Introduction

Signaling pathways that lead to neurite outgrowth and the establishment of neuronal polarity remain poorly understood. However, dynamic rearrangements of microtubules and actin filaments at the tips of growing axons are observed during neurite sprouting and elongation, suggesting that molecules that coordinate microtubule and actin microfilament dynamics play key roles in neurite extension and neuronal polarity determination (Fukata et al, 2002; Baas and Buster, 2004). Cytoplasmic dynein light-chain component, Tctex-1, was demonstrated recently to play key roles in initial neurite sprouting, axonal specification and elongation of hippocampal neurons in culture (Chuang et al, 2001, 2005). Cytoplasmic dynein is a microtubule-based, minus-end-directed motor complex involved in various cellular activities, including retrograde trafficking in neurons, Golgi maintenance, breakdown of the nuclear envelope and mitosis (Hirokawa, 1998; Sakato and King, 2004). Dynein comprises two 530 kDa heavy chains (DHCs) with ATPase and motor activities, two or three 74-kDa intermediate chains (DICs), and a group of accessory polypeptides including light intermediate chains and light chains (DLCs) (Vallee et al, 2004). DICs link accessory proteins, including the DLCs and the dynactin complex, to the DHC (Waterman-Storer et al, 1995; King, 2000). Three distinct DLC families have been identified: Tctex-1 (DYNTL1; Pfister et al, 2005), LC8 and LC7/Roadblock (Vallee et al, 2004).

Several lines of evidence suggest that Tctex-1 might function independently from dynein. First, there is strong biochemical evidence that a dynein-free pool of Tctex-1 exists independently of the dynein complex-associated Tctex-1 (Tai et al, 1998). Secondly, Tctex-1 is abundantly expressed in postmitotic, young neurons and was demonstrated recently to play a key role in neuritogenesis in hippocampal neurons in culture (Chuang et al, 2001, 2005). Cultured hippocampal neurons develop multiple abnormally long neurites when Tctex-1 is overexpressed and fail to develop neurites when Tctex-1 is suppressed (Chuang et al, 2005). The function of Tctex-1 in neuritogenesis was demonstrated to be dynein independent, since a mutant of Tctex-1 (Tctex-1 T94E), which failed to bind to DIC and therefore could not get incorporated into the dynein complex, induced a similar phenotype as the wild-type Tctex-1 protein (Chuang et al, 2005). Finally, Tctex-1 interacts with proteins besides DIC, including rhodopsin (Tai et al, 1999), parathyroid hormone receptor (PTHR) (Sugai et al, 2003), poliovirus receptor CD155 (Ohka et al, 2004), Herpes virus capsid protein VP26 (Douglas et al, 2004), bone morphogenetic receptor type II (BMPR-II) (Machado et al, 2003), the voltage-dependent anion channel (VDAC) (Schwarzer et al, 2002), Fyn kinase (Kai et al, 1997; Mou et al, 1998) and Trk neurotrophin receptor (Yano et al, 2001; Yano and Chao, 2004). How dynein-associated and dynein-free pools of Tctex-1 are maintained in the cell and how the assembly of Tctex-1 into the dynein complex is regulated is not known.

Tctex-1 was independently identified as an activator of G protein signaling 2 (AGS2), a receptor-independent activator of heterotrimeric guanine-nucleotide-binding regulatory proteins (G proteins), in a functional yeast screen in yeast (Takesono et al, 1999). AGS molecules can generally be divided into three subgroups: those that directly activate G, those that modulate G–G interaction by binding to G, and those that modulate G–G interaction by binding to G (Lanier, 2004). These AGS molecules have since demonstrated novel, non-canonical roles for G protein subunits in cell development and differentiation. For example, silencing of AGS3, a receptor-independent activator of G signaling, resulted in defects in mitotic spindle orientation and cleavage plane determination of neural progenitors in the developing neocortex demonstrating a novel role for G proteins in regulating neuronal cell fate (Sanada and Tsai, 2005).

AGS2/Tctex-1 was reported to bind G, but the molecular mechanism and functional consequences of the putative Tctex-1–G interaction have remained unknown (Takesono et al, 1999). Here, we studied the role of G in regulating the dynein-independent function of Tctex-1 in neuritogenesis. An endogenous G–Tctex-1 complex can be isolated from embryonic brain lysates and G overlaps with the 'dynein-free' Tctex-1 in cell fractionation experiments. Subcellular distribution of G and Tctex-1 overlap in the cell bodies as well as the growth cones of nascent axons in stage 3 primary cultured hippocampal neurons. Both Tctex-1 and G overexpression elicit similar phenotypes in primary hippocampal neurons. Interfering with G function inhibits neuritogenesis and diminishes the ability of Tctex-1 to induce neurite outgrowth. All known G isoforms contain an identical consensus Tctex-1-binding motif first described in DIC (Mok et al, 2001). We show that full-length G1, as well as a G-peptide corresponding to the Tctex-1-binding region of G1, compete with DIC for Tctex-1 binding. We propose that G binds to Tctex-1 to regulate the dynein-independent pool of Tctex-1 and its incorporation into the dynein motor complex. Finally, we demonstrate that G–Tctex-1 complex plays a key role in neuritogenesis in an established model of hippocampal neuron differentiation.

