The Aspergillus nidulans CENP-E kinesin motor KipA interacts with the fungal homologue of the centromere-associated protein CENP-H at the kinetochore

Authors


E-mail reinhard.fischer@KIT.edu; Tel. (+49) 721 608 44630; Fax (+49) 721 608 44509.

Summary

Chromosome segregation is an essential process for nuclear and cell division. The microtubule cytoskeleton, molecular motors and protein complexes at the microtubule plus ends and at kinetochores play crucial roles in the segregation process. Here we identified KatA (KipAtarget protein, homologue of CENP-H) as a kinesin-7 (KipA, homologue of human CENP-E) interacting protein in Aspergillus nidulans. KatA located at the kinetochore during the whole cell cycle and colocalized with KipA and partially with the putative microtubule polymerase AlpA (XMAP215) during mitosis. Deletion of katA was lethal at 37°C and caused severe growth and morphology defects at room temperature. KipA was shown before to play an important role in growth directionality determination and our new results suggest a second function of KipA in the interaction between the microtubule plus ends and the kinetochores during mitosis.

Introduction

Molecular motors are fascinating mechanochemical enzymes, which convert chemical energy into force. They generate the movement of a wide variety of materials in eukaryotic cells and thus are involved in many different cellular and developmental functions such as organelle movement, localization of polarity marker proteins, mitosis or meiosis (Schliwa and Woehlke, 2003). The transport machinery consists of polar cytoskeletal tracks, motors as engines, ATP as fuel, accessory factors as regulators and specific cargoes. Molecular motors are classified into microtubule-dependent kinesins and dynein and actin-dependent myosins. All eukaryotes contain motors of all three classes, although the number of motors of each class varies in different organisms. In the case of kinesin, Saccharomyces cerevisiae harbours only six, the filamentous fungus Aspergillus nidulans contains 11 and mammalian genomes encode about 50 putative kinesins. Kinesins are classified into 17 different families, which used to be named after, e.g. the founding members. However, a novel nomenclature has been introduced recently and the families were distinguished by numbers. The family of conventional kinesin is therefore now called kinesin-1 family, and the Kip2 family was re-named to kinesin-7 (Wickstead and Gull, 2006).

In S. cerevisiae kinesin-7 (Kip2) is required for microtubule plus-end accumulation of Bik1 (CLIP-170 homologue) (Carvalho et al., 2004; Wu et al., 2006). In S. pombe kinesin-7 (Tea2) performs the same function as Kip2 in S. cerevisiae, the transport of a CLIP-170 homologue, Tip1, to the microtubule plus end (Browning et al., 2003; Busch et al., 2004). The system could even be reconstituted in vitro, clearly showing that no other components such as membranous vesicles are required for the transportation (Bieling et al., 2007). In addition, Tea2 transports the cell end marker protein Tea1 (Browning et al., 2003). Tea1 is anchored at the membrane at the growing tip through the interaction with the membrane associated Mod5 protein and forms a larger protein complex with other proteins, which ultimately recruit formin. Formin then catalyses the formation of actin cables, which serve as tracks for myosin-dependent vesicle secretion, necessary for membrane and cell wall extension (Snaith and Sawin, 2003; Snaith et al., 2005). Deletion of tea2 caused bent instead of straight-growing cells (Browning et al., 2000). This phenotype is due to the lack of transportation of the cell end marker protein. In order to further unravel the polarity determination machinery, a yeast two-hybrid screening was performed using Tea2 as bait (Busch et al., 2004). The authors obtained 54 candidate genes, 12 of which encoded Tip1. The other potential cargoes have not yet been described.

In A. nidulans kinesin-7 (KipA) is – like in S. pombe– involved in polarity determination (Rischitor et al., 2004; Konzack et al., 2005; Fischer et al., 2008). In addition to Tea2, other cell end factors from S. pombe are also conserved in this filamen tous fungus. Tea1 (A. nidulans TeaA), Mod5 (TeaR), Tea4 (TeaC) and For3 (SepA) have been characterized and a role in determining the zone for polarized growth has been shown (Sharpless and Harris, 2002; Takeshita et al., 2008; Higashitsuji et al., 2009). However, TeaA and TeaC are apparently transported to the cortex independent of KipA (Takeshita et al., 2008; Higashitsuji et al., 2009).

Kinesin-7 is also present in mammals and other higher eukaryotes where it is called CENP-E. It is undetectable in non-dividing tissues and before late G2 in cycling cells. CENP-E accumulates at the end of G2 and is degraded after mitosis (Brown et al., 1994). Depletion of this kinesin causes chromosome segregation errors due to a weaker spindle assembly checkpoint (Weaver et al., 2003) and an impaired interaction between the centromeres and the microtubules of the mitotic spindle (Putkey et al., 2002). The interaction partners of the CENP-E motor protein found to date are the kinase BubR1, the spindle microtubule-associated protein CENP-F (Chan et al., 1998), Nuf2, a member of the Ndc80 complex (Liu et al., 2007), Skp1, which is implicated in CENP-E degradation after mitosis (Liu et al., 2006), and Mitosin/Septin 7 (Zhu et al., 2008). CENP-E is responsible for the localization and the activity of BubR1 at the kinetochore. BubR1 phosphorylates the checkpoint component Mad2 resulting in an arrest of mitosis in metaphase. This mitotic delay is necessary for correct attachment of the sister kinetochores to the spindle, without CENP-E the checkpoint is inactive resulting in a mis-segregation of sister chromatids (Putkey et al., 2002).

In order to identify kinesin-7 interacting proteins in A. nidulans, we performed a yeast two-hybrid assay and isolated several putative new KipA-binding proteins, one of which was KatA, a putative homologue of CENP-H. CENP-H is a constitutive component of the inner kinetochore plate and is required for the targeting of other centromeric proteins (Okada et al., 2006). We propose an anchoring rather than a transportation function of KipA at the kinetochore and KatA as the centromeric-interacting partner of KipA during mitosis.

