Novel Small Molecule Inhibitors of Choline Kinase Identified by Fragment-Based Drug Discovery.

Choline kinase α (ChoKα) is an enzyme involved in the synthesis of phospholipids and thereby plays key roles in regulation of cell proliferation, oncogenic transformation, and human carcinogenesis. Since several inhibitors of ChoKα display antiproliferative activity in both cellular and animal models, this novel oncogene has recently gained interest as a promising small molecule target for cancer therapy. Here we summarize our efforts to further validate ChoKα as an oncogenic target and explore the activity of novel small molecule inhibitors of ChoKα. Starting from weakly binding fragments, we describe a structure based lead discovery approach, which resulted in novel highly potent inhibitors of ChoKα. In cancer cell lines, our lead compounds exhibit a dose-dependent decrease of phosphocholine, inhibition of cell growth, and induction of apoptosis at low micromolar concentrations. The druglike lead series presented here is optimizable for improvements in cellular potency, drug target residence time, and pharmacokinetic parameters. These inhibitors may be utilized not only to further validate ChoKα as antioncogenic target but also as novel chemical matter that may lead to antitumor agents that specifically interfere with cancer cell metabolism.

Preclinical study of treatment response in HCT-116 cells and xenografts with (1) H-decoupled (31) P MRS.

The topoisomerase I inhibitor, irinotecan, and its active metabolite SN-38 have been shown to induce G(2) /M cell cycle arrest without significant cell death in human colon carcinoma cells (HCT-116). Subsequent treatment of these G(2) /M-arrested cells with the cyclin-dependent kinase inhibitor, flavopiridol, induced these cells to undergo apoptosis. The goal of this study was to develop a noninvasive metabolic biomarker for early tumor response and target inhibition of irinotecan followed by flavopiridol treatment in a longitudinal study. A total of eleven mice bearing HCT-116 xenografts were separated into two cohorts where one cohort was administered saline and the other treated with a sequential course of irinotecan followed by flavopiridol. Each mouse xenograft was longitudinally monitored with proton ((1) H)-decoupled phosphorus ((31) P) magnetic resonance spectroscopy (MRS) before and after treatment. A statistically significant decrease in phosphocholine (p = 0.0004) and inorganic phosphate (p = 0.0103) levels were observed in HCT-116 xenografts following treatment, which were evidenced within twenty-four hours of treatment completion. Also, a significant growth delay was found in treated xenografts. To discern the underlying mechanism for the treatment response of the xenografts, in vitro HCT-116 cell cultures were investigated with enzymatic assays, cell cycle analysis, and apoptotic assays. Flavopiridol had a direct effect on choline kinase as measured by a 67% reduction in the phosphorylation of choline to phosphocholine. Cells treated with SN-38 alone underwent 83 ± 5% G(2) /M cell cycle arrest compared to untreated cells. In cells, flavopiridol alone induced 5 ± 1% apoptosis while the sequential treatment (SN-38 then flavopiridol) resulted in 39 ± 10% apoptosis. In vivo (1) H-decoupled (31) P MRS indirectly measures choline kinase activity. The decrease in phosphocholine may be a potential indicator of early tumor response to the sequential treatment of irinotecan followed by flavopiridol in noninvasive and/or longitudinal studies.

Multiple isoforms of choline kinase from Caenorhabditis elegans: cloning, expression, purification, and characterization.

Choline kinase is the first enzymatic step in the CDP-choline pathway for phosphatidylcholine biosynthesis. The genome of the nematode, Caenorhabditis elegans, contains seven genes that appear likely to encode choline and/or ethanolamine kinases. We cloned five and expressed four of these genes, and purified or partially purified three of the encoded enzymes. All expressed proteins had choline kinase activity; those that most closely resemble the mammalian choline kinases were the most active. CKA-2, a very active form, was purified to near homogeneity. The K(m) values for CKA-2 were 1.6 and 2.4 mM for choline and ATP, respectively, and k(cat) was 74 s(-1). CKA-2 was predominantly a homodimer as assessed by glycerol gradient sedimentation and dynamic light scattering. CKB-2, which was less similar to mammalian choline kinases, had K(m) values for choline and ATP of 13 and 0.7 mM, and k(cat) was 3.8 s(-1). Both of these highly purified enzymes required magnesium, had very alkaline pH optima, and were much more active with choline as substrate than with ethanolamine. These results provide a foundation for future studies on the structure and function of choline kinases, as well as studies on the genetic analysis of the function of the multiple isoforms in this organism.

