HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 28, 25493-25501, July 12, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/28/25493    most recent M202651200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Knockaert, M. Articles by Meijer, L. Search for Related Content PubMed PubMed Citation Articles by Knockaert, M. Articles by Meijer, L. Social Bookmarking                      What's this?

Intracellular Targets of Paullones

IDENTIFICATION FOLLOWING AFFINITY PURIFICATION ON IMMOBILIZED INHIBITOR^*

Marie Knockaert^§, Karen Wieking^¶, Sophie Schmitt, Maryse Leost, Karen M. Grant, Jeremy C. Mottram, Conrad Kunick^¶, and Laurent Meijer^**

From the  Station Biologique de Roscoff, CNRS, BP 74, 29682 Roscoff Cedex, Bretagne, France, ^¶ Institut für Pharmazie, Universität Hamburg, Bundesstrasse 45, D-20146 Hamburg, Germany, and  Wellcome Center for Molecular Parasitology, The Anderson College, University of Glasgow, 56 Dumbarton Road, Glasgow G11 6NU, Scotland, United Kingdom

Received for publication, March 19, 2002, and in revised form, April 5, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Numerous inhibitors of cyclin-dependent kinases and glycogen synthase kinase-3 (GSK-3) are being developed in view of their potential applications against cancers and neurodegenerative disorders. Among these, paullones constitute a family of potent and apparently selective cyclin-dependent kinase and GSK-3 inhibitors. However, their actual intracellular targets remain to be identified. To address this issue we have immobilized a paullone, gwennpaullone, on an agarose matrix. Extracts from various cell types and tissues were screened for proteins interacting with this matrix. This approach validated GSK-3 and GSK-3 as major intracellular paullone targets and also mitochondrial, but not cytoplasmic, malate dehydrogenase (MDH). Mitochondrial MDH was indeed inhibited by micromolar concentrations of paullones. Mitochondrial MDH was the major paullone-binding protein in the parasitic protozoon Leishmania mexicana, and paullones inhibited growth of the parasite. This simple batchwise affinity chromatography approach constitutes a straightforward method for the identification of intracellular targets of this particular class of novel anti-mitotic compounds. It has revealed an unexpected target, mitochondrial MDH, the inhibition of which may participate in the pharmacological effects of paullones.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The search for selective inhibitors of protein kinases has intensified over the last few years because of their involvement in essentially all major physiological pathways and because of their implications in many human diseases. These efforts have led to the identification of several families of potent and rather selective kinase inhibitors (reviewed in Refs. 1-4). In our laboratory we have focused on CDKs¹ and GSK-3 as screening targets. Identified CDK inhibitors include olomoucine (5), roscovitine (6), purvalanol (7), hymenialdisine (8), indirubins (9, 10), and paullones (11-13) (reviewed in Refs. 14-20). GSK-3 inhibitors include indirubins (10), paullones (13), maleimides (21, 22), and lithium (23).

The selectivity of these inhibitors constitutes a major problem. This question is usually approached by testing the compounds on a panel of purified, usually recombinant, kinases. This approach is rather unsatisfactory for several reasons. First, it requires the expression, purification, and assay of each individual kinase, a very tedious process. Second, only a limited number (from 5 to 40) of kinases can be reasonably tested. This represents a very minor subset of the estimated total number of kinases in the human genome (>800). Third, other potential, non-kinase targets are not evaluated. For these reasons another method for investigating the selectivity of a given inhibitor has been designed (24). It is based on the purification of potential targets by affinity chromatography. Basically, the inhibitor is immobilized on an agarose matrix through a linker. The choice of the inhibitor orientation with respect to the matrix is derived from the co-crystal structure of the inhibitor with CDK2, which shows the area of the inhibitor that is orientated toward the outside of the enzyme, the area where the linker should be attached. Extracts are prepared from tissues or cells and incubated with the affinity beads. The matrix is then washed, and the bound proteins are analyzed by SDS-PAGE and identified by microsequencing. This approach has allowed the discovery of several unexpected targets of purvalanol (24) and flavopiridol (25-27) (see "Discussion").

Paullones have been identified as CDK1/CDK2/CDK5 inhibitors using the COMPARE analysis of a data base of compounds tested in the NCI, National Institutes of Health, in vitro cancer cell line screening (12). A structure/activity relationship study has led to the more active compounds kenpaullone and alsterpaullone (11). During a classical selectivity study, we found that paullones were also excellent GSK-3 inhibitors (13). Kenpaullone and alsterpaullone are about 10-fold more potent at inhibiting GSK-3 compared with CDK1 (13). Nevertheless, the true selectivity of paullones needs to be defined. We describe here the synthesis of two paullones with a linker and their immobilization on a matrix. The immobilized gwennpaullone allowed the purification of GSK-3 and GSK-3 from various tissues. Unexpectedly, mitochondrial malate dehydrogenase (MDH) was identified as another major target of paullones.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Chemicals and Antibodies

