NIMA-related kinases (Neks) belong to a large family of Ser/Thr kinases that have critical roles in coordinating microtubule dynamics during ciliogenesis and mitotic progression. The Nek kinases are also expressed in neurons, whose axonal projections are, similarly to cilia, microtubule-abundant structures that extend from the cell body. We therefore investigated whether Nek kinases have additional, non-mitotic roles in neurons. We found that Nek3 influences neuronal morphogenesis and polarity through effects on microtubules. Nek3 is expressed in the cytoplasm and axons of neurons and is phosphorylated at Thr475 located in the C-terminal PEST domain, which regulates its catalytic activity. Although exogenous expression of wild-type or phosphomimic (T475D) Nek3 in cultured neurons has no discernible impact, expression of a phospho-defective mutant (T475A) or PEST-truncated Nek3 leads to distorted neuronal morphology with disturbed polarity and deacetylation of microtubules via HDAC6 in its kinase-dependent manner. Thus, the phosphorylation at Thr475 serves as a regulatory switch that alters Nek3 function. The deacetylation of microtubules in neurons by unphosphorylated Nek3 raises the possibility that it could have a role in disorders where axonal degeneration is an important component.
Mammalian NIMA (never in mitosis A)-related kinases (Nek1-Nek11) are a family of Ser-Thr kinases that constitute approximately 2% of the entire human kinome (Fry and Nigg, 1995; O'Connell et al., 2003; O'Regan et al., 2007; Quarmby and Mahjoub, 2005; Schultz et al., 1993). They share approximately 40-45% identity within their N-terminal catalytic kinase domains, but the C-terminal non-catalytic regions are highly divergent (Fry and Nigg, 1995; Fry et al., 1997). The founding member of the Nek family, NIMA, is essential for entry into mitosis in Aspergillus nidulans (O'Connell et al., 2003). Other homologues of NIMA, such as Fin1 and Kin3 from yeast, also have important cell cycle functions via regulation of microtubule organization and mitotic spindle assembly (Jones and Rosamond, 1990; Krien et al., 1998). In addition, the Chlamydomonas NIMA-family kinases Fa2p and Cnk2p regulate ciliary function and length by promoting cilia disassembly (Bradley and Quarmby, 2005; Bradley et al., 2004; Finst et al., 2000; Finst et al., 1998; Mahjoub et al., 2002; Mahjoub et al., 2004; Parker et al., 2007; Quarmby and Mahjoub, 2005), and Nek from Crithidia displays tubulin polyglutamylation activity (Westermann and Weber, 2002; Westermann and Weber, 2003). Finally, Nek kinases influence organ development and morphogenesis in Arabidopsis through their effects on microtubules (Cloutier et al., 2005; Motose et al., 2008; Sakai et al., 2008; Vigneault et al., 2007). Thus, a common theme in Nek function is the regulation of microtubule structures.
The functions of mammalian Nek proteins are largely unknown; however, similarly to their ancestral homologues, they are associated with cilia, centrosomes and microtubule organization (O'Regan et al., 2007; Quarmby and Mahjoub, 2005). For example, Nek2 is crucial for proper mitosis through its effects on centrosomes (Fry et al., 1998; Prigent et al., 2005), and Nek1 and Nek8 are involved in primary cilium formation and polycystic kidney diseases (Mahjoub et al., 2005; Shalom et al., 2008; Upadhya et al., 2000). Nek9 phosphorylates Bicd2, a protein that is important for microtubule maintenance and axonal stability (Holland et al., 2002; Teuling et al., 2008). Nek9 also modulates mitotic progression (Roig et al., 2002) and can activate Nek6 and Nek7 (Belham et al., 2003), which act on centrosomes and modulate microtubule nucleation (Kim et al., 2007; Yin et al., 2003). Nek6 can phosphorylate the kinesin Eg5, which regulates microtubule movement and is necessary for spindle bipolarity during mitosis (Rapley et al., 2008). The involvement of Nek kinases in cell cycle control in particular has stimulated intense scrutiny of their roles in cancer (Hernandez and Almeida, 2006; Kokuryo et al., 2007; Malumbres and Barbacid, 2007; McHale et al., 2008; Miller et al., 2007; Miller et al., 2005; Simpson et al., 2008).
