|Home | About | Journals | Submit | Contact Us | Français|
Intraflagellar transport (IFT) has been studied for decades in model systems such as Chlamydomonas and Caenorhabditis elegans. More recently, IFT has been investigated using genetic approaches in mammals using the mouse as a model system. Through such studies, a new appreciation of the importance of IFT and cilia in mammalian signal transduction has emerged. Specifically, IFT has been shown to play a key role in controlling signaling by Sonic and Indian Hedgehog (Hh) ligands. The effects of mutations in IFT components on Sonic Hh signaling in the embryo are complex and differ depending on the nature of the components, alleles, and tissues examined. For this reason, we provide a basis for analyzing the phenotype as a guide for those investigators who study IFT in cell culture or use invertebrate systems and wish to extend their studies to include development of the mouse embryo. We provide an overview of Sonic Hh-dependent tissue patterning in the developing neural tube and limb buds, the two systems in which it has been studied most extensively, and we show examples of how this patterning is disrupted by mutations in mouse IFT components.
Primary cilia are microtubule-based organelles found on nearly all mammalian cell types in interphase. As discussed elsewhere in this volume, assembly and disassembly of cilia and flagella is driven by intraflagellar transport (IFT), a bidirectional microtubule motor-driven transport process [for a review, see Pedersen and Rosenbaum (2008)]. IFT directs the anterograde transport of cargo-loaded protein complexes from the minus ends of axonemal microtubules at the base of cilia (the basal bodies) toward microtubule plus ends, where the cargo is used for ciliary assembly at their growing ends. IFT also mediates retrograde transport of protein complexes back to the cell body, allowing for recycling or degradation of the cargo and IFT machinery. IFT occurs through the movement of plus and minus end-directed motors, primarily kinesin II and cDynein1b. These motors function together with the IFT proteins, which are thought to act as adaptors linking cargo proteins to the motors. The IFT proteins have been biochemically classified as members of two large complexes, complex A and complex B (Cole et al., 1998). Based on their mutant phenotypes in a variety of systems, complex B IFT proteins are believed to have a more crucial role in anterograde transport, whereas complex A IFT proteins are thought to act primarily in retrograde transport (Blacque et al., 2006; Piperno et al., 1998), although this distinction is probably an oversimplification.
A relationship between IFT and mammalian Hedgehog (Hh) signaling was first uncovered in 2003 through analysis of mouse mutations affecting embryonic development that proved to be mutations in IFT proteins (Huangfu et al., 2003). Hedgehog is a secreted, lipid-modified protein that signals to cells in most metazoans to control tissue patterning, growth, and differentiation during embryogenesis (Jiang and Hui, 2008). Mammalian genomes harbor three Hh ligands: Sonic hedgehog (Shh), Indian hedgehog (Ihh), and desert hedgehog (Dhh). These signals primarily control transcriptional responses in the cells of target tissues by binding to a twelve-pass membrane receptor, Patched, thereby alleviating its repression of a seven-pass transmembrane protein called Smoothened (Smo, Fig. 1). Once activated, Smo regulates the Gli transcription factors (e.g., Gli1, Gli2, and Gli3 in mammals), which bind regulatory elements of target genes and regulate their expression. The Glis can act both as activators and as repressors of pathway targets, depending on the status of upstream Smo activity. For example, in the absence of Hh signals, Gli3 is proteolytically processed to a smaller transcriptional repressor form, Gli3Rep. Stimulation of the pathway blocks this processing and converts the Glis into transcriptional activators, collectively referred to as GliAct (Pan et al., 2006; Wang et al., 2000).
A growing body of evidence strongly suggests that IFT components function in mammalian Hh signaling in the context of primary cilia [reviewed in Eggenschwiler and Anderson (2007)]. This is supported by genetic data showing that at least five IFT proteins, subunits of both IFT motors, and a protein controlling the axonemal structure of the cilium (Arl13b), play key roles in Hh pathway regulation (Caspary et al., 2007; Huangfu and Anderson, 2005; Huangfu et al., 2003; Liu et al., 2005; May et al., 2005). Moreover, many established Hh-signaling components, such as Ptch1, Smo, SuFu, Gli1, Gli2, and Gli3, and several newly identified regulators, such as C2cd3, Tulp3, Ftm, and Evc, localize to primary cilia and/or basal bodies (Corbit et al., 2005; Haycraft et al., 2005; Hoover et al., 2008; Norman et al., 2009; Rohatgi et al., 2007; Ruiz-Perez et al., 2007; Vierkotten et al., 2007). Interestingly, for at least two key signaling components, Ptch1 and Smo, localization to primary cilia is controlled by exposure to Hh ligands. In the absence of signaling, Ptch1 localizes to the cilium and Smo localizes to intracellular vesicles, whereas this pattern is reversed when Ptch1 binds Hh ligand (Corbit et al., 2005; Rohatgi et al., 2007).
