Phosphoinositide 3-kinase and Akt are essentialfor Sonic Hedgehog signalingNatalia A. Riobo*, Ke Lu†, Xingbin Ai†, Gwendolyn M. Haines*, and Charles P. Emerson, Jr.†‡

*Department of Cell and Developmental Biology, University of Pennsylvania School of Medicine, 1157 BRBII�III, 421 Curie Boulevard, Philadelphia, PA 19104;and †Boston Biomedical Research Institute, 64 Grove Street, Watertown, MA 02472

Edited by Lewis C. Cantley, Harvard Institutes of Medicine, Boston, MA, and approved January 18, 2006 (received for review May 25, 2005)

Hedgehogs (Hhs) are key signaling regulators of stem cell main-tenance and tissue patterning in embryos, and activating muta-tions in the pathway that increase Gli transcriptional activity arecausal in a diversity of cancers. Here, we report that phosphoino-sitide 3-kinase (PI3-kinase)-dependent Akt activation is essentialfor Sonic Hedgehog (Shh) signaling in the specification of neuronalfates in chicken neural explants, chondrogenic differentiation of10T1�2 cells, and Gli activation in NIH 3T3 cells. Stimulation ofPI3-kinase�Akt by insulin-like growth factor I potentiates Gli acti-vation induced by low levels of Shh; however, insulin-like growthfactor I alone is insufficient to induce Gli-dependent transcription.Protein kinase A (PKA) and glycogen synthase kinase 3� sequen-tially phosphorylate Gli2 at multiple sites, identified by mutagen-esis, thus resulting in a reduction of its transcriptional activity. Gli2mutant proteins in which the major PKA and glycogen synthasekinase 3� phosphorylation sites were mutated to alanine remainfully transcriptionally active; however, PKA-mutant Gli2 functionsindependently of Akt signaling, indicating that Akt positivelyregulates Shh signaling by controlling PKA-mediated Gli inactiva-tion. Our findings provide a basis for the synergistic role ofPI3-kinase�Akt in Hh signaling in embryonic development andHh-dependent tumors.

glycogen synthase kinase 3� � protein kinase A � Gli2 � phosphorylation

Hedgehog (Hh) signaling regulates tissue patterning and stemcell maintenance in vertebrate and invertebrate embryos

(1). Hh, by binding to its receptor Patched (Ptc), releases Ptcinhibition of Smoothened (Smo), a membrane protein related toG protein-coupled receptors, which then transduces a signal foractivation and nuclear translocation of a family of transcriptionfactors, Ci in Drosophila and Glis (Gli1, Gli2, and Gli3) invertebrates. Although Smo signal transduction mechanisms arenot well understood, Smo is known to control Ci phosphoryla-tion. In the absence of Hh, Smo is localized mostly in vesicles (2)and Ci is phosphorylated by protein kinase A (PKA) at multiplesites, which prime additional phosphorylation at interspersedsites by the glycogen synthase kinase 3 (GSK-3�) homologueshaggy and casein kinase-I (3, 4). Hyperphosphorylated Ci istargeted for proteasomal degradation to generate a repressorform. During pathway activation, Smo is enriched at the plasmamembrane, and Ci phosphorylation is prevented, leading tostabilization and nuclear translocation of full-length Ci. The roleof phosphorylation in the regulation of vertebrate Gli proteinshas not yet been defined, although PKA is a known inhibitor ofvertebrate Hh signaling (5, 6).

A diversity of human cancers are caused by mutations thatlead to inappropriate Hh pathway activation, including loss-of-function mutations in Ptc, gain-of-function mutations in Smo, orGli gene amplification and overproduction of Hh ligand (7).Genetic studies in mice reveal that the insulin-like growth factor(IGF)�phosphoinositide 3-kinase (PI3-kinase)�Akt pathwayprovides a synergistic signal for Hh tumor formation (8, 9).Interestingly, there is a high incidence of loss-of-function mu-tations of PTEN, a negative regulator of PI3-kinase activity, in�39% of Hh-dependent human pancreatic cancers (10), which

apparently strictly depend on PI3-kinase for proliferation (11).Thus, we hypothesized that the Hh pathway and the PI3-kinasepathway could be part of a cross-regulatory signaling network.Our results establish that PI3-kinase and Akt are indeed essen-tial for Hh signaling and provide a basis for the increased tumorformation potential when the Hh pathway is activated in tissueswith enhanced PI3-kinase signaling.

