEVELOPMENTAL DYNAMICS 236:991–999, 2007

© 2007 Wiley-Liss, Inc.

protein starts after cellularization atthe blastoderm stage, when majorgastrulation movements begin. Fur-ther experiments demonstrated thatthe 5-HT2Dro receptor is functionalduring early developmental stages.Concomitant with 5-HT2Dro receptorexpression, there is also a detectabletransient peak of 5-HT synthesis(Colas et al., 1995). Moreover, the5-HT2Dro receptor mRNA is expressedin a pattern similar to the pair-rulegene Fushi tarazu (Pankratz andJackle, 1993). In contrast to otherpair-rule genes whose expression sur-rounds the embryo entirely, the sevenstripes of 5-HT2Dro mRNA are re-stricted to the presumptive ectoderm.

We previously demonstrated thatembryos homozygous for the defi-ciency Df(3R)HTRI, that removes the5-HT2Dro receptor locus, display ab-normal morphogenetic movementsduring the germband extension (GBE)and die at the end of embryogenesis orearly first-instar larval stage (Colas etal., 1999b). However, a contribution ofother genes present in the 55 kb ofDNA deleted in Df(3R)HTRI could notbe completely ruled out. In this work,we looked for a point mutant in the5-HT2Dro receptor gene to validate thefunctions of 5-HT in GBE process.

In wild-type gastrulae, peaks ofboth 5-HT2Dro receptor expression and5-HT synthesis coincide precisely withthe onset of GBE. We previously re-ported that the peak of 5-HT synthesisis dependent on the maternal deposi-tion of biopterins, cofactors of trypto-phan hydroxylase (the limiting en-zyme in 5-HT synthesis) and on thezygotic expression of both tryptophanhydroxylase and DOPA decarboxylase(Colas et al., 1999a). Mutant embryoswith an impaired peak of 5-HT syn-thesis display a particular cuticularorganization called “double line” (Co-las et al., 1999b), which was similarlyobserved in Df(3R)HTRI homozygousembryos and in shotgun embryos mu-tated in the DE-cadherin locus (Colaset al., 1999b). Whether the cuticularphenotype is a direct consequence of agastrulation defect remained to be de-termined.

GBE results in an approximate dou-bling in length of the anteroposterioraxis of the Drosophila embryo (Zallenand Wieschaus, 2004). As in other spe-cies, this process of convergent exten-

sion relies mostly on cell intercalation(Schoenwolf and Alvarez, 1989). Ir-vine and Wieschaus proposed that ec-todermal cell intercalation, the mainprocess in GBE, could be driven byadhesion strength, which would bedifferent in adjacent segments (Irvineand Wieschaus, 1994). Differences inadhesion strength between cells of al-ternate segments would control cellintercalation speed. In fact, cells donot delaminate or migrate across thetissue but, instead, rearrange specificcontacts after an ordered spatial–tem-poral pattern of junction remodeling.It has been proposed that axis elonga-tion in Drosophila occurs through ste-reotyped cell-shape changes across auniform field (Bertet et al., 2004).However, this model stipulates a sin-gle round of intercalation, whereasmultiple rounds are required for fullelongation. Recently, the asymmetricdistribution of cytoskeletal and junc-tional proteins has been proposed tocontribute directly to polarized cell be-havior during tissue elongation (Ber-tet et al., 2004; Zallen and Wieschaus,2004; Blankenship et al., 2006).

DE-cadherins localize to the apicalpart of the lateral membranes of ecto-dermal cells and are involved in celladhesion (Tepass et al., 1996; Oda etal., 1998; Greaves et al., 1999). DE-cadherin regulation seems to be im-portant for the GBE process. Zallenand Wieschaus have demonstratedthat striped expression of the pair-rule genes are both necessary and suf-ficient to orient planar polarity (Zallenand Wieschaus, 2004). The F-actinasymmetry was reported as the pri-mary planar polarity in the GBE, fol-lowed by the segregation of Myosin IIand Bazooka into complementary sur-face domains. DE-cadherin and Arma-dillo/beta-catenin would then be re-cruited to sites of new cell contactsbefore Bazooka association (Blanken-ship et al., 2006). The local destabili-zation of DE-cadherin–based adhe-sion indeed may cause the localdestabilization of polarized cell behav-ior during tissue elongation. Regula-tors of these phenomenon remain tobe identified.

In this work, we screened for muta-tions in the 5-HT2Dro receptor geneand identified one point mutant thatgenerates an increase in 5-HT affinityfor the receptor. Homozygous embryos

for this mutation present a defect incell speed movements during GBE ac-companied by cuticular defects.

RESULTS

l(3)82CDbC1644 Encodes aMutant 5-HT2Dro ReceptorWith Increased Affinity for5-HT

To understand the embryonic func-tions of 5-HT, we became interestedin mapping mutants in the 5-HT2Dro

receptor. The gene encoding thisreceptor is localized on the right armof the third chromosome at 82C5(Colas et al., 1995). Using EMS,Meagraw and Kaufman (FlyBase)screened several homozygous lethalmutants in the 82C4-D5 region. Fivemutants (l(3)82CDbC1644, l(3)82CDcC931,l(3)82CDdC142, l(3)82CDdC360, andl(3)82CDcl6) were mapped in this in-terval, because they all complementwith Df(3R)6-7 (82D5-82F3-6) but notwith Df(3R)110 (82C4-82F3). To checkwhether some of these point mutantsaffect the 5-HT2Dro receptor gene, werefined further the mapping usingDf(3R)HTRI (82C4-82D5) that de-letes 5-HT2Dro and the Df(3R)HTR6(82D2-82D5) that leaves the gene in-tact (Fig. 1A; Colas et al., 1999b).The l(3)82CDbC1644, l(3)82CDcC931,l(3)82CDdC142, and l(3)82CDdC360

were lethal with Df(3R)HTRI butviable with Df(3R)HTR6, indicatingthat they are in the region of interest.However, l(3)82CDcl6 that is viablewith Df(3R)HTRI, does not mapat the 5-HT2Dro locus and was notfurther analyzed. We found thatl(3)82CDdC142 and l(3)82CDdC360 de-fine a distinct complementation groupin 82C5, because they fail to comple-ment the lethal insertion j3A4 local-ized in the vicinity of 5-HT2Dro (Colaset al., 1999b).

