The Affinity of DNA-Microtubule Protein Complexes and their Disruption by Tubulin Binding Drugs

Abstract

Using the gel shift assay system, we have measured the apparent affinity constant for the interaction of two different DNAs with MAP proteins found in both total calf brain microtubules and heat stable brain preparations. Both DNAs studied contained centromere/kinetochore sequences- one was enriched in the calf satellite DNA; the other was a large restriction fragment containing the yeast CEN11 DNA sequence. Complexes formed using both DNAs had similar K_app values in the range of 2.1×10⁷ M^?1 to 2.0×10⁸ M^?1. CEN11 DNA-MTP complexes had by far the highest K_app value of 2.0×10⁸ M^?1. The CEN11 DNA sequence is where the yeast kinetochore of chromosome 11 is formed and where the single yeast microtubule is bound in vivo. The CEN11 conserved region II known binding sites -(dA/dT)n runs- for mammalian MAP2 protein, are in good agreement with this higher K_app value. The effects of the classical tubulin binding drugs colchicine, podophyllotoxin and vinblastine on the DNA-MAP protein complex stability were investigated by determining the drug concentrations where the complexes were destabilized. Only the complexes formed from total microtubule protein (tubulin containing) were destabilized over a wide drug concentration range. Heat stable brain protein complexes (no tubulin) were largely unaffected. Furthermore, it took 10–100 fold higher drug concentrations to disrupt the CEN 11 DNA complexes compared to the calf thymus satellite DNA enriched complexes. These data support our previous results suggesting that there is a DNA sequence dependent interaction with MAP proteins that appears to be conserved in evolution (Marx et. al., Biochim. Biophys. Acta. 783, 383–392,1984; Marx and Denial, Molecular Basis of Cancer 172B,65-15 1985). In addition, these results imply that the classical tubulin binding drugs may exert their biological effects in cells at least in part by disrupting DNA-Protein complexes of the type we have studied here.     

Point centromeres contain more than a single centromere-specific Cse4 (CENP-A) nucleosome

Cse4 is the budding yeast homologue of CENP-A, a modified histone H3 that specifies the base of kinetochores in all eukaryotes. Budding yeast is unique in having only one kinetochore microtubule attachment site per centromere. The centromere is specified by CEN DNA, a sequence-specific binding complex (CBF3), and a Cse4-containing nucleosome. Here we compare the ratio of kinetochore proximal Cse4-GFP fluorescence at anaphase to several standards including purified EGFP molecules in vitro to generate a calibration curve for the copy number of GFP-fusion proteins. Our results yield a mean of ~5 Cse4s, ~3 inner kinetochore CBF3 complexes, and ~20 outer kinetochore Ndc80 complexes. Our calibrated measurements increase 2.5-3-fold protein copy numbers at eukaryotic kinetochores based on previous ratio measurements assuming two Cse4s per budding yeast kinetochore. All approximately five Cse4s may be associated with the CEN nucleosome, but we show that a mean of three Cse4s could be located within flanking nucleosomes at random sites that differ between chromosomes.     

Study of binding of some amino acid derivatives of 2,2'-bipyridineplatinum(II) to calf thymus DNA

The interactions of glycine, alanine, valine, leucine, proline, methionine, and asparagine derivatives of 2,2'-bipyridineplatinum(II) with calf thymus DNA, polyadenylic acid, and polyguanylic acid have been studied by difference absorption spectral technique. The association constants (Kapp) and number of binding sites per 100 nucleotides (n) have been obtained from binding isotherms which were constructed by treatment of data according to Scatchard equation. The Kapp values (2.2 X 10(4) to 4.6 X 10(5) M-1) of binding of platinum complexes with DNA are comparable to Kapp values (5.6 X 10(4) to 2.9 X 10(5) M-1) of binding of platinum complexes with polyguanylic acid. However, the polyadenylic acid does not exhibit any binding to platinum complexes. The hydrogen bonding is, however, the most probable mode of bonding involved in stabilizing the DNA-amino acid complexes.     