Sodium orthovanadate, EGTA, EDTA, MOPS, -glycerophosphate, DTT, sodium fluoride, p-nitrophenyl phosphate, leupeptin, aprotinin, soybean trypsin inhibitor, benzamidine, formaldehyde, -mercaptoethanol, HEPES, Na₂CO₃, phenyl phosphate, Amido Black, Tween 20, CNBr-activated Sepharose 4B, BSA, Trizma base, histone H1 (type III-S), porcine heart cytoplasmic (M7383) and mitochondrial MDH (M410-13; M2634), malic acid, oxaloacetic acid, -NAD, and -NADH were obtained from Sigma. Nonidet P-40 was from Fluka. Electrophoresis reagents, acrylamide/bisacrylamide (30/0.8), protein assay reagent and horseradish peroxidase-coupled anti-mouse antibodies were purchased from Bio-Rad. Reactigel® 6× (1,1'-carbonyldiimidazole-activated 6% cross-linked beaded agarose) was obtained from Pierce. The GS-1 peptide (YRRAAVPPSPSLSRHSSPHQSpEDEEE), a specific GSK-3 substrate, was synthesized by the Peptide Synthesis Unit, Institute of Biomolecular Sciences, University of Southampton, UK. Hymenialdisine, indirubin-3'-monoxime, flavopiridol, and aminopurvalanol were gifts from Dr. G. R. Pettit (Cancer Research Institute, Arizona State University, Tempe, AZ), Dr. G. Eisenbrand (Department of Chemistry, Food Chemistry, and Environmental Toxicology, University of Kaiserslautern, Kaiserslautern, Germany), Dr. D. Zaharevitz (NCI, Developmental Therapeutics Program, National Institutes of Health, Bethesda), and Dr. N. Gray (Genomics Institute of the Novartis Research Foundation, San Diego), respectively. Staurosporine, roscovitine, and olomoucine were purchased from Calbiochem.

Anti-sea urchin egg CDK5 antibodies were a kind gift from Dr. A. Picard (Laboratoire Arago, Station Marine, Banyuls, France). Polyclonal rabbit anti-mammalian CDK5 (sc-173) was obtained from Santa Cruz Biotechnology. Polyclonal sheep anti-MDH antibodies (porcine mitochondrial MDH) were purchased from Biogenesis. A mouse monoclonal anti-GSK-3/ antibody (KAM-ST002C) was obtained from StressGen Biotechnologies Corp.

Synthesis and Immobilization of Paullones

General procedures and intermediary steps to the synthesis of paullones with a linker are being published elsewhere (28). Other paullones were synthesized as described previously (11). The synthesis and characterization of gwennpaullone (2-(4-aminobutoxy)-9-bromo-7,12-dihydroindolo[3,2-d](1)benzazepin-6(5H)-one hydrochloride or C-2-paullone) is outlined below.

Preparation is performed according to general procedure 4 (28) from 9-bromo-2-(4-phthalimidobutoxy)-7,12-dihydroindolo[3,2-d](1)benzazepin-6(5H)-one (544 mg, 1 mmol) yields a beige powder (258 mg, 54%), m.p. 169 °C (dec.); IR (KBr), 1620 cm¹ (C=O); ¹H NMR ([D₆]-Me₂SO, 400 MHz),  (ppm) = 1.73-1.82 (m, 4H, CH₂-CH₂), 2.87-2.88 (m, 2H, CH₂), 3.47 (s, 2H, CH₂), 4.07-4.11 (m, 2H, CH₂), 6.70 (dd, 1H, 8.7/2.5 Hz, arom. H), 7.18 (d, 1H, 9.2 Hz, arom. H), 7.28 (dd, 1H, 8.4/1.8 Hz, arom. H), 7.30 (d, 1H, 2.6 Hz, arom. H), 7.41 (d, 1H, 8.6 Hz, arom. H), 7.87 (s, ³H, NH₂·HCl), 7.90 (s, 1H, arom. H), 9.93 (s, 1H, NH), 11.90 (s, 1H, NH); ¹³C NMR ([D₆]-Me₂SO, 100.6 MHz):  (ppm) = 23.8, 25.7, 31.23, 55.9, 67.4 (CH₂), 111.4, 113.4, 115.7, 120.3, 123.7, 124.4 (tert.), 107.1, 111.5, 123.3, 128.1, 129.1, 134.0, 135.9, 154.6, 171.0 (quart.).

(Eq. 1)

Paullones were coupled in 0.2 M NaHCO₃, 0.2 M NaCl, pH8.5, to Reactigel®-agarose beads at a calculated final concentration of 10-50 µmol/ml of resin. They were stored at 4 °C as a 50% (v/v) slurry in bead buffer.

Buffers

Homogenization buffer contained 60 mM -glycerophosphate, 15 mM p-nitrophenyl phosphate, 25 mM MOPS, pH 7.2, 15 mM EGTA, 15 mM MgCl₂, 1 mM DTT, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM phenyl phosphate, 10 µg of leupeptin/ml, 10 µg of aprotinin/ml, 10 µg of soybean trypsin inhibitor/ml, and 100 µM benzamidine. Buffer C was homogenization buffer with 5 mM EGTA and without NaF or protease inhibitors. Bead buffer contained 50 mM Tris, pH 7.4, 5 mM NaF, 250 mM NaCl, 5 mM EDTA, 5 mM EGTA, 0.1% Nonidet P-40, 10 µg of leupeptin/ml, 10 µg of aprotinin/ml, 10 µg of soybean trypsin inhibitor/ml, and 100 µM benzamidine. TBST contained 50 mM Tris, pH 7.4, 150 mM NaCl, 0.1% Tween 20. Coupling buffer contained 200 mM NaHCO₃, pH 8.5, 200 mM NaCl. L.m. medium contained 1× M199 (Invitrogen), 20 mM sodium bicarbonate, 200 mM HEPES, pH 7.4, 0.5 mM adenine, 0.0025% hemin.

