Antibacterial Activity of 2'-Esters of Erythromycin

Transcription

1 APPLIED MICROBIOLOGY, Aug. 1969, p Copyright ( 1969 American Society for Microbiology Vol. 18, No. 2 Printed in U.S.A. Antibacterial Activity of 2'-Esters of P. L. TARDREW, J. C. H. MAO, AND D. KENNEY Biological Development, Abbott Laboratories, Scientific Divisions, North Chicago, Illinois Received for publication 21 March 1969 The effect of esterification at the 2'-position of desosamine on the antibacterial activity of was investigated by determining the bacteriostatic and bactericidal activities of and a number of its 2'-esters on S. aureus and relating these activities to the hydrolysis rates of the esters. These studies, together with comparison of the inhibition of protein synthesis in a cell-free system isolated from S. aureus, lead to the conclusion that 2'-esters of are inactive until hydrolyzed. Loss of activity appears to result from inability of esters to bind to bacterial ribosomes and thus inhibit synthesis of protein. 2'-Esters of were found by the different esters were established. Second, the Murphy (5) to possess less antimicrobial activity bactericidal activity of and some of than the parent compound,. Although the 2'-esters of are less ac- inhibits bacterial growth by its 2'-esters was determined as a function of time. tive than the parent compound, they are not blocking protein synthesis (4, 7). At saturation, completely inactive, and the question arises a 1:1 complex between and ribosomes is formed (3), and available evidence (8) whether the low level of activity observed was due to formed by ester hydrolysis indicates that binding of to ribosomes is a prerequisite for inhibition of protein or whether the 2-esters possess intrinsic antibacterial activity. Kavanagh (1) states that esters (for example, 2'-propionyl or 2'-ethyl to ribosomes was therefore tested synthesis. The binding of several 2'-esters of carbonyl ) have little if any antibacterial activity until hydrolyzed. and certain other active derivatives. and compared with the binding of When is administered orally in the form of its 2-esters (2'-ethyl succinyl or 2'-ethyl carbonyl ) or as (970 fig per mg) and 2'-ethyl suc- MATERIALS AND METHODS ester salts ( estolate), the compound cinyl (840 Ig per mg) were manufactured at Abbott Laboratories, North Chicago, Ill. is absorbed, distributed in the body, and excreted in large part in the ester form (R. G. Wiegand, Propionyl (855 p&g per mg) was purchased from the Eli Lilly & Co., Indianapolis, Ind. et. al. in preparation). Since hydrolysis of the ester is essentially complete under conditions Acetyl (600 ug per mg) and benzoyl routinely used for analysis of antibacterial activity (about 514 pg per mg) were synthesized in serum, the serum concentrations reported in by the method of Murphy (5). Prior to use, the latter the literature for these dosage forms are a measure two esters were purified by countercurrent distribution by using the system described by of total concentration (base plus ester). This Pettinga, et. al. (6). B, C, and N-ethyl would not be a true measure of the antibacterial were all prepared in the Department of activity in the serum in vivo if, indeed, the Organic Chemistry, Abbott Laboratories. 2'-esters of are inactive. Thus Hydrolysis studies. Hydrolysis rates were determined in 0.1 M phosphate buffer at ph values of 6.5, more definitive data on the activity of esters was sought. 7.0, and 7.5, in Difco Brain Heart Infusion (BHI) In this study, several approaches were used to broth (37.0 g/liter in deionized water adjusted to ph determine the effect of esterification in the 2'-position on the antibacterial activity of. human serum. The procedure was as follows. A known 6.5 before sterilization for 30 min at 125 C) and in First, the inhibition of growth of S. aureus volume, usually 50 ml of by buffer, BHI broth, or human serum was warmed in a water bath to 37 C. A and a number of 2'-esters was determined as a function of time and drug concen- 50 ;liter of a 10 mg/ml of solution) was added to the known quantity of ester in acetone solution (usually tration, and the relationships between these determined activities and the hydrolysis rates of mediately after mixing and placed in a capped test heated solution, and a 10-ml sample was taken im- 159