Results

Yeast two-hybrid screening

In order to identify cargos for the kinesin-7 KipA we screened a cDNA library by using the Matchmaker Library Construction and Screening Kit (Clontech). We used the C-terminal half of KipA as bait in order to avoid interactions of the motor domain with cytoskeletal elements such as tubulin. The bait spans from the beginning of the coiled-coil domain to the stop codon and it was cloned into the bait vector (pGBKT7). The screening was performed twice, and ∼ 90 yeast colonies were isolated, one-third of which were able to grow again in the same medium (SD-Leu-Trp-His-Ade + 2 mM 3AT) and produced a blue colour in the X-gal assay. Plasmids from these colonies were sequenced, and 14 genes identified (Table 1) (Fig. 1A). We have chosen one candidate, KatA (KipAtarget), for a detailed analysis. The interaction between KatA and KipA was confirmed by co-immunoprecipitation (Fig. 1B).

Table 1.  KipA-interacting proteins detected in the yeast two-hybrid screening.
Identification No. at BROAD InstituteAmino acids/size of the isolated cloneNo. of clones in screeningConserved domainsAssigned function in other organisms
AN0925958/3194Sac1 domain, synaptojanin like phosphataseControl of endocytosis, cell vesicle trafficking and secretion
AN2054642/1924Alk1, haspin-like kinaseCell-cycle regulated kinase. Histone H3 phosphorylation
AN2768245/2451Spc34Dam–DASH complex subunit, a kinetochore complex
AN2886238/2382CENP-HCentromeric protein
AN10061881/4481F-box domain, CAAX motifCell separation and septation, localizes to septa and cell tips. Recycling pathway
AN2181112/1125TF2ATranscription initiation factor IIA gamma subunit
AN4739301/1372SAICAR synthetasePurine biosynthetic pathway
AN10182346/3462eIF3C, Mov34/MPN/PAD-1 familyTranslation initiation factor
AN2530182/1821Hsp2030 kDa Heat shock protein
AN4594116/22120S40S ribosomal subunit protein
AN8866476/4761ACT domaind-3-phosphoglycerate dehydrogenase
AN8953956/1191Alpha glucosidase 
AN370395/951unknown 
AN8343136/1361FKBP-typePeptidyl-propyl cis-trans isomerase
Figure 1.

Identification of KatA as interaction partner of KipA.
A. Mapping of the KipA cargo-binding domain. Yeast two-hybrid assay of KipA variants (depicted on the left) and the cargo proteins (indicated on the right). The KipA variants were cloned into the plasmid pGBKT7 (Clontech) and transformed into the S. cerevisiae strain AH109. Transformed yeast was crossed to the compatible strain Y187 containing the corresponding KipA-binding protein indicated above. After mating, 10 µl drops of yeast suspension with an OD600 of 0.1 were inoculated on a SD-Leu-Trp-His 2 mM 3AT plate and incubated at 30°C for 48 h. The blue boxes indicate the coiled-coil domains.
B. Co-immunoprecipitation of KatA and KipA. Western-blot with anti-GFP antibody of the probes indicated above the panel. Mycelium of the strains SSH27 (GFP–KipA), SSH46 (3xHA–KatA) and SSH47 (both) was cultivated at 37°C for 28 h in Thr-supplemented medium. GFP–KipA was co-precipitated with HA–KatA using the anti-HA antibody.
C. CustalW alignment of A. nidulans KatA, human CENP-H and S. pombe Fta3. Identical amino acids were shaded in black and similar amino acids are in grey. The sequence comparison is restricted to the region similar to CENP-H detected by the Pfam software (http://pfam.sanger.ac.uk/). Scheme of the exons of the katA gene (green) and the position and size of the CENP-H motif (yellow).

In order to map the interaction site of KipA, we made some overlapping deletions in the C-terminus of the KipA bait (Fig. 1A). For comparison, four further proteins, obtained after the yeast two-hybrid screening with the full-length bait, were included into the study as KipA-binding partners. RcyA corresponds to AN10061, SacA to AN0925, AlkA to AN2054 and KatA to AN2886 (for more details see Table 1). The yeast two-hybrid assay between these proteins and the kipA deletion mutants showed that the cargo-binding domain is located at the C-terminus (Fig. 1A, KW4). Deletion of the coiled-coil domain (Fig. 1A, KW2) also affected the interaction with RcyA, SacA and AlkA, suggesting that the region is necessary to keep the conformation of the cargo-binding domain. KatA interacted only with the full-length bait. In this series of experiments KipA self-interaction was also observed (data not shown). This is consistent with the current model for kinesin motor proteins, which proposes that dimerization through coiled-coil regions is necessary for the progression along the microtubule lattice.

KatA is the fungal homologue of the centromeric protein H (CENP-H)

The katA cDNA was found twice in the kipA screening as full sequence including 5′ and 3′ UTRs. The ORF has a length of 717 bp and codes for a polypeptide of 238 amino acids. The predicted two introns of 45 and 69 bp, respectively, were confirmed by cDNA sequencing (Fig. 1C). Database searches with the KatA protein revealed that the protein is conserved in filamentous fungi, none of which has been analysed before. ClustalW analysis of KatA showed that the best-conserved region was a motif close to the C-terminus (Fig. 1C). blast searches at the GeneDB revealed that the most similar protein in S. pombe is SPAC17H9.06c (e-value 0.00047). This is a hypothetical protein of 601 amino acids. The second one, with an e-value of 0.0022, was SPBP8B7.12c, which has been characterized as Fta3/Sma3 (Liu et al., 2005). Fta3 has a similar size as KatA (220 amino acids) and is associated to the Sim4 and Mal2 centromeric complex. Further analysis of KatA at the Pfam website indicated the presence of a partial CENP-H motif close to the C-terminus corresponding to the sequence 214-IIVGSGVDWAEDEDLTALVL-233 (key residues in bold). This motif spans the same region that is best conserved in comparison with other filamentous fungi, and both are indicated in the alignment in Fig. 1C. In the alignment of KatA with the human CENP-H and S. pombe Fta3 the key residues are better conserved in A. nidulans than in S. pombe. In a recent analysis S. pombe Fta3 has been assigned as CENP-H homologue whereas before Sim4 was considered as homologue of CENP-H (Pidoux et al., 2003; Meraldi et al., 2006; Przewloka and Glover, 2009). In any case Fta3 is also associated to the Sim4 complex (Liu et al., 2005).