ResultsTo investigate the role of the PI3-kinase�Akt pathway in thecontrol of Hh signaling, we tested the effect of IGF-I on SonicHedgehog (Shh) signaling in LIGHT cells. Recombinant Shh(N-Shh) induces a dose-dependent activation of Gli-luciferase,and this response is enhanced �3-fold by IGF-I at submaximalN-Shh concentrations (Fig. 1A). IGF-I strongly induces Aktphosphorylation, an effect that is sensitive to the specific PI3-kinase inhibitor 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002). However, IGF-I is not sufficient to driveGli-luciferase activation (Fig. 1B). N-Shh alone also induces theactivation (�2-fold) of PI3-kinase-dependent Akt phosphoryla-tion in LIGHT cells within 5 min (Fig. 1C), in agreement withprevious findings in endothelial cells (12). Notably, LY294002treatment inhibits Shh-activated Gli-luciferase expression to thesame extent as cyclopamine, a potent Smo inhibitor (Fig. 1D).LY294002 also blocks activation of the Gli-luciferase reporterelicited by SmoM2, an oncogenic mutant of Smo (Fig. 5, whichis published as supporting information on the PNAS web site)(13). These data indicate that PI3-kinase activity, stimulated byShh or by other ligands, is essential for Gli activation.

To investigate whether PI3-kinase regulates Shh signalingthrough Akt, we reduced endogenous Akt levels with an Akt1small interfering RNA (siRNA) or impaired its activity with anonphosphorylatable Akt mutant that acts as dominant negative(dnAkt). Akt1 siRNA reduces Shh-induced luciferase activity inLIGHT cells by 60% (Fig. 1E). The �20% difference in efficacyof LY294002, which totally blocks PI3-kinase signaling throughAkt, and Akt1 siRNA to suppress Shh signaling is likely causedby the coexpression of low levels of Akt2 and Akt3, which likelyfunction redundantly. Moreover, dnAkt potently inhibits Gli-luciferase induced by Shh, SmoM2, Gli1, and Gli2 in NIH 3T3cells (Fig. 1F), establishing that Akt is essential for Gli function.Similar inhibition was observed with a kinase-dead dnAkt (datanot shown). In addition, overexpression of a constitutively activeform of Akt (myr-Akt) in LIGHT cells rescues activation of theGli-luciferase reporter by N-Shh in the presence of LY294002

Conflict of interest statement: No conflicts declared.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: Hh, Hedgehog; Shh, Sonic Hedgehog; N-Shh, recombinant Shh; PI3-kinase,phosphoinositide 3-kinase; PKA, protein kinase A; PSM, PKA site mutations; GSK-3�,glycogen synthase kinase 3�; GSM, GSK-3� site mutations; IGF, insulin-like growth factor;LY294002, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one; Ptc, Patched; Smo,Smoothened; siRNA, small interfering RNA; dnAkt, dominant negative Akt; AP, alkalinephosphatase; RLU, relative luciferase units.

‡To whom correspondence should be addressed. E-mail: emersonc@bbri.org.

© 2006 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0504337103 PNAS � March 21, 2006 � vol. 103 � no. 12 � 4505–4510

DEV

ELO

PMEN

TAL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

22, 2

020

(Fig. 1G). Rapamycin, a specific inhibitor of mTOR pathwaydownstream of PI3-kinase, does not affect N-Shh-induced Gli-luciferase (Fig. 6, which is published as supporting informationon the PNAS web site). These results support a requirement ofAkt for Shh signaling and rule out a contribution of the mTORpathway.

To test whether PI3-kinase and Akt play a role in Shh-regulated developmental signaling, we analyzed the effects ofLY294002 on Shh-dependent dorsal-ventral patterning in em-bryonic neural tube explants cocultured with its associated,Shh-producing notochord. LY294002 treatment of neural tube-notochord explants reduces phospho-Akt levels and causesneural patterning defects consistent with Shh loss of function.These patterning defects include a 50% decrease in Islet1-expressing motor neuron progenitors, which normally form inthe ventral neural tube in response to high concentrations of Shhfrom the floor plate and notochord, and a reciprocal �300%increase in dorsally localized Pax7-expressing progenitors (Fig.2 A and B), which are normally repressed by Shh (14, 15).Cyclopamine causes identical defects in Shh-induced neural tubepatterning in this experimental model (16). LY294002 treatmentof neural tube explants does not disrupt Shh expression in thefloor plate (Fig. 2 A) and does not affect cell proliferation orapoptosis in the neural tube, as determined by BrdUrd incor-poration and caspase-3 cleavage (Fig. 2B), suggesting a specificeffect of LY294002 on Shh-induced neural tube patterning.

