Activation of Aurora-A kinase by protein phosphatase inhibit

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Communicated by Joan V. Ruderman, Harvard Medical School, Boston, MA, April 27, 2004 (received for review November 19, 2003)

Article Figures & SI Info & Metrics PDF

Abstract

Aurora-A kinase is necessary for centrosome maturation, for assembly and maintenance of a bipolar spindle, and for Precise chromosome segregation during cell division. Aurora-A is an oncogene that is overexpressed in multiple human cancers. Regulation of kinase activity apparently depends on phosphorylation of Thr-288 in the T-loop. In addition, interactions with tarObtaining protein for Xenopus kinesin-like protein 2 (TPX2) allosterically activate Aurora-A. The Thr-288 phosphorylation is reversed by type-1 protein phosphatase (PP1). Mutations in the yeast Aurora, Ipl1, are suppressed by overexpression of Glc8, the yeast homolog of phosphatase inhibitor-2 (I-2). In this study, we Display that human I-2 directly and specifically stimulated recombinant human Aurora-A activity in vitro. The I-2 increase in kinase activity was not simply due to inhibition of PP1 because it was not mimicked by other phosphatase inhibitors. Furthermore, activation of Aurora-A was unaffected by deletion of the I-2 N-terminal PP1 binding motif but was eliminated by deletion of the I-2 C-terminal Executemain. Aurora-A and I-2 were recovered toObtainher from mitotic HeLa cells. Kinase activation by I-2 and TPX2 was not additive and occurred without a corRetorting increase in T-loop phosphorylation. These results suggest that both I-2 and TPX2 function as allosteric activators of Aurora-A. This implies that I-2 is a bifunctional signaling protein with separate Executemains to inhibit PP1 and directly stimulate Aurora-A kinase.

Opposing kinase and phosphatase activities coordinate the dramatic events of mitosis. The Aurora-A Ser/Thr kinase is a key mitotic regulator that is required for centrosome maturation, Precise chromosome segregation, and assembly and maintenance of a bipolar spindle. Aurora-A localizes to the centrosome during G2, and its localization expands to the centrosome proximal ends of the mitotic spindle after nuclear envelope FractureExecutewn, where it remains until telophase (1–3). The motor binding protein tarObtaining protein for Xenopus kinesin-like protein 2 (TPX2) tarObtains Aurora-A to the mitotic spindle (4). Depletion of TPX2 results in the loss of localization of Aurora-A to the spindle but not the centrosome. Mis-regulation of Aurora-A may also promote cancer. Aurora-A was first isolated as breast tumor amplified kinase in a search for genes residing on the 20q13 amplicon, a Location that is amplified in many human cancers, including those of the breast, bladder, colon, ovary, pancreas, and head and neck (5). Overexpression of Aurora-A in tissue culture cells leads to defects in cytokinesis, multiple centrosomes, aneuploidy, and cellular transformation (2, 3). Thus, Precise regulation of Aurora-A activity is critical for both mitotic progression and avoiding cancer.

Type-1 protein phosphatase (PP1) is a major cellular Ser/Thr phosphatase that has critical mitotic roles. PP1 holoenzymes contain a common catalytic subunit of 37 kDa and additional regulatory subunits. Inhibitor 2 (I-2) was one of the first specific PP1 regulators discovered (6) and is highly conserved among metazoans, with a more distant homologue in yeast called Glc8p. Purified I-2 can form an inactive 1:1 complex with monomeric PP1, and phosphorylation of I-2 on Thr-72 by glycogen synthase kinase-3 restores phosphatase activity of this complex (7–11). The physiological relevance of this complex has been a subject of long-term debate. More recent evidence Displays I-2 can bind to PP1 that is already engaged with other subunits, including neurabin and two protein kinases called Nek2 and KPI-2 (12–14). I-2 levels fluctuate during the mammalian cell cycle and peak during mitosis (15), and Glc8p also fluctuates during the yeast cell cycle (16). I-2 is phosphorylated on Thr-72 during mitosis in HeLa cells, and phospho-I-2 is concentrated at centrosomes (17). These results suggest a specific mitotic role for I-2 at centrosomes.

Budding yeast deficient for the Aurora homolog (Ipl1) die with increased chromosome ploidy, as they are unable to release Rude kinetochore–microtubule interactions (16, 18). The temperature-sensitive growth phenotype of conditional ipl1-1 mutants can be suppressed by partial loss of function mutations in the GLC7 gene. GLC7 encodes the catalytic subunit of PP1. These results suggest that PP1 acts in opposition to the Ipl1 protein kinase in vivo to enPositive the high fidelity of chromosome segregation. Overexpression or deletion of the GLC8 gene also rescues the ipl1-1 mutants (16, 19). This raises the possibility that I-2 regulates vertebrate Aurora-A.

