The Kinetochore Receptor for the Cohesin Loading Complex

Abstract

The ring-shaped cohesin complex brings together distant DNA domains to maintain, express, and segregate the genome. Establishing specific chromosomal linkages depends on cohesin recruitment to defined loci. One such locus is the budding yeast centromere, which is a paradigm for targeted cohesin loading. The kinetochore, a multiprotein complex that connects centromeres to microtubules, drives the recruitment of high levels of cohesin to link sister chromatids together. We have exploited this system to determine the mechanism of specific cohesin recruitment. We show that phosphorylation of the Ctf19 kinetochore protein by a conserved kinase, DDK, provides a binding site for the Scc2/4 cohesin loading complex, thereby directing cohesin loading to centromeres. A similar mechanism targets cohesin to chromosomes in vertebrates. These findings represent a complete molecular description of targeted cohesin loading, a phenomenon with wide-ranging importance in chromosome segregation and, in multicellular organisms, transcription regulation.

Keywords: cell cycle; centromere; cohesin; kinetochore.

Graphical abstract

Figure 1

DDK Phosphorylates Ctf19 (A) Ctf19 subcomplexes and their members (left) listed next to a schematic of the kinetochore with the Ctf19 complex shown in blue (right). (B) In vitro phosphorylation of the Ctf19-Mcm21 dimer by DDK. The indicated substrate proteins were incubated with purified DDK and γ-³²[P]-ATP. Reaction products were resolved by SDS-PAGE and visualized by autoradiography. (C) Diagram of Ctf19 showing the fragment resolved crystallographically (dark blue, PDB: 3ZXU) and the unstructured N-terminal region (light blue). Sequence alignment shows the N-terminal fragment. Red boxes indicate candidate phosphorylation motifs. Asterisks mark residues mutated in the ctf19-9A allele. (D) Ctf19 alleles used in this work. Yellow boxes indicate residues mutated to alanine. (E) DDK phosphorylates residues in the Ctf19 N-terminal region. Kinase assay performed as in (B). Substrates are indicated below. See also Figures S1–S3 and Table S1.

Figure S1

Reconstitution of DDK Association with Ctf19 Complex Proteins, Related to Figure 1 Dbf4 and Cdc7 were translated in vitro in the presence of ³⁵S-labeled methionine and incubated together with the indicated 6His-tagged Ctf19 complex members. Bead-bound proteins after Ni²⁺-affinity pulldowns were analyzed by SDS-PAGE and autoradiography (top). Purified bait proteins were analyzed separately by Coomassie stain (bottom). DDK associates with the Ctf3 trimer and with the Ctf19-Mcm21 dimer.

Figure S2

Determinants of Ctf19 Phosphorylation by DDK, Related to Figure 1 (A) Products of phosphorylation reactions containing Ctf19-Mcm21, analog-sensitive DDK, γ-³²[P]-ATP, and increasing concentrations of PP1. (B) Reactions carried out as in panel A, except wild-type DDK was used as in Figure 1B. After phosphorylation, reactions were subjected to a brief cleavage reaction by incubation with trypsin (top, autoradiography). Identical cleavage reactions were carried out in the absence of DDK and ATP to identify Ctf19-Mcm21 cleavage fragments (bottom; Coomassie). A schematic of the cleavage reaction is shown below (PDB: 3ZXU). (C) Phosphorylation of mutant Ctf19-Mcm21 complexes by purified DDK. Phosphorylation reactions were carried out as in panel A except that PP1 was not included. Substrates are indicated below. (D) The Ctf3 complex does not enhance Ctf19-Mcm21 phosphorylation in vitro. Reactions were carried out as in (B) in the presence or absence of an equimolar amount of purified Ctf3 complex.

Figure 2

Observation of Ctf19 Phosphorylation In Vivo (A) Ctf19-3FLAG, purified from exponentially growing cells expressing either wild-type CTF19 (SMH372) or ctf19-9A (SMH378) was resolved on a phostag gel and detected by western blot (arrow, phosphorylated Ctf19; ^∗, non-specific band). (B) Ctf19 phosphorylation was assessed as in (A) for strains expressing CTF19, ctf19-9A, ctf19-2A, or ctf19-3A-2 (SMH372, SMH378, SMH449, and SMH450, respectively). (C) Ctf19 phosphorylation assessed as in (A) for strains expressing CTF19, ctf19-9A, or CTF19 in a chl4Δ background (SMH372, SMH378, and SMH401, respectively). Cells were untreated or arrested in G1 as indicated (arrows, phosphorylated Ctf19; ^∗, non-specific band). (D) Strains bearing Ctf19-3FLAG and an inducible DDK expression cassette (Cdc7-6His, Dbf4 untagged) were not induced (left) or induced to express DDK (right) for 2 hr. Ctf19 phosphorylation was assessed in strains lacking CHL4 (SMH429) or expressing ctf19-9A (SMH430).

