Current Biology 16, 1895–1904, October 10, 2006 ª2006 Elsevier Ltd All rights reserved DOI 10.1016/j.cub.2006.08.057

ArticleThe bantam MicroRNA Is a Targetof the Hippo Tumor-Suppressor Pathway

Riitta Nolo,1 Clayton M. Morrison,1,3 Chunyao Tao,1

Xinwei Zhang,1 and Georg Halder1,2,3,*1Department of Biochemistry and Molecular Biology2Program in Genes and DevelopmentUniversity of Texas M. D. Anderson Cancer CenterHouston, Texas 770303Program in Developmental BiologyBaylor College of MedicineHouston, Texas 77030

Summary

Background: The Hippo tumor-suppressor pathwayhas emerged as a key signaling pathway that controlstissue size in Drosophila. Hippo signaling restricts tissuesize by promoting apoptosis and cell-cycle arrest, andanimals carrying clones of cells mutant for hippo de-velop severely overgrown adult structures. The Hippopathway is thought to exert its effects by modulatinggene expression through the phosphorylation of thetranscriptional coactivator Yorkie. However, how Yorkieregulates growth, and thus the identities of downstreamtarget genes that mediate the effects of Hippo signaling,are largely unknown.Results: Here, we report that the bantam microRNA isa downstream target of the Hippo signaling pathway.In common with Hippo signaling, the bantam microRNAcontrols tissue size by regulating cell proliferation andapoptosis. We found that hippo mutant cells had ele-vated levels of bantam activity and that bantam wasrequired for Yorkie-driven overgrowth. Additionally,overexpression of bantam was sufficient to rescuegrowth defects of yorkie mutant cells and to suppressthe cell death induced by Hippo hyperactivation. Hipporegulates bantam independently of cyclin E and diap1,two other Hippo targets, and overexpression of bantammimics overgrowth phenotypes of hippo mutant cells.Conclusions: Our data indicate that bantam is an es-sential target of the Hippo signaling pathway to regulatecell proliferation, cell death, and thus tissue size.

Introduction

During normal development, the number of cells ingrowing tissues is controlled through regulation of cellproliferation and apoptosis [1–3]. In Drosophila, theHippo (Hpo) tumor-suppressor pathway controls thesize of adult tissues by coordinately regulating theseprocesses [4–14]. Hpo signaling restricts tissue growthby promoting termination of cell proliferation and bystimulating apoptosis during development. hpo mutanttissues thus develop into severely overgrown adultstructures because mutant cells continue to proliferateafter organs reach their normal size and because mutant

*Correspondence: ghalder@mdanderson.org

cells are resistant to proapoptotic signals that normallyeliminate extra cells.

Several components of the Hpo pathway have beenidentified, including the membrane-associated 4.1, ez-rin, radixin, moesin (FERM)-domain-containing proteinsExpanded (Ex) and Merlin (Mer) [13]; the serine/threo-nine kinases Hippo (Hpo) and Warts (Wts) [4–8, 15, 16];the adaptor molecule Salvador (Sav), which binds toHpo [9, 10]; Mats (Mob as a tumor suppressor) [11],a regulator of Wts activity; and Yorkie (Yki), a transcrip-tional coactivator [12]. Genetic and biochemical experi-ments have indicated that Ex and Mer act upstream ofHpo, which in turn regulates the activity of the Warts(Wts) kinase [4, 6, 8, 13]. Wts together with its cofactorMats then suppresses the transcriptional activity ofYki, possibly through phosphorylation [11, 12]. Ex,Mer, Hpo, Sav, Wts, and Mats are negative regulatorsof growth, and mutations in these genes result in dra-matically overgrown tissues containing an excess num-ber of cells. Yki, on the other hand, is a positive regulatorof growth, and overexpression of Yki causes severeovergrowths that resemble the loss-of-function pheno-types of the other pathway members, whereas cellsmutant for yki grow poorly [12]. Yki is required for theovergrowth of hpo or wts mutant cells in vivo, and itwas thus postulated that Yki mediates most if not all ofthe growth effects of Hpo signaling, presumably by driv-ing the expression of transcriptional target genes [12].When Hpo signaling is reduced, for example by muta-tions in hpo, Yki is hypophosphorylated and active, driv-ing the expression of target genes that promote cell pro-liferation and suppress apoptosis, resulting in tissueovergrowth. Normally, the Hpo and Wts kinases limitthe growth of tissues through the inactivation of Yki.

Although the last few years have witnessed the identi-fication of several components of the Hpo signal trans-duction pathway, we have only limited information aboutthe transcriptional target genes that mediate the effectsof Hpo signaling. We and others have previously re-ported that cells mutant for hpo cell-autonomously up-regulate the transcription of cyclin E and diap1 (Dro-sophila inhibitor of apoptosis protein-1) [4–13]. CyclinE is a limiting factor for S phase entry in imaginal-disccells [17], and Cyclin E overexpression is sufficient todrive ectopic cell divisions [17, 18]. DIAP1 is an antia-poptotic protein, and extra DIAP1 likely protects cellsfrom apoptosis induced during development [19]. How-ever, these two target genes alone are insufficient toaccount for the hpo mutant phenotype because overex-pression of Cyclin E together with apoptosis inhibitiondoes not cause tissue overgrowth [18]. Thus, critical tar-gets of Hpo signaling remain to be identified.

