The NIMA-family kinase, Nek1 affects the stability of centrosomes and ciliogenesis
© White and Quarmby; licensee BioMed Central Ltd. 2008
Received: 05 December 2007
Accepted: 04 June 2008
Published: 04 June 2008
Skip to main content
© White and Quarmby; licensee BioMed Central Ltd. 2008
Received: 05 December 2007
Accepted: 04 June 2008
Published: 04 June 2008
Mutations in Nek1 (NIMA-Related Kinase 1) are causal in the murine models of polycystic kidney disease kat and kat2J. The Neks are known as cell cycle kinases, but recent work in protists has revealed that in addition to roles in the regulation of cell cycle progression, some Neks also regulate cilia. In most cells, cilia are disassembled prior to mitosis and are regenerated after cytokinesis. We propose that Neks participate in the coordination of ciliogenesis with cell cycle progression. Mammalian Nek1 is a candidate for this activity because renal cysts form in response to dysfunctional ciliary signalling.
Here we report that over-expression of full-length mNek1 inhibited ciliogenesis without disrupting centrosomes in the murine renal epithelial cell line IMCD3. In contrast, over-expression of the kinase domain with its associated basic region, but without the acidic domain, caused loss of centrosomes. As expected, these cells also failed to grow cilia. Both defective ciliogenesis in response to too much mNek1 and disassembly of centrosomes in response to expression of the kinase lacking the presumptive regulatory domain was abrogated by kinase-inactivating mutations or by removal of the coiled-coil domain. We observed that kinase-inactive, C-terminal truncations of mNek1 retaining the coiled-coil domain localized to the cilium, and we define a ciliary targeting region within the coiled-coil domain.
Based on our data, we propose that Nek1 plays a role in centrosome integrity, affecting both ciliogenesis and centrosome stability.
N imA-R e lated K inase 1 (Nek1) is a member of the kinase family sharing sequence identity to catalytic domain of the essential Aspergillus nidulans gene nimA (N ever I n M itosis A; ). Mutations in Nek1 are causal in the murine models of polycystic kidney disease kat and kat2J, resulting in a recessive pleiotropic phenotype that includes progressive cystic kidneys, male sterility, dwarfing, abnormal olfactory lobes, facial dysmorphism, choroid plexus cysts, hydrocephalus, uremia and anemia [2–4]. Although it has been suggested that Nek1 may play roles in cell-cycle control, meiosis, gametogenesis, and in the nervous system [1, 5] the cellular function of Nek1 has remained elusive.
The Neks are considered to be cell cycle kinases , but some members of the family play additional roles in ciliary function. Our studies in Chlamydomonas revealed both cell cycle and ciliary functions for two Neks: Fa2p plays a role at the G2/M transition of the cell cycle and is essential for the deflagellation response [7, 8], whereas Cnk2p is a ciliary protein that participates in the assessment of cell size prior to commitment to enter mitosis and plays a role in regulation of flagellar length . A survey of the Neks encoded by the genomes of various organisms supported our idea that the Nek family may have co-evolved with the alternating use of centrioles as basal bodies and as the foci of spindle poles . These results lead us to speculate that dual roles in ciliary and cell cycle regulation might be a hallmark of the Nek family.
We recently completed a comprehensive phylogenetic analysis of the Nek family across the eukaryotes and discovered that they are an ancient family . Our analyses suggest that six or seven different Neks were already present in the ciliated unicellular organism that gave rise to all of the eukaryotic lineages. The further observation that Neks representative of each extant clade localize to centrosomes and/or cilia supports our idea that the original expansion of the Neks was associated with dual roles in the regulation of cilia and the cell cycle, possibly co-ordinating ciliogenesis with cell cycle progression . Dual roles in cilia and cell cycle have not yet been demonstrated for any of the mammalian Neks.
There is some indication that the mammalian kinase, Nek1 may subserve both roles. Aberrant ciliary signaling and cell cycle regulation are both implicated in the etiology of renal cysts [12, 13] and mutations in Nek1 are causal for kidney cyst formation. Additionally, we have previously localized Nek1 to centrosomes . We thus set out to explore the cellular activities of Nek1 in order to further test our hypothesis about the dual roles of this kinase family.