The two systems in the developing mouse embryo where Shh signaling has been studied the most are the developing neural tube, which gives rise to the brain and spinal cord, and the limbs (Jiang and Hui, 2008). In both cases, Shh is produced by a small group of cells and acts as a long-range signal within the target tissues to control cell fate and tissue growth.
During neural tube patterning, Shh is initially produced by the notochord that lies ventral to the neural tube and it forms a ventral (high) to dorsal (low) gradient within the neural tissue. Cells of the neural tube adopt distinct identities as a function of the Shh concentration to which they are exposed. The most ventral cells within the neural tube are specified to become the floor plate, which acts as a secondary source of Shh ligands, whereas cells positioned further away adopt other fates such as motor neurons or interneurons. Shh acts by inducing and repressing the expression of a number of transcription factors, such as members of the homeobox and bHLH families, within proliferating progenitor cells. In turn, these transcription factors act in a combinatorial fashion to direct distinct neural differentiation programs.
In the limb buds, Shh is used to control anterior–posterior (i.e., thumb to pinky) digit identity and to control limb bud growth (McGlinn and Tabin, 2006). Shh is produced by a group of mesenchymal cells in the posterior limb bud. This region is termed the zone of polarizing activity because of its ability to organize pattern throughout the rest of the tissue. In the limb bud, cells in proximity to the source of Shh (or cells which experience Shh signaling for longer periods of time) express different sets of genes and adopt different digit identities than those further away experiencing little or no Shh signaling. Shh controls differential gene expression in the limb primarily through inhibiting the formation of Gli3 repressor, Gli3Rep, although the Shh-dependent gene expression profile is also controlled by the GliAct forms (te Welscher et al., 2002; Vokes et al., 2008).
The regulation of gene expression by Shh signaling in the mouse embryo is quite complex. For example, some targets are induced by weak Shh signaling, yet are repressed by higher levels, whereas others are only induced by intermediate or high levels. In addition, targets can be positively regulated by the activator forms of the Gli transcription factors (GliAct), negatively regulated by the Gli repressor forms (Gli3Rep, in particular), or they may be regulated by both modes. Finally, some of the genes encoding components of the Shh signal transduction apparatus, such as Shh, Gli1, Gli3, and Ptch1, are also direct or indirect transcriptional targets of the pathway. Such regulation is thought to provide feedback within the system. Here we outline these expression patterns and present examples of how they are affected in mouse mutants with defects in IFT components.
The effects of IFT mutations on Shh signaling vary widely depending on the component and the nature of the mutations. Here we show three examples to illustrate these effects:
Null mutations in components required for anterograde IFT typically disrupt ciliogenesis entirely (e.g., Fig. 2D, H, L) and lead to lethality around embryonic day 9.5 (E9.5) (Marszalek et al., 1999; Murcia et al., 2000). This allows for an analysis of early Shh-dependent neural tube patterning, but largely precludes the study of limb patterning as limb development has just begun at this stage. An example presented below (Kif3atm1Gsn) is a null mutation in Kif3a, encoding a subunit of kinesin II, but other mutants, such as Ift88 null homozygotes, show similar phenotypes (Marszalek et al., 1999; Murcia et al., 2000). In such mutants, the Gli activators are rendered mostly or entirely nonfunctional (Haycraft et al., 2005). However, Shh targets are not entirely suppressed in these embryos because formation of Gli3Rep is also disrupted, leading to partial derepression of target genes normally responding to weak Shh signaling (Huangfu and Anderson, 2005; Liu et al., 2005).
Hypomorphic mutations that partially disrupt antergograde IFT, such as Ift88fxo, reduce the frequency of ciliogenesis and/or result in shorter cilia (Fig. 2C, Liu et al., 2005; Taulman et al., 2001). In such mutants, Gli activator function is reduced but not abolished and Gli3 repressor formation is impaired (Liu et al., 2005). These mutants survive to later embryonic stages (beyond E10.5), allowing analysis of Shh signaling in the limbs, as well as the neural tube. Here we show as examples the limb and neural tube-patterning phenotypes of Ift88fxo/Ift88null compound heterozygotes.
