Normal human serum (NHS) with two copies of FHR1, serum with one chromosomal copy of FHR1 (HS+/Δ) and FHR1-FHR-3 deficient serum (HSΔ/Δ) were collected from healthy individuals, upon informed consent. FHR1 copy numbers were determined by multiplex ligation dependent probe amplification (MLPA); FHR1 and Factor H levels were evaluated by Western Blotting and ELISA. FHR1 was detected either with a monoclonal antibody JHD-7.10.1 (4, 59) or with polyclonal rabbit FHR1 antiserum, which reacts with FHR1 and Factor H. Factor H was detected with M16 antibody (11, 60) polyclonal goat anti-C3 antiserum and polyclonal rabbit anti-C3a antiserum were from CompTech (Tyler, TX).

HUVECs (CRL-1730; ATCC) were cultivated as described (53). Necrosis and apoptosis were induced as described in DMEM lacking FBS. After washing with PBS, necrotic, apoptotic, or intact cells were incubated with FHR1-variants (10 µg/ml), Factor H or BSA for 30 min at 37°C and evaluated by to flow cytometry. CHO-K1 cells (ATCC CCL-61) were cultivated in Ham’s F-12 medium (61).

Recombinant FHR1_YVQ was generated using FHR1 H36.2 cDNA (22). Mutations were introduced H36.1 and H36.2 cDNAs by site-directed mutagenesis according to the manufacturer’s protocol (Stratagene, LaJolla, CA). The following primers were used: FHR1_YVQ_Fwd: 5’-CCGCCCACAGTACAAAATGCTTATATAGTGTCGAGACAGAT-3’ and FHR1_YVQ_Rev: 5’-GGTGAGAGAGTACGTTATCAATGTAGGAGCCCTTAT-3’. Purified Factor H, C3b, C3d, and Factor I were from Complement Technology, Inc (Tyler, TX). In addition, the mutants FHR1_YVQ-SV and FHR1_HLE-SV were generated using FHR1_LA_Fwd: 5’-GCCAAACAGAAGCTTTATTCGAGAACAGGTGAATCAGTTGAATTTGTGTGTAAACGG-3’. After sequencing, the four proteins were expressed in Pichia pastoris and purified as described (62).

Heparin (Fluka), heparan-sulfate, chondroitin-sulfate-A and chondroitinsulfate-4 (Sigma-Aldrich, Taufkirchen) (50 µg/ml in DPBS), were immobilized onto heparin binding plates (EpranEx, BD). C3b, C3d, C3 or iC3b (5 µg/ml) were coated to Maxisorp ELISA plates (Nunc) or heparin binding plates (6). After washing with PBS and blocking (2%; v/v) BSA-PBS) human serum (NHS, HS+Δ, HSΔΔ), FHR1 isotypes or mutants, Factor H or BSA were added at the indicated concentrations. The mixtures were incubated at 37°C for 1h. Bound FHR1 or Factor H were detected using JHD 7.10.1, diluted (1:1000) followed by the corresponding anti-mouse secondary antiserum (1:1000; Dako). OD values were measured at 450nm.

Protein extracts from cell lysates or human serum were lysed with RIPA-buffer containing 20 mM Tris, 150 mM NaCl, 1% Triton X-100, supplemented with a cocktail of protease inhibitors (Thermofisher, #PIA32953) separated by 10% SDS-PAGE and proteins were transferred onto nitrocellulose membranes (GE Healthcare) as described (5).

Total RNA was extracted from cells with RNeasy (Qiagen, #75144) and from tissues with TRIzol (Thermofisher, #15596018). 1 μg of total RNA was reverse transcribed into cDNA using the iScript™ cDNA Synthesis Kit (BioRad, #1708890). RT-PCR analysis was performed as previously described (5).

Binding of FHR1 variants, Factor H, or FHR1 and Factor H C-terminals fragmentsto intact and necrotic HUVECs was determined as described (61). Briefly, cells were incubated with 10 μg FHR1 variants or BSA for 30 min at 37 °C in PBS-BSA 1%. Cells were washed with PBS and bound FHR1 was assessed by staining with JHD-7.10.1 in combination with Alexa Fluor 488 labeled donkey anti-mouse IgG (1:500, Invitrogen). Fluorescence was measured by flow cytometry (LSRII according to FSC/SSC and VD/Annexin).

