MAPping the kinetochore MAP functions required for stabilizing microtubule attachments to chromosomes during metaphase

Abstract

In mitosis, faithful chromosome segregation is orchestrated by the dynamic interactions between the spindle microtubules (MTs) emanating from the opposite poles and the kinetochores of the chromosomes. However, the precise mechanism that coordinates the coupling of the kinetochore components to dynamic MTs has been a long-standing question. Microtubule (MT)-associated proteins (MAPs) regulate MT nucleation and dynamics, MT-mediated transport and MT cross-linking in cells. During mitosis, MAPs play an essential role not only in determining spindle length, position and orientation but also in facilitating robust kinetochore-microtubule (kMT) attachments by linking the kinetochores to spindle MTs efficiently. The stability of MTs imparted by the MAPs is critical to ensure accurate chromosome segregation. This review primarily focuses on the specific function of non-motor kinetochore MAPs, their recruitment to kinetochores and their MT-binding properties. We also attempt to synthesize and strengthen our understanding of how these MAPs work in coordination with the kinetochore-bound Ndc80 complex (the key component at the MT-binding interface in metaphase and anaphase) to establish stable kMT attachments and control accurate chromosome segregation during mitosis.

Introduction

1. Outer kinetochores at a glance - structural and functional components

The kinetochore is a complex multiprotein structure that assembles at the centromere region of chromosomes. Kinetochores connect dynamic mitotic spindle microtubules to sister chromatids to drive equational partitioning of genetic material between the two daughter cells during mitosis (Alushin and Nogales, 2011; Cheeseman and Desai, 2008). Failure or anomalies in this attachment process arrest or delay the mitotic exit via the activation of the spindle assembly checkpoint (SAC) - a fail-safe surveillance system that synchronizes the attachment of the chromosomes to the spindles during cell cycle progression (Bakhoum and Compton, 2012; Foley and Kapoor, 2013). Among the multitude of proteins that populate mitotic kinetochores, an evolutionarily conserved network of protein complexes, the KMN network (Knl1-Mis12-Ndc80), functions as the lead conductor in mediating the robust attachment of kinetochores to spindle microtubules (Cheeseman and Desai, 2008; Varma and Salmon, 2012). Within the KMN network, Ndc80 is a heterotetrameric dumbbell-shaped complex comprising four subunits - Hec1, Nuf2, Spc24 and Spc25, and a major contributor to the MT-binding function of the kinetochore (Umbreit hot, 2012; Wei et al., 2007); Knl1 follows suit as the second but less distinctive facilitator (Espeut et al., 2012; Ghongane et al., 2014). The dimer Hec1/Nuf2 has been shown to bind to MTs even in the absence of Spc24/25. This interaction is attributed to the characteristic folding of the calponin homology (CH) domains located at the N-terminal globular regions of Hec1/Nuf2 (Ciferri et al., 2008; DeLuca et al., 2006; DeLuca and Musacchio, 2012; Wei et al., 2007). The CH domain is a 100 amino acid (aa) long tetrahelical structural motif that is commonly found in proteins that bind to the cytoskeleton (Korenbaum and Rivero, 2002). In Fimbrin or α-Actinin, several CH-domains are arranged in tandem to promote high-affinity actin binding, which differs from the single CH-domain in the EB1 protein, which mediates attachment to MTs (Gimona et al., 2002). In addition to the CH domain, an intrinsically unstructured, highly positively charged ~80 aa N-terminal tail in the Hec1 protein was also required for high-affinity MT-binding (Alushin et al., 2010; Cheeseman et al., 2006; DeLuca et al., 2006; Guimaraes et al., 2008; Zaytsev et al., 2015). However, the involvement of the Ndc80 tail in MT attachments varies across species. In mammalian cells, deletion of the N-terminal tail destabilized kinetochore-microtubule (kMT) attachments, but no obvious perturbation in chromosome segregation was observed in worms or budding yeast (Cheerambathur et al., 2017; Tooley and Stukenberg, 2011).

Notwithstanding the fact that the Ndc80 complex of the KMN network has been repeatedly demonstrated as the primary contributor to the MT-binding function, Knl1/Blinkin/AF15q14 has also been shown to function as a conserved MT-binding kinetochore protein that is critical for effective MT attachment; however, its role in this context remains unclear (Cheeseman and Desai, 2008; Varma and Salmon, 2012).

2. Looping in on the (Ndc80) loop – how much do we know?

It has been observed that the depletion of Ndc80 did not completely abolish kMT attachment in some experimental systems, even though the attachments were rudimentary and were not able to sustain chromosomal congression and eventual segregation (Ciferri et al., 2007). This result is suggestive of the idea that additional MT-binding proteins at kinetochores must function either independently or in conjunction with the Ndc80 complex to orchestrate efficient contact with kMTs. The structural analysis revealed that the entire central part of the Ndc80 complex is composed of a coiled-coil structure with a kink/loop approximately 16 nm from the Hec1/Nuf2 dimer head (Wang et al., 2008). While the sequence of this loop is not strictly conserved, its position within the coiled region is relatively fixed. The loop primarily imparts some degree of flexibility by introducing a break in the otherwise long and rigid coiled-coil structure of Ndc80 (Nilsson, 2012). Although the loop does not contribute to MT-binding directly and is separated by distance from the globular Hec1/Nuf2 heads that partake in MT-binding activity, it is interesting that there has been a recent upsurge in the study of this internal loop region, particularly its role in establishing end-on attachments with MTs for accurate chromosomal segregation (Nilsson, 2012; Tang and Toda, 2013; Zhang et al., 2012). Thus, it is imperative to understand the intricacies of how the loop domain of Ndc80 participates in the modulation of spindle MTs during cell division.

There is a plethora of evidence available from studies on organisms from yeast to humans suggesting that the loop domain of Ndc80 functions as a docking hub for several proteins that can bind to MTs at kinetochores (Hsu and Toda, 2011; Maure et al., 2011; Tang and Toda, 2013; Varma et al., 2012; Zhang et al., 2012): for example, the Dam1/DASH complex in budding yeast, Dis1 in fission yeast, and the Ska complex and Cdt1 in humans are all important components recruited to kinetochores with the assistance of the loop domain. A common theme that emerges from these findings is that the loop domain may function to recruit several MT-binding proteins to kinetochores where they can optimally augment the MT-binding ability of the Ndc80 complex. Thus, the loop region of Ndc80 can be compared to an integrating, multichannel circuit where additional “loops” simultaneously or independently feed into each other to regulate or amplify the output. While some of the proteins that dock to the Ndc80 loop region have been identified and studied in great detail, additional proteins that may feed into the Ndc80 loop to possibly provide temporal control of kMT attachments in a stage-specific manner during mitosis have yet to be discovered. It is interesting that, whether via the loop or without it, all these proteins absolutely need the Ndc80 complex to localize to kinetochores, and their ultimate role is to assist Ndc80, especially by augmenting its MT-binding function.

