Title: Regions of Byr4, a Regulator of Septation in Fission Yeast, That Bind Spg1 or Cdc16 and Form a Two-component GTPase-activating Protein with Cdc16
Abstract: In the fission yeast Schizosaccharomyces pombe, septation and constriction of the actomyosin ring for cell division are positively regulated by the Spg1 GTPase, a member of the Ras superfamily. Spg1 is negatively regulated by Byr4 and Cdc16, which together form a two-component GTPase-activating protein for the Spg1 GTPase. To better understand how Byr4 regulates septation, Byr4 mutants were tested for in vitro functions. This analysis revealed that Byr4 contained one Cdc16-binding site and four Spg1-binding sites (SBS), designated SBS1–SBS4. Although mutants with a single SBS bound Spg1 and inhibited GTP dissociation, the equilibrium binding affinity of these mutants was 28–280-fold weaker than Byr4. Because some Byr4 mutants with multiple SBSs bound Spg1 tighter than the corresponding mutants with a single SBS, multiple SBSs probably interact to cause the high affinity binding of Byr4 to Spg1. A region of Byr4 that bound Spg1, SBS4, and the region that bound Cdc16, Cdc16-binding site, was necessary and sufficient to form Cdc16-dependent Spg1GAP activity that was similar to that of wild-type Byr4 with Cdc16. In the fission yeast Schizosaccharomyces pombe, septation and constriction of the actomyosin ring for cell division are positively regulated by the Spg1 GTPase, a member of the Ras superfamily. Spg1 is negatively regulated by Byr4 and Cdc16, which together form a two-component GTPase-activating protein for the Spg1 GTPase. To better understand how Byr4 regulates septation, Byr4 mutants were tested for in vitro functions. This analysis revealed that Byr4 contained one Cdc16-binding site and four Spg1-binding sites (SBS), designated SBS1–SBS4. Although mutants with a single SBS bound Spg1 and inhibited GTP dissociation, the equilibrium binding affinity of these mutants was 28–280-fold weaker than Byr4. Because some Byr4 mutants with multiple SBSs bound Spg1 tighter than the corresponding mutants with a single SBS, multiple SBSs probably interact to cause the high affinity binding of Byr4 to Spg1. A region of Byr4 that bound Spg1, SBS4, and the region that bound Cdc16, Cdc16-binding site, was necessary and sufficient to form Cdc16-dependent Spg1GAP activity that was similar to that of wild-type Byr4 with Cdc16. In many eukaryotes, including Schizosaccharomyces pombe, cell division results from the constriction of an actomyosin ring that is perpendicular to the mitotic spindle (reviewed in Ref. 1Gould K. Simanis V. Genes Dev. 1997; 11: 2939-2951Crossref PubMed Scopus (158) Google Scholar). In fission yeast, a medial ring composed of F-actin and other proteins forms as cells enter mitosis. This ring is formed at the future site of cell division (2Marks J. Hagan I. Hyams J. Spec. Publ. Soc. Gen. Microbiol. 1987; 23: 119-135Google Scholar). Following anaphase, this actomyosin ring constricts (2Marks J. Hagan I. Hyams J. Spec. Publ. Soc. Gen. Microbiol. 1987; 23: 119-135Google Scholar, 3Jochova J. Rupes I. Streiblova E. Cell Biol. Int. Rep. 1991; 15: 607-610Crossref PubMed Scopus (46) Google Scholar, 4Fankhauser C. Reymond A. Cerutti L. Utzig S. Hofmann K. Simanis V. Cell. 1995; 82: 435-444Abstract Full Text PDF PubMed Scopus (222) Google Scholar, 5McCollum D. Balasubramanian M. Pelcher L. Hemmingsen S. Gould K. J. Cell Biol. 1995; 130: 1-11Crossref PubMed Scopus (141) Google Scholar, 6Chang F. Drubin D. Nurse P. J. Cell Biol. 1997; 137: 169-182Crossref PubMed Scopus (332) Google Scholar, 7Kitayama C. Sugimoto A. Yamamoto M. J. Cell Biol. 1997; 137: 1309-1319Crossref PubMed Scopus (177) Google Scholar), a primary septum is deposited, and secondary septa are formed on both sides of the primary septum (8Johnson B. Yoo B. Calleja G. J. Bacteriol. 