CENP-H is an essential protein in human (Orthaus et al., 2006) and chicken cells (Fukagawa et al., 2001). It is a constitutive component of the inner kinetochore plate and is required for the targeting of other centromeric proteins. Although it was proposed that CENP-E, the vertebrate homologue of kinesin KipA, bridges the spindle microtubules to the kinetochore inner plate (Yao et al., 1997; Kim et al., 2008), to date it was not found which kinetochore protein is the direct partner of CENP-E at the inner plate. We propose that KatA serves this function and is the kinetochore anchor for KipA in A. nidulans. To proof such a function, we studied the localization of KatA.

KatA localizes in the mitotic spindle

KatA was tagged at the N-terminus with GFP, integrated at the katA locus and expressed from the native promoter or the alcA promoter with moderate induction (Experimental procedures). We confirmed that the localization of KatA is identical in both cases (Fig. S1). For further investigations, the transformants expressing GFP–KatA from the alcA promoter were analysed. The GFP–KatA fusion protein appeared as dots distributed regularly along the hyphae KatA was tagged at the N-terminus with GFP, integrated at the katA locus and expressed from the alcA promoter with moderate induction (Experimental procedures). In the transformed hyphae, the GFP–KatA fusion protein appeared as dots distributed regularly along the hyphae. The dots corresponded to nuclei (Fig. 2A). The GFP–KatA dots localized to the periphery of the nucleus close to the nuclear envelope, as shown in a strain with mRFP1–KatA/GFP–nup49, a marker for the nuclear pore complex (Fig. 2B) (kindly provided by S. Osmani) (Liu et al., 2009). The same localization pattern was observed when KatA was expressed from the natural promoter (Fig. S1).

Figure 2.

Localization of KatA during interphase.
A. Strain SSH04 was grown on minimal medium with glycerol as a carbon source and stained with DAPI. KatA localized at the nucleus as shown in the DAPI stained hyphae. GFP-tagged KatA appeared associated to a single cluster in every nucleus.
B. The mRFP1-tagged KatA cluster localized at the periphery of the interphase nucleus, close to the nuclear envelope. The nuclear pore complexes are visualized with Nup49-GFP. Hyphae have a diameter of 2–3 µm.

In order to study further KatA during mitosis, we crossed strain SSH16 (mRFP1–KatA) to a strain with GFP-labelled microtubules to visualize the mitotic spindles. In A. nidulans, the mitotic spindle appears as compact microtubule bundle, short and thick at the beginning of mitosis and long and thin during telophase. At the beginning of mitosis (prophase-metaphase), KatA was distributed along the spindle in a dot-like pattern with the main population of the protein at the middle of the spindle (Fig. 3A, upper row). During late mitosis (anaphase-telophase), KatA disappeared completely from the equator of the spindle and appeared as a couple of dots segregated at the spindle poles (Fig. 3A, lower row). This dynamics is typical for centromeric DNA and kinetochore proteins (Sanchez-Perez et al., 2005). The same localization of KatA was found when expressed from the natural promoter (Fig. S1). Next, we have chosen AlpA as another marker for the mitotic spindle. AlpA is a microtubule-associated protein homologous to Dis1/XMAP215. It localizes at the microtubule-plus end during interphase and is probably involved in microtubule polymerization (N. Takeshita et al., pers. comm.) (Enke et al., 2007). We observed a weak AlpA signal in metaphase at the mid-zone of the spindle, exactly in the same position where the mRFP1–KatA signal was strong (Fig. 3B, metaphase). After that, AlpA appeared to dissociate from the microtubules and localized to four or five dots along the spindle (Fig. 3B, anaphase) (Enke et al., 2007). This pattern has previously also been described in fission yeast (Garcia et al., 2001). The central dots, which colocalized with KatA, correspond to the kinetochores and the distal ones are probably associated to the spindle pole bodies (SPBs). During telophase, a rest of AlpA, associated to the microtubules was again visible while the main part was concentrated at the spindle poles (Fig. 3B, telophase).

Figure 3.

Localization of KatA during mitosis.
A. Colocalization of GFP-α-tubulin-tagged microtubules and mRFP1–KatA during metaphase (upper row) and anaphase-telophase (lower row). Fluorescence images of the mitotic spindle in SSH18. Spores were incubated overnight in minimal medium with 2% glycerol.
B. Colocalization of AlpA and KatA. Fluorescence images of the mitotic spindle in SSH20 during different phases of mitosis. Spores were incubated with minimal medium with 2% glycerol in a FluoroDish tissue culture dish. Hyphae have a diameter of 2–3 µm.

In S. pombe the AlpA homologue Alp14 has been described as kinetochore-associated protein (Garcia et al., 2001). Therefore, we studied also the functional relation between AlpA and KatA. To test for genetic interaction of alpA and katA in A. nidulans, we constructed a double-mutant strain. Depletion of KatA in combination with deletion of alpA had an additive effect and only very small colonies were formed (data not shown).

KatA localizes at kinetochore clusters

The subcellular localization of KatA at nuclei during interphase was reminiscent of SPBs. To test this localization, we expressed mRFP1–KatA along with GFP–ApsB (Veith et al., 2005) as a marker for SPBs and analysed the signals. KatA localized close to ApsB, but the signals did not overlap suggesting distinct structures (Fig. 4A). It has been described that the kinetochores of the eight chromosomes of A. nidulans appear as a single cluster during interphase (Yang et al., 2004). This cluster is tightly associated to the SPB (De Souza et al., 2009). The GFP–KatA signal thus could represent the kinetochore cluster. The localization of KatA in a cluster close to SPBs and the analysis of the protein sequence indicate that KatA is a kinetochore protein homologous to CENP-H.

Figure 4.

KatA localizes in the vicinity of the SPB during interphase.
A. Comparison of the localization of the spindle pole body (SPB) protein ApsB (GFP) and KatA (mRFP1). The insert was enlarged to show the distinct localizations.
B. Localization of the kinetochore protein GFP–SpcA, and mRFP1–KatA (strain SSH38) during metaphase (upper row) and anaphase (lower row). Hyphae have a diameter of 2–3 µm.

