The Antibiotic Activity of N-Pentylpantothenamide Results from Its Conversion to Ethyldethia-Coenzyme A, a Coenzyme A Antimetabolite*

Pantothenic acid (vitamin B5) is the natural precursor of coenzyme A (CoA), an essential cofactor in all organisms. The pantothenic acid antimetaboliteN-pentylpantothenamide inhibits the growth ofEscherichia coli with a minimum inhibitory concentration of 2 μm. In this study, we examine the mechanism of this inhibition. Using the last five enzymes of the CoA biosynthetic pathway in E. coli we demonstrate thatN-pentylpantothenamide does not inhibit the CoA biosynthetic enzymes but instead acts as an alternative substrate, forming the CoA analog ethyldethia-CoA. We show thatN-pentylpantothenamide is converted to ethyldethia-CoA 10.5 times faster than CoA is biosynthesized from pantothenic acid, demonstrating that ethyldethia-CoA biosynthesis can effectively compete with CoA biosynthesis in the cell. We conclude that the mechanism of toxicity of N-pentylpantothenamide is most likely due to its biosynthetic conversion to the CoA analog ethyldethia-CoA, which may act as an inhibitor of CoA- and acetyl-CoA-utilizing enzymes.

carboxylate of pantothenic acid 2a is replaced by a sulfonate, does not inhibit the growth of Escherichia coli because it is not transported by pantothenate permease (6); and N-pentylpantothenamide 2b, which inhibits E. coli with a minimum inhibitory concentration (MIC) of ϳ2 M (7).
In this study we considered two mechanisms for the antibiotic activity of N-pentylpantothenamide 2b. In the first mechanism, inhibition of CoA biosynthesis by N-pentylpantothenamide 2b or one of its metabolites would result in a decreased amount of CoA being formed in the cell and thus would retard bacterial cell growth. In the second mechanism, we considered the possibility that 2b is biosynthetically converted to the CoA analog ethyldethia-CoA, which might function as a CoA antimetabolite and inhibit CoA and acetyl-CoA-utilizing enzymes. Since the five enzymes required for the conversion of pantothenic acid to CoA (Fig. 1) have all recently been overexpressed and characterized (8 -11), we were able to differentiate between these two mechanisms in this paper.

Materials
General-All chemicals and buffer components were purchased from Aldrich, Sigma or Fisher Scientific and used without further purification. 4Ј-Phosphopantetheine 5 was synthesized as previously described (12). 1 H NMR spectra were measured on a Varian INOVA 400 MHz instrument. All HPLC analyses were performed on a HP series 1100 HPLC system with HPLC grade solvents using a Supelcosil LC-18-T 3 m, 15 cm ϫ 4.6 mm ID column (Supelco). ESI-MS analyses were performed at the Cornell Biotechnology Resource Center on a Bruker Esquire-LC ESI-ion trap mass spectrometer by direct infusion of the analyte mixture into the instrument at a rate of 1 l/min. Synthesis of N-Pentylpantothenamide (2b)-Sodium pantothenate (2.0 g, 8.3 mmol) was dissolved in deionized water, and the solution was passed through an Amberlite IR-120 (H ϩ ) column. The column was subsequently washed with two column volumes of deionized water, and the combined eluates were lyophilized. The free acid obtained in this manner was dissolved in dry dimethyl formamide (10 ml), and amylamine (1.16 ml, 10 mmol) and diphenylphosphoryl azide (2.24 ml, 10 mmol) were added. The solution was cooled to 0°C and triethylamine (1.39 ml, 10 mmol) was added. The solution was stirred at 0°C for 2 h, followed by stirring at room temperature overnight. The volume of solvent was reduced by rotary evaporation with external heating, and the residue applied to a short silica gel column, eluting with 95:5 CH 2 Cl 2 /methanol and concentrating the eluate in vacuo to remove final traces of dimethyl formamide. The product was purified by flash column chromatography on silica gel with 95:5 CH 2 Cl 2 /methanol, and the fractions containing 2b were collected and were concentrated in vacuo yielding the product as a colorless oil. The oil was dissolved in deionized water and lyophilized to give 2b as a white powder (1.56 g, 65% yield). 