Here, we report that transient over-expression of mNek1 in the ciliated renal epithelial cell line IMCD3 results in failure to assemble cilia. This effect depends upon the kinase activity of Nek1. In contrast, expression of dominant negative forms of Nek1 caused disruption of centrosomes, as indicated by loss of γ-tubulin and centrin foci. These disruptive effects require the coiled-coil domain of Nek1, in addition to kinase activity. We show that the coiled-coil domain of Nek1 is sufficient for nuclear, centrosomal and ciliary localization of the protein. We propose that Nek1 activity in differentiated kidney epithelial cells contributes to the dynamic regulation of the centrosome and the primary cilium, a sensory apparatus critical for the maintenance of tubules and in the prevention of renal cytogenesis.
Cells reabsorb their cilia prior to mitosis and regrow them after mitosis (reviewed in ). If the only cells expressing eGFP in the first few hours after transfection are those that have recently completed mitosis, then we predicted that the extent of ciliation of this population of cells will increase with time, corresponding to post-division ciliogenesis. As expected, ciliation of eGFP-expressing cells increases between 2 and 12 hours post-transfection (see Figure 2B).
In order to determine whether inhibition of ciliogenesis requires kinase activity, we generated the mutation K33R in the invariant lysine residue within the ATP-binding domain . The cognate mutation inactivates kinase function in Nek2 and Nek11 [24, 25]. We refer to constructs containing this mutation as presumptive kinase-dead. In contrast to cells expressing mNek1 with a lysine at position 33, cells expressing mNek1 containing the kinase-inactivating K33R mutation show no abnormal phenotype (data not shown).
Several kinases, such as FAK and Plk, are regulated in part through inhibition of kinase activity by the C-terminal, non-kinase domain of the protein [26, 27]. The C-terminus of Nek1 can bind the N-terminal kinase domain , and we sought to determine whether the acidic C-terminal domain of mNek1 influenced the ability of overexpressed eGFP-mNek1 to disrupt the primary cilium.
Expression of mNek1 truncations disrupt centrosomal γ-tubulin.
mNek1 aa residues expressed with GFP tag
Subcellular localization of GFP-truncated Nek1
Percent of transfected cells without γ-tubulin foci
8 ± 1
68 ± 3
64 ± 3
64 ± 4
In order to determine which of the predicted coiled-coils is necessary for the depletion of γ-tubulin foci, we expressed a panel of eGFP-tagged, truncated mNek1 constructs that sequentially remove the four coiled-coils of the coiled-coil domain. We observe that the constructs containing both coiled-coils 1 and 2 (residues 1–510), the first three coiled-coils (residues 1–599), and all four (residues 1–686) cause depletion of γ-tubulin foci, whereas the constructs containing either none (residues 1–352), or only the first coiled-coil (residues 1–407) fail to deplete γ-tubulin foci. These data suggest that the second coiled-coil of mNek1 (possibly in combination with the first coiled-coil) is required for the depletion of γ-tubulin foci, whereas the third and forth coiled-coils are not involved in this activity.
To evaluate whether transfected cells with compromised centrosomes continue to progress through the cell cycle, we co-stained with an antibody against PCNA (proliferating cell nuclear antigen), an accessory protein for polymerase δ often used as a marker for cells in S phase . At 24 hours post-transfection with myc-mNek1 (1–686), we observe that transfected cells lacking γ-tubulin foci are negative for PCNA staining, whereas 10% of myc-mNek1 (1–686, K33R) expressing cells are positive for PCNA (Figure 7B, C). We conclude that mNek1-induced centrosomal loss results in failure to progress through the cell cycle. This is consistent with a recent report that the integrity of the centrosomes is a requirement for progression through the cell cycle .
