Research ArticleNeuroscience

Akt and PP2A Reciprocally Regulate the Guanine Nucleotide Exchange Factor Dock6 to Control Axon Growth of Sensory Neurons

See allHide authors and affiliations

Science Signaling  05 Mar 2013:
Vol. 6, Issue 265, pp. ra15
DOI: 10.1126/scisignal.2003661

Abstract

During neuronal development, axons navigate long distances, eventually forming precise connections with such targets as peripheral tissues. Dock6 is a guanine nucleotide exchange factor (GEF) that activates the Rho family guanosine triphosphatases Rac1 and Cdc42 to regulate the actin cytoskeleton. We found that phosphorylation of Ser1194 in Dock6 inhibited its GEF activity and suppressed axonal growth of embryonic sensory neurons and axon regeneration of postnatal sensory neurons in vitro and in vivo. At early developmental stages, when axons are growing, the protein phosphatase PP2A interacted with and dephosphorylated Dock6, thereby increasing the activity of Dock6. At later developmental stages, the abundance of the kinase Akt increased, resulting in the binding of Akt to Dock6 and the phosphorylation of Dock6 at Ser1194. In dorsal root ganglion neurons from mice lacking Dock6, reintroduction of Dock6 with a nonphosphorylatable S1194A mutation rescued axon extension but not branch number, whereas reintroduction of Dock6 with a phosphomimetic S1194E mutation resulted in premature branching. Thus, the phosphorylation status of Dock6 at Ser1194 determines whether it promotes axon extension or branching in sensory neurons, revealing interplay between kinase and phosphatase action on a Rho-GEF during axon growth.

Introduction

Axon morphogenesis, which includes neurite outgrowth, axon guidance, branching, and synapse formation, is an orchestrated developmental process eventually leading to the establishment of the neuronal circuits of the two nervous systems in mammals. Peripheral nervous system (PNS) development in mammals is unique in two ways. First, an interaction between two different cell types, namely, neurons and Schwann cells, is responsible for part of PNS development (1, 2). Second, sensory neurons such as dorsal root ganglion neurons exhibit comparatively long axons, such as the ones that innervate limbs. For this reason, they are used as a model to study axon morphological changes. The mechanisms underlying sensory axon navigation are mediated by signaling molecules that control continuous and complex cytoskeletal changes, such as the Rho family of guanosine triphosphatases (GTPases), which includes Rac and Cdc42 (36).

Guanine nucleotide exchange factors (GEFs) catalyze the replacement of guanosine diphosphate (GDP) with guanosine 5′-triphosphate (GTP) to activate GTPases and integrate the upstream signals to determine the timing and the specificity of GTPase activation (713). We previously characterized Dock6, the 200-kD kiaa1395 gene product (14), as an atypical Dock180-related GEF for Rac and Cdc42 (15), consistent with the fact that Dock6 is closely related to Dock7 and Dock8 (13, 16). Dock6 is a GEF that has a unique catalytic domain because it differs from the catalytic Dbl homology domains of Dbl-family GEFs for Rho family GTPases (713). Here, we describe the phosphorylation-dependent regulation of Dock6 in dorsal root ganglion neurons. We found that Akt-mediated phosphorylation of Dock6 inhibited axon extension. During the early stages of axon extension, protein phosphatase 2A (PP2A) bound to Dock6, which attenuated phosphorylation of Dock6. In this state, Dock6 was active as a GEF and mediated axon extension. During later stages of axon extension, the abundance of Akt was increased. Akt bound to and phosphorylated Dock6 at Ser1194 to inactivate Dock6. This phosphorylation event led to decreased axon growth, providing evidence for the role of Dock6 in the development of PNS neurons and of the GEF as the substrate of Akt as well as PP2A.

Results

Dock6 is required for axon extension

Primary dorsal root ganglion neurons are a good in vitro model for the study of axon growth because they extend axons in response to various growth factors, including nerve growth factor (NGF). We detected Dock6 in dorsal root ganglion neurons by immunofluorescence with an affinity-purified antibody specific to Dock6 (fig. S1, A and B). Endogenous Dock6 was widely distributed throughout cell bodies, axons, and branches in dorsal root ganglion neurons and partially colocalized with the GTPase effector Rac1 (fig. S1A). To investigate the role of Dock6 in dorsal root ganglion neurons, we constructed plasmids encoding nonoverlapping short hairpin RNAs (shRNAs) for Dock6, Rho GTPases, or luciferase as a control (fig. S1, C to E) and a fluorescent protein (ZsGreen) as a marker (representative morphologies of neurons transfected with each shRNA plasmid are depicted in Fig. 1A). Ordinarily, dorsal root ganglion neurons extend axons by ~300 μm over 48 hours in the presence of NGF, but knockdown of Dock6 or either of the small GTPases resulted in a decrease in total axon length or in the length of the longest axon by more than 50% (Fig. 1B). Knockdown of Dock6 or Rac1 had a greater effect on axon length than that of Cdc42 (Fig. 1B, upper panels). Knockdown of Dock6 or Rac1 also decreased axon number and the number of axonal branch points, whereas Cdc42 knockdown decreased branching but did not have a significant effect on axon number (Fig. 1B, lower panels). Moreover, knockdown of Dock6 resulted in inactivation of Rac1 but did not have a detectable effect on active GTP-bound Cdc42, both of which were measured by a pull-down assay using the Cdc42 and Rac interactive binding (CRIB) domain of Pak1 (Fig. 1, C and D). We concluded that, in dorsal root ganglion neurons, Dock6 preferentially activates a Rac1 signal; accordingly, we focused on Dock6 and Rac1 in the subsequent experiments.

Fig. 1

Dock6 is required for axon growth in dorsal root ganglion neurons. (A) Dorsal root ganglion neurons were transfected with control luciferase (Luci), Dock6#1, Dock6#2, Rac1, or Cdc42 shRNA and incubated with NGF. Neurons were fixed and immunostained with an anti-neurofilament antibody. Scale bar, 100 μm. (B) The total axon length and the length of the longest axon were measured. The number of axons and the number of branch points were counted. Error bars show ±SD from four independent experiments. Data were evaluated by one-way analysis of variance (ANOVA). *P < 0.01, n = 30 to 48 neurons. (C and D) Dorsal root ganglion neurons were transfected with control luciferase or Dock6#1 shRNA and lysed. GTP-bound Rac1 or Cdc42 was affinity-precipitated with glutathione S-transferase (GST)–CRIB, which specifically binds to GTP-bound Rac1 and Cdc42, from lysates and immunoblotted with an anti-Rac1 or anti-Cdc42 antibody. Total Rac1, Cdc42, or Dock6 is also shown. Amounts of GTP-bound Rac1 and Cdc42 were normalized to the amount of Rac1 and Cdc42. Data were evaluated using Student’s t test. **P < 0.05, n = 4 experiments.

Knockdown of Dock6 in vivo affects axon extension both in developmental stages and after injury

To address the in vivo relevance of the in vitro results regarding Dock6, we generated genetically modified mice lacking Dock6. We injected the DNA construct encoding Dock6#1 shRNA to produce small interfering RNA for Dock6 into fertilized mouse eggs according to standard methods (17, 18) and generated two Dock6 shRNA transgenic mouse lines (TG#1 and TG#2) (fig. S2, A and B). Copy numbers of the transgene for TG#1 and TG#2 mice were calculated to be 100.2 ± 16.16 and 19.24 ± 6.314. Dorsal root ganglion neurons from both TG#1 and TG#2 mice displayed an effective knockdown of Dock6 protein as did central nervous system (CNS) tissues such as the cerebrum and spinal cord (fig. S2, C and E). The abundance of other proteins, such as the Cα subunit of PP2A, Akt1, Rac1, and β-actin, was comparable between transgenic and nontransgenic tissues. Dorsal root ganglion neurons from TG#1 and TG#2 mice displayed comparable decreases in total axon length and number of axons (fig. S2, D and F). Because knockdown efficiencies and the phenotypes in cells are not likely to depend on transgene copy numbers, we used TG#1 mice in the following experiments to analyze the effects of Dock6 knockdown on developmental stages and injury in mice.

