Mechanistic basis for Sgo1-mediated centromere localization and function of the CPC

Abad and Gupta et al. reveal the molecular basis for the interaction between the CPC and Sgo1, two essential regulators of chromosome segregation. Their work provides a rationale for the kinetochore–proximal centromere recruitment of the CPC and highlights its requirement for high-fidelity chromosome segregation.


Introduction
Equal and identical segregation of chromosomes to daughter cells during mitosis requires physical attachment of duplicated sister chromatids (via their kinetochores) to microtubules emanating from opposite spindle poles and subsequent alignment of chromosomes at the metaphase plate, a state known as biorientation (Musacchio and Desai, 2017). Chromosome biorientation is achieved and monitored by several processes including sister chromatid cohesion and quality control mechanisms known as error correction and spindle assembly checkpoint (SAC), all controlled by the spatiotemporal regulation of kinases and phosphatases (Funabiki and Wynne, 2013;Gelens et al., 2018;Saurin, 2018).
Error correction is a mechanism that destabilizes incorrect kinetochore-microtubule (KT-MT) attachments, such as syntelic (two sister kinetochores attached to microtubules from the same spindle pole) or merotelic (a single kinetochore attached to microtubules emanating from both spindle poles) attachments and stabilizes correct bipolar attachments. The chromosomal passenger complex (CPC), consisting of Aurora B kinase, inner centromere protein (INCENP), Borealin, and Survivin, is one of the key players regulating this process (Carmena et al., 2012). The CPC, via its Aurora B enzymatic core, destabilizes aberrant KT-MT attachments by phosphorylating outer kinetochore substrates such as the Knl1 complex/Mis12 complex/Ndc80 complex network so that new attachments can be formed (Cheeseman et al., 2006;Cimini et al., 2006;DeLuca et al., 2006;Lampson et al., 2004;Welburn et al., 2010). Sgo1 has also been shown to regulate KT-MT attachments via PP2A-B56 recruitment that balances Aurora B activity at the centromeres (Meppelink et al., 2015). In addition to error correction, the CPC is also involved in the regulation of the SAC, a surveillance mechanism that prevents anaphase onset until all kinetochores are attached to microtubules (Foley and Kapoor, 2013;Musacchio, 2015).
During (pro)metaphase, the CPC predominantly localizes in the centromeric region between the sister kinetochores, and multiple independent studies recently suggested that the evolutionary conserved Haspin and Bub1 kinases can recruit independent pools of the CPC along the interkinetochore axis. Both recruitment pathways appear redundant for KT-MT error correction and can support faithful chromosome segregation (Bekier et al., 2015;Broad et al., 2020;Hadders et al., 2020;Liang et al., 2020). Haspin mediates phosphorylation on histone H3 Thr3 (H3T3ph), which is recognized by the BIR domain (baculovirus inhibitor of apoptosis repeat domain) of Survivin (Dai et al., 2005;Du et al., 2012;Jeyaprakash et al., 2011;Kelly et al., 2010;Niedzialkowska et al., 2012;Serena et al., 2020;Wang et al., 2010;Yamagishi et al., 2010). Bub1 phosphorylates Thr120 of Histone H2A (H2AT120ph) that is recognized by Sgo1, which in turn is suggested to interact with Borealin via its coiled-coil domain (Bonner et al., 2020;Kawashima et al., 2007;Kawashima et al., 2010;Liu et al., 2015;Tsukahara et al., 2010;Yamagishi et al., 2010). However, our earlier work showed that the histone H3-like Sgo1 N-terminal tail can also interact with the Survivin BIR domain using a binding mode that is nearly identical to that of the histone H3 tail phosphorylated at Thr3 (Jeyaprakash et al., 2011). This suggests that a direct interaction between Survivin and Sgo1 is possible. H3T3ph and H2AT120ph appear to localize to distinct regions within the mitotic centromeres, with H3T3ph localizing to the inner centromere and H2AT120ph to the KT-proximal centromere (Broad et al., 2020;Hadders et al., 2020;Liang et al., 2020;Liu et al., 2013a;Yamagishi et al., 2010). While Sgo1 is known to play a role in the recruitment of the CPC to centromeres, the structural and molecular basis for how the CPC and Sgo1 interact and how these interactions contribute to the localization and function of the specific CPC pools remain unclear. Here, we address these questions by combining biochemical, structural, biophysical, and cellular approaches.

Sgo1 makes multipartite interactions with CPC subunits
Previous studies have suggested that the Sgo1-CPC interaction is mediated via the N-terminal coiled-coil of Sgo1 and Borealin (Bonner et al., 2020;Tsukahara et al., 2010). However, our structural data revealed that the very N-terminus of Sgo1 can interact with the BIR domain of Survivin (Jeyaprakash et al., 2011). Together, these studies suggest that multipartite interactions between Sgo1 and different CPC subunits could facilitate CPC-Sgo1 complex formation. To gain further structural insights, we performed chemical cross-linking of the CPC ISB10-280 -Sgo1 1-415 complex using a zero-length cross-linker, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), followed by mass spectrometry analysis (Fig. S1 E). Cross-linking-mass spectrometry (CLMS) data showed that (1) consistent with our previous observations (Jeyaprakash et al., 2011), the N-terminal region of Sgo1 (amino acids 1-34) makes extensive contacts with Survivin BIR domain (amino acids 18-89); (2) the N-terminal coiled-coil of Sgo1 (amino acids 10-120) interacts with the CPC triple helical bundle; (3) consistent with previous findings (Bonner et al., 2020), the N-terminal coiled-coil also contacts the Borealin dimerization domain; and (4) the Sgo1 region beyond the N-terminal coiled-coil region, which is predicted to be unstructured, contacts both Survivin and Borealin, with most contacts confined to the Sgo1 central region spanning amino acids 180-300 (Fig. 2, A and B). Thus, our cross-linking results suggest that Sgo1 interacts with the CPC mainly via two regions, the N-terminal coiled-coil domain and the unstructured central region (Fig. 2, A and B).