Pea choline kinase: purification, properties and isolation of a cDNA.

Choline kinase has been partially purified from pea seedlings and its properties studied. Using sequence information from soya bean and other choline kinases, we have also isolated a cDNA encoding the enzyme. It encodes a protein of 343 amino acids (calculated molecular mass of 39785 Da), which shows 82% homology with the soya bean choline kinase. The protein has been expressed in Escherichia coli with very good activity and high expression levels.

Expression, purification, and characterization of choline kinase, product of the CKI gene from Saccharomyces cerevisiae.

In the yeast Saccharomyces cerevisiae, choline kinase (ATP:choline phosphotransferase, EC 2.7.1.32) is the product of the CKI gene. Choline kinase catalyzes the committed step in the synthesis of phosphatidylcholine by the CDP-choline pathway. The yeast enzyme was overexpressed 106-fold in Sf-9 insect cells and purified 71.2-fold to homogeneity from the cytosolic fraction by chromatography with concanavalin A, Affi-Gel Blue, and Mono Q. The N-terminal amino acid sequence of purified choline kinase matched perfectly with the deduced sequence of the CKI gene. The minimum subunit molecular mass (73 kDa) of purified choline kinase was in good agreement with the predicted size (66.3 kDa) of the CKI gene product. Native choline kinase existed in oligomeric structures of dimers, tetramers, and octomers. The amounts of the tetrameric and octomeric forms increased in the presence of the substrate ATP. Antibodies were raised against the purified enzyme and were used to identify choline kinase in insect cells and in S. cerevisiae. Maximum choline kinase activity was dependent on Mg2+ ions (10 mM) at pH 9.5 and at 30 degrees C. The equilibrium constant (0.2) for the reaction indicated that the reverse reaction was favored in vitro. The activation energy for the reaction was 6.26 kcal/mol, and the enzyme was labile above 30 degrees C. Choline kinase exhibited saturation kinetics with respect to choline and positive cooperative kinetics with respect to ATP (n = 1.4-2.3). Results of the kinetic experiments indicated that the enzyme catalyzes a sequential Bi Bi reaction. The Vmax for the reaction was 138.7 micromol/min/mg, and the Km values for choline and ATP were 0.27 mM and 90 microM, respectively. The turnover number per choline kinase subunit was 153 s-1. Ethanolamine was a poor substrate for the purified choline kinase, and it was also poor inhibitor of choline kinase activity. ADP inhibited choline kinase activity (IC50 = 0.32 mM) in a positive cooperative manner (n = 1.5), and the mechanism of inhibition with respect to ATP and choline was complex. The regulation of choline kinase activity by ATP and ADP may be physiologically relevant.

Partial purification of two forms of choline kinase and separation of choline kinase from sphingosine kinase of rat brain.

Choline kinase of rat brain was purified approximately 200,000 fold using acid precipitation, ammonium sulphate fractionation, Q-Sepharose, Octyl-Sepharose and AH-Sepharose chromatography. The ability of this enzyme to catalyze the phosphorylation of choline, ethanolamine (Etn), monomethylethanolamine (MeEtn), dimethylethanolamine (Me2Etn) and sphingosine was investigated. Choline kinase was separated from sphingosine kinase. The fraction with highly purified choline kinase had four major polypeptides with different molecular masses and possessed activities towards choline, Etn, MeEtn and Me2Etn. Two forms of choline kinase were obtained when the enzymatically active fractions eluted from the Q-Sepharose column were subjected to a horizontal isoelectrofocusing electrophoresis. One form focused around pH 4.7 and is able to phosphorylate choline, Etn, MeEtn and Me2Etn. The other form focused around pH 10 and possessed only choline kinase activity. The latter form of choline kinase did not display classical Michaelis-Menten's mechanism but revealed a positive co-operative pattern for two choline binding sites. This form was purified to apparent homogeneity with a approximate molecular mass of 14.4 kDa.