In 10T1�2 mesodermal cells, Shh induces the late expressionof alkaline phosphatase (AP), a marker of chondrogenic differ-entiation (17). Using this model, we found that LY294002completely inhibits AP induction by N-Shh (Fig. 2C), which isrescued by ectopic expression of myr-Akt (Fig. 2D), further

demonstrating the essential function of PI3-kinase and Akt in Hhdevelopmental signaling. Finally, dnAkt also reduces the acti-vation of a Gli-luciferase reporter by Shh, SmoM2, Gli1, and Gli2in 10T1�2 cells (Fig. 2E), demonstrating a broad requirement ofAkt signaling in Gli-dependent Hh signaling.

As Akt is known to phosphorylate and inhibit GSK-3� (18), weinvestigated the role of PI3-kinase�Akt in the regulation of Gli2transcriptional activity by PKA and GSK-3� phosphorylation.PKA represses Gli2 transcriptional activity in NIH 3T3 cells,which is further reduced by GSK-3� (Fig. 3A), thus establishingthat PKA and GSK-3� are negative regulators of Gli2, as shownfor Drosophila Ci (3, 4). However, GSK-3� alone is insufficientto inhibit Gli2-mediated activation of the Gli-luciferase reporter,consistent with an obligatory priming role of PKA for GSK-3�phosphorylation (3). Immunoprecipitated Gli2 is phosphory-lated directly by PKA, and indirectly by GSK-3�, after PKApriming (Fig. 3B). Comparison of Ci and Gli2 sequences revealsfour clusters of consensus PKA phosphorylation sites inter-spersed with GSK3-� and casein kinase I sites within conserveddomains (Fig. 3C), as well as a single PKA site located N-terminal of cluster I. The functionality of each putative PKA andGSK-3� phosphorylation site was tested directly by in vitrophosphorylation by using 23-mer peptides corresponding to thefour WT clusters and modified peptides containing Ser�Thr toAla substitutions in each site. These experiments reveal thatS805, S817, and S956 in clusters I, II, and IV are functional PKAsites. However, only S801 and S813 in clusters I and II, and notT952, appear to be functional GSK-3� phosphorylation sites inthe peptide assay. Phosphorylation of the GSK-3� sites strictlydepends on the phosphorylation of the neighboring PKA sites(Fig. 3D). To test the functionality of those candidate sites, we

Fig. 1. PI3-kinase signaling through Akt is essential for Shh signal transduction. (A) Gli-luciferase activity in LIGHT cells induced for 24 h by increasingconcentrations of N-Shh in the absence or presence of 50 nM IGF-I. (B) Gli-luciferase activity in LIGHT cells after 24 h of treatment with IGF-I or N-Shh (5 �g�ml).Phosphorylation of Akt (Ser-473) after 15 min of treatment with increasing concentrations of IGF-I was determined by Western blot (phospho-Akt, 1:1,000; totalAkt 1:2,000). (C) Effect of 5 �g�ml N-Shh on Akt phosphorylation at different time points, assessed as in B. Densitometry values representing the increase in P-Aktare indicated at the bottom. (D) Effect of LY294002 (15 �M) and KAAD-cyclopamine (100 nM) on Gli-luciferase induced by Shh-conditioned medium (Shh-CM)or N-Shh in LIGHT cells at 24 h. (E) LIGHT cells were transfected with Akt1 or control siRNA, and 48 h later they were stimulated with N-Shh for 24 h. (F) Effectof dnAkt on Gli-luciferase activity in NIH 3T3 cells transiently transfected with the indicated constructs. (G) LIGHT cells stably transfected with LacZ or myr-Aktwere stimulated with 5 �g�ml N-Shh with or without LY294002 (15 �M) for 24 h and assayed for Gli-luciferase activity, normalized to the baseline in the absenceof N-Shh. Data represent the mean � SEM of three experiments. *, P � 0.01. RLU, relative luciferase units.