1  was filtered and added drop-wise to a solution of 2b (577 mg, 2.0 mmol) in dry pyridine (9 ml) at Ϫ40°C (dry ice/acetonitrile). After stirring at Ϫ40°C for 2 h, the mixture was placed in a Ϫ20°C freezer overnight. The reaction was allowed to warm to room temperature and was subsequently quenched with water (3 ml). The solvent was removed in vacuo and ethyl acetate (25 ml) was added. The resulting suspension was washed with 1 M H 2 SO 4 (2 ϫ 5 ml) and 1 M NaHCO 3 (2 ϫ 5 ml), saturated Na 2 SO 4 (1 ϫ 5 ml), and dried (Na 2 SO 4 ), and then the solvent was removed. The product was purified by flash column chromatography on silica gel with 95:5 CH 2 Cl 2 /methanol, and the fractions containing 7 were collected and concentrated in vacuo to give the product as a colorless oil (271 mg, 25% yield). 1  Synthesis of 4Ј-Phospho-N-pentylpantothenamide (3b)-Pd/C (10%, 20 mg) was added to a solution of 7 (200 mg, 365 mol) in 9:1 methanol/ water (10 ml), and the suspension was hydrogenated at atmospheric pressure and room temperature for 2 h. The solution was filtered, and the solvent removed in vacuo to give 3b as a clear glass (110 mg, 82% yield). The product was dissolved in H 2 O, titrated to pH ϳ6.0 with 1 M NaOH and stored as frozen aliquots of a stock solution (60 mM) at Ϫ20°C. 1

Measurement of the Minimal Inhibitory Concentration: Growth Curves
Minimal medium (0.8 mM MgSO 4 , 10.0 mM citric acid, 60 mM K 2 HPO 4 , 20 mM NaNH 4 HPO 4 , 0.5% glucose) (50 ml) was inoculated with 50 l of an overnight culture of E. coli B and subsequently grown at 37°C in the presence of varying concentrations of 2b. The A 600 was periodically determined until the stationary growth phase was reached in each respective culture. The minimal concentration of 2b that retarded the onset of logarithmic growth under these conditions was regarded as the minimal inhibitory concentration (MIC).
CoaD-Phosphopantetheine adenylyltransferase from E. coli (CoaD) has previously been overexpressed from plasmid pUC/coaD (a gift from D. Drueckhammer) and purified (8). However, as this plasmid did not encode a His-tagged protein, pUC/coaD was treated with NdeI and BamHI, and the resulting coaD-containing fragment was cloned into NdeI/BamHI-digested pET28a expression vector (Novagen). The sequence of the resulting plasmid, designated pESC106, was verified by automated DNA sequencing (Cornell Biotechnology Resource Center) and was transformed into E. coli BL21(DE3) (Novagen). To obtain pure protein E. coli BL21(DE3) pESC106 was grown at 37°C in 500 ml of LB broth supplemented with 15 g/ml kanamycin sulfate to A 600 ϳ0.6 and induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. After growing overnight at 37°C, the cells were harvested, suspended in sonication buffer (5 mM imidazole, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9, 10 ml/g cell paste), disrupted by sonication, and centrifuged at 35,000 ϫ g for 30 min to clarify the cell-free extract. This was applied to a 2-ml His-Bind column (Novagen). Weakly bound proteins were removed by washing with sonication buffer, followed by sonication buffer containing 60 mM imidazole. CoaD was eluted by increasing the imidazole concentration to 0.5 M. The chromatography was monitored at A 280 .