In order to test whether KatA indeed is a kinetochore component in A. nidulans, we compared the KatA localization with the kinetochore marker SpcA. SpcA is a Spc34 homologue, which is a well-conserved kinetochore protein only found in fungi and is part of the Dam–DASH complex. This complex forms a 16-mer ring around the microtubule lattice at the kinetochore (Miranda et al., 2005). Interestingly, SpcA was also isolated in our yeast two-hybrid screening as interacting protein of KipA (Table 1). Full-length SpcA was cloned as GFP-fusion protein under alcA promoter control and transformed in A. nidulans. In contrast to KatA, and according to the model of microtubule-induced self-assembly of the Dam–DASH ring complex at the kinetochore (Wang et al., 2007), SpcA was not detectable in interphase. During mitosis, however, SpcA was found in the nucleus as dots like mRFP1–KatA (Fig. 4B). GFP and mRFP1 signals overlapped during metaphase-anaphase (upper row) as well as at the end of mitosis prior to disassembly or degradation of the DASH complex (lower row). At the end of mitosis only one SpcA spot was detected at each nucleus (Fig. 4B, lower panels; and Fig. S2). The fact that KatA was detectable during all cell cycle phases indicates that kinetochore association of KatA is independent of the interaction of the kinetochore with microtubules, and strengthens the idea that KatA is a homologue of CENP-H, a kinetochore protein closely associated with the centromere.

katA is an essential gene

In order to determine the function of katA, we aimed to construct a katA deletion strain. We were not able to obtain such a strain with our standard methods. However, we found that strains expressing KatA under alcA-promoter control grew more slowly under repressing conditions. Therefore, we anticipated that katA is an essential gene and used the heterokaryon rescue technique to obtain a knock-out strain (Osmani et al., 2006a). We transformed protoplasts with the deletion cassette (Fig. S3A). Deletion of katA was confirmed by Southern blot (Fig. S3B). Heterokaryotic transformants grew normal but single-cell spores obtained from the heterokaryon could not grow on selective medium at 37°C (Fig. 5A) or in minimal medium at any temperature (data not shown). Conidiospores were only able to form small colonies with few conidia at low temperature in complete medium. These results suggest that deletion of katA is temperature-sensitive lethal at 37°C and growth is largely affected at room temperature. The residual growth at low temperature allowed us to study the cellular effects of the deletion in detail.

Figure 5.

Phenotype of a kat-A deletion strain.
A. Colony phenotype of wild-type (SO451), ΔkipA (SSK44) and ΔkatA (SSH29) strains. Complete medium agar plates were incubated for 48 h at 37°C or 72 h at 28°C.
B. DIC images of germlings of the ΔkatA mutant. Spores were incubated overnight at RT on coverslips with 0.5 ml of complete medium. Size bars are indicated in the figure.
C. Fluorescence images of the cell wall of SSH29 (ΔkatA) stained with Calcofluor white. Probes were incubated for one day on complete medium. Calcofluor white was added at 2 µg ml−1 five minutes before the microscopy.
D. Nuclear number and distribution were highly variable in different hyphae of the ΔkatA mutant. Nuclei were stained with DAPI. Size bars represent 5 µm.

KatA depletion causes pleiotropic changes in hyphal morphology

Microscopic inspection of germinated spores of the ΔkatA mutant revealed multiple growth defects and many aberrant forms. Some of the hyphae appeared normal, others displayed an irregular width (Fig. 5B, panels 1, 2 and 5). In addition, the morphology of nuclei was altered. In many cases, nuclear distribution along the hyphae was like in wild-type. In other hyphae the number, distribution and size of the nuclei varied largely. We observed that abnormal nuclear localization is usually accompanied by aberrant hyphae. We observed an abnormal accumulation of cell wall material at the tips and along hyphae (see arrows at panels 1, 2 and 3, Fig. 5B), which were stained with Calcofluor white (Fig. 5C). Septation was also highly abnormal with regards to the form and spacing of septa, in addition, some hyphae were without any septa (Fig. 5B, panels 2, 4 and 5). Some germinated spores were very large (Fig. 5B, panels 4 and 5), suggesting uncontrolled isotropic growth prior to or after germ tube formation. However, we observed also conidiospores with abnormal size in spore suspensions obtained from colonies on agar plates (data not shown), suggesting uncontrolled growth during spore formation. Furthermore, deletion of katA affected polarity initiation. In wild-type, spores form a single germ tube and a second one at the opposite site. In contrast, ΔkatA spores produced three, four or even five germ tubes at random positions (Fig. 5B, panels 7 and 8). Nuclear number and distribution were also highly variable in different hyphae (Fig. 5D). The role of KatA in hyphal growth was also studied in a strain where katA was placed under the control of the alcA promoter (SSH04). Conidia were germinated in glucose-containing media (repressing conditions) at 37°C and germlings observed in the microscope. As in the case of the katA deletion strain, misshapen germlings were observed, although the number was lower and the phenotype less severe (Fig. S4). This may be explained by residual amounts and the stability of the KatA protein in conidia. At room temperature no phenotype was visible.

KipA colocalizes with KatA in the spindle equator

To study the relation between KipA and KatA, we studied their localization in dependence of each other. We crossed the strains SSH16 (mRFP1–KatA) and SSK92 (GFP–KipA) and obtained strain SSH25 with both proteins tagged. In interphase, KipA localized at growing microtubule plus ends as previously described (Konzack et al., 2005), and KipA was absent from the nucleus (Fig. 6A). In comparison, the location of KatA was restricted to the kinetochore with a weak signal in the entire nucleus (Fig. 6A).

Figure 6.

Colocalization of KipA and KatA.
A. Fluorescence microscopy images of strain SSH25, which expresses GFP–KipA and mRFP1–KatA.
B. Localization of GFP–KipA and mRFP1–KatA at the mitotic spindle in strain SSH25.
C. Localization of GFP–KipA in the ΔkatA background. A piece of agar from the colony border of strain SSH31 was placed on a coverslip with complete medium containing 2% glycerol and incubated for one day.
D. GFP–KatA localization in strain SSH11 (ΔkipA).

At the mitotic spindle, KatA was observed at the kinetochores while GFP–KipA stained the entire spindle (Fig. 6B). However, KipA was not regularly distributed along the spindle. It was mainly located at the spindle poles and at the equator where it colocalized with KatA. The localization of KipA at the extreme of the spindle corresponded to the SPBs as shown by KipA comets, which indicate growing microtubule plus ends of astral microtubules and which originated from these points (movie 1). The localization at the middle of the spindle suggests that the interaction between KipA and KatA take place at the kinetochores during mitosis.