We further analyzed the contribution of different Sgo1 regions for CPC binding using a LacO-LacI tethering assay. For this, we made use of U-2 osteosarcoma (OS) cells harboring a LacO array on the short arm of chromosome 1, to which we could recruit Sgo1 fragments as LacI-GFP fusions (U-2 OS-LacO cells; Janicki et al., 2004). To exclude any contribution from H3T3ph on CPC recruitment, we made use of a Haspin CRISPR mutant (CM) cell line that displays no discernible Haspin activity (Hadders et al., 2020). Constructs containing Sgo1 1-130 and fulllength Sgo1 1-527 recruited endogenous Aurora B (Fig. 2 C) and Borealin ( Fig. 2 D) to the LacO foci, at comparable levels. This is in line with previous data that suggested the Sgo1 N-terminal region as a major CPC binding site (Bonner et al., 2020;Jeyaprakash et al., 2011;Tsukahara et al., 2010). Surprisingly, Sgo1 130-280 fused to LacI-GFP was also able to recruit endogenous Aurora B (Fig. 2 C) and Borealin (Fig. 2 D) to the LacO foci, although at lower levels compared with Sgo1 1-130 and Sgo1 1-527 . In contrast, the Sgo1 fragments Sgo1 274-415 and Sgo1 415-527 (Sgo1 274-415 -LacI-GFP and Sgo1 415-527 -LacI-GFP) failed to recruit either Aurora B or Borealin (Fig. 2,C and D). Taken together, the LacO-LacI tethering data confirm that the main CPC-interacting regions of Sgo1 lie within the N-terminal coiled-coil region of Sgo1 (Sgo1 1-130 ) and the adjacent unstructured region (Sgo1 130-280 ).
The Survivin interaction with the Sgo1 N-terminal tail is essential for CPC-Sgo1 assembly Our previous study identified a histone H3-like N-terminal tail in Sgo1 (Ala1-Lys2-Glu3-Arg4) that interacted with the Survivin BIR domain with affinity similar to that of the histone H3 tail (Jeyaprakash et al., 2011). Further crystal structure analysis revealed that the mode of Sgo1 tail binding is near identical to that of the histone H3 tail with phosphorylated threonine 3 (Ala1-Arg2-     , CPC ISB10-280 K62A , and CPC ISB10-280 H80A , respectively, and no measurable binding affinity for CPC ISB10-280 3A ). Conversely, when we mixed CPC ISB10-280 with Sgo1 1-155 Nmut or Sgo1 1-415 Nmut and tested their interaction by SEC (Fig. 3, D-F) and ITC (Figs. S2, D and E; and S3 D), neither Sgo1 1-155 Nmut , nor the longer Sgo1 1-415 Nmut , which includes the second CPC interacting region (aa 130-280), were able to interact with the CPC. These data agree with the tethering assays in which Sgo1 1-130 Nmut -LacI-GFP showed a drastic reduction in its ability to recruit Aurora B (Fig. S3 B) and Borealin (Fig. S3 C) to the LacO array compared with Sgo1 1-130 -LacI-GFP. Together, our results reveal that the Survivin-Sgo1 interaction is essential for CPC binding to Sgo1 and that the Sgo1 N-terminal tail acts as a hotspot whose perturbation abolishes the ability of CPC to form a complex with Sgo1.
Borealin and INCENP are required for a high-affinity CPC-Sgo1 interaction To assess how the different CPC subunits contribute to the high-affinity Sgo1 interaction, we performed a series of ITC experiments with either Survivin on its own or CPC ISB containing different Borealin truncations. Sgo1 1-130 interacted with Survivin with mid-nanomolar affinity (K D of 255 ± 33 nM; Figs. 4 A and S3 D). This, together with our previous observation that a Sgo1 N-terminal tail peptide bound Survivin with ∼1 μM affinity (Jeyaprakash et al., 2011) suggests that, although the interaction between the alanine and the Survivin BIR domain is essential for Sgo1/Survivin complex formation, the Sgo1-Survivin interaction extends beyond Sgo1 N-terminal tail. Sgo1 1-130 bound CPC ISB10-280 with a K D of 57.4 ± 7.9 nM, an approximately fivefold higher affinity compared with the affinity for Survivin alone (Figs. 4 B and S3 D). This observation together with the CLMS analysis suggests that further interactions involving Borealin, and possibly INCENP, strengthen the affinity between the CPC and Sgo1. Consistent with our CLMS analysis (Fig. 2, A and B) and a previous study (Bonner et al., 2020), CPC ISB lacking the Borealin dimerization domain (CPC ISB10-221 ) bound Sgo1 1-130 with a threefold lower affinity compared with the CPC ISB10-280 (K D = 163 ± 15.9 nM vs. 57.4 ± 7.9 nM), highlighting the contribution of the Borealin dimerization domain for binding to Sgo1 (Figs. 4 C and S3 D). The measured affinity of CPC  binding to the near-full-length Sgo1 (Sgo1 1-415, K D = 52.8 ± 6.95 nM; Fig. 1 C) is almost identical to that for Sgo1 1-130 (Fig. 4 B; K D = 57.4 ± 7.9 nM). This confirms that the first 130 amino acids of Sgo1 represent the main CPC-interacting region in vitro. Furthermore, the observation that the affinity goes from a micromolar range for the Sgo1 N-terminal tail (AKER peptide) with Survivin (Jeyaprakash et al., 2011) to the low nanomolar range for the CPC ISB10-280 -Sgo1 1-130 complex indicates that although the interaction between the CPC and Sgo1 depends on the Sgo1 N-terminal tail binding to Survivin, the high-affinity interaction requires Sgo1 binding to Borealin and possibly IN-CENP. Overall, the ITC data indicate that the interaction between the Sgo1 N-terminal tail and Survivin is electrostatically driven (Fig. S3, D and E), while the high-affinity interaction between the rest of the Sgo1 regions and the CPC is strengthened by entropic contributions that could be due to a release of water molecules associated with the surface and/or a conformational rearrangement upon binding (Fig. S3 D). These data together suggest that a weak micromolar affinity electrostatic interaction between Survivin and the Sgo1 N-terminal tail is required to establish a high-affinity CPC-Sgo1 interaction mediated by multiple interprotein contacts and hydrophobic effects.
