Group A streptococci induce stronger M protein-fibronectin interaction when specific human antibodies are bound

Abstract

Group A streptococcus (GAS) is a highly adapted, human-specific pathogen that is known to manipulate the immune system through various mechanisms. GAS' M protein constitutes a primary target of the immune system due to its spatial configuration and dominance on the bacterial surface. Antibody responses targeting the M protein have been shown to favor the conserved C region. Such antibodies (Abs) circumvent antigenic escape and efficiently bind to various M types. The ability of GAS to bind to fibronectin (Fn), a high molecular weight glycoprotein of the extracellular matrix, has long been known to be essential for the pathogen's evolutionary success and fitness. However, some strains lack the ability to efficiently bind Fn. Instead, they have been found to additionally bind Fn via the A-B domains of their M proteins. Here, we show that human Abs can induce increased Fn-binding affinity in M proteins, likely by enhancing the weak A-B domain binding. We found that this enhanced Fn binding leads to a reduction in Ab-mediated phagocytosis, indicating that this constitutes a GAS immune escape mechanism. We could show that the Fc domain of Abs is necessary to trigger this phenomenon and that Ab flexibility may also play a key role. We, moreover, saw that our Abs could enhance Fn binding in 3 out of 5 emm type strains tested, belonging to different clades, making it likely that this is a more generalizable phenomenon. Together our results suggest a novel synergistic interplay of GAS and host proteins which ultimately benefits the bacterium.

Keywords: adaptive immunity; antibodies; co-evolution; fibronectin; group A – beta hemolytic streptococcus; immune subversion.

Figure 1

Convalescent patient plasma enhances Fn-binding to GAS. (A) GAS was incubated in plasma from patients who had recently recovered from a severe GAS infection (convalescent donors) as well as in plasma from healthy controls. This was done at three different dilutions (50, 10, and 1%). All healthy donor plasma, except one, led to greatly reduced Fn-binding to GAS compared to the convalescent donor plasma. Each point represents a distinct donor, and the line shows the respective median. (B) GAS incubated in plasma derived from the original donor of AB25, 32, and 49 led to a dose-dependent increase in bound Fn compared to both untreated bacteria and pooled plasma (MFI of bound Fn displayed in the upper right corner). GAS was incubated in three different plasma dilutions in PBS (50, 10, and 1%). Each histogram corresponds with 20,000 bacterial events as assessed by flow cytometry and they are presented as normalized to mode. All data for this figure was acquired by flow cytometry.

Figure 2

Certain monoclonal Abs induce increased Fn-binding to M1 GAS in an M protein-dependent manner. (A) The M-specific monoclonals Ab25 and 49 lead to a significant increase in Fn-binding to SF370 GAS compared to the untreated control (UT, no additional Ab treatment). The Fc-binding control (IgG Fc ctrl), pooled intravenous Abs (IVIG), and the M-specific monoclonal Ab32 did not. (B) The M1 protein knock-out mutant SF370 ΔM was treated with Fn alone (UT) or in combination with the Fc control mAb (IgG Fc ctrl), IVIG, and the M-specific monoclonals (Ab25, 32, and 49). All Abs and Fn were added at 20 μg/ml. No significant difference in Fn-binding was observed for any of the treatments compared to the untreated control. (C) Lower Ab background levels (80% saliva and 1% serum) result in an increased GAS Fn-binding compared to the untreated control (UT, no additional Ab treatment) when Ab25 and 49 are substituted while this was not seen with other Ab treatments. Higher Ab titers found in 80 and 10% serum reduced the effect substituted Abs had on GAS Fn-binding (all Abs were added at 20 μg/ml). (D) Fn binding affinity increases in the presence of Ab25. Binding curves of Fn with and without 20 μg/ml Ab25 give an estimate of affinity to protein M1. Fibronectin binding was measured using flow cytometry. The figure shows measured binding with fitted ideal binding curves as a function of the total fibronectin concentration. N = 3 for all concentration points. Dissociation constants (K_D) for the curves are given in the plot, together with a confidence interval calculated using the Bootstrap method. (E) A two-site binding model was applied to the same data. High and low estimates for the K_D values are given in the plot. No K_D value was attainable as a low estimate for Fn alone. Normalization for panels (A–C) was performed by dividing the MFI values by the mean MFI of corresponding untreated samples. Each data point for panels (A–C) represents the normalized measurement of a bacterial mid-log culture each grown from a distinct colony. Error bars represent the SEM. Statistical significance was assessed using Kruskal–Wallis combined with Dunn’s multiple comparisons test and *denotes p ≤ 0.05 and **for p ≤ 0.005.