2 160 TARDREW, MAI,0, AND KENNEY APPL. MICROBIOL. tube containing 7 ml of ether held in an ice bath. The sample was extracted immediately by the procedure outlined below. The solution of ester was shaken at 37 C in the water bath, and, at appropriate intervals (15 min for rapidly hydrolyzing esters, 30 to 60 min for more slowly hydrolyzing esters), samples were withdrawn and extracted immediately. The test tube containing the sample and ether were shaken (120 to 140 cycles per minute) on a reciprocating shaker for 5 min, and the phases were separated by centrifuging in an International model V centrifuge for 5 min. With the aid of a syringe and pipette, the ether was removed and put into separate clean test tubes. A 7 ml portion of ether and 10 ml of phosphate buffer (ph 9.5; 600 g/liter) were added to the test tubes containing the samples, and the extraction procedure was repeated. The ether was then removed and combined with the ether from the first extraction, and the extraction procedure was repeated with 7 ml of ether. A small quantity of anhydrous sodium sulfate was added to the combined ether extract. The ether was decanted into a second clean test tube and evaporated to dryness with an air jet at about 30 C. Two procedures were used to determine the ratio of ester to base in the extracted samples. In the first of these, the air-dried ether extracts were dissolved in 0.1 ml of Analytical-Reagent (AR) grade acetone. Of this solution, 50 pliters were applied to thin-layer chromatographic plates prepared by applying 30 g of Silica Gel G (E. Merck AG Darmstadt, Germany) in 60 ml of water to a thickness of 0.25 mm and drying at 110 C for 1 hr. An base and ester standard was applied to each plate. The chromatograms were developed with a 2:7:1 mixture of diethyl carbinol, ethyl acetate, and dimethyl formamide. In this system the RF of is 0.35, whereas the esters fall in the range of Rp 0.65 to After development the plates were removed and dried for 30 min at 75 C in a vacuum oven. While the plates were still hot, they were sprayed with a mixture of arsenomolybdate reagent and 6 N H2SO4 in the ratio of 1:2 (v/v). and its esters are detected as blue spots on a white background. After the plates had cooled, the ester and base zones were scraped off into clean Pyrex test tubes (20 by 20 mm) to which 0.1 ml of AR grade chloroform was added. Five ml of the arsenomolybdate, 6 N H2SO4 reagent in a ratio of 1:2 was added to each test tube which was then placed in a boiling-water bath for 17 min. After cooling, the tubes were centrifuged at 1,000 X g on an International centrifuge size Z model V for 10 min, and the optical density (OD) was determined at 660 nm with the use of a DU spectrophotometer (Beckman Instruments Inc., Fullerton, Calif.). Per cent ester in the sample was calculated from the equation: % ester = OD ester zone/(od ester zone + OD base zone) X 100 The arsenomolybdate reagent used in the spray reagent for detection of ester and base on the developed thin-layer chromatographic plates and also in the analytical reagent was prepared as follows. Of ammonium molybdate [(NH4)6Mo7O24-4H2O], 50 g was dissolved in 850 ml of deionized water. Of concentrated H2SO4, 42 ml was added with constant stirring of the solution. The solution was cooled during addition of acid so that its temperature did not exceed 50 C. A 6-g portion of sodium arsenate (Na2HAsO4) 7H20) dissolved in 75 ml of deionized water was added with stirring to the ammonium molybdate solution. Water was then added to a total volume of 1 liter. The solution was stored in an amber bottle and was not used until 24 hr after preparation. In the second procedure, ester-base ratios were determined bioautographically. Sufficient acetone solution (prepared as described in the thin-layer procedure) to give 200,ug of activity was applied to strips of Eaton Dikeman 613 paper 5/16-inch (0.8 cm) wide. The strips were developed at 28 C by the upflow procedure for 3 hr by using a solvent prepared by dissolving 25 g of AR ammonium chloride and 50 ml of p-dioxane in 850 ml of distilled water, adjusting to a ph of 5.5 with ammonium hydroxide, and diluting to 1 liter. The solvent front was marked after development and the strips were air-dried. Bioautographs were produced by placing the developed strips on a single layer of Streptomycin Assay Agar (Difco) seeded with a known volume of a spore suspension of Bacillus subtilis and incubated overnight at 28 C. The developed bioautographs were photostated, the ratio of the diameters of the ester zone (RF approximately 0.5) to the base zone (RF approximately 0.7) was measured, and the per cent ester was determined from a standard curve. A typical standard curve is shown in Fig. 1. Bacteriostatic studies. Bacteriostatic activity of and its esters were determined by the following procedure. A 3.0-ml portion of BHI broth (ph 6.5) was inoculated with 0.5 ml of an actively growing subculture (1 to 2 hr) of an 18-hr broth cul- I 4A z o A PERCENT ESTER FIG. 1. Ratios of diameters of the zones of inhibition of Bacillus subtilis I produced by ethyl succinyl and after bioautography plotted as a function of per cent ester.