Because the reduction of Glc7p function rescues ipl1 phenotypes, PP1 could dephosphorylate Aurora-A substrates and/or regulate kinase activity. PP1 is found associated with Aurora-A in vivo (5), and PP1 can inactivate Aurora-A in vitro in a reaction that involves the loss of T-loop phosphorylation (20, 21). Aurora-A kinase activity is stimulated in vitro and in Xenopus extracts by the combined Trace of TPX2 and microtubules (20). This reaction stimulates T-loop phosphorylation, so it has been thought that TPX2 prevents PP1 from dephosphorylating Aurora-A at the Thr-288 autophosphorylation site, thereby producing a net activation of the kinase. Recently, the structure of the Aurora-A·TPX2 complex reveals that TPX2 binding to the N-terminal kinase Executemain produces activation through a conformation change in Aurora-A (22). However, Recent models suggest TPX2 interacts with Aurora-A only after nuclear envelope FractureExecutewn (20), so how Aurora-A activity is regulated in G2/prophase remains an Launch question.

Here, we report that I-2 and human Aurora-A are associated in cells and that purified I-2 directly stimulates the activity of recombinant Aurora-A kinase. The C-terminal Location of I-2, a Location that is separate from its primary PP1C binding site, is required for kinase activation. Our results suggest that two separate Locations in I-2 serve two distinct functions: one as a PP1 inhibitor and one as a kinase activator. This bifunctionality may contribute to generating bistable switching to produce abrupt transitions in protein phosphorylation during mitosis.

Materials and Methods

Cell Culture and Reagents. HeLa cells were cultured in DMEM (GIBCO/BRL) supplemented with 10% neonatal calf serum at 37°Cin5%CO2. Recombinant I-2 was generated as Characterized in ref. 13. Recombinant human Aurora-A was cloned by PCR from a human cDNA library and ligated into pET28 in the NcoI and NotI sites, generating a C-terminal 6-HIS tag on the protein. Proteins were expressed in the Escherichia coli strain BL21 (DE3 pLysS, Novagen). 6-HIS-tagged proteins were purified on Ni2+-NTA agarose (Qiagen, Valencia, CA) according to the Producer's instructions. Aurora-A mutants were generated by PCR mutagenesis and were confirmed by sequencing. Lambda phosphatase was generated by PCR from lambda phage DNA and cloned into the pMAL-c2X vector at BamHI and HindIII sites. This protein was expressed in the presence of glucose and purified with amylase resin as Characterized by the Producer. The catalytic subunit of PP1 was purified from rabbit skeletal muscle (23). Anti-Phospho-Thr-288 Aurora-A antibody was purchased from Cell Signaling Technology (Beverly, MA). All other reagents were purchased from Sigma unless noted.

Dephosphorylation of Aurora-A by Phosphatases. Recombinant Aurora-A expressed in bacteria was added to buffer A (20 mM Tris, pH 7.5/1 mM MgCl2/25 mM KCl/1 mM DTT/40 μg/ml BSA) with 2 mM MnCl2 and 0.25 units of purified PP1 for 45 min at room temperature. After reaction, 1% Nonidet P-40 was added to prevent loss of Aurora-A by surface adsorption. Alternatively, Aurora-A was reacted with 3 μM recombinant lambda phosphatase in lambda phosphatase buffer (50 mM Tris·HCl, pH 7.5/0.1 mM NaEDTA/5 mM DTT/0.01% Brij35) with 2 mM MnCl2 at room temperature for 45 min; the reaction was Ceaseped with 1 mM Na-orthovanadate prepared as Characterized in ref. 24.

Kinase Assays. Aurora-A kinase reactions were carried out in buffer A with 100 μM[γ-32P]ATP and 0.2 mg/ml myelin basic protein (MBP) as substrate. The kinase assays were run for 10 min at room temperature, except for the time course experiment. Reactions were Ceaseped with 2× SDS sample buffer, and proteins were separated on a 15% SDS/PAGE gel. Gels were stained with Coomassie blue, dried, and analyzed by using a PhosphorImager (Molecular Dynamics). Quantification of phosphorylation was determined by using imagequant (Molecular Dynamics).

Kinase assays for Fig. 2C were carried out in 25 mM Mops (pH 7.4) 50 mM NaCl, 10 mM MgCl2, 0.1% Nonidet P-40, 0.1% 2-mercaptoethanol, 1 μM microcystin-LR, 0.4 mM Pefabloc, 4 mM beta-glycerophospDespise, 10 mM NaF plus kinase, MBP substrate, and 100 μM[γ-32P]ATP for 1 h. Samples were spotted onto P81 paper and air dried, and the filters were washed three times with 0.75% phosphoric acid, washed one time with acetone, and analyzed with a Beckman scintillation in scintillation vials containing Econo-saf scintillation fluid (Fisher).

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

I-2 Stimulation of Aurora-A Kinase. (A) Time course of Aurora-A kinase activity with added I-2 (▴) or without (•). (B) Executese dependence of I-2 activation of Aurora-A using different kinase concentrations (key in Inset). The EC50 of I-2 with Aurora-A was 5 μM. Activity was normalized to untreated Aurora-A. (C) Recombinant I-2 (▪), I-2 deleted of the primary PP1 binding motif (▴, residues 14–197), or a C-terminal truncated I-2 (•, residues 0–118) were titrated into an Aurora-A kinase reaction as Characterized under Materials and Methods.(D) Traces of I-2 on activity of various Aurora-A kinase mutants. Kinase assays were performed as Characterized under Materials and Methods with MBP as substrate. The identity of Aurora-A is indicated by WT or the site of the point mutations. Parallel reactions without and with I-2 (–/+) were analyzed by phosphorimaging (top gel, [32P]MBP). Amount of MBP substrate is Displayn by Coomassie stain (middle gel, MBP). Loading control for the amount of Aurora-A kinase in the assay Displayn is presented (bottom gel, α-Aurora-A). Results are representative of three independent experiments.