Results

Characterization of G–Tctex-1 interaction

Tctex-1 was shown to interact specifically with G and not with the G subunit of heterotrimeric G proteins (Takesono et al, 1999). We confirmed the interaction of G with Tctex-1 using purified components. We generated GST-tagged Tctex-1 and purified visual G protein transducin, G_t. GST–Tctex-1 was incubated with G_t and the bound G_t subunit was detected using an anti-G antibody. As shown in Figure 1A, GST–Tctex-1 but not GST alone, specifically bound G_t. We also studied the G–Tctex-1 interaction in a mammalian cell expression system. We cotransfected human embryonic kidney (HEK) cells with Tctex-1 and G1 expression plasmids. Confirming the in vitro data, we showed that Tctex-1 could robustly co-immunoprecipitate (co-IP) with G (Figure 1B). Tctex-1 did not co-IP with overexpressed G_t in HEK cells (data not shown). The G subunit consists of six different isoforms, including G1 through 5 and a splice variant of G5 called G5L (long form). G1–4 are about 80% identical to each other, whereas G5 is only about 50% identical to the others. As shown in Figure 1C, all FLAG-tagged G subunits (G1, 2, 3, 5 and 5L) co-immunoprecipitated with Tctex-1 using an anti-Tctex-1 antibody. The expression of G1 or G2 together with G1 did not affect the ability of Tctex-1 to associate with G1. G1 and G2 expression was confirmed by stripping and reprobing the membranes with anti-G1 and anti-G2 antibodies, respectively (data not shown). Even though both G and Tctex-1 are expressed in HEK 293 cells, we failed to co-IP a native endogenous G–Tctex-1 complex from these cells. Since Tctex-1 is abundantly expressed in postmitotic young neurons and several G combinations are found in brain, we employed embryonic mouse brain lysates in pull-down assays (Chuang et al, 2001). As shown in Figure 1C, we were able to pull down G along with Tctex-1, demonstrating a native endogenous G–Tctex-1 complex in mouse embryonic brain lysate.

To map the G-binding site on Tctex-1, we constructed N- and C-terminal truncation mutants of Tctex-1. FLAG-tagged full-length Tctex-1 and various truncation mutants of Tctex-1 were cotransfected with G1 in HEK cells (Figure 2A). As shown in Figure 2B, full-length and the N-terminal truncated mutants of Tctex-1 were able to co-IP G, whereas the C-terminal truncation mutant of Tctex-1, 1–92, which is missing the last 21 amino acids of Tctex-1, failed to co-IP G (Figure 2B). Taken together, these results suggest that the C-terminal tail of Tctex-1 is required for the G–Tctex-1 interaction.

Consensus binding motif on G is required for Tctex-1 binding

An 11-amino-acid peptide derived from DIC was used to identify residues on Tctex-1 that are involved in the DIC–Tctex-1 interaction (Mok et al, 2001). Comparison of this peptide sequence of DIC with sequences of other Tctex-1 target proteins helped to identify a motif of basic residues, R/K-R/K-X-X-R/K, as a consensus Tctex-1-binding motif (Mok et al, 2001). Additionally, a second Tctex-1-binding motif, V-S-K/H-T/S-X-V/T-T/S-N/Q-V, has also been identified in a subset of Tctex-1-interacting proteins (Sugai et al, 2003). We found a potential Tctex-1-binding motif in all six G subunits, which maps to the outermost -strand of the seventh blade of the G propeller (Figure 3A and B). We generated a G-peptide corresponding to amino acids 40–57 of G1, which includes the basic Tctex-1-binding motif and surrounding residues either with no tag or with a thiol-specific, environmentally-sensitive, fluorescent compound, MIANS. We performed fluorescence anisotropy measurements to determine the binding affinity of the G-peptide to Tctex-1. Keeping the MIANS-labeled G-peptide constant and adding increasing concentrations of purified Tctex-1, we observed enhanced binding, which leveled off at high concentrations of G-peptide (Figure 3C). The experimentally determined K_d for the G-peptide–Tctex-1 interaction was 1.35 M.

