INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

For the last 5 decades, antibiotics have revolutionized medicine by providing cures for life-threatening infections (2). However, the 1990s has been a period of growing anxiety about the emergence and management of antibiotic-resistant bacteria (1, 4, 10-12). The need to discover and develop more effective antibiotics with unique modes of action is greater than ever. With this goal, we have continued to screen diverse groups of organisms for antimicrobial activity, using a variety of conventional and target-based screening methods. In this paper, we report biological and mechanistic studies of two novel heptadecaglycoside antibiotics, saccharomicins A and B (Fig. 1), which were isolated and characterized from culture broths of the rare actinomycete Saccharothrix espanaensis LL-C19004 (5, 6). This culture (NRRL 15764) was originally isolated from a soil sample collected in Puerto Liano, Spain.

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Structure of saccharomicins A (1) and B (2). Abbreviations: Agl, aglycon; Fuc, fucose; Sac, saccharosamine; Rha, rhamnose; Eva, 4-epi-vancosamine; Dig, digitoxose.

(A preliminary report of these studies was presented at the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy held in San Francisco, Calif., 1999.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains. Clinical isolates were collected from various medical centers in the United States, and quality control strains were obtained from the American Type Culture Collection (ATCC), Rockville, Md. Bacillus subtilis trpC2 BGSC1A1 and Escherichia coli imp BAS849 were obtained from the Bacillus Genetic Stock Center, Columbus, Ohio, and S. A. Benson (13), respectively. Identification of each culture was done by conventional methods: gram-negative rods by API 20E (Analytab Products, Plainview, N.Y.) and NF systems (Remel, Lenexa, Kans.) and staphylococci by Staph Trac (Analytab Products). All clinical isolates were stored frozen in skim milk at 70°C, and other strains were stored frozen at 80°C in 20% dimethyl sulfoxide.

Media. All media were prepared in distilled deionized (DI) water. Mueller-Hinton (MH) media was purchased from Becton Dickinson Microbiology Systems, Cockeysville, Md. Modified minimal medium (used for mechanistic studies in E. coli imp) contained the following (per liter): dextrose, 4 g; NH₄Cl, 1 g; KH₂PO₄, 3 g; Na₂HPO₄, 6 g; MgSO₄ · 7H₂O, 0.25 g; FeSO₄ · 7H₂O, 0.5 mg; and vitamin-free Casamino Acids, 2 g. ATCC medium 21 (K₂HPO₄, 0.5 g; ferric ammonium citrate, 0.5 g; MgSO₄ · 7H₂O, 0.5 g; glycerol, 20 g; citric acid, 2 g; L-glutamic acid, 4 g per liter [pH 7.4]), supplemented with 20 mg of L-tryptophan/liter, was used for studies with B. subtilis. All ingredients in the minimal medium were purchased from Sigma Chemical Co., St. Louis, Mo., with the exception of the Casamino Acids, which were purchased from Difco Laboratories, Detroit, Mich.

Chemicals. [³H]thymidine ([³H]TdR, TRK.686, 90 Ci/mmol), [³H]uridine ([³H]UdR, TRK.410, 49 Ci/mmol), and ³H-labeled amino acids (TRK.550; a mixture of leucine, lysine, phenylalanine, proline, and tyrosine, with specific activities of 135, 83, 123, 103, and 118 Ci/mmol, respectively) were purchased from Amersham Corp., Arlington Heights, Ill. All reference antimicrobial agents and all other medium components were purchased from Sigma Chemical Co.

In vitro antibacterial activity. The in vitro antibacterial activities were determined by agar or microdilution methods as described earlier (9, 14). MH II broth or agar was used for nonfastidious aerobic bacteria, and the medium was supplemented with 5% lysed horse blood for Streptococcus spp. The inoculum of each culture was approximately 1 × 10⁴ to 5 × 10⁴ CFU applied with the Steers multiple-inocula replicator to plates containing the antibiotic in MH agar. Wilkins-Chalgren agar was used for the anaerobes. The plates were incubated at 35°C for 18 h. The agar MIC was defined as the lowest concentration of antimicrobial agent that completely inhibited visible growth of the organism. Broth MICs were determined by adding 5 µl of an overnight broth bacterial culture (adjusted to a density of 1 × 10⁷ to 5 × 10⁷ CFU/ml) to 0.1 ml of broth medium in polystyrene plates containing the drug at 0.03 to 128 µg/ml. The broth MIC was defined as the lowest concentration of antibiotic which prevented visual turbidity after 18 h of incubation at 37°C. Since mechanistic studies were done with exponential-phase cultures with an absorbance at 600 nm (A₆₀₀) of 0.20, the growth-inhibitory concentration of each drug was also determined at this inoculum level by measuring the absorbance of drug-treated cultures after 3 h of incubation at 37°C with agitation. The minimum concentration of drug preventing any increase in the absorbance was called the MIC_3h. MH broth was supplemented with the desired levels of divalent cations to study their effects on the antibacterial activity. Calcium chloride, magnesium chloride, ferrous sulfate, and copper sulfate were used as the divalent cations.