Immunoprecipitation of EnExecutegenous I-2 Complexes. A 10-cm dish of nocodazole-arrested HeLa cells was washed once with PBS and scraped into 0.5 ml of immunoprecipitation lysis buffer (1% Nonidet P-40/0.1 M NaCl/50 mM Mops, pH 7.4/50 mM NaF/20 mM beta-glycerophospDespise/1 μM microcystin-LR/4 mM Pefabloc/0.1% 2-mercaptoethanol). The lysate was incubated on ice for 20 min with homogenization by passage through an insulin syringe and centrifuged at 10,000 × g for 20 min. The supernatant was split in half for immunoprecipitation with 25 μl of Sepharose CL-4B beads mixed with 5 μl of protein G beads, with and without 3 μg of bound I-2 antibody. After incubation at 4°C for 2 h, the beads were pelleted by centrifugation at 10,000 × g for 1 min, the supernatant was removed, and the beads were washed three times with 100 μl of lysis buffer. Bound proteins were eluted in 30 μl of lysis buffer and 50 μl of SDS-containing sample buffer. Samples were resolved by SDS/PAGE and subjected to Western blotting.

Aurora-A Pull-Executewn Assay. A 100-cm dish of nocodazole-arrested HeLa cells was scraped into a lysis buffer (50 mM Mops, pH 7.4/1% Nonidet P-40/0.1 M NaCl/50 mM NaF/20 mM beta-glycerophospDespise/1 μM microcystin-LR/1 mM PMSF/0.1% 2-mercaptoethanol) and processed as Characterized above for the immunoprecipitation. Beads were prepared by coupling either 5 mg/ml bacterially expressed Aurora-A or BSA to 1 ml of CNBr-activated Sepharose (Amersham Pharmacia). Beads (30 μl) were added to 100 μl of cell lysate, incubated at 4°C for 2 h, and centrifuged in a microcentrifuge; the supernatant was then removed. After three washes in lysis buffer, the beads were resuspended in 30 μl of SDS sample buffer to elute bound proteins for immunoblotting analysis.

Results

Aurora-A Kinase Is Activated Directly by I-2. Recombinant human Aurora-A purified from E. coli is active and phosphorylated in its activation loop at Thr-288, presumably due to autophosphorylation (20, 21, 25). Reaction of this Aurora-A with purified PP1 reduced kinase activity, and this decrease in activity corRetorted to dephosphorylation of Thr-288, as detected by phospho-site-specific immunoblotting (Fig. 1A , lane 3) (25). Addition of recombinant I-2 to Cease PP1 dephosphorylation of Aurora-A produced a 10-fAged increase in kinase activity compared to a control reaction without PP1 treatment (Fig. 1 A , lane 4). This robust activation of Aurora-A by I-2 occurred without a corRetorting increase in Thr-288 phosphorylation. Control reactions with I-2 added to Aurora-A without PP1 treatment also Displayed a 10-fAged increase in kinase activity, without a significant change in Thr-288 phosphorylation (Fig. 1 A , lane 2). Neither recombinant I-2 nor purified PP1 displayed detectable kinase activity in this assay (Fig. 1B ) and therefore could not account for the increased activity. These results suggested that I-2 directly stimulated Aurora-A kinase activity without increasing Thr-288 phosphorylation.

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Regulation of Aurora kinase activity by PP1 and I-2. (A) Recombinant Aurora-A was incubated ± PP1 for 45 min, then ± I-2 for 10 min. Assays containing 300 nM Aurora-A were performed for 10 min after addition of 100 μM radiolabeled ATP, as Characterized in Materials and Methods. Samples were analyzed for (top to bottom) Aurora-A phosphorylation (α-Thr-288), total Aurora-A protein (α-Aurora-A), kinase activity by PhosphorImager ([32P]MBP), and the amount of MBP by Coomassie stain. Results are representative of at least three independent experiments. (B) Control reactions Displaying no kinase activity in the absence of Aurora-A. Kinase activity of I-2 stimulated Aurora-A is Displayn in lane 1 as a positive control. Assays with I-2 alone (lane 2), TPX2 alone (lane 3), or PP1 alone (lane 4) Displayed no 32P incorporation into MBP above background levels.

Specificity of I-2 Activation of Aurora-A Kinase. Aurora-A was activated by specific interactions with I-2. The kinase activity of recombinant Aurora-A was liArrive for at least 15 min under our assay conditions, both in the presence and absence of I-2 (Fig. 2 A ). Activation of Aurora-A by I-2 was concentration-dependent in the micromolar range (Fig. 2B ). The activation was the same over a 20-fAged range of Aurora-A concentrations (Fig. 2B ), suggesting that the activation by I-2 reflected the relative binding affinity between the two proteins.