CoaE-Dephospho-CoA kinase from E. coli (CoaE) has previously been overexpressed as a non-fusion protein using pET/coaE (a gift from D. Drueckhammer) (11). To obtain the His-tagged protein, the coaE (previously yacE) gene was amplified by PCR using as the forward primer: 5Ј-CCGGGAATATAACATATGAGGTATATAGTTGCC-3Ј, to introduce an NdeI site (underlined) at the start of the gene, and as the reverse primer: 5Ј-AAAGGACCTGGATCCGCATTACGG-3Ј to introduce a BamHI site (underlined) at the end of the gene. The resulting PCR product was cloned into NdeI/BamHI-digested pET28a expression vector (Novagen). The sequence of the resulting plasmid, designated pESC124, was verified by automated DNA sequencing (Cornell Biotech- nology Resource Center) and was transformed into E. coli BL21(DE3) (Novagen). To obtain pure protein, E. coli BL21(DE3) pESC124 was grown at 37°C in 500 ml of LB broth supplemented with 15 g/ml kanamycin sulfate to A 600 ϳ0.6, and induced with 100 M isopropyl-1thio-␤-D-galactopyranoside. After growing overnight at 37°C, the cells were harvested, suspended in sonication buffer (5 mM imidazole, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9, 10 ml/g cell paste), disrupted by sonication, and centrifuged at 35,000 ϫ g for 30 min to clarify the cell-free extract. This was applied to a 2.5-ml His-Bind column (Novagen). Weakly bound proteins were removed by washing with sonication buffer, followed by sonication buffer containing 60 mM imidazole. CoaE was eluted by increasing the imidazole concentration to 0.5 M. The chromatography was monitored at A 280 .
Biosynthesis and Purification of Ethyldethia-CoA (1b)-A 600-l reaction mixture contained 3b (10.5 mM), ATP (20 mM), MgCl 2 (5.0 mM), CoaD (100 g) and CoaE (150 g) in 50 mM Tris-HCl buffer (pH 7.6). Reactions were initiated by addition of the biosynthetic enzymes, incubated for 2 h at 37°C and stopped by transferring the reaction to 95°C for 5 min, and the precipitated protein was removed by centrifugation (13,000 rpm ϫ 5 min). The supernatant was loaded onto a DEAEcellulose column (1 ϫ 25 cm) pre-equilibrated with NH 4 HCO 3 (50 mM), and the column was eluted with a 600-ml gradient of NH 4 HCO 3 (50 -300 mM). The chromatography was monitored at A 254 . The product eluted as the last fraction from the column at ϳ180 mM NH 4 HCO 3 . The product-containing fractions were combined and lyophilized, dissolved in water and lyophilized again. This was repeated until a constant weight of product was achieved. Yield: 4.4 mg (tetra-ammonium salt) (83%). 1  Biosynthesis and Purification of 3Ј-Dephospho-ethyldethia-coenzyme A (6b)-A 600-l reaction mixture contained 3b (10.5 mM), ATP (20 mM), MgCl 2 (5.0 mM), CoaD (100 g) and inorganic pyrophosphatase (2 units) in 50 mM Tris-HCl buffer (pH 7.6). Reactions were initiated by addition of the biosynthetic enzymes. Four identical reaction mixtures were treated as for the purification of 1b, their supernatants combined and loaded onto a single DEAE-cellulose column (1 ϫ 25 cm) preequilibrated with NH 4 HCO 3 (50 mM). The column was eluted with a 500-ml gradient of NH 4 HCO 3 (50 -250 mM). The chromatography was monitored at A 254 . The product eluted as the first fraction from the column at ϳ90 mM NH 4 HCO 3 . The product-containing fractions were combined and lyophilized, dissolved in water and lyophilized again. This was repeated until a constant weight of product was achieved. Yield: 14.8 mg (di-ammonium salt) (84%). 1

Enzyme Assays
General-All enzyme assays were based on the decrease of NADH concentration, as monitored by changes in absorbance at 340 nm. An extinction coefficient of 6220 M Ϫ1 ⅐cm Ϫ1 was used for NADH. Reactions were performed at 25°C in a Hitachi U-2010 Spectrophotometer. Kinetic parameters were determined by fitting the obtained data to the Michaelis-Menten equation using Origin 6.0 (Microcal).