3. A functional overview of the mitotic MAPs required for chromosome alignment and segregation

During mitosis, an entire array of molecular machinery, i.e., the mitotic spindle, is built by MTs and plays an essential role throughout the whole process of chromosome capture, congression, and segregation followed by successful cytokinesis (Desai and Mitchison, 1997). The ability of MTs to perform a plethora of diverse functions originates from their intrinsic dynamic behavior, which is regulated and coordinated by a dedicated family of proteins referred to as microtubule-associated proteins (MAPs) (Akhmanova and Steinmetz, 2008). By virtue of their ability to regulate MT turnover, stability and dynamics, MAPs are involved in a wide variety of processes, such as guiding MTs towards specific cellular locations, cross-linking MTs, and facilitating the interactions of other proteins with MTs. Mitotic MAPs can be broadly categorized into 4 types based on their function: (1) MAPs that promote and stabilize MT polymerization, (2) MAPs that lead to MT destabilization or severing, (3) MAPs functioning as spindle MT cross-linkers, and (4) MAPs that participate in kinetochore motility. These proteins may also function as motor or nonmotor MAPs based on their structural and enzymatic properties, as reviewed by Petry S (Petry, 2016).

(a). Motor MAPs

MAPs serving as motors are mechanoenzymes that utilize energy from ATP hydrolysis to physically walk along the length of the MTs. The most widely and exhaustively studied mitotic motor MAPs are those of the mitotic kinesin superfamily, including the MT-depolymerizing kinesin 13 family (e.g., Kif2s, MCAK, and Klp10A), the kinesin 8 family (e.g., Kif3 and Kif18A), (Wordeman, 2010) and the kinesin 14 family (e.g., Kar3, Ncd, KlpA, HSET and Xctk2) (Daire and Pous, 2011; Furuta et al., 2008; Wordeman, 2010). Yet another class of motor MAPs that associates with the chromosome arm rather than kinetochores and have been proposed to be required for the stabilization the kMT attachments, are the kinesin-10 family members including the chromokinesin Kid in humans and Nod in Drosophila (Cane et al., 2013; Drpic et al., 2015).

Another extremely well-studied category of motor MAPs are the kinetochore-associated MT motors, including cytoplasmic dynein and the kinesin-7 subfamily member centromeric protein E (CENP-E), which move towards MT minus-ends and plus-ends, respectively (Kapoor et al., 2006; Kardon and Vale, 2009; Kim et al., 2010). In addition to the diverse array of mitotic functions carried out by cytoplasmic dynein, such as maintaining normal mitotic spindle architecture and chromosome alignment, it is required for generating robust end-on kMT attachments during metaphase (Varma et al., 2008; Yang et al., 2007). CENP-E has also been reported to stably associate with both the assembling and disassembling MT tips and thus plays an active role in linking kinetochores with dynamic MT tips after the establishment of end-on kMT attachments (Gudimchuk et al., 2013).

In summary, the motor MAPs use MTs as tracks to enable the motility of chromosomes in the ‘+’, anterograde and ‘−’, retrograde direction as required during chromosome alignment and segregation. Since several other reviews have already been published on mitotic motors (Maiato et al., 2017; Welburn, 2013) and the primary goal of this review is to highlight the function of the nonmotor MAPs required for stabilizing kMT attachments during metaphase and anaphase, we refrain from an extensive discussion on the function of motor MAPs during mitosis in this review (Maiato et al., 2017; Welburn, 2013).

(b). +TIPs

Unlike the motor MAPs that bind along the entire surface of the MT lattice, other MAPs popularly called +TIPs distinctively associate with the plus-ends of MTs and are suggested to be inherently required for stabilizing the attachments between MT plus-ends and kinetochores in metaphase (Ferreira et al., 2014; Tamura and Draviam, 2012). The first reported plus-end tracking protein was the 170 kDa cytoplasmic linker protein CLIP-170. Since its discovery, more than 20 different +TIP families have been identified. Some +TIPs, including EB1 (end-binding 1), chTOG (known as XMAP215 in Xenopus and CKAP5 in humans), CLIP-170, CLIP-associated proteins (CLASPs) and their homologs, promote polymerization of MT plus-ends, while others, including kinesin-13/MCAK, depolymerize MTs at the plus-ends (Akhmanova and Steinmetz, 2008; Akhmanova and Steinmetz, 2010). EB1 and chTOG are unique +TIPs since they can autonomously accumulate at the MT plus-ends without the assistance of other MAPs (Akhmanova and Steinmetz, 2008; Akhmanova and Steinmetz, 2010). Since EB1 of the EB family proteins can recruit most of the other +TIPs and several other MAPs to the MT plus-ends through direct protein-protein interactions, it is often considered “the master +TIP”.

EB family proteins consist of four functional regions: the N-terminal calponin homology (CH) domain, which is typically found in actin-binding and signaling proteins and is responsible for MT-binding (Akhmanova and Steinmetz, 2008); the central coiled-coil domain for homodimerization; the EB homology (EBH) domain; and a disordered acidic tail with the C-terminal EEY/F motif (Matsuo et al., 2016). The EBH domain specifically binds to a variety of other +TIPs that contain the SxIP (x=any aa) motif, including CLASPs, adenomatous polyposis coli (APC), microtubule-actin cross-linking factor (MACF) and MCAK. The short linear SxIP motif serves as a general MT tip localization signal for these proteins (Honnappa et al., 2009). The C-terminal EEY/F motifs of the EB family proteins guide certain other MAPs, including CLIP-170 and p150^glued to MT plus-ends, by binding to cytoskeleton-associated protein-glycine rich (CAP-Gly) domains that have a well-conserved GKNDG sequence motif (Li et al., 2002; Saito et al., 2004; Weisbrich et al., 2007). The LxxPTPh motif (x=any aa and h=any hydrophobic aa) was recently identified as a third EB-binding motif that enables major +TIPs, including yeast Kar9 and human TACC1, to interact with the EB proteins at the MT ends (Kumar et al., 2017). However, there is currently only limited evidence to suggest that EB1 or any other factor that it recruits to MT plus-ends is required for stabilizing kMT attachments. Further, since our primary focus is on the nonmotor kinetochore MAPs required for stable kMT attachments, an elaborate discussion of EB1 and its binding partners is beyond the scope of this review.

In contrast, another class of +TIPs that has been clearly demonstrated to be required for stabilizing kMT attachments is the XMAP215 family of proteins. These proteins localize to the extreme MT plus-ends, in contrast to members of the EB1 family, which bind to the extended region, and localize to mitotic kinetochores (Al-Bassam and Chang, 2011). The functions of the XMAP215 family members are discussed in detail in the succeeding section on the nonmotor kinetochore MAPs that are required for stabilizing kMT attachments during mitosis.