Immunologically and enzymatically distinct rat choline kinase isozymes.

The purification and characterization of choline kinase R1 expressed in Escherichia coli, using expression plasmid pKCK(H) constructed with rat liver choline kinase cDNA [Uchida, T. and Yamashita, S. (1992) J. Biol. Chem. 267, 10156-10162], were carried out. Choline kinase R1 was chromatographically, enzymatically, and immunologically distinct from rat brain choline kinase [Uchida, T. and Yamashita, S. (1990) Biochim. Biophys. Acta 1043, 281-288]. This report, for the first time, definitively demonstrates the presence of two types of choline kinase isozymes. Rat choline kinase was immunologically classified into choline kinases P and R. Choline kinase P was purified from rat brain as described previously. Choline kinase R was designated as the choline kinase generated from the choline kinase R gene for choline kinase R1. Choline kinases from cytosol were chromatographically separated into two fractions, that is, one passed through a Blue-Toyopearl column and the other was retained on it. The choline kinase retained on the Blue-Toyopearl column was highly purified from rat liver cytosol and characterized in this study, being found to comprise isoforms of choline kinase R, and a choline kinase which was immunologically similar to choline kinase P. Choline kinases P and R are co-distributed in rat tissues. The carcinogen, 3-methylcholanthrene, and hepatotoxic carbon tetrachloride increased the enzyme activity, and the transcript and protein levels of choline kinase R in rat liver. The relation of choline kinase isozymes to previously purified enzymes was discussed.

Molecular cloning, characterization, and expression in Escherichia coli of a cDNA encoding mammalian choline kinase.

A rat liver lambda gt11 cDNA library was screened with antibody against rat liver choline kinase. One of the putative phage clones obtained was found to express choline kinase activity in infected Escherichia coli cells and used for the isolation of fully extended cDNA clones by hybridization. The obtained clones had identical overlapping nucleotide sequences and gave a composite cDNA sequence of 2540 bases. Poly(A)+ tails were found at three different sites of the cDNA sequence. Within the sequence there was a large open reading frame encoding 435 amino acids with a molecular mass of 49,743 Da. When this open reading frame was placed after the trc promoter and introduced into E. coli cells, fully active choline kinase was produced. This enzyme was partially purified using a published choline kinase purification method and was found to mediate the phosphorylation of choline, N,N-dimethylethanolamine,N-monomethylethanolamine, and ethanolamine. Similar to the case with choline kinases from various sources, Km decreased as the number of N-methyl groups increased. The deduced amino acid sequence significantly resembled the yeast choline kinase sequence. It also showed local sequence similarity to protein kinases and some bacterial phosphotransferases. Northern blot analysis revealed that the cDNA probe hybridized to 2.8- and 1.8-kilobase RNA in various rat tissues. The sizes of the transcripts were fairly consistent with that of cDNA. The order of mRNA abundance was testis, brain, lung, kidney, and liver, but did not coincide with the order of the activity levels in these tissues. These results show that the cloned cDNA encodes one of the choline kinase isoforms present in mammalian tissues.

Repression of choline kinase by inositol and choline in Saccharomyces cerevisiae.