We further dissected the binding motif to identify which particular residues within the binding motif are critical for the G–Tctex-1 interaction. The cluster of basic residues within the Tctex-1-binding motif (R/K-R/K-X-X-R/K) prompted us to evaluate the role of charge-based regulation of the interaction. We generated site-directed mutants of the Tctex-1-binding motif in G, where the Arg residues were replaced either individually or in combination, and tested their ability to interact with Tctex-1 in co-IP experiments. In accordance with our hypothesis, we found that mutation of all three basic residues (R48, R49 and R52) to alanines within the full-length G protein (G1 AAA) significantly decreased the ability of G to interact with Tctex-1 (Figure 3D).

G and Tctex-1 distribution overlaps in cultured hippocampal neurons

Tctex-1 was recently shown to regulate neurite outgrowth in primary hippocampal neurons (Chuang et al, 2005). We decided to investigate whether the endogenous G–Tctex-1 complex, isolated from mouse brain lysates, could play a role in Tctex-1-induced neuritogenesis. To this end, we first examined the expression pattern of G in cultured primary hippocampal neurons. Primary hippocampal neurons adopt the characteristic polarized morphology in a well-defined sequence of developmental stages (Dotti et al, 1988; Fukata et al, 2002). Within 24 h of plating, hippocampal neurons send out several processes of relatively equal length (stage 2). Even though phenotypically the neurons still appear non-polarized, proteins known to be involved in cell polarization already start to display a polarized distribution pattern (Bradke and Dotti, 2000; Fukata et al, 2002; Banker, 2003; Schwamborn and Puschel, 2004). At this stage, G displays diffuse labeling throughout the cell body and the minor processes (Figure 4A–C). As the neurons develop through stages 2–3 and reach stage 3, they adopt the final differentiated polarized neuronal phenotype with one long neurite (the nascent axon) and several minor processes. Approximately 40% of stage 3 neurons examined (n>50) displayed G immunoreactivity in the central region of the axonal growth cones (Figure 4D–K), whereas the remaining neurons failed to show an enhanced labeling intensity of G in the growth cones (Figure 4L–N).

We then examined the expression patterns of G and Tctex-1 in hippocampal neurons of various stages. Tctex-1 and G are expressed essentially homogenously throughout the cell body and processes of stage 2 neurons (Figure 5Aa–c). As the neurons develop through stages 2–3, G and Tctex-1 continue to overlap in the perinuclear Golgi region within the cell body. We also observed a strong colocalization of Tctex-1 and G within the axonal growth cones of majority of stage 3 neurons (70%, n=100) (Figure 5Ad–i and 5B).

Role of G and Tctex-1 in neuronal differentiation

To study the role of G in hippocampal neuronal differentiation, we used a G-specific antagonist, ARKct, to perform loss-of-function experiments. ARKct contains the G-binding domain of -adrenergic receptor kinase 1 and when expressed in intact cells, it inhibits G-dependent signaling by binding to and sequestering G (Koch et al, 1994). Hippocampal neurons transfected with EGFP expression vector alone were indistinguishable from non-transfected cells and developed normally through stage 2 and stage 3 (Figure 6A). Quantification of the different stages showed that a significant fraction of neurons transfected with EGFP alone were found in stage 2 (52%) and stage 3 (40%) (Table I). To study the effect of ARKct on neurite outgrowth, hippocampal neurons were transfected either 2 h after plating or 12 h after plating with expression vectors for ARKct together with EGFP (Figure 6B and Supplementary Figure 1). As seen in Figure 6B, neurons that were transfected with the EGFP plus ARKct 2 h after plating failed to develop neurites and seemed to be arrested in stage 1 (78%) (Figure 6B and Supplementary Figure 1). The neighboring, untransfected cells (observed within the same field) appear healthy and proceed normally through development reaching stage 3 (Figure 6B). As compared with EGFP alone or un-transfected cells, a significantly smaller fraction of ARKct-transfected cells reached stage 2 (17%) or stage 3 (4%) (Table I). Neurons transfected 12 h after plating seemed to be arrested in stage 2 (Figure 6B). This result is very similar to what was seen with loss of Tctex-1 function wherein most Tctex-1-suppressed neurons had segmented lamellipodia but neither typical neurites nor growth cones (Chuang et al, 2005).