Although Nek kinases are most closely associated with cell cycle regulation and ciliogenesis, recent reports of high Nek expression in the nervous system, including post-mitotic neurons, indicate that they might participate in other cellular processes (Arama et al., 1998; Feige and Motro, 2002; Tanaka and Nigg, 1999). For example, Nek3 levels are higher in quiescent cells, and inhibition of Nek3 activity via antibody injection or mutation did not affect cell cycle progression (Tanaka and Nigg, 1999). Furthermore, although most Nek kinases are located in the centrosome or in cilia, Nek3 is found in the cytoplasm (Tanaka and Nigg, 1999). Nek3 bears a conserved NIMA-related kinase domain at its N-terminus, and the C-terminal non-catalytic region contains two PEST (proline-, glutamate-, serine- and threonine-rich) domains, a protein motif that is implicated in promoting protein turnover or degradation as well as protein-protein interactions (Pongubala et al., 1992; Rogers et al., 1986). Interestingly, a screen for phosphorylated proteins in developing mouse brain identified Nek3 (Ballif et al., 2004). The phosphorylated residue was identified as Thr475, which is located within the C-terminal PEST domain. This residue and the PEST domains are phylogenetically conserved, suggesting that they are important for Nek3 function.
In the present study, we found that several Nek kinases, including Nek3, are highly expressed in neurons from the central and peripheral nervous systems. Given the role of Nek kinases in microtubule regulation, we investigated whether Nek3 regulates microtubules and therefore cytoskeletal dynamics in neurons. We found that cells expressing Nek3 mutants lacking either the conserved PEST domain or containing mutations of Thr475 had severely decreased levels of acetylated α-tubulin. Experiments with deacetylase inhibitors showed that Nek3-mediated alterations in tubulin acetylation resulted from HDAC6 modulation. Neurons expressing these mutants had abnormal morphology and multiple axons, indicating disrupted cell polarity. These results suggest that phosphorylation of Nek3 at T475 is an important regulator of Nek3 activity. The regulation of neuronal microtubule acetylation by Nek3 raises the possibility that it could be involved in disorders where axonal degeneration is an important component.
Nek3 is expressed in post-mitotic neurons
NIMA-related kinases are crucial for regulating microtubules during mitosis; however, several Nek kinase mRNAs are present in distinct neuronal structures in mouse brain and spinal cord (Arama et al., 1998; Feige and Motro, 2002; Tanaka and Nigg, 1999). As microtubule dynamics are important for the formation of neuronal projections, these findings suggested that Nek kinases might regulate microtubule assembly in neurons. We examined the expression of all Nek kinases in adult mouse brain and detected high levels of mRNA encoding Nek1, Nek3, Nek6, Nek7 and Nek9 (data not shown). We also found high expression of these Nek kinases in glia-free dorsal root ganglion (DRG) sensory neuron cultures, indicating that these kinases are expressed in post-mitotic neurons (Fig. 1A).
To examine the potential impact of Nek kinases in post-mitotic neurons, we focused on Nek3 because of its high neuronal expression and its lack of involvement in cell cycle progression or centrosome function (Tanaka and Nigg, 1999). To further characterize Nek3 expression in the nervous system, we examined brain lysates from embryonic and postnatal mice by western blotting. We found that Nek3 was highly expressed during embryogenesis and early postnatal life, but was expressed at lower levels in adult mice (Fig. 1B), indicating it has an important role during early brain development. Furthermore, Nek3 was detected in lysates made from primary cultured DRG and hippocampal neurons (Fig. 1C).
The subcellular distribution of Nek3 in neurons was examined using immunohistochemistry and western blot. Similarly to previous reports using 3T3 fibroblasts, we found that endogenous Nek3 was primarily located in the cytoplasm of DRG neurons (Fig. 1E). To determine whether Nek3 is present in axons, we analyzed lysates from distal axons compared with neuronal cell bodies from mouse DRG explants cultured for 14 days in vitro (DIV), a time when axonal extension allows for a complete separation of distal axons from cell bodies. We found that Nek3 was present at similar levels in both compartments (Fig. 1D), indicating that it is transported into axonal processes. To further examine Nek3 distribution, we used lentivirus infection to express Myc-tagged Nek3 (Myc-Nek3) in cultured hippocampal and DRG neurons. Using anti-Myc immunofluorescence, Nek3 was detected in cell bodies and the axonal shaft, and strong focal Nek3 signals were detected in growth cones (Fig. 1F).