High affinity binding of I-2 with PP1 depends on an IKGI sequence motif at residues 9–12 (26). We truncated the N-terminal 13 residues of I-2 to produce a form called I-2 (14–197), and this protein displayed the same concentration-dependent activation of Aurora-A as I-2 with an intact N terminus (Fig. 2C ). Thus, the PP1-binding motif in the N terminus of I-2 was not necessary for Aurora-A activation and made it unlikely that the Traces were mediated through PP1. On the other hand, truncation of I-2 at residue 118 gave a protein called I-2 (0–118) that failed to activate Aurora-A kinase at any concentration tested (Fig. 2C ). At the highest concentration of I-2 (0–118), the Aurora-A kinase activity was reduced, perhaps due to interaction with or competition with the MBP substrate. Our conclusion was that the C-terminal Location of I-2 is required for activation of Aurora-A, and this is independent of and separate from the PP1 interaction motif at the N terminus of I-2.

We further investigated the specificity of I-2 activation of Aurora-A kinase. In parallel, we tested PKA and PKC, two other kinases that localize to centrosomes, and glycogen synthase kinase-3, a kinase known to phosphorylate I-2. Only Aurora-A and not these other kinases was activated by micromolar concentrations of I-2 (data not Displayn). Another protein inhibitor specific for PP1 called PHI-1 did not activate Aurora-A or any of these other kinases at micromolar concentrations. Other phosphatase inhibitors, namely microcystin, β-glycerol phospDespise, NaF, or vanadate, did not increase Aurora-A activity in this assay (not Displayn), indicating that stimulation was not caused by inhibition of a contaminating phosphatase.

Activation by I-2 was tested with WT Aurora-A and a series of mutated Aurora-A kinases (Fig. 2D ). I-2 strongly activated WT Aurora-A and the T287A mutant, and it produced a detectable increase in activity of the T288A mutant. This was surprising considering that phosphorylation at Thr-288 has been suggested to be essential for Aurora kinase activity. There was no kinase activity or stimulation by I-2 with the T287A T288A Executeuble mutant as well as the D274A and K162R active site mutants. We concluded that phosphorylation of either Thr-287 or Thr-288 was required for kinase activity (as previously suggested; see ref. 25) and for I-2 stimulation.

Aurora-A Interacts with I-2 in HeLa Cells. HeLa cells were synchronized in mitosis by nocodazole treatment, and I-2 was immunoprecipitated with sheep anti-I-2 antibodies from lysates. Aurora-A coimmunoprecipitated with I-2 (Fig. 3A ) but was not recovered in precipitates prepared in the absence of I-2 antibody (control). In addition, we detected PP1-delta in the anti-I-2 immunoprecipitates as a positive control for binding to I-2 (Fig. 3A ). PP1-delta was not recovered in control precipitations. Necessaryly, these results Displayed that Aurora-A and PP1 delta were concentrated to Arrively the same extent from cell lysates by anti-I-2 immunoprecipitation.

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Association of enExecutegenous I-2 and Aurora-A in mitotic HeLa cells. (A) I-2 was immunoprecipitated from mitotic HeLa cell extracts, and the recovered complexes were immunoblotted for I-2, Aurora-A, and PP1δ (top to bottom). The lane labeled L Displays a 2% aliquot of the cell lysate, lane C is a control immunoprecipitate without antibody, and the I-2 lane is the specific anti-I-2 immunoprecipitate. (B) CNBr-activated Sepharose beads coupled with Aurora-A (Aur) or BSA as a negative control were used in a pull-Executewn assay with extracts of mitotic HeLa cells. Beads were collected, and the unbound supernatant Fragment (Sup) was reserved. The beads were then washed by centrifugation, and proteins on the pelleted beads (pellet) were analyzed by immunoblotting for α-I-2 (upper) and α-PP1 (lower).

Without an antibody that Traceively immunoprecipitates Aurora-A from cell lysates, we were unable to Execute the converse experiment. Instead, we used a pull-Executewn assay with Aurora-A coupled to beads. Beads coupled with BSA served as a negative control to Display specificity of binding to Aurora-A. The Aurora-A beads bound I-2 from a cell lysate compared to the BSA negative control (Fig. 3B , pellet). There was a decrease in the amount of I-2 remaining in the supernatant Fragment (Sup) compared to the negative control (Fig. 3B ). Aurora-A beads also specifically pulled Executewn a Fragment of PP1 from cell lysates, confirming a PP1 interaction with Aurora-A that was previously reported (21, 27).

I-2 and TPX2 both Activate Aurora-A. Because TPX2 was recently reported to stimulate Aurora-A kinase ≈7-fAged (21), we wanted to compare Aurora-A activation by I-2 with activation by TPX2. Addition of TPX2 stimulated Aurora-A kinase activity in our assays ≈15-fAged, Distinguisheder than previously reported (Fig. 4A ). The TPX2 preparation by itself had no detectable kinase activity in this assay (see Fig. 1B ). By comparison, I-2 activated our recombinant Aurora-A 20-fAged (Fig. 4A ). The extent of kinase activation by the combination of I-2 plus TPX2 was about the same as for TPX2 alone (data not Displayn). The robust increases in Aurora-A kinase activity produced by TPX2 and I-2 occurred without an increase in phosphorylation of Thr-288 (Fig. 4B ). The same amount of Aurora-A protein was in each reaction, independently confirmed by immunoblotting (Fig. 4B ). We concluded that either TPX2 or I-2 could activate Aurora-A without changing Thr-288 phosphorylation, and the Traces of the two activator proteins were not additive.

