Nanomolar Inhibitors of Trypanosoma brucei RNA Triphosphatase

ABSTRACT Eukaryal taxa differ with respect to the structure and mechanism of the RNA triphosphatase (RTPase) component of the mRNA capping apparatus. Protozoa, fungi, and certain DNA viruses have a metal-dependent RTPase that belongs to the triphosphate tunnel metalloenzyme (TTM) superfamily. Because the structures, active sites, and chemical mechanisms of the TTM-type RTPases differ from those of mammalian RTPases, the TTM RTPases are potential targets for antiprotozoal, antifungal, and antiviral drug discovery. Here, we employed RNA interference (RNAi) knockdown methods to show that Trypanosoma brucei RTPase Cet1 (TbCet1) is necessary for proliferation of procyclic cells in culture. We then conducted a high-throughput biochemical screen for small-molecule inhibitors of the phosphohydrolase activity of TbCet1. We identified several classes of chemicals—including chlorogenic acids, phenolic glycopyranosides, flavonoids, and other phenolics—that inhibit TbCet1 with nanomolar to low-micromolar 50% inhibitory concentrations (IC50s). We confirmed the activity of these compounds, and tested various analogs thereof, by direct manual assays of TbCet1 phosphohydrolase activity. The most potent nanomolar inhibitors included tetracaffeoylquinic acid, 5-galloylgalloylquinic acid, pentagalloylglucose, rosmarinic acid, and miquelianin. TbCet1 inhibitors were less active (or inactive) against the orthologous TTM-type RTPases of mimivirus, baculovirus, and budding yeast (Saccharomyces cerevisiae). Our results affirm that a TTM RTPase is subject to potent inhibition by small molecules, with the caveat that parallel screens against TTM RTPases from multiple different pathogens may be required to fully probe the chemical space of TTM inhibition.

The fact that the tertiary structures, active sites, and chemical mechanisms of the TTM-type RTPases are completely different from those of mammalian cysteine-phosphatase-type RTPases highlights TTM RTPases as promising targets for antifungal, antiprotozoal, and antiviral drug discovery predicated on interdicting the capping of the pathogen's mRNAs while sparing the host's capping pathway (27). The TTM RTPases Cet1, Pct1, and TriA are essential for the growth of the budding yeast S. cerevisiae, the fission yeast S. pombe, and the human-pathogenic fungus Aspergillus fumigatus, respectively (5,28,29). To our knowledge, it has not been determined whether a TTM RTPase is essential for a protozoon. Here, we employed RNA interference (RNAi) to show that T. brucei RTPase Cet1 (TbCet1) is necessary for proliferation of procyclic cells in culture.
Yeast ceg1⌬ strains that rely for viability on the activity of either fungal, viral, and protozoan TTM RTPases or mammalian cysteine-phosphatase RTPases afford enabling tools for cell-based screens and counterscreens for cytotoxins that inhibit TTM RTPases (11,30,31). In vitro screening for inhibitors of TTM RTPases is simplified by their signature biochemical property of hydrolyzing nucleoside triphosphates (NTPs) to nucleoside diphosphates (NDPs) and inorganic phosphate (P i ) in the presence of manganese (2), thereby avoiding the need to prepare triphosphate-terminated RNAs as the substrates.
In the present study, we conducted a biochemical screen for small-molecule inhibitors of the T. brucei RTPase TbCet1. Kinetoplastid protozoan parasites of the genus Trypanosoma are major zoonotic pathogens of humans. Trypanosoma cruzi is the cause of Chagas disease, endemic in South America. T. brucei, transmitted by tsetse flies, causes sleeping sickness in sub-Saharan Africa. The drugs used to treat trypanosomiasis (pentamidine, suramin, melarsoprol, and eflornithine) either are old, have many undesirable side effects, are not effective against late-stage disease, or are cumbersome to administer (32). There is a need for new therapeutic approaches.
TbCet1 is a 252-amino-acid (aa) monomeric TTM protein that can function in cap formation in vivo in S. cerevisiae in lieu of yeast Cet1 (11,12). Recombinant TbCet1 has vigorous manganesedependent ATPase activity (k cat , 59 s Ϫ1 ) that is effaced by alanine mutations of conserved active site residues in the triphosphate tunnel. TbCet1 is exceptionally thermostable. Tripolyphosphate is a potent competitive inhibitor of TbCet1 (K i , 1.4 M) that binds more avidly to the active site than the ATP substrate (K m , 25 M). Synergistic activation of the TbCet1 triphosphatase by manganese and magnesium suggested a two-metal mechanism of catalysis (12). Recent crystallographic studies of other TTM family members indicate that two-metal catalysis is a core feature of the TTM clade (33).