Biological Material

Sea Urchin Eggs-- The sea urchins Sphaerechinus granularis were obtained by diving in Northern Brittany. Egg spawnings were induced by injection of 0.5 ml of 0.2 M acetylcholine through the perribuccal membrane. Eggs were washed 3 times with Millipore-filtered natural seawater, pelleted by centrifugation, and frozen in liquid nitrogen as described previously (29). Only unfertilized eggs were used in this study.

Xenopus Eggs-- Unfertilized eggs were obtained from female Xenopus laevis induced to ovulate following injection of 500 units of human chorionic gonadotropin. The eggs were dejellied in 2% cysteine solution (30), immediately frozen in liquid nitrogen, and stored at 80 °C until protein extraction and affinity chromatography.

Rat Tissues-- The liver, brain, heart, muscles, lungs, and spleen of mature rats (Rattus norvegicus) were provided by Dr. P. Loyer (INSERM U49, Unité de Recherches Hépatologiques, Hôpital Pontchaillou, Rennes, France). They were stored at 80 °C until protein extraction and affinity chromatography.

Porcine Brain-- Pork brains were obtained from a local slaughterhouse and either directly homogenized and processed for affinity chromatography or stored at 80 °C prior to use.

The rat pheochromocytoma cell line PC12 was cultured in Dulbecco's modified Eagle's medium supplemented with 10% FCS, 5% horse serum, 4 mM L-glutamine, and gentamycin at 37 °C in an atmosphere of 95% air, 5% CO₂. PC12 cell pellets were lysed in homogenization buffer just prior to affinity chromatography.

Leishmania mexicana-- L. mexicana mexicana (MNYC/BZ/62/M379) promastigotes were cultured at 25 °C in HOMEM medium, as described previously (31), supplemented with 5-10% heat-inactivated fetal calf serum. The parasites were harvested at mid-log phase (5 × 10⁶ to 1 × 10⁷ cells/ml), washed with phosphate-buffered saline, and the cell pellets frozen at 80 °C until required. The cell pellets were extracted with homogenization buffer just prior to affinity chromatography.

Preparation of Extracts

Tissues were weighed, homogenized, and sonicated in homogenization buffer (2 ml per g of material). Homogenates were centrifuged for 10 min at 14,000 × g and 4 °C. The supernatant was recovered, assayed for protein content (Bio-Rad protein assay), and immediately loaded batchwise on the affinity matrices. Cells were directly sonicated in homogenization buffer and processed the same way.

Affinity Chromatography of Interacting Proteins

Just before use, 20 µl of packed gwennpaullone beads were washed with 1 ml of bead buffer and resuspended in 600 µl of this buffer. The cell/tissue extract supernatant (2 mg of total protein) was then added; the tubes were rotated at 4 °C for 30 min. After a brief spin at 10,000 × g and removal of the supernatant, the beads were washed four times with bead buffer before addition of 50 µl of 2× Laemmli sample buffer (32). When larger samples were prepared for microsequencing, up to 30 mg of proteins were incubated with 200-500 µl of packed beads and 25 ml of bead buffer. The beads were processed as described above, and the bound proteins were recovered with 300 µl of 2× Laemmli sample buffer (32).

Immobilized CDK1/2-binding protein p9^CKShs1 (33), immobilized purvalanol (95) (24), and immobilized axin (34) were also used for comparison. Cell extracts (2 mg of total proteins) were incubated with 10 µl of packed beads + 600 µl of bead buffer for 30 min. The beads were processed as described for the gwennpaullone beads.

Electrophoresis and Western Blotting

Proteins bound to the paullone matrix, 95-matrix, axin matrix, or p9^CKShs1-Sepharose beads were recovered with 2× Laemmli sample buffer (32). Following heat denaturation for 3 min, the bound proteins were separated by 10% SDS-PAGE (0.7 mm thick gels) followed by immunoblotting analysis or silver staining. Silver staining was performed according to a "home recipe" (fixative, 250 µl of 37% formaldehyde in 250 ml of 50% methanol; rinsing with MilliQ water containing 35 µM DTT, followed by 0.1% AgNO₃ in MilliQ water (w/v); developer, 12 g of Na₂CO₃ in 400 ml of MilliQ water containing 200 µl of 37% formaldehyde). When samples were prepared for microsequencing, gels (1.5-mm thick) were stained with Amido Black (3 mg/100 ml methanol/acetic acid/water, 25/5/20). For immunoblotting, proteins were transferred to 0.1-µm nitrocellulose filters (Schleicher & Schuell). These were blocked with 5% low fat milk in Tris-buffered saline/Tween 20 (TBST), incubated with anti-GSK-3/ (1/1000; 1 h), anti-sea urchin CDK5 (1/1000; 1 h), anti-mammalian CDK5 (1/500; 1 h) or anti-MDH (1/1000 in 1% BSA in TBST; 1 h) antibodies, and analyzed by ECL (Amersham Biosciences). In the case of MDH Western blotting, membranes were blocked with 3% BSA in TBST.

