Download Caspary T, Larkins CE, Anderson KV. Dev Cell. 2007 May;12(5):767-78. The graded response to Sonic Hedgehog depends on cilia architecture.

Developmental Cell
Article
The Graded Response to Sonic Hedgehog
Depends on Cilia Architecture
Tamara Caspary,1,2 Christine E. Larkins,2,3 and Kathryn V. Anderson1,*
1
Developmental Biology Program, Sloan-Kettering Institute, 1275 York Avenue, New York, NY 10021, USA
Department of Human Genetics, Emory University School of Medicine, 615 Michael Street, Suite 301, Atlanta, GA 30322, USA
3
Graduate Program in Biochemistry, Cell and Developmental Biology, Emory University, Atlanta, GA 30322, USA
*Correspondence: [email protected]
DOI 10.1016/j.devcel.2007.03.004
2
SUMMARY
Several studies have linked cilia and Hedgehog
signaling, but the precise roles of ciliary proteins in signal transduction remain enigmatic.
Here we describe a mouse mutation, hennin
(hnn), that causes coupled defects in cilia structure and Sonic hedgehog (Shh) signaling. The
hnn mutant cilia are short with a specific defect
in the structure of the ciliary axoneme, and the
hnn neural tube shows a Shh-independent expansion of the domain of motor neuron progenitors. The hnn mutation is a null allele of Arl13b,
a small GTPase of the Arf/Arl family, and the
Arl13b protein is localized to cilia. Double mutant analysis indicates that Gli3 repressor activity is normal in hnn embryos, but Gli activators
are constitutively active at low levels. Thus, normal structure of the ciliary axoneme is required
for the cell to translate different levels of Shh
ligand into differential regulation of the Gli
transcription factors that implement Hedgehog
signals.
INTRODUCTION
Nonmotile primary cilia are microtubule-based organelles
present on nearly every interphase cell in vertebrates
whose functions are only beginning to be understood
(Scholey and Anderson, 2006; Wheatley, 1995; Wheatley
et al., 1996). Some specialized nonmotile cilia such as
olfactory cilia and photoreceptor outer segments act as
chemosensory antennae, and others act as mechanosensors (Eley et al., 2005). Recent data argue that the broadly
distributed, nonspecialized primary cilia are important for
Hedgehog signal transduction. Cilia depend on intraflagellar transport (IFT) proteins for their assembly and maintenance (Scholey, 2003), and the ability to respond to
Hedgehog ligands is lost in embryos that lack any one of
six different IFT proteins (Houde et al., 2006; Huangfu
and Anderson, 2005; Huangfu et al., 2003; Liu et al.,
2005; May et al., 2005). Hedgehog-responsive cells are
ciliated, and Hedgehog signaling depends on both the anterograde and retrograde IFT motors, which suggested
that the cilium is an organelle required for Hh signaling
(Huangfu and Anderson, 2005). This hypothesis was supported by the demonstration that several components of
the classical Sonic hedgehog (Shh) signaling pathway
are enriched in cilia. Smoothened (Smo), a transmembrane
protein that is essential for Hh signaling, is enriched in the
cilia of Hh responding cells in the mouse embryo (Corbit
et al., 2005). The Gli transcription factors are the effectors
of vertebrate Hh signaling. Overexpressed Gli1, Gli2, and
Gli3 proteins are localized to cilia of primary limb bud cells
(Haycraft et al., 2005). Endogenous Gli3 and Suppressor
of Fused (SuFu), a negative regulator of Hedgehog signaling, have also been detected in limb bud cell cilia
(Dunaeva et al., 2003; Haycraft et al., 2005; Merchant
et al., 2004).
Patterning of the cell types in the ventral neural tube depends on a graded response to Shh ligand. Shh produced
in the notochord moves to the ventral spinal cord, where it
induces a series of ventral cell types in a concentrationdependent manner (Echelard et al., 1993). High concentrations of Shh induce the floor plate at the ventral midline
of the neural tube, and progressively lower levels of Shh
can specify a series of five different types of neurons in
a concentration-dependent manner, including motor neurons and four different classes of interneurons (Briscoe
et al., 2000; Ericson et al., 1997). Mutants that lack all
Hh signaling do not specify a floor plate and lack all five
classes of ventral neurons (Chiang et al., 1996; Litingtung
and Chiang, 2000). The gradient of Shh activity in the spinal cord can be visualized in the graded expression pattern of Patched1 (Ptch1), a direct Shh target gene that is
expressed in a ventral-to-dorsal gradient in the spinal
cord (Eggenschwiler et al., 2001; Goodrich et al., 1997).
All vertebrate Hh signaling is mediated by the Gli family
of transcription factors: Gli1, Gli2, and Gli3. Proteolytically
processed forms of Gli proteins (Gli-repressor) can inhibit
transcription of target genes in the absence of ligand, and
full-length Gli proteins can be modified in the presence of
Hh ligands (Gli-activator) to promote target gene transcription (Dai et al., 1999; Litingtung et al., 2002; Ruiz i
Altaba, 1998; Sasaki et al., 1997; Wang et al., 2000). In
the ventral neural tube Gli2 is predominantly an activator;
Gli2 mutants lack the extreme ventral neural cell types that
depend on high levels of Shh (Ding et al., 1998; Matise
et al., 1998). Gli3 mutants have subtle defects in patterning of lateral cell types (Persson et al., 2002), but Gli3
Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc. 767
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
repressor and activator activities are revealed in Gli3 mutants that also lack Shh, Smo, or Gli2 (Bai et al., 2004; Lei
et al., 2004; Litingtung and Chiang, 2000; Motoyama et al.,
2003; Wijgerde et al., 2002). Shh controls the balance of
the Gli activator and repressor, thereby generating a Gli
activity gradient that is sufficient to define all ventral neural
fates (Stamataki et al., 2005). IFT proteins are required for
both activity of Gli activator forms and for the proteolytic
processing that generates Gli3 repressor (Huangfu and
Anderson, 2005; May et al., 2005; Liu et al., 2005; Haycraft
et al., 2005); as a result, IFT mutants lack all responses to
Hh ligands.
Here we describe hennin, an ENU-induced mutation
that has an unprecedented effect on dorsal-ventral
patterning of the mouse neural tube. In hnn embryos,
the ventrolateral domain of motor neuron progenitors is
expanded at the expense of both the most ventral and
the most dorsal neural cell types. We find that the mutation
responsible for the hnn phenotype disrupts Arl13b, a member of the small GTPase superfamily. Arl13b (formerly
called Arl2l1) is also disrupted in the zebrafish scorpion
mutant where it is required for the formation of kidney cilia
(Sun et al., 2004), although no role in neural patterning was
described. We find that the mouse Arl13b protein is localized to cilia, and in its absence, cilia are short and display
a specific structural defect in the ciliary axoneme. The
data indicate that production of the Gli activity gradient
that defines the spatial organization of ventral neural cell
types depends on cilia structure.
RESULTS
hennin Mutant Embryos Have Defects in Neural Tube
Patterning, Limbs, and Eyes
hnn was identified in a recessive ENU screen that identified mutations that disrupted the morphology of the e9.5
mouse embryo (Garcı́a-Garcı́a et al., 2005). At that stage,
hnn mutants showed several phenotypes, including an
open neural tube in the head and caudal spinal cord,
and randomized heart looping. hnn mutants survived until
e13.5–14.5, when they also displayed abnormal eyes and
axial polydactyly (Figures 1A–1E).
