The Inhibitory Effects of a Factor B-Binding DNA Aptamer Family Supersede the Gain-of-Function of Factor B Variants Associated with Atypical Hemolytic Uremic Syndrome

Abstract

Aptamers are short, single-stranded oligonucleotides that selectively bind to target biomolecules. Although they generally exhibit good binding specificity, their affinities are often limited due to the relative lack of hydrophobic groups in nucleic acids. To address this obstacle, chemically modified nucleotides incorporating hydrophobic structures into uracil have been synthesized. Modified DNA aptamers containing such non-standard nucleotides have been developed for over twenty different complement proteins. These modified aptamers show increased affinity and enhanced serum stability and have potential value as therapeutic agents. We recently conducted a structure/function study on a family of modified DNA aptamers that bind specifically to complement factor B (FB). This work revealed that these aptamers selectively inhibit the complement alternative pathway (AP) by preventing the formation of the AP C3 proconvertase complex, C3bB. Certain patients with atypical hemolytic uremic syndrome (aHUS) express gain-of-function variants of FB that enhance the formation of the proconvertase complex and/or decrease the efficacy of endogenous regulators against the C3 convertases they form. To investigate whether these FB-binding aptamers could override the effects of disease-causing mutations in FB, we examined how they interacted with several FB variants, including D279G, F286L, K323E, and K350N, in various assays of complement function. We found that the inhibitory effect of the FB-binding aptamers superseded the gain-of-function mutations in FB, although the aptamers could not dissociate preformed C3 convertases. These findings suggest that FB-binding aptamers could be further developed as a potential treatment for certain aHUS patients or those with other diseases characterized by excessive complement activity.

INTRODUCTION

Complement is a critical component of the innate immune system that helps remove pathogens and damaged cells, clear circulating immune complexes, and promote downstream inflammatory reactions (1, 2). There are three canonical routes for complement activation known as the classical (CP), lectin (LP), and alternative (AP) pathways; although each pathway is triggered by distinct biochemical cues, they coalesce upon the activation of complement component C3 (C3) (3). When serving as an initiation route, the AP begins with spontaneous hydrolysis of the intramolecular thioester found in C3. Subsequent conformational change in hydrolyzed C3 (C3(H₂O)) allows it to bind the pro-protease Factor B (FB), generating a fluid-phase proconvertase complex (C3(H₂O)B); activation of this initial proconvertase by Factor D (FD) yields a fluid-phase C3 convertase (C3(H₂O)Bb) that can proteolytically cleave C3 into its bioactive fragments, C3a and C3b. Whereas C3a serves as a chemotactic molecule, C3b can covalently attach itself to nearby surfaces and biomaterial. This leads to the formation of the surface-bound AP C3 convertase (C3bBb) through an analogous process to that described above. Surface-bound C3 convertases are more efficient enzymes than their fluid-phase counterparts (4). This is due partly to the effects of the positive regulator, properdin (FP), which promotes formation of surface-bound convertases and increases their stability (5). Moreover, localized generation of high C3b levels on surfaces facilitates self-amplification of the AP (6, 7). Thus, the AP provides the primary means of complement amplification under most settings (8, 9). Indeed, one well-known study has suggested that the AP contributes over 80% of downstream C5 activation products when complement is initiated by the CP (8).

The self-amplifying nature of the AP requires that its activity be strictly controlled to avoid damaging otherwise healthy cells and tissues. This is achieved through conceptually distinct yet ultimately overlapping mechanisms (1–3). On one hand, the AP C3 convertase itself is intrinsically labile and dissociates irreversibly with a half-life of approximately one minute under physiological conditions (4). On the other hand, a dedicated suite of soluble and cell surface-exposed proteins act as negative regulators of AP activity (10). These so-called Regulators of Complement Activation (RCA) accelerate the irreversible decay of preformed convertases into their individual components. They also act as cofactors for Factor I (FI), which cleaves C3b into iC3b and C3c, neither of which support the formation of new AP C3 convertases. Among the RCA proteins are Factor H (FH), which, though soluble, binds to host cells through sialylated carbohydrates, as well as Decay Accelerating Factor (DAF) and Membrane Cofactor Protein (MCP), which are membrane-spanning glycoproteins. Together, these intrinsic and extrinsic regulatory mechanisms help shield host cells from indiscriminate complement attack (11).

Atypical hemolytic uremic syndrome (aHUS) is a rare, chronic, and potentially life-threatening thrombotic microangiopathy characterized by thrombocytopenia, hemolytic anemia, and acute kidney injury (12). Whereas typical HUS is often triggered by Shiga toxin-producing bacterial infections, the central pathogenic mechanism in aHUS is dysregulation of the complement AP (13, 14). Mutations in several RCA proteins are most commonly associated with the development of aHUS, and have highlighted the role of excessive AP activity in pathogenesis of this disease (15–17). Several mutations in FH have been identified that diminish binding of this regulator to endothelial cells, thereby resulting in insufficient complement regulation at the cell surface (18–20). Mutations in MCP (CD46) that disrupt its C3b-binding and cofactor activities have also been described (21). Finally, mutations in FI have also been reported that either lead to loss of its proteolytic activity toward C3b or lack of FI protein altogether (22). While these examples describe loss-of-function in negative regulators of the AP, gain-of-function mutations in other AP components have also been linked to aHUS. Several mutations in C3 have been identified that lead to reduced binding to RCA proteins, including FH and MCP (23). aHUS-associated mutations have also been described in FB (24). Although these are rare and account for less than 2% of all aHUS patients (17), molecular analyses of these disease-linked FB variants have revealed interesting features including an enhanced rate of formation of the C3 proconvertase (i.e. C3bB) and/or formation of a fully active C3 convertase (i.e. C3bBb) that is resistant to both intrinsic and regulatory decay (25–27).

The treatment of aHUS has evolved significantly over the years. The initial first-line therapy was plasma treatment, which is based upon plasma infusion and exchange (28). This method aimed to supply functional complement regulators and to remove abnormal complement components. However, the introduction of eculizumab - a monoclonal antibody that directly binds with C5, prevents C5 cleavage, and the formation of the terminal complement complex - revolutionized aHUS treatment (29). Eculizumab showed high efficacy and safety and quickly became the gold standard for managing aHUS (30, 31). Whereas terminal complement inhibition via eculizumab is effective for treating aHUS resulting from a variety of molecular lesions, including gain-of-function mutation in FB, one clinical study reported ongoing C5 cleavage and terminal pathway activation in a patient with an FB mutation after receiving eculizumab (32). An additional concern with anti-C5 treatment is its lack of complement pathway selectivity, that can reduce the overall protection against bacterial infection (28), although this can be managed through preventative vaccination against problematic species like Neisseria. Nevertheless, the recent regulatory approvals of molecules targeting upstream complement components have renewed interest in the use of proximal complement inhibitors for the treatment of aHUS. Indeed, the small-molecule FB inhibitor iptacopan is currently undergoing a phase III trial for this indication (33).