In this report, we present evidence that the bantammicroRNA (miRNA) is a key target of the Hpo signalingpathway. The bantam miRNA is a positive growth regu-lator, and tissues that overexpress bantam are largerthan wild-type tissues [20–22]. Conversely, animalswith mutations that inactivate bantam are smaller than

Current Biology1896

normal [20–22]. bantam-induced tissue growth iscaused by an increase in cell number as a result of an in-crease in the rate of cell proliferation and defects in ap-optosis [20–22]. Unlike Ras, Myc, and Insulin receptorsignaling, which primarily affect cellular growth andcell size [1, 23–25], bantam specifically affects cell num-ber without affecting cell size, cell differentiation, or pat-tern formation [20–22]. bantam thus acts as a balancedgrowth regulator, coordinately driving cell growth andcell-cycle progression, so that bantam-overexpressingcells proliferate faster but keep their normal size [20,22]. Similar to bantam, Cyclin D-Cdk4 control tissuesize by balanced regulation of cell growth and cell divi-sion, so that cells overexpressing Cyclin D and Cdk4have an increased rate of proliferation but no changesin cell size [26, 27]. In contrast to bantam, however, Cy-clin D-Cdk4 do not regulate apoptosis, and double-mu-tant and epistasis analyses showed that bantam and Cy-clin D-Cdk4 act in parallel pathways [20]. Interestingly,the effects of bantam overexpression resemble the phe-notypes of mutations in hpo that also show acceleratedcell proliferation through balanced growth and reducedapoptosis [4–13]. The striking similarities between theeffects of bantam and Hpo signaling prompted us to in-vestigate their relationship during development.

Results

Hippo Signaling Regulates the Expression

of the bantam miRNATo test whether the activity of the bantam miRNA wasregulated by Hpo signaling, we made use of a GFP ban-tam sensor that reports the spatial activity of bantam[21]. This bantam sensor expresses GFP under the con-trol of a ubiquitously active tubulin promoter and hastwo perfect bantam target sites in its 30 UTR. When pres-ent, the bantam miRNA reduces GFP expressionthrough its RNAi effect [21]. The expression pattern ofGFP is thus a negative image of the activity pattern ofthe bantam miRNA. In third-instar wing imaginal discs,the bantam sensor is expressed in a complex patternwith higher levels along the presumptive wing margin,in the anterior compartment along the anteroposteriorcompartment boundary, and in several patches in thethorax region (Figure 1A, [21]). A control sensor thatlacks the bantam target sites is expressed at muchhigher levels overall, indicating that the bantam miRNAis expressed throughout the disc, although at varyinglevels (data not shown, [21]). Overexpression of the ban-tam miRNA in the developing wing eliminated the GFPexpression of the bantam sensor in the correspondingregion, demonstrating that the expression of GFP is in-deed under the control of bantam (Figure 1B, [21]). In de-veloping eye discs, the bantam sensor is also broadlyexpressed, with higher levels in differentiating photore-ceptor cells (Figure 1C). As in wing discs, overexpres-sion of bantam downregulated GFP expression in eyediscs (not shown). The bantam sensor thus reflects theactivity of the bantam miRNA in eye and wing discs [21].

To address whether Hpo signaling regulates the activ-ity of the bantam miRNA, we monitored GFP expressionof the bantam sensor in imaginal discs that had defectsin Hpo signaling. We found that hpo or wts mutantcells had lower levels of bantam-sensor-driven GFP

expression throughout the mutant clones (Figures 1D–1F). Significantly, hpo and wts mutant clones showedlower levels of GFP in multiple tissues, including thewing, antenna, and eye imaginal discs. In eye imaginaldiscs, wts clones affected the bantam sensor anteriorto the morphogenetic furrow, where cells are still un-committed as well as posterior to the furrow in differen-tiating photoreceptor cells (Figure 1F). In all cases, theregulation of the bantam sensor was cell autonomous(Figures 1D–1F). In addition, wing imaginal discs thatoverexpressed Yki had lower levels of bantam sensorexpression in the entire region of Yki overexpression(Figure 1G). The control sensor, which lacks the bantamtarget sites, was not affected by overexpression of Ykior in wts mutant clones (not shown). In summary, weconclude that Hpo signaling generally regulates bantamexpression in multiple imaginal discs and cell types.

Because bantam expression is strongly correlatedwith cell proliferation [21], the regulation of bantam byHpo may not be a specific effect of Hpo signaling butmay be a secondary consequence of the extra growthand proliferation induced in hpo mutant cells. We thustested whether bantam was also regulated by other reg-ulators of cell growth and cell proliferation. We assayedthe expression of GFP from the bantam sensor in wingdiscs that overexpressed Cyclin D-Cdk4, Myc, orRasV12 in the posterior compartment or that carriedclones of cells mutant for TSC1, manipulations knownto cause overgrowth and overproliferation [26, 28–30].We found no significant change of GFP expressionfrom the bantam sensor in cells overexpressing CyclinD-Cdk4, Myc, or RasV12 (Figures 2A, 2B, and 2D), andTSC1 mutant cells even had higher levels of bantam sen-sor expression, indicating that bantam activity was sup-pressed in TSC1 mutant clones (Figure 2C). Promotingcellular growth and accelerating cell division, throughoverexpression of Cyclin D-Cdk4, Myc, or RasV12 orloss of TSC1, was therefore not sufficient to induce ban-tam expression. The regulation of bantam expression byHpo signaling is thus not simply the consequence of thegrowth and proliferation induced in hpo mutant cells,but is a specific downstream effect of Hpo signaling.

bantam Is Required for Yorkie-Driven Overgrowth

The finding that Hpo regulates the activity of bantam ir-respective of the type of imaginal disc and broadlywithin discs suggested that bantam is an importantdownstream target of Hpo signaling. Hpo signalingmay thus restrict tissue size at least in part by suppress-ing the expression of bantam. When Hpo signaling is de-fective, for example in tissues overexpressing Yki, ban-tam expression is induced, which may be required forthe overgrowth of these tissues. To test this hypothesis,we asked whether bantam is required for overexpressedYki to induce tissue overgrowth in the retina. Wild-typepupal retinae show exquisitely precise hexagonal arraysof the different cell types, and the retina has served asa sensitive model system to analyze the effects of Hpoand other growth-control pathways on cell numberand apoptosis (Figure 3A, [4, 6–10, 12, 13, 31, 32]). Over-expression of Yki in the developing eye resulted in thegeneration of excess interommatidial pigment and bris-tle cells, resembling the phenotypes observed in hpoand wts loss-of-function clones (Figure 3B, [4–10, 12,

Hippo Signaling Targets the bantam miRNA1897

Figure 1. Hippo Signaling Regulates the Ac-

tivity of the bantam Sensor in Eye and Wing

Imaginal Discs

Confocal images of wing (A, B, D, E, G, and H)

and eye-antennal (C and F) imaginal discs

from third-instar larvae.

