Human- and plant-pathogenic Pseudomonas species produce bacteriocins exhibiting colicin M-like hydrolase activity towards peptidoglycan precursors

Abstract

Genes encoding proteins that exhibit similarity to the C-terminal domain of Escherichia coli colicin M were identified in the genomes of some Pseudomonas species, namely, P. aeruginosa, P. syringae, and P. fluorescens. These genes were detected only in a restricted number of strains. In P. aeruginosa, for instance, the colicin M homologue gene was located within the ExoU-containing genomic island A, a large horizontally acquired genetic element and virulence determinant. Here we report the cloning of these genes from the three Pseudomonas species and the purification and biochemical characterization of the different colicin M homologues. All of them were shown to exhibit Mg(2+)-dependent diphosphoric diester hydrolase activity toward the two undecaprenyl phosphate-linked peptidoglycan precursors (lipids I and II) in vitro. In all cases, the site of cleavage was localized between the undecaprenyl and pyrophospho-MurNAc moieties of these precursors. These enzymes were not active on the cytoplasmic precursor UDP-MurNAc-pentapeptide or (or only very poorly) on undecaprenyl pyrophosphate. These colicin M homologues have a narrow range of antibacterial activity. The P. aeruginosa protein at low concentrations was shown to inhibit growth of sensitive P. aeruginosa strains. These proteins thus represent a new class of bacteriocins (pyocins), the first ones reported thus far in the genus Pseudomonas that target peptidoglycan metabolism.

FIG. 1.

Sequence alignment of the different colicin M homologues. (A) Sequences of E. coli colicin M (UniProtKB/Swiss-Prot accession number P05820) (Ecoli) and its homologues from P. aeruginosa (UniProtKB/TrEMBL accession number Q1W548) (Paeru), P. fluorescens (GenBank accession number EU982300; Dmitri Mavrodi, personal communication) (Pfluo), and P. syringae (UniProtKB/TrEMBL accession number Q88A25) (Psyri) were aligned using the ClustalW program. Amino acids that are invariant in the four proteins are indicated by the one-letter abbreviation for the amino acid below the sequences, and amino acids that are conserved in three or two of the proteins are indicated by asterisks and colons below the sequences, respectively. Gaps introduced to maximize alignment are indicated by dashes. (B) Percentages of identity of residues in the N- and C-terminal regions (N-ter and C-ter, respectively) of the ColM homologues are indicated for pair-wise comparisons. The black triangle in panel A indicates the limit of these regions that was chosen for this comparison.

FIG. 2.

Purification of ColM homologues from Pseudomonas species. The different ColM homologues were overproduced in His₆-tagged form and were purified by affinity chromatography on Ni²⁺-NTA agarose. The lanes contain ColM from E. coli (Ecoli) (12) and purified preparations of its three homologues from P. aeruginosa (Paeru), P. fluorescens (Pfluo), and P. syringae (Psyri). Lane M contains molecular mass standards, namely, bovine serum albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), and soybean trypsin inhibitor (20 kDa). Staining was performed with Coomassie brilliant blue R250.

FIG. 3.

In vitro degradation of the peptidoglycan lipid II intermediate by ColM-like proteins from Pseudomonas species. The activity of ColM and its different homologues was tested under standard assay conditions using ¹⁴C-lipid II labeled in the GlcNAc moiety as the substrate. Conditions that result in partial (∼20 to 50%) conversion of lipid II into 1-pyrophospho-MurNAc(-pentapeptide)-GlcNAc product were used. The radiolabeled substrate and product were separated by TLC (R_f values were 0.7 and 0.3, respectively,) and spots were visualized with the PhosphorImager. Lane A, lipid II incubated in the absence of ColM; lanes B to E, lipid II incubated in the presence of E. coli ColM (0.4 μg) or one of the ColM homologues from P. aeruginosa (2 ng), P. fluorescens (28 μg), and P. syringae (0.1 μg), respectively.

FIG. 4.

Identification of lipid I degradation product by HPLC. l-[¹⁴C]alanine-labeled lipid I was digested to completion by colicin M and the three homologues from Pseudomonas species, and the reaction mixtures were analyzed by HPLC as described in Materials and Methods. The solid line (detection at 215 nm) depicts analysis of a standard mixture of 1-pyrophospho-MurNAc-pentapeptide (peak 1), 1-phospho-MurNAc-pentapeptide (peak 2), and MurNAc-pentapeptide (peaks 3 and 4; anomeric forms β and α, respectively). The broken line (detection of radioactivity) depicts analysis of the reaction mixture following incubation of ¹⁴C-labeled lipid I with the P. aeruginosa enzyme, showing the formation of a single radiolabeled product comigrating with authentic 1-pyrophospho-MurNAc-pentapeptide. The same results were obtained with ColM and the P. fluorescens and P. syringae homologue proteins, and no radiolabeled product was observed in the absence of enzyme (data not shown).

FIG. 5.

Effect of the ColM homologue from Pseudomonas aeruginosa on the growth of P. aeruginosa DET08. (A) Two microliters of the purified bacteriocin (4 μg) were loaded in duplicate at the surface of a 2YT agar plate that was previously inoculated with the indicator strain P. aeruginosa DET08. Diffusion of the enzyme inhibited growth, resulting in darker zones. (B) P. aeruginosa DET08 was grown at 37°C in 2YT medium, and the purified bacteriocin was added at various concentrations (mg/ml) as indicated by the arrow. OD, optical density.