Application of a Fluorescence Recovery-based Polo-like Kinase 1 Binding Assay to Polo-like Kinase 2 and Polo-like Kinase 3

Summary

Assay systems for evaluating compound protein-binding affinities are essential for developing agonists and/or antagonists. Targeting individual members of a protein family can be extremely important and for this reason it is critical to have methods for evaluating selectivity. We have previously reported a fluorescence recovery assay that employs a fluorescein-labelled probe to determine IC₅₀ values of ATP-competitive type 1 inhibitors of polo-like kinase 1 (Plk1). This probe is based on the potent Plk1 inhibitor BI2536 [fluorescein isothiocyanate (FITC)-polyethylene glycol (PEG)-lysine (Lys) (BI2536) 1]. Herein, we extend this approach to the highly homologous Plk2 and Plk3 members of this kinase family. Our results suggest that this assay system is suitable for evaluating binding affinities against Plk2 and Plk3 as well as Plk1. The new methodology represents the first example of evaluating KD affinities of Plk2 and Plk3. It represents a simple and cost-effective alternative to traditional kinase assays to explore the KD-binding compounds against Plk2 and Plk3 as well as Plk1.

INTRODUCTION

Achieving binding selectivity among members of a protein family can be an important consideration when developing agonists and/or antagonists.^1,2) For this reason, it is essential to have in hand methods for determining selectivity. Recently, the PROteolysis Targeting Chimera (PROTAC) approach has focussed attention on targeting the destruction of proteins that are classified as undruggable or difficult to target for drug development.^3,4) PROTAC molecules typically consist of a ligand that binds to the target protein, a linker and an E3 ligase ligand. Together these bring an E3 ligase into close proximity with a target protein and so promote its subsequent ubiquitination and degradation by the host cell ubiquitin-proteasome system. This provides one example of how ligand affinity to a target protein can be important for drug development.

The serine/threonine-specific polo-like kinase 1 (Plk1) is a key cell cycle regulator that plays crucial roles in mitosis. The location and catalytic activity of this kinase are spatiotemporally restricted at critical points in the cell cycle.^5,6) Over-expression of Plk1 occurs in a variety of cancer cells, including breast cancers, colon cancers, non-small cell lung cancer, and prostate cancers and this can be associated with aggressiveness and poor prognosis.^7–13) Therefore, Plk1 is an attractive target for anti-cancer therapeutic development.^14–16) The Plk family is composed of five members (Plk1–5), with Plk1 and Plk4 being considered as potential targets for anti-cancer drug development.^17,18) In contrast, Plk3 has been reported to be a tumor suppressor,¹⁸⁾ while the function of Plk2 is mixed, such that it both functions a tumor suppressor and simultaneously represents a target for cancer therapy.¹⁹⁾ Finally, Plk5 is an inactive kinase that lacks a kinase domain.²⁰⁾ Given the distinct structures of Plk4 and Plk5 relative to Plk’s 1–3, achieving selectivity among these latter three proteins is considered paramount in developing Plk1 inhibitors.

Plk1 is composed of an N-terminal catalytic kinase domain (KD) and a C-terminal polo-box domain (PBD), which are tethered by a flexible internal domain linker (IDL).^5,21) A primary function of the PBD is to recognize and bind to phosphoserine/phosphothreonine-containing sequences, which leads to spatiotemporal regulation of Plk1 function. The PBD also interacts with its own KD (resulting in “autoinhibition”) through mechanisms that are not clearly understood.^5,22) Accordingly, both the KD and PBD of Plk1 are recognized as molecular targets for drug development. The discovery of bioavailable KD-targeting inhibitors has been proven to be more tractable than PDB-targeting inhibitors.^15,16) A number of KD inhibitors have entered clinical trials and some remain so.¹⁵⁾ Yet there are no clinically used Plk1 kinase inhibitors and continued efforts in this field are warranted. To date, several assay methods have been reported for evaluating kinase inhibitory activities of Plk1. However, these methods typically require expensive assay kits, specific instruments or complex assay procedures to optimize factors such as ATP concentration, concentrations of target proteins and/or substrates, and enzymatic reaction time.^23,24) Given these shortcomings, it is highly desirable to develop simple and rapid assay protocols to explore Plk1 KD inhibitors. Along these lines, we have developed a fluorescence recovery-based binding assay to evaluate KD affinities of test compounds against Plk1. The assay employs a fluorescent probe based on the Plk1 selective type-1 (ATP-competitive) kinase inhibitor, BI2536 [[fluorescein isothiocyanate (FITC)-polyethylene glycol (PEG)-lysine (Lys) (BI2536) 1] (Fig. 1).²⁵⁾ Using this methodology, we have successfully evaluated the binding affinities of a variety of test compounds and we have been able to distinguish type-1 kinase inhibitors²⁶⁾ from other classes of Plk1 inhibitors.