(A) Expression of the bantam sensor in a wild-

type wing imaginal disc.

(B) bantam-sensor expression in a wing

imaginal disc overexpressing bantam in the

posterior compartment (asterisk) by the en-

Gal4 driver.

(C) bantam sensor is expressed ubiquitously

in eye-antennal discs with higher levels in dif-

ferentiating photoreceptor cells (arrow).

(D–F00) Expression of the bantam sensor (red

in [D], [E], and [F] and separately shown in

gray in [D0], [E0], and [F0]) in wing and eye

imaginal discs that have hpo and wts mutant

clones (arrowheads) marked by the absence

of bGal expression (green in [D], [E], and [F]

and gray in [D00], [E00], and [F00]).

(G–G00) bantam-sensor expression in a wing

disc that overexpressed Yki in the posterior

compartment under the control of en-Gal4

(asterisks). The anterior compartment is

marked by the expression of Ci (green in [G]

and gray in [G00]). Arrows point to the A/P

compartment boundary.

(H–H00) bantam-sensor expression in a wing

disc overexpressing Hpo in clones of cells

marked by Hpo staining (green in [H], gray

in [H00]). The bantam sensor is induced in

Hpo-expressing cells (arrowheads). Anterior

is to the left in all panels.

13]). Loss of bantam suppressed the induction of extrainterommatidial cells in the retina caused by Yki overex-pression, whereas it did not affect the pattern of cells inotherwise wild-type retinae (Figures 3A and 3B). Tomeasure the number of interommatidial pigment cells,we defined regions corresponding to one side of an

ommatidial hexagon between two vertices and thencounted the cells in these regions. We found that suchregions overexpressing Yki had 5.9 6 1.0 interommati-dial cells, whereas regions that were also mutant forbantam had only about half as many cells (3.1 6 0.5cells) (Figure 3). Removal of bantam thus significantly

Current Biology1898

Figure 2. Cyclin D-Cdk4, Myc, and Ras Do

Not Regulate bantam, and TSC1 Negatively

Regulates bantam Activity

Confocal images showing the expression of

the bantam sensor (red in [A], [B], [C], and

[D] and gray in [A0], [B0], [C0], and [D0]) in

wing imaginal discs.

(A and B) Discs with en-Gal4 driving expres-

sion of Cyclin D-Cdk4 or Myc in the posterior

compartments (green in [A] and [B] and gray

in [A00] and [B00]).

(C) Disc containing TSC1 clones marked by

the absence of bGal expression (green in [C]

and gray in [C00]).

(D) Disc with en-Gal4 driving expression of

RasV12. The anterior compartment is labeled

with a-Ci antibody (green in [D] and gray in

[D00]). Arrowheads point to the A/P compart-

ment boundaries. Arrow points to TSC1 mu-

tant clone. Anterior is to the left and ventral

is up in all panels.

reduced the number of extra interommatidial cells in ret-inae that overexpressed Yki, which was, however, stillhigher than wild-type retinae, which have 1.67 cells inthe corresponding regions (2 3 1/3 of a tertiary pigmentcell at the vertex and 1 secondary pigment cell). bantamis thus required for Yki-induced cell proliferation, andthe loss of bantam limits the amount of tissue growththat can be induced by Yki overexpression.

bantam Expression Rescues Apoptosis and Growth

Defects in Cells with Increased Hippo SignalingNext, we wanted to test whether overexpression of ban-tam was sufficient to rescue the apoptosis and growthdefects induced by hyperactivation of Hpo signaling.Hyperactivation of Hpo signaling, for example by over-expression of Hpo, causes phenotypes opposite tothose observed in hpo mutants: It suppresses cell prolif-eration and induces cell death, resulting in smaller thannormal adult structures [4–6, 8, 13]. The level of Hpo sig-naling is thus in a balance in wild-type cells, and it can beup- or downregulated. This balance is reflected at thelevel of bantam expression. yki mutant cells or clonesof cells with hyperactivated Hpo signaling due to over-expressed Hpo cell-autonomously induced bantam

sensor expression (Figure 1H, data not shown), indicat-ing that the expression of bantam was suppressed inthese cells. Together with the induction of bantam inhpo mutant cells, these results indicate that in wild-type cells, the levels of bantam expression can be pos-itively or negatively modulated by the Hpo pathway.

If suppression of bantam expression by Hpo is impor-tant for hyperactivated Hpo signaling to induce apopto-sis and suppress growth, then re-expression of bantammay suppress the phenotypes caused by hyperactiva-tion of Hpo signaling. To test this hypothesis, we usedHpo gain-of-function phenotypes in the eye and wing.Overexpression of Hpo during eye development inducesapoptosis and results in small and rough eyes (Figures4A and 4B, [4, 6, 8]). Coexpression of bantam signifi-cantly rescued the eye defects caused by overexpres-sion of Hpo (Figures 4A–4D) and eliminated the Hpo-in-duced apoptosis (Figures 4B00 and 4C00). Similarly,coexpression of bantam rescued the defects causedby overexpression of Ex in developing eyes (not shownand [20]). We extended these experiments to thewing, where overexpression of Hpo similarly causesa strong reduction in tissue size (Figures 4E and 4H,[4]). Coexpression of bantam rescued the small-wing

Hippo Signaling Targets the bantam miRNA1899

Figure 3. bantam Is Required for Yorkie-

Driven Overgrowth in the Retina

(A and B) Panels show mid-pupal retinae

stained with Discs large (Dlg) antibodies to vi-

sualize cell outlines.