4. Kinetochore MAPs required for robust kMT attachment during metaphase and anaphase

Kinetochores establish a stable association with dynamic MT ends, which undergo constant assembly and disassembly to achieve proper chromosome alignment and segregation. In this section, we comprehensively analyze the characteristics of several proteins that are recruited to kinetochores by the Ndc80 complex and contribute to the robust kMT attachment during metaphase and anaphase.

(a). The Ska complex

The heterotrimeric spindle and kinetochore-associated (Ska) complex, formed by the association of three subunits (Ska1, Ska2, and Ska3/Rama1), is enriched at the kinetochores of aligned chromosomes and has been shown to bind to MTs (Fig. 1A1, Fig. 2A) (Gaitanos et al., 2009; Raaijmakers et al., 2009; Schmidt et al., 2012; Welburn et al., 2009). The Ska complex has also been implicated in SAC silencing (Daum et al., 2009; Hanisch et al., 2006) by recruiting PP1 phosphatase, thereby facilitating mitotic exit (Sivakumar et al., 2014; Sivakumar et al., 2016). The core of the Ska complex is W-shaped, composed of the coiled-coil dimers generated by the intertwined interactions between Ska1, Ska2, and Ska3 (Jeyaprakash et al., 2012). Each Ska subunit contributes uniquely and specifically to ensure that the complex functions as a complete unit. While the N-terminal region of Ska1 mediates the interaction between Ska2 and Ska3, its C-terminus serves as the main MT-binding component of the complex, and the Ska2/Ska3 subunits mainly constitute the scaffold of the dimerization interface (Jeyaprakash et al., 2012). [A detailed description of the winged-turn-helix (WTH) MT-binding domain of Ska1 is provided later in this review.] The C-terminus of Ska3 has also been shown to provide additional MT-binding ability (Gaitanos et al., 2009). The C-terminal domains of Ska1 and Ska3 subunits protrude symmetrically on both sides of the W-shaped central core to contact the MTs, but this interaction is possible only when the subunits they are connected at the central core (Jeyaprakash et al., 2012). These findings clearly suggest that the central core of Ska1 and Ska3 by itself is not capable of MT binding, but it serves to position the peripheral MT-binding domains of the complex in the appropriate geometry.

Figure 1: A detailed view of various kinetochore MAPs across phyla, dissociating their kinetochore and microtubule localizations and functions.

A schematic representation of several microtubule-associated proteins (MAPs) ranging from those in yeast to those in humans based on their function or classification within a family. Multiple interactions are shown to depict the localization of each MAP either directly or indirectly at both the kinetochores and MTs. Kinetochore localization specifically refers to the ability of the indicated MAPs to bind to the Ndc80 complex of the KMN network unless otherwise specified. The binding to the human Ska complex to kinetochores is thought to require the loop domain of Ndc80 (A1). Ska binds to MTs directly, while it has been reported to bind to the plus-ends with the help of the +TIP, EB1 (A2). Human Cdt1 also localizes to kinetochores with the help of the loop domain of the Ndc80 complex (B1), while it binds to MTs directly (B2). MT-binding of Cdt1 is regulated by Aurora B kinase phosphorylation (B2). Human Astrin/SKAP localizes to kinetochores but whether it binds to the Ndc80 complex directly or indirectly is not clear (C1). Human Astrin binds to MTs through its binding partner, SKAP (C2). For chTOG, MT-binding has been shown to be dependent on another MAP, TACC3, and is also thought to require Cyclin B/Cdc2 function (D2). However, it is not yet clear whether Cyclin B binds directly to the Ndc80 complex to recruit chTOG to kinetochores (D1). In the context of the XMAP215 family members, fission yeast Dis1 has been shown to be directly dependent on the loop domain for localizing to kinetochores (E1), while Alp14 localization to the loop is dependent on its binding partner, Alp7/TACC (H1). However, for budding yeast Stu2, the region of the Ndc80 complex that is required for kinetochore localization is not clear (E1). As far as MT-binding is concerned, it has been well established that all XMAP215 members, including chTOG in humans, bind directly to MTs (E2, G2 & H2). For Stu2, binding to plus-ends also depends on the yeast EB1 homolog (E2). The Dam1 ring complex in budding yeast binds to an MT directly (F2) and localizes to the kinetochore by binding to multiple Ndc80 regions, including the loop domain (F1). The proteins and protein complexes, which are color-coded identically to those presented in Fig. 2, are drawn in arbitrary shapes for the purpose of depiction only and are not to the scale. The Ska complex constitutes three subunits, Ska1, 2 and 3, but for brevity, clarity and ease of reading, the complex is represented as a singular unit (yellow ovals). Additionally, the region within Ndc80, where the MAPs bind, is not precisely marked because the exact details of the interaction interface have not been resolved in many cases.

Figure 2: A summary of the MAP functions at the kMT interface that are required for the stabilization of kMT attachment in yeast and humans.

The formation of robust kMT attachments is enabled by the MAPs recruited to kinetochores through the interaction with the Ndc80 complex. A. In humans, Cdt1 binds to Ndc80 through its loop domain. Ska also directly binds to Ndc80 through both the loop and CH domains. Both Cdt1 and Ska enhance the formation of Ndc80-mediated stable kMT attachments at the kMT interface. The C-terminus of Astrin in complex with SKAP binds to a yet unidentified region of kinetochore Ndc80, while the Astin-SKAP complex binds to a MT through the SKAP MT-binding domain. The XMAP125 homolog chTOG binds to MTs either directly or by the mediation of another MAP, TACC3. The interaction between chTOG and the Ndc80 complex is likely indirect and mediated through Cyclin B1. Cyclin B1 is recruited to kinetochores by Ndc80, but it is also not yet known whether it exhibits a direct interaction. chTOG, TACC3 and clathrin have been found to be important for recruiting each other to spindle MTs. B. In budding yeast, Dam1/DASH complexes form oligomers and/or ring structures around a MT after being recruited to kinetochores by binding to the loop and CH domains of Ndc80. The XMAP125 homolog Stu2 directly binds to kinetochore Ndc80 and the MT-tracking protein EB1, but the details of how Stu2 interacts with Ndc80 remain unclear. C. In fission yeast, the XMAP125 homolog Dis1 is loaded onto kinetochores through direct interaction with the loop domain where it interacts with EB1 located at the MT plus-end. EB1 interacts with Ndc80 in vitro, but the functional relevance of this interaction in vivo is still unclear. The other XMAP125 homolog, Alp14, forms a complex with Alp7/TACC and is recruited to kinetochores by interacting with the Ndc80 loop domain. Black arrows indicate the interaction between proteins, but the site on Ndc80 has not yet established. Red arrows indicate that the interaction between two proteins that has not yet been characterized.