Mutations in IFT components can also have qualitatively distinct effects on ciliogenesis and Hh signaling. For example, components of IFT complex A, such as the mouse Ift139 homolog THM1/aln, appear to have a more important role in retrograde IFT in comparison to anterograde IFT (Tran et al., 2008). Here we provide another example of a mutation in a complex A IFT protein, a null allele of Ift122 (Ift122sopb). Ift122sopb mutants harbor a missense mutation in the start codon preventing protein translation (J. Qin and J. T. E., in preparation). In Ift122sopb mutants, cilia are generated at, or near, normal frequencies, but the tips of their cilia are swollen and show accumulation of IFT components and other ciliary proteins (Fig. 1B, F, J and data not shown). In these mutants, Gli2Act is constitutively active independently of Shh, leading to an overall hyperactivation of the pathway (J. Qin and J.T.E., in preparation). Such mutants typically survive until E12.5, allowing for neural tube and limb-patterning analysis.
Patterning of cell fates along the dorsal–ventral axis of the developing mouse neural tube is governed to a large extent by graded activity of the Shh pathway. Shh signaling represses the expression of genes marking dorsal neural fates and induces the expression of markers for ventral fates. Because Shh acts as a morphogen, each marker is induced over a distinct range of Shh signaling. The Shh-dependent gene expression patterns in the neural tube presented below are summarized in Fig. 3.
Figure 4 shows an analysis of such markers in different mutant neural tubes as examined by immunofluorescence (IF) in transverse cryosections. An advantage to this method is that it is relatively simple and double or triple labeling can reveal the spatial relationships of multiple distinct cell types simultaneously in the same tissue. These markers can be assayed in the E10.5 neural tube but, in cases where certain mutants arrest earlier, they can also be studied at E9.5. It is often convenient to counterstain these sections with a nuclear stain such as DAPI to delineate the neural tube. Here we show the expression of markers for proliferating progenitor cells of the ventricular zone, but because these progenitors give rise to distinct types of postmitotic neurons, the expression of neuronal markers, such as Isl1 (motor neurons), Chx10 (V2 interneurons), En1 (V1 interneurons), and Evx1 (V0 interneurons), may also be examined at stages when neurogenesis is well underway (E10.5 or later).
The floor plate markers FoxA2 (previously called Hnf3β) and Shh are induced at the ventral pole of the neural tube by the highest levels of Shh signaling (Marti et al., 1995; Roelink et al., 1995). These markers show the highest sensitivity to diminished Shh signaling and neither marker is expressed in the neural tissues of Kif3a−/− or Ift88fxo/− mutants (Fig. 4C, E, H, J). Shh is also expressed in the notochord, lying ventral to the neural tube, and its initial expression here does not rely on Shh pathway activity (Epstein et al., 1999). In IFT mutants with hyperactive Shh signaling, such as Ift122sopb, the domains of Shh and FoxA2 expand dorsally beyond the ventral midline (Fig. 4B and G). These effects can vary with respect to the rostral-caudal level of the neural tube. For example, inappropriate ventralization of cell fates is limited to the lumbar regions of the spinal neural tube in Ift122sopb mutants.
The expression of homeobox and bHLH transcription factors such as Nkx2.2, Nkx6.1, Olig2, Dbx1, Irx3, and Pax6 is controlled by distinct levels of Shh signaling and thus provide sensitive readouts for this pathway in neural tissues (Dessaud et al., 2008). Nkx6.1 (Fig. 4P, red) is normally expressed in the ventral third of the neural tube in response to high and intermediate levels of Shh signaling in progenitors giving rise to floor plate, V3 and V2 interneurons, and motor neurons (Briscoe et al., 2000), whereas Nkx2.2 (Fig. 4K, red) is normally induced only by high levels in a progenitor domain adjacent to the floor plate that gives rise to V3 interneurons (Briscoe et al., 1999). The motor neuron progenitor marker Olig2 (Fig. 4K, green) is induced by intermediate levels of signaling and is therefore expressed in a narrow band of cells just dorsal to the Nkx2.2+ domain (Takebayashi et al., 2002). Pax6 (Fig. 4P, green) is upregulated by weak Shh signaling in progenitors giving rise to V0, V1, and V2 interneurons, but its expression is downregulated or fully repressed in cells experiencing higher levels of Shh signaling (Ericson et al., 1997). Msx1 and Msx2 (Msx1/2) are expressed in the most dorsal neural progenitors (e.g., Fig. 4U) in response to BMP signaling (Liem et al., 2000). Although the Msx1/2 expression domain in the neural tube is not affected by loss of Shh signaling, increased Shh signaling downregulates Msx expression (Cho et al., 2008; Liem et al., 2000).