CHO cells were challenged with complement active HSΔΔ (20%; diluted in activation buffer (20 mM Hepes, 144 mM NaCl, 7 mM MgCl₂, 10 mM EGTA, pH 7.4) in presence of the FHR1 variants (75 µg/ml), Factor H (50 µg/ml) or BSA (50 µg/ml) for 30 min at 37°C. After incubation, and washing and deposited C3b was analyzed by flow cytometry using a FITC-conjugated goat F(ab’)2 anti−human C3 (1:200 (Protos Immunoresearch, CA). C5b9 deposition was determined with mAb C5b (1:200) in combination with appropriate secondary antiserum.

Chicken erythrocytes (CRBC) or sheep erythrocytes (both at 5x10⁶/ml) were treated with FHR1 variants (75 µg/ml) added to complement active HSΔ/Δ (20%) (40). After 30 min incubated at 37°C hemoglobin release was recorded at 414 nm.

Apoptotic HUVECs were incubated with FHR1 and Factor H. The JHD10 antibody and Factor H antiserum were added followed by gold labeled specific secondary antiserum. Then the cells were fixed with 2.5% glutaraldehyde, dehydrated and embedded in Epon as described. Samples were then processed as previously reported (63).

Standard immunohistochemistry was performed in paraffin fixed biopsies obtained from patients with Lupus nephritis and ANCA nephritis (61). Recombinant FHR1 was added to the biopsies and following washing specific FHR1-antiserum was added and detected with appropriate immunolabeled secondary antiserum.

A computational to-scale model of FHR1 and Factor H was developed to analyze their respective action regions. This involved an analytical approach to compute the action regions approximated by geometric shapes and millions of Monte-Carlo simulations to explore probable locations of N-terminal SCRs and their distance to C-terminal SCRs for both proteins. Detailed information can be found in the Supplementary Information .

Data are represented as mean ± SD. The groups were compared by Student’s t-test or one-way analysis of variance (ANOVA). P-values <0.05 were considered statistically significant. All statistical analyses were performed using Prism 9-software (GraphPad-Software).

We first assayed binding of FHR1 and Factor H to heparin. For NHS, i.e., serum from individuals with two alleles of FHR1 (HS++), heparin binding was set at 100% ( Table 1 ). For serum from an individual carrying one FHR1 allele, i.e., HS+Δ, FHR1 binding was reduced by 50% ( Table 1 ) and with serum lacking FHR1 (HSΔΔ), binding was reduced to background-levels. Factor H binding increased by 35% with HS+Δ, and by more than 5-fold with HSΔΔ. Thus, FHR1 levels strongly influenced Factor H binding to heparin ( Table 1 ).

The binding of FHR1 to C3b::Heparin and C3d::Heparin combinations was further tested. Using NHS, FHR1 bound 6-fold more to C3b::Heparin and almost 7-fold more to the C3d::Heparin surface ( Table 1 ) than to heparin alone. Factor H binding to the C3b::Heparin combination increased almost 4-fold, but its binding to C3d::Heparin was lower (1.7-fold) ( Table 1 ). FHR1 clearly preferred C3d::Heparin while Factor H preferred C3b::Heparin, suggesting a selective role for surface-attached C3-fragments.

Binding of FHR1_YVQ and Factor H to late apoptotic HUVECs. Apoptotic HUVECs showed morphologic alterations and bleb formation ( Figure 2A ). FHR1_YVQ bound to apoptotic cells, and adjacent dots revealed that FHR1 bound as a dimer ( Figure 2A ). Factor H also bound to apoptotic cells, but single and adjacent dots were observed. The double dots could potentially indicate that Factor H binds as a dimer to apoptotic surfaces ( Figure 2A ). When FHR1_YVQ and Factor H were combined, and proteins were visualized using different sized gold particles, double-dots of different sizes were observed, indicating that FHR1 and Factor H colocalize at surfaces. Altogether, the results suggest that FHR1_YVQ bind to apoptotic surfaces as a homodimer, and that it could potentially form heterodimers with Factor H ( Figure 2A ).

Confocal microscopy revealed that FHR1_YVQ bound at low levels to intact HUVECs and more to necrotic HUVECs. FHR1 binding was most prominent at sites where membrane integrity had been lost, i.e., sites lacking WGA-staining ( Figure 2B , Supplementary Figure S1 ).