We then focused on the two remain-ing mutants l(3)82CDbC1644 andl(3)82CDcC931. To assess whetherthey correspond to mutant alleles ofthe 5-HT2Dro gene, we extractedgenomic DNA from a single embryo tocheck by polymerase chain reaction(PCR) the presence of the balancerchromosome. We took advantage of areporter gene (twi-LacZ) of the bal-ancer chromosome to distinguish the

992 SCHAERLINGER ET AL.

homozygous mutant embryos from theothers. We sequenced all the codingexons of the 5-HT2Dro receptor gene of

homozygous l(3)82CDbC1644 andl(3)82CDcC931 embryos. The sequencerevealed that the l(3)82CDbC1644

chromosome had a mutation in thefirst coding exon of 5-HT2Dro gene(Exon2), while the l(3)82CDcC931 is-

Fig. 1. l(3)82CDbC1644 is an allele of 5-HT2Dro that encodes a P52S mutant receptor. A: The l(3)82CDbC1644, l(3)82CDcC931, l(3)82CDdC142,l(3)82CDdC360, and l(3)82CDcl6 complementation groups are represented. Their locations were ordered and refined by deletion mapping withDf(3R)HTRI and Df(3R)HTR6. Only l(3)82CDbC1644 and l(3)82CDcC931 mapped in the vicinity of 5-HT2Dro. B: The sequence of the second exon of the5-HT2Dro gene (first coding exon) is presented for wild-type (WT) and 5-HT2Dro

C1644 chromosomes. A C to T transition in l(3)82CDbC1644 changes theproline 52 into a serine within the long N-terminus domain of the receptor. C: Schematic drawing indicating the location of the point mutation in thereceptor molecule of the l(3)82CDbC1644 strain. TM, transmembrane domain.

POINT MUTATION IN 5-HT2DRO GENE 993

sued from the same genetic screen hada wild-type sequence (Fig. 1B). Thepoint mutation in l(3)82CDbC1644

changes the proline 52 into a serine inthe N-terminus of the receptor(Fig. 1B,C).

Until now, the role of the N-termi-nus part of the GPCR that recognizesbiogenic amine had not established. Inmammals, the N-terminus domain ofthe 5-HT receptors is not directly in-volved in ligand binding. Therefore, itwas intriguing that a mutation in thisdomain leads to homozygous lethalmutation. We investigated how thispoint mutation influenced the recep-tor function. We performed transientexpression of the mutated 5-HT2Dro

receptor in COS-1 cells and bindingexperiments on membrane prepara-tions. This experiment showed thatseveral of the ligands tested, including5-HT, the endogenous physiological li-gand, have an increased affinity forthe mutant receptor compared withwild-type receptor. The Ki of 5-HTwas 200 nM for the wild-type5-HT2Dro receptor and only 11 nM forthe mutant receptor, while expressionwas similar (Table 1). These data in-dicate that the mutation in the N-ter-minus domain of the 5-HT2Dro recep-tor confers to the receptor a gain of5-HT affinity.

5-HT2DroC1644 Homozygous

Embryos Display SpecificCuticular Alterations

The l(3)82CDbC1644 was EMS-inducedon a Rucuca chromosome that presentsseveral nonlethal mutations: ru h1 th stcu sr es ca. To analyze the phenotypeinduced by the mutation in the5-HT2Dro receptor, we first generated arecombinant chromosome that was onlymutated in the 5-HT2Dro gene. After re-combination of both arms of the thirdchromosomes, we verified that the re-combinant 5-HT2Dro

C1644 chromosomewas still lethal against the originall(3)82CDbC1644 strain and Df(3R)HTRI.This recombined 5-HT2Dro

C1644 chro-mosome was then used for all subse-quent experiments. We analyzed thetiming of the lethality in 5-HT2Dro

C1644

mutant embryos.To distinguish the phenotypes, we

used a green fluorescent protein(GFP) -labeled balancer chromosometo follow the fate of each embryo.With this technique, we observedthat approximately 20% of homozy-gous 5-HT2Dro

C1644 embryos hatchedand died shortly during the first in-star larvae (Table 2). The majority ofthe homozygous 5-HT2Dro

C1644 em-bryos died at the embryonic stages.We analyzed the cuticular pheno-type of the embryos that never

hatched in the 5-HT2DroC1644 strain

(Fig. 2; Table 2). We verified that thehomozygous balancer embryos dis-played mainly wild-type cuticles (Ta-ble 2). The 5-HT2Dro

C1644 homozy-gous embryos presented either“double line” or “ghost” phenotypes(Fig. 2C,D, 2E,F; Table 2). Among theunhatched homozygous 5-HT2Dro

C1644

embryos, 43% present wild-type cuticle,30% double-line and 22% ghost cuticles.The “double line” phenotype is charac-terized by the presence of two rows ofdenticles in each segment (Fig. 2E,F)and has been previously observed in ho-mozygous Df(3R)HTRI embryos thatlack the 5-HT2Dro receptor (Colas et al.,1999b). Moreover, we observed that the“ghost” phenotype with no denticle atthe surface of the embryo was mainlyrepresented in 5-HT2Dro

C1644 embryos(Fig. 2C,D; Table 2). This phenotypecould be distinguished from the nonfer-tilized embryos as the cuticle had beensecreted. Interestingly, the transhet-erozygous 5-HT2Dro

C1644/Df(3R)HTRIembryos hatched but led to small larvaethat all died quickly with no obviouscuticular defects, showing a partial cu-ticular rescue. These data revealed thatthe 5-HT2Dro

C1644 mutation signifi-cantly affects cuticular development.

