Structural evidence for Scc4-dependent localization of cohesin loading

Stephen M Hinshaw; Vasso Makrantoni; Alastair Kerr; Adèle L Marston; Stephen C Harrison

doi:10.7554/eLife.06057

. 2015 Jun 3;4:e06057. doi: 10.7554/eLife.06057

Structural evidence for Scc4-dependent localization of cohesin loading

Stephen M Hinshaw ¹, Vasso Makrantoni ², Alastair Kerr ², Adèle L Marston ², Stephen C Harrison ^1,^3,^*

Editor: Leemor Joshua-Tor⁴

PMCID: PMC4471937 PMID: 26038942

Abstract

The cohesin ring holds newly replicated sister chromatids together until their separation at anaphase. Initiation of sister chromatid cohesion depends on a separate complex, Scc2^NIPBL/Scc4^Mau2 (Scc2/4), which loads cohesin onto DNA and determines its localization across the genome. Proper cohesin loading is essential for cell division, and partial defects cause chromosome missegregation and aberrant transcriptional regulation, leading to severe developmental defects in multicellular organisms. We present here a crystal structure showing the interaction between Scc2 and Scc4. Scc4 is a TPR array that envelops an extended Scc2 peptide. Using budding yeast, we demonstrate that a conserved patch on the surface of Scc4 is required to recruit Scc2/4 to centromeres and to build pericentromeric cohesion. These findings reveal the role of Scc4 in determining the localization of cohesin loading and establish a molecular basis for Scc2/4 recruitment to centromeres.

DOI: http://dx.doi.org/10.7554/eLife.06057.001

Research organism: S. cerevisiae

eLife digest

DNA replication copies the genetic information contained in a cell's chromosomes. A ring-like protein complex, cohesin, holds together each pair of newly-replicated chromosomes, known as ‘sister chromatids’. When the cell divides, cohesin is cleaved; this allows sister chromatids to separate, so that each daughter cell receives one member of each sister chromatid pair and thereby inherits a full complement of genes. Defects in this process result in severe developmental abnormalities. Moreover, the genes that underlie this process are among the most frequently mutated in cancer.

Cohesin is enriched at centromeres: the chromosomal points of attachment to the apparatus (called the ‘mitotic spindle’) that segregates the sister chromatids into the daughter cells. The protein complex that loads cohesin onto chromosomes determines this preferential localization. The two components of the loading complex, Scc2 and Scc4, associate as a 1:1 pair.

Hinshaw et al. used a technique called X-ray crystallography to determine the structure of Scc4 bound with a large fragment of Scc2. The result showed that the elongated Scc4 twists around the fragment of Scc2, forming an extended groove. Scc2 snakes through the Scc4 groove and emerges at both ends.

Hinshaw et al. then performed a series of experiments in yeast cells to probe how Scc4 determines the location at which cohesin loads onto chromosomes. These experiments revealed that a region on the surface of Scc4 targets both Scc4 and Scc2 to centromeres. The amino-acid sequence of this centromere-targeting patch on the surface of Scc4 is conserved across species. Thus, the mechanism by which Scc4 localizes cohesin to centromeres may be similar in all eukaryotic organisms.

DOI: http://dx.doi.org/10.7554/eLife.06057.002

Introduction

Tight association between sister chromatids is crucial for successful chromosome segregation in eukaryotic cell division. Cohesin, a ring-shaped protein complex that wraps around sister chromatids (Gruber et al., 2003; Haering et al., 2008) is the molecular agent of sister chromatid cohesion, which persists from the time of DNA replication until anaphase. A distinct protein complex containing Scc2^NIPBL and Scc4^Mau2 (Scc2/4) initiates linkage of cohesin with DNA (Ciosk et al., 2000), and an in vitro reconstitution of cohesin loading by Scc2/4 suggests that the product of this reaction is a topological protein–DNA linkage (Murayama and Uhlmann, 2013). In metazoans, cohesin deposition by Scc2/4 is required for normal development (Dorsett et al., 2005; Kawauchi et al., 2009), and mutations in NIPBL, the human homolog of Scc2, are dominant and causally linked to a severe developmental disorder, Cornelia de Lange Syndrome (CDLS) (Krantz et al., 2004).

In addition to initiating a connection between cohesin and chromatin, Scc2/4 determines the timing and location of cohesin loading (Ciosk et al., 2000; Kogut et al., 2009). Cohesin enrichment at mitotic centromeres and pericentromeres results in tension across sister chromatids when paired kinetochores attach to opposite spindle microtubules (Tanaka, 2000). Mitotic spindle checkpoint signaling senses the tension between sister centromeres to ensure correct kinetochore–microtubule attachments (Stern and Murray, 2001). Defective centromeric cohesion therefore leads to elevated rates of chromosome missegregation (Eckert et al., 2007; Fernius and Marston, 2009).

Centromeric cohesion depends on recruitment of Scc2/4 to centromeres in late G1/early S phase (Hu et al., 2011; Fernius et al., 2013). A group of conserved kinetochore proteins—the Ctf19 complex in yeast (homologous to the human CCAN)—participates in this recruitment pathway, along with the S phase kinase complex, DDK (Fernius and Marston, 2009; Hu et al., 2011; Natsume et al., 2013). Deletion of any of several Ctf19 complex members leads to impaired centromeric cohesion and chromosome missegregation (Eckert et al., 2007; Fernius and Marston, 2009; Ng et al., 2009; Hu et al., 2011), but whether individual components make direct contact with Scc2/4 is not yet known.

The cohesin loading activity of Scc2/4 in vitro requires only Scc2 (Murayama and Uhlmann, 2013). Scc4 is essential in yeast, however, and in humans, de novo Mau2^Scc4 missense and nonsense mutations are significantly underrepresented in exome sequences, indicating that disruption of Mau2^Scc4 function is probably dominant and lethal (Table 1). Thus, Scc2 activity in vivo must depend on Scc4 in ways not recapitulated by the in vitro loading reaction. We report here the structure of yeast Scc4 in complex with an N-terminal fragment of Scc2 and demonstrate that Scc4 determines cohesin localization through a conserved patch on its surface. These findings show that Scc4 targets cohesin loading to a specific genomic locus and that this function is separable from its essential role in establishing sister chromatid cohesion across the genome.

Table 1.

De novo mutation profiles for human NIPBL and Mau2

DOI: http://dx.doi.org/10.7554/eLife.06057.019

	Observed	Expected
NIPBL
Synonymous	58	58.4
Missense	88	160.9
Loss of function	0	18.4
Mau2
Synonymous	25	26.6
Missense	13	52.7
Loss of function	0	3.8

Open in a new tab

Results

Structure of an Scc2^1–181/Scc4 complex

We prepared full-length Scc2/4 by co-expressing both proteins in baculovirus-infected insect cells. Scc2 is a 1493 amino acid residue protein predicted to have an unstructured N-terminal segment and a C-terminal HEAT (Huntington, EF3, PP2A, TOR1)/ARM (Armadillo) repeat domain (Figure 1A). Scc4, also conserved among nearly all eukaryotes, is predicted to have an extensive TPR (TetratricoPeptide Repeat) architecture. Negative stain electron microscopy of the full-length complex showed two large globular structures, variably positioned relative to each other, which we interpret as corresponding to the Scc2 HEAT repeat module and Scc2^N-Scc4 (Figure 1B). We used limited proteolysis and mass spectrometry to identify an Scc4-containing subcomplex (Figure 1—figure supplement 1A). An N-terminal fragment of Scc2 (residues 1–181 or 1–205) is sufficient for stable association with full-length Scc4 (Figure 1C). Both the truncated complex of Scc2^1–181/Scc4 and full-length Scc2/4 are heterodimers in solution (Figure 1—figure supplement 1B).

Figure 1. — (A) Domain organization of Scc2/4. Dotted lines show the Scc2–Scc4 interaction. An arrow indicates the position of a regulated cleavage site (Woodman et al., 2014). (B) Negatively stained Scc2/4 visualized by electron microscopy. Individual particles are shown. (C) Gel filtration chromatograms and SDS-PAGE show that Scc2^FL/Scc4 (magenta, left inset) and Scc2^1–181/Scc4 (purple, right inset) form stable complexes (* marks an Scc2 cleavage product).

**DOI:** http://dx.doi.org/10.7554/eLife.06057.003

Figure 1—figure supplement 1. — (A) Domain organization of Scc2/4. Dotted lines show the Scc2–Scc4 interaction. An arrow indicates the position of a regulated cleavage site (Woodman et al., 2014). (B) Negatively stained Scc2/4 visualized by electron microscopy. Individual particles are shown. (C) Gel filtration chromatograms and SDS-PAGE show that Scc2^FL/Scc4 (magenta, left inset) and Scc2^1–181/Scc4 (purple, right inset) form stable complexes (* marks an Scc2 cleavage product).

**DOI:** http://dx.doi.org/10.7554/eLife.06057.003

We obtained crystals of full-length Scc4 in complex with an N-terminal, 181-residue fragment of Scc2. Diffraction data collected from selenomethionine-substituted derivatives of these crystals allowed us to determine initial phases by single wavelength anomalous dispersion (SAD). We used an incomplete model, built into a 2.8 Å resolution map, to obtain phase information by molecular replacement for diffraction data from crystals of a native protein complex, extending to a minimum Bragg spacing of 2.1 Å. Our final model includes residues 5–383, 390–527, and 536–622 of Scc4 and residues 1–64, 73–95, and 106–132 of Scc2 (Figure 2A).

Figure 2—figure supplement 1. — (A) Rotated views of the Scc2^1–181/Scc4 complex. Scc2 is shown in gray as a cartoon and transparent surface. Individual Scc4 repeats (R1-13) are colored as indicated in (B).

**DOI:** http://dx.doi.org/10.7554/eLife.06057.005

Scc4 is a superhelical array of 13 TPR modules with a beta ribbon insertion between repeats 6 and 7 (Figure 2B). This tightly wrapped solenoid brings successive turns in contact with each other, and the concave surface becomes an axial groove. Repeat 8, which has a particularly long first helix, lacks a second helix and has instead an extended segment with a disordered surface loop. This irregularity divides the solenoid into two subdomains, TPR^N (Scc4^1–384) and TPR^C (Scc4^391–624) (Figure 2A).

Residues 10–50 of Scc2 snake along the continuous inner cavity of Scc4 and emerge at both ends, with the N-terminus of Scc2 close to the C-terminus of Scc4. Examples of similar TPR-peptide interfaces include the interaction of the kinesin light chain with cargo peptides (Figure 2—figure supplement 1A) (Pernigo et al., 2013) and the interaction of the cell polarity-determining LGN protein with its binding partners nuclear mitotic apparatus protein 1 (NuMA) and mInscutable (Figure 2—figure supplement 1B) (Zhu et al., 2011). The extended conformation of Scc2^1–181 clearly depends on its contacts with the surrounding Scc4; it is likely to be unstructured in the absence of its partner. The Scc2–Scc4 interaction has two unusual features. First, the concave surface of Scc4 is entirely enclosed, and Scc2 dissociation would therefore require either unfolding or proteolysis of Scc2 or Scc4. In fact, regulated proteolysis of Scc2 has been reported recently (Woodman et al., 2014). Second, residues 58–132 of Scc2 extend beyond the axial groove and make extensive contact with the external surface of Scc4.