Fig. 1.

Structures of compounds discussed in the text.

At the heart of our method is our finding that the FITC-labeled probe 1, shows fluorescence auto-quenching in the absence of Plk1. Fluorescence intensity increases in dose-dependent manner by the addition of Plk1. We interpret this phenomenon to reflect fluorescence recovery by binding of the probe 1 to Plk1 that results in cancelling of auto-quenching. We used this to successfully develop a type-1-selective Plk1 KD binding assay. More recently, we reported bivalent inhibitors of Plk1 (2) that simultaneously target both the KD and PBD by means of tethering BI2536 with our PBD-binding phosphorylated pentapeptide 3 using PEG linkers (Fig. 1).^27,28) Bivalent inhibitor 2 demonstrated 100-fold enhanced Plk1 PBD affinity and greater kinase selectivity as compared to the parent peptide 3. When we examined BI2536 in a panel of kinase assays, that included the three Plk’s, we found that BI2536 showed similar IC₅₀ values among Plk1 (IC₅₀ = 4.4 nM), Plk2 (IC₅₀ = 3.6 nM), and Plk3 (IC₅₀ = 5.3 nM). These values are somewhat in keeping with previous reports [Plk1 (IC₅₀ = 0.83 nM, 0.78 nM); Plk2 (IC₅₀ = 3.5 nM, 5.9 nM) and Plk3 (IC₅₀ = 9.0 nM, 13.8 nM)].^29,30) Surprisingly, the bivalent inhibitor 2 showed low Plk-selectivity, with IC₅₀ values of 2.0 nM for Plk1, 1.6 nM for Plk2, and 7.2 nM for Plk3.²⁸⁾ Because kinase inhibition values may depend on assay conditions and proteins used, we sought alternate approaches to determine selectivity of the inhibitors among the three Plk’s. As we report herein, using as precedence our previously reported fluorescence recovery-based Plk1 binding assay, we have developed binding assays for the KDs of Plk2 and Plk3.

MATERIALS AND METHODS

Compounds 1–3 were prepared as previously reported.^25,27,28) Plk1 (Human PLK1, His Tag Recombinant Protein, catalog # PR5304B), Plk2 (Human PLK2, GST Tag Recombinant Protein, catalog # PR7592A), and Plk3 (Human PLK3, GST Tag Recombinant Protein, catalog # PR7316B) were purchased from Invitrogen (Waltham, MA, U.S.A.).

Kinase enzymatic assays using recombinant Plk1, Plk2 and Plk3.

The Z’-LYTE^™ Kinase Assay Kit - Ser/Thr 16 Peptide (Invitrogen) was used according to manufacturer’s instructions with the recombinant proteins. The assay conditions were as follows: concentration of the proteins at their EC₅₀ values, 20 μM ATP, 2 μM substrate (or 2 μM phosphorylated substrate as 100% control) in the presence of 1% dimethyl sulfoxide (DMSO) and 1 mM dithiothreitol (DTT). For validation assays, the Dose-response – Stimulation, log(agonist) vs. response – variable slope (four parameters) equation was used to determine the EC₅₀ value of each protein, and for kinase assays, the Dose-response – Inhibition, log(inhibitor) vs. response – variable slope (four parameters) equation was used to determine the IC₅₀ value of each compound in GraphPad Prism 10. IC₅₀ values represent average ± standard error of the mean (S.E.M.) from three independent experiments.