the CPC subunits are shown. Intermolecular contacts of INCENP, Survivin, and Borealin with Sgo1 are shown as yellow, green, and purple lines, respectively. XiNet (Kolbowski et al., 2018) was used for data visualization. Autovalidation filter was used. (B) Cartoon representation of the crystal/nuclear magnetic resonance structures of the CPC (CPC core; PDB accession no. 2QFA; Jeyaprakash et al., 2007; Borealin dimerization domain; PDB accession no. 2KDD; Bourhis et al., 2009) and domain architecture of Sgo1 highlighting the regions involved in the CPC-Sgo1 contacts observed in A. Borealin residues in the circular view are annotated to match data deposited to the ProteomeXchange Consortium via the PRIDE repository (1-271 is equivalent to 10-280). (C and D) Representative immunofluorescence images (top) and quantification (bottom) for the analysis of the recruitment of endogenous Aurora B (C) and Borealin (D) to the LacO array in U-2 OS-LacO Haspin CM cells expressing different Sgo1-LacI-GFP constructs: LacI-GFP (n = 22 for Aurora B; n = 22 for Borealin), Sgo1 1-527 -LacI-GFP (n = 22 for Aurora B; n = 22 for Borealin), Sgo1 1-130 -LacI-GFP (n = 22 for Aurora B; n = 22 for Borealin), Sgo1 130-280 -LacI-GFP (n = 22 for Aurora B; n = 22 for Borealin), Sgo1 274-415 -LacI-GFP (n = 22 for Aurora B; n = 22 for Borealin), and Sgo1 415-527 -LacI-GFP (n = 22 for Aurora B; n = 22 for Borealin). Representative immunofluorescence images in C and D show Aurora B and Borealin signal for the same cell, thus, DAPI and GFP in C and D are the same. The graphs show the intensities of Aurora B and Borealin over GFP (dots) and the means (red bar). Data are representative of two biological replicates. Scale bar, 5 μm. One-way ANOVA with Dunnett's multiple comparison test (****, P < 0.0001).

Abad et al.
Journal of Cell Biology 5 of 16 Molecular basis for CPC-Sgo1 interaction https://doi.org/10.1083/jcb.202108156    Fig. S4, A and B) and that perturbing the central region interaction did not significantly reduce the measured CPC-binding affinity of Sgo1 1-415 by ITC (Fig. S4, D and E), we conclude that the Sgo1 central region does not make a significant contribution to CPC binding in vitro. However, as the same Sgo1 mutant (Sgo1 130-280 4A -LacI-GFP) is sufficient to perturb CPC-Sgo1 interaction in cells, we propose that Sgo1 central region requires one or more yet-unidentified posttranslational modifications to facilitate its interaction with the CPC, either in the Sgo1 region and/or in the CPC. As our analysis identified two CPC-interacting regions within Sgo1 (the N-terminal 130 aa including the N-terminal tail and the conserved coiled-coil, and the conserved hydrophobic region between aa 188 and 191), we next evaluated their contribution for CPC recruitment in the context of full-length Sgo1 using the LacO-LacI tethering assay. Consistent with our in vitro data, full-length Sgo1, harboring the N-terminal mutation (Sgo1 1-527 Nmut -LacI-GFP), recruited less Aurora B or Borealin (Fig. 4 E) compared with the Sgo1 1-527 -LacI-GFP. Similarly, the 4A mutation in the full-length context (Sgo1 1-527 4A -LacI-GFP) also reduced the recruitment of Aurora B and Borealin, while the double mutant (Sgo1 1-527 Nmut/4A -LacI-GFP) showed an even stronger reduction of endogenous Aurora B and Borealin recruitment to the LacO array (Fig. 4 E). Collectively, these data demonstrate the contribution of both Sgo1 regions for CPC binding in cells.
