Expression and purification of the NgLptDE complex
N. gonorrhoeae wild-type full-length LptD (UniProt: Q5F651) and LptE (UniProt: Q5F9V6), with a hex-histidine tag on the C-terminus, were cloned into the expression vector pBAD22A (graciously provided by Professor Yihua Huang). NgLptDE complex was cotransformed in SF100 E. coli cells (graciously provided by Professor Yihua Huang). After transformation, preculture was started, and cells were grown overnight at 37 °C with 100 µg/ml ampicillin and 50 µg/ml kanamycin. Large LB-broth culture was then inoculated at a starting OD600nm = 0.05 and grown at 37 °C with 100 µg/ml ampicillin. When OD600nm = 0.8, temperature was switched to 20 °C and induction was performed at OD600nm = 1 with 0.4% arabinose. After overnight induction at 20 °C, cells were harvested and resuspended in lysis buffer consisting of 200 mM NaCl, 50 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 10 µg/ml DNAse I, 100 µg/ml AEBSF [4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride], and protease-inhibitor cocktail (Complete, Roche). Cells were cracked by multiple passages through a microfluidizer system using a pressure of 18,000 psi, and the lysate was centrifuged at 7500 × g for 10 min to remove the cell debris. The supernatant was collected, and inner membranes were solubilized by incubation with 2% Triton X-100 for 30 min at 4 °C under gentle agitation. The outer-membrane fraction was collected by centrifugation at 100,000 × g for 30 min at 4 °C. The pellet containing the outer-membrane fraction was resuspended in 200 mM NaCl, 50 mM Tris-HCl, pH 7.5, 20 mM imidazole, and 1% lauryl-dimethylamine oxide (LDAO), supplemented with protease-inhibitor cocktail (Complete, Roche), and incubated under gentle agitation for 12 h at 4 °C. Insoluble material was removed by centrifugation at 100,000 × g for 1 h at 4 °C. The supernatant was incubated in batches with ~5 ml NiNTA resin for 2 h at 4 °C under gentle agitation. The resin was subsequently washed by gravity flow with 10 column volumes (CV) of Wash buffer A (200 mM NaCl, 20 mM Tris-HCl pH 8.0, 20 mM imidazole, 1% LDAO), 10 CV of Wash buffer B (150 mM NaCl, 20 mM Tris-HCl, pH 8.0, 40 mM imidazole, and 0.5% LDAO), 10 CV of Wash buffer C (150 mM NaCl, 20 mM Tris-HCl, pH 8.0, 40 mM imidazole, and 0.2% LDAO), and 10 CV of Wash buffer D (150 mM NaCl, 20 mM Tris-HCl, pH 8.0, 40 mM imidazole, and 0.1% lauryl maltose neopentyl glycol (LMNG)). Elution was performed with 5 CV of Elution buffer (150 mM NaCl, 20 mM Tris-HCl pH 8.0, 300 mM imidazole, 0.01% LMNG). Eluted material was desalted against 150 mM NaCl, 20 mM Tris-HCl, pH 8.0, and 0.005% LMNG using disposable PD-10 desalting columns (GE Healthcare) and was concentrated in 100 kDa centrifugal concentrators (Millipore). Concentrated sample was further purified by SEC on a Superdex 200 Increase column, equilibrated to 150 mM NaCl, 20 mM Tris-HCl, pH 8.0, and 0.005% LMNG. The peak fractions corresponding to the monomeric or dimeric LptDE complex were kept separately, concentrated to ~1 mg/ml, and used for sybody generation, as well as biophysical and structural studies.
Protein used for sybody generation and GCI characterization
N. gonorrhoeae (Zopf) Trevisan (ATCC 700825) LptD (1–801)-3C-His10 and LptE (1–159)–AVI complex was produced using E. coli SF100 cells and purified as described above, and an in vitro biotinylation step was added after the SEC step34. Briefly, LptDE protein was concentrated to 10 µM concentration and mixed with 40 µg of BirA E. coli enzyme, 5 mM ATP, 10 mM MgAc, and 15 µM biotin. The mixture was incubated for 16 h at 4 °C and a second SEC step was performed to desalt the sample and remove the BirA enzyme and free biotin in the following buffer: 20 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 0.005% LMNG. Fractions corresponding to the monomeric complex peak were pooled, concentrated, and flash-frozen in liquid nitrogen for subsequent use.