Identification of Purified Proteins

Proteins were identified either by Western blotting with specific antibodies or by microsequencing of internal peptides. The Amido Black-stained bands were excised from the gel and dried under vacuum. They were sent to the Pasteur Institute, Protein Sequencing Laboratory, Paris, France (Dr. J. D'Alayer and M. Davi). The proteins were digested with endolysine C, and the generated peptides were separated by high pressure liquid chromatography and microsequenced. Sequences were compared with available protein sequences using Blastp (Blast 2.0).

Kinase Assays

Kinases activities were assayed in buffer C, at 30 °C, at a final ATP concentration of 15 µM. Blank values were subtracted and activities calculated as picomoles of phosphate incorporated for a 10-min incubation and expressed in percent of the maximal activity, i.e. without inhibitors. Controls were performed with appropriate dilutions of Me₂SO. IC₅₀ values were estimated from the concentration-response curves.

Native starfish oocyte CDK1/cyclin B, recombinant mammalian CDK5/p25 (constructs kindly provided by Dr. J. H. Wang), and recombinant GSK-3 were expressed, purified, and assayed as described previously (6). Histone H1 and GS-1 peptide were used as substrates, for CDKs and GSK-3, respectively.

Malate Dehydrogenase Assays

MDH assays were run in a Uvikon 930 spectrophotometer equipped with a constant temperature cell holder maintained at 28 °C. Mitochondrial and cytoplasmic MDH activities were assayed either in the forward direction (malate  oxaloacetate) or in the reverse direction (oxaloacetate  malate), by following the increase or decrease in absorption at 340 nm, due to NAD reduction or NADH oxidation, respectively. Assays were run, at 28 °C, in the presence of 100 mM Trizma base, pH 8.5, 500 mM malate, and 16 mM NAD (forward reaction) or 100 mM Trizma base, pH 8.5, 10 mM oxaloacetate, and 16 mM NADH (reverse reaction). The A₃₄₀ change was measured against a blank (same mixture of reagents but no enzyme). Reactions were performed in polymethylmethacrylate-disposable cuvettes. Addition of the substrate was used to start the reaction. The amount of MDH was chosen to provide a linear reaction during the first 2.5 min of the reaction. Initial slope rates were used to calculate MDH activities, which are expressed in  A₃₄₀/min/µl MDH or as percent of the maximal activity (measured in the absence of inhibitor). Me₂SO had no effect on MDH activity at the highest concentration used (5%).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Synthesis, Characterization, and Immobilization of Paullones on a Matrix-- Paullones have not been co-crystallized with CDK2. However, they have been docked into the ATP-binding pocket of a homology model of CDK1/cyclin B developed from the crystal structure of CDK2/cyclin A (35). This has allowed us to select ring A for attachment of a side chain (Fig. 1) as it appears to be orientated toward the outside of the ATP-binding pocket. Two paullones, analogous to kenpaullone but modified with a side chain on C-2 or C-3, were synthesized (Fig. 1). These two paullones, C-3-paullone and C-2-paullone (gwennpaullone), were tested on CDK1/cyclin B, CDK5/p25, and GSK-3 (Table I). This revealed that both paullones have a strong preference for GSK-3 over CDKs. The paullones were then immobilized on agarose beads, as described under "Experimental Procedures."

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Structure and kinase inhibitory properties of the paullones used in this study. GSK-3 and CDKs were assayed using the GS-1 peptide or histone H1 as substrates, respectively, with 15 µM ATP and in the presence of increasing concentrations of gwennpaullone (blue) or C-3-paullone (red). Activity is presented as percent of maximal activity (no inhibitors).

Immobilized Paullone Binds GSK-3 and Mitochondrial Malate Dehydrogenase-- Extracts were prepared from various tissues and cells and batch-loaded on the matrices, as described under "Experimental Procedures." Beads were extensively washed with bead buffer before addition of Laemmli sample buffer (32). The bound proteins were resolved by SDS-PAGE (Figs. 2-9) and either silver-stained, excised for digestion followed by sequencing of internal peptides, or analyzed by Western blotting.

We first tested the two types of beads using a porcine brain extract (Fig. 2). Results clearly show that gwennpaullone provided the best binding of proteins from porcine brain, suggesting that the linker had been positioned in the most appropriate place. We thus used immobilized gwennpaullone for the rest of the work.

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Paullone-binding proteins in porcine brain: GSK-3, GSK-3, and MDH. A porcine brain extract was loaded on control (ethanolamine) beads, immobilized paullones, axin beads, or aminopurvalanol beads. The bound proteins were analyzed by SDS-PAGE followed by silver staining (A) or Western blotting using anti-GSK-3 (B), anti-MDH (C), and anti-CDK5 (D) antibodies. The 38-kDa protein (1) bound to immobilized gwennpaullone was identified as mitochondrial MDH by microsequencing (Table II).

