LptM promotes oxidative maturation of the lipopolysaccharide translocon by substrate binding mimicry

Abstract

Insertion of lipopolysaccharide (LPS) into the bacterial outer membrane (OM) is mediated by a druggable OM translocon consisting of a β-barrel membrane protein, LptD, and a lipoprotein, LptE. The β-barrel assembly machinery (BAM) assembles LptD together with LptE at the OM. In the enterobacterium Escherichia coli, formation of two native disulfide bonds in LptD controls translocon activation. Here we report the discovery of LptM (formerly YifL), a lipoprotein conserved in Enterobacteriaceae, that assembles together with LptD and LptE at the BAM complex. LptM stabilizes a conformation of LptD that can efficiently acquire native disulfide bonds, whereas its inactivation makes disulfide bond isomerization by DsbC become essential for viability. Our structural prediction and biochemical analyses indicate that LptM binds to sites in both LptD and LptE that are proposed to coordinate LPS insertion into the OM. These results suggest that, by mimicking LPS binding, LptM facilitates oxidative maturation of LptD, thereby activating the LPS translocon.

Fig. 1. LptM interacts with the LPS translocon.

a The envelope fractions of wild-type (WT, lptM⁺) and ΔlptM transformed with pLptDE^His cells were solubilized using the mild detergent DDM and subjected to Ni-affinity purification. The purified OM LPS translocon was analyzed by SDS-PAGE (lanes 1 and 2) and BN-PAGE (lanes 3 and 4) followed by Coomassie Brilliant Blue staining. Data are representative of three independent experiments. b DDM-solubilized LptM^His purified from ΔlptM cells transformed with pLptDEM^His was analyzed by SDS-PAGE and Coomassie Brilliant Blue staining. Data are representative of three independent experiments. c Native mass-spectrometry analysis of LptDE^His and of LptDEM^His purified as in (a) and (b), respectively. A schematic diagram of each complex is shown for each spectrum; LptD in gray, LptE in yellow, LptM in purple. Top: LptDE^His formed a heterodimeric complex of 108,425 ± 11 Da (LptD, theoretical mass of the mature protein = 87,080 Da; LptE^His, theoretical mass of the mature and tri-acylated protein = 21,348 Da). Bottom: LptDEM^His formed a heterotrimeric complex of 114,325 ± 14 Da (LptD, theoretical mass of the mature protein = 87,080 Da; LptE, theoretical mass of the mature and tri-acylated protein = 20,248 Da; LptM^His, theoretical mass of the mature and tri-acylated protein = 6,988 Da). The mass difference between the two complexes is 5,900 Da ± 25 Da, corresponding to the mass of mature tri-acylated LptM 5888 Da. d ΔlptM cells transformed with pLptDEM^His derivative (*) plasmids that carry pBpa at distinct positions in LptD were subjected to UV irradiation as indicated prior to DDM-solubilization and Ni-affinity purification of LptM^His. Upon SDS-PAGE of the elution fractions, proteins were revealed by Coomassie Brilliant Blue staining or Western blotting using an antiserum against LptA. Data are representative of three independent experiments. e Logoplot representation of the LptM amino acid sequence in Enterobacteriaceae. The plot was obtained from the multiple alignment of 37 amino acid sequences of LptM in representative genomes of a restricted Enterobacteriaceae family reported in Supplementary Fig. 4 (see also Supplementary Information). The location of PF13627 is indicated above the logoplot.

Fig. 2. LptM facilitates assembly of the LPS translocon by the BAM complex.

a Drop dilution growth test of wild-type E. coli and the indicated derivative strains deleted of bamB, lptM or both under the indicated conditions. vanc, vancomycin; SDS, sodium dodecyl sulfate. b The DDM-solubilized and purified LPS translocon obtained from lptM⁺ or ΔlptM cells harboring pLptDE^His was analyzed by Coomassie Brilliant Blue staining to visualize the bait protein (lanes 1 and 2) used for Western blotting with the indicated antisera to identify co-eluted proteins (lanes 3-6). Load: 1.8%; Elution: 100%. Data are representative of three independent experiments.