Expression of a panel of constructs that encode fusion proteins removing the four coiled-coils revealed that constructs containing coiled-coils 1 and 2 localize to the primary cilium (compare Figure 8C and 8D) whereas coiled-coils 3 and 4 are dispensable for localization to the primary cilium (compare Figure 8D–F). Furthermore, the coiled-coil domain alone is sufficient for localization to the primary cilium (Figure 8G).
The observation of Nek1 localization to the primary cilium is novel. We have not observed endogenous mNek1 in the primary cilium (data not shown; ), nor do we observe any of our other tagged-Nek1 constructs in the cilium. Based on these observations, we propose that Nek1 may normally cycle through the ciliary compartment but only kinase-dead Nek1 is retained in the cilia long enough (or in sufficient quantities) to be visualized.
Nek1 is a centrosome-associated protein and mice without functional Nek1 develop polycystic kidneys [2–4]. A link between centrosomal or ciliary dysfunction and the formation of kidney cysts is well-established and generally thought to be causal, although the molecular details have not yet been worked out [12, 13]. Kidney cysts are associated with aberrant cell proliferation, differentiation, apoptosis and loss of proper orientation of mitotic spindles along the axis of the renal tubules [33–37]. This is consistent with a growing body of evidence that many signals for proliferation and differentiation are received and processed by the primary cilium [38, 39]. Consequently, Nek1 provides a promising opportunity to test our prediction that some mammalian Nek proteins play dual roles in the regulation of cilia and cell cycle progression [10, 22].
We have shown here that transient over-expression of mNek1 inhibits ciliogenesis and overexpression of a Nek1 construct lacking its inhibitory domain causes disassembly of centrosomes. These data suggest that the normal function of the endogenous protein might involve regulation of centrosomal integrity and ciliogenesis. Unfortunately, we were unable to complement this work with studies of mNek1 knock-down because none of our mNek1 RNAi experiments yielded greater than 50% knock-down of mNek1 protein expression, and we did not detect any mutant phenotypes in these cells (data not shown). It is unlikely that this failure to generate knock-down cells is a consequence of some essential function of mNek1 because the kat and kat2J mice are viable as homozygotes. Nevertheless, the over-expression studies allow us to deduce some aspects of the cellular activities of Nek1.
We found that overexpression of full-length Nek1 and certain truncated forms of the protein inhibit ciliogenesis (see Figures 3 and 5). We think that the over-expression phenotype is a consequence of excessive mNek1 kinase activity. This conclusion is supported by the observation that over-expression of presumptive kinase-dead forms of mNek1 do not inhibit ciliogenesis. Over-expression of full-length kinase is expected to add substantial Nek1 kinase activity to the cell, even if the bulk of the protein is inhibited. An excess of Nek1 activity could inhibit ciliogenesis by blocking post-mitotic upregulation of ciliary proteins, by interfering with assembly or docking of pre-ciliary complexes at the centrosome, or by stimulating excessive rates of ciliary disassembly.
There are precedents for Nek activity to enhance rates of ciliary disassembly. Ciliary assembly, disassembly and length control are all mediated by a balance of the rates of anterograde and retrograde IFT (intraflagellar transport; ). The Chlamydomonas Nek Cnk2 is an axonemal protein that affects ciliary length by modulating the rate of disassembly and the delay before onset of assembly . Stable overexpression results in cilia that are shorter than normal, whereas RNAi-based knockdown results in long cilia. The shorter flagella were shown to be a consequence of accelerated rates of disassembly in the Chlamydomonas cells that were overexpressing Cnk2. Importantly, the longer cilia in cells with reduced Cnk2 were not a consequence of decreased rates of disassembly. Rather, when levels of Cnk2 are reduced there is a decrease in the lag time preceding ciliary assembly. Taken together, these data suggest that activated Cnk2 promotes ciliary disassembly both during reabsorption of cilia and in the initial stages of ciliogenesis, perhaps providing a ciliary assembly checkpoint which coordinates ciliogenesis with exit from mitosis. In the multi-ciliated protist Tetrahymena, overexpression of the Cnk2 orthologs also result in short cilia, suggesting conservation of function . Nek1 may directly regulate the IFT machinery because it directly interacts with Kif3a, a subunit in the heterotrimeric molecular motor kinesin-2 responsible for anterograde IFT .