In Ift88fxo/− mutants, the level of Shh pathway activation is partially reduced. This level is sufficient to induce Olig2 and Nkx6.1 expression. However, Nkx2.2 expression is lost and the ventral repression of Pax6 expression is incomplete. Kif3a−/− mutants, analyzed at E9.5, show a further reduction in the level of Shh signaling. This is manifested by the loss of Olig2 expression, restriction of Nkx6.1 expression to a just few cells in the ventral midline, and complete derepression of Pax6 throughout the ventral neural tube. As the ventral border of Msx1/2 expression is not normally set by Shh signaling, the Msx½+ domain remains unchanged in Ift88fxo/− and Kif3a−/− mutants.
These expression patterns are quite different in Ift122sopb mutants. Rather than being repressed or ventrally restricted, the Nkx2.2+ and Nkx6.1+ domains expand toward the dorsal pole, similar to the Shh+ and FoxA2+ domains. The Olig2+ domain dorso-lateral the shifts to regions, while ventral border of the Pax6+ domain is restricted to the dorsal third of the neural tube. The increase in Shh signaling in such mutants also disrupts specification of dorsal cell fates, as shown by the absence of Msx1/2 expression in the dorsal neural tube.
The expression patterns of some genes are not easily monitored by immunofluorescent methods due to lack of availability of suitable commercially available antibodies. In such cases, section in situ hybridization provides a viable alternative. As an example, in situ hybridization analysis of neural identity genes, such as Irx3 and Dbx1, and Shh signaling pathway components is shown in Figs. Figs.55 and and66.
Irx3 and Dbx1 (Fig. 5) are normally repressed in the ventral neural tube by intermediate to high levels of Shh signaling (Briscoe et al., 2000; Pierani et al., 1999). Hence, they are frequently expressed in ventrally expanded domains in Kif3a−/− mutants showing strongly attenuated Shh signaling, while they are expressed in dorsally shifted/restricted domains in Ift122sopb mutants. The expression patterns of Dbx1 and Irx3 are largely unchanged in Ift88fxo/− mutants because they retain a higher level of Shh pathway activity in comparison to Kif3a−/− mutants.
Transcriptional targets of the Shh pathway are not restricted to effector genes, such as the homeobox transcription factors that specify cell fate, but also include components of the signaling pathway itself. The expression of such targets is shown in Fig. 6. As described above, Shh from the notochord induces its own expression in the ventral midline of the neural tube (Fig. 6A and D). In addition, expression of the genes encoding the Shh receptor Ptch1 and the transcription factor Gli1 are induced by Shh signaling in the ventral neural tube, although they are not expressed in floor plate cells experiencing the highest levels of Shh signaling (Fig. 6F, I, K, N). In contrast, expression of Gli3 is restricted dorsally in the neural tube due to repression by Shh signaling (Fig. 6P). Because of the lack of suitable antibodies, monitoring Ptch1, Gli1, and Gli3 expression is best performed through in situ hybridization.
Consistent with the effects described above, expression of Gli1 and Ptch1 in the neural tubes of Ift88fxo/− mutants is weaker or is ventrally restricted when compared to wild type, and is reduced to basal levels in Kif3a−/− mutants. In contrast, the expression domains of Gli1 and Ptch1 are either shifted or expanded dorsally in Ift122sopb mutants due to hyperactivation of the Shh pathway. The level of Shh pathway activity in Ift88fxo/− and Kif3a−/− mutants is not reduced enough to appreciably affect Gli3 expression ventrally, but the increase in pathway activity is sufficient to downregulate Gli3 in the dorsal domains of Ift122sopb mutant neural tubes. Thus, the expression of these Shh targets can provide an independent measure of the effects of IFT mutations on the activity of the Hh pathway in the neural tube.