Although depleting the Ska subunits does not affect the localization of Ndc80 or any other KMN component, it severely impairs kMT attachment; the Ska complex, on the other hand, absolutely requires Ndc80 to localize to the kinetochores (Fig. 1A1, Fig. 2A) (Gaitanos et al., 2009; Welburn et al., 2009). A recent study showed that the Ska complex directly binds to the Ndc80 complex and increases the affinity of the Ndc80 complex for MTs, thus enabling the formation of stable kMT attachments (Helgeson et al., 2018). As expected, both the Aurora B kinase phosphomimetic Hec1 N-terminal tail mutant and a mutant Ndc80 complex lacking the entire unstructured N-terminal 80-aa Hec1 tail formed weak attachments to MTs compared to the wild-type Ndc80 complex. However, upon the addition of the free Ska complex, the stability of the MT attachments was not only regained but was also reinforced 5-fold. This finding also suggests that the Ska complex-mediated strengthening of the MT attachments is independent of the Hec1 N-terminal tail (Helgeson et al., 2018).

Jeyaprakash et al. eloquently elaborate on the interesting parallels between the overall design and arrangement of the Ska and Ndc80 complexes(Jeyaprakash et al., 2012). The central core of both complexes is composed of coiled-coil regions decorated with peripheral domains on either end. While the Ndc80 coiled-coil has MT-binding domains at one-end and kinetochore-binding domains at the other, the Ska complex has a more symmetrical architecture in the sense that it has MT-binding domains on each end. This symmetrical nature of the Ska complex provides it with the unique and well-suited ability to contact a MT at two sites simultaneously. This flexibility is typical of many MT-binding proteins and enables unbiased diffusional motility (Cooper and Wordeman, 2009).

In general, MT tracking proteins are primarily considered for their association with the MT lattice or the polymerizing ends, but the depolymerizing ends are also potential sites of interaction. The Ska complex has been shown to associate with not only the depolymerizing MT ends but also equally efficiently with the straight MT lattice and curved MT protofilaments. On the other hand, the human Ndc80 complex binds only to the MT lattice and does not track the activity of the polymerizing or depolymerizing MT ends (Schmidt et al., 2012). Interestingly, however, the yeast Ndc80 complex has been shown to be essential for the formation of load-bearing attachments during both MT assembly and disassembly (Powers et al., 2009).

Analysis of the Ska1 sequence across various organisms shows that, similar to EB1 protein, Ska1 also contains a nearly perfect and conserved “microtubule tip localization signal” (MtLS)-SxIP motif (Honnappa et al., 2009), “SHLP”, at its N terminus (Thomas et al., 2016). Interestingly, the analysis of the Cdt1 sequence did not reveal the presence of any such SxIP or related motif(s). The C-terminal coiled-coil EB1 domain has been shown to interact with numerous +TIPs through their conserved SxIP motifs. Interestingly, Ska1 recruitment to the plus-ends and its interaction with EB1 were shown to involve the same C-terminal EB1 region that interacts with the SxIP motifs of other +TIPs (Fig. 1A2) (Thomas et al., 2016). Several plus-end tip-tracking proteins that associate with elongating MTs through the end-binding (EB) family of proteins can simply be considered as “hitchhiker” molecules and not “autonomous” tip trackers. However, in vitro analyses indicate that the Ska1 complex is capable of tracking polymerizing MTs independent of EB1 (Monda et al., 2017). Monda et al. observed bright GFP-Ska fluorescence at the growing MT ends (Fig. 1A2). In addition, when MT depolymerization was induced by the removal of soluble tubulin, the Ska1 complex was also enriched at the shortening MT ends (Monda et al., 2017). Thus, the Ska complex exhibits the remarkable ability to remain associated with both the polymerizing and depolymerizing MT plus-ends, a property that has thus far been reported only for one other human kinetochore MAP, the CENP-E motor (Gudimchuk et al., 2013).

(b). Cdt1

The Cdt1 (Cdc10-dependent transcript 1) protein primarily functions in the licensing of the origins during DNA replication in the G1 phase of the cell cycle. Therein, Cdt1 serves as a “courier” that delivers and loads soluble minichromosome maintenance (MCM) complexes to the DNA-bound origin recognition complex (ORC)/cell division cycle 6(Cdc6) complex, marking the completion of the presumptive prereplication complex (pre-RC) (Pozo and Cook, 2016). Interestingly, our previously published work demonstrated a second essential role of Cdt1 during the M-phase of the cell cycle, wherein Cdt1 was shown to localize to kinetochores through an interaction with the Hec1 component of the Ndc80 complex (Varma et al., 2012). Two-hybrid screening in yeast and in vivo GST pull-down assays using asynchronous HeLa cell lysates confirmed the direct interaction between Cdt1 and Hec1 (Varma et al., 2012), which was further substantiated in our recent study using purified proteins in a blot overlay assay (Agarwal et al., 2018).

High-resolution microscopy data indicated that the average separation, Delta, between the two ends of the Ndc80 complex was considerably lower in Cdt1-depleted cells as compared to the control metaphase cells in which kinetochores attain a full MT complement (Varma et al., 2012). Although these data suggested that Cdt1 binding to the Ndc80 loop domain provided an additional attachment site, which was essential for an extended Ndc80 configuration and stable kMT attachment, the exact mechanism of how Cdt1 functioned to generate the observed phenotypes has not been delineated. Our recent study explicitly revealed the mechanistic details of how Cdt1 contributes to the stabilization of the kMT attachments (Agarwal et al., 2018). In vitro and in vivo studies show that Cdt1 directly binds to MTs through an extended segment comprising its middle and C-terminal regions (Fig. 1B1, Fig. 2A). This segment, while largely unstructured, consists of two winged-turn-helix (WTH) domains, one in the middle and another at the C-terminus, that are predicted to be involved in MT-binding. On the other hand, the N-terminal region of Cdt1 binds to the Ndc80 complex (Agarwal et al., 2018; Varma et al., 2012), thus serving to bridge the loop domain of the complex to the spindle MTs. The 2^nd MT-binding interface of the Ndc80 complex, which is mediated by the loop domain and Cdt1, likely contributes to the more robust kMT attachments formed by the complex (Fig. 1B2, Fig. 2A). Similar to the Ndc80 complex, Aurora B kinase negatively regulates the MT-binding activity of Cdt1 (Fig. 1B2). Functional rescue experiments using Aurora B phosphomimetic mutants of Cdt1 in vivo showed the generation of severely defective KMT attachments and impaired normal mitotic progression (Agarwal et al., 2018), mirroring the Cdt1 loss-of-function phenotype in mitotic cells (Varma et al., 2012).