Here, we implemented a colorimetric assay of phosphate release from ATP by TbCet1 for the purpose of high-throughput screening (HTS) for small-molecule inhibitors. The initial questions were (i) can potent TbCet1 inhibitors be identified, and (ii) if so, do they illuminate common features? The answer to both questions is yes, as described below. Part of the allure of TTM RTPases as anti-infective drug targets is the prospect of obtaining an inhibitor with broad-spectrum potency against fungal, protozoal, and viral RTPases. To put this idea to the test, we surveyed the newly identified inhibitors of TbCet1 for their effects on exemplary fungal and viral TTM RTPases and on a non-RTPase bacterial TTM enzyme (34,35). The results indicated that the inhibitors generally displayed greatest potency and selectivity for the TbCet1 enzyme against which they were screened, signifying that parallel screening against TTM RTPases from pathogens of interest may be required to best explore the chemical space of TTM inhibition.

RESULTS
Inducible RNAi knockdown of TbCet1 arrests the growth of T. brucei in culture. TbCet1 was depleted in T. brucei procyclic cells by using an RNAi system in which the synthesis of doublestranded TbCet1 RNA by T7 RNA polymerase is tetracycline inducible. In the experiment shown in Fig. 1A, the cells were inoculated into medium containing 1.0 g/ml tetracycline (ϩTet) to induce TbCet1 double-stranded RNA (dsRNA) production and into a parallel control culture lacking tetracycline (ϪTet). The control uninduced cells maintained logarithmic growth over a 14-day period. In contrast, the induction of TbCet1 RNAi by Tet acutely curtailed cell proliferation after an initial 5-day interval of logarithmic growth. RNAi depletion of cellular TbCet1 protein was verified by Western blot analysis of total protein from ϩTet T. brucei 29.13-TbCet1RNAi transfectants were inoculated into medium containing 1.0 g/ml tetracycline (ϩTet) to induce TbCet1 dsRNA production and into a parallel control culture lacking tetracycline (ϪTet). Cell density was monitored by microscopy and maintained between 1 ϫ 10 6 and 1 ϫ 10 7 cells/ml by dilution into fresh medium. The growth curves display on the y axis the log of the direct cell count multiplied by the dilution factor. Induction of RNAi against TbCet1 arrested T. brucei growth after 5 days. (B) Western blotting verifies knockdown of TbCet1 protein by RNAi induction. An anti-TbCet1 immunoblot assay of total protein from cells harvested from a ϩTet culture is shown. The positions and sizes (kilodaltons) of marker polypeptides are indicated on the left. The immunoreactive TbCet1 polypeptide, denoted by the arrowhead at left, is depleted after RNAi induction. A nonspecific crossreacting 50-Da polypeptide, indicated by the asterisk at right, is unchanged.
Smith et al. cells (Fig. 1B). The level of TbCet1 protein was serially decreased to 26% of the initial value after 1 day of incubation in tetracycline, to 12% after 2 days, and to Յ8% for the rest of the 14-day period of RNAi induction. These results indicate that TbCet1 is essential for T. brucei growth, and they encourage screening for TbCet1 inhibitors as potential antitrypanosomal drugs.
Test library screening for inhibitors of TbCet1 ATPase. We applied a colorimetric assay using malachite green reagent to gauge the release of inorganic phosphate from ATP by recombinant TbCet1 in the presence of manganese. The Z= score for this assay was 0.88. Z= is a measure of the signal-to-noise ratio intrinsic to an experiment and is calculated by comparing the typical variance in values for positive and negative controls with the dynamic range seen between the two. Values above 0.5 indicate that an assay is well suited for high-throughput screening (36).