We then performed gain-of-function studies to examine the phenotype of neurons overexpressing G1. Primary hippocampal neurons were transfected 2 h after plating with FLAG-G1 and EGFP expression vectors. In contrast to cells expressing EGFP alone, a significantly greater number of FLAG-G1-transfected neurons developed multiple long neurites (Figure 6A and Table I). Similar to the case of G1 overexpression and as demonstrated previously (Chuang et al, 2005), Tctex-1 overexpression also elicited the same phenotype of multiple long neurites (Figure 6A and Table I), demonstrating that ectopic overexpression of either G1 or Tctex-1 promotes neurite outgrowth. Coexpression of Tctex-1 along with the G-sequestering reagent ARKct seems to partially revert the phenotype induced by ARKct by 24 h after transfection (Figure 6A, Table I and Supplementary Figure 1).

G overlaps with the 'dynein-free' pool of Tctex-1 and regulates assembly of Tctex-1 into the dynein motor complex

Previous studies demonstrated that Tctex-1-mediated neurite outgrowth proceeds via a dynein-independent mechanism (Chuang et al, 2005). However, how dynein-bound and dynein-free pools of Tctex-1 are regulated in neurons remains unclear. We are interested in testing whether the G–Tctex-1 complex may modulate the partition of Tctex-1 into dynein-associated and dynein-free pools. Toward this end, we first examined whether G overlaps with the dynein-associated or dynein-independent pool of Tctex-1. Using velocity density gradient sedimentation of lysates generated from embryonic E15 mouse brains, we demonstrated that, as shown previously, in addition to co-sedimenting with the DIC-containing, 19–20S fractions, Tctex-1 can also be found in the dynein-free, lighter-density fractions (Figure 7A). Interestingly, we show that the majority of G-containing fractions overlap with the lighter, non-DIC containing fraction of Tctex-1. Very little, if any, G can be found in the DIC-containing, higher-density fractions (Figure 7A).

Tctex-1 binding to DIC is required for its incorporation into the dynein motor complex. We observed that all known G isoforms contain a DIC-like, consensus basic Tctex-1-binding motif (R/K-R/K-X-X-R/K) (Figure 3B) and (Mok et al, 2001). Since G contains the same Tctex-1-binding motif as DIC, we investigated whether Tctex-1 can simultaneously bind to G and DIC. We investigated the effect of overexpression of full-length FLAG-tagged G1 on the ability of Tctex-1 to interact with DIC. Figure 7B shows that G1 overexpression results in dramatically reduced Tctex-1 binding to endogenous DIC in HEK 293 cells. We further assessed the DIC–Tctex-1 interaction in the presence of a G-peptide corresponding to amino acids 40–57 of G1, which includes the basic Tctex-1-binding motif and surrounding residues. As shown in Figure 7C, the presence of increasing concentration of the G-peptide leads to decreased amounts of Tctex-1 that can be co-immunoprecipitated with DIC in an anti-DIC pull down.

Discussion

The recent discoveries of novel and non-canonical modes of G protein signaling and regulation have expanded the classical repertoire of receptor-dependent G protein activation (Cismowski and Lanier, 2005; McCudden et al, 2005). AGS proteins are a functionally defined group of proteins that activate G protein signaling in the absence of a classical G protein-coupled receptor (GPCR). It has been proposed that AGS proteins play important roles in the generation or positioning of signaling complexes and may serve as alternative binding partners for G protein subunits. Notably, receptor-independent activation of G proteins has been implicated in proper positioning of the mitotic spindle and asymmetric cell division (Willard et al, 2004). Silencing of AGS3, a receptor-independent activator of G signaling, resulted in defects in mitotic spindle orientation and cleavage plane determination of neural progenitors in the developing neocortex affecting neuronal cell fate (Sanada and Tsai, 2005). AGS2 was identified as a G-binding protein and was shown to be identical to cytoplasmic dynein light-chain Tctex-1 (Takesono et al, 1999). Interestingly, Tctex-1 has been shown to have both dynein-dependent and dynein-independent functions, and there is biochemical evidence for a dynein-free pool of Tctex-1 (Tai et al, 1998; Chuang et al, 2005). It is completely unclear how the dynein-associated and dynein-free pools of Tctex-1 are maintained in the cell.

In this study, we have investigated the functional role of the AGS2/Tctex-1–G complex. We demonstrate that both G and Tctex-1 distribute in primary hippocampal neurons similarly within the cell body and in the growth cone of the future axon of differentiated neurons. As with Tctex-1, ectopic expression of G induces multiple neurites, whereas interfering with G signaling, using a G-sequestering reagent ARKct, inhibits neurite outgrowth. We demonstrate that all known isoforms of G contain a DIC-like Tctex-1-binding motif and ectopic expression of G1 interferes with the Tctex-1–DIC interaction. Moreover a G-peptide, spanning the Tctex-1-binding motif, competes with DIC for Tctex-1 binding, suggesting that G–Tctex-1 interaction may modulate the dynein–Tctex-1 motor complex formation. Since the neuritogenic effects of Tctex-1 are dynein motor activity independent, G binding to Tctex-1 provides a mechanistic model for recruiting and maintaining a dynein-free pool of Tctex-1 and the role of this complex in Tctex-1-mediated neurite outgrowth.