Figure S3

In Vivo Analysis of Ctf19 Phosphorylation, Related to Figure 2 (A) Ctf19 phosphorylation in arrested cultures (2h in arresting conditions; G1 – alpha factor; S phase – hydroxyurea; Metaphase – nocodazole and benomyl; arrows – phosphorylated Ctf19; ^∗ – non-specific band). For lanes 1-3, the indicated strains were harvested during asynchronous growth (CTF19 – SMH372, ctf19-9A – SMH378). For lanes 4-7, the CTF19-3FLAG strain used in lane 2 was harvested after the indicated arrests. (B) DNA content analysis for cultures used to produce samples 4-7 in panel B. (C) Ctf19 phosphorylation was assessed for strains expressing wild-type CTF19 (SMH425), ctf19-9A (SMH426), or CTF19 and dbf4-9myc (SMH427) and arrested in G1. Dbf4-9myc blot is shown for pre-immunoprecipitation samples. (D) Ctf19 phosphorylation was assessed for asynchronous cultures of the indicated strains (SMH372, SMH378, SMH401, SMH404, SMH402, and SMH403).

Figure 3

Ctf19 Phosphorylation Is Not Required for DDK Localization or Early Origin Activation (A and B) DDK localization is not perturbed by ctf19-9A. Wild-type and ctf19-9A cells carrying GFP-DBF4 (AM21871 and AM21872, respectively (A), or CDC7-6His-3FLAG (AM21110 and AM21234, respectively (B), together with an untagged control strain (AM1176) were arrested in G1 with alpha factor for 3 hr before harvesting for ChIP-qPCR (average of three biological repeats, error bars show SE; ns, not significant, Student’s t test). (C and D) Dbf4 or Cdc7 protein levels are unaltered in ctf19-9A strains. Representative anti-GFP and anti-HA immunoblots are shown (Pgk1, loading control). (E and F) Live-cell imaging to assess Sld7 and Cdc7 localization. Strains expressing Spc110-mCherry to mark spindle pole bodies and either Sld7-GFP or Cdc7-GFP were examined for the presence of GFP foci in G1 (SMH477, SMH478, and SMH480 for Sld7-GFP; SMH473, SMH474, and SMH476 for Cdc7-GFP). Sld7-GFP-expressing strains were arrested in G1 with alpha factor, and Cdc7-GFP-expressing strains were observed during unperturbed growth. (E) Images of wild-type CTF19 cells in G1 are shown for Sld7-GFP (top) and Cdc7-GFP (bottom; scale bar, 3 μm). (F) Fraction of cells with foci in the indicated strains (error bars, +/−SD for three experiments; ^∗∗p < 0.01; ^∗∗∗p < 0.001; ns, not significant; Student’s t test versus the CTF19 strain for the indicated GFP fusion protein, two tails, unequal variance). See also Figure S4.

Figure 4

Ctf19 Phosphorylation Is Required for Scc2 Localization at Centromeres (A and B) Model for the mechanism of Scc2/4 (gray) recruitment by the Ctf19 complex (blue) and DDK (red). Ctf3 is shown in green, and phosphate modifications are depicted as yellow circles. (A) Ctf3 recruits DDK, which phosphorylates Ctf19 to provide a docking site for Scc4. (B) Cells lacking the Ctf19 phosphorylation sites lack the Scc4 docking site, despite normal DDK recruitment. (C and D) Ctf19 phosphorylation is required for efficient localization of Scc2 to centromeres. Strains of the indicated genotypes expressing Scc2-GFP and Mtw1-tdTomato (SMH353, SMH361, SMH354, SMH355, SMH451, SMH452, SMH459, SMH460, SMH416, and SMH417 listed in the order shown) were imaged during unperturbed vegetative growth and scored for the presence of centromeric Scc2-GFP foci. Values are the averages of at least three independent experiments. At least 25 cells were scored per strain per experiment (error bars, +/−SD for the independent experiments; ^∗∗∗p < 0.001; ^∗∗p < 0.01; ns, not significant; Student’s t test versus CTF19 strain, two tails, unequal variance). (D) Representative images from experiments quantified in (C) (scale bar, 3 μm). Time after the beginning of image acquisition is shown. (E and F) Scc2 and Scc1 localization require Ctf19 phosphorylation. Cells were arrested in metaphase following treatment with nocodazole and benomyl for 2.5 hr (AM1176, untagged control strain). Values are means from four independent experiments (error bars, SE; ^∗p < 0.05; ^∗∗p < 0.01; Student’s t test for indicated comparisons, unpaired, two-tails). (E) ChIP-qPCR for Scc2 performed using wild-type (AM6006), CTF19 (AM20613), ctf19-9A (AM20619), ctf19-6A (AM20622), and ctf19-2A (AM21908) cells, carrying SCC2-6His-3FLAG. (F) ChIP-qPCR for Scc1 performed using wild-type (AM1145), CTF19 (AM20629), ctf19-9A (AM23456), and ctf19-2A (AM21958) cells, carrying SCC1-6HA. (G and H) Immunoblots showing levels of Scc2-6His-3FLAG, Scc1-6HA, and Pgk1 (loading control) from a representative experiment.

