Fe-S cluster coordination of the chromokinesin KIF4A alters its subcellular localization during mitosis

Recent studies have attempted to expand the list of genes contributing to chromosome instability (CIN) in yeast. These studies resulted in the identification of many yeast CIN genes that can be organized into a number of functional groups. One of the associated biological pathways is the biogenesis of iron-sulfur (Fe-S) proteins (Ben-Aroya et al., 2010, 2008; Lev et al., 2013; Stirling et al., 2011).

Fe-S clusters are small inorganic cofactors found in hundreds of proteins that are required in all kingdoms of life from bacteria to humans (Lill and Mühlenhoff, 2008; Netz et al., 2014). These clusters exist in multiple chemical forms, and may readily accept or donate electrons, for example, in redox reactions catalyzed by certain proteins. Fe-S clusters also serve in enzyme catalysis, regulation of gene expression and stabilization of protein structures (Lill, 2009; Lill and Mühlenhoff, 2008; Netz et al., 2014). The components involved in the biogenesis of Fe-S proteins are conserved in evolution from bacteria to humans, and many insights into the process of eukaryotic Fe-S protein biogenesis arose from studies in the yeast Saccharomyces cerevisiae and human cell culture. The biosynthesis of cellular Fe-S clusters is initiated in mitochondria, which harbor the iron-sulfur cluster (ISC) assembly machinery that was inherited from bacteria during evolution (Netz et al., 2014; Paul and Lill, 2014). The core components of the ISC system include the cysteine desulfurase complex Nfs1–Isd11–Acp1, the scaffold Isu1, the ferredoxin Yah1 and the Hsp70 chaperone Ssq1, and are required for maturation of both [2Fe-2S] and [4Fe-4S] proteins. Additional late-acting ISC proteins are specific for the assembly of mitochondrial [4Fe-4S] proteins. Mitochondria further contribute to the maturation of cytosolic and nuclear Fe-S proteins, as they export a still unknown sulfur-containing component to the cytosol. This compound is used by the cytosolic iron-sulfur protein assembly (CIA) machinery to generate a transiently bound [4Fe-4S] cluster on the Cfd1–Nbp35 scaffold complex requiring the electron transfer chain NADPH–Tah18–Dre2 (human NDOR1–CIAPIN1) (early-acting CIA components) (Netz et al., 2014; Paul and Lill, 2014).

Studies in both yeast and human cells have revealed that the late-acting CIA proteins Mms19 (also known as Met18; human MMS19), Cia1 (human CIAO1) and Cia2 (human CIA2B, also termed FAM96B and MIP18) comprise the so-called CIA targeting complex, that specifically transfers and inserts the [4Fe-4S] cluster into recipient apoproteins (Gari et al., 2012; Netz et al., 2014; Paul and Lill, 2014; Stehling et al., 2012). During this transfer reaction the CIA component Nar1 (human IOP1, also termed NARFL) connects the early and late-acting CIA factors through a yet unknown mode of action (Balk et al., 2005, 2004; Gari et al., 2012; Hausmann et al., 2005; Seki et al., 2013; Song and Lee, 2008; Stehling et al., 2013, 2012).

Many Fe-S proteins have been identified only recently. The most rapidly growing class of Fe-S proteins are enzymes involved in nucleic acid metabolism, such as DNA repair proteins (Klinge et al., 2007), DNA helicases (Rudolf et al., 2006; Wu and Brosh, 2012), DNA polymerases (Netz et al., 2012) and primases (Klinge et al., 2007). The function of Mms19, Cia1 and Cia2 in the maturation of these crucial components of DNA metabolism explains the complex phenotypes associated with the downregulation of their genes in yeast and human cells, including sensitivity to genotoxic stress [e.g. methyl methanesulfonate (MMS) or UV radiation] (Lauder et al., 1996; Prakash and Prakash, 1977), the presence of extended telomeres, abnormal mitotic spindles (Askree et al., 2004) and chromosome mis-segregation (Ben-Aroya et al., 2008; Ito et al., 2010; Yuen et al., 2007).

The importance of the CIA targeting complex in delivering the Fe-S cluster to recipient apoproteins is evident from the impressive list of Fe-S proteins that have been found to associate with its human constituents MMS19, CIAO1 and CIA2B (Stehling et al., 2013, 2012). These results make it likely that additional Fe-S proteins are required for the maintenance of nuclear genome integrity and other cellular functions. For most of these Fe-S candidates, it is currently unclear whether their CIA interactions are physiologically meaningful, hence detailed functional characterization is required.

Here, we show that part of the CIA targeting complex colocalizes with components of the mitotic machinery, and that the knockdown of this complex impairs the mitotic cycle in mammalian cells. We also show that members of the CIA targeting complex colocalize and physically interact with the chromosome-associated kinesin KIF4A, a protein that plays an essential role in the formation of the midzone and midbody during telophase and cytokinesis. Importantly, our results identify a conserved cysteine-rich domain (CRD) at the C-terminus of KIF4A as a Fe-S cluster-binding site, representing a previously unidentified KIF4A post-translational modification. Furthermore, a KIF4A mutant specifically deleted in its CRD domain recapitulated some of the known phenotypes seen upon deletion of the KIF4A C-terminal cargo-docking tail (encompassing the CRD) (Hu et al., 2011; Kurasawa et al., 2004; Mazumdar et al., 2011; Midorikawa et al., 2006; Wu and Chen, 2008; Zhu and Jiang, 2005).

Collectively, our study demonstrates that a fraction of the CIA targeting complex colocalizes with components of the mitotic machinery in order to facilitate the transfer of Fe-S clusters to KIF4A and probably other targets, such as specific mitotic apoproteins. These findings suggest that the Fe-S cluster is an important determinant of correct KIF4A localization to its functional sites during mitosis. Furthermore, we propose that impaired Fe-S cluster delivery to KIF4A can in part explain the mitotic defects associated with the downregulation of the CIA targeting complex genes in human cells.