5-HT2DroC1644 Embryos

Display Altered Cell SpeedMovements During GBE

A previous study showed thatDf(3R)HTRI homozygous embryospresented defects in GBE (Colas etal., 1999b). We, thus, investigatedthe GBE in 5-HT2Dro

C1644 homozy-gous and heterozygous mutant em-bryos. Following the previously de-scribed method (Colas et al., 1999b),we video recorded the rapid phase ofthe GBE and followed the move-ments of ectodermal cells around themidline of single embryos. After re-cording, each embryo was PCR geno-typed. The absence of the balancerchromosome marker was indicativeof homozygous mutant chromosome.Using image analysis of the digitalvideo recording, we assessed thespeed of cell movements, near theventral and dorsal midline of the em-bryo. This analysis was made by us-ing a software developed to allow the“peeling off” of a single superficiallayer of pixels from each recorded

TABLE 1. 5-HT2Dro Pharmacological Propertiesa

Transfection 5-HT2Dro cDNA 5-HT2DroC1644 cDNA

AgonistsDexfenfluramine 17 � 1 5.0 � 0.3#

5-HT 200 � 16 11 � 1#

�-methyl 5-HT 420 � 23 370 � 185-CT 1700 � 72 28 000 � 900*Tryptamin 2040 � 83 1950 � 62

AntagonistsRitanserin 8.1 � 1.1 7.0 � 0.8Ketanserin 40 � 5 8.1 � 0.7#

Mesulergine 78 � 6 71 � 5Bufotenine 200 � 12 1350 � 45*Yohimbine 310 � 18 340 � 16

aAffinity determined as competition of [125I]DOI is expressed as Ki � SEM(nanomolar) for each compound for wild-type 5-HT2Dro or mutant 5-HT2Dro

C1644

receptor cDNA expressed transiently in COS-1 cells. #Significant increase comparedwith wild-type; *significant decrease compared with wild-type; P � 0.05, n � 9. Aftertransfection of identical amounts of plasmid encoding cDNAs, the expression levelsof wild-type and mutant receptors were not significantly different (similar Bmax).5-HT, 5-hydroxytryptamine; 5-CT, carboxytryptamine.

994 SCHAERLINGER ET AL.

video image, then laying the layerflat after opening from the most ros-tral point. Accumulating these lay-ers over time allowed the visualiza-tion and quantification of the speedand extent of individual cell move-ment at the surface of the embryoand the evaluation of the synchroni-zation of these movements. Taking

the initiation of extension move-ments as a reference time point, thespeed and synchronization of the ex-tension movements can be directlyevaluated. Figure 3 displays repre-sentative results of these investiga-tions on an individual genotyped em-bryo. As the X-axis represents theposition of the cell along the embryo

and the Y-axis the time scale, theslope of the curves represents theirspeed. We followed the migration ofectodermal cells ventral to the polecells because they were easy to dis-tinguish in each of the embryos (Fig.3D). Of interest, we observed an ini-tial contraction of the ventral ecto-derm, reflecting the mesoderm in-vagination in wild-type embryos,before ectodermal cells start moving(Fig. 3A), which could not be de-tected in 5-HT2Dro

C1644 homozygousembryos (Fig. 3B). For cells close tothe pole cells, the highest cell speed(V2) was recorded in 5-HT2Dro

C1644

homozygous embryos and was fourtimes higher than in the wild-typeembryos (V2 � 12.5 � 1.8, n � 6 vs.3.3 � 0.5 �m/min, n � 6; Fig. 3E).Calculating the speed of similar cellsin 5-HT2Dro

C1644/TM3 heterozygousembryos (V2 � 5.9�0.3 �m/min, n �5) showed that it was only two timesfaster (Fig. 3E). The data show that5-HT and 5-HT2Dro receptors are in-volved in cell speed regulation dur-ing the GBE. Moreover, the interme-diary phenotype observed for theheterozygous 5-HT2Dro

C1644/TM3embryos supports the notion that themutation leads to a gain of functionfor the 5-HT2Dro receptor.

Locally Overexpressing 5-HT2Dro Receptor EmbryosDisplay Altered Cell SpeedMovements During GBE

A transgenic strain that overexpressesa wild-type 5-HT2Dro receptor underthe control of the Kruppel promoterhas been previously produced (Kr-Gal4,UAS-5-HT2Dro; Colas et al.,1999b). This transgenic strain dis-plays an overexpression of the wild-type 5-HT2Dro receptor in the centralregion of the embryos, where the en-dogenous receptor expression is low(Colas et al., 1999b). Cell speed calcu-lation in Kr-Gal4,UAS-5-HT2Dro em-bryos during GBE showed that themaximal speed was similar to het-erozygous 5-HT2Dro

C1644/TM3 em-bryos (V2 � 5.7 � 0.7, n � 3 vs. 5.9 �0.3 �m/min, n � 5; Fig. 3C–E). Thesedata reveal that defects of GBE in het-erozygous 5-HT2Dro

C1644 mutant em-bryos can be phenocopied by 5-HT2Dro

local overexpression.

Fig. 2. Cuticular phenotype of 5-HT2DroC1644 mutant embryos. A,B: Cuticular phenotype of wild-

type embryos. C,D: “Ghost” cuticular phenotype of 5-HT2DroC1644 embryos. “Ghost” embryos

present no denticle but clear differentiated cuticle. E,F: “Double line” phenotype of 5-HT2DroC1644

embryos. Note that in each segment, there are only two rows of denticles. Scale bar � 80 �m inA,C,E, 20 �m in B,D,F.

