Biophysical and structural studies of fibulin-2

Fibulin-2 is a multidomain, disulfide-rich, homodimeric protein which belongs to a broader extracellular matrix family. It plays an important role in the development of elastic fiber structures. Malfunction of fibulin due to mutation or poor expression can result in a variety of diseases including synpolydactyly, limb abnormalities, eye disorders leading to blindness, cardiovascular diseases and cancer. Traditionally, fibulins have either been produced in mammalian cell systems or were isolated from the extracellular matrix, a procedure that results in poor availability for structural and functional studies. Here, we produced seven fibulin-2 constructs covering 62% of the mature protein (749 out of 1195 residues) using a prokaryotic expression system. Biophysical studies confirm that the purified constructs are folded and that the presence of disulfide bonds within the constructs makes them extremely thermostable. In addition, we solved the first crystal structure for any fibulin isoform, a structure corresponding to the previously suggested three motifs related to anaphylatoxin. The structure reveals that the three anaphylatoxins moieties form a single-domain structure.

The 2Fo-Fc map (blue, countered at 0.6 σ) and the Fo-Fc map (positive density, green, countered at +2.0 σ) around C500 of chain F. The loop region between α2 and α3 including C500 is flexible in all chains of this crystal structure.In chain F, the region just before the C500 (E494-T499) has some continuous electron density at low sigma levels (shown with dashed line before C500(F)), but the quality was not good enough for reliable model building.The corresponding loop region in the chain E (the dimer partner of chain F), has also a very weak electron density shown only at low sigma levels (the missing region, D496-V502, shown with a dashed line).However, the electron densities do support the concept of a disulfide bond between C500(F) and C500(E).

Figure S1 .
Figure S1.AlphaFold predicted structures of fibulin-2.Ribbon representations of the domains extracted from the AlphaFold predicted monomeric model of full-length mouse fibulin-2 protein (accession code: AF-P37889-F1).a) N-terminal part of region 1 which is composed of the cysteine rich Na domain (grey) and the cysteine lacking Nb domain (predicted to be fully unstructured, red).b) The C-terminal part of region I containing the three anaphylatoxin-like motifs (grey).The predicted structure is very similar and the disulfide bond architecture the same as in the crystal structure presented in this study.c) Region 2 containing the 11 EGF-like domains (grey) along with two linker regions with no predicted regular secondary structure (red).d) The Domain III of region 3 predicted to consist of mainly β-sheet structure (grey).All the secondary structure predications have a high confidence level (above 90 %).Disulfide bonds (yellow) are shown as ball and stick representation.

Figure
Figure S2.SEC-MALS analysis of the S427-G545 fragment of mouse fibulin 2. The light scattering signals of the wild type (red) and C500L mutant (blue) are shown.In addition, the molecular mass distribution profiles of both runs are shown for the peak regions (horizontal lines).The molar masses (~26.3 kDa for the wild type and ~25.7 kDa for the C500L mutant) were calculated from the peak fractions and using the refractive index (RI) and light scattering signals with Astra software (Wyatt Technologies).

Figure S4 .
Figure S4.Electron density map of the crystal structure of the S427-G545 fragment of mouse fibulin 2. a) The 2Fo-Fc map around the C436-C469 disulfide (1-SS2) bond in chain A contoured at 1.0 σ. b)The 2Fo-Fc map (blue, countered at 0.6 σ) and the Fo-Fc map (positive density, green, countered at +2.0 σ) around C500 of chain F. The loop region between α2 and α3 including C500 is flexible in all chains of this crystal structure.In chain F, the region just before the C500 (E494-T499) has some continuous electron density at low sigma levels (shown with dashed line before C500(F)), but the quality was not good enough for reliable model building.The corresponding loop region in the chain E (the dimer partner of chain F), has also a very weak electron density shown only at low sigma levels (the missing region, D496-V502, shown with a dashed line).However, the electron densities do support the concept of a disulfide bond between C500(F) and C500(E).

Figure S5 .
Figure S5.Structural alignment analysis.Ribbon representations of anaphylatoxin structures with C3a anaphylatoxin domain (PDB code 4HW5) [29] in green as a reference structure.a) Superimposed with C5a (PDB code 4P3A) [30] in blue, with a r.m.s.d.value of 1.408 Å. b) Superimposed with module 1 of solved structure having an r.m.s.d.value of 1.172 Å. c) Superimposed with module 2 of solved structure having r.m.s.d. of 5.476 Å. d) Superimposed structure with module 3, having a r.m.s.d.value of 0.727 Å. Disulfide bonds are shown as yellow stick representations.

Figure S6 .
Figure S6.Homodimer model of mouse fibulin-2.A cartoon representation of one possible homodimer model of the full-length mouse fibulin-2 protein as calculated by AlphaFold multimer.Chain A and chain B of the Na and anaphylatoxin-like (AT) domains are shown in dark and light grey and red, respectively.The EGF-like domains are shown in blue and the two, unassigned sequence regions, linker 1 (between AT and EGF-like domain I) and linker 2 (between the first and the second EGF-like domains) are shown in black.The C-terminal domain III is shown in green.The long and fully unstructured domain Nb is not shown for both chains.The coloring scheme is the same as in Figure 1.In this model, the key dimeric interactions are made between the Na and anaphylatoxin regions, as well as the linker 1 region.The anaphylatoxin dimer of the full-length fibulin-2 model shown here is highly similar to the experimentally determined crystal structure of the anaphylatoxin dimer described in this paper.

Table S1 : Data collection and statistics of the data processing and refinement of the structure of the S427-G545 fragment of mouse fibulin-2 Data collection
1.The numbers in parentheses refer to the outer shell.2. Non-hydrogen atoms.