Fluorescence recovery assays using recombinant Plk1, Plk2 and Plk3.

Fluorescence recovery assays were performed as previously described.²⁵⁾ In order to determine K_d values for each protein, recombinant protein was diluted to a 1.33x working dilution in assay buffer [N-(2-hydroxyethyl)-piperazine-N-2-ethanesulfonic acid (HEPES) buffered saline (HBS) containing 0.05% Tween-20 (HBST), 1 mM dithiothreitol (DTT) and 1 mM ethylenediaminetetraacetic acid (EDTA)]. To each well of a 384-well plate was added 30 μL of 1.33x protein solution. Fluorescent probe 1 was diluted to 80 nM (4x) in assay buffer and then 10 μL was added to each well. The plate was allowed to equilibrate at room temperature for 30 min with shaking. The fluorescence intensity was read using a BioTek Synergy 2 plate reader with 485/20 excitation and 528/20 emission. The fluorescence intensity values were obtained in duplicate and background signal intensity (no protein) was subtracted. The average values were plotted versus concentration of each protein and analyzed using non-linear regression in GraphPad Prism 10 [log(agonist) vs response – variable slope (four parameter) model]. K_d values represent concentrations of half-maximum signal intensity shown as average ± S.E.M. from two independent experiments.

For binding assays, recombinant protein was diluted to a 2x working dilution in assay buffer (HBST, 1 mM DTT, and 1 mM EDTA) with the final protein concentration representing the approximate K_d values as determined for the probe 1. Inhibitors were serially diluted to generate 4x working dilutions in assay buffer. To each well of a 384-well plate was added 20 μL of 2x protein solution (0% binding controls received 20 μL of assay buffer). A total of 10 μL of the 4x inhibitor solution (or DMSO blank) was added to corresponding wells and allowed to pre-incubate at room temperature for 30 min with shaking. Fluorescent probe 1 were diluted to 80 nM (4x) in assay buffer and then 10 μL was added to each well. The plate was allowed to equilibrate at room temperature for 30 min with shaking. The fluorescence intensity was read using a BioTek Synergy 2 plate reader with 485/20 excitation and 528/20 emission. The fluorescence intensity values were obtained in triplicate and normalized to 100% (no inhibitor) and 0% binding (no protein) controls. Normalized values were plotted versus concentration and analyzed using non-linear regression in GraphPad Prism 10 [log(inhibitor) vs response – variable slope (four parameter) model]. IC₅₀ values represent average ± standard error of the mean (SEM) from three independent experiments.

Fluorescence polarization (FP) assays using probes based on PBD-selective phosphopeptide sequences and recombinant Plk1, Plk2 and Plk3.

Recombinant protein was diluted to a 1.33x working dilution in assay buffer (HBST, 1 mM DTT, and 1 mM EDTA). To each well of a 384-well plate was added 30 μL of 1.33x protein solution. Fluorescent probe for each protein was diluted to 40 nM (4x) in assay buffer and then 10 μL was added to each well. The following fluorescent dye-labeled sequences were utilized as fluorescent probes: 5-carboxyfluorescein (5CF)-GPMQSpTPLNG-NH₂ for Plk1, 5CF-GPMQTSpTPKNG-NH₂ for Plk2, and 5CF-GPLATSpTPKNG-NH₂ for Plk3.^27,32,33) These probes had been developed for PBD selectivity against Plk1, Plk2 or Plk3 using isolated PDBs of these three kinases.^32,33) The plate was allowed to equilibrate at room temperature for 30 min with shaking. The FP signal was read using a BioTek Synergy 2 plate reader with 485/20 excitation and 528/20 emission. The FP values were obtained in duplicate and background signal intensity (no protein) was subtracted. The average values were plotted versus concentration of each protein and analyzed using non-linear regression in GraphPad Prism 10 [log(agonist) vs response – variable slope (four parameter) model]. K_d values represent concentration of half-maximum signal intensity shown as average ± standard error of the mean (SEM) from two independent experiments.