We tested whether the abnormal morphology of the hnn
neural tube was caused by a disruption of patterning of
cell types along the dorsal-ventral axis. Motor neurons
are a Shh-dependent cell type and normally arise in a ventrolateral region of the neural tube. In the caudal neural
tube of hnn mutants, the domain expressing the motor
neuron markers HB9, Lhx3, and Isl1/2 was expanded
both ventrally and dorsally to include the ventral twothirds of the neural tube (Figures 1H, 1M, 5B, 5C, 5E and
5F), a phenotype that has not been described in other mutants that affect neural patterning.
The increased number of motor neurons seen in hnn
embryos (Figures 1H and 1M) could have been due to
either overproliferation of motor neuron precursors or to
an early expansion of the domain that gives rise to motor
neurons. We examined expression of the cell cycle
markers phospho-histone H3 and Ki67 and found they
were indistinguishable in wild-type and hnn neural tubes
from e8.5–e10.5 (see Figure S1 in the Supplemental
Data available with this article online; data not shown). In
contrast, Olig2, an early marker of motor neuron precursors, was expressed in an expanded domain of e9.5 hnn
embryos (Figures 1G and 1L). Thus the hnn phenotype is
the result of abnormal cell type specification rather than
abnormal proliferation.
The Gradient of Sonic Hedgehog Activity
Is Shallow in the hnn Neural Tube and Independent
of Shh Ligand
Ptch1 is a direct transcriptional target of Shh signaling and
therefore reflects Shh activity (Agren et al., 2004; Marigo
et al., 1996). In wild-type embryos, there is a ventralto-dorsal gradient of Ptch1 expression, which can be assayed by the expression of a Ptch1-lacZ reporter allele
(Eggenschwiler et al., 2001). In hnn mutants, there was
only a shallow gradient of Ptch1-lacZ expression: the level
of Ptch1-lacZ at the ventral midline was lower than in wildtype embryos, and the domain of Ptch1-lacZ expression
extended further dorsally than normal (Figures 1I and 1N).
The altered pattern of expression of Ptch-lacZ in hnn
mutants indicated that the effective domain of Shh signaling in the neural tube was expanded while the highest level
of response to Shh was lost. One possible explanation for
this Ptch-lacZ expression pattern was that the Shh morphogen spread further from its source in the notochord.
To test if there was an altered distribution of Shh ligand
in hnn, we analyzed neural patterning in the Shh hnn double mutants. Patterning in the caudal neural tube of Shh
hnn double mutants was indistinguishable from that in
hnn single mutants (Figures 1J and 1O; data not shown);
for example, motor neurons markers were expressed in
the same expanded domain in the double as in the single
mutants. As this indicates that hnn does not affect the
distribution of Shh ligand in the neural tube, Arl13b must
affect the Hedgehog signal transduction pathway at
a point downstream of Shh ligand.
The hennin Mutation Disrupts Arl13b,
which Encodes a Ciliary Protein
Using meiotic recombination, we mapped the hnn mutation to a 2 MB interval on mouse chromosome 16 that contained five predicted transcripts. We sequenced all five
transcripts and found one mutation: a T-to-G transversion
in the splice acceptor site of exon 2 of Arl13b (Figure 2A).
RT-PCR analysis confirmed that the Arl13b transcripts
in hnn did not contain exon 2. The remaining open reading frame lacked most of a predicted GTPase domain,
including the four consensus nucleotide binding sites
(Figure 2A).
Immunofluorescent staining revealed that Arl13b is expressed in the ventricular zone of the neural tube in a punctate pattern reminiscent of cilia (Experimental Procedures;
Figures 2C and 3D). We confirmed the ciliary localization
of Arl13b by double labeling with acetylated a-tubulin,
which is enriched in the stable microtubules of the ciliary
axoneme. Arl13b and acetylated a-tubulin colocalized in
768 Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc.
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Figure 1. The hnn Phenotype
(A–C) Whole e10.5 wild-type (A) and hnn (B and C) embryos, which show exencephaly and spina bifida.
(D–E) Cartilage staining of e12.5 wild-type (D) and hnn (E) forelimbs shows an extra digit in the mutant.
(F–H and K–M) Expression of Shh (F and K), Pax6 (green) and Olig2 (red) (G and L), and HB9 (H, M, J, and O) in the caudal neural tube of e10.5 wild-type
(F–H), hnn (K–M), Shh (J), and Shh hnn (O) embryos. In contrast to the defects in neural patterning shown here, patterning in the rostral spinal cord
appeared to be normal (data not shown).
(I and N) Expression of Ptch1-lacZ shows a steep gradient of b-galactosidase activity in the wild-type neural tube (I) and a dorsally expanded gradient
of b-galactosidase activity in the hnn neural tube (N). The identical phenotype of Shh hnn double mutants (O) and hnn single mutants (M) indicates that
the altered Shh activity gradient in hnn mutants (N) is Shh-independent.
the neural tube, primary mouse embryonic fibroblasts
(PMEFs), and the e8.0 node (Figures 3A–3L), indicating
that Arl13b is indeed enriched in cilia. To determine how
the hnn splice site mutation, which eliminated exon 2, affected the protein, we performed immunofluorescence
and western blotting in the hnn mutants with the antibody
raised against the C-terminal half of Arl13b (Figures 2B–
2D). Affinity-purified antibody specifically recognized the
predicted 48 kDa band on a western blot of wild-type
e10.5 whole embryo protein extract, as well as a larger
band which is likely to be a posttranslationally modified
Arl13b protein (see Experimental Procedures). Although
a transcript including the C-terminal domain was detected
in the hnn mutants by RT-PCR, both specific bands on the
western were absent in extracts from hnn mutant em-
bryos. In addition, no staining in cilia was detected in
hnn mutant PMEFs by immunofluorescence (Figures
3J–3L). These results indicate that the hnn mutation is
a null allele of the Arl13b gene.
Arl13b is a member of the Arl (ADP ribosylation factor
[Arf] related) subfamily of the Ras superfamily of small
GTPases. Specific members of the Arf/Arl group have
been shown to be important for microtubule dynamics,
lipid metabolism, or vesicle trafficking (Antoshechkin and
Han, 2002; Hoyt et al., 1990; Kahn et al., 2006; Li et al.,
2004; Radcliffe et al., 2000; Zhou et al., 2006), although
the functions of most of the members of this group are
not known. Arl13b is unusual among the Arf/Arl family,
as most members are !20 kDa in size and composed
only of their Arf domain, whereas Arl13b is a 48 kDa
Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc. 769
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Figure 2. hnn Is a Null Allele of Arl13b
(A) Schematic of the Arl13b transcript with the
position of the hnn splice site mutation marked
(*) and the corresponding intron (capital letters)
-exon (lowercase letters) sequence depicted
above. RT-PCR from hnn embryos detected
no transcripts containing exon 2, which would
cause the deletion of the putative nucleotide
binding sites in the ARF domain.
(B) Western analysis of whole embryo protein
extracts from e10.5 wild-type and hnn mutant embryos using affinity-purified antibody
against Arl13b and actin (42 kDa) as a loading
control. The anti-Arl13b antibody detects two
specific bands at !60 kDa and 48 kDa.
Arl13b is predicted to be a 48 kDa protein,
and the 48 kDa and !60 kDa bands are not
detectable in hnn extracts. As the 60 kDa band is specific and many ARL proteins are posttranslationally modified, the !60 kDa band is likely to
represent a modified form of Arl13b. The 70 kD band is an unrelated crossreacting protein (see Experimental Procedures).
(C) Arl13b protein is expressed in the ventricular zone of the wild-type e9.5 neural tube.
(D) No protein is detected by immunofluorescence in the hnn neural tube (identical exposure to wild-type).
protein that contains a C-terminal tail in addition to the Arf
domain. The C-terminal domain contains no recognizable
domains or motifs; it is present in vertebrate orthologs of
Arl13b but not in related genes in Drosophila or C. elegans.