Nek3 is constitutively phosphorylated within the PEST motif at Thr475
Within the Nek3 C-terminal PEST domain, Thr475 is highly conserved and is phosphorylated in embryonic mouse brain (http://phospho.elm.eu.org) (Ballif et al., 2004). To examine the phosphorylation of this residue directly, we first immunoprecipitated Myc-tagged wild-type Nek3 and a phosphorylation-defective mutant [Nek3(T475A)] (as a control) from transfected HeLa cells and analyzed the phosphorylation of these proteins by western blotting using anti-phosphorylated threonine (Thr-P) antibody. We readily detected phosphorylated wild-type Nek3 but no signal was obtained with Nek3(T475A) (Fig. 2A). To further examine the phosphorylation of Nek3 at Thr475, we raised antibodies against a Nek3 peptide containing this phosphorylated Thr residue, and tested against lysates of Nek3-transfected cells (Fig. 2B). This antisera recognized only Nek3 molecules with a Thr at position 475. To confirm that the Nek3 Thr475-P antibody specifically recognized Nek3 phosphorylated at this residue, we treated wild-type Nek3 with calf intestinal alkaline phosphatase (CIP). The anti-Thr475-P signal detected in wild-type Nek3 was lost following CIP treatment (Fig. 2C). Next, we performed a peptide competition study, in which the antibody was pre-incubated with the Nek3 phosphopeptide used as an immunogen or its non-phosphorylated counterpart. Preincubation with the Nek3 peptide containing Thr475-P resulted in a loss of reactivity with Nek3, whereas incubation with the non-phosphorylated peptide did not (Fig. 2D).
Autophosphorylation is an important regulatory step for many kinases and has been observed in several NIMA-family kinases (Fry et al., 1999; Fry et al., 1995; Graf, 2002; Helps et al., 2000; Holland et al., 2002; Lu et al., 1993; Roig et al., 2002). To determine whether Thr475 is phosphorylated in this manner, we expressed Nek3 in HeLa cells and found that this residue was phosphorylated even in a Nek3 mutant (i.e. kdNek3) that lacks intrinsic kinase activity because of an Asp143 to Ala mutation in the kinase domain (Tanaka and Nigg, 1999) (Fig. 2B), indicating that this residue is unlikely to be modified by autophosphorylation. This Thr residue conforms to a casein kinase II (CKII) consensus site (http://scansite.mit.edu/) (Obenauer et al., 2003), and we attempted to demonstrate phosphorylation of Nek3 by CKII, but were unsuccessful (data not shown). Nek3 Thr475 is clearly phosphorylated by another kinase, but its identity remains unknown.
Nek3 mutation affects neuronal morphology and polarity
Nek kinases often contain PEST sequences, a motif associated with regulatory functions and/or protein stability. For example, the PEST domain of fission yeast NIMA homologue, fin1, controls its protein level (and thus function) during the cell cycle (Krien et al., 1998). Similarly, removal of the PEST domain in Aspergillus NIMA also resulted in a stable protein that promotes a lethal premature condensation of chromatin (O'Connell et al., 1994). Inspection of Nek3 sequence revealed two PEST domains in the Nek3 C-terminus (PEST scores, + 9.21 N-terminal, +14.64 C-terminal; threshold of significance, +5.0) (Fig. 3A) (https://emb1.bcc.univie.ac.at/toolbox/pestfind/) (Rogers et al., 1986). To assess the role of Nek3 in neurons and the importance of Nek3 PEST domains, Myc-tagged Nek3 truncation mutants lacking either one (Nek3-T1) or both (Nek3-T2) PEST domains were generated and expressed via lentivirus infection in DRG neurons (Fig. 3A). DRG neurons expressing Myc-tagged Nek3 (WT), Nek3-T1, or Nek3-T2 were examined for alterations in neuronal morphology by immunohistochemistry using anti-Myc antibody 5 days after infection. We found that both wild-type and mutant Nek3 were localized to the neuronal cytoplasm and processes; however, although neurons expressing wild-type Nek3 appeared normal, those expressing Nek3 truncation mutants were distinctly abnormal, with an aberrant neuronal soma and multiple neurite-like outgrowths (Fig. 3B). These results indicate that Nek3 influences neuronal morphology and that the PEST domains have important roles in regulating Nek3 protein activities.