To study the localization of KipA in the absence of KatA, we crossed the deletion strain SSH29 with SSH27, which expresses the GFP-tagged KipA and obtained strain SSH31. The colonies with a ΔkatA phenotype expressing GFP–KipA were not able to form conidiospores in glucose-supplemented medium. As previously described for the ΔkatA-hyphal morphology, KipA localization was sometimes similar to wild-type (Fig. 6C, movie 2) or strongly altered. The KipA signal was very different in intensity and localization in every hypha and even in different compartments of the same hypha. Because the location of KipA depends primarily on microtubules, the delocalization of the kinesin in the katA deletion mutant is primarily the result of changes in the cytoskeleton observed in Fig. 7A.

Figure 7.

Effect of katA deletion on the microtubule cytoskeleton and the frequency of chromosome loss.
A. Microtubule shaping in the SSH33 strain (GFP-MT, ΔkatA) during interphase.
B. Secondary bundles of microtubules observed in SSH33 at the mitotic spindle with free end (arrow) or bound to both poles (arrowhead).
C. Quantification of the haploidization events at increasing concentrations of benomyl indicated in the graphic. The graphic represents the number of white or yellow sectors in the colonies of diploid strains after 2 weeks at 37°C. Ten colonies of the kipA/kipA strain and 12 of the ΔkipA/ΔkipA were grown on plates with MM supplemented with the auxotrophic markers present in the parental strains.

In contrast to the dependence of correct KipA localization on KatA, KatA localization appeared to be unaffected in the absence of KipA (Fig. 6D). As suggested previously KatA seems to be tightly associated to the centromere even in interphase when the microtubules and KipA are absent from the nucleus. The interaction between both proteins was analysed by bimolecular–fluorescence complementation (BiFC) (data not shown). The YFP signal was observed at the kinetochore clusters during interphase, although KipA alone has never been observed in the nucleus during interphase. While our data support a KatA–KipA interaction, the persistent presence of KipA in the kinetochore in interphase, and through the cell cycle, is likely an artefact due to the interaction between the two halves of YFP being irreversible.

Frequency of chromosome loss in diploid strains

To investigate an effect of katA deletion on nuclear division, we compared the wild-type and the ΔkatA mutant for the integrity of the MT cytoskeleton and chromosome loss during mitosis. In the katA knock-out strain the organization of the microtubule cytoskeleton was severely altered (Fig. 7A). Microtubules appeared with slightly modified lengths or distribution in some hyphal compartments. In other compartments an ordered microtubule array was completely absent and sometimes no GFP signals were observed at all. These alterations probably explain the strong effects of the katA deletion on hyphal morphology described in Fig. 5B. Deletion of katA affected also the morphology of the mitotic spindle. Instead of compact spindles, additional microtubule bundles were observed. They had a free end (Fig. 7B, left panel) or were connected to both poles giving the spindle the appearance of a security needle or a bow (Fig. 7B, right panel).

To study the effect of the kipA deletion on chromosome loss, we used a diploid strain. Diploid spores are formed at a low frequency and can be obtained from conidia that are generated in a heterokaryon. A diploid strain is stable when the selection pressure is maintained. In the absence of the selection pressure chromosome loss can be observed by using spore colour markers. Chromosome loss can be stimulated by the application of low concentrations of benomyl and can be visualized by the formation of sectors or colony areas with different spore colours (Rischitor et al., 2004). For the present study we crossed the strains GR5 (white conidia) and RMS011 (yellow conidia) and selected a diploid with the kipA wild-type copy (kipA/kipA, green conidia). The ΔkipA strain SS44 (white conidia) was also crossed to RMS011 to obtain the heterozygous strain, kipA/ΔkipA. A homozygous ΔkipA diploid strain was generated by crossing SSK44 to GR5 and RMS011 to obtain ΔkipA strains with the appropriate colour and auxotrophic markers. Those strains were crossed to obtain the ΔkipA/ΔkipA strain. We compared the frequency of haploidization by counting the number of sectors with different colour (Fig. 7C). The number of haploidization events was very low, less than one per colony. The number of sectors increased at benomyl concentrations of 0.4 and 0.5 µg ml−1. The number of sectors at low concentrations of benomyl was slightly higher in the ΔkipA/ΔkipA strain than in the wild-type (Fig. 7C). This suggests a role of KipA during mitosis. This finding is further supported by the finding that in a ΔkipAΔkipB double mutant spindles with an abnormal shape have been observed, although at low frequency (Fig. S5).

Discussion

The motor protein kinesin-7 is a component of the centromere-kinetochore interaction machinery in vertebrates and is involved in the determination of growth directionality in yeast and filamentous fungi. To date, there was no report about the function of a CENP-E-homologue in polar growth and there was no evidence for the presence of the yeast CENP-H homologue, Kip2, in the nucleus. In A. nidulans KipA is involved in polarity maintenance and the protein localized also in the mitotic spindle (Konzack et al., 2005). However, the exact function of KipA in mitosis remained elusive. In the present work, we identified KatA, a KipA-binding protein, as the CENP-H homologue in A. nidulans. In addition, our results suggest that mammalian CENP-E may connect mitotic microtubules to the inner plate of kinetochores through interaction with CENP-H.

The depletion of CENP-H in vertebrate cells is lethal because of aberrant chromosomes and mis-segregation of chromatids (Fukagawa et al., 2001; Orthaus et al., 2006). In A. nidulans deletion of katA caused a temperature-sensitive phenotype. Whereas at high temperature no growth was observed, small colonies were produced at lower temperatures. At permissive conditions, pleiotropic phenotypes were observed, which may be explained through the main role of KatA during mitosis. The altered morphology of the mitotic spindle (Fig. 7B) in the ΔkatA mutant supports the idea of abnormal nuclear divisions. A. nidulans is a syncytial organism and the loss of one chromosome could be compensated by its presence in a neighbour nucleus. However, after a couple of nuclear divisions the gene dosage of an increasing number of genes must be altered and probably causes various phenotypes and finally cell death. A snapshot of hyphae thus will show a mixture of normal and severely impaired hyphae and all stages in between (Fig. 5B and C). However, most of the phenotypic effects of the katA deletion are related to polar growth and the cytoskeleton: cell wall and septum formation, nuclear distribution, germ tube formation and branching. A third possibility is that KatA participates in the structure of the SPB, because of its proximity during interphase. As a result of its deletion there is a cytoskeletal disruption that affects polar growth and generates the described aberrant forms.