3 VOL. 18, 1969 ANTIBACTERIAL ACIIVITY OF ESTERS 161 ture of Staphylococcus aureus 209P at 37 C. Antibiotics dissolved in acetone were added 47 to 75 min later (ODa, = 0.04 to 0.06) to give the desired concentration. An equivalent volume of acetone containing no antibiotic was added to the control. Changes in OD at 420 nm were determined to measure growth, by using a Beckman model DU spectrophotometer, at intervals of 15 mi. Bactericidal studies. The bactericidal test was run at 37 C in BHI broth (ph 6.5) on a reciprocating shaker. The tubes (10 ml each) were inoculated with an actively growing subculture (1 to 2 hr old) of an 18-hr broth culture of S. aureus Smith. At 0 hr, freshly weighed samples of the antibiotics were dissolved in acetone and added to the inoculated test broth so that a level of 2.5 pg of activity per ml was obtained. At 1-hr intervals, the tubes were removed from the shaker; 0.1 ml was removed from each tube, and the viable cells remaining were determined by using the pour plate technique. The counting medium utilized was composed of tryptone 0.3%, beef extract 0.37%, dextrose 0.1%, yeast extract 0.1%, and agar 1.5%. The results were recorded as viable cells per milliliter. Both the bacteriostatic and bactericidal tests were carried out at a ph of 6.5 to minimize hydrolysis of the esters. Inhibition of protein synthesis in the cell-free system. The procedures for growth of S. aureus 209P and preparation of the cell-free protein-synthesizing extract have been reported (2). The reaction mixture for protein synthesis contained the following components per milliliter: tris(hydroxymethyl)aminomethane(tris)-hydrochloride buffer (ph 7.5), 10 pmoles; magnesium acetate, 16 pmoles; ammonium acetate, 50 pmoles; dithiothreitol, 0.1 ;mole; ATP, 1 ;mole; GTP, CTr, and UPT, 0.05 pmole of each; phosphoenolpyruvate, 5 pmole; pyruvate kinase, 40 p&g; a mixture of 20 Lo C-amino acids, excluding the radioactive amino acid, 0.03 pmoles of each; an L-14Camino acid, 0.25 pc, and cell-free extract, in a total volume of 0.5 ml. Assays were performed in the absence of added polyuridylic acid (endogenous synthesis) and in the presence of 100 pg of polyuridylic acid. After incubation at 37 C for 30 min, reactions were stopped by the addition of trichloroacetic acid to a final concentration of 5%. Samples were heated at 90 C for 15 min. The precipitates were deposited on glass fiber discs (Reeves Angel, 934AH) and washed under suction with three 3-ml volumes of 5% trichloroacetic acid and twice with 3 ml of ethyl alcohol. The discs were placed in scintillation vials, dried, and counted with 10 ml of scintillation fluid in a liquid scintillation spectrometer with counting efficiency of about 80%. Binding of 14C- to ribosomes. The TABLE 1. Hydrolysis of 2'-esters of at 37 C (thin-layer chromatographic analysis) Ester Conditions (1) ph Hydrlysi 95% Confidence limits min 2'-Ethyl succinyl BHIa PBb PB PB Human serum '-Acetyl BHI '-Propionyl BHI PB PB PB Human serum '-Ethyl carbonyl BHI PB PB PB '-Benzoyl BHI 6.5 1, PB PB PB a Brain Heart Infusion broth. b 0.1 M Phosphate buffer.