(A and A0) Retina with bantam mutant clones.

(B and B0) Retina with bantam mutant clones

that overexpressed Yki in the entire retina by

using the GMR-Gal4 driver. Mutant clones are

marked by the absence of GFP expression

(red/green in [A] and [B]). Cell outlines from

the bantam mutant cells were colored yellow

in [A0] and [B0].

(C) Quantification of cell numbers in pupal

retinae. The number of interommatidial cells

in regions corresponding to one side of an

ommatidial hexagon between two vertices

were counted. Wild-type retinae or retinae

with bantam mutant clones had 1.67 cells in

such regions, whereas retinae overexpress-

ing Yki had 5.9 6 1.0 interommatidial cells

(n = 10), and regions that overexpressed Yki

but were mutant for bantam had 3.1 6 0.5

cells (n = 8).

Error bars represent standard deviations.

phenotype caused by overexpressed Hpo and the re-sulting wings had nearly normal size (Figures 4H and4I). In addition, loss of one copy of the bantam geneenhanced the small-wing phenotype of overexpressedHpo (Figures 4H and 4J). Removing one copy of bantamalso enhanced the eye phenotypes caused by overex-pression of Ex (not shown and [20]), which also hyperac-tivates Hpo signaling [13]. We conclude that the Hpo-induced phenotypes are sensitive to bantam levelsbecause removing one copy of bantam enhances,whereas expressing bantam suppresses, Hpo-inducedphenotypes.

Hyperactivation of Hpo signaling leads to induction ofthe proapoptotic gene head involution defective (hid) [4].Similarly, yki mutant cells had slightly higher levels of Hid(Figure S1A in the Supplemental Data available online).bantam has been shown to suppress the translation ofhid mRNA and thus to downregulate Hid protein levels[21]. Indeed, expression of bantam also suppressedthe elevated levels of Hid in yki mutant cells (Figure S1B).bantam may thus suppress cell death in Hpo-overex-pressing or yki mutant cells at least in part through thesuppression of Hid expression.

bantam Expression Rescues Growth Defects

of yorkie Mutant CellsNext, we asked whether overexpression of bantamcould also rescue the growth defects of yki mutant cells.To test this hypothesis, we did two sets of experiments.First, we induced yki mutant clones in wing discs thatoverexpressed bantam in the posterior compartment.We found that yki mutant clones in the anterior compart-ment grew poorly and were very small (Figure 5A, arrow).In contrast, yki mutant clones in the posterior (bantam-expressing) compartment grew to much larger sizes(Figure 5A, arrowhead). bantam expression thus en-hanced the growth of yki mutant cells.

In the second set of experiments, we induced bantamexpression specifically in yki mutant clones and com-pared their sizes with those of their correspondingtwin clones. Because the wild-type twin clone is bornin the same cell division as the mutant clone, the sizeof the twin clone is a good normalizing measurementfor the growth rate of the cells in the mutant clone. Wefound that in developing wing discs, clones of cells mu-tant for yki grew poorly and only rarely produced clonesthat contained more than a couple of cells, even after

Current Biology1900

Figure 4. bantam Expression Rescues Hippo

Hyperactivation Phenotypes

(A)–(D) show SEM images of eyes of adult

flies and (A0)–(D0) show higher-magnification

images of the panels above them.

(A00)–(D00) show third-instar eye discs stained

with a-Drice to mark apoptotic cells. The ge-

notypes of the animals are indicated above

the panels. GMR-Gal4 drives overexpression

of UAS transgenes in the eye. Expression of

bantam suppresses the reduced- and

rough-eye phenotype as well as the apopto-

sis caused by overexpression of Hpo.

(E–J) Wings of adult flies of the indicated ge-

notypes. The expression of bantam causes

wing enlargement (F) and rescues the small-

wing phenotype caused by expression of

Hpo (compare [H] and [I]). Removal of one

copy of the bantam gene enhances the

small-wing phenotype caused by overex-

pressed Hpo (compare [H] and [J]), whereas

heterozygous bantam mutants have wild-

type wing size (G).

3 days of growth (Figures 5B and 5E). In contrast, yki mu-tant cells that overexpressed bantam readily formedclones of many cells, although these clones were notas large as their twin clones (Figures 5C and 5E). Thisrescue of yki mutant clones is unlikely to be due solelyto the suppression of cell death by bantam, because ex-pression of the caspase inhibitor p35, which suppressescell death, only weakly rescued the size of yki mutantclones (Figures 5D and 5E). This indicates that bantamcan rescue growth defects in yki mutant cells, ratherthan simply prevent the cells from death. This growthrescue, however, was incomplete because yki mutantclones that overexpressed bantam did not grow aswell as wild-type clones (Figure 5E) and were alsosmaller than wild-type clones overexpressing bantam,which were generally larger than their twin clones(Figure 5E). Together with the rescue of the small-eyeand -wing phenotypes caused by hyperactivated Hpo,these results show that overexpression of bantam res-cues growth defects as well as the induction of apopto-sis in cells with hyperactivated Hpo signaling.

bantam, cyclin E, and diap1 Are Independent Targetsof Hpo Signaling

The findings that Yki regulates the expression of bantamand that overexpression of bantam significantly rescuesthe growth defects of yki mutant cells indicate that ban-tam is a critical downstream effector of Yki. However,

because the rescue of the growth defects of yki mutantcells by bantam was not complete (Figure 5), we wantedto know whether bantam is sufficient to mediate all func-tions of Yki activity or instead whether bantam is one ofseveral target genes regulated by Yki. To distinguish be-tween these possibilities, we tested the relationship ofbantam to other known targets of Hpo signaling. If ban-tam were mediating all functions of Hpo, then we wouldexpect bantam to regulate the expression of other Hpotarget genes. Hpo is known to negatively regulate theexpression of ex, mer, diap1, and cyclin E at the tran-scriptional level [4–7, 9, 10, 13]. We found that clonesof cells overexpressing bantam did not affect the ex-pression of lacZ enhancer-trap reporters in the ex anddiap1 genes (Figures S2A and S2B). Furthermore, ban-tam overexpression did not significantly affect DIAP1and Ex protein levels and had only very subtle effectson Cyclin E levels (Figure S3 and not shown), unlikeYki expression, which induces robust expression of Cy-clin E [12]. We conclude that Hpo signaling regulates theexpression of cyclin E, diap1, and bantam indepen-dently of one another.