Within the C-terminal PEST domain, Thr475 is highly conserved and was found to be phosphorylated in mouse brain (Ballif et al., 2004) and in cultured cells (Fig. 2). As protein activities are often regulated by post-translational modification, we explored whether phosphorylation of Thr475 regulates Nek3 function. To this end, we generated a Nek3 mutant in which Thr475 was changed to aspartic acid [Nek3(T475D)] to mimic Thr-P. DRG neurons expressing Nek3(T475D), similarly to those expressing wild-type Nek3, appeared morphologically normal; however, those expressing Nek3(T475A) showed cell body deformation and multiple processes similar to those expressing Nek3 PEST domain truncation mutants (Fig. 3B). These results suggest that phosphorylation of Thr475 within the PEST domain is a crucial step in regulating Nek3 activity, and supports the idea that the PEST domains participate in Nek3 processes that regulate neuronal morphology.
A key aspect of morphology in neurons is neuronal polarity. Because Nek3 truncation mutants and a Nek3 phosphorylation defective mutant induce abnormal neuronal morphology, we analyzed their effects on neuronal polarity in cultured hippocampal neurons, a well characterized system for studying neuronal polarity (Dotti et al., 1988). In normally polarized hippocampal neurons, only one axonal process is present, whereas multiple axons are evidence of disturbed neuronal polarity. We infected these neurons with lentiviruses expressing Nek3, Nek3(T475A), Nek3-T1 or Nek3-T2 and stained them 3 days later with neuron-specific β III tubulin (Tuj1) antibody to analyze their processes and morphology. Control neurons and those expressing wild-type Nek3 bore only a single long axon, characteristic of a normally polarized hippocampal neuron (Fig. 4A). By contrast, Nek3(T475A) and Nek3 truncation mutants induced the growth of multiple long processes. The identity of these processes as axons was verified by staining with the axonal marker Tau-1 (Fig. 4B), thus demonstrating that Nek3 mutation disturbed the polarity of these neurons.
To determine whether the effects of the Nek3 mutants on neuronal morphology and polarity required kinase activity, we generated additional mutants in which a crucial residue in the kinase domain, Asp143, was mutated to Ala to produce a catalytically dead Nek3 protein (Fig. 3A) (Tanaka and Nigg, 1999). This mutation (D143A) was engineered into the Nek3-T2 mutant to create kinase-dead Nek3-T2 (kdNek3-T2). This mutant was expressed along with Nek3-T2 in hippocampal neurons and analyzed as above. We found that neuronal abnormalities induced by Nek3-T2 were not present in neurons expressing kinase-dead Nek3-T2 (Fig. 4A,C), indicating that kinase activity is required for mutant Nek3 to induce abnormalities in neuronal morphology and polarity. The observation that these Nek3 mutant-mediated alterations in morphology and polarity require intact kinase activity further indicates that these effects are not caused by non-specific effects due to mis-folded mutant Nek3.
Nek3 mutants alter levels of acetylated α-tubulin
Post-translational modifications of microtubules, such as acetylation, tyrosination, phosphorylation and others, have crucial roles in establishing and maintaining neuronal polarity and morphology (Fukushima et al., 2009). For example, reducing the level of microtubule acetylation causes abnormalities in dendritic branching of cultured neurons (Ohkawa et al., 2008) and branch formation of cortical neurons in the developing mouse brain (Creppe et al., 2009). The effects of Nek kinases on microtubule dynamics during mitosis and ciliogenesis (O'Regan et al., 2007; Quarmby and Mahjoub, 2005) along with our observation that Nek3 mutants affect neuronal morphology and polarity prompted us to examine whether they might also alter neuronal microtubule acetylation. DRG neurons expressing wild-type Nek3, Nek3(T475A), kdNek3(T475A), Nek3-T2 and kdNek3-T2 were analyzed with an antibody specific for acetylated α-tubulin (Ac-tubulin), a marker of stable microtubules (Cambray-Deakin and Burgoyne, 1987; Schulze et al., 1987). We found that neurons expressing Nek3-T2 and Nek3(T475A) had greatly reduced levels of Ac-tubulin (Fig. 5A). However, Nek3 mutants lacking kinase activity had normal levels of Ac-tubulin, again indicating the necessity for kinase activity in these Nek3-mutant-mediated effects.