We previously characterized a class of FB-binding modified DNA aptamers that are potent inhibitors of the AP (34). These aptamers bind directly to FB, prevent its interaction with C3b, and thereby block the formation of the C3 proconvertase complex, C3bB (34). In turn, this leads to decreased levels of the enzymatically active C3 convertase, C3bBb (34). Since certain FB variants linked to aHUS have been shown to increase the formation rate of the C3 convertase and/or generate C3 convertases resistant to decay, we considered whether the inhibitory effect of these FB-binding aptamers could counteract the gain-of-function caused by these mutations in FB. In this study, we tested this concept using a panel of FB mutants, all of which showed either an increased affinity for C3b, an increased rate of proconvertase formation, and/or an increased stability of the active C3 convertase (25–27). Our findings demonstrate that the FB-binding modified DNA aptamers effectively blocked these gain-of-function FB mutants from binding to C3b, thereby preventing the formation of the C3 convertase and counteracting the increased complement activity caused by these gain-of-function mutations. Our results suggest that it should be possible to use these FB-binding aptamers therapeutically in a subset of aHUS patients who express gain-of-function variants in FB. Considering their potency, use of these aptamers could also be extended to other disease settings where excessive complement amplification via the AP is a contributing factor (35).

MATERIALS AND METHODS

Protein Samples

Samples of complement components FB, FD, FH, FP, and C3b that were purified from human serum were obtained commercially from Complement Technologies.

Plasmids for expression of recombinant wild-type and mutant forms of FB were obtained from GeneScript. DNA fragments encoding proteins of interest were synthesized with a C-terminal octahistidine-tag and then subcloned into pcDNA3.1+. Sufficient plasmid for large-scale transient transfection was prepared using a QIAGEN Plasmid Maxi kit. Individual expression plasmids were transiently transfected into Human Embryonic Kidney cells (HEK293T) using polyethyleneimine (PEI) MAX according to the general methods described previously (36). Following transfection, approximately 1 L of conditioned cultured medium was collected and concentrated by tangential flow filtration to a volume of 150 mL. The buffer was then exchanged into native binding buffer (20 mM Tris [pH 8.0], 20 mM imidazole, 500 mM NaCl) in preparation for immobilized metal ion affinity chromatography. The sample was applied to a 5-mL HisTrap HP column using an AKTA FPLC system (Cytiva Life Sciences). Following extensive washing using native binding buffer, the bound protein was eluted using a linear gradient to full-strength native elution buffer (20 mM Tris [pH 8.0], 500 mM imidazole, 500 mM NaCl) while collecting 1 mL fractions. Fractions containing recombinant FB protein were collected, concentrated, and further purified by gel filtration chromatography on a Superdex 75 26/60 column using HEPES-buffer saline (HBS; 20 mM HEPES [pH 7.40], 140 mM NaCl) as a mobile phase. The column eluent was collected in 2 mL fractions, which were subsequently analyzed by SDS-PAGE. Fractions containing purified, recombinant FB were pooled, concentrated, quantitated, and stored at either 4°C or −80°C until further use.

Aptamer Synthesis

The modified, 31-base DNA aptamer SL1100 was used for all studies described here (34). Its sequence is 5’-dC-dG-dC-2NEdU-dG-dA-dG-dA-dA-2NEdU-dA-dG-dA-dA-dG-2NEdU-dA-dG-dG-dA-dG-2NEdU-dA-2NEdU-dG-dC-2NEdu-2NEdU-dG-dC-dG-3’. SL1100 was synthesized by standard DNA phosphoramidite methods as described previously (37). The phosphoramidite for incorporating 5-(N-2-naphthylethylcarboxyamide)-29-deoxyuridine (2NEdU) was likewise prepared according earlier publications (38, 39). Following synthesis, SL1100 was purified using ion-pairing reversed-phase liquid chromatography, desalted using a Hi-Trap Sephadex G-25 column, and evaporated to dryness. The purified SL1100 product was analyzed using ultra-performance liquid chromatography and its identity was confirmed using MALDI-TOF mass spectrometry. A form of SL1100 that was used for FB-binding studies, which incorporated a 5’-modified base containing both a polyethylene glycol linker and a biotin moiety, was also prepared using the methods outlined above.

Surface Plasmon Resonance Binding Studies

Surface plasmon resonance (SPR) experiments were performed on a BiaCore T-200 instrument at 25°C using HBS-T (20 mM HEPES [pH 7.4], 140 mM NaCl, 0.005% [v/v] Tween 20, and 10 mM MgCl₂) as the running buffer and a flow rate of 30 μL/min. A streptavidin-derivatized SAD200L chip (Xantec Bioanalytics; Dusseldorf, Germany) was used to capture human C3b that had been site-specifically biotinylated at Cys-1010 via EZ-link Maleimide-PEG2-Biotin (Thermo Fisher Scientific) according to a previously published method (40). A reference surface was prepared by 3 injections of biotin alone (10 nM) on one flow cell (fc1). Replicate experimental surfaces containing 2158 (fc2), 2150 (fc3), and 2027 (fc4) RU of C3b-biotin were prepared thereafter by injecting dilute solutions of C3b-biotin (8 ng/ml) across the remaining flow cells. Binding of various forms of FB to C3b was assessed by injecting a buffer blank, followed by a 2-fold dilution series of FB across seven different concentrations spanning 15.625 nM to 1 μM. Each injection cycle consisted of a 1-minute association phase, a 2-minute dissociation phase, and three consecutive pulses of a mildly acidic, high salt regeneration buffer (10 mM sodium acetate [pH 4.0], 2M NaCl).