As in the neural tube, signaling by Shh controls the expression of a large number of target genes in the developing mouse limb buds and its role in limb patterning and growth has been extensively studied (McGlinn and Tabin, 2006). These target genes can be positively or negatively regulated by Shh signaling and the individual target genes have different response thresholds. Shh is expressed in a small group of mesenchymal cells in the posterior region of the limb, and it signals to cells throughout much of the limb bud along the anterior–posterior axis. In contrast to the neural tube, much of the patterning is established by the activity of the Gli3 repressor, whose formation is inhibited by Shh signaling (te Welscher et al., 2002; Vokes et al., 2008; Wang et al., 2000). Gene expression in limb buds is easily assayed by whole-mount in situ hybridization. Because patterning of the limbs by Shh is not easily seen at earlier stages (before E10.5), this type of analysis is suitable for those mutants surviving until E10.5 or later (e.g., Ift88fxo/− and Ift122sopb mutants), as opposed to those arresting around E9.5 (e.g., Kif3a−/−mutants).
Figure 7 shows examples of forelimb buds of Ift88fxo/− and Ift122sopb mutants at E10.5 analyzed by in situ hybridization. At this stage, the size and the shape of Ift88fxo/− limb buds are normal, although at later stages the mutant limb buds exhibit preaxial polydactyly (extra digits on the anterior side of the limbs). In contrast, Ift122sopb mutant forelimb buds are shorter and broader than in wild-type, although the overall size of Ift122sopb mutants is normal at this stage. The altered gene expression patterns in IFT mutant forelimbs described below are also observed in the hindlimb buds (not shown).
The Hoxd genes are expressed in the limb bud mesenchyme in nested patterns (Robert and Lallemand, 2006). The Gli3 repressor form, Gil3Rep, whose production is differentially inhibited along the anterior–posterior axis by Shh signaling, controls these expression patterns. Thus, loss of Shh leads to overproduction of Gli3Rep and strong downregulation of Hoxd genes such as Hoxd11 and Hoxd13, whereas loss of Gli3Rep results in their ectopic expression in anterior regions (te Welscher et al., 2002; Zakany et al., 2004). Gremlin1 (Grem1) is also induced by Shh signaling in the mesenchyme, but its expression is excluded from the most posterior regions where Shh is expressed (Nissim et al., 2006). Grem1 expression is regulated by both GliAct and Gli3Rep (Vokes et al., 2008). As a result of Grem1 upregulation by Shh, Fgf4 expression is upregulated in the apical ectodermal ridge (AER) overlying the mesenchyme through inhibition of Bmp signaling (Zuniga et al., 1999).
Anterograde IFT is required not only for activation of GliAct forms but also for appropriate generation of Gli3Rep. Thus, although the most prominent phenotypic features observed in Ift88fxo/− mutant neural tubes are explained by attenuated Shh pathway activity, many of the altered gene expression patterns observed in the limbs are explained by ectopic pathway activity. This can be seen by the anterior expansion of the Hoxd11, Hoxd13, Grem1, and Fgf4 expression domains in Ift88fxo/− mutants (Fig. 7B, E, H, K). Despite their abnormal shapes, Ift122sopb mutant limb buds exhibit similar anterior expansion and/or shifting of these expression domains (Fig. 7C, F, I, L).
The expression of Shh pathway components in the limb buds also serves as an informative readout of the pathway. Figure 8 shows expression of such targets in the limb buds. Shh expression is initiated in the posterior limb bud independently of Shh pathway activity but is maintained by positive feedback via ectodermal FGFs (Niswander et al., 1994). In Ift88fxo/− mutants, Shh expression is limited to the posterior domain, whereas in Ift122sopb mutant limb buds, it is slightly expanded. Gli1 and Ptch1 are induced by Shh signaling in the limbs as they are in the neural tube. These genes are induced broadly in the posterior regions of normal limb buds (Fig. 8D, and G). In contrast to Hoxd11 and Hoxd13, the expression domains of Ptch1 and Gli1 appear to be set primarily by GliAct forms, rather than by Gli3Rep (te Welscher et al., 2002; Vokes et al., 2008). As a result, both genes are strongly downregulated and posteriorly restricted in Ift88fxo/− mutant limb buds (Fig. 8E and H). In Ift122sopb mutants, GliAct (Gli2Act, in particular) appears to be constitutively active. Thus, both Gli1 and Ptch1 are ectopically expressed in anterior regions in these mutants (Fig. 8F and I). Gli3 expression is normally inhibited by strong Shh signaling and is therefore repressed in the most posterior regions of the limb buds. Gli3 expression in Ift88fxo/− mutants expands slightly toward the posterior side in comparison to wild type, while it is more anteriorly restricted in Ift122sopb limb buds (Fig. 8K and L). Thus, constitutive activity of the GliAct forms in Ift122sopb mutant limb buds results in a general expansion of posterior cell fates at the expense of anterior fates. By contrast, the combined effects of diminished formation of the Gli3Rep form and reduced activity of GliAct forms in Ift88fxo/− mutants result in ectopic expansion of some Shh targets (e.g., Hoxd genes and Grem1), as well as downregulation/posterior restriction of others (e.g., Ptch1 and Gli1).