Porcine brain extracts yielded three major gwennpaullone-binding proteins, with molecular masses of 38, 48, and 55 kDa (Fig. 2A). None of the three bands were recovered on control ethanolamine beads (Fig. 2A). The two upper proteins were identified as GSK-3 and GSK-3 from their size, their cross-reactivity with anti-GSK-3 antibodies (Fig. 2B), and their depletion by axin beads (Fig. 3A), a GSK-3 selective affinity binding matrix (34). The third protein was excised from the gel and microsequenced. An internal peptide revealed that this 38-kDa protein (1) is mitochondrial malate dehydrogenase (MDH) (Table II). Indeed it cross-reacted with anti-MDH antibodies (Fig. 2C). When the extract was first depleted of GSK-3 on axin beads prior to loading on gwennpaullone beads, MDH was still recovered on the gwennpaullone matrix (Fig. 3). This demonstrates that MDH directly binds to gwennpaullone and not indirectly through a complex with GSK-3 or GSK-3. No CDK5 was found on gwennpaullone (Fig. 2D), even when the extract had been pre-cleared of GSK-3 on axin beads (Fig. 3). CDK5 was readily detected when the same amount of porcine brain extract was loaded on purvalanol-Sepharose (Fig. 2, A and D). Furthermore, purified recombinant CDK5/p25 or native CDK1/cyclin B, purified from starfish oocytes (33), did not bind to gwennpaullone beads (data not shown).

View larger version (62K): [in this window] [in a new window]   Fig. 3.   Binding of MDH on immobilized gwennpaullone is independent from GSK-3. A porcine brain extract was loaded on immobilized gwennpaullone before and after extensive depletion of GSK-3/ on axin beads. The bound proteins were analyzed by SDS-PAGE followed by silver staining (left panel) or Western blotting using anti-GSK-3 antibodies (right panel).

When extracts were supplemented with increasing amounts of alsterpaullone prior to loading on gwennpaullone beads, a concentration-dependent reduction of bound GSK-3 and GSK-3, but not of MDH, was observed (Fig. 4A). This suggests that alsterpaullone and gwennpaullone interact with GSK-3 at the same site. The lack of effect on MDH is in agreement with the similar efficacy of alsterpaullone and gwennpaullone in inhibiting MDH (Table II, see below). When extracts were supplemented with increasing concentrations of ATP, a concentration-dependent reduction of bound GSK-3 and GSK-3, but not of MDH, was seen (data not shown). This confirms the ATP competitive nature of the paullone/GSK-3 interaction (13). MDH levels remained constant whatever the ATP level, in agreement with the lack of an ATP-binding site on MDH. Finally, when extracts were supplemented with increasing concentrations of NAD, a concentration-dependent reduction of bound MDH, but not of GSK-3 and GSK-3, was observed (Fig. 4B), suggesting that gwennpaullone and NAD directly compete for binding to the same site on MDH (see below).

View larger version (92K): [in this window] [in a new window]   Fig. 4.   Alsterpaullone inhibits GSK-3/ binding to immobilized gwennpaullone (A), whereas NAD inhibits MDH binding (B). Porcine brain extracts were supplemented with increasing concentrations of either alsterpaullone (A) or NAD (B), prior to loading on gwennpaullone beads. The bound proteins were analyzed by SDS-PAGE followed by silver staining.

We next loaded extracts from different rat organs and tissues on gwennpaullone beads (Fig. 5). Analysis of the bound proteins shows a diversity of targets, which are nevertheless dominated by a 38-kDa protein coinciding with MDH (Fig. 5, A and C), and a doublet of proteins which are GSK-3 and GSK-3, as revealed by Western blotting (Fig. 5, A and B). Extracts of the rat pheochromocytoma cell line PC12 were also loaded on gwennpaullone beads (Fig. 6). The bound proteins are dominated by GSK-3 and GSK-3 as shown by Western blotting (Fig. 6, left), and by a 38-kDa protein, identified as mitochondrial MDH by Western blotting. These proteins were absent from control ethanolamine beads (Fig. 6).

View larger version (50K): [in this window] [in a new window]   Fig. 5.   Gwennpaullone-binding proteins in rat tissues and organs. Extracts (2 mg of proteins) were prepared from different rat organs and tissues and loaded on immobilized gwennpaullone. The bound proteins were analyzed by SDS-PAGE followed by silver staining (A) or Western blotting using anti-GSK-3 (B) and anti-MDH (C) antibodies.

View larger version (55K): [in this window] [in a new window]   Fig. 6.   Gwennpaullone-binding proteins in rat PC12 cells. An extract from PC12 cells was loaded on control (ethanolamine) beads or on immobilized gwennpaullone. The bound proteins were analyzed by SDS-PAGE followed by silver staining (left panel), or Western blotting using anti-GSK-3 antibodies (middle panel), and anti-MDH (right panel) antibodies.

X. laevis egg (metaphase II-arrested) extracts were also loaded on gwennpaullone beads. SDS-PAGE (Fig. 7) and microsequencing (Table II) revealed that the major paullone-interacting protein (2) from Xenopus eggs is mitochondrial MDH.

View larger version (38K): [in this window] [in a new window]   Fig. 7.   Gwennpaullone-binding proteins in X. laevis eggs. Egg extracts were loaded on immobilized gwennpaullone. The bound protein (2) were analyzed by SDS-PAGE followed by silver staining and identified as mitochondrial MDH by microsequencing (Table II).