Sybody generation
Sybodies were generated as described previously24,35 with one notable difference. NgLptDE could not be produced in large amounts, which did not allow to use a large excess of nonbiotinylated NgLptDE for off-rate selection. Therefore, we pooled all the sybodies present after the first round of phage display and used them as competitors during the second round of phage display to perform an off-rate selection. To this end, the pDX_init plasmid outputs (of the concave, loop, and convex library) of the first round of phage display were purified by miniprep (Qiagen). FX cloning was performed to transfer the sybody pool from the pDX_init to the pSb_init expression plasmid using 2 µg of pDX_init pool and 1 µg of pSb_init. The cloning reaction was subsequently transformed into electrocompetent E. coli MC1061 cells (>10 mio cfu). The sybody pools were expressed as described for single sybodies using the pSb_init construct35. After expression of the pools in 600 ml cultures of TB medium, the sybodies were extracted from the cells by periplasmic extraction, purified by IMAC, and dialyzed overnight against Tris Buffered Saline (TBS). Precipitation was removed by centrifugation at 20,000 × g for 15 min. The pools were used at a concentration of approximately 100 µM to perform an off-rate selection for 2 min in the second round of phage display.
Fluorescent labeling of sybodies
To perform site-specific labeling of the sybodies, a glycine located between SapI-restriction site and myc tag on the pSb_init backbone was mutated to cysteine via Quick change mutagenesis, thereby adding the following amino acids to the C-terminus of the sybody: GRACEQKLISEEDLNSAVDHHHHHH. The sybodies were expressed and purified as previously described35, except that 1 mM DTT was added to all buffers used for purification. Subsequently, DTT was removed and the sybody was rebuffered to degassed PBS using a PD10 desalting column and immediately mixing the sybody with Alexa Fluor 647 C2 maleimide (ThermoFisher Scientific) at a molar ratio of 1:3.6. The labeling reaction was carried out for 1 h at 4 °C. Excess label was removed by desalting the labeled sybody with a PD10 column.
Cellular-binding assay
For cellular-binding assays, overnight cultures of E. coli SF100 cells with and without overexpression of NgLptDE were used. The number of cells was normalized by adjusting 1 ml of culture to an OD600 of 3. The cells were harvested by centrifugation washed three times with 500 µl of PBS containing 0.5% BSA (PBS–BSA), and subsequently blocked for 20 min in the same buffer. After an additional wash with 500 µl of PBS–BSA, the cells were incubated for 20 min in 100 µl of PBS–BSA containing 1 µM of the Alexa Fluor 647-labeled sybodies. After three washes with 500 µl of PBS, cells were resuspended in 100 µl of PBS and transferred to a microtiter plate with nontransparent walls. Fluorescence was measured in a plate reader with excitation of 651 nm and emission of 671 nm.
Antibiotic-susceptibility assay
N. gonorrhoeae (Zopf) Trevisan (ATCC 700825) was streaked from a glycerol stock on blood agar and incubated for 24 h at 37 °C with 5% CO2 atmosphere. Colonies were scraped off the agar and resuspended in Fastidious broth at a density of McFarland 0.536. The cells were further diluted 1:100 in Fastidious broth. In 96-well plates, dilution series of vancomycin with and without sybodies in Fastidious broth were prepared and mixed with the diluted culture. The plates were incubated without shaking at 37 °C with 5% CO2 atmosphere for 24 h. About 100 µl of a 0.04 mg/ml resazurin stock solution in PBS was added to the cells and incubated for one hour at 37 °C with 5% CO2 atmosphere. Fluorescence was measured at 571 nm excitation and 585 nm emission.