1973; 115: 358-366Crossref PubMed Google Scholar). The primary septum is then degraded to yield two cells. Characterization of fission yeast mutants that affect these processes suggests they represent the following four categories: 1) division site selection mutants, 2) actomyosin ring assembly mutants, 3) actomyosin ring constriction and septation initiation mutants, and 4) cell separation mutants (reviewed in Ref. 1Gould K. Simanis V. Genes Dev. 1997; 11: 2939-2951Crossref PubMed Scopus (158) Google Scholar). cdc7− and spg1− mutants form a medial ring but do not constrict this ring or deposit a septum, leading to elongated, multinucleate cells (9Schmidt S. Sohrmann M. Hofmann K. Woollard A. Simanis V. Genes Dev. 1997; 11: 1519-1534Crossref PubMed Scopus (177) Google Scholar, 10Fankhauser C. Simanis V. EMBO J. 1994; 13: 3011-3019Crossref PubMed Scopus (138) Google Scholar, 11Balasubramanian M. McCollum D. Chang L. Wong K. Naqvi N. He X. Sazer S. Gould K. Genetics. 1998; 149: 1265-1275Crossref PubMed Google Scholar). In contrast,cdc16− and byr4−mutants undergo repeated rounds of septation (12Minet M. Nurse P. Thuriaux P. Mitchison J. J. Bacteriol. 1979; 137: 440-446Crossref PubMed Google Scholar, 13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar). These phenotypes suggest that Cdc7 and Spg1 are positive regulators of septation, whereas Cdc16 and Byr4 are negative regulators of septation. Consistent with this notion, Spg1 and Cdc7 overexpression causes arrest of cells with multiple septa (9Schmidt S. Sohrmann M. Hofmann K. Woollard A. Simanis V. Genes Dev. 1997; 11: 1519-1534Crossref PubMed Scopus (177) Google Scholar, 10Fankhauser C. Simanis V. EMBO J. 1994; 13: 3011-3019Crossref PubMed Scopus (138) Google Scholar), whereas Byr4 overexpression causes arrest of cells with multiple nuclei (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar). Spg1 is a GTPase of the Ras superfamily that is constitutively localized to spindle-pole bodies where it presumably regulates septation (9Schmidt S. Sohrmann M. Hofmann K. Woollard A. Simanis V. Genes Dev. 1997; 11: 1519-1534Crossref PubMed Scopus (177) Google Scholar, 14Sohrman M. Schmidt S. Hagan I. Simanis V. Genes Dev. 1998; 12: 84-94Crossref PubMed Scopus (159) Google Scholar). In its GTP-bound form, Spg1 binds the Cdc7 protein kinase and causes it to translocate to spindle poles (10Fankhauser C. Simanis V. EMBO J. 1994; 13: 3011-3019Crossref PubMed Scopus (138) Google Scholar, 14Sohrman M. Schmidt S. Hagan I. Simanis V. Genes Dev. 1998; 12: 84-94Crossref PubMed Scopus (159) Google Scholar). Consistent with their role as negative regulators of septation, Byr4 and Cdc16 form a two-component GTPase-activating protein (GAP) 1The abbreviations used are: GAP, GTPase-activating protein; CBS, Cdc16-binding site; SBS, Spg1-binding site; CRIB, Cdc42/Rac interactive binding; Gpp(NH)p, guanosine 5′-(β,γ-imido)triphosphate.1The abbreviations used are: GAP, GTPase-activating protein; CBS, Cdc16-binding site; SBS, Spg1-binding site; CRIB, Cdc42/Rac interactive binding; Gpp(NH)p, guanosine 5′-(β,γ-imido)triphosphate. (1Gould K. Simanis V. Genes Dev. 1997; 11: 2939-2951Crossref PubMed Scopus (158) Google Scholar) for the Spg1 GTPase (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). Although a role for Cdc16 in Spg1GAP activity is not surprising because Cdc16 has sequence similarity to proteins with YptGAP activity (9Schmidt S. Sohrmann M. Hofmann K. Woollard A. Simanis V. Genes Dev. 1997; 11: 1519-1534Crossref PubMed Scopus (177) Google Scholar, 16Neuwald A. Trends Biochem. Sci. 1997; 22: 243-244Abstract Full Text PDF PubMed Scopus (86) Google Scholar), the requirement for two proteins for GAP activity is unexpected because GAPs for other Ras family GTPases contain one component (17Boguski M.S. McCormick F. Nature. 