The Survivin interaction with the Sgo1 N-terminal tail is essential for the centromeric localization of the CPC and proper chromosome segregation We next evaluated how the different Sgo1 regions we identified as important for the CPC-Sgo1 interaction contribute to the centromeric levels of the CPC in cells. Endogenous Sgo1 was depleted by siRNA in HeLa Kyoto cells transiently expressing either wild-type Sgo1 (Sgo1-GFP) or mutant Sgo1 (Sgo1 Nmut -GFP, Sgo1 4A -GFP, or Sgo1 Nmut/4A double mutant), and centromeric levels of Borealin were analyzed by quantitative immunofluorescence microscopy ( Fig. 5 A; Fig. S4, F and G; and Fig. S5 A). Consistent with previous observations (Broad et al., 2020;Kawashima et al., 2007;Meppelink et al., 2015;Tsukahara et al., 2010;van der Waal et al., 2012;Wang et al., 2010), depletion of Sgo1 led to a twofold reduction in the centromeric levels of Borealin. As expected, expression of wild-type Sgo1 (Sgo1-GFP) rescued the centromeric abundance of Borealin (Fig. 5 A). In line with our in vitro binding studies and cellular tethering data, expression of Sgo1 mutants (Sgo1 Nmut -GFP, Sgo1 4A -GFP, or Sgo1 Nmut/4A -GFP), aimed to perturb either the Sgo1-N-terminal tail-Survivin interaction or the Sgo1 188-191 -Borealin interaction, did not rescue the centromeric levels of Borealin, demonstrating that these regions directly contribute to the efficient centromere recruitment of the CPC (Fig. 5 A).
Considering the ability of Sgo1 to bind H2AT120ph and to recruit CPC to the kinetochore-proximal centromere, we analyzed the precise localization of Borealin using chromosome spreads of nocodazole-arrested HeLa cells expressing the Sgo1 mutants. Control HeLa cells or Sgo1 depletion in Sgo1-GFP-expressing HeLa cells displayed Borealin localization at the inner centromere with a small pool localized at the kinetochore-proximal centromere (Fig. 5 C), consistent with the previously described pattern of CPC localization in unperturbed mitotic cells (Bekier et al., 2015;Hadders et al., 2020;Liang et al., 2020). In contrast, depletion of Sgo1 on Sgo1 Nmut -GFP or Sgo1 4A -GFP expressing cells, Borealin was enriched as a single focus between the two sister ACA dots, similar to the inner centromere localization previously observed for Borealin dimerization mutants that bind less well to Sgo1 (Bekier et al., 2015). Quantification of the full width at half maximum values for the Borealin intensity profiles obtained from the line plots of the chromosome spreads confirmed that rescue of Sgo1 depletion with Sgo1-GFP expression generated a broader Borealin signal at the centromere (most likely the result of the combination of inner centromere and kinetochore-proximal centromere pools), while expression of Sgo1 mutants (Sgo1 Nmut -GFP and Sgo1 4A -GFP) generated narrower Borealin profiles consistent with CPC localized at the inner centromere only (Fig. 5 D). These data reveal that the interaction of CPC with H2AT120ph-bound Sgo1 is responsible for the kinetochore-proximal centromere pool of the CPC.

Discussion
Concentration of the CPC near centromeres during early mitosis facilitates accurate chromosome congression and segregation in many organisms (Carmena et al., 2012;Dai et al., 2005;Krenn and Musacchio, 2015;Liu et al., 2009;Tanaka et al., 2002;van der Horst and Lens, 2014;Wang et al., 2010;Wang et al., 2012;Wang et al., 2011;Welburn et al., 2010). Two histone phosphorylation marks, Haspin-mediated H3T3ph and Bub1-mediated H2AT120ph, ensure the inner and kinetochore-proximal centromere enrichment of the CPC, respectively (Broad et al., 2020;Hadders et al., 2020;Liang et al., 2020). Several independent studies have provided molecular and structural understanding of how the Survivin subunit of the CPC directly recognizes the H3T3ph mark and its flanking amino acid residues, including the free amino terminus (Du et al., 2012;Jeyaprakash et al., 2011;Kelly et al., 2010;Niedzialkowska et al., 2012;Serena et al., 2020;Wang et al., 2010). Unlike the H3T3ph mark, the H2AT120ph is indirectly recognized by the CPC via Sgo1, which is capable of directly binding H2AT120ph via its C-terminal Sgo motif (Liu et al., 2015). As far as the CPC-Sgo1 interaction is concerned, the coiled-coil (Tsukahara et al., 2010) and dimerization domains of Borealin (Bonner et al., 2020) and Survivin BIR domain 4A -GFP) and depleted of endogenous Sgo1 using siRNA oligonucleotides (right). Immunofluorescence of endogenous Borealin and ACA. DAPI was used for DNA staining. Scale bar, 10 µm. Quantification of Borealin levels at the centromeres using ACA as reference channel (left). Values normalized to Sgo1 siRNA/Sgo1-GFP condition. Three independent experiments, n ≥ 50 cells analyzed in total per treatment, mean ± SD, Kruskal-Wallis with Dunn's multiple comparisons test; ****, P ≤ 0.0001. The values from the three independent replicates are represented in three different symbols. (B) Quantification of chromosome alignment of cells subjected to biorientation assay. Transfected cells were treated with 100 μM Monastrol for 16 h and released into medium containing 5 μM MG132 for 1 h. Representative examples of the alignment categories: complete alignment, mild misalignment (with one to three misaligned chromosomes), and severe misalignment (with more than three misaligned chromosomes) are found in the left panel. Representative images of the conditions expressing the three Sgo1 mutants showing pairs of CENP-C foci (red; right panel). DAPI was used to visualize DNA. Scale bar, 5 μm. Three independent experiments, n ≥ 100 of metaphases analyzed; mean ± SD. (C) Line plots depicting normalized fluorescence intensity levels of Borealin and ACA, measured along a line across the two sister ACA signals of the interkinetochore axis. Scale bar, 2 μm. Left, representative images of kinetochore pairs represented in the line plots. (D) Quantification of the full width at half maximum for the Borealin signal obtained in the line plots. Three independent experiments, n ≥ 49 kinetochores, mean ± SD, Kruskal-Wallis with Dunn's multiple comparisons test; ****, P ≤ 0.0001.

