Nature Structural Biology
9, 719 - 724 (2002)
Published online: 23 September 2002; | doi:10.1038/nsb848
The Sak polo-box comprises a structural domain sufficient for mitotic subcellular localizationGenie C. Leung1, 2, John W. Hudson1, Anna Kozarova1, Alan Davidson2, James W. Dennis1, 2
& Frank Sicheri1, 21 Program in Molecular Biology and Cancer, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto, Ontario M5G 1X5, Canada. 2 Department of Molecular and Medical Genetics, University of Toronto, Toronto, Ontario M5S 1A8, Canada.
Correspondence should be addressed to Frank Sicheri sicheri@mshri.on.caThe small family of polo-like kinases (Plks) includes Cdc5 from Saccharomyces cerevisiae, Plo1 from Schizosaccharomyces pombe, Polo from Drosophila melanogaster and the four mammalian genes Plk1, Prk/Fnk, Snk and Sak. These kinases control cell cycle progression through the regulation of centrosome maturation and separation, mitotic entry, metaphase to anaphase transition, mitotic exit and cytokinesis. Plks are characterized by an N-terminal Ser/Thr protein kinase domain and the presence of one or two C-terminal regions of similarity, termed the polo box motifs. These motifs have been demonstrated for Cdc5 and Plk1 to be required for mitotic progression and for subcellular localization to mitotic structures. Here we report the 2.0 Å crystal structure of a novel domain composed of the polo box motif of murine Sak. The structure consists of a dimeric fold with a deep interfacial cleft and pocket, suggestive of a ligand-binding site. We show that this domain forms homodimers both in vitro and in vivo, and localizes to centrosomes and the cleavage furrow during cytokinesis. The requirement of the polo domain for Plk family function and the unique physical properties of the domain identify it as an attractive target for inhibitor design.
The polo-like kinases (Plks) play several overlapping roles in cell cycle progression (reviewed in refs 1,
2,
3). Mutation of the Plk proteins Polo in Drosophila, Plo1 in Schizosaccharomyces pombe and Cdc5 in Saccharomyces cerevisiae causes mitotic defects, including monopolar spindles, aberrant chromosome segregation and failure of cytokinesis4,
5,
6,
7,
8. The targeted disruption of Sak in mouse is embryonic lethal at gastrulation, with cells arresting in late stage mitosis and displaying failure of cytokinesis9. In S. cerevisiae, mitotic defects arising from the loss of cdc5 function can be rescued by the heterologous expression of mammalian Plk1 (ref. 10) or Prk/Fnk11.
The Plks localize to characteristic mitotic structures during cell cycle progression, presumably to promote the interaction of the enzymes with specific substrates and effectors. Plk1, Prk/Fnk, Cdc5, Plo1, Polo and Sak localize to centrosomes in early M phase and/or to the cleavage furrow or mother bud neck during cytokinesis9,
12,
13,
14,
15,
16,
17. Mutational analyses of Cdc5 and Plk1 have demonstrated a requirement and sufficiency of the polo box motifs for subcellular localization13,
15. In addition, these studies have demonstrated a requirement of proper subcellular localization for Plk family function. Interestingly, although most Plks possess two polo box motifs, the Sak orthologs possess only one. Because the subcellular localization of Sak conforms to that of the other Plks, the functional relevance of this difference remains to be determined.
To further our understanding of the structure and function of the polo box motif in Plk family subcellular localization, we have expressed and characterized a protein fragment encompassing the polo box motif of Sak (residues 839−925). Using limited proteolysis and mass spectrometry, we have found that the polo box motif comprises an autonomously folding unit, which we designate the polo domain, that behaves as a dimer in solution, as indicated by size exclusion chromatography and static light scattering (SLS) analysis (SLS molecular weight is 22.6  0.9 kDa versus the predicted monomer molecular weight of 10.8 kDa). We have crystallized this domain and determined its structure using the selenomethione (SeMet) MAD method. A comprehensive structure-based sequence alignment of the Plk family polo domains is shown (Fig. 1).