Figure 3

Ab enhanced Fn binding inhibits monocyte phagocytosis. THP-1 cells were incubated with increasing MOPs (multiplicity of prey) of heat-inactivated SF370 bacteria [opsonized with 20 μg/ml of the IgG Fc ctrl, Ab25, or Ab25 + Fn (both 20 μg/ml)]. The THP-1 cells were allowed to associate with and internalize the bacteria for 30 min before flow cytometric analysis. (A) Representative flow cytometry plots for each treatment where cells were exposed to opsonized GAS at an MOP of 40. ‘Internalized’ signifies the percentage of cells with internalized prey and ‘total associated’ signifies percentage of cells with both attached and internalized prey. (B) The curves represent the percentage of cells associated with bacteria as a function of the MOP. The inset to the top left displays the MOP₅₀ for each opsonization condition. R² represents the coefficient of determination for the respective treatment curve. MOP₅₀ CI represents the 95% confidence interval for the respective MOP₅₀ values. (C) MFI values of prey (GAS) associated with THP-1 cells when cells were exposed to 20 GAS per FIGURE 3 (Continued)cell (MOP 20). Ab25 led to a significant increase in associated prey compared to both the IgG Fc ctrl and Ab25 + Fn. (D) The curves represent the percentage of cells with internalized prey. Statistical significance was assessed using one-way ANOVA combined with Dunnett’s multiple comparisons test and * denotes p ≤ 0.05. The curves represent the percentage of cell with internalized prey as a function of the MOP. The presented data represents four independent experiments done on separate days. For panels (A,D) the data points represent the mean and the error bars represent the SEM. For panel (C) the bars represent the mean, data points represent separate experiments, and the error bars represent the SEM. All data was acquired with flow cytometry.

Figure 4

Fn binds to the upper domain in M proteins and Ab flexibility as well as intact Fc domains are necessary for enhanced Fn binding. (A) Fn and fibrinogen (Fg) binding site distance measurements from the bacterial surface are displayed as violin plots. Fn-binding resulting from CAb treatment and low-affinity Fn-binding, achieved by incubating the bacteria in high concentrations of Fn are displayed next to the results from a binding distance control Fg. Both Fn-binding measurements showed that Fn consistently bound below Fg on the M protein. (B) Violin plots showing binding site distance measurements of Fn and corresponding Ab used to induce said Fn-binding (Ab25, Ab49). Regardless of the Ab treatment, the measured Fn-binding distance was very similar and was in both cases larger than the Ab binding distance. (C) Ab25 F(ab’)2 fragments lead to no significant increase in Fn binding compared to the non-binding IgG Fc control (left panel). Full digestion of Ab25 into F(ab’)2 fragments was assessed by SDS page (right panel) (D), Ab25 PFA crosslinked (x-linked) on the surface of GAS led to no significant increase in Fn binding compared to the an Ab untreated control.

Figure 5

Monoclonal Abs induce enhanced Fn-binding in several tested M types. (A) Five GAS strains expressing distinct M types (M1, M5, M12, M79, and M89) were exposed to Fn in combination with Ab treatments. All strains showed variable levels of background Fn-binding when no treatment (UT) or no specific Ab (IgG Fc ctrl) was present. For 3 of the 5 tested strains (M1, M79, and M89), treatment with the monoclonals Ab25 or 49 resulted in a significant increase in Fn-binding compared to the untreated control. All Abs were added at 20 μg/ml. The data for this graph was acquired by flow cytometry and each data point represents the normalized measurement of a bacterial mid-log culture each grown from a distinct colony. (B) Amino acid sequence alignment of the various tested M types. The sequences include the signal peptide (blue), A/B regions (orange), and the binding epitope sequences of Ab25 and 49 are framed in magenta and green, respectively. (C) Illustration depicting the GAS M protein, divided into its various regions whereby not all M types feature an S region. The binding sites of Ab25 and 49 are shown in magenta and green respectively, whereby Ab25 binds to the M protein with both of its Fab domains in dual-Fab cis binding conformation (Bahnan et al., 2022). Site localization analysis (Figures 4A,B) revealed that Fn binds above both Abs. (D) Percent identity matrix showing the similarity of the aligned amino acid sequence of all tested M types. The data is shown as a heatmap. Normalization was performed by dividing the MFI values by the mean MFI of corresponding untreated samples. Error bars represent the SEM. Statistical significance was assessed using one-way ANOVA combined with Dunnett’s multiple comparisons test and *denotes p ≤ 0.05, **for p ≤ 0.005, ***for p ≤ 0.001, and ****for p ≤ 0.0001.

Figure 6

Hypothetical model of antibodies inducing Fn-binding to M protein. Illustration showing a hypothetical binding model of Abs inducing Fn-binding to surface-bound M protein. Specific antibodies are required to bind at a particular distance from the bacterial surface. Intact Ab Fc domains are required to enhance Fn binding. Flexible antibodies increase likeliness of Fc domains being able to interact with Fn. The scale bars represent relative binding distances attained from binding distance measurements (Figure 4B).