|Home | About | Journals | Submit | Contact Us | Français|
The mitotic spindle is the apparatus upon which chromosomes are segregated during cell division. We have discovered new roles for two members of the NIMA-related kinase (NEK) family in different molecular processes of spindle assembly. Moreover, loss of these proteins leads to segregation errors that drive cancer progression.
Spindle assembly is a complex multistep process that begins upon mitotic entry. In prophase, the two centrosomes separate and move around the nucleus to form opposing spindle poles. This requires dissolution of the intercentrosomal linker that holds interphase centrosomes together and cross-linking and antiparallel sliding of microtubules nucleated from opposing poles. In prometaphase the nuclear envelope breaks down, giving microtubules access to the condensed chromosomes where they attach to the kinetochores of the sister chromatids. These microtubules become bundled into kinetochore (K)-fibers by intermicrotubule K-fiber bridging proteins, including transforming acidic coiled‐coil protein 3 (TACC3) and cytoskeleton-associated protein 5 (CKAP5, also known as ch‐TOG). These contribute to the robust kinetochore attachments required for chromosome congression. At metaphase, all chromosomes become aligned at the cell center with bi-orientated K-fibers. Only once this has been achieved is the spindle assembly checkpoint (SAC) inactivated and cells allowed to enter anaphase (Fig. 1A).
The NIMA-related kinases (NEKs) are a family of 11 serine/threonine kinases in humans.1 They are related to the Aspergillus NIMA (never in mitosis) kinase, which is essential for mitotic entry in this filamentous fungus. Several human NEKs contribute to mitotic progression suggesting conservation of function. Yet, compared to other highly conserved mitotic kinases, such as the cyclin-dependent kinases (CDKs), polo-like kinases (PLKs) and Aurora kinases, the functions of the NEKs remain poorly understood. In two studies published recently in the Journal of Cell Biology, we made exciting discoveries into how two members of this family, NEK5 and NEK6, contribute to the timing and vigor of spindle assembly.2,3
The first study on NEK5 revealed that cells depleted of this kinase by RNAi showed defective chromosome segregation as a result of delayed centrosome separation in prophase.2 The reason why centrosome separation was delayed in these cells was 2-fold. First, dissolution of the intercentrosomal linker was impaired. Linker disassembly is achieved through phosphorylation of linker components, including centrosomal NEK2-associated protein 1 (C-NAP1) and ciliary rootlet coiled-coil protein (CROCC, more commonly known as rootletin), by NEK2.4 However, in the absence of NEK5 there was less NEK2 at prophase centrosomes, which resulted in the retention of these linker proteins. Second, as cells enter mitosis centrosome maturation normally takes place, during which the centrosomes double in size through the recruitment of proteins such as γ-tubulin, centrosomal protein 192 kDa (CEP192), CDK5 regulatory subunit associated protein 2 (CDK5RAP2), and pericentrin.5 This dramatically increases the microtubule nucleating capacity of the centrosome facilitating cross-linking and sliding of microtubules that drive centrosomes apart. In cells depleted of NEK5 less γ-tubulin, CEP192, CDK5RAP2, and pericentrin was recruited to mitotic centrosomes, leading to a reduced rate of microtubule nucleation. As a consequence, there was a reduced capacity to generate the microtubule-dependent forces required to separate centrosomes (Fig. 1B).
The substrates that NEK5 phosphorylates to promote centrosome separation and maturation remain to be identified. The mitotic consequences of NEK5 depletion may result from loss of centrosome integrity in interphase, as cells depleted of NEK5 had reduced levels of C-NAP1, NEK2, γ-tubulin, CEP192, CDK5RAP2, and pericentrin. NEK5 may therefore promote the recruitment of these centrosome components throughout the cell cycle. Conversely, loss of NEK5 resulted in accumulation of excess rootletin and a moderate increase in intercentrosomal distance. This may be a consequence of less centrosomal NEK2 being present to regulate linker organization via basal phosphorylation,4 or a compensatory response to resist centrosome splitting due to reduced levels of C-NAP1. Whatever the cause, it will be of great interest to identify the substrates through which NEK5 regulates the centrosomal levels of these proteins and whether NEK2 and NEK5 directly cooperate to achieve disassembly of the centrosome linker.