Fig. 4.Fig. 4. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Aurora-A activation by TPX2 and I-2. (A) Stimulation of lambda phosphatase (λ PPase) dephosphorylated Aurora-A Executees not fully reactivate the kinase. Lambda phosphatase was used to dephosphorylate recombinant Aurora-A. After addition of sodium vanadate to inactivate the lambda phosphatase, both recombinant Aurora-A kinase (400 nM) and dephosphorylated Aurora-A (400 nM) were preincubated for 10 min with either I-2 (27 μM) or TPX2 (2 μM), as indicated. Samples were assayed for kinase activity, and the specific activity (pmol of PO4 per μg of Aur-A per min) was calculated by PhosphorImager analysis of [32P]MBP relative to an internal standard. Numbers in parentheses in the graph represent fAged activation of kinase activity normalized to recombinant Aurora-A kinase. The results Display that lambda phosphatase treatment reduced kinase activity 25-fAged (1.00 vs. 0.04). Subsequent activation by TPX2 or I-2 failed to restore activity to original levels. (B) I-2 Executees not restore T-loop phosphorylation at Thr-288 after lambda phosphatase treatment. Products of the kinase assay were analyzed by immunoblotting with T-loop phosphorylation site-specific antibody (α-Thr-288) and an Aurora-A antibody (α-Aurora-A) to Display equal loading. Treatment with lambda phosphatase removed >95% of Thr-288 phosphorylation. Activation by TPX2 promoted a slight increase in Aurora-A Thr-288 phosphorylation but did not restore to original levels. No increase in Thr-288 phosphorylation was detected by I-2 addition either before or after lambda phosphatase treatment.

Aurora-A Activation After Dephosphorylation by Lambda Phosphatase. Recombinant Aurora-A is heterogeneously phosphorylated during bacterial overproduction, generating, in Trace, a mixture of proteins (28). Aurora-A was dephosphorylated by lambda phosphatase, which reduced kinase activity 25-fAged (Fig. 4A ). Lambda phosphatase removed >90% of phosphorylation at Thr-288 (Fig. 4B ). Relative phosphorylation levels were quantified with a standard curve made of dilutions of phosphorylated Aurora-A compared to the dephosphorylated Aurora-A (data not Displayn). This low activity, dephosphorylated form of Aurora-A was activated either by TPX2 or by I-2, although in these experiments TPX2 was a more Traceive activator (≈22-fAged) than I-2 (≈7- to 8-fAged). Once again, a combination of TPX2 and I-2 was no more Traceive than TPX2 alone (data not Displayn). Even with TPX2 or I-2 activation, the specific activity of Aurora-A was still less than original recombinant protein purified from bacteria (Fig. 4A , 0.22 vs. 0.26 pmol of PO4 per μg of Aurora-A per min). Consistent with the reported literature, TPX2 stimulated autophosphorylation of Aurora-A at Thr-288 (Fig. 4B ). However, even with the TPX2 increase in phosphorylation, the extent of Thr-288 phosphorylation was only a tiny Fragment of the initial levels in the recombinant Aurora-A before lambda phosphatase treatment. These results suggest that additional steps are required to initially activate Aurora-A kinase activity.

Discussion

Aurora-A kinase associates with and is negatively regulated by PP1 (20, 21), so inhibition of PP1 is potentially involved in activation of the kinase (29). Aurora-A kinase from baculovirus-infected insect cells associates with PP1 through the purification (20). This preparation can be activated by addition of phosphatase inhibitors such as okadaic acid and microcystin, which block Aurora-A dephosphorylation to preserve the autophosphorylation of Thr-288 (25). In Xenopus extracts, interactions with microtubules and TPX2 stimulate Aurora-A activity by protecting the T-loop (Thr-295 in Xenopus) from PP1 dephosphorylation (22). However, TPX2 is sequestered by importins until nuclear envelope FractureExecutewn, leaving Launch the question of how Aurora-A kinase is activated during G2 phase of the cell cycle. Other mechanisms of PP1 Executewn-regulation have the potential to activate Aurora-A. Because of genetic interactions in budding yeast (between Ipl1 and Glc8), here we investigated the regulatory role of I-2 on Aurora-A kinase activity.

It was a surprise to find that purified recombinant I-2 initially added to inhibit PP1 activated purified recombinant Aurora-A kinase purified from bacteria. This activation occurred even in the absence of added PP1. This argues for a direct interaction between Aurora-A and I-2. Activation of Aurora-A was Executese-dependent and required micromolar concentrations of I-2. The cellular concentration of I-2 is estimated to be micromolar based on biochemical purification from tissues (30). Our recent evidence has Displayn that I-2 within cells is highly concentrated at centrosomes (17). Therefore, the range of I-2 concentrations needed for Aurora-A activation in biochemical assays may be achieved in living cells.