The cellular microtubular network is particularly well visualized in HeLa cells; therefore, to further analyze the effects of Nek3 on the Ac-tubulin network, we analyzed HeLa cells expressing wild-type Nek3, Nek3(T475A) or kdNek3(T475A). Immunofluorescent microscopy showed that HeLa cells expressing Nek3(T475A), similarly to DRG neurons, had a markedly diminished Ac-tubulin cytoplasmic network (Fig. 5B). Decreased levels of Ac-tubulin were also observed in cells expressing Nek3-T1 or Nek3-T2, but no changes were seen in cells expressing wild-type Nek3 or kinase-dead versions of the Nek3 mutants (Fig. 5B, data not shown). We also noted the absence of cilia in cells expressing Nek3(T475A) or Nek3-T2, but not kdNek3-T2 or wild-type Nek3 (supplementary material Fig. S1), suggesting that the Nek3 mutation is detrimental to cilia integrity. In contrast to levels of Ac-tubulin, the overall pattern of the microtubule network highlighted by anti-α-tubulin immunofluorescence was not affected by Nek3(T475A) (Fig. 5C), indicating that Nek3(T475A) affects the acetylation of α-tubulin, but not the α-tubulin network per se.
In addition to diminished levels of Ac-tubulin, HeLa cells expressing Nek3(T475A) or Nek3 truncation mutants assumed a highly elongated shape and became aligned with neighboring cells (Fig. 6A). Interestingly, the cells expressing Nek3(T475A) appeared even more deformed at 72 hours, with neurite-like projections (data not shown). Notably, phalloidin staining, which detects polymerized actin, demonstrated that these Nek3-mutant-mediated morphological alterations were not accompanied by changes in the actin filament network (data not shown). Furthermore, no changes in proliferation were induced by expression of wild-type or mutant Nek3 (data not shown), which is consistent with previous reports that Nek3 does not directly affect cell cycle progression (Tanaka and Nigg, 1999).
The alterations in α-tubulin acetylation could drive the changes in neuronal morphology and polarity induced by Nek3 mutation. Two major effectors of α-tubulin acetylation state are the deacetylases, HDAC6 (Hubbert et al., 2002; Matsuyama et al., 2002) and SirT2 (North et al., 2003). To explore whether they are involved in Nek3-mutant-mediated decreases in Ac-tubulin, we treated HeLa cells expressing Nek3(T475A) with nicotinamide, a sirtuin inhibitor, trichostatin A (TSA), a broad-spectrum HDAC inhibitor, or tubacin, an HDAC6-specific inhibitor (Haggarty et al., 2003). After 24 hours, immunofluorescence microscopy was used to monitor the levels of Ac-tubulin. We found that Ac-tubulin levels remained very low in cells expressing Nek3(T475A) treated with nicotinamide (a SirT2 inhibitor) (North et al., 2003), indicating that sirtuins are not involved in this deacetylation process [Fig. 6B; compare Ac-tubulin levels in cells denoted by arrow (transfected) and asterisk (non-transfected) in nicotinamide-treated cells]. This is very similar to DMSO-treated cells, where Nek3(T475A)-expressing cells showed greatly diminished Ac-tubulin. However, addition of either TSA or tubacin, restored microtubule acetylation in Nek3(T475A)-transfected cells to levels present in non-transfected cells [Fig. 6B; compare Ac-tubulin levels in cells denoted by one asterisk (transfected) and two asterisks (non-transfected) in tubacin- or TSA-treated cells]. These results clearly indicate that HDAC6 activity is required for loss of Ac-tubulin observed in cells expressing mutant Nek3.