At embryonic day 11, Dock6 shRNA transgenic mice exhibited a phenotype characterized by shortened peripheral neuronal fibers, which extend from the ganglia to the ventral roots, compared with those in control nontransgenic littermates (Fig. 2, A and B), suggesting the involvement of Dock6 in the formation of peripheral neuronal fibers. In earlier or later embryonic stages, the phenotypes of axons were comparable between transgenic mice and their control littermates (fig. S3, A and B). We speculate that Dock6 is involved in regulating axon extension in the embryonic stage, when the axons are growing rapidly. To analyze the effects of Dock6 knockdown on sciatic nerve generation, we measured axonal extension after a nerve crush (19) and observed delayed extension of neuronal fibers in transgenic mice 3 and 7 days after injury (Fig. 2, C to F). The axon numbers in sciatic nerves of transgenic mice were comparable to those in control littermates in newly born pups (fig. S3C). Together, these results suggest that Dock6 regulates axon extension both in normal developmental and after injury.

Fig. 2

Effect of Dock6 knockdown on peripheral neuronal fiber and on axon neuronal fiber extension after injury. (A) Cross sections along the spinal cord (SC) and ventral root (VR) regions of a Dock6 shRNA transgenic TG#1 mouse (TG) or nontransgenic littermate (NTG) on embryonic day 11 were immunostained with an anti–βIII-tubulin antibody. Arrows indicate the positions of neuronal fibers linking ganglia to ventral roots. Scale bar, 100 μm. (B) The number of ganglia with or without neuronal fibers to ventral roots was counted. Data were evaluated using Student’s t test. *P < 0.01, n = 155 slices from two independent littermates for NTG, n = 118 slices from two independent littermates for TG. (C and E) Representative images of Dock6 shRNA transgenic TG#1 mouse (TG) or nontransgenic littermate (NTG) sciatic nerve longitudinal sections on day 3 (C) or day 7 (E) after nerve injury are shown. Sections were immunostained with an anti-CGRP antibody. Scale bar, 100 μm. The position is depicted in (C). (D and F) The length of the gap remaining between the outgrowing axons and the distal side of the lesion was measured. Data were evaluated using Student’s t test. *P < 0.01, **P < 0.05, n = 5 mice.

Dock6 is phosphorylated by Akt

The closely related molecule Dock7 is activated by tyrosine phosphorylation by the cell surface receptor ErbB2 (16); however, the ErbB2 phosphorylation site is not conserved in Dock6. Instead, Dock6 contains several potential Akt phosphorylation (RXRXXS/T) sites, leading us to determine whether Akt binds to Dock6. Akt1 coimmunoprecipitated with full-length Dock6 (Fig. 3A). To identify the domain of Dock6 that binds to Akt, we cotransfected the plasmids encoding wild-type Akt1 and each Dock6 domain into 293T cells (fig. S4A). The DHR-1 (Dock homology region 1) region of Dock6 specifically coimmunoprecipitated with Akt1 (fig. S4B), and a form of Dock6 lacking the DHR-1 region (ΔDHR-1) did not coimmunoprecipitate with Akt1 and was not phosphorylated by Akt1 (fig. S4C). In vitro binding assays using recombinant Akt1 and recombinant DHR-1 region purified from 293T cell lysates (20) revealed that the dissociation constant (Kd) was ~61.7 nM (fig. S4D), indicating that the association of Akt1 with the DHR-1 region of Dock6 was direct and of high affinity. To determine the region of Akt that bound to the DHR-1 region of Dock6, either full-length Akt1 or a mutant lacking the PH domain (ΔPH) was incubated with the DHR-1 region. Both the full-length and the ΔPH form of Akt1 bound to the DHR-1 region of Dock6 (fig. S4E). These results indicated that Akt bound to the DHR-1 region of Dock6 through its kinase domain.

Fig. 3

Dock6 is phosphorylated at Ser1194 by Akt and dephosphorylated by PP2A. (A) FLAG immunoprecipitates from 293T cells expressing full-length Akt1 and full-length Dock6 were immunoblotted with an anti-myc antibody. Cell lysates were also immunoblotted with an anti-myc or anti-FLAG antibody. Data are representative of two experiments. (B) FLAG immunoprecipitates from 293T cells expressing Akt1 and each region of Dock6 were immunoblotted with an anti–phospho-Akt substrate antibody. Cell lysates were also immunoblotted with an anti-FLAG or anti-myc antibody. Data are representative of three experiments. (C) The amino acid sequence of the middle2 region showing the two serine residues. (D) FLAG immunoprecipitates from cells expressing the wild-type (WT), S1194A, or S1334A forms of the middle2 region were incubated with recombinant Akt1 and adenosine 5′-triphosphate (ATP) and immunoblotted with an anti–phospho-Akt substrate antibody. Serine phosphorylation of the constructs and their abundance are also shown. Data are representative of three experiments. (E) Immobilized full-length FLAG-Dock6, FLAG–Dock6 S1194A, or FLAG–Dock6 S1194E was incubated with recombinant Akt1 and ATP and immunoblotted with an anti–(pS1194) Dock6 antibody. Cell lysates were also immunoblotted with an anti-FLAG antibody. Data are representative of three experiments. (F) Comparison of the amino acid sequences surrounding the Akt phosphorylation sites (red) of mammalian Dock6 with other homologous proteins. Black, conserved amino acids; gray, nonconserved amino acids. (G) Immobilized, phosphorylated Dock6 protein was incubated with recombinant PP2A and immunoblotted with an anti–(pS1194) Dock6 antibody. Data are representative of three experiments. (H) FLAG immunoprecipitates from 293T cells expressing FLAG-Dock6 or FLAG–Dock6–ΔDHR-2 with myc-Akt1 and V5-PP2A Cα were immunoblotted with an anti–(pS1194) Dock6 or anti-V5 antibody. Cell lysates were immunoblotted as indicated. Data are representative of three experiments. (I) FLAG immunoprecipitates from 293T cells expressing V5-PP2A Cα, FLAG-Dock6, and varying amounts of myc-Akt1 were immunoblotted with an anti-V5 (lane 1), anti-myc (lane 3), or anti–(pS1194) Dock6 (lane 5) antibody. Cell lysates were also immunoblotted with an anti-V5 (lane 2), anti-myc (lane 4), or anti-FLAG (lane 6) antibody. (J) Precipitated PP2Ac was normalized to the amount of total PP2Ac. Data were evaluated by one-way ANOVA. *P < 0.01, n = 3 experiments.

We thus investigated whether Akt phosphorylated Dock6 in vitro. Immobilized full-length Dock6 was incubated with the recombinant active Akt1 protein and immunoblotted with an antibody that specifically recognizes the consensus phosphorylated Akt target sequence. The addition of Akt1 protein resulted in phosphorylation of full-length Dock6 (15) in vitro (fig. S4F). To further identify the site on Dock6 that was phosphorylated by Akt1, we performed in vitro kinase reactions with constructs encoding the Dock6 domains purified from transfected 293T cells (fig. S4G). Although Akt1 weakly phosphorylated the N-terminal region, it phosphorylated the middle2 region robustly in vitro. In cells cotransfected with plasmids encoding Akt1 and each Dock6 domain into 293T cells, immunoblotting with the antibody against phosphorylated Akt substrate showed that Akt1 coexpression increased phosphorylation of the middle2 region (Fig. 3B). This region contains two serine residues in Akt phosphorylation consensus sequences (Fig. 3C). We made two constructs containing alanine substitutions in the middle2 region, Ser1194→Ala (S1194A) and Ser1334→Ala (S1334A). The mutation in Ser1194 abolished phosphorylation by the recombinant Akt1 protein (Fig. 3D). In full-length Dock6, it also abolished phosphorylation by Akt1, as detected by the antibody against phosphorylated Akt substrate (fig. S4H), as well as an affinity-purified antibody specific for (pS1194) Dock6 (Fig. 3E). The phosphomimetic Ser1194→Glu mutant of Dock6 (S1194E) was weakly recognized by these antibodies, regardless of the presence or absence of the Akt1 protein (Fig. 3E and fig. S4H). Collectively, these results indicate that Akt primarily phosphorylates Dock6 at Ser1194, which is uniquely conserved in mammalian Dock6 proteins (Fig. 3F).