Pro-Macrobody generation
Pro-Macrobodies (PMbs) were produced in E. coli as described earlier for the original macrobodies20. Briefly, two PCR-amplified fragments, the specific sybody and the C-terminal MBP, were cloned simultaneously into the expression vector pBXNPH3M (Addgene #110099)37,38,39 using FX cloning40. N-terminally of the resulting PMb insert, the plasmid expresses a pelB leader sequence followed by a deca-His tag, an MBP, and a 3C-protease site. The insert is fused during cloning through an overlapping proline-encoding CCG codon introduced by reverse and forward primers at the 3′- and 5′-end of the sybodies and MBP, respectively, and released by digestion with the type IIS restriction enzyme SapI (NEB). The second proline of the linker is encoded in the forward primer of the MBP (3′ of the overlapping CCG codon) and replaces the natural lysine. The resulting amino acid sequence of the linker is VTVPPLVI (VTV is the conserved C-terminus of sybodies, PP in italics/bold denotes the linker and underlined the truncated N-terminus of processed E. coli malE starting at Leu7). For the sybodies 21 and 51, to convert them into Pro-Macrobodies, forward primers 5′-TA TATAGCTCTTCTAGTCAGGTTCAGCTGGTTGAGAGCGGTGGTGGCC-3′ (Sb21), 5′-TATATAGCTCTTCT AGTCAAGTCCAGCTGGTGGAATCGGGTGGTGGTAG-3′ (Sb51), and reverse primers 5′-TATATA GCTCTTCTCGGCACAGTCACTTGGGTACCTTGG-3′ (Sb21) and 5′-TATATAGCTCTTCTCGGAACGGTAACTT GGGTGCCCTG-3′ (Sb51) were used. MBP for PMbs was amplified from pBXNPHM3 with forward primer 5′-TATATAGCTCTTCTCCGCCTCTGGTAATCTGGATTAACGG-3′ and reverse primer 5′-TATATAGCTCTTCTTGCA CCCGGAGTCTGCGCGTCTTTC-3′. The resulting amplicons for the sybodies and MBP, respectively, were mixed in a 1:1 molar ratio and cloned into pBXNPHM3. PMbs were expressed in terrific broth in MC1061 E. coli cells at 37 °C by induction with 0.02% arabinose at an OD600 = 0.7. After 3.5 h, cells were harvested and resuspended in lysis buffer consisting of 150 mM NaCl, 50 mM Tris-HCl pH 8, 20 mM imidazole, 5 mM MgCl2, 10% glycerol, 10 µg/ml DNAse I, and protease inhibitors (Complete, Roche). Cells were cracked and the lysate was centrifuged for 30 min at 147,000 × g in a Beckman 45Ti rotor. Subsequently, the supernatant was incubated in batch with ~3 ml NiNTA resin/6 L of culture. The resin was washed with 150 mM KCl, 40 mM imidazole, pH 7.6, and 10% glycerol and eluted with 150 mM KCl, 300 mM imidazole, pH 7.6, and 10% glycerol. The N-terminal MBP with the deca-His tag was removed by cleavage with 3 C protease overnight during dialysis against 150 mM KCl, 10 mM Hepes–NaOH, 20 mM imidazole, pH 7.6, and 10% glycerol. After removal of the His-tagged MBP by Re-IMAC, the unbound material was concentrated in 50 kDa centrifugal concentrators (Millipore) and subjected to SEC on a Superdex 200 Increase column (equilibrated to 150 mM NaCl, 10 mM Hepes–Na, pH 7.6) connected to an AKTA system (GE) operated with Unicorn software v7.1. The eluted fractions with the PMbs were supplemented with 20% glycerol, concentrated to 3–8 mg/ml, and aliquots were flash-frozen in liquid nitrogen for subsequent use. Macrobody versions of sybodies 21 and 51 with the original VK-linker were expressed and purified in the same way.
SEC analysis/purification of LptDE–PMb complexes
To identify ternary and quaternary complexes of NgLptDE with various PMbs against NgLptDE, monomeric NgLptDE was first separated from dimers by SEC on a Superdex 200 Increase 10/300 column in 150 mM NaCl, 20 mM Tris-Cl, pH 8, and 0.005% LMNG. A complex was formed by addition of 3- to 4-fold molar excess of PMb51 to NgLptDE sample. The uncomplexed and PMb-bound samples were analyzed on an Agilent 1260 Infinity II HPLC operated with Chemstation 2.12.26 using a Superdex 200 Increase 5/150 column and Trp-fluorescence detection after 10–15 min incubation time on ice. Binding was indicated by a shift to earlier elution volumes and increase in the UV absorption and Trp fluorescence. The quaternary complex was formed by adding PMb21 in 3- to 4-fold molar excess to the preformed NgLptDE–PMb51 complex. The quaternary NgLptDE–PMb51–PMb21 complex showed the most distinct shift and fluorescence increase and was therefore chosen for scale-up and cryo-EM analysis.