Recently, we searched for novel yeast CIN genes by screening the yeast temperature-sensitive (Ts) mutant collection for a CIN phenotype (Ben-Aroya et al., 2010, 2008; Stirling et al., 2011). One functional group of identified genes encompassed components involved in cytosolic and nuclear Fe-S protein biogenesis. To test whether this phenotype is conserved in evolution, we examined whether diminished levels of members of the human CIA targeting complex cause a chromosomal mis-segregation phenotype in cell lines during the mitotic cycle. The subcellular distribution of α-tubulin and chromosomes was examined by immunofluorescence and DAPI staining, respectively. In order to analyze the effect of CIA targeting complex depletion on chromosome segregation, we exploited mitotic HEK293 cell lines expressing shRNAs constructs for silencing MMS19 (Stehling et al., 2012), combined with siRNA to target CIA2B (Stehling et al., 2013). Western blot analysis confirmed the efficient knockdown of both CIA proteins (Fig. 1A). Combined knockdown of two members of the CIA targeting complex led to significant defects during metaphase, anaphase and the telophase stages of the mitotic cycle (Fig. 1B,C). In the MMS19/CIA2B knockdown cells, 24% of the mitotic cells showed abnormal multipolar spindles in metaphase (versus 6% in the control siRNA- and control shRNA-transfected cells), supporting previous results that were obtained in HCT116 cells (Ito et al., 2010). Moreover, we also observed an uncoordinated movement of single chromosomes within the spindles during anaphase in 25% of the cells, and the formation of abnormal midbodies in 22% of the cells (versus 5% and 6%, respectively, in the control samples) (Fig. 1C).

These results indicated that the CIA targeting complex is required for the proper progression of the mitotic cycle. To further examine this finding, we followed the subcellular distribution of the CIA targeting factors MMS19 and CIA2B in exponentially growing HeLa cells during the mitotic cycle. For MMS19, we stained the cells with anti-MMS19 antibody. Because no anti-CIA2B antibodies suitable for immunodetection were available, we determined the cellular distribution of a GFP–CIA2B fusion protein. During prophase, MMS19 was enriched on the centrosomes (Fig. 2A, top). Furthermore, and in agreement with a previous study (Ito et al., 2010), both MMS19 and CIA2B localized to the mitotic spindle during metaphase. During telophase MMS19 surrounded the compacted midzone microtubules (Fig. 2A, bottom), and CIA2B was detected on the midbody (Fig. 2B, bottom).

Taken together, the CIA targeting complex components CIA2B and MMS19 colocalize with elements of the mitotic machinery.

As mentioned above, during the final step of cytosolic-nuclear Fe-S protein maturation, the Fe-S cluster is inserted into the target apoproteins via the late-acting CIA targeting complex (CIAO1–CIA2B–MMS19). Preceding steps of cytosolic Fe-S protein biosynthesis include the early-acting Cfd1–Nbp35 scaffold complex, and the electron transfer chain NADPH–Tah18–Dre2 (human NDOR1–CIAPIN1) (Netz et al., 2014; Paul and Lill, 2014). Immunofluorescence analysis of HeLa cells stained with anti-α-tubulin, and anti-NBP35, -CFD1 or -CIAPIN1 primary antibodies showed that, in contrast to the CIA targeting complex, these early-acting CIA proteins were not enriched on elements of the mitotic machinery, yet were detected in the cytoplasm along the mitotic cycle (Fig. 2C). These results suggest that mitotic localization of the CIA targeting complex components CIA2B and MMS19 (Fig. 2A,B) is specifically required to facilitate Fe-S cluster transfer to selected mitotic components. However, none of the known Fe-S proteins identified so far was specifically linked to chromosome segregation.

In order to identify such a mitotic Fe-S protein, we relied on our recent affinity purification studies of the human CIA targeting complex components, followed by proteomic analysis (Stehling et al., 2012), which identified a list of CIA targeting complex-associated proteins. Among other proteins, we recognized the chromosome-associated kinesin KIF4A, which plays an essential role in chromosome congression and cytokinesis (Zhu et al., 2006, 2005). KIF4A also controls microtubule dynamics during anaphase and telophase, a process important for proper spindle midzone formation and successful cytokinesis (Hu et al., 2011; Wandke et al., 2012). To confirm the putative physical interaction between CIA targeting complex components and the chromokinesin KIF4A, we co-expressed FLAG–CIA2B or FLAG–MMS19 with GFP–KIF4A in HEK293 cells. Affinity purification of the FLAG-tagged CIA components revealed a co-purification of exogenous GFP-tagged KIF4A (Fig. 3A,B). The physical interaction was further confirmed by the co-precipitation of the endogenous KIF4A with FLAG-tagged CIA2B and MMS19 (Fig. 3C,D). Moreover, immunofluorescence analysis of HeLa cells expressing GFP–CIA2B and immunostained with an anti-KIF4A antibody showed that while KIF4A associates with mitotic chromosomes and CIA2B colocalizes with tubulin on the mitotic spindle during early mitosis, they both colocalize to the midzone and midbody during telophase and cytokinesis (Fig. 3E). These data suggest that the colocalization and physical interaction of KIF4A and the CIA targeting complex components is of physiological relevance.

Previous studies have revealed that Fe-S proteins transiently associate with the CIA targeting complex mainly in their apo-forms, that is, before Fe-S cluster incorporation (Gari et al., 2012; Vashisht et al., 2015). The association is stabilized when Fe-S cluster assembly is impaired, for example, upon the depletion of intracellular iron ions. This leads to impaired Fe-S protein biogenesis, and hence a more stable association of target apoproteins with the CIA targeting complex (Gari et al., 2012; Vashisht et al., 2015). Indeed, similar to what is seen with the Fe-S proteins XPD (also known as ERCC2), FANCJ (also known a BRIP1) and Pol δ (Gari et al., 2012), a 2.2 times higher level of KIF4A was found to co-immunoprecipitate with MMS19 in HEK293 cells treated with the iron chelator deferoxamine (Fig. 3F). Taken together, these results indicate that the chromosome-associated kinesin KIF4A colocalizes and physically interacts with the CIA targeting complex, suggesting that KIF4A is a candidate Fe-S protein.