1993; 366: 643-654Crossref PubMed Scopus (1749) Google Scholar). In addition to its role in Spg1GAP activity, Byr4 binds Spg1, and this binding inhibits GTP dissociation and hydrolysis but does not affect GDP dissociation (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). Cdc16 binds Byr4 but does not bind Spg1 or affect the nucleotide bound to Spg1 in the absence of Byr4. Byr4-Spg1 binding and Byr4-Cdc16 Spg1GAP activity are specific for Spg1 because Byr4 and Cdc16 do not interact with Ypt family GTPases, which are the most similar to Spg1 by primary sequence comparison. Genetic experiments support the biochemical interactions between Spg1, Byr4, and Cdc16 (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar, 18Jwa M. Song K. Molecules Cells. 1998; 8: 240-245PubMed Google Scholar). Mild Byr4 overexpression suppresses the temperature-sensitive growth ofcdc16–116 mutants. This suppression is allele-specific because Byr4 does not suppress the growth defect ofcdc16 − mutants. Whereas higher Byr4 overexpression is lethal, Spg1 overexpression suppresses this lethality. A Byr4-like regulator and a two-component GAP are unique to the Spg1 GTPase. Novel regulators for Spg1 are consistent with classification of Spg1 in a separate subfamily of the Ras superfamily (9Schmidt S. Sohrmann M. Hofmann K. Woollard A. Simanis V. Genes Dev. 1997; 11: 1519-1534Crossref PubMed Scopus (177) Google Scholar) because regulators for different subfamilies are usually distinct (17Boguski M.S. McCormick F. Nature. 1993; 366: 643-654Crossref PubMed Scopus (1749) Google Scholar). TheSaccharomyces cerevisiae Tem1 GTPase is currently the only other member of the Spg1 subfamily (9Schmidt S. Sohrmann M. Hofmann K. Woollard A. Simanis V. Genes Dev. 1997; 11: 1519-1534Crossref PubMed Scopus (177) Google Scholar). Unlikespg1− mutants, tem1−mutants arrest in late mitosis perhaps because of an inability to degrade mitotic cyclins (19Shirayama M. Matsui Y. Toh-e A. Mol. Cell. Biol. 1994; 14: 7476-7482Crossref PubMed Scopus (159) Google Scholar, 20Jaspersen S. Charles J. Tinker-Kulberg R. Morgan D. Mol. Biol. Cell. 1998; 9: 2803-2817Crossref PubMed Scopus (261) Google Scholar). Whereas tem1−mutants do not undergo septation or cytokinesis, it is unclear whether this failure results from their cell cycle arrest or reflects a requirement for Tem1 to regulate septation and cytokinesis. TheS. cerevisiae Bub2 protein is the likely homologue of Cdc16 because these proteins are globally similar, and BUB2suppresses the temperature-sensitive growth defect ofcdc16–116 mutants (21Fankhauser C. Marks J. Reymond A. Simanis V. EMBO J. 1993; 12: 2697-2704Crossref PubMed Scopus (139) Google Scholar). cdc16−mutants and bub2− mutants are similar in that they are defective in the spindle-assembly checkpoint, but differ in that BUB2 is not essential for viability (21Fankhauser C. Marks J. Reymond A. Simanis V. EMBO J. 1993; 12: 2697-2704Crossref PubMed Scopus (139) Google Scholar, 22Hoyt M.A. Totis L. Roberts B. Cell. 1991; 66: 507-517Abstract Full Text PDF PubMed Scopus (891) Google Scholar). To further understand how Byr4 regulates septation, Byr4 mutants were assayed for binding to Cdc16 or Spg1 and for Spg1GAP activity with Cdc16. A single Cdc16-binding site (CBS) and four Spg1-binding sites (SBS) were identified. A C-terminal fragment of Byr4, which contained SBS4 and CBS, was necessary and sufficient for Spg1GAP activity with Cdc16. pbyr4/ET14b, pcdc16/GEX, and pspg1/GEX were previously described (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar, 15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). byr4 mutants were constructed by amplification of the byr4 cDNA using 5′-oligonucleotides with anNdeI site and 3′-oligonucleotides with a BamHI site. Amplified products were digested with NdeI andBamHI and ligated with pET14b that was similarly digested. Byr4 mutants, Gst-Cdc16, and Gst-Spg1 were expressed in BL21(DE3)pLysE, NM522, and XL1-Blue bacteria, respectively. Methods for purification of Spg1, Cdc16, and Byr4, Cdc16-Byr4 