RESULTS

Evaluation of kinase enzymatic inhibitory potencies against Plk1, Plk2 and Plk3.

We used the Z’-LYTE^™ Kinase Assay Kit - Ser/Thr 16 Peptide (Invitrogen) to evaluate the enzyme catalytic activities of Plk1, Plk2 and Plk3. We determined the inhibitory selectivity of BI2536 and the bivalent inhibitor 2 against these kinases (Table 1, Supplementary Figs. S1, S2). It is known that the Plk1 KD interacts in an intramolecular fashion with its PBD to spatiotemporally regulate activity (auto-inhibition). Recently, we reported that addition of Plk1 PBD-binding peptide 3 to the reaction medium enhances signal intensity in fluorescence recovery assays using full-length Plk1. We interpret this to result from binding of 3 to the PBD, which reduces autoinhibition by the PBD, thereby allowing the probe 1 to more easily access its binding pocket.²⁸⁾ Accordingly, we also conducted kinase assays in the presence of excess quantities of peptide 3.

Table 1.

Results of kinase enzymatic inhibition assays against full-length Plk1, Plk2 and Plk3 using BI2536 and bivalent ligand 2 in the absence or presence of excess peptide 3.

Compound	Plk1	Plk2	Plk3
	IC50 (nM)
BI2536	65 ± 8.0	140 ± 2.0	110 ± 7.1
BI2536 + 10 μM 3	14 ± 0.53	16 ± 0.25	16 ± 0.59
2	110 ± 6.6	120 ± 14	78 ± 20
2 + 100 μM 3	3.2 ± 0.27	14 ± 1.0	3.8 ± 0.89

Evaluation of KD binding affinities against Plk1, Plk2 and Plk3.

Because BI2536 has comparable inhibitory potencies against all three kinases, we applied to Plk2 and Plk3 our in-house fluorescence recovery assay that we had used to evaluate Plk1 KD binding affinity. This assay is based on the binding of the BI2536-derived fluorescent probe 1. Initially, the fluorescence recovery assays were conducted using probe 1 and serially diluted Plk2 and Plk3 proteins to confirm binding as indicated by increases in fluorescence intensity in a protein concentration-dependent manner (Fig. 2). After determining the K_d values for the binding of probe 1 to each protein, fluorescence recovery assays were conducted to evaluate the kinase selectivity of BI2536 and the bivalent inhibitor 2 in the absence or presence of excess peptide 3 (Fig. 3, Table 2). In order to show the robustness of this assay method, the raw data in the absence of inhibitor (100% binding) and background in Fig. 3 were plotted and the signal-to-noise ratio (S/N), signal-to-background ratio (S/B), and Z-factor values for each protein were calculated (Supplementary Fig. S3, Table S1).³⁵⁾

Fig. 2.

Results from fluorescence recovery assays to determine K_d values of full-length Plk1, Plk2 and Plk3 with the probe 1. The X axis represents protein concentration (nM) and the Y axis represents fluorescence intensity (Ex: 485 nm, Em: 528 nm). The binding curves are shown using [Binding - Saturation, Specific binding with Hill slope] regression in GraphPad Prism 10. The K_d value of each protein is calculated as follows: Plk1 (43 ± 4.0 nM), Plk2 (240 ± 70 nM), Plk3 (33 ± 0.59 nM). Exp.: experiment.

Fig. 3.

Fluorescence recovery binding plots used to determine the data shown in Table 2. Plots show the ability of inhibitors to compete with the probe 1 for binding to full-length Plk1, Plk2 and Plk3. The X axis represents inhibitor concentration (log M) and the Y axis represents relative probe binding based on fluorescence intensity (Ex: 485 nm, Em: 528 nm) of no inhibitor (100%) and blank (no protein, 0%). Data points represent average ± SEM from three independent experiments and fit using non-linear regression in GraphPad Prism 10.

Table 2.

Results of fluorescence recovery assays of BI2536 and bivalent ligand 2 against full-length Plk1, Plk2 and Plk3 in the absence or presence of excess peptide 3.