Based on the above findings, we asked whether KIF4A is able to bind a Fe-S cluster. The binding motifs for Fe-S clusters are not highly conserved, yet these cofactors are usually coordinated via iron binding to cysteine sulfhydryl groups or, in rare cases, histidinyl residues (Lill and Mühlenhoff, 2008). KIF4A and its vertebrate homologues contain a CRD at the C-terminus, in addition to the highly conserved ATPase motor domain at the N-terminus and the long coiled-coil middle region (Mazumdar and Misteli, 2005; Wu and Chen, 2008) (Fig. 4A). Multi-sequence alignments identify nine conserved cysteine residues in the KIF4A CRD that in principle could qualify as Fe-S cluster coordination sites (Fig. 4B). To test whether this CRD could bind a Fe-S cluster in vivo, we expressed a C-terminal HA-tagged fragment of KIF4A containing (HA–KIF4A-F1) or lacking (HA–KIF4A-F1ΔCRD) the CRD in wild-type yeast cells (for details see Fig. 4A), and employed a well-established ⁵⁵Fe radiolabeling assay to directly follow Fe-S cluster insertion (Pierik et al., 2009). After radiolabeling, both KIF4A proteins were immunoprecipitated, and the associated radioactivity was estimated by scintillation counting. Because Fe-S cluster coordination to KIF4A turned out to be quite labile, the cell lysis and immunoprecipitation steps were performed under anaerobic conditions. HA–KIF4A-F1 bound a significant amount of ⁵⁵Fe (Fig. 4C). In contrast, only background values of ⁵⁵Fe were bound to HA–KIF4A-F1ΔCRD, despite the fact that the protein was expressed in amounts similar to HA–KIF4A-F1 (Fig. 4C, bottom). We next tested whether the ⁵⁵Fe incorporated into HA-KIF4A-F1 depended on known components of the mitochondrial (ISC) and cytosolic (CIA) Fe-S protein assembly machineries. We first depleted the mitochondrial core ISC components Nfs1, Yah1 and Ssq1 using yeast strains in which the respective genes were under the control of a glucose-repressible and galactose-inducible (GAL) promoter (Pierik et al., 2009). ISC protein depletion resulted in a strong diminution of ⁵⁵Fe binding to HA–KIF4A-F1 to values similar to the background radioactivity associated with the mutant HA–KIF4A-F1ΔCRD (Fig. 4C). This result suggests that the C-terminal fragment of KIF4A indeed binds a Fe-S cluster.

To assess the contribution of the CIA machinery to the potential Fe-S cluster formation on KIF4A, we performed the ⁵⁵Fe radiolabeling experiments with GAL-promoter yeast strains in which the different CIA components were depleted by growth in glucose (see above). Surprisingly, depletion of Grx4, Tah18, Nar1, Cia2 or Mms19 in the respective Gal-CIA cells was not accompanied by a diminished radioactivity bound to HA–KIF4A-F1 (Fig. S1A). Because in rare cases heterologously expressed proteins are targeted to mitochondria, we analyzed the subcellular localization of HA–KIF4A-F1 and its CRD-truncated version in wild-type yeast cells (Fig. S1B). Cell fractionation into post-mitochondrial supernatant and mitochondrial pellet fractions revealed that, in contrast to our expectation, both HA fusion proteins were predominantly found in the mitochondrial compartment. The reason for the mislocalization remains unknown, since no typical mitochondrial-targeting sequence was identified in the truncated KIF4A versions. However, the mitochondrial localization of HA–KIF4A-F1 and HA–KIF4A-F1ΔCRD explains why only depletion of ISC and not CIA components diminished ⁵⁵Fe association. In conclusion, our ⁵⁵Fe radiolabeling analysis reveals that the cysteine-rich domain of KIF4A is able to bind a Fe-S cluster upon expression in yeast.

To further verify our conclusion that KIF4A-F1 binds Fe-S cofactors, a His₆-tagged version was produced in E. coli, isolated via affinity chromatography (Fig. S2) and analyzed via several spectroscopic techniques (Fig. 5). Purified KIF4A-F1 displayed a UV–visible spectrum with a slight shoulder at 420 nm, hinting towards the coordination of a Fe-S cluster (Fig. 5A, ‘as is’). Upon chemical reconstitution of the purified protein with 4- and 8-fold molar excess of iron and sulfide ions, a strong absorption at ∼420 nm was observed (Fig. 5A, ‘1:4 rec.’ and ‘1:8 rec.’). The absorption signal was immediately lost upon reduction with sodium dithionite (Fig. 5A, ‘+DT’), indicating Fe-S cluster breakdown. Bands characteristic for Fe-S cluster binding were also evident by circular dichroism (CD) spectroscopy. While purified KIF4A-F1 (Fig. 5B, ‘as is’) had no distinct CD spectrum, chemical reconstitution of the recombinant protein led to CD spectra with a positive band at 450 nm and a negative band at 365 nm (Fig. 5B, ‘1:4 rec.’ and ‘1:8 rec.’). Biochemical reconstitution of KIF4A-F1 with four iron and sulfide ions each per polypeptide chain resulted in the binding of 3.9 Fe and 3.7 S. Reconstitution with twice the amounts of iron and sulfide ions did not substantially increase the stoichiometry. This saturable behavior may indicate the binding of one Fe-S cluster, presumably of the [4Fe-4S] type.

We attempted to characterize the type of Fe-S cluster bound to KIF4A by electron paramagnetic resonance (EPR) and Mössbauer spectroscopy (Pandelia et al., 2015; Freibert et al., 2018). EPR spectra recorded at 10 K temperature with a reconstituted and a dithionite-reduced sample (2 mM) showed a rather weak but distinct signal with an average g value below 2, which is typical for reduced [4Fe-4S] clusters [Fig. 5C, g=(2.077, 1.925, 1.891), spin concentration ∼0.03 mM]. The signal vanished upon raising the temperature from 10 K to 50 K (Fig. 5D). This behavior is typical for [4Fe-4S]⁺ clusters, and usually not seen for [2Fe-2S]⁺ clusters, and can be assigned to fast and strong temperature-dependent spin relaxation. An additional isotropic signal (found at g=4.3 that also remained at higher temperatures) indicates the presence of a minor impurity of mononuclear Fe(III) ions (S=5/2, spin concentration <0.005 mM), which is not unusual for sensitive iron-binding proteins.

Mössbauer spectroscopy of a ⁵⁷Fe-reconstituted sample without reduction revealed a complex zero-field spectrum that could be disentangled into four more- or less-resolved sub-spectra. The predominant quadrupole doublet with 34% intensity (Fig. 5E, green line) can be attributed to oxidized [4Fe-4S]²⁺ clusters, because of its typical intermediate isomer shift and quadrupole splitting (δ=0.44 mm/s, ΔE_Q=1.06 mm/s), indicating valence delocalized Fe(II/III) oxidation states. Another doublet with low isomer shift (blue line, δ=0.3 mm/s, ΔE_Q=0.52 mm/s, 20% intensity) can be assigned to valence-localized Fe(III) sites with tetrahedral sulfur coordination. Additionally, a Fe(II) component with sulfur coordination can be identified from its larger isomer shift [labeled Fe(II)_a, pink line, δ=0.65 mm/s, ΔE_Q=3.17 mm/s, 9% intensity]. Finally, the fourth doublet, Fe(II)_b (orange line, δ=1.18 mm/s, ΔE_Q=2.69 mm/s, 38% intensity) clearly indicates the presence of ferrous non-cluster iron ions with hard (N- or O-) ligands because of the typical high isomer shift. It can be assumed that the latter is a leftover of non-coordinated Fe(II) from the in vitro reconstitution of the protein. In principle the Fe-S-related, the Fe(II)_a component and a fraction of Fe(III) could come from reduced [2Fe-2S]⁺ clusters, but an aliquot of the EPR sample did not show the typical EPR spectrum of [2Fe-2S]⁺ clusters with its distinct low g values. Therefore, we suggest that the Fe(III) component most probably indicates oxidized all-ferric [2Fe-2S]²⁺ clusters, which are EPR silent, and the Fe(II)_a component arises from mononuclear iron centers, such as the rubredoxin-type [Fe-4S] or from Fe-S precipitates, which are also EPR silent.

