Goldberg–Shprintzen syndrome protein KIF1BP is a CITK interactor implicated in cytokinesis

Congenital microcephaly (CM) is a heterogeneous group of disorders characterized by reduced head circumference at birth by at least 3 s.d. below the mean (Passemard et al., 2013; Woods and Parker, 2013). CM can be the result of non-genetic conditions, such as viral infections and toxic exposure, or it can be generated by rare genetic disorders (Passemard et al., 2013). CM is also one of the leading features of several complex syndromes, comprising structural brain abnormalities, seizures, palsy, ataxia, short stature, skeletal abnormalities and cancer predisposition (Abuelo, 2007; Passemard et al., 2013; Woods and Parker, 2013). At the cellular level, CM is characterized by reduced expansion of neural progenitors, anticipated exit of the cell cycle and/or increased apoptosis, resulting in a reduced number of neuronal and glial cells.

Primary hereditary microcephaly (MCPH) is the simplest form of genetic CM, in which brain size reduction is accompanied by grossly normal brain architecture and mild to moderate intellectual disability (Mahmood et al., 2011; Passemard et al., 2013). Functional analysis of proteins encoded by MCPH genes has revealed that they are involved in different biological processes, such as centriole biogenesis, centrosome dynamics, centromere and kinetochore function, transmembrane or intracellular transport, Wnt signaling, autophagy, apical polarity complex localization, DNA replication and repair and cytokinesis (Boonsawat et al., 2019; Jayaraman et al., 2018; Naveed et al., 2018; Passemard et al., 2013; Zhou et al., 2020; Iegiani et al., 2021). Microtubules are crucial for many of these functions and their role is increasingly being recognized in phenotypes of cells characterized by reduced function of MCPH genes (Gai et al., 2016; Jin et al., 2017; Pagnamenta et al., 2019; Pilaz et al., 2016; Sgrò et al., 2016).

Citron kinase (CITK) is the main product of the MCPH17 (CIT) gene (Harding et al., 2016; Li et al., 2016). Loss of CITK affects the dynamics of microtubules (Gai et al., 2016; Sgrò et al., 2016) and leads to many cellular phenotypes associated with MCPH, including abnormal spindle orientation, DNA damage, apoptosis and cytokinesis failure (Bassi et al., 2011; Bianchi et al., 2017b; Gai et al., 2011; Sgrò et al., 2016). CITK is a well-known regulator of abscission, the end-stage of cytokinesis in which the microtubule bridge, which links daughter cells, is resolved (D'Avino, 2017). After cleavage furrow ingression, the two cells form a structure called midbody. This is divided in a central region called the midbody core, around which several proteins form a midbody ring. The flanking regions of this bulge are called midbody arms (Hu et al., 2012). CITK localizes at the midbody ring and is important to maintain midbody architecture and stability through multiple mechanisms (Bianchi et al., 2017a; Capalbo et al., 2019; D'Avino, 2017; Dema et al., 2018). In particular, CITK contributes to midbody maturation by bridging chromosomal passenger complex (CPC) proteins, such as Aurora B (AURKB) and INCENP, with midbody ring kinesins KIF14 and KIF23 (McKenzie et al., 2016). This complex interaction network is involved in midbody organization and abscission.

Goldberg–Shprintzen disease (GOSHS, OMIM 609460) is a rare microcephaly syndrome characterized by intellectual disability, dysmorphic facial features and peripheral neuropathy, as well as Hirschsprung disease (Brooks et al., 2005; Dafsari et al., 2015; Drévillon et al., 2013; Valence et al., 2013). Recessive mutations in the gene encoding KIF1BP (also known as KIFBP) are associated with GOSHS (Brooks et al., 2005). Consistently, KIF1BP knockout mice show a small and flat head, among other neurological features (Hirst et al., 2017). KIF1BP has been implicated in the organization of neuronal microtubules, as well as axonal growth and maintenance (Drerup et al., 2016; Kevenaar et al., 2016; Lyons et al., 2008). Moreover, it has a direct role in mitochondria distribution (Wozniak et al., 2005) and mitochondrial biogenesis in mouse embryonic stem cells (Donato et al., 2017). During mitosis, KIF1BP regulates the congression and alignment of chromosomes by buffering KIF15 and KIF18 kinesins (Brouwers et al., 2017; Malaby et al., 2019). KIF1BP inhibits the activity of a subset of kinesins involved in axonal transport and mitotic functions (Kevenaar et al., 2016; Malaby et al., 2019) by sequestering the tubulin-binding surface of the kinesin motor domain (Atherton et al., 2020). Nevertheless, it is not understood how KIF1BP loss could generate microcephaly.