Bactericidal activity (kill kinetics). The test organism was grown overnight, transferred to fresh broth, and incubated at 35°C and 200 rpm to logarithmic phase. The culture (50 ml) was then added to a 250-ml Erlenmeyer flask containing 50 ml of antibiotic solution in broth, yielding a final concentration of 10⁶ CFU/ml and an antibiotic concentration of 4 or 10 times the MIC. The flasks were incubated for 24 h with shaking at 35°C. Viable cell counts were performed at various time points.

In vivo activity. In vivo activity was assessed in female strain CD-1 mice (Charles River Laboratories, Kingston, N.Y.), weighing 20 ± 2 g each, which were infected intraperitoneally with sufficient bacterial cells suspended in broth or with about 5% mucin to kill 95 to 100% of the untreated mice within 48 h. Antibiotic was administered in single doses 0.5 h after infection. Seven-day survival ratios from three or four separate tests, each with five dose levels and five animals per dose level, were pooled for the determination of the median effective dose (ED₅₀) by probit analysis.

Incorporation of radiolabeled precursors. Macromolecular synthesis in E. coli imp and B. subtilis was studied by measuring the incorporation of appropriate radiolabeled precursors into trichloroacetic acid (TCA)-precipitable material (14). Overnight cultures were diluted 1:500 in fresh modified minimal medium (50 ml of medium in a 250-ml Erlenmeyer flask) and incubated at 37°C and 200 rpm to an A₆₀₀ of 0.20. Aliquots of 100 ml were dispensed into microtiter wells containing 5 µl of antibiotic, and plates were incubated for 2 to 14 min at 37°C with vigorous agitation. Cells were pulse-labeled for 5 min by adding one of the following radiolabeled precursors at the indicated final concentrations: for E. coli imp, [³H]TdR (1 µCi/ml, with 0.05 µg of unlabeled thymidine/ml), [³H]UdR (1 µCi/ml, with 0.12 µg of unlabeled uridine/ml), or ³H-labeled amino acids (10 µCi/ml); for B. subtilis, [³H]TdR (2 µCi/ml), [³H]UdR (2 µCi/ml, with 0.35 µg of unlabeled uridine/ml), or ³H-labeled amino acid mixture (2 µCi/ml, with 0.035 µg each of leucine, lysine, phenylalanine, proline, and tyrosine/ml). To determine specific incorporation into DNA, RNA, and protein, 100 µl of chilled (4°C) TCA (10%) supplemented with 0.5 mg of unlabeled precursors per ml was added to each well, and the plate was immediately refrigerated for 1 h. The precipitate was collected on a glass fiber filter (Wallac filtermat B, Wallac 1205-404) using a Skatron 96-well cell harvester (model 11050) programmed for a 3-s prewet with chilled DI water, a 12-s wash with 5% chilled TCA, and a 5-s drying cycle. To assess the effects of the drugs on cellular uptake of radiolabeled precursors, the step with addition of TCA to the microtiter plate was eliminated, and the contents of each well were harvested onto a glass fiber filter by the Skatron 96-well cell harvester programmed for a 3-s prewet with chilled DI water, a 10-s wash with chilled normal saline (0.9% NaCl in DI water), and a 5-s drying cycle. Filter mats were dried for 7 min at high power in a microwave oven (Quasar, 700 W), solid scintillant (Meltilex B/HS, Wallac 1205-442) was applied, and the isotope retained on the filter was quantitated in an LKB Betaplate scintillation counter (Wallac 1205). The levels of incorporation of [³H]TdR, [³H]UdR, and ³H-labeled amino acids are expressed as the percentage relative to the untreated control.

Effect on intracellular potassium of E. coli imp. Effects on intracellular potassium in E. coli imp were studied in a saline buffer (10 mM HEPES buffer containing 150 mM NaCl and 0.1 mM KCl, pH 7.0). Each log-phase culture was washed twice with saline buffer, and the pellet was resuspended in the same buffer to an optical density at 600 nm of 2.00. One milliliter of the bacterial suspension was treated with the test compounds at various concentrations for 1 h, and cells were pelleted by centrifugation (at 10,000 × g for 2 min). The resulting supernatant was diluted 1:10 in high-pressure liquid chromatography-grade water and analyzed for potassium ion by atomic absorption spectrophotometry (Instrumentation Laboratories 551). For determination of the total potassium level, 1 ml of the culture was hydrolyzed in 2 M sulfuric acid by heat (100°C for 1 h), chilled for 1 h, and centrifuged (at 10,000 × g for 2 min). The supernatant was then diluted 1:10 and analyzed for potassium ion concentration.