Taken together, different spectroscopic analyses of recombinant and chemically reconstituted KIF4A-F1 unanimously reveal the binding of a Fe-S cluster to this protein. The majority is of the [4Fe-4S] cluster type, while the presence of a [2Fe-2S] type may indicate the rapid breakdown of the [4Fe-4S] cluster. These in vitro findings are fully consistent with the in vivo data demonstrating the ISC-dependent Fe-S cluster assembly of KIF4A in yeast mitochondria. The rather labile character of this Fe-S cluster may be crucial for its dynamic role in the progress of mitosis.

As mentioned above, KIF4A localizes to chromosomes in early mitosis, and re-localizes to spindle midzone in anaphase (Kurasawa et al., 2004; Mazumdar and Misteli, 2005). There, it physically interacts with the protein required for cytokinesis (PRC1). Together, the PRC1–KIF4A complex plays an essential role in determining the size of the antiparallel microtubule overlap in the midzone during cytokinesis (Bieling et al., 2010; Zhu and Jiang, 2005; Zhu et al., 2006). Previous studies have established the critical role of the C-terminal cargo-docking tail of KIF4A (harboring the CRD with its Fe-S cluster-binding site) in regulating the cellular localization of the protein and its interactions with other partner proteins (Hu et al., 2011; Kurasawa et al., 2004; Mazumdar et al., 2011; Midorikawa et al., 2006; Wu and Chen, 2008; Zhu and Jiang, 2005). These include the proper association of KIF4A with chromosomes and with the midzone/midbody during mitotic exit (Wu and Chen, 2008). Other studies have shown that various large C-terminal KIF4A deletions expanding from 126 to 881 amino acids, decreased its association with PRC1 (Hu et al., 2011; Kurasawa et al., 2004; Zhu and Jiang, 2005). To test the connection between the presence of the Fe-S cluster-binding domain (CRD) of KIF4A and its functions in mitosis, we examined the mitotic distribution of exogenously expressed GFP–KIF4A or a CRD deletion variant (GFP–KIF4AΔCRD) in KIF4A-knockout (KO) cells (generated by the Stephan Geley laboratory; Wandke et al., 2012). Consistent with previous studies (Kurasawa et al., 2004; Lee et al., 2001; Mazumdar et al., 2004; Wu and Chen, 2008; Zhu and Jiang, 2005), we show that the distribution of GFP–KIF4A is similar to that of the endogenous immunostained KIF4A (Fig. 6A, top and middle panels). Thus, we conclude that the GFP tagging of KIF4A provides a feasible approach to examine the cellular distribution of KIF4A in mammalian cells. In contrast to GFP–KIF4A, GFP–KIF4AΔCRD was diffuse or mis-localized from the midzone/midbody during anaphase and telophase (Fig. 6A, bottom). Co-immunoprecipitation experiments in HEK293 cells expressing Myc-tagged PRC1 as a bait, and either GFP–KIF4A or GFP–KIF4AΔCRD, reveal that this mis-localization was not the result of impaired association with PRC1 (Fig. S2B). Furthermore, it should be noted that some Fe-S apoproteins, such as XPD, are unstable in their apoform (Gari et al., 2012). In our analysis, the mis-localization of GFP–KIF4AΔCRD cannot be attributed to a putative destabilization, since the levels of GFP–KIF4AΔCRD were not affected when compared to those of the GFP–KIF4A control (Fig. S2C). Next, we tested how the mitotic profile is affected in cells expressing KIF4AΔCRD. Previous studies have shown that the knockdown of KIF4A is associated with prometaphase arrest (Mazumdar et al., 2004; Wandke et al., 2012; Wu and Chen, 2008). Geley and colleagues revealed a much stronger effect when combining the KIF4A KO with a knockdown of the human kinesin KID (also known as KIF22) (Wandke et al., 2012). Consistent with this, we show that the siRNA-mediated knockdown of KID in KIF4A KO cells results in an increased frequency of prometaphase cells (Fig. 6B). However, while expression of GFP–KIF4A partially reversed the prometaphase arrest of these cells, the GFP–KIF4AΔCRD variant protein had no effect (Fig. 6B).

Collectively, the phenotypes obtained when deleting large fragments of the KIF4A including the CRD (i.e. the Fe-S cluster-binding site) were recapitulated upon the specific deletion of the CRD. We conclude that these phenotypic effects can be attributed directly to the 59-residue-long segment.

The phenotypes obtained when removing the KIF4A CRD could have consequences for the structure of the KIF4A tail. To further support the notion that the observed phenotypes are the result of impaired Fe-S cluster binding rather than of an unstructured protein, we employed an in vitro yeast chromatin fractionation assay (Liang and Stillman, 1997). Herein, we show that the expected impairment of Fe-S cluster transfer upon the depletion of Nfs1, which is part of the mitochondrial ISC assembly machinery, abrogated the recruitment of the human HA–KIF4A-F1 fragment to chromatin in yeast cells (Fig. 6C). Similar results were obtained when specifically mutating KIF4A cysteine residues 1106, 1110 and 1112 into alanine, showing that alterations of individual cysteine residues rather large structural changes potentially accompanying the deletion of the entire CDR are causative of these effects (Fig. 6D).

Finally, we show that the expected impairment of Fe-S cluster transfer to the full-length KIF4A in IOP1-knockdown cells (Balk et al., 2004; Song and Lee, 2008) results in reduced association of the endogenous KIF4A with the midzone/midbody (Fig. 6E–G).

Taken together, these results imply that the Fe-S cluster-binding CRD is an important determinant for the recruitment and physical interaction of KIF4A with particular mitotic components. These findings support the notion that the mitotic defects associated with the downregulation of CIA targeting complex components can, at least in part, be attributed to a reduced level of Fe-S cluster transfer and insertion into the CRD of KIF4A.