Dual role of Tctex-1

Tctex-1 has been identified as a light-chain subunit of the cytoplasmic dynein motor complex (Vallee et al, 2004), where it acts as a cargo adaptor for the dynein motor (Tai et al, 1999). However, Tctex-1 seems to have two functional roles—one as a dynein motor component and the second as a dynein-independent regulator of cell fate. A dynein-free pool of Tctex-1 has been confirmed biochemically (Tai et al, 1998; Li et al, 2004). Dynein motor activity is dispensable for Tctex-1-mediated modulation of cortical microfilament dynamics and Rac1 activity in embryonic hippocampal neurons (Chuang et al, 2005). G overlaps with dynein-free, less-dense fractions of Tctex-1 in embryonic brain lysates, suggestive of a dynein-independent role for the G–Tctex-1 complex.

We have identified a common Tctex-1-binding motif, which was initially defined by comparing the minimum Tctex-1-interacting sequence of DIC with other known Tctex-1-interacting proteins (Mok et al, 2001), in all known isoforms of G. The sequence alignment demonstrates that all known G subunits contain a DIC-like Tctex-1-binding motif. This region of G maps to the outermost -strand of the seventh blade of the G propeller and has been previously shown to contain residues important for interaction with other known G effectors, including PLC2 and RACK1 (Panchenko et al, 1998; Buck and Iyengar, 2001; Chen et al, 2005). A second Tctex1-binding motif has also been identified, although it was often found that gaps of varying lengths had to be inserted and parts of the sequence were missing (Sugai et al, 2003). Overall, Tctex-1-interacting proteins fall into following three broad subsets: proteins that contain only the core basic R/K-R/K-X-X-R/K consensus motif; proteins that contain only the second Tctex-1-binding motif and proteins that contain both binding motifs. G belongs to the first subset. We show both full-length G and a G-peptide, corresponding to amino acids derived from G1 spanning the putative Tctex-1-binding motif, compete with DIC for Tctex-1 binding.

According to a model proposed by Wu et al (2005), DIC and cargo proteins would bind to opposite ends of the Tctex-1 dimer. However, Tctex-1-interacting proteins that contain only the core, DIC-like Tctex-1-binding motif, present a paradox, since these proteins must interact with Tctex-1 in the same region as DIC (Wu et al, 2005). It is therefore unclear whether proteins containing only the same core binding motif can be bound to the dynein–Tctex-1 complex and serve as cargo for transport by the dynein motor complex. Our data demonstrate that G competes with DIC for Tctex-1 binding. We propose that the G–Tctex-1 complex cannot itself be incorporated in the dynein motor complex. However, G regulates incorporation of Tctex-1 into the dynein motor complex and thereby regulates dynamics of the dynein-free pool of Tctex-1. Structural studies of the Tctex-1–G co-complex are underway to help clarify whether G and DIC can simultaneously bind Tctex-1.

Role of G and Tctex-1 in regulating neuronal differentiation

Several lines of evidence suggest a role for the G–Tctex-1 interaction in neuronal differentiation. Tctex-1 was recently shown to play an important role in multiple steps of neuronal development, including neurite sprouting and extension, axonal polarity and dendritic arbor elaboration (Chuang et al, 2005). Moreover, Tctex-1 is highly abundant in fetal brains and in postmitotic young neurons in adult brain (Chuang et al, 2001). We report that G and Tctex-1 share overlapping expression patterns in primary hippocampal neurons. Both Tctex-1 and G localize in the growth cone of the axons in stage 3 neurons. Disruption of Tctex-1 expression and interfering with G function results in inhibition of neurite outgrowth, whereas ectopic expression of G and Tctex-1 results in multiple axon-like long neurites. Moreover, interfering with G function partially reduced the ability of Tctex-1 to induce its neuritogenic effects, suggesting that the G–Tctex-1 complex is required for full activity of Tctex-1. But how does G regulate Tctex-1-induced neuronal effects? We show that G competes with DIC for Tctex-1 binding and thus regulates the incorporation of Tctex-1 into the dynein motor complex. Since motor activity is dispensable for Tctex-1-mediated neurite outgrowth and it is the dynein-free pool of Tctex-1 that promotes its neuritogenic effects, according to our model (see Figure 7D), G binding to Tctex-1 prevents its incorporation into the dynein complex and thereby promotes its dynein-independent activity.