To analyze mouse embryonic phenotypes, crosses of heterozygous mice are set up and vaginal plugs are checked each morning to establish the day of conception (noon of the day when a plug is found is considered to be E0.5). After a suitable period of gestation, pregnant female mice are sacrificed by regulated CO2 inhalation, and the abdominal cavity is opened so that the uterine horns (containing individual decidua) may be removed into a dish with chilled 1× PBS containing 0.4% bovine serum albumin (PBS/BSA). Decidua are dissected from the uterus with watchmaker forceps in PBS/BSA. Individual decidua are then dissected to remove the embryos and their surrounding yolk sacs. The latter are removed and saved individually for genotyping by PCR. The thin amniotic membrane surrounding the embryo is also removed, after which embryos are fixed in 4% paraformaldehyde.
For in situ hybridization (ISH), embryos are fixed overnight (o/n) at 4°C. For whole-mount ISH, embryos are washed 3× [5 min each at room temperature (r.t.)] in DEPC-treated PBSw and dehydrated through a methanol series (25, 50, and 75% MeOH in DEPC-PBSw, 5 min each), washed twice (5 min each) in 100% MeOH, and stored in MeOH at −20°C until subjected to ISH. For section ISH, embryos are washed 3× (5 min each) in DEPC-treated PBS (DEPC-PBS), immersed o/n in 30% sucrose (prepared in DEPC-PBS) at 4°C, and then embedded in Tissue-Tek O.C.T. mounting medium in plastic molds. Embryos are positioned upright in the O.C.T. for transverse sections through the spinal neural tube and the molds are transferred to dry ice to freeze. Blocks prepared this way may be stored at −80°C for at least 2 years.
For IF, embryos are fixed in 4% paraformaldehyde for 2 h at 4°C, washed twice in PBS (5 min each) and then immersed o/n in 30% sucrose solution in PBS at 4°C. The following day, embryos are embedded in O.C.T. blocks as described above, and stored at −80°C.
Ten- to twelve-micrometer-thick sections of embryos for section ISH or IF are prepared using a cryostat (e.g., Leica CM3000). Sections should be collected on positively charged microscopy slides (e.g., Superfrost Plus, Fisher Scientific, Pittsburgh, PA, USA), labeled, and stored in slide boxes. Slides prepared this way may be stored for at least 1 month before analysis.
Primary cilia may be stained in embryo sections using the protocol described above (see Materials for description of antibodies). Acetylated alpha tubulin is a commonly used marker for cilia. However, it is difficult to use it for staining cilia of the neural progenitors, as these cilia stain weakly with this marker, while there is intense staining throughout the cell body. For this reason, we have chosen to image primary cilia on cells of the lateral plate mesoderm (Fig. 2). Due to their small size (~1 μm long), primary cilia are best imaged by confocal microscopy using 60 or 100× (1.4 NA) objectives. Here, we have used a Leica TCS SP5 microscope and Volocity software (Improvision, Waltham MA, USA) and we have shown 5-μm-thick Z-stack projections.