The regulation of choline kinase (EC 2.7.1.32), the initial enzyme in the CDP-choline pathway, was examined in Saccharomyces cerevisiae. The addition of myo-inositol to a culture of wild-type cells resulted in a significant decrease in choline kinase activity. Additional supplementation of choline caused a further reduction in the activity. The coding frame of the choline kinase gene, CK1, was joined to the carboxyl terminus of lacZ and expressed in Escherichia coli as a fusion protein, which was then used to prepare an anti-choline kinase antibody. Upon Western (immuno-) and Northern (RNA) blot analyses using the antibody and a CK1 probe, respectively, the decrease in the enzyme activity was found to be correlated with decreases in the enzyme amount and mRNA abundance. The molecular mass of the enzyme was estimated to be 66 kilodaltons, in agreement with the value predicted previously from the nucleotide sequence of the gene. The coding region of CK1 was replaced with that of lacZ, and CK1 expression was measured by assaying beta-galactosidase. The expression of beta-galactosidase from this fusion was repressed by myo-inositol and choline and derepressed in a time-dependent manner upon their removal. The present findings indicate that yeast choline kinase is regulated by myo-inositol and choline at the level of mRNA abundance.

Purification and properties of choline kinase from rat brain.

A blue-dye column separated rat brain choline kinase (EC 2.7.1.32) into two peaks, very likely corresponding to distinct isozymes. The major-peak enzyme was purified 15,000-fold to homogeneity. The final specific activity was approx. 40 mumol.min-1.mg-1. This is 10-times higher than that reported for the enzymes from lung and kidney. The purified enzyme gave a single 44 kDa protein band on sodium dodecyl sulfate polyacrylamide gel electrophoresis. Analytical gel-filtration showed that the native enzyme had a molecular weight of 90,000 and a Stokes radius of 4.2 nm. The sedimentation coefficient was deduced to be 4.8 S and the molecular weight 87,600 by sucrose-density-gradient centrifugation. Hence, the native enzyme appears to be a dimer. The apparent Km values for ATP and choline were 1.0 mM and 14 microM, respectively. At high choline concentrations, the enzyme showed deviation from Michaelis-Menten kinetics. The enzyme was active in a high pH range and utilized a variety of amino alcohols structurally related to choline, including ethanolamine, N-methylethanolamine and N,N-dimethylethanolamine as substrates. Spermine and spermidine stimulated the enzyme by decreasing the apparent Km for ATP and increasing Vmax. Although less efficiently, monovalent cations such as NH4+, K+, Li+ and Na+ and quaternary amines such as carpronium, chlorocholine and acetylcholine were also stimulatory.

Purification and characterization of choline/ethanolamine kinase from rat liver.

Choline kinase, the first enzyme in the CDP-choline pathway for phosphatidylcholine biosynthesis, was purified 26,000-fold from rat liver to a specific activity of 143,000 nmol.min-1.mg-1 protein. The subunit molecular mass was 47 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, while the apparent native molecular mass was 160 kDa by size exclusion chromatography, suggesting a tetrameric structure. Two peaks of choline kinase activity were obtained by chromatofocusing. These isoforms eluted at pH 4.7 (CKI) and 4.5 (CKII). CKII appeared to be homogeneous by sodium dodecyl sulfate gel electrophoresis. Peptide mapping of two isoforms indicated a high degree of similarity, although there were peptides not common to both. Ethanolamine kinase activity copurified with both isoforms. The ratio of choline to ethanolamine kinase activity was 3.7 +/- 0.7 throughout the purification procedure. Choline and ethanolamine were mutually competitive inhibitors. The respective Km values, 0.013 and 1.2 mM, were similar to the Ki values of 0.014 and 2.2 mM. An antibody raised against CKII immunoprecipitated both choline and ethanolamine kinase activities from liver cytosol at the same titer. These data suggest that both activities reside on the same protein and occur at the same active site. Similarly, both activities were immunoprecipitated from brain, lung, and kidney cytosols. Western blot analysis showed both purified liver isoforms, as well as brain, lung and kidney enzymes, to have a molecular mass of 47 kDa.

Complete co-purification of choline kinase and ethanolamine kinase from rat kidney and immunological evidence for both kinase activities residing on the same enzyme protein(s) in rat tissues.

