Biosynthesis and Mechanism of Action of the Cell Wall Targeting Antibiotic Hypeptin

Abstract

Hypeptin is a cyclodepsipeptide antibiotic produced by Lysobacter sp. K5869, isolated from an environmental sample by the iChip technology, dedicated to the cultivation of previously uncultured microorganisms. Hypeptin shares structural features with teixobactin and exhibits potent activity against a broad spectrum of gram-positive pathogens. Using comprehensive in vivo and in vitro analyses, we show that hypeptin blocks bacterial cell wall biosynthesis by binding to multiple undecaprenyl pyrophosphate-containing biosynthesis intermediates, forming a stoichiometric 2:1 complex. Resistance to hypeptin did not readily develop in vitro. Analysis of the hypeptin biosynthetic gene cluster (BGC) supported a model for the synthesis of the octapeptide. Within the BGC, two hydroxylases were identified and characterized, responsible for the stereoselective β-hydroxylation of four building blocks when bound to peptidyl carrier proteins. In vitro hydroxylation assays corroborate the biosynthetic hypothesis and lead to the proposal of a refined structure for hypeptin.

Keywords: antibiotic; cell wall; cyclodepsipeptide; hydroxylase; lipid II.

Figure 1

Gene organization of the hyn BGC and biosynthetic pathway of hypeptin (1). The NRPS HynAB assemble a linear octapeptide which is finally released and cyclized by HynB_TE. The tailoring hydroxylases HynC and HynE (green) modify the building blocks during assembly. Has: 3‐Hydroxyasparagine Hty: 3‐Hydroxytyrosine Hle: 3‐Hydroxyleucine.

Figure 2

Results of in vitro assays to test NRPS and hydroxylase activities. On the left, the A domain specificity towards the substrate amino acid and the dependency of the MbtH‐like protein (MLP) HynMLP was examined for each module via γ¹⁸O₄‐ATP exchange assay. On the right, extracted LC‐MS traces of the hydroxylation assays of the module construct with HynC and HynE show formation of hydroxylated amino acids in comparison with the respective negative controls. At the bottom, the formation of the hydroxylated amino acid is summarized. a) The A domain of module 4 activates l‐Asn and is independent of HynMLP. HynE then hydroxylates the bound amino acid, leading to the formation of 3‐hydroxyasparagine (Has) (m/z=147.0). b) The A domain in module 5 activates l‐Asn only in presence of HynMLP. HynE then hydroxylates the bound amino acid, leading to the formation of 3‐hydroxyasparagine (Has) (m/z=147.0). c) The A domain of module 6 activates l‐Tyr in the presence of HynMLP. Subsequently, HynC hydroxylates the amino acid (m/z=196.1). d) The A domain of module 7 activates l‐Leu independently of HynMLP. HynC then hydroxylates the amino acid (m/z=146.1).

Figure 3

1 shows excellent bactericidal activity against S. aureus. Time‐dependent killing of early‐exponential (a) and late‐exponential phase‐grown (b) cells treated with 1 at 1× (open circles) and 2× MIC (circles), with teixobactin (diamonds) and vancomycin (triangles) both at 10× MIC. Cells left untreated are shown with squares. Data are representative of three independent experiments. c) 1‐induced lysis is mediated by the major autolysin AtlA in S. aureus. Deletion of atlA results in markedly reduced autolysis after treatment with 1 and TEIX.

Figure 4

1 targets bacterial cell wall biosynthesis. a) B. subtilis bioreporter strains with selected promotor‐lacZ gene fusions were used to identify interference with major biosynthesis pathways including cell wall (P_ypuA), DNA (P_yorB), RNA (P_yvgS), and protein (P_yhel). A blue halo at the edge of the inhibition zone demonstrates induction of a specific stress response by β‐galactosidase expression. Antibiotics vancomycin (VAN), ciprofloxacin, rifampicin, and clindamycin were used as positive controls. b) Treatment with 1 (1×MIC, open circles) strongly induced P_lial as observed by expression of the lux operon from Photorhabdus luminescens in B. subtilis P_liaI‐lux. VAN (triangles) and clindamycin (CLI, squares) were used as control antibiotics. c) Phase‐contrast microscopy of B. subtilis confirmed impairment of cell wall integrity as severe cell‐shape deformations and characteristic blebbing were observed following 1 treatment. Cell wall active antibiotics teixobactin (TEIX), VAN, plectasin (PLEC), ampicillin (AMP), and lysozyme (LYS) were used as controls. Scale bar=2 μm. d) Intracellular accumulation of the cell wall precursor UDP‐MurNAc‐pentapeptide after treatment of S. aureus with 1 (5×MIC). Untreated and VAN‐treated (5×MIC) cells were used as controls. Experiments are representative of 3 independent experiments each.

Figure 5

1 binds to undecaprenyl pyrophosphate‐containing cell wall precursors. a) 1 interferes with membrane‐associated steps of PGN and WTA synthesis in vitro. The antibiotic was added in molar ratios from 0.5 to 10 with respect to the amount of the lipid substrate C₅₅P, C₅₅PP, lipid II, or lipid III_WTA used in the individual test system. The amount of reaction product synthesized in the absence of 1 was taken as 100 %. Mean values from three independent experiments are shown. Error bars represent standard deviation. b) 1 forms extraction‐stable complexes with C₅₅PP‐containing purified cell wall precursors including the PGN precursors lipid I and lipid II, the WTA intermediate lipid III_WTA and C₅₅PP. Cell wall intermediates are fully locked in a complex at a twofold molar excess of 1. No complex formation was observed with C₅₅P. Binding of 1 is indicated by a reduction of the amount of free lipid intermediates visible on the TLC. The chromatograms are representative of two independent experiments.