Aurora kinases are a family of serine/threonine protein kinases that play essential roles in mitosis and cytokinesis. AurB (Aurora B kinase) has shown a clear link to cancer and is being pursued as an attractive cancer target. Multiple small molecules targeting AurB have entered the clinic for the treatment of cancer. A protein cofactor, INCENP (inner centromere protein), regulates the cellular localization and activation of AurB. In the present study, we examined the effect of INCENP on the activation kinetics of AurB and also elucidated the kinetic mechanism of AurB-catalysed substrate phosphorylation. We have concluded that: (i) substoichoimetric concentrations of INCENP are sufficient for AurB autophosphorylation at the activation loop residue Thr232, and hence INCENP plays a catalytic role in AurB autophosphorylation; (ii) AurB/INCENP-catalysed phosphorylation of a peptide substrate proceeds through a rapid equilibrium random Bi Bi kinetic mechanism; and (iii) INCENP has relatively minor effects on the specific activity of AurB using a peptide substrate when compared with its role in AurB autoactivation. These results indicate that the effects of INCENP, and probably accessory proteins in general, may differ when enzymes are acting on different downstream targets.
Aurora kinases are a family of serine/threonine protein kinases that play critical roles in mitosis and cytokinesis [1–4]. In mammals, three Aurora kinases, Aurora A, B and C kinase (AurA, AurB and AurC respectively), have been identified. These share a high degree of sequence homology, especially in the catalytic domain. Yet, the isoenzymes differ in their cellular functions, subcellular localization and in vivo binding partners. Among the three Aurora kinases, AurA and AurB have been more extensively studied. Gene amplification and overexpression of AurA and AurB have been observed in a variety of cancer cell lines. Inhibition of AurA and/or AurB has been shown to induce tumour regression in mouse xenograft models. Currently, a number of Aurora kinase inhibitors are in clinical trials for the treatment of cancer [1,5–7].
AurA is mainly associated with centrosomes and nearby microtubules and is involved in centrosome separation and maturation, as well as spindle assembly and orientation during mitosis [8,9]. TPX2 (target protein for Xenopus kinesin-like protein 2) binds to AurA, alters its interactions with substrates and inhibitors, and regulates its biological activities [10,11]. AurB plays essential roles in both mitosis and cytokinesis [2,4]. During mitosis, AurB is required for maintaining proper chromosome alignment, separation and segregation. It is also involved in activating the anaphase checkpoint which prevents the onset of anaphase in cells with misattached kinetochores. During cytokinesis, AurB is concentrated in the cortex at the cleavage furrow and the midbody. Here it regulates a number of critical processes involved in cytokinesis and co-ordinates chromosome separation and segregation during cell division. AurC is highly expressed in the testis. It plays a role in meiosis and may function in mitosis and cytokinesis by complementing AurB activities [12,13].
Given the critical and diverse roles of AurB in mitosis and cytokinesis, its cellular functions require proper regulation. AurB is regulated through cell-cycle-dependent expression and degradation, through phosphorylation-mediated activation/inactivation and through its dynamic subcellular localizations during mitosis and cytokinesis. In addition, AurB is regulated through interactions with a number of its cellular protein-binding partners, including INCENP (inner centromere protein) and Survivin [14–16].
INCENP is both a substrate and an accessory protein of AurB and is involved in AurB cellular targeting as well as its activation. INCENP has several functional domains. Sequence alignment revealed a conserved sequence motif, the IN box, on the C-terminal region of this protein. Further studies indicated that the IN box region [including the adjacent TSS (Thr-Ser-Ser) motif] is required for interaction with AurB and stimulation of AurB kinase activity [14–16].
INCENP-mediated AurB activation is a complex process and requires the phosphorylation of not only the Thr232 of AurB (human sequence) in the activation loop, but also the two serine residues of the TSS motif of INCENP. Previous studies have suggested a two-step model for INCENP-mediated AurB activation that begins with the binding of unphosphorylated INCENP to phospho-AurB to produce a complex with intermediate kinase activity. AurB-mediated phosphorylation of INCENP on the TSS motif then elicits full kinase activity [14,16]. However, the impact of INCENP on the kinetics of AurB activation has not been characterized.