and Spg1-Byr4 binding assays, and Spg1GAP assays were previously described (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar,15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). The affinity of Byr4 mutants for Spg1 was determined by measuring the effect of Byr4 on Spg1 GTPase properties. Radioactivity is lost from Spg1-[γ-32P]GTP by GTP dissociation and hydrolysis. The fraction of the initial radioactivity bound to Spg1 is described in terms of the GTP dissociation rate, kt, the GTP hydrolysis rate, kh, and time, t, by the following equation. Spgl−[γ−32P]GTP(Spgl−[γ−32P]GTP)t=0=exp−kh+kttEquation 1 Because the Byr4-Spg1-GTP complex had negligible rates of GTP dissociation and hydrolysis (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar), the effect of Byr4 on the observed rate of radioactivity lost from Spg1-GTP, ko, was used to measure the Byr4-Spg1-GTP equilibrium binding constant,Kb. Kb is related toko by the following equation. ko=kh+kt1+Byr4KbEquation 2 If Byr4 mutants inhibited only GTP dissociation, then the observed rate of radioactivity lost from Spg1-GTP would be as follows. ko=kh+kt1+Byr4KbEquation 3 In either case, 50% maximal inhibition occurs when the Byr4 concentration equals Kb. Ras-Raf-1 binding was measured similarly (23Herrmann C. Martin G. Wittinghofer A. J. Biol. Chem. 1995; 270: 2901-2905Abstract Full Text Full Text PDF PubMed Scopus (316) Google Scholar). Our experience with this method revealed that the measured equilibrium constants varied by less than 50% between different Byr4 preparations. Byr4 contains 665 amino acids, and the only motifs detected by primary sequence inspection were two imperfect direct repeats of 43 amino acids (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar) (Fig.1 A). To identify Byr4 regions that bind Cdc16, Byr4 deletion mutants were expressed inEscherichia coli, purified, and assayed for binding to Gst-Cdc16 (Fig. 1 A). Byr4-D2, a mutant with residues 533–665, bound Gst-Cdc16, although Byr4-D6, a mutant with residues 1–535, did not bind Gst-Cdc16 (Fig. 1 B). To further limit the CBS, mutants containing the N- and C-terminal portions of Byr4-D2 were analyzed. These assays showed that Byr4-D10, a mutant with residues 595–665, bound Gst-Cdc16, although Byr4-D14 did not bind Gst-Cdc16 (Fig. 1 B). As with wild-type Byr4 (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar), Byr4-D2 and Byr4-D10 degradation products that copurified with the full-length proteins did not bind Gst-Cdc16 (Fig. 1 B). This result was consistent with a requirement for the Byr4 C terminus to bind Cdc16 because these degradation products were missing C-terminal residues (data not shown). We conclude that the Byr4 C terminus is necessary and sufficient for Byr4-Cdc16 binding. The ability of Byr4 to inhibit the dissociation and hydrolysis of GTP that is bound to Spg1 was used to identify SBS in Byr4 (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). This assay was advantageous because it was quantitative, and binding was measured under equilibrium conditions. The equilibrium binding constant,Kb, is the Byr4 concentration where 50% of the maximal inhibition of Spg1 GTPase properties occurs (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar, 23Herrmann C. Martin G. Wittinghofer A. J. Biol. Chem. 1995; 270: 2901-2905Abstract Full Text Full Text PDF PubMed Scopus (316) Google Scholar). Four SBSs, designated SBS1, SBS2, SBS3, and SBS4, were identified using this assay (Fig. 2 A). SBS1 was contained within Byr4-D3, which had residues 1–200. Byr4-D3 inhibited loss of radioactivity from Spg1-[γ-32P]GTP about 50% at saturating concentrations with a 140 nm Kb (Fig. 2, A and B) and from Spg1-[γ-35S]GTP (Fig. 2 C), showing that SBS1 inhibited GTP dissociation. Because Byr4-D3 only partially inhibited loss of radioactivity from Spg1-[γ-32P]GTP, Byr4-D3 may not affect GTP hydrolysis. Mutants that deleted