The Ciliary Axoneme Is Abnormal in hnn Mutants
Because Arl13b localized to cilia, we examined the structure of the cilia in the node of e8.0 embryos, which are
relatively long and easy to visualize. Scanning electron
microscopy (SEM) showed that hnn nodal cilia were approximately half the length of wild-type cilia (48% ±
19%; Figures 4A and 4B). In transmission electron microscopy (TEM) cross-sections, wild-type nodal cilia have nine
peripheral doublet microtubules located around the
circumference of the axoneme (Figures 4C and 4E). In
wild-type cilia, the A-tubule of the doublet is composed
of 13 tubulin protofilaments, and the B-tubule has 10 tubulin protofilaments and an 11th filament that connects the
B-tubule to the A-tubule (Linck and Stephens, 2007; Tilney
et al., 1973). All sections of hnn mutant nodal cilia showed
disruptions of the outer doublet organization; in many
outer doublets the B-tubule was not closed or attached
to the A-tubule (Figures 4E and 4F). Thus Arl13b appears
to have a specific role in the assembly of the outer doublet.
It has previously been reported that, unlike most motile
cilia, the motile cilia of the mouse node lack the central pair
of microtubules (Takeda et al., 1999). It was therefore surprising that a central pair was clearly present in some sections of both wild-type and hnn mutant nodal cilia (Figures
4E and 4F); this finding would be consistent with the observation that there are motile and immotile cilia present
in the node (McGrath et al., 2003) as well as recent data
showing the presence of a central pair in rabbit nodal cilia
(Feistel and Blum, 2006). We speculate that we observed
central pairs because of the rapid fixation procedure used
(see Experimental Procedures), as the central pair microtubules are more labile than the outer doublet (Behnke and
Forer, 1967; Stephens, 1970). Because hnn cilia showed
a specific defect in the B-tubule in sections that contained
the labile central pair, we conclude that the electron micrographs reflected the true structure of the hnn axoneme.
The Role of Arl13b in Neural Patterning Depends
on Cilia
The hnn mutant has striking defects in both cilia structure
and activity of the Shh pathway. To test whether the defect
in cilia structure could be responsible for the signaling defect, we analyzed embryos that lack both Arl13b and cilia.
wimple (wim) mice lack the function of IFT172 and therefore have no cilia (Huangfu et al., 2003). Patterning of the
caudal neural tube of wim and hnn wim double mutants
was indistinguishable: there was no floor plate, no V3 interneurons, very few motor neurons, and the domain of
Pax6 expression extended across the ventral midline (Figure 5). Because Arl13b was localized in cilia and required
for cilia structure and loss of Arl13b had no effect on neural
patterning if cilia were not present, the simplest interpretation of the findings is that the primary function of Arl13b is
in the construction of the ciliary axoneme, and that the
correct structure of the axoneme is required for normal
specification of cell types in the neural tube.
Arl13b Acts Downstream of Patched and Smo
To better define the role of Arl13b in neural patterning, we
examined neural cell type specification in the hnn neural
tube in greater detail. As described above (Figure 1), the
domain of motor neurons was expanded both dorsally
and ventrally in the hnn neural tube, paralleling the change
in the expression pattern of Ptch1. Shh, which is normally
expressed in the notochord and induces the formation of
the floor plate at the ventral midline of the neural tube,
was expressed in the hnn notochord. However, markers
of the floor plate, Shh and FoxA2, were not expressed in
the neural tube (Figures 1F, 1K, 6A, and 6B). Nkx2.2, which
is normally expressed in the p3 domain between the floor
plate and the motor neuron progenitor domain, was
770 Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc.
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Figure 3. Arl13b Is Localized to Cilia
Arl13b (red; A, D, G, J, and M), acetylated a-tubulin (green; B, E, H, and K) and g-tubulin (green; N) in e8.0 wild-type node (A–C), e10.5 ventral neural
tube (D–F), wild-type PMEFs (G–I), hnn PMEFs (J–L), and mouse A9 fibroblasts (from ATCC) (M–O); merged images (C, F, I, L, and O).
(M–O) g-tubulin, a marker of the basal body, does not overlap with Arl13b in mouse A9 fibroblasts. The apparent nuclear staining seen in fibroblasts
under these fixation conditions was also seen with preimmune sera and therefore does not represent nuclear Arl13b protein.
expressed across the ventral midline of hnn mutants and
overlapped with expression of Olig2, a motor neuron progenitor marker (Figures 6E, 6F, 6I, and 6J). Pax6 is normally expressed at high levels in progenitors of lateral in-
terneurons, but the Pax6 expression domain was shifted
to the dorsal third of the hnn neural tube (Figures 1G, 1L,
5A, and 5D) and lateral Chx10- and En1-positive (V2 and
V1) interneurons were shifted dorsally (data not shown).
Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc. 771
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Figure 4. Cilia in hnn Node
(A and B) SEM analysis of embryonic node of E8.0 wild-type (A) and
hnn (B) embryos. The hnn cilia are, on average, half the length of the
wild-type; size bar = 1 mm. (C–F) TEM analysis of cilia from embryonic
node of e8.0 wild-type (C and E) and hnn (D and F) embryos; size bar =
0.1 mm. A central pair is visible in some wild-type and hnn cilia (E and F).
The B-tubule of the hnn outer doublets hnn cilia is frequently open
(D and F).
Accompanying the shift of Pax6 to the dorsal neural tube,
dorsal cell types were disrupted in hnn embryos: expression of Wnt1, a roof plate marker, was discontinuous in
the caudal hnn embryo, and dorsal progenitors that normally express Math1 and Mash1 were not properly specified (Figure S2). Thus the expansion of the motor neuron
domain in hnn embryos occurred at the expense of both
more dorsal and more ventral neural cell types.
To help define how Arl13b affected Shh signaling, we
generated double mutants that lacked both Arl13b and
Smo, the membrane protein that promotes the response
to Hh ligands. Smo mutants lack both Shh and Ihh signaling and have a slightly stronger effect on neural patterning than Shh mutants (Caspary et al., 2002; Zhang et al.,
2001). Like hnn, the Smo hnn double mutants expressed
Olig2 in the ventral two-thirds of the neural tube (Figures
6I, 6J, 6M, and 6N). Thus Arl13b must act at a step downstream of Smo in the cytoplasmic signal transduction
pathway.
As expected because Ptch1 acts upstream of Smo, and
Smo appeared to act upstream of Arl13b, loss of Arl13b
modified the Ptch1 phenotype. In Ptch1 mutant embryos,
the entire neural plate is specified as floor plate, the most
ventral cell type (Goodrich et al., 1997). In Ptch1 hnn double mutant embryos, floor plate cells (expressing FoxA2)
and motor neuron precursors (expressing Olig2) were
both specified, intermixed at all positions along the dorsal-ventral axis (Figures 6A–6D and 6I–6L). Thus loss of
Arl13b blocks the complete activation of the pathway
caused by loss of Ptch1, and Arl13b must act downstream
of Ptch1. As floor plate cells were not specified in hnn single mutants, the specification of floor plate cells in Ptch1
hnn double mutants indicates that hnn embryos retain
a limited ability to respond to changes in the activity of
upstream components of the Hh pathway.
Figure 5. Dorsal-Ventral Neural Tube Patterning in wim hnn
Double Mutants
Expression of Pax 6 (A, D, G, and J), Lhx3 (B, E, H, and K), and Isl1/2 (C,
F, I, and L) in e11.5 neural tube sections from wild-type (A–C), hnn
(D–F), wim (G–I), and hnn wim (J–L) double mutant embryos (G–L).
The hnn wim double mutant is indistinguishable from the wim single
mutant.
