Antimicrobial Evaluation of Nocathiacins, a Thiazole Peptide Class of Antibiotics

Abstract

Nocathiacins are cyclic thiazolyl peptides with inhibitory activity against gram-positive bacteria. BMS-249524 (nocathiacin I), identified from screening a library of compounds against a multiply antibiotic-resistant Enterococcus faecium strain, was used as a lead chemotype to obtain additional structurally related compounds. The MIC assay results of BMS-249524 and two more water-soluble derivatives, BMS-411886 and BMS-461996, revealed potent in vitro activities against a variety of gram-positive pathogens including methicillin-resistant Staphylococcus aureus, penicillin-resistant Streptococcus pneumoniae, vancomycin intermediate-resistant S. aureus, vancomycin-resistant enterococci, Mycobacterium tuberculosis and Mycobacterium avium. Analysis of killing kinetics revealed that these compounds are bactericidal for S. aureus with at least a 3-log₁₀ reduction of bacterial growth within 6 h of exposure to four times the MICs. Nocathiacin-resistant mutants were characterized by DNA sequence analyses. The mutations mapped to the rplK gene encoding the L11 ribosomal protein in the 50S subunit in a region previously shown to be involved in the binding of related thiazolyl peptide antibiotics. These compounds demonstrated potential for further development as a new class of antibacterial agents with activity against key antibiotic-resistant gram-positive bacterial pathogens.

Resistance rates of bacterial species of clinical importance to existing classes of antibiotics continue to increase (1, 2, 6, 8, 15, 28). Problem organisms include such gram-positive pathogens as staphylococci, pneumococci, and enterococci affecting both community and hospital settings. Many of these strains are resistant to multiple antibiotics, which severely limits treatment options and constitutes a great public health threat worldwide. Few new antibacterials with novel chemotypes have entered the market over the past 30 years, and relatively few new classes of agents are in development (3). Thus, there is an urgent need for the discovery of new chemotypes with antibacterial activity to overcome existing resistance mechanisms and to effectively combat these human pathogens that can cause life-threatening infections.

BMS-249524 (nocathiacin I) (5, 12) is the prototype of a new nosiheptide-class of tricyclic peptide antibiotics called nocathiacins and was identified in a cell-based screen of natural product extracts with a multiply antibiotic-resistant strain of Enterococcus faecium as an indicator strain (14). This natural product lead compound was originally isolated from a Nocardia sp. (ATCC 202099) fermentation broth extraction. Previously reported related compounds in this class include thiostrepton and micrococcin (4, 13, 19, 23, 24, 27, 30) and more recently the thiazole antibiotics MJ347-81F4 A and B (25), GE2270A (9), and GE37468 A (26). The mode of action of the nocathiacins on bacterial protein synthesis is closely related to that of thiostrepton, as these antibiotics bind to the 23S rRNA of the 50S ribosomal subunit at the same site as the L11 ribosomal protein (13, 19-21, 24, 27, 29, 30). This prevents the normal conformational transition that occurs from 23S-L11 interaction and results in a stalling of translation affecting the elongation step of bacterial protein synthesis. These compounds, although possessing antibacterial activity and a unique mechanism of action, thus far have not been developed for use in humans due to pharmacokinetic and solubility issues (18). The nocathiacins described here, prototype BMS-249524 and the more water-soluble derivatives BMS-411886 and BMS-461996, possess potent in vitro activity against gram-positive bacteria, including many key antibiotic-resistant clinical pathogens. These compounds offer the potential for further development as a new class of antibiotics for serious gram-positive infections.

MATERIALS AND METHODS

Bacterial strains and antibiotics.

Bacterial strains used in this work are either clinical isolates from the Bristol-Myers Squibb Culture Collection, designated by an A prefix, or ATCC strains (American Type Culture Collection, Manassas, Va.). BMS-249524 was isolated and purified at Bristol-Myers Squibb (Wallingford, Conn.) (14). BMS-411886 and BMS-461996 were analogs obtained by chemical modification of BMS-249524 (10, 14, 16, 22; B. N. Naidu, M. E. Sorenson, T. P. Connolly, J. Witchtowski, Y. Ueda, J. J. Bronson, Y. Zhang, O. Kim, W. Li, K. S. Lam, M. J. Pucci, J. M. Clark, G. A. Warr, K. L. DenBleyker, T. M. Stickle, D. Taylor, L. M. Lamb, I. A. Medina, R. Macci, S. Venkatesh, L. Fung, J. Alberts, 224th Am. Chem Soc. Nat. Meet., abstr. MEDI-209, 2002). Vancomycin and streptomycin were purchased from Sigma-Aldrich (St. Louis, Mo.). Trovafloxacin and levofloxacin were extracted and purified from commercially available tablets and were determined to be >95% pure by high-performance liquid chromatography. Clarithromycin was a kind gift from Abbott Co. (Abbott Park, Ill.).

In vitro susceptibility testing.

Whole-cell antibacterial activity was determined by broth microdilution according to methods recommended by NCCLS (17). Test compounds were dissolved in dimethyl sulfoxide and diluted 1:10 in water to produce a stock solution at 256 μg/ml. In a 96-well microtiter plate, 50 μl of the stock solution was serially diluted into cation-adjusted Mueller-Hinton broth (Becton-Dickinson, Cockeysville, Md.) except Haemophilus influenzae (Haemophilus test medium). After the compounds were diluted, 50-μl aliquots of the test organism (each, ∼5 × 10⁻⁵ CFU/ml) were added to the appropriate wells of the microtiter plate. Inoculated plates were incubated aerobically at 35°C for 18 to 24 h. The MIC was determined to be the lowest concentration of compound that inhibited visible growth. MICs were found to be equivalent in either Mueller-Hinton broth or brain heart infusion broth (BHI; Difco, Detroit, Mich.). Susceptibility tests for Mycobacterium spp. were performed by a macrodilution method in 7H9 broth (Difco) with a final inoculum concentration of 2 × 10⁷ CFU/ml. Tubes were incubated at 35°C in 5% CO₂. Inhibition of growth was monitored for 3 weeks. The MIC was defined as the lowest concentration of drug that inhibited visible growth after 3 weeks of incubation. All work involving mycobacteria was done under a biological safety cabinet in a high-containment facility. Susceptibility testing of Legionella pneumophila was done by agar (buffered starch yeast extract) dilution and incubated at 35°C for 24 h. The MICs for Chlamydia trachomatis were determined with McCoy (ATCC) and HL cells (Washington Research Foundation, Seattle, Wash.) in a microtiter format. Chlamydial suspensions prepared in a maintenance medium (Eagle's minimal essential medium [Gibco] supplemented with 10% fetal bovine serum) contained approximately 100 to 1,000 inclusion-forming units per ml, with 0.05 ml of the suspension added per well. The plates were centrifuged at 1,000 × g for 60 min. Twofold serial dilutions of compounds in maintenance medium were added (each dilution, 100 μl), and plates were incubated for 48 h in 5% CO₂, fixed with methanol, and stained with 0.02 ml of fluorescein-conjugated monoclonal antibody specific to the chlamydial lipopolysaccharide. The presence of chlamydial inclusion bodies was detected with an inverted fluorescent microscope at 100× magnification. The MIC was defined as the lowest compound concentration with no observable inclusion bodies. MICs for Mycoplasma pneumoniae were performed by broth microdilution where the test medium was PPLO medium (Difco) supplemented with 2.5% yeast extract, 20% horse serum, and 1% dextrose. The inoculum contained 10⁴ to 10⁵ color-changing units (the minimum inoculum required for growth as indicated by a color change in 0.002% phenol red indicator) per ml and 100 μl of the diluted compound in each well. The MIC was defined as the lowest dilution of compound that inhibited growth as indicated by the lack of a color change, relative to the growth control with color change (7).

In vitro killing curve assay.

The bacteriostatic and bactericidal effects of BMS-249524, BMS-411886, and BMS-461996 against Staphylococcus aureus A27223, a homogeneous methicillin-resistant clinical isolate, were determined by a time-kill assay. Cultures were grown overnight at 35°C in BHI broth with aeration, diluted into fresh BHI, and allowed to grow for several generations at 35°C with aeration until the cultures reached approximately 10⁶ CFU/ml. At this point, 5-ml aliquots of culture were removed, and antibiotics were added at one, two, and four times the MIC, and the cultures were further incubated at 35°C with aeration. Samples of 100 μl (each) were withdrawn at 30-min intervals for 8 h, diluted, and plated in duplicate on BHI agar plates. Plates were incubated for 24 h at 35°C, and colonies on each plate were counted. The CFU per ml were calculated for each time interval at each drug concentration.

Mutation analyses.

Mutants were isolated on BHI agar containing compound at two to four times the MIC. Mutations were determined by PCR cloning of the gene encoding the L11 ribosomal protein, followed by DNA sequence analyses of the PCR fragments. Chromosomal DNA was isolated from 2 ml of overnight culture grown in BHI broth. Cells were pelleted and washed with TES buffer (50 mM EDTA, 50 mM NaCl [pH 8.0]) and resuspended in 1 ml of TES. Lysostaphin (Sigma) was added to a final concentration of 20 μg/ml, RNase A (Sigma) was added to a final concentration of 200 μg/ml, and cells were incubated at 37°C for 60 min. This was followed by the addition of STEP buffer (0.5% sodium dodecyl sulfate, 50 mM Tris, 0.4 M EDTA, 1 mg of proteinase K [Sigma]/ml) and further incubation at 40°C for 60 min. Samples were then extracted twice with each of the following: phenol, phenol-chloroform, and chloroform. A 1/10 volume of 3 M sodium acetate and 2 volumes of ethanol were added to precipitate the chromosomal DNA. The DNA was spooled out with a glass rod and dissolved in water. The following primers flanking the staphylococcal L11 gene were used to PCR clone the gene: L11-R (5′ AGTTAAGAGCAGACAACAGAAG 3′) and L11-L (5′ AGTGTTAAAATTATGTGGTCGCG 3′). Vent polymerase (New England Biolabs, Beverly, Mass.) was used to amplify the L11 sequences. The PCR products were purified with a Qiaquick PCR purification kit (QIAGEN, Valencia, Calif.) and sequenced with ABIprism BigDye (Perkin-Elmer, Wellesley, Mass.). DNA traces were analyzed with Lasergene software (DNASTAR, Madison, Wis.).

RESULTS

In vitro potency.

The MICs of BMS-249524, BMS-411886, and BMS-461996 (Fig. 1) for several bacterial species were determined (Table 1). These compounds were found to be very potent (MICs ranging from 0.01 to 0.1 μg/ml) for gram-positive pathogens including representative isolates of penicillin-resistant Streptococcus pneumoniae, vancomycin-resistant E. faecium, and methicillin-resistant S. aureus. These MICs were minimally affected by human serum. In contrast, vancomycin MICs were 0.25 to 4.0 μg/ml for the same bacterial strains (64 μg/ml for the VRE strain). Excellent potency was observed against vancomycin-intermediate-resistant S. aureus strains, for which MICs were equivalent to those observed for susceptible strains (0.007 μg/ml) (Table 2). These compounds were also active against gram-positive anaerobes such as Clostridium difficile and Clostridium perfringens (Table 3), for which MICs compared favorably with MICs of the fluoroquinolone trovafloxacin. BMS-249524 compared well with linezolid, clarithromycin, streptomycin, and levofloxacin for activity against Mycobacterium tuberculosis and Mycobacterium avium (Table 4). Similarly, BMS-411886 and BMS-461996 were very active against L. pneumophila and M. pneumoniae, compared to linezolid and clarithromycin (Table 5), and BMS-249524 had good activity against C. trachomatis (MICs, <0.005 μg/ml) (Table 6). These nocathiacins were not inhibitory against gram-negative organisms, including H. influenzae, Bacteroides fragilis, and Escherichia coli (Table 1).

FIG. 1.

Chemical structures of the nocathiacins discussed in this work.

TABLE 1.

MICs of nocathiacins and vancomycin for selected bacterial pathogens

Organism (resistance)a	Strain	MIC (μg/ml)
		BMS-249524	BMS-411886	BMS-461996	Vancomycin
S. pneumoniae	A9585	0.0005	0.015	0.003	0.25
S. pneumoniae (PenI)	A27881	0.0005	0.015	0.003	0.25
S. pneumoniae (PenR)	A28272	0.0005	0.015	0.003	0.5
Enterococcus faecalis	A20688	0.03	0.03	0.03	1
E. faecalis + seruma	A20688	0.06	0.125	0.125	4
E. faecium	A24885	0.03	0.03	0.03	0.5
E. faecium (VanR)	A28142	0.03	0.015	0.03	64
S. aureus (Pen+)	A15090	0.007	0.007	0.015	1
S. aureus + serum	A15090	0.015	0.015	0.06	2
S. aureus (MRSA)	A27223	0.007	0.007	0.06	1
S. aureus + serum	A27223	0.015	0.015	0.06	1
Staphylococcus epidermidis	A24548	0.015	0.015	0.007	2
Staphylococcus haemolyticus	A27298	0.015	0.015	0.06	1
Moraxella catarrhalis (Pen+)	A22344	0.5	0.5	0.5	32
H. influenzae (Pen−)	A20191	>16	>64	>64	>64
H. influenzae (Pen+)	A21515	>16	>64	>64	>64
E. coli	A29522	>16	>64	>64	>64
B. fragilis	A27850	>64	>64	>64	>64

^aAbbreviations: PenI, penicillin intermediate resistant; PenR, penicillin resistant; VanR, vancomycin resistant; Pen+, penicillinase-producing; Pen−, no penicillinase; MRSA, methicillin-resistant; serum, 50% human serum.

TABLE 2.

MICs of BMS-249524 and vancomycin for vancomycin-intermediate resistant S. aureus

Strain	MIC (μg/ml)
	BMS-249524	Vancomycin
A24407 (susceptible)	0.007	1
A29342	0.007	4
A29510	0.007	4
A29511	0.007	8
A29512	0.007	8

TABLE 3.

MICs of BMS-259524 and trovafloxacin for selected anaerobic bacteria

Organism (no. of strains)	MIC (μg/ml)
	BMS-259524	Trovafloxacin
C. difficile (5)	0.06-1	0.25-1
C. perfringens (5)	0.13-0.25	0.13-0.25
Peptostreptococcus spp. (5)	0.06-0.13	0.06-0.25
Eubacterium lentium (2)	0.015	1
Bacteroides spp. (1)	>8	0.25-0.5

TABLE 4.

MICs of BMS-249524 and other drugs for mycobacteria

Organism	Strain	MIC (μg/ml)
		BMS- 249524	Line- zolid	Clari- thromycin	Strepto- mycin	Levo- floxacin
M. tuberculosis	ATCC 35828	≤0.008	0.5	1	≤0.25	0.5
M. avium	A26778	0.06	4	≤0.25	16	1
M. avium	A26640	0.25	2	≤0.25	4	2

TABLE 5.

MIC of nocathiacins and other drugs for M. pneumoniae and L. pneumoniae

Organism	MIC (μg/ml)
	BMS-411886	BMS-461996	Linezolid	Clarithromycin
M. pneumoniaea	0.06	0.5	>32	0.015
L. pneumophilab	0.17	0.02	2	0.01

^aModel MICs, based on four strains.

^bMIC at which 90% of isolates tested were based on 12 strains.

TABLE 6.

MICs of nocathiacins and other drugs for C. trachomatis^a

C. trachomatis strain	MIC (μg/ml)
	BMS-249524	Levofloxacin	Ciprofloxacin
CH3025	0.001	0.25	1
887VR	0.004	0.25	1
10882	0.001	0.25	0.5
11073	0.002	0.25	1

^aMICs were determined in McCoy cell lines (7).

In vitro killing curve assay.

To determine the killing capability of the nocathiacins, killing curve assays were conducted. As demonstrated in the results shown in Fig. 2, BMS-249524 showed killing kinetics to similar to that of vancomycin against S. aureus A27223 with approximately a 3-log₁₀ decrease in CFU per ml over 6 h at concentrations of two- and four times the MIC. Similar results were seen with other nocathiacins (data not shown). This was in contrast to killing kinetics seen with the bacteriostatic antibiotic erythromycin, where a maximum decrease of only 1 log₁₀ was observed at 6 h at concentrations of two- and four times the MIC. The nocathiacins inhibited the growth of enterococci and were essentially bacteriostatic (data not shown).

FIG. 2.

Time-kill curves of nocathiacin-treated S. aureus A27223. Experimental details are found in Materials and Methods. Addition of compound BMS-249524 was at time zero. Symbols: ▪, untreated control culture; ♦, one times the MIC (0.01 μg/ml); ▾, two times the MIC (0.02 μg/ml); ▴, four times the MIC (0.04 μg/ml).

Mutation analysis.

Sequence analyses of several S. aureus mutants, selected on medium containing nocathiacins at two times and four times the MICs, revealed mutations in the rplK gene encoding the L11 ribosomal protein, consisting of base pair changes, insertions, and deletions (Table 7). Some of the same mutations were discovered in additional staphylococcal strains as well as enterococcal strains in different experiments. The mutants listed here were obtained at a frequency of 10⁻⁹; however, mutation frequency ranges of 10⁻⁷ to 10⁻⁹ have been observed. Mutant colonies were generally smaller in size and found to have reduced growth rates, as much as twofold lower than the parental strains with doubling times of ∼60 min in BHI broth (data not shown). These mutations were localized near several key proline residues present in the antibiotic binding cleft of the L11-23S rRNA region of the 50S ribosome (Fig. 3) in a region previously shown to be involved in thiostrepton binding (20, 29).

TABLE 7.

RplK mutations found in nocathiacin-resistant mutants

S. aureus Strain	DNA change	L11 protein changea
A15090 and A27223	None (wild type)	P(20)APPVGPA(27)
	3-bp deletion	P(20)AP-VGPA(27)
	12-bp deletion	P(20)AP-----A(27)
A27223	None (wild type)	K(87)GSGEPNKTKVAT(99)
	1-bp deletion	K(87)GSGEPNKTKVLQ(99) STOP
	2-bp deletion	K(87)GSGEPNKTKVLQ(99) STOP
	None (wild type)	K(71)TPPAPVLLK(80)
	1-bp change	K(71)TPPAPVLL(79) STOP
A27223 and A24407	57-bp deletion	G(32)---------L(52)
A24407	None (wild type)	K(87)GSGEPNKTKVAT(99)
	1-bp insertion	K(87)RFRRTKQN(95)

^aSTOP, stop codon (protein termination).

FIG. 3.

Putative nocathiacin binding site on the 50S E. coli ribosome. Labeled amino acid residues and nucleotides indicate previously reported mutations that confer thiostrepton and micrococcin resistance. Mutations are clustered around a cleft (arrow) between RNA (left) and a proline-rich helix in the L11 N terminus (29). The figure was drawn with InsightII (Accelrys, San Diego, Calif.) with identification of secondary structures as previously described (11).

DISCUSSION

Infections caused by antibiotic-resistant bacteria are a significant problem in the world today, particularly in regard to nosocomial infections. In addition, resistance rates for community-acquired bacterial pathogens have also been on the increase in recent years (1, 3, 28). One way to combat this continuous threat of bacterial resistance is to discover and develop new antibacterial agents, especially those possessing novel chemotypes. These compounds would be less likely to encounter cross-resistance with existing antibiotics in clinical use today. The nocathiacin class described here has the potential to become such a valuable new weapon in our antibacterial arsenal.

The initial compound, BMS-249524, discovered from a screen against a multiply antibiotic-resistant E. faecium strain, was found to be extremely potent against gram-positive pathogens, including several antibiotic-resistant isolates. The aqueous solubility of BMS-249524 was found to be insufficient for use as an intravenous agent in humans. Therefore, an effort was made to investigate modifications of this class of compounds that would maintain the biological activity of nocathiacins while increasing their water solubility (Naidu et al., 224th Am. Chem. Soc. Nat. Meet.; 10). BMS-411886 and BMS-461996 represent examples of such compounds that display improved aqueous solubility profile while retaining good antibacterial activity.

All three nocathiacins described here show improved in vitro potency compared with vancomycin, currently an important antibiotic for the treatment of nosocomial gram-positive bacterial infections. Beginning in the 1980s, increased use of vancomycin resulted in the selection of resistant strains, initially of enterococci but recently of some S. aureus strains (28). The nocathiacins compared favorably in vitro with other classes of antibiotics including oxazolidinones, macrolides, fluoroquinolones, and β-lactams against various pathogens including Mycobacteria spp., L. pneumophila, M. pneumoniae, and gram-positive anaerobes. Although most bacterial protein synthesis inhibitors described to date are bacteriostatic against staphylococci, these nocathiacins displayed bactericidal activity against staphylococci. However, as with many other antibacterials, bacteriostatic activity was observed against enterococci.

Nocathiacin-resistant staphylococcal and enterococcal mutants obtained were single-step, high-level resistant organisms. DNA sequence analyses revealed several different types of mutations that mapped to the rplK gene encoding the ribosomal L11 protein in the region of several key proline residues in the L11 binding domain, consistent with a region shown to be involved in thiostrepton binding (20, 29). These mutants were also cross-resistant with thiostrepton (data not shown). That the mutations were found predominantly in the rplK gene, resulting in a modified L11 protein, and not in the 23S rRNA may indicate some differences in the binding of these compounds versus that of thiostrepton and micrococcin. Mutant colonies were small in size and grew significantly more slowly in liquid medium than the wild-type strain. It is unclear at this point how quickly resistance would emerge in the clinic or if the lower growth rates is an indication of the cost that resistance imposes on the fitness of the bacteria.

We believe that the nocathiacins described here represent a novel group of compounds related to thiostrepton that inhibit bacterial protein synthesis and merit further investigation. They exhibit excellent potency against important gram-positive pathogens, including antibiotic-resistant strains. Because they target a region of the ribosome that is unique in comparison to currently used drugs, no cross-resistance to existing agents is likely. The nocathiacins have the potential to become new, effective agents for the treatment of serious gram-positive bacterial infections.