sequences in Byr4-D3, such as Byr4-D17 and Byr4-D18, did not bind Spg1 (Fig.2 A). SBS2 was contained within Byr4-D27, which had residues 200–475. Byr4-D27 inhibited loss of radioactivity from Spg1-[γ-32P]GTP with a 175 nm Kb (Fig. 2, A and B) and from Spg1-[γ-35S]GTP (Fig. 2 C), showing that Byr4-D27 inhibited both GTP hydrolysis and dissociation. Mutants that deleted sequences in Byr4-D27, such as Byr4-D22 and Byr4-D21, significantly weakened binding to Spg1 (Fig. 2 A). SBS3 and SBS4 were contained within Byr4-D13 and Byr4-D14, respectively (Fig.2 A). Byr4-D13 and Byr4-D14 inhibited loss of radioactivity from Spg1-[γ-32P]GTP with 1400 and 300 nm Kb, respectively (Fig. 2 B), and from Spg1-[γ-35S]GTP (Fig. 2 C), showing that Byr4-D13 and Byr4-D14 inhibited both GTP hydrolysis and dissociation. The boundaries of these mutants corresponded to the imperfect, direct repeats identified by primary sequence analysis (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar). Byr4-D10, which contained CBS, did not affect Spg1 GTPase properties (Fig.2 B). Hence, Byr4 contained four SBSs with equilibrium binding constants from 140 to 1400 nm. These binding affinities were much weaker than that of Byr4, which had a 5 nm Kb (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). The affinities of Byr4 mutants with multiple SBSs were measured to determine which Byr4 regions were required for maximal binding to Spg1. Two mutants with two sequential SBSs bound tighter than the corresponding mutants with a single SBS. Byr4-D4, a mutant with SBS1 and SBS2, and Byr4-D7, a mutant with SBS2 and SBS3, had 50–60 nm Kb, which were about 3-fold lower than mutants with the corresponding single SBS (Fig. 2 D). Interestingly, Byr4-D2, a mutant with SBS4 and CBS, had a 6-fold lowerKb than a mutant with only SBS4. Whether the higher affinity of Byr4-D4, Byr4-D7, and Byr4-D2 resulted from simultaneous binding of multiple SBSs or increased binding of a single SBS is unknown. In contrast, Byr4-D11, a mutant with SBS3 and SBS4, had a 300 nm Kb, which was similar to that of mutants with only SBS4 (Fig. 2 D). The lack of cooperativity between SBS3 and SBS4 was not surprising because these domains share primary sequence similarity (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar). All mutants tested with three or four sequential SBSs and/or CBS had 30–50 nm Kb (Fig. 2 D), showing that these combinations did not significantly increase Spg1 binding. Because a 5 nm Kb for Byr4-Spg1 was measured (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar), we conclude that SBS1, CBS, and perhaps SBS2, SBS3, or SBS4 are required for maximal binding affinity. Byr4 binding to Spg1-GDP could not be measured using changes in GDP dissociation because Byr4 did not affect this GTPase property (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). Instead, the ability of Byr4 mutants to bind immobilized Spg1-GDP or Spg1-Gpp(NH)p was used to test the effect of the Spg1 nucleotide state on Byr4 binding. These assays showed that Byr4-D3 and Byr4-D27, mutants with only SBS1 or SBS2, respectively, bound Spg1-Gpp(NH)p slightly better than Spg1-GDP (Fig. 2 E). Byr4-D13 and Byr4-D14, mutants with only SBS3 or SBS4, respectively, bound Spg1-Gpp(NH)p much better than Spg1-GDP (Fig. 2 E). These binding preferences were unexpected because Byr4 bound equally to Spg1-Gpp(NH)p and Spg1-GDP (15Furge K. Wong K. Armstrong J. Balasubramanian M. Albright C.F. Curr. Biol. 1998; 8: 947-954Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). To understand the basis for the nucleotide specificity difference between the individual SBSs and Byr4, Byr4 mutants with multiple SBSs were assayed. Byr4-D1, a mutant with SBS2, SBS3, and SBS4, bound Spg1-Gpp(NH)p slightly better than Spg1-GDP (Fig.2 E). Because the nucleotide specificity of Byr4-D1 binding was similar to that of Byr4-D27, SBS2 appeared to dominate the