Overexpression of the mitotic kinase Aurora A leads to the loss of cilia in mammalian cells . Although the disappearance of cilia was attributed to disassembly, inhibition of post-mitotic ciliogenesis was not ruled out. Nevertheless, loss of cilia in the context of excess Aurora A was shown to be mediated by activation of the tubulin deacetylase HDAC6. Pugacheva and colleagues showed that cilia are retained in cells with excess Aurora A, if these cells are treated with HDAC inhibitors. In contrast, we could not rescue Nek1-dependent cilia loss by incubation with the same HDAC inhibitors, although acetylation of microtubules was greatly increased in these cells (data not shown). We conclude that Nek1 is affecting ciliogenesis by a different pathway than Aurora A.
We next considered the possibility that Nek1 was affecting ciliogenesis as a consequence of disruptive effects on the centrosomes/basal body apparatus, the docking site for protein complexes en route to the cilium. We did not detect disruption of localization of centrosomal components in cells overexpressing Nek1 (Figure 3 and data not shown), nevertheless these cells failed to assemble cilia. It is possible that there may have been subtle centrosomal defects not detected because when truncation mutants lacking the presumptive inhibitory C-terminal domain were expressed, we detected rapid loss of γ-tubulin at the centrosomes, followed by loss of the centrin signal (Table 1 and Figures 5 and 6). Similar to the inhibition of ciliogenesis by the full-length protein, these effects depended upon the kinase activity of Nek1 because constructs carrying the kinase-disrupting mutation K33R did not disrupt centrosomal integrity and ciliogenesis occurred normally (Table 1 and Figure 6). We interpret these data to reflect a more severe augmentation of Nek1 kinase activity by the truncation mutants, which may be constitutively active. If the centrosome disassembly induced by these mutant forms of the protein reflects a pathological manifestation of a more subtle centrosome-remodelling activity of the endogenous protein, then these data would support the idea that the inhibition of ciliogenesis by over-expression of full-length mNek1 might be secondary to effects on the integrity of the centrosome.
We observed that presumptive kinase-dead Nek1 truncation mutants localized to the cilium and the nucleus (Figure 8). Kinase-active proteins were also found in the nucleus, but these cells did not have cilia (Table 1). We used presumptive kinase-dead constructs to identify a ciliary targeting domain in the first two coiled-coils of mNek1 (Figure 8). This same domain is also required for the centrosome/primary cilium disruption caused by expression of constructs that encode the wild-type kinase domain, suggesting that Nek1 might act directly on the centrosome and possibly also within the cilium. One caveat is that although the coiled-coil domain of mNek1 is sufficient for localization to the primary cilium, fusion proteins composed of the entire C-terminus of mNek1 (residues 258–1203) fail to localize to the primary cilia (not shown). This could indicate that the C-terminus prevents ciliary accumulation. Although we demonstrate the necessity of a functional ciliary targeting sequence for inhibition of ciliogenesis (Table 1 and Figure 8), we can not rule out the possibility that the effect was via nuclear activity because each of fusion proteins that disrupted centrosomes also localized to the nucleus.
Based on our data, we conclude that the cellular roles of endogenous Nek1 likely include regulation of ciliogenesis, possibly via effects on centrosome remodelling. We interpret the results of this study with caution because they are based exclusively on over-expression phenotypes of tissue culture cells. Nevertheless, the pattern of phenotypes caused by the array of mutant forms of Nek1 used in this work is informative. Our data strongly indicate that the phenotypes are related to Nek1 kinase activity. Furthermore, the phenotypes are specific to over-expression of mNek1. For example, over-expression of mNek8 does not affect ciliogenesis, centrosomes or cell cycle progression [43, 44]. Therefore, from the current work we conclude that an excess of Nek1 kinase activity inhibits assembly of cilia. Considering our data in the context of other studies (see Discussion) we further speculate that Nek1 participates in signalling between the primary cilium and the nucleus and effects the coordination of ciliogenesis with cell cycle progression.