In the present study, we carried out detailed kinetic analysis of INCENP-mediated AurB activation and also elucidated the reaction mechanism of AurB catalysis. Experiments with stoichiometric and substoichiometric amounts of an INCENP peptide (amino acids 826–919) containing both the IN box and the TSS motif reveal a catalytic role for INCENP in AurB phosphorylation during activation. Our results also indicate a rapid equilibrium random Bi Bi kinetic mechanism for AurB-mediated phosphorylation of a peptide substrate.
MATERIALS AND METHODS
ATP was purchased from Sigma–Aldrich. [γ-33P]ATP (10 mCi/ml in 10 mM Tricine) was purchased from PerkinElmer. The peptide substrate and inhibitors (Table 1) were synthesized by 21st Century Biochemicals. Freeze-dried peptides were resuspended in deionized water, divided into 10 μl aliquots and stored at −80 °C. Concentrations were determined by amino acid analysis. INCENP [human sequence, residues 826–919, GST (glutathione transferase)-tagged] was obtained from the University of Dundee, Dundee, U.K. Staurosporine was purchased from Sigma–Aldrich. All other reagents were analytical grade or higher.
|Peptide name||Peptide sequence|
|Peptide name||Peptide sequence|
Construct design and expression of full-length AurB-(2–344)
A gene encoding full-length human AurB-(2–344) was used to construct the plasmid pDest8-FLAG-His6-↓-AurB-(2–344) using the Gateway™ cloning system (Invitrogen). N-terminal FLAG and His6 tags were added along with a thrombin (↓) cleavage site. Protein expression was carried out in Sf9 insect cells using the Bac-To-Bac Baculovirus expression system (Invitrogen) with the addition of 5 nM okadaic acid (Calbiochem) 4 h before harvest.
Purification of AurB-(2–344)
Baculovirus Sf9 cells containing the expressed FLAG-His6-↓-AurB-(2–344) construct were resuspended in lysis buffer [50 mM Tris/HCl (pH 8.0), 250 mM NaCl, 5 mM 2-mercaptoethanol, 1 mM tris-(3-hydroxypropyl)phosphine, 5 nM okadaic acid, 10 mM NaF, 10 mM glycerol phosphate, 0.1 mM sodium vanadate, protease and phosphatase inhibitor cocktails (Sigma P8849 and P2850 respectively)] and lysed by sonication. After centrifugation at 30000 g for 30 min to remove cell debris, the supernatant was incubated with Ni-NTA (Ni2+-nitrilotriacetate) resin and incubated at 4 °C for 1 h. The mixture was column-packed, and the unbound impurities were removed by washing with a buffer containing 50 mM Tris/HCl (pH 8.0) and 5 mM 2-mercaptoethanol. Bound protein was eluted with a 0–300 mM imidazole gradient in the same buffer. Fractions were analysed by SDS/PAGE (10% gels), and peak fractions of FLAG-His6-↓-AurB-(2–344) were pooled and dialysed against buffer A [50 mM Tris/HCl (pH 8.0), 250 mM NaCl and 2 mM DTT (dithiothreitol)] to remove imidazole. The protein recovered was 65% pure and was stored in buffer A plus 10% (v/v) glycerol. Protein concentration was determined by the Bradford method, and purity was estimated from SDS/PAGE analysis. Edman sequencing confirmed the correct N-terminus of AurB. No phosphorylation on Thr232 of the activation loop was observed by either MS or Western blotting.
AurB activation reactions were carried out in the activation buffer containing 30 mM Tris/HCl (pH 8.0), 2 mM magnesium acetate, 400 μM ATP, 0.1 mM EGTA, 0.1% 2-mercaptoethanol and 0.1 mM sodium vanadate. Typically, 2 μM AurB and various amounts of INCENP were added to the above activation buffer. The reaction was allowed to proceed at 30 °C for 3 h and dialysed for 4 h against a storage buffer containing 50 mM Tris/HCl (pH 7.5), 270 mM sucrose, 150 mM NaCl, 0.1 mM EGTA, 0.1% 2-mercaptoethanol, 1 mM benzamidine and 0.2 mM PMSF.