nucleotide specificity of Byr4-D27. Byr4-D4, a mutant with SBS1 and SBS2, bound equally to Spg1-Gpp(NH)p and Spg1-GDP, suggesting that SBS1 and SBS2 interact to affect both binding specificity and affinity (Fig.2 E). The nucleotide-independent binding of Byr4-D4 to Spg1 probably explains the binding behavior of Byr4. Hence, Byr4 contains four SBSs and some of these SBSs interact to cause high affinity, nucleotide-independent binding to Spg1. Because a two-component GAP is unique to Byr4 and Cdc16, we determined what regions of Byr4 were required for this reaction. This analysis showed that Byr4-D2, which contained SBS4 and CBS, was the minimal mutant that had Spg1GAP activity with Cdc16 (Fig.3 A). Neither Byr4-D16, which lacked CBS, nor Byr4-D10, which contained only CBS, had detectable Spg1GAP activity. Byr4 mutants without CBS had at least 100-fold less Spg1GAP activity than Byr4 because Spg1GAP activity was detected with 2 nm Byr4 but not with 200 nm Byr4-D16 (Fig.3 B and data not shown). Spg1GAP assays with a range of Byr4-D2 and Byr4 concentrations revealed that Byr4 and Byr4-D2 had similar Spg1GAP activity (Fig. 3 B). Hence, a region of Byr4 that binds Spg1 (SBS4) and a region that binds Cdc16 (CBS) are necessary and sufficient to form Cdc16-dependent Spg1GAP activity. To better understand how Byr4 regulates septation, Byr4 mutants were assayed for binding to Spg1 or Cdc16 and for formation of Spg1GAP activity with Cdc16. This analysis revealed that Byr4 contained one CBS and four SBSs, designated SBS1–SBS4. SBS4 and CBS were necessary and sufficient to form Spg1GAP activity with Cdc16, which was very similar to that of Byr4. Biochemical and sequence analyses suggest that the four SBSs in Byr4 represent three classes of SBSs. One SBS class includes SBS3 and SBS4, which correspond to the imperfect direct repeats of 43 amino acids identified by primary sequence analysis (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar). Like Byr4, SBS3 and SBS4 inhibited GTP dissociation and hydrolysis by Spg1. Unlike Byr4, SBS3 and SBS4 bound Spg1-GTP with a 280- and 60-fold, respectively, weaker affinity than Byr4 and bound Spg1-GTP better than Spg1-GDP. Although unrelated in primary sequence, SBS3 and SBS4 have some similarities to the Cdc42/Rac interactive binding (CRIB) motif, a sequence that binds the Cdc42 and Rac1 GTPases (24Manser E. Leung T. Salihuddin H. Zhao Z. Lim L. Nature. 1994; 367: 40-46Crossref PubMed Scopus (1280) Google Scholar) and is found in several proteins (25Burbelo P. Drechel D. Hall A. J. Biol. Chem. 1995; 270: 29071-29074Abstract Full Text Full Text PDF PubMed Scopus (549) Google Scholar). Like CRIB motifs, SBS3 and SBS4 are relatively short in length and bind the GTP-bound form of the GTPase better than the GDP-bound form of the GTPase. Strengthening of SBS4 binding to Spg1-GTP by CBS is similar to results found with the CRIB motif from the Wiskott-Aldrich syndrome protein. In the case of Byr4, the relatively weak binding of SBS4 to Spg1-GTP, characterized by an equilibrium binding affinity of 300 nm, was strengthened 5-fold by sequences containing CBS, which did not detectably interact with Spg1 without SBS4. In a similar manner, the relatively weak binding of Cdc42-GTP to the CRIB peptide from Wiskott-Aldrich syndrome protein, characterized by an equilibrium binding affinity of 480 nm, was strengthened about 6-fold by adjacent sequences that did not bind Cdc42-GTP in isolation (26Rudolph M. Bayer P. Abo A. Kuhlmann J. Vetter I. Wittinghofer A. J. Biol. Chem. 1998; 273: 18067-18076Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar). Perhaps there will be additional similarities between SBS3/SBS4 and CRIB motifs. It will be particularly interesting to test whether SBS3 and SBS4, like the Wiskott-Aldrich syndrome protein-CRIB motif (26Rudolph M. Bayer P. Abo A. Kuhlmann J. Vetter I. Wittinghofer A. J. Biol. Chem. 1998; 273: 