For NEK6, it was already known that RNAi-mediated depletion or expression of kinase-inactive mutants leads to fragile spindles incapable of satisfying the SAC.6,7 However, NEK6 does not localize strongly to the spindle apparatus and its only reported substrate is kinesin family member 11 (KIF11). NEK6 phosphorylates KIF11 on S1033, thereby targeting it the centrosome to promote centrosome separation.8 An S1033A phosphonull mutant resulted in some mitotic delay, but did not lead to monopolar spindles. Therefore, KIF11 is unlikely to be the only mitotic substrate of NEK6. To identify additional NEK6 substrates, phosphorylation and interaction-based screens were performed. As described in the second report, these led to identification of the cytoplasmic chaperone Heat Shock 70 kDa Protein 1A (HSPA1, also known as HSP72) as a novel mitotic binding partner and substrate for NEK6.3
If HSP72 is a physiologic substrate of NEK6, it would be expected to have a similar role in mitotic progression. Consistent with this, cells in which HSP72 function was blocked, either by RNAi depletion or chemical inhibition, exhibited poorly organized spindles, chromosome congression defects, and delayed anaphase onset. Detailed analysis revealed weakened K-fibers with substantially reduced levels of TACC3 and ch-TOG. Alongside a strongly active SAC and an inability to maintain metaphase chromosome alignment, this points to a failure to generate stable kinetochore-MT attachments and thus build robust K-fibers. Co-precipitation experiments revealed that the HSP72 inhibitor reduced interaction of TACC3 with ch-TOG, implying that HSP72 may specifically facilitate their assembly into a complex and recruitment to K-fibers (Fig. 1C).
How then does NEK6 regulate HSP72 function? HSP72 is phosphorylated by NEK6 on T66 within the nucleotide-binding domain; this phosphorylation, and the interaction with Nek6, is mitosis-specific. Importantly, either NEK6 depletion or expression of a HSP72-T66A phosphonull mutant led to reduced K-fibers, whereas a HSP72-T66E phosphomimetic mutant could rescue the K-fiber defects that arose from NEK6 depletion. These data provide evidence that a major function of NEK6 is to promote K-fiber stabilization through phosphorylation of HSP72. However, while this can be explained in part by facilitation of ch-TOG/TACC3 complex assembly, our data suggest an additional mechanism. Total HSP72 localizes to spindle poles and spindle fibers similarly to ch-TOG and TACC3. However, Nek6-phosphorylated HSP72 is clearly detected not only at spindle poles but also at sites of kinetochore-microtubule attachment. This argues for a direct role for phosphorylated HSP72 in promoting plus-end attachment of microtubules to kinetochores. Furthermore, loss of HSP72 function led to reduced interpolar distances that, together with reduced astral microtubules and misoriented spindles, suggests a defect in cortical attachment. Thus, HSP72 may well have a dual function in supporting spindle assembly through promotion of K-fiber stability via ch-TOG/TACC3 and ensuring that microtubule plus ends make robust attachments to kinetochores and the cell cortex.
Chromosome segregation errors drive cancer progression and promote tumor heterogeneity. The assembly of a robust mitotic spindle to which the chromosomes attach in a stable and timely manner is therefore fundamental to preventing such errors. The studies on NEK5 and NEK6 demonstrate that both kinases are crucial to this process. In the absence of NEK5, centrosome separation is delayed in prophase and instead occurs via the prometaphase pathway that uses kinetochore microtubules and motor proteins to drive spindle pole separation. However, this promotes merotelic chromosome attachments that lead to a high frequency of segregation errors, including chromosome fragmentation, lagging chromosomes, and unresolved sister chromatids.9 Significantly, merotely does not trigger SAC-dependent arrest and therefore contributes directly to segregation errors as cells progress through mitosis unchecked.10 Through targeting HSP72 to kinetochores, NEK6 contributes to the formation of stable kinetochore-microtubule attachments that are essential for chromosome congression and thus mitotic progression. In the absence of NEK6, fragile spindles are formed, chromosome congression fails and the SAC remains active. Inhibition of Nek6 in cells with a robust SAC offers a therapeutic avenue to block spindle assembly and thus cell division. Furthermore, inhibition of NEK6 in tumors that have a weakened SAC could lead to mitotic catastrophe and selective death of cancer cells.
No potential conflicts of interest were disclosed.