Compound	Plk1	Plk2	Plk3
	IC50 (nM)
BI2536	47 ± 9.3	17 ± 1.8	22 ± 1.5
BI2536 + 100 μM 3	33 ± 1.5	16 ± 1.2	16 ± 0.78
2	11 ± 0.85	7.0 ± 1.9	5.1 ± 0.096
2 + 100 μM 3	15 ± 0.59	15 ± 3.2	9.2 ± 2.2

Determination of probe binding to PBDs of Plk1, Plk2 and Plk3.

In our efforts to develop PBD-binding inhibitors, we have examined the selectivity of PBD-binding ligands against Plk1, Plk2 or Plk3. We employed fluorescence polarization (FP) assays using isolated PBDs of Plk1, Plk2 and Plk3. The following fluorescent dye-labeled peptides were utilized as their probes in these assays: 5CF-GPMQSpTPLNG-NH₂ for Plk1; 5CF-GPMQTSpTPKNG-NH₂ for Plk2 and 5CF-GPLATSpTPKNG-NH₂ for Plk3.^27,32,33) In FP assays with full-length Plk1, we observed that the Plk1 probe (5CF-GPMQSpTPLNG-NH₂) did not show significant FP signal enhancement even at the highest concentration of full-length Plk1. In contrast, when we performed the assay with a fluorescent probe based on peptide 3 (FITC-3, Fig. 1)²⁸⁾ we observed a remarkable protein concentration-dependent response. The PBD-binding affinity of 3 is 160-fold higher than for MAGPMQSpTPLNGAKK peptide which is the parent peptide of 5CF-GPMQSpTPLNG-NH₂, which has greater affinity than for the GPMQSpTPLNG.^27,32,34) This suggested that the affinity of the GPMQSpTPLNG for the Plk1 PBD was not sufficient to overcome autoinhibition by the KD. In contrast, peptide 3 was able to overcome KD autoinhibition and bind to the PBD of full-length Plk1. We hypothesized that similar phenomena could occur in assays against full-length Plk2 and Plk3 when we used probes based on their respective PBD-binding phosphopeptides 5CF-GPMQTSpTPKNG-NH₂ and 5CF-GPLATSpTPKNG-NH₂.

DISCUSSION

Initially, we evaluated the kinase inhibitory selectivity of BI2536 and the bivalent inhibitor 2 against Plk1, Plk2 and Plk3 using a commercially available kinase assay kit (Z’-LYTE^™ Kinase Assay Kit - Ser/Thr 16 Peptide, Invitrogen, Table 1, Fig. S1 and Fig. S2). The compounds showed low selectivity among these kinases (IC₅₀ values of BI2536 for Plk1: 65 nM, Plk2: 140 nM, Plk3: 110 nM, and 2 for Plk1: 110 nM, Plk2: 120 nM, Plk3: 78 nM), which is consistent with our previous findings.²⁸⁾ In the presence of excess peptide 3 the enzyme inhibitory potencies improved for both BI2536 (IC₅₀ values of BI2536 for Plk1: 14 nM, Plk2: 16 nM, Plk3: 16 nM) and for 2 (Plk1: 3.2 nM, Plk2: 14 nM, Plk3: 3.8 nM). This suggests that excess peptide 3 results in high occupancy of the PBD and subsequent disruption of the autoinhibition that would normally be incurred by the intramolecular interaction of the KD and PBD. This could facilitate access of inhibitors to the KD catalytic pocket.