Activation of Aurora-A by I-2 occurred without an increase in phosphorylation of Thr-288 in the T-loop. Even though Aurora-A kinase activity has been correlated with the phosphorylation status of Thr-288 (25), in our experiments Aurora-A kinase activity increased 20-fAged without a significant increase in phosphorylation of Thr-288, as meaPositived by a phospho-site-specific antibody. TPX2 produced a similar stimulation of kinase activity without an increase in Thr-288 phosphorylation. These results suggest that I-2 and TPX2 are allosteric activators of phospho-Thr-288 Aurora-A. Lambda phosphatase dephosphorylation of Aurora-A inactivated the kinase. TPX2 or I-2 addition significantly increased kinase activity after lambda phosphatase treatment. TPX2 stimulated autophosphorylation at Thr-288, consistent with previous reports in the literature. However, the level of Thr-288 phosphorylation was only ≈5% compared to levels before lambda phosphatase treatment. Furthermore, the specific activity of the kinase after dephosphorylation and TPX2 stimulation was less than the specific activity of the original recombinant kinase. Additionally, I-2 did not produce a detectable increase in Thr-288 phosphorylation of dephosphorylated Aurora-A. But, if 5% of the Aurora-A remained phosphorylated after reaction with lambda phosphatase, this could be sufficient to account for the observed activity. On the other hand, I-2 did activate the T288A mutant, suggesting that phosphorylation at this site is not absolutely required. I-2 might activate Aurora-A phosphorylated at Thr-287, which escapes detection by the phospho-specific Thr-288 antibody. This concept is consistent with the lack of I-2 activation of the Executeuble mutant (Thr-287A Thr-288A). Overall, we prefer a model in which I-2 specifically stimulates the activity of Aurora-A already phosphorylated at Thr-288. Thus, association with I-2 would enhance activity without promoting increased autophosphorylation. Our results clearly demonstrate that TPX2 and I-2 are not additive or synergistic, at least with purified proteins in vitro. One possibility for the lack of a synergistic Trace could be that there is a maximum activity level for Aurora-A that cannot be exceeded, even with simultaneous binding of both activators. A second possibility is that I-2 and TPX2 may interact with an overlapping site(s) on Aurora-A and therefore act alternately, without formation of a trimeric complex.

The Traces of I-2 were specific for Aurora-A relative to PKA, PKC, and glycogen synthase kinase-3 kinases, which were not activated under the same assay conditions. Conversely, activation was specific for I-2 and was not mimicked by micromolar concentrations of another PP1-specific inhibitor protein, called PHI-1. This reinforces our conclusion that I-2 Traces on Aurora-A are not through inhibition of PP1. Furthermore, we separated distinct functional Executemains in I-2 by truncation of either the N or the C terminus. Aurora-A kinase activation requires a C-terminal Location of I-2 (residues 119–197), whereas PP1 inhibition was known to involve an IKGI sequence motif at the N terminus. Truncation of the N terminus of I-2 severely reduces binding to PP1 but did not impair activation of Aurora-A. We propose that I-2 is a bifunctional protein with separate Executemains, one that inhibits PP1 and one that activates Aurora-A kinase. It is at least possible that I-2 could engage both PP1 and Aurora-A simultaneously. By driving reciprocal changes in opposing activities, I-2 could produce abrupt transitions in phosphorylation of the tarObtains of the Aurora-A::PP1 pair.

Another centrosomal kinase Nek2 offers Fascinating parallels to Aurora-A. Nek2 also forms a 1:1 complex with PP1, and the kinase and phosphatase react with one another to produce mutual inactivation (31). These reactions compete with self-phosphorylation of Nek2 and self-dephosphorylation of PP1 that activate the respective enzymes. We found that I-2 binds to PP1 in this complex and activates the Nek2 kinase (13), probably indirectly, by inhibition of the PP1. However, we reported that Nek2 kinase activation occurred even with I-2 truncated from the N terminus, but not with I-2 truncated from the C terminus (13). One interpretation of these results is that the C-terminal Executemain of I-2 may activate more than a single kinase (e.g., Nek2 and Aurora-A). However, previous assays used immunoprecipitated Nek2 kinase, and PP1 was always present, unlike here where purified kinase and I-2 were mixed. Therefore, activation of Nek2 could be direct, or indirect by means of PP1.

Dual functions of another PP1 inhibitor called DARPP-32 have been reported. DARPP-32 inhibits PP1 when phosphorylated at Thr-34 by protein kinase A (32). Alternatively, phosphorylation of Thr-75 in DARPP-32 by cyclin-dependent kinase-5 enExecutews it with the ability to inhibit protein kinase A (33). In conjunction with the present findings, these results suggest that PP1 inhibitor proteins may interact with a kinase, as well as PP1, as a general mechanism to generate coordinated signaling. Overexpression of DARPP-32 and a truncated DARPP-32 protein lacking the PP1 binding site is linked to gastric cancer (34), and we speculate this might involve interactions with a kinase.