To monitor the extent of AurB activation, aliquots of 1 μl were taken at various times from the activation reaction and quenched by 100-fold dilution into the reaction buffer for the activity assay (see the next section below). It was shown that no further activation of AurB took place after quenching. A 20 μl portion of the quenched activation assay aliquot was then mixed with 20 μl of the substrate mixture containing 5 μM ATP and 10 μM peptide substrate. The extent of AurB activation was determined by measuring the AurB kinase activity as described below. All experiments on AurB activation and activity were performed with n≥2.
AurB kinase assay
The AurB kinase assay was carried out in 50 mM Hepes buffer (pH 7.5) containing 2 mM MgCl2, 25 mM NaCl, 2 mM DTT, 0.15 mg/ml BSA, 0.01% Tween 20 and phosphatase inhibitor cocktail (Sigma P2850). Concentrations of ATP, the peptide substrate and inhibitors are indicated in the Figure and Table legends. [γ-33P]ATP (10 mCi/ml) was included at a specific radioactivity of 500–2000 c.p.m./pmol. To measure the extent of AurB activation, AurB/INCENP was added to the reaction mixture at a final concentration of 10 nM, and the reactions were allowed to run for 60 min. To measure the kinetic mechanism of AurB/INCENP catalysis, 2.5 nM fully activated AurB/INCENP 1:1 complex was tested, and the reactions were stopped after 15 min. Reactions were analysed using a filter-binding format as reported previously .
Phosphorylation of AurB and INCENP was monitored by two methods. Total phosphorylation of both proteins was determined by measuring the incorporation of the 33P radiolabel during activation with [γ-33P]ATP. The AurB activation reactions were carried out as described above and quenched at various times by the addition of SDS/PAGE loading buffer. Quenched samples were analysed by SDS/PAGE (12% gels), and the amount of 33P incorporation in AurB and INCENP was quantified using a PhosphorImager (Storm, Molecular Dynamics).
Site-specific phosphorylation of AurB on Thr232 and of INCENP on Ser884/Ser885 in the TSS motif was determined by LC (liquid chromatography)–MS/MS (tandem MS). Protein samples were separated by SDS/PAGE (12% gels) and stained with colloidal Coomassie Blue. Aurora B and INCENP bands were excised, reduced, alkylated and digested with either trypsin (AurB) or Lys-C (INCENP). The resulting peptides from both proteins were pooled and analysed by LC–MS/MS using MRM (Multiple Reaction Monitoring)  on a Sciex 4000 QTRAP triple-quadrupole mass spectrometer coupled to a nanolitre flow HPLC (75 μm PepMap C18 column). The abundance of each phosphorylated and non-phosphorylated peptide was determined by monitoring the production of fragment ions specific for each phosphorylated and non-phosphorylated species and summing the individual fragment ion abundances for each respective species. The apparent stoichiometry at each site of phosphorylation was derived from the ratio of the abundance of phosphorylated peptide to total peptide (phosphorylated plus non-phosphorylated).
In the case of AurB, the phosphorylated peptide was two residues longer than the non-phosphorylated peptide because trypsin cleavage at Lys231 was inhibited by phosphorylation on the adjacent residue, Thr232. To account for the likely differences in the ionization efficiency of these two peptides, we prepared a synthetic standard for each and measured the ratio of their intensities as the pair was mixed in varying proportions. From these measurements, we determined a response factor which was used to correct the apparent stoichiometry of the native peptides.
where [E0], [A] and [B] represent the enzyme, ATP and the peptide substrate concentrations respectively, kcat is the turnover number, Ka and Kb are the Michaelis constant for ATP and the peptide substrate respectively, and Kia is the dissociation constant for ATP.
where Kis and Kii are the dissociation constants for inhibitors from either free enzyme or the enzyme–substrate complex respectively, and [I] is the inhibitor concentration. The α (Kis/Kii) value is generated from this fitting. The inhibitor is considered competitive, mixed-type or uncompetitive when α>10, α=0.1–10 or α<0.1 respectively.