Conserved Scc2–Scc4 contacts

Human Mau2^Scc4 binds the N-terminus of human NIPBL^Scc2 (Braunholz et al., 2012), but the primary sequence of NIPBL bears little resemblance to that of Scc2^1–181. We compiled sequence alignments for the N-terminus of Scc2 from divergent eukaryotes, including yeasts and humans, and mapped amino acid conservation onto its structure (Figure 2—figure Supplement 2A,B). Despite their low level of overall sequence conservation, Scc2/NIPBL proteins have conserved amino acid residues at buried positions that contact Scc4. Several Scc2–Scc4 contacts on the external face of Scc4 are also conserved, including Scc2^F86, which stacks onto Scc4^Y40-Y41, and Scc2^112–120, which forms a helix that fits over a hydrophobic surface on Scc4 TPR^N. We found that Scc2 variants mutated at conserved residues contacting either the concave or the external surface of Scc4 were less effective than wild-type Scc2 in restoring viability to an Scc2-degron (SCC2-AID) strain under depletion conditions (Figure 2—figure supplement 2C). The phenotype of these mutants was comparable to the phenotype we observed when we removed the entire Scc2 fragment visible in the crystal structure (Δ1–138).

If cohesin loading can occur in the absence of Scc4 in vitro (Murayama and Uhlmann, 2013) and to some extent in the absence of the Scc4-binding region of Scc2 in vivo, why is SCC4 an essential gene? We found that supplementing the chromosomal copy of SCC2 with a plasmid coding for Scc2 or Scc2^Δ1–138 complemented transcriptional repression of SCC4 (pGAL1-SCC4). That is, increasing the SCC2 gene copy number bypassed the effect on viability of transcriptional repression of SCC4, and the bypass did not require the Scc4-binding region of Scc2 (Figure 2—figure supplement 3). Because this region of Scc2 is probably unstructured when not bound by Scc4, it is a plausible cause of instability or aggregation when expressed unprotected. One reason why Scc4 is essential may therefore be that Scc2 is unstable in its absence.

A conserved surface patch on Scc4

Mapping Scc4 sequence conservation onto our structure revealed a cluster of extremely conserved, solvent-facing residues, some of them invariant among all eukaryotes we examined (Figure 3A,B). We found unaccounted-for electron density in this patch, into which we modeled a sulfate group. Because bound sulfates often mark phosphate binding sites in crystal structures (Bax et al., 2001), we tested the effects of mutating conserved residues that contribute to this patch. Strains in which the endogenous copy of SCC4 had been replaced by SCC4 coding for mutations at these residues were viable but displayed a plasmid missegregation phenotype (Figure 3—figure supplement 1A). Combined mutation of seven of the most conserved residues (scc4^{L256L; Y298A; K299D; Y313A; F324A; K327D; K331D}; scc4^m7) did not inactivate Scc4, as strains bearing these substitutions were fully viable and had a plasmid segregation phenotype comparable in strength to that of strains bearing mutations only at positions 324, 327, and 331 (scc4^{F324A; K327D; K331D}; scc4^m3) (Figure 3—figure supplement 1B,C). Recombinant Scc2^1–181-Scc4 complexes bearing these mutations behaved identically to wild-type preparations (Figure 1—figure supplement 1B), indicating that perturbation of the conserved Scc4 patch does not impair the stability or folding of the rest of the protein.

Figure 3. — (A) Surface view of Scc4 colored according to primary sequence conservation across eukaryotes. Inset shows the Scc4-conserved patch with mutated residues labeled and colored according to their effect on plasmid segregation fidelity (Figure 3—figure supplement 1). (B) Multiple sequence alignment of Scc4 and homologs from fungi and metazoans. Alignment is colored by conservation according to the color scheme in (A). (C) Plasmid missegregation was measured for the indicated strains (*scc4*^m3—*scc4*^{F324A; K327D; K331D}; error bars indicate SD; * p < 0.05, Student's t-test vs WT, two tails; n.s. indicates p > 0.05.). The dotted line shows the rate of plasmid segregation in a WT background. (D) Spindle length measurements for each indicated strain arrested in S phase with hydroxyurea. The dotted line shows the WT mean spindle length (error bars indicate SD; * p < 0.05, Student's t-test vs WT, two tails).

**DOI:** http://dx.doi.org/10.7554/eLife.06057.009

Figure 3—figure supplement 1. — (A) Surface view of Scc4 colored according to primary sequence conservation across eukaryotes. Inset shows the Scc4-conserved patch with mutated residues labeled and colored according to their effect on plasmid segregation fidelity (Figure 3—figure supplement 1). (B) Multiple sequence alignment of Scc4 and homologs from fungi and metazoans. Alignment is colored by conservation according to the color scheme in (A). (C) Plasmid missegregation was measured for the indicated strains (*scc4*^m3—*scc4*^{F324A; K327D; K331D}; error bars indicate SD; * p < 0.05, Student's t-test vs WT, two tails; n.s. indicates p > 0.05.). The dotted line shows the rate of plasmid segregation in a WT background. (D) Spindle length measurements for each indicated strain arrested in S phase with hydroxyurea. The dotted line shows the WT mean spindle length (error bars indicate SD; * p < 0.05, Student's t-test vs WT, two tails).

**DOI:** http://dx.doi.org/10.7554/eLife.06057.009

Strains lacking the Ctf19 complex subunit Chl4 (CENP-N in humans) are defective in centromeric cohesin loading because they cannot preferentially localize Scc2/4 to centromeres, although cohesin loads normally elsewhere in the genome (Fernius et al., 2013). To determine whether the Scc4-conserved patch functions in the same pathway as Chl4, we introduced the scc4^m3 mutation into a chl4Δ strain and tested for its effect on plasmid segregation (Figure 3C). Inclusion of scc4^m3 does not augment the severe plasmid loss phenotype caused by CHL4 deletion. This relationship also holds true for scc4^m7 (Figure 3—figure supplement 1B). Strains lacking Chl4 have extended metaphase spindles, and this phenotype corresponds to weakened centromeric cohesion (Fernius and Marston, 2009; Laha et al., 2011). We found that an scc4^m3-bearing strain exhibited increased inter-spindle pole distances and that the scc4^m3 mutation did not exacerbate the spindle extension phenotype of a chl4Δ strain (Figure 3D). Moreover, GFP-labeled sister centromeres (Figure 4—figure supplement 1A), but not chromosome arms (Figure 4—figure supplement 1D–G), were separated more frequently and to greater distances in strains mutated at 5 positions in the conserved patch on Scc4 (scc4^{F324A; K327A; K331A; K541A; K542A}; scc4^m35). We conclude that the conserved Scc4 patch promotes centromeric cohesion and that it does so in a manner that may also depend on CHL4.

Scc4 mutations disrupt centromeric cohesin loading

These results suggest that interactions at the Scc4-conserved patch target Scc2/4 specifically to centromeres. To test this hypothesis, we measured Scc2 and cohesin localization by chromatin immunoprecipitation (ChIP). Perturbation of the conserved patch (either scc4^m35 or scc4^m7) eliminates centromeric localization of Scc2 in mitotic cells and reduces association of the cohesin subunit, Scc1, with the centromere and pericentromere, but not with chromosome arms (Figure 4B,C, Figure 4—figure supplement 2A–D). We also observed this pattern of Scc2 localization in cells progressing through S phase (Figure 4—figure supplement 2E,F), the stage at which cohesin loading is initiated (Kogut et al., 2009; Natsume et al., 2013).

Figure 4. — (A) Sister centromeres (*+2.4CEN4-GFP*) are separated earlier and more frequently in an *scc4*^m35 strain entering the cell cycle after a G1 arrest. *pMET-CDC20* strains of the indicated genotypes (gray—wild type; magenta—*SCC4* integrated; blue—*scc4*^m35 integrated) were arrested in G1 with alpha factor and released into the cell cycle in the presence of methionine to repress *CDC20* expression. Solid lines show the percent of cells with separated *CEN4* dots, and dotted lines show the percent of cells with separated spindle pole bodies (Spc42-tdTomato). (B) Strains of the indicated genotypes and either Scc2-His6-3FLAG (top) or Scc1-6HA (bottom) were arrested in metaphase of mitosis following treatment with nocodazole and benomyl (to depolymerize microtubules). Cells were harvested after 2 hr. Anti-FLAG or anti-HA antibodies were used for ChIP, and pulldown samples were analyzed by qPCR. Mean values of four independent experiments are shown (error bars indicate ±SD; * p < 0.05; **p < 0.01 paired two-tailed t-test). (C) Schematic of a fragment of chromosome IV showing the location of qPCR amplicons used in (B). (D) Scc1 enrichment in a 50-kb domain surrounding all 16 budding yeast centromeres is shown for wild-type and *scc4*^m35 cells. For both wild type and *scc4*^m35, the ratio of reads (normalized to RPM) over input in a 100-bp window was calculated separately for each chromosome at the indicated position. The median count value for each window was then plotted to give a composite view of all 16 pericentromeres. (E) Scc1 enrichment along chromosome V together with a magnification of a 50-kb region including the centromere is shown. The number of reads at each position was normalized to the total number of reads for each sample (RPM: reads per million) and shown in the Integrated Genome Viewer from the Broad Institute (Robinson et al., 2011). (F) Live cell imaging of homozygous diploid cells expressing Scc2-GFP, Mtw1-tdTomato to mark centromeres, and the indicated version of Scc4 (left—wild type; right *scc4*^m7). GFP dots observed in wild-type cells are marked with white arrows in the first frame in which they are visible, and these foci were not observed in *scc4*^m7 cells. Time is given relative to the start of the imaging session (hr:mm). (G) Quantification of live cell imaging. At least 10 budding cells per field (three fields of view each for each strain for three separate experiments) were scored for the presence of an Scc2-GFP focus (error bars indicate ±SD; *p < 0.005, two-tailed t-test).

**DOI:** http://dx.doi.org/10.7554/eLife.06057.011

Figure 4—figure supplement 1. — (A) Sister centromeres (*+2.4CEN4-GFP*) are separated earlier and more frequently in an *scc4*^m35 strain entering the cell cycle after a G1 arrest. *pMET-CDC20* strains of the indicated genotypes (gray—wild type; magenta—*SCC4* integrated; blue—*scc4*^m35 integrated) were arrested in G1 with alpha factor and released into the cell cycle in the presence of methionine to repress *CDC20* expression. Solid lines show the percent of cells with separated *CEN4* dots, and dotted lines show the percent of cells with separated spindle pole bodies (Spc42-tdTomato). (B) Strains of the indicated genotypes and either Scc2-His6-3FLAG (top) or Scc1-6HA (bottom) were arrested in metaphase of mitosis following treatment with nocodazole and benomyl (to depolymerize microtubules). Cells were harvested after 2 hr. Anti-FLAG or anti-HA antibodies were used for ChIP, and pulldown samples were analyzed by qPCR. Mean values of four independent experiments are shown (error bars indicate ±SD; * p < 0.05; **p < 0.01 paired two-tailed t-test). (C) Schematic of a fragment of chromosome IV showing the location of qPCR amplicons used in (B). (D) Scc1 enrichment in a 50-kb domain surrounding all 16 budding yeast centromeres is shown for wild-type and *scc4*^m35 cells. For both wild type and *scc4*^m35, the ratio of reads (normalized to RPM) over input in a 100-bp window was calculated separately for each chromosome at the indicated position. The median count value for each window was then plotted to give a composite view of all 16 pericentromeres. (E) Scc1 enrichment along chromosome V together with a magnification of a 50-kb region including the centromere is shown. The number of reads at each position was normalized to the total number of reads for each sample (RPM: reads per million) and shown in the Integrated Genome Viewer from the Broad Institute (Robinson et al., 2011). (F) Live cell imaging of homozygous diploid cells expressing Scc2-GFP, Mtw1-tdTomato to mark centromeres, and the indicated version of Scc4 (left—wild type; right *scc4*^m7). GFP dots observed in wild-type cells are marked with white arrows in the first frame in which they are visible, and these foci were not observed in *scc4*^m7 cells. Time is given relative to the start of the imaging session (hr:mm). (G) Quantification of live cell imaging. At least 10 budding cells per field (three fields of view each for each strain for three separate experiments) were scored for the presence of an Scc2-GFP focus (error bars indicate ±SD; *p < 0.005, two-tailed t-test).