Binding analytics by grating coupled interferometry
Initial screen with ELISA-positive sybodies was performed at 20 °C on the Wave delta instrument from Creoptix in Tris 25 mM pH 7.5, NaCl 300 mM, and DDM 0.1% as running buffer. Biotinylated protein was immobilized on a 4PCP-S (streptavidin) chip, conditioned with 1 M NaCl, 0.1 M sodium borate, at levels between 600 and 700 pg/mm2 to avoid any mass-transport limitation. One injection of each sybody (200 nM) was performed and binding responses were evaluated with the Wave control software. The ten best sybodies in terms of slower Koff as well as quality of binding signals obtained were characterized further with dose–response analysis in order to determine accurate kinetic parameters. For each ten-sybody, 8 concentrations were recorded (serial 2-fold dilution) in duplicate injections, and equilibrium as well as kinetic data were analyzed and fitted with 1:1 model. Fits were of high quality (black curves) and recapitulate the experimental data (red curves). At higher concentration of sybodies, bulk (RI) effects could be observed, but this effect did not influence data analysis. PMb21 showed a slightly decreased affinity (3 to 4-fold) compared with the respective sybody 21. This loss can be explained by a faster Koff of the PMb21.
Sample preparation and cryo-EM data acquisition
Quantifoil (1/2) 200-mesh copper grids were glow-discharged for 20 sec prior to sample freezing. About 3 µl of NgLptDE–PMb51–PMb21 complex at a concentration of 1 mg/ml were placed on the grid, blotted for 3.0 s, and flash-frozen in a mixture of liquid propane and liquid ethane cooled with liquid nitrogen using a Vitrobot Mark IV (FEI) operated at 4 °C and under 100% humidity.
The EM data-collection statistics in this study are reported in Supplementary Table 1. Data were recorded on a FEI Titan Krios transmission-electron microscope, operated at 300 kV and equipped with a Quantum-LS energy filter (slit width 20 eV, Gatan Inc.) containing a K2 Summit direct electron detector. Data were automatically collected using the software SerialEM41. Dose-fractionated exposures (movies) were recorded in electron-counting mode, applying 60 electrons per square angstrom (e−/Å2) over 45 frames, or 50 e−/Å2 over 35 frames for, respectively, the NgLptDE (apo) or the NgLptDE–PMb51–PMb21 samples. A defocus range of −0.8 to −2.8 µm was used and the physical pixel size was 0.64 Å/pixel for the NgLptDE and 0.82 Å/pixel for the NgLptDE–PMb51–PMb21 datasets. Recorded data were online-analyzed and preprocessed using FOCUS42, which included gain normalization, motion correction, and calculation of dose-weighted averages with MotionCor243, as well as estimation of micrograph defocus with CTFFIND444.
Image processing
The following processing workflows were used for the samples in the study. The aligned movies were imported into CryoSPARC V245. A set of aligned averages with a calculated defocus range of −0.6 to –3.0 μm was selected from which averages with poor CTF-estimation statistics were discarded. Automated particle picking in CryoSPARC V2 resulted in 815,057 particle locations for the NgLptDE–PMb51–PMb21 sample. After several rounds of 2D classification, 490,743 particles were selected and subjected to 3D classification using the multi class ab initio refinement process (5 classes, 0.4 similarity) and heterogeneous refinement. The best-resolved class consisting of 184,206 particles was finally subjected to 3D nonuniform refinement. The overall resolution of the resulted map was estimated at 3.40 Å based on the Fourier shell correlation (FSC) at 0.143 cutoff46. To visualize the LptD–NTD-dimerization interface, another round of 3D heterogeneous refinement was performed and a subset consisting of 80,140 particles was selected. Those particle coordinates used re-extraction of particle images with an increased box size. Particles were recentered by 2D classification in order to process the dimeric LptDE complex. Ab initio reconstruction and nonuniform refinement on this set of particles resulted in a DIMERNgLptDE–PMb51–PMb21 map with an overall resolution of 5.27 Å. As for the opened state, a multiclass ab initio refinement process and heterogeneous refinement was performed for the NgLptDE–PMb51–PMb21 sample. Of the five 3D classes, one class consisting of 93,151 particles was further refined by computationally removing the density corresponding to the detergent micelle with the particle-subtraction tool within CryoSPARC V2, followed by 3D local refinement. The resulting map had an estimated overall resolution of 4.72 Å as judged by FSC at 0.143 cutoff. Analysis of the apo NgLptDE was performed similarly to the NgLptDE-PMb21-PMb51. Briefly, a set of aligned averages with a calculated defocus range of −0.6 to –3.0 μm was selected, from which averages with poor CTF-estimation statistics were discarded. Automated particle picking in CryoSPARC V2 resulted in 1,395,392 particle locations for the NgLptDE. After several rounds of 2D classification, 196,182 particles were selected and subjected to 3D classification using the multiclass ab initio refinement process (3 classes, 0.1 similarity) and heterogeneous refinement for the 2 best classes. The best-resolved class consisting of 119,115 particles was finally subjected to 3D nonuniform refinement. The overall resolution of the resulted map was estimated at 4.6 Å based on the Fourier shell correlation (FSC) at 0.143 cutoff.
