Factor H-Inspired Design of Peptide Biomarkers of the Complement C3d Protein

Abstract

C3d is a hallmark protein of the complement system, whose presence is critical to measure the progression of several immune diseases. Here, we propose to directly target C3d through small peptides mimicking the binding of its natural ligand, the complement regulator Factor H (FH). Through iterative computational analysis and binding affinity experiments, we establish a rationale for the structure-based design of FH-inspired peptides, leading to low-micromolar affinity for C3d and stable binding over microsecond-length simulations. Our FH-inspired peptides call now for further optimization toward high-affinity binding and suggest that small peptides are promising as novel C3d biomarkers and therapeutic tools.

Figure 1

(A) Binding modes of Factor H (FH) to the complement protein C3d, as from crystallographic structures. FH is composed of the Complement Control Proteins (CCP) 19 and 20, which bind at three distinct sites of C3d: Site 1, referred as the “acidic patch”, Site 2, and Site 3. C3d is shown as gray ribbons, while FH is shown highlighting the CCP19 (orange) and CCP20 (violet) in different colors. A close-up view of the three sites is reported, showing critical interactions. (B) Change in free energy of binding for the C3d–FH complex (compared to the wild-type) upon alanine-scan of FH (top) and C3d (bottom panel). Two X-ray structures are used for analysis: 3OXU and 2XQW, considering several chains. Negative values (red) correspond to gain of binding upon alanine mutations, while positive values (blue) correspond to loss of binding. Adapted with permission from ref (10). Copyright 2015 Wiley-VCH Verlag GmbH & Co. KGaA.

Figure 2

Thermophoretic response of fluorescently labeled C3d is observed at varying concentrations of the S1P1 (left) and S1P2 (right) peptides (sequence on top of the graphs). For S1P1, the dose–response curve suggests binding, but is not fully formed, due to low affinity of S1P1 for C3d. The dissociation constant (K_d) is estimated at 2.1 ± 1.6 mM. For S1P2, the dose–response curve is fully formed and K_d is estimated as 61.6 ± 20.0 μM. Units of the inset for K_d are M.

Figure 3

(A) Root Mean Square Fluctuations (RMSF) for the S1P1 peptide residues, computed over ∼1 μs of MD of the peptide in solution and within C3d. (B) Conformations of S1P1 bound to C3d, as from MD simulations. Representative snapshots from the three most populated clusters are shown (cluster populations are in brackets). (C) Time evolution along MD of critical ionic interactions, which act as “ionic anchors” and are established between the R3–R6 residues of S1P1 and negatively charged residues of C3d.

Figure 4

(A) Conformational space adopted by C3d (gray), S1P1 (green), and S1P2 (blue) during MD simulations, as detected by plotting the two first Principal Components (PC1 vs PC2). (B) Conformations of S1P2 bound to C3d, as from MD simulations. Representative snapshots from the two most populated clusters are shown (cluster populations are in brackets). Cluster 1 is superposed to FH, highlighting the overlap of S1P2 with FH.