Peptidoglycan Remodeling Enables Escherichia coli To Survive Severe Outer Membrane Assembly Defect

Abstract

Gram-negative bacteria have a tripartite cell envelope with the cytoplasmic membrane (CM), a stress-bearing peptidoglycan (PG) layer, and the asymmetric outer membrane (OM) containing lipopolysaccharide (LPS) in the outer leaflet. Cells must tightly coordinate the growth of their complex envelope to maintain cellular integrity and OM permeability barrier function. The biogenesis of PG and LPS relies on specialized macromolecular complexes that span the entire envelope. In this work, we show that Escherichia coli cells are capable of avoiding lysis when the transport of LPS to the OM is compromised, by utilizing LD-transpeptidases (LDTs) to generate 3-3 cross-links in the PG. This PG remodeling program relies mainly on the activities of the stress response LDT, LdtD, together with the major PG synthase PBP1B, its cognate activator LpoB, and the carboxypeptidase PBP6a. Our data support a model according to which these proteins cooperate to strengthen the PG in response to defective OM synthesis.IMPORTANCE In Gram-negative bacteria, the outer membrane protects the cell against many toxic molecules, and the peptidoglycan layer provides protection against osmotic challenges, allowing bacterial cells to survive in changing environments. Maintaining cell envelope integrity is therefore a question of life or death for a bacterial cell. Here we show that Escherichia coli cells activate the LD-transpeptidase LdtD to introduce 3-3 cross-links in the peptidoglycan layer when the integrity of the outer membrane is compromised, and this response is required to avoid cell lysis. This peptidoglycan remodeling program is a strategy to increase the overall robustness of the bacterial cell envelope in response to defects in the outer membrane.

Keywords: Escherichia coli; cell envelope; lipopolysaccharide; peptidoglycan; stress response.

FIG 1

LDTs prevent cell lysis upon defective OM assembly. (A to D) Cells of the araBplptC conditional strain (A and B) and the isogenic mutants with ldtD, ldtE, and ldtF deleted (C and D) were grown in the presence of 0.2% arabinose to an OD₆₀₀ of 0.2, harvested, washed three times, and resuspended in an arabinose-supplemented (+ Ara) or arabinose-free (no Ara) medium. (A and C) Growth was monitored by OD₆₀₀ measurements (top panels) and by determining CFU (bottom panels). Growth curves shown are representative of at least three independent experiments. At t = 120, 210, and 270 min (arrows), samples were imaged (araBplptC [B]; isogenic mutant deleted for ldtD, ldtE, and ldtF [D]). Phase-contrast images (top) and fluorescence images (bottom) are shown. Bars, 3 μm. (E) PG sacculi purified from araBplptC cells grown in the presence of arabinose or after 210 min (2) or 270 min (3) growth in the absence of arabinose were digested with cellosyl, and the muropeptide composition was determined by HPLC. The graph shows the relative abundance of TetraTetra (with a 4-3 cross-link) and TetraTri(3-3) (with a 3-3 cross-link) muropeptides. The latter significantly increased upon depletion of LptC. (F) Cells of the araBplptC conditional strain and isogenic mutants deleted for every ldt gene alone or in all possible combinations were grown in an arabinose-free medium as indicated above. Growth phenotypes are summarized as the slope of growth curves measured between 180 and 390 min. Positive and negative values indicate cell growth and cell lysis, respectively. Values are means plus standard deviations (SD) (error bars) from three independent experiments. The mean slope calculated from growth curves in arabinose-supplemented medium for the araBplptC conditional strain and isogenic ldt mutants was 0.56 ± 0.03. The ldt genes are indicated by their loci shown by capital letters.

FIG 2

Ectopic expression of LdtD and LdtE-LdtF results in 3-3 cross-links. (A) Muropeptide profiles of E. coli BW25113Δ6LDT cells containing either no plasmid, empty plasmid (pJEH12), or plasmid with ldtD (pJEH12-ldtD), ldtE (pAMS01-ldtE), ldtF (pAMS02-ldtF), ldtE-ldtF (pAMS01-ldtE and pGS124), or ldtD-ldtF (pJEH12-ldtD and pGS124) grown in the presence of inducer. A.U., arbitrary units. (B) Structures of major peaks numbered in the top chromatogram in panel A. LDT products are muropeptides containing 3-3 cross-links (peaks 4 to 7), tripeptides (peaks 1, 5, and 7) and glycine at position 4 (Gly4, peaks 2 and 4). G, N-acetylglucosamine; M(r), N-acetylmuramitol; L-Ala, L-alanine; D-Glu, D-glutamic acid; D-Ala, D-alanine; m-DAP, meso-diaminopimelic acid. The detected muropeptides with tripeptides or glycine at position 4 (peaks 2 and 4) are typical products of side reactions in PG from cells with active LDTs (due to LD-CPase and Ala-Gly exchange reactions, respectively).