Digoxigenin-labeled riboprobes are synthesized from plasmid templates (see Materials) using a digoxigenin-UTP labeling mix (Roche, Branford, CT, USA) following the protocol provided by the manufacturer. Briefly, plasmids are linearized at the 5′ ends of the inserts (for generating antisense probes) and 20 μl in vitro transcription reactions are set up using SP6, T7, or T3 RNA polymerases and DIG labeling mix with 1 μl RNasin for 2 h at 37°C. One microliter of RNase-free DNase is added to the reactions (to digest template DNA), and the reaction is incubated for an additional 15 min at 37°C. Two microliters of the reaction is combined with loading buffer, heated at 95°C for 5 min, and run with a DNA ladder on an agarose gel to check synthesis. Probes are precipitated by the addition of 1 μl glycogen (20 mg/ml), 2 μl 0.2 M EDTA, 2.5 μl 4M LiCl, and 75 μl ethanol, followed by incubation at −20°C for 2 h and centrifugation for 15 min. The pellets are washed with 75% ethanol, air dried, and the probes are resuspended in DEPC-treated dH20. The probe concentration is determined with a spectrophotometer.
We follow a modified version of previously developed protocols for section ISH (Schaeren-Wiemers and Gerfin-Moser, 1993) as follows: Unless otherwise stated, all procedures are performed at r.t.
We follow a modified version of previously developed protocols for whole-mount ISH developed by Domingos Henrique and David Ish-Horowitz as follows: Unless otherwise stated, all procedures are performed at r.t.
Dissection buffer: 1× PBS 0.4% BSA. Chill to 4°C.
Fixative to: 4% Paraformaldhyde (Sigma Aldrich, St. Louis, MO, USA, Cat. #6148) made in 1× PBS (heat to 60°C to dissolve). Store in 10-ml aliquots at −20°C. Thaw to room temperature just before use.
Embedding: Tissue-Tek O.C.T. mounting medium (Sakura Finetek, Torrence, CA, USA, Cat. #4583), plastic embedding molds (Peel-A-Way, Polysciences, Warrington, PA, USA, Cat. #18986).
Wash buffer: 1× PBS + 0.1% Triton X-100 and 1% heat-inactivated goat serum (HINGS) or horse serum (HIHS). Heat inactivate serum at 55°C for 25 min.
DAPI: Make 500× stock of 4′,6-diamidino-2-phenylindole (Sigma Aldrich, Cat. #D9542, 100 μg/ml in dH20). Store as aliquots at −20°C.
FMM: 90 ml glycerol, 10 ml 10× PBS, 5 g n-propyl gallate (Sigma Aldrich, Cat. #P3130). Store at 4°C.
Primary antibodies: Mouse monoclonal antibodies against Shh (clone 5E1), Nkx2.2 (74.5A5), FoxA2 (4C7), Nkx6.1 (F55A10), and Msx1/2 (4G1) are available from the Developmental Studies Hybridoma Bank (DSHB), University of Iowa. Tissue culture supernatants are typically used at a 1:10 dilution. Rabbit anti-Pax6 and goat anti-Olig2 polyclonal antibodies are from Covance (Berkeley, CA, USA, Cat. #PRB-278P, use at 1:1000) and R&D Systems (Minneapolis, MN, USA, Cat. #AF2418, use at 1:200), respectively. Other useful antibodies for neural patterning analysis include α-Pax7 (DSHB, Pax7), α-Dbx2 (Abcam, Cambridge, MA, USA, ab25554), α-MNR2/HB9 (DSHB, 81.5C10), α-Isl1/2 (DSHB, 39.4D5), α-Lhx3 (DSHB, 67.4E12), α-En1 (DSHB, 4G11), and α-Evx1 (DSHB, 99.1-3A2). Primary cilia are stained with mouse antiacetylated alpha tubulin monoclonal antibodies from Sigma (Cat. #T6793), rabbit α-IFT88 polyclonal antiserum (generous gift from Dr Gregory Pazour, used at 1:1000), and rabbit α-Arl13b polyclonal antiserum (generous gift from Dr Tamara Caspary, used at 1:3000).
Secondary antibodies: Donkey Cy3-labeled antimouse IgG (Cat. #715-165-151), Donkey Cy2-labeled antirabbit IgG (Cat. #711-225-152), donkey Cy2-labeled antigoat IgG (Cat. #705-225-147) from Jackson ImmunoResearch (West Grove, PA, USA).
Reagents for riboprobes synthesis: T3, T7, or SP6 polymerases (Roche), 10× DIG labeling mix (Roche, Cat. #11277073910) and RNasin (Promega, Madison, WI, USA, Cat. #N251A), Rnase-free DNase (Roche, Cat. #776785).