Dock6 is dephosphorylated by PP2A

PP2A is a ubiquitously distributed, evolutionarily conserved heterotrimeric serine and threonine phosphatase that regulates diverse cellular functions and shows broad substrate specificity. PP2A consists of a structural A subunit, a regulatory B subunit, and a catalytic C subunit. The core unit is composed of the A and C subunits (21). Available proteome bioinformatics data (at the Human Protein Reference Database, http://www.hprd.org/) show that Dock7, a molecule closely related to Dock6, forms a complex with the Aα and Cβ subunits of PP2A.

First, we tested whether PP2A dephosphorylated Dock6 that had been phosphorylated by Akt by incubating the recombinant PP2A heterodimer, which is composed of the A and C subunits, with Dock6 phosphorylated by recombinant Akt1 in vitro. Immunoblotting with an antibody that recognized Dock6 phosphorylated at Ser1194 [anti–(pS1194) Dock6] showed that phosphorylated Dock6 was dephosphorylated by PP2A (Fig. 3G), indicating that PP2A directly dephosphorylates phosphorylated Ser1194 in Dock6. Next, to clarify whether Dock6 binds to PP2A, we cotransfected the plasmids encoding the PP2A Cα subunit and each of the Dock6 domains into 293T cells. The DHR-2 region of Dock6 specifically coimmunoprecipitated with the PP2A Cα subunit (fig. S5A). The ΔDHR-2 mutant of Dock6, which lacks the DHR-2 domain, was not detectably dephosphorylated by the PP2A Cα subunit and did not coimmunoprecipitate with it (Fig. 3H), suggesting that the DHR-2 domain is an interaction site with PP2A. On the basis of the in vitro binding assay between the recombinant PP2A and the recombinant DHR-2 region, the Kd value was calculated as ~256 nM (fig. S5B).

We next determined whether the PP2A Cα subunit and Akt1 competed for binding to Dock6 in transfected 293T cells. As the amount of transfected Akt1 was increased, the amount of PP2A Cα subunit in Dock6 immunoprecipitates decreased (Fig. 3, I and J). Although the crystal structure of Dock6 has not been determined, these results agree with the fact that the Kd value of the PP2A C subunit (fig. S5B) is one order of magnitude smaller than that of Akt1 (fig. S4D). Furthermore, to verify the effect of PP2A on the interaction between Dock6 and Rac1, we changed the expression of the PP2A Cα subunit in 293T cells. As the amount of transfected PP2A Cα subunit was increased, both Dock6 in Rac1 immunoprecipitates and, reciprocally, Rac1 in Dock6 immunoprecipitates also increased (fig. S5C), suggesting that PP2A promotes the binding of Dock6 to Rac1.

Dock6 colocalizes with active Akt and PP2A, and the association of Akt and PP2A with Dock6 changes during development

We performed immunohistochemistry on embryonic day 13.5 dorsal root ganglion neurons for Dock6 phosphorylated at Ser1194, PP2A, and active Akt (Akt phosphorylated at Ser473) (Fig. 4, A and B). The PP2A C subunit exhibited a broad but weak distribution in dorsal root ganglia and ventral roots, whereas phosphorylated Akt was found in more frontal regions of axons in ventral roots. The amount of PP2A C subunit that was present in Dock6 immunoprecipitates from dorsal root ganglion neurons gradually decreased as a function of time in culture without a substantial change in its abundance (Fig. 4C). In contrast, phosphorylated active Akt as well as total Akt1 in Dock6 immunoprecipitates increased (Fig. 4C). These immunoblotting data from dorsal root ganglion cultures are consistent with the immunostaining data. Furthermore, treatment with the PP2 inhibitor fostriecin decreased axon extension and the number of axons but increased the number of axon branches (fig. S5, D and E). Additionally, we found that fostriecin increased the basal phosphorylation of Dock6 at Ser1194 (fig. S5F). We therefore suggest that PP2 is required for the dephosphorylation of Ser1194 of Dock6 in dorsal root ganglion neurons. Together, these results are consistent with our hypothesis that Dock6 is a substrate of Akt, which antagonizes PP2A in the axonal growth of dorsal root ganglion neurons.

Fig. 4

Phosphorylated Dock6 colocalizes with active Akt and PP2A C subunit, and the association of active Akt or PP2A C subunit and Dock6 changes during development. (A and B) Embryonic day 13.5 mouse cross sections along the spinal cord (SC), dorsal root ganglion (DRG), and ventral root (VR) regions were coimmunostained with anti–PP2A C subunit (red) and (pS1194) Dock6 (green) antibodies (A) or anti–(pS473) Akt (red) and (pS1194) Dock6 (green) antibodies (B). Their positions are depicted in (A). Scale bar, 200 μm. Data are representative of 12 slices of two independent experiments. (C) Dock6 immunoprecipitates from primary rat dorsal root ganglion neurons cultured for 2, 8, or 15 days were immunoblotted with an anti–PP2A C subunit (lane 1) or anti–(pS473) Akt antibody (lane 3). Total amounts of PP2A C subunit (lane 2) and Akt1 (lane 4) are also shown. Data are representative of three experiments.

Dock6 is phosphorylated at Ser1194 after axon growth

We next examined whether phosphorylation of Ser1194 in Dock6 occurred in dorsal root ganglion neurons. Immunoblotting analysis showed that this phosphorylation event, as recognized with the anti–(pS1194) Dock6 antibody, increased ~3.0- and 9.0-fold at 8 and 15 days in culture, respectively (Fig. 5, A and B). Reciprocally, we performed an affinity precipitation assay with guanine nucleotide–free Rac1G15A, which preferentially interacts with catalytically active GEFs (15, 22) and can detect GEF activity in cells. Active Dock6 decreased by ~50 and ~20% at 8 and 15 days in culture, respectively (Fig. 5, A and C). In contrast, the abundance of Dock6 was comparable (Fig. 5A). These observations are consistent with the results that as dorsal root ganglion neurons grow, the amount of active Akt1 in Dock6 immunoprecipitates increased, whereas that of the PP2A C subunit in Dock6 immunoprecipitates gradually decreased (Fig. 4C).

Fig. 5

The phosphorylation status of Dock6 at Ser1194 changes during development. (A) Dock6 immunoprecipitates from dorsal root ganglion neurons cultured for 2, 8, or 15 days were immunoblotted with an anti–(pS1194) Dock6 antibody (lane 1). Cell lysates were affinity-precipitated with nucleotide-free GST-Rac1G15A, which binds to active Dock6, and immunoblotted with an anti-Dock6 antibody (lane 2). Total Dock6 or β-actin is also shown. (B and C) Phosphorylation of Dock6 (B; n = 5 experiments) or precipitated active Dock6 (C; n = 8 experiments) was normalized to the amount of total Dock6. Data were evaluated by one-way ANOVA. *P < 0.01. (D and E) Embryonic day 11.5 (D) or 13.5 (E) mouse cross sections along the spinal cord, dorsal root ganglion, and ventral root regions were coimmunostained with an anti-Dock6 (green) antibody or with (pS1194) Dock6 (green) and anti–βIII-tubulin antibodies (red). Scale bar, 200 μm. Data are representative of nine slices of two independent experiments.