CoaA-Pantothenate kinase activity was determined using a continuous spectrophotometric assay that coupled the production of ADP to the consumption of NADH. Each 500-l reaction mixture contained ATP (1. CoaD-Phosphopantetheine adenylyltransferase activity was assayed by the method of O'Brien (13) using the pyrophosphate detection kit available from Sigma (catalog no. P7275) that couples the production of pyrophosphate to the consumption of NADH. Each 500-l reaction mixture contained 200 l of pyrophosphate reagent, ATP (1.5 mM), DTT (1.0 mM), MgCl 2 (10 mM), KCl (20 mM), and CoaD (5 g) in 50 mM Tris-HCl buffer (pH 7.6). The reaction was initiated by addition of substrate (5 or 3b, at concentrations between 10 and 250 M. CoaE-Dephospho-CoA kinase activity was determined by a modification of the published assay method (11). Each 500-l reaction mixture contained ATP (5 mM), NADH (0.3 mM)), phospho(enol)pyruvate (0.5 mM), MgCl 2 (10 mM), KCl (20 mM), pyruvate kinase (5 units), lactic dehydrogenase (5 units), and CoaE (5 g) in 50 mM Tris-HCl buffer (pH 7.6). The reaction was initiated by addition of substrate (6a or 6b, at concentrations between 0.13 and 2.0 mM). At set time intervals 40-l aliquots were withdrawn and immediately placed in a microcentrifuge tube preheated to 95°C. After 5 min at 95°C, the sample was centrifuged (13,000 rpm ϫ 5 min) and 30 l of the supernatant was added to 6 l of guanosine (6 mM in 0.5 M HCl) as internal standard. The solution was mixed well, and 25 l was injected onto the HPLC column for analysis using the same conditions as described above. The concentrations of 1a and 1b were determined using a standard curve obtained using known amounts of pure CoA co-injected with the same internal standard.

RESULTS AND DISCUSSION
The pantothenic acid analog N-pentylpantothenamide 2b and its phosphate 3b were obtained by the improved synthetic route shown in Fig. 2. We found that compound 2b inhibited the growth of E. coli in minimal medium (MIC ϭ 2 M) as previously reported (7). Analysis of the growth rate of E. coli grown in minimal medium and in the presence of varying amounts of 2b shows that the inhibitor retarded the onset of logarithmic growth, but did not abolish it entirely (Fig. 3). Furthermore, regardless of the initial amount of inhibitor present, all cultures grew to similar final cell densities. Taken together, these results show that compound 2b is bacteriostatic rather than bacteriocidal and that E. coli can overcome the inhibitory effect of N-pentylpantothenamide over time.
We considered two hypotheses for the toxicity of 2b: either it inhibits CoA biosynthesis, thus limiting the amount of CoA available to cells, or it is biosynthetically converted to ethyldethia-CoA, a potential CoA antimetabolite that could inhibit CoA utilizing enzymes.
To differentiate between these mechanisms the last five enzymes involved in CoA biosynthesis were isolated from overexpression strains and incubated with the natural precursor 2a or its analog 2b, and the reaction products were analyzed by HPLC. The chromatograms from these reaction mixtures are shown in Fig. 4. When 2a, the natural biosynthetic intermediate, was used as the substrate, the major new product formed was identified as CoA since it co-eluted with an authentic sample of 1a (Fig. 4A). When the analog 2b was used as the substrate, a new compound, eluting with a longer retention time than CoA, was formed (Fig. 4B). To determine the identity of the new compound authentic ethyldethia-CoA 1b was prepared synthetically using 3b as the starting material, which allows its purification from reaction mixtures in multimilligram quantities by anion exchange chromatography. The structure of the material obtained in this way was confirmed as being ethyldethia-CoA 1b by 1 H NMR and ESI-MS analyses.  The compound first identified by HPLC analysis was thus shown to be ethyldethia-CoA 1b by co-elution with this authentic sample. In the same way, authentic dephospho-ethyldethia-CoA 6b was prepared by replacing the CoaE in the reaction mixtures with inorganic pyrophosphatase, and the synthetic material used to determine the identity of a compound eluting after 27 min, also by co-elution (not shown). These results demonstrate that the E. coli CoA biosynthetic enzymes (CoaABCDE) can catalyze the conversion of N-pentylpantothenamide 2b to ethyldethia-CoA 1b.