The winged-turn-helix (WTH) domain, the predicted MT-binding domain of Cdt1 and Ska1, is one of the core components of the transcription machinery and has the ability to bind DNA and to mediate protein-protein interactions (Gajiwala and Burley, 2000; Teichmann et al., 2012). The canonical WTH domain (~110 aa) typically has a compact α/β structure consisting of two wings or loops (W1 and W2), three α helices (H1, H2, and H3) and three β-sheets (S1, S2, and S3) arranged as H1-S1-H2-H3-S2-W1-S3-W2 (Gajiwala and Burley, 2000). The kinetochore-localized Ska1 protein was the first in which the WTH domain was shown to function as a MT-binding domain (Abad et al., 2014; Jeyaprakash et al., 2012). The WTH domain of Ska1 differs from that of a canonical WTH domain in that it has two extra elements [module I comprising two α helices (α1 and α2) at the N-terminal and module II comprising three α helices (α4, α5 and α6) in the middle] in addition to the core elements (Abad et al., 2014). The Cdt1 WTH domains, on the other hand, are more similar to the canonical WTH domains except that the C-terminal WTH domain has an extra α helix in addition to the core elements (Agarwal et al., 2018). It is interesting that DNA binding motifs can be shared and used as MT-binding motifs because both the DNA and MTs are negatively charged moieties and thus are recognized by proteins via electrostatic interactions. These domains thus represent a contemporary class of unconventional MAP motifs that can impart MT-binding function to MAPs.

The present understanding is that, in the G1 phase, Cdt1 associates only indirectly with the nuclear chromatin through the origin recognition complex (ORC) and coordinates with Cdc6 to load the MCM2-7 helicase complex onto the chromatin to license DNA replication (Symeonidou et al., 2012; Tsakraklides and Bell, 2010). However, one study has shown that an anti-Cdt1 antibody efficiently immunoprecipitated DNA sequences, thus suggesting the possibility of direct interaction between Cdt1 and DNA (Sugimoto et al., 2011). At this point, we can only surmise that this direct interaction is likely mediated by the WTH domains of Cdt1; however, without any experimental validation, the finer details of Cdt1 DNA-binding remain unclear. Additionally, since Ska is an outer kinetochore protein that has not been reported to localize to the nucleus in interphase, in contrast to Cdt1, it is expected that the WTH domain of Ska exclusively bind MTs. To the best of our knowledge, there is no available evidence showing that DNA-binding proteins are also capable of binding MTs. In summary, there is limited evidence currently showing that the “conventional” DNA-binding WTH and the “unconventional” MT-binding WTH domains can be used by the same protein to bind to both DNA and MTs interchangeably.

Even though both Cdt1 (Agarwal et al., 2018) and Ska (Schmidt et al., 2012) demonstrated lower speed and bidirectional diffusive motion when interacting with MTs, this characteristic was substantially different from the directed, high-speed motility exhibited by motor MAPs such as kinesins. It is also fascinating that even though Cdt1 and Ska do not share any substantial sequence or structural similarities, they both have functional similarities with respect to binding to the Ndc80 loop for kinetochore recruitment and interacting with MTs. It is still unclear which regions these proteins use to dock to the loop and whether one is required for the assembly of other proteins. For the recruitment of the Ska complex to the kinetochores, although Ndc80 has been recognized as a receptor (Chan et al., 2012; Gaitanos et al., 2009; Raaijmakers et al., 2009; Welburn et al., 2009), the specific involvement of the loop domain within the Ndc80 complex has been clearly demonstrated (Zhang et al., 2012; Zhang et al., 2017). In support of the above findings, Cdk1-mediated phosphorylation of Ska3 has been shown to promote its association with the Ndc80 loop domain. Concomitantly, the Ska3 mutants deficient in Cdk1 phosphorylation failed to localize to kinetochores but retained their MT localization (Zhang et al., 2017). Contrary to these studies, fairly recently, an alternate mechanism for Ska recruitment to kinetochores was demonstrated. via the N-terminal unstructured tail of Ndc80 (Janczyk et al., 2017). Another study demonstrated that the direct interaction of the Ska complex (Ska1 subunit) with the MT plus-end binding EB1 protein was responsible for the localization of Ska1 to the kMT interface (Thomas et al., 2016). At this point, it suffices to say that, despite several studies, the architecture of the interaction interface between the Ndc80 and Ska complexes and how these components assemble at the kMT interface remain poorly defined. The most recent contribution in this context is provided by a study from Helgeson et al., who used cross-linking and mass spectrometry of the Ska complex, Ndc80 complex, and Taxol-stabilized MTs to find that the unstructured Ska3 C terminal region (residues 102–412) cross-linked robustly with the Ndc80 complex in the regions that are predicted to form coiled-coils (Helgeson et al., 2018). Further mapping revealed that few other interactions were also distributed among the 4 subunits of the Ndc80 complex; however, some cross-links were observed with the CH domains of Hec1 and Nuf2 and the RWD domains of Spc24 and Spc25 (Helgeson et al., 2018). In summary, the data from these studies suggest that multiple Ndc80-dependent mechanisms may function in parallel to recruit Ska to kinetochores.

Another interesting question worthy of deliberation is whether the kinetochore binding ability of the MAPs such as Ska and Cdt1 influence their ability to interact with a specific region of spindle MTs, such as the plus-ends. Interestingly, it has been shown that the Ska complex is able to localize to kinetochores only after kMT attachments have been formed (Chan et al., 2012; Hanisch et al., 2006). However, no such correlation has yet been discerned for Cdt1 since Cdt1 was able to localize normally to kinetochores (in both normal and nocodazole-treated cells) as well as to spindle MTs (Agarwal et al., 2018; Varma et al., 2012). In fact, the staining of Cdt1 was more evident at the spindle poles compared to the rest of the spindle, which could be due to the high density of MTs at that location. Similar to Cdt1, myc-tagged Ska1 has been shown to localize to kinetochores as well as colocalize partially with spindle MTs and the spindle poles in mitotic cells (Hanisch et al., 2006). These observations suggest that the localization of Ska or Cdt1 to kinetochores is not required for their MT-binding. Since Cdt1 localization to kinetochores and to MTs seemingly independent events, the binding of Cdt1 to one site does not impede its localization to the other site. In fact, we feel that this unique ability of Cdt1 to localize to both the kinetochores and spindle MTs might be crucial for its function in robust kMT attachment formation.