TABLE 2. Cuticular Phenotypic Classes of Nonhatched Embryos (%)a

5-HT2DroC1644 5-HT2Dro

C1644/Df(3R)HTRI l(3)82CDcC931

n 204 208 487Wild-type 79 96 95Ghost 8 4 1Double line 11 0 2Others 2 0 2

aCuticles were prepared as described in the Experimental Procedures section andobserved with a brightfield microscope. The fly strains were maintained with thesame TM3 balancer chromosome. The l(3)82CDcC931 embryos were used as control ofthe homozygous balancer phenotype because homozygous l(3)82CDcC931 are latelarval lethal. Transheterozygous 5-HT2Dro

C1644/Df(3R)HTRI and homozygousl(3)82CDcC931embryos are hatching as weak first-instar larvae that are dying withapparent wild-type cuticle. Part (62.5% for 5-HT2Dro

C1644) and the near totality(5-HT2Dro

C1644/Df(3R)HTRI and l(3)82CDcC931) of the nonhatched embryoscorrespond to homozygous TM3 balancer chromosome on the same geneticbackground. The percentage of homozygous 5-HT2Dro

C1644 embryos that hatch anddie as first-instar larvae is, thus, approximately 20%, and the percentage ofunhatched homozygous 5-HT2Dro

C1644 embryos presenting wild-type, double-line, orghost cuticular phenotypes is approximately 43, 30, and 22%, respectively. Using agreen fluorescent protein-labeled balancer chromosomes, we confirmed that the“ghost” phenotype represented approximately one fourth of all 5-HT2Dro

C1644

homozygous embryos. Wild-type, “ghost,” and “double line” phenotypes are describedin Figure 2. Values are presented as percentage of embryos with a cuticularphenotype in relation to those that hatch 72 hr after egg laying, and n correspondsto the total number of embryos counted.

POINT MUTATION IN 5-HT2DRO GENE 995

DISCUSSION

Mutation in the N-terminalDomain of 5-HT2Dro ReceptorIs a Gain of 5-HT Affinity

We identified a lethal mutation in aDrosophila strain characterized by apoint mutation in the 5-HT2Dro recep-tor gene that changes proline 52 toserine in the N-terminal domain of thereceptor, increasing 20-fold its affinityfor 5-HT. The complete deletion of theN-terminal domain of the 5-HT2Dro re-ceptor also led to a gain of 5-HT affin-ity (Colas et al., 1997). These data in-dicate that the N-terminal domain ofthe 5-HT2Dro receptor is involved inthe binding of 5-HT. Classically, bio-genic amines including 5-HT bind toamino acid side chains within thetransmembrane domains of their re-ceptors (Manivet et al., 2002). How-ever, in Drosophila, most biogenicamine receptors have a long N-termi-nal domain (Onai et al., 1989; Saudouet al., 1990; Witz et al., 1990; Colas etal., 1995; Gotzes and Baumann, 1996).Few studies have been performed tounderstand the role of these long do-mains. Our study demonstrates forthe first time a role for the long N-terminal domain of the serotonergicreceptor 5-HT2Dro. Because this mu-tant receptor is more sensitive for itsendogenous ligand 5-HT, 20-fold less5-HT could be sufficient to reach sig-naling activity in the mutant receptor,comparable to that of the fully stimu-lated wild-type receptor.

Mutant 5-HT2Dro ReceptorGBE Phenotype Can BePhenocopied by Wild-Type5-HT2Dro Receptor EctopicExpression

5-HT2DroC1644 homozygous mutant

embryos display an increase in GBEmovements compared with wild-type.As previously shown, 5-HT is tran-siently synthesized in Drosophila em-bryos, with a peak of synthesis at 3 hr15 min after fertilization (Colas et al.,1999a). Because the mutated 5-HT2Dro

receptor found in the 5-HT2DroC1644

strain could be activated by 20-foldless 5-HT, it could, thus, reach athreshold of activity earlier than wild-type receptor, have stronger signal-ing, and/or its activity may last longer.

Fig. 3. Quantification of cell movements during germband extension (GBE) in wild-type, 5-HT2DroC1644,

and Kr-Gal4, UAS-5-HT2Dro embryos. The gastrulation of embryos was video recorded, and image analysisallowed the direct visualization and quantification of ectodermal cell movements along their relative syn-chronization The X-axis is the flat projection of a pixel layer at the periphery of the embryos for each videoframe, that is, at a determinate time. This process allows the visualization of the location of ectodermal cellspresent around the midline of the embryo before cell movements start (one pixel layer corresponds to onepicture). As we recorded one picture every 15 seconds, we accumulated all flat projections along the Y-axis(thus, being the time scale). As a consequence, each line on the figures corresponds to the trajectory of onecell along the periphery of the embryos. We analyzed three to five ectodermal cells directly close to the polecells (bold lines) to calculate the cell speed movements from the beginning of the GBE. One pixel corre-sponds to 0.2 �m on the X-axis and to 15 sec on the Y-axis (see Supplementary Material). A: On arepresentative wild-type (w1118) embryo, mesoderm invagination is first visualized by slight ventral ectodermcontraction before the ectodermal extension starts (time t � 0 min). Then, ventral ectodermal cells begin tomove from the anterior to the posterior position. The cephalic furrow (cf) can easily be observed. Cellssurrounding the pole cells (PC) correspond to the ventral (VE) and dorsal (DE) endoderm that will form theposterior midgut after invagination. B: In the video recording of a representative 5-HT2Dro