FIG 3

Inhibition of LPS synthesis causes lysis in ldtD deleted cells and activates the ldtD promoter. (A) E. coli BW25113 (left) and BW25113ΔldtD (right) cells were grown in LB-Lennox medium. At t = 150 min, cells were treated with 0.031 µg/ml (1×MIC) of LPC-058 (short arrow) or not treated with LPC-058. Cell growth was monitored by OD₆₀₀ measurements. When cells reached late exponential phase, cultures were diluted to an OD₆₀₀ of 0.1 (long arrow), and growth was monitored by OD₆₀₀ measurements. (B) Cells of araBplptC (left panel) and araBplptC ΔldtD (middle and right panels) were grown in the presence of 0.2% arabinose to an OD₆₀₀ of 0.2, harvested, washed three times, and resuspended in an arabinose-supplemented (+ Ara) or arabinose-free (no Ara) medium. Cell growth was then monitored by OD₆₀₀ measurements. At t = 150 min, cells were treated with 0.006 µg/ml (0.75×MIC) of LPC-058 (short arrow) or not treated with LPC-057, and afterwards growth was monitored by OD₆₀₀ measurements. When araBplptC and araBplptC ΔldtD cells grown in the presence of arabinose and treated with LPC-058 (right panel) reached late exponential phase, the cultures were diluted to an OD₆₀₀ of 0.1 (long arrow), and growth was monitored by OD₆₀₀ measurements. Growth curves shown are representative of at least three independent experiments. (C) BW25113 cells carrying plasmids expressing ldtDp-lacZ and ldtEp-lacZ fusions were grown in LB Lennox broth. At t = 150 min cells were treated with 0.031 µg/ml (1×MIC) LPC-058 or not treated. β-Galactosidase specific activity was determined from cells collected at 210 min (OD₆₀₀ of 0.5), 270 min (60 min after dilution), and 330 min (120 min after dilution). Light gray bars show strain BW25113, and gray bars show strain BW25113 treated with LPC-058. Note that ldtE expression is not affected by LPC-058.

FIG 4

The ldtD promoter is activated under envelope stress conditions, and ldtE and ldtF are RpoS-regulated genes. Wild-type strain BW25113 (lptC⁺) and isogenic mutants with every ldt gene deleted alone and in all possible combinations were transformed with plasmids expressing ldtDp-lacZ (A), ldtEp-lacZ (B), or ldtFp-lacZ (C) fusions. Cells were grown in LD medium. β-Galactosidase specific activity was calculated from cells collected at 120 min (OD₆₀₀ of ∼0.2) (light gray bars), 180 min (OD₆₀₀ of ∼0.8) (gray bars), and 210 min (OD₆₀₀ of ∼2.0) (dark gray bars) (left side). The araBplptC conditional strain was transformed with plasmids expressing ldtDp-lacZ (A), ldtEp-lacZ (B), or ldtFp-lacZ (C). Cells were grown with 0.2% arabinose to an OD₆₀₀ of 0.2, harvested, washed three times, and resuspended in an arabinose-supplemented (+ Ara) or arabinose-free (− Ara) medium. Samples for determination of β-galactosidase specific activity were collected at the time point at which the strains cultivated under nonpermissive conditions arrested growth (white bars) and 30 min (gray bars) and 60 min (black bars) afterwards (+Ara and no Ara conditions, right side). (D) BW25113 ΔrpoS cells carrying plasmids expressing ldtDp-lacZ, ldtEp-lacZ, or ldtFp-lacZ fusions were grown in LD broth. β-Galactosidase specific activity was determined from cells collected at 120 min (OD₆₀₀ of 0.2), 180 min (OD₆₀₀ of 0.8), and 210 min (OD₆₀₀ of 2.0). Strains BW25113 (light gray bars) and BW25113ΔrpoS (gray bars) are shown. Note that ldtD expression is not affected in a ΔrpoS background. The values are the means ± SD from at least three independent experiments. All mutants were also transformed with the empty plasmid, and the mean of β-galactosidase specific activity calculated from cells grown in any condition was 249 ± 30 (min⁻¹mg⁻¹). In panels A to C, the ldt genes are indicated by their loci shown in capital letters.