4506 � www.pnas.org�cgi�doi�10.1073�pnas.0504337103 Riobo et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

22, 2

020

generated Gli2 mutants with triple PKA site mutations (PSM) inclusters I, II, and IV or triple GSK-3� site mutations (GSM).After expression in 293 cells and immunoprecipitation, the Gli2PSM and Gli2 GSM mutants show a significant reduction inPKA- and GSK-3�-dependent phosphorylation, respectively(Fig. 3E). Mutation of all five putative PKA sites results in totallack of phosphorylation by PKA (data not shown). We thensought to determine whether Gli2 is phosphorylated in intactcells by endogenous PKA at those sites. Indeed, PKA stimulationwith forskolin in [32P]orthophosphate-labeled 293 cells inducesa significant increase in 32P incorporation in WT Gli2, whereasphosphorylation of the PSM mutant is markedly decreased(Fig. 3F).

To test whether PKA and GSK-3� regulate Gli2 transcrip-tional activity by phosphorylation, mutants and WT Gli2 werecoexpressed with PKA and GSK-3�. Single or double mutationsof the PKA phosphorylation sites in all clusters did not alter Gli2activity or its sensitivity to repression by PKA or GSK-3� (datanot shown). However, Gli2 PSM is completely resistant torepression by PKA or PKA plus GSK-3� (Fig. 4A), demonstrat-ing the functionality of the PKA kinase sites identified above.Both Gli2 PSM and Gli2 GSM are expressed at higher levels thanWT Gli2 (Fig. 4A), because of reduced degradation (Fig. 7,which is published as supporting information on the PNAS website), establishing that PKA and GSK-3� phosphorylation func-tion cooperatively to control Gli2 turnover. Moreover, WT Gli2and Gli2 PSM are similarly localized primarily in the cytoplasm(Fig. 4C). These results suggest that the apparent elevatedtranscriptional activity of the mutants is likely a result ofincreased stability rather than of differential subcellular local-ization. In addition, Gli2 with mutations of the GSK-3� sites inclusters I and II (GSM I and II) or all four clusters (GSM I–IV)are inhibited by PKA, but are resistant to additional repression

by coexpressed GSK-3� (Fig. 4A), demonstrating the function-ality of those sites in intact cells.

Mutation of S230 in Gli2, a putative Akt phosphorylation site(RXRXXS, ref. 19), does not prevent dnAkt repression (Fig. 8,which is published as supporting information on the PNAS website). We then tested whether inhibition of Gli2 by dnAkt isexerted through the control of PKA- or GSK-3�-dependent Gli2phosphorylation and�or degradation. Specifically, we tested theeffect of dnAkt on the transcriptional activity of the Gli2phosphorylation mutants. Gli2 PSM, which is resistant to phos-phorylation and repression by PKA, is also fully resistant toinhibition by dnAkt (Fig. 4B). In contrast, dnAkt still repressesGli2 GSM (I–IV) activity (Fig. 4B). The reduction of WT Gli2activity by dnAkt is associated to a significant decrease in Gli2expression levels, likely caused by increased degradation,whereas Gli2 PSM and Gli2 GSM levels are less affected bydnAkt (Fig. 4 B and C and Fig. 9, which is published assupporting information on the PNAS web site), thus indicatingthat Akt does not regulate Gli2 turnover directly, but mediatesits positive function through control of PKA phosphorylation.Also, mutation of all conserved putative casein kinase I sitesdoes not prevent repression by dnAkt (Fig. 8). In addition, dnAktdoes not change the subcellular localization of WT or PSM Gli2(Fig. 4C). Altogether, these studies establish that Gli2 turnoverand its transcriptional activity are regulated by phosphorylationby PKA and GSK-3�, and that PI3-kinase, acting through Akt,interferes either with PKA phosphorylation of Gli2 or PKAphosphorylation-dependent events.

DiscussionOur findings provide evidence that PI3-kinase and Akt activitiesare essential for Gli-dependent Shh signaling. Furthermore,stimulation of PI3-kinase�Akt by IGF-I potentiates Gli tran-

Fig. 2. PI3-kinase signaling is required for Shh-induced neural patterning and chondrogenesis. (A Upper) Treatment of neural tube�notochord explants for48 h with LY294002 (30 �M) blocked Akt phosphorylation, as assayed by Western blotting, and blocked Shh signaling, as assayed by whole-mountimmunostaining for Pax7 and Islet1 expression. (A Lower) Explants were photographed dorsal side up for Pax7 and ventral side up for Islet1. High-magnificationinsets show nuclear localization. Lowest panel shows staining of Shh in the floor plate (fp). (Scale bar: 100 �m.) (B) Quantification of Pax7 and Islet1 expression,BrdUrd incorporation, and activated caspase-3 in control and LY294002-treated neural tube explants. (C) Effects of LY294002 (15 �M) on AP induction by 5 �g�mlN-Shh in 10T1�2 cells after 4 days. RU, relative units. (D) 10T1�2 cells stably transfected with empty vector or p-myr-Akt treated as in C. (E) Effect of dnAkt onGli-luciferase activity in 10T1�2 cells transiently transfected with the indicated constructs. Data are the mean � SEM of three experiments performed in triplicate.