**DOI:** http://dx.doi.org/10.7554/eLife.06057.011

We further analyzed Scc1 localization by ChIP followed by high-throughput sequencing (ChIP-seq) and found that all 16 centromeres were specifically depleted of Scc1 in the scc4^m35 background (Figure 4—figure supplement 3). Scc1 depletion extended roughly 10 kilobases to either side of the core centromere but not to chromosome arms (Figure 4D,E). Moreover, centromeric Scc2-GFP signal, normally visible in wild-type cells and eliminated by CHL4 deletion (Fernius et al., 2013), was lost upon mutation of the Scc4-conserved patch (Figure 4F,G). These results indicate that, like Chl4, a specific interaction surface on Scc4 is required to target cohesin loading to centromeres.

Discussion

Cohesin targeting to specific genomic locations is required in all species for robust centromeric cohesion and for tissue-specific transcriptional programs in multicellular organisms (Kawauchi et al., 2009; Fay et al., 2011). DNA sequences per se do not drive cohesin loading (Onn and Koshland, 2011; Murayama and Uhlmann, 2013). Instead, targeting likely depends on interactions between the loading complex and chromatin-associated factors. We have determined the structure of Scc4 bound to a minimal fragment of Scc2, and we have used this structure to derive separation of function alleles that uncouple cohesin loading from cohesin targeting to centromeres. These experiments demonstrate that cohesin targeting in yeast depends on Scc4. This finding is consistent with unpublished genetic evidence for an Scc4-dependent cohesin localization pathway (Nasmyth, personal communication). Because the pathway that targets cohesin to centromeres requires Scc4 residues that are, in some cases, invariant across diverse eukaryotes, we suggest that the Scc4-dependent cohesin targeting we describe is a general feature of the control of cohesin loading.

Consistent with previous reports (Fernius et al., 2013; Natsume et al., 2013), we find that the centromeric enrichment of cohesin loading is not essential for viability and that arm cohesion is unaffected in strains in which this pathway is compromised. These results probably reflect two modes for cohesin loading: one that depends on a conserved patch of Scc4 and happens at centromeres, and a second opportunistic mode that happens everywhere on the chromosome and is sufficient to support viability in the absence of the first mode. If so, Scc4-conserved patch mutations would not eliminate cohesin loading but would redistribute loading events, resulting in similar cohesin levels at centromeres and on chromosome arms. This prediction is borne out both by strains bearing Scc4-conserved patch mutations and by strains lacking CHL4 (Fernius and Marston, 2009).

Recent reports have shown that cohesin targeting to specific locations along the chromosome arms of fission yeast (Mizuguchi et al., 2014) and flies (Oliveira et al., 2014) is critical for normal three-dimensional chromosome structure. In addition to findings from studies conducted in mammalian cells (Dowen et al., 2014), one of these reports (Mizuguchi et al., 2014) suggests that cohesin complexes residing at specified sites in the genome function in gene looping and in defining chromosome territories. Because our ChIP-seq experiments show that Scc4 directs cohesin localization to broad centromeric domains but does not affect the local distribution of cohesin peaks, we suggest that the Scc4-dependent localization pathway we discuss here is overlaid upon local determinants of cohesin positioning on chromosomes. These local determinants would be a second level of cohesin regulation, operating at roughly the resolution of the transcriptional units in yeast. Integration of these spatial cues with temporal cues, including the kinase activity of DDK, could then generate the final cohesin distribution observed in metaphase cells. The Scc4-conserved patch presents a possible explanation for how broad cohesin-dense domains may be specified by the cohesion-loading complex, and it provides a molecular foundation for the study of chromatin-associated factors, both known and unknown, that localize cohesin loading.

Materials and methods

Exome analysis

Analysis of exome data sets was performed as described (Samocha et al., 2014).

Plasmids

Coding sequences for Scc2 and Scc4 were amplified from yeast genomic DNA and inserted into a modified version of pFastbac (Life Technologies, Carlsbad, CA) suitable for ligation-independent cloning. The expression vector contains an N-terminal 6-His tag followed by a TEV protease sequence. Full-length Scc4 and fragments of the Scc2 coding sequence were cloned by the same procedure into a bacterial expression vector for expression from a single mRNA. The coding sequences for both genes were augmented such that each contains an N-terminal 6-His tag followed by a TEV protease cleavage site.

For complementation experiments, the Scc2 or Scc4 locus was amplified from Saccharomyces cerevisiae genomic DNA by PCR and cloned by restriction digest into a plasmid containing a CEN-ARS cassette and a selectable auxotrophic marker (LEU2). Genomic regions included were as follows: Scc2—chrIV:820797–825978; Scc4—chrV:465380–462755. We used PCR stitching and isothermal assembly to generate mutated versions of these constructs.

Yeast strains and culture conditions

Yeast strains bearing Scc4 point mutations or gene deletions were constructed using PCR methods as previously described (Longtine et al., 1998). All Scc4-conserved patch mutations described in this text were achieved by replacement of the native SCC4 locus. Viability of s288c strains expressing SCC4 mutants was first confirmed by complementation of pGAL1-SCC4 repression as shown for Figure 2—figure supplement 3 using plasmids bearing the SCC4 chromosomal locus. Viability of the scc4^m7 strain was determined by FACS analysis of homozygous diploid cells and during strain construction by sequential sporulation of the diploid imaging strains. To generate w303 strains carrying scc4 alleles, diploid strain (AM14499) carrying a heterozygous deletion (scc4Δ::KanMX6) was transformed with a PCR product corresponding to full-length SCC4 (or its mutant derivatives) and a downstream marker (HIS3). G418-sensitive, histidine prototrophs were sporulated, and SCC4 mutations were confirmed in the haploids by sequencing. All Scc2-GFP strains are derivatives of Scc2-GFP from the Yeast GFP Clone Collection (Huh et al., 2003). Strains for live cell imaging were constructed by integration of an Mtw1-tdTomato PCR into the Scc2-GFP strain followed by mating to achieve the final diploid strains. Plasmid segregation experiments were performed essentially as described previously (Hinshaw and Harrison, 2013).

Auxin-inducible degron-tagged Scc2 (SCC2-AID) was generated using PCR methods (Nishimura et al., 2009). For complementation assays, SCC2-AID cells bearing a CEN-ARS (Chr VI) plasmid encoding Scc2 flanked by its native control elements were grown to mid-log phase in synthetic complete (SC) medium lacking leucine (to select for the plasmid). Cells were plated in a fivefold dilution series on a solid SC medium lacking leucine and supplemented with the indicated amount of 1-Naphthaleneacetic acid (Auxin; Sigma-Aldrich, St. Louis, MO). Benomyl and nocodazole were used at 30 μg/ml and 15 μg/ml, respectively.

Protein expression and purification

Recombinant baculoviruses for His6-Scc2 and His6-Scc4 were amplified separately in Sf21 cells (Life Technologies) for three passages. For protein expression, Trichoplusia ni cells grown in suspension in Ex-Cell405 medium (Sigma–Aldrich) were infected with equal amounts of both viruses, and cells were pelleted and resuspended for freezing in lysis buffer (40 mM HEPES pH 7.5, 20 mM imidazole, 50 mM NaCl, 10% glycerol, and 2 mM β-mercaptoethanol) after 72 hr. Upon thawing and addition of protease inhibitors (4 μM aprotinin, 1 μM leupeptin, 1.4 μM pepstatin, and 1 mM PMSF), NaCl was added to a final concentration of 800 mM, and cells were broken by Dounce homogenization and sonication. Insoluble material was pelleted by centrifugation for 30 min at 18,000 rpm in a JA-20 rotor (Beckman-Coulter, Pasadena, CA). 6-His-tagged Scc2/4 complexes were isolated from the supernatant by Co²⁺ affinity chromatography followed by ion exchange chromatography (HiTrap SP HP, GE Healthcare, UK) and size exclusion chromatography (Superdex 200 20/16, GE) in gel filtration buffer (20 mM Tris–HCl pH 8.5, 200 mM NaCl, 1 mM TCEP).

To isolate Scc2^N–Scc4 complexes, polycistronic expression vectors (described above) encoding these proteins under the control of a single T7 promoter were transformed into the Escherichia coli strain Rosetta 2(DE3)pLysS (Millipore, Billerica, MA), and protein expression was induced with 400 μM IPTG at an OD600 of approximately 0.5. Bacterial cultures were further incubated overnight at 18°C. Cells were resuspended and frozen in lysis buffer containing 800 mM NaCl. Upon thawing and addition of protease inhibitors, cells were lysed by sonication, and 6-His-tagged proteins were purified as described for full-length Scc2/4. Selenomethionine-derivatized (SeMet) Scc2¹⁸¹-Scc4 samples were prepared as described for native protein samples with the exception that growth medium was prepared as described previously (Hinshaw and Harrison, 2013).

Electron microscopy and SEC-MALS

Purified Scc2/4 was diluted in gel filtration buffer and adsorbed to glow discharged carbon-coated copper grids. After staining with 0.75% (wt/vol) uranyl formate, grids were imaged using a CM10 electron microscope (Philips, Amsterdam).

For size exclusion chromatography coupled to multiple angle light scattering (SEC-MALS), experiments were performed essentially as described previously (Hinshaw and Harrison, 2013) with the exception that a 3-ml size exclusion column was used for analysis of truncated Scc2^N–Scc4 complexes (Superdex 200 5/150 GL; GE Healthcare).

Crystallization and structure determination

Crystals of Scc2^1–181-Scc4 formed overnight at 18°C. For native crystals, the protein was concentrated in gel filtration buffer to 18 mg/ml and mixed in a 1-to-1 ratio (vol:vol) with crystallization buffer (0.2M ammonium sulfate, 16% [wt:vol] PEG 3350). Crystals were washed first in wash buffer (160 mM NaCl, 16 mM Tris–HCl pH 8.5, 1 mM TCEP, 14.4% PEG 3350, and 0.16M ammonium sulfate) and then in wash buffer supplemented with 30% (vol:vol) glycerol before flash freezing in liquid nitrogen. SeMet-derivative crystals were concentrated to 18 mg/ml, and diffracting crystals formed in crystallization buffer with 20% (wt:vol) PEG 3350. These crystals were frozen as described for native versions. All diffraction data were collected on NE-CAT beamline 24ID-E. Data were indexed and scaled with XDS (SeMet) (Kabsch, 2010) or HKL2000 (native data) (Otwinowski and Minor, 1997).