The protein sequence of KatA is only very weakly conserved in higher eukaryotes but the localization at kinetochores strengthens the idea that it is a homologue of CENP-H. It was proposed that CENP-E links spindle microtubules to kinetochores (Yao et al., 2000), but none of the interacting partners identified functions at the inner kinetochore plate. Assuming that KatA is a homologue of CENP-H, we identified KipA as an interacting partner. This suggests CENP-E as an interacting partner of CENP-H in higher eukaryotes. In agreement with this hypothesis is that depletion of CENP-H by RNAi reduced CENP-E binding to the kinetochore to 72% in human cells (Orthaus et al., 2006). Another connection between CENP-H and E is established through the protein Nuf2, a member of the tetrameric complex Ndc80, which can bind both CENP-H (Mikami et al., 2005) and CENP-E (Liu et al., 2007).

If KipA fulfils the same function in A. nidulans than CENP-E in vertebrates, it has to be considered, that the coiled-coil region of CENP-E is very long (230 nm) and could span the distance between the spindle microtubules and the kinetochore inner plate. The KipA coiled-coil domain is considerably shorter than the one of CENP-E and a function as centromere-kinetochore linker seems to be unlikely in fungi. However, the structure of the fungal kinetochore is clearly different from that of higher eukaryotes. All eight kinetochores form a single cluster and there are exclusive fungal complexes like the Dam–DASH ring, which bind KipA too (see below and Table 1). On the other hand, CENP-E acts as regulator of the checkpoint activity (Weaver et al., 2003). We do have further results that suggest such a function also in A. nidulans. We identified other kinetochore-associated proteins (Table 1), such as AN2768 and AN2054, in the yeast two-hybrid screening. The first one was SpcA a well-conserved fungal protein integrated in the DASH complex that forms a ring around the microtubule lattice at the kinetochore (Miranda et al., 2005). SpcA was located at the kinetochore during mitosis and the GFP–SpcA signal overlapped with that of mRFP1–KatA (Fig. 4B). The interaction between KipA and SpcA suggests that the Dam–DASH ring could be pulled along microtubules by the kinesin or that it participates as a tension sensor. The second one is very similar to mammalian haspin and less similar to yeast Alk1 and Alk2 (Nespoli et al., 2006). We named it AlkA and it localizes in the nucleus during mitosis (Fig. S2). Haspin is a kinase that phosphorylates the centromeric histone H3 and is required for normal mitosis (Dai et al., 2006). At the centromeric nucleosome the histone H3 is replaced by CENP-A, which is closely associated to the CENP-H complex (Okada et al., 2006). Again, we have another possible relationship between KipA and KatA indirectly through AlkA. Taking these data together, we propose that KipA is implicated in the control of the kinetochore-centromere attachment in fungi, like CENP-E in higher eukaryotes. In agreement with such a role for KipA is also the recent finding that CENP-H, as part of the CENP-A NAC/CAD complex, has been proposed to be a direct regulator of kinetochore-microtubule dynamics (Amaro et al., 2010).

In vertebrates, the kinesin CENP-E localized to the outer kinetochores and is essential for the establishment and correct functioning of the spindle checkpoint (Putkey et al., 2002). The spindle associated checkpoint arrests mitosis in metaphase until all kinetochores are correctly attached to microtubules. It was previously described in A. nidulans a hypersensitivity to microtubule destabilization of sldA744 mutants (Efimov and Morris, 1998). SldA is the homologue of the budding yeast spindle assembly checkpoint gene BUB1. A ΔkipA mutant is also more sensitive to benomyl than wild-type and it was also suggested a redundant function of KipA and the mitotic kinesin KipB (Konzack et al., 2005). In the diploid strains, the number of haploidization sectors was only at low concentrations of benomyl slightly higher in the ΔkipA/ΔkipA mutant than in the wild-type (Fig. 7C). This indicates that the absence of KipA produces a slight increase in aneuploidy events, suggesting a small role in the kinetochore similar to CENP-E in vertebrates. Deletion of kipA in A. nidulans results in a curved-hyphal phenotype but the colony phenotype is not altered (Konzack et al., 2005 and Fig. 5A). It is surprising that altered polar growth has no effect on the colony size. However, A. nidulans strains with an inactive spindle assembly checkpoint display a faster mitosis than wild-type (De Souza et al., 2009). If KipA is necessary for this checkpoint activity, like CENP-E, the deletion of kipA should result in faster nuclear divisions and probably faster growth that could compensate the longer distance of a curved hypha.

Our investigation of the localization of AlpA at the mitotic spindle provides evidence for a new function of AlpA after metaphase. It was proposed that Alp14, the homologue of AlpA in S. pombe, localizes at the kinetochore in metaphase and is responsible for the activation of the checkpoint (Garcia et al., 2001). Our results suggest that in the moment the kinetochores divide and segregate to the poles, AlpA migrates to the kinetochores and moves polewards together with them (Fig. 3B, anaphase). AlpA is still observable at the kinetochore in telophase while KipA has disappeared completely from the spindle (Fig. S6). Recently, it was proposed that XMAP215, the Xenopus homologue, is a microtubule polymerase and promotes microtubule growth (Brouhard et al., 2008). This activity is reversible and XMAP215 can also accelerate microtubule depolymerization. In addition, it has been reported that Alp14/Dis1 and the microtubule-destabilizing kinesin Klp5/Klp6 play cooperative roles in S. pombe and the double knock-out is synthetically lethal (Garcia et al., 2002). We propose that the role of AlpA in mitosis is related to microtubule dynamics, promoting anaphase depolymerization of the spindle microtubules, and not to the checkpoint activity. On the contrary, it may be that the kinetochore components switch AlpA from a microtubule-stabilizing protein to a depolymerase at the end of metaphase.

In summary, KipA is a kinesin-7 motor protein with two faces: a previously reported role in polar growth and a new role during mitosis related to that of CENP-E.

Experimental procedures

Strains, plasmids and culture conditions

Supplemented minimal (MM) and complete media (CM) for A. nidulans were prepared as described, and standard strain construction procedures are described by Hill and Käfer (2001). A list of A. nidulans strains used in this study is given in Table 2. Standard laboratory Escherichia coli strains (XL-1 blue, Top 10 F′) were used. Plasmids are listed in Table 3.

