Characterization of the Key Determinants of Phd Antitoxin Mediated Doc Toxin Inactivation in Salmonella

In the search for novel antimicrobial therapeutics, toxin-antitoxin (TA) modules are promising yet underexplored targets for overcoming antibiotic failure. The bacterial toxin Doc has been associated with the persistence of Salmonella in macrophages, enabling its survival upon antibiotic exposure. After developing a novel method to produce the recombinant toxin, we have used antitoxin-mimicking peptides to thoroughly investigate the mechanism by which its cognate antitoxin Phd neutralizes the activity of Doc. We reveal insights into the molecular detail of the Phd–Doc relationship and discriminate antitoxin residues that stabilize the TA complex from those essential for inhibiting the activity of the toxin. Coexpression of Doc and antitoxin peptides in Salmonella was able to counteract the activity of the toxin, confirming our in vitro results with equivalent sequences. Our findings provide key principles for the development of chemical tools to study and therapeutically interrogate this important class of protein–protein interactions.


S4
The pET28-FLAG vector was generated by cloning the sequence for the FLAG tag into pET28 upstream of the multiple cloning site. The sequence of Phd 1-73 protein was PCR-amplified from S. Typhimurium genomic DNA and cloned into pET28-FLAG vector using NheI (5' end) and BamHI (3' end) restriction sites. For expression of Phd 1-73 protein variants and Phd 52-73 peptides in the growth rescue experiments, respective Phd sequences were cloned into pCA24N vector 2 using MfeI (5' end) and NotI (3' end) restriction sites. For the co-expression of EF-Tu with Doc, the EF-Tu sequence was PCR-amplified from S. Typhimurium genomic DNA and cloned into modified pBAD33 vector using SacI (5' end) and HindIII (3' end) restriction sites and standard DNA ligation. A ribosomal binding site (RBS) was subsequently inserted into the pBAD33 plasmid upstream of the cloned EF-Tu sequence using 5' phosphorylated primers. For generation of the EF-Tu-His protein construct, the PCR-amplified EF-Tu sequence was cloned into the pET24a vector using NdeI (5' end) and XhoI (3' end) restriction sites. Generation of the EF-Tu T383A mutant in both pBAD33 and pET24a plasmids was accomplished by PCR amplification of the EF-Tu insert using the respective forward cloning primer and a designated mutagenesis reverse primer containing the T383A mutation, followed by reinsertion of the amplified sequence into empty, modified pBAD33 plasmid and pET24a plasmid using SacI/HindIII (pBAD33) and NdeI/XhoI (pET24a) restriction sites, respectively. EF-Tu T383V mutant in both pBAD33 and pET24a plasmids was accomplished by amplification into a linear vector using non-overlapping primers (just the forward primer containing the mutation). The linear strand was phosphorylated at the 5' ends and ligated using standard protocols, followed by a transformation where successfully mutated vector was selected after DNA sequencing of colonies. Details of each protein construct and plasmid, as well as the used primer sequences, can be found in the ESI file section 3.
S5 grow overnight at 37 °C. Successfully transformed cells were then selected from the observed colonies.
Before introducing the plasmid containing the Doc protein sequence, these cells were made chemically competent again using the following procedure: colonies from the previous transformation were used to inoculate 10 mL of LB medium (100 µg/mL of Amp). After shaking overnight at 37 °C the resulting high-density culture was transferred to a 2 L conical flask containing 1 L of LB medium (100 µg/mL of Amp). The