4 162 TARDREW, MAO, AND KENNEY standard binding reaction was performed in a total volume of 0.1 ml containing 0.4 mg of S. aureus ribosomes in the standard buffer [0.01 M Tris, M Mg C12, 0.05 M NH4CI, and M dithiothreitol TABLE 2. Hydrolysis of 2'-esters of at 37 C (bioautographic procedure) Ester Condi- tions ph Hydrolysis half-life min 2'-Ethyl succinyl BHI '-Propionyl BHI '-Ethyl Carbonyl BHI '-Benzoyl BHI 6.5 >400 E 0 C~4 I.- at 0 APPL. MICROBIOL. (ph 7.5)] containing 1.4 X 106 M 14C- (15 c/mole). After 5 min of incubation at 34 C, the reaction mixtures were diluted with 3 ml of cold standard buffer, filtered through a nitrocellulose filter (Schleicher and Schuell Co., Keene, N. H.; type B-6) and washed with three 3-ml portions of cold buffer. The radioactivity on the filter was determined by liquid scintillation counting. Since radioactive derivatives of were not available, a direct test of the binding of derivatives to ribosomes was not possible. The alternative way was to test the ability of derivatives to reduce the binding of '4C- to ribosomes. A reduction of '4C- binding means that derivatives can compete with for the binding site on the ribosome. In these binding experiments, the unlabeled derivatives were dissolved in 95% ethyl alcohol (because of the low solubility of derivatives and the danger of hydrolysis of esters in water). A preliminary experiment established that 5 Kliters of ethyl alcohol in the reaction mixture has no effect on TIME (MINUTES) FIG. 2. Effect of and certain of its 2'-esters on the growth of S. aureus 209P as measured by the increase in OD at 420 nm.

5 VOL. 18, 1969 ANTIBACTERIAL ACTIVITY OF ESTERS '163 the binding. The 5 pliters of ethyl alcohol solution which contained 1.4 X 10-1, 3.5 X 10-10, or 7 X 1010 moles of derivatives were mixed with 1.4 X moles of 14C- in 0.1 ml of standard buffer, and then 0.4 mg of ribosomes were added. After 5 min of incubation, samples were filtered, washed, and counted as described earlier. RESULTS AND DISCUSSION To determine if 2'-esters of are active it is necessary to first know the rate of ester hydrolysis, since base released during the determination of the bactericidal or bacteriostatic activity of the esters will inhibit growth of the test organism. Hydrolysis rates were determined by direct analysis of ester-base ratios. Determination of the rate of hydrolysis at constant ph in the range 6.5 to 7.5 by titration was not used, because the change in basicity of the ester on hydrolysis results in artificially low rate constants. In Tables 1 and 2 different esters are listed in order of hydrolysis rates as determined by both thin-layer chromatographic and bioautographic procedures. Esters can be listed in order of in- M < z. I- 1.0 / ETHYL CARBONYL / ERYTHROMYCIN SENZOYL ERYTHOMYCIN U CONCENTRATION (MICROGRAMS/m.) FIG. 3. Inhibition of growth of S. aureus 209P by and certain of its 2'-esters at different concentrations. Per cent inhibition at 2 hr after the addition of drug is plotted as a function of drug concentration. 0 3 FIG. 4. Bactericidal activity of and certain of its 2'-esters on S. aureus Smith. creasing stability as follows: 2'-ethyl succinyl, 2'-acetyl, 2'-propionyl, 2'-ethyl carbonyl, 2'-benzoyl. As expected, esters are hydrolyzed more rapidly at higher ph values. Rates in BHI broth are similar to those in 0.1 M phosphate buffer at similar ph values. However, rates in human serum are approximately 30 to 50% of those observed in phosphate buffer under similar conditions, suggesting that 2-esters of are stabilized in human serum possibly by binding to serum protein. Typical inhibition curves obtained when different 2'-esters of and were compared by the bacteriostatic procedure described are shown in Fig. 2. In Fig. 3 the apparent per cent inhibition after 2 hr of exposure to antibiotic is plotted as a function of concentration. As can be seen from Fig. 2, apparent inhibition of growth is not complete even at the highest concentration of studied (2.5 ug/ml), at which time has been shown to be bactericidal (Fig. 4). This incomplete inhibition probably reflects an increase in absorbance due to increase in the size of the bacterial cells even though cell division has ceased. It is evident from the data in Fig. 2 and 3 that 2'-esters are less active than and that the apparent activity decreases in the order:

6 164 TARDREW, MAO, AND KENNEY APPL. MICROBIOL. TABLE 3. Relative antibacterial activities of and its 2'-esters Compund ~ 6~iPercePt E (base) Hydrolysis Compound inhibi- equivalent" half-life '-Ethyl Succinyl '-Propionyl '-Ethyl Carbonyl '-Benzoyl Of S. aureus 209P after 2 hr at 1.0 pg/ml. 6 E(base) equivalent = concentration of base in micrograms per milliliter causing a percentage inhibition of growth of S. aureus 209P equivalent to 1.0 jug/ml of the 2'-ester. TABLE 4. Effect of and 2'-propionyl on protein synthesis Compound Concn Incorpurated Incorporate Inoprtu min 14C-Phenylg alanine Incorporated Radio- Inhi- Radioactivity bition activity Inhibitioi mi u % counts/ % min mins Control 8,780 9, , , , , , , '-Propionyl 1 8, '-Propionyl 0.1 5, , '-Propionyl , , '-Propionyl ,530 14, 2'-ethyl succinyl, 2'-propionyl, 2'-ethyl carbonyl, 2'-benzoyl. Thus apparent activity decreases with increased stability of the ester to hydrolysis. Also, with, inhibition of growth is observed almost immediately after addition of antibiotic to the culture, whereas, with the esters, inhibition is delayed. The time of delay increases with different esters in the order: 2'-ethyl succinyl, 2'-propionyl, 2'-ethyl carbonyl, 2'-benzoyl. This effect is most marked at a concentration of 2.5 TABLE 5. Displacement of 14C- binding to ribosomes by unlabeled derivatives - 8 U 60- u 40- E Per cent 14C-Erythro- Relamydn displaced by tive unlabeled compounds anti- Unlabeled compounds at concn of bacterial 1.4 X 3.5X 7 X aictv x 10-6 M 10-6 s b B b C b N-Ethyl b Anhydro b N-Oxide b 2'-Acetyl c 2'-Propionyl c 2'-Benzoyl e u ce!2 a 4 Expressed as micrograms per millimole. b Determined by bioassay. c Calculated from the data in Fig RYTHROMYCIN ERYTHROMYCIN C N-ETHYL ERYTHROMYCIN 2-ACETYL ERYTHROMYCIN 2'-PROPIONYL ERYTHROMYCIN 2'-BENZOYL ERYTHROMYCIN ERYTHROMYCIN B RELATIVE ACTIVITY I m*g per p mole) Fio. 5. Displacement of '4C- bound to ribosomes prepared from S. aureus Smith by different derivatives of.,gg/ml. These results are consistent with the hypothesis that 2'-esters of are inactive per se and that activity results from formed by ester hydrolysis. In Table 3, apparent inhibition of S. aureus by different esters at 2 hr and at a concentration of 1.0 jig/ml is compared with the half-times of hydrolysis determined under similar conditions (BHI broth at ph of 6.5 and 37 C). The relative activities of the esters, defined as that concentra- 800

7 VOL. 18, 1969 ANTIBACTERIAL ACI1VITY OF ESTERS 165 tion of base which will produce the same apparent inhibition in 2 hr as 1.0~g per ml of ester, are also shown in Table 3. From these data, 2'-ethyl succinyl appears to be 10% as active as, 2'-propionyl 4%, and 2'-ethyl carbonyl and 2'-benzoyl less than 1%. Under the conditions used in these experiments, 2'-ethyl succinyl is approximately 90% and 2'-propionyl approximately 50% hydrolyzed after 2 hr. This means that the observed inhibition is due in large part to formed by hydrolysis rather than to any intrinsic activity of the 2'-esters. 