Table 2. A. nidulans strains used in this study.
StrainGenotypeSource
  1. All strains harbour in addition the veA1 mutation. In strains used for crosses only the relevant mutation are listed.

TN02A3pyrG89; argB2, ΔnkuA::argB; pyroA4Nayak et al. (2006)
RMS011pabaA1, yA2; ΔargB::trpCΔB; trpC801, veA1Stringer et al. (1991)
SJW02wA3; pyroA4; alcA(p)::GFP::tubA; ΔargB::trpCΔBJ. Warmbold, Marburg, Germany
SSK44pabaA1; wA3;ΔargB::trpCΔB;ΔkipA::pyr4; veA1kipA)Konzack et al. (2005)
SSK92wA3; pyroA4; alcA(p)::GFP::kipAKonzack et al. (2005)
SSK28pabaA1; wA3; ΔkipB::argB; pyroA4, ΔkipA::pyr4Konzack et al. (2005)
SCE05ΔargB::trpCΔB; pyroA4; alcA(p)::GFP::alpAEnke et al. (2007)
SDV83pabaA1, yA2, pyrG89; ΔalpA::pyr4Enke et al. (2007)
SNZ59TN02A3 transformed with pNZ-SI38 [apsB(p)::GFP::apsB]N. Zekert. Karlsruhe, Germany
SO451pyrG89; wA3; argB2; pyroA4; ΔnkuA::argB; sE15, nirA14, chaA1, fwA1Nayak et al. (2006)
CDS394wA3; argB2; pyroA4; Nup49::GFP::pyrGOsmani et al. (2006b)
SSH03TN02A3 transformed with pSH17 (GFP-AlkA) pyroA4This study
SSH04TN02A3 transformed with pSH18 (GFP–KatA) pyroA4This study
SSH08TN02A3 transformed with pNT (YFPN–KatA) and pNT36 (YFPC–KipA) pyroA4This study
SSH11SSK44 (ΔkipA) crossed to SSH04 (GFP–KatA) wA3This study
SSH15TN02A3 transformed with pSH32 (mRFP1-AlkA) pyrG89This study
SSH16TN02A3 transformed with pSH33 (mRF1P–KatA) pyrG89This study
SSH17SJW02 (GFP–TubA) crossed to SSH15 (mRFP1-AlkA)This study
SSH18SJW02 (GFP–TubA) crossed to SSH16 (mRFP1–KatA)This study
SSH20SCE05 (GFP–AlpA) crossed to SSH16 (mRFP1–KatA)This study
SSH24SNZ59 (GFP–ApsB) crossed to SSH16 (mRFP1–KatA)This study
SSH25SSK92 (GFP–KipA) crossed to SSH16 (mRFP1–KatA)This study
SSH26CDS394 (GFP–Nup49) crossed to SSH16 (mRFP1–KatA)This study
SSH27SSK92 (GFP–KipA) crossed to RMS011; wA3, yA2, ΔargB::trpC ΔBThis study
SSH29SO451 transformed with the katA deletions cassette. wA3; pyrG89; argB2; pyroA4, ΔnkuA::argB; sE15; ΔkatA::pyrGThis study
SSH31SSH27 (GFP–KipA) crossed to SSH29 (ΔkatA)This study
SSH32SJW02 (GFP–TubA) transformed with pSH34 (mRFP1–KipA) wA3, ΔargB::trpCΔB; trpC801This study
SSH33SSH32 (GFP–TubA) crossed to SSH29 (ΔkatA) wA3This study
SSH34TN02A3 transformed with pSH21 (GFP-Spc34); argB2, ΔnkuA::argB; pyroA4; alcA(p)::GFP::spcAThis study
SSH38SSH16 (mRFP1–KatA) crossed to SSH34 (GFP-spc34)This study
SSH46TN02A3 transformed with pSH42 (3xHA–KatA) argB2, ΔnkuA::argB; pyrG89; alcA(p)::3xHA::katAThis study
SSH47SSH27 (GFP–KipA) crossed to SSH46 (HA–KatA); yA2This study
SSH48Diploid strain katA/katA of cross between SO451 x RMS011This study
SSH49Diploid strain katA/ΔkatA of cross between SSH29 x RMS011This study
SSH50SSK44 crossed to RMS011, yA2, ΔkipA, ΔargB::trpCΔB, trpC801This study
SSH51SSH58 crossed to RMS011; yA2; ΔkipA, pyroA4This study
SSH52TN02A3 transformed with katA-GFP cassette; katA(p)::katA::GFP; argB2, ΔnkuA::argB; pyroA4This study
SSH58SSK44 (ΔkipA) crossed to GR5; wA3, ΔkipA, pyroA4This study
SSH59kipA/kipA diploid strain from cross of GR5 to RMS011This study
SSH60ΔkipA/ΔkipA diploid strain from cross of SSH51 to SSK44This study
SSH61Δ kipA/kipA diploid strain from cross of SSK44 to RMS011This study
Table 3.  Plasmids used in this study.
PlasmidsConstructionSource
pCR2.1-TOPOCloning vectorInvitrogen (NV Leek, The Netherlands )
pCMB17apxalcA(p)::GFP, for N-terminal fusion of GFP to proteins of interest; contains N. crassa pyr4V. Efimov (Piscataway, USA)
pGBKT7Yeast Two-Hybrid bait vector, Gal4-BDClontech
pGADT7Yeast Two-Hybrid prey vector, Gal4-ADClontech
pDV7GFP replaced N-terminal half of YFP in pCMB17apxTakeshita et al. (2008)
pDV8GFP replaced C-terminal half of YFP in pCMB17apxTakeshita et al. (2008)
pCE05alpA in pMT-mRFP1Enke et al. (2007)
pSH091.1 kb kipA C-terminal fragment in pGBKT7This study
pSH170.7 kb alkA fragment in pGBKT7apxThis study
pSH180.3 kb katA fragment in pCMB17apxThis study
pSH21spcA (full length) in pCMB17apxThis study
pSH25kipA deletion KW2 in pSH09This study
pSH26kipA deletion KW3 in pSH09This study
pSH27kipA deletion KW4 in pSH09This study
pSH28kipA fragment from pSH34 in pDV7This study
pSH29katA fragment from pSH18 in pDV8This study
pSH320.7 kb alkA fragment in pCMB17apx, with mRFP1 instead of GFP and pyroA instead of pyr4This study
pSH330.3 kb katA fragment in pCMB17apx, with mRFP1 instead of GFP and pyroA instead of pyr4This study
pSH341.0 kb 5′-fragment of kipA in pCMB17 apx with mRFP1 instead of GFP and pyroA instead of pyr4This study
pSH420.3 kb 5′-fragment of katA in pCMB17 apx with 3xHA instead of GFP and pyroA instead of pyr4This study