The structure of Scc2^1–181-Scc4 was initially determined by SAD. To locate selenium atoms, we used SHELXD as implemented by HKL2MAP (Pape and Schneider, 2004). A search for 20 heavy atom sites with a resolution cutoff of 4 Å yielded a solution with 28 heavy atom positions. A truncated list of coordinates for 18 heavy atoms was used to generate an initial map at 3 Å resolution using Phenix Autosol (Adams et al., 2010). After density modification using Resolve (as implemented by Phenix), the map displayed extensive density corresponding to alpha helices. Placement of ideal helices and refinement using Phenix Refine yielded a partial structure, which was then used as a search model for phase determination by molecular replacement using a high-resolution native data set and Phaser-MR. During later stages of refinement, riding hydrogens were included, and TLS groups were invoked for Scc4 (Painter and Merritt, 2006). Crystallography statistics are shown in Table 2, and the coordinates have been deposited in the Protein Data Bank, accession number 4XDN.

Table 2.

Crystallographic data collection and refinement statistics

DOI: http://dx.doi.org/10.7554/eLife.06057.020

	Scc2^1–181; Scc4 (SeMet)	Scc2^1–181; Scc4 (Native)
Data collection
Resolution (Å)	30.0–2.8	178–2.0
Wavelength (Å)	0.979210	0.979240
Space group	P2₁2₁2₁	P2₁
Unit cell dimensions (a, b, c) (Å)	58.6, 89.0, 178.0	51.9, 178.1, 52.7
Unit cell angles (α, β, γ) (°)	90, 90, 90	90, 111.7, 90
I/σ (last shell)	11.6 (1.9)	6.0 (1.3)
R_sym (last shell) (%)	14.1 (92.3)	11.0 (72.9)
Completeness (last shell) (%)	99.7 (90.0)	93.0 (89.7)
Number of reflections	168241	154940
unique	23460	50878
Number of Se sites	18	–

Refinement
Resolution (Å)	–	28.7–2.1
Number of reflections	–	47188
working	–	45322
free	–	1866
R_work (last shell) (%)	–	18.5 (28.7)
R_free (last shell) (%)	–	21.0 (28.0)

Structure Statistics
Number of atoms (protein)	–	5845
sulfate	–	24
solvent	–	301
r.m.s.d. bond lengths	–	0.004
r.m.s.d. bond angles	–	0.661

Open in a new tab

Spindle pole separation assay

Log-phase cultures grown in SC medium were arrested in S phase with 10 mg/ml hydroxyurea for 90 min. Cells were fixed at room temperature for 10 min with 3.7% formaldehyde, washed twice with phosphate buffered saline, pH 8.5, and resuspended in wash buffer containing 1.2 M sorbitol. Cells were immobilized on concanavalin-A-coated cover slips and imaged using a Nikon Ti motorized inverted microscope with a 60× objective lens (NA 1.4) and a Hamamatsu ORCA-R2 cooled digital camera. Z-stacks (11 × 0.3 µm) were acquired with MetaMorph image acquisition software, and maximum z-projections were generated with ImageJ.

To calculate spindle pole distances, we wrote a Matlab script that identifies Spc110-mCherry foci and calculates a distance to the nearest neighbor for each instance. The list of distances was filtered to remove redundant measurements and to remove measurements arising from S phase spindles that straddled the edge of the image (distance measurements surpassing 32.9 μm).

CENIV dot and Spc42 separation

Cell growth and measurements were carried out as described previously (Fernius et al., 2013). Strain genotypes are listed in the strain table (Supplementary file 1).

Live cell imaging

Cells were grown in an SC medium overnight and diluted 1:20 (vol:vol) the next morning. After 6 hr, cells were immobilized on concanavilin A-coated cover slips and a fresh SC medium was applied. Experimental and control strains were loaded in adjacent imaging chambers, and cells were maintained at 30°C with high humidity using a Tokai Hit stage top incubator. Live cell images were captured using the imaging setup described above with the exception that Z-stacks (8 × 0.3 µm) were acquired every 8 min. We used exposure times of 10 ms (tdTomato) and 200 ms (GFP) for each image. Maximum intensity projections were generated with ImageJ for each timepoint, and figures were created with Nikon Elements software using identical processing steps and settings for each image.

Chromatin immunoprecipitation and qPCR

ChIP-qPCR and sequencing experiments were carried out as described previously (Fernius et al., 2013; Verzijlbergen et al., 2014). Scripts, data files, and workflows used to create the ChIP-Seq figures can be found on the github repository at https://github.com/AlastairKerr/Hinshaw2015. ChIP-Seq data sets have been deposited with the NCBI Gene Expression Omnibus under the accession number GSE68573.

FACS analysis

Flow cytometry was performed as previously described (Fernius et al., 2013). 5,000 cells were analyzed for each sample.

Acknowledgements

We thank Jonathan Schuermann and the staff at NE-CAT for help with data collection. We thank Simon Jenni for help with data processing and Kevin Corbett for critical reading of the manuscript. Microscopy experiments were performed with assistance from the Nikon Imaging Center at Harvard Medical School. We thank Bianka Baying at Genecore EMBL for library preparation and sequencing. This work was supported by funding from the National Science Foundation (SMH), HHMI (SCH), and the Wellcome Trust [090903, 092076, 096994].

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grants:

Wellcome Trust 090903 to Vasso Makrantoni, Alastair Kerr, Adèle L Marston.
National Science Foundation (NSF) Graduate student fellowship to Stephen M Hinshaw.
Howard Hughes Medical Institute (HHMI) to Stephen C Harrison.
Wellcome Trust 096994 to Vasso Makrantoni, Alastair Kerr, Adèle L Marston.
Wellcome Trust 092076 to Vasso Makrantoni, Alastair Kerr, Adèle L Marston.

Additional information

Competing interests

SCH: Reviewing editor, eLife.

The other authors declare that no competing interests exist.

Author contributions

SMH, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article.

VM, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article.

AK, Analysis and interpretation of data.

ALM, Conception and design, Analysis and interpretation of data, Drafting or revising the article.

SCH, Conception and design, Analysis and interpretation of data, Drafting or revising the article.

Additional files

Supplementary file 1.

Yeast strains used in this study.

DOI: http://dx.doi.org/10.7554/eLife.06057.015

elife06057s001.docx^{(131.2KB, docx)}

DOI: 10.7554/eLife.06057.015

Supplementary file 2.

Primers for ChIP-qPCR experiments in this study.

DOI: http://dx.doi.org/10.7554/eLife.06057.016

elife06057s002.docx^{(55.8KB, docx)}

DOI: 10.7554/eLife.06057.016

Major datasets

The following dataset was generated:

Kim H, Grunkemeyer TJ, Modi C, Chen L, Fromme R, Matz MV, Wachter RM, 2013, Crystal Structure of a reconstructed Kaede-type Red Fluorescent Protein, Least Evolved Ancestor (LEA), http://www.rcsb.org/pdb/explore/explore.do?structureId=4DXN, Publicly available at RCSB Protein Data Bank (Accession No: 4DXN).

The following previously published datasets were used:

Samocha, 2013, database of Genotypes and Phenotypes (dbGaP), http://www-ncbi-nlm-nih-gov.ezp-prod1.hul.harvard.edu/projects/gap/cgi-bin/study.cgi?study_id=phs000298.v1.p1, Publicly available to query here: http://atgu.mgh.harvard.edu/webtools/gene-lookup/.

Pernigo S, Lamprecht A, Steiner RA, Dodding MP, 2013, Crystal structure of the TPR domain of kinesin light chain 2 in complex with a tryptophan-acidic cargo peptide, http://www.rcsb.org/pdb/explore/explore.do?structureId=3ZFW, Publicly available at RCSB Protein Data Bank (Accession No: 3ZFW).

Zhu J, Wen W, Zheng Z, Shang Y, Wei Z, Xiao Z, Pan Z, Du Q, Wang W, Zhang M, 2011, Structures of the LGN/NuMA complex, http://www.rcsb.org/pdb/explore/explore.do?structureId=3RO2, Publicly available at RCSB Protein Data Bank (Accession No: 3RO2).