In this study, we identified a mitotic component, the chromokinesin KIF4A, as a protein that is able to bind a Fe-S cluster at its conserved CRD (Fig. 7). Moreover, we show that decreases in Fe-S cluster association to KIF4A may contribute to genomic instability via hampering the mitotic pathway. To date, Fe-S cluster-mediated genomic instability has been mainly attributed to the Fe-S cluster-containing DNA polymerases and DNA helicases (Paul and Lill, 2015). Their functional impairment upon Fe-S cluster biogenesis defects leads to defects in the DNA maintenance pathway.

By using in vivo and in vitro approaches in yeast and E. coli combined with various spectroscopic methods, we show that the cysteine-rich motif of the chromokinesin motor protein KIF4A carries a Fe-S cluster. Whereas the C-terminal cargo-docking tails of kinesin-4 proteins are divergent, they all contain a set of nine, regularly spaced and highly conserved cysteine residues. This CRD was believed to form a zinc-finger structure or a protein–protein interaction domain (Mazumdar and Misteli, 2005; Vernos and Karsenti, 1995) but, as shown here, is able to coordinate a Fe-S cluster.

Our in vitro and yeast expression data showed the presence of a Fe-S cluster, presumably of the [4Fe-4S] type, bound to the CRD of KIF4A. When expressed in yeast, only the CRD-containing F1 fragment of KIF4A was able to bind radioactive ⁵⁵Fe. This occurred in an ISC assembly machinery-dependent fashion, clearly identifying the KIF4A-bound radioactive ⁵⁵Fe as part of a Fe-S cluster. The purified recombinant and chemically reconstituted KIF4A-F1 was able to bind ∼4 Fe and 4 S on average per polypeptide. A number of spectroscopic techniques (UV–visible, CD, EPR and Mössbauer) indicated the predominant presence of a [4Fe-4S] cluster at the CRD. However, we note that the sum of our analyses also show the presence of [2Fe-2S] species. This could be due to rapid decay of the [4Fe-4S] clusters in vitro, a behavior frequently seen for labile [4Fe-4S] clusters (Netz et al., 2016). According to our biochemical studies, the KIF4A Fe-S cluster belongs to the more unstable representatives of these metal cofactors. This labile cluster binding behavior might be physiologically relevant and allow the dynamic assembly and disassembly of the Fe-S cluster in order to regulate or assist the mitotic process.

The CRD with its nine conserved cysteine residues does not conform to any known Fe-S cluster-binding motif, including that of ferredoxins. Many of the recently identified Fe-S proteins were shown to contain non-canonical binding motifs, demonstrating the high versatility of polypeptide chains to accommodate Fe-S clusters (Netz et al., 2012; Wu and Brosh, 2012). The presence of nine cysteine residues in the KIF4A-CRD is reminiscent of eukaryotic replicative DNA polymerases with eight conserved cysteine residues (Netz et al., 2012). However, only four of these residues are used for the assembly of a [4Fe-4S] cluster, whereas the other four bind zinc ions as indicated by crystal structures (Klinge et al., 2009). It seems formally possible that KIF4A could bind another metal cofactor (including zinc ions or a second Fe-S cluster) at these extra cysteine residues within the CRD. Elucidation of this interesting question will require both mutational studies to define the important cysteine residues in vivo and further biochemical and structural studies.

MMS19 deletion has long been known to increase the sensitivity of cells to DNA damage. Several more-recent studies have revealed that these phenotypes are the result of the role of MMS19 as part of the CIA targeting complex in facilitating Fe-S cluster insertion into cytoplasmic Fe-S apoproteins (Gari et al., 2012; Stehling et al., 2012; Veatch et al., 2009). Impaired Fe-S cluster delivery in MMS19-deficient cells has been shown to inactivate transcription factor TFIIH-mediated nucleotide excision repair and transcription, by destabilizing the levels of the Fe-S cluster-coordinating TFIIH subunit XPD (Rudolf et al., 2006). Other Fe-S cluster-containing targets of the CIA machinery, such as DNA2, FANCJ and RTEL1 also play key roles in maintaining genome stability and the response to other types of DNA damage (i.e. DNA double-strand breaks and DNA interstrand crosslinks) (Pokharel and Campbell, 2012; Rudolf et al., 2006; Wu and Brosh, 2012). Hence, the deficiency of Fe-S protein biogenesis is considered to promote genomic instability mainly by simultaneous inactivation of multiple DNA repair pathways. However, our results indicate that the impairment of the DNA damage repair pathways provides only a partial mechanistic basis for the genomic instability associated with the downregulation of the CIA targeting complex, and points to an additional failure of the mitotic pathway. Indeed, here we show that the downregulation of the CIA targeting complex led to the impairment of stages of the mitotic cycle. Tanaka and colleagues had previously shown that the knockdown of MMS19 and CIA2B, then identified as components of the MMXD complex (MMS19–CIA2B–XPD), led to a poor alignment of chromosomes to the metaphase plate and improper localization of key mitotic factors (Ito et al., 2010). When published, it was unclear how these phenotypes were elicited by the knockdown of the MMXD complex. Based on later studies (Gari et al., 2012; Stehling et al., 2012), and our results, we now propose that the significant mitotic defects and chromosome aneuploidy associated with the impairment of the CIA targeting complex (MMS19–CIA2B–CIAO1) result, at least in part, from the loss of Fe-S cluster delivery to KIF4A.

As was mentioned above, the knockdown of both MMS19 and CIA2B results in a pleiotropic effect, namely impaired Fe-S cluster delivery to numerous proteins that play different roles in maintaining genome integrity. Hence, it would be challenging to distinguish which of the complex mitotic phenotypes associated with impairment of the CIA targeting complex can be specifically attributed to KIF4A. It is well established that during the final steps of cytosolic and nuclear Fe-S proteins, Fe-S cluster transfer is accomplished by the three CIA-targeting complex components CIAO1, CIA2B and MMS19 (Paul and Lill, 2014). However, recent studies have suggested that various CIA subcomplexes are involved in a target-specific delivery of the [4Fe-4S] clusters (Stehling et al., 2013; Paul and Lill, 2014). Thus, owing to the dynamic assembly of the CIA subcomplexes, our result showing that MMS19 and CIA2B do not necessarily colocalize on each of the elements of the mitotic machinery is not surprising (Fig. 2A,B). This could suggest that there are additional mitotic Fe-S targets that still await discovery. Different CIA subcomplexes and adaptor proteins (such as Yae1-Lto1) mediate maturation (Paul et al., 2015; Stehling et al., 2013).