Nek kinases were initially studied for their roles in regulating cilia, centrosomes and microtubule dynamics during cell cycle. The discovery that some Nek kinases are expressed in the adult brain raised the possibility that they participate in additional, non-mitotic roles in neurons. In Chlamydomonas, a NIMA-family kinase is crucial for cilia severing because it promotes the disassembly of microtubules as cells progress towards mitosis (Bradley et al., 2004; Finst et al., 2000; Finst et al., 1998; Fliegauf et al., 2007; Mahjoub et al., 2002; Mahjoub et al., 2004). Given the similarities between cilia and neuronal axons (both are microtubule-rich structures extending from the cell body) and the role of Nek kinases in microtubule regulation, we initiated studies to test whether mammalian Nek kinases regulate microtubules in neurons. We found that Nek3, which is not involved in mitosis, was expressed at high levels in developing mouse brain and in cultured post-mitotic neurons. The expression of Nek3 mutants lacking a key regulatory phosphorylation site (Thr475) caused changes in cellular morphology and polarity of both primary cultured central and peripheral neurons as well as epithelial cells. These cellular changes induced by Nek3 mutations are dependent on intact kinase activity, because Nek3 mutant versions with an additional mutation in the active kinase site itself did not produce these cellular changes. Interestingly, Nek3 mutants with an intact kinase domain also stimulated dramatic increases in α-tubulin deacetylation that were consistent with the effects of reduced microtubule stability, such as that induced by NIMA-family kinase during deflagellation in Chlamydomonas.
Cells overexpressing PEST truncated or phosphorylation-defective (Thr475) Nek3 mutants have dramatic alterations in their microtubule cytoskeleton and greatly reduced acetylated microtubules in the cytoplasm. These changes are strikingly similar to those induced by knocking down another cilia-related protein BBIP10 (Loktev et al., 2008). BBIP10 is a subunit of a stable protein complex, the BBSome, that participates in the function and vesicular trafficking of the primary cilium (Berbari et al., 2008; Nachury et al., 2007). Similarly to the NIMA-family kinases, components of the BBSome have been proposed to be important for microtubule and centrosome-based functions, because mutations in BBSome proteins underlie a variety of genetic disorders referred to as ciliopathies (Badano et al., 2006). The strikingly similar phenotypes induced by overexpressed mutant Nek3 and loss of BBIP10 raise the possibility that Nek3 and BBSIP10 could negatively regulate each other and influence the same pathways. Indeed, in addition to a marked reduction in cytoplasmic microtubule acetylation, NIH3T3 cells expressing the Nek3(T475A) mutant also show dramatic decreases in Ac-tubulin signal for cilia, which again is similar to cells lacking BBIP10 (Loktev et al., 2008). Other connections between Nek kinases and BBS proteins include centrosome splitting, which occurs in cells overexpressing Nek2 (Fry et al., 1998) or lacking BBS8 (Loktev et al., 2008). Better understanding of the interplay between Nek kinases and BBS proteins might lead to further insights into ciliary function and its role in disease.
Structure-function experiments revealed that the C-terminal PEST domains and, in particular, the phosphorylation of Thr475 within the PEST domain were critical modulators of Nek3 function. Nek3 mutants lacking these motifs or defective in Thr475 phosphorylation stimulated a signaling pathway that ultimately led to the activation of HDAC6 and α-tubulin deacetylation, a process that leads to reduced microtubule stability (Cambray-Deakin and Burgoyne, 1987; Matsuyama et al., 2002; Schulze et al., 1987; Tran et al., 2007). Interestingly, HDAC6 has also been implicated in cilia resorption or disassembly (Pugacheva et al., 2007) and α-tubulin deacetylation induced by depletion of BBIP10 (Loktev et al., 2008). How the PEST domain or the dephosphorylation of Thr475 regulates Nek3 function is an important question. The PEST domain was initially identified as a protein motif that promoted rapid protein degradation (Rogers et al., 1986); however, it is now known to also mediate protein-protein interactions (Pongubala et al., 1992; Zhang et al., 2007). Removal of the PEST domain from NIMA-family kinase in fission yeast resulted in elevated activity because of increased protein stability and higher protein levels (Krien et al., 1998; O'Connell et al., 1994). However, the truncation of the PEST domains in Nek3 led to decreased levels of protein, indicating that increased Nek3 levels are not responsible for the abnormalities caused by these mutants. The Nek3 PEST truncation mutants produced abnormalities that were identical to those induced by Nek3(T475A), suggesting that they both adopt a similar protein conformation. Based on the abnormalities induced by these mutants and on the predicted structural flexibility of the Nek3 PEST regions (http://imtech.res.in/raghava/apssp/), we propose that these domains modulate Nek3 activity through an allosteric effect that alters inter- or intramolecular interactions depending on the phosphorylation state of Thr475 within the PEST domain.