The hnn Phenotype Is Associated with Altered
Gli Activity
Because the specification of the ventral neural cell
types affected in hnn embryos depends on the Gli2 and
Gli3 transcription factors, we investigated whether
Arl13b was required for normal regulation of Gli2, Gli3,
or both.
Gli2, the primary transcriptional activator for Hh signaling in the neural tube, is required for specification of the
floor plate and the normal number of Nkx2.2-expressing
V3 interneuron progenitors (Ding et al., 1998; Matise
et al., 1998). hnn mutants lack a floor plate, which suggested that hnn embryos lacked the high level of Gli2
activator required for floor plate specification. The number
of Nkx2.2-expressing cells was decreased in the Gli2 hnn
double mutant compared to hnn single mutants (Figures 7B and 7F), which indicated that Gli2 had some activity in hnn mutants.
As in hnn single mutants, the domain of motor neurons
and their progenitors expanded dorsally in the Gli2 hnn
neural tube (Figure 7H and data not shown); thus the level
of Gli2 activator required for the initial specification of
772 Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc.
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Figure 6. Dorsal-Ventral Patterning in Ptch1 hnn and Smobnb hnn Double Neural Tubes
Expression of FoxA2 (A–D), Nkx2.2 (E–H), and Olig2 (I–N) in e9.5 neural tube sections from wild-type (A, E, I), hnn (B, F, J), Ptch1 (C and K), Ptch1 hnn
double mutant (D and L), Smo (G and M), and Smo hnn double mutant (H and N) embryos. In both Ptch1 hnn and Smo hnn double mutants, motor
neuron progenitors and more ventral cell types are intermixed in the ventral two-thirds of the neural tube.
motor neurons was present in the hnn neural tube. However, there were fewer mature motor neurons in the expanded motor neuron domain of the Gli2 hnn double mutants than in hnn single mutants (Figures 7D and 7H). It has
been shown that the activator functions of both Gli2 and
Gli3 promote the differentiation of motor neurons (Bai
et al., 2004; Lei et al., 2004). We therefore think it likely
that the reduction in the number of mature motor neurons
in Gli2 hnn double mutants reflects the loss of Gli2 coupled
with decreased levels of Gli3 activator.
Double mutants that lacked both Gli3 and hnn showed
a greater expansion of ventrolateral cell types than hnn
single mutants: the motor neuron domain in the double
mutants expanded to include the entire neural tube (Figures 7D and 7L). Similarly, the Nkx2.2 expression domain
of Gli3 hnn double mutants expanded to include all but
the most dorsal neural tube (Figure 7J). Thus in the
absence of both Gli3 and Arl13b, there is essentially no
patterning of cell types along the dorsal-ventral axis, and
nearly all neural cells assumed one of two ventrolateral
identities, that of motor neurons or V3 interneurons. The
double mutant phenotype suggests that the residual polarity present in hnn embryos is due to asymmetric activity
of Gli3 repressor, with less Gli3 repressor ventrally than
dorsally.
In the absence of Shh, Gli3 is processed to make a
transcriptional repressor that prevents transcription of target genes in the neural tube. Unprocessed Gli3 is present
in cilia (Haycraft et al., 2005), and Gli3 processing is
blocked in mutants that lack cilia due to IFT mutations
(Haycraft et al., 2005; Huangfu and Anderson, 2005; Liu
et al., 2005). However, processing of Gli3 appeared to
be normal in hnn embryos (Figure 7), despite their abnormally structured cilia. Thus while normal activity of Gli2
requires Arl13b, Gli3 repressor activity does not depend
on Arl13b.
DISCUSSION
The hennin Phenotype Is Due to Loss of Function
of Arl13b
The hnn mutant was identified on the basis of its abnormal
morphology and abnormal neural patterning. Several lines
of evidence demonstrate that the hnn mutant phenotypes
are the result of loss of function of Arl13b. The hnn mutation mapped to an interval containing only five genes, and
a mutation was found only in Arl13b. This splice site mutation does not result in the production of stable Arl13b protein as demonstrated by the lack of protein in hnn mutants
by both immunofluorescence and western blot analysis.
Furthermore, the Arl13b protein is localized to cilia, and
cilia are defective in the hnn mutant. Additional support
for the role of Arl13b in cilia comes from the zebrafish scorpion mutation, which is caused by a retroviral insertion into
the zebrafish ARL13B gene; the scorpion mutation causes
cystic kidneys, and kidney cilia were not detectable in
scorpion mutants (Sun et al., 2004).
The Function of Arl13b in Cilia Structure
The Arf/Arl family within the Ras superfamily of small
GTPases is defined by sequence similarity and not by
a common cellular function (Kahn et al., 2006). Of the 30
Arf/Arl mammalian proteins, several have been shown to
affect vesicle trafficking, and three have been implicated
in cilia and microtubule dynamics (Hoyt et al., 1990;
Kahn et al., 2006; Van Valkenburgh et al., 2001; Vitale
et al., 1998). Arl3 is localized to cilia (Zhou et al., 2006),
and genetic and biochemical studies have suggested
that Arl2 regulates microtubule assembly (Antoshechkin
and Han, 2002; Bhamidipati et al., 2000; Cuvillier et al.,
2000; Li et al., 2004; Radcliffe et al., 2000; Zhou et al.,
2006). Arl3 regulates flagellar synthesis in flagellated protozoans (Cuvillier et al., 2000). Arl6 is mutated in human
Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc. 773
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Mice that lack the transcription factor Rfx3 have shortened node cilia with normal ultrastructure (Bonnafe
et al., 2004). These animals show randomized left/right
asymmetry, but Shh signaling must be intact as these
animals are viable. Therefore, it is likely that the abnormal
axoneme structure caused by loss of Arl13b, rather
than shorter length of cilia, leads to the disruption of Hh
signaling.
The open B-tubule in the cilia of hnn mutants indicates
that Arl13b is required for a specific aspect of axoneme
structure. Structural and biochemical studies have identified several nontubulin proteins that are enriched at the
junction of the A- and B-tubules of the outer doublet,
and suggest that specific protein complexes are required
to link the B-tubule to the A-tubule (Linck and Norrander,
2003; Nicastro et al., 2006; Stephens, 2000; Sui and
Downing, 2006). Arl13b could be a protein in the 11th filament of the B-tubule (Linck and Stephens, 2007; Tilney
et al., 1973), or it may regulate the assembly of a protein(s)
that links the two tubules of the axonemal outer doublet.
Figure 7. Dorsal-Ventral Neural Tube Patterning in Shh hnn,
Gli2 hnn, and Gli3 hnn Double Mutants
Expression of Nkx2.2 (A, B, E, F, I, and J) and HB9 (C, D, G, H, K, and L)
in e10.5 neural tube sections from wild-type (A and C), hnn (B and D),
Gli2 (E and G), Gli2 hnn (F and H), Gli3 (I and K), and Gli3 hnn (J and L)
embryos. Western analysis of Gli3 processing in protein extracts from
e10.5 wild-type, hnn, and Gli3 embryos. Similar amounts of full-length
Gli3 (190 kDa) are cleaved to the repressor form (83 kDa) in the wildtype and hnn extracts. No Gli3 is detected in extracts from Gli3
mutants.
Bardet-Biedl Syndrome 3 (BBS3), a disease associated
with defects in basal bodies and cilia (Fan et al., 2004).
An Arl13b-related protein in C. elegans localizes to ciliated
neurons (Fan et al., 2004), although this protein does not
include the long C-terminal tail of Arl13b. The precise biochemical connections between Arl proteins and microtubules have not been defined, but the association of several Arl proteins with cilia suggests that some of these
proteins regulate specific aspects of cilia structure.