The susceptibility of TbCet1 to chemical inhibition was tested via an initial screen of a library of 2,879 structurally diverse compounds comprising pharmaceuticals, natural products, and industrial chemicals. We identified 22 compounds from this collection that reproducibly inhibited the TbCet1 ATPase activity by Ն75% at a 10 M concentration ( Table 1). Nineteen of the 22 inhibitors were phenolic compounds that fell into discrete chemical classes, either flavonoids (11/22), phenolic glucopyranosides (2/22), or nonflavonoid phenolics (6/22). The chemical structures of the flavonoids and selected phenolics are shown in Fig. 2. In a previous study (37), a screen of the same test library for TbCet1 ATPase inhibitors using a fluorescence polarization-based assay (Transcreener) for detection of ADP formation yielded 23 confirmed positives with 50% inhibitory concentrations (IC 50 s) of less than 10 M. Sixteen of the TbCet1 inhibitors identified in the present screen using a phosphate release assay were identified independently by assaying inhibition of ADP release.
HTS for TbCet1 inhibitors. We next applied the colorimetric high-throughput screening (HTS) assay to survey 191,050 compounds sourced from four different commercial libraries, of which 543 (0.28%) inhibited the ATPase activity of TbCet1 by at least 50% at a 10 M concentration. A set of 216 nonredundant compounds were resupplied by the vendors, 83 (38%) of which showed reproducible inhibition of TbCet1 upon retesting and were thus designated confirmed "hit" compounds. The hit compounds were then subjected to IC 50 determination via the HTS assay. Nine compounds had IC 50 s of 1.1 M or less ( Table 2) Each of these nine compounds was then retested for TbCet1 inhibition via a direct assay of the release of 32 P i from [␥-32 P]ATP, in which the reaction products were analyzed by polyethyleneimine (PEI)-cellulose thin-layer chromatography (TLC). Logarithmic plots of inhibition versus compound concentration are shown in Fig. 4 for tetracaffeoylquinic acid (compound 1A; IC 50 , 13 nM) and rosmarinic acid (compound 3A; IC 50 , 120 nM), as well as for three less potent compounds (to be discussed below). The IC 50 s for the radiolabeled ATP hydrolysis assay agreed fairly well with those from the colorimetric phosphate release assay employed for HTS (Table 2). In the radioactive phosphate release assay, the TbCet1, ATP, and Mn 2ϩ concentrations were 2.2 nM, 0.2 mM, and 2 mM, respectively. Thus, the nanomolar to micromolar inhibition observed for the 9 hit compounds cannot be attributable to sequestration of metal cofactor or nucleotide substrate. The direct radiolabel phosphate release assay was employed for all further studies.
Inhibition of TbCet1 by chlorogenic acids and their derivatives. Chlorogenic acids are compounds wherein cinnamic acid or one of its derivatives is esterified to a quinic acid scaffold (Fig. 3). Chlorogenic acids often contain multiple ester substitutions on the quinic acid hydroxyl groups. The two most potent inhibitors of TbCet1 identified by HTS were the chlorogenic acids tetracaffeoylquinic acid (compound 1A) and 3,5-dicaffeoylquinic acid (compound 1B). Caffeic acid is the dihydroxy derivative of cinnamic acid (Fig. 3). To ascertain the inhibitory properties of compounds similar to these, we screened an additional 14 chlorogenic acids and 4 structurally related quinic acid esters for inhibition of TbCet1 (compounds 1C to 1T [ Table 3]). Four of these compounds (1D, 1E, 1G, and 1T) contained caffeic acid substituents. Two compounds (1I and 1Q) contained cinnamic acid. Eight others (1H, 1J, 1L, 1M, 1N, 1P, 1R, and 1S) included three different derivatives of cinnamic acid: coumaric acid, ferulic acid, and p-ferulic acid (chemical structures shown in Fig. 3). Coumaric acid differs from caffeic acid by the removal of the meta hydroxyl moiety. Ferulic acid and p-ferulic acid are O-methyl derivatives of caffeic acid at the meta and para positions, respectively (Fig. 3). Though not chlorogenic acids, the 4 additional quinic acid esters had related phenolic substituents in place of a cinnamic acid de-  Rosmarinic acid (compound 3A) resembles a chlorogenic acid but contains no central scaffold analogous to quinic acid; rather, rosmarinic acid can be considered a partially hydrogenated dimer of caffeic acid. Whereas rosmarinic acid inhibited TbCet1 with an IC 50 of 120 nM, a simple caffeic acid monomer (compound 3B) had an IC 50 of 29 M (Fig. 4). Quinic acid showed no inhibition of TbCet1, nor did gallic acid.