Lysis of human RBCs. One milliliter of freshly pooled human blood was centrifuged (at 10,000 × g for 2 min), the pellet was washed four times with normal saline by repeated resuspension and centrifugation, and then the pellet was resuspended in 1 ml of red blood cell (RBC) buffer (10 mM Na₂HPO₄-150 mM NaCl-1 mM MgCl₂, pH 7.4). Twenty-five microliters of the RBC suspension was added to a microcentrifuge tube containing 1 ml of drug solution (final concentration ranging from 1 to 128 µg/ml) prepared in duplicate in RBC buffer. After 2 h of treatment, the contents of the tube was centrifuged (at 10,000 × g for 2 min) and absorbance (at 540 nm) of the supernatant was measured. For 100% lysis, 25 µl of the RBC suspension was added to 1 ml of water.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

A systematic reevaluation of our antibiotic compound library of the last 5 decades provided some extracts and complexes with very interesting bioactivity and novel chemistry. LL-C19004 was one such complex, produced by the rare actinomycete Saccharothrix espanaensis. Further characterization of this complex led to the discovery of two novel, high-molecular-mass (2,800 Da) heptadecaglycoside antibiotics, saccharomicins A and B (Fig. 1).

In vitro activity. The antibacterial activity of saccharomicins against a diverse group of clinical isolates is presented in Tables 1 to 3. Although saccharomicins A and B are slightly different in chemical structure, both were mostly equipotent against gram-positive and gram-negative bacteria (Table 1). They exhibited good activity against gram-positive organisms and good to moderate activity against gram-negative organisms. The MICs for gram-negative bacteria were 8 to 16 times higher in the medium supplemented with divalent cations, which suggested some sort of an antagonistic effect of cations on this antibiotic (Table 3). Saccharomicins exhibited good activity against methicillin-susceptible and methicillin-resistant staphylococci, with the majority of the MICs being between <0.12 and 0.5 µg/ml. Over 50% of the methicillin-resistant staphylococci tested were also resistant to minocycline, gentamicin, and erythromycin, and several staphylococcal isolates were multiresistant to antibiotics like -lactams, tetracyclines, streptomycin, tobramycin, chloramphenicol, quinolones, and rifampin (data not shown). All of these antibiotic-resistant isolates were susceptible to saccharomicins (MIC range, <0.12 to 0.5 µg/ml). Saccharomicins were slightly less active against enterococcus species (MIC range, 0.25 to 16 µg/ml), but they were equally active against vancomycin-resistant and -susceptible isolates (Tables 1 and 2). These data on activity of saccharomicins suggest that cross-resistance to other classes of antibiotics would not be expected. The presence of 5% lysed horse blood in the medium did not increase the MICs for the staphylococcal and streptococcal strains tested. Generally, antibiotics with strong protein binding properties have much higher MICs under such conditions. Activity against gram-negative bacteria was good to moderate in cation-deficient medium, but the presence of divalent cations drastically affected the activity (Table 3). The decreased activity against gram-negative bacteria may also be due to the presence of the outer membrane, since saccharomicins exhibited two- to eightfold-higher activity against an E. coli imp strain (Tables 1 and 3). The imp mutation increases the permeability of the cell membrane and renders the bacteria much more susceptible to many antibacterial agents, especially larger molecules (10). These antibiotics were inactive against Candida albicans. Although more active against gram-positive bacteria, saccharomicin A was bactericidal for both gram-positive and gram-negative organisms. Killing of Staphylococcus aureus Smith was more rapid with saccharomicin A than with vancomycin (Fig. 2). Killing of E. coli imp was more rapid in stationary-phase cells suspended in an isotonic saline buffer than in those in minimal medium (Fig. 3). This reduced bactericidal activity of saccharomicin A in medium containing magnesium and possible sources of calcium suggested that divalent cations may diminish the activity of the antibiotic.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Kill kinetics of S. aureus Smith.

View larger version (128K): [in this window] [in a new window]   FIG. 3.   Kill kinetics of E. coli imp under stationary (left; cells suspended in saline buffer) and growing (right; cells in minimal medium) conditions. OD, optical density.