Abad et al.
Journal of Cell Biology 10 of 16 Molecular basis for CPC-Sgo1 interaction https://doi.org/10.1083/jcb.202108156 (Jeyaprakash et al., 2011) have been implicated in direct Sgo1 binding. However, whether these interactions (Borealin-Sgo1 and Survivin-Sgo1) take place in the context of the CPC and their relative contribution for CPC-Sgo1 binding and centromere localization has remained unresolved.
Here we show that Sgo1 forms a tight complex with the CPC ISB10-280 in vitro and that the interaction between the Histone H3-like N-terminal region of Sgo1 and the BIR domain of Survivin is crucial for CPC-Sgo1 complex formation, while the interaction of ∼120 amino acid residues downstream of the Sgo1 N-terminal tail with Borealin, and possibly INCENP, are required for high-affinity binding. We previously showed that the CPC binds H3T3ph nucleosomes with a K D of ∼90 nM (Abad et al., 2019). This value is comparable to the measured K D for CPC-Sgo1 binding, and notably, both interactions rely on a micromolar affinity interaction involving the Survivin BIR domain and the N-terminal tails of H3 and Sgo1, respectively. In addition, we also identified a hydrophobic region in Sgo1, comprising aa 188-191, that was required for the CPC-Sgo1 interaction. Mutations of this motif in Sgo1 abrogated the interaction with the CPC as well as CPC centromere recruitment in cells. However, while a Sgo1 fragment surrounding aa 188-191 could bind the CPC ISB10-280 in vitro, this region did not appear to further contribute to the Sgo1 1-415 -CPC ISB10-280 interaction in vitro. Therefore, we deem it likely this interaction is mediated by yetunidentified posttranslational modification(s) on Sgo1 and/ or CPC.
Several independent studies proposed that Haspin-mediated H3T3ph and Bub1-mediated H2AT120ph recruit the CPC to centromeres independently and as distinct spatial pools, inner centromere and kinetochore-proximal centromere pools, respectively (Bekier et al., 2015;Broad et al., 2020;Hadders et al., 2020;Liang et al., 2020). Our detailed molecular mapping of the CPC-Sgo1 interaction provided an excellent opportunity to test this model using Sgo1 separation-of-function mutants. Importantly, it has been previously suggested that the kinetochore proximal pool is independent of H2AT120ph/Sgo1 (Bekier et al., 2015). However, we observed that Sgo1 mutations that specifically perturb CPC binding (Sgo1 Nmut and Sgo1 4A ) mainly affect the kinetochore-proximal centromere pool of the CPC while leaving the inner centromere pool largely intact, indicating Sgo1 as a main kinetochore-proximal centromere receptor for the CPC. This is in line with the observation that inhibition of Bub1 leads to loss of kinetochore-proximal centromere CPC in Haspin KO cells (Hadders et al., 2020). Finally, the aforementioned Sgo1 mutations led to chromosome misalignment and segregation errors. These observations suggest that the H2AT120phmediated kinetochore-proximal centromere pool of the CPC could indeed play a role in error correction (Hadders et al., 2020), in addition to a proposed role for this pool in SAC inhibition (Liang et al., 2020). Our data will guide future research that aims to couple specific CPC functions to the distinct CPC pools.
The observation that the Sgo1 and histone H3 N-terminal tails exploit the same binding site in Survivin suggests that these interactions could be mutually exclusive and may explain why the Bub1-dependent CPC pool exists as a kinetochore-proximal centromere pool that is spatially distinct from the Haspindependent inner centromere CPC pool (Broad et al., 2020;Hadders et al., 2020;Liang et al., 2020). Moreover, our observation that the H3T3ph binding deficient Survivin BIR mutants (K62A and H80A) retain Sgo1 binding is in line with previous findings from Liang et al. (2020), showing that these mutants retain their ability to form a kinetochore-proximal CPC pool. It further suggests that subtle differences in H3T3ph and Sgo1 binding mediated by Survivin BIR domain might contribute to the spatiotemporal control of the CPC pools along the intersister KT axis.