In addition, we analyzed the distribution and phosphorylation of the Dock6 protein in transverse sections of mouse embryos during development. On embryonic days 11.5 and 13.5, Dock6 protein was broadly distributed in dorsal root ganglia and ventral roots, as well as in spinal cords, whose positions were determined by an antibody against βIII-tubulin, a neuron marker (Fig. 5, D and E, upper panels). In contrast, staining with an anti–(pS1194) Dock6 antibody was concentrated in ventral roots rather than in dorsal root ganglia during development (Fig. 5, D and E, lower panels). These results suggest that (pS1194) Dock6, which exhibits low GEF activity, is increased in the late stages of development and that the phosphorylation may occur in more frontal regions of axons, where axons have almost completed their extension.

Phosphorylation of Dock6 occurs downstream of TrkA and phosphoinositide 3-kinase

The major signal downstream of the TrkA receptor is mediated by phosphoinositide 3-kinase (PI3K) to Akt signaling in many types of cells (23). Phosphorylation of Dock6 at Ser1194 was decreased in dorsal root ganglia isolated from trkA+/− mice, which show reduced abundance of TrkA (Fig. 6, A and B) (24, 25). Additionally, phosphorylation of Dock6 was increased after NGF stimulation by ~1.8-fold (Fig. 6, C and D). Treatment of dorsal root ganglion neurons with the PI3K inhibitor LY294002 resulted in a decrease in NGF-mediated phosphorylation of Dock6 at Ser1194 (Fig. 6, C and D), suggesting that Dock6 acts downstream of TrkA and PI3K.

Fig. 6

Phosphorylation of Dock6 occurs downstream of TrkA and PI3K, and the phosphorylation status of Dock6 determines its ability to promote axon growth. (A) GST-CRIB precipitates from dorsal root ganglion neurons from TrkA mutant mice (+/−) or control littermates (−/−) were immunoblotted with an anti-Rac1 antibody. Total Rac1, (pS1194) Dock6, Dock6, TrkA, and β-actin proteins in dorsal root ganglion neurons are also shown. (B) The phosphorylation of Dock6 was normalized to the amount of total Dock6. Data were evaluated using Student’s t test. **P < 0.05, n = 3 experiments. (C) Dock6 immunoprecipitates from dorsal root ganglion neurons incubated with NGF in the absence or presence of LY294002 were immunoblotted with an anti–(pS1194) Dock6 antibody. Total Dock6 is also shown. (D) Phosphorylation of Dock6 was normalized to the amount of total Dock6 measured. Data were evaluated by one-way ANOVA. *P < 0.01, n = 5 experiments. (E) Rat dorsal root ganglion neurons expressing control vector, FLAG-shRNA–resistant WT Dock6, FLAG-shRNA–resistant Dock6 S1194A, or FLAG–shRNA-resistant Dock6 S1194E as well as ZsGreen-luciferase or Dock6#1 shRNA were cultured with NGF and immunostained with an anti-Dock6 antibody (red). Scale bar, 100 μm. (F) The total axon length and the length of the longest axon were measured. The number of axons and the number of branch points were counted. Error bars show ±SD from four independent experiments. Data were evaluated by one-way ANOVA. *P < 0.01, n = 20 to 27 neurons.

The phosphorylation status of Dock6 determines its ability to promote axon growth

We next asked whether axon morphogenesis was associated with the phosphorylation of Dock6. We constructed plasmids encoding forms of Dock6 that were resistant to Dock6#1 shRNA and that had either a wild-type Akt phosphorylation site, the phosphorylation-deficient mutation (S1194A), or the phosphorylation mimetic mutation (S1194E). We cotransfected the shRNA-resistant Dock6 constructs together with a control luciferase shRNA or Dock6#1 shRNA into dorsal root ganglion neurons. The expression of Dock6#1 shRNA–resistant constructs did not change with cotransfection with Dock6#1 shRNA, as revealed by immunoblotting (fig. S6A) as well as immunofluorescence (Fig. 6E, lower third, fourth, and fifth panels, compared with lower second panel). Expression of the shRNA-resistant form of Dock6 with the wild-type Akt phosphorylation site reversed the Dock6#1 shRNA–mediated decreases in axon extension, the number of axons, and the number of branches (Fig. 6, E and F). Expression of the shRNA-resistant form of Dock6 with the S1194A mutation reversed the Dock6#1 shRNA–mediated decreases in axon extension and the number of axons but not the decrease in axon branching. In contrast, expression of the shRNA-resistant form of Dock6 with the S1194E mutation increased axon branching but could not rescue the Dock6#1 shRNA–mediated decreases in axon extension and the number of axons (Fig. 6, E and F).

Moreover, in dorsal root ganglion neurons isolated from Dock6 shRNA transgenic mice, the reintroduction of Dock6#1 shRNA–resistant S1194A or S1194E Dock6 resulted in effects on axon length, number of axons, and number of branch points similar to those observed in the transfection studies on rat dorsal root ganglion neurons (fig. S6, B and C). These results suggest that Dock6 plays a key role in axon extension and that transgenic mouse neurons also preserve the regulatory mechanism through the phosphorylation of Ser1194 in Dock6.

Akt inactivates Dock6 by phosphorylating Ser1194

Next, to clarify whether Akt changes the Rac1-GEF activity of Dock6 after its phosphorylation, we performed an affinity precipitation assay with guanine nucleotide–free Rac1G15A. In 293T cells, coexpression with myc-Akt1 decreased the amount of Dock6 that precipitated with Rac1G15A by ~50% (fig. S7, A and B). To further assess the effect of PP2A on the activity of phosphorylated Dock6, we coexpressed the PP2A Cα subunit with Dock6 and Akt1 and performed affinity precipitation with Rac1G15A. Coexpression of the PP2A Cα subunit with Dock6 and Akt1 increased the affinity precipitation by ~2.4-fold, compared to coexpression of Dock6 and Akt1 (Fig. 7, A and B). Collectively, these results suggest that PP2A binds to Dock6 and dephosphorylates phosphorylated Ser1194 to attenuate the inhibitory effect of Akt, indicating that Akt-phosphorylated Dock6 is a substrate for PP2A.

Fig. 7

Active Dock6 stimulates the activity of Rac1. (A) GST-Rac1G15A precipitates from 293T cells expressing myc-Akt1 and FLAG-Dock6 with or without V5-PP2A Cα were immunoblotted with an anti-FLAG antibody. Total amount of each protein is also shown. (B) The amount of precipitated Dock6 was normalized to the amount of total Dock6 measured. Data were evaluated using Student’s t test. *P < 0.01, n = 3 experiments. (C) GST-CRIB precipitates from dorsal root ganglion neuron lysates were immunoblotted with an anti-Rac1 (lane 1) or anti-Cdc42 (lane 3) antibody. Total Rac1 (lane 2) and Cdc42 (lane 4) are also shown. Data are representative of three experiments. (D) Cell lysates of primary dorsal root ganglion neurons were prepared on day 2, 8, or 15 in culture in the absence or presence of Akt inhibitor or fostriecin with NGF. GST-CRIB precipitates from cell lysates were immunoblotted with an anti-Rac1 antibody. Total Rac1 is also shown. (F) GST-Rac1G15A precipitates from lysates in (D) were immunoblotted with an anti-Dock6 antibody. Total Dock6 is also shown. (E and G) The amount of precipitated active Rac1 (E; n = 3 experiments) or precipitated active Dock6 (G; n = 4 experiments) was normalized to the amount of total Rac1 (E) or Dock6 (G). Data were evaluated by one-way ANOVA. *P < 0.01, **P < 0.05. (H) Dorsal root ganglion neurons expressing pCMV–FLAG–shRNA-resistant WT Dock6, pCMV–FLAG–shRNA-resistant Dock6 S1194A, or pCMV–FLAG–shRNA-resistant Dock6 S1194E and pSIREN-ZsGreen-Dock6#1 shRNA were cultured with NGF. GST-Rac1G15A precipitates from these cells were immunoblotted with an anti-Dock6 antibody. Total Dock6 is also shown. (I) The amount of active Dock6 was normalized to that of total Dock6 protein. Data were evaluated by one-way ANOVA. *P < 0.01, n = 3 experiments. (J and L) GST-CRIB precipitates from lysates were immunoblotted with an anti-Rac1 (J) or anti-Cdc42 (L) antibody. Total Rac1 or Cdc42 is also shown. (K and M) The amount of active Rac1 or Cdc42 was normalized to that of total Rac1 or Cdc42, respectively. Data were evaluated by one-way ANOVA. **P < 0.05, n = 3 experiments. (N) Schematic diagram describing the regulatory mechanism of axon growth.