Overexpression of bantam Mimics hippo Mutant

PhenotypesThe identification of bantam as a gene acting down-stream of Hpo signaling raised the question of whetheroverexpression of bantam alone or together with other

Hippo Signaling Targets the bantam miRNA1901

Figure 5. bantam Expression Rescues

Growth Defects of yorkie Mutant Cells

(A–A00) Confocal images of wing imaginal

discs that contain yki mutant clones and

also express bantam in the posterior com-

partment by hh-Gal4. Clones are marked by

the absence of GFP (green in [A] and [A00]

and gray in [A0]), and a-Ci marks the anterior

compartment (red in [A] and [A00]). The

expression of bantam in the posterior com-

partment allows much more clone growth

(arrowhead) than that seen in the anterior

compartment (arrow). TO-PRO (blue in [A])

shows that the yki clones are not void of cells.

(B–D) Confocal images of wing imaginal discs

with (B) yki mutant clones, (C) yki mutant

clones expressing bantam, and (D) yki mutant

clones expressing p35. Clones are positively

marked by the expression of GFP (green)

and negatively marked by the absence of

CD2 expression (red). The twin clones can

be identified because they express twice as

much CD2 as heterozygous cells. Arrow-

heads point to clones, and asterisks mark

twin clones. The clones in (B)–(D) were pro-

duced with the MARCM system [42].

(E) Quantification of the size of clones com-

pared to their twin clones. n = 54 for yki

clones, n = 28 for yki clones overexpressing

bantam, n = 32 for yki clones expressing

p35, n = 23 for wild-type clones, and n = 23

for wild-type clones overexpressing bantam.

Hpo target genes would be sufficient to recapitulate thephenotypes of hpo mutant cells. To test this, we overex-pressed bantam, DIAP1, and Cyclin E in the developingeye and compared the resulting phenotypes with thosecaused by Yki overexpression. Overexpression of ban-tam caused the generation of extra interommatidial pig-ment and bristle cells (Figures 6A and 6I). This pheno-type is strikingly similar to, although weaker than, thephenotypes seen when Yki is overexpressed (Figure 6H,[12]). In contrast, overexpression of DIAP1 or Cyclin Ecaused much weaker phenotypes and resulted in onlya slight increase in the number of interommatidial cells(Figures 6B and 6C). Addition of Cyclin E, but notDIAP1, to bantam enhanced the extra-cell phenotype,so that it more closely resembled the phenotype causedby overexpression of Yki (Figures 6D and 6E). Coexpres-sion of DIAP1 and Cyclin E generated many small cells,and expression of bantam, Cyclin E, and DIAP1 togethercaused a large increase in interommatidial cells as wellas cone and photoreceptor cells (Figures 6F and 6G).This phenotype is stronger than that caused by Yki ex-pression, possibly because overexpression by GMR-Gal4 drives higher levels of bantam, Cyclin E, andDIAP1 than that induced by Yki overexpression. In sum-mary, overexpression of bantam alone phenocopies hy-pomorphic hpo loss-of-function phenotypes, and coex-pression of Cyclin E and DIAP1 enhances these

phenotypes. These results support a model whereinthe deregulation of bantam in hpo mutant clones con-tributes to the hpo overgrowth phenotype.

Discussion

On the basis of the presented data, we postulate a modelin which bantam is an essential target of the Hpo signal-ing pathway to regulate cell proliferation, cell death, andthus tissue size (Figure 7). Our model is based on severalobservations. First, we found that bantam is regulatedby Hpo signaling broadly and in various tissues. Thisregulation is a specific downstream effect of Hpo signal-ing and is not simply the consequence of the cell prolif-eration induced in hpo mutant cells. Second, bantam isrequired for Yki to drive tissue overgrowth, because re-moval of bantam suppressed the overgrowth pheno-types caused by overexpression of Yki in the retina.Third, overexpression of bantam rescued the cell deathinduced by overexpressed Hpo and significantly res-cued growth defects of yki mutant cells. And fourth, ban-tam overexpression mimics the phenotypes of hypo-morphic hpo mutations. Taken together, our datasupport a model in which bantam is an important down-stream target of the Hpo pathway.

The finding that Hpo signaling regulates the expres-sion of bantam raises the question of how important

Current Biology1902

Figure 6. Overexpression of Hippo Target

Genes in the Developing Eye

Panels show mid-pupal retinae stained with

Discs large (Dlg) antibodies to visualize cell

outlines. bantam, DIAP1, Cyclin E, and Yki

were overexpressed by using the GMR-Gal4

driver that expresses in all cells in the devel-

oping eye beginning at the third-instar stage

when cell-type differentiation starts. Identity

of the overexpressed genes is indicated in

each panel. The GMR-Gal4 driver line by itself

has a weak-eye phenotype shown in panel (I).