References

Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW, McCoy AJ, Moriarty NW, Oeffner R, Read RJ, Richardson DC, Richardson JS, Terwilliger TC, Zwart PH. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallographica. Section D, Biological Crystallography. 2010;66:213–221. doi: 10.1107/S0907444909052925. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bax B, Carter PS, Lewis C, Guy AR, Bridges A, Tanner R, Pettman G, Mannix C, Culbert AA, Brown MJ, Smith DG, Reith AD. The structure of phosphorylated GSK-3beta complexed with a peptide, FRATtide, that inhibits beta-catenin phosphorylation. Structure. 2001;9:1143–1152. doi: 10.1016/S0969-2126(01)00679-7. [DOI] [PubMed] [Google Scholar]
Braunholz D, Hullings M, Gil-Rodriguez MC, Fincher CT, Mallozzi MB, Loy E, Albrecht M, Kaur M, Limon J, Rampuria A, Clark D, Kline A, Dalski A, Eckhold J, Tzschach A, Hennekam R, Gillessen-Kaesbach G, Wierzba J, Krantz ID, Deardorff MA, Kaiser FJ. Isolated NIBPL missense mutations that cause Cornelia de Lange syndrome alter MAU2 interaction. European Journal of Human Genetics. 2012;20:271–276. doi: 10.1038/ejhg.2011.175. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ciosk R, Shirayama M, Shevchenko A, Tanaka T, Toth A, Shevchenko A, Nasmyth K. Cohesin's binding to chromosomes depends on a separate complex consisting of Scc2 and Scc4 proteins. Molecular Cell. 2000;5:243–254. doi: 10.1016/S1097-2765(00)80420-7. [DOI] [PubMed] [Google Scholar]
Dorsett D, Eissenberg JC, Misulovin Z, Martens A, Redding B, McKim K. Effects of sister chromatid cohesion proteins on cut gene expression during wing development in Drosophila. Development. 2005;132:4743–4753. doi: 10.1242/dev.02064. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dowen JM, Fan ZP, Hnisz D, Ren G, Abraham BJ, Zhang LN, Weintraub AS, Schuijers J, Lee TI, Zhao K, Young RA. Control of cell identity genes occurs in insulated neighborhoods in mammalian chromosomes. Cell. 2014;159:374–387. doi: 10.1016/j.cell.2014.09.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
Eckert CA, Gravdahl DJ, Megee PC. The enhancement of pericentromeric cohesin association by conserved kinetochore components promotes high-fidelity chromosome segregation and is sensitive to microtubule-based tension. Genes & Development. 2007;21:278–291. doi: 10.1101/gad.1498707. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fay A, Misulovin Z, Li J, Schaaf CA, Gause M, Gilmour DS, Dorsett D. Cohesin selectively binds and regulates genes with paused RNA polymerase. Current Biology. 2011;21:1624–1634. doi: 10.1016/j.cub.2011.08.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fernius J, Marston AL. Establishment of cohesion at the pericentromere by the Ctf19 kinetochore subcomplex and the replication fork-associated factor, Csm3. PLOS Genetics. 2009;5:e1000629. doi: 10.1371/journal.pgen.1000629. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fernius J, Nerusheva OO, Galander S, Alves Fde L, Rappsilber J, Marston AL. Cohesin-dependent association of scc2/4 with the centromere initiates pericentromeric cohesion establishment. Current Biology. 2013;23:599–606. doi: 10.1016/j.cub.2013.02.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gruber S, Haering CH, Nasmyth K. Chromosomal cohesin forms a ring. Cell. 2003;112:765–777. doi: 10.1016/S0092-8674(03)00162-4. [DOI] [PubMed] [Google Scholar]
Haering CH, Farcas AM, Arumugam P, Metson J, Nasmyth K. The cohesin ring concatenates sister DNA molecules. Nature. 2008;454:297–301. doi: 10.1038/nature07098. [DOI] [PubMed] [Google Scholar]
Hinshaw SM, Harrison SC. An Iml3-Chl4 heterodimer links the core centromere to factors required for accurate chromosome segregation. Cell Reports. 2013;5:29–36. doi: 10.1016/j.celrep.2013.08.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hu B, Itoh T, Mishra A, Katoh Y, Chan KL, Upcher W, Godlee C, Roig MB, Shirahige K, Nasmyth K. ATP hydrolysis is required for relocating cohesin from sites occupied by its Scc2/4 loading complex. Current Biology. 2011;21:12–24. doi: 10.1016/j.cub.2010.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
Huh WK, Falvo JV, Gerke LC, Carroll AS, Howson RW, Weissman JS, O'Shea EK. Global analysis of protein localization in budding yeast. Nature. 2003;425:686–691. doi: 10.1038/nature02026. [DOI] [PubMed] [Google Scholar]
Kabsch W. Xds. Acta Crystallographica. Section D, Biological Crystallography. 2010;66:125–132. doi: 10.1107/S0907444909047337. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kawauchi S, Calof AL, Santos R, Lopez-Burks ME, Young CM, Hoang MP, Chua A, Lao T, Lechner MS, Daniel JA, Nussenzweig A, Kitzes L, Yokomori K, Hallgrimsson B, Lander AD. Multiple organ system defects and transcriptional dysregulation in the Nipbl(+/-) mouse, a model of Cornelia de Lange syndrome. PLOS Genetics. 2009;5:e1000650. doi: 10.1371/journal.pgen.1000650. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kogut I, Wang J, Guacci V, Mistry RK, Megee PC. The Scc2/Scc4 cohesin loader determines the distribution of cohesin on budding yeast chromosomes. Genes & Development. 2009;23:2345–2357. doi: 10.1101/gad.1819409. [DOI] [PMC free article] [PubMed] [Google Scholar]
Krantz ID, McCallum J, DeScipio C, Kaur M, Gillis LA, Yaeger D, Jukofsky L, Wasserman N, Bottani A, Morris CA, Nowaczyk MJ, Toriello H, Bamshad MJ, Carey JC, Rappaport E, Kawauchi S, Lander AD, Calof AL, Li HH, Devoto M, Jackson LG. Cornelia de Lange syndrome is caused by mutations in NIPBL, the human homolog of Drosophila melanogaster Nipped-B. Nature Genetics. 2004;36:631–635. doi: 10.1038/ng1364. [DOI] [PMC free article] [PubMed] [Google Scholar]
Laha S, Das SP, Hajra S, Sanyal K, Sinha P. Functional characterization of the Saccharomyces cerevisiae protein Chl1 reveals the role of sister chromatid cohesion in the maintenance of spindle length during S-phase arrest. BMC Genetics. 2011;12:83. doi: 10.1186/1471-2156-12-83. [DOI] [PMC free article] [PubMed] [Google Scholar]
Longtine MS, McKenzie A, III, Demarini DJ, Shah NG, Wach A, Brachat A, Philippsen P, Pringle JR. Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast. 1998;14:953–961. doi: 10.1002/(SICI)1097-0061(199807)14:10<953::AID-YEA293>3.0.CO;2-U. [DOI] [PubMed] [Google Scholar]
Mizuguchi T, Fudenberg G, Mehta S, Belton JM, Taneja N, Folco HD, FitzGerald P, Dekker J, Mirny L, Barrowman J, Grewal SI. Cohesin-dependent globules and heterochromatin shape 3D genome architecture in S. pombe. Nature. 2014;516:432–435. doi: 10.1038/nature13833. [DOI] [PMC free article] [PubMed] [Google Scholar]
Murayama Y, Uhlmann F. Biochemical reconstitution of topological DNA binding by the cohesin ring. Nature. 2013;505:367–371. doi: 10.1038/nature12867. [DOI] [PMC free article] [PubMed] [Google Scholar]
Natsume T, Muller CA, Katou Y, Retkute R, Gierlinski M, Araki H, Blow JJ, Shirahige K, Nieduszynski CA, Tanaka TU. Kinetochores coordinate pericentromeric cohesion and early DNA replication by Cdc7-Dbf4 kinase recruitment. Molecular Cell. 2013;50:661–674. doi: 10.1016/j.molcel.2013.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ng TM, Waples WG, Lavoie BD, Biggins S. Pericentromeric sister chromatid cohesion promotes kinetochore biorientation. Molecular Biology of the Cell. 2009;20:3818–3827. doi: 10.1091/mbc.E09-04-0330. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nishimura K, Fukagawa T, Takisawa H, Kakimoto T, Kanemaki M. An auxin-based degron system for the rapid depletion of proteins in nonplant cells. Nature Methods. 2009;6:917–922. doi: 10.1038/nmeth.1401. [DOI] [PubMed] [Google Scholar]
Oliveira RA, Kotadia S, Tavares A, Mirkovic M, Bowlin K, Eichinger CS, Nasmyth K, Sullivan W. Centromere-independent accumulation of cohesin at ectopic heterochromatin sites induces chromosome stretching during anaphase. PLOS Biology. 2014;12:e1001962. doi: 10.1371/journal.pbio.1001962. [DOI] [PMC free article] [PubMed] [Google Scholar]
Onn I, Koshland D. In vitro assembly of physiological cohesin/DNA complexes. Proceedings of the National Academy of Sciences of USA. 2011;108:12198–12205. doi: 10.1073/pnas.1107504108. [DOI] [PMC free article] [PubMed] [Google Scholar]
Otwinowski Z, Minor W. Processing of x-ray diffraction data collected in oscillation mode. Methods in Enzymology. 1997;276:307–326. doi: 10.1016/S0076-6879(97)76066-X. [DOI] [PubMed] [Google Scholar]
Painter J, Merritt EA. Optimal description of a protein structure in terms of multiple groups undergoing TLS motion. Acta Crystallographica. Section D, Biological Crystallography. 2006;62:439–450. doi: 10.1107/S0907444906005270. [DOI] [PubMed] [Google Scholar]
Pape T, Schneider TR. HKL2MAP: a graphical user interface for macromolecular phasing iwth SHELX programs. Journal of Applied Crystallography. 2004;37:843–844. doi: 10.1107/S0021889804018047. [DOI] [Google Scholar]
Pernigo S, Lamprecht A, Steiner RA, Dodding MP. Structural basis for kinesin-1:cargo recognition. Science. 2013;340:356–359. doi: 10.1126/science.1234264. [DOI] [PMC free article] [PubMed] [Google Scholar]
Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP. Integrative genomics viewer. Nature Biotechnology. 2011;29:24–26. doi: 10.1038/nbt.1754. [DOI] [PMC free article] [PubMed] [Google Scholar]
Samocha KE, Robinson EB, Sanders SJ, Stevens C, Sabo A, McGrath LM, Kosmicki JA, Rehnström K, Mallick S, Kirby A, Wall DP, MacArthur DG, Gabriel SB, DePristo M, Purcell SM, Palotie A, Boerwinkle E, Buxbaum JD, Cook EH, Jr, Gibbs RA, Schellenberg GD, Sutcliffe JS, Devlin B, Roeder K, Neale BM, Daly MJ. A framework for the interpretation of de novo mutation in human disease. Nature Genetics. 2014;46:944–950. doi: 10.1038/ng.3050. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stern BM, Murray AW. Lack of tension at kinetochores activates the spindle checkpoint in budding yeast. Current Biology. 2001;11:1462–1467. doi: 10.1016/S0960-9822(01)00451-1. [DOI] [PubMed] [Google Scholar]
Tanaka T, Fuchs J, Loidl J, Nasmyth K. Cohesin ensures bipolar attachment of microtubules to sister centromeres and resists their precocious separation. Nature Cell Biology. 2000;2:492–499. doi: 10.1038/35019529. [DOI] [PubMed] [Google Scholar]
Tonkin ET, Wang TJ, Lisgo S, Bamshad MJ, Strachan T. NIPBL, encoding a homolog of fungal Scc2-type sister chromatid cohesion proteins and fly Nipped-B, is mutated in Cornelia de Lange syndrome. Nature Genetics. 2004;36:636–641. doi: 10.1038/ng1363. [DOI] [PubMed] [Google Scholar]
Verzijlbergen KF, Nerusheva OO, Kelly D, Kerr A, Clift D, de Lima Alves F, Rappsilber J, Marston AL. Shugoshin biases chromosomes for biorientation through condensin recruitment to the pericentromere. eLife. 2014;3:e01374. doi: 10.7554/eLife.01374. [DOI] [PMC free article] [PubMed] [Google Scholar]
Woodman J, Fara T, Dzieciatkowska M, Trejo M, Luong N, Hansen KC, Megee PC. Cell cycle-specific cleavage of Scc2 regulates its cohesin deposition activity. Proceedings of the National Academy of Sciences of USA. 2014;111:7060–7065. doi: 10.1073/pnas.1321722111. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zhu J, Wen W, Zheng Z, Shang Y, Wei Z, Xiao Z, Pan Z, Du Q, Wang W, Zhang M. LGN/mInsc and LGN/NuMA complex structures suggest distinct functions in asymmetric cell division for the Par3/mInsc/LGN and Galphai/LGN/NuMA pathways. Molecular Cell. 2011;43:418–431. doi: 10.1016/j.molcel.2011.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

eLife. 2015 Jun 3;4:e06057. doi: 10.7554/eLife.06057.017

Decision letter

Editor: Leemor Joshua-Tor¹

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

Thank you for sending your work entitled “Structural evidence for Scc4-dependent localization of cohesin loading” for consideration at eLife. Your article has been favorably evaluated by James Manley (Senior editor) and three reviewers, including Kim Nasmyth and Leemor Joshua-Tor, who is a member of our Board of Reviewing Editors. However, a number of points require attention, as described below.

The Reviewing editor and the other reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.

This paper reports on a crystal structure of a complex, between Scc4 and the N-terminus of Scc2. The Scc2/4 complex is essential for loading cohesin onto chromosomes. It has also been implicated in regulating chromosome structure and in controlling gene expression. Haplo-insufficiency of its Scc2 subunit (Nipbl) causes Cornelia de Lange syndrome in humans, an effect thought but not yet proven to be due to subtle defects in cohesin's genomic distribution. It has been proposed but not yet proven that cohesin loading corresponds to entrapment of DNAs by cohesin rings. If so, the function of Scc2/4 must be to facilitate DNA entry, presumably by opening transiently one of the interfaces connecting its Smc1, Smc3, and kleisin subunits. It has been reported that Scc2's ortholog in S.pombe, Mis4, can promote the stable association of cohesin with DNA in vitro in the absence of Scc4 but whether this association is truly topological is not known. How then does Scc2/4 facilitate cohesin loading and what is the function of Scc4 if it is not directly involved in the loading reaction? These are questions of central importance for the cohesin field and, because cohesin is just one of several related Smc/kleisin complexes, a topic of broad general interest for the field of chromosome biology.