The murine renal epithelial cell line derived from inner medullary collecting duct (IMCD3) was maintained in a 1:1 mixture of DMEM and Ham's F12 medium supplemented with 10% fetal bovine serum (DMEM-F12 (+), Invitrogen, Carlsbad, CA). Experiments were performed using cells below passage ten. Transient transfections were carried out with 4 μg of plasmid using Lipofectamine 2000 (Invitrogen) in OPTI-MEM reduced-serum media, according to the manufacturer's protocol. After 6 hours, cells were washed and transferred to pre-warmed DMEM-F12 (+). We set this as time zero post-transfection. For immunofluorescence, cells grown on coverslips were fixed with -20°C methanol for 10 minutes and rehydrated in PBS. The coverslips were incubated in primary antibody diluted in PBS for 1 hour, followed by two 15-minute PBS washes. The coverslips were then incubated in secondary antibody for 1 hour, followed by two 15-minute PBS washes. Nuclei were stained by a 10 minute, 1 μg/mL 4', 6-diamidino-2-phenylindole (DAPI) incubation at room-temperature. Stained coverslips were mounted in mowiol (Calbiochem, Dan Diego, CA). Immunofluorescence microscopy was performed using the Delta Vision system (Applied Precision, Isaaquah, WA) as described previously .
The primary antibodies used include mouse monoclonal anti-γ tubulin (clone GTU-88; 1:1000; Sigma-Aldrich, St. Louis, MO), rabbit polyclonal anti-γ tubulin (1:1000; Sigma-Aldrich), mouse monoclonal anti-acetylated tubulin (clone 6-11B-1; 1:2000; Sigma-Aldrich), mouse monoclonal anti-myc (Clone 9E10; 1:2000; Clontech, Palo Alto, CA), mouse monoclonal anti-α-tubulin-FITC conjugate (Clone DM1A; 1:1000; Sigma-Aldrich), mouse monoclonal anti-centrin (1:1000; from J. Salisbury, Mayo Clinic, Rochester, MN; ), mouse monoclonal anti-PCNA (1:200; Calbiochem), rabbit polyclonal anti-mouse cleaved PARP (1:100; Cell Signaling Technologies, Danvers, MA), rabbit polyclonal anti-mNek1 (1:100; from Dr. Y. Chen, The University of Texas Health Science Center, San Antonio), human autoimmune serum M4491 (which is reactive to multiple components of the proximal pericentrosomal tube, ; 1:3000; from J. Rattner, University of Calgary, Alberta), rabbit polyclonal anti-pericentrin (1:1000; Covance Research Products, Denver, PA) and rabbit polyclonal anti-BBS6 (1:200; from M. Leroux, Simon Fraser University, Burnaby, BC; ). The secondary antibodies used include Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:500; Molecular Probes, Eugene, OR), Alexa Fluor 594-conjugated goat anti-human IgG (1:1000; Molecular Probes), Alexa Fluor 594-conjugated goat anti-mouse IgG (1:2000; Molecular Probes) and Cy5-conjugated goat anti-rabbit IgG (1:500; Southern Biotech, Birmingham, AL).
mNek1 (accession number AY850065) was cloned into the SacI-SalI sites of pEGFP-C2 (Clontech), and truncations were made using standard molecular biology techniques. To generate myc-mNek1, the full-length mNek1 sequence was cloned into the Sal1 and Not1 sites of the pCMV-myc plasmid.
This work was funded by grants from the Canadian Institutes of Health Research (MOP-37861) and the Kidney Foundation of Canada to L.M.Q. We thank Michel Leroux and his laboratory members for use of their tissue culture facilities. We are grateful to Jeffery Salisbury, Jerome Rattner, and Michel Leroux for the centrin antibody, M4491 serum, and the BBS6 antibody, respectively. The intellectual engagement of members of the Quarmby lab is also appreciated. We are especially grateful to Jeremy Parker for his comments on the manuscript.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.