Growth curves in E. coli
Chemically competent BL21-AI E. coli cells were transformed with the following set of plasmids using standard protocols, generating in total four transformants: 1) pET28 Doc; 2) pET28 Doc + pBAD EF-Tu; 3) pET28 Doc + pBAD EF-Tu T383A ; 4) pET28 Doc + pBAD EF-Tu T383V . Successfully transformed cells were used to inoculate 15 mL of LB medium with suitable antibiotics (ESI Table S6) in a 50 mL conical flask and left shaking overnight at 37 °C. S7 significantly improved reproducibility. This flask was left in an incubator overnight at 30 °C shaking at 250 rpm. The temperature was then increased to 37 °C and the flask was left shaking at same speed. Typically, after 1.5-2.5 h the bacterial culture reached an OD600 of 1.0-1.5. At this stage the temperature was lowered to 18 °C and IPTG was added to a final concentration of 0.5 mM. The cells were left shaking at 250 rpm overnight and were harvested by centrifugation at 3000 x g (30 minutes at 4 °C).
The cell pellets resuspended in ice-cold IMAC binding buffer (5 mL per gram of cell paste), followed by the addition of cOmplete™ mini EDTA-free protease inhibitor cocktail (Roche, 1 tablet per 50 mL of lysate), Benzonase® Nuclease 250 U/µL (Sigma-Aldrich, 1 µL per 100 mL of lysate) and MgCl2 to a final concentration of 5 mM. This suspension was left shaking for 30 minutes and the cells were lysed with two passages at 25 kpsi and 5 °C in a CF1 Cell Disrupter (Constant Systems Ltd.) coupled with a ThermoFlex1400 (ThermoFisher Scientific) cooling unit. The lysate was clarified by centrifugation at 18,000 rpm (45 minutes at 4 °C) in a Sorvall LYNX 4000 superspeed centrifuge (ThermoFisher Scientific) with a Fiberlite™ F21-8 x 50y fixed-angle rotor attached (ThermoFisher Scientific).
Using an ÄKTA pure (Cytiva) protein purification system, the supernatant was loaded into a 5 mL HiTrap TALON® crude column (Cytiva) previously equilibrated with the IMAC binding buffer. Two or three columns were attached in tandem when lysate volumes exceeded 100 mL to avoid Doc precipitation after elution (poor solubility when too concentrated). After the sample application was completed, the column was washed with 3-5 column volumes (CV) of 0.5% (v/v) of IMAC elution buffer in IMAC binding buffer. Bound proteins were eluted with a 0.5-100% (v/v) gradient of IMAC elution buffer in IMAC binding buffer over 8 CV, with Doc typically being eluted at 25-75% (v/v) of the gradient as a broad peak with low absorbance at 280 nm (A280).
Doc containing fractions (verified by SDS-PAGE) were combined, transferred to a 3.5K MWCO SnakeSkin™ Dialysis Tubing (ThermoFisher Scientific) and dialysed overnight in a beaker containing 2 L of Doc IEC binding buffer. The sample was filtered (0.2 µm) and loaded in a 5 mL HiTrap Q HP anion exchange column (Cytiva) previously equilibrated with the Doc IEC binding buffer. After the sample loading, the column was washed with 10 CV of Doc IEC binding buffer and bound proteins were eluted with a 0-75% gradient of Doc IEC elution buffer in Doc IEC binding buffer over 20 CV. In these conditions, Doc does not bind to the column and is collected with high purity during the sample application and column wash steps, while most of the impurities were collected in the elution step. Attempts to increase the pH to retain S8 Doc in the column for the elution step were not preferred as those led to protein with poor activity in the biophysical assays.
Doc containing fractions were concentrated using a 3000 MWCO PES Vivaspin® 20 centrifugal concentrator (Sartorius) to approximately 25-50 µM, filtered (0.2 µm) and loaded in a HiLoad 16/600 Superdex 75pg column (Cytiva) previously equilibrated with Doc SEC buffer. After an isocratic elution of Doc SEC buffer over 1 CV, Doc containing fractions were eluted with high purity in between retention volumes of 75-90 mL. Using the manufacturer's calibration curves, the eluted sample corresponded to a molecular weight of approximately 15 kDa, showing that the Doc protein was obtained in a monomeric form (expected 14.8 kDa). The lysis and IMAC purification steps were performed using the same procedures, reagents and buffers as described for Doc. The Phd protein eluted at 10-65% (v/v) of the gradient as a broad peak and fractions were combined. To remove the hexahistidine (His6) tag, 250 µL of 10 kU/mL Thrombin from bovine plasma (Sigma-Aldrich, diluted in water from the lyophilised powder) were added to the pooled fractions. This mixture was transferred to a 3.5K MWCO SnakeSkin™ Dialysis Tubing (ThermoFisher Scientific) and dialysed overnight in a beaker containing 2 L of IMAC binding buffer. After this period the sample was submitted to the same IMAC purification step. Due to the cleavage of the His6 tag, the Phd protein was collected during the sample application and column wash steps, while most of the impurities were collected in the elution step.