(c). The Dam1/DASH complex

In the budding yeast Saccharomyces cerevisiae, the heterodecameric Dam1 or DASH complex couples kinetochores to MT ends (Tanaka and Desai, 2008). Interestingly, while the Dam1/DASH complex is essential for the survival of budding yeast, it is dispensable in fission yeast (Joglekar et al., 2010). This ten-protein complex contains Dam1p, the main MT-binding component(Hofmann et al., 1998) and nine other essential proteins: Duo1p, Dad1p, Dad2p, Spc19p, Spc34p, Ask1p, Dad3p, Dad4p and Hsk3p (McAinsh et al., 2003). Mutations in DASH subunits resulted in unstable or broken spindles that consequently led to weakened kMT attachments (Cheeseman et al., 2001). Purified Dam1/DASH complexes have been shown to coalesce into stable ring-like structures or open spirals around MTs in vitro, enabling them to slide passively along the MT lattice (Fig. 1F2, Fig. 2B) (Westermann et al., 2005; Westermann et al., 2006). This unique property of the Dam1/DASH complex to form rings around the MT lattice and its preferential binding to GTP-tubulin makes it distinct from a number of the other MAPs studied thus far. The Dam1 complex, but not the Ndc80 complex, was shown to possess an intrinsic plus-end–tracking ability. However, Dam1 was shown to mediate the continuous Ndc80 tip association with dynamic MTs, suggesting that, in yeast, Dam1 complexes are the primary sources of kinetochore plus-end–tracking activity, whereas Ndc80 complexes structurally bridge MT ends with chromosomes (Lampert et al., 2010). Moreover, the Ndc80 complex is absolutely essential for the recruitment and assembly of the Dam1 complex to the kinetochore (Fig. 1F1, Fig. 2B) (Janke et al., 2001). The Ndc80 loop region has been shown to be essential for the Ndc80-Dam1 interaction and kinetochore loading of the Dam1 complex (Maure et al., 2011). Evidence for an interaction between these two complexes has been provided through localization and two-hybrid studies (Joglekar et al., 2009; Shang et al., 2003). Phosphorylation of Dam1 at Ser20 by the yeast Aurora B kinase homolog Ipl1 has been shown to reduce its affinity for the MT lattice (Gestaut et al., 2008). Further, phosphorylation of the components of the Dam complex (Dam1p, Ask1p, and Spc34p) has been shown to negatively impact its interaction with the Ndc80 complex (Shang et al., 2003; Tien et al., 2010); however, the specific phosphoprotein component of the Dam complex that contributes to this disrupted interaction is still unclear. Recently, an interesting study by Kim et al. revealed that the Ndc80 complex can simultaneously bind and bridge two Dam1 complex rings, not one (Kim et al., 2017), as was presumed before (Joglekar et al., 2009; Tanaka et al., 2007). Each component within this tripartite interaction was shown to be regulated by Aurora B kinase. Further, mutating any of the Dam1-Ndc80 interacting regions to generate defective Ndc80 rods that can bind to only one Dam ring in vitro resulted in erroneous chromosomal segregation and weakened kMT attachments in vivo (Kim et al., 2017). The study was extended to show that the specific distance between the two rings, bridged by the extended Ndc80 complex, is also crucial for accurate cell division (Kim et al., 2017). Interestingly, based on its functional similarities and dependence on the Ndc80 complex for kinetochore recruitment, the Ska complex has been cited as a functional equivalent of the Dam1/DASH complex in humans.

(d). XMAP215 family

The XMAP215 family of proteins was first identified as a group comprising MAPs that regulate the dynamic properties of MTs (Mack and Compton, 2001). The XMAP215 family orthologs include colonic hepatic tumor overexpressed gene (chTOG) in humans (Charrasse et al., 1998), XMAP215 (the most studied member of the family) in Xenopus, Stu2 in budding yeast, Dis1/Alp14 in fission yeast, Zyg9 in worms, and Mini spindles (MSPS) in Drosophila (Al-Bassam and Chang, 2011). These proteins are localized at MT plus-ends, MT-organizing centers, kinetochores and along MT lattices. A characteristic feature of this family of proteins is a repeating N-terminal structure and a more-or-less conserved C-terminal nonrepeating domain (Hsu and Toda, 2011; Lee et al., 2001; Li et al., 2011; Spittle et al., 2000; van der Vaart et al., 2011). Each N-terminal repeat region comprises ∼200 aa residues (TOG domains) and consists of as many as five HEAT motifs. The HEAT motifs are thought to be the protein-protein interaction domains that allow these proteins to interact with other proteins and, by virtue of these interactions, enable defined localization during specific stages, thereby modulating MT dynamics in a spatially and temporally controlled manner. The C-terminal nonrepeat regions are likely responsible for MT binding and centrosome, spindle pole body (SPB) and kinetochore localization. The number of TOG domains varies from two to five depending on the model organism. Humans, Xenopus and Drosophila have five TOG domains, worms have three and yeasts have one or two TOG domains. The TOG domain arrays in the XMAP215 family of proteins bind to free tubulin dimers and promote their stable incorporation onto MT plus-ends, while the basic linker regions separating the TOG domains promote the association of the protein to the MT lattice (Ayaz et al., 2014; Ayaz et al., 2012; Currie et al., 2011; Fox et al., 2014; Widlund et al., 2011). Thus, the TOG array usually promotes MT polymerization and is required to generate mitotic spindles of appropriate spindle length in the different model systems studied, including Drosophila and Xenopus (Brittle and Ohkura, 2005; Currie et al., 2011; Reber et al., 2013).

At centrosomes, XMAP215 orthologs function to stabilize MTs and contribute to the formation of proper centrosomal asters (Cassimeris and Morabito, 2004; Cullen et al., 1999; Gergely et al., 2003; Popov et al., 2002). XMAP215 family proteins generally function as MT polymerases at MT plus-ends to accelerate MT assembly by promoting growth and suppressing catastrophe (Al-Bassam et al., 2012; Al-Bassam et al., 2006; Brouhard et al., 2008; Podolski et al., 2014; Widlund et al., 2011). TIRF microscopy has shown that the budding yeast Stu2 binds preferentially to MT plus-ends in vitro and that the fission yeast Dis1 weakly associates with the MT lattice and accumulates specifically at the growing MT plus-ends (Fig. 1E2, Fig. 2B & 2C). In contrast, human chTOG accumulates at both the growing and shrinking MT plus-ends, and it is thought that MT polymerase activity promotes MT assembly at both growing and shrinking (also called rescue) MT tips (Fig. 1D2, Fig. 2A) (Matsuo et al., 2016; Podolski et al., 2014; Roostalu et al., 2015; van Breugel et al., 2003). Budding yeast Stu2 and Xenopus XMAP215 have also been reported as MT-destabilizing factors at the plus-ends, possibly by inducing catastrophe (Shirasu-Hiza et al., 2003; van Breugel et al., 2003). To date, human chTOG has not been reported to play a role in MT destabilization but rather in spindle pole organization and spindle MT assembly (Cassimeris et al., 2009; Gergely et al., 2003). A study using TIRF microscopy has also shown that human chTOG functions in vitro as a MT polymerase and acts synergistically with another MT binding protein, TPX2, for nucleating MTs (Roostalu et al., 2015).

At kinetochores, the XMAP215 orthologs are required for the regulation of the kMT attachments (Garcia et al., 2002; Tanaka et al., 2005). Stu2 is required for MTs to form from kinetochores, thereby facilitating the attachment of kinetochores to spindle MTs in budding yeast (Fig. 1E1 & E2, Fig. 2B) (Kitamura et al., 2010). A recent study has shown that Stu2 interacts directly with the kinetochore Ndc80 complex to control the stability of the Ndc80-mediated kMT attachments at the dynamic MT tips (Fig. 1E2); however, the mode of their interaction is still unclear (Miller et al., 2016). It was observed that Stu2 can either stabilize or destabilize Ndc80-mediated kMT attachments, depending on the assembly or disassembly status of the MT tips. These activities of Stu2 are defined by the level of tension imparted on the kinetochores (Miller et al., 2016). These studies strongly support the phenotypes observed (chromosome alignment defects and unattached kinetochores) in cells lacking Stu2 or human chTOG (Gandhi et al., 2011; Gillett et al., 2004; Kitamura et al., 2010; Kosco et al., 2001; Marco et al., 2013; Severin et al., 2001).