The binding of various forms of FB to 5’-biotinylated SL1100 aptamer was also studied using a separate SAD200L chip. A reference surface was prepared using the method described above (fc1), while biotinylated SL1100 was immobilized on experimental flow cells with an initial capture level of 59 (fc2) 59 (fc3), and 34 (fc4) RU. Interactions were assessed using single-cycle kinetics with 5 concentrations of FB (i.e. 62.5 nM, 125 nM, 250 nM, 500 nM, and 1000 nM). Surfaces were regenerated between analyte series using three consecutive pulses of 4 M MgCl₂, and the experimental flow cells were reloaded with biotinylated SL1100 as described above. The performance of the surface across multiple analytes was assessed by injecting the same analyte (i.e. FB-K350N) at the beginning and end of the entire experiment. These internal standard data showed excellent agreement with one another, yielding apparent K_D values of 5.9±2.8 and 7.0±4.4 for the initial and final injections, respectively.

Reference-subtracted sensorgrams from all experiments were analyzed using Biacore T-200 Evaluation software. Data were fitted to a two-state reaction model to derive an apparent K_D using a mechanism that incorporates a single second-order rate constant (i.e. Ka-1) and three first-order rate constants (i.e. K_a-2, K_d-1, and K_d-2). Values from fits are presented as the mean plus or minus the standard deviation from 3 technical replicates (Table I).

Table I.

Interaction parameters of various FB proteins with immobilized C3b or SL1100^a,b

Analyte	Biotinylated-Ligand	Ka-1 (104 M−1s−1)	Kd-1 (10−3 s−1)	Ka-2 (10−3 s−1)	Kd-2 (10−4 s−1)	KD (nM)	Rmax (RU)
FB-Native	C3b	16.8±1.1	70.2±0.7	7.1±0.2	24.9±2.4	109±15	579±38
	SL1100	5.2±1.2	0.1±0.0	2.1±2.7	89.2±82.5	14.6±7.5	651±126
FB-WT	C3b	32.5±1.2	79.6±5.0	5.0±0.6	138.3±10.2	180±7	641±23
	SL1100	24.2±7.4	3.9±1.4	5.2±1.1	2.4±1.2	7.3±4.0	270±35
FB-D279G	C3b	30.4±1.7	37.6±1.5	7.0±0.7	94.1±4.4	71±1	659±33
	SL1100	26.2±3.4	2.7±0.7	2.7±1.0	2.3±0.6	8.3±3.2	228±9
FB-F286L	C3b	55.1±7.3	22.9±9.6	28.9±8.5	83.0±6.5	9±0	710±42
	SL1100	25.1±9.8	2.4±2.5	3.5±2.8	1.7±0.7	4.1±1.8	240±90
FB-K323E	C3b	28.0±1.6	80.7±4.2	5.9±0.5	72.3±9.2	159±4	690±28
	SL1100	18.3±6.4	2.4±2.2	4.5±2.6	2.5±1.5	6.0±3.3	270±99
FB-K350N	C3b	13.1±0.5	99.5±3.7	5.7±0.1	41.9±2.6	322±11	594±27
	SL1100	14.7±1.9	1.2±0.4	3.2±1.4	2.4±1.1	5.9±2.8	229±60
FB-2Mut	C3b	22.5±0.9	52.6±1.6	7.4±0.2	45.0±3.8	88±4	607±30
	SL1100	26.0±6.4	2.8±1.6	2.9±1.9	2.4±1.2	9.1±5.7	207±78
FB-3Mut	C3b	25.6±0.8	46.9±1.9	6.3±0.5	83.6±6.3	104±2	620±31
	SL1100	15.8±4.1	2.2±1.0	3.7±1.9	1.7±0.8	6.9±4.1	273±65

^aAll values are presented as the mean ± standard deviation obtained for a minimum of three replicate experiments.

^bValues result from fitting to a two-state binding model.

SPR Competition Binding Studies

Competition binding assays were conducted on the same C3b-immobilized experimental surface prepared above using a method described previously (34). A two-fold dilution series of various forms of FB (spanning seven different concentrations from 15.625 nM to 1 μM), either in the absence or in the presence of an increasing concentration of SL1100, was injected over the surface. Analysis of the reference-corrected sensorgram series was carried out as described above. Data were normalized by assigning a value of 100% to the signal resulting from the injection of the competitor-free, 1 μM FB sample immediately before the end of the association phase. Each experimental flow cell was normalized independently to account for the slight differences in C3b density. All plots were generated using GraphPad Prism v10.0.0.

SPR Measurements of Alternative Pathway C3 Convertase Formation and Decay

The AP C3 convertase formation assay was performed using the same surface described above using a previously described SPR-based method (41). Each experiment consisted of two separate injection events: the first injection of a 100 nM mixture of FD and FB occurred for 1 min, followed by a 15-second dissociation phase and a brief stabilization period, whereas the second injection occurred for 1 min, followed by a 2-min dissociation phase. For the second injection, HBS-T was used as a negative control, whereas FH served as a positive control. 3 concentrations (100 nM, 500 nM, and 1000 nM) of SL1100 aptamer were injected to test for any potential effects on convertase stability and/or decay. Regeneration of the surface between experimental cycles was achieved by three consecutive pulses of a mildly acidic, high salt buffer (10 mM sodium acetate [pH 4.0], 2M NaCl).

Additional SPR experiments were conducted to investigate the effects of FP on AP C3 convertase formation and decay. These studies were performed using the same general methodology described above, with the exception that streptavidin-modified SPR surfaces were generated by coupling NeutrAvidin (ThermoFisher Scientific) to a CMD200 sensor chip (Xantec Bioanalytics) as previously described (40). The reference surface, fc1, was prepared by 3 injections of biotin (10 nM), while experimental surfaces were prepared by injecting C3b-biotin (5 ng/ml) to achieve 978 RU on fc2 and 970 RU on fc3. FP (100 nM) was injected for 90 sec over fc2 to saturate the surface prior to any experimental injections over other flow cells. Each experiment was comprised of two sequential injection events. The first injection consisted of a 100 nM mixture of FB and FD lasting for 1 min, and was followed by a 3 min dissociation phase. The second injection consisted of 100 nM FH lasting for 1 min, and was followed by a 3 min dissociation phase. Regeneration of the surface between experimental cycles was achieved by three consecutive pulses of a mildly-acidic, high salt buffer (10 mM sodium acetate [pH 4.0], 2M NaCl). This same general procedure was modified to assess binding differences of various FB and FB/SL1100 complexes in the presence or absence of FP. In those experiments, the only experimental injection consisted of either 250 nM FB or a solution of FB/SL1100 containing 250 nM of each component. Surface regeneration between experimental cycles was achieved using three consecutive pulses of 2M NaCl, 10 mM EDTA.