Reversal of activity by divalent cations. MICs of the saccharomicins for all gram-positive bacteria and E. coli imp were unaffected by the presence of divalent cations (Ca²⁺ and Mg²⁺) in the medium, but 8- and 16-fold increases in MICs were observed for E. coli and P. aeruginosa, respectively (Table 3). Growth inhibition of P. aeruginosa by saccharomicin A was neutralized by calcium and magnesium ions in a concentration-dependent manner (Table 4). The control antibiotic, polymyxin B, behaved similarly. Calcium ion was about twofold more effective than magnesium ion in neutralizing the activity. The concentration needed to neutralize the antibacterial activity of saccharomicin A also depended upon bacterial inoculum density and the concentration of the drug used. Higher inoculum density required a lower cation concentration for neutralization, and higher drug concentrations needed higher cation concentrations for the reversal of antibacterial activity (data not shown). Fe²⁺ and Cu²⁺ had no effect (data not shown). These data suggest that the saccharomicins do not act as general chelators like EDTA, but it is more likely that they interact with outer membrane-bound cations of gram-negative bacteria, which contribute to the maintenance of stability of this outermost layer of the cell wall. By this mechanism, the high-molecular-weight saccharomicins may be expected to self-promote their uptake into these organisms in a manner similar to polymyxin B, aminoglycosides, and quinolones (3, 7, 8). The self-promoted uptake of these antibiotics involves an interaction with the binding sites of the divalent cations on the lipopolysaccharide, and the addition of calcium and magnesium ions counteracts the displacement of the cations by these antibiotics. Binding of the saccharomicins to such sites on the lipopolysaccharide has not been established, but the observed effect of divalent cations upon antibacterial activity may involve a similar mechanism. This hypothesis is further supported by the fact that an E. coli imp mutant which is deficient in this permeability barrier was two- to eightfold more sensitive to saccharomicins than the wild-type strains, and the addition of cations did not interfere with the susceptibility of this organism (Table 3).

Incorporation of radiolabeled precursors. Inhibition of DNA, RNA, and protein syntheses was determined by measuring the incorporation of [³H]TdR, [³H]UdR, and ³H-labeled amino acids, respectively, into the TCA-precipitable material of log-phase cultures of E. coli imp and B. subtilis. The effects of drugs on the cellular uptake of radiolabeled precursors were determined by measuring radioactivity retained in saline-washed cells under the same experimental conditions. For each drug tested at one to two times the MIC_3h, uptake of the three radiolabeled precursors was unaffected relative to the specific inhibition of incorporation into TCA-precipitable material. Since both saccharomicins were equally active in MIC tests, saccharomicin A was used for mechanistic studies in bacteria. Within 5 min of saccharomicin A treatment at 4 µg/ml (twice the MIC_3h), all three macromolecular syntheses were shut down in E. coli imp (Table 5). At 2 µg/ml (MIC_3h), uptake and incorporation of amino acids appeared to be more affected than those of the other two precursors. During the same period, three control drugs, ciprofloxacin, rifampin, and chloramphenicol, predominantly inhibited DNA, RNA, and protein synthesis, respectively. Polymyxin B inhibited uptake and incorporation of all three precursors tested.

View larger version (107K): [in this window] [in a new window]   FIG. 4.   Effects on macromolecular synthesis in an exponential-phase culture of B. subtilis treated with 0.03, 0.06 (MIC), and 0.12 µg of saccharomicin A/ml. Data are percentages of incorporation relative to untreated control for 2 to 14 min of antibiotic treatment and 5 min of pulse-labeling with [3H]TdR, [3H]UdR, and 3H-labeled amino acids (AA).

Effects on prokaryotic and eukaryotic membranes. The membrane-damaging effects of the antibiotics on E. coli imp and human RBCs were studied by measuring the leakage of intracellular potassium and hemolysis, respectively. Saccharomicin A did not cause hemolysis of RBCs, but it led to extensive leakage of intracellular potassium from E. coli (Table 6). These data are consistent with the membrane-damaging effects of saccharomicin A in bacteria but not in eukaryotic cells.

In vivo efficacy and toxicity. The ability of saccharomicins to protect mice from lethal challenges by various pathogenic bacteria was compared to that of vancomycin (Table 7). The ED₅₀s for saccharomicin activity against gram-positive infections by subcutaneous administration were in the range of 0.06 to 2.6 mg/kg, which was comparable to the range for vancomycin. The 50% lethal dose (LD₅₀) for the saccharomicins varied from 2 to 360 mg/kg, depending upon the route of administration. Saccharomicins administered orally and subcutaneously protected the mice from a lethal challenge by gram-positive bacteria, but they were unable to protect mice from a lethal challenge of gram-negative bacteria at desirable dosages. The ED₅₀ of saccharomicins against S. aureus Smith by the subcutaneous route was 260 times lower than the LD₅₀. However, the ED₅₀-to-LD₅₀ ratio for other gram-positive organisms ranged from 1:5 to 1:18, a narrow therapeutic window for this antibiotic compared to that of vancomycin.