The reviewers felt that the structure is an important advance to the field; however, the functional experiments are not performed using state of the art methodologies, as will be elaborated upon below, and the conclusions based on them do not appear to be justified by the data presented in the paper. Many of these should be straightforward to rectify and the authors are encouraged to consider resubmission upon completion of the following:

1) More extensive mutagenesis of the surface on Scc4.

2) Since ChIP with Scc2 is unreliable, results should be backed up by live cell imaging, which is more reliable and a better reflection of what is going on in the cell. This should be done using state of the art imaging (see point 8).

3) Whole genome comparisons of ChIP data would be a more rigorous way of demonstrating that Scc4 is concerned with loading cohesin around centromeres and not along arms. Such a conclusion is hard to make merely on the basis of analyzing a few sequences.

Major comments:

The structure reveals that Scc4 forms a superhelical array of TPR modules, as predicted from sequence analyses, and that the N-terminal residues of Scc2 snake along its inner cavity, which is entirely novel information. As might be expected, mutation of conserved residues within Scc2's NTD making contact with Scc4 compromise viability, as does deletion of the entire NTD. This is not unexpected as Scc4 itself is an essential gene. Mapping sequence variation onto the Scc4 structure revealed a cluster of highly conserved surface residues, mainly basic and hydrophobic. One might think therefore that this would prove to be a key surface through which Scc4 facilitates cohesin loading. Strangely, mutation of five residues within this conserved patch was not lethal (though the documentation here was inadequate). It did however appear to compromise recruitment of Scc2 to centromeres and to reduce the loading of cohesin onto peri-centric DNAs but not onto chromosome arms. Curiously, the anomaly of the most conserved part of Scc4 not being essential for loading cohesin onto chromosomes is not discussed in the manuscript, which is odd. The paper concludes that Scc4's conserved patch is required for peri-centric cohesin loading but not for a second “opportunistic” mode that occurs elsewhere on the chromosome (note that it is this second so called opportunistic mode that seems essential!). The paper also concludes that one of the functions of Scc4 is to mediate Scc2 stabilization, a statement for which there was no adequate justification. Though the structure appears to imply that this would be the case, this should be supported by some biochemical data (see below).

The claim that Scc4's conserved patch is not in fact essential for its function is a surprising one. Bold claims require exceptional documentation and sadly this is lacking here. From the structural figure supplied and the accompanying sequence variation, it would appear that L256, Y298, K299, Y313, F324, K327, and possibly K331 are the key surface conserved residues. And yet, for some reason the mutant Scc4 with the most extensive alterations in its conserved patch was scc4^m35, which contained F324A, K324A (presumably this should be K327A -typo?), K331A, K541A, and K542A. In other words, several of the key residues within the patch were left unmutated while residues that appear less conserved (K541 and K542, only conserved in S. cerevisiae like yeasts according to the figure) were mutated. If one wished to establish whether or not the conserved patch really is non-essential and has a specialized role in recruiting cohesin to centromeres, then one really needs to obliterate the nature of this surface and see whether the same results are obtained. What if mutating L256, Y298, K299, Y313, F324, K327, and K331 simultaneously were lethal? Would this not greatly alter the thrust of the paper and the interpretation of the results? Would it not imply that in fact, scc4^m35 is a hypomorph (with regard to the function of the conserved patch) and that the selective reduction in cohesin loading at centromeres arises because efficient loading at this location is more sensitive to a partial loss of function than loading along chromosome arms? This is a serious issue but not one that would require an inordinate amount of work to address. Making such mutants, characterizing their (non-?) effect on Scc2/4 complex formation in vivo, and a far more careful analysis of their phenotypes rigorously at the genomic level and by live imaging should be a high priority.

Specific comments:

1) The analysis of the effect of mutations within Scc2 that alter conserved Scc4 contact residues (Results) is not state of the art and raises important issues. First of all, the figure is of poor quality. Second, one should not use plasmids to do this type of analysis. As the authors subsequently point out, plasmid borne SCC2 supposedly bypasses the need for SCC4. Why, if the N-terminal domain of Scc2 is concerned solely with Scc4 binding and if plasmid borne SCC2 can suppress lethality due to lack of Scc4, do plasmid born mutant SCC2s that supposedly merely compromise Scc4 binding not suppress lethality due to SCC2-AID? Something is badly wrong here, unless there is a misunderstanding of some sort.

Note also that the method of using plasmids is likely to under-estimate the importance of residues that have been mutated. These experiments should be performed by integrating single copy wild type and mutant SCC2 at an ectopic site and testing function by tetrad dissection in a cross heterozygous for an scc2 deletion. Lastly, it would seem appropriate to test whether the mutations actually affect binding of Scc2's NTD to Scc4. This is easy to do and surprising that this was not done.

2) The claim that SCC2 over-expression bypasses the complete lack of Scc4 is interesting but maybe not too surprising (though see comment 1), but the experiment performed does not prove this. What was shown was that plasmid borne SCC2 can suppress the lethality of GAL-SCC4 cells on glucose. This experiment cannot exclude the possibility that there is minor expression of SCC4 from the GAL promoter and that this contributes to the suppression of lethality. It would be trivial to do this properly, that is, to show that plasmid born SCC2 can suppress the null. This should be done using tetrad analysis not just plasmid shuffle, which can easily lead to the isolation of secondary suppressors.

3) The statement that finding that SCC2 over-expression suppresses the lack of Scc4 is consistent with the essential function of Scc4 is to stabilize Scc2 is fairly meaningless. If the authors think that this is what is going on, then they need to measure protein stability.

4) Results. The authors raise the rather interesting possibility that Scc4's conserved patch might be involved in binding phosphate groups. This is particularly interesting given that the Cdc7 kinase is required for loading cohesin at centromeres. It is surprising that the paper does not address this issue further. Might the conserved patch be involved in binding a phosphorylated target on the Ctf19 complex?

5) Also in the Results. The plasmid mis-segregation assay is not a good one, the data not particularly impressive, and adds very little. These low quality experiments rather detract from the fine structural biology. Why not measure chromosome loss properly if this is so important?

6) In the same section. There appears to be little or no documentation of how the mutations in Scc4's conserved patch were made, how expressed, and how their non-lethality was initially determined. This is crucial part of the paper and the main text must explain carefully how this was done.

7) In the Results. The conclusion that mutations like scc4^m35 affect establishment of centromeric cohesion through a Chl4-dependent process is not rigorous. Is it not possible that other hypomorphic mutations of the Scc2/4 complex might have very similar phenotypes according to the experiments described and yet we know that Scc2/4 is required for loading throughout the genome? For example, it would be surprising if scc2-4 would not also have very similar phenotypes on centromeric cohesion. If so, would you conclude that this mutation affected in a specific manner the formation of centromeric cohesion? These results per se do not “strongly suggest that the Scc4 conserved patch targets Scc2/4 specifically to centromeres”. A close look at the data in Figure 4A suggests a slightly longer inter-SBP distance with the double mutant compared to single mutants. But there is nothing to compare a synthetic effect with, such as another mutant that would provide a longer inter-SPB distance when coupled with a chl4 deletion. The phenotype of chl4 mutants is stronger than that of scc4^m3 mutants and it is therefore hardly surprising that chl4 is epistatic to scc4^m3.

Such a result could arise whether or not scc4^m3 worked exclusively via Chl4 or not.

8) The experiments claiming that scc4^m35 compromises Scc2's recruitment to centromeres and selectively reduces peri-centric cohesin loading are not state of the art. First, there are major problems with analysing the distribution of Scc2 using ChIP. This problem afflicts data from a variety of organisms, not just yeast. The problem is that Scc2 is not stably associated with chromatin like cohesin. The only reliable method of documenting Scc2's localization at centromeres is to use live imaging, which avoids the countless pitfalls associated with ChIP. Fortunately, this can be readily done with S. cerevisiae where Scc2-GFP co-localizes with centromeres for much of the cell cycle and the centromeres are conveniently all clustered. The analysis performed here is of very poor quality. The images are not believable. It should be possible, especially if one uses diploids, to observe clear images in all cells post S phase and pre-anaphase. To distinguish mutant and wild type strains, these need to be differentially marked so that they can be placed on the same slide and imaged together and thereby compared rigorously. Moreover, the centromeres need to be marked separately in an unambiguous manner. The second problem concerns the use of QPCR ChIP to analyse cohesin's genomic distribution. In this day and age, it is not sufficient to draw major conclusions about the genomic distribution of proteins without analysing this using ChIP-seq. This would be easy to do.

eLife. 2015 Jun 3;4:e06057. doi: 10.7554/eLife.06057.018

Author response

1) More extensive mutagenesis of the surface on Scc4.

We have constructed strains bearing substitutions at seven of the most conserved amino acid residues that line the surface of the Scc4 conserved patch. We have examined several phenotypes relating to sister chromatid cohesion in these strains. We find that the changes at these seven positions, which extensively modify the wild type protein surface at this site on Scc4, yield strains that are fully viable: the only phenotypes we observe are consistent with a specific inability to enrich cohesin at centromeres and pericentromeres. Moreover, we find that recombinant Scc4-Scc2^1-181 protein complexes expressed in E. coli behave essentially identically regardless of the status of the Scc4 conserved patch. We have included these experiments in the revised version of our manuscript (Figure 1–figure supplement 1B; Figure 3–figure supplement 1B-C; Figure 4F-G; Figure 4–figure supplement 2D). We conclude that the conserved surface of Scc4 has a specific role in Scc2/4 localization but is not otherwise required for cohesin loading.

To address this point, we have generated new imaging strains and examined Scc2 localization to centromeres in the wild type and Scc4-mutant backgrounds (Figure 4F). Consistent with our initial observations, we find that centromeric Scc2-GFP foci are observed only in wild type cells and not in cells with the native copy of SCC4 replaced by scc4^m7. A more detailed discussion of these experiments follows in our response to point 8 of the reviewer's comments, below.

To analyze cohesin localization genome-wide, we have sequenced chromosomal DNA that purifies with Scc1 (ChIP-seq) in strains expressing either wild type or mutant Scc4 from its endogenous locus (Figure 4D-E; Figure 4–figure supplement 3). We find that Scc1 localization across the genome is identical in the wild type and scc4^m35 backgrounds with the exception of centromeres and ∼10kb of surrounding DNA on either side. These experiments provide evidence for a specific function of the conserved surface of Scc4 in enrichment of cohesin at centromeric regions.

One of the reviewers had extensive and very helpful comments, to which we respond in detail below.

Specific comments:

1) The analysis of the effect of mutations within Scc2 that alter conserved Scc4 contact residues (Results) is not state of the art and raises important issues. First of all, the figure is of poor quality. Second, one should not use plasmids to do this type of analysis. As the authors subsequently point out, plasmid borne SCC2 supposedly bypasses the need for SCC4. Why, if the N-terminal domain of Scc2 is concerned solely with Scc4 binding and if plasmid borne SCC2 can suppress lethality due to lack of Scc4, do plasmid born mutant SCC2s that supposedly merely compromise Scc4 binding not suppress lethality due to SCC2-AID? Something is badly wrong here, unless there is a misunderstanding of some sort.

There is indeed some misunderstanding. Our experiments show that an increase in SCC2 copy number suppresses SCC4 repression and that restoration of normal Scc2 protein rescues SCC2-AID depletion. These observations show that Scc4 likely supports the execution of Scc2’s normal function. In the absence of Scc4, more Scc2 is needed to achieve normal function, and with normal levels of Scc2, Scc2-Scc4 contact is required for normal function.