Hemolytic Assays for Alternative Pathway Activity

AP hemolysis assays were conducted in a 100 μL reaction mixture consisting of 12 μL FB-depleted serum (Complement Technologies), 53 μL of 0.1% (w/v) gelatin veronal buffer (GVB; Complement Technologies), 10 μL of 0.1 M MgEGTA (Complement Technologies), 5 μL of FB solution in GVB to achieve desired concentration, and 20 μL of rabbit erythrocytes (3*10⁸/mL in GVB; Complement Technologies). Following a 30-min incubation at 37°C, samples were centrifuged for 10 min at 200 xg, the supernatant was diluted 4-fold in ddH₂O, and the absorbance was determined at 412 nm. The percentage of lysis was determined by the following equation: [100*(OD_{412, sample}−OD_{412, buffer}/(OD_{412, water}−OD_{412, buffer})], and the EC₅₀ was determined by a four-parameter fit using GraphPad Prism v10.0.0.

A similar procedure was employed for studying inhibition by SL1100, except that the concentrations of the 5 μL FB stock solutions added were adjusted to give the desired final concentration (i.e. 20 nM and/or 60 nM), and the GVB buffer volumes were reduced to allow for 5 μL SL1100 stock solution to achieve the final inhibitor concentration. The percentage of lysis inhibition was determined by the following equation: 100-[100*(OD_{412, sample}-OD_{412, buffer}/(OD_{412, water}-OD_{412, buffer})]. The IC₅₀ was determined by a four-parameter fit using GraphPad Prism v10.0.0.

Finally, an analogous protocol was also employed for comparing inhibition of hemolysis by SL1100 in either normal human serum (NHS; Innovative Research) or FP-depleted serum (FP-Dep; Complement Technologies). In these experiments, the final concentration of both sera was increased to 20% (v/v). The reaction using FP-Dep was also incubated at 37°C for 3 h, as opposed to 30 min for NHS. The inhibition percentage of lysis was determined as described above.

RESULTS

Generation and Characterization of aHUS-associated FB mutants

FB is a 93 kDa, single-chain glycoprotein comprised of two distinct functional regions (Fig. 1A): the N-terminal Ba region consists of three consecutive complement control protein (CCP) domains followed by a 40-amino acid linker region, while the C-terminal Bb region consists of a metal-binding von Willebrand factor (vWF) domain followed by a serine protease (SP) domain (42). By contrast, the FB-binding aptamers described previously are approximately 11 kDa and are between 29 to 31 bases in length (34). Crystallographic analysis of two members of this aptamer family, SL1102 and SL1103, bound to human FB revealed that these aptamers bind FB at a groove-like surface lying at the juncture of the CCP1 and CCP3 domains from Ba and the vWF domain form Bb (Fig. 1B). Approximately half of the FB-associated surface area in these complexes is derived from the CCP1 domain with the remainder divided equally between the CCP3 and vWF domains; however, these aptamers do not bind well to either the isolated Ba or Bb regions themselves. Nevertheless, because the Ba region is responsible for the initial interaction of FB with C3b, binding of these aptamers at this site inhibits the formation of the C3 proconvertase through a steric hindrance-type mechanism (Fig. 1C; (34)).

FIGURE 1. Description and Characterization of Gain-of-Function Variants in FB.

(A) The structure of human FB as drawn from the PDB entry 7JTN is shown as a ribbon diagram. The individual domains of FB are colored as follows: CCP1, sand; CCP2, yellow; CCP3, tan; vWF, magenta; SP, purple. (B) Structure of the modified DNA aptamer, SL1103, bound to human FB as drawn from the PDB entry 7JTN. SL1103 is shown in ball-and-stick convention with carbon atoms colored teal. FB is shown as a molecular surface with domains colored identically to panel A. The locations of residues D279, F286, K323, and K350 are highlighted in blue. Missing residues were added via Modeller, using the FB structure from PDB entry 7JTN as a starting point. Note that the CCP3 domain and S699 are not visible in this orientation. (C) Illustrated mechanism for AP inhibition by members of a FB-binding aptamer family. These aptamers block C3 proconvertase formation by occupying a site normally required for FB binding to C3b. (D) Samples of serum-purified and recombinant FB proteins were processed under reducing conditions, separated by SDS-PAGE, and visualized by Coomassie Blue staining. Lane M, molecular mass marker; lane 2, FB-Native; lane 3, FB-WT; lane 4, FB-D279G; lane 5, FB-F286L; lane 6, FB-K323E; lane 7, FB-K350N; lane 8, FB-2Mut; lane 9, FB-3Mut.

Several gain-of-function forms of FB associated with aHUS have been described. These variants result in excessive AP activity by increasing the rate of formation and/or overall stability of the C3 convertases they form (25–27). Since many of these aHUS-linked mutations map to FB regions outside the aptamer binding site (Fig. 1B), we hypothesized that these specific mutations would not diminish aptamer binding to these FB variants. To test this possibility, we first designed expression vectors to produce a panel of FB mutants for further studies. We focused on four single mutations, FB-D279G, FB-F286L, FB-K323E, and FB-K350N, identified in aHUS patients (25–27, 43). We also designed a combinatoric double mutant, FB-D279G/K350N (FB-2Mut); although this variant has not been described in any aHUS patient, we sought to test it for any synergistic effects that might arise from incorporating two gain-of-function mutations simultaneously. Finally, we created a triple mutant, FB-D279G/K350N/S699A (FB-3Mut), to serve as a negative control since this variant lacked the catalytic residue within the serine protease domain (i.e. S699). We expressed wild-type FB (FB-WT) and all six FB mutants in transfected HEK293T cells and purified the recombinant proteins from conditioned culture medium. Each form of recombinant FB maintained its integrity and displayed a molecular weight comparable to FB purified from human serum (FB-Native), as judged by SDS-PAGE (Fig. 1D).