In dorsal root ganglion neurons, Rac1 activity remained steady from days 2 to 15 in culture, whereas Cdc42 activity increased starting at day 8 (Fig. 7C). The activity of Rac1 was maintained in culture and increased by pretreatment with an Akt inhibitor (Fig. 7, D and E). Similarly, the Akt inhibitor resulted in increased Dock6 activity, which was increased especially at 15 days (Fig. 7, F and G). Decreased Rac1 activity was also seen in trkA+/− mice (Fig. 6A), indicating that Akt signaling inhibits the activity of Dock6 in dorsal root ganglion neurons. Pretreatment with the PP2 inhibitor fostriecin significantly decreased Rac1 activity (Fig. 7, D and E). Similarly, fostriecin decreased the activity of Dock6 (Fig. 7, F and G), suggesting that PP2A suppresses Rac1 activation in dorsal root ganglion neurons through Dock6.

To determine the activity of the Dock6 mutants in dorsal root ganglion neurons, we cotransfected Dock6#1 shRNA–resistant wild-type, S1194A, or S1194E Dock6 with Dock6#1 shRNA into dorsal root ganglion neurons and measured Dock6, Rac1, or Cdc42 activity by pull-down assay. Expression of the S1194A mutant increased the activity of Dock6 as well as that of Rac1 (Fig. 7, H to K). In contrast, expression of the S1194E mutant resulted in decreased activity of Dock6 as well as that of Rac1 (Fig. 7, H to K). The activity of Cdc42 was not appreciably altered because of the expression of any mutant (Fig. 7, L and M). These results suggest that the S1194A mutant is catalytically active, whereas the S1194E mutant is catalytically inactive in dorsal root ganglion neurons, consistent with the finding that Dock6 activates Rac1 in neurons.

To confirm whether Ser1194 of Dock6 is critical for its GEF activity, we assayed direct nucleotide exchange reaction using a fluorescent N-methylanthraniloyl (mant)–guanine nucleotide compound. Dock6 was immunoprecipitated from unstimulated or NGF-stimulated 293T cells expressing TrkA and either the wild-type Dock6 or the S1194A mutant. Both Dock6 from unstimulated cells (closed circles) and Dock6 from the S1194A mutant (open squares), but not Dock6 from NGF-stimulated cells (open circles), promoted nucleotide release from mant-GDP–loaded Rac1 protein (fig. S7C). The purified, catalytic DHR-2 domain of Dock6 (closed squares; a positive control) promoted nucleotide release from mant-GDP–loaded Rac1 (fig. S7D). Recombinant Dock6 itself promoted nucleotide release (fig. S7D, closed circles); however, the amount of nucleotide released by Dock6 phosphorylated by the recombinant Akt1 was comparable to that seen with control vehicle (fig. S7D, open circles). Furthermore, the S1194A mutant (closed squares), but not the S1194E mutant (open circles), showed GEF activities for both Rac1 and Cdc42 (fig. S7, E and F), suggesting that Dock6 has GEF activities for Rac1 and Cdc42 in vitro, as seen previously (15). Together, these results demonstrate the inhibitory effect of Ser1194 phosphorylation on the GEF activity of Dock6 in vitro.

To investigate whether the axon phenotypes caused by each of these mutants of Dock6 in dorsal root ganglion neurons were related to those mediated through Rac1 or Cdc42 or both, we cotransfected an active form of Rac1 (Rac1G12V) or Cdc42 (Cdc42G12V) with Dock6#1 shRNA into dorsal root ganglion neurons. Transfection of Rac1G12V increased axon extension and number but not the number of branch points (fig. S7G). In contrast, transfection of Cdc42G12V increased the number of branch points and modestly increased the extension of axons but did not affect axon number (fig. S7G). These results suggest that both Rac1 and Cdc42 are involved to varying degrees in axon extension, and Rac1 affects axon number, whereas Cdc42 affects branch point number. The S1194A mutant promotes axon extension and increases the number of axons, similar to the Rac1G12V mutant, suggesting that the S1194A mutant, which is catalytically active, can promote Rac1 signaling in dorsal root ganglion neurons. In contrast, the S1194E mutant increases the number of branch points but does not have a strong effect on axon extension or axon number, similar to the Cdc42G12V mutant. The S1194E mutant is catalytically inactive in dorsal root ganglion neurons. Thus, instead of a decreased Rac1 signal, a basal Cdc42 signal in dorsal root ganglion neurons may emerge as the dominant signal. We speculate that the S1194E mutant is not a simple inactive form; rather, it may have aspects of a weak or mild dominant-negative variant of the Rac1 signal in dorsal root ganglion neurons.

In earlier developmental stages, axons grow and elongate; in later stages, axons contact their targets and arborize to form mature connections with the targets (2). In particular, the phenotype seen with the S1194E mutant is similar to that seen in later stages of axon development. This observation is in agreement with immunofluorescence data showing an accumulation of Dock6 phosphorylated at Ser1194 in ventral roots at embryonic day 13.5 (Fig. 5E). These results are again consistent with an inhibitory role of phosphorylation of Ser1194 in Dock6’s GEF activity.

Discussion

Genetic studies indicate that Dbl-family GEFs are key regulators of neurogenesis (26, 27). Drosophila mutants deficient in the gene trio exhibit defects in axon guidance and motility in peripheral neurons as well as in central neurons (28). The Caenorhabditis elegans unc-73 mutant also has defects in axon guidance and motility in motoneurons (29). The Drosophila trio and C. elegans unc-73 genes encode Rac1-GEFs, and their mammalian orthologues, Trio and Kalirin, are also Rac1-GEFs that play important roles in neuronal development (30, 31). Trio-deficient mice exhibit defects throughout neuronal tissue organization (30). Kalirin is also critical in neuronal morphogenesis (31). In addition, genetic studies have identified ced-5, the gene in C. elegans that is orthologous to mammalian dock180, as a Dock180-related protein underlying sensory neuron development in C. elegans amphids (32, 33). Thus, Trio, Kalirin, and Dock180 are conserved GEFs that coordinately control neuronal morphogenesis ranging from Drosophila and C. elegans to mammals. These results also suggest that GEFs for Rho GTPases including Dock180-related proteins could be involved in morphological changes in mammalian neurons. Here, we have characterized the GEF Dock6 as a modulator of PNS axon growth.

The navigation of PNS neuronal axons involves active axon elongation, and in turn branching, which occurs when axons arrive at their target fields. The process requires continuous morphological changes, yet its underlying molecular mechanisms still remain to be understood. Here, we demonstrate that, in dorsal root ganglion neurons, axon extension and branching are regulated by phosphorylation and dephosphorylation of the single serine residue of Dock6 (Fig. 7N). We found that, during initial axon outgrowth, Dock6 forms a complex with PP2A, which inhibits the phosphorylation of Dock6 by Akt, allowing axons to extend. As axons grow, Akt abundance is increased, and Akt binds to Dock6 instead of PP2A and phosphorylates Dock6 at Ser1194, thereby decreasing the GEF activity of Dock6 and inhibiting axon growth. Accordingly, expression of Dock6 harboring the S1194A mutation in dorsal root ganglion neurons results in normal axon extension with less branching, whereas neurons expressing the S1194E mutant exhibit decreased axon extension and increased branching. Furthermore, the difference in the Kd values of PP2 and Akt for Dock6, which are ~256 and ~61.7 nM, respectively, may explain the finding that Akt eventually replaces PP2A and interacts with Dock6 after axon growth. Liu et al. (34) have reported that, in hippocampal neurons at an early developmental stage, PP2A plays a key role in initial axon extension. In contrast, other studies have focused on the contribution of Akt to axon branching rather than axon extension. Expression of constitutively active Akt increases axon branching in hippocampal neurons (35), and expression of myristoylated Akt gives rise to many axon branches in sensory neurons (36), an observation that agrees with our findings that PP2A and Akt antagonistically regulate axon growth through Dock6.