Making an effective switch at the kinetochore by phosphorylation and dephosphorylation

The kinetochore, the proteinaceous structure on the mitotic centromere, functions as a mechanical latch that hooks onto microtubules to support directional movement of chromosomes. The structure also brings in a number of signaling molecules, such as kinases and phosphatases, which regulate microtubule dynamics and cell cycle progression. Erroneous microtubule attachment is destabilized by Aurora B-mediated phosphorylation of multiple microtubule-binding protein complexes at the kinetochore, such as the KMN network proteins and the Ska/Dam1 complex, while Plk-dependent phosphorylation of BubR1 stabilizes kinetochore–microtubule attachment by recruiting PP2A-B56. Spindle assembly checkpoint (SAC) signaling, which is activated by unattached kinetochores and inhibits the metaphase-to-anaphase transition, depends on kinetochore recruitment of the kinase Bub1 through Mps1-mediated phosphorylation of the kinetochore protein KNL1 (also known as Blinkin in mammals, Spc105 in budding yeast, and Spc7 in fission yeast). Recruitment of protein phosphatase 1 to KNL1 is necessary to silence the SAC upon bioriented microtubule attachment. One of the key unsolved questions in the mitosis field is how a mechanical change at the kinetochore upon microtubule attachment is converted to these and other chemical signals that control microtubule attachment and the SAC. Rapid progress in the field is revealing the existence of an intricate signaling network created right on the kinetochore. Here we review the current understanding of phosphorylation-mediated regulation of kinetochore functions and discuss how this signaling network generates an accurate switch that turns on and off the signaling output in response to kinetochore–microtubule attachment.     

Molecular architecture of the kinetochore-microtubule attachment site is conserved between point and regional centromeres

Point and regional centromeres specify a unique site on each chromosome for kinetochore assembly. The point centromere in budding yeast is a unique 150-bp DNA sequence, which supports a kinetochore with only one microtubule attachment. In contrast, regional centromeres are complex in architecture, can be up to 5 Mb in length, and typically support many kinetochore-microtubule attachments. We used quantitative fluorescence microscopy to count the number of core structural kinetochore protein complexes at the regional centromeres in fission yeast and Candida albicans. We find that the number of CENP-A nucleosomes at these centromeres reflects the number of kinetochore-microtubule attachments instead of their length. The numbers of kinetochore protein complexes per microtubule attachment are nearly identical to the numbers in a budding yeast kinetochore. These findings reveal that kinetochores with multiple microtubule attachments are mainly built by repeating a conserved structural subunit that is equivalent to a single microtubule attachment site.     

Synthesis of azidotubulin: a photoaffinity label for tubulin-binding proteins

A photoaffinity label for the identification of tubulin-binding proteins was synthesized from phosphocellulose-purified bovine brain tubulin and (N-hydroxysuccinimidyl)-4-azidosalicylic acid. The azidotubulin derivative retained the ability to undergo temperature-dependent microtubule assembly and disassembly. When incubated with purified tau protein, the azidotubulin and tau formed cross-linked complexes upon photoactivation. When 125I-labeled azidotubulin was used to photoaffinity label tubulin-binding proteins within the kinetochore of isolated mammalian chromosomes, a 130-kDa band was identified on autoradiographs of SDS-polyacrylamide gels of the 125I-labeled azidotubulin/chromosome preparations. The 130-kDa complex was isolated by antitubulin affinity chromatography and analyzed by immunoblotting using both antitubulin and kinetochore-specific sera obtained from human patients with the autoimmune disease scleroderma CREST. The immunoblots demonstrated that the 130-kDa band that was observed on autoradiographs was a complex of a subunit of the tubulin dimer and an 80-kDa CREST-specific kinetochore protein. The binding of azidotubulin to the 80-kDa kinetochore protein was significantly decreased when chromosomes were treated with a mixture of 9 parts underivatized tubulin to 1 part azidotubulin prior to photolysis. The formation of the 130-kDa azidotubulin/kinetochore protein complex was not inhibited by pretreating the chromosomes with CREST serum prior to incubation with azidotubulin. Azidotubulin should be a useful probe for the identification and characterization of tubulin-binding proteins.     

Isolation and characterization of a unique 15 kilodalton trypanosome subpellicular microtubule-associated protein.

A protein of 15 kDa (p15) was isolated from Trypanosoma brucei subpellicular microtubules by tubulin affinity chromatography. The protein bound tubulin specifically both in its native form and after SDS-PAGE in tubulin overlay experiments. p15 promoted both the in vitro polymerization of purified calf brain tubulin and the bundling of preformed mammalian microtubules. Immunolabeling identified p15 at multiple sites along microtubule polymers comprising calf brain tubulin and p15 as well as on the subpellicular microtubules of cryosectioned trypanosomes. Antibodies directed against p15 did not cross react with mammalian microtubules. It is suggested that p15 is a trypanosome-specific microtubule-associated protein (MAP) that contributes to the unique organization of the subpellicular microtubules.     