To test whether these FB variants exhibited increased affinity to C3b, we characterized the interaction of each FB protein with site-specifically biotinylated C3b that had been captured on an SPR surface (Fig. 2, Fig. S1A, & Table I). As has been reported previously (3, 44, 45), we employed a two-state binding model to evaluate FB/C3b binding and determined the apparent affinity for each form of FB. Both serum-isolated FB-Native and recombinant FB-WT exhibited similar affinities of 109 nM and 180 nM, respectively (Fig. 2A–B). The apparent affinities of FB-D279G and FB-K323E were likewise within a similar range as FB-Native and FB-WT at 71 nM and 159 nM, respectively (Fig. 2C & 2E). FB-F286L displayed a higher apparent affinity for C3b of 9 nM (Fig. 2D), which was driven primarily by an enhanced association rate with C3b (Fig. S1A & Table I). In contrast, FB-K350N displayed a weaker apparent affinity for C3b of 322 nM (Fig. 2F). Both the FB-2Mut and FB-3Mut variants exhibited a similar affinity to FB-D279G at 88 nM and 104 nM, respectively (Fig. 2G–H). This observation suggested that the effect of the D279G mutation was dominant to that of K350N. Furthermore, it also revealed that there was also no synergistic effect between these two mutations insofar as C3b binding was concerned.

FIGURE 2. Direct Binding Studies of Various FB Proteins to C3b.

Site-specifically biotinylated C3b was captured on three experimental flow cells of streptavidin-coated biosensor, leaving one flow cell unmodified for reference subtraction. Protein-protein interactions were monitored by injecting increasing concentrations of FB in multi-cycle mode, where 1 μM was used as the highest concentration. Reference-subtracted sensorgrams (black lines) were analyzed using a two-state kinetic model (red lines). (A) Binding of FB-Native to C3b. (B) Binding of FB-WT to C3b. (C) Binding of FB-D279G to C3b. (D) Binding of FB-F286L to C3b. (E) Binding of FB-K323E to C3b. (F) Binding of FB-K350N to C3b. (G) Binding of FB-2Mut to C3b. (H) Binding of FB-3Mut to C3b. For panels A-H, the data shown represents one of three independent replicates.

The FB-binding Aptamer SL1100 Binds Gain-of-Function Forms of FB with Nanomolar Affinity

With a panel of aHUS-linked forms of FB now available, we examined whether members of the previously described family of FB-binding DNA aptamers retained the ability to bind these forms of FB (34). To test this possibility, we first synthesized a 5’-biotinylated version of the modified DNA aptamer, SL1100 (34). SL1100 is a 31-base, single-stranded DNA oligonucleotide that contains seven 5-(N-2-naphthylethylcarboxyamide)-2’-deoxyuridine (2NEdU) residues in lieu of deoxythymidine (34); it is closely related to aptamers SL1102 and SL1103, which have been crystallized bound to FB, although it lacks the DNA backbone 2’-O-methylation at certain positions designed to increase serum-stability of SL1102 and SL1103 (34). We then characterized the interaction of each FB protein with biotinylated SL1100 that had been captured on an SPR surface using single-cycle kinetics (Fig. 3, Fig. S1B, & Table I). We again used a two-state binding model to analyze the experimental sensorgrams and obtained two sets of association and dissociation rate constants, along with an apparent affinity. Significantly, we found that each form of FB displayed an apparent affinity for SL1100 in the range from 4.1 nM to 14.6 nM (Table I). Although serum-isolated FB-Native had a higher maximal response value than expected (i.e. Rmax), all recombinant forms of FB studied here, including FB-WT, had a similar Rmax value to the others (Table I). Together, these data supported our hypothesis by showing that SL1100 retained high-affinity binding to these aHUS-linked variants in FB.

FIGURE 3. Direct Binding Studies of Various FB Proteins with Modified DNA Aptamer SL1100.

5’-biotinylated SL1100 was captured on three experimental flow cells of a streptavidin-coated biosensor, leaving one flow cell unmodified for reference subtraction. Protein-aptamer interactions were monitored by injecting increasing concentrations of FB in single-cycle mode, where 1 μM was used as the highest concentration. Reference-subtracted sensorgrams (black lines) were analyzed using a two-state kinetic model (red lines). (A) Binding of FB-Native to SL1100. (B) Binding of FB-WT to SL1100. (C) Binding of FB-D279G to SL1100. (D) Binding of FB-F286L to SL1100. (E) Binding of FB-K323E to SL1100. (F) Binding of FB-K350N to SL1100. (G) Binding of FB-2Mut to SL1100. (H) Binding of FB-3Mut to SL1100. For panels A-H, the data shown represents one of three independent replicates.

SL1100 Blocks C3 Proconvertase Formation by Gain-of-Function Variants of FB but Does Not Affect C3 Convertase Stability

Our earlier work established that the FB-binding aptamer family typified by SL1100 inhibits the AP by blocking the formation of the C3 proconvertase complex, C3bB (Fig. 1C; (34)). However, most of the aHUS-associated FB variants described in previous publications exhibit increased affinity for C3b and/or increase the stability of the AP C3 convertases they form ((25–27); Fig 2 & Table I). Therefore, we considered whether the high-affinity binding by SL1100 could supersede the functional consequences of these FB mutations. To investigate this possibility, we employed two different SPR-based methods that monitor the formation of the C3 proconvertase and the stability of the already formed C3 convertase, respectively (34, 41).

We initially examined whether SL1100 could effectively block C3b-binding by the gain-of-function forms of FB (Fig. 4). We monitored FB binding to site-specifically immobilized C3b across a dilution series consisting of seven different FB concentrations, and as a function of increasing molar ratios of SL1100 competitor relative to FB (i.e. no competitor, 0.25:1, 0.5:1, 0.75:1, and 1:1 stoichiometry; (34)). Following normalization of the signal for each data point relative to the signal generated by the highest concentration of FB in the absence of any competitor, we plotted the residual SPR signal as a function of FB concentration and the level of SL1100 present. We found that increasing concentrations of SL1100 resulted in progressively greater inhibition of FB binding to C3b. Regardless of the specific mutation under investigation, we observed that SL1100 could completely block FB binding to C3b when present at equimolar concentrations. Together, these results showed that SL1100 reduced the effective concentration of FB, thereby inhibiting C3 proconvertase formation by each of the FB variants available.

FIGURE 4. Modified DNA Aptamer SL1100 Inhibits C3 Proconvertase Formation by Gain-of-Function FB Variants.

Site-specifically biotinylated C3b was captured on three experimental flow cells of streptavidin-coated biosensor, leaving one flow cell unmodified for reference subtraction. A two-fold dilution series of FB spanning seven different concentrations from 15.625 nM to 1 μM was injected over the surface, either in the absence or presence of increasing molar ratios of SL1100 to FB. SPR signals were normalized for each flow cell by treating the response immediately prior to injection stop for the highest concentration of FB without any inhibitor as 100%. The residual signals were then plotted as a function of FB for each ratio of SL1100 to FB. (A) Normalized SPR response of FB-Native binding to C3b. (B) Normalized SPR response of FB-WT binding to C3b. (C) Normalized SPR response of FB-D279G binding to C3b. (D) Normalized SPR response of FB-F286L binding to C3b. (E) Normalized SPR response of FB-K323E binding to C3b. (F) Normalized SPR response of FB-K350N binding to C3b. (G) Normalized SPR response of FB-2Mut binding to C3b. (H) Normalized SPR response of FB-3Mut binding to C3b. Increasing aptamer levels diminished FB binding to the surface at each FB concentration examined. For panels A-H, the data shown represents one of three independent replicates.