We next analyzed sea urchin eggs, a frequently used cell cycle model. Unfertilized, G₁ phase eggs were obtained by acetylcholine-induced spawning of the sea urchin S. granularis. Extracts were prepared in homogenization buffer and loaded on gwennpaullone-agarose, and only a major, 33-kDa, protein (3) was found to bind specifically to this matrix (Fig. 8). Microsequencing of an internal peptide revealed that it is a mitochondrial MDH (Table II). This protein was absent from control ethanolamine beads loaded with the same amount of sea urchin egg extract. No GSK-3 was detected either on gwennpaullone or axin beads. This kinase may only be expressed later during the sea urchin development. Finally, although CDK1/2 and CDK5 are present in the sea urchin eggs, as detected on purvalanol and p9^CKShs1-Sepharose beads, none of these CDKs was found to bind to the gwennpaullone matrix (Fig. 8). Similarly, no CDK1/cylin B was recovered on gwennpaullone beads loaded with starfish oocyte extracts, a very rich source of this kinase which readily binds to purvalanol (24) or p9^CKShs1-Sepharose (33) beads (data not shown), demonstrating that CDKs do not bind to immobilized gwennpaullone.

View larger version (61K): [in this window] [in a new window]   Fig. 8.   Gwennpaullone-binding proteins in sea urchin eggs. A sea urchin egg extract was loaded on control (ethanolamine) beads, immobilized gwennpaullone, axin beads, aminopurvalanol beads, or p9CKShs1-Sepharose beads. The bound proteins were analyzed by SDS-PAGE followed by silver staining (left panel) or Western blotting using anti-CDK5 antibodies (right panel). The unique 33-kDa protein (3) bound to immobilized gwennpaullone was identified as mitochondrial MDH by microsequencing (Table II).

Finally, we analyzed the paullone-binding proteins of two major unicellular parasites of man, L. mexicana (Fig. 9) and Plasmodium falciparum (data not shown). Two Leishmania proteins were found to bind to gwennpaullone beads (Fig. 9A). The major one (36.5 kDa) (4) was identified by microsequencing as mitochondrial MDH (Table II). This form corresponds to the particulate MDH purified from L. mexicana (36, 37). The other protein has yet to be identified. In contrast, no specific gwennpaullone-binding proteins were recovered from P. falciparum extracts (data not shown). This result is in agreement with the apparent lack of mitochondrial MDH in P. falciparum (38, 39). Plasmodial GSK-3, PfGSK-3, appears to be only weakly sensitive to paullones,² and this may explain its absence from immobilized gwennpaullone (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 9.   Gwennpaullone-binding proteins in L. mexicana (A) and effects of paullones on L. mexicana growth (B). A, protozoan extracts were loaded on control (ethanolamine) beads or immobilized gwennpaullone. The bound proteins were analyzed by SDS-PAGE followed by silver staining. The major 36.5-kDa protein (4) bound to immobilized gwennpaullone was identified as mitochondrial MDH by microsequencing (Table II). B, L. mexicana promastigotes were seeded at a density of 1 × 106 cells ml1 and incubated in the presence of alsterpaullone. Cell density was determined at 24-h time points and the mean values (of triplicate cultures) plotted versus time.

To test the effect of paullones on L. mexicana, promastigotes were seeded at a density of 1 × 10⁶ ml¹ and incubated in the presence of a range of concentrations of alsterpaullone. The drug inhibited growth of Leishmania in a concentration-dependent manner (Fig. 9B). 3 µM resulted in complete growth arrest, whereas lower concentrations only partially inhibited growth. This growth arrest was fully reversible on removal of the drug. 10 µM alsterpaullone killed the parasites after 5 days (not shown).

Effects of Paullones on Malate Dehydrogenase Activity-- MDH appeared to be a major gwennpaullone-binding protein in all tissues studied here. Therefore, we next investigated the effects of paullones on MDH activity. MDH catalyzes the NAD-dependent oxidation of malic acid to oxaloacetate, as well as the reverse reaction, the NADH-dependent reduction of oxaloacetate to malate (reviewed in Refs. 41 and 42). Purified porcine heart mitochondrial and cytoplasmic MDH were obtained from a commercial source and assayed in the presence of increasing concentrations of various paullones (Fig. 10). These experiments showed that mitochondrial MDH, and to a lower extent cytoplasmic MDH, are inhibited by paullones in a concentration-dependent manner (Fig. 10, A and B, and Table I). IC₅₀ values for paullone, kenpaullone, gwennpaullone, and alsterpaullone are in the micromolar range. These values correlate only weakly with the IC₅₀ values of these paullones tested against CDKs and GSK-3 (Table I). Mitochondrial MDH activities were tested in both directions (forward, malate  oxaloacetate, and reverse, oxaloacetate  malate). Reactions in both directions were equally sensitive to inhibition by gwennpaullone (Fig. 10A) and other paullones (data not shown). Kinetic studies demonstrate that paullones act by competing with the binding of NAD/NADH to MDH (Fig. 10C). Finally, we tested the effects on mitochondrial MDH of several CDK/GSK-3 inhibitors (hymenialdisine, indirubin-3'-monoxime, flavopiridol, staurosporine, aminopurvalanol, roscovitine, and olomoucine). At a final concentration of 10 µM, none of these compounds had a major inhibitory effect (data not shown). However, we cannot rule out that other kinase inhibitory scaffolds may cross-react with mMDH or other NAD-dependent enzymes.