DNA tris-intercalation: first acridine trimer with DNA affinity in the range of DNA regulatory proteins. Kinetic studies

A trimer made up of three acridine chromophores linked by a positively charged poly(aminoalkyl) chain was synthesized as a potential tris-intercalating agent. The length of the linking chain was selected to allow intercalation of each chromophore according to the excluded site model. 1H NMR studies have shown that, at 5 mM sodium, pH 5, the acridine trimer occurred under a folded conformation stabilized by stacking interactions between the three aromatic rings. DNA tris-intercalation of the dye at a low dye/base pair ratio was shown by measurements of both the unwinding of PM2 DNA and the lengthening of sonicated rodlike DNA. The trimer exhibits a high DNA affinity for poly[d(A-T)] (Kapp = 8 X 10(8) M-1, 1 M sodium) as shown by competition experiments with ethidium dimer. Kinetic studies of both the association with poly[d(A-T)] and the exchange between poly[d(A-T)] and sonicated calf thymus DNA have been performed as a function of the ionic strength. In 0.3 M sodium the on-rate constant (k1 = 2.6 X 10(7) M-1 s-1) is similar to that reported for other monoacridines or bis(acridines), whereas the off-rate constant is much smaller (k-1 = 1.2 X 10(-4) s-1), leading to an equilibrium binding constant as large as Kapp = 2.2 X 10(11) M-1. A plot of log (k1/k-1) as a function of log [Na+] yielded a straight line whose slope shows that 5.7 ion pairs (out of 7 potential) are formed upon the interaction with DNA. From this linear relationship a Kapp value of 10(14) M-1 in 0.1 M sodium can be estimated. Such a value reaches and even goes beyond that of some DNA regulatory proteins. This acridine trimer appears to be the first synthetic ligand with such a high DNA affinity.     

High affinity DNA-microtubule associated protein interaction

We have isolated the MAP/tau proteins from twice-cycled chick brain microtubule preparations and demonstrated that they are responsible for the nitrocellulose DNA binding activity we and others have measured. Using the isolated MAP/tau proteins we then measured the apparent affinity constant K_app for the homologous chick DNA interaction and found evidence for two equilibrium affinity classes-a K_app = 6 × 10⁷ M^–1, responsible for the bulk of the DNA binding activity and a small (< 10%) higher affinity K_app = 10⁸ – 10⁹ M^–1, likely due to sequence specific binding protein species. Using the same chick brain MAP-tau protein, a heterologous interaction with D. melanogaster DNA, was found to possess just the lower affinity class-K_app = 2 × 10⁷ M^–1. Under stringent binding conditions we carried out equilibrium nitrocellulose filter binding experiments in a ternary reaction mixture at constant MAP/tau protein and ³⁵S radiolabelled chick DNA concentration using increasing and excess concentrations of competitor DNAs of different sources. The order of competitor strengths found was-chick DNA > mouse DNA > D. melanogaster = E. coli. DNA. These data and specifically the homologous DNA: protein case being the strongest competitor corroborate our previous studies using total microtubule protein and provide new evidence for a conserved interaction of a small DNA sequence class with MAP/tau protein species. Moreover, these data allow us to conclude that the conserved DNA sequence: MAP/tau protein interactions do not critically depend upon any energetic feature co-involving tubulin for their properties since tubulin is absent from these preparations.     

Identification of two distinct microtubule binding domains on recombinant rat MAP 1B.

A set of partially overlapping cDNA clones covering 9 kb of continuous sequence encoding the high molecular weight microtubule-associated protein (MAP) 1B, was isolated from a rat brain library in lambda gt11. The protein encoded was immunoreactive with monoclonal antibodies raised against calf MAP 1B, rat MAP 1X, and rat MAP 5, as shown by immunoblotting. Using Northern blot analysis, it was shown that the level of MAP 1B mRNA increased dramatically upon nerve growth factor-induced PC12 cell differentiation. The expression of polypeptides encoded by cDNA constructs, in conjunction with microtubule binding assays, revealed two separate microtubule binding domains, corresponding to sequences at the 5' and 3' end of the mRNA. As shown by DNA sequencing, the binding domain encoded by 5' terminal sequences consisted of the basic repeat motif KKEE(I/V), previously identified in mouse MAP 1B (Noble, M., S. A. Lewis, N. J. Cowan, J. Cell Biol. 109, 3367-3376 (1989)). The second binding domain, too, was found to be basic, but without any apparent repeat structure. It is concluded that single proteolytically unprocessed MAP 1B molecules would have the potential to function as microtubule cross-linkers.     