C1644 homozy-gous mutant embryo, ectodermal cells start their extension before contractions triggered by mesoder-mal invagination could be observed. C: A representative Kr-Gal4,UAS-5-HT2Dro embryo displays milderdefects than homozygous 5-HT2Dro

C1644 embryos. The bold line starting initially close to the ventralendoderm (VE) corresponds to the cell tracing used to calculate speed presented in E. D: In an embryothat undergoes GBE, A, B, C, D, and E correspond to the position of the cell at the time of a change inits speed (Vi) during elongation. These points are noted on the X-axis of the graph on the right. E: Inwild-type (w1118) embryos, at the beginning of GBE, cells move slowly and accelerate until pole cellsinvaginate and then cell speed decreases until posterior cells reach the dorsal cephalic furrow; V2

corresponds to the maximal speed of the cell. The asterisk corresponds to significantly different speedsfrom wild-type (P � 0.05). All embryos have been individually polymerase chain reaction genotypedafter video recording using balancer chromosome markers. False colors correspond to different graylevels of the recorded images. See Supplementary Material for the corresponding video recordings.

996 SCHAERLINGER ET AL.

The 5-HT2DroC1644 homozygous em-

bryos seem to respond to an early signalinducing cell movements, because GBEstarts before detectable mesodermal in-vagination. However, the precise mea-sure of the time elapsed between fertil-ization and the onset of GBE is hardlyaccessible, thus, limiting the possibilityto document this putative time shift.

A transgenic fly strain with a re-stricted ectopic expression of 5-HT2Dro

cDNA under the Kruppel gene pro-moter (Kr-Gal4, UAS-5-HT2Dro) alsogenerates gastrulation defects, sup-porting a role of 5-HT signaling inGBE (Colas et al., 1999b). This trans-genic strain displays (1) an overex-pression of the 5-HT2Dro receptor inthe central region of the embryos,where the endogenous receptor ex-pression is low with a local disruptionof the receptor segmental expressionand (2) a change in the timing of the5-HT2Dro receptor expression as Krup-pel is expressed before 5-HT2Dro. InKr-Gal4, UAS-5-HT2Dro embryos, thefirst defects appear as abnormal ante-rior initiation of the extension move-ments before the posterior initiation ofmovements (Colas et al., 1999b). Cellmovement studies during GBE showthat cell speed defects of heterozygous5-HT2Dro

C1644/TM3, and Kr-Gal4,UAS-5-HT2Dro embryos are similar.The cuticular phenotype of Kr-Gal4,UAS-5-HT2Dro embryos is weaker than5-HT2Dro

C1644. Indeed, they display ei-ther wild-type or mild denticle alter-ations, like heterozygous 5-HT2Dro

C1644/TM3 embryos. Moreover, the Kr-Gal4,UAS-5-HT2Dro strain is viable, indi-cating that the overexpression of thereceptor mimics the phenotype of5-HT2Dro

C1644/TM3 embryos. The phe-notype of homozygous point mutant is,thus, stronger than the overexpressionof the wild-type 5-HT2Dro

C1644. The dif-ference between homozygous and het-erozygous 5-HT2Dro

C1644 embryos andthe similar phenotype of heterozygous5-HT2Dro

C1644 and Kr-Gal4, UAS-5-HT2Dro strongly support the notion thatthe mutation in 5-HT2Dro confers a gainof function to the receptor with respectto GBE movements. In mutants de-pleted of 5-HT (synthesis mutant; Colaset al., 1999a) or of 5-HT2Dro receptor{Df(3R)HTRI} (Colas et al., 1999b), al-though altered, GBE movements stilloccur. The 5-HT and the 5-HT2Dro re-ceptor seem, thus, to be involved in reg-

ulating the speed of the cell movementsduring the GBE process but apparentlynot the GBE or mesodermal invagina-tion movements themselves.

GBE Defects in 5-HT2DroC1644

Could Be Due to CellAdhesion Defects

The change in cell speed in the5-HT2Dro receptor mutant most prob-ably reflects alterations in ectodermalcell intercalation that was initiallyproposed to depend upon the differ-ence of adhesion strength betweenneighboring cells (Irvine and Wie-schaus, 1994). Cell intercalation of ep-ithelial cells would be driven by sim-ple remodeling of contacts in anordered directional pattern, allowingprogressive exchange of places withneighboring cells (Bertet et al., 2004;Zallen and Wieschaus, 2004). In thismodel, axis elongation in Drosophilaoccurs through stereotyped cell-shapechanges across a uniform field (Bertetet al., 2004). However, multiplerounds are required for full elonga-tion, and intercalating cells locally or-ganize to generate multicellular struc-tures (rosettes) that form and resolvein a directional manner. Directional-ity is disrupted in mutants that lackanteroposterior patterning and fail toelongate. Multicellular structures,and not individual cells or cell inter-faces, represent, thus, the functionalunits of cell behavior during tissueelongation. The asymmetric distribu-tion of cytoskeletal and junctional pro-teins has been proposed to contributedirectly to polarized cell behavior dur-ing tissue elongation (Blankenship etal., 2006). F-actin enrichment at an-teroposterior interfaces is the earliestevidence of planar polarity. Myosin IIsubsequently accumulates at these in-terfaces and may coordinate the con-traction of linked edges to drive mul-ticellular (rosette) formation. Thesetransient structures resolve in a direc-tional manner to create contact be-tween cells that were previously sep-arated along the dorso–ventral axis.DE-cadherin association is an earlystep in new contact formation and co-incides with formation of a transientF-actin structure. Bazooka may be re-cruited later to new interfaces duringtheir eventual stabilization (Blanken-ship et al., 2006).