Given their different binding strengths to C3d/C3d::Heparin combinations, we directly compared FHR1 and Factor H binding to C3b and C3d. FHR1_YVQ bound to immobilized C3b, but it bound more strongly to C3d ( Figure 2C ) and Factor H showed the opposite profile bound more strongly to C3b and less to C3d ( Figure 2C ).

FHR1 deposition was evaluated in diseased glomeruli. FHR1 revealed granular positivity in close proximity to necrotic segments in ANCA-associated necrotizing glomerular areas ( Figure 3A ). In Lupus nephritis FHR1 bound strongly to tissue areas within the immune complexes ( Figure 3B ). FHR1 deposition at specific sites in glomeruli of the two kidney diseases agrees with a local action of this effector protein.

FHR1_YVQ and FHR1_HLE differ by three amino acids in SCR3. Both variants are common in the European population but whether they are functionally different is unclear ( Figure 4A , Supplementary Figure 1 ) (25). To determine whether the amino acid differences alter ligand binding and complement action, we first assessed their binding to heparin and GAGs. Both isoforms bound to heparin, heparan-sulfate, chondroitin-sulfate-A, and chondroitin-sulfate-4 ( Figure 4B , Supplementary Table 1 ). The FHR1_YVQ variant bound slightly better than the FHR1_HLE variant to all four ligands, and the higher binding of FHR1_YVQ to heparin, heparan-sulfate, and chondroitin-sulfate-A was significant. In addition, both variants bound to necrotic HUVECs. Again, FHR1_YVQ bound better to necrotic HUVECs than FHR1_HLE (~25%) ( Figure 4C ). Binding to C3 and C3-activation fragments revealed that both isoforms bound better to C3 and to C3d than to C3b and iC3b ( Figure 4D ). FHR1_HLE bound less to C3d than FHR1_YVQ. The differences for heparin and C3d binding demonstrate that SCR3 also contributes to ligand binding.

To evaluate complement regulation, each variant was added to active HSΔ/Δ serum and following incubation, C3b and C5b-9 deposition were recorded. Neither isoform affected C3b deposition ( Figure 4A ). Factor H, used as a control, reduced C3b deposition (data not shown). However, both isoforms reduced C5b-9 deposition and chicken erythrocytes lysis ( Figures 4B, C ). FHR1_HLE showed ~20% higher inhibitory effect than FHR1_YVQ. Thus, SCR3 contributes to ligand binding, surface recognition, and complement control.

To compare their proinflammatory effects FHR1-variants were added to HSΔ/Δ and the supplemented serum was added to THP1-macrophages. Both isoforms induced IL-1β and TNFα secretion and FHR1_HLE was around 25% more effective than FHR1_YVQ ( Figure 4D ). Similarly, when added together with LPS, both isoforms enhanced cytokine releases and again, FHR1_HLE was more effective ( Figure 4D ). Thus, SCR3 influences FHR1 mediated C5b-9 deposition, hemolysis, and secretion of proinflammatory cytokines.

The biological effects of the C-terminal fragments, which lack the N-terminal interaction and regulatory-regions, i.e., FHR1_3-5 with LA and Factor H_18-20 with SV-motifs, were compared ( Figure 5A , Supplementary Figure 1B ).

FHR1_3-5 or Factor H_18-20 were added to HSΔ/Δ and following incubation, C3b-deposition was examined. FHR1_3-5 increased C3b-deposition, but Factor H_18-20 was more effective ( Figure 5B ). However, FHR1_3-5 reduced and Factor H_18-20 increased C5b-9 deposition ( Figure 5C ). These opposite effects highlight the relevance of the LA and SV-residues for target access and complement control. FHR1_3-5 and Factor H_18-20 at 3.0 ug/ml had similar effects on erythrocyte-lysis. At higher doses, lysis by FHR1_3-5 reached a plateau, but Factor H_18-20 continued to increase lysis dose dependently ( Figure 5D ). Furthermore, both fragments induced CCL2, CXC10, IL6 transcription in human monocytes ( Figure 5E ). However, Factor H_18-20 induced a 2-fold higher CCL2 expression levels than FHR1_3-5.

