Designed α-sheet peptides suppress amyloid formation in Staphylococcus aureus biofilms

Abstract

Nosocomial infections affect hundreds of millions of patients worldwide each year, and ~60% of these infections are associated with biofilm formation on an implanted medical device. Biofilms are dense communities of microorganisms in which cells associate with surfaces and each other using a self-produced extracellular matrix composed of proteins, polysaccharides, and genetic material. Proteins in the extracellular matrix take on a variety of forms, but here we focus on functional amyloid structures. Amyloids have long been associated with protein misfolding and neurodegenerative diseases, but recent research has demonstrated that numerous bacterial species utilize the amyloid fold to fortify the biofilm matrix and resist disassembly. Consequently, these functional amyloids, in particular the soluble oligomeric intermediates formed during amyloidogenesis, represent targets to destabilize the extracellular matrix and interrupt biofilm formation. Our previous studies suggested that these amyloidogenic intermediates adopt a non-standard structure, termed "α-sheet", as they aggregate into soluble oligomeric species. This led to the design of complementary α-sheet peptides as anti-α-sheet inhibitors; these designs inhibit amyloidogenesis in three unrelated mammalian disease-associated systems through preferential binding of soluble oligomers. Here we show that these anti-α-sheet peptides inhibit amyloid formation in Staphylococcus aureus biofilms. Furthermore, they inhibit aggregation of pure, synthetic phenol soluble modulin α1, a major component of Staphylococcus aureus functional amyloids. As it aggregates phenol soluble modulin α1 adopts α-helix then α-sheet and finally forms β-sheet fibrils. The binding of the designed peptide inhibitors coincides with the formation of α-sheet.

Fig. 1

Screening of designed peptides for inhibition of amyloid formation in S. aureus biofilms. a Schematic of protocol for testing designed α-sheet peptides in S. aureus biofilm cultures with an illustration of the main-chain structure of the AP90 α-sheet design. b A panel of designed α-sheet peptides (AP90, AP401, and AP407), as well as random coil (RC) and β-hairpin controls (β), was tested against two S. aureus strains, MN8 (hashed bars) and SH1000 WT (white bars). ThT fluorescence fluorescence values indicate the extent of amyloid formation in the EM and are shown as the percent of peptide-free control conditions. c Dose-response curve for the designed peptide AP90 against S. aureus MN8 biofilms reveals a significant decrease in EM amyloid content as the concentration of AP90 is increased. Error bars in b and c represent the standard error of the mean for experiments performed in triplicate

Fig. 2

S. aureus biofilm structures become less robust when grown in the presence of designed peptide inhibitors. a S. aureus MN8 + mCherry biofilms were grown on glass substrates for 24 h and then cells were washed and fixed. The peptide inhibitors (80 μM) caused cells to detach during the wash step. Images are representative of triplicate wells. For quantification of the images see Supplementary Fig. S2. b In S. aureus SH1000 WT biofilms grown in regular LB medium, PSM amyloid fibrils were visible as deposits in spaces between cells (left). Upon addition of the designed peptide AP90 (80 μM), no extracellular fibril deposits were observed (right)

Fig. 3

CD measurements capture structural transitions of PSMα1. a CD spectra of PSMα1 samples (30 μM, 1.3% HFIP, 50 mM potassium phosphate buffer, pH 5.5) were taken periodically during aggregation. At early time points (t = 0, 48, 84 h), negative peaks at ~208 and ~220 nm represent α-helical secondary structure. At intermediate time points (t = 130 h), featureless spectra indicate formation of α-sheet, and by the end of the time course (t = 188 h) a negative peak at ~218 nm signals the presence of β-structure. b Close-up view of characteristic CD spectra for α-helix (0 h, purple), α-sheet (red, 130 h), and β-sheet (green, 188 h). c Aggregation of synthetic PSMα1 peptide (30 μM, same conditions as for CD) was tracked over time by ThT fluorescence in a microtiter plate. Error bars in c represent the standard deviation of the mean of four samples

Fig. 4

AFM images of synthetic PSMα1 amyloid fibrils. a PSMα1 peptide samples solubilized with DMSO were allowed to aggregate at high concentrations (437 μM), and the resulting fibrils exhibited extensive surface coverage, with each fibril measuring ~10 nm in diameter and 0.1–4 μM in length. b PSMα1 was also allowed to aggregate at 44 μM, yielding comparable but less dense fibrils

Fig. 5

Aggregation of synthetic PSMα1 is inhibited by designed α-sheet peptides. a Synthetic PSMα1 peptide (30 μM, 0.34% DMSO, pH 5) was allowed to aggregate alone and in the presence of AP90 (1:4 molar ratio) and RC (1:4 molar ratio). Two different solvent conditions were used (0.34% DMSO in water, blue bars; LB medium + 0.34% DMSO, red bars) and aggregation was monitored by ThT fluorescence. Inhibition values for each peptide are reported as a percentage of the peptide-free control samples (0% inhibition). Error bars represent the standard error of the mean of 3–6 replicates. b ThT fluorescence curves monitoring the aggregation kinetics of PSMα1 under the two different solvent conditions in panel a (water = blue curve; LB medium = red curve). c The contribution of individual LB medium components (10 g/L peptone, 5 g/L yeast extract, and 85 mM NaCl) to PSMα1 aggregation kinetics was also investigated. Fluorescence values in NaCl solution were quite high, so its curve is shown as an inset. All solutions contain 0.34% DMSO for consistency and solubilization of PSMα1. Values in b and c are averages of 3 samples, corrected by relevant PSMα1-free controls, with error bars to represent the standard deviation of the mean and curves are corrected by subtracting blanks lacking PSMα1

Fig. 6

Designed α-sheet peptides preferentially bind α-sheet-rich PSMα1 over fresh or fibrillar PSMα1. a Synthetic PSMα1 peptides (30 μM, 0.34% DMSO, pH 5) were allowed to aggregate as in Fig. 5, and matched samples were removed periodically from the plate for binding assessment using an agarose bead, resin-based assay and biolayer interferometry. Error bars represent the standard deviation of six samples. b In the resin-based assay, AP193-functionalized beads preferentially bound α-sheet rich PSMα1 (48 h) over earlier time points (0 and 24 h). Error bars represent the standard error of the mean over six samples. c In biolayer interferometry experiments, the equilibrium dissociation constant, K_D, indicates preferential binding between AP90 and α-sheet-rich PSMα1 (~48 h) as opposed to α-helix-rich (~0 h) or β-sheet-rich states (~150 h). Note that 150 h is still in the sigmoidal region of the transition and some α-sheet is present