We identified two defects in the structure of nodal cilia
that lack Arl13b: they are half the normal length and the axoneme has an abnormal structure. Unlike IFT proteins,
which are enriched at the base of the cilium and overlap
in distribution with the basal body protein g-tubulin (Deane
et al., 2001; Taulman et al., 2001), Arl13b is found along
the length of the axoneme and does not overlap with gtubulin (Figures 3M–3O). This localization suggests that
Arl13b plays a direct role in the assembly or maintenance
of the axoneme.
The Function of Arl13b in Hh Signaling
The hnn phenotype does not correspond to either a simple
decrease or increase in the activity of the Hh pathway, as
the cells that require the highest Hh activity fail to be specified, while cells that require intermediate Hh activity are
found in an expanded domain, a phenotype that has not
been seen in other mouse mutants. It has been suggested
that cilia may be important for signaling pathways other
than the Hh pathway (Davis et al., 2006), so the complexity
of the hnn phenotype could, in principle, reflect the
disruption of more than one signaling pathway. Although
we cannot rule out the possibility that the hnn mutation
affects multiple signaling pathways, the altered patterning
of neural cell types seen in hnn embryos parallels the expanded, shallow gradient of Ptch-lacZ expression in the
mutant neural tube, and Ptch-lacZ expression is a direct
readout of Hh pathway activity. We therefore conclude
that the most parsimonious interpretation of the hnn phenotype is that loss of Arl13b specifically disrupts the Shh
signaling pathway so that the pathway has an intermediate level of activity in an expanded domain.
Although the shallow Ptch1 expression gradient and
expanded domain of motor neurons in the hnn neural
tube could arise if the Shh ligand were spread over a larger
region of the neural tube, the phenotype of the Shh hnn
double mutant shows that the hnn phenotype does not
depend on the presence of Shh ligand. Furthermore, the
double mutants with Ptch1 and Smo indicate that Arl13b
acts downstream of Smo, the same step in the pathway
that requires IFT proteins (Huangfu et al., 2003).
The hnn phenotype is reminiscent of that caused by loss
of the zebrafish iguana gene, a component of the zebrafish
Hh pathway. The iguana gene encodes DZIP1, a large
protein with coiled-coil domains and a single zinc finger
(Sekimizu et al., 2004; Wolff et al., 2004). hnn and iguana
mutants both show both a failure to activate the
highest responses to Hh and an expanded domain of Hh
response. Like hnn, the iguana phenotype is ligand
774 Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc.
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
independent. It has been proposed that the complex phenotype of iguana embryos arises because DZIP1 controls
nuclear import of both Gli activator and repressor forms,
or that DZIP1 acts in the cytoplasm to modulate that activity of Sufu, a negative regulator of Hh signaling (Sekimizu
et al., 2004; Wolff et al., 2004).
The neural tube of the mouse embryo provides a good
context to understand phenotypes of the hnn phenotype
because it is known that the identity of ventral neural cell
types is defined by the ratio of Gli activator/Gli repressor
and that changes in the effective level of either the activator or the repressor can alter the Gli activity gradient. The
floor plate and the highest level of Ptch1 expression both
depend on high levels of Gli2 activator (Ding et al., 1998;
Eggenschwiler et al., 2006, 2001; Matise et al., 1998)
and both are lost in hnn mutants, which indicates that
the highest level of Gli2 activator is not achieved in hnn
embryos. In contrast, the dorsal expansion of motor neuron progenitors and Ptch1 expression would correspond
to an increase in the ratio of Gli activator/Gli repressor in
lateral and dorsal neural cells. The dorsal expansion of
motor neurons cannot be due simply to ectopic Gli2
activity, as the expanded domain is still present in hnn
Gli2 double mutants. The dorsal expansion is therefore
likely to be due to changes in the activity of both Gli2
and Gli3.
Gli3 processing appears to be normal in hnn embryos,
even though IFT proteins (and therefore cilia) are required
for normal Gli3 processing (Huangfu and Anderson, 2005;
Liu et al., 2005; May et al., 2005). Although the hnn mutation affects neural tube patterning only at posterior positions, decreased Gli3 processing was detected in mutants
that lack Rab23 (Eggenschwiler et al., 2006), which also
affects patterning in only the caudal neural tube. Therefore, it appears that hnn cilia can promote Gli3 processing
despite their abnormal structure. In addition, the strong
enhancement of the hnn phenotype seen in hnn Gli3 double mutant embryos demonstrates that Gli3 repressor is
functional in hnn embryos. The data are thus consistent
with the hypothesis that neither Gli3 repressor production
nor Gli3 repressor activity is affected by loss of Arl13b. We
therefore conclude it is likely that the dorsal expansion of
the motor neuron domain is due to ectopic activity of both
Gli2 and Gli3 activators and that the hnn phenotype is the
result of abnormal regulation of Gli activators, without
changes in Gli3 repressor. This indicates that two ciliadependent events, the events that promote formation of
Gli activators and those that promote production of Gli3
repressor, can be regulated independently: the events
that lead to formation of Gli3 repressor can occur in the
abnormal cilia that form in the absence of Arl13b, but
those that regulate the production of Gli activators cannot.
In mutants that lack cilia altogether, such as Ift172 and
Ift88 mutants, no Gli activator is produced (Haycraft et al.,
2005; Huangfu and Anderson, 2005; Liu et al., 2005). In
contrast, in hnn mutants, it appears there is a constitutive
low level of Gli activator in all cells in the neural tube. We
therefore propose that there are two cilia-dependent
steps in the formation of Gli activators. We propose that
full-length Gli proteins are modified within the cilium in
a Hh-independent process to have a low level of activator
function, and the low-level activator is normally tethered in
the cilium in the absence of ligand. The idea that there are
ligand-independent steps in Hh signaling that depend on
cilia is not novel: processing of Gli3 to make Gli3 repressor
is also cilia dependent and occurs in the absence of ligand
(Huangfu and Anderson, 2005). In response to Hh, fulllength Gli proteins can be further modified to create
high-level activator, which is then released from the cilium
to the nucleus. In the abnormal cilia that lack Arl13b, the
modification that produces full-length Gli protein with
low-level activity takes place, but this low-level activator
is not effectively tethered in the cilium and is released inappropriately to the nucleus in the absence of Hh ligand.
This step retains a small amount of sensitivity to upstream
signals in hnn mutant, as the phenotype of Ptch1 hnn double mutants is not identical to the hnn phenotype.
Two alternative views of the relationship between cilia
and Hh signaling have been proposed (Scholey and
Anderson, 2006). In the simpler view, cilia represent
a site where Hh pathway components are enriched, and
the high local concentration of the proteins allows efficient
signaling transduction. Alternatively, dynamic trafficking
within the cilium may allow a sequence of protein interactions that promote the activity of the pathway. The
hnn phenotype supports the latter model, as it reveals a
complexity of events that occur within the cilium.
EXPERIMENTAL PROCEDURES
Mouse Strains
We identified hnn in a screen for recessive N-ethyl, N-nitrosourea
mutations that caused morphological defects at e9.5 (Garcı́a-Garcı́a
et al., 2005). Other mouse strains used were: Shh, Ptch1tm1Mps,
Ptch1 (lacZ reporter, D allele, M.P. Scott, personal communication),
Smobnb, Rab23opb2, wim, Gli2, and Gli3Xt-J (Caspary et al., 2002;
Chiang et al., 1996; Eggenschwiler and Anderson, 2000; Goodrich
et al., 1997; Huangfu et al., 2003; Hui and Joyner, 1993; Matise
et al., 1998). All phenotypes were analyzed in the C3H background,
and crosses and genotyping were performed as described (Caspary
et al., 2002; Chiang et al., 1996; Eggenschwiler and Anderson, 2000;
Goodrich et al., 1997; Huangfu et al., 2003; Hui and Joyner, 1993;
Matise et al., 1998).