Magnification of all panels is the same.

this effect is for Hpo signaling to control tissue size. Re-moval of bantam suppressed the induction of extra in-terommatidial cells in the retina by Yki overexpressionbut did not cause a general elimination of retinal cellsin a wild-type background. These data indicate thatthe regulation of bantam is an essential downstream ef-fect of Hpo signaling to regulate tissue size. However,loss of bantam only partially suppressed the effects ofYki overexpression, indicating that Yki regulates othertargets in addition to bantam. We found that Hpo regu-lates bantam independently of cyclin E and diap1, twoother genes known to be regulated by Hpo signaling.bantam is thus not a component of the Hpo signal trans-duction pathway itself, but is one of several downstreamtarget genes (Figure 7). Yki must have targets in additionto bantam, cyclin E, and diap1, because overexpressionof bantam, Cyclin E, and DIAP1 together did not inducethe amount of overgrowth caused by Yki overexpressionin wing discs (not shown). Nevertheless, overexpressionof bantam alone caused phenotypes resembling hypo-morphic situations for Hpo signaling, indicating thatbantam is a critical mediator of Hpo function. Whetherthe regulation of bantam by a Yki-containing transcrip-tion factor complex is direct remains to be determined.However, the fact that Hpo regulates bantam cell auton-omously and in multiple tissues is consistent with sucha model.

bantam expression is spatially modulated, and pat-terning signals such as Wg and Dpp also regulate the ex-pression of bantam to generate its expression pattern[21, 33]. These patterning signals regulate specific as-pects of the bantam expression pattern, and they havedifferent effects on cell proliferation as well as bantamactivity in different regions in various imaginal discs

[21, 33]. In contrast, hpo mutant cells upregulate bantamactivity independently of cell type and in multiple imag-inal discs, indicating an intimate relationship. Hpo isthus a more general and ubiquitous regulator of bantamexpression in imaginal discs. An important question thatremains to be answered is how these patterning signalsregulate tissue growth and bantam expression andwhether they regulate bantam expression directly andindependently of Hpo signaling or through the regulationof Hpo activity.

Surprisingly, opposite to hpo mutant cells, TSC1 mu-tant cells had lower levels of bantam activity althoughthese cells overgrow [30], indicating that TSC1 mutantcells induce growth independently of bantam. NeitherMyc, Ras, nor Cyclin D-Cdk4 expression induced ban-tam, although they induce cell growth and proliferation[26, 28, 29]. bantam is thus not simply a part of thecell-intrinsic machinery that executes cell growth and di-vision but rather acts as an upstream component to in-struct cells to proliferate. In summary, although Hpo isa key regulator of bantam expression, bantam is alsoregulated by other pathways potentially integrating theeffects of several growth-regulatory and patterningpathways.

miRNAs and their target genes often show mutuallyexclusive expression patterns, and miRNAs inducedduring differentiation tend to target messages thatwere abundant in the previous developmental stage[34, 35]. miRNAs may thus provide a rapid and effectivemeans to suppress expression of residual, unwantedmRNAs while the transcriptional program in a cell ischanging [34, 35]. Hpo signaling is involved in regulatingcell proliferation and apoptosis in developing imaginaldiscs. Cell lineages and cell proliferation show

Hippo Signaling Targets the bantam miRNA1903

significant plasticity in growing imaginal discs, whichcan rapidly respond to surgical ablation or genetic in-sults by regenerating missing (eliminated) cells or by ab-lating unwanted (extra) cells [1, 36]. This adjustment ofcell proliferation and apoptosis requires a mechanismthat can rapidly change the growth properties of a cell.Yki appears to regulate cell number on the one handby inducing the expression of positive regulators ofcell proliferation and cell survival and on the otherhand by inducing the expression of bantam, which post-transcriptionally suppresses the expression of proteinsthat inhibit cell proliferation and induce apoptosis. Anexample of such cooperative action of Yki and bantamis the regulation of Hid: Yki suppresses the expressionof hid, but also induces bantam, which then suppressesthe translation of hid mRNAs that may still be present ina cell. The induction of bantam by Yki may also acceler-ate the repression of negative growth regulators,thereby enabling a cell to more quickly and robustly ad-just its rate of cell proliferation. It will be interesting toelucidate how bantam regulates growth and how itsgrowth targets are integrated with other targets of Hposignaling.

Experimental Procedures

Drosophila Stocks

Mutant clones were induced with the flipase/flipase recombination

target (FLP/FRT) system [37]. For analysis of the GFP expression

from the bantam sensor [21] in hpo, wts, yki, or TSC1 clones, the

Figure 7. Model for Hippo Signaling

Logic of the interactions of known Hippo signaling components. The

FERM-domain proteins Expanded (Ex) and Merlin (Mer) act up-

stream of the Hippo (Hpo) kinase, which in turn activates the Warts

(Wts) kinase with help from the Salvador (Sav) adaptor protein.

Warts, together with its cofactor, Mats, suppresses the activity of

the transcriptional coactivator Yorkie (Yki). Yki drives the expression

of the downstream target genes bantam, cyclin E, diap1, ex, and

mer. The expression of these effector genes regulates cell growth

and cell-cycle progression as well as cell death. The regulation of

the expression of ex and mer provides a negative feedback loop in

the pathway. The effector genes are transcriptionally regulated

and are shown on a shaded background.

following alleles were flipped against corresponding arm-lacZ-

marked FRT chromosomes: hpo42-47 [6], wtsx1 [16], ykiB5 [12], and

TSC1IQ69 [4]. Overexpression was achieved with the Gal4-UAS sys-

tem [38] and the following stocks: UAS-hpo [4], UAS-DIAPI (from B.