Ectopic expression of Tctex-1 mostly reverts the dramatic inhibition of neurite outgrowth observed in the presence of the G-sequestering reagent, ARKct. Since both G and Tctex-1 interact with several proteins and regulate various signaling pathways, it is probable that the G–Tctex-1 complex could be one of several signaling complexes leading to regulation of neurite induction and neuronal differentiation. For example, Tctex-1 stimulation of actin dynamics in neurons seems to be Rac dependent (Chuang et al, 2005). Tctex-1 may directly interact with a Rac-GEF to modulate Rac activity. Rho family GTPases are key regulators of actin dynamics that lead to actin-based assemblies associated with the structure and motility of cells. Rho GTPases have emerged as common regulators of actin dynamics that drive growth cone motility. The activity of Rho GTPases is regulated by a variety of proteins that either promote GTP uptake (GEFs) or stimulate hydrolysis (GAPs). Sequential activation of Rap1B and Cdc42 was shown to be essential for axonal specification and the establishment of neuronal polarity in hippocampal neurons (Schwamborn et al, 2004, 2006). G drives the membrane recruitment and activation of PAK-associated PIX, an exchange factor for Cdc42 (Li et al, 2003). This G-PAK1/PIX/Cdc42 complex and phosphatidylinositol 3-kinase (PI3K) product PI(3,4,5)P3 (PIP3) were demonstrated to be essential for actin polymerization at the leading edge and persistent directional migration of myeloid cells (Li et al, 2003). G subunit might serve to link extracellular cues to localized regulation of actin and microtubule dynamics. Rho GTPases and their modulators provide a direct link between cell surface receptors, including growth factor receptors and GPCRs, and cytoskeletal rearrangements, leading to cell motility and cell polarity.

Tctex-1, but not the other dynein subunits, is highly abundant in the two germinal zones of adult brain, namely dentate gyrus and subventricular zone (SVZ) (Dedesma et al, 2006). In the developing neocortex, neural stem cells at the ventricular zone (VZ) either symmetrically divide into two identical daughter cells (e.g., progenitors or postmitotic neurons), or asymmetrically divide into one progenitor and one postmitotic neuron. Sanada and Tsai (2005) showed that perturbation of G signaling resulted in significantly decreased frequency of asymmetric cell division causing an overproduction of postmitotic neurons at the expense of mitogenic neural stem cells (Sanada and Tsai, 2005). Disruption of AGS3, which binds to Gi-GDP, preventing the reassociation of G with G to reform the heterotrimer, leads to a phenotype resembling perturbation of G signaling (Sanada and Tsai, 2005). In the cortical neural progenitor cells, interfering with G function resulted in defects associated with mitotic spindle orientation and cleavage plane determination. But exactly how G regulates mitotic spindle alignment and cleavage orientation and thereby asymmetric cell division of neural stem cells is still unclear. Taken together, it can be concluded that G proteins and molecules that regulate G protein signaling are important players in neuronal development. We show specifically that G and Tctex-1, which individually are important in various pathways involving neuronal development, act in concert to regulate neuritogenesis.

Materials and methods

Antibodies, chemical reagents and plasmid constructs

DIC mAb (clone 74.1) was obtained from Chemicon (Temecula, CA), anti-FLAG pAb and tyrosinated-tubulin (Tyr-Tubulin) (clone TUB1A2) mAb from Sigma-Aldrich (St Louis, MO), anti-pan G antibody, (cat. no. sc-378), which recognizes all G isoforms and anti-G1 and anti-G2, from Santa Cruz Biotechnology (Santa Cruz, CA), horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit IgG from Jackson Immunoresearch Laboratories (Bar Harbor, ME) and Alexa 488-conjugated goat anti-rabbit and Alexa 594-conjugated goat anti-mouse secondary antibodies from Invitrogen (Carlsbad, CA). Isopropyl-1-thio--D-galactopyranoside (IPTG) was obtained from US Biological (Swampscott, MA). Affinity-purified anti-Tctex-1 pAb was described in Tai et al (1998), and additionally, an anti-Tctex-1 mAb antibody was also generated (see Supplementary Figure 2, and J-Z Chuang and C-H Sung, unpublished data). Untagged Tctex-1 cloned into pET-3a was a gift from Dr Zhang (Mok et al, 2001). Expression vectors encoding EGFP, pEGFP-C1 (Clontech Laboratories Inc., Mountain View, CA), untagged and FLAG-tagged Tctex-1 (Chuang et al, 2001), untagged and FLAG-tagged G1 and FLAG-tagged G2, G3, G5 and G5L (UMR cDNA Resource Center, www.cdna.org), and control vector pcDNA3.1 (+) (Invitrogen, Carlsbad, CA) were used.