We then applied our Plk1 fluorescence recovery assay to evaluate the kinase domains of Plk2 and Plk3. We determined the K_d values against each kinase with 20 nM of probe 1. We observed that the fluorescence signal intensity of probe 1 increased with protein concentration (Fig. 2). Based on these results we calculated the K_d values against full-length Plk1 (43 nM), Plk2 (240 nM), and Plk3 (33 nM). Using protein concentrations consistent with the determined K_d values, we performed fluorescence recovery assays to determine binding affinities of BI2536 and the bivalent inhibitor 2 for the KDs of Plk1, Plk2 and Plk3 (Fig. 3, Table 2). In contrast to the results of kinase enzyme inhibition assays, BI2536 showed low nanomolar binding affinity to all three proteins both in the absence and presence of excess peptide 3. This suggested that the BI2536-based probe 1 has sufficiently high affinity to overcome PBD autoinhibition and that relief of autoinhibition by 3 has little effect. The bivalent inhibitor 2 showed higher affinity than BI2536 against all three proteins. However, the affinities of 2 for the proteins were slightly decreased in the presence of excess peptide 3. This suggests that binding of peptide 3 inhibited simultaneous binding of 2 to the KD and PBD of Plk1, Plk2 and Plk3, thereby resulting in similar IC₅₀ values of the bivalent BI2536-containing inhibitor 2 with monovalent BI2536. Additionally, as shown in Fig. S3 and Table S1, these fluorescence recovery assays using Plk1, Plk2, and Plk3 demonstrated moderate to excellent Z-factor values suggesting the worth of this assay method for exploring inhibitors of Plk1, Plk2, or Plk3.

Finally, we examined probes employing sequences that are specific for individual PBDs (Fig. 4). Although several assays have been reported to evaluate PBD binding affinities against full-length Plk1,^27,28,32,36) no assays have been reported to evaluate PBD-binding affinities of full-length Plk2 or Plk3. In our current work we examined binding to full-length proteins using the previously reported fluorescent probes that had been used for binding to the isolated PBDs of Plk1, Plk2 and Plk3. We found that none of the probes used in isolated PBD binding assays showed significant FP signal enhancement. This indicated that the binding affinities of the phosphopeptide-based probes may be insufficient to overcome KD-PBD autoinhibition in full-length proteins.

Fig. 4.

Results from FP assays to determine K_d values using full-length Plk1, Plk2 and Plk3 with the indicated probes. The X axis represents protein concentration (nM) and the Y axis represents FP signal intensity (mFP, Ex: 485 nm, Em: 528 nm). The binding curves are shown using [Binding - Saturation, Specific binding with Hill slope] regression in GraphPad Prism 10. The K_d value of Plk1 with FITC-3 is calculated as 13 ± 0.31 nM. Exp.: experiment.

In order to confirm that fluorescence response does not occur with the same concentrations of proteins other than Plk1, Plk2, or Plk3, we also performed the fluorescence recovery assays and FP assays using bovine serum albumin (BSA) with serial dilutions (Supplementary Fig. S4). As expected, BSA did not increase signal intensities in either the fluorescence recovery assays or the FP assays.

CONCLUSION

In our current work we have extended our fluorescence recovery-based Plk1 KD binding assay to full-length Plk2 and Plk3. As we reported previously, this assay system represents a unique method for determining Plk1 KD binding affinities. To the best of our knowledge, no methods have yet been described to determine binding affinities for the KDs of Plk2 or Plk3. The assay system that we now report is the first to evaluate binding affinities for the KDs of Plk2 and Plk3. Although, we did not demonstrate kinase selectivity by these assay systems due to low selectivity of tested compounds, BI2536 and bivalent inhibitor 2, the evaluation of selectivity against KDs of Plk1, Plk2 and Plk3 will be performed using these assay methods in future work. In addition, methods have yet to be reported for determining binding affinities for the PBDs of full-length Plk2 or Plk3. We conducted FP-based PBD-binding assays against full-length Plk2 and Plk3 using fluorescent probes based on phosphopeptides that had previously shown to exhibit binding selectivity for the isolated PBDs of Plk1, Plk2 and Plk3. However, these did not show significant signal enhancement in the presence of the full-length proteins. This suggests that higher affinity probes and/or novel assay systems may be needed to determine PDB-binding affinities of these full-length proteins. These assays could potentially facilitate efforts to develop selective PDB-binding inhibitors. Work is in progress to devise such assays and utilize them for advancing discovery of selective Plk1 inhibitors are ongoing.

Supplementary Material

This article contains supplementary materials.