Hay), UAS-p35 (from B. Hay), UAS-cyclin E [17], bantam EP3622 [20],

UAS-yki [12], UAS-Rasv12 [39], UAS-Myc [28], UAS-CycD, UAS-Cdk4

[26], GMR-Gal4, engrailed-Gal4, and nubbin-Gal4. Clones over-

expressing bantam, Cyclin E, and DIAP1 were induced with hsFLP;

act>y+>GAL4, UAS-GFPS65T and yw P{Act>CD2>Gal4}; hsFLP

MKRS/TM6B [40]. Other stocks used were FRT80B bandelta1/

TM6B [20], ex697/CyO [41], and diap1-lacZ [19]. Bantam-overex-

pressing MARCM ykiB5 clones were made with yw FLP122

tub-GAL4 UAS-GFP-6xmyc-NLS; FRT42D tub-GAL80 hsp70-CD2

y+/CyO [42].

Scanning Electron Microscopy and Immunohistochemistry

Scanning electron microscopy (SEM) of adult flies was carried out by

following the hexamethyldisilazane (HMDS) method [10]. Antibody

stainings of imaginal discs were performed as described previously

[10] by using the following antibodies (dilutions and source in paren-

theses): mouse a-CD2 (1:1000; Serotec), mouse a-bGalactosidase

(1:2000; Promega), rat a-Ci (1:150; a gift from R. Holmgren), mouse

a-Dlg (1:300; Developmental Studies Hybridoma Bank), mouse a-

Cyclin E (1:40; a gift from Helena Richardson), mouse a-Cyclin D

(1:100; a gift from Wei Du), mouse a-DIAP1 (1:200; a gift from Bruce

Hay), rabbit a-Drice (1:2000; a gift from Bruce Hay), guinea pig a-Hid

(1:50; a gift from Hyung Don Ryoo and Herman Steller), and guinea

pig a-Hpo (1:2000; [13]). Secondary antibodies used were donkey

Fab2 fragments from Jackson ImmunoResearch (West Grove, Penn-

sylvania); TO-PRO was from Molecular Probes.

Supplemental Data

Supplemental Data include three figures and are available with this

article online at: http://www.current-biology.com/cgi/content/full/

16/19/1895/DC1/.

Acknowledgments

We thank S. Cohen, W. Du, B. Edgar, R.G. Fehon, B. Hay, R. Holmg-

ren, M. Kango-Singh, D.J. Pan, H. Richardson, H.D. Ryoo, A. Singh,

H. Steller, G. Struhl, N. Tapon, the Bloomington Drosophila Stock

Center, and the Developmental Studies Hybridoma Bank for fly

stocks and antibodies. We thank K. Dunner for help with SEM at

the M.D. Anderson core facility, which is supported by Cancer Cen-

ter Core Grant CA16672. We thank all members of the Halder lab for

discussions. This publication was made possible by a grant from the

National Institutes of Health (GM067997) to G.H.

Received: May 21, 2006

Revised: July 30, 2006

Accepted: August 18, 2006

Published online: August 31, 2006

References

1. Johnston, L.A., and Gallant, P. (2002). Control of growth and or-

gan size in Drosophila. Bioessays 24, 54–64.

2. Conlon, I., and Raff, M. (1999). Size control in animal develop-

ment. Cell 96, 235–244.

3. Hipfner, D.R., and Cohen, S.M. (2004). Connecting proliferation

and apoptosis in development and disease. Nat. Rev. Mol. Cell

Biol. 5, 805–815.

4. Udan, R.S., Kango-Singh, M., Nolo, R., Tao, C., and Halder, G.

(2003). Hippo promotes proliferation arrest and apoptosis in

the Salvador/Warts pathway. Nat. Cell Biol. 5, 914–920.

5. Pantalacci, S., Tapon, N., and Leopold, P. (2003). The Salvador

partner Hippo promotes apoptosis and cell-cycle exit in Dro-

sophila. Nat. Cell Biol. 5, 921–927.

6. Wu, S., Huang, J., Dong, J., and Pan, D. (2003). hippo encodes

a Ste-20 family protein kinase that restricts cell proliferation

and promotes apoptosis in conjunction with salvador and warts.

Cell 114, 445–456.

Current Biology1904

7. Harvey, K.F., Pfleger, C.M., and Hariharan, I.K. (2003). The Dro-

sophila Mst ortholog, hippo, restricts growth and cell prolifera-

tion and promotes apoptosis. Cell 114, 457–467.

8. Jia, J., Zhang, W., Wang, B., Trinko, R., and Jiang, J. (2003). The

Drosophila Ste20 family kinase dMST functions as a tumor sup-

pressor by restricting cell proliferation and promoting apopto-

sis. Genes Dev. 17, 2514–2519.

9. Tapon, N., Harvey, K., Bell, D., Wahrer, D., Schiripo, T., Haber,

D., and Hariharan, I. (2002). salvador promotes both cell cycle

exit and apoptosis in Drosophila and is mutated in human can-

cer cell lines. Cell 110, 467–478.

10. Kango-Singh, M., Nolo, R., Tao, C., Verstreken, P., Hiesinger,

P.R., Bellen, H.J., and Halder, G. (2002). Shar-pei mediates cell

proliferation arrest during imaginal disc growth in Drosophila.

Development 129, 5719–5730.

11. Lai, Z.C., Wei, X., Shimizu, T., Ramos, E., Rohrbaugh, M., Niko-

laidis, N., Ho, L.L., and Li, Y. (2005). Control of cell proliferation

and apoptosis by mob as tumor suppressor, mats. Cell 120,

675–685.

12. Huang, J., Wu, S., Barrera, J., Matthews, K., and Pan, D. (2005).

The Hippo signaling pathway coordinately regulates cell prolif-

eration and apoptosis by inactivating Yorkie, the Drosophila ho-

molog of YAP. Cell 122, 421–434.

13. Hamaratoglu, F., Willecke, M., Kango-Singh, M., Nolo, R., Hyun,

E., Tao, C., Jafar-Nejad, H., and Halder, G. (2006). The tumour-

suppressor genes NF2/Merlin and Expanded act through Hippo

signalling to regulate cell proliferation and apoptosis. Nat. Cell

Biol. 8, 27–36.