Colchicine binding to an oligomer of tubulin

An oligomeric form of tubulin present in microtubule protein prepared from mammalian brain, the 36S double ring containing tau protein, is reported to bind colchicine. Colchicine binds to each individual 6S tubulin subunit in the 36S ring without apparent effect on quarternary structure. The colchicine-oligomer complex forms by colchicine binding directly to the tubulin ring; alternatively, complexes formed by colchicine with 6S tubulin subunits associate in the presence of tau protein to form the colchicine-oligomer complex.     

A 240 kd multisubunit protein complex, CBF3, is a major component of the budding yeast centromere

A key protein component (CBF3) of the budding yeast (S. cerevisiae) centromere/kinetochore has been purified and characterized. CBF3 is a 240 kd multisubunit protein complex that binds specifically to the yeast wild-type centromere DNA (CEN), but not to nonfunctional CEN DNA containing a single base substitution in the critical CDEIII consensus sequence. When purified by affinity chromatography, CBF3 contains three protein components: CBF3A (110 kd), CBF3B (64 kd), and CBF3C (58 kd). Highly purified CBF3 requires the presence of a separate assembly factor or chaperone activity to bind to CEN DNA. Treatment with phosphatase inactivates CBF3, indicating that at least one of the CBF3 subunits must be phosphorylated for DNA binding to occur. A 56 bp region including the 26 bp CDEIII consensus is protected from DNAase I cleavage in the CBF3-CEN DNA complex.     

Binding of gossypol to purified tubulin and inhibition of its assembly into microtubules

Gossypol is a polyphenolic pigment, which is employed as a male antifertility drug. It inhibits, among other reported effects, the growth of cultured mammalian cells, spermiogenesis, flagellar motility in Trypanosoma and sperm, dynein ATPase and the lactate dehydrogenase X (LDH-X) isozyme. We have characterized the non-covalent binding of gossypol to purified calf brain tubulin in 10 mM phosphate buffer, 0.1 mM GTP pH 7.0 at 25 degrees C. Equilibrium measurements were performed by difference spectroscopy. A peak at 435 nm was produced by the perturbation of gossypol light absorption upon binding to tubulin. The experimental isotherm was fitted by 1.96 +/- 0.06 gossypol binding sites per tubulin molecule, with identical apparent equilibrium binding constants of (7.5 +/- 1.1) X 10(4) M-1. The complex formed could be separated from free gossypol by gel chromatography. Binding of gossypol was independent of the presence of 0.1 mM GTP in the buffer. Gossypol did not affect the binding of ligands to the colchicine site. Gossypol interacted with vinblastine but apparently did not bind to the vinblastine sites of tubulin. Gossypol did not displace anilinonaphthalene sulphonate (ANS) bound to tubulin, but caused a strong (fivefold) quenching of its fluorescence. This indicated that gossypol probably binds in the vicinity of the ANS site of tubulin. Gossypol inhibited in vitro microtubule assembly at the same concentration range employed in the binding studies. An increase in the critical protein concentration required for polymerisation was observed, most simply interpreted by a stoichiometric mechanism. Gossypol did not induce any noticeable distortion of the microtubules observed under the electron microscope. This compound constitutes a new tubulin ligand and an inhibitor of microtubule assembly in vitro.     

Tubulin-G protein interactions involve microtubule polymerization domains

It has been suggested that elements of the cytoskeleton contribute to the signal transduction process and that they do so in association with one or more members of the signal-transducing G protein family. Relatively high-affinity binding between dimeric tubulin and the alpha subunits of Gs and Gi1 has also been reported. Tubulin molecules, which exist in solution as alpha beta dimers, have binding domains for microtubule-associated proteins as well as for other tubulin dimers. This study represents an attempt to ascertain whether the association between G proteins and tubulin occurs at one of these sites. Removal of the binding site for MAP2 and tau from tubulin by subtilisin proteolysis did not influence the association of tubulin with G protein, as demonstrated in overlay studies with [125I]tubulin. A functional consequence of that association, the stable inhibition of synaptic membrane adenylyl cyclase, was also unaffected by subtilisin treatment of tubulin. However, ring structures formed from subtilisin-treated tubulin were incapable of effecting such inhibition. Stable G protein-tubulin complexes were formed, and these were separated from free tubulin by Octyl-Sepharose chromatography. Using this methodology, it was demonstrated that assembled microtubules bound G protein quite weakly compared with tubulin dimers. The alpha subunit of Gi1 and, to a lesser extent, that of Go were demonstrated to inhibit microtubule polymerization. In aggregate, these data suggest that dimeric tubulin binds to the alpha subunits of G protein at the sites where it binds to other tubulin dimers during microtubule polymerization. Interaction with signal-transducing G proteins, thus, might represent a role for tubulin dimers which is independent of microtubule formation.     