We previously reported that the5-HT2Dro receptor is expressed with apair-rule–like pattern in ectodermalcells and that the apicobasal localiza-tion of Armadillo (�-catenin ortholog),in the ectodermal cells, that links cad-herins with the cytoskeleton is alteredin homozygous Df(3R)HTRI embryos(Colas et al., 1999b). Moreover, GBEdefects are detected both in the5-HT2Dro

C1644 gain of function strainand in the Df(3R)HTRI-deficientstrain. The 5-HT by means of the5-HT2Dro receptor appears, thus, to beable to influence the polarized adhe-sion signals controlling the GBE pro-cess. Because DE-cadherin and Arma-dillo/�-catenin are recruited to sites ofnew cell contacts before Bazooka asso-ciation (Blankenship et al., 2006), the5-HT–dependent DE-cadherin–basedadhesion may, thus, regulate the sta-bilization/destabilization of polarizedcell behavior during tissue elongation.

Putative Link Between GBEDefects and 5-HT2Dro

Receptor

In embryos injected with Y-27632, apharmacological inhibitor of Rho-ki-nase and specific regulator of Myosin-II, intercalation fails to occur (Bertetet al., 2004). Of interest, the 5-HT2Dro

receptor is orthologous to mammalian5-HT2A, B, C receptors, which are in-volved in smooth muscle cell contrac-tion by means of a Rho-dependent reg-ulation of myosin light chain kinase(Nishikawa et al., 2003). The 5-HT2Dro

receptor may, thus, transduce extra-cellular signal (5-HT) to the cytoskel-eton and regulate either the timing orintensity of cell intercalation, Rho/Rho kinase being a likely effector ofthe 5-HT2Dro receptor, although theprecise transduction of this pathwayremains to be determined in Drosophila.

Many 5-HT2DroC1644 homozygous

embryos that did not hatch display a“ghost” phenotype that is nearly ab-sent in other strains studied. The pre-viously described “ghost” and “doubleline” cuticular phenotypes inDf(3R)HTRI (Colas et al., 1999b) havebeen proposed to result from varia-tions in intensity of GBE defects, the“ghost” phenotype reflecting earlierand/or stronger defects. In bothstrains, alterations of differential ad-hesion between cell expressing, or not,

POINT MUTATION IN 5-HT2DRO GENE 997

the 5-HT2Dro receptor may explain thecuticular phenotypes. Of interest, byin situ hybridization, 5-HT2Dro probeshowed no 5-HT2Dro mRNA misex-pression in 5-HT2Dro

C1644 mutant em-bryos and by immunohistochemistry,we observed no defects in Fushitarazu or Engrailed, expression in5-HT2Dro

C1644 embryos at differentdevelopmental stages (data notshown). These observations supportan indirect role of 5-HT2Dro in cuticu-lar formation. It was noticed that cellspeed appears more variable in5-HT2Dro

C1644 homozygous embryos(higher SEM; Fig. 3D), suggestinghigher variability in the control ofGBE. Together, these observationssupport the notion that this gain offunction mutation in the 5-HT2Dro re-ceptor increases variability in the em-bryonic development and that 5-HT islikely involved in setting the timeand/or speed of the GBE process bypolarizing adhesion signals by meansof the 5-HT2Dro receptor transduction.

EXPERIMENTALPROCEDURES

Fly Stocks

The following strains have been usedin this study. Point mutant lines:l(3)82CDbC1644 Stocks 0004830,l(3)82CDcC931 Stocks 0004831,l(3)82CDdC142 Stocks 0004832,l(3)82CDdC360 Stocks 0004833,l(3)82CDcl6 Stocks 0004839. Deficiencylines: Df(3R)HTRI, Df(3R)HTR6,Df(3R)110, Df(3R)6-7 (Colas et al.,1999b). Blue balancers: TM3 Sb SerP[hb-lacZ] and TM3 Ser P[twi-lacZ];Green balancers: TM3 Sb Ser P[hb-GFP].

To obtain 5-HT2DroC1644 strain, we

used kar2ry506, ru,h1,th,st,cu,sr,es,Pr,ca(Ruprica), Ki,kar2ry506, D1,kar2 ry506,and w1118 background chromosomes.

We performed the crossings as fol-lows to obtain the 5-HT2Dro

C1644

strain:(1) Recombination of the left arm of

the third chromosome: l(3)82CDbC1644/TM3 Ser and kar2ry506/kar2ry506generated l(3)82CDbC1644/kar2ry506;l(3)82CDbC1644/kar2ry506 and Ru-prica/TM6 Bri,Tb generated recom-bined l(3)82CDbC1644*/Ruprica; andrecombined l(3)82CDbC1644*/Ruprica

and Df(3R)HTRI/TM3 Ser, Sb gener-ated l(3)82CDbC1644*/TM3 Ser, Sb.

(2) Preliminary crossing to preparethe second recombination: pr, pwnFLP38/CyO; Ki, kar2ry506 andkar2ry506/kar2ry506 generated �/CyO;Ki, kar2ry506/kar2ry506.

(3) Recombination of the right arm ofthe third chromosome: l(3)82CDbC1644*/TM3 Ser, Sb X and �/CyO ; Ki, kar2ry-506/kar2ry506 generated l(3)82CD-bC1644*/Ki, kar2ry506; l(3)82CDbC1644*/Ki, kar2ry506 and D, kar2ry506/TM3 Ser,Sb generated recombined l(3)82CD-bC1644**/D, kar2ry506; l(3)82CDbC1644**/D, kar2ry506 and Df(3R)HTRI/TM3 Sergeneratedl(3)82CDbC1644**kar2ry506/TM3 Ser.

This stock was used in this study as5-HT2Dro

C1644.