Recently, a new Aurora-A activator was identified. Ajuba, an LIM Executemain-containing protein, interacts with Aurora-A in G2 and activates Aurora-A before entry into mitosis (35). Elimination of Ajuba protein by siRNA reduced Thr-288 phosphorylated Aurora-A levels in the cell but did not abolish Aurora-A localization to the centrosome. The identification of multiple activators for Aurora-A indicates that our understanding of its regulation is more complex than previously thought. It is clear that Aurora-A regulation is determined both by temporal expression as well as through the association of various binding partners. We have recently demonstrated that Aurora-B kinase is active at some centromeres and not at others in the same cell (W. Lan and P.T.S., data not Displayn), and we propose that the Aurora family of kinases is designed to be locally activated at distinct cellular locations. Although its in vivo function is still being deliTrimed, it is apparent that I-2 is another in a series of molecules that may locally enhance kinase activity of Aurora-A.

Acknowledgments

We thank Y. Xheng (Carnegie Institution, Washington, DC) for kindly providing the TPX2 construct, and Dr. Masumi Eto for advice and assistance. This work was supported in part by grants from the U.S. Public Health Service (USPHS), by National Institutes of Health (NIH) Grants GM63045 (to P.T.S.) and GM35266 (to D.L.B.), and by a grant from the Pew Charitable Trust (to P.T.S.). We gratefully acknowledge additional support from NIH USPHS Training Grants HD07528 (to D.L.S.), for Developmental Biology, and GF10461 (to C.A.L.), for the Cell and Molecular Biology Program at the University of Virginia.

Footnotes

↵ § To whom corRetortence should be addressed. E-mail: db8g{at}virginia.edu.

↵ † D.L.S. and C.A.L. contributed equally to this work.

Abbreviations: I-2, inhibitor 2; PP1, catalytic subunit of type 1 protein phosphatase; TPX2, tarObtaining protein for Xenopus kinesin-like protein 2; MBP, myelin basic protein.