Figure 5

Ctf19 Phosphorylation Is Required for Centromeric Cohesion Establishment (A) Schematic depicting sister centromere separation at metaphase. Distance between centromere-associated GFP foci (green stars) was measured (red line). (B and C) Sister centromeres are further apart in metaphase in ctf19-9A cells compared to wild-type cells. (B) Strains bearing a lacO array at CEN15 (1.8 kb to the left of CEN15) and expressing LacI-GFP were imaged after arrest in metaphase by depletion of Cdc20 (pMET-HA3-CDC20) for 2h. Average centromere separation was determined for wild-type (SMH159), CTF19 (SMH412), ctf19Δ (SMH395), ctf19-9A (SMH396), and ctf19-3A-1 (SMH397) strains. Values were binned into categories (0 μm for no separation, 0–1 μm, and >1 μm–4 μm) and expressed as a percentage of all cells scored across three experiments. The fraction of cells in each bin was compared to the wild-type using Fishers exact test (^∗∗p < 0.0001). (C) Phospho-null but not phospho-mimetic mutations in phosphorylation sites closest to the Ctf19 N terminus cause a centromere cohesion defect. Wild-type (AM4643), ctf19-2A (AM22083), ctf19-2D (AM22481), and ctf19-2E (AM22484) strains, all carrying +2.4 CEN4-GFP (2.4 kb to the right of CEN4) were arrested in G1 before release into medium containing methionine to deplete CDC20 for 2 hr before imaging and analyzed as in (B). (D and E) Schematic showing the cohesin loading defect in ctf3Δ cells (D) and rescue of this defect by artificial tethering of Dbf4-FRB to the kinetochore (E). (F) Scc2 localization was assessed by ChIP-qPCR in the presence or absence of DDK tethering, CTF3, and Ctf19 phosphorylation sites. Wild-type (AM21613), ctf19-9A (AM21681), ctf19-2A (AM22360), ctf3Δ (AM21614), ctf19-9A ctf3Δ (AM21617), ctf19-2A ctf3Δ (AM22672), and control (AM21486) strains, all expressing DBF4-FRB and CTF19-FKBP12, were arrested in G1 for 3 hr before release into medium containing nocodazole, benomyl, and either DMSO (no DDK tethering) or rapamycin (DDK tethering; 1 μM) for 2.5 hr. Mean values of at least four independent experiments are shown. Error bars indicate SE (^∗p < 0.05; ^∗∗p < 0.01; Student’s t test, unpaired, two-tails; ns, not significant). (G) Immunoblots showing relative levels of Scc2-6His-3FLAG and Pgk1 (loading control) in a representative experiment.