Peptide synthesis and purification

Unlabeled G-peptide, ⁴⁰-VGRIQMRTRRTLRGHLAK-⁵⁷ and labeled G-peptide, C(MIANS)-⁴⁰VGRIQMRTRRTLRGH-LAK-⁵⁷ [MIANS=2-(4'-maleimidylanilino)napthalene-6-sulfonic acid] (corresponding to the amino-acid sequence of G1) were obtained from the Proteomics Resource Center of Rockefeller University. Peptides were synthesized using F-moc amino acids and had amidated C-terminus. Molecular mass was confirmed by MALDI mass spectrometry, purified to greater than 95% purity by reverse-phase high-performance liquid chromatography (RP-HPLC) and was highly soluble in water.

Protein expression and purification

Tctex-1 was expressed and purified both as a GST-tagged protein as well as an untagged protein as described in Mok et al (2001). Briefly, a three-step FPLC-based column chromatography procedure was used. pET-3a vector encoding mouse Tctex-1 cDNA transformed into Escherichia coli BL21(DE3) host cells and Tctex-1 expression was induced by the addition of 400 M IPTG. The cells were harvested and lysed 3 h after induction and the cleared lysate was loaded onto a HiPrep 16/10 DEAE Sepharose (GE Healthcare) column and the protein was eluted with a linear NaCl gradient of 0–0.3 M. Fractions containing Tctex-1 were pooled, concentrated and loaded onto a sizing column (HiLoad 16/60 Superdex 75 gel filtration column) and excess salt was removed by desalting chromatography (Hi Prep 26/10 desalting column). Fractions from the desalting column were loaded on to a MonoQ HR 5/5 and eluted with a gradient of 0–0.35 M NaCl. The purified Tctex-1 was dialyzed and snap frozen in working aliquots and stored at -80°C. Transducin (G_t) was purified from frozen bovine retinas (Min et al, 1993; Marin et al, 2000). The purified proteins were snap frozen in working aliquots and stored at -80°C.

Fluorescence anisotropy

Interaction between MIANS-labeled G-peptide and Tctex-1 was monitored by fluorescence anisotropy. Anisotropy values were measured as described previously (Krishna et al, 2002). Briefly, the peptide was excited at 325 nm and the fluorescence anisotropy was measured in the absence or presence of Tctex-1 at 440 nm. Ludox was used as a standard in the anisotropy experiments (with an anisotropy value of 0.97).

Cell culture and transfection

HEK 293 cells were grown in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% (v/v) fetal calf serum. All cell cultures were maintained in 5% CO₂ at 37°C. For immunoprecipitation assays, HEK293 cells were plated in 10-cm plates and transfected with 3.5 g DNA/plate using Lipofectamine Plus (Invitrogen). Cell extracts were generated 48 h after transfection.

Immunoprecipitation and peptide competition assay

Cells were lysed in lysis buffer (50 mM Tris–HCl, pH 7.5, 150 mM NaCl, EDTA, 2 mM EDTA, 1.0% (v/v) Triton X-100) plus protease inhibitor mixture (1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 2 g/ml leupeptin and 0.7 g/ml pepstatin). Cell lysates were cleared by centrifugation and the supernatant fractions were then incubated with antibody and Protein A–agarose beads (Repligen, Cambridge, MA) for 2 h at 4°C. For G-peptide competition assay, cell lysates were incubated with either anti-DIC antibody alone or in the presence of increasing concentration of unlabeled G-peptide (1, 10, 25, 50, 100 or 1000 nM). The immunoprecipitates were washed three times with cold lysis buffer. The proteins were released from beads by boiling in SDS sample buffer and separated by SDS–PAGE. Proteins were transferred to nitrocellulose and detected using anti-G (1:1000), anti-Tctex-1 (1:2000), anti-FLAG (1:5000) or anti-DIC (1:1000) antibodies.