 | | Figure 1. Structure-based sequence alignment of the Plk family polo domains. | | |  | The polo domains from Sak orthologs are shown on top, and polo domains 1 and 2 from all other Plks are shown in the middle and bottom, respectively. The secondary structure of the polo domain of Sak is indicated above the alignment. Residue numbers for the start of each amino acid sequence are shown on the left. Conserved hydrophobic core residues are green or yellow (green denotes hydrophobic residues conserved in all polo domains and yellow denotes hydrophobic residues conserved within the first or second polo domain); Asp residues, red; Asn residues, orange; Lys residues, blue; and Arg residues, turquoise. Positions are identified as conserved if >85% of residues are identical or are hydrophobic in nature. Conserved dimer interface (red arrow), pocket (filled circle) and cleft (open triangle) positions are indicated. The linker regions between polo domains 1 and 2 are outlined in purple. Species notation is as follows: m = Mus musculus, h = Homo sapiens, Dm = Drosophila melanogaster, Dr = Danio rerio, r = Rattus norvegicus, Ce = Caenorhabditis elegans, u = Hemicentrotus pulcherrimus, Tb = Trypansoma brucei and asterisk = partial EST sequences available only.
 Full Figure and legend (157K) |
| |
Structure description
The crystal structure of the polo domain of Sak is dimeric, consisting of two  -helices and two six-stranded  -sheets (Fig. 2a,b). Analysis by VAST18 identifies this structure as a new protein fold. The topology of one polypeptide subunit of the dimer consists of, from its N- to C-terminus, an extended strand segment (Ex1), five -strands (1−5), one -helix (A) and a C-terminal -strand (6). -strands 6, 1, 2 and 3 from one subunit form a contiguous antiparallel -sheet with -strands 4 and 5 from the second subunit. The two -sheets pack with a crossing angle of 110°, orienting the hydrophobic surfaces inward and the hydrophilic surfaces outward. Helix A, which is colinear with -strand 6 of the same polypeptide, buries a large portion of the non-overlapping hydrophobic -sheet surfaces. Interactions involving helices A comprise a majority of the hydrophobic core structure and also the dimer interface. The total surface area buried by dimer formation is 2,448 Å2. Overall, the dimeric structure is clam-like (60 Å  44 Å 22 Å), hinged at one end through the seamless association of -strand 3 from each subunit (Fig. 2b). Extending inwards from the mouth of the structure is a deep cavity of  17 Å 8 Å 12 Å (Fig. 2a,b). The entry to this cavity is divided in two by the contact of the Trp 853 side chains on -strand 1 from each polypeptide of the dimer. Strands Ex1 from each polypeptide designate the proximal ends of the cleft (Fig. 2b).
| | |
Residues of Sak that comprise much of the polo domain hydrophobic core are highly conserved across the Plks (Fig. 1). Mutation of one hydrophobic core position, L427A in Plk1 (equivalent to Leu 857 in Sak), disrupts the ability of Plk1 to complement the cdc5-1 temperature-sensitive mitotic arrest phenotype in yeast13. We predict that this mutation disrupts the overall polo domain fold. A large proportion of the conserved hydrophobic core residues (13 out of 19) also participate in dimer formation. Only two charged residues, equivalent to Asp 868 and Lys 906 in Sak, are conserved among most polo domains, and these residues participate in dimerization through a 2.6 Å intermolecular salt bridge in the crystal structure (Fig. 2a,b). Together, these observations suggest that the dimeric fold revealed by the crystal structure may be a functionally relevant conformation accessible by all polo domains.