Acetylation of microtubules is critical for neuronal development and function. For example, increased microtubule acetylation via inhibition of Sirt2 results in decreased α-synuclein-mediated neurotoxicity (Outerio et al., 2007) and HDAC6 inhibitors reverse transport deficits in neurons expressing mutant Huntington disease protein (Dompierre et al., 2007). By contrast, decreased microtubule acetylation has been found to lead to abnormal neuronal shape and migration (Creppe et al., 2009; Ohkawa et al., 2008). Similarly, we found that neurons expressing mutant Nek3 also have dramatic abnormalities in neuronal polarity and morphology as well as reduced levels of microtubule acetylation. Although the mechanism responsible for these abnormalities is still unknown, increased microtubule acetylation is commonly associated with their stability, thus decreased microtubule stability could be responsible for the neuronal abnormalities caused by Nek3 mutants.
Nek3 can also regulate prolactin-stimulated actin-mediated cell mobility through Vav1 or Vav2 and Rac1 activation in breast cancer cells (Miller et al., 2007; Miller et al., 2005). However, we found that Nek3 mutants could still induce morphological alterations in neurons derived from Vav1-Vav2-Vav3 triple-deficient mice (unpublished data), seemingly in contradiction to these experiments. Furthermore, in contrast to results in breast cancer cells, Nek3-induced alterations in neurons occur through microtubule alterations, with no evidence for changes in actin polymerization. Finally, kdNek3 expression causes apoptosis in breast cancer cells (Miller et al., 2005), but Nek3-regulated increases in neuronal cell death were not observed in our experiments.
In summary, we discovered that the PEST domains and Thr475 are critical for Nek3 activity and that Nek3 affects neuronal morphology and polarity. These effects are presumably mediated by Nek3 mutant stimulation of HDAC6 α-tubulin deacetylation. The presence of Nek3 in axons and its ability to regulate microtubule acetylation suggests that it could be involved in processes linked to axonal projection and degeneration. The identification of the kinase(s) that regulate Nek3 through phosphorylation of Thr475, as well as characterization of the downstream substrates of Nek3, will help shed new light on its role in the nervous system.
Materials and Methods
Antibodies were routinely used at a 1:1000 dilution unless otherwise stated. Anti-Myc (9B11) (Cell Signaling 2276), anti-Nek3 (BD Transduction Laboratories 611716; dilution 1:250), anti-β-tubulin (Developmental Studies hybridoma; dilution 1:5000), anti-phosphothreonine (42H4) (Cell Signaling 9386), Nek3 Thr475-P (in house; dilution 1:1000 from 200 μg/ml), anti-Tau-1 (Chemicon MAB3420), anti-neuronal class III B-tubulin (TujI) (Covance PRB-435P), anti-α-tubulin (B5-1-2) (Sigma T5168), anti-acetylated α-tubulin (6-11B-1) (Sigma T6793). All secondary antibodies were from Jackson ImmunoResearch. Rabbit IgG-Cy3 (111-166-045), mouse IgG-Cy3 (115-165-003), rabbit IgG-HRP (111-035-144), mouse IgG-HRP (115-035-003).
Plasmid and lentivirus production
Full-length mouse Nek3 cDNA was obtained from Open Biosystems (Huntsville, AL) and a Myc tag was added to the N-terminus after the initiator ATG using PCR. Nek3 mutants were generated by PCR mutagenesis, cloned into FUIV (Araki et al., 2004), a lentiviral vector containing the human ubiquitin promoter along with EGFP cassette expressed via an IRES so that infected cells can be identified. Lentivirus was produced as described previously (Araki et al., 2004).