*, P � 0.01.

Riobo et al. PNAS � March 21, 2006 � vol. 103 � no. 12 � 4507

DEV

ELO

PMEN

TAL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

22, 2

020

scriptional activity in the presence of submaximal amounts ofShh, which, by itself, also stimulates PI3-kinase and Akt signalingat a lower magnitude. Such signaling cross talk between IGF-1and Hh at the level of Akt provides a mechanism to producegraded Hh signaling responses in embryonic and adult tissues fortissue patterning.

The precise mechanisms by which Akt controls PKA functionto regulate Gli-mediated Hh signaling remain unclear. However,our results indicate that Akt is required to antagonize PKA-dependent Gli2 inactivation. It is conceivable that Akt regulatesthe accessibility of PKA to Gli2, a frequent theme in Hhsignaling: Drosophila Hh interferes with the formation of acomplex containing Costal-2, Ci, PKA, GSK-3, and casein kinaseI (20), and at the same time, PKA phosphorylates Smo C-tail as

a requisite for activation (21–23). Vertebrate Hh signaling,however, has significant differences: the key intracellular trans-ducer, Costal-2, is not conserved, and mammalian and fish Smoare phosphorylated by GRK-2 upon activation, not by PKA, andthen internalized through a �-arrestin2-dependent mechanism(24, 25). Thus, although a Costal-2-like mechanism for regula-tion of Gli phosphorylation is unlikely, the target of Akt for Hhsignaling could be a protein of the cytoplasmic complex thatretains Gli outside of the nucleus and makes it accessible to PKA.Akt-mediated inhibition of this molecule(s) could promotenuclear translocation of Gli, escaping PKA phosphorylation.Alternatively, Akt could target protein(s) that regulate PKAdirectly, either through control of its activity and�or accessibilityto Gli2. A model to explain our findings would be that binding

Fig. 3. Gli2 has functional sites for PKA and GSK-3� phosphorylation. (A) NIH 3T3 cells were cotransfected with Gli2 and empty vector, PKA, GSK-3� or bothand assayed for Gli-luciferase activity 48 h later. (B Upper) Immunoprecipitated myc-Gli2 was incubated in 40 mM Tris�HCl (pH 7.4), 20 mM magnesium acetate,and 0.2 mM [�-32P]ATP (1 �Ci��mol) (control), with 10 units of PKA catalytic subunit or with PKA and 10 units of protein kinase inhibitor (PKI) for 30 min at 30°C.Beads were washed with 50 mM Tris�HCl (pH 7.5) and subjected to 5% SDS�PAGE and autoradiography. (B Lower) myc-Gli2 was incubated in 20 mM Tris�HCl (pH7.5), 10 mM MgCl2, 5 mM DTT, and 200 �M [�-32P]ATP (1 �Ci��mol) without (control) or with 500 units of GSK-3� for 30 min at 30°C, or Gli2 was phosphorylatedfirst with PKA and cold ATP, as described above, followed by 10 units of PKI, 500 units of GSK-3�, and 200 �M [�-32P]ATP and incubated for 30 min. Beads wereprocessed as before. (C) Sequence of the Gli2 peptides corresponding to the four clusters with the conserved PKA and GSK-3� phosphorylation sites highlightedin boxes. (D) Phosphorylation of peptides shown in C and variant peptides in which each candidate residue was substituted by alanine. Peptides (1 mM) weretreated with PKA and GSK-3� or both enzymes sequentially, as described for Gli2 in B. Reactions then were stopped by spotting onto P81 paper and washed with50 mM H3PO4, and 32P incorporation was determined by liquid scintillation. PKA phosphorylation of kemptide was taken as 100%. (E) Gli2 WT, Gli2 PSM, andGli2 GSM were expressed in 293 cells, immunoprecipated with Xpress antibody, and used as substrates for phosphorylation by PKA and GSK-3� as described inB. Phosphorylation and expression were assayed by autoradiography (Autorad.) and Western blot (WB) (Xpress Ab 1:5,000). Densitometric analysis of threeexperiments is shown in the graph. (F) Intact 293 cells were transfected with Gli2 WT or Gli2 PSM, labeled with [32P]orthophosphate, and stimulated with 10 �Mforskolin for 2 h. Gli2 was immunoprecipitated with Xpress Ab, and 32P incorporation was determined by autoradiography. Quantification by densitometry ofthree experiments was performed by using the basal level of phosphorylation of Gli2 WT as 100%.

