NMR reveals the interplay between SilE and SilB model peptides in the context of silver resistance

SilE and SilB are both proteins involved in the silver efflux pump found in Gram-negative bacteria such as S. typhimurium. Using model peptides along with NMR and CD experiments, we show how SilE may store silver ions prior to delivery and we hypothesize for the first time the interplay between SilB and SilE.

shift perturbation (CSP), where a series of 1 H-15 N HSQC spectra have been recorded for each peptide while adding small volumes of silver stock solution (390 mM AgNO 3 ). No labelled peptides have been used for this study and NMR spectra were acquired at 15 N natural abundance. To counteract the low abundancy of nitrogen, sofast 1 H-15 N HSQC with 140 scans per spectra were recorded. The combination of homonuclear and heteronuclear experiments recorded at different concentrations of silver allowed us to assign completely 1 H-15 N spectra of each peptide. Those spectra are used as fingerprint of the two peptides for the interaction studies.
CD experiments. CD experiments were acquired at 293 K on a Chirascan spectrometer (Applied Photophysics). Peptide concentration was 10 µM. Each sample was prepared in 10 mM acetate ammonium solution at pH 6.4. Five repetitions have been recorded for each ratio peptide / Ag + .
Mass spectrometry. Peptides were diluted (100-500 μM) in 10 mM solution of ammonium acetate at pH 7.0 and 7.8 and silver nitrate was added (0-4 eq.). Mass spectra at pH 6.4 were also run for consistency and the same complexes were observed. Due to the low signals of the complexes compared to protonation, no spectra at this pH are shown.
Sequence alignment. The alignment has been performed by PRALINE 1-2 . The scoring scheme works from 0 for the least conserved alignment position, up to 10 for the most conserved alignment position. The peptide sequence SilB 401-430 has been displayed using a purple rectangle on each alignment.          Table S2. Overview of the binding constants obtained by 1 H NMR spectroscopy in HEPES buffer (20 mM, pD=7.8) for the two different peptides SilB-p1 and SilB-p2. The binding constants were derived from the titration curves obtained in Fig. S6,7 by means of a 1:1 model (see Appendix 1 below).

Peptide
K d SilB-p1 8 ± 2 µM SilB-p2 2 ± 1 µM Figure S8. CD experiments during the silver titration of SilB-p (10 µM) in ammonium acetate solution (10 mM, pH = 6.4). Five repetitions have been recorded for each ratio.  Figure S9. Hypothetical mechanism of the interplay between SilB and SilE derived from our observations. At low silver concentration, the C-terminus of SilB may accommodate two silver ions with a possible further transfer to SilC. When the silver concentration significantly increases, the system triggers a rapid remodeling with SilE acting as a regulator to avoid saturation of the efflux pump.

Appendix 1. Dissociation constants calculation
To derive the corresponding binding constant for the SilE-p/Ag + interaction, CSPs were analyzed by calculating the combined amide chemical shift perturbation () as =[(( H ) 2 +( N /5) 2 )/2] 1/2 . By considering a two sites interaction and a 2:1 stoichiometry, we assume that two silver ions can bind to either of the two SilE-p binding sites. The perturbations observed on SilE-p do not discriminate between the two sites so that the observed chemical shift perturbation is a weighted average between the two extreme values corresponding to the free (=0) and ligand-bound state (= LB ). For a 2:1 binding model, considerations based on partitioning between the free and various ligand-bound states of SilE-p give : where [P 0 ] and [L 0 ] are the total molar concentrations of SilE-p and Ag + respectively.
The dissociation constant K d and  LB were fitted with non-linear regression by using an inhouse Matlab (The MathWorks, Inc) based program. Errors were estimated by sampling 100 initial guesses, assuming 10% error on the protein and ligand concentrations.
To derive the dissociation constant corresponding to the SilB-p/Ag + interaction, we have used a two sites sequential model that can be described by: where P stands for the free protein, PL 1 the partly bound protein and PL 2 the totally bound protein. The two dissociation constants can be written as: Assuming fast exchange, the chemical shift perturbation can be rewritten as:   p P  P  p PL 1  PL 1  p PL 2  PL 2 where p i is the corresponding populations of the different complexes. The concentrations of the different complexes are: where P T stands for the total protein concentration.  can be recast as: where the two extreme values correspond to the free ( P =0) and totally bound protein (= PL2 ) and [L] is obtained by solving the following cubic equation: L