Molecular techniques

Standard DNA transformation procedures were used for A. nidulans (Yelton et al., 1984), E. coli (Sambrook and Russel, 1999) and S. cerevisiae. For PCR experiments, standard protocols were applied using a Biometra Personal cycler (Biometra, Göttingen) for the reaction cycles. DNA sequencing was done commercially (eurofins-MWG-operon, Ebersberg, Germany). Genomic DNA was extracted from A. nidulans with the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany). DNA analyses (Southern hybridizations) were performed as described by Sambrook and Russel (1999). Western blot was performed with the MiniProtean system (Bio-Rad) following the manufacturer instructions.

Deletion of katA

katA flanking regions were amplified by PCR using genomic DNA and the primers Kat1.2 (5′-AGGAATGTTTTCGAGGTTAGC-3′) and pyrG–KatA–LB (5′-gaagagcattgtttgaggcgAGAGCAAGGCCTCGTGGCAGG-3′) for the upstream region of katA and pyrG–KatA–RB (5′-atcagtgcctcctctcagacagTGCGGTACAGGGATATTTTGG-3′) and kat3 (5′-ACCGCCCAGATGGAGCTGGTC-3′) for the downstream. The pyrG-gene from plasmid pFNO3 (S. Osmani, Ohio, USA) was amplified by PCR and used as template together with katA-flanking regions for the fusion-PCR. The deletion cassette was amplified with the fusion-PCR method (Szewczyk et al., 2006) with the primers kat1.2 and kat3. The resulting PCR-product was transformed into the pyrG89-auxotrophic A. nidulans strain SO451 (kindly provided by S. Osmani, Ohio State University, USA) and the transformants were selected by the heterokaryon rescue method (Osmani et al., 2006a).

Transformants were screened by PCR for the homologous integration event. Single integration of the construct was confirmed by Southern blotting (Fig. S3). One katA deletion strain was selected from the transformants and named SSH29.

Tagging of proteins with GFP

For fluorescence microscopy a 0.3 kb fragment of the katA gene and the full length of spcA, respectively, were cloned into two versions of the pCMB17apx plasmid one with GFP as tag and pyr4 as marker, and the other one with mRFP1 and pyroA. The resulting plasmids are listed in Table 1. The primer set used for katA was AN2886-AscI (5′-GGCGCGCCTATGACATCTTCCAAAGGGCG) and AN2886-PacI (5′-TTAATTAATAATGAGCACGGTGCTGACCG). The restiction sites AscI and PacI are underlined. For spcA the primer set was AN2768-Asc (5′-ATACCAAAAGGCGCGCCGATGTCGTTGCTAGAAAGCCAC) and Spc34-rev-pac2 (5′-TTAATTAATCACGAAAGGCCGCCCAAATC). All of these plasmids were transformed into the uracil-auxotrophic TN02A3 (ΔnkuA). The integration events were confirmed by PCR and Southern blotting (results not shown).

Yeast two-hybrid analysis

The yeast two-hybrid analysis was performed using the Matchmaker Library Construction & Screening system (BD Clontech). For bait generation, a kipA cDNA fragment corresponding to the C-terminal half of KipA (505-889 amino acids) with primers KipA-F(NcoI) (5′-CCATGGATTCGCTGGTCAGTATTCTTTG) and KipA_SR (5′-GCGCGTCGACTCATGCACTTGGACGACTG-3′) was amplified and cloned in the pGBKT7 vector, which contains the GAL4 DNA-BD and the TRP1 marker (BD Clontech). cDNA from Aspergillus strain SRF200 were amplified and cloned in the pGADT7-Rec vector, under the manufacturer instructions, which contains the GAL4 DNA-AD and the LEU2 marker (BD Clontech). pGBK7-associated plasmids were transformed in yeast AH109 (mating type MATa) and pGADT7-associated plasmids were transformed in yeast Y187 (mating type MATα). The system utilizes two reporter genes (HIS3 and LacZ) under the control of the GAL4-responsive UAS. β-Galactosidase activity was analysed by colony-lift filter assay using X-gal (X-nitrophenyl β-d-galactopyranoside (Karl Roth) as substrate.

For the truncated versions of KipA used in the yeast two-hybrid assay, primers with phosphorylated 5′-ends were used to amplify the entire bait vector except the deleted region. The primers used for the KW2 mutant were KW2F (P-5′-GATGGAAAGGCGAGTGCTCAA) and KW1R (P-5′-AAGCTCTTTCTCGGCCTGCGC; for the KW3 were KW3F (P-5′-AACCAATCGTTACCAAAAGAG) and KW2R (P-5′-GTGGAATTCTTGGCTATGGAC); and for the KW4 were KW4F (P-5′-TAAGTCGACCTGCAGCGGCCG) and KW3R (P-5′-GGAGCGAAGAGCTGTCAGCAT).

Light/fluorescence microscopy

For live-cell imaging of germlings and young hyphae, cells were grown on coverslips in 0.4 ml MM + 2% glycerol (induction of the alcA promoter), MM + 2% glucose (repression of the alcA promoter). Cells were incubated at room temperature for 1–2 days. For pictures of young hyphae of each gene deletion strain, the spores were inoculated on microscope slides coated with MM + 2% glucose + 0.8% agarose and grown at 30°C for 1 day. Images were captured at room temperature using an Axiophot microscope (Zeiss, Jena, Germany). Images were collected and analysed with the AxioVision system (Zeiss).

Acknowledgements

We kindly thank Aysha and Stephen Osmani (Ohio State University, USA) and their colleagues for their help with the heterokaryon rescue technique and the generous gift of various plasmids and strains. This work was supported by the special program ‘Lebensmittel und Gesundheit’ from the Baden-Württemberg Stiftung. N.T. was a fellow of the Humboldt Society.

Ancillary