The physiological role of a CRD-bound Fe-S cluster in regulating KIF4A subcellular localization and protein–protein interactions is currently unknown. Fe-S clusters were previously shown to play an important physiological role in protein stability, for instance in the case of the XPD C190S mutant, which displayed diminished Fe-S cluster incorporation in mammalian cells (Gari et al., 2012), or for stability of the Pol δ subunit Pol3 in yeast cells impaired in Fe-S cluster biogenesis (Netz et al., 2012). Our results exclude a similar possibility for KIF4A, since its levels were relatively stable both upon depletion of the CIA targeting complex and in the KIF4AΔCRD mutant.

Another study describes a two-step activation of XPD, showing that apo-XPD is first recruited to the CIA machinery via its interaction with the CIA targeting complex (Vashisht et al., 2015). The resulting acquisition of the Fe-S cluster then releases matured holo-XPD from the CIA machinery, and allows its assembly into the TFIIH complex (Vashisht et al., 2015). Consistent with this sequential mechanism, we show that cellular depletion of iron with the iron chelator deferoxamine, and hence impaired Fe-S protein biogenesis, results in a more efficient association between the CIA machinery and KIF4A. Hence, it is intriguing to speculate that KIF4A only transiently associates with the CIA machinery in its apoform, and that, only after Fe-S cluster transfer, is the mature protein released to associate with chromatin or KIF4A-interacting proteins, depending on the particular mitotic phase. In this case, lack of Fe-S cluster delivery to KIF4A may promote the association with the CIA targeting components, and lead to the impairment of many of its essential protein-protein interactions and proper subcellular localization. As a result, each stage of the mitotic cycle can be severely affected, with phenotypes similar to those seen upon the complete loss of KIF4A, including hypercondensation of mitotic chromosomes, various mitotic defects and chromosome mis-segregation (Hu et al., 2011; Mazumdar et al., 2004; Wandke et al., 2012).

In summary, our study identifies KIF4A as the first mitosis component modified by a Fe-S cluster, and adds the mitotic pathway to the list of Fe-S cluster-dependent processes. We show that two of the CIA targeting complex components colocalize with components of the mitotic machinery in order to facilitate the transfer of Fe-S clusters to KIF4A and probably other specific mitotic apoproteins. Furthermore, we propose that impaired Fe-S cluster delivery to KIF4A can, at least in part, explain the mitotic defects associated with the downregulation of the CIA targeting complex genes in human cells. Our work provides the framework for future dissection of the mechanistic role of the Fe-S cluster in modulating the activity of KIF4A.

Saccharomyces cerevisiae strain W303-1A was used as the wild-type strain (for information on all used yeast strains see Table S1). Yeast strains were cultivated in rich (YP) or minimal (SC) medium containing the required carbon sources at a concentration of 2% (w/v) (Sherman, 2002).

For information on all used expression vectors, see Table S2.

For ⁵⁵Fe radiolabeling experiments, genes encoding KIF4A-F1 and KIF4A-F1ΔCRD were subcloned into the yeast vector p424-TDH3 (Mumberg et al., 1995) containing an N-terminal 3HA tag sequence (see Table S2). To increase the stability of the fusion proteins the N-termini were expressed with a MAST-sequence (Varshavsky, 2008). In vivo radiolabeling of transformed yeast cells with ⁵⁵FeCl₃ was performed as described previously (Pierik et al., 2009). Lysis of cells after labeling and immunoprecipitation of KIF4A-F1 and the KIF4A-F1ΔCRD mutant were performed under anaerobic conditions as the Fe-S cluster stability was increased under this condition.

A pET-15b plasmid (Novagen) was used for heterologous expression of N-terminal His₆-tagged KIF4A-F1 in C41 (DE3) Escherichia coli host cells (see Table S2). For protein expression, 100 ml LB medium with ampicillin was inoculated with colonies harboring the plasmid and incubated 16 h at 37°C under shaking (170 rpm). 1% of the preculture was diluted into 2 liters of LB medium containing ampicillin and incubated at 37°C and 170 rpm. When the culture reached an optical density at 600 nm (OD₆₀₀) of 0.5 the shaker temperature was adjusted to 30°C, and overexpression of His₆–KIF4A-F1 was induced by the addition of isopropyl-β-D-thiogalactopyranoside (1 mM final concentration). After 4 h cells, were harvested by centrifugation (7000 g for 10 min), shock-frozen and stored at −80°C until use. Protein purification was initiated by resuspending E. coli cells in lysis buffer [50 mM Tris-HCl pH 8.0, 200 mM NaCl, 10 mM imidazole, 5% (v/v) glycerol] supplemented with Complete Protease inhibitor cocktail (Roche), PMSF, DNase and lysozyme. Disruption of cells was performed at 4°C by sonication (2 s pulse followed by 2 s pause, 20 min). After centrifugation (100,000 g, 90 min, 4°C) the supernatant was loaded on the Ni-NTA resin (Amintra) pre-equilibrated with lysis buffer. The column was washed with 10 bed volumes of lysis buffer containing 20 mM imidazole, followed by a second washing step with lysis buffer containing 30 mM imidazole. The sample was eluted with lysis buffer plus 300 mM imidazole. Desalting of purified proteins was carried out with a PD-10 column (GE Healthcare) equilibrated with 50 mM Tris-HCl pH 8.0, 150 mM NaCl and 5% (v/v) glycerol. Purified protein was stored at −80°C until use.

For chemical Fe-S cluster reconstitution, purified His₆–KIF4A-F1 (100 µM) was reduced in an anaerobic chamber (Coy Laboratory Products, Ann Arbor, MI) with 3 mM DTT for 3 h at 4°C. Anaerobic stock solutions of ferric ammonium citrate and Li₂S (200 mM in water) were prepared freshly. Reconstitution was started by adding reconstitution buffer (50 mM Tris-HCl pH 8.0) and a 4- or 8-fold molar excess of ferric ammonium citrate. After 5 min, Li₂S was slowly added at 4- or 8-fold molar excess, and samples incubated for 2 h at room temperature. Reconstituted protein was desalted on a PD-10 column equilibrated with 50 mM Tris-HCl pH 8.0, 150 mM NaCl and 10% (v/v) glycerol. Incorporation of Fe-S clusters into apoproteins was monitored by UV–visible spectroscopy (V-550, Jasco, Inc.). Circular dichroism (CD) spectra were recorded on a J815 CD spectrometer (Jasco).