We then explored whether SL1100 could affect the stability of pre-formed C3 convertases generated with each FB variant (Fig. 5). We used an SPR-based assay to monitor C3 convertase formation and decay in real time (41). In this first phase of such experiments, a mixture of FB and FD are injected over captured C3b-biotin to generate the C3 convertase; following a brief period wherein the rate of intrinsic C3 convertase decay is measured, the second phase is begun by injecting proteins (e.g., FH) or other molecules to determine if they accelerate the rate of C3 convertase decay. We validated this approach by showing that both serum-purified (i.e., FB-Native) and recombinant FB (i.e., FB-WT) formed transiently stable C3 convertases whose decay was significantly accelerated by injecting FH over each surface (Fig. 5A–B). However, we did not detect appreciable decay-accelerating activity when we injected SL1100 at 100 nM, 500 nM, or 1000 nM instead of FH over C3 convertases formed by FB-Native or FB-WT.

FIGURE 5. Modified DNA Aptamer SL1100 Does Not Alter C3 Convertase Stability.

Site-specifically biotinylated C3b was captured on three experimental flow cells of a streptavidin-coated biosensor, leaving one flow cell unmodified for reference subtraction. Reference-corrected sensorgrams were used to follow the formation and decay of the AP C3 convertase in real-time. The first experimental cycle involved injection of a solution of 100 nM FB/FD followed by HBS-T buffer alone to monitor formation and decay of the C3 convertase, respectively (black trace). The second cycle involved FB/FD injection followed by FH to illustrate decay acceleration of the C3 convertase (red trace). The third, fourth, and fifth cycles involved FB/FD injection followed by 100 nM SL1100 (blue trace), 500 nM SL1100 (magenta trace), or 1000 nM SL1100 (orange trace) to determine the effect of SL1100 on C3 convertase decay. Arrows indicate the start and stop points of the FB/FD injection and the decay injection sequences. (A) C3 convertase formed by FB-Native. (B) C3 convertase formed by FB-WT. (C) C3 convertase formed by FB-D279G. (D) C3 convertase formed by FB-F286L. (E) C3 convertase formed by FB-K323E. (F) C3 convertase formed by FB-K350N. (G) C3 convertase formed by FB-2Mut. (H) C3 convertase formed by FB-3Mut. For panels A-H, the data shown represents one of three independent replicates.

We conducted analogous experiments on the remaining FB variants in our panel. Although FB-D279G displayed only a ~2.5-fold increase in affinity for C3b (Table I), we found that its fully formed C3 convertase was significantly more stable than that of FB-WT (Fig. 5C). The FB-D279G convertase was only mildly decayed by FH, and not all when SL1100 was injected instead. By contrast, the increased rate of proconvertase formation seen for FB-F286L did not translate into an increased level of convertase stability, as we found that its fully formed C3 convertase decayed faster than all other FB variants studied here (Fig. 5D). In fact, the FB-F286L convertase seemed insensitive to decay acceleration by both FH and SL1100, but this was most likely due to an already significant decay rate in the absence of any exogenous factors. Whereas both FB-K323E (Fig. 5E) and FB-K350N (Fig. 5F) behaved similarly to each other and to FB-WT insofar as C3 convertase formation was concerned, the FB-K323E convertase was more resistant to decay acceleration by FH; however, neither of these convertases could be dissociated by injections of SL1100. Finally, the C3 convertases formed by FB-2Mut (Fig. 5G) and FB-2Mut (Fig. 5H) exhibited the increased stability found in the FB-D279G convertase but appeared to be fully resistant to decay acceleration by FH and SL1100. Although the FB-K350N convertase was not as resistant to decay acceleration as the FB-D279G convertase, including both mutations concurrently in the FB-2Mut and FB-3Mut variants seemed to have an additive effect in promoting convertase stability. While our observations here largely confirmed previous work showing that these FB variants generated intrinsically stabilized C3 convertases and/or convertases resistant to decay (25–27), we further found that SL1100 could not diminish the stability of any C3 convertases once they had formed.

Inhibition by SL1100 Supersedes the Increased Alternative Pathway Activity Resulting from FB Gain-of-Function

Although the results presented above showed that SL1100 blocked proconvertase formation by gain-of-function variants of FB in purified settings (Fig. 4), we sought further insight into whether this could also be seen in more complex assays of complement activity. To investigate this issue, we monitored hemolysis of rabbit erythrocytes resulting from AP activity in the absence or presence of SL1100 using FB-depleted human serum reconstituted with various forms of FB (34). We first evaluated hemolytic activity as a function of FB concentration using 12% (v/v) FB-depleted serum (Fig. 6A). We found that FB-Native and FB-WT had similar activities to one another as judged by EC₅₀ values of 19.8 nM and 15.8 nM, respectively. Both FB-F286L and FB-K323E exhibited a ~4-fold increase in potency, as judged by their EC₅₀ values of 3.6 nM and 3.8 nM, respectively, while FB-K350N was ~8-fold more potent with an EC₅₀ value of 2.0 nM. The most potent hemolytic activities were observed for FB-D279G and FB-2Mut, with EC₅₀ values of 0.9 nM and 1.0 nM, respectively. Curiously, neither of these FB variants achieved full hemolysis at the highest concentrations of FB examined; they even showed some evidence of inhibition at concentrations above 100 nM in this assay. FB-3Mut did not support any hemolytic activity, as expected, since this variant contains a serine protease catalytic site mutation.

FIGURE 6. Inhibition by the Modified DNA Aptamer SL1100 Supersedes the Increased Alternative Pathway Activity Resulting from FB Gain-of-Function.