Figure S4

DDK Recruitment to Kinetochores Is Abolished in Cells Lacking CTF3 or with C-Terminally Tagged Dbf4, Related to Figure 3 (A and B) DDK localization was assessed by ChIP-qPCR for wild-type and ctf3Δ strains. Wild-type and ctf3Δ cells carrying GFP-DBF4 (AM21871 and AM23275, respectively; panel A) or CDC7-6His-3FLAG (AM21110 and AM23268, respectively; panel B), together with an untagged control strain (AM1176) were arrested in G1 for 3h before harvesting for ChIP-qPCR. Means of three independent experiments are shown (error bars – standard error). (C and D) Immunoblots showing levels of GFP-Dbf4, Cdc7-FLAG and Pgk1 (loading control) in the samples from a representative experiment. (E) Ctf19 localization was assessed by ChIP-qPCR for wild-type and ctf3Δ strains. Wild-type and ctf3Δ strains carrying CTF19-6His-3FLAG were arrested in metaphase by treatment with nocodazole and benomyl and ChIP-qPCR experiments were performed as in Figure 5F. (F) Immunoblot showing levels of Ctf19-6His-3FLAG and Pgk1 (loading control) for the indicated strains. (G and H) DDK localization was assessed by ChIP-qPCR for strains expressing C-terminally tagged Dbf4. Wild-type cells and those in which Dbf4 bears a C-terminal 13Myc or FRB tag and carrying GFP-DBF4 (AM21871, AM22226, and AM22230, respectively; panel E) or CDC7-FLAG (AM21110, AM22538, and AM22540, respectively; panel F), along with an untagged control strain (AM1176), were arrested in G1 for 3h before harvesting for ChIP-qPCR. The means of three independent experiments are shown (error bars – standard error; ^∗ – p < 0.05; ^∗∗ – p < 0.01; Student’s t test, paired, two-tails). (I and J) Immunoblots showing relative levels of GFP-Dbf4, Cdc7-FLAG and Pgk1 (loading control) in the samples from a representative experiment.

Figure 6

Scc4 Recognizes Phosphorylated Ctf19 (A) Reconstitution of DDK-dependent Scc4-Ctf19 interaction in vitro. Recombinant Ctf19-6His-Mcm21 complexes were phosphorylated with purified DDK and then used for pulldowns with recombinant FLAG-Scc2^1–181-Scc4. Proteins were immunopurified on anti-FLAG beads, resolved by SDS-PAGE, and detected by immunoblot against an 6His-Mcm21 (top) or by Coomassie stain (total protein, bottom; ∗, degradation product). (B) Synthetic Ctf19^1–6 peptide (MDFpTpSD) was tested for binding to Scc2^1–181-Scc4^WT or Scc2^1–181-Scc4^m3. Mean average scaled response units are shown for three independent experiments for each data point. (C) Structure of phosphorylated Ctf19^1–6 bound to Scc2^1–181-Scc4. An overview of the complex (bottom) is shown as a cartoon with Scc4 in gray and Scc2^1-181 in rainbow (violet, N terminus; red, C terminus). The Ctf19-Scc4 interaction is shown in detail above. Individual residues mutated in the scc4-m7 allele are in purple. Other residues contributing to the interaction are in light blue. Ctf19 is purple with non-carbon atoms colored by element (red, oxygen; blue, nitrogen; yellow, phosphorous). Omit map density contoured to 0.8 sigma is shown for the Ctf19 peptide. See also Figure S5 and Tables S2 and S3.

Figure S5

Synthetic Ctf19^1–11 binds Scc2^1–181-Scc4, and Binding Requires Phosphorylation of Ctf19 Residue T4 or S5, Related to Figure 6 Experiments were carried out as in Figure 6B (error bars – +/−SD for three independent experiments, obscured by data points where not visible). m3 refers to immobilized Scc2^1-181-Scc4^m3. Ctf19 peptides and Scc2^1-181-Scc4 complexes used are listed in the table along with measured dissociation constants (K_d; SE – standard error for best-fit binding curve; single site, specific binding; Phos. – phosphorylated residues).

Figure 7

Scc4-Dependent Scc2/4 Localization in Vertebrates (A) Multiple sequence alignment for Scc4 homologs in yeasts and animals with residues mutated in the scc4-m3 allele shown. (B) Chromatin association of X. laevis Scc2^1–1024-Scc4 depends on the Scc4 conserved patch. X. laevis high-speed supernatant (HSS) was incubated with sperm chromatin and recombinant Scc2^1–1024-Scc4. Chromatin-associated proteins were recovered by sucrose sedimentation and detected by autoradiography (Scc2^1–1024-Scc4) or western blot (Orc2; ^∗non-specific band). (C) Schematic showing idealized kinase activity during the cell cycle. (D) Cartoon comparing DDK-dependent targeted cohesin loading at the yeast centromere with a similar process in vertebrates. See also Figure S6.

Figure S6

Phylogenetic Analysis of Candidate Ctf19 Phosphorylation Sites, Related to Figure 7 Candidate Ctf19 phosphorylation sites are widespread in yeasts and largely absent from chordates. N-terminal fragments of Ctf19 homologs for all listed species were examined for the presence of adjacent serine/threonine residues (SS, ST, TS, or TT). Blue boxes immediately surrounding the species names indicate the presence of candidate sites, with darker blue corresponding to multiple sites.