In fission yeast, the XMAP215 homolog Dis1 has been reported to associate with the loop domain of Ndc80, where it contribute to kMT attachments (Fig. 1G1, Fig. 2C) (Hsu and Toda, 2011; Tang et al., 2013). Also, in fission yeast, another XMAP215 homolog, Alp14, forms a stable complex with Alp7/TACC (transforming acidic coiled-coil) family of proteins, another MAP that binds to mitotic spindle MTs (Fig. 1H1, Fig. 2C) (Sato and Toda, 2007; Sato et al., 2004). Both Alp14 and Alp7 are recruited interdependently to spindle pole bodies and spindle MTs (Fig. 1H2). They are also recruited to mitotic kinetochores through the loop domain of Ndc80, where they have been reported to promote kMT attachment (Fig. 1H2) (Garcia et al., 2001; Tang et al., 2013; Tang and Toda, 2015). In humans, chTOG and the Alp7 homolog TACC3 interact with each other to form a chTOG-TACC3 complex, which is recruited to centrosomes and proximal mitotic spindles as mediated by the Aurora A kinase-dependent phosphorylation of TACC3 (Fig. 1D2, Fig. 2A) (Thakur et al., 2014). The chTOG-TACC3 complex associates with Clathrin, which is localized to kinetochore fibers and the mitotic spindle during mitosis, for its recruitment to spindle MTs, and acts as an inter-MT bridge to stabilize kinetochore fibers (Fig. 1D2, Fig. 2A) (Booth et al., 2011; Compton, 2000; Royle, 2012). The components of the chTOG-TACC3-Clathrin complex are dependent on each other for recruitment to spindle MTs (Booth et al., 2011). The mechanism of how Clathrin, together with chTOG and TACC3, regulates the stability of kinetochore fibers is still unknown, and further studies are necessary to address whether this function is related to kinetochore proteins, including the Ndc80 complex.

The XMAP215 ortholog in Xenopus and yeast, XMAP215/Dis1/Stu2 interacts with the MT plus-end tracking protein EB1 (Kronja et al., 2009; Matsuo et al., 2016; Wolyniak et al., 2006) to regulate MT dynamics at the kMT interface (Fig. 1E1 & 1E2, Fig. 1G1 & 1G2, Fig. 2B & 2C). An in vitro study using the purified Ndc80 complex from fission yeast showed that the complex interacted with EB1 to track the growing MT plus-ends, but no detectable binding was observed between the complex and Dis1 (Matsuo et al., 2017). The studies described thus far suggest the possibility that the Ndc80 complex could bind either directly to Dis1 or indirectly to Alp14 via the mediation of Alp7 to contribute to the establishment of a dynamic kMT interface. It is noteworthy that a direct interaction between Ndc80 and chTOG has not yet been observed in humans.

The analysis of the MTs isolated from Xenopus egg extracts demonstrated an association between XMAP215 and Cyclin B1. A coimmunoprecipitation assay using the MT fraction from Xenopus egg extract and HeLa cells determined that XMAP215 and its human ortholog chTOG interact with cyclin B1. An in vitro cosedimentation assay showed that cyclin B1 binds to MTs and that its MT-binding is enhanced by XMAP215/chTOG (Fig. 1D2) (Charrasse et al., 2000). Cyclin B1 is recruited to unattached kinetochores via Ndc80, and it has been reported to be required for proper kMT attachments (Fig. 1D1, Fig. 2A); however, it is not clear how Cyclin B1 contributes to this function (Bentley et al., 2007; Chen et al., 2008). These observations point to the possibility that chTOG might play a role in facilitating Ndc80-mediated kMT attachment indirectly through cyclin B1.

In humans, chTOG is predicted to play a role in kMT attachment, as observed for the yeast XMAP215 orthologs (Fig. 1D1 & 1D2, Fig. 2A) (Gergely et al., 2003; Meraldi et al., 2004; Miller et al., 2016). In contrast, another study has suggested that this might not be the case as chTOG-depleted cells have stable kMT attachments under cold conditions or in the presence of high calcium levels (Cassimeris et al., 2009) Thus, further studies aimed at addressing outstanding questions, such as whether human chTOG interacts with Ndc80 directly or indirectly (i.e., through other factors) and how the activities of chTOG and Ndc80 are coordinated, are required to shed light on the function of chTOG in kMT attachments in vertebrates.

(e). Astrin and SKAP

Astrin is a spindle-associated MAP (Chang et al., 2001; Mack and Compton, 2001) that localizes to spindle MTs, spindle poles and outer kinetochores. Astrin is instrumental in normal mitotic progression and the maintenance of chromatid cohesion and centrosome integrity (Dunsch et al., 2011; Gruber et al., 2002; Thein et al., 2007). Astrin forms a complex with a +TIP, CLASP1, to promote kMT stability and chromosome alignment during metaphase (Manning et al., 2010). Small kinetochore-associated protein (SKAP), identified by proteome analysis of vertebrate kinetochores, has also been shown to localize to spindle MTs and spindle poles through its N-terminal region and to outer kinetochores through its C-terminal region (Fig. 1C1, Fig. 2A) (Dunsch et al., 2011; Schmidt et al., 2010; Wang et al., 2012). Experiments conducted by two independent groups, using immunoprecipitation followed by mass spectrometry, demonstrated that Astrin interacts with SKAP to form the Astrin-SKAP complex (Dunsch et al., 2011; Friese et al., 2016; Kern et al., 2017; Schmidt et al., 2010). The targeting of the Astrin-SKAP complex to the bioriented kinetochores depends on the KMN network, Aurora B kinase activity and MT motor CENP-E (Huang et al., 2012; Manning et al., 2010; Schmidt et al., 2010; Wang et al., 2012). Results from a cosedimentation assay showed that Astrin and SKAP bind directly to individual MTs as well as to a complex (Fig. 1C2) (Schmidt et al., 2010). While the N-terminal region of Astrin binds to MTs through its interaction with SKAP, the C-terminal region is important for its kinetochore recruitment, the mechanism of which is unclear (Dunsch et al., 2011; Kern et al., 2017). In cells, the Astrin N-terminus (aa 1–693) was shown to localize to MTs and weakly to kinetochores. In contrast, the Astrin C-terminus (aa 694–1193), which lacks the SKAP binding site, localized to kinetochores, but not MTs (Kern et al., 2017), indicating that, similar to that of Ska1, the MT-binding and kinetochore binding domains of the Astrin-SKAP complex are independent and separable (Kern et al., 2017). It has been reported that the MT-binding domain in SKAP stimulates the growth rate of the MTs, possibly through direct interaction with tubulin, suggesting that the Astrin-SKAP complex has MT polymerase activity (Friese et al., 2016). This complex localizes to MT plus-ends where it interacts with both EB1 and EB3 through the SxIP motif (Akhmanova and Steinmetz, 2010; Honnappa et al., 2009) to form a stable MT plus-end tracking complex (Tamura et al., 2015). TIRF microscopy shows that SKAP binds both to growing MT plus-ends in an EB1-dependent manner and to the MT lattice in vitro (Friese et al., 2016; Wang et al., 2012). As previously mentioned, Astrin-SKAP may also stabilize MT plus-end dynamics at kinetochores through its interaction with CLASP1 (Manning et al., 2010). These studies suggest that this complex is required to form robust kinetochore fibers and is a promoter of stable kMT attachment (Dunsch et al., 2011; Thein et al., 2007).