The following comments may further clarify. First, the failure of Scc2^∆138 to rescue SCC2-AID depletion is not total (compare Scc2^∆138 to vector in Figure 2–figure supplement 2C). Second, while transcriptional repression of SCC4 does not lead to a complete loss of viability, as pointed out by the reviewers, rescue by increasing SCC2 dosage does not require the Scc2-Scc4 interaction surface, indicating that an increased dosage of complete Scc2/4 complexes is not a likely explanation for this observation.

The experiment we present in Figure 2–figure supplement 2 shows that these residues are important, if not essential. We agree with the reviewer that investigation of precise phenotypes arising in cells bearing crippled versions of Scc2 should be the subject of future work, but the data we have presented shows that the interface between Scc2 and Scc4 is critical.

Lastly, it would seem appropriate to test whether the mutations actually affect binding of Scc2's NTD to Scc4. This is easy to do and surprising that this was not done.

We attempted to perform Scc2-Scc4 association studies in vitro but were not successful. Closure of the Scc2-Scc4 interaction surface, as the structure shows, probably means that these proteins fold together. The mode of association would in any case complicate measurement of Scc2-Scc4 binding upon mixing of individual recombinant components, because it would be hard to establish a condition of equilibrium.

The experiment we present in Figure 2–figure supplement 3 shows that compromised Scc4 can be rescued by increased SCC2 gene dosage and that this rescue does not require the N-terminus of SCC2. This result shows that Scc4 plays a supportive role in Scc2 function. While we offer some speculation on this point, the nature of this support should be the subject of future studies.

We have modified the main text so that it reflects our observations more accurately and to make clear that the comment is speculative (please see the subsection headed “Conserved Scc2-Scc4 contacts”).

We decided that including a thorough discussion of this possibility in our current manuscript would be overly speculative without further evidence for such a mechanism. We have included a brief reference to this model in our updated Discussion section.

5) Also in the Results. The plasmid mis-segregation assay is not a good one, the data not particularly impressive, and adds very little. These low quality experiments rather detract from the fine structural biology. Why not measure chromosome loss properly if this is so important?

We have moved our previous Figure 2B to Figure 3–figure supplement 1A, leaving only one panel for the plasmid loss phenotype in the main figures. This assay, a measure of chromosome loss in a physiological sense, gives a robust measure of the key downstream mitotic outcome of the pathway under investigation.

We have updated the Methods section of our manuscript to make clearer our strain construction procedures (please see the subsection headed “Yeast strains and culture conditions”).

7) Results. The conclusion that mutations like scc4^m35affect establishment of centromeric cohesion through a Chl4-dependent process is not rigorous. Is it not possible that other hypomorphic mutations of the Scc2/4 complex might have very similar phenotypes according to the experiments described and yet we know that Scc2/4 is required for loading throughout the genome? For example, it would be surprising if scc2-4 would not also have very similar phenotypes on centromeric cohesion. If so, would you conclude that this mutation affected in a specific manner the formation of centromeric cohesion?

In this manuscript, we provide the first documentation of clear separation of function alleles that are specifically deficient in their inability to designate centromeres as special domains for cohesin loading. We view this as an important contribution, if only because these alleles will allow further investigation of this pathway. Although we suppose it is unlikely that scc2-4 would show an identical phenotype, this remains to be tested.

These results per se do not “strongly suggest that the Scc4 conserved patch targets Scc2/4 specifically to centromeres”. A close look at the data in Figure 4A suggests a slightly longer inter-SBP distance with the double mutant compared to single mutants. But there is nothing to compare a synthetic effect with, such as another mutant that would provide a longer inter-SPB distance when coupled with a chl4 deletion. The phenotype of chl4 mutants is stronger than that of scc4^m3mutants and it is therefore hardly surprising that chl4 is epistatic to scc4^m3.

Such a result could arise whether or not scc4^m3 worked exclusively via Chl4 or not.

We agree that our CHL4 epistasis experiments are not conclusive proof that the Scc4 conserved patch and Chl4 function in the same pathway. The revised version of our manuscript reflects this uncertainty (please see the subsection headed “A conserved surface patch on Scc4”). We considered the usefulness of a second “off-pathway” mutant to test the range of signal detectable in in these assays. However, we decided that these experiments would be difficult to interpret given limited knowledge of genetic pathways leading to preferential centromeric cohesin loading and their interplay with DNA replication.

We have repeated our measurements of the spindle pole separation phenotype in triplicate and found no observable difference between the chl4Δ and scc4^m3 chl4Δ strains. An updated figure is included in the current version of the manuscript (Figure 3D).

8) The experiments claiming that scc4^m35 compromises Scc2's recruitment to centromeres and selectively reduces peri-centric cohesin loading are not state of the art. First, there are major problems with analysing the distribution of Scc2 using ChIP. This problem afflicts data from a variety of organisms, not just yeast. The problem is that Scc2 is not stably associated with chromatin like cohesin. The only reliable method of documenting Scc2's localization at centromeres is to use live imaging, which avoids the countless pitfalls associated with ChIP. Fortunately, this can be readily done with S. cerevisiae where Scc2-GFP co-localizes with centromeres for much of the cell cycle and the centromeres are conveniently all clustered. The analysis performed here is of very poor quality. The images are not believable. It should be possible, especially if one uses diploids, to observe clear images in all cells post S phase and pre-anaphase. To distinguish mutant and wild type strains, these need to be differentially marked so that they can be placed on the same slide and imaged together and thereby compared rigorously. Moreover, the centromeres need to be marked separately in an unambiguous manner.

We generated diploid yeast strains carrying homozygous copies of Scc2-GFP and Mtw1-tdTomato for simultaneous imaging of kinetochores and Scc2. Diploid cells showed clear GFP foci in nearly all WT cells we examined. In analogous strains bearing homozygous copies of the scc4^m7 allele, we found that these foci were absent or diminished beyond the limit of detection. This observation confirms our hypothesis that Scc4 guides Scc2 localization through its conserved patch.

While we were unable to incorporate a third label to allow imaging of both strains in the same field of view, we performed all imaging experiments by placing both strains immediately next to each other on the same cover slip and imaging during the same session with identical microscope and image collection settings. We also repeated the experiment using independently derived strains for each final diploid (distinct haploid precursors), and we made the same observation each time. We therefore conclude that Scc2 is not efficiently recruited to centromeres in strains bearing the scc4^m7 mutation.

The second problem concerns the use of QPCR ChIP to analyse cohesin's genomic distribution. In this day and age, it is not sufficient to draw major conclusions about the genomic distribution of proteins without analysing this using ChIP-seq. This would be easy to do.

We carried out ChIP-seq experiments using a 6xHA-tagged version of Scc1 expressed from its endogenous locus. These experiments show that the SCC4 mutations we report specifically impact cohesin recruitment to centromeres and the surrounding regions. We have included these experiments in the revised version of our manuscript (Figure 4D-E; Figure 4–figure supplement 3). We have also included a short note on the implications of these experiments in our updated Discussion.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary file 1.

Yeast strains used in this study.

DOI: http://dx.doi.org/10.7554/eLife.06057.015

elife06057s001.docx^{(131.2KB, docx)}

DOI: 10.7554/eLife.06057.015

Supplementary file 2.

Primers for ChIP-qPCR experiments in this study.

DOI: http://dx.doi.org/10.7554/eLife.06057.016

elife06057s002.docx^{(55.8KB, docx)}

DOI: 10.7554/eLife.06057.016

[bib1] Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW, McCoy AJ, Moriarty NW, Oeffner R, Read RJ, Richardson DC, Richardson JS, Terwilliger TC, Zwart PH. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallographica. Section D, Biological Crystallography. 2010;66:213–221. doi: 10.1107/S0907444909052925. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] Bax B, Carter PS, Lewis C, Guy AR, Bridges A, Tanner R, Pettman G, Mannix C, Culbert AA, Brown MJ, Smith DG, Reith AD. The structure of phosphorylated GSK-3beta complexed with a peptide, FRATtide, that inhibits beta-catenin phosphorylation. Structure. 2001;9:1143–1152. doi: 10.1016/S0969-2126(01)00679-7. [DOI] [PubMed] [Google Scholar]

[bib3] Braunholz D, Hullings M, Gil-Rodriguez MC, Fincher CT, Mallozzi MB, Loy E, Albrecht M, Kaur M, Limon J, Rampuria A, Clark D, Kline A, Dalski A, Eckhold J, Tzschach A, Hennekam R, Gillessen-Kaesbach G, Wierzba J, Krantz ID, Deardorff MA, Kaiser FJ. Isolated NIBPL missense mutations that cause Cornelia de Lange syndrome alter MAU2 interaction. European Journal of Human Genetics. 2012;20:271–276. doi: 10.1038/ejhg.2011.175. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] Ciosk R, Shirayama M, Shevchenko A, Tanaka T, Toth A, Shevchenko A, Nasmyth K. Cohesin's binding to chromosomes depends on a separate complex consisting of Scc2 and Scc4 proteins. Molecular Cell. 2000;5:243–254. doi: 10.1016/S1097-2765(00)80420-7. [DOI] [PubMed] [Google Scholar]

[bib5] Dorsett D, Eissenberg JC, Misulovin Z, Martens A, Redding B, McKim K. Effects of sister chromatid cohesion proteins on cut gene expression during wing development in Drosophila. Development. 2005;132:4743–4753. doi: 10.1242/dev.02064. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] Dowen JM, Fan ZP, Hnisz D, Ren G, Abraham BJ, Zhang LN, Weintraub AS, Schuijers J, Lee TI, Zhao K, Young RA. Control of cell identity genes occurs in insulated neighborhoods in mammalian chromosomes. Cell. 2014;159:374–387. doi: 10.1016/j.cell.2014.09.030. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] Eckert CA, Gravdahl DJ, Megee PC. The enhancement of pericentromeric cohesin association by conserved kinetochore components promotes high-fidelity chromosome segregation and is sensitive to microtubule-based tension. Genes & Development. 2007;21:278–291. doi: 10.1101/gad.1498707. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] Fay A, Misulovin Z, Li J, Schaaf CA, Gause M, Gilmour DS, Dorsett D. Cohesin selectively binds and regulates genes with paused RNA polymerase. Current Biology. 2011;21:1624–1634. doi: 10.1016/j.cub.2011.08.036. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] Fernius J, Marston AL. Establishment of cohesion at the pericentromere by the Ctf19 kinetochore subcomplex and the replication fork-associated factor, Csm3. PLOS Genetics. 2009;5:e1000629. doi: 10.1371/journal.pgen.1000629. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] Fernius J, Nerusheva OO, Galander S, Alves Fde L, Rappsilber J, Marston AL. Cohesin-dependent association of scc2/4 with the centromere initiates pericentromeric cohesion establishment. Current Biology. 2013;23:599–606. doi: 10.1016/j.cub.2013.02.022. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] Gruber S, Haering CH, Nasmyth K. Chromosomal cohesin forms a ring. Cell. 2003;112:765–777. doi: 10.1016/S0092-8674(03)00162-4. [DOI] [PubMed] [Google Scholar]

[bib12] Haering CH, Farcas AM, Arumugam P, Metson J, Nasmyth K. The cohesin ring concatenates sister DNA molecules. Nature. 2008;454:297–301. doi: 10.1038/nature07098. [DOI] [PubMed] [Google Scholar]