The periodic association of MAP2 with brain microtubules in vitro

Several high molecular weight polypeptides have been shown to quantitatively copurify with brain tubulin during cycles of in vitro assembly-disassembly. These microtubule-associated proteins (MAPs) have been shown to influence the rate and extent of microtubule assembly in vitro. We report here that a heat-stable fraction highly enriched for one of the MAPs, MAP2 (mol wt approximately 300,000 daltons), devoid of MAP1 (mol wt approximately 350,000 daltons), has been purified from calf neurotubules. This MAP2 fraction stoichiometrically promotes microtubule assembly, lowering the critical concentration for tubulin assembly to 0.05 mg/ml. Microtubules saturated with MAP2 contain MAP2 and tubulin in a molar ratio of approximately 1 mole of MAP2 to 9 moles of tubulin dimer. Electron microscopy of thin sections of the MAP2-saturated microtubules fixed in the presence of tannic acid demonstrates a striking axial periodicity of 32 +/- 8 nm.     

High affinity DNA-microtubule interactions: evidence for a conserved DNA-MAP interaction involving unusual high CsCl density repetitious DNA families

We have examined high affinity interactions of chick brain microtubule proteins with ³⁵S labelled tracer DNAs from chick, mouse and D. melanogaster under equilibrium conditions by the nitrocellulose filter binding technique. Ternary reaction mixtures of the above two components and a third component, an excess of unlabelled competitor DNA from either E. coli., mouse, D. melanogaster or chick, were used to measure small fractions of DNA in each case (1–4%) bound to microtubule protein under high stringency- large competitor DNA concentration and 0.5 M NaCl. As seen in part previously (Marx, K.A. and Denial, T. (1985) in The Molecular Basis of Cancer, 172B, 65–75 (Rein, ed), A. Liss, N.Y.) the measured order of competitor DNA strengths was identical for all three tracer DNAs. That is: chick > mouse > D. melanogaster > E. coli competitor DNA. Since the homologous interaction, chick competitor DNA with chick brain microtubule protein, is always the strongest interaction measured, we interpret this as evidence for a conserved protein-DNA sequence interaction. ³⁵S chick DNA tracer sequences, isolated from nitrocellulose filters following the stringent binding in the presence of 0.9 mM^–1 E. coli. competitor DNA, was used in driven reassociation reactions with total chick driver DNA. This fraction was found to be significantly enriched in repetitive chick DNA sequences. Since we have observed a similar phenomenon in mouse, we then compared the stringent binding mouse sequences and showed that the bulk of these sequences did not cross-hybridize with total chick DNA. Finally, all three ³⁵S tracer DNAs binding to nitrocellulose were isolated and sedimented to equilibrium on CsCl density gradients. The CsCl density distributions from all three DNAs showed significant (100-fold) enrichment in classical satellite DNAs as well as higher enrichment in two very unusual high CsCl density families of DNA (1.720–1.740 g/cm³; 1.750–1.765 g/cm³). These families are never observed as distinct bands in total DNA CsCl gradients, nor could we isolate them in purified tubulin control binding experiments. This apparently general phenomena may be identifying some of the sequence families involved in the high affinity microtubule interaction, which appears to be conserved in evolution.     

Molecular architecture of a kinetochore-microtubule attachment site

Kinetochore attachment to spindle microtubule plus-ends is necessary for accurate chromosome segregation during cell division in all eukaryotes. The centromeric DNA of each chromosome is linked to microtubule plus-ends by eight structural-protein complexes. Knowing the copy number of each of these complexes at one kinetochore-microtubule attachment site is necessary to understand the molecular architecture of the complex, and to elucidate the mechanisms underlying kinetochore function. We have counted, with molecular accuracy, the number of structural protein complexes in a single kinetochore-microtubule attachment using quantitative fluorescence microscopy of GFP-tagged kinetochore proteins in the budding yeast Saccharomyces cerevisiae. We find that relative to the two Cse4p molecules in the centromeric histone, the copy number ranges from one or two for inner kinetochore proteins such as Mif2p, to 16 for the DAM-DASH complex at the kinetochore-microtubule interface. These counts allow us to visualize the overall arrangement of a kinetochore-microtubule attachment. As most of the budding yeast kinetochore proteins have homologues in higher eukaryotes, including humans, this molecular arrangement is likely to be replicated in more complex kinetochores that have multiple microtubule attachments.     