Effect of INCENP on AurB phosphorylation during activation
The effect of INCENP on the kinetics of AurB activation was examined using a GST-tagged truncated INCENP (residues 826–919). This INCENP construct contained the IN box domain and the TSS motif shown to be necessary for interacting with AurB and stimulation of its kinase activity [14,16]. To determine the optimal conditions for AurB activation, various buffer components were titrated and their effects on AurB activation were determined (results not shown). The final optimized activation buffer system contained 30 mM Tris/HCl (pH 8.0), 2 mM magnesium acetate, 400 μM ATP, 0.1 mM EGTA, 0.1% 2-mercaptoethanol and 0.1 mM sodium vanadate.
The activation reactions were carried out with 2 μM AurB and 2 or 0.2 μM INCENP. At each time point, AurB kinase activity, as well as phosphorylation of AurB and INCENP, were determined. In the absence of INCENP, the amount of AurB autoactivation as measured by the ability of AurB to phosphorylate a peptide substrate is negligible. At both INCENP concentrations, there is significant activation of the AurB kinase activity over time which reached a plateau at approx. 120 min (Figure 1A). Both time courses of AurB activation include an initial lag phase, suggesting that the autoactivation reaction is intermolecular in nature. To confirm the effect of INCENP on AurB activation, a 66-mer INCENP peptide corresponding to INCENP-(835–900) was also tested and found to give an activating effect very similar to those observed with the GST–INCENP construct (results not shown).
Activation of AurB and phosphorylation of AurB and INCENP at different AurB/INCENP ratios
The amount of total phosphorylation of AurB as monitored by 33P incorporation increased over time and was similar at both INCENP concentrations (Figure 1B). An early lag phase was observed, and the phosphorylation time course of AurB closely resembled its activation time course. Furthermore, INCENP also underwent rapid and extensive phosphorylation during the reaction.
Since multiple phosphorylation sites are present in both proteins , the phosphorylation of Thr232 on AurB and Ser884/Ser885 in the TSS motif of INCENP was monitored over time by MS under the same experimental conditions as described above. After purification of the reactions by SDS/PAGE, the proteins were excised and digested with either trypsin or Lys-C, and the digests were analysed by LC–MS/MS. The overall phosphorylation profile of AurB Thr232 over the 180 min time course (Figure 2A) is similar to the total phosphorylation profile, using either stoichiometric (2 μM) or substoichiometric (0.2 μM) amounts of INCENP. The initial rate of AurB Thr232 phosphorylation is lower at 2 μM INCENP, possibly due to INCENP acting as a competing substrate in the AurB phosphorylation reaction. The final stoichiometry of AurB Thr232 phosphorylation at 180 min is determined to be 70 and 90% for the 2 and 0.2 μM INCENP reactions respectively. The overall phosphorylation profile of the INCENP TSS motif (Figure 2B) takes into consideration contributions from both the mono- and di-phosphorylated forms of the peptide. Final phosphorylation of the INCENP TSS motif is determined to be 90 and 60% at 2 and 0.2 μM INCENP respectively.
Site-specific analysis of AurB Thr232 and INCENP Ser884/Ser885 phosphorylation by MS
The results from these experiments indicate that a substoichiometric amount of INCENP is able to elicit nearly complete phosphorylation of Thr232 on AurB. Therefore the role of INCENP in AurB autophosphorylation appears to be catalytic; that is, stoichiometric binding of INCENP to AurB is not required for AurB autophosphorylation. One possible explanation of these results is that the fraction of AurB that is in the native conformation is low. Therefore INCENP is saturating even at 0.2 μM. However, our phosphorylation analysis shows that ∼80% of AurB is phosphorylated on Thr232 after activation, indicating that the majority of AurB is in the native conformation. Taken together, these data suggest a catalytic role of INCENP in AurB autophosphorylation during activation.
We have also observed that 2 μM INCENP induces a modest 2-fold increase in activity compared with the activation reaction carried out in the presence of 0.2 μM INCENP (Figure 1A), suggesting that INCENP binding imparts additional AurB kinase activity, possibly through the introduction of further conformational changes. Furthermore, INCENP binding increases the stability of the AurB enzyme. Although the 1:1 AurB–INCENP complex retains its activity after freeze–thaw, AurB activated with a substoichiometric amount of INCENP loses ∼50% of activity after one freeze–thaw cycle (results not shown).