Rabbit erythrocytes were incubated for 30 min at 37 °C with 12% (v/v) FB-depleted serum that had been fortified with various forms of FB, either in the absence or presence of SL1100. Hemolysis was assessed spectrophotometrically, and the data were fit to dose-response curve (please see Materials and Methods). (A) AP-driven hemolysis using 12% (v/v) FB-depleted serum fortified with increasing concentrations of various FB proteins. Observed EC₅₀ values were as follows: FB-Native, 19.8 nM; FB-WT, 15.8 nM; FB-D279G, 0.9 nM; FB-F286L, 3.6 nM; FB-K323E, 3.8 nM; FB-K350N, 2.0 nM; FB-2Mut, 1.0 nM; FB-3Mut, no activity. (B) Inhibition of hemolysis using 12% (v/v) FB-depleted serum fortified with 60 nM FB-Native or FB-WT in the presence of increasing concentrations of SL1100. Observed IC₅₀ values for SL1100 were as follows: FB-Native, 36.7 nM; FB-WT, 31.7 nM. (C) Identical to panel B, except that 12% (v/v) FB-depleted serum was fortified with 20 nM of each gain-of-function FB variant. Observed IC₅₀ values for SL1100 were as follows: FB-D279G, 33.9 nM; FB-F286L, 32.4 nM; FB-K323E, 26.5 nM; FB-K350N, 36.9 nM; FB-2Mut, 23.8 nM). For panels A-C, note that normalized values from six technical replicates were expressed as a function of FB protein or SL1100 concentration and were fit to a four-parameter dose-response curve.

To examine whether SL1100 could supersede the enhanced AP activity induced by these gain-of-function FB variants, we evaluated its inhibitory potency by adding increasing concentrations of SL1100 into hemolysis experiments that used 12% (v/v) FB-depleted human serum fortified with FB. We used 60 nM FB-Native and FB-WT to ensure sufficient hemolytic activity in the absence of any inhibitor and found that SL1100 blocked the AP with IC₅₀ values of 36.7 nM and 31.7 nM, respectively (Fig. 6B). We repeated these experiments for the gain-of-function variants, with the exception that we used 20 nM FB for these forms of FB since they showed increased AP activity relative to wild-type (Fig. 6A). We found that SL1100 inhibited the AP activity arising from each FB variant within a relatively narrow IC₅₀ range from 23.8 nM to 33.9 nM. We also carried out these experiments using the same concentration of FB (i.e. 60 nM) for both the wild-type FB and various mutants (Fig. S2A). These results were in good agreement with the previous experiment, and showed that increased levels of SL1100 were required to inhibit the gain-of-function arising from mutations in FB. It should be noted that these experiments collectively reflected several functionally critical parameters, including the affinity of each FB variant for C3b (Fig. 2), the affinity of SL1100 for each form of FB (Fig. 3), and the intrinsic features of the C3 convertases these FB variants form (Fig. 5). Nevertheless, these observations were consistent with the prediction that SL1100 counteracts the enhanced hemolytic activity induced by gain-of-function FB mutants.

DISCUSSION

aHUS is characterized by hemolytic anemia, thrombocytopenia, and renal failure, but a dysregulated complement AP is widely acknowledged as the molecular basis of this disease (13, 27). Genetic studies have established that aHUS is most often associated with loss of function in AP regulatory proteins, including FH, MCP, and DAF (14–16, 27). However, a smaller subset of aHUS patients expresses gain-of-function mutations in AP components themselves, including FB (25–27). Several of these FB mutations increase the rate at which the C3 convertase forms or convertase stability thereafter (25–27). This suggests that inhibiting the initial interaction of FB with C3b could be an effective strategy for preventing excessive AP activity resulting from these gain-of-function mutations in FB. In this report, we provided proof-of-concept for this approach. We showed that a prototypic member of a previously described family of FB-binding modified DNA aptamers effectively bound to disease-causing gain-of-function FB variants (Fig. 3, Fig. S1 & Table I), thereby blocking C3 proconvertase formation (Fig. 4 & Fig. S4) and inhibiting activity of the AP (Fig. 6 & Fig. S2).

Molecular studies on FB variants identified from aHUS patients have been a rich source of information on the formation, decay, and regulation of the AP C3 convertase (25–27). Because most of these FB variants were described by different investigator teams at separate times, there has not been a concurrent, systematic investigation of these FB variants until now. Here we compared four aHUS-linked FB variants using well-established assays to interrogate relevant properties of the AP C3 convertase, including the rate of proconvertase formation (Fig. 2) and how efficiently it is activated (Fig. 5), its intrinsic stability and sensitivity to decay acceleration (Fig. 5), and its ability to cleave C3 and support downstream complement activity (Fig. 6). Our results largely agree with the observations from other investigators, with a few potentially interesting exceptions. First, we found that the FB-D279G and FB-K350N variants have weaker apparent affinities for C3b than earlier work might have predicted (25). Roumenina et al. reported that FB-D279G and FB-K350N bound to hydrolyzed C3 (i.e. C3(H₂O)) with apparent affinities of 1.4 nM and 1.8 nM, while wild-type FB bound with an apparent affinity of 105 nM (25); here, we found that FB-D279G and FB-K350N bound to C3b with apparent affinities of 71 nM and 322 nM, while wild-type FB had an apparent affinity of 180 nM (Table I). Although C3(H₂O) and C3b are not structurally identical to one another (46), the long-standing observation that C3(H₂O) and C3b share many functional properties suggests that the values should have been more similar (47, 48). Second, the apparent affinity for C3b of FB-K350N we observed here was surprisingly weaker than that of wild-type FB (Table I). Earlier work reported the apparent affinity of FB-K350N for C3(H₂O) was comparable to that of FB-D279G and nearly ~60-fold tighter than wild-type FB. The reasons for this disparity are unclear since the FB-K350N convertase was partly resistant to decay by FH (Fig. 5) and exhibited gain-of-function in hemolytic assays as expected (Fig. 6).

A final noteworthy difference we observed while characterizing FB variants pertains to the unusual functional properties of FB-D279G. As mentioned in the paragraph above, we found that FB-D279G was not as dramatically enhanced in the rate of proconvertase formation as previously reported ((25, 26); Table I). Once the FB-D279G convertase was formed, however, we found that it was remarkably stable and nearly completely resistant to decay acceleration by FH (Fig. 5); these observations extend those of Rodríguez de Córdoba et al., who showed that the FB-D279G convertase was significantly stabilized against decay acceleration by DAF (27). We further found that simultaneous D279G and K350N mutations in FB variants resulted in convertases that appeared completely resistant to FH decay, as judged by studies on FB-2Mut and FB-3Mut (Fig. 5). This suggested that the convertase-stabilizing attributes of these individual mutations are additive. Perhaps most unusually, though, we noticed that FB-D279G did not support full activity in hemolytic assays of complement function (Fig. 6); in fact, we even found evidence of dose-dependent inhibition in these experiments at FB-D279G concentrations above 30 nM. Both features were shared by FB-2Mut, suggesting these effects were likely not the result of an artifact. While we cannot provide a definitive explanation for these observations at this time, it seems the effective levels of FB available for productive convertase formation were reduced under these conditions. This could have been caused by the binding of FB-D279G and FB-2Mut to alternative ligands like iC3b (25), which may not generate catalytically efficient C3 convertases. Future studies could explore this possibility in greater detail.