Dock180 is the prototypical member of the Dock180-related protein family and mediates multiple roles in axon development by interacting with various signaling complexes. For example, in the developing spinal cords, netrin-1, a factor secreted from the floor plate, increases the interaction between Dock180 and deleted in colorectal cancer (DCC), a netrin receptor in commissural neurons (37). Dock180 is necessary for netrin-1–induced activation of Rac1 and mediates axon attraction. Dock180 colocalizes with ephrin-B3–positive hippocampal granule cells (38), and the reverse signal from EphB2 to ephrin-B3–positive cells induces the interaction of Dock180 with the cytoplasmic adaptor protein Nck2 (also called Nckβ) and causes axon pruning and retraction in these cells. Dock180, as well as the closely related proteins Dock3 and Dock4, also functions as a component in a complex with GTP-bound RhoG, Elmo, and the adaptor protein CrkII downstream of certain neuronal receptors (39, 40). Similarly, Dock3 and Dock4, molecules that are closely related to Dock180, are contained in this complex, although Dock3 is escorted by WAVE2 in some cases (40, 41). Dock3 also binds to WAVE2 and acts downstream of neurotrhophin receptor TrkB (41). In addition, Dock7 and Dock9 are involved in determining the polarity of hippocampal neurons (42, 43), but how they themselves are regulated remains unclear.

Although various GEFs link signaling through Rho GTPases to many aspects of morphological changes in CNS neurons, such a molecular link in PNS neurons is still not sufficiently understood. Here, we identified Dock6 as a modulator of axon growth in dorsal root ganglion neurons. We previously identified a Dbl-family GEF involved in morphological changes in chick dorsal root ganglion neurons and named it FRG (also called FARP2 or FIR) (44, 45). It is possible that Dock6 and FRG act cooperatively with axon morphogenesis of dorsal root ganglion in mammals. In addition, we identified Dock6 as a substrate for both Akt and PP2A. Also, with respect to the similarities to the role of the Dock7 to Rac1 signaling pathway in the CNS (42), it would be interesting to examine whether Dock6 regulates axon morphological changes of dorsal root ganglion through cytoskeletal rearrangement by Rac1 effector kinase–mediated phosphorylation of stathmin subfamilies or related proteins. Further studies on the activation and inactivation mechanisms of Dock6 would increase our understanding of its role not only in axon morphological changes but also in interactions with Schwann cells. Such studies may clarify its possible involvement in nerve regeneration.

Materials and Methods

Antibodies and inhibitors

The following antibodies were purchased: anti-Rac1 and anti–β-actin from BD Biosciences Pharmingen; anti-FLAG from Sigma-Aldrich Biosciences; anti–phospho-(Ser473) Akt and anti–phospho-Akt substrate from Cell Signaling Technology; anti-myc, anti-hemagglutinin, and anti-V5 from Nacalai Tesque; anti–PP2A C subunit from Millipore; anti-Akt1 from Santa Cruz Biotechnology Inc.; anti-TrkA and anti-CGRP from Abcam; anti–βIII-tubulin from R&D Systems; and horseradish peroxidase–conjugated anti-mouse and anti-rabbit secondary immunoglobulin Gs (IgGs) from GE Healthcare Bio-Sciences. The Alexa Fluor–conjugated anti-mouse or anti-rabbit IgGs were obtained from Invitrogen. The rabbit polyclonal affinity-purified anti-Dock6 antibody was produced as previously described (15). The rabbit antiserum for phospho-(Ser1194) Dock6 was affinity-purified using a phosphorylated peptide RRSREpSPFGNQ-conjugated resin from nonadsorbed fractions of a nonphosphorylated peptide RRSRESPFGNQ-conjugated resin. Akt inhibitor, fostriecin, and LY294002 were purchased from Merck Chemicals.

Plasmids

The coding region of human Dock6 and its fragments encoding the N-terminal (amino acids 1 to 540), DHR-1 (amino acids 541 to 786), middle1 (amino acids 787 to 1025), middle2 (amino acids 1026 to 1488), DHR-2 (amino acids 1489 to 2048), ΔDHR1 (amino acids 1 to 540 combined with amino acids 787 to 2048), and ΔDHR-2 (amino acids 1 to 1488) of Dock6 were ligated into the mammalian expression vector pCMV-FLAG, as previously described (15). The plasmid pCMV–FLAG–full-length Dock7 was constructed, as previously described (16). Full-length Dock8 was made on the basis of the human complementary DNA (cDNA) clones FLJ00346 and FLJ36253 (National Institute of Technology and Evaluation, Chiba, Japan) and ligated into pCMV-FLAG (16). The constructs of the full-length Dock6 harboring the Ser1194-to-Ala (S1194A) or Ser1194-to-Glu (S1194E) mutation were created from pCMV-FLAG-Dock6 using the QuickChange Site-Directed Mutagenesis Kit (Stratagene). S1194A or S1334A mutants of the middle2 region were produced from the pCMV-FLAG-Dock6 middle2 region plasmid. Constructs of the full-length Rac1 or Cdc42 harboring Gly12-to-Val (Rac1G12V or Cdc42G12V) were created from pCMV-Rac1 or pCMV-Cdc42 plasmids (15, 16, 20). The region encoding the PP2A Cα subunit was amplified from cDNA of rat dorsal root ganglion neurons and inserted into pCMV-V5. The Escherichia coli GST tag expression plasmids pET42a-Rac1·GTP binding domain (CRIB domain) of Pak1, pET42a-Rac1, pET42a-Cdc42, and pET42a–guanine nucleotide–free Rac1 (Rac1G15A) were constructed as previously described (20). The pCMV–myc-Akt1 plasmid was provided by U. Kikkawa (Kobe University, Kobe, Japan). All sequences were confirmed using automatic sequencers (Applied Biosystems).

Preparation of plasmids encoding sequences for shRNAs

The oligonucleotides used were the following: Dock6#1 (starting from nucleotide 661 of rat Dock6) sense oligonucleotide, 5′-GATCCGCCCTCAGGAGGCAGCACCTTCAAGAGAGGTGCTGCCTCCTGAGGGCTTTTTTACGCGTG-3′; Dock6#1 antisense oligonucleotide, 5′-AATTCACGCGTAAAAAAGCCCTCAGGAGGCAGCACCTCTCTTGAAGGTGCTGCCTCCTGAGGGCG-3′; Dock6#1 for generation of Dock6 shRNA TG mice (starting from nucleotide 661 of mouse Dock6) sense oligonucleotide, 5′-GATCCGCCCTCAGGCGGCAGCACCTTCAAGAGAGGTGCTGCCGCCTGAGGGCTTTTTTACGCGTG-3′; Dock6#1 for generation of Dock6 shRNA TG mice antisense oligonucleotide, 5′-AATTCACGCGTAAAAAAGCCCTCAGGCGGCAGCACCTCTCTTGAAGGTGCTGCCGCCTGAGGGCG-3′; Dock6#2 (starting from nucleotide 1069 of rat Dock6) sense oligonucleotide, 5′-GATCCCCCTACATGGTGATGAAGGTTCAAGAGACCTTCATCACCATGTAGGGTTTTTTACGCGTG-3′; Dock6#2 antisense oligonucleotide, 5′-AATTCACGCGTAAAAAACCCTACATGGTGATGAAGGTCTCTTGAACCTTCATCACCATGTAGGGG-3′; Rac1 (starting from nucleotide 187 of rat Rac1) sense oligonucleotide, 5′-GATCCGATTATGACAGACTGCGTCTTCAAGAGAGACGCAGTCTGTCATAATCTTTTTTACGCGTG-3′; Rac1 antisense oligonucleotide, 5′-AATTCACGCGTAAAAAAGATTATGACAGACTGCGTCTCTCTTGAAGACGCAGTCTGTCATAATCG-3′; Cdc42 (starting from nucleotide 204 of rat Cdc42) sense oligonucleotide, 5′-GATCCGATTATGACAGACTGCGTCTTCAAGAGACTGTGGATAACTTAACGGTTTTTTTACGCGTG-3′; Cdc42 antisense oligonucleotide, 5′-AATTCACGCGTAAAAAAGATTATGACAGACTGCGTCTCTCTTGAACTGTGGATAACTTAACGGTG-3′; and the control Photinus pyralis luciferase sense oligonucleotide, 5′-GATCCGGCCATTCTATCCTCTAGAGTTCAAGAGACTCTAGAGGATAGAATGGCCTTTTTTAGATCTC-3′; luciferase antisense oligonucleotide, 5′-AATTCAGATCTAAAAAAGGCCATTCTATCCTCTAGAGTCTCTTGAACTCTAGAGGATAGAATGGCCG-3′. The luciferase nucleotide sequence does not have any notable homology to mammalian gene sequences. Each of the sense oligonucleotides for Dock6 #1, Dock6 #2, Rac1, Cdc42, and control luciferase was each annealed with its corresponding antisense oligonucleotide. The annealed duplexes were ligated into pSIREN-RetroQ-ZsGreen1 (Takara Bio).