4508 � www.pnas.org�cgi�doi�10.1073�pnas.0504337103 Riobo et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

22, 2

020

of Shh to Ptc increases surface localization of Smo for activationof PI3-kinase and Akt. Phospho-Akt would then oppose theeffects of PKA on Gli2 (and possibly Gli3) required for itsproteasomal degradation, followed by Gli2�3 nuclear transloca-tion to activate target genes (Fig. 4D). Cells responding to bothShh and additional PI3-kinase activating signals such as IGF-Iwould be less dependent on Smo for activation of PI3-kinase,rendering these cells more responsive to Shh. In this way, Shhpathway could integrate a diversity of negative and positivePI3-kinase regulatory signals from surrounding tissues to mod-ulate the final Hh signaling output through Gli function. Inter-estingly, IGF-2 and IGF binding proteins are themselves tran-scriptional targets of Shh signaling through Gli (26, 27),suggesting that the activation of Akt by Shh signaling throughSmo could become amplified and sustained through an IGF-2autocrine loop.

Our finding that PI3-kinase�Akt activation is essential for Hhsignaling sheds light on the synergy of IGF-2, PI3-kinase, andAkt with Hh in developmental processes (28, 29) and in Hh-dependent cancers (8, 9). Loss of PTEN or overexpression ofIGF in Hh cancers would up-regulate PI3-kinase�Akt activity tostimulate even low-level ligand-dependent or ligand-indepen-dent Hh signaling caused by mutations in Hh pathway compo-nents. The essential role of PI3-kinase�Akt in mammalian Hh

signaling suggests their potential use as therapeutic targets in thetreatment of Hh-dependent malignancies.

Materials and MethodsMaterials. Rabbit anti-phospho (Ser-473) Akt, rabbit anti-Akt,and rabbit anti-cleaved caspase-3 were from Cell SignalingTechnology (Beverly, MA). Mouse anti-BrdUrd was from BDBiosciences (Franklin Lakes, NJ). Mouse anti-Shh 5E1, mouseanti-Islet1 39.4D5, and mouse anti-Pax7 were from Develop-mental Studies Hybridoma Bank (Iowa City, IA).

Cell Culture and Transfection. Details are in Supporting Text, whichis published as supporting information on the PNAS web site.

Constructs and Mutagenesis. Full-length Xpress-tagged mGli1 andmGli2 expression constructs and the Gli reporter vectorp8XGBS-luc were provided by H. Sasaki (Riken Center forDevelopmental Biology, Kobe, Japan). Myc-tagged mGli1 andmGli2 were constructed by subcloning the XhoI–HindIII andBamHI–SalI fragments of mGli1 and mGli2, respectively, intopAG3-myc. Mutation of mGli2 Ser�Thr residues correspondingto consensus phosphorylation sites to Ala were done by usingoverlap extension. pRK5-Shh and pGE-SmoM2 were kindlyprovided by P. Beachy (The Johns Hopkins University, Balti-

Fig. 4. Mechanism of Akt-mediated Shh signaling. (A) Gli-luciferase activity in NIH 3T3 cells transiently transfected with WT Gli2 (WT), Gli2 GSM (clusters I–IV),and Gli2 PSM (clusters I, II, and IV), either alone or in combination with PKA or both PKA and GSK-3� expression vectors. Expression of the Gli2 variants wasdetermined by Western blot. (B) Gli2 WT, Gli2 PSM, or Gli GSM were cotransfected with empty vector or dnAkt in NIH 3T3 cells and assayed for Gli-luciferaseactivity after 48 h. Expression levels of Gli2 variants in the presence of dnAkt were determined by Western blot. (C) Subcellular localization of Gli2 WT (Left) andGli2 PSM (Right) in NIH 3T3 cells in the absence (Top) and the presence (Middle and Bottom) of dnAkt. Gli2 was detected with Xpress Ab (1:250)�Alexa Fluor 546(red) and dnAkt with Akt Ab (1:120)�Alexa Fluor 488 (green). DAPI staining of the nuclei is in blue. Asterisks indicate cells coexpressing dnAkt and Gli2. (D) Modelof Shh signaling showing the requirement PI3-kinase�Akt activation by Smo or other proliferative pathways in the regulation of Gli phosphorylation by PKA.All phosphorylation events are represented by �P . Details are described in the text.