Structure-activity relationships for chlorogenic acids. The importance of both the size of the quinic acid ester substituents and their content of hydroxyl groups is underscored by comparing 5-galloylquinic acid (compound 1O; IC 50 , 125 M) with 5-galloyl-galloylquinic acid (compound 1C; IC 50 , 110 nM), whereby the latter is 1,100-fold more potent than the former as a TbCet1 inhibitor. Correlation between substituent size and hydroxyl content is also evident when all hydroxy groups are subtracted from the most potent inhibitor, tetracaffeoylquinic acid (compound 1A; IC 50 , 13 nM), to yield the 120-fold-less-potent tetracinnamoylquinic acid (compound 1I; IC 50 , 1.6 M). This theme is fortified by comparing 3,5-dicaffeoylquinic acid (compound 1B; IC 50 , 70 nM) with 3,5-dicoumaroylquinic acid (compound 1M; IC 50 , 24 M), whereby the subtraction of one meta hydroxyl group on each of the esterified substituents results in a 340-fold drop in the potency of TbCet1 inhibition.
Hydroxyl content per se is not predictive of potency for the 5-O-monoesters of quinic acid. In this series, a 5-caffeoyl group in compound 1T (2 hydroxyls) conferred no activity (IC 50 , Ͼ500 M), and a 5-galloyl moiety in compound 1M (3 hydroxyls) and a 5-cinnamoyl group (no hydroxyls) in compound 1Q conferred similarly feeble inhibition (IC 50 , 125 to 210 M), whereas 5-feruoylquinic acid (compound 1L; IC 50 , 12 M) and 5-pferuoylquinic acid (compound 1J; IC 50 , 2.4 M) were much more potent. The comparison of inactive 5-caffeoylquinic acid with active 5-feruoyl and 5-feruoyl derivatives illustrates the salutary effect of a methoxy group versus a hydroxyl at the meta or para positions. The position of monoesterification to the quinic acid scaffold is highlighted as a governor of inhibitor potency in the series of feruoylquinic acid isomers, whereby 4-feruoylquinic acid (compound 1P; IC 50 , 160 M) is 13-fold less effective than 5-feruoylquinic acid and 1-feruoylquinic acid (compound 1R) is inactive.
Selectivity of TbCet1 inhibitors. The promise of TTM RTPases as anti-infective drug targets would be greatest if one could identify inhibitors with broad-spectrum potency against fungal, protozoal, and viral TTM RTPases. To see if this property applied to any of the submicromolar TbCet1 inhibitors identified currently, we surveyed their effects on (i) S. cerevisiae Cet1; (ii) the RNA triphosphatase component of mimivirus capping enzyme (MimiCE) (18); (iii) the RNA triphosphatase component of the baculovirus Autographa californica nucleopolyhedrovirus (AcNPV) capping enzyme Lef4 (20)(21)(22); and (iv) Clostridium thermocellum TTM (CthTTM), a bacterial TTM family member that hydrolyzes inorganic polyphosphates, NTPs, and dinucleotide polyphosphate substrates (34,35). The IC 50 s of 25 compounds against the five ATPase enzymes surveyed are compiled in Table 5.

DISCUSSION
The present study affirms that a protozoan TTM RTPase, TbCet1 from T. brucei, is essential for growth of the parasite in culture and is subject to inhibition in vitro by small molecules at lowmicromolar and nanomolar concentrations. Applying HTS methods to an in-house library of nearly 200,000 compounds, we have identified several classes of chemicals (including chlorogenic acids, phenolic glycopyranosides, flavonoids, and other phenolics) that inhibit TbCet1. We subsequently confirmed the activity of the HTS hits, and various analogs thereof, by direct manual assays of phosphohydrolase activity.
The majority of the TbCet1 inhibitors that we identified showed strong specificity for the trypanosome enzyme and were manyfold less active (or inactive in some cases) against the orthologous TTM-type RTPases of mimivirus, baculovirus, and budding yeast (or the homologous bacterial phosphohydrolase CthTTM). Thus, (i) potent inhibition of a particular RTPase is no guarantee that the inhibitory compounds will have a quantitatively similar effect on another member of the TTM family, and (ii) an optimal search for inhibitors of the TTM-type mRNA cap-  ping enzymes of infectious pathogens will entail screening against RTPases from each organism or virus of interest. From the data in Table 5, we can infer certain trends in the responsiveness of other TTM enzymes to TbCet1 inhibitors. For example, SceCet1 was the least responsive, insofar as 13/25 TbCet1 inhibitors had no effect on the yeast enzyme (IC 50 , Ͼ500 M). MimiCE was unaffected by 7/25 inhibitors. CthTTM was also unaffected by 7/25 inhibitors, six of which overlapped with the compounds inactive against MimiCE. Lef4 was unaffected by 5/25 inhibitors, all five of which were also inactive versus MimiCE and CthTTM.