In this report, we show that KIF1BP is a prominent component of the CITK interactome, which we characterized by means of affinity chromatography and mass spectrometry (MS). Functional analysis of KIF1BP showed that it shares with CITK a role in midbody maturation and abscission during cytokinesis. The implications of these findings for GOSHS are discussed.

To identify new CITK partners, we resorted to an interactomic approach. We overexpressed in 293-T cells the Cherry protein (control) and the following Cherry-tagged versions of CITK: wild type (CITK), K126A inactive mutant (CITKD) (Di Cunto et al., 1998) and CITN, a natural product of the CIT gene completely lacking the kinase domain (Camera et al., 2003). The reason for using kinase-inactive or kinase-lacking proteins was that they may be more efficient at trapping CITK interactors, especially those that may be direct substrates of its kinase activity (Macías-Silva et al., 1996). Overexpressed proteins were purified by affinity chromatography. SDS-PAGE of immune-complexes (Fig. S1) revealed the presence of specific interactors in samples obtained with CITKD expression. MS-based proteomic analysis of the eluates identified 53 different proteins significantly associated with at least one of the CITK baits compared to the negative control (Table S1). Half of the co-precipitating proteins were common to the three baits and a significant fraction of them was shared with the interactors identified by previous studies (Table S1). These proteins were enriched for cytoskeletal regulators involved in mitosis and cytokinesis (Fig. S2). They included several molecules previously connected to CITK function, such as KIF14 and KIF23 (Bassi et al., 2013), ANLN (Gai et al., 2011), RACGAP1 (Chalamalasetty et al., 2006) and PPP1R12A (MYPT1) (Capalbo et al., 2019; Kawano et al., 1999). These data strongly validate our biochemical approach. Interestingly, the majority of the identified proteins was significantly enriched or exclusively detected in the immunoprecipitate of CITKD. One of these proteins was the kinesin-binding protein KIF1BP, previously implicated in GOSHS microcephaly syndrome (Brooks et al., 2005).

We evaluated in more detail the interaction of CITK with KIF1BP, as the latter protein may regulate different CITK-interacting kinesins, such as KIF14 and KIF23 (Kevenaar et al., 2016; D'Avino, 2017). Moreover, KIF1BP is involved in the microcephaly syndrome GOSHS (Brooks et al., 2005; Valence et al., 2013), suggesting that a functional interaction with CITK could be relevant for the phenotypes common to GOSHS and MCPH17.