The effects of the C-terminal fragments suggest that residues 290/296 of FHR1 influence complement functions. To show this more clearly, we generated FHR1_YVQ and FHR1_HLE mutants containing the Factor H_SV motive in SCR5 (FHR1_YVQ-SV and FHR1_HLE-SV) ( Figure 6A , Supplementary Figures S1A, B ). FHR1_YVQ-SV and FHR1_YVQ bound more strongly to heparin, heparan-sulfate, chondroitin-sulfate-A, and chondroitin sulfate-4 than did FHR1_HLE and FHR1_HLE-SV ( Figures 5B , 6B , Supplementary Table 2 ), confirming the importance of SCR3 and SCR5 for heparin and GAG binding.

FHR1_YVQ-SV and FHR_HLE-SV mutants bound more strongly to necrotic HUVECs than did wild-type equivalents ( Figures 5C , 6C, D ). The FHR1-mutants bound less (~20%) to C3b and C3d than did their wild-type counterparts ( Figures 5D , 6C, D ), confirming that the LA-SV residues in SCR5 influence FHR1 binding to necrotic surfaces and C3b/C3d.

FHR1_YVQ-SV mutant promotes complement activation and inflammation. Both proteins with the YVQ-motif did affect C3a-generation to a larger extent than the YVQ forms ( Figures 7A, B ). Full-length Factor H reduced C3a-generation. By contrast, both FHR1_SV mutants increased C3b-deposition (by 35% and 30%), while the FHR1-isoforms showed minor effects ( Figure 7C ).

Both FHR1_SV-mutants also increased C5b-9 deposition (35% and 30%), whereas the FHR1 isoforms reduced C5b-9 deposition ( Figure 7D ). Factor H reduced C5b-9 deposition by 80% ( Figure 7D ). Similar effects were observed for erythrocyte lysis. Both FHR1-isoforms reduced chicken erythrocyte lysis by 25% and 30%, whereas both FHR1_SV-mutants increased chicken erythrocyte lysis by 25% and 18% ( Figure 7E ). The increase in lysis by the FHR1_SV-mutants was dose-dependent ( Figure 7F ). These opposite effects of the FHR1 variants and FHR1_SV-mutants show that the LA-SV residues play an important role in complement control at surfaces.

The effect of the FHR1 variants and FHR1 mutants were tested on complement activation with FHR1 deficient complement active human serum. C3a generation upon addition of FHR1 _YVQ , FHR1 _HLE and also of FHR1 _{YVQ- SV,} FHR1_HLE-SV to HSΔΔ resulted in C3a ( Figures 8A , B ). Factor H showed reduction. Similarly, the FHR1 variants and mutants influence C3b deposition and the effects were different to Factor H. ( Figure 8C ). The FHR1 variants were less efficient in C3b generation as compared to the FHR1 mutants, with the SV motive in the most C-terminal SCR. This effect can be interpreted by a more efficient competition of Factor H. Highly related effects were observed for C5b-9 deposition ( Figure 8D ) and lysis of both chicken and sheep erythrocytes ( Figures 8E–G ).

When bound via their C-termini FHR1 and Factor H control different surface areas or three-dimensional hemispheres. Our modeling approach revealed that FHR1-dimers attached via its C-termini control a rectangular area of 72 nm² and a hemicylinder with a volume of 509 nm³. Its N-terminal region occupy about two-fifth (39.3%) and one-tenth (11.1%) of its corresponding surface area and volume respectively ( Table 2 , Figure 9A ). The FHR1*-dimer, an FHR1-dimer with only one C-terminal domain attached, can bend in two dimensions and controls twice the surface area and volume of the FHR1-dimer, yet occupies only half of its corresponding surface area and volume ( Table 2 , Figure 9B ). By contrast, Factor H which exposes a flexible 17SCRs-long segment, controls a large circular surface area of 8,638 nm² and a hemisphere with a volume of 303,023 nm³. The N-terminal SCRs of one Factor H occupy 0.33% of its surface area and only 0.02% of its volume ( Table 2 , Figure 9C ). However, Factor H is able to reach over 100 (50) times the surface area and over 500 (250) times the volume of FHR1 (or FHR1*) ( Table 2 ). Interestingly, probable locations of N-terminal SCRs of FHR1 and FHR1* are uniformly distributed regarding their distance to the C-terminal SCRs ( Figure 9D , Supplementary Videos 1, 2 ), while for Factor H, distances from C-terminal SCRs to probable locations of N-terminal SCRs form a normal distribution ( Figure 9D , Supplementary Video 3 ). This reveals compelling differences in the action radius of both proteins.