An Antibiotic That Inhibits a Late Step in Wall Teichoic Acid Biosynthesis Induces the Cell Wall Stress Stimulon in Staphylococcus aureus

Abstract

Wall teichoic acids (WTAs) are phosphate-rich, sugar-based polymers attached to the cell walls of most Gram-positive bacteria. In Staphylococcus aureus, these anionic polymers regulate cell division, protect cells from osmotic stress, mediate host colonization, and mask enzymatically susceptible peptidoglycan bonds. Although WTAs are not required for survival in vitro, blocking the pathway at a late stage of synthesis is lethal. We recently discovered a novel antibiotic, targocil, that inhibits a late acting step in the WTA pathway. Its target is TarG, the transmembrane component of the ABC transporter (TarGH) that exports WTAs to the cell surface. We examined here the effects of targocil on S. aureus using transmission electron microscopy and gene expression profiling. We report that targocil treatment leads to multicellular clusters containing swollen cells displaying evidence of osmotic stress, strongly induces the cell wall stress stimulon, and reduces the expression of key virulence genes, including dltABCD and capsule genes. We conclude that WTA inhibitors that act at a late stage of the biosynthetic pathway may be useful as antibiotics, and we present evidence that they could be particularly useful in combination with beta-lactams.

INTRODUCTION

Staphylococcus aureus is a pathogenic Gram-positive bacterium that is increasingly difficult to treat due to the rise of drug resistance (45, 60). Many S. aureus strains are resistant to all beta-lactams, the most effective class of antibiotics in history (20, 59). Some of these methicillin-resistant S. aureus (MRSA) strains have begun to acquire resistance to vancomycin, the drug of last resort (27, 59, 61). Two new compound classes have been introduced to treat S. aureus infections since 2000, and the emergence of resistance to those is now a concern (40, 61). Thus, novel compounds and strategies to overcome multiply resistant S. aureus infections are needed.

Wall teichoic acid (WTA) biosynthesis is a proposed new antibiotic target in S. aureus (70). WTAs are phosphate-rich, sugar-based polymers that make up a large portion of the cell wall mass in many Gram-positive bacteria (4, 48). In Bacillus subtilis, WTAs are critical for the maintenance of the rod-shaped morphology (14). In S. aureus, WTAs play an important role in cell division and are required for establishing infections in animal hosts (10, 62, 67–69).

S. aureus WTAs are synthesized on a bactoprenol carrier lipid on the inner leaflet of the cytoplasmic membrane (28). In the first enzymatic step of WTA biosynthesis, TarO transfers an N-acetylglucosamine-1-phosphate (GlcNAc-phosphate) residue to bactoprenol phosphate to generate a GlcNAc-diphospholipid, which is transformed by the glycosyltransferase TarA into a lipid-anchored disaccharide [ManNAc(β1,4)GlcNAc] (Fig. 1) (22). The C4 hydroxyl of this disaccharide is primed with two glycerol-phosphate units, and then 20 to 40 ribitol-phosphate repeats are added to form the main chain polymer (9). The free hydroxyls on the ribitols are heavily modified with α- or β-O-GlcNAc residues (62). The mature polymers are transported to the external surface of the bacterial membrane by a two-component ABC transporter, TarGH, and are subsequently attached to peptidoglycan (PG) by an unknown mechanism (Fig. 1) (32). The main chains are further tailored with d-alanyl esters, which affect their overall charge and increase resistance to cationic antibiotics (56).

Fig 1.

Simplified schematic of the S. aureus WTA biosynthetic pathway. Tunicamycin inhibits TarO and targocil inhibits the transmembrane component (TarG) of the ABC transporter (TarGH) responsible for WTA export. Nonessential early steps in the WTA pathway are indicated by green arrows; essential late steps are indicated by red arrows. Nonessential tailoring steps are omitted for clarity.

A striking feature of the WTA biosynthetic pathway is its mixed gene dispensability pattern (15, 62). The first two genes in the pathway, tarO and tarA, are not required for in vitro growth; however, WTA deletion strains are nonpathogenic due to defects in host adhesion and dysregulated cell division (10, 13). Hence, these early steps are proposed targets for anti-virulence factor agents. In contrast, most of the downstream genes cannot be deleted unless flux into the biosynthetic pathway is prevented (e.g., by deleting tarO) (15). The conditional essentiality of the late acting genes is proposed to result from toxicity of accumulated intermediates and/or depletion of the bactoprenol-phosphate carrier lipid, which is also used for PG biosynthesis (66). Therefore, the late, essential steps in the WTA biosynthetic pathway have been suggested as novel targets for antibiotics.

The natural product tunicamycin (Fig. 1) is a selective inhibitor of TarO, which catalyzes the first step in the S. aureus WTA biosynthetic pathway, and we have previously characterized its cellular effects (10). While tunicamycin has only a modest effect on cell growth and gene expression at concentrations that inhibit TarO, it profoundly alters cellular morphology. Like ΔtarO strains, tunicamycin-treated cells have major defects in septal placement and other aspects of cell division. Since WTAs are necessary for the expression of beta-lactam resistance in MRSA, tunicamycin also resensitizes MRSA strains to this important family of antibiotics (10). Hence, our previous studies have validated TarO as a promising target for novel anti-infectives for use alone or in combination with beta-lactams to treat S. aureus infections.

Here we characterize the effects on S. aureus of a specific inhibitor of one of the late, essential steps in the WTA biosynthetic pathway. The inhibitor, targocil, was discovered via high-throughput screening (36, 63) and shown to block TarG, the transmembrane component of the ABC transporter that exports WTA polymers to the cell surface (58, 63). Deletion of tarO or inhibition of TarO by tunicamycin antagonizes the activity of targocil, consistent with the conditional essentiality of TarG (63). We report here that TarG blockade by targocil has very different effects on S. aureus than TarO inhibition. Targocil inhibits bacterial growth, strongly induces the cell wall stress stimulon, and downregulates the expression of numerous virulence factors. Transmission electron microscopy (TEM) studies show morphological abnormalities, but they are distinct from those observed for tunicamycin and suggest a response to cell envelope damage and osmotic stress rather than global defects in cell division. Unlike tunicamycin, targocil does not sensitize MRSA strains to beta-lactams; however, subinhibitory concentrations of beta-lactams suppress the development of targocil-resistant mutants. These observations and their implications for the use of WTA-active antibiotics are discussed.

MATERIALS AND METHODS

Growth conditions for electron microscopy samples.

S. aureus SH1000, a fully sequenced methicillin-sensitive derivative of NCTC8325 (51) in which a defect in the alternative sigma factor sigB has been corrected (26), was used in the present study. A single colony of S. aureus SH1000 was inoculated into tryptic soy broth (TSB) and shaken overnight at 30°C. This overnight culture was then diluted 1/500 into 60 ml of fresh TSB and shaken at 37°C for 2 h (optical density at 600 nm [OD₆₀₀] of ∼0.17). The culture was split into 10 5-ml aliquots, each in a 50-ml conical tube. Targocil (5.0 μg/ml, 8× MIC) was added to half of the cultures, which continued to shake at 37°C until harvesting. Samples were spun down (7,500 × g, 8 min) each hour following compound treatment (at 1, 2, 3, 4, and 8 h). Pellets were resuspended in 0.25 ml of TSB, and 0.25 ml of glutaraldehyde fixative solution was added to the sample. After 30 min at room temperature, the fixed samples were spun down and submitted to the Harvard Medical School EM Facility for further processing.

Autolysis procedure.

Autolysis experiments were performed as described previously (35). Briefly, an overnight culture of S. aureus SH1000 was diluted 1/100 into TSB (20 ml) in six 100-ml Erlenmeyer flasks and grown at 37°C to an OD₆₀₀ of ∼0.3. Cultures of four flasks were treated with 8× MIC of targocil dissolved in dimethyl sulfoxide (DMSO). A similar volume of DMSO (2%) was added to the other two flasks as a control. After 30 min, two control and two targocil-treated cultures were harvested by centrifugation. Two targocil-treated culture flasks were harvested after 120 min. The cells were washed by suspending in cold sterile water, followed by centrifugation. The cell pellets were resuspended in 0.05 M Tris-Cl (pH 7.2) containing 0.05% (vol/vol) Triton X-100 to an OD₆₀₀ of ∼0.6. The cell cultures were incubated at 37°C with shaking, and the OD₆₀₀ was determined at 30-min intervals.

Growth conditions for microarray analysis.

For transcriptional profiling, an overnight S. aureus SH1000 culture was diluted (2% [vol/vol]) into 20 ml of TSB medium in a 50-ml Erlenmeyer flask and grown at 37°C with shaking at 200 rpm. For the targocil treatment experiments, S. aureus was grown to an OD₆₀₀ of ∼0.4 and challenged with 8× MIC of targocil dissolved in DMSO for 30 min. Simultaneously, an equal amount of DMSO (final concentration 2% [vol/vol]) was added to the control culture, which was incubated along with the treated cultures. After 30 min, 5-ml aliquot was collected and used as a control for the 30-min treated culture.

RNA preparation and probe synthesis.

For total RNA extraction, 5 ml of culture was used and mixed with 5 ml of bacterial RNA Protect solution (Qiagen, Valencia, CA) and was further processed, and probes were synthesized as described by Muthaiyan et al. (46). Three independent bacterial cultures were prepared as biological replicates for RNA isolation of both test and control samples.

Microarray hybridization, processing, and data analysis.

Labeled c-DNA was hybridized with S. aureus Genome Microarray (version 7.0), provided by the Pathogen Functional Genomics Resource Center of the National Institutes of Allergy and Infectious Diseases, National Institutes of Health. The full genome array consists of 4,589 open reading frames (ORFs) from S. aureus reference strains COL, Mu50, MW2, N315, MRSA252, USA300-FPR3757, and pLW043 (http://pfgrc.jcvi.org/index.php/microarray/array_description/staphylococcus_aureus/version7.htmL). Each ORF is printed in four replicates on the array. Hybridization signals were scanned, and fold changes in gene expression were calculated as described previously (46). A cutoff of 2.0-fold for over- and underexpressed ORFs was used except where noted. The gene data sets were found to be significant at a false discovery rate of <1% (i.e., q < 0.01 [q represents the estimated probability of a false positive]). Genes with altered transcription levels were functionally categorized using our in-house Gene Sorter software according to the categories described in the comprehensive microbial resource of TIGR (http://cmr.tigr.org/tigr-scripts/CMR/shared/Genomes.cgi). The data discussed here will be available in NCBI's Gene Expression Omnibus (GEO) under accession number GSE34441.

RESULTS

Effects of targocil on S. aureus growth and morphology.

The targocil MIC against most S. aureus strains is about 1 μg/ml (36, 63). Based on counting CFUs as a function of time, we previously reported that targocil has bacteriostatic activity against S. aureus (36, 62, 63). Since then, we have found that the OD₆₀₀ of S. aureus does not remain constant following the addition of targocil to exponentially growing cultures but continues to rise gradually, albeit at a greatly reduced rate compared to the untreated control. Hence, the biomass of cultures treated with targocil increases even though the number of viable colonies does not change (see Fig. S1 in the supplemental material).

We used TEM to characterize the effects of targocil on the morphology of S. aureus SH1000, a methicillin-sensitive strain (see Materials and Methods). Samples were isolated each hour for 8 h from cultures of exponentially growing S. aureus SH1000 treated with targocil, and TEM images were acquired for each sample (Fig. 2 and see Fig. S2 in the supplemental material). After 2 h, septa are present in 26.1% ± 2.6 of wild-type cells but in 67.4% ± 4.4 of targocil-treated cells. This observation initially suggested a block in cell division. However, after 2 h we observed clusters of cells encased in a continuous layer of dark-staining surface material in the targocil-treated samples (Fig. 2). The number of cells in each cluster increased over time, and the septal material became thicker but stained more lightly, suggesting decreased levels of electron dense WTAs (43). The cells also increased in size, with the average area of individual cells doubling within 4 h (wild type, 0.02 ± 0.002; treated [no septa], 0.05 ± 0.005; treated [septa], 0.082 ± 0.02). Based on the TEM studies, we concluded that the increased biomass observed by OD₆₀₀ measurement is due to increased cell size and slow ongoing cell division, which is not reflected in increased CFUs because daughter cell separation is severely impaired (see Fig. S3 and Table S1 in the supplemental material).

Fig 2.

Targocil treatment causes impaired cell separation and increased cell size. TEM images of S. aureus SH1000 after 1 to 8 h of treatment with 8× MIC targocil show a shutdown of cell separation leading to the formation of multicellular clusters. (a) Untreated control; (b) 1 h; (c) 2 h; (d) 3 h; (e) 4 h; (f) 8 h. Scale bars, 400 nm. For septa percentages, we counted four fields of cells using the program ImageJ and averaged the data. The difference was statistically significant.

Effects of targocil on autolytic activity.

Autolysins are cell wall-degradative enzymes required for cell division in S. aureus (71). They facilitate cell separation following division by degrading old PG. Triton X-100 induces autolysin activity in S. aureus by unknown mechanisms, and this is detected as a decrease in OD₆₀₀ as the cells lyse (24). Since cell separation is impaired in targocil-treated cells, we tested the effects of targocil on induced autolytic activity. Cultures of S. aureus SH1000 were treated with targocil for 30 or 120 min, and then the cells were transferred to buffer containing Triton X-100. Whereas untreated cells were fully lysed within 2 h, targocil-treated cells showed minimal lysis after 2 h (Fig. 3). Since cell separation requires activation of autolytic enzymes, the decreased autolytic activity upon Triton X-100 treatment is consistent with the TEM images showing that daughter cell separation is inhibited.

Fig 3.

Targocil treatment protects the bacteria against Triton X-100-induced autolysis. Autolysis of S. aureus SH1000 treated without (●) or with 8× MIC targocil for 30 min (■) and 120 min (▲).

Effects of targocil on gene transcripts level in S. aureus.

Gene expression profiling studies in S. aureus SH1000 were performed to identify genes for which transcript levels changed significantly upon targocil treatment (i.e., up- or downregulated genes). Microarray analysis was carried out after 30 min of targocil treatment, and the data reported represent the averages of three independent biological replicates. A total of 119 genes were upregulated at least 2-fold after 30 min, whereas 96 genes were downregulated (Tables 1 and 2; see also Tables S2 and S3 in the supplemental material). Excluding genes of unknown or hypothetical function, the major categories encode proteins involved in transport and binding, cell envelope, energy metabolism, cellular processes, regulatory functions, protein fate, nucleotide synthesis, and central metabolism (Fig. 4). Within each category, the number of genes up- and downregulated is similar except in the protein fate category, which contains almost exclusively upregulated genes. The pattern of significantly affected and minimally affected categories is consistent with the effects of a molecule that targets the cell wall. This conclusion is supported by more detailed analysis of specific changes, which are described below.

Table 1.

Genes upregulated by targocil treatment in S. aureus SH1000^a

Locus	Gene	Protein	Description	Fold changeb
Cell envelope
SAV2556	cwrA	Hypothetical protein	Induced by cell wall stress	26.5
SACOL1066	fmt	Beta-lactamase fold	Biosynthesis and degradation of murein sacculus and peptidoglycan	4.5
SACOL2509	fnbB	Fibronectin binding protein B	Cell adherence	4.0
SACOL2511	fnbA	Fibronectin-binding protein A	Cell adherence	3.6
SAS2388	NAc	Fibronectin-binding protein precursor	Cell adherence	3.5
SACOL1932	sgtB	Transglycosylase domain protein	Biosynthesis and degradation of murein sacculus and peptidoglycan	3.5
SAV2442	NA	Similar to dTDP-glucose 4,6-dehydratase	Biosynthesis and degradation of surface polysaccharides and lipopolysaccharides	3.0
SAV1450	pbp2	Class A penicillin-binding protein	Biosynthesis and degradation of murein sacculus and peptidoglycan	2.5
SAV2124	murZ	UDP-N-acetylglucosamine 1-carboxyvinyltransferase	Biosynthesis and degradation of murein sacculus and peptidoglycan	2.1
SAV2133	hmrA	Similar to amidase	Surface structures	2.2
SACOL0239	tarF	CDP-glycerol-3-phosphate primase	WTA biosynthesis	2.3
SAV0253	tarK	CDP-ribitol-5-phosphate polymerase	WTA biosynthesis	2.2
SACOL0694	tarH	Teichoic acid export protein ATP-binding subunit	WTA biosynthesis	1.5*
SAV0256	tarJ	Xylitol dehydrogenase	WTA biosynthesis	1.7*
SAV0212	NA	Peptidase	Biosynthesis and degradation of murein sacculus and peptidoglycan	5.4
Regulatory function
SAV0524	mcsB	Arginine phosphotransferase	Derepresses CtsR, a repressor of heat-shock stress response gene expression	5.8
SAS0479	ctsR	Putative DNA-binding protein	Represses heat-shock stress response genes clpC, clpB, and clpP transcription	4.9
SAV1582	hrcA	Heat-inducible transcriptional repressor	Negative regulator of class I heat shock genes	3.0
SACOL1942	vraR	DNA-binding response regulator VraR	Regulates cell wall stress response genes	4.2
SAR1975	vraS	Histidine kinase sensor	Senses cell wall stress and transduces signal	3.0
Protein fate
SAV0975	clpB	ATP-dependent Clp protease, ATP-binding subunit ClpB	Involved in the recovery of the cell from heat-induced damage	6.4
SACOL0570	clpC	ATP-dependent Clp protease, ATP-binding subunit ClpC	Affects virulence gene expression, including cap operon	4.7
SAV2030	groES	Chaperonin, 10 kDa	Protein folding and stabilization	5.4
SAV2029	groEL	Chaperonin, 60 kDa	Protein folding and stabilization	3.4
SAV1581	grpE	Heat shock protein GrpE	Protein folding and stabilization	3.8
SAV1841	prsA	Peptidylprolyl cis/trans isomerase	Essential folding factor for secreted proteins	3.5
SAV1423	msrB	Methionine sulfoxide reductase B	Protein modification and repair	3.0
SACOL1636	dnaJ	DnaJ protein	Protein folding and stabilization	2.7
SACOL1637	dnaK	DnaK protein	Protein folding and stabilization	1.6
SAS0834	NA	Putative signal peptidase Ia	Protein and peptide secretion and trafficking	2.9
SAS1654	NA	Putative protease	Degradation proteins, peptides, and glycopeptides	2.6
SACOL0968	spsA	Signal peptidase IA, inactive	Protein and peptide secretion and trafficking	2.5
SAS0835	NA	Signal peptidase Ib	Protein and peptide secretion and trafficking	2.5
SACOL2385	NA	Heat shock protein, Hsp20 family	Protein folding and stabilization	2.5
SACOL0595	NA	Peptidase, M20/M25/M40 family	Degradation of proteins, peptides, and glycopeptides	2.5
SACOL0833	clpP	ATP-dependent Clp protease, proteolytic subunit ClpP	Responsible for adaptation to multiple stresses by degrading accumulated and misfolded proteins	2.3
SAV1253	clpQ	ATP-dependent protease peptidase subunit	Responsible for adaptation to multiple stresses by degrading accumulated and misfolded proteins	2.0
SAV1153	NA	Hypothetical protein	Degradation proteins, peptides, and glycopeptides	2.1
Transport and binding proteins
SAS0283	NA	Putative amino acid transport system	Amino acids, peptides, and amines	3.4
SAV0838	NA	Similar to ABC transporter, permease protein homolog	Amino acids, peptides, and amines	3.3
SAV2450	NA	Amino acid transporter	Amino acids peptides and amines	2.4
SAR2537	opuCB	Putative glycine betaine/carnitine/choline transport system permease protein	Responsible for the translocation of the compatible solutes across the membrane	2.3
SAV2446	opuCC	Glycine betaine/carnitine/choline ABC transporter	Responsible for the translocation of the compatible solutes across the membrane	2.3
SACOL2449	NA	Drug transporter	Toxin production and resistance	2.2
SAV1866	NA	ABC transporter homolog	Amino acids peptides and amines	2.4
SAS2251	hrtB	Putative permease protein	Hemin import	12.8
SAV2359	hrtA	Similar to ABC transporter (ATP-binding protein)	Hemin import	12.7
Energy metabolism
SAV0523	NA	Hypothetical protein	ATP-proton motive force interconversion	4.3
SAV2536	NA	Similar to thioredoxin	Electron transport	2.8
SAV0452	ndhF	NADH dehydrogenase subunit L	Electron transport	4.3
SAV2187	NA	Similar to quinone oxidoreductase	Electron transport	2.7
SAS0410	NA	NADH dehydrogenase subunit L	Electron transport	2.5
SAV1086	NA	Cytochrome d ubiquinol oxidase subunit 1 homolog	Electron transport	2.1
SAR0400	NA	Nitroreductase family protein	Electron transport	2.1
SAS0525	NA	Hypothetical protein	Electron transport	2.0
SACOL2280	ureA	Urease, gamma subunit	Catalyzes the hydrolysis of urea to form ammonia	2.7
SAV2289	ureB	Urease, beta subunit	Catalyzes the hydrolysis of urea to form ammonia	2.6
SACOL2282	ureC	Urease, alpha subunit	Catalyzes the hydrolysis of urea to form ammonia	2.7
SAV2294	ureD	Urease accessory protein UreD	Required for maturation of urease	2.7
SACOL2283	ureE	Urease accessory protein UreE	Required for maturation of urease	2.6
SACOL2284	ureF	Urease accessory protein UreF	Required for maturation of urease	2.3
Hypothetical proteins
SACOL0625	NA	Hypothetical protein	Conserved	7.3
SAS1587	NA	Hypothetical protein	Hypothetical proteins	6.2
SACOL1945	NA	Hypothetical protein	Conserved	4.6
SACOL1705	NA	Hypothetical protein	Hypothetical proteins	4.3
SAS0657	NA	Hypothetical protein	Hypothetical proteins	4.1
SAR0453	NA	Hypothetical protein	Conserved	4.1

^aData presented here are the combined results of three independent biological replicates.

^bThe cutoff was 2-fold for upregulated genes. *, A fold change of <2-fold.

^cNA, not applicable.

Table 2.

Genes downregulated by targocil treatment in S. aureus SH1000^a

Locus	Gene	Protein	Description	Fold changeb
Cell envelope
SAV0932	dltA	d-Alanine-d-alanyl carrier protein ligase	Biosynthesis and degradation of murein sacculus and peptidoglycan	–2.7
SACOL0936	dltB	Activated d-alanine transport protein	Biosynthesis and degradation of surface polysaccharides and lipopolysaccharides	–1.7*
SACOL0937	dltC	d-Alanine-poly(phosphoribitol) ligase subunit 2	Biosynthesis and degradation of surface polysaccharides and lipopolysaccharides	–2.0
SAV0935	dltD	Poly(glycerophosphate chain) d-alanine transfer protein	Biosynthesis and degradation of murein sacculus and peptidoglycan	–3.3
SAR0151	capA	Similar to S. aureus capsular polysaccharide synthesis enzyme Cap5A	Involved in capsular polysaccharide synthesis	–2.2
SAV0160	capL	Capsular polysaccharide synthesis enzyme Cap5L	Involved in capsular polysaccharide synthesis	–2.1
SACOL0148	cap5M	Capsular polysaccharide synthesis enzyme galactosyltransferase Cap5M	Involved in capsular polysaccharide synthesis	–2.0
SACOL0137	cap5B	Capsular polysaccharide biosynthesis protein Cap5B	Involved in capsular polysaccharide synthesis	–1.9*
SAV0154	capF	Capsular polysaccharide synthesis enzyme reductase Cap5F	Involved in capsular polysaccharide synthesis	–2.4
Regulatory function
SACOL1812	rot	Repressor of toxins	Virulence gene regulator	–6.1
SACOL1210	pyrR	Uracil phosphoribosyltransferase	Pyrimidine regulatory protein PyrR	–6.7
SACOL1328	glnR	Glutamine synthetase repressor	Toxin production and resistance	–4.0
SACOL2287	sarR	Staphylococcal accessory regulator R	Regulate agr expression	–3.4
SACOL2026	agrA	Accessory gene regulator protein A	Regulates the expression of many virulence genes	–3.3
SA1248	Truncated arlR	Truncated (putative response regulator ArlR)	Pathogenesis	–2.5
SACOL2056	rsbV	Anti-anti-sigma factor RsbV	Key regulators of anti-sigma factors	–2.0
Transport and binding proteins
SAS0164	NAc	Glucose-specific PTS transporter protein, IIABC component	Glucose specific	–9.2
SAV1199	pyrP	Uracil permease	Nucleosides, purines, and pyrimidines	–7.7
SAV0189	glcA	PTS enzyme II	Glucose specific	–7.2
SAS0665	NA	PTS transport system, fructose-specific IIABC component	Carbohydrates, organic alcohols, and acids	–5.5
SAV0389	pbuX	Xanthine permease	Nucleosides, purines, and pyrimidines	–5.2
SACOL0689	NA	ABC transporter, permease protein	Involved in the uptake of siderophores, heme, vitamin B12, or the divalent cations	–2.9
SACOL2292	nhaC	Na+/H+ antiporter NhaC	Cations and iron carrying compounds	–2.6
SAV1099	potA	ATP-binding protein	Spermidine/putrescine ABC transporter	–2.3
SAV0621	NA	Putative monovalent cation/H+ antiporter subunit A	Involved in resistance to high concentrations of Na+, K+, Li+, and/or alkali	–2.1
Energy metabolism
SAV0699	fruB	Fructose 1-phosphate kinase	Carbohydrate transport and metabolism	–4.0
SACOL1321	glpD	Aerobic glycerol-3-phosphate dehydrogenase	Other	–2.6
SAV0320	geh	Glycerol ester hydrolase	Biosynthesis and degradation of polysaccharides	–2.2
SA2102	NA	Hypothetical protein	Fermentation	–2.0
SAV1413	odhA	Oxoglutarate dehydrogenase	TCA cycle	–2.0
SAS2105	NA	Alpha-acetolactate decarboxylase	Fermentation	–2.0
SA1533	ackA	Acetate/propionate kinase	Fermentation	–2.0
SAV2607	mqo2	Malate:quinone oxidoreductase	TCA cycle	–2.0
Cellular processes
SAV0111	spa	Staphylococcal protein A	Virulence factor	–2.5
SACOL2576	crtM	Dehydrosqualene desaturase	Involved in staphyloxanthin biosynthesis	–1.8*
SAV2561	crtN	Squalene synthase	Involved in staphyloxanthin biosynthesis	–1.5*
Purines, pyrimidines, nucleosides, and nucleotides
SAV1202	pyrAA	Carbamoyl-phosphate synthase small subunit	Pyrimidine ribonucleotide biosynthesis	–8.2
SAV1201	pyrC	Dihydroorotase	Pyrimidine ribonucleotide biosynthesis	–5.4
SACOL1217	pyrE	Orotate phosphoribosyltransferase	Pyrimidine ribonucleotide biosynthesis	–5.1
SACOL1216	pyrF	Orotidine 5′-phosphate decarboxylase	Pyrimidine ribonucleotide biosynthesis	–4.8
SAV1203	pyrAB	Carbamoyl-phosphate synthase large subunit	Pyrimidine ribonucleotide biosynthesis	–4.4
SAS1134	NA	Aspartate carbamoyltransferase catalytic subunit	Pyrimidine ribonucleotide biosynthesis	–5.1
SAV0699	fruB	Fructose 1-phosphate kinase	Sugar-nucleotide biosynthesis and conversions	–4.0
SAV0390	guaB	Inositol-monophosphate dehydrogenase	Purine ribonucleotide biosynthesis	–3.0
SAR0406	xpt	Xanthine phosphoribosyltransferase	Salvage of nucleosides and nucleotides	–2.8
SACOL2606	pyrD	Dihydroorotate dehydrogenase	Pyrimidine ribonucleotide biosynthesis	–2.8

^aData presented here are the combined results of three independent biological replicates.

^bThe cutoff was 2-fold for downregulated genes. *, A fold change of <2-fold.

^cNA, not applicable.

Fig 4.

Numbers of genes in major categories affected after 30 min of targocil treatment (up/down). Fewer than 15 genes involved in central intermediary metabolism (11/2) and purine/pyrimidine biosynthesis (1/13) were affected; categories are not displayed. No significant changes were observed in the expression of genes responsible for amino acid biosynthesis, cofactor biosynthesis, DNA metabolism, fatty acid and phospholipid synthesis, protein synthesis, signal transduction, or transcription, and these categories are not displayed.

Cell envelope genes.

Cell wall-active antibiotics induce characteristic changes in numerous genes that comprise the “cell wall stress stimulon” (30, 34, 64). Many of these genes are regulated by a two-component system composed of a sensor, VraS, and a response regulator, VraR, which correspond to LiaRS in the model organism B. subtilis (41, 42). Targocil treatment upregulated both the vraR and vraS genes (Table 1). Correlated changes in other cell wall stress stimulon genes were also observed upon targocil treatment (Table 1). For example, cwrA, a very sensitive reporter of cell wall stress, which encodes a small peptide of unknown function, was upregulated by 26.5-fold (5, 44). The sgtB gene, which encodes a monofunctional glycosyltransferase involved in PG biosynthesis and is known to be induced by cell wall stress (34), was also upregulated. Other upregulated cell envelope-related genes include the following: SAV0212, a predicted peptidase; fmt, a gene implicated in both methicillin resistance and regulation of autolysis; pbp2, the sole essential class A penicillin-binding protein in S. aureus; and murZ, a UDP-N-acetylglucosamin-1-carboxyvinyl transferase that catalyzes the first committed step in PG precursor synthesis.

Unlike typical cell wall-active antibiotics, targocil treatment also upregulates a number of WTA pathway genes in addition to PG synthesis genes (Table 1). These genes include the CDP-glycerol-3-phosphate primase tarF, a CDP-ribitol-5-phosphate polymerase tarK (55, 56), and tarGH, which encode the two-component ABC transporter that exports WTAs to the cell surface. TarG is the target of targocil (36, 58, 63). The upregulation of a number of WTA genes is consistent with the proposal that targocil blocks the maturation of WTAs (58, 63).

Genes involved in virulence factor expression or synthesis.

S. aureus expresses a large number of nonessential proteins, polysaccharides, lipids, and small molecules that have been implicated in virulence (8, 17). Analysis of the transcriptome shows that targocil downregulates several genes encoding transcription factors involved in virulence factor production (Table 2). These downregulated transcription factor genes include sarR (staphylococcal accessory regulator R), agrA (accessory gene regulator A), arlR (a DNA-binding response regulator R), and rsbV (anti anti-sigma factor V) (1, 7, 18, 53). In S. aureus, the agr operon is expressed as a single polycistronic unit and other genes of this operon, agrC2, agrB, and agrD are also moderately downregulated upon targocil treatment (see Table S3 in the supplemental material). The downregulation of some of these transcription factor genes may explain decreased expression of several virulence genes including a probable adhesin, SAV0631, the lipase geh and staphylococcal protein A.

Targocil treatment also reduces the expression of capsular polysaccharide genes including capA, capF, capH, capL, and cap5M (Table 2). Although changes in the levels of capsule gene mRNA levels were modest (ranging from −1.9- to −2.4-fold), similar values were observed for multiple genes in the operon, indicating a consistent association between targocil treatment and decreased capsule production. Capsule is important for the efficient colonization and infection of certain animal models (6). The cap genes may be downregulated due to reduced expression of arlR and agrA during an SOS response (11). ClpC has also been shown to repress capsule production (39), and its expression is upregulated in the targocil transcriptome (Table 1). The dltABCD operon showed similar reductions in gene expression (ranging from −1.7- to −3.3-fold), which also supports a functional connection between targocil treatment and cell wall changes (Table 2). The products of the dlt genes catalyze d-alanylation, a tailoring modification of both lipoteichoic acids and WTAs that has been implicated in animal infection, colony spreading, and resistance to cationic antibiotics (31, 56, 69). Due to the roles of d-alanylation in virulence, the enzymes in the d-alanylation pathway are proposed therapeutic targets, and downregulation of the dlt genes is a potentially advantageous feature of late stage WTA inhibitors.

The staphyloxanthin biosynthetic genes ctrM and ctrN were also modestly downregulated (−1.9- and −1.5-fold, respectively; Table 2), and the cell pellets following targocil treatment had lost their characteristic golden color, which is due to staphyloxanthin production (see Fig. S4 in the supplemental material). Staphyloxanthin is a carotenoid that increases fitness by protecting S. aureus from oxidative stress and neutrophil-based killing (12) and, since its loss increases the susceptibility of S. aureus to innate immune clearance in a mouse infection model (38, 50), staphyloxanthin biosynthesis is a proposed anti-virulence factor target.

Protein fate genes.

A major category of affected genes are those involved in protein folding and degradation, which are almost exclusively upregulated in the targocil transcriptome (Table 1). These include the following genes: grpE, dnaJ, and dnaK, which encode proteins that work together to prevent the aggregation of stress-denatured proteins (49); clpB, groES, and groEL, which encode proteins that help ensure the proper folding of nascent peptides (16, 19); clpC, which encodes a protease that degrades denatured proteins (37); and psrA, which encodes a peptidyl-prolyl-cis-trans isomerase that participates in protein folding (23, 25). Genes that regulate protein fate genes also show increased expression. For example, mcsB, a gene encoding an unusual arginine kinase that negatively regulates the stress gene repressor CtsR, is highly upregulated (21). Likewise, the gene encoding the heat-inducible transcriptional repressor HcrA, which coordinates the expression of many of these protein quality control elements (47) is also upregulated. Although PG biosynthesis inhibitors also induce a heat shock-like stress response due to a proposed accumulation of misfolded and aberrant proteins (2), the response to targocil treatment is particularly strong and suggests that misfolded and aberrant proteins may accumulate to a greater extent upon targocil treatment than other cell wall-active antibiotics. Alternatively, the accumulated WTA polymers may be sensed as defective/misfolded proteins.

Transport/binding proteins.

Genes responsible for transport and binding proteins make up the largest functional category in S. aureus (other than hypothetical proteins), and the greatest number of gene expression changes occurs in this category upon targocil treatment (Fig. 4). For example, the expression of hrtA and hrtB, which encode a two-component ABC exporter putatively involved in hemin efflux, are each increased 12.7-fold. These genes modulate virulence, prevent toxicity due to the accumulation of hemin, and redirect central metabolism to increase iron availability (3, 29). Genes SAS0283, SAV0838, and SAV2450, encoding several putative amino acid transporters, are also upregulated. The genes opuCB and opuCC encoding the compatible solute glycine betaine and carnitine transporters are upregulated as well (Table 1).

Transport genes negatively affected by targocil treatment include the following: the uracil permease gene pyrP and its regulator, pyrR; pbuX (xanthin permease), the fructose-specific transporter; SACOL0689 (metal cation ABC transporter permease); and nhaC (Na⁺/H⁺ antiporter). Genes encoding the phosphotransferase system, SAS0665 (fructose specific) and glcA (glucose specific) are also downregulated. Changes in the expression of transport genes are observed with other cell wall-active antibiotics and suggest altered envelope integrity.

Metabolism and other genes.

Some electron transport genes are upregulated in S. aureus upon targocil treatment. For example, ndhF and SAV2536 transcript levels increase 4.3- and 2.8-fold, respectively. The product of ndhF protects bacteria from oxidative stress and has been implicated in a theory that links bacterial killing by antibiotics to oxidative stress (33). Several urease pathway genes are also upregulated (Table 2). These genes function to produce ammonia and could be acting to neutralize and/or buffer the negatively charged polymers accumulating within the cell upon TarG inhibition.

Subinhibitory concentrations of beta-lactams suppress the development of targocil resistance.

We have previously reported that inhibiting TarO with tunicamycin restores beta-lactam sensitivity to MRSA (10). For example, the MIC of oxacillin decreases by 128-fold in some MRSA strains when WTA synthesis is blocked (10). To determine whether late-stage inhibition of WTA synthesis also sensitizes MRSA strains to beta-lactams, we tested the oxacillin MICs of several MRSA strains in the presence of targocil using a checkerboard assay (data not shown). Synergy was not observed in the checkerboard assay since the MICs for each compound remained unchanged. Despite the lack of synergy, however, we observed that oxacillin concentrations 5-fold below the MIC suppressed the growth of targocil-resistant mutants in culture plates over a period of several days. This effect was recapitulated on agar plates. For example, when MRSA strains 1784A and MW2 were streaked onto plates containing targocil at 8× MIC, resistant mutants grew at high frequencies (>1 in 10⁶ CFUs). However, no colonies were visible if oxacillin at 0.2× MIC (5 μg/ml) was also included in the plates, even though a lawn of MRSA is typically observed at this beta-lactam concentration (Fig. 5 and see Fig. S5 in the supplemental material). We conclude from these observations that beta-lactams suppress the development of targocil-resistant mutants in MRSA strains.

Fig 5.

Beta-lactams suppress the development of targocil resistance. (a) Clinical MRSA isolate 1784A plated on TSB agar containing DMSO shows a lawn of bacteria. (b) 1784A plated on 8× MIC targocil shows a high frequency of resistant mutants after overnight incubation. (c) 1784A plated on 0.2× MIC (5 μg/ml) oxacillin shows a lawn of bacteria after overnight incubation. (d) 1784A plated on a combination of targocil (8× MIC) and oxacillin (0.2× MIC) shows no colonies.

DISCUSSION

New antibacterials to treat MRSA infections are desperately needed, and the WTA biosynthetic pathway is a proposed target for such agents. Genetic studies have identified two possible classes of targets in the pathway: (i) antivirulence targets that are not essential but affect pathogenicity and (ii) antibacterial targets that are essential (10, 15, 36, 62, 63, 67). Specific inhibitors of members of both classes of targets are available, and we have been investigating the pharmacological effects of blocking early and late steps of WTA synthesis using expression analysis and phenotypic characterization.

In previous work we showed that tunicamycin, which blocks the first step in the WTA pathway, causes profound morphological defects even though it does not significantly affect in vitro growth rates and has only a modest effect on gene expression (10). The morphological defects include aberrations in septal placement, a high frequency of duplicate septa, and an inability to separate following the completion of new septa. These defects imply that WTAs are required for properly coordinated cell division and suggest a link between PG and WTA biosynthesis. Properly regulated cell division is important for viability under stressful conditions and the extent of the cell division defects may help to explain why S. aureus ΔtarO strains are avirulent in animal models.

We have examined here the effects of a late-stage WTA inhibitor on S. aureus. This inhibitor, targocil, is a synthetic small molecule that was identified through high-throughput screening (36, 63). Unlike tunicamycin, targocil inhibits bacterial growth by affecting PG biosynthesis and also induces a strong cell wall stress response. For example, many genes in the cell wall stress stimulon are upregulated. Since tunicamycin does not have a similar effect on gene expression at the concentrations required to inhibit TarO (10), the observed cell wall stress response is not a result of preventing WTA expression, but of blocking its maturation at a late stage. Such a block would lead to the accumulation of bactoprenol-linked intermediates and prevent recycling of the bactoprenol phosphate carrier lipid (15), which is also required for PG biosynthesis. Since an intact PG layer is required for cell survival, damage to the cell envelope occurs over time if PG precursor pools are limited.

In addition to affecting the cell wall stress stimulon, targocil treatment upregulates “heat shock” genes such as groES, groEL, and grpE, which are involved in protein quality control. It is known that induction of heat shock proteins decreases the rate of autolysis in S. aureus cells (57). Furthermore, in Listeria monocytogenes, induction of a heat shock response leads to downregulation of both cell division genes and autolysins, preventing cell division (65). We observed significant decreases in both autolytic activity and cell division in S. aureus upon targocil treatment. These effects may be linked to the observed induction of a strong “heat shock-like” stress response.

Targocil treatment also downregulates genes involved in the production of several virulence factors, including capsule. Capsule is a nonessential cell wall glycopolymer that plays an important role in S. aureus pathogenesis (52). Like PG and WTAs, capsule is synthesized on the bactoprenol carrier lipid (72). The downregulation of genes involved in capsule formation may reflect an attempt to conserve the carrier lipid for use in the essential process of PG biosynthesis. This explanation could also be consistent with the observed downregulation of staphyloxanthin biosynthetic genes since staphyloxanthin is produced from a polyisoprenoid intermediate (54). We speculate that S. aureus has a mechanism for sensing PG precursor and can downregulate nonessential molecules that utilize the carrier lipid or its building blocks to ensure flux into PG synthesis. Downregulation of capsule may be a useful aspect of late stage WTA inhibitors.

In addition to examining changes in gene expression, cell morphology, and autolytic activity upon blocking TarG, we have also looked at the effects on MRSA of combining targocil with a beta-lactam. Although the TarO inhibitor tunicamycin strongly sensitizes MRSA to beta-lactams, no synergy between targocil and beta-lactams was found in a checkerboard assay. Since beta-lactams are known to be effective only against rapidly growing cells, the lack of synergy reflects targocil's growth-inhibitory activity. However, we observed that resistant mutants appear at a greatly reduced frequency when targocil is used in combination with oxacillin. A major mechanism by which resistance arises to targocil in vitro is through null mutations in tarO and tarA, which encode the first two, nonessential genes in the WTA biosynthetic pathway. We have previously shown that blocking TarO in MRSA strains restores beta-lactam susceptibility. Beta-lactams may suppress the formation of targocil resistance by eliminating the tarO/A-null mutant population (58, 63). Another mechanism for resistance suppression may be related to the viability of multicellular clusters in the presence of beta-lactams. We have shown that S. aureus cells continue to divide very slowly without separating in the presence of targocil alone. Resistant mutants may emerge from cells within the resulting multicellular clusters. If simultaneous beta-lactam treatment reduces the viability/growth of these clusters, e.g., by further weakening the cell wall due to dysregulated synthetic and degradative activity, then mutants may not emerge. Regardless of the mechanism, the finding implies that the combined use of beta-lactams and targocil may be useful in limiting the spread of resistance.

Conclusions.

The findings reported here show that targocil, the first identified late-stage inhibitor of WTA biosynthesis, acts as a cell wall-active antibiotic. It causes damage to the cell envelope, probably due to an inhibitory effect on PG biosynthesis since the bactoprenol carrier lipid is sequestered in WTA intermediates. It also induces a strong heat shock-like stress response due to the accumulation of misfolded proteins and the possible missensing of WTA intermediates as damaged proteins. Targocil decreases expression of several virulence factor genes and their regulators, which may be advantageous for its use in vivo. Finally, although beta-lactams do not synergize with targocil, as they do with the TarO inhibitor tunicamycin, subinhibitory concentrations of beta-lactams suppress the formation of targocil resistant mutants. Based on these results, we propose that the combined effects of a late-stage WTA inhibitor may lead to effective killing of S. aureus, particularly when used in combination with a beta-lactam.