Primary neuronal cultures and lentivirus infection
Dissociated dorsal root ganglia (DRG) neurons from E13 mouse embryos were cultured in Neurobasal medium with B27 (Invitrogen) as described previously (Chen et al., 2008). The medium contained the anti-mitotic reagent 5-Fluoro-2′-deoxyuridine (FDU) (Sigma) and uridine (Sigma) at final concentration of 20 μM for the first 3 days to eliminate proliferating, non-neuronal cells. Hippocampal neurons were isolated from E15 embryos and cultured as above, except no FDU was added to these cultures. Neurons were infected with lentiviruses at the time of plating for hippocampal cultures and 3 days after plating for DRG cultures using ∼107 colony-forming units (CFU)/ml as previously described (Araki et al., 2004). More than 95% of the neurons were infected using this procedure.
Immunoblotting and immunofluorescence
Cells were lysed on ice for 30 minutes in NEB buffer (50 mM HEPES-KOH, pH 7.5, 5 mM MnCl2, 10 mM MgCl2, 2 mM EDTA, 100 mM NaCl, 5 mM KCl, 0.1% NP-40, 20 mM β-glycerophosphate, 20 mM sodium fluoride, 2 mM sodium orthovanadate and proteinase inhibitor from Roche) and lysates were clarified by centrifugation at 13,000 × g for 10 minutes at 4°C. Western blotting was performed following standard procedures. For immunofluorescence, neurons or HeLa cells grown on coverslips (Carolina; ww-63-3029) were fixed with 4% paraformaldehyde in PBS pH 7.4 for 20 minutes. The cells were permeabilized/blocked by incubation in 0.3% Triton X-100, 10% goat serum (Invitrogen) in PBS for 1 hour at room temperature. Cells were then incubated with the indicated primary antibody in 0.1% Triton X-100, 10% goat serum in PBS for 16 hours at 4°C. The signals were visualized by incubation with Cy3-conjugated fluorescent secondary antibodies in 0.1% Triton X-100, 10% goat serum in PBS for 1 hour at room temperature.
Nek3 Thr475-P antibody production
To produce antibodies specific for Nek3 phosphorylated at Thr475, rabbits were immunized with peptides corresponding to Nek3 residues 466-481 [Cys-FEPRLDEEDT(-P)DFEEDN, where T(-P) represents phosphothreonine] at Pacific Immunology. The anti-Nek3 antisera were sequentially affinity purified using SulfoLink Coupling Gel (Pierce 20401) conjugated with non-phospho-peptide (Cys-FEPRLDEEDTDFEEDN) and phospho-peptide according to the manufacturer's protocol.
Cell culture and transfection
HeLa cervical carcinoma cells were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum, 100 IU/ml penicillin, 100 μg/ml streptomycin and 20 mM L-glutamine at 37°C in a 5% CO2 atmosphere. Transfections were performed using FuGENE 6 (Roche Diagnostics) according to the manufacturer's instructions (1:3 of plasmid and FuGene reagent ratio was used). About 80-90% cells were transfected. Where indicated, the cells were treated 48 hours after transfection with nicotinamide (20 mM), tubacin (20 μM) or TSA (500 nM) for the indicated time.
In vitro phosphatase assay and peptide competition assay
Lysates of HeLa cells expressing Myc-Nek3 were prepared, clarified by centrifugation and incubated with 20 U calf intestinal phosphatase (CIP) for 1 hour in the recommended buffer (New England Biolabs). Reactions were stopped by addition of 6× Laemmli buffer and the lysates were analyzed by western blotting. For peptide competition assays, the Nek3 Thr475-P antibody was incubated with a fivefold molar excess of the respective peptide before use in western blotting.
Statistical analysis was carried out using two-way ANOVA using Graphpad Prism.
We thank Ta-Chiang Liu, Yo Sasaki, Jason Gustin and Bin Zhang for reading this manuscript. This work was supported by NIH Neuroscience Blueprint Center Core Grant P30 NS057105 to Washington University, the HOPE Center for Neurological Disorders, and National Institutes of Health Grants CA111966, NS040745, AG13730 (to J.M.), and MDA grant 3972 (J.M.). R.H.B. is supported by grants from the NIH (K08NS055980) and the Muscular Dystrophy Association and holds a Career Award for Medical Scientists from the Burroughs Wellcome Fund. Deposited in PMC for release after 12 months.