Cell culture

Rodents (rats and genetically modified and unmodified mice) were cared for in accordance with the protocol approved by the Japanese National Research Institute for Child Health and Development Animal Care Committee and were monitored by the Laboratory Animal Facility of the Japanese National Research Institute for Child Health and Development. Dorsal root ganglion neurons were dissociated from embryonic day 15 (E15) rats or E13 mice and plated on collagen I–coated 22-mm coverslips as previously described (46). Dorsal root ganglion neurons were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) and NGF (50 ng/ml) with or without 10 μM Akt inhibitor, 10 μM fostriecin, or 10 μM LY294002. For plasmid transfection and single-cell immunofluorescence studies, dorsal root ganglion neurons were fully dissociated into single cells, plated onto 22-mm coverslips coated with collagen I, and cultured with NGF (50 ng/ml) for 48 hours. We randomly selected 10 to 20 cells per 22-mm cover glass in the independent experiments. Neurites longer than two cell bodies were counted as axons. The longest neurite on each cell was considered the longest axon and was traced and measured from the cell body to its distal end using ImageJ software (National Institutes of Health; http://rsb.info.nih.gov/ij/). Total axon length was also manually measured. The total number of branches exceeding 10 μm in length was counted and divided by the number of axons. The mean of this ratio serves as an estimation of the number of branching points per main neurite. 293T cells were cultured in tissue culture dishes in DMEM containing 10% FBS, penicillin (50 U/ml), and streptomycin (50 μg/ml). To confirm cell viability under these conditions, we stained dorsal root ganglion neurons and 293T cells with 0.4% trypan blue. Attached trypan blue–incorporating cells were less than 2.5% of cells in each experiment at 72 hours after transfection.

Plasmid transfection

For primary dorsal root ganglion neurons, various shRNAs or plasmids were transfected using the Nucleofector II Device (Lonza) and the Basic Neuron Nucleofector Transfection Kit (Lonza) according to the manufacturer’s instructions. To confirm cell viability under these experimental conditions, we stained dorsal root ganglion neurons with trypan blue. Trypan blue–positive attached cells in culture dishes numbered <5% at 48 hours after shRNA transfection (0.7 ± 0.02 for control luciferase shRNA, 0.7 ± 0.08 for Dock6 #1 shRNA, 0.9 ± 0.09 for Dock6 #2 shRNA, 0.8 ± 0.06 for Rac1 shRNA, and 0.7 ± 0.02 Cdc42 shRNA). For 293T cells, plasmids were cotransfected using the CalPhos transfection reagent according to the manufacturer’s instructions (Takara Bio).

Generation of Dock6 shRNA transgenic mice

The DNA fragment (~0.3 kb) containing a Dock6 shRNA transcription unit of the pSIREN-RetroQ-ZsGreen shDock6#1 plasmid (base number 6306-2) was digested from the vector backbone with Bgl II and Eco RI, purified, and injected into fertilized C57BL/6JJmc oocytes. Transgenic founder mice and offspring transgenic mice were identified with genomic polymerase chain reaction (PCR) using tail DNA [5′-GAGGGCCTATTTCCCATGATTCCTTCATATTTGCATATACG-3′ and 5′-GGATCCTCGTCCTTTCCACAAGATATATAAAGCCAAG-3′ for the human U6 promoter to amplify the transgene, and glyceraldehyde-3-phosphate dehydrogenase primers (Applied Biosystems) for the gapdh gene as the positive control amplification]. Genomic PCR was performed in 40 cycles, each consisting of denaturation at 94°C for 1 min, annealing at 65°C for 1 min, and extension at 72°C for 1 min. Transgenic founders were mated to wild-type C57BL/6JJms, and the transgene was stably transmitted to offspring. The transgenic mice (TG line#1 and line#2), as well as their nontransgenic control littermates, were normally fertile under standard breeding conditions.

TrkA knockout mice

TrkA knockout mice (on a mixed 129/S2 × C57BL/6J background) were obtained from the Jackson Laboratory. Heterozygous offspring were mated to wild-type C57BL/6JJms mice and used for experiments. The genomic PCR for identification of the knockout allele was performed according to the Jackson Laboratory’s standard protocol.

Quantitative real-time PCR

Total RNA was extracted with ISOGEN (NIPPON GENE CO.) from tissues of Dock6 shRNA transgenic mice. The cDNAs were prepared from 1 μg of total RNA using the QuantiMir RT Kit Small RNA Quantitation System (System Biosciences). Quantitative real-time PCR was performed with a Thermal Cycler Dice Real Time System and SYBR Premix Ex Taq (Takara). Quantification of PCR products was performed by means of the absolute quantification method according to the manufacturer’s instruction. The primers 5′-GAGGGCCTATTTCCCATGATTCCTTCATATTTGCATATACG-3′ and 5′-GGATCCTCGTCCTTTCCACAAGATATATAAAGCCAAG-3′ were used for the human U6 promoter to amplify the transgene. Primers for the housekeeping gene gapdh for normalization were purchased from Applied Biosystems.

Surgery for sciatic nerve injury

Six- to 8-week-old Dock6 shRNA transgenic mice and their littermate controls were used for the experiments. Under general anesthesia, the left sciatic nerve was crushed for 10 s with fine forceps. Three or 7 days after surgery, sciatic nerves were collected, and their longitudinal sections were processed for immunohistochemical study.

Immunofluorescence

Dorsal root ganglion neurons on collagen-coated coverslips were fixed with 4% paraformaldehyde (16, 47, 48). The fixed cultures were permeabilized, blocked, and then incubated with primary antibodies and fluorescence-labeled secondary antibodies. The coverslips were mounted with the Vectashield reagent (Vector Laboratories). The confocal images were collected with an IX81 microscope with a laser-scanning FV500 or FV1000D system (Olympus) and analyzed with FluoView software (Olympus). The fluorescence images were captured with a fluorescence microscope system (DMI4000B, Leica) and analyzed with AF6000 software (Leica).

Immunohistochemistry

Tissues were fixed with 4% paraformaldehyde and embedded in Tissue-Tek reagent (Sakura Finetechnical). Microtome sections were blocked, stained with antibodies, and mounted with Vectashield reagent. The confocal images were captured with an IX81 microscope with a laser-scanning FV500 or FV1000D system (Olympus) and analyzed with FluoView software (Olympus). The fluorescence images were captured with a BX51 microscope system (Olympus) and analyzed with DP2-BSW software (Olympus).

