Solution Structure of CCP Modules 10–12 Illuminates Functional Architecture of the Complement Regulator, Factor H

The 155-kDa plasma glycoprotein factor H (FH), which consists of 20 complement control protein (CCP) modules, protects self-tissue but not foreign organisms from damage by the complement cascade. Protection is achieved by selective engagement of FH, via CCPs 1–4, CCPs 6–8 and CCPs 19–20, with polyanion-rich host surfaces that bear covalently attached, activation-specific, fragments of complement component C3. The role of intervening CCPs 9–18 in this process is obscured by lack of structural knowledge. We have concatenated new high-resolution solution structures of overlapping recombinant CCP pairs, 10–11 and 11–12, to form a three-dimensional structure of CCPs 10–12 and validated it by small-angle X-ray scattering of the recombinant triple‐module fragment. Superimposing CCP 12 of this 10–12 structure with CCP 12 from the previously solved CCP 12–13 structure yielded an S-shaped structure for CCPs 10–13 in which modules are tilted by 80–110° with respect to immediate neighbors, but the bend between CCPs 10 and 11 is counter to the arc traced by CCPs 11–13. Including this four-CCP structure in interpretation of scattering data for the longer recombinant segments, CCPs 10–15 and 8–15, implied flexible attachment of CCPs 8 and 9 to CCP 10 but compact and intimate arrangements of CCP 14 with CCPs 12, 13 and 15. Taken together with difficulties in recombinant production of module pairs 13–14 and 14–15, the aberrant structure of CCP 13 and the variability of 13–14 linker sequences among orthologues, a structural dependency of CCP 14 on its neighbors is suggested; this has implications for the FH mechanism.


Supplementary Data
Table S1 Comparison of new structures with existing CCP structures.

Table S2
Structural statistics of FH 10-12 model calculated according to a combined SAXS and NMR dataset (see Methods).

Figure S1
Comparison of 1 H, and FH 13 (black). The spectrum of FH 13-14 (lower) shows little evidence of folded structure. Only a few cross peaks in the FH 13-14 spectrum correspond to folded FH 13, based on an overlay of the spectra (upper).

Figure S2
Comparison of slowly exchanging amides in 10-11 and 11-12. Slowly exchanging amides for FH [10][11]. In each case, peaks in a 1 H, 15 N HSQC spectrum that are detectable after a 15-minute exposure to D 2 O are identified.

Figure S3
Extrapolating to triple-module structures. (A) CCP 11 of the FH 10-11 structure and CCP 11 from the FH 11-12 structure overlay well. (B) Overlays (orthogonal views) of the filtered average DAMMIF shape envelopes (mesh representation) with the concatenated FH10-12 structural model (data and CRYSOL fits shown in Fig. 6C in main text). (C) CCP 12 of the FH 11-12 structure and CCP 12 from the FH 12-13 structure also overlay well.

Figure S4
Comparisons with the most similar CCP structures in the database Sequence alignments and overlays of NMR-derived structures of FH 10, 11 and 12 with the CCP module structures to which they are most similar (see Table S1); details in text.

Figure S5
Relaxation data for bi-modules. (A) From the top, T 1 , T 2 and heteronuclear NOE values, for FH 10-11, are plotted versus residue number. In each plot the mean, +/-standard deviation is indicated, calculated for residues with heteronuclear NOEs > 0.6. β-strands and the position of the linker are shown by the labeled, shaded boxes. Inverted triangles on the x-axis indicate Pro positions while star shapes denote residues that were excluded from the analysis due to overlap or low signal. (B) As in A, but for FH 11-12. The sequential occurrence of residues with unusual relaxation parameters may be interpreted in terms of motions on various timescales. For example the BD loop of CCP 11 (both in FH 10-11 and FH 11-12 is probably undergoing slow exchange as suggested by low T 2 values (and unremarkable T 1 values). The BD loop is less mobile (ns-timescale) in FH 10-11 (as evidenced by larger heteronuclear NOEs), presumably because it is stabilized by the interface with CCP 10. Likewise, the region just prior to strand D of CCP 12 is characterized by high heteronuclear NOEs, perhaps because it is participating in the interface with CCP 11. Notably, the two intermodular linkers are not markedly more mobile than other regions. Taking all this data together, there is little evidence of intermodular flexibility in these two bi-modules on the time-scales probed by NMR.

Figure S7
An inferred salt bridge at the interface between CCPs 10 and 11. The side chains of participating residues are shown as sticks in the enlargement; cysteine side chains are shown as spheres. Color coding as in Figure 4C in main text.   Table S1. Shown are the results of pair-wise comparisons of FH 10, FH 11 a (where a = from FH 10-11), FH 11 b (where b = from FH 11-12) and FH 12 with all other individual CCPs of known structure within complement proteins based upon alpha-carbon (RMSD) values, using structural alignment program Combinatorial Extension (see Methods). For each CCP, inclusive module boundaries were one residue before Cys-I and the third residue after Cys-IV. Where structures were solved by both NMR and X-ray diffraction, the higher resolution X-ray structure was used for comparison. Where both liganded and unliganded structures were available, the highest resolution unliganded X-ray or NMR structure was used. Colour key used in