PMC

Electrostatic potential surfaces of wild type Efb-C (A), N138A mutant (B), R131A mutant (C), and C3d (D). Iso-surfaces are shown at ±1 kT/e. Red indicates negative charge and blue indicates positive charge. The molecules are drawn as ribbons.

Energetic contribution of individual residues to the binding energy of the C3d/Efb-C complex: (A) C3d, (B) wild-type Efb-C, (C) Efb-C N138A, (D) Efb-C R131A. Notice the difference in scale in the y-axis. Notable contributions are indicated in the figures. Only the C3d molecule of the wild-type Efb-C/C3d complex is shown in panel A. The C3d molecules of the N138A and R131A mutants were omitted, as they show a very similar pattern.

Evolution of the RMSD and potential energy of the different components of the wild type and the two mutant complexes during the simulation. (A) Potential energy plot of the three complexes during the simulation. The RMSD of C3d (B) and Efb-C (C) were measured with respect to the minimized C3d and Efb-C of each one of the complexes, respectively. Hydrogen atoms were not included in the measurement. There is a difference in scale on the y-axis between A, B, and C.

Characterization of the binding properties between C3d and three Efb-C proteins (wild type, R131A, and N138A) by ITC (A) and SPR (B–D). (A) Calorimetric determination of the binding enthalpy and affinity by injecting Efb-C proteins (150 μM) into C3d (12 μM). (B–D) Kinetic profiling of Efb-C wild type (B), R131A (C), and N138A (D) by injecting a twofold C3d dilution series (0.27–22 nM) over immobilized Efb-C proteins (215–240 RU). Black lines represent processed SPR binding signals that have been fitted to Langmuir 1:1 kinetic models (red lines). Binding constants from both the ITC and SPR experiments are summarized in .

Contribution of pairwise interactions to the binding energy of the native C3d/Efb-C. Each panel depicts the interactions involving one of the three binding loops of C3d, i.e., residues 1029–1050 (A), 1089–1098 (B), and 1157–1166 (C). Favorable interactions are displayed in different shades on the red–blue scale with blue shades indicating lowest energy (see color bar). While interactions with energies higher than −1.5 kcal/mol are not shown, some very low energy values were truncated at −8 kcal/mol up to allow clearer display. (D) Some of the favorable interactions between C3d (pink) and wild-type Efb-C (green). These interactions are marked in boldface in .

Impact of alanine mutations at Efb-C residues Arg-131 and Asn-138 on the crystal structure of the Efb-C/C3d complex. (A) Cartoon representation of the cocrystal (PDB accession code 2GOX) between human C3d (pale blue) and wild-type Efb-C (spectrum). The three helices of Efb-C (α1, α2, α3) and the sites of mutation (R131 and N138; stick representation) are highlighted in the structure. (B) Aligned backbone traces of the crystallized C3d complexes with Efb-C wild type (blue) as well as mutants R131A (green) and N138A (red). No significant changes in the overall structure could be detected between the three complexes, suggesting that the mutations do not induce any significant conformational changes at the Efb-C/C3d interface. Snapshots of the Efb-C conformations in the equilibrated state (blue) and after 20 nsec of the simulation (red): (C) wild type, (D) R131A, and (E) N138A. Although the entire complex was simulated, only Efb-C is shown here for clarity, as C3d exhibited relatively little change during the simulation.

Electrostatic titration experiment using SPR. C3d (22 nM) was injected over immobilized Efb-C proteins (wild type, R131A, and N138A) in 10 mM phosphate buffer (pH 7.4) of increasing sodium chloride concentration (75, 150, 300, 600 mM NaCl). The influence of buffer ionicity is clearly visible in the absolute signal intensity of the SPR response at constant C3d concentration (left panels). In order to better visualize the differential effect on association and dissociation phase, the signals have been normalized at injection end (300 sec) by adjusting the maximum intensity to 100 RU (right panels). In the case of R131A, the curve at highest salt concentration has not been normalized due to the lack of detectable binding.