We confirmed the presence of endogenous KIF1BP in immunocomplexes of overexpressed Cherry-CITK and Cherry-CITKD by performing western blot analysis with KIF1BP-specific antibodies (Fig. 1A). Consistent with MS results, the amount of KIF1BP co-immunoprecipitated with Cherry-CITKD was higher than with Cherry-CITK (Fig. 1A). The binding was not dependent on using the Cherry epitope, as similar results were obtained by immunoprecipitating Myc-tagged wild type and kinase-dead CITK (Fig. 1B). Moreover, we were able to detect co-immunoprecipitation of KIF1BP using only the citron coiled-coil domain (CITcc) (Fig. 1B), which is involved in the binding of most CITK interactors so far identified (D'Avino, 2017). Reciprocal co-immunoprecipitation was obtained in 293 cells co-transfected with Myc-tagged CITK and FLAG-tagged KIF1BP by performing immunoprecipitation with anti-FLAG antibodies (Fig. 1C). Even the reciprocal co-immunoprecipitation assay revealed increased association of KIF1BP with CITKD compared to the kinase-active protein (Fig. 1C). Importantly, we detected co-immunoprecipitation of endogenous CITK and KIF1BP from lysates of HeLa cells growing in asynchronous cultures or synchronized in different stages of mitosis (Fig. 1D). In particular, we analyzed cells in the interval between 30 min and 180 min after release of mitotic block, in which cells pass from metaphase to telophase (Fig. S3). As expected for CITK (Liu et al., 2003), the expression of KIF1BP peaks at 90 min from release, i.e. before mitotic exit (Fig. 1D). The association between KIF1BP and CITK was easily detectable both at early and late mitotic times. Immunoprecipitates of synchronized cultures also contained KIF23 and KIF14, although the association was not easily detectable in asynchronous cells (Fig. 1D). The amount of the two kinesins in immunoprecipitates peaked at 30 min and decreased at later times, although the association was still detectable at 180 min (Fig. 1D). Considering the stronger association between catalytically inactive CITK and KIF1BP (Fig. 1), we investigated whether the latter could be a substrate for the isolated CITK kinase domain. In non-radioactive in vitro kinase assays, purified KIF1BP stimulated ATP consumption similarly to MYPT1 (Fig. S4A), a known substrate of myotonic dystrophy kinase family members, which include CITK (Zhao and Manser, 2015). In vitro phosphorylation of KIF1BP by CITK was also confirmed by immunoblotting with anti-phospho threonine antibodies (Fig. S4B). These data demonstrate that KIF1BP and CITK can be part of a physical complex under physiological conditions, and show that their association can be modulated by CITK catalytic activity.

KIF1BP is required for the alignment of chromosomes at the metaphase plate and for the assembly of stable kinetochore fibers of correct length (Brouwers et al., 2017; Malaby et al., 2019). Considering the KIF1BP-CITK interaction (Fig. 1), we investigated whether KIF1BP may also share a role with CITK in cytokinesis.

Immunofluorescence analysis of wild-type HeLa cells revealed that, in late cytokinesis, KIF1BP associates with the midbody (Fig. 2A), where CITK is known to reach its highest concentration (Bassi et al., 2013). This midbody staining specifically disappeared in cells treated with siRNAs specific for KIF1BP (Fig. 2A,B). The association of KIF1BP with the midbody and colocalization with CITK at the midbody ring were confirmed by overexpressing an N-terminal GFP-fused version of KIF1BP in HeLa cells (Fig. 2C). Interestingly, in cells overexpressing KIF1BP, the levels of CITK at the midbody ring were consistently reduced (Fig. 2C,D). Conversely, KIF1BP depletion did not inhibit CITK localization at the midbody but, on the contrary, tended to increase its levels (Fig. 2E,F). On the other hand, CITK knockdown consistently increased the localization of KIF1BP at the midbody and in particular its association with the midbody ring (Fig. 2G,H). These data show that KIF1BP is specifically localized to the midbody and that CITK and KIF1BP may antagonistically regulate their association with the midbody ring.

Considering the specific midbody localization of KIF1BP, we next investigated whether it may play a functional role in HeLa cell abscission. Loss of KIF1BP leads to a significant increase of binucleated cells (Fig. 3A,B), which is suggestive of cytokinesis failure. Dynamic microscopy of mitotic divisions revealed that KIF1BP-depleted cells display significant delay (data not shown), as reported previously (Brouwers et al., 2017; Malaby et al., 2019). Cells that are capable of progressing into mitosis display increased frequency of cytokinesis failure (see Movies 1, 2). In particular, the cytokinesis failure of KIF1BP knockdown cells is very similar to the phenotype described after CITK loss of function (Bassi et al., 2013; Gai et al., 2011): the contractile ring ingresses normally and midbody constriction progresses to an advanced stage, but cytokinesis fails afterwards (Fig. 3C,D), with reopening of the cellular bridge and daughters fusing into a binucleated cell.