S9
Phd containing fractions were concentrated using a 3000 MWCO PES Vivaspin® 20 centrifugal concentrator (Sartorius) to approximately 2.0-5.0 mL, filtered (0.2 µm) and loaded in a HiLoad 16/600 Superdex 75pg column previously equilibrated with Phd SEC buffer. After an isocratic elution of Phd SEC buffer over 1 CV, Phd containing fractions were eluted with high purity in between retention volumes of 60-75 mL. Using the manufacturer's calibration curves, the eluted sample corresponded to a molecular weight of approximately 30 kDa (for an ordered globular protein), showing that the Phd protein (10 kDa) was likely obtained as the The lysis and TALON purification steps were performed using the same procedures, reagents and buffers described for Doc. The protein eluted at 10-65% (v/v) of the gradient as a broad peak. Fractions were combined, transferred to a 3.5K MWCO SnakeSkin™ Dialysis Tubing (ThermoFisher Scientific) and dialysed overnight in a beaker containing 2 L of EF-Tu IEC buffer.
The sample was filtered (0.2 µm) and loaded in a double 5 mL HiTrap Q HP anion exchange column (Cytiva) previously equilibrated with the EF-Tu IEC binding buffer. After the sample loading, the column was washed with 5 CV of EF-Tu IEC binding buffer and bound proteins were eluted with a 0-75% gradient of EF-Tu IEC elution buffer in EF-Tu IEC binding buffer over 10 CV. EF-Tu containing fractions were recovered in the elution phase at 35-50% of the gradient, combined and concentrated using a 3000 MWCO PES Vivaspin® 20 centrifugal concentrator (Sartorius) to approximately 2.0-5.0 mL.

S10
The sample was filtered (0.2 µm) and loaded in a HiLoad 16/600 Superdex 75pg column previously equilibrated with EF-Tu SEC buffer. After an isocratic elution in the same buffer over 1 CV, EF-Tu containing fractions were eluted with high purity in between retention volumes of 55-65 mL. Using the manufacturer's calibration curves, the eluted sample corresponded to a molecular weight of approximately 45 kDa, showing that the EF-Tu proteins were obtained in a monomeric form (expected 44.3 kDa). Fractions were combined and concentrated to approximately 260 µM, flash frozen in liquid nitrogen and stored at -80 °C. For the EF-Tu wild-type and T383V proteins, the approximate yield of per litre of TB medium was, respectively, 16 mg and 5 mg.

Circular dichroism
Lyophilized peptides were dissolved in a 20 mM potassium phosphate buffer (pH 7.4) or in 30% (v/v) trifluoroethanol (TFE) in 20 mM potassium phosphate buffer to concentrations of 25-50 µM. Peptide solutions (300 μL) were transferred into a Hellma QS quartz cuvette (1 mm) and CD spectra were recorded with a Chirascan V100 CD spectrometer (Applied Photophysics) in the 190-260 nm far UV range at 25 ºC with 1 nm band width, 1 nm step size, 1 sec per step and five scans per measurement. In addition, a background measurement of buffer or 30% TFE in buffer was performed. Scans for each measurement were averaged, the averaged spectra were smoothened using Savitsky-Golay smoothing (window size of 5) and the buffer spectra were subtracted from the respective peptide spectra. CD values in mdeg were converted to mean residue ellipticity [θ] (deg. • cm 2 • dmol -1 ) and the peptide spectra were plotted using GraphPad Prism 5.

Supplementary figures and tables
S12 Figure 1, 3, 4, 6, 7, 8, 10, 13, 14, 17, 19, 20, 21, 22, 23, 24 and 25. All peptides were tested at eight concentrations, ranging from 10 µM to 5 nM (3-fold dilutions).    Table S8. Full amino acid sequences of the proteins purified in this work. The desired protein sequence is underlined, mutated residues are highlighted in red, tags are highlighted in purple and cleavage sites are indicated with a slash in the middle of the sequence.

Protein Sequence
Doc 1-122 -His6   Figure S9. Melting curves of DocSTm at 5 µM. Replicates are split into three graphs for better visualisation of the curves.  1500 s, respectively. The surface was regenerated between each cycle and the ligand density on the surface varied between replicates (lowest and highest densities correspond to replicates 1 and 3, respectively) to verify the reproducibility of the results. Due to the slow association rate and mass-transport limitation at higher analyte concentrations, the data was fitted with a constant RMAX, which was calculated based on the RMAX of a positive control (peptide 1).   1.610 ± 0.002 × 10 6 9.520 ± 0.047 × 10 -5 59.1 29.40 ± 0.01 1.670 ± 11.200 × 10 13 0.157 3 4