Chemically reconstituted His₆–KIF4A-F1 was reduced under anaerobic conditions by addition of sodium dithionite (2 mM final concentration) and samples were shock-frozen after 2 min. Electron paramagnetic resonance (EPR) spectra were recorded at cryogenic temperatures with a Bruker ESP 300E X-band spectrometer equipped with an Oxford Instruments ESR910 helium flow cryostat, or with a Bruker E500 ELEXSYS spectrometer with the Bruker dual-mode cavity ER4116DM and an Oxford Instruments helium flow cryostat ESR 900. The microwave bridge of the latter was the Bruker high-sensitivity ER-049X Super-X unit with integrated microwave frequency counter. The magnetic field controller ER032T had been externally calibrated with a Bruker NMR field probe ER035M.

For Mössbauer studies, protein was reconstituted anaerobically with the ⁵⁷Fe isotope following the protocol above. Mössbauer spectra were recorded on a conventional spectrometer with an alternating constant acceleration of the γ-source. The minimum experimental line width was 0.24 mm/s (full width at half-height). The sample temperature was maintained constant in an Oxford Instruments Variox cryostat. The γ-source (⁵⁷Co/Rh, 1.8 GBq) was kept at room temperature. Isomer shifts are quoted relative to iron metal at 300 K.

The cell lines used in this study were HeLa cells, HEK293 cells and stable HEK293 cells expressing doxycycline-inducible shRNAs designed to silence expression of MMS19 (Gari et al., 2012). All cell lines used were maintained in Dulbecco's modified Eagle's medium (DMEM, GIBCO) supplemented with 10% fetal bovine serum (Biological Industries), 2 mM L-glutamine (GIBCO) and 1% penicillin-streptomycin (GIBCO). Cells were maintained in a humidified incubator, at 37°C and 5% CO₂. Stable shRNA cell lines were treated with doxycycline (1 μg/ml) for 7 days to turn on shRNA expression and attain sustained depletion of the genes of interest. Fresh doxycycline was added to the culture medium every 24 h. KIF4A-KO cells were generated, and kindly provided by the Stephan Geley laboratory (Biocenter, Division of Molecular Pathophysiology, Innsbruck Medical University, A-6020 Innsbruck, Austria) (Wandke et al., 2012). This laboratory previously demonstrated that the combined loss of function of KID (KIF22) and KIF4A is much stronger than the depletion of the individual chromokinesins. We thus added KID siRNA in KIF4A-KO cells, which resulted in an increased frequency of prometaphase cells (for details see Wandke et al., 2012).

Cells fixed in 4% paraformaldehyde were treated with 0.5% Triton X-100 at room temperature, and then washed twice with PBS. After blocking with 1% BSA in PBS for 60 min, cells were incubated for 1 h with antibodies (see later section), washed with PBS containing 0.1% Tween 20, and incubated with appropriate secondary antibodies conjugated to Alexa Fluor 488 or Alexa Fluor 594 (Invitrogen A11001, A11012). The DNA was visualized with 5 µg/ml DAPI (Sigma-Aldrich), and the coverslips were mounted on slides with Immu-Mount mounting solution (Thermo Shandon) and sealed. The samples were examined with an inverted microscope (Axio Observer.Z1; Carl Zeiss) equipped with an automated stage (MS-2000; Applied Scientific Instrumentation), a 300-W xenon light source (Lambda DG-4 LS; Sutter Instrument), a 63× oil, 1.4 NA Plan Apochromat objective lens, and a six-position filter cube turret with a GFP filter (excitation, BP470/40; emission, BP525/50; Beamsplitter, FT495), a HcRed filter (excitation, BP592/24; emission, BP675/100; Beamsplitter, FT615), and a DAPI filter (excitation, G365; emission, BP445/50; Beamsplitter, FT395) from Chroma Technology Corp. Images were acquired using a camera (CoolSnap HQ2; Roper Scientific). The microscope, camera and shutters (Uniblitz) were controlled by AxioVision Rel. 4.8.2 (Carl Zeiss).

For determination of the mitotic index, fixed HeLa or HCT116 cells were imaged using a LD 40×/NA objective. 20 images (1200–1600 cells) were analyzed per experiment, and mitotic cells identified in the UV channel by their condensed DNA content.

siRNAs against human mRNAs of CIA2B and a mix of scrambled siRNAs (control) with predesigned sequences were purchased from DharmaFECT (ON-TARGETplus SMARTpool siRNA) as follows: (1) sense, 5′-GUUUCAAGAUGGACGUGCAtt-3′; antisense, 5′-UGCACGUCCAUCUUGAAACgc-3′; (2) sense, 5′-AGUUGAACGUAGUAGAGCAtt-3′; antisense, 5′-UGCUCUACUACGUUCAACUcc-3′, and (3) sense, 5′-CGCGAGAUCUUCGAUCUGAtt-3′; antisense, 5′-UCAGAUCGAAGAUCUCGCGtg-3′.

Depletion of CIA2B was achieved through three consecutive transfections at a 3 day interval, applying a pool of siRNAs as described previously (Stehling et al., 2013).

Transfections were carried out using a standard calcium phosphate method (HeBS at pH 7.05, 0.25 M CaCl₂), or Metafectene reagent (Biontex Laboratories, Germany) according to manufacturer's instructions. After transfection, cells were incubated for 24–36 h before harvest or fixation.

Antibodies used in this study were: anti-Flag antibody directly conjugated to horseradish peroxidase (HRP) (1:2500, Abcam, ab1238), tubulin antibodies (1:2500, Abcam ab18251, ab7291), KIF4A rabbit polyclonal (1:2500, Bethyl, A301-074A), MMS19 mouse monoclonal (1:1500, EUROMEDEX, MMS-3H10), CIA2B rabbit polyclonal (1:500, Sigma HPA041501), GFP mouse monoclonal (1:2500, Roche, 11 814 460 001), and monoclonal antibodies against the HA tag (1:2500, sc-57592, Santa Cruz Biotechnology). Antibodies against CFD1, NBP35 and CIAPIN1 were raised in rabbits using recombinant purified protein. The specificity of the anti-CFD1 (1:200), -NBP35 (1:250), and -CIAPIN1 (1:200) antibodies used in immunofluorescence experiments was verified using respective RNAi depletion cells (see Stehling et al., 2008 for anti-NBP35).