We conducted preliminary studies on the effects of tetracaffeoylquinic acid and rosmarinic acid on cultured procyclic T. brucei and found that exposure to 10 M tetracaffeoylquinic acid or 10 M rosmarinic acid had no effect on cell proliferation (data not shown). The lack of antitrypanosomal activity of tetracaffeoylquinic acid and rosmarinic acid could arise from either poor cellular uptake of the compounds, rapid efflux of the compounds, or their rapid intracellular catabolism.
Critical insights into the basis for TbCet1 inhibition by the various classes of phenolic compounds would be afforded by a crystal structure of RTPase-inhibitor complexes. However, our strenuous efforts to crystallize TbCet1, as apoenzyme, or in complex with Mn 2ϩ and/or ATP, or as inhibitor complexes with tetracaffeoylquinic acid or rosmarinic acid, have not met with success. Absent an atomic structure, we speculate that small-molecule inhibition of TbCet1 (or other TTM enzymes) could entail plugging of the entrance of the triphosphate tunnel to prevent ingress of NTP or triphosphate-terminated RNA to the active site within the tunnel cavity.

MATERIALS AND METHODS
Inducible RNAi knockdown of TbCet1. TbCet1 was depleted in T. brucei by using a tetracycline-inducible RNAi system as described previously (45). A 352-bp DNA fragment of the TbCET1 open reading frame (from nucleotides 19 to 371) was PCR amplified and inserted into the p2T7-177 vector (46), wherein expression of double-stranded RNA is under the control of opposing tetracycline-inducible bacteriophage T7 promoters. The RNAi construct was electroporated into procyclic T. brucei 29.13 cells, which constitutively express T7 RNA polymerase and Tet repressor (47). The procyclic cell line was cultivated in SDM 79 medium supplemented with 5% fetal bovine serum, 50 g/ml hygromycin B, and 15 g/ml G418, and the TbCet1-RNAi transfectants were selected with the  same medium plus 2.5 g/ml phleomycin. To gauge the effects of inducible RNAi knockdown of TbCet1 on cell proliferation, equal aliquots of the transfectants were inoculated into medium containing 1.0 g/ml tetracycline (ϩTet) to induce TbCet1 dsRNA production and into a parallel control culture lacking tetracycline (ϪTet). Cell density over a 14-day period was monitored by microscopy and was maintained between 1 ϫ 10 6 and 1 ϫ 10 7 cells/ml by dilution into fresh medium. Growth curves (Fig. 1A) display on the y axis the log of the direct cell count multiplied by the dilution factor. Cellular levels of TbCet1 during RNAi induction were gauged by Western blotting as follows. Aliquots (30 g of protein) of whole-cell lysates of TbCet1-RNAi transfectants harvested prior to addition of tetracycline (day 0) and after 1, 2, 6, 10, or 14 days of exposure to tetracycline were resolved by 12% SDS-PAGE. The polypeptides were transferred to a polyvinylidene difluoride membrane. TbCet1 was detected by incubating the membrane in a 1:500 dilution of affinity-purified rabbit polyclonal serum raised against recombinant TbCet1 protein (at Pocono Rabbit Farm and Laboratory, PA), followed by a 1:5,000 dilution of horseradish peroxidase-conjugated secondary antibody. Immune complexes were visualized by using an enhanced chemiluminescence (ECL) kit from Pierce. The intensity of the immunoreactive TbCet1 polypeptide was determined by scanning the autoradiogram with a Kodak Gel Logic 100 system and quantifying the signal intensity with molecular imaging software.