Culture and immunofluorescence analyses of hippocampal neurons

Cultures of dissociated embryonic hippocampal neurons were prepared as described previously (Goslin and Banker, 1991; Chuang et al, 2005). Briefly, embryonic E18 rat brains were dissected, the isolated hippocampi were incubated in 1 HBSS (Invitrogen) containing 0.25% trypsin for 15 min at 37°C and dissociated by pipetting in plating medium (minimum essential medium (MEM) supplemented with 10% horse serum, glucose and 100 U/ml penicillin/streptomycin; Invitrogen). Freshly trypsin-dissociated neurons were plated onto glass coverslips coated with poly-L-lysine (Sigma P2636) and cultured at 37°C and 5% CO₂. For endogenous expression analysis, neurons were allowed to attach to the substrate followed by incubation in maintenance medium (MEM supplemented with ovalbumin, B27 and N2.1 supplement, pyruvate, and 100 U/ml penicillin/streptomycin; Invitrogen). Neurons were fixed at 1–3 days in vitro (d.i.v.) with 4% paraformaldehyde/sucrose solution in phosphate-buffered saline (PBS) for 15 min at room temperature (RT) and processed for immunohistochemistry. For transfection studies, neurons were transfected either at 2 h after plating or 12 h after plating using Lipofectamine 2000 (Invitrogen) following the manufacturer's protocol, with minor adjustments. Total DNA amount for each transfection was kept constant using the empty backbone control vector, pcDNA3.1(+). Neurons were incubated with the DNA:lipid mixture for 2 h at 37°C and the medium was then removed and replaced with maintenance medium. Neurons were fixed at the indicated times and processed for immunohistochemistry following published protocols (Chuang et al, 2005). Primary antibody was followed by goat anti-rabbit or anti-mouse secondary antibodies coupled to Alexa 488 or Alexa 594 (Invitrogen). The coverslips were mounted in Vectashield mounting medium (Vector Labs Inc., Burlingame, CA) with DAPI. All immunostained cells were analyzed either using a Zeiss confocal microscope or a Zeiss Axiovert 200 M equipped with an Axiocam HR camera. At least three independent experiments were conducted for each manipulation and 50–100 cells on 15–40 coverslips were examined in each experiment. Different stages of neuritic development were classified based on established guidelines (Dotti et al, 1988; Chuang et al, 2005). Quantification of labeling intensities and morphometric analyses were carried out using Metamorph software (Universal Imaging Co., Downingtown, PA) as described (Chuang et al, 2005) and the images were processed using Adobe Photoshop.

Velocity density gradient sedimentation

Embryonic E15 mouse brains were harvested and rinsed with ice-cold PBS twice. Cells were broken using the Dounce homogenizer (Kimble Kontes, Vineland, NJ) with pestle A (10 times) followed by pestle B (20 times) in 1 ml PEM buffer (80 mM PIPES, pH 6.8, 1 mM EGTA, 1 mM MgSO₄, 0.5 mM DTT and protease inhibitors) on ice. Homogenates were cleared by centrifugation at 2000 r.p.m. for 5 min to remove nuclei followed by centrifugation at 100 000 g at 4°C for 1 h. The high-speed supernatant fraction was fractionated in a 5–20% linear sucrose gradient in 11 ml of Tris–KCl buffer (20 mM Tris, pH 7.5, 50 mM KCl, 5 mM MgCl₂, 0.5 mM EGTA, 0.5 mM DTT and protease inhibitors) using a SW41 Ti rotor (Beckman) at 32 000 r.p.m. in 4°C for 16 h. The S-values were determined by measuring the reflective index of each fraction.

Supplementary data

Supplementary data are available at The EMBO Journal Online (http://www.embojournal.org).

Acknowledgements

PS was supported earlier by the NIH Training Grant EY07138 and is currently a fellow of the Murray Foundation. TPS was formerly an Ellison Medical Foundation senior scholar. This work was supported in part by NIH EY11307 to C-HS and the Alene Reuss Memorial Trust and the Howard Hughes Medical Institute.

References

Baas PW, Buster DW (2004) Slow axonal transport and the genesis of neuronal morphology. J Neurobiol 58: 3–17 | Article | PubMed | ChemPort |

Banker G (2003) Pars, PI 3-kinase, and the establishment of neuronal polarity. Cell 112: 4–5 | Article | PubMed | ChemPort |

Bradke F, Dotti CG (2000) Establishment of neuronal polarity: lessons from cultured hippocampal neurons. Curr Opin Neurobiol 10: 574–581 | Article | PubMed | ISI | ChemPort |

Buck E, Iyengar R (2001) Modular design of Gbeta as the basis for reversible specificity in effector stimulation. J Biol Chem 276: 36014–36019 | Article | PubMed | ChemPort |

Chen S, Lin F, Hamm HE (2005) RACK1 binds to a signal transfer region of Gbeta gamma and inhibits PCLbeta 2 activation. J Biol Chem 280: 33445–33452 | Article | PubMed | ChemPort |

Chuang JZ, Milner TA, Sung CH (2001) Subunit heterogeneity of cytoplasmic dynein: differential expression of 14 kDa dynein light chains in rat hippocam