FIG 5

The GTase activity of PBP1B is required to prevent cell lysis upon defective OM assembly. Cultures of araBplptC ΔmrcA (A) or araBplptC ΔmrcB (C) strains lacking PBP1A and PBP1B, respectively, were grown with 0.2% arabinose to an OD₆₀₀ of 0.2, harvested, washed three times, and resuspended in an arabinose-supplemented (+ Ara) or arabinose-free (no Ara) medium. Cell growth was then monitored by OD₆₀₀ measurements. At t = 120 min, 210 min, and 270 min (arrows 1, 2, and 3, respectively), samples from araBplptC ΔmrcA (B) and araBplptC ΔmrcB (D) strains were collected for imaging. Phase-contrast images (top) and fluorescence images (bottom) are shown. Bars, 3 μm. (E) Complementation of the araBplptC ΔmrcB lysis phenotype by ectopic expression of wild-type mrcB (GT⁺TP⁺), mrcB with mutated GTase (GT*), TPase (TP*), or both (TP*GT*). All mutants were grown in the presence of 0.2% arabinose at 30°C to an OD₆₀₀ of 0.2, harvested, washed three times, and resuspended in an arabinose-free medium. The growth of the araBplptC ΔmrcB strain in arabinose-supplemented medium is shown as a control. Cell growth was monitored by OD₆₀₀ measurements. Growth curves shown are representative of at least three independent experiments.

FIG 6

The DD-CPase PBP6a prevents cell lysis upon defective OM assembly. Cells of the araBplptC ΔdacA (A) or araBplptC ΔdacC (C) strain lacking PBP5 or PBP6a, respectively, were grown in the presence of 0.2% arabinose to an OD₆₀₀ of 0.2, harvested, washed three times, and resuspended in an arabinose-supplemented (+ Ara) or arabinose-free (no Ara) medium. Cell growth was then monitored by OD₆₀₀ measurements. Growth curves shown are representative of at least three independent experiments. At t = 120 min, 210 min, and 270 min (arrows 1, 2, and 3, respectively), samples from araBplptC ΔdacA (B) and araBplptC ΔdacC (D) strains were collected for imaging. Phase-contrast images (top) and fluorescence images (bottom) are shown. Bars, 3 μm.

FIG 7

LdtD interacts with PBP1B in vitro and in vivo. (A) Coomassie blue-stained SDS-PAGE gel showing the pulldown of proteins to Ni²⁺-NTA beads. LdtD bound to the beads and was present in the elution fraction (lanes E) only in the presence of oligohistidine-tagged PBP1B, and not in the presence of oligohistidine-tagged LpoB, oligohistidine-tagged PBP1A, or in the absence of another protein. A, applied sample. (B) Microscale thermophoresis curves showing that LdtD interacts with PBP1B and not with PBP1A. The K_D value for the LdtD-PBP1B interaction is indicated. Values are means ± SD from three independent experiments. (C). BW25113 ldtD-his and araBplptC ldtD-his cells grown with and without arabinose (Ara) were treated with the DTSSP cross-linker. Cell-free extract was prepared, and LdtD-His was purified onto a Ni-NTA resin. PBP1B, LdtD-His, and LptE (as loading control) were immunodetected after SDS-PAGE and Western blotting. WT, wild type. (D) In vivo interaction between PBP1B and LdtD-His by cross-linking/coimmunoprecipitation assay. araBplptC ldtD-his cells were treated with cross-linker DTSSP. The membrane fraction was prepared, and PBP1B was precipitated by specific antibody (the control sample received no antibody). LdtD-His was detected by Western blotting using specific anti-oligohistidine-tag antibody. M, membrane extract; S, supernatant; E, elution.

FIG 8

LdtD shows LD-TPase activity with different PG substrates. (A) HPLC chromatograms showing the formation of TetraTri(–3) dimers by LdtD incubated with glycan chains harboring monomeric tetrapeptides (DS-tetra chains) or PG from BW25113Δ6LDT cells lacking all six ldt genes. Samples were digested with cellosyl and, reduced with sodium borohydride before HPLC analysis. (B) HPLC chromatograms obtained from samples upon incubating [¹⁴C]GlcNAc-labeled lipid II and PG from strain BW25113Δ6LDT and the proteins indicated to the right (I, II, and III indicate the different samples). Samples were digested with cellosyl, reduced with sodium borohydride, and subjected to HPLC analysis with detection of both UV signal (black traces) and radioactivity (red traces). PBP1B (TP*) is PBP1B with an inactive transpeptidase site due to the replacement of Ser-510 by Ala. Tetra-P and Penta-P originate from the hydrolysis of the respective pentapeptide and tetrapeptide versions of lipid II prior to HPLC analysis. (C) Proposed structures of muropeptides present in the fractions in panels A and B. G, N-acetylglucosamine; M, N-acetylmuramic acid; M(r), N-acetylmuramitol; M-P, N-acetylmuramic acid-1-phosphate; L-Ala, L-alanine; D-Glu, D-glutamic acid; D-Ala, D-alanine; m-DAP, meso-diaminopimelic acid.

FIG 9

Role of a PG repair machine. (A) Nonperturbed LPS transport to the OM. (B) Upon LptC depletion, PBP1B-LpoB, LdtD, and PBP6a work in concert to repair the PG, synthesizing it locally with 3-3 cross-links (CL) (red line). Components of the Lpt machine are colored blue and indicated by capital letters.