18067-18076Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar), lack a defined domain structure in the absence of the GTPase. Cdc7, a protein kinase that is likely an effector of Spg1 (14Sohrman M. Schmidt S. Hagan I. Simanis V. Genes Dev. 1998; 12: 84-94Crossref PubMed Scopus (159) Google Scholar), may use a motif similar to SBS3 and SBS4 to bind Spg1 because Cdc7 contains a sequence similar to SBS3 and SBS4 (13Song K. Mach K.E. Chen C.Y. Reynolds T. Albright C.F. J. Cell Biol. 1996; 133: 1307-1319Crossref PubMed Scopus (89) Google Scholar). If this is the case, then Byr4 and Cdc7 may bind competitively to Spg1, and Byr4 might negatively regulate Spg1 by inhibiting effector binding as well as by forming Spg1GAP activity with Cdc16. SBS1 and SBS2 likely define distinct classes of SBSs. SBS1 and SBS2 differ from SBS3/SBS4 in that they lack the conserved sequence motifs found in SBS3 and SBS4. Furthermore, SBS1 and SBS2 bind similarly to Spg1-GTP and Spg1-GDP. SBS1 and SBS2 differ in that SBS1 inhibits GTP hydrolysis weakly, if at all, whereas SBS2 inhibits GTP hydrolysis. Consistent with SBS1, SBS2, and SBS3/SBS4 representing different classes of SBSs, mutants with SBS1/SBS2 or SBS2/SBS3 bind Spg1 better than the corresponding mutants with only one SBS. Multiple SBSs were unexpected because of the other GTPase-binding proteins; only Ran-binding protein 2 contains multiple binding sites for the same GTPase (27Wu J. Matunis M. Kraemer D. Blobel G. Coutavas E. J. Biol. Chem. 1995; 270: 14209-14213Crossref PubMed Scopus (395) Google Scholar, 28Yokoyama N. Hayashi N. Seki T. Pante N. Ohba T. Nishii K. Kuma K. Hayashida T. Miyata T. Aebi U. Fukui M. Nishimoto T. Nature. 1995; 376: 184-188Crossref PubMed Scopus (410) Google Scholar). Multiple SBSs may be needed to increase Byr4-Spg1 affinity because all deletion mutants had at least 10-fold weaker affinity than wild-type Byr4. In this case, different SBSs may interact with different regions of Spg1 to form the high affinity binding observed between Byr4 and Spg1. The increased binding affinity of mutants with more than one SBS is consistent with this notion. Multiple SBSs may also allow Byr4 to bind more than one Spg1 molecule in vivo. Multivalent binding could dramatically increase the effective Byr4-Spg1 affinity and allow different SBSs to perform distinct functions. For instance, some SBSs may target Byr4 to spindle poles, whereas other SBSs may form Spg1GAP activity with Cdc16. In vivo analysis of Byr4 mutants will be needed to understand the physiologic roles of the different SBSs. Sequences containing SBS4 and CBS were necessary and sufficient for Spg1GAP activity with Cdc16. GAPs are unusual enzymes in that two proteins, the GTPase and the GAP, interact transiently and contain residues important for catalysis. GAPs are thought to stimulate GTP hydrolysis by stabilizing switch regions in the GTPase that orient the GTPase catalytic machinery and by supplying an external arginine residue that stabilizes the charge in the transition state (reviewed in Ref. 29Scheffzek K. Ahmadian M. Wittinghofer A. Trends Biochem. Sci. 1998; 23: 257-262Abstract Full Text Full Text PDF PubMed Scopus (347) Google Scholar). Byr4 and Cdc16 may further divide these GAP functions. SBS4, one part of Byr4 required for Spg1GAP activity, binds Spg1-GTP, affects its GTPase properties, and may stabilize the switch regions of Spg1. CBS, another part of Byr4 required for Spg1GAP activity, binds Cdc16 and is presumably required for Cdc16 to interact with Spg1. Once near Spg1, Cdc16 may supply the arginine residue presumably required for Spg1GAP activity. Byr4 would thus provide both binding and catalytic functions required for Spg1GAP activity in this model. The division of traditional GAP activities into two proteins may allow additional regulation of Spg1GAP activity. We thank Kathleen Mach for helpful comments on the manuscript.