Kinetic mechanism of AurB–INCENP-mediated phosphorylation of a peptide substrate
ATP/peptide double titration
We then characterized the kinetic mechanism of AurB-(2–344) activated by a stoichiometric amount of INCENP-(826–919) under the optimized activation conditions.
Initial rates of AurB, activated in the presence of INCENP, were measured at various ATP and peptide substrate concentrations. The data were fitted to eqn (1), and the fitted lines intersect in the double-reciprocal plots, indicating that the reaction follows a sequential, rather than a Ping Pong, mechanism (Figure 3). Km values of 1.1±0.08 μM for the peptide substrate and 13.6±0.8 μM for ATP were obtained from this analysis, with Kd values of 1.0±0.2 and 12.0±2.6 μM respectively. A kcat of 23.3±0.4 min−1 was obtained using eqn (1). These results indicate the formation of a ternary complex along the reaction pathway, but do not differentiate between the random and ordered mechanisms . To determine the order of substrate addition, product and dead-end inhibition studies were carried out. Results were generated by fitting to eqn (2) and are presented in Table 2.
|Variable substrate||Inhibitor||Kis (μM)||Kii (μM)||α||Inhibition pattern|
|Variable substrate||Inhibitor||Kis (μM)||Kii (μM)||α||Inhibition pattern|
ATP/peptide double-titration experiment for AurB/INCENP (1:1)
Inhibition by ADP
The kinetics of AurB–INCENP-catalysed reactions were examined in the presence of ADP, one of the two products of the reaction. When the peptide substrate was fixed at its Km and ATP was varied, the double-reciprocal plot generated at various fixed ADP concentrations indicates a competitive pattern of ADP against ATP (α=14.2) (Table 2). When ATP was fixed at its Km and the peptide substrate was varied, the inhibition pattern of ADP is non-competitive against peptide (α=1.1) (Table 2). These inhibition patterns are consistent with three possible kinetic mechanisms: a rapid-equilibrium random mechanism, a compulsory ordered mechanism with ATP binding first and ADP releasing last, or a Theorell–Chance mechanism with peptide binding first and P-peptide (phospho-peptide) releasing last .
Dead-end inhibition by A-peptide (alanine-peptide) and staurosporine
Dead-end inhibition is a very useful tool in determining the order of substrate binding and product release. The serine residue in the substrate peptide is replaced by alanine in the A-peptide to prevent phosphorylation. Inhibition of the A-peptide is competitive against the peptide substrate (α=25.2) and non-competitive against ATP (α=0.8) (Table 2). The observed competitive inhibition mechanism when ATP is the variable substrate indicates that the peptide substrate can bind before ATP, thus eliminating the ordered mechanism with ATP binding first.
The dead-end inhibitor staurosporine was used as an ATP analogue in the inhibition study. The results indicate that staurosporine is competitive against ATP (α>>10) and non-competitive against the peptide substrate (α=2.8) (Table 2). These results therefore indicate that ATP can bind to the enzyme before the peptide substrate and exclude the Theorell–Chance mechanism with peptide binding first. Taken together, results from the product and dead-end inhibition studies are most consistent with a rapid equilibrium random Bi Bi kinetic mechanism .
Product inhibition by P-peptide
With a random mechanism, inhibition by the product P-peptide, in which the serine residue is replaced by a phosphoserine residue, should give a non-competitive inhibition pattern against ATP and a competitive pattern against the substrate peptide. Data from the P-peptide inhibition experiment confirm the non-competitive pattern against ATP (α=0.8, Kis=246.0±44.1 nM). P-peptide inhibition against the substrate peptide was fitted to eqn (2) and generated an α value of 11, indicating an essentially competitive mode of inhibition (Table 2). Therefore our collective inhibition data with both product and dead-end inhibitors are most consistent with a rapid equilibrium random Bi Bi mechanism for AurB–INCENP (1).
Rapid equilibrium random kinetic mechanism of AurB activated by INCENP
AurB is essential for a number of processes during mitosis and cytokinesis, including chromosome condensation and segregation, and the spindle-assembly checkpoint. Similarly to AurA, AurB has shown a clear link to cancer and is being pursued as an attractive cancer target. Small molecules targeting AurB have entered clinical trials for the treatment of cancer [1,5–7]. A better understanding of the activation and the kinetic mechanism of AurB at the biochemical level will facilitate the design and evaluation of small-molecule inhibitors of AurB.