Immunoprecipitation and immunoblotting

Cells or homogenized tissues were lysed in lysis buffer B [50 mM Hepes-NaOH (pH 7.5), 20 mM MgCl2, 150 mM NaCl, 1 mM dithiothreitol, 1 mM phenylmethane sulfonylfluoride (PMSF), leupeptin (1 μg/ml), 1 mM EDTA, 1 mM Na3VO4, 10 mM NaF, and 0.5% NP-40] and centrifuged as previously described (16, 47, 48). The supernatants were mixed with protein G resin that was preadsorbed with various antibodies. The immunoprecipitates or proteins in the lysates were denatured and then separated on SDS-polyacrylamide gels. The electrophoretically separated proteins were transferred onto polyvinylidene difluoride membranes, blocked with the Blocking One kit (Nacalai Tesque), and immunoblotted first with primary antibodies and later with peroxidase-conjugated secondary antibodies. The bound antibodies were detected using the Chemi-Lumi One L kit (Nacalai Tesque) or ImmunoStar LD (Wako). The band images were captured with a GT-7300U scanner (Epson) and analyzed with ImageJ software. At least three separate experiments were carried out, and a representative experiment is shown in each of the figures.

The recombinant proteins

FLAG-tagged DHR-1, DHR-2, wild-type, and S1194A proteins of Dock6 were purified from 293T cells, as previously described (15, 16). Briefly, cells were lysed in lysis buffer A [50 mM Hepes-NaOH (pH 7.5), 3 mM MgCl2, 150 mM NaCl, 1 mM dithiothreitol, 1 mM PMSF, leupeptin (1 μg/ml), 1 mM EDTA, and 0.5% NP-40] and centrifuged. The supernatants were mixed with protein G resin (GE Healthcare) that was preadsorbed with an anti-FLAG antibody. Bound FLAG-Dock6 proteins were washed with buffer A containing 50 mM EDTA and eluted with lysis buffer A containing 20 mM FLAG peptide (Sigma-Aldrich). The buffer contained in the fractions was exchanged with reaction buffer [20 mM Hepes-NaOH (pH7.5), 5 mM MgCl2, 150 mM NaCl, 1 mM dithiothreitol, 1 mM PMSF, leupeptin (1 μg/ml), and 1 mM EDTA]. Activated Akt1, ΔPH of Akt1, and PP2A were purchased from Millipore. GST-tagged Rac1, Cdc42, guanine nucleotide–free Rac1, and CRIB were purified using Escherichia coli BL21 (DE3) pLysS (Takara Bio) as previously described (15, 16). Briefly, the transformed E. coli cells were treated with 0.4 mM isopropyl-1-thio-β- d-galactopyranoside at 30°C for 2.5 hours and harvested by centrifugation. A cell-free extract was made with lysozyme (500 μg/ml) and DNase I (100 μg/ml) in extraction buffer [50 mM tris-HCl (pH 7.5), 5 mM MgCl2, 1 mM dithiothreitol, 1 mM PMSF, leupeptin (1 μg/ml), 1 mM EDTA, and 0.5% NP-40]. All purification steps were performed at 4°C. The lysates were centrifuged, and the supernatants were mixed with glutathione resin (GE Healthcare). Bound proteins were washed with extraction buffer and eluted with extraction buffer containing 20 mM glutathione. The buffer contained in elution fractions was dialyzed against reaction buffer with 0.1 μM GDP for GST-Rac1 and GST-Cdc42 or against that without 0.1 μM GDP for GST-CRIB and GST-Rac1G15A.

In vitro phosphorylation reaction

Purified immobilized FLAG-Dock6 proteins (100 to 250 ng) were incubated with 20 μM cold ATP in the presence or absence of 100 ng of active Akt1 or 0.01 to 0.05 U of PP2A in 30 μl of reaction buffer at 30°C for 20 min and then chilled on ice. Phosphorylated proteins were detected by immunoblotting.

Affinity precipitation of active Rho-GEFs

Dock6 proteins in the cell lysates were affinity-precipitated with 20 μg of GST-Rac1G15A, which is a guanine nucleotide–free Rho GTPase. A Gly-to-Ala mutation of residue 15 in Rac1 decreases its nucleotide binding (22). Active GEFs preferentially interact with guanine nucleotide–free forms of the small GTPases (22). Affinity-precipitated GEFs were detected by immunoblotting (16).

Affinity precipitation of active Rho GTPases

To detect active GTP-bound Rac1 or Cdc42 in the cell lysates, we performed affinity precipitation with 20 μg of GST-CRIB (15, 16). Each of the affinity-precipitated Rho GTPases was detected through immunoblotting with an anti-Rho GTPase antibody.

Guanine nucleotide exchange assays

For the guanine nucleotide assay, mant-GDP–bound GST-Rac1 or GST-Cdc42 was obtained by incubation with reaction buffer containing Rac1 (125 ng/μl) or Cdc42, bovine serum albumin (BSA) (250 ng/μl), and 100 μM mant-GDP at 30°C for 90 min, followed immediately by chilling on ice. Immobilized FLAG–Dock6–DHR-2 domain (125 ng) or 250 ng of FLAG-Dock6 proteins was incubated in 30 μl of reaction buffer containing GST–Rac1–mant-GDP (16 ng/μl) or GST–Cdc42–mant-GDP, BSA (33 ng/μl), and 100 μM GTPγS at 30°C for 0 to 15 min. Changes in mant-GDP fluorescence (excitation, 295 nm; emission, 335 nm) were monitored on an Infinite M200 reader (Tecan Systems Inc.).

Statistical analysis

Values shown represent the means ± SD from separate experiments. Student’s t test was carried out for intergroup comparisons (*P < 0.01; **P < 0.05). A one-way ANOVA was followed by a Fisher’s protected least significant difference or Bonferroni test as a post hoc comparison (*P < 0.01; **P < 0.05).

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/6/265/ra15/DC1

Fig. S1. Estimation of the specificity of an anti-Dock6 antibody and efficiency of knockdown of various shRNAs in rat dorsal root ganglion neurons.

Fig. S2. Generation of Dock6 shRNA transgenic mice.

Fig. S3. Analysis of peripheral neuronal fibers and the number of CGRP-positive neurons in Dock6 shRNA transgenic mice.

Fig. S4. Binding of the DHR-1 region of Dock6 to Akt, binding of the kinase region of Akt to Dock6, and identification of the phosphorylation site of Dock6 targeted by Akt.

Fig. S5. Binding of the DHR-2 region of Dock6 to PP2Ac, the effect of PP2Ac on Dock6 binding to Rac1, and requirement of PP2 in axon growth and Dock6 dephosphorylation.

Fig. S6. Reintroduction of Dock6 constructs into Dock6 shRNA transgenic mouse dorsal root ganglion neurons.

Fig. S7. Attenuation of the GEF activity of Dock6 by phosphorylation of Ser1194 and the effect of active Rho GTPases in dorsal root ganglion neurons.

References and Notes

Acknowledgments: We thank E. M. Shooter (Stanford University), Y. Kaziro (Kyoto University), H. Koga (Kazusa DNA Research Institute), and H. Saito and K. Nakamura (National Research Institute for Child Health and Development) for their participation in insightful discussions and for their encouragement. We thank K. Spicer and W. Fumanski for reading this manuscript. Funding: This work was supported by Grants-in-Aid for Scientific Research from both the Japanese Ministry of Education, Culture, Sports, Science, and Technology (to J.Y.) and the Japanese Ministry of Health, Labor, and Welfare (to Y.M., A.T., and J.Y.) and partially by research grants from the Astellas Foundation (to J.Y.), the Kowa Life Science Foundation (to Y.M.), the Mochida Foundation (to Y.M. and J.Y.), the Takeda Science Foundation (to Y.M.), and the Uehara Memorial Foundation (to Y.M.). Author contributions: Y.M. and J.Y. designed the experiments. Y.M., J.Y., T.T., and N.Y. performed the experiments. Y.M. and T.O. performed the surgery for sciatic nerve injury. Y.M. and J.Y. made the figures. Y.M. and J.Y. wrote the paper. Y.M., J.Y., and A.T. analyzed the data. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article