Riobo et al. PNAS � March 21, 2006 � vol. 103 � no. 12 � 4509

DEV

ELO

PMEN

TAL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

22, 2

020

more). Akt1 (T308A and S473A) and myr-Akt1 were gifts of Y.Mitsuuchi (Fox Chase Cancer Center, Philadelphia). pmyr-Akt-hygro was constructed by subcloning of the HindIII–XbaI myr-Akt1 fragment into pcDNA3.1-hygro. The vector encodingthe catalytic subunit of PKA (pMT-CEV) was a gift from S.McKnight (University of Washington, Seattle), and pGSK-3�was kindly provided by P. Klein (University of Pennsylvania).

AP Assay. 10T1�2 cells were cultured in DMEM with 0.5% FBSwith 5 �g�ml N-Shh or vehicle and treated with or without 15�M LY294002. After 5 days, cells were sonicated in 0.1 MTris�HCl (pH 7.5) and 0.1% Triton X-100. AP activity in lysateswas determined by incubation at 37°C in 0.1 M Tris�HCl (pH 9.5),1 mM MgCl2, and 1 mg�ml p-nitrophenylphosphate. Phosphaterelease was determined spectrophotometrically at 405 nm. Re-sults were expressed as AP relative units per mg of protein.

Neural Plate Explant Culture. Fertilized chick eggs were incubatedat 38°C in a humidified incubator until developmental stage 11.Embryos were collected into L15 medium, and a region of theneural plate adjacent to newly formed four somites and pre-somitic mesoderm was isolated. The dissected neural plate wastreated with Dispase (Roche Molecular Biochemicals; 2 mg�mlin L15) at 37°C for 6–10 min to remove somites, presomiticmesoderm, and ectoderm. Neural plate�notochord explantswere cultured for 48 h in Neurobasal medium with B-27 (In-

vitrogen) in 24-well plates precoated with 2 �g�ml fibronectin.For BrdUrd labeling, 1 mg�ml BrdUrd was added to the neuralplate culture for 1 h.

Immunostaining of Cultured Neural Plate Explants. Cultured ex-plants were fixed with 4% paraformaldehyde for 2 h at 4°C andwashed three times for 30 min each in PBS (pH 7.4). Explantswere blocked in 20% goat serum in PBS, 0.2% Triton X-100, and0.2% Tween 20 (pH 7.4) for 2 h at room temperature and thenincubated with primary and secondary antibodies overnight at4°C with gentle agitation. Then explants were washed in PBSfour times for 1 h at room temperature. After staining, explantswere incubated in DAPI (Roche Molecular Biochemicals;1:10,000 in PBS) for 20 min to visualize nuclei. To quantify cellnumbers, explants were embedded in OCT compound at �25°C,cut in serial 10-�m sections, and immunostained with antibodiesand DAPI. For BrdUrd staining, sections were washed in PBS,treated with 2 M HCl for 20 min, and incubated with theantibodies. Staining was examined on a Zeiss Axiophot epif luo-rescense microscope.

We thank Phil Beachy, Hiroshi Sasaki, Yasuhiro Mitsuuchi,Stan McKnight, and Peter Klein for molecular reagents; Renne Lu(Boston Biomedical Research Institute) and Paul Leavis (Boston Bio-medical Research Institute) for peptides; and Morris Birnbaum andLucia Rameh for helpful discussion. This work was supported by aNational Cancer Institute research grant (to C.P.E.).

1. Ingham, P. W. & McMahon, A. P. (2001) Genes Dev. 15, 3059–3087.2. Zhu, A. J., Zheng, L., Suyama, K. & Scott, M. P. (2003) Genes Dev. 17,

1240–1252.3. Price, M. A. & Kalderon, D. (2002) Cell 108, 823–835.4. Jia, J., Amanai, K., Wang, G., Tang, J., Wang, B. & Jiang, J. (2002) Nature 416,

548–552.5. Epstein, D. J., Marti, E., Scott, M. P. & McMahon, A. P. (1996) Development

(Cambridge, U.K.) 122, 2885–2894.6. Wang, B., Fallon, J. F. & Beachy, P. A. (2000) Cell 100, 423–434.7. Beachy, P. A., Karhadkar, S. S. & Berman, D. M. (2004) Nature 432, 324–331.8. Hahn, H., Wojnowski, L., Specht, K., Kappler, R., Calzada-Wack, J., Potter, D.,

Zimmer, A., Muller, U., Samson, E., Quintanilla-Martinez, L. & Zimmer, A.(2000) J. Biol. Chem. 275, 28341–28344.