The presence of two polo domains in all Plks other than the Sak orthologs raises an interesting possibility for an intramolecular mode of polo domain dimerization. In support of this possibility is a covariance in primary structure across paired polo domains involving the conserved salt bridge (Asp 868 and Lys 906) and a dimer interface residue equivalent to Val 846 in Sak (Fig. 2a,b). Val 846, which lies in close proximity to the conserved salt bridge, is substituted with an Asp residue in the first, but not the second, polo domain of the Plks. This hydrophobic-to-charged amino acid substitution seems to be compensated by the K906R substitution in the second polo domain. Our modeling studies suggest that this concerted substitution would allow for the formation of a bidentate salt interaction between the Arg and two Asp residues, facilitated by the increased hydrogen-bonding capacity of the guanidinium group of the Arg residue (Fig. 2a, inset). In further support of the possibility for an intramolecular mode of dimerization, the linker region between tandem polo domains is sufficiently long (21−37 amino acids) in all Plks to bridge the 36 Å distance that separates the N- and C-termini of opposing dimer chains in the polo domain crystal structure.
Although less conserved than the hydrophobic core and dimer interface structure, the interfacial cleft and pocket display properties suggestive of a functionally important surface. Of the 19 conserved hydrophobic positions in the polo domain alignment, nine contribute side chains to the outer cleft and inner pocket (Fig. 1). Modeling of the polo domain sequences of Fnk/Prk, Snk and Plk1 to form an intramolecular dimer shows that the approximate dimensions and hydrophobic character of the pocket and cleft region are also generally preserved (data not shown). Polo domain mutations in Plk1 and Cdc5 that disrupt localization or the ability to complement the cdc5-1 temperature-sensitive mutation in yeast map mostly to the interfacial cleft region13,
15. These include the mutations W414F and V415A in Plk1, or W517F and V518A in Cdc5 (equivalent to Lys 844 and Ser 845 in Sak), which locate within or just prior to strand Ex1 at the proximal ends of the cleft. Indeed, the cdc5-1 temperature-sensitive mutation itself (P511L) maps to the region proceeding strand Ex1. A third mutation in Plk1, N437D (equivalent to Asn 867 in the 2−3 linker of Sak), is positioned to influence the conformation of strand Ex1. In the polo domain structure of Sak, Asn 867 forms intramolecular hydrogen bonds with backbone amino and carbonyl groups of the Ex1 strand residues Phe 847 and Ser 845. These observations suggest that the interfacial cleft and pocket region is functionally important, possibly composing a ligand-binding site.
Polo domain self-association in vivo
To investigate the ability of the polo domain of Sak to dimerize in vivo, we have generated differentially tagged mammalian expression constructs and tested them for self-association in vivo using a coimmunoprecipitation assay. The Myc-tagged polo domain of Sak (Sakpd) was coimmunoprecipitated with a FLAG-tagged polo domain when both constructs were transfected into NIH 3T3 cells (Fig. 3a). This confirms the potential of the isolated domain to self-associate in vivo. To determine whether full-length Sak can self-associate and whether self-association is polo domain dependent, immunoprecipitations were performed with similarly tagged expression constructs (Fig. 3b). Immunoprecipitation of FLAG-tagged, full-length Sak yielded Myc-tagged Sak, confirming the self-association of full-length Sak in vivo (lane 6, Fig. 3c). However, deletion of the polo domain (Sak pd) did not abolish this association (lane 7), whereas it was eliminated by a more extensive C-terminal deletion, Sak(pd+241) (lane 8). Further analysis revealed that the 241-amino acid region N-terminal to the polo domain, Sak241, was sufficient for self-association (lane 10) and was also able to associate with regions N-terminal (lane 9), but not C-terminal (lane 11), to itself. A BLAST19 analysis of the primary structure of Sak241 reveals high sequence conservation among Sak orthologs but not other Plk family members, and analysis with SMART20 and PROSITE21 reveals no similarity to known motifs or domains involved in protein−protein interaction. Together these data suggest that the polo domain of Sak can self-associate in vivo but regions N-terminal to the polo domain can also mediate the self-association of the full-length molecule.