To establish the extent to which the CoA biosynthetic enzymes differentiate between the natural substrates and their respective analogs, we have determined k cat /K m for both possible substrates of each enzyme (Table I). This was done using standard kinase assays for CoaA and CoaE, while an assay based on pyrophosphate release was used in the case of CoaD. These experiments demonstrate that, except in the case of CoaA where the difference is small, the analog is a better substrate than the natural precursor. In the case of CoaD, 3b is adenylylated ϳ3 times faster than 5, while CoaE exhibits a slightly larger preference, with the phosphorylation of 6b to give ethyldethia-CoA 1b occurring 4.6 times faster than the phosphorylation of 6a. In addition, treatment of a mixture of the natural substrate 3a (50 M) and the analog 3b with the bifunctional enzyme phosphopantothenoylcysteine synthetase/ decarboxylase (10) failed to reveal any inhibition of the synthetase activity of this enzyme with concentrations of the analog as high as 3 mM, as determined using the pyrophosphate release assay. These results demonstrate that inhibition of the CoA biosynthetic enzymes cannot be the mode of antibiotic action of N-pentylpantothenamide 2b.
To test the prediction from the kinetic parameters in Table I that the biosynthesis of ethyldethia-CoA 1b is faster than the biosynthesis of CoA, 2a and 2b (500 M) each) were incubated with purified CoaABCDE and the formation of 1a and 1b was followed by HPLC (Fig. 5). Based on initial rates, the biosynthesis of ethyldethia-CoA (1b) occurs 10.5 times faster than the biosynthesis of CoA. Since pantothenate kinase does not significantly differentiate between 2a and 2b, this rate difference is due to the preferential adenylylation of 3b over 5, the preferential phosphorylation of 6b over 6a, and the fact that the conversion of 3b to 1b requires two enzyme activities (CoaBC) less than the conversion of 3a to 1a. Thus, while the relative concentrations of the biosynthetic enzymes and the ratio of 2a to 2b in vivo is likely to be different from what was used here, this experiment suggests that the biosynthesis of ethyldethia-CoA is likely to compete effectively with the biosynthesis of CoA in the cell.
In conclusion our data show that the CoA biosynthetic en-zymes are not inhibited by N-pentylpantothenamide 2b; rather this compound functions as a substrate of these enzymes, as shown by its preferential conversion to ethyldethia-CoA 1b in the competition experiment. This suggests that the toxicity of N-pentylpantothenamide results from the inhibition of CoAutilizing enzymes by the CoA analog ethyldethia-CoA 1b. The proposed antimetabolite activity of 1b is supported by the demonstration that desulfo-CoA, which differs from 1b only in the length of the alkyl chain of the N-substituted amide, inhibits carnitine acetyl transferase, phosphotransacetylase, citrate synthase, ␤-hydroxy-␤-methylglutaryl-CoA synthase and ␣-ketoglutarate dehydrogenase (14). Our mechanistic proposal for the mode of action of 2b is analogous to the recently discovered mechanism of action of bacimethrin, which is converted in vivo to methoxythiamin pyrophosphate, a thiamin pyrophosphate antivitamin (15). The facile in vivo conversion of N-pentylpantothenamide to ethyldethia-CoA overcomes the permeability problem of getting CoA analogs into cells; as a result N-pentylpantothenamide may be a useful compound for studying the physiological consequences of depriving cells of CoA and for trapping intermediates on complex enzymes such as fatty acid synthase, polyketide synthases, and non-ribosomal polypeptide synthetases, all of which utilize the CoA-derived 4Ј-phosphopantetheine cofactor.