View larger version (16K): [in this window] [in a new window]   Fig. 10.   Effects of paullones on MDH. A, gwennpaullone concentration-response curve: mitochondrial (mMDH) and cytoplasmic (cMDH) malate dehydrogenases were both assayed in the forward (malate  oxaloacetate) and reverse (oxaloacetate  malate) directions, in the presence of increasing concentrations of gwennpaullone. Activity is presented as percent of maximal activity (no inhibitor). B, effects of different paullones tested as in A. C, kinetics of mMDH inhibition by gwennpaullone in the presence of variable NAD concentration (see text).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Paullones were initially identified as CDK inhibitors (11, 12). However, during extensive selectivity studies, we found that paullones were also excellent inhibitors of GSK-3 (13). The affinity chromatography approach used here confirmed the strong interaction between paullones and GSK-3/. This is not a surprise because GSK-3/, which are relatively abundant, are most sensitive to paullones. Our results therefore demonstrate that the cellular and physiological effects of paullones can be, at least partially, explained by inhibition of GSK-3/. This was recently illustrated by the fact that alsterpaullone is able to inhibit the in vivo phosphorylation of tau at Alzheimer's disease- and GSK-3/-specific sites (13).

In a previous paper (13) we showed that alsterpaullone inhibits the CDK5-specific phosphorylation of DARPP-32 on Thr-75 in isolated rat striatum slices (43). This demonstrated that paullones are also able to inhibit CDKs in a cellular context. Surprisingly, CDKs were not detected on gwennpaullone beads loaded with CDK1- (starfish oocyte) or CDK5 (porcine brain and sea urchin eggs)-rich extracts, or even with purified CDKs. CDKs are less abundant than GSK-3/ and about 10 times less sensitive to paullones than GSK-3/, but this is unlikely to explain their total lack of binding to immobilized gwennpaullone. The absence of CDK binding to immobilized gwennpaullone is also certainly not due to saturation of the beads by the relatively more abundant GSK-3/ and MDH. Indeed depletion of GSK-3/ (by axin beads) does not lead to CDK5 binding. Furthermore, purified CDK5/p25 or CDK1/cyclin B do not bind to immobilized gwennpaullone, despite the fact that these very same enzymes are inhibited by and thus interact with free gwennpaullone (Table I and Fig. 1). One possibility is that CDKs are unable to bind to immobilized gwennpaullone because of some steric hindrance due to the proximity of the beads that may be too close to the ligand. As an alternative explanation the inhibitor may not be perfectly orientated with respect to the linker (we have no crystal structure of CDK2 with the paullones, and can only "guess" from models). Perhaps the orientation of the paullones in the CDK-binding pocket is subtly different from binding to the GSK-3s, and thus the linker has no effect on GSK-3 binding but prevents CDK binding.

Quite unexpectedly, mitochondrial MDH was found to be a major paullone-interacting enzyme. This was observed in a wide range of tissues and cells, including mammalian tissues and cell lines, sea urchin eggs, Xenopus eggs, and a parasitic protozoon L. mexicana. MDH plays a role in a variety of metabolic pathways including the citric acid (Krebs) cycle (44), tricarboxylic acid cycle, glyoxylate bypass, amino acid synthesis, gluconeogenesis, etc. (reviewed in Refs. 41 and 42). In higher eukaryotes, MDH occurs in two forms, a cytoplasmic form (332 amino acids) and a mitochondrial form (314 amino acids). Both assemble in homodimers of 2 × 35 and 2 × 33 kDa, respectively. Both cytoplasmic MDH (45) and mitochondrial MDH (46, 47) have been crystallized. Despite important divergence of sequence between the two forms (only 20-25% identity), the three-dimensional structures of cytoplasmic and mitochondrial MDH are quite similar (48, 49).

The interaction of gwennpaullone (and other paullones) with mitochondrial MDH appears to be due to a direct competition with NAD/NADH. This raises a number of questions. What is the molecular basis behind the selectivity of paullones for mitochondrial MDH among other NAD-binding enzymes (NAD kinase, lactate dehydrogenase, and many others)? What is the basis for a preference of paullones for mitochondrial versus cytoplasmic MDH? These questions could be addressed by a comparison of their NAD-binding pockets as both enzymes have been crystallized (45-49). To what extent does the ATP-binding pocket of CDKs share some similarity with the NAD-binding domain of mitochondrial MDH? Among the paullones that have been synthesized and tested against CDKs and GSK-3 (13), are there more potent MDH inhibitors? What are the structural determinants for selectivity toward MDH versus CDKs or GSK-3/?