All cell lines used in this study were washed twice in cold PBS and lysed in cold lysis buffer (10% NP40, 1 M Tris-HCl pH 7.6, 1 M NaCl pH 7.5, 0.5 M EDTA pH 8.0 and a mixture of phosphatase inhibitors consisting of 10 mM sodium fluoride, 20 mM P-nitrophenylphosphate, 20 mM β-glycerophosphate, 1 mM sodium orthovanadate and 1 mM protease inhibitor PMSF). The extract was incubated on ice for 30 min and the non-soluble components were then pelleted by centrifugation at 14,000 g for 45 min. The protein concentration in each sample was determined by Bradford protein assay (Bio-Rad). Protein extracts were separated by SDS-PAGE on 4–12% gels, transferred to nitrocellulose membranes, and blocked for 1 h at room temperature. The membranes were incubated with primary antibodies, washed with 0.1% Tween-20 in Tris-buffered saline, incubated with secondary antibody and finally visualized by exposing the membranes to imageQuant LAS 4000 mini.

For immunoprecipitations, lysates were incubated with anti-GFP (1:2500, 11 814 460 001, Roche) antibody for 24 h at 4°C. Then 25 µl of protein A–agarose-conjugated beads were added, and the mixture was incubated for 1.5 h at 4°C. Beads were recovered by centrifugation (1699 g, 1 min), washed two times with TGET buffer (20 mM Tris-HCl pH 7.5, 10% glycerol, 0.1% Triton X-100, 1 mM EDTA) supplemented with 150 mM NaCl, and one time with TGET buffer supplemented with 75 mM NaCl. For Flag immunoprecipitations, lysates were incubated with Flag-beads for 4 h. Beads were recovered by centrifugation, washed two times with TGET buffer as described above and then eluted with 3×Flag peptide. The beads were then boiled in SDS sample buffer for 5 min and briefly pelleted at 1699 g before the supernatant was loaded for electrophoresis.

Full-length human MMS19 cDNA was subcloned into the NotI and SalI sites of the mammalian expression vector pCMV-Myc with Flag tag by using forward primer 5′-TTACAGTCGACTGACTACAAGGACGACGATGACAAGATGGCCGCTGCCGCGGCTGT-3′, and reverse primer 5′-TATTGCGGCCGC TCAGCTGCCAGGGCTCCCCA-3′. The full-length human CIA2B was subcloned into the NotI and SalI sites of the mammalian expression vector pCMV-Myc with Flag tag by using forward primer 5′-TTACAGTCGACTGACTACAAGGACGACGATGACAAGATGGTAGGCGGCGGCGGG-3′, and reverse primer 5′-TATTGCGGCCGCTCAGGAGCGGGCTGACAGGC-3′. The coding sequence for full-length KIF4A fused to GFP (GFP–KIF4A) was a gift from Phang-Lang Chen, University of California, CA (Wu and Chen, 2008). The GFP–KIF4A plasmid was used as a template for PCR site-directed mutagenesis reaction to delete the CRD (amino acids 1086–1144) of KIF4A by using forward primer 5′-GAAGAACATCCAAGGGCCTACCGAGGTGACCC-3′ and reverse primer 5′-GGGTCACCTCGGTAGGCCCTTGGATGTTCTTC-3′.

The chromatin fractionation assay was performed as previously described (Liang and Stillman, 1997). Cells were grown to OD₆₀₀ 0.5 in 50 ml culture. Samples were spun down (865 g) in 50 ml conical tubes for 5 min, resuspended in 3 ml of 100 mM PIPES/KOH pH 9.4, 10 mM DTT, 0.1% Na-Azide, and incubated for 10 min at room temperature. Samples were then spun (385 g) for 2 min. Supernatant was aspirated off, and samples were resuspended in 2 ml of 50 mM KPi (potassium phosphate buffer), pH 7.4, containing 0.6 M sorbitol and 10 mM DTT, and transferred to 2 ml microfuge tubes. A 10 μl aliquot was then diluted in 990 μl H₂O in a cuvette. 4 μl of 20 mg/ml Zymolase T-100 was added for 10 min, incubated in 37°C water bath (tubes were gently inverted every 2–3 min). After 1 min, a 10 μl aliquot was used to measure the OD₆₀₀ (for hypotonic lysis). The OD of the 1:100 dilutions after spheroplasting was less than 10% of the value before. From this point on everything was done in a cold room. Tubes were spun for 1 min (1699 g), cells were then washed with 1 ml of 50 mM HEPES/KOH pH 7.5, 100 mM KCl, 2.5 mM MgCl₂ and 0.4 M sorbitol. Tubes were spun for 1 min (1699 g), and resuspended in equal pellet volume of EB (∼80 μl) (EB, 50 mM HEPES/KOH pH 7.5, 100 mM KCl, 2.5 mM MgCl₂, 1 mM DTT, 20 μg/ml leupeptin, 2 mM benzamidine, 2 μg/ml aprotinin, 0.2 mg/ml bacitracin, 2 μg/ml Pepstatin A, 1 mM PMSF, added just before use). A 1/40 volume of 10% Triton X-100 (0.25% final, e.g. 4 μl for 160 μl suspension), was added and cells were incubated for 3 min for lysis on ice, (vortexing occasionally). This sample represents the whole-cell extract (WCE). A 20 μl sample of the WCE was removed and 20 μl of SDS loading buffer was added. Then, 100 μl EBX-S was prepared in separate microfuge tubes (EBX-S, EB+0.25% Triton X-100 plus 30% sucrose). 100 μl of WCE was laid onto the EBX-S, and microfuge tubes were spun for 10 min (15,294 g). The resulting fractions represent a white chromatin pellet (CHR), the clear sucrose layer and, above, a yellow supernatant fraction (SUP). 20 μl of SDS loading buffer was added to 20 μl of the SUP fraction. The rest of supernatant and sucrose buffer were then aspirated. The chromatin pellet was resuspended in 100 μl EBX (EBX, EB plus 0.25% Triton X-100), and spun (10,621 g) for 5 min. Supernatant was aspirated off, and chromatin pellet was resuspended again in 100 μl EBX. 20 μl sample was then removed and added to 20 μl of SDS loading buffer.

A recent publication (Nag et al., 2018) has shown the essential role of the Drosophila MMS19 as a mitotic gene. The midbody localization of Mms19::GFP in telophase during the embryonic cell cycle may be explained, at least in part, by the Drosophila KIF4A target Fe-S protein matured by MMS19.

We thank Sebastian Bänfer (Core Facility ‘BioImaging’ of Philipps-Universität Marburg) for help with immunofluorescence, and for Stephan Geley for providing the KIF4A-KO cell lines.