Model building and refinement
An initial LptDE model was generated using SWISS-MODEL47, using as templates the KpLptD structure (PDB-ID 5IV9) and the EcLptE structure (PDB-ID 4RHB). The template for building the PMb21 coordinates into the EM map was based on the X-ray structure that was solved specifically for this study (PDB-ID 7OMT). The same structure was also used to model the maltose-binding protein region of PMb51.
Rigid-body fitting was initially done in Chimera48 followed by manually rebuilding of the model in Coot49. The remaining clashes between sidechains were detected using Schrodinger version 2019-4 (Maestro, Schrödinger, LLC, New York, NY, 2021), and remodeled using prime50. Manual inspection of missing H-bonds in the model was used to refine sidechain positions. Finally, real-space refinement was performed in Phenix version 1.17–3644, applying Ramachandran plot restraints51.
Molecular-dynamics simulations
Possible linkers connecting the VHH to MBP were assessed by all-atom MD simulations. After 300 ns of equilibration, conformational space was explored by conducting 500 ns MD trajectories initiated from a structure solved with a (Val122 and Lys123) VL linker (PDB-ID 6HD8), as well as the same set of coordinates, but modeled with an engineered (Pro122–Pro123) PP linker. Both trajectories were later aligned on the VHH structure (residues 1–120) to show the larger conformational flexibility of the VK linker. The macrobody interdomain (VHH to MBP) angle was also monitored during MD trajectories. This angle was defined as the angle between the geometrical centers of residue 1–120 (Nb), residue 122–123 (linker), and residue 124–486 (MBP). All MD simulations were conducted in explicit water, at room temperature (300 K, or 26.85 °C), and with Desmond standard parameterization52 and the OPLS3e force field53.
Crystallization and structure determination of PMb21
Purified PMb21 in 150 mM NaCl, 10 mM HEPES–Na, pH 7.5, concentrated to 10 mg/ ml and supplemented with 2.5 mM D-(+)-maltose, was subjected to crystallization screens. The protein crystallized in 0.2 M MgCl2, 0.1 M Hepes, pH 7.5, and 30% PEG400. For cryoprotection, crystals were soaked in 0.2 M MgCl2, 0.1 M Hepes, pH 7.5, 35% PEG400, and 2.5 mM D-(+)-maltose. Crystals were flash-frozen in liquid N2 and data were collected at the PXII (X10SA) beamline of the Swiss Light Source (SLS, Villigen).
The dataset was integrated and scaled with XDS (built=20190806). The structure was solved by PHASER version 2.8.3 (CCP4 7.0.077) using MBP (PDB 1N3W) and a nanobody (PDB 5FWO) as molecular replacement-search model. The structure was initially refined using REFMAC version 5.8.0257 (CCP4 7.0.077) and later with PHENIX version 1.16, together with iterative model building in COOT using 2FoFc and FoFc map.
Figure preparation
Figures were prepared using the programs Chimera X (http://www.rbvi.ucsf.edu/chimerax/54), Chimera (http://www.cgl.ucsf.edu/chimera/48), PyMOL (http://www.pymol.org/) and Schrödinger (www.schrodinger.com).
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