As CITK influences midbody architecture and stability by modulating microtubule dynamics (Sgrò et al., 2016), and KIF1BP is implicated in the organization of neuronal microtubules (Drerup et al., 2016), we wondered whether KIF1BP may have a similar function in cytokinesis. Therefore, we measured the ratio of acetylated versus tyrosinated tubulin at the midbody, as stable microtubules are characterized by higher levels of acetylated tubulin (Witte et al., 2008). Compared to control cells, KIF1BP knockdown cells showed a general significant reduction of acetylated/tyrosinated tubulin ratio (35±5%; mean±s.e.m.), which was even more evident at the midbody bridge (Fig. 3E,F). These data show that KIF1BP shares with CITK a functional role in midbody stabilization and abscission at the end of cytokinesis.

KIF1BP binds to and inactivates a specific subset of kinesins, including KIF23 and KIF14. These proteins are involved in cytokinesis and are known to cooperate with CITK to ensure the correct localization of the CPC at the midbody (McKenzie et al., 2016; Sgrò et al., 2016). In line with this finding, we wondered whether KIF1BP knockdown may alter the localization and/or the local levels of all these proteins.

In control cells, AURKB is enriched in the midbody arms (Fig. 4A). After KIF1BP knockdown, AURKB showed strongly reduced levels at the midbody and failed to form the two distinct bands in ∼40% of divisions (Fig. 4A,B). Similarly, the midbody signal of INCENP was reduced in KIF1BP knockdown telophases (Fig. 4C,D). In addition, INCENP localized mostly in the midbody core instead of its normal localization at the midbody arms (Fig. 4C,D). The focusing at the midbody of KIF23 and KIF14 was not affected by KIF1BP depletion (Fig. 4E,G). However, levels of the two proteins were consistently altered, with an increase of KIF23 (Fig. 4E,F) and a decrease of KIF14 (Fig. 4G,H). Altogether, these results indicate that KIF1BP plays a functional role in midbody maturation, regulating microtubule stability, CPC components and the localization of kinesins.

KIF1BP loss leads to GOSHS syndrome, which is characterized by dysmorphic facial features and multiple alterations of the central and peripheral nervous system (Brooks et al., 2005; Dafsari et al., 2015; Drévillon et al., 2013; Valence et al., 2013). In this report, we provided new molecular insight regarding KIF1BP function. We found that KIF1BP can be physically associated with CITK, not only upon overexpression but also at the endogenous levels reached in mitotic cells. Previous studies have shown that KIF1BP plays a mitotic role before anaphase, promoting the congression and alignment of chromosomes (Brouwers et al., 2017; Malaby et al., 2019). Here, we demonstrated that KIF1BP is also important to complete cytokinesis. In particular, CITK (Gai et al., 2011) or KIF1BP RNAi (Fig. 3C; Movie 2) produce similar cytokinesis phenotypes, with normal cleavage furrow ingression, apparently normal midbody formation but frequent midbody bridge re-opening (Fig. 3C; Movie 2). Consistent with this phenotype, we found that KIF1BP is enriched at the midbody and colocalizes with CITK at the midbody ring, suggesting that the two proteins may functionally interact to control abscission.

Previous studies clarified that CITK associates with microtubules (Bassi et al., 2011; Eda et al., 2001), promotes the stabilization of microtubules (Sgrò et al., 2016) and interacts with KIF14 and KIF23, as well as with microtubule bundling protein PRC1 (Bassi et al., 2013). The finding that immunocomplexes formed by inactive CITK contain increased levels of KIF1BP and three tubulin subunits (Table S1) suggests that CITK activity may control microtubule dynamics and may also impinge on KIF1BP localization. Accordingly, CITK depletion by RNAi leads to increased levels of KIF1BP at the midbody (Fig. 2G,H). On the other hand, KIF1BP overexpression reduces CITK midbody levels, whereas KIF1BP depletion does not increase them significantly. We found that KIF1BP is required to concentrate CPC proteins AURKB and INCENP in the midbody arms, as well as to exclude them from the midbody core (Fig. 4). This result is very interesting considering the previous identification of a cross-regulatory loop between AURKB and CITK. The latter is crucial for the orderly arrangement of many proteins at the midbody, including kinesins KIF23 and KIF14 (Bassi et al., 2013; McKenzie et al., 2016), which are also KIF1BP interactors (Kevenaar et al., 2016). KIF1BP loss increased KIF23 levels at the midbody but at same time reduced KIF14 levels. These results, together with the co-immunoprecipitation of KIF1BP and CITK in mitotic cells (Fig. 1D), suggest that the two proteins form a physical complex that regulates midbody maturation through proper recruitment of CPC and by modulating KIF14 and KIF23, especially in early cytokinesis. This scenario would be consistent with the previously reported feedback loop between CITK, CPC and KIF14 (McKenzie et al., 2016). Moreover, the findings that KIF1BP overexpression reduces CITK, and that CITK depletion increases KIF1BP midbody levels, suggest a dynamic cross-regulatory interaction between the two proteins, which may also impinge on KIF23 localization. Deeper elucidation of the molecular details underlying the dynamic interactions between all of these proteins in the final stages of cytokinesis will be an interesting subject for future studies. In particular, a better clarification of all the phosphorylation events controlled by CITK will be crucial because kinase-inactivating mutations are sufficient to produce severe microcephaly in humans (Harding et al., 2016; Li et al., 2016). In this regard, it will be particularly interesting to map and validate KIF1BP residues that can be phosphorylated in vitro by the isolated CITK catalytic domain (Fig. S3).