14. Edgar, B.A. (2006). From cell structure to transcription: Hippo

forges a new path. Cell 124, 267–273.

15. Justice, R.W., Zilian, O., Woods, D.F., Noll, M., and Bryant, P.J.

(1995). The Drosophila tumor suppressor gene warts encodes

a homolog of human myotonic dystrophy kinase and is required

for the control of cell shape and proliferation. Genes Dev. 9, 534–

546.

16. Xu, T., Wang, W., Zhang, S., Stewart, R.A., and Yu, W. (1995).

Identifying tumor suppressors in genetic mosaics: The Drosoph-

ila lats gene encodes a putative protein kinase. Development

121, 1053–1063.

17. Richardson, H., O’Keefe, L.V., Marty, T., and Saint, R. (1995). Ec-

topic cyclin E expression induces premature entry into S phase

and disrupts pattern formation in the Drosophila eye imaginal

disc. Development 121, 3371–3379.

18. Neufeld, T.P., de la Cruz, A.F., Johnston, L.A., and Edgar, B.A.

(1998). Coordination of growth and cell division in the Drosophila

wing. Cell 93, 1183–1193.

19. Hay, B.A., Wassarman, D.A., and Rubin, G.M. (1995). Drosophila

homologs of baculovirus inhibitor of apoptosis proteins function

to block cell death. Cell 83, 1253–1262.

20. Hipfner, D.R., Weigmann, K., and Cohen, S.M. (2002). The ban-

tam gene regulates Drosophila growth. Genetics 161, 1527–

1537.

21. Brennecke, J., Hipfner, D.R., Stark, A., Russell, R.B., and Cohen,

S.M. (2003). bantam encodes a developmentally regulated mi-

croRNA that controls cell proliferation and regulates the proa-

poptotic gene hid in Drosophila. Cell 113, 25–36.

22. Raisin, S., Pantalacci, S., Breittmayer, J.P., and Leopold, P.

(2003). A new genetic locus controlling growth and proliferation

in Drosophila melanogaster. Genetics 164, 1015–1025.

23. Prober, D.A., and Edgar, B.A. (2001). Growth regulation by onco-

genes—new insights from model organisms. Curr. Opin. Genet.

Dev. 11, 19–26.

24. Tapon, N., Moberg, K.H., and Hariharan, I.K. (2001). The cou-

pling of cell growth to the cell cycle. Curr. Opin. Cell Biol. 13,

731–737.

25. Potter, C.J., and Xu, T. (2001). Mechanisms of size control. Curr.

Opin. Genet. Dev. 11, 279–286.

26. Datar, S.A., Jacobs, H.W., de la Cruz, A.F., Lehner, C.F., and Ed-

gar, B.A. (2000). The Drosophila cyclin D-Cdk4 complex pro-

motes cellular growth. EMBO J. 19, 4543–4554.

27. Meyer, C.A., Jacobs, H.W., Datar, S.A., Du, W., Edgar, B.A., and

Lehner, C.F. (2000). Drosophila Cdk4 is required for normal

growth and is dispensable for cell cycle progression. EMBO J.

19, 4533–4542.

28. Johnston, L.A., Prober, D.A., Edgar, B.A., Eisenman, R.N., and

Gallant, P. (1999). Drosophila myc regulates cellular growth dur-

ing development. Cell 98, 779–790.

29. Prober, D.A., and Edgar, B.A. (2000). Ras1 promotes cellular

growth in the Drosophila wing. Cell 100, 435–446.

30. Potter, C.J., Huang, H., and Xu, T. (2001). Drosophila Tsc1 func-

tions with Tsc2 to antagonize insulin signaling in regulating cell

growth, cell proliferation, and organ size. Cell 105, 357–368.

31. Baker, N.E. (2001). Cell proliferation, survival, and death in the

Drosophila eye. Semin. Cell Dev. Biol. 12, 499–507.

32. Brachmann, C.B., and Cagan, R.L. (2003). Patterning the fly eye:

The role of apoptosis. Trends Genet. 19, 91–96.

33. Martin, F.A., Perez-Garijo, A., Moreno, E., and Morata, G. (2004).

The brinker gradient controls wing growth in Drosophila. Devel-

opment 131, 4921–4930.

34. Stark, A., Brennecke, J., Bushati, N., Russell, R.B., and Cohen,

S.M. (2005). Animal MicroRNAs confer robustness to gene ex-

pression and have a significant impact on 30UTR evolution.

Cell 123, 1133–1146.

35. Farh, K.K., Grimson, A., Jan, C., Lewis, B.P., Johnston, W.K.,

Lim, L.P., Burge, C.B., and Bartel, D.P. (2005). The widespread

impact of mammalian MicroRNAs on mRNA repression and evo-

lution. Science 310, 1817–1821.

36. Day, S.J., and Lawrence, P.A. (2000). Measuring dimensions:

The regulation of size and shape. Development 127, 2977–2987.

37. Xu, T., and Rubin, G.M. (1993). Analysis of genetic mosaics in de-

veloping and adult Drosophila tissues. Development 117, 1223–

1237.

38. Brand, A.H., and Perrimon, N. (1993). Targeted gene expression

as a means of altering cell fates and generating dominant pheno-

types. Development 118, 401–415.

39. Lee, T., Feig, L., and Montell, D.J. (1996). Two distinct roles for

Ras in a developmentally regulated cell migration. Development

122, 409–418.

40. Pignoni, F., and Zipursky, S.L. (1997). Induction of Drosophila

eye development by decapentaplegic. Development 124, 271–

278.

41. Boedigheimer, M., and Laughon, A. (1993). Expanded: A gene in-

volved in the control of cell proliferation in imaginal discs. Devel-

opment 118, 1291–1301.

42. Lee, T., and Luo, L. (1999). Mosaic analysis with a repressible

cell marker for studies of gene function in neuronal morphogen-

esis. Neuron 22, 451–461.