DNA Extraction, PCR,Sequencing

Unique embryos have been genotypedby PCR. After collection of the em-bryos (from 3 to 5 hr), DNA was ex-tracted as described (Colas et al.,1999b) and 1/10 of the DNA from asingle embryo was amplified usingregular PCR amplification conditionsfor 38 cycles. To identify homozygousmutants, embryos were genotyped byPCR using oligonucleotides for Atonal(5�CGTTGGATCCAGCAACATAACA-CCACCATA3� and 5�ATCGGGATC-CGGGAATTCAGCGCAGCAATC3�)or twist-�-galactosidase (5�GCGT-CAGTTGCGTTCCGTAAGTGC3� and5�CAGTTTGAGGGGACGACGACA-GTA3�) genes.

To identify point mutations in ho-mozygous embryos, each exon of the5-HT2Dro receptor had been amplifiedby PCR using the following amplimers:exon 1, 5�GGAATTCCAAAACAAGTA-TTGTTGATGCTG3� and 5�GGGGTAC-CTTATCTACTATATTTTACTTGC3�;exon 2, 3 and 4, 5�GGGGTACCAAT-TTTTTTTGAGATTCACATATA3� and5�GGGGTACCGCAGTAAGTCACC-ACCACGTAATT3�; exon 5, 5�GGT-GGAAAATTATTCGAAAGT3� and5�GCGATTCACCACCAAAATAAT-TTA3�; and exon 6, 5�GGGGTACCC-TTAACTTTTTGGTACTACTAATA3�and 5�GGGGTACCTGGCACTGGAT-CCGCGTCTGGGCT3�.

Embryos, whose DNA was not am-plified by twist �-galactosidase am-plimers but by atonal (a marker of the

third chromosome) primers, were con-sidered as homozygous for a mutationon the third chromosome. Indeed,twist �-galactosidase is present on thebalancer chromosome. This techniqueallowed the unambiguous genotypingof all individual embryos. PCR prod-ucts of the 5-HT2Dro receptor were pu-rified with the Geneclean kit and di-rectly sequenced.

Binding Experiment

[125I]DOI was used as a radioligand toinvestigate the pharmacological prop-erties of membrane fractions from5-HT2DroR transiently transfectedCOS-1 cells (Colas et al., 1995).Briefly, cell membranes were pre-pared by four cycles of homogeniza-tion (Brinkman P10 disrupter) andcentrifugation (48,000 g for 15 min).The assay was established to achievesteady state conditions and to opti-mize specific binding (Kellermann etal., 1996). Fifty micrograms of mem-brane proteins were incubated with 5nM [125I]DOI at 4°C for 60 min. Theincubation medium (200 �l) contained50 �l of radioligand, 50 �l of buffer orof competing drug, and 100 �l of mem-brane suspension (protein concentra-tion, 50 �g/ml). The mixture was incu-bated at 30°C for 30 min. Assays wereterminated by vacuum filtrationthrough glass fiber filters (GF/B),which had been pretreated with 0.1%polyethyleneimine, followed by fourwashes of 5 ml of ice-cold buffer. Fil-ters were dried rapidly, and their ra-dioactivity was determined by liquidscintillation counting. Nonspecificbinding, determined in the presence of10 �M unlabeled DOI, representedapproximately 30% of total binding.Competition studies for [125I]DOIbinding were performed by adding in-creasing concentrations of test drug tothe reaction. Data were analyzed us-ing GraphPad software. Pharmacolog-ical profiles of transiently expressedreceptors were determined over threeindependent experiments performedin triplicate.

Cuticle Preparation

Embryos were laid on an agar plateand observed at 25°C, 60% humidity.The percentage of hatching larvae wasevaluated after 24, 48, and 72 hr

998 SCHAERLINGER ET AL.

(25°C) after egg laying and expressedas the percentage of nonhatched em-bryos in relation to the total numberembryos and larvae. After 72 hr, thenumber of nonhatched embryos wascounted. Approximately 25% ofl(3)82CDcC931 embryos (larval lethalon the same chromosome), 25% of5-HT2Dro

C1644/Df(3R)HTRI, and 40%of 5-HT2Dro

C1644 embryos died beforehatching. Cuticles from nonhatchedembryos were prepared as describedby Colas et al. (1999b). Percentages inTable 2 correspond to the proportionof a phenotypic class in relation to thetotal number of nonhatched embryos.

Time-Lapse VideoExperiment

Flies laid on apple juice agar for 1 hr,aged for 2 hr, then overlaid with halo-carbon 3S oil (Prolabo Voltalef) werestaged according to Campos-Ortega andHartenstein (1985). The development ofrandomly selected single embryos wasvideo recorded for 1 hr (one picture ev-ery 15 sec, i.e., 241 pictures/hr) using aLeica MZ12 binocular and digital Sonycamera under transmitted light. A sur-rounding superficial pixel layer of eachpicture of the video-recorded embryoswas selected and, after opening at themost rostral point and flattening, accu-mulated over time (for the detailed pro-cedure, see Supplementary Material,which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). This strategy allows a directvisualization of the movements of indi-vidual cells initially located around themidline, the quantification of theirspeed during gastrulation movements,and their relative synchronization(speed � slope of the curve). We fol-lowed a set of ventral ectodermal cellsclose to the invaginating endoderm be-cause they can be easily distinguish oneach figures. We illustrated the differ-ent speed of individual cells (V1 to V4)as changes in the slope of the curve. Themain inflexion point of the curve, themost reliably identifiable point, corre-sponds to V2. Each embryo was individ-ually PCR-genotyped after video record-ing using balancer chromosomes asmarkers (Colas et al., 1999b). False col-ors correspond to different gray levels ofthe recorded images. Statistical signifi-cance was assessed by analysis of vari-ance and Neuman–Keul’s post hoc test.