By using an affinity chromatography approach similar to ours, glycogen phosphorylase was recently identified (25, 27) as a major flavopiridol-binding protein. This CDK inhibitor indeed inhibits glycogen phosphorylase by binding to the purine-inhibitory site of the enzyme (26, 27). This unexpected result suggests that the anti-proliferative (reviewed in Ref. 50) and anti-tumor properties (reviewed in Refs. 51 and 52) of flavopiridol may be due to inhibition of intracellular mechanisms other than CDK activity. These include inhibition of glycogen phosphorylase (26, 27), GSK-3/ (10), and cytosolic aldehyde dehydrogenase (25) as well as interaction with multidrug resistance protein 1 (53, 54). Similarly, an affinity chromatography approach recently allowed us to identify Erk2 as a major intracellular target of the CDK inhibitor purvalanol (24). The anti-proliferative properties of purvalanol are clearly related to inhibition of Erk2, in addition to inhibition of CDKs (55). The most important question following the discovery of mMDH as a major paullone target is the link between this target and the cellular effects of paullones. In other words, does inhibition of mMDH contribute, and to what extent, to the anti-mitotic properties of paullones? How does inhibition of the other identified targets (CDK1, CDK2, GSK-3, and GSK-3) also contribute to the phenotype caused by addition of paullones to cells, is there some additive phenomena? This is not a trivial question at all, especially in view of the lack of inhibitors strictly selective for each of these targets. It would be best approached in cells where mMDH really appears to be the major, if not unique, target of paullones and which are highly synchronous. This is the case for sea urchin eggs (Fig. 8) and Xenopus oocytes (Fig. 7). Answering this question is beyond the scope of this work, but it would necessarily need to be addressed were other kinase inhibitors to inhibit mMDH or were paullones to be developed into clinically important drugs. It would then be critical indeed to understand the contribution of mMDH inhibition to the anti-tumor activity as mMDH inhibition could either be an advantage for the curative effects (in such case mMDH may even become a screening target per se), or, in contrast, cause secondary and undesired effects, or reduce the anti-tumor properties (in such case paullones analogues should be developed with reduced mMDH inhibitory properties).

In the case of parasitic protozoa for which new anti-proliferative drugs are desperately needed, our affinity chromatography approach has allowed us to identify mitochondrial MDH as the major target of paullones in L. mexicana. L. mexicana contains three forms of MDH, one cytoplasmic, one in the mitochondrion, and one in an unusual peroxisome-like organelle, the glycosome (36, 37). In Trypanosoma cruzi at least two forms of MDH exist, a tetrameric glycosomal MDH and a dimeric mitochondrial MDH (56). The mitochondrial MDH gene has been cloned and characterized from Trypanosoma brucei (57). The encoded protein has only 55% sequence identity with the glycosomal enzyme (57, 58). Only one MDH sequence is available to date from the incomplete Leishmania major genome project, and this is very closely related to the mitochondrial MDH of trypanosomes providing strong evidence that the peptide sequence derived from the major paullone-binding protein of L. mexicana is a mitochondrial MDH. P. falciparum only contains a cytoplasmic MDH (38, 39), as does its host red blood cell (59, 60). The finding that paullones do not inhibit cytoplasmic MDH efficiently provides an explanation as to why we did not recover MDH on gwennpaullone beads from Plasmodium extracts. Alsterpaullone was found to inhibit growth of L. mexicana promastigotes in culture (Fig. 9A) and to prevent the parasite replicating in macrophages in vitro (not shown). It remains to be established if this growth arrest is due to inhibition of mitochondrial MDH, Cdc2-related protein kinases of which L. mexicana has several (31, 40), or a combination of both. Whatever the in vivo target proves to be, these compounds provide promising leads for anti-leishmanial drug design.

We feel that the affinity chromatography approach, exemplified here with the paullones, is the best approach to identify the targets (sometimes unexpected!) of a given family of compounds with definitive pharmacological interest. Once the targets have been identified, their contribution to the pharmacological properties of the compounds can be evaluated. As a consequence of these studies, the optimization parameters can be modified (such as introduction of the newly identified target enzyme among the screening targets). Finally, target identification through the affinity chromatography approach may be used to anticipate, and possibly to circumvent, potential toxicity problems associated with the compounds.

	ACKNOWLEDGEMENTS

We thank the fishermen of the "Station Biologique de Roscoff" for collecting the sea urchins. We are grateful to Dr. Lénaïck Détivaud for providing the Xenopus eggs; Dr. André Picard for anti-CDK5 antibodies; Dr. Pascal Loyer for rat tissues; Dr. Jerry Wang for the CDK5/p25 constructs; Dr. Daniel Parzy for Plasmodium falciparum pellets; and Dr. George R. Pettit, Dr. Gerhard Eisenbrand, Dr. Daniel Zaharevitz, and Dr. Nathanael Gray for reagents. We are grateful to Dr. Jacques D'Alayer and Marilyne Davi from the Pasteur Institute for expert microsequencing of proteins.

	FOOTNOTES

^* This work was supported in part by "Association pour la Recherche sur le Cancer" Grants ARC 5343 and ARC 5732 (to L. M.) and the INCO-DC Program Contract IC18-CT97-0217 (to L. M.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Supported by a Ph.D. fellowship from the "Ministère de la Recherche." To whom correspondence and reprint requests may be addressed. Tel.: 33-2-98-29-23-39; Fax: 33-2-98-29-23-42; E-mail: knockaer@sb-roscoff.fr.

^** To whom correspondence and reprint requests may be addressed: Station Biologique, 29682 Roscoff Cedex, France. Tel.: 33-2-98-29-23-39; Fax: 33-2-98-29-23-42; E-mail: meijer@sb-roscoff.fr.

Published, JBC Papers in Press, April 18, 2002, DOI 10.1074/jbc.M202651200

² E. Droucheau, A. Primot, D. Mattei, M. Knockaert, P. Alano, A. Jafarshad, B. Baratte, C. Kunick, D. Parzy, C. Doerig, and L. Meijer, submitted for publication.

	ABBREVIATIONS

The abbreviations used are: CDKs, cyclin-dependent kinases; BSA, bovine serum albumin; DTT, dithiothreitol; GSK-3, glycogen synthase kinase-3; MDH, malate dehydrogenase; MOPS, 3-(N-morpholino) propanesulfonic.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES