Biosynthetic tailoring of microcin E492m: post-translational modification affords an antibacterial siderophore-peptide conjugate

Abstract

The present work reveals that four proteins, MceCDIJ, encoded by the MccE492 gene cluster are responsible for the remarkable post-translational tailoring of microcin E492 (MccE492), an 84-residue protein toxin secreted by Klebsiella pneumonaie RYC492 that targets neighboring Gram-negative species. This modification results in attachment of a linearized and monoglycosylated derivative of enterobactin, a nonribosomal peptide and iron scavenger (siderophore), to the MccE492m C-terminus. MceC and MceD derivatize enterobactin by C-glycosylation at the C5 position of a N-(2,3-dihydroxybenzoyl)serine (DHB-Ser) moiety and regiospecific hydrolysis of an ester linkage in the trilactone scaffold, respectively. MceI and MceJ form a protein complex that attaches C-glycosylated enterobactins to the C-terminal serine residue of both a C10 model peptide and full-length MccE492. In the enzymatic product, the C-terminal serine residue is covalently attached to the C4' oxygen of the glucose moiety. Nonenzymatic and base-catalyzed migration of the peptide to the C6' position affords the C6' glycosyl ester linkage observed in the mature toxin, MccE492m, isolated from bacterial cultures.

Figure 1

Structures of enterobactin (Ent), linearized enterobactin (lin-Ent), monoglycosylated enterobactin (MGE), linearized MGE (lin-MGE) and a truncated depiction of microcin E492m (MccE492m) showing the last eleven amino acids of the microcin peptide. Diglycosylated (DGE) and triglycosylated enterobactin (TGE) have glucose moieties attached to the C5 positions of either two or three DHB-Ser units, respectively.

Figure 2

(a) The MccE492 gene cluster: mceA, 312 bp; mceB, 288 bp; mceC, 1113 bp; mceD, 1245 bp; mceE, 345 bp; mceF, 540 bp; mceG, 2097 bp; mceH, 1242 bp; mceI, 492 bp; mceJ, 1575 bp (ref. 1). (b) Amino acid sequences for the serine-rich C-termini of MccE492, MccH47, MccI47 and MccM. The sequence of the C-terminus of MccE492 was determined experimentally. MccH47, MccI47 and MccM have not been isolated and their C-terminal sequences were deduced from the respective genes.

Figure 3

SDS-Page gel (4–15% Tris-HCl) of purified His₆ fusions of MceC (1, C-terminal His₆ fusion), MceD (2, C-terminal His₆ fusion; 3, N-terminal His₆ fusion) and MceIJ (4, MceJ bears a N-terminal His₆ tag).

Figure 4

MceC C-glycosylates Ent. (a) Representation of the MceC-catalyzed converstion of Ent to MGE and DGE. (b) HPLC analysis of the MceC-catalyzed glycosylation of Ent. (c) Kinetic traces for MceC-catalyzed glycosylation of Ent, MGE and lin-Ent. The corresponding k_cat and K_m values are listed in Table 1.

Figure 5

MceD hydrolyzes apo Ent, MGE and DGE. (a) HPLC analysis of MceD-catalyzed hydrolysis of Ent. Dimer and monomer refer to the DHB-Ser dimer and DHB-Ser monomer, respectively, (b) HPLC analysis of MceD-catalyzed hydrolysis of MGE. Glc-Dimer is the monoglycosylated DHB-Ser dimer and “Monomers” refer to the Glc-DHB-Ser (elution time ~ 2 min) and DHB-Ser (elution time ~2.5 min) monomers, the former of which results from hydrolysis of the Glc-DHB-Ser dimer. (c) Kinetic traces for the MceD-catalyzed hydrolysis of Ent, lin-Ent, MGE, lin-MGE, and DGE. Corresponding kinetic parameters are listed in Table 2. All reactions were carried out at room temperature (75 mM Hepes, pH 7.5). (d) Schematic of the MceD-catalyzed hydrolysis of MGE illustrating the regioselectivity of the cut as determined by 2-D NMR. HPLC traces for assays with DGE are given in Figure S2 and the m/z data for Ent, MGE and DGE hydrolysis products are listed in Table S1.

Figure 6

HPLC monitoring of a reaction containing 550 μM C₁₀, 100 μM MGE and 2 μM MceIJ, 5 mM ATP and 5 mM MgCl₂ (75 mM Tris-HCl pH 8, 2.5 mM TCEP). The two new peaks, 1 and 2, both have a mass equivalent to C₁₀+MGE-H₂O, which suggests attachment of MGE to the C₁₀ peptide through an ester linkage.

Figure 7

(a) Section of the 2D-¹H homonuclear NOESY spectra for C₁₀-C4′-MGE (peak 1). (b) Section of the 2D-¹H homonuclear NOESY spectra for C₁₀-C6′-MGE (peak 2). Both spectra were acquired using a W5 pulse sequence for water suppression and a 250 msec mixing time. The dotted lines indicate the positions of the sugar protons and the corresponding NOE correlations. The proton assignments are listed in Table S2. (c) Analytical HPLC (220 nm absorption) of purified C₁₀-C6′-MGE. (d) Analytica HPLC trace (220 nm absorption) of purified C₁₀-C4′-MGE. A gradient of 0 to 40% MeCN in 8 min was employed for both samples, (e) Numbering scheme for the glucose and DHB moieties.

Scheme 1

(a) Attachment of MGE to the MccE492m C-terminal decapeptide (C₁₀) by MceIJ through the C4′ hydroxyl and subsequent migration to the C6′ position, (b) Formation of the six-membered intermediate between the C4′ and C6′ positions.

Scheme 2

Maturation of MccE492m as defined by in vitro studies. MceA is the peptide precursor to MccE492. It undergoes cleavage at amino acid 15 or 19 during export to yield the 84-residue MccE492(m).