Choline kinase and ethanolamine kinase were completely co-purified from rat kidney cytosol through acid treatment, ammonium sulfate fractionation, DEAE-cellulose column chromatography, Sephadex G-150 gel filtration followed by choline-Sepharose affinity chromatography. The final preparation appeared to be highly homogeneous with respect to both native and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Ishidate, K., Nakagomi, K. and Nakazawa, Y. (1984) J. Biol. Chem. 259, 14706-14710). Throughout the purification steps, the ratio of ethanolamine kinase activity to choline kinase activity was almost constant in a range of 0.3-0.4, which strongly indicated that, in rat kidney, both activities could reside on a single enzyme protein. The rabbit polyclonal antibody raised against highly purified rat kidney choline (ethanolamine) kinase protein inhibited both choline and ethanolamine kinase activities in a parallel manner in crude enzyme preparations not only from rat kidney, but also from rat liver, lung and intestinal cytosols. The results, together with our previous findings, suggested strongly that, in rat tissues, at least large portions of both kinase activities are present on the same enzyme protein(s). The kinetic properties of both kinase reactions with the highly purified kidney enzyme were compared in some detail and the overall result suggested that choline kinase and ethanolamine kinase activities may not have a common active site on a single enzyme protein.

Evidence for the existence of isozymes of choline kinase and their selective induction in 3-methylcholanthrene- or carbon tetrachloride-treated rat liver.

New evidence is provided that rat liver choline kinase exists in several distinct forms (choline kinases I, II and III) which differ in isoelectric point, molecular size and antigenicity against anti-rat kidney choline kinase IgG. Remarkable and selective induction of the choline kinase II and choline kinase III forms of choline kinase was caused similarly by the administration of polycyclic aromatic hydrocarbon carcinogen, 3-methylcholanthrene or hepatotoxic carbon tetrachloride (CCl4). The immunochemical approach further indicated that the elevation in the activity of choline kinase in the 3-methylcholanthrene- or CCl4-treated rat liver was not accompanied by a parallel increase in the amount of choline kinase II enzyme protein, compatible with the induction of either a small amount of new enzyme protein(s) with very high specific activity or another enzyme which might catalyze post-translational modification of choline kinase.

Evidence for the existence of multiple forms of choline (ethanolamine) kinase in rat tissues.

In order to characterize the form of choline kinase in rat tissues, both electrophoretic and gel chromatographic patterns of choline kinase activity were compared in the liver, kidney, lung, whole intestine and carbon tetrachloride-induced liver cytosols. Kinetic parameters of the reaction were also compared for the main forms of choline kinase protein from these tissues. The overall results suggested strongly that choline kinase does not exist in one particular active form but exists in multiple forms in rat tissues. In the study present here, the electrophoretic patterns of both choline kinase and ethanolamine kinase activities were compared in rat liver, kidney, lung and intestinal cytosols. The results strongly supported the view that both kinase activities are represented on the same enzyme protein(s) in each of the rat tissues examined.

Complete purification of choline kinase from rat kidney and preparation of rabbit antibody against rat kidney choline kinase.

Choline kinase was purified from rat kidney to apparent homogeneity with respect to both native and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified enzyme showed a minimum molecular weight of 42,000 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. On the other hand, the molecular size of 75,000-80,000 was estimated through Sephadex G-150 gel filtration, indicating that the enzyme in rat kidney exists most likely in a dimeric form. Specific antibody was raised in rabbit against the highly purified rat kidney choline kinase protein, then immunochemical cross-reactivity was investigated between rabbit antiserum and choline kinase preparations from various rat tissues. The antiserum inhibited choline kinase activity almost completely in the crude preparation not only from kidney but also from lung, intestine, and normal untreated liver cytosol, but it could inhibit only partially the activity from either 3-methylcholanthrene- or carbon tetrachloride-induced rat liver cytosol. The overall results demonstrated that, although choline kinase protein appears to exist in multiple forms in rat tissues, most of them are immunochemically identical, and that either 3-methylcholanthrene- or carbon tetrachloride-inducible form(s) of choline kinase in rat liver could be quite different from a form or forms existing in normal untreated rat liver cytosol.