| | Figure 3. The polo domain of Sak can self-associate in vivo but Sak may use several mechanisms for self-association. | | |  | a, The polo domain of Sak can self-associate in vivo. NIH 3T3 cells were transfected with FLAG3-tagged polo domain (FLAG−Sakpb), Myc-tagged polo domain (Myc−Sakpb) or both, as indicated. Immunoprecipitations were performed using an antibody to FLAG and probed with anti-Myc. Myc−Sakpb coimmunoprecipitated with FLAG−Sakpb from cells that were transfected with both constructs but not those that were singly transfected. Reciprocal immunoprecipitations revealed identical results (data not shown). b, Sak constructs generated for coimmunoprecipitation assays. Numbers indicate the first and last amino acids for each construct. The kinase domain and polo domain are illustrated by the hatched and black regions, respectively. N-terminal tagged Myc and N-terminal tagged FLAG3 constructs were generated for each construct. (+) or (-) indicate association or lack of association as observed by coimmunoprecipitations shown in (c). c, Full-length Sak can dimerize in a polo domain-independent manner. NIH 3T3 cells were transfected with the constructs illustrated in (b), as indicated. Untransfected and single transfected Myc-tagged controls are shown in lanes 1−5, and double transfected coimmunoprecipitation experiments are shown in lanes 6−11. Immunoblots of the lysates demonstrate that all constructs are expressed. Immunoprecipitations were performed using anti-FLAG and probed with anti-Myc. As shown in lane 6, Myc-tagged Sak coimmunoprecipitated with FLAG3-tagged Sak, showing that full-length Sak can self associate. Deletion of the polo domain (Sakpd) did not abolish this association (lane 7), showing that self-association of full-length Sak does not require the polo domain. A larger C-terminal deletion of an additional 241 residues, Sak(pd+241), did not self-associate by coimmunoprecipitation (lane 8). The signal in lane 8, which is larger than the predicted 72 kDa mass for Myc-Sak(pd+241), is a result of overflow from lane 7. Lanes 9 and 10 illustrate coimmunoprecipitation of the 241 amino acid region, Sak241, with Sak(pd+241) (lane 9) and with itself (lane 10). Myc-tagged Sak241 did not coimmunoprecipitate with the polo domain, Sakpd (lane 11). Immunoprecipitation of the single-transfected Myc-tagged constructs with anti-FLAG confirmed that the observed interactions were not due to nonspecific binding of the Myc-tagged constructs (lanes 2−5). The asterisk indicates the positions of -Myc cross-reactive bands at 21 kDa and 50 kDa.
Full Figure and legend (90K) |
| |
Polo domain subcellular localization
To investigate the role of the polo domain in the subcellular localization of Sak, enhanced green fluorescent protein (EGFP) fusion constructs of Sak, Sakpd, Sak(pd+241), Sak241 and Sakpd were transiently transfected into NIH 3T3 cells and examined using immunofluorescence. EGFP−Sak colocalizes in cells with  -tubulin and actin, which indicate the positions of centrosomes and the cleavage furrow, respectively (panel i, Fig. 4a,c). Localization to these structures has been demonstrated for full-length Plk1, Cdc5 and Sak9,
13,
15. Our experiments show that the isolated polo domain of Sak localizes to centrosomes and the cleavage furrow (Fig. 4a, panel iii, Fig. 4c, panel ii), which is consistent with previous observations for larger C-terminal protein fragments encompassing the polo domains of Cdc5 and Plk1 (refs 15,22). Unexpectedly, deletion of the polo domain (Sakpd) did not abolish the subcellular localization of Sak (Fig. 4a, panel ii), although the larger of two C-terminal deletions, Sak(pd+241), did reduce the efficiency of localization to centrosomes from 93 to 24% in comparison to full length Sak (Fig. 4b). Sak241 also localizes efficiently to centrosomes, demonstrating that residues 596−836 of Sak are also sufficient for subcellular localization (Fig. 4b). These observations conflict with the results of mutational studies of Plk1 and Cdc5 in yeast in which the polo domains seem to be essential for localization13,
15. This discrepancy may reflect the presence of a second localization domain unique to Sak or alternatively may reflect the ability of regions outside of the polo domain to promote an association with endogenous Sak in NIH 3T3 cells.