Sgo1 fragments (Sgo1 1-415 , Sgo1 1-155 , and Sgo1 1-130 ) were cloned in the pTYB11 vector (IMPACT system; New England Biolabs) which contains an Intein tag with an embedded chitinbinding domain. The Intein tag is a DTT-induced self-cleavable tag that allows purification of proteins with a native N-terminus, as it leaves no extra amino acids after cleavage. Cloning of the Sgo1 in the pTYB11 vector with an N-terminal Intein-tag allowed the purification of an Sgo1 with a native N-terminus, leaving the initiator methionine exposed to be excised by methionine aminopeptidases (Giglione et al., 2004). Sgo1 fragments were expressed in the BL21 (DE3) Gold Escherichia coli strain. Cells were grown at 37°C to OD 1.5 and induced overnight at 18°C with 0.35 mM IPTG. Cells were resuspended in lysis buffer containing 20 mM Tris-HCl, pH 7.5, 500 mM NaCl, and 1 mM EDTA and supplemented with complete EDTA-free cocktail tablets (1 tablet/50 ml cells; Roche), 0.01 mg/ml DNase (Sigma-Aldrich), and 1 mM PMSF. The lysate was sonicated for 8 min and centrifuged at 58,000 g for 50 min at 4°C, and the protein was batch purified using chitin beads (New England Biolabs). Protein-bound chitin beads were washed with lysis buffer and high salt buffer (20 mM Tris-HCl, pH 7.5, 1 M NaCl, 1 mM EDTA, and 1 mM ATP) and eluted with 20 mM Tris-HCl, pH 7.5, 500 mM NaCl, and 50 mM DTT overnight at room temperature. The eluted protein was then dialyzed into 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 50 mM glutamate, 50 mM arginine, and 2 mM DTT overnight at 4°C and loaded onto a HiTrap SP-HP (Cytiva) ion exchange column. The samples containing Sgo1 were pooled, concentrated, and run in a Superdex 200 Increase 10/300 column (Cytiva) pre-equilibrated with 25 mM Hepes, pH 8, 250 mM NaCl, 5% glycerol, and 2 M DTT. Sgo1 130-280 was cloned in a pEC-S-CDF-His vector as N-terminally His-tagged. Sgo1 130-280 4A was generated using the Quickchange site-directed mutagenesis method (Stratagene). The vectors were transformed in BL21 Gold cells and grown and induced as described above. Cells were resuspended in lysis buffer containing 20 mM Tris-HCl, pH 8, 500 mM NaCl, 35 mM imidazole, and 2 mM β-mercaptoethanol and supplemented with complete EDTA-free cocktail tablets (1 tablet/50 ml cells; Roche), 0.01 mg/ml DNase (Sigma-Aldrich), and 1 mM PMSF. The protein was purified using a HisTrap column (Cytiva). The protein-bound column was washed with lysis buffer and high salt buffer (20 mM Tris-HCl, pH 8, 1 M NaCl, 35 mM imidazole, and 2 mM β-mercaptoethanol) and eluted with 20 mM Tris-HCl, pH 8, 200 mM NaCl, 400 mM imidazole, and 2 mM β-mercaptoethanol. The eluted protein was then dialyzed into 20 mM Tris-HCl, pH 8, 200 mM NaCl, and 1 mM DTT overnight at 4°C and loaded onto a HiTrap Q (Cytiva) ion exchange column. The samples containing Sgo1 were pooled, concentrated, and run in a Superdex 200 Increase 10/300 column (Cytiva) preequilibrated with 25 mM Hepes, pH 8, 150 mM NaCl, 5% glycerol, and 1 mM DTT.
Interaction studies using SEC All SEC experiments for the purified recombinant proteins were performed on an AKTA Pure 25 HPLC unit (Cytiva) with sample collector. For all interaction studies, a Superdex 200 10/300 GL 24 ml column (Cytiva) was used at 4°C. Before sample injection, the column was pre-equilibrated in 25 mM Hepes, pH 7.5, 150 mM NaCl, 4 mM DTT, and 5% glycerol (vol/vol) for interaction experiments involving Sgo1 1-155 or pre-equilibrated in 25 mM Hepes, pH 8, 250 mM NaCl, 2 mM DTT, and 5% glycerol (vol/vol) for interaction experiments involving Sgo1 1-415 . 0.5-ml fractions were collected with a 0.2-column volume delayed fractionation setting. UV 280-and 260-nm wavelengths were monitored. A 1.5× to 2× molar excess of Sgo1 was used in all interaction studies with CPC. Proteins were mixed and incubated at 4°C for 1 h before being injected to the size exclusion column.

Chemical cross-linking and MS analysis
Cross-linking experiments of Sgo1 1-415 and CPC ISB10-280 were performed using EDC (Thermo Fisher Scientific) in the presence of N-hydroxysulfosuccinimide (Thermo Fisher Scientific). 25 μg of gel-filtrated protein complex was cross-linked with 20 μg EDC and 44 μg of N-hydroxysulfosuccinimide in 25 mM Hepes, pH 6.8, and 150 mM NaCl for 1 h 30 min at room temperature. The cross-linking was stopped by the addition of 100 mM Tris-HCl, and cross-linking products were briefly resolved using 4-12% Bis-Tris NuPAGE (Thermo Fisher Scientific). Bands were visualized by short Instant Blue staining (Abcam), excised, reduced with 10 mM DTT for 30 min at room temperature, alkylated with 5 mM iodoacetamide for 20 min at room temperature, and digested overnight at 37°C using 13 ng/μl trypsin (Promega). Digested peptides were loaded onto C18-Stage-tips (Rappsilber et al., 2007). Liquid chromatography with tandem mass spectrometry (LC-MS/MS) analysis was performed using an Orbitrap Fusion Lumos Tribrid Mass Spectrometer (Thermo Fisher Scientific) applying a "high-high" acquisition strategy. Peptide mixtures were injected for each mass spectrometric acquisition. Peptides were separated on a 75 µm × 50 cm PepMap EASY-Spray column (Thermo Fisher Scientific) fitted into an EASY-Spray source (Thermo Fisher Scientific), operated at 50°C column temperature. Mobile phase A consisted of water and 0.1% vol/vol formic acid. Mobile phase B consisted of 80% vol/ vol acetonitrile and 0.1% vol/vol formic acid. Peptides were loaded at a flow-rate of 0.3 μl/min and eluted at 0.2 μl/min using a linear gradient going from 2% mobile phase B to 40% mobile phase B over 139 (or 109) min, followed by a linear increase from 40 to 95% mobile phase B in 11 min. The eluted peptides were directly introduced into the mass spectrometer. MS data were acquired in the data-dependent mode with the top-speed option. For each 3-s acquisition cycle, the mass spectrum was recorded in the Orbitrap with a resolution of 120,000. The ions with a precursor charge state between 3+ and 8+ were isolated and fragmented using higher-energy collisional dissociation (HCD) or electron-transfer/HCD (EThcD). The fragmentation spectra were recorded in the Orbitrap. Dynamic exclusion was enabled with single repeat count and 60-s exclusion duration.

Mass photometry
High-precision microscope coverslips (no. 1.5, 24 × 50 mm) were cleaned with Milli-Q water, 100% isopropanol, Milli-Q water, and dried. Silicone gaskets (103250; Grace BioLabs) were placed on the coverslips. Samples were cross-linked with 0.01% glutaraldehyde for 5 min at 4°C and quenched by addition of 50 mM Tris-HCl, pH 7.5, for 1 h at 4°C. Immediately before mass photometry measurements, samples were diluted to 100 nM in buffer containing 25 mM Hepes, pH 8, 250 mM NaCl, and 2 mM DTT. For each acquisition, 20 nM of diluted protein was measured following manufacturer's instructions. All data was acquired using a One MP mass photometer instrument (Refeyn) and AcquireMP software (Refeyn, v2.4.1). Videos were recorded in the regular field of view using default settings. Data was analyzed using Discover MP software (Refeyn, v2.4.2).

Tethering assays
The LacO tethering assays were performed essentially as described before (Hadders et al., 2020). U-2 OS LacO Haspin CM cells (Hadders et al., 2020) were seeded on glass coverslips and directly transduced with recombinant baculovirus expressing LacI-GFP fusion proteins. After ∼4-6 h, S-trityl-L-cysteine (20 μM) was added and left to incubate overnight. The next morning cells were fixed in 4% PFA for 15 min and permeabilized with ice-cold methanol. Before staining, cells were blocked in PBS supplemented with 0.01% Tween20 (PBST) and 3% BSA for 30 min followed by staining with primary antibodies in PBST + 3% BSA for 2-4 h. Coverslips were then washed three times with PBST followed by staining with secondary antibodies and DAPI (1 μg/ml) for 1 h. After another three washes with PBST, coverslips were mounted using Prolong Diamond. Cells were imaged on a DeltaVision system. The following antibodies were used for indirect immunofluorescence: anti-Aurora B (mouse monoclonal; 1:1,000; 611083; BD Transductions), anti-Borealin (1:1,000; rabbit polyclonal; a kind gift from Dr. S. Wheatley, School of Life Sciences, Medical School, Queen's Medical Centre, University of Nottingham, Nottingham, UK), and GFP-Booster were processed by constrained iterative deconvolution using SoftWoRx 3.6 software package (Applied Precision), and the centromere intensity of Borealin was quantified using an ImageJ plugin (https://doi.org/10.5281/zenodo.5145584). Briefly, the plugin quantifies the mean fluorescence signal of Borealin and Sgo1 in a 3-pixel-wide ring immediately outside the centromere, defined by the ACA staining. For background subtraction, a selected area within the cytoplasm signal was selected. To compare data from different replicates, values obtained after background correction were averaged and normalized to the mean of Borealin intensity in the Sgo1-GFP rescue condition. Statistical significance of the difference between normalized intensities at the centromere region was established by a Kruskal-Wallis test with Dunn's multiple comparisons test using Prism 7.0. Quantification of anaphases displaying chromosome bridges or lagging chromosomes was performed 24 h after HeLa Kyoto cells were transfected with the siRNA oligonucleotides. For the Monastrol assay, HeLa Kyoto cells were synchronized with 100 μM Monastrol for 16 h and released into 5 μM MG132 for 1 or 2 h. Observed metaphases were classified as complete alignment, mild misalignment (one to three unaligned chromosomes), and severe misalignment (more than three unaligned chromosomes). Quantification of chromosome alignment errors in unperturbed asynchronous cells was performed as described above. The experiments were performed in triplicate, and a minimum of 85 cells per condition were quantified. For the chromosome spreads, 8 h after siRNA oligonucleotide transfection, HeLa cells were treated with 50 ng/ml Nocodazole. 16 h after Nocodazole treatment, cells were collected by mitotic shake-off and incubated in hypotonic buffer (75 mM KCl) at 37°C for 10 min. After attachment to glass coverslips using Cytospin at 1,800 rpm for 5 min, chromosome spreads were extracted with ice-cold PBS/0.2% Triton X-100 for 4 min and fixed with 4% PFA. The immunofluorescence was performed as described below. Three replicates were performed, and a minimum of 49 kinetochores were analyzed. The centroids of kinetochores were detected in ImageJ using Speckle TrackerJ (Smith et al., 2011) software. A custom ImageJ script (https://doi.org/10.5281/ zenodo.5235670) was then used to assign kinetochore pairs as closest neighbors, with a maximum separation of 1.5 µm. The fluorescence intensities along 2-µm line regions of interest through the centroids and centered on the midpoint of the pair were taken in both the channels. Full width at half maximum values for the Borealin line plots were calculated by linear interpolation using a combination of the point-slope formula and the slope formula. Statistical significance of the difference between the full width at half maximum values between different Sgo1 constructs was established by a Kruskal-Wallis test with Dunn's multiple comparisons test using Prism 7.0.