Genetic Mapping and Molecular Identification of hnn
MIT SSLP markers were used to map hnn on mouse chromosome 16.
Additional polymorphic markers were generated for high-resolution
mapping (http://mouse.ski.mskcc.org). In a mapping cross of 1283
opportunities for recombination, hnn was mapped to a 2.08 MB interval between D16SKI147 (D16SKI147F: AATGCCTCAAGTGCCTCTTT;
D16SKI147R: GGGACTCATCTTTGGGAACA) and D16SKI174
(D16SKI174F: TGTGGGTGGCATATGTAGGA; D16SKI174R: GCTAGC
TATTTTCTGTTGCTGGA). The only sequence change detected in the
interval was a T-to-G substitution in the splice acceptor site of exon
2 of the Arl13b transcription unit.
Antibodies were raised in rabbit to a purified bacterially expressed
protein of GST fused to amino acids 208–428 of Arl13b. Serum was
used at 1:1500 dilution for immunohistochemistry. The serum was preabsorbed against the fusion protein for western analysis and used
at 1:5000 along with an anti-actin Ab (Sigma, 1:2000). Three bands
were detected in e10.5 whole-embryo extracts and purified antigen
Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc. 775
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
specifically competed away the 48 kDa and !60 kDa bands, indicating
that the remaining !70 kDa band in hnn extracts is not specific.
Antoshechkin, I., and Han, M. (2002). The C. elegans evl-20 gene is
a homolog of the small GTPase ARL2 and regulates cytoskeleton dynamics during cytokinesis and morphogenesis. Dev. Cell 2, 579–591.
Phenotypic Analysis
In situ hybridization and X-gal staining on whole-mount embryos,
Alcian blue staining, immunofluorescence, and scanning electron microscopy (SEM) were done as previously described (Huangfu et al.,
2003; Shen et al., 1997). For SEM, the embryos were viewed with
a LEO1550 Field Emission Scanning Electron Microscope at 2.5 kv.
For transmission electron microscopy (TEM), samples were dissected
in fresh 2% paraformaldehyde/2.5% glutaraldehyde/0.1 M cacodylate
buffer (pH 7.4) and fixed overnight at room temperature. After rinsing
with 0.1 M cacodylate buffer, the embryos were postfixed with 1%
OsO4/0.1 M cacodylate buffer for 1 hr on ice, dehydrated with a graded
alcohol series and propylene oxide, and embedded in Durcapan (EM
Sciences, Fort Washington, PA). Ultrathin sections were cut on
a Reichert Ultracut E, poststained with Uranyl Acetate and Reynold’s
lead, and viewed with a JEOL CX100 at 80 kv.
Antibodies used were: Olig2 (gift of D. Rowitch, 1:1000), acetylated
a-tubulin (Sigma, 1:1000), g-tubulin (Sigma, 1:1000), phospho histone
H3 (Upstate, 1:250), Shh, FoxA2, Nkx2.2, Pax6, Isl1/2, Lhx3, HB9
(Developmental Studies Hybridoma Bank), all were used at a dilution
of 1:10, except Nkx2.2 which was used at a dilution of 1:5.
Gli3 processing was analyzed in e10.5 whole-embryo extracts as
described previously (Huangfu and Anderson, 2005).
Bai, C.B., Stephen, D., and Joyner, A.L. (2004). All mouse ventral spinal
cord patterning by hedgehog is Gli dependent and involves an activator function of Gli3. Dev. Cell 6, 103–115.
Primary Mouse Embryonic Fibroblast Lines
PMEFs were isolated from e12.5 wild-type and hnn mutant mouse and
grown on gelatinized plates. At confluence, cells were split 1:5 into
culture dishes containing polylysine-coated coverslips. After reaching
confluence, cells were serum starved for 24 hr followed by processing
and antibody staining.
Supplemental Data
Supplemental Data include two figures and can be found with this
article online at http://www.developmentalcell.com/cgi/content/full/
12/5/767/DC1/.
ACKNOWLEDGMENTS
We thank Nina Lampen and Eleana Sphicas for technical support for
the electron microscopy; Michael Hillman and Danwei Huangfu for
western analysis; Matthew Scott, Phillip Beachy, and Alexandra Joyner for mice; Danwei Huangfu, David Katz, Jeffrey Lee, Richard Linck,
Richard Kahn, Polloneal Jymmiel Ocbina, and Andrew Rakeman for
helpful comments on the manuscript. Monoclonal antibodies were obtained from the Developmental Studies Hybridoma Bank, which was
developed under the auspices of the National Institute of Child Health
and Human Development and is maintained by The University of Iowa,
Department of Biological Sciences. Genome sequence analysis used
Ensembl, the UCSC Genome Browser, and the Celera Discovery
System, made possible in part by the AMDeC Foundation. T.C. was
a recipient of a Burroughs-Wellcome Fund Hitchings Elion Career
Development Fellowship and a Muscular Dystrophy Association
Development Grant. This work was supported by NIH grant NS44385
to K.V.A.
Received: November 30, 2006
Revised: February 28, 2007
Accepted: March 5, 2007
Published: May 7, 2007
REFERENCES
Agren, M., Kogerman, P., Kleman, M.I., Wessling, M., and Toftgard, R.
(2004). Expression of the PTCH1 tumor suppressor gene is regulated
by alternative promoters and a single functional Gli-binding site.
Gene 330, 101–114.
Behnke, O., and Forer, A. (1967). Evidence for four classes of microtubules in individual cells. J. Cell Sci. 2, 169–192.
Bhamidipati, A., Lewis, S.A., and Cowan, N.J. (2000). ADP ribosylation
factor-like protein 2 (Arl2) regulates the interaction of tubulin-folding
cofactor D with native tubulin. J. Cell Biol. 149, 1087–1096.
Bonnafe, E., Touka, M., AitLounis, A., Baas, D., Barras, E., Ucla, C.,
Moreau, A., Flamant, F., Dubruille, R., Couble, P., et al. (2004). The
transcription factor RFX3 directs nodal cilium development and leftright asymmetry specification. Mol. Cell. Biol. 24, 4417–4427.
Briscoe, J., Pierani, A., Jessell, T.M., and Ericson, J. (2000). A homeodomain protein code specifies progenitor cell identity and neuronal
fate in the ventral neural tube. Cell 101, 435–445.
Caspary, T., Garcı́a-Garcı́a, M.J., Huangfu, D., Eggenschwiler, J.T.,
Wyler, M.R., Rakeman, A.S., Alcorn, H.L., and Anderson, K.V. (2002).
Mouse Dispatched homologue1 is required for long-range, but not
juxtacrine, Hh signaling. Curr. Biol. 12, 1628–1632.
Chiang, C., Litingtung, Y., Lee, E., Young, K.E., Corden, J.L., Westphal,
H., and Beachy, P.A. (1996). Cyclopia and defective axial patterning in
mice lacking Sonic hedgehog gene function. Nature 383, 407–413.
Corbit, K.C., Aanstad, P., Singla, V., Norman, A.R., Stainier, D.Y., and
Reiter, J.F. (2005). Vertebrate Smoothened functions at the primary
cilium. Nature 437, 1018–1021.
Cuvillier, A., Redon, F., Antoine, J.C., Chardin, P., DeVos, T., and
Merlin, G. (2000). LdARL-3A, a Leishmania promastigote-specific
ADP-ribosylation factor-like protein, is essential for flagellum integrity.