Previous studies have shown that the Astrin-SKAP complex interacts with dynein light chain DYNLL1, a component of the MT motor protein complex cytoplasmic dynein. Thus, this complex possibly promotes dynein-mediated assembly, enhances the stability of the MT array, and advances the movement of the kinetochores from the MT plus-ends towards the spindle poles (Dunsch et al., 2011; Sharp et al., 2000; Wittmann et al., 2001). This complex has also been shown to interact with the nonmotor NuMA (nuclear mitotic apparatus) protein (Chu et al., 2016), another dynein binding partner, possibly to promote NuMA-mediated bundling of spindle MTs at centrosomes and contribute to spindle assembly (Haren et al., 2009; Merdes et al., 1996; Saredi et al., 1996). However, whether the interaction between Astin-SKAP and dynein contributes to the formation of robust kMT attachment remains unresolved.

How the Astrin-SKAP complex coordinates in Ndc80-mediated kMT attachment was unclear until a recent study was carried out in vitro using a biochemically reconstituted Astrin-SKAP complex (Kern et al., 2017). In this study, kinetochore cross-linking immunoprecipitation and mass spectrometry were used to show that the Astrin-SKAP complex interacts with the Ndc80 complex in humans (Fig. 1C1, Fig. 2A). However, the details of this interaction are still not clear. Although the Astrin-SKAP complex docks to an unknown kinetochore location through the C-terminal region of Astrin, the complex binds to Ndc80 through the N-terminal region of Astrin and to MTs through the SKAP MT-binding domain. However, unlike EB1, the MT-binding domain of SKAP lacks the SxIP motif that was found to promote the plus-end tracking and MT polymerization activity of EB1 (Friese et al., 2016). Further, the Astrin-SKAP complex binds to MTs synergistically with the Ndc80 complex to form an integrated and dynamic kMT surface that stabilizes the kMT attachments. The Astrin-SKAP complex displayed enhanced binding to MTs in the presence of a truncated version of the Ndc80 complex, (Ndc80 Bonsai), as was evident by the significant reduction in the apparent K_d of the Astrin-SKAP complex for MTs in the presence of Ndc80 Bonsai (1.4 μM) compared to that of Astrin-SKAP alone (~3.2 μM) (Kern et al., 2017). Taken together, these findings suggest that the Astrin-SKAP complex may facilitate processive interactions between the outer kinetochore and MT plus-ends, as observed for other kinetochore-localized MT binding proteins, including the Ska complex (or the Dam1 complex in yeast), and chTOG (Gergely et al., 2003; Lampert et al., 2010; Meraldi et al., 2004; Miller et al., 2016; Schmidt et al., 2012; Welburn et al., 2009). It was previously reported that SKAP physically interacts with the Mis13/Dsn1 subunit of the kinetochore Mis12 complex, a component of the KMN network, to serve as a link to dynamic spindle MTs (Wang et al., 2012).

5. Conclusions

All the aforementioned kinetochore proteins that function as MAPs, including Cdt1, the Ska complex, chTOG and the Astrin-SKAP complex, coordinate with the outer kinetochore complex Ndc80 to reshape the major MT binding interface of the human metaphase kinetochores. In this manuscript, we present a detailed overview of the functional relationship of these kinetochore MAPs with the Ndc80 complex and their targeting to the kMT interface in diverse model systems, including yeast and humans (Figs. 1 & 2, see details in legends). Briefly, in yeast, the stabilization of the kMT attachments by the Ndc80 complex is promoted by XMAP125 homologs (Stu2 in budding yeast and Dis1 and Alp14 in fission yeast, Fig. 1E1 & 1E2, Fig. 1G1 & 1G2, Fig. 2B & 2C) and the Dam/DASH complex (in budding yeast, Fig. 1F1 & 1F2, Fig. 2B). The Dam/DASH complex plays a role in kMT attachment by interacting with the loop domain and the unstructured tail of Ndc80 (Fig. 1F1, Fig. 2B). Alp7/Alp14 and Dis1 in fission yeast also promote kMT stabilization by interacting with the loop domain of Ndc80 (Fig. 1H1, Fig. 2C). In mammals, the Ndc80-mediated kMT attachment is enhanced by the Ska complex and Cdt1 in conjunction with the loop domain of Ndc80 (Fig. 1A1 & 1A2, Fig. 1B1 & 1B2, Fig. 2A). Ska also contributes to kMT stabilization by interacting with the unstructured N-terminal tail of Ndc80 (Fig. 1A1, Fig. 2A). In contrast, it is still unclear how human chTOG and the Astrin-SKAP complex associate with the Ndc80 complex to enhance the stability of the kMT attachments. Moreover, how these different MAPs coordinate with each other and with the Ndc80 complex, the master regulator, is an arena for further research that will most likely unveil several layers of mechanistic aspects related to the formation, maturation and stabilization of the kMT attachments during mitosis. Further, the details of whether these different kinetochore MAPs coordinate to stabilize kMT attachments or act independently of each other remain elusive. In addition, other MT-binding kinetochore components, such as the kinetochore motors CENP-E and dynein, and the MT plus-end tracking proteins, including EB1, CLIP-170, and CLASP1/2, also contribute to the formation of stable kMT attachments (Akhmanova et al., 2001; Brunner and Nurse, 2000; Cheeseman and Desai, 2008; Nakamura et al., 2001; Santaguida and Musacchio, 2009; Schuyler and Pellman, 2001). Studies unraveling whether these kinetochore proteins contribute to kMT attachment in coordination with the Ndc80 complex or the kinetochore MAPs will facilitate a detailed and comprehensive understanding of how kinetochores are stably coupled to the ends of dynamic MTs to prevent chromosome missegregation.

Abbreviations:

MT(s)

Microtubule(s)

kMT

Kinetochore-microtubule

MAP(s)

Microtubule-associated proteins

WTH

Winged-turn-helix

CH

Calponin homology