Recombinant TTM proteins. Recombinant full-length T. brucei RTPase TbCet1 (11,12), the N-terminal RTPase domain (amino acids [aa] 1 to 237) of mimivirus capping enzyme (MimiCE) (18), the N-terminal RTPase domain (aa 1 to 236) of AcNPV Lef4 capping enzyme (21), the biologically active C-terminal catalytic domain (aa 241 to 539) of S. cerevisiae RTPase (SceCet1) (4), and full-length C. thermocellum TMM (CthTTM) (34) were produced in Escherichia coli and purified as described in the publications cited. TbCet1, MimiCE, Lef4, and CthTTM were produced as N-terminal HisSmt3-tagged fusions, and SceCet1 was produced as an N-terminal His-tagged fusion. After nickel-affinity purification of the tagged proteins from soluble bacterial lysates, the HisSmt3 tags were removed by the Smt3 protease Ulp1, and the tag-free recombinant proteins were recovered in the flowthrough fraction during a second round of nickel-affinity chromatography. The tag-free enzymes and Histagged SceCet1 were further purified by gel filtration through a HiPrep Sephacryl S300 size exclusion column (GE Healthcare) equilibrated in buffer C (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 10% glycerol, 0 to 5 mM dithiothreitol [DTT]). Peak fractions were pooled and adjusted to attain solutions of 10 M TbCet1, 35 M MimiCE, 42 M SceCet1, 500 M CthTTM, and 7 M Lef4. The proteins were stored at Ϫ80°C in single-use aliquots.
HTS assay for inhibitors of TbCet1 ATPase. The high-throughput screening (HTS) assay was based on the detection of P i release from ATP via the use of malachite green reagent. P i is quantified by measuring the A 630 of a chromophore complex formed by malachite green and phosphomolybdate. For HTS procedures, 1 l of test compound in 10% dimethyl sulfoxide (DMSO) was predispensed into flat-bottomed transparent 384well plates to which was added 4 l of ATPase master mix containing the following reagents to attain the final concentrations indicated: 50 mM Tris-HCl (pH 8.0), 20 mM MgCl 2 , 2 mM DTT, 4 mM MnCl 2 , and 2 mM ATP. Reactions were initiated with the addition of 5 l of a solution containing 50 mM Tris-HCl (pH 8.0), 28.6 nM TbCet1, 0.02 mg/ml bovine serum albumin, 2 mM DTT, 10% glycerol, 0.005% Tween 80. "Noinhibitor" control reaction mixtures received 1 l of 10% DMSO alone in lieu of test compound. "Complete-inhibition" control reaction mixtures received 1 l of 400 mM EDTA in 10% DMSO. The plates were covered and incubated for 1 h at room temperature. The reactions were then quenched with 40 l of malachite green solution, made by mixing 1 volume of 0.125% (wt/vol) malachite green in 2 M H 2 SO 4 with 4 volumes of 0.47% (wt/vol) ammonium molybdate in 0.055% (vol/vol) Tween 20. The quenched plates were incubated for 1 h at room temperature and then scanned to quantify A 630 for each reaction well. Z= for the assay was gauged by comparing three 384-well plates of no-inhibitor control reactions with three 384-well plates of complete-inhibition control reactions. The effect of any test compound on TbCet1 was quantified by comparison of its A 630 value for P i release with those of the no-inhibitor (0% inhibition) and complete-inhibition (100% inhibition) controls. IC 50 s in the HTS assay format were determined by testing a series of 12 serial 2-fold dilutions of the test compound in 10% DMSO.
TLC assay for 32 P i release from [␥-32 P]ATP. Reaction mixtures (20 l) contained 100 mM Tris-HCl, pH 8.0, 10 mM NaCl, 1% (vol/vol) glycerol, 2 mM MnCl 2 , 200 M [␥-32 P]ATP, 5% DMSO, TTM enzymes (either 2.2 nM TbCet1, 7 nM MimiCE, 25 nM SceCet1, 50 nM Lef4, or 500 nM CthTTM), and inhibitors as specified. The inhibitors were serially diluted in DMSO, and 1 l was included in each ATPase reaction mixture. The reactions were initiated by addition of enzyme. After incubation for 10 min at 37°C (for TbCet1, MimiCE, Lef4, and CthTTM) or 30°C (for SceCet1), the reactions were quenched by adjustment to 1 M formic acid. Aliquots of the mixtures were applied to polyethyleneimine cellulose TLC plates and analyzed by ascending TLC with 200 mM ammonium sulfate as the mobile phase. Radiolabeled ATP and P i were visualized and quantified by scanning the TLC plate with a Fuji imager. The concentrations of input enzyme were chosen so as to attain hydrolysis of~30% of the input ATP substrate in the "no-inhibitor" control reaction mixtures containing 5% DMSO. The extents of ATP hydrolysis were plotted versus inhibitor concentration after normalization to the "no-inhibitor" control values. IC 50 s were obtained using the "dose-response" curve fitting algorithms in Prism.