Page 47 - Read Online
P. 47

Page 6 of 17              Mezhyrova et al. Microbiome Res Rep 2023;2:28  https://dx.doi.org/10.20517/mrr.2023.28

               either stained or directly thawed bacterial culture was pipetted on the agarose pad and a cover slip was
               placed on top of the pad and fixated with fimo clay. Confocal fluorescence images were acquired with a laser
               scanning microscope (Zeiss LSM 700) using a Plan-Apochromat 63 × 1.40 Oil DIC objective and Zeiss LSM
               software. LED lasers [488 nm for green fluorescent protein (GFP), 555 nm for mScarlet and PI] were used at
               0.2%-4% intensity and the signal was captured with a PMT detector using a SP555 or LP560 filter,
               respectively. Gain settings were individually adjusted for each sample, ranging between 500 and 900. The
               pinhole was set to 1 airy unit (45 μm) and images were acquired with a pixel size of 99 nm and a pixel dwell
               time of 1.58 μs. To maximize the signal-to-noise ratio, averaging of two scans was performed line by line. All
               images were processed with the ImageJ software.


               Cryo-electron microscopy sample preparation and data collection
               MS2-L toxin expression in E. coli Lemo21 (DE3) cells was induced with 1 mM IPTG at OD  = 0.27. 2 mL
                                                                                             600
               of cells were taken 60 min after induction, harvested, and resuspended in 250 µL LB-medium. 3.5 µL of
               sample was deposited onto glow-discharged Quantifoil R3.5/1, 200-mesh Cu holey carbon coated grids and
               vitrified in liquid ethane using a Vitrobot (FEI, Eindhoven, Netherlands) at 100% humidity and 4 °C. Data
               was collected with a Titan Krios transmission electron microscope (FEI, Eindhoven, Netherlands) operating
               at 300  kV with a post-column energy filter in zero-loss peak mode (Gatan Inc., GIF Quantum, Pleasanton,
               California). Electron micrographs were recorded with the software SerialEM v3.8.0beta24 on a K2 Summit
               direct detector (Gatan Inc., GIF Quantum, Pleasanton, California).


               RESULTS
               CF expression and membrane insertion of MS2-L derivatives
               CF expression eliminates problems with toxic effects upon overproduction of MS2-L and similar toxins. In
               addition, the synthesized hydrophobic proteins can already be co-translationally inserted into provided lipid
               bilayers such as NDs. The combination of CF expression and ND technology thus enables the generation of
               purified samples of toxins inserted into native-like lipid environments for in vitro biochemical analysis. The
               coding sequences of the full-length toxin MS2-L as well as of a series of five nested deletions of the
               N-terminal soluble domain starting with deletion of the first 22 amino acids up to position 35 were cloned
               into pET expression vectors [Supplementary Figure 1]. For MS2-L, a derivative containing a C-terminal
               GFP fusion was additionally constructed. Furthermore, an N-terminal mScarlet-MS2-L fusion protein was
               designed. All constructs contained either C-terminal His  or StrepII-tags for purification and immunoblot
                                                               10
               detection.


               The peptides were first CF synthesized in the precipitate forming (P-CF) mode without the supply of
               hydrophobic agents. All proteins were efficiently expressed, with estimated synthesis yields of between 0.5
               and 0.8 mg protein per mL of reaction mixture (RM) as determined by SDS-PAGE analysis [Supplementary
               Figure 2]. The generated precipitates could be solubilized in the relatively harsh detergent 0.75%
               1-myristoyl-2-hydroxy-sn-glycero-3-[phospho-rac-(1-glycerol)] (LMPG), while milder detergents such as
               DPC or n-dodecyl-β-D-maltopyranoside (DDM) were ineffective. We therefore proceeded to co-
               translationally solubilize the synthesized proteins in the presence of various supplied detergents in the
               detergent-based (D-CF) mode. The peptides were synthesized in the presence of 0.4% of either Brij78,
               Brij35, or glyco-diosgenin (GDN). The MS2-L derivatives were almost completely soluble in Brij78 and
               approx. to 70% in Brij35 and to 40% in GDN, respectively. Brij78 was therefore selected for further studies
               using D-CF synthesized MS2-L derivatives.

               In a third approach, MS2-L and MS2-Lp  were synthesized in the presence of supplied preformed
                                                     23
               nanodiscs (NDs). The NDs were preformed with either DMPC or DMPG in vitro and added to the CF
   42   43   44   45   46   47   48   49   50   51   52