2'-Benzoyl and 2'-ethyl carbonyl are only slightly hydrolyzed (approximately 22% and 5%, respectively) under these conditions. The low inhibition observed with 2'-ethyl carbonyl and 2'-benzoyl again strongly supports the conclusion that esterification of in the 2'-position causes loss of antibacterial activity. To confirm these conclusions the bactericidal activities of and its ethyl succinyl, propionyl, ethyl carbonyl, and benzoyl esters were determined. Results are summarized in Fig. 4. As with the bacteriostatic activities, the bactericidal activities of the esters as measured by this technique are a function of the hydrolysis rates of the esters, and these esters appear to possess little if any bactericidal activity until hydrolyzed. Since the primary action of on susceptible organisms is inhibition of protein synthesis (4, 7), the ability of 2'-propionyl to inhibit protein synthesis was compared with that of in the cell-free system derived from S. aureus (Table 4). Although it seems that 2'-propionyl can inhibit protein synthesis, the inhibition is substantially lower than with. However, the protein synthesis assay was performed at 37 C for 30 min in buffer at a ph of 7.5. Again appreciable amounts of 2'-propionyl are hydrolyzed under these conditions, suggesting that the inhibition observed with 2'-propionyl is due entirely to the produced. Recently, it has been reported that forms a 1:1 complex with bacterial ribosomes (3). Also, evidence indicates that binding of to ribosomes is a prerequisite for the inhibition of protein synthesis (8). Several derivatives were tested for their ability to reduce '4C- binding to ribosomes at a concentration equal to and at 2.5 and 5 times the concentration of 4C-. A correlation between antibacterial activity and inhibition of the 14C- binding is clearly shown in Table 5. Compounds with high antibacterial activity reduce 14C- binding significantly, whereas compounds with low antibacterial activity are unable to inhibit 14C- binding or inhibit it only slightly. If per cent of displacement of 14C- is plotted against relative antibacterial activity (Fig. 5), it is possible to estimate the relative antibacterial activity of these three esters. By this method, the estimated activities are 60, 35, and 15,g/,umole for 2'-acetyl, 2'-propionyl, and 2'-benzoyl, respectively. In summary, then, esterification of in the 2' position causes substantial reduction in, and probably complete loss of, antibacterial activity. This loss of activity apparently results from the inability of the 2'-esters, in contrast to, to combine with bacterial ribosomes and thus inhibit synthesis of protein by the affected organism. This conclusion has considerable significance in relation to the use of these derivatives of in the chemotherapy of bacterial disease. ACKNOWLEDGMENTS We are grateful for the support of various personnel in the Scientific Divisions, Abbott Laboratories, particularly L. F. Graham, Jr., who performed most of the analytical work. LMRATURE CITED 1. Kavanagh, F Analytical microbiology, p Academic Press Inc., New York. 2. Mao, J. C.-H Protein synthesis in a cell-free extract from Staphylococcus aureus. J. Bacteriol 94: Mao, J. C.-H The stoichiometry of binding to ribosomal particles of Staphylococcus aureus. Biochem. Pharmacol Mao, J. C.-H., and R. G. Wiegand Mode of action on macrolides. Biochim. Biophys. Acta 157: Murphy, H. W Esters of. Antibiot. Ann , p Pettinga, C. W., W. M. Stark, and F. R. Van Abeele The isolation of a second crystalline antibiotic from Streptomyces erythreus. J. Amer. Chem. Soc. 76: Taubman, S. B., A. G. So, F. E. Young, E. W. Davie, and J. W. Corcoran Effect of on protein biosynthesis in Bacillus subtilis. Antimicrobial Agents and Chemotherapy-1963, p Wilhelm, J. M., N. L. Oleinick, and J. S. Corcoran Interaction of antibiotics with ribosomes: structure-function relationships and a possible common mechanism for the antibacterial action of the macrolides and lincomycin. Antimicrobial Agents and Chemotherapy-1967, p