Previously, biochemical analysis of the effects of INCENP on AurB was carried out mainly with AurB kinase that was already phosphorylated [14,16] or with pre-formed AurB–INCENP complex . In one study, INCENP was titrated into unphosphorylated AurB, and an increase in AurB catalytic activity was observed towards protein substrates . However, the effect of INCENP on the kinetics of AurB activation was not examined in these studies. In the present study, we have specifically examined the role of INCENP on the kinetics of AurB phosphorylation/activation by using unphosphorylated AurB from baculoviral expression. We also analysed the time course of phosphorylation of INCENP during AurB activation.
Our findings indicate that INCENP has a catalytic role in AurB phosphorylation during activation. Our kinetic analysis suggests that AurB–INCENP follows a rapid equilibrium random Bi Bi mechanism, the same as that for AurA–TPX2 that we reported previously . In this mechanism, ATP and the peptide substrate bind to AurB–INCENP in a random order, and an enzyme–ATP–peptide ternary complex is formed before the phosphate transfer step.
Our kinetic analysis on INCENP-mediated AurB phosphorylation/activation suggests a model that agrees with the previous two-step activation model. In this model, INCENP first binds to and is phosphorylated by a small amount of active phospho-AurB in the initial inactive enzyme preparation. The dually phosphorylated AurB–INCENP complex exhibits a higher activity in phosphorylating free AurB and INCENP. Therefore AurB phosphorylation readily proceeds without INCENP being bound. The role of INCENP is to increase the activity of AurB towards phosphorylating INCENP and AurB rather than to render AurB a better substrate for phosphorylation.
In addition, we have also observed that the AurB with substoichiometric amount of INCENP at the 10:1 ratio shows only a 2-fold fall in phosphorylating the peptide substrate compared with AurB with 1:1 ratio of INCENP. Because similar Thr232 phosphorylation is observed for both enzyme forms by MS, our data suggest that, in addition to phosphorylation, INCENP binding to AurB may induce further conformational changes, resulting in enhanced AurB kinase activity. On the other hand, this result is distinct from the role of INCENP in AurB activation. During AurB activation, basal AurB autophosphorylation is extremely low in the absence of INCENP. The addition of even a catalytic amount of INCENP results in a >100-fold increase in the activation of AurB. The crystal structure of AurB/INCENP has revealed limited interaction of INCENP with the active site of AurB . Therefore it is likely that no direct interaction exists between INCENP and the peptide substrate. However, INCENP could facilitate the binding of a protein substrate such as AurB that requires interactions beyond the substrate-binding pocket. Our data suggest that INCENP can induce higher AurB activity when a protein target, rather than a peptide, serves as the substrate.
In summary, our kinetic analysis of AurB activation and catalysis reveals a number of features that have not been reported previously. First, INCENP catalytically increases AurB autophosphorylation and activation. On the basis of this result, a model of INCENP-mediated AurB phosphorylation is proposed which is consistent with the previously proposed two-step activation mechanism. Secondly, AurB phosphorylation of substrate occurs through a random Bi Bi kinetic mechanism. Finally, INCENP has relatively minor effects on the specific activity of AurB towards a peptide substrate when compared with its role in AurB autoactivation. These results indicate that the effects of INCENP, and probably kinase accessory proteins in general, may differ when acting on different downstream targets. This may reflect the diverse functions and the intricate regulatory mechanisms of protein kinases in vivo.
We thank Ms Amy Calamari and Dr Robert Kirkpatrick (both at GlaxoSmithKline) for reagents. We are also grateful to Mr Octerloney McDonald, and Dr Mary Ann Hardwicke, Dr Denis Patrick and Dr Jerry Adams for helpful discussions. All authors are current or former employees of GlaxoSmithKline.
Aurora A kinase
Aurora B kinase
Aurora C kinase
inner centromere protein
Multiple Reaction Monitoring
target protein for Xenopus kinesin-like protein 2
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.