ACKNOWLEDGMENTSWe are indebted to the Drosophilastocks centers of Bloomington, Szeged,Exelixis, and Umea, Drs. A. Spradling,J. O’Donnell, J. Roote, S. DiNardo, P.Heitzler, and G. Reuter for providing flystrains. We acknowledge M. Nullans forfly cares and D. Hentsch for help withvideo recordings and image analysis.We thank Drs. V. Setola and P. Heitzlerfor critical reading of the manuscript,and for helpful discussions. B.S. re-ceived a Region Alsace fellowship.

REFERENCES

Bertet C, Sulak L, Lecuit T. 2004. Myosin-dependent junction remodelling controlsplanar cell intercalation and axis elonga-tion. Nature 429:667–671.

Blankenship JT, Backovic ST, Sanny JS,Weitz O, Zallen JA. 2006. Multicellularrosette formation links planar cell polar-ity to tissue morphogenesis. Dev Cell 11:459–470.

Campos-Ortega JA, Hartenstein V. 1985.The embryonic development of Drosoph-ila melanogaster. Berlin: Springer-Ver-lag.

Colas JF, Launay JM, Kellermann O, Ro-say P, Maroteaux L. 1995. Drosophila5-HT2 serotonin receptor: coexpressionwith fushi-tarazu during segmentation.Proc Natl Acad Sci U S A 92:5441–5445.

Colas JF, Choi DS, Launay JM, MaroteauxL. 1997. Evolutionary conservation ofthe 5-HT2B receptors. Ann N Y Acad Sci812:148–150.

Colas JF, Launay J-M, Maroteaux L.1999a. Maternal and zygotic control ofserotonin biosynthesis are both neces-sary for Drosophila germband extension.Mech Dev 87:67–76.

Colas JF, Launay J-M, Vonesch J-L, HickelP, Maroteaux L. 1999b. Serotonin syn-chronises convergent extension of ecto-derm with morphogenetic gastrulationmovements in Drosophila. Mech Dev 87:77–91.

Gotzes F, Baumann A. 1996. Functionalproperties of Drosophila dopamine D1-receptors are not altered by the size ofthe N-terminus. Biochem Biophys ResCommun 222:121–126.

Greaves S, Sanson B, White P, Vincent JP.1999. A screen for identifying genes in-teracting with armadillo, the Drosophilahomolog of beta-catenin. Genetics 153:1753–1766.

Irvine K, Wieschaus E. 1994. Cell interca-lation during Drosophila germband ex-tension and its regulation by pair-rulesegmentation genes. Development 120:827–841.

Kellermann O, Loric S, Maroteaux L, Lau-nay J-M. 1996. Sequential onset of three5-HT receptors during the 5-hy-droxytryptaminergic differentiation ofthe murine 1C11 cell line. Br J Pharma-col 118:1161–1170.

Manivet P, Schneider B, Smith JC, ChoiDS, Maroteaux L, Kellermann O, Lau-nay JM. 2002. The serotonin binding siteof human and murine 5-HT2B receptors:molecular modeling and site-directedmutagenesis. J Biol Chem 277:17170–17178.

Nishikawa Y, Doi M, Koji T, Watanabe M,Kimura S, Kawasaki S, Ogawa A, SasakiK. 2003. The role of rho and rho-dependentkinase in serotonin-induced contractionobserved in bovine middle cerebral artery.Tohoku J Exp Med 201:239–249.

Oda H, Tsukita S, Takeichi M. 1998. Dy-namic behavior of the cadherin-based cell-cell adhesion system during Drosophilagastrulation. Dev Biol 203:435–450.

Onai T, Fitzgerald MG, Arakawa S, Go-cayne JD, Urquhart DA, Hall LM,Fraser CM, McCobbie WR, Venter JC.1989. Cloning, sequence analysis andchromosome localization of a Drosophilamuscarinic acetylcholine receptor. FEBSLett 255:219–225.

Pankratz MJ, Jackle H. 1993. Blastodermsegmentation. In: Bate M, Martinez-Arias A, editors. The development ofDrosophila melanogaster. Cold SpringHarbor, NY: Cold Spring Harbor Labora-tory Press. p 467–516.

Park SK, George R, Cai Y, Chang HY,Krantz DE, Friggi-Grelin F, Birman S,Hirsh J. 2006. Cell-type-specific limita-tion on in vivo serotonin storage follow-ing ectopic expression of the Drosophilaserotonin transporter, dSERT. J Neuro-biol 66:452–462.

Saudou F, Amlaiky N, Plassat J-L, BorrelliE, Hen R. 1990. Cloning and characteri-sation of a Drosophila tyramine receptor.EMBO J 9:3611–3617.

Saudou F, Boschert U, Amlaiky N, PlassatJL, Hen R. 1992. A Family of Drosophilaserotonin receptors with distinct intra-cellular signalling properties and expres-sion patterns. EMBO J 11:7–17.

Schoenwolf GC, Alvarez IS. 1989. Roles ofneuroepithelial cell rearrangement anddivision in shaping of the avian neuralplate. Development 106:427–439.

Tepass U, Gruszynski-DeFeo E, Haag TA,Omatyar L, Torok T, Hartenstein V.1996. shotgun encodes Drosophila E-cad-herin and is preferentially required dur-ing cell rearrangement in the neurecto-derm and other morphogenetically activeepithelia. Genes Dev 10:672–685.

Witz P, Amlaiky N, Plassat J-L, MaroteauxL, Borrelli E, Hen R. 1990. Cloning andcharacterization of a Drosophila seroto-nin receptor that activates adenylate cy-clase. Proc Natl Acad Sci U S A 87:8940–8944.

Wright TR. 1987. The genetics of biogenicamine metabolism, sclerotization, andmelanization in Drosophila melano-gaster. Adv Genet 24:127–222.

Zallen JA, Wieschaus E. 2004. Patternedgene expression directs bipolar planarpolarity in Drosophila. Dev Cell 6:343–355.

POINT MUTATION IN 5-HT2DRO GENE 999