Although the best understood function of CITK is abscission control (Bassi et al., 2011; Dema et al., 2018; Gai et al., 2011), it is not excluded that the CITK-KIF1BP interplay may regulate other microcephaly relevant phenomena. Analysis of cellular and animal models of MCPH17 revealed that CITK is required for normal spindle positioning, which may impinge on asymmetric division (Gai et al., 2016), as well as for preventing the accumulation of DNA double-strand breaks, which is critical for p53-dependent apoptosis (Bianchi et al., 2017b; Pallavicini et al., 2020). KIF1BP could also cooperate with CITK in these cases. In this regard, it is worth noting that KIF1BP knockdown leads to a reduction of the ratio between acetylated and tyrosinated tubulin, not only at the midbody but also in the cytoplasm. Finally, besides contributing to a better understanding of microcephaly, a deeper elucidation of the KIF1BP-CITK interplay could also provide insight for developing new anti-tumor strategies, considering the anti-neoplastic effects determined by CITK inactivation in epithelial tumors (Fu et al., 2011; Meng et al., 2019) and medulloblastoma (Pallavicini et al., 2018, 2019).

Unmodified HeLa cells were originally obtained from American Type Culture Collection, and a new batch was thawed after five passages. All cells were routinely screened for mycoplasma contamination. HeLa cells were cultured in RPMI medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/ streptomycin. The HeLa cell line expressing α-tubulin-GFP was maintained in Dulbecco's modified Eagle medium (DMEM)-GlutaMAX (Invitrogen) supplemented with 10% FBS, 100 U ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin, 200 μg ml⁻¹ Geneticin (Sigma-Aldrich, St Louis, MO, USA) and 0.5 μg ml⁻¹ puromycin. All cells were cultured in a humidified 5% CO₂ incubator at 37°C. For synchronization, asynchronous cultures were supplemented with 25 μg/ml aphidicolin (Sigma-Aldrich) and maintained under these conditions for 24 h, and then cultured for a further 16 h in fresh complete medium in the presence of 50 ng/ml nocodazole (Sigma-Aldrich) to block cells at prometaphase. Finally, cells were washed three times with fresh medium and were allowed to progress through mitosis/cytokinesis for the indicated times. Mitotic progression was controlled by immunofluorescence microscopy for α-tubulin and DAPI.

HEK293T cells were cultured in DMEM supplemented with 10% fetal calf serum and 1% penicillin/streptomycin (Life Technologies, Thermo Fisher Scientific, Waltham, MA, USA). For both cell lines, the number of in vitro passages from thawing of the original aliquots to experiments was between five and eight. Cells were routinely analyzed for morphological features and tested for mycoplasma contamination with the following oligonucleotide sequences: MycO1, 5′-ACTCCTACGGGAGGCAGCAGTA-3′; MycO2, 5′-TGCACCATCTGTCACTCTGTTAACCTC-3′.