In all cases, cells were fixed in 4% PFA 48 h after DNA transfection and 24 h after oligonucleotide transfection. Cells were then permeabilized with permeabilization buffer (0.2% Triton X-100 in 1× PBS) for 10 min, blocked with 3% BSA in permeabilization buffer for 1 h, and incubated with primary and secondary antibodies in blocking buffer for 1 h each.
All experiments were performed in triplicate.

Western blot
To study Sgo1 levels after siRNA oligo treatment and to test the expression levels of each of the Sgo1-GFP constructs, HeLa Kyoto cells were transfected in 12-well dishes as described above for the rescue experiments and fluorescence microscopy, lysed in 1× Laemmli buffer, boiled for 5 min, and analyzed by SDS-PAGE followed by Western blotting. The antibodies used for the immunoblot were rabbit anti-Sgo1 antibody (1:1,000; a gift from Ana Losada's laboratory, Spanish National Cancer Research Centre, Madrid, Spain; Serrano et al., 2009), mouse anti-tubulin (1:10,000; ab18251; Abcam), and rabbit anti-GFP (1:1,000; Abcam). Secondary antibodies used were goat anti-mouse 680, donkey anti-rabbit 800, and donkey anti-mouse 800 (LI-COR) at 1:2,000 dilution. Immunoblots were imaged using the Odyssey CLx system, and band intensities were quantified using ImageJ, uncalibrated OD values. Values were then corrected by the corresponding tubulin levels (loading control) and normalized to siRNA control values. Three experimental replicates were analyzed.

Statistical methods
In the graphs corresponding to the Sgo1 siRNA and rescue experiments with Sgo1-GFP WT and mutants, mean ± SD was plotted. Data derived from the different conditions were compared using either a Kruskal-Wallis test with Dunn's multiple comparisons test, a χ 2 test, or a Student's two-tailed unpaired t test using Prism 7.0. When parametric tests were used, normality was tested using a Shapiro-Wilk normality test using Prism 7.0. The tethering assays were analyzed using a one-way ANOVA with Dunnett's multiple comparisons test. Data were considered statistically different at P ≤ 0.05 with a single asterisk, at P ≤ 0.01 with two asterisks, at P ≤ 0.001 with three asterisks, and at P ≤ 0.0001 with four asterisks.
Online supplemental material Fig. S1 shows the sequence alignment of Sgo1 orthologues and the mass photometry histograms and kernel density estimates that support Fig. 1. Fig. S1 also shows the crosslinking SDS-PAGE that supports Fig. 2 and a cartoon representation of the Survivin-Sgo1 AKER structure and Survivin/Sgo1 isotherms that support Fig. 3. Fig. S2 shows SEC profiles and ITC isotherms that highlight the importance of the BIR domain of Survivin and the N-terminus of Sgo1 for CPC/Sgo1 interaction, supporting Fig. 3. Fig. S3 shows the in vivo LacO-LacI tethering assays and the ITC data that support the contribution of the BIR domain of Survivin and the N-terminus of Sgo1 for CPC/Sgo1 interaction, supporting Fig. 3. Fig. S4 shows SEC profiles and ITC isotherms corresponding to the Sgo1 4A mutant, supporting Fig. 4. Fig. S4 also shows the Western blots corresponding to Sgo1 depletion and Sgo1 transient expression, supporting Fig. 5. Fig. S5 shows centromere localization of Sgo1 constructs (right) and quantification of Sgo1 intensities at centromeres (left), supporting Fig. 5. It also shows the quantification of chromosome alignment and segregation defects observed for Sgo1 mutants, supporting Fig. 5 Figure S1. CPC and Sgo1 interact in vitro. (A) Sequence alignment of Sgo1 orthologues from Homo sapiens (hs), Bos taurus (bt), Mus musculus (mm), Gallus gallus (gg), Danio rerio (dr), and Xenopus laevis (xl). The conservation score is mapped from red (highly conserved) to yellow (poorly conserved). Predicted secondary structure elements are shown below the sequence alignment. Multiple sequence alignment was performed with Clustal Omega (EMBL-EBI) and edited with Jalview 2.11.0 (Waterhouse et al., 2009). Highlighted with boxes are the N-terminal coiled-coil domain of Sgo1, the highly conserved 188-191 region, and the Sgo motif. The N-terminal AKER motif of Sgo1 is well conserved in most higher vertebrates. . Observed metaphases were classified as complete alignment, mild misalignment (with one to three misaligned chromosomes), and severe misalignment (with more than three misaligned chromosomes). Three independent experiments, n ≥ 100 of metaphases analyzed; mean ± SD. (D) Quantification of anaphase cells with lagging chromosomes or chromosome bridges for the siRNA-rescue assay of the Sgo1-GFP constructs: Sgo1-GFP, Sgo1 Nmut-GFP , Sgo1 4A-GFP , or Sgo1 Nmut/4A-GFP . Right: Representative examples of lagging chromosomes and chromosome bridges quantified. Three independent experiments, n ≥ 300 of anaphases analyzed; mean ± SD; χ 2 test for differences between the indicated groups and the control, for % complete alignment; **, P ≤ 0.01; ***, P ≤ 0.001). Scale bar, 10 µm.