Linkage between ligand binding and control of tubulin conformation.

The effect of both antimitotic drugs and nucleotide analogues on the magnesium-induced self-association of purified tubulin into 42S double rings has been examined by sedimentation velocity. In the absence of magnesium, all complexes sedimented as the 5.8S species. The binding of colchicine to tubulin led to a small but consistent (-0.1 to -0.2 kcal/mol) enhancement in the self-association of tubulin alpha-beta dimers. In the absence of nucleotide at the exchangeable site, tubulin retained a weak ability (K2 = 7.5 x 10(3) M-1) to self-associate, which was unchanged by the addition of guanosine or GMP. Analogues with altered P-O-P bonds (GMPPCP, GMPPNP) did not support ring formation at the protein concentrations examined, although GMPPCP supported microtubule assembly. When the exchangeable site was occupied by nucleotides altered on the gamma-phosphate (GTP gamma S, GTP gamma F), rings were formed; tubulin-GTP gamma F formed rings to an extent slightly greater than did tubulin-GTP, and tubulin-GTP gamma S to about the same extent as tubulin-GDP. Both of these analogues are inhibitors of microtubule assembly. These results are consistent with a model [Melki, R., Carlier, M.-F., Pantaloni, D., & Timasheff, S. N. (1989) Biochemistry 28, 9143-9152] in which an equilibrium exists between straight (microtubule-forming) and curved (ring-forming) conformations of tubulin. Furthermore, the present results indicate that the "switch" which controls the nature of the final polymeric product via free energy linkages is the occupancy of the gamma-phosphate binding locus of the exchangeable site by a properly coordinated metal-nucleotide complex.     

Purification of the yeast centromere binding protein CP1 and a mutational analysis of its binding site

CP1 is a yeast protein which binds to the highly conserved DNA element I (CDEI) of yeast centromeres. We have purified CP1 to near homogeneity; it is comprised of a single polypeptide of molecular weight 58,400. When bound to yeast CEN3 DNA, CP1 protects a 12-15-base pair region centered over CDEI. Methylation interference experiments show that methylations of residues located outside of the 8-base pair CDEI sequence have no detectable effect on CP1 binding, suggesting that the DNA sequences important for CP1 recognition are confined to the CDEI octanucleotide. The equilibrium constant for CP1 binding to CEN3 DNA is relatively low, 3 x 10(8) M-1. Using a novel method to determine relative DNA binding constants, we analyzed the effect of CDEI mutations on CP1 binding. A C to T point mutation at position 5 (CO1) reduces the equilibrium constant about 35-fold, while the insertion of an additional T at this position (CAT) reduces the equilibrium constant 1,400-fold. The effect of these mutations on mitotic centromere function in vivo was assessed using a plasmid stability assay. While the CO1 mutation had a slight effect, the CAT mutation significantly impaired function, implying that CP1 binding is required for the optimal mitotic function of yeast centromeres.     

Binding of the essential Saccharomyces cerevisiae kinetochore protein Ndc10p to CDEII

Chromosome segregation at mitosis depends critically on the accurate assembly of kinetochores and their stable attachment to microtubules. Analysis of Saccharomyces cerevisiae kinetochores has shown that they are complex structures containing >/=50 protein components. Many of these yeast proteins have orthologs in animal cells, suggesting that key aspects of kinetochore structure have been conserved through evolution, despite the remarkable differences between the 125-base pair centromeres of budding yeast and the Mb centromeres of animal cells. We describe here an analysis of S. cerevisiae Ndc10p, one of the four protein components of the CBF3 complex. CBF3 binds to the CDEIII element of centromeric DNA and initiates kinetochore assembly. Whereas CDEIII binding by Ndc10p requires the other components of CBF3, Ndc10p can bind on its own to CDEII, a region of centromeric DNA with no known binding partners. Ndc10p-CDEII binding involves a dispersed set of sequence-selective and -nonselective contacts over approximately 80 base pairs of DNA, suggesting formation of a multimeric structure. CDEII-like sites, active in Ndc10p binding, are also present along chromosome arms. We propose that a polymeric Ndc10p complex formed on CDEII and CDEIII DNA is the foundation for recruiting microtubule attachment proteins to kinetochores. A similar type of polymeric structure on chromosome arms may mediate other chromosome-spindle interactions.