| | Figure 4. Subcellular localization of EGFP-fusion proteins demonstrate that the polo domain of Sak is sufficient for localization. | | |  | a,c, Localization of EGFP−Sak, EGFP−Sakpd and EGFP−Sakpd. Cells were stained with anti--tubulin or TRITC−phalloidin to indicate the positions of the centrosomes and actin cleavage furrow respectively. EGFP−Sak localizes to centrosomes (a, panel i) and the cleavage furrow (c, panel i). Deletion of the polo domain (Sakpd) does not abolish subcellular localization (a, panel ii) and the polo domain itself localizes to centrosomes (a, panel iii) and the cleavage furrow (c, panel ii). Localization of Sak(pd+241), Sak241, and EGFP control are not shown but quantified results are shown in (b). b, Bar graph representing the percentage of cells showing centrosomal localization with a sample population of n = 100, scored in triplicate.
Full Figure and legend (106K) |
| |
Summary and implications
We have shown that the polo domain of Sak forms dimers both in vitro and in a crystal environment, self-associates in vivo and localizes to mitotic structures. The conservation of the hydrophobic core and dimer interface residues, the presence of two copies of the polo domain in most Plks and the covariance across tandem polo domains in most Plks suggest that the ability to adopt a dimeric conformation may be a general characteristic of all polo domains and that dimerization may occur in an intramolecular manner for some family members.
The identity of the interacting partner(s) that mediate the localization of the polo domain to specific mitotic structures are currently unknown. The identification of these partners will be key to resolving issues of polo domain-binding specificity and regulation that impact Plk family function. Indeed, Plk1 and Polo, unlike Sak, localize to centromeres, suggesting that polo domain specificity may be at play. Knowledge of the interacting partners will also help to address whether the dimeric conformation of the polo domain is functionally competent for localization or whether it represents a transient regulatory state. Numerous candidate partners for the Plk polo domains have been identified, including septins8; Spc72; Smc1, Smc3 and Irr1 (ref. 23); -, - and -tubulin24; Bfa1 (ref. 25); Mid1p26; cyclin B1; Scc1; Cdc16; Cdc 27; MKLP-1 and Hsp90 (reviewed in ref. 1). Although the subcellular localizations of many of these proteins overlap with that of the Plks, additional work is required to resolve which of these proteins interact directly with the Plks through their polo domains.
The de-regulation of Plks alters mitotic checkpoints, chromosome stability and can lead to tumor development27,
28. Indeed, Plk1 is overexpressed in many human tumors29,
30,
31,
32 and causes malignant transformation when overexpressed in NIH 3T3 cells33. In addition, over expression of a kinase-deficient form of Plk1 results in cell death, an apparent dominant-negative effect that is more pronounced in tumor cells than nontransformed cells34. This identifies the Plks as potential targets for cancer therapy. The requirement of the polo domain for Plk family function and, in contrast to the catalytic domain, its exclusive presence in this small family of proteins that regulate mitotic progression suggest that the polo domain itself may serve as a good target for intervention. Indeed, the large semi-enclosed cleft and pocket with its partial hydrophobic character seems well suited for the design of small molecule inhibitors.
Methods
Protein expression, mutagenesis and purification.
The polo domain of Sak (residues 839−925), which was delimited by proteolysis and mass spectrometry, was expressed in Escherichia coli as a GST-fusion protein using the pGEX-2T vector (Pharmacia). The QuikChange kit (Stratagene) was used to generate the double site-directed mutant C909L/V874M to improve long-term protein stability and for phasing purposes. Protein was purified by affinity chromatography using glutathione-sepharose (Pharmacia). Bound protein was eluted by cleavage with thrombin (Sigma). Eluate was applied to a HiQ ion-exchange column under low salt conditions. The flow-through containing the polo domain was concentrated to 1 mM and then applied to a Superdex 75 gel filtration column (Pharmacia) for final purification and characterization by static light scattering as described35.
Crystallization and data collection.
Hanging drops containing 1  l of 50 mg ml-1 native or mutant protein in 20 mM HEPES, pH 8.0, and 5 mM dithiothreitol (DTT) were mixed with equal volumes of reservoir buffer containing 100 mM Tris, pH 7.0, 32.5% (v/v) Jeffamine M-600 (Hampton) and 200 mM MgCl2. Hexagonal-like crystals of 0.10 0.10 0.03 mm were obtained overnight for both native and mutant proteins. The asymmetric unit of the crystals consists of two polypeptides that form an interdigitated dimer. The crystals belong to the space group P3212 (a = b = 51.782 Å and c = 146.941 Å).
MAD diffraction data was collected on frozen crystals at the Structural Biology Center 19-BM and BIOCARS 14-BMC at the Advanced Photon Source at Argonne National Laboratory. Data processing and reduction was carried out using HKL2000 (ref. 36). Heavy atom sites were identified using CNS37, and phasing, density modification and experimental electron density map calculation was performed using SHARP38.
Model building and refinement.
Model building was performed using O39. A starting model composed of 85% of the polypeptide sequence was refined using CNS37. Bulk solvent correction was applied during refinement, and simulated annealing protocols were used. The remaining structure was built into 2|Fo - Fc| electron density maps generated with CNS. The final refinement statistics are shown in Table 1. The first and last six residues of the polo domain fragment are disordered (residues 839−844 and residues 920−925) and have not been modeled. Analysis by Procheck40 indicated that no residues occupy disallowed regions of the Ramachandran plot and 94% occupy the most favored regions.
Sak protein localization.
Full length Sak (residues 1−925), Sakpb (residues 1−823), Sak241 (residues 596−836), Sak(pb+241) (residues 1−595) and Sakpb (residues 824−925) were fused to EGFP in the vector pEGFP-C1 (Clontech). NIH 3T3 murine fibroblast cells were maintained in DMEM containing 10% FBS. For transient gene expression, cells at 20−30% confluence on glass cover slips were transiently transfected with pEGFP-Sak, pEGFP-Sakpb, pEGFP-Sak(pb+241), Sak241, pEGFP-Sakpb or pEGFP-C1 with Effectene (Qiagen). Cells were released from 48 h of serum starvation by addition of fresh media containing 10% FBS and fixed at intervals as they proceeded through the cell cycle. Cells were processed by rinsing twice in PBS, fixed with 3.7% para-formaldehyde in PBS for 12 min and permeabilized for 5 min in PBS 0.5% Triton X-100. Actin microfilaments were stained with a 1:100 dilution of phalloidin-TRITC (Sigma) in PBS. -tubulin was stained with a 1:200 dilution of anti--tubulin (Sigma) in Tris/saline + 0.1% Tween20 at 20 °C for 40 min. Cells were washed three times in Tris/saline + 0.1% Tween20 and incubated in a 1:500 dilution of rhodamine-conjugated goat anti-mouse (Pierce) for 40 min. Nuclei were stained with Hoechst 33258 (Molecular Probes) in PBS for 1 min. Images were obtained using an Olympus IX-70 inverted microscope equipped with a Princeton CCD camera and Deltavision Deconvolution microscopy software (Applied Precision).
Quantification of EGFP-fusion proteins showing centrosomal localization was performed by counting 3 independent populations of 100 cells. Because of the inability to generate large populations of cells undergoing cytokinesis, the quantification of EGFP-fusion protein localization to the cleavage furrow was not scored. The Sakpd construct (residues 1−823) fused to EGFP differed from the FLAG- and Myc-tagged Sakpd construct (residues 1−836) prepared for coimmunprecipitation studies by a deletion of 13 amino acids from the C-terminus. The Sakpd construct (residues 824−925) fused to EGFP differs from the FLAG- and Myc-tagged Sakpd (residues 819−925) prepared for coimmunoprecipitation studies by the deletion of five amino acids at the N-terminus.
Immmunoprecipitation.
NIH 3T3 murine fibroblast cells were maintained in DMEM containing 10% FBS. For transient gene expression, cells at 30−40% confluence were transfected using Effectene (Qiagen). After 24 h post transfection, cells were lysed in 50 mM Tris, pH 7.5, 100 mM NaCl, 1 mM EDTA and 0.5% Triton-X 100. Immunoprecipitations were performed using antibody to FLAG (Sigma) and Protein G Sepharose (Pharmacia) according to product specifications. The Protein G Sepharose matrix was washed three times with lysis buffer. Western blots were performed using a 1:200 dilution of anti-Myc (Santa Cruz Biotech) or a 1:4,000 dilution of anti-FLAG (Sigma).
Coordinates.
The Sak polo domain coordinates and structure factor files have been deposited in the Protein Data Bank (accession code 1MBY).
Received 27 June 2002; Accepted 28 August 2002; Published online: 23 September 2002.
REFERENCES
- Donaldson, M.M., Tavares, A.A., Hagan, I.M., Nigg, E.A. & Glover, D.M. J. Cell Sci. 114, 2357−2358 (2001). | PubMed | ISI | ChemPort |
- Glover, D.M., Hagan, I.M. & Tavares, A.A. Genes Dev. 12, 3777−3787 (1998). | PubMed | ISI | ChemPort |
- Nigg, E.A. Curr. Opin. Cell Biol. 10, 776−783 (1998). | Article | PubMed | ISI | ChemPort |
- Sunkel, C.E. & Glover, D.M. J. Cell. Sci. 89, 25−38 (1988). | PubMed | ISI |
- Carmena, M. et al. J. Cell Biol. 143, 659−671 (1998). | Article | PubMed | ISI | ChemPort |
- Ohkura, H., Hagan, I.M. & Glover, D.M. Genes Dev. 9, 1059−1073 (1995). | PubMed | ISI | ChemPort |
- Simchen, G., Kassir, Y., Horesh-Cabilly, O. & Friedmann, A. Mol. Gen. Genet. 184, 46−51 (1981). | Article | PubMed | ISI | ChemPort |
- Song, S. & Lee, K.S. J. Cell. Biol. 152, 451−469 (2001). | Article | PubMed | ISI | ChemPort |
- Hudson, J.W. et al. Curr. Biol. 11, 441−446 (2001). | Article | PubMed | ISI | ChemPort |
- Lee, K.S. & Erikson, R.L. Mol. Cell. Biol. 17, 3408−3417 (1997). | PubMed | ISI | ChemPort |
- Ouyang, B. et al. J. Biol. Chem. 272, 28646−28651 (1997). | Article | PubMed | ChemPort |
- Golsteyn, R.M., Mundt, K.E., Fry, A.M. & Nigg, E.A. J. Cell Biol. 129, 1617−1628 (1995). | Article | PubMed | ISI | ChemPort |
- Lee, K.S., Grenfell, T.Z., Yarm, F.R. & Erikson, R.L. Proc. Natl. Acad. Sci. USA 95, 9301−9306 (1998). | Article | PubMed | ChemPort |
- Wang, Q. et al. Mol. Cell. Biol. 22, 3450−3459 (2002). | Article | PubMed | ISI | ChemPort |
- Song, S., Grenfell, T.Z., Garfield, S., Erikson, R.L. & Lee, K.S. Mol. Cell. Biol. 20, 286−298 (2000). | Article | PubMed | ISI | ChemPort |
- Mulvihill, D.P., Petersen, J., Ohkura, H., Glover, D.M. & Hagan, I.M. Mol. Biol. Cell 10, 2771−2785 (1999). | PubMed | ISI | ChemPort |
- Adams, R.R., Tavares, A.A., Salzberg, A., Bellen, H.J. & Glover, D.M. Genes Dev. 12, 1483−1494 (1998). | PubMed | ISI | ChemPort |
- Madej, T., Gibrat, J.F. & Bryant, S.H. Proteins 23, 356−369 (1995). | PubMed | ISI | ChemPort |
- Wheeler, D.L. et al. Nucleic Acids Res. 30, 13−16 (2002). | Article | PubMed | ISI | ChemPort |
- Schultz, J., Milpetz, F., Bork, P. & Ponting, C.P. Proc. Natl. Acad. Sci. USA 95, 5857−5864 (1998). | Article | PubMed | ChemPort |
- Falquet, L. et al. Nucleic Acids Res. 30, 235−238 (2002). | Article | PubMed | ISI | ChemPort |
- Jang, Y.J., Lin, C.Y., Ma, S. & Erikson, R.L. Proc. Natl. Acad. Sci. USA 99, 1984−1989 (2002). | Article | PubMed | ChemPort |
- Ho, Y. et al. Nature 415, 180−183 (2002). | Article | PubMed | ISI | ChemPort |
- Feng, Y. et al. Biochem. J. 339, 435−442 (1999). | Article | PubMed | ISI | ChemPort |
- Hu, F. et al. Cell 107, 655−665 (2001). | Article | PubMed | ISI | ChemPort |
- Bahler, J. et al. J. Cell Biol. 143, 1603−1616 (1998). | Article | PubMed | ISI | ChemPort |
- Smits, V.A. et al. Nature Cell Biol. 2, 672−676 (2000). | Article | PubMed | ISI | ChemPort |
- Simizu, S. & Osada, H. Nature Cell Biol. 2, 852−854 (2000). | Article | PubMed | ISI | ChemPort |
- Wolf, G. et al. Oncogene 14, 543−549 (1997). | Article | PubMed | ChemPort |
- Dietzmann, K., Kirches, E., von Bossanyi, P., Jachau, K. & Mawrin, C. J. Neurooncol. 53, 1−11 (2001). | Article | PubMed | ChemPort |
- Tokumitsu, Y. et al. Int. J. Oncol. 15, 687−692 (1999). | PubMed | ISI | ChemPort |
- Knecht, R. et al. Cancer Res. 59, 2794−2797 (1999). | PubMed | ISI | ChemPort |
- Smith, M.R. et al. Biochem. Biophys. Res. Commun. 234, 397−405 (1997). | Article | PubMed | ISI | ChemPort |
- Cogswell, J.P., Brown, C.E., Bisi, J.E. & Neill, S.D. Cell Growth Differ. 11, 615−623 (2000). | PubMed | ISI | ChemPort |
- Luo, Y. et al. Nature Struct. Biol. 8, 1031−1036 (2001). | Article | PubMed | ISI | ChemPort |
- Otwinowski, Z. & Minor, W. Methods Enzymol. 276, 307−326 (1997). | Article | PubMed | ISI | ChemPort |
- Brünger, A.T. et al. Acta. Crystallogr. D 54, 905−921 (1998). | Article | PubMed | ISI |
- de La Fortelle, E. & Bricogne, G. 6 Methods Enzymol. 276, 472−494 (1997). | ChemPort |
- Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. Acta. Crystallogr. A 47, 110−119 (1991). | Article | PubMed | ISI |
- Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton, J.M. J. Appl. Crystallgr. 26, 283−291 (1993). | Article | ChemPort |
- Carson, M. J. Appl. Crystallgr. 24, 958−961 (1991). | Article |
- Nicholls, A., Sharp, K.A. & Honig, B. Proteins Struct. Funct. Genet. 11, 281−296 (1991). | PubMed | ChemPort |
Acknowledgments
We thank I. Blasutig, S.H. Ong, S. Oster, L. Harrington, D. Durocher and P. Plant for helpful discussion and assistance with the coimmunoprecipitation experiments, and P. Taylor and B. Larsen for instruction on mass spectrometry. We also thank the BioCars and Structural Biology Centre staff at the Advanced Photon Source at Argonne National Laboratories, where diffraction data were collected. This work was supported by grants from the National Cancer Institute of Canada to F.S. and J.D. F.S. is a recipient of a National Cancer Institute of Canada Scientist award. G.C.L is a recipient of a Natural Sciences and Engineering Research Council of Canada award.
Competing interests statement:
The authors declare that they have no competing financial interests. |
. Final model is shown in stick representation. This panel was generated using O