9. Rao, G., Pedone, C. A., Valle, L. D., Reiss, K., Holland, E. C. & Fults, D. W.(2004) Oncogene 23, 6156–6162.

10. Okami, K., Wu, L., Riggins, G., Cairns, P., Goggins, M., Evron, E., Halachmi, N.,Ahrendt, S. A., Reed, A. L., Hilgers, W., et al. (1998) Cancer Res. 58, 509–511.

11. Perugini, R. A., McDade, T. P., Vittimberga, F. J. & Calleri, M. P. (2000)J. Surg. Res. 90, 39–44.

12. Kanda, S., Mochizuki, Y., Suematsu, T., Miyata, Y., Nomata, K. & Kanetake,H. (2003) J. Biol. Chem. 278, 8244–8249.

13. Taipale, J., Chen, J. K., Cooper, M. K., Wang, B., Mann, R. K., Milenkovic, L.,Scott, M. P. & Beachy, P. A. (2000) Nature 406, 1005–1009.

14. Ericson, J., Morton, S. E., Kawakami, A., Roelink, H. & Jessell, T. M. (1996)Cell 87, 661–673.

15. Ericson, J., Rashbass, P., Schedl, A., Brenner-Morton, S., Kawakami, A., vanHeyningen, V., Jessell, T. M. & Briscoe, J. (1997) Cell 90, 169–180.

16. Incardona, J. P., Gaffield, W., Kapur, R. P. & Roelink, H. (1998) Development(Cambridge, U.K.) 125, 3553–3562.

17. Nakamura, T., Aikawa, T., Iwamoto-Enomoto, M., Iwamoto, M., Higuchi, Y.,Pacifici, M., Kinto, N., Yamaguchi, A., Noji, S., Kurisu, K. & Matsuya, T. (1997)Biochem. Biophys. Res. Commun. 237, 465–469.

18. Cross, D. A., Alessi, D. R., Cohen, P., Andjelkovich, M. & Hemmings, B. A.(1995) Nature 378, 785–789.

19. Obata, T., Yaffe, M. B., Leparc, G. G., Piro, E. T., Maegawa, H., Kashiwagi,A., Kikkawa, R. & Cantley, L. C. (2000) J. Biol. Chem. 275, 36108–36115.

20. Zhang, W., Zhao, Y., Tong, C., Wang, G., Wang, B., Jia, J. & Jiang, J. (2005)Dev. Cell 8, 267–278.

21. Jia, J., Tong, C., Wang, B., Luo, L. & Jiang, J. (2004) Nature 432, 1045–1050.22. Apionishev, S., Katanayeva, N. M., Marks, S. A., Kalderon, D. & Tomlinson,

A. (2005) Nat. Cell Biol. 7, 86–92.23. Zhang, C., Williams, E. H., Guo, Y., Lum, L. & Beachy, P. A. (2004) Proc. Natl.

Acad. Sci. USA 101, 17900–17907.24. Wilbanks, A. M., Fralish, G. B., Kirby, M. L., Barak, L. S., Li, Y. X. & Caron,

M. G. (2004) Science 306, 2264–2267.25. Chen, W., Ren, X. R., Nelson, C. D., Barak, L. S., Chen, J. K., Beachy, P. A.,

de Sauvage, F. & Lefkowitz, R. J. (2004) Science 306, 2257–2260.26. Ingram, W. J., Wicking, C. A., Grimmond, S. M., Forrest, A. R. & Wainwright,

B. J. (2002) Oncogene 21, 8196–8205.27. Lipinski, R. J., Cook, C. H., Barnett, D. H., Gipp, J. J., Peterson, R. E. &

Bushman, W. (2005) Dev. Dyn. 233, 829–836.28. Kenney, A. M., Widlund, H. R. & Rowitch, D. H. (2004) Development

(Cambridge, U.K.) 131, 217–228.29. Pirskanen, A., Kiefer, J. C. & Hauschka, S. D. (2000) Dev. Biol. 224, 189–203.

4510 � www.pnas.org�cgi�doi�10.1073�pnas.0504337103 Riobo et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

22, 2

020