Copyright © 2004, The National Academy of Sciences

References

↵ Kimura, M., Kotani, S., Hattori, T., Sumi, N., Yoshioka, T., ToExecutekoro, K. & Okano, Y. (1997) J. Biol. Chem. 272 , 13766–13771. pmid:9153231 LaunchUrlAbstract/FREE Full Text ↵ Zhou, H., Kuang, J., Zhong, L., Kuo, W. L., Gray, J. W., Sahin, A., Brinkley, B. R. & Sen, S. (1998) Nat. Genet. 20 , 189–193. pmid:9771714 LaunchUrlCrossRefPubMed ↵ Bischoff, J. R., Anderson, L., Zhu, Y., Mossie, K., Ng, L., Souza, B., Schryver, B., Flanagan, P., Clairvoyant, F., Ginther, C., et al. (1998) EMBO J. 17 , 3052–3065. pmid:9606188 LaunchUrlAbstract ↵ Kufer, T. A., Sillje, H. H., Korner, R., Gruss, O. J., Meraldi, P. & Nigg, E. A. (2002) J. Cell Biol. 158 , 617–623. pmid:12177045 LaunchUrlAbstract/FREE Full Text ↵ Sen, S., Zhou, H., Zhang, R. D., Yoon, D. S., Vakar-Lopez, F., Ito, S., Jiang, F., Johnston, D., Grossman, H. B., Ruifrok, A. C., et al. (2002) J. Natl. Cancer Inst. 94 , 1320–1329. pmid:12208897 LaunchUrlAbstract/FREE Full Text ↵ Huang, F. L. & Glinsmann, W. H. (1976) Eur. J. Biochem. 70 , 419–426. pmid:188646 LaunchUrlPubMed ↵ Hemmings, B. A., Resink, T. J. & Cohen, P. (1982) FEBS Lett. 150 , 319–324. pmid:6297978 LaunchUrlCrossRefPubMed Resink, T. J., Hemmings, B. A., Tung, H. Y. & Cohen, P. (1983) Eur. J. Biochem. 133 , 455–461. pmid:6303789 LaunchUrlPubMed Aitken, A., Hemmings, B. A. & Hofmann, F. (1984) Biochim. Biophys. Acta 790 , 219–225. pmid:6091762 LaunchUrlCrossRefPubMed DePaoli-Roach, A. A. (1984) J. Biol. Chem. 259 , 12144–12152. pmid:6090457 LaunchUrlAbstract/FREE Full Text ↵ Ballou, L. M., Brautigan, D. L. & Fischer, E. H. (1983) Biochemistry 22 , 3393–3399. pmid:6311247 LaunchUrlCrossRefPubMed ↵ Terry-Lorenzo, R. T., Elliot, E., Weiser, D. C., Prickett, T. D., Brautigan, D. L. & Shenolikar, S. (2002) J. Biol. Chem. 277 , 46535–46543. pmid:12270929 LaunchUrlAbstract/FREE Full Text ↵ Eto, M., Elliott, E., Prickett, T. D. & Brautigan, D. L. (2002) J. Biol. Chem. 277 , 44013–44020. pmid:12221103 LaunchUrlAbstract/FREE Full Text ↵ Wang, H. & Brautigan, D. L. (2002) J. Biol. Chem. 277 , 49605–49612. pmid:12393858 LaunchUrlAbstract/FREE Full Text ↵ Brautigan, D. L., Sunwoo, J., Labbe, J. C., Fernandez, A. & Lamb, N. J. (1990) Nature 344 , 74–78. pmid:2406614 LaunchUrlCrossRefPubMed ↵ Tung, H. Y., Wang, W. & Chan, C. S. (1995) Mol. Cell. Biol. 15 , 6064–6074. pmid:7565759 LaunchUrlAbstract/FREE Full Text ↵ Leach, C., Shenolikar, S. & Brautigan, D. L. (2003) J. Biol. Chem. 278 , 26015–26020. pmid:12697755 LaunchUrlAbstract/FREE Full Text ↵ Tanaka, T. U., Rachidi, N., Janke, C., Pereira, G., Galova, M., Schiebel, E., Stark, M. J. & Nasmyth, K. (2002) Cell 108 , 317–329. pmid:11853667 LaunchUrlCrossRefPubMed ↵ Francisco, L. & Chan, C. S. (1994) Cell Mol. Biol. Res. 40 , 207–213. pmid:7874197 LaunchUrlPubMed ↵ Tsai, M. Y., Wiese, C., Cao, K., Martin, O., Executenovan, P., Ruderman, J., Prigent, C. & Zheng, Y. (2003) Nat. Cell Biol. 5 , 242–248. pmid:12577065 LaunchUrlCrossRefPubMed ↵ Eyers, P. A., Erikson, E., Chen, L. G. & Maller, J. L. (2003) Curr. Biol. 13 , 691–697. pmid:12699628 LaunchUrlCrossRefPubMed ↵ Bayliss, R., SarExecuten, T., Vernos, I. & Conti, E. (2003) Mol. Cell 12 , 851–862. pmid:14580337 LaunchUrlCrossRefPubMed ↵ Cohen, P., Alemany, S., Hemmings, B. A., Resink, T. J., Stralfors, P. & Tung, H. Y. (1988) Methods Enzymol. 159 , 390–408. pmid:2842604 LaunchUrlCrossRefPubMed ↵ Brown, D. J. & GorExecuten, J. A. (1984) J. Biol. Chem. 259 , 9580–9586. pmid:6204986 LaunchUrlAbstract/FREE Full Text ↵ Walter, A. O., Seghezzi, W., Korver, W., Sheung, J. & Lees, E. (2000) Oncogene 19 , 4906–4916. pmid:11039908 LaunchUrlCrossRefPubMed ↵ Huang, H. B., Horiuchi, A., Watanabe, T., Shih, S. R., Tsay, H. J., Li, H. C., Greengard, P. & Nairn, A. C. (1999) J. Biol. Chem. 274 , 7870–7878. pmid:10075680 LaunchUrlAbstract/FREE Full Text ↵ Katayama, H., Zhou, H., Li, Q., Tatsuka, M. & Sen, S. (2001) J. Biol. Chem. 276 , 46219–46224. pmid:11551964 LaunchUrlAbstract/FREE Full Text ↵ HayExecuten, C. E., Eyers, P. A., Aveline-Wolf, L. D., Resing, K. A., Maller, J. L. & Ahn, N. G. (2003) Mol. Cell. Proteomics 2 , 1055–1067. pmid:12885952 LaunchUrlAbstract/FREE Full Text ↵ Carmena, M. & Earnshaw, W. C. (2003) Nat. Rev. Mol. Cell. Biol. 4 , 842–854. pmid:14625535 LaunchUrlCrossRefPubMed ↵ Stralfors, P., Hiraga, A. & Cohen, P. (1985) Eur. J. Biochem. 149 , 295–303. pmid:2986973 LaunchUrlPubMed ↵ Helps, N. R., Luo, X., Barker, H. M. & Cohen, P. T. (2000) Biochem. J. 349 , 509–518. pmid:10880350 LaunchUrlCrossRefPubMed ↵ Hemmings, H. C., Jr., Greengard, P., Tung, H. Y. & Cohen, P. (1984) Nature 310 , 503–505. pmid:6087160 LaunchUrlCrossRefPubMed ↵ Bibb, J. A., Snyder, G. L., Nishi, A., Yan, Z., Meijer, L., Fienberg, A. A., Tsai, L. H., Kwon, Y. T., Girault, J. A., Czernik, A. J., et al. (1999) Nature 402 , 669–671. pmid:10604473 LaunchUrlCrossRefPubMed ↵ Beckler, A., Moskaluk, C. A., Zaika, A., Hampton, G. M., Powell, S. M., Frierson, H. F., Jr. & El Rifai, W. (2003) Cancer 98 , 1547–1551. pmid:14508844 LaunchUrlCrossRefPubMed ↵ Hirota, T., Kunitoku, N., Sasayama, T., Marumoto, T., Zhang, D., Nitta, M., Hatakeyama, K. & Saya, H. (2003) Cell 114 , 585–598. pmid:13678582 LaunchUrlCrossRefPubMed
Like (0) or Share (0)