Fig. 3. LptM promotes LptD oxidative maturation.

a Schematic representation of mature LptD primary sequence indicating the position and oxidation state of Cys residues in LptD^Red (top, left), LptD^C1-C2 (middle, left) or fully oxidized LptD^Ox (bottom, left). A typical migration pattern corresponding to different oxidative states of LptD by non-reducing SDS-PAGE is represented on the right. b The total protein contents of wild-type (lptM⁺) and ΔlptM strains transformed with an empty vector pCtrl, and the complementation strain ΔlptM transformed with pLptM^His were separated by non-reducing SDS-PAGE and analyzed by Western blotting using the indicated antisera. **Indicates a non-identified protein band. Data are representative of three independent experiments. c The total protein contents of wild-type and the indicated mutant strains were separated by reducing or non-reducing SDS-PAGE and analyzed by Western blotting as in (b). Red, empty arrowheads indicate reduced LptD (LptD^Red), whereas filled, gray arrowheads indicate LptD^C1-C2. **Indicates a non-identified protein band that migrates slightly faster than LptD^Red. Data are representative of three independent experiments. d Purification of LptE^His from ΔlptM cells that overproduced LptDE^His (lane 1) or from wild-type cells that overproduced LptDME^His (lane 2) or ΔlptM cells that overproduced the indicated LptD mutant variant of LptDE^His (3-5). *Refers to the LptD Cys-to-Ser variant: wild-type LptD^CCCC, LptD^CCSS, LptD^CSCS or LptD^SCSC, as specified on the top of each gel lane. Data are representative of three independent experiments. The signals of any LptD forms in lanes 1 and 2 were quantified. The amount of intermediate forms (LptD^Red + LptD^C1-C2 + LptD^C2-C4) was normalized to the overall amount of all LptD forms (LptD^Red + LptD^C1-C2 + LptD^C2-C4 + LptD^ox). Data are represented as means ± s.e.m. (n = 3 independent experiments). Source data are provided as Source Data file. e LptDE^His containing either wild-type (LptD^CCCC) or LptD^CCSS co-overproduced together with LptM or alone in wild-type or ΔlptM cells were Ni-affinity purified and resolved by BN-PAGE, prior to Coomassie Brilliant Blue staining. *Indicates the plasmid-encoded LptD variant, wild-type LptD^CCCC, LptD^CCSS, as specified on the top of each gel lane. Data are representative of three independent experiments.

Fig. 4. LptM functions synergistically with the oxidative folding machinery.

a Drop dilution growth test of E. coli wild-type and the indicated derivative strains lacking lptM, dsbA or both under different conditions as indicated. vanc, vancomycin; SDS, sodium dodecyl sulfate. b LPS was extracted from the indicated strains, resolved by SDS-PAGE and silver stained. Results are representative of four independent experiments. c Drop dilution growth test of the indicated derivative strains lacking lptM, dsbC or both transformed with pDsbC on media lacking or supplemented with 0.02% arabinose (ara). d Drop dilution growth test of lptM⁺ or ΔlptM strains transformed with the indicated plasmids pLptD*E^His or pLptD*ME^His on media lacking or supplemented with 100 μM IPTG. * indicates variants of plasmid-encoded LptD, the wild-type LptD^CCCC or the LptD^CCSS variant, as specified on the top of each dilution test. e Schematic representation of the LptD oxidative maturation pathway. In wild-type cells the four cysteines of LptD are oxidized stepwise: LptD^Red is first oxidized to form LptD^C1-C2. Disulfide bond shuffling in the latter generates LptD^C2-C4. A final event of oxidation forms a disulfide between C1 and C3, thus generating LptD^Ox. LptM acts synergistically with the disulfide bond formation machinery in facilitating proper oxidation of LptD. In the absence of LptM (ΔlptM, red arrow), a considerable fraction of LptD accumulates as off-pathway oxidation intermediates, most prominently LptD^C1-C2. In this mutant, the assembly of LptD together with LptE stalls at the BAM complex and DsbC becomes essential for cell viability, suggesting that disulfide bond isomerization helps to rescue LPS translocon assembly.

Fig. 5. LptM binds the membrane-embedded portion of the LPS translocon, influencing LPS-interaction sites.

a View of top ranking AlphaFold2 model of LptDE-LptM. LptD is shown as gray cartoon, LptE as yellow cartoon, and the LptM backbone is shown as purple spheres. b Zoom in on the LptD hinge between the β-taco and β-barrel in either the LptDE (top) or LptDE-LptM (bottom) systems. Image from a snapshot following 500 ns of MD simulation. Several interactions are made between the β-taco and β-barrel in LptDE but broken when LptM is present (purple spheres and green sticks), including between the residues shown. c View of the LptD lateral gate with quantification of hydrogen (H)-bond number from 3 × 500 ns simulations for the LptDE and LptDE-LptM systems using gmx hbond. Representative H-bonds are shown as dashed red lines, as computed using VMD. d Differential deuterium uptake between the LptDE translocon in the presence or absence of LptM after statistical analysis with Deuteros, showing in blue regions that are significantly protected upon LptM binding. e View of the AlphaFold2 model plus modeled LptM tri-acylation, showing the position of the LptM N-terminus and tri-acylation (purple spheres and green sticks, respectively) in relation to the LptD 212-240 peptide, which is protected from deuteration by LptM in the HDX-MS experiments. Residues from the 212-240 peptide which are in contact with LptM (<0.3 nm) are shown as lines.