Previously published CITK double-stranded RNAs were used (Gai et al., 2011). For knockdown of KIF1BP, KIA1269 ONTARGETplus Dharmacon SMART pools were used. D-001810-10 non-targeting pool was used as a negative control (Dharmacon, Lafayette, CO, USA). HeLa cells plated on six-well plates were transfected using 6.25 μl of the required siRNA (20 μM) together with 1.5 μl Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instructions. HEK293T cells were transfected through use of CaCl₂. CITK Cherry-tagged and Myc-tagged constructs used in this article were previously published (Gai et al., 2011). CITcc construct was obtained from CIT N plasmid (Gai et al., 2011) by deleting the cDNA portion encoding the C-terminal part of the protein after amino acid 1302. FLAG-tagged and GFP-tagged KIF1BP constructs were kindly provided by the Pagano Lab (Donato et al., 2017).

For all immunoprecipitations, cells were extracted with lysis buffer containing 150 mM NaCl, 1 mM MgCl₂, 50 mM Tris (pH 7), 1% NP40, 5% glycerol protease inhibitors (Roche, Basel, Switzerland) and 1 mM phenylmethylsulfonyl fluoride. Antibodies, according to the manufacturer's protocols, and 7 μl of Dynabeads protein G (GE Healthcare Life Science, Little Chalfont, UK) were added to 1 mg of cleared lysates, and incubated for 2 h at 4°C. RFP-TRAP affinity resin from Chromotek (rtma-10 Chromotek, Planegg, Germany) was used according to the manufacturer's instructions. Pellets were washed four times with lysis buffer and analyzed by SDS-PAGE. For immunoblots and immunoprecipitates, equal amounts of proteins from total cell lysates were resolved by reducing SDS-PAGE, and transferred to nitrocellulose filters that were then incubated with the indicated antibodies.

Cleared lysates (1 mg) were incubated with RFP-TRAP affinity resin for every condition; eluted proteins were stacked in the top of an SDS-PAGE gel (4-12% NuPAGE, Life Technologies), stained with Coomassie Blue R-250 (Bio-Rad) before in-gel digestion using modified trypsin (Promega, sequencing grade) as described previously (Salvetti et al., 2016). Resulting peptides were analyzed by online nanoliquid chromatography coupled to tandem MS (UltiMate 3000 and LTQ-Orbitrap Velos Pro, Thermo Scientific). Peptides were sampled on a 300 µm×5 mm PepMap C18 precolumn, and separated on a 75 µm×250 mm PepMap C18 column (Thermo Scientific) using a 120-min gradient. MS and MS/MS data were acquired using Xcalibur (Thermo Scientific). Peptides and proteins were identified using Mascot (version 2.6) through concomitant searches against the UniProt database (Homo sapiens taxonomy), the classical contaminant database (homemade) and the corresponding reversed databases. Proline software (Bouyssié et al., 2020) was used to filter the results: conservation of rank 1 peptides, peptide score ≥25, peptide length ≥6 amino acids, peptide-spectrum-match identification false discovery rate <1% as calculated on scores by employing the reverse database strategy, and a minimum of one specific peptide per identified protein group. Proline was then used to perform a compilation, grouping and spectral counting-based comparison of the protein groups identified in the different samples. Proteins from the contaminant database and additional keratins were discarded from the final list of identified proteins. To be considered as a potential CITK interactor, a protein must be identified in CITK, CITN or CITKD eluates with a minimum of three specific spectral counts and not in negative control eluate, or enriched at least ten times in CITK, CITN or CITKD eluates compared to negative control eluate.

Cells were fixed in methanol at −20°C for 10 min, saturated in 5% bovine serum albumin (BSA) in PBS for 30 min and incubated with a primary antibody for 1 h at room temperature. For GFP overexpression, cells were fixed with PBS with 4% paraformaldehyde for 10 min, permeabilized with PBS 0,1% Triton X-100 for 10 min, and finally saturated in 5% BSA for 30 min and incubated with a primary CITK antibody for 1 h at room temperature. Primary antibodies were detected with anti-rabbit IgG Alexa Fluor 488 or 568 (Molecular Probes, Invitrogen) and anti-mouse IgG Alexa Fluor 488 or 568 (Molecular Probes, Invitrogen) used at a 1:1000 dilution for 30 min. Cells were counterstained with 0.5 μg/ml DAPI for 10 min and washed with PBS. To quantify midbody fluorescence signals, we used integrated density from Fiji, which is the sum of the values of the pixels in the image or selection, subtracting the cytoplasmic background; we then calculated the mean of control midbodies and used that value as reference for all midbodies of the same experiment (controls and treated). Background subtraction was not feasible for KIF1BP immunofluorescence because cytoplasmic signal was too high and heterogeneous compared to the specific midbody signal; in this case we used the absolute level of the integrated density. All quantifications were repeated at least three times.