[bib13] Hinshaw SM, Harrison SC. An Iml3-Chl4 heterodimer links the core centromere to factors required for accurate chromosome segregation. Cell Reports. 2013;5:29–36. doi: 10.1016/j.celrep.2013.08.036. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] Hu B, Itoh T, Mishra A, Katoh Y, Chan KL, Upcher W, Godlee C, Roig MB, Shirahige K, Nasmyth K. ATP hydrolysis is required for relocating cohesin from sites occupied by its Scc2/4 loading complex. Current Biology. 2011;21:12–24. doi: 10.1016/j.cub.2010.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] Huh WK, Falvo JV, Gerke LC, Carroll AS, Howson RW, Weissman JS, O'Shea EK. Global analysis of protein localization in budding yeast. Nature. 2003;425:686–691. doi: 10.1038/nature02026. [DOI] [PubMed] [Google Scholar]

[bib16] Kabsch W. Xds. Acta Crystallographica. Section D, Biological Crystallography. 2010;66:125–132. doi: 10.1107/S0907444909047337. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] Kawauchi S, Calof AL, Santos R, Lopez-Burks ME, Young CM, Hoang MP, Chua A, Lao T, Lechner MS, Daniel JA, Nussenzweig A, Kitzes L, Yokomori K, Hallgrimsson B, Lander AD. Multiple organ system defects and transcriptional dysregulation in the Nipbl(+/-) mouse, a model of Cornelia de Lange syndrome. PLOS Genetics. 2009;5:e1000650. doi: 10.1371/journal.pgen.1000650. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] Kogut I, Wang J, Guacci V, Mistry RK, Megee PC. The Scc2/Scc4 cohesin loader determines the distribution of cohesin on budding yeast chromosomes. Genes & Development. 2009;23:2345–2357. doi: 10.1101/gad.1819409. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] Krantz ID, McCallum J, DeScipio C, Kaur M, Gillis LA, Yaeger D, Jukofsky L, Wasserman N, Bottani A, Morris CA, Nowaczyk MJ, Toriello H, Bamshad MJ, Carey JC, Rappaport E, Kawauchi S, Lander AD, Calof AL, Li HH, Devoto M, Jackson LG. Cornelia de Lange syndrome is caused by mutations in NIPBL, the human homolog of Drosophila melanogaster Nipped-B. Nature Genetics. 2004;36:631–635. doi: 10.1038/ng1364. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib20] Laha S, Das SP, Hajra S, Sanyal K, Sinha P. Functional characterization of the Saccharomyces cerevisiae protein Chl1 reveals the role of sister chromatid cohesion in the maintenance of spindle length during S-phase arrest. BMC Genetics. 2011;12:83. doi: 10.1186/1471-2156-12-83. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] Longtine MS, McKenzie A, III, Demarini DJ, Shah NG, Wach A, Brachat A, Philippsen P, Pringle JR. Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast. 1998;14:953–961. doi: 10.1002/(SICI)1097-0061(199807)14:10<953::AID-YEA293>3.0.CO;2-U. [DOI] [PubMed] [Google Scholar]

[bib22] Mizuguchi T, Fudenberg G, Mehta S, Belton JM, Taneja N, Folco HD, FitzGerald P, Dekker J, Mirny L, Barrowman J, Grewal SI. Cohesin-dependent globules and heterochromatin shape 3D genome architecture in S. pombe. Nature. 2014;516:432–435. doi: 10.1038/nature13833. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] Murayama Y, Uhlmann F. Biochemical reconstitution of topological DNA binding by the cohesin ring. Nature. 2013;505:367–371. doi: 10.1038/nature12867. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] Natsume T, Muller CA, Katou Y, Retkute R, Gierlinski M, Araki H, Blow JJ, Shirahige K, Nieduszynski CA, Tanaka TU. Kinetochores coordinate pericentromeric cohesion and early DNA replication by Cdc7-Dbf4 kinase recruitment. Molecular Cell. 2013;50:661–674. doi: 10.1016/j.molcel.2013.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib25] Ng TM, Waples WG, Lavoie BD, Biggins S. Pericentromeric sister chromatid cohesion promotes kinetochore biorientation. Molecular Biology of the Cell. 2009;20:3818–3827. doi: 10.1091/mbc.E09-04-0330. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] Nishimura K, Fukagawa T, Takisawa H, Kakimoto T, Kanemaki M. An auxin-based degron system for the rapid depletion of proteins in nonplant cells. Nature Methods. 2009;6:917–922. doi: 10.1038/nmeth.1401. [DOI] [PubMed] [Google Scholar]

[bib27] Oliveira RA, Kotadia S, Tavares A, Mirkovic M, Bowlin K, Eichinger CS, Nasmyth K, Sullivan W. Centromere-independent accumulation of cohesin at ectopic heterochromatin sites induces chromosome stretching during anaphase. PLOS Biology. 2014;12:e1001962. doi: 10.1371/journal.pbio.1001962. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] Onn I, Koshland D. In vitro assembly of physiological cohesin/DNA complexes. Proceedings of the National Academy of Sciences of USA. 2011;108:12198–12205. doi: 10.1073/pnas.1107504108. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] Otwinowski Z, Minor W. Processing of x-ray diffraction data collected in oscillation mode. Methods in Enzymology. 1997;276:307–326. doi: 10.1016/S0076-6879(97)76066-X. [DOI] [PubMed] [Google Scholar]

[bib30] Painter J, Merritt EA. Optimal description of a protein structure in terms of multiple groups undergoing TLS motion. Acta Crystallographica. Section D, Biological Crystallography. 2006;62:439–450. doi: 10.1107/S0907444906005270. [DOI] [PubMed] [Google Scholar]

[bib31] Pape T, Schneider TR. HKL2MAP: a graphical user interface for macromolecular phasing iwth SHELX programs. Journal of Applied Crystallography. 2004;37:843–844. doi: 10.1107/S0021889804018047. [DOI] [Google Scholar]

[bib32] Pernigo S, Lamprecht A, Steiner RA, Dodding MP. Structural basis for kinesin-1:cargo recognition. Science. 2013;340:356–359. doi: 10.1126/science.1234264. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP. Integrative genomics viewer. Nature Biotechnology. 2011;29:24–26. doi: 10.1038/nbt.1754. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] Samocha KE, Robinson EB, Sanders SJ, Stevens C, Sabo A, McGrath LM, Kosmicki JA, Rehnström K, Mallick S, Kirby A, Wall DP, MacArthur DG, Gabriel SB, DePristo M, Purcell SM, Palotie A, Boerwinkle E, Buxbaum JD, Cook EH, Jr, Gibbs RA, Schellenberg GD, Sutcliffe JS, Devlin B, Roeder K, Neale BM, Daly MJ. A framework for the interpretation of de novo mutation in human disease. Nature Genetics. 2014;46:944–950. doi: 10.1038/ng.3050. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] Stern BM, Murray AW. Lack of tension at kinetochores activates the spindle checkpoint in budding yeast. Current Biology. 2001;11:1462–1467. doi: 10.1016/S0960-9822(01)00451-1. [DOI] [PubMed] [Google Scholar]

[bib36] Tanaka T, Fuchs J, Loidl J, Nasmyth K. Cohesin ensures bipolar attachment of microtubules to sister centromeres and resists their precocious separation. Nature Cell Biology. 2000;2:492–499. doi: 10.1038/35019529. [DOI] [PubMed] [Google Scholar]

[bib37] Tonkin ET, Wang TJ, Lisgo S, Bamshad MJ, Strachan T. NIPBL, encoding a homolog of fungal Scc2-type sister chromatid cohesion proteins and fly Nipped-B, is mutated in Cornelia de Lange syndrome. Nature Genetics. 2004;36:636–641. doi: 10.1038/ng1363. [DOI] [PubMed] [Google Scholar]

[bib38] Verzijlbergen KF, Nerusheva OO, Kelly D, Kerr A, Clift D, de Lima Alves F, Rappsilber J, Marston AL. Shugoshin biases chromosomes for biorientation through condensin recruitment to the pericentromere. eLife. 2014;3:e01374. doi: 10.7554/eLife.01374. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] Woodman J, Fara T, Dzieciatkowska M, Trejo M, Luong N, Hansen KC, Megee PC. Cell cycle-specific cleavage of Scc2 regulates its cohesin deposition activity. Proceedings of the National Academy of Sciences of USA. 2014;111:7060–7065. doi: 10.1073/pnas.1321722111. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] Zhu J, Wen W, Zheng Z, Shang Y, Wei Z, Xiao Z, Pan Z, Du Q, Wang W, Zhang M. LGN/mInsc and LGN/NuMA complex structures suggest distinct functions in asymmetric cell division for the Par3/mInsc/LGN and Galphai/LGN/NuMA pathways. Molecular Cell. 2011;43:418–431. doi: 10.1016/j.molcel.2011.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Structural evidence for Scc4-dependent localization of cohesin loading

Stephen M Hinshaw

Vasso Makrantoni

Alastair Kerr

Adèle L Marston

Stephen C Harrison

Roles

Abstract

eLife digest

Introduction

Table 1.

Results

Structure of an Scc21–181/Scc4 complex

Figure 1. Purification of the cohesin-loading complex.

Figure 1—figure supplement 1. Purification and characterization of an Scc2N–Scc4 complex.

Figure 2. Crystal structure of the Scc21–181/Scc4 complex.

Figure 2—figure supplement 1. Comparison of Scc2N-Scc4 with structural homologs.

Figure 2—figure supplement 2. Structure and conservation of Scc2N.

Figure 2—figure supplement 3. Complementation of Scc4 repression by Scc2.

Conserved Scc2–Scc4 contacts

A conserved surface patch on Scc4

Figure 3. A conserved patch on the surface of Scc4.

Figure 3—figure supplement 1. (A) Plasmid segregation defects in Scc4-conserved patch mutants.

Scc4 mutations disrupt centromeric cohesin loading

Figure 4. Defective centromeric cohesin loading in Scc4 conserved patch mutants.

Figure 4—figure supplement 1. Spindle pole and CEN separation but not chromosome arm separation in an Scc4-conserved patch mutant.

Figure 4—figure supplement 2. Scc2 and Scc1 association with centromeres and chromosome arms.

Figure 4—figure supplement 3. Scc1 is reduced around all 16 individual centromeres in scc4m35 cells.

Discussion

Materials and methods

Exome analysis

Plasmids

Yeast strains and culture conditions

Protein expression and purification

Electron microscopy and SEC-MALS

Crystallization and structure determination

Table 2.

Spindle pole separation assay

CENIV dot and Spc42 separation

Live cell imaging

Chromatin immunoprecipitation and qPCR

FACS analysis

Acknowledgements

Funding Statement

Funding Information

Additional information

Competing interests

Author contributions

Additional files

Major datasets

References

Decision letter

Roles

Author response

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Structure of an Scc2^1–181/Scc4 complex

Figure 1—figure supplement 1. Purification and characterization of an Scc2^N–Scc4 complex.

Figure 2. Crystal structure of the Scc2^1–181/Scc4 complex.

Figure 2—figure supplement 1. Comparison of Scc2^N-Scc4 with structural homologs.

Figure 2—figure supplement 2. Structure and conservation of Scc2^N.

Figure 4—figure supplement 3. Scc1 is reduced around all 16 individual centromeres in scc4^m35 cells.