The hemolytic assays we performed utilized FB-depleted serum that was fortified with various forms of FB under investigation (Fig. 6). This serum contained otherwise normal levels of complement components, including FP, which is the only positive regulator of the AP (5). Since the activities of FP oppose those of the FB-binding aptamers studied here, we sought additional information on how SL1100 might impact FP function. We found that the concentration of SL1100 needed to inhibit the hemolytic activity of normal human serum (NHS) was only slightly greater than that needed to inhibit FP-depleted serum (Fig. S2B), even though the rate at which FP-depleted serum lysed rabbit erythrocytes was considerably slower than NHS. Furthermore, while FP promoted C3 convertase formation for each FB variant examined and stabilized their respective convertases after activation (Fig. S3), we observed that inclusion of SL1100 in the initial stage of this experiment completely blocked C3 proconvertase formation regardless of whether FP was present or not (Fig. 4). This indicates that the inhibitory effect of SL1100 supersedes the stimulatory activities of FP. It is also in agreement with our earlier observations that the IC₅₀ of SL1100 in hemolytic assays reflects the levels of FB present in the serum (Fig. 6 & Fig. S2).

Whereas earlier work suggested that the affinity of these aptamers for FB lay in the 10-100 pM range (34), here we found that the interaction was weaker in the realm of 1-10 nM (Fig. 3 & Table I). This discrepancy could be due to differences in the solution-based bead-binding assays used previously instead of the solid-phase SPR method used here. Differences could also be attributed to the specific 5’-biotinylated linker we used, which may have somehow influenced FB binding to the immobilized aptamer. Nevertheless, the SL1100 aptamer used in our current studies showed comparable binding to all forms of FB independent of any mutations they contained (Fig. 3 & Table I). We attribute the generally similar affinities we observed to the fact that the aptamer binding site on FB is removed from the region of its vWF domain where these aHUS-linked mutations cluster. In hindsight, assuming this would have been the case seems reasonable. However, we thought it was essential to rule out any differences in affinity due to unexpected changes in protein dynamics or other minor structural alterations caused by these mutations. Regardless of the genuine affinity of these aptamers for FB, the concentration of FB in human plasma is relatively high (~200 μg/mL or ~2.1 μM); thus, the concentrations of aptamer needed to approach stoichiometric equivalence would almost certainly yield FB/aptamer complexes, resulting in potent inhibition of C3 proconvertase formation and blocking the AP. Because of this, affinity seems less of a limiting factor to potentially using these aptamers for therapeutic purposes than their pharmacologic properties, which are currently unknown.

Considering that aptamers offer several significant advantages over traditional protein therapeutics, including synthetic accessibility, high solubility, and an apparent lack of immunogenicity (49, 50), the developmental landscape of aptamer drugs has been relatively weak for the past two decades. The first aptamer-based drug, Macugen (pegaptanib), was approved by the FDA in 2004 (50). Macugen is a pegylated, 27-mer RNA aptamer that targets and inhibits vascular endothelial growth factor for treatment of wet age-related macular degeneration (AMD), a major cause of vision loss. Although Macugen foreshadowed the potential of aptamers as therapeutics, it took till 2023 for the next aptamer-based drug, Izervay (avacinaptad), to receive FDA approval (51). Coincidentally, Izervay is also used to treat AMD. This pegylated, 39-mer RNA aptamer is a complement C5 inhibitor that prevents C5 activation, leading to the blockade of the terminal complement pathway (52). The limited portfolio of aptamer-based drugs likely reflects technical challenges they face during development, including metabolic instability, susceptibility to rapid renal filtration due to their small size, and unpredictable pharmacokinetic properties, all of which can diminish their efficacy in vivo (49, 50, 53, 54). Indeed, both approved aptamer drugs are modified RNAs that include polyethylene glycol chains to extend their half-lives in vivo, and inverted thymidine nucleotides, fluorinated and O-methylated ribose bases to increase their resistance to degradation. Although the FB-binding DNA aptamer we investigated here (i.e. SL1100) lacks such modifications, other members of this aptamer family (i.e. SL1102 and SL1103) contain 2’-O-methylations that improve their serum stability without significantly affecting their potency (34). Moving forward, the extensive structure/function information available for this aptamer family should facilitate the design of pegylation strategies compatible with preserving their bioactivity (34).

While effective therapies for aHUS are already in the clinic (31), further development of these FB-binding aptamers could provide valuable new tools for treating other complement-driven diseases aside from aHUS (35). In this regard, there are many other diseases where excessive or uncontrolled complement activity is believed to play a contributing, if not causal role (35). Since the AP is the main driver of complement amplification (8), molecules that block any components of its convertase (i.e. C3b, FB, and FD) or its substrate (i.e. C3) could be useful therapeutic modalities for these diseases. Indeed, inhibitors of C3, FD, and FB have received FDA approvals in recent years for various conditions. Although the FB-binding aptamers we describe here would not be unique in their ability to block complement amplification via the AP, their synthetic nature, high solubility, and potent activities may impart certain advantages that justify continued research and development of these unusual molecules throughout the coming years.

KEY POINTS.

• Gain of function forms of Factor B cause AP dysregulation, leading to disease.

• Factor B-binding aptamers block C3 convertase formation and inhibit the AP.

• Inhibition by Factor B-binding aptamers supersedes gain of function mutations.

ABBREVIATIONS

aHUS

atypical hemolytic uremic syndrome

AMD

age-related macular degeneration

AP

alternative pathway of complement

C3

complement component C3

C3b

complement component C3b

C4b

complement component C4b

C5

complement component C5

CCP

complement control protein domain

CP

classical pathway of complement

DAF

decay accelerating factor

FB-2Mut

recombinant FB with D279G and K350N

FB-3Mut

recombinant FB with D279G , K350N, and S699A

FB-D279G

recombinant FB with D279G mutation

FB