Probe templates: Plasmids used for synthesis are Shh (Echelard et al., 1993), Gli1 (Hui et al., 1994), Ptch1 (Goodrich et al., 1996), Gli3 (Bulgakov et al., 2004), Dbx1 (Takahashi and Osumi, 2002), Irx3 (Bosse et al., 1997), Hoxd11 (Dolle et al., 1989), Hoxd13 (Dolle et al., 1989), Fgf4 (Hebert et al., 1990), and Grem1 (Pearce et al., 1999).
Baskets for whole mount ISH: 15 mm Netwell Insert, Corning, Lowell, MA, USA, Cat. #3478
DEPC-treated solutions: add 1 ml diethyl pyrocarbonate (Sigma Aldrich, Cat. #D5758) per liter solution, stir o/n at 37°C, autoclave to inactivate DEPC.
PBSw: DEPC-treated 1× PBS + 0.1% Tween-20.
Hybridization buffer A (sections): 50% formamide, 5× SSC, 5× Denhardt’s solution (Invitrogen, Carlsbad, CA, USA), 0.25 mg/ml Bakers yeast RNA (Sigma Aldrich), 0.5 mg/ml boiled/sonicated Salmon Sperm DNA (Sigma Aldrich). All stock solutions made in DEPC-treated dH20.
Hybridization buffer B (whole mount): Sock solutions should be made in DEPC-treated dH20. For 50 ml, combine 25 ml formamide, 12.5 ml 20× SSC (pH 7.0), 0.5 g Boehringer Blocking Reagent (Roche, Cat. #109676). Heat at 65°C for 1 h with stirring to dissolve. Add 6 ml DEPC-dH20, 5 ml 10 mg/ml Bakers yeast RNA (Sigma Aldrich), 100 μl 50 mg/ml heparin (Sigma Aldrich), 250 μl 20% Tween-20, 500 μl 10% CHAPS, 500 μl 0.5 M EDTA. Filter solution and store at −20°C.
Acetylation buffer: 246 ml DEPC-treated dH20, 3.3 ml triethanolamine (Sigma Aldrich), 0.44 ml concentrated HCL. Add 0.625 ml acetic anhydride (Sigma Aldrich) and mix immediately before use.
MAB: 100 mM maleic acid, 150 mM NaCl, pH 7.5.
Blocking Solution: Dissolve Boehringer Blocking Reagent to 1% in PBSw with 10% HINGS by heating to 65°C and vortexing repeatedly. Store at −20°C.
B1 solution: 0.1 M Tris-HCl pH 7.5, 0.15 M NaCl.
B2a solution: B1 solution + 10% HINGS.
B2b solution: B1 solution: + 1% HINGS, and alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche, Cat. #11093274910) diluted 1:1000.
B3 solution: 0.1 M Tris-HCl pH 9.5, 0.1 M NaCl, 50 mM MgCl2.
B4 solution: B3 solution + 5 μl/ml Nitro blue tetrazolium chloride (NBT, Roche, Cat. #11383213001), 3.75 μl/ml 5-bromo-4-chloro-3-indolyl phosphate (BCIP, Roche, Cat. #11383221001).
Developmental geneticists studying mammalian Hh signaling and cell biologists studying IFT and ciliogenesis have been unexpectedly brought together by recent findings. As a result, both groups are expanding their research efforts to explore the fascinating connection between the two systems. Here, we have attempted to provide a framework for analyzing Shh signaling in mouse tissues for those cell biologists less familiar with the topic. Due to the complexities of the Shh signaling pathway, the patterning phenotypes of mutants defective for IFT components are complex and distinct from one another, as illustrated by the examples shown here. However, this very complexity provides clues for understanding how Shh signal transduction components interact with one another in the context of the highly organized structure of the primary cilium.
We thank Jian Qin, who identified the Ift122sopb mutation and conducted much of the initial characterization, and Yulian Lin for technical assistance. We thank K. Anderson, N. Murcia, T. Caspary, G. Pazour, L. Niswander, X. Sun, G. Fishell, and A. McMahon, for mouse strains, antibodies, and plasmids for ISH. Monoclonal antibodies were obtained from the DSHB developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biology, Iowa City. Work on cilia and Hh signaling is funded by Pennsylvania State University start-up funds (to A.L.), the New Jersey Commission on Spinal Cord Research (to H.W.K. and J.T.E.), and the National Institutes of Health, grant #5R01HD050761 (to J.T.E.).