We incubated, for 30 min at 30°C in kinase buffer [50 mM HEPES (pH 7.4), 10 mM MgCl₂ 5 mM MnCl₂ and 1 mM DTT] and 100 nM of ATP, 150 ng of recombinant CITK (Abcam) and different amounts of substrate, as indicated in Fig S4. For ATP-consumption assays, the samples were processed using the ADP-Glo Kinase Assay (V6930, Promega), according to the manufacturer's specifications. For immunodetection of phosphorylated serine and threonine, samples were loaded in an acrylamide gel and processed as described above.

Imaging was performed using a Leica TCS SP5 confocal system equipped with a 405 nm diode, an argon ion, a 561 nm DPSS and a HeNe 633 nm laser. Fixed cells were imaged using a HCX PL APO 63×/1.4 NA oil immersion objective. For live imaging, time lapses were recorded overnight with an interval of 5 min using a 40× PlanApo N.A. 1.4 oil immersion objective on the cells kept in the microscope incubator at 37°C and 5% CO₂. Super-resolution images were obtained using a Leica SP8 confocal system with HyVolution 2 equipped with an argon ion, a 561 nm DPSS and a HeNe 633 nm laser. Fixed cells were imaged using a HCX PL APO 63×/1.4 NA oil immersion objective. A series of x-y-z images (typically 0.04×0.04×0.106 μm³ voxel size) were collected.

The following antibodies were used: mouse monoclonal anti-citron (611377 Transduction Laboratories, BD Biosciences, San Jose, CA, USA); rabbit polyclonal anti-KIF1BP (NBP1-84143, Novus Biologicals, CO, USA); mouse monoclonal anti-KIF1BP (H12) for endogenous immune precipitation (sc-390449, Santa Cruz Biotechnology, Dallas, TX, USA); rabbit polyclonal anti-β-actin (A2066, Sigma-Aldrich); mouse anti-FLAG (F3165, Sigma-Aldrich); rabbit monoclonal anti-KIF14 (ab71155, Abcam); rabbit monoclonal anti-KIF23 (ab174304, Abcam); mouse anti-Myc TAG (05-419, Sigma-Aldrich); RFP-Trap Magnetic Agarose (rtma-10 Chromotek, Planegg, Germany) for immunoprecipitation of Cherry-tagged proteins and mouse monoclonal homemade mCherry for blots; mouse monoclonal anti-AURKB (A78720, Transduction Laboratories, BD Biosciences, San Jose, CA, USA); rabbit polyclonal anti-INCENP (2807, Cell Signaling Technology, Danvers, MA, USA); rabbit polyclonal anti-tyrosinated tubulin (ABT171, Sigma-Aldrich); mouse anti-acetylated tubulin (T6199, Sigma-Aldrich); mouse anti-tubulin (T5168, Sigma-Aldrich); polyclonal rabbit anti-phosphothreonine (71-8200, Zymed, Invitrogen); and polyclonal rabbit phospho-(Ser/Thr) Phe antibody (9631, Cell Signaling Technology). The following human recombinant proteins were used: CITK human recombinant protein (ab179954, Abcam); KIF1BP human recombinant protein (ab161903, Abcam); and MYPT1 (654-880) (12-457, Merck, Sigma-Aldrich).

Statistical analyses were performed using Microsoft Office Excel or GraphPad. The mean values shown represent the average of at least three independent experiments and data are mean±s.e.m.

We thank Valerio Donato and Michele Pagano (New York University, NY, USA) for kindly supplying plasmids.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.250902