J. Cell Sci. 113, 2065–2074.
Dai, P., Akimaru, H., Tanaka, Y., Maekawa, T., Nakafuku, M., and Ishii,
S. (1999). Sonic Hedgehog-induced activation of the Gli1 promoter is
mediated by GLI3. J. Biol. Chem. 274, 8143–8152.
Davis, E.E., Brueckner, M., and Katsanis, N. (2006). The emerging
complexity of the vertebrate cilium: new functional roles for an ancient
organelle. Dev. Cell 11, 9–19.
Deane, J.A., Cole, D.G., Seeley, E.S., Diener, D.R., and Rosenbaum,
J.L. (2001). Localization of intraflagellar transport protein IFT52 identifies basal body transitional fibers as the docking site for IFT particles.
Curr. Biol. 11, 1586–1590.
Ding, Q., Motoyama, J., Gasca, S., Mo, R., Sasaki, H., Rossant, J., and
Hui, C.C. (1998). Diminished Sonic hedgehog signaling and lack of
floor plate differentiation in Gli2 mutant mice. Development 125,
2533–2543.
Dunaeva, M., Michelson, P., Kogerman, P., and Toftgard, R. (2003).
Characterization of the physical interaction of Gli proteins with SUFU
proteins. J. Biol. Chem. 278, 5116–5122.
Echelard, Y., Epstein, D.J., St-Jacques, B., Shen, L., Mohler, J.,
McMahon, J.A., and McMahon, A.P. (1993). Sonic hedgehog, a
member of a family of putative signaling molecules, is implicated in
the regulation of CNS polarity. Cell 75, 1417–1430.
Eggenschwiler, J.T., and Anderson, K.V. (2000). Dorsal and lateral
fates in the mouse neural tube require the cell-autonomous activity
of the open brain gene. Dev. Biol. 227, 648–660.
Eggenschwiler, J.T., Bulgakov, O.V., Qin, J., Li, T., and Anderson, K.V.
(2006). Mouse Rab23 regulates Hedgehog signaling from Smoothened
to Gli proteins. Dev. Biol. 290, 1–12.
Eggenschwiler, J.T., Espinoza, E., and Anderson, K.V. (2001). Rab23 is
an essential negative regulator of the mouse Sonic hedgehog signalling pathway. Nature 412, 194–198.
Eley, L., Yates, L.M., and Goodship, J.A. (2005). Cilia and Disease.
Curr. Opin. Gen. Dev. 15, 308–314.
776 Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc.
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Ericson, J., Briscoe, J., Rashbass, P., van Heyningen, V., and Jessell,
T.M. (1997). Graded Sonic Hedgehog signaling and the specification
of cell fate in the ventral neural tube. Cold Spring Harb. Symp. Quant.
Biol. 62, 451–466.
Fan, Y., Esmail, M.A., Ansley, S.J., Blacque, O.E., Boroevich, K., Ross,
A.J., Moore, S.J., Badano, J.L., May-Simera, H., Compton, D.S., et al.
(2004). Mutations in a member of the Ras superfamily of small GTPbinding proteins causes Bardet-Biedl syndrome. Nat. Genet. 36,
989–993.
Feistel, K., and Blum, M. (2006). Three types of cilia including a novel
9+4 axoneme on the notochordal plate of the rabbit embryo. Dev.
Dyn. 235, 3348–3358.
Garcı́a-Garcı́a, M.J., Eggenschwiler, J.T., Caspary, T., Alcorn, H.L.,
Wyler, M.R., Huangfu, D., Rakeman, A.S., Lee, J.D., Feinberg, E.H.,
Timmer, J.R., and Anderson, K.V. (2005). Analysis of mouse embryonic
patterning and morphogenesis by forward genetics. Proc. Natl. Acad.
Sci. USA 102, 5913–5919.
Goodrich, L.V., Milenkovic, L., Higgins, K.M., and Scott, M.P. (1997).
Altered neural cell fates and medulloblastoma in mouse Patched
mutants. Science 277, 1109–1113.
Haycraft, C.J., Banizs, B., Aydin-Son, Y., Zhang, Q., Michaud, E.J.,
and Yoder, B.K. (2005). Gli2 and Gli3 localize to cilia and require the intraflagellar transport protein polaris for processing and function. PLoS
Genet. 1, e53.
Houde, C., Dickinson, R.J., Houtzager, V.M., Cullum, R., Montpetit, R.,
Metzler, M., Simpson, E.M., Roy, S., Hayden, M.R., Hoodless, P.A.,
and Nicholson, D.W. (2006). Hippi is essential for node cilia assembly
and Sonic hedgehog signaling. Dev. Biol. 300, 523–533.
Hoyt, M.A., Stearns, T., and Botstein, D. (1990). Chromosome instability mutants of Saccharomyces cerevisiae that are defective in microtubule-mediated processes. Mol. Cell. Biol. 10, 223–234.
Huangfu, D., and Anderson, K.V. (2005). Cilia and Hedgehog responsiveness in the mouse. Proc. Natl. Acad. Sci. USA 102, 11325–11330.
Huangfu, D., Liu, A., Rakeman, A.S., Murcia, N.S., Niswander, L., and
Anderson, K.V. (2003). Hedgehog signalling in the mouse requires intraflagellar transport proteins. Nature 426, 83–87.
Hui, C.C., and Joyner, A.L. (1993). A mouse model of greig cephalopolysyndactyly syndrome: the extra-toesJ mutation contains an intragenic deletion of the Gli3 gene. Nat. Genet. 3, 241–246.
Kahn, R.A., Cherfils, J., Elias, M., Lovering, R.C., Munro, S., and Schurmann, A. (2006). Nomenclature for the human Arf family of GTP-binding proteins: ARF, ARL, and SAR proteins. J. Cell Biol. 172, 645–650.
Lei, Q., Zelman, A.K., Kuang, E., Li, S., and Matise, M.P. (2004). Transduction of graded Hedgehog signaling by a combination of Gli2 and
Gli3 activator functions in the developing spinal cord. Development
131, 3593–3604.
Li, Y., Kelly, W.G., Logsdon, J.M., Jr., Schurko, A.M., Harfe, B.D., HillHarfe, K.L., and Kahn, R.A. (2004). Functional genomic analysis of the
ADP-ribosylation factor family of GTPases: phylogeny among diverse
eukaryotes and function in C. elegans. FASEB J. 18, 1834–1850.
Liu, A., Wang, B., and Niswander, L.A. (2005). Mouse intraflagellar
transport proteins regulate both the activator and repressor functions
of Gli transcription factors. Development 132, 3103–3111.
Marigo, V., Davey, R.A., Zuo, Y., Cunningham, J.M., and Tabin, C.J.
(1996). Biochemical evidence that patched is the Hedgehog receptor.
Nature 384, 176–179.
Matise, M.P., Epstein, D.J., Park, H.L., Platt, K.A., and Joyner, A.L.
(1998). Gli2 is required for induction of floor plate and adjacent cells,
but not most ventral neurons in the mouse central nervous system.
Development 125, 2759–2770.
May, S.R., Ashique, A.M., Karlen, M., Wang, B., Shen, Y., Zarbalis, K.,
Reiter, J., Ericson, J., and Peterson, A.S. (2005). Loss of the retrograde
motor for IFT disrupts localization of Smo to cilia and prevents the
expression of both activator and repressor functions of Gli. Dev.
Biol. 287, 378–389.
McGrath, J., Somlo, S., Makova, S., Tian, X., and Brueckner, M. (2003).
Two populations of node monocilia initiate left-right asymmetry in the
mouse. Cell 114, 61–73.
Merchant, M., Vajdos, F.F., Ultsch, M., Maun, H.R., Wendt, U.,
Cannon, J., Desmarais, W., Lazarus, R.A., de Vos, A.M., and de Sauvage, F.J. (2004). Suppressor of fused regulates Gli activity through
a dual binding mechanism. Mol. Cell. Biol. 24, 8627–8641.
Motoyama, J., Milenkovic, L., Iwama, M., Shikata, Y., Scott, M.P., and
Hui, C.C. (2003). Differential requirement for Gli2 and Gli3 in ventral
neural cell fate specification. Dev. Biol. 259, 150–161.
Nicastro, D., Schwartz, C., Pierson, J., Gaudette, R., Porter, M.E., and
McIntosh, J.R. (2006). The molecular architecture of axonemes
revealed by cryoelectron tomography. Science 313, 944–948.
Persson, M., Stamataki, D., Te Welscher, P., Andersson, E., Bose, J.,
Ruther, U., Ericson, J., and Briscoe, J. (2002). Dorsal-ventral patterning of the spinal cord requires Gli3 transcriptional repressor activity.
Genes Dev. 16, 2865–2878.
Radcliffe, P.A., Vardy, L., and Toda, T. (2000). A conserved small GTPbinding protein Alp41 is essential for the cofactor-dependent biogenesis of microtubules in fission yeast. FEBS Lett. 468, 84–88.
Ruiz i Altaba, A. (1998). Combinatorial Gli gene function in floor plate
and neuronal inductions by Sonic hedgehog. Development 125,
2203–2212.
Sasaki, H., Hui, C., Nakafuku, M., and Kondoh, H. (1997). A binding site
for Gli proteins is essential for HNF-3beta floor plate enhancer activity
in transgenics and can respond to Shh in vitro. Development 124,
1313–1322.
Scholey, J.M. (2003). Intraflagellar transport. Annu. Rev. Cell Dev. Biol.
19, 423–443.
Scholey, J.M., and Anderson, K.V. (2006). Intraflagellar transport and
cilium-based signaling. Cell 125, 439–442.
Sekimizu, K., Nishioka, N., Sasaki, H., Takeda, H., Karlstrom, R.O., and
Kawakami, A. (2004). The zebrafish iguana locus encodes Dzip1,
a novel zinc-finger protein required for proper regulation of Hedgehog
signaling. Development 131, 2521–2532.
Linck, R.W., and Norrander, J.M. (2003). Protofilament ribbon compartments of ciliary and flagellar microtubules. Protist 154, 299–311.
Shen, J., Bronson, R.T., Chen, D.F., Xia, W., Selkoe, D.J., and Tonegawa, S. (1997). Skeletal and CNS defects in Presenilin-1-deficient
mice. Cell 89, 629–639.
Linck, R.W., and Stephens, R.E. (2007). Functional Protofilament
Numbering of Ciliary, Flagellar and Centriolar Microtubules. Cell Motil.
Cytoskeleton, in press. Published online March 15, 2007. 10.1002/cm.
20202.
Stamataki, D., Ulloa, F., Tsoni, S.V., Mynett, A., and Briscoe, J. (2005).
A gradient of Gli activity mediates graded Sonic Hedgehog signaling in
the neural tube. Genes Dev. 19, 626–641.
Litingtung, Y., and Chiang, C. (2000). Specification of ventral neuron
types is mediated by an antagonistic interaction between Shh and
Gli3. Nat. Neurosci. 3, 979–985.
Litingtung, Y., Dahn, R.D., Li, Y., Fallon, J.F., and Chiang, C. (2002).
Shh and Gli3 are dispensable for limb skeleton formation but regulate
digit number and identity. Nature 418, 979–983.
Stephens, R.E. (1970). Thermal fractionation of outer fiber doublet
microtubules into A- and B-subfiber components. A- and B-tubulin.
J. Mol. Biol. 47, 353–363.
Stephens, R.E. (2000). Preferential incorporation of tubulin into the
junctional region of ciliary outer doublet microtubules: a model for
treadmilling by lattice dislocation. Cell Motil. Cytoskeleton 47, 130–
140.
Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc. 777
Developmental Cell
Cilia Architecture and Sonic Hedgehog Signaling
Sui, H., and Downing, K.H. (2006). Molecular architecture of axonemal
microtubule doublets revealed by cryo-electron tomography. Nature
442, 475–478.
Wang, B., Fallon, J.F., and Beachy, P.A. (2000). Hedgehog-regulated
processing of Gli3 produces an anterior/posterior repressor gradient
in the developing vertebrate limb. Cell 100, 423–434.
Sun, Z., Amsterdam, A., Pazour, G.J., Cole, D.G., Miller, M.S., and
Hopkins, N. (2004). A genetic screen in zebrafish identifies cilia genes
as a principal cause of cystic kidney. Development 131, 4085–4093.
Wheatley, D.N. (1995). Primary cilia in normal and pathological tissues.
Pathobiology 63, 222–238.
Takeda, S., Yonekawa, Y., Tanaka, Y., Okada, Y., Nonaka, S., and
Hirokawa, N. (1999). Left-right asymmetry and kinesin superfamily protein KIF3A: new insights in determination of laterality and mesoderm
induction by kif3A"/" mice analysis. J. Cell Biol. 145, 825–836.
Taulman, P.D., Haycraft, C.J., Balkovetz, D.F., and Yoder, B.K. (2001).
Polaris, a protein involved in left-right axis patterning, localizes to basal
bodies and cilia. Mol. Biol. Cell 12, 589–599.
Tilney, L.G., Bryan, J., Bush, D.J., Fujiwara, K., Mooseker, M.S.,
Murphy, D.B., and Snyder, D.H. (1973). Microtubules: evidence for
13 protofilaments. J. Cell Biol. 59, 267–275.
Van Valkenburgh, H., Shern, J.F., Sharer, J.D., Zhu, X., and Kahn, R.A.
(2001). ADP-ribosylation factors (ARFs) and ARF-like 1 (ARL1) have
both specific and shared effectors: characterizing ARL1-binding
proteins. J. Biol. Chem. 276, 22826–22837.
Vitale, N., Horiba, K., Ferrans, V.J., Moss, J., and Vaughan, M. (1998).
Localization of ADP-ribosylation factor domain protein 1 (ARD1) in
lysosomes and Golgi apparatus. Proc. Natl. Acad. Sci. USA 95,
8613–8618.
Wheatley, D.N., Wang, A.M., and Strugnell, G.E. (1996). Expression of
primary cilia in mammalian cells. Cell Biol. Int. 20, 73–81.
Wijgerde, M., McMahon, J.A., Rule, M., and McMahon, A.P. (2002). A
direct requirement for Hedgehog signaling for normal specification of
all ventral progenitor domains in the presumptive mammalian spinal
cord. Genes Dev. 16, 2849–2864.
Wolff, C., Roy, S., Lewis, K.E., Schauerte, H., Joerg-Rauch, G., Kirn,
A., Weiler, C., Geisler, R., Haffter, P., and Ingham, P.W. (2004). iguana
encodes a novel zinc-finger protein with coiled-coil domains essential
for Hedgehog signal transduction in the zebrafish embryo. Genes Dev.
18, 1565–1576.
Zhang, X.M., Ramalho-Santos, M., and McMahon, A.P. (2001).
Smoothened mutants reveal redundant roles for Shh and Ihh signaling
including regulation of L/R symmetry by the mouse node. Cell 106,
781–792.
Zhou, C., Cunningham, L., Marcus, A.I., Li, Y., and Kahn, R.A. (2006).
Arl2 and Arl3 regulate different microtubule-dependent processes.
Mol. Biol. Cell 17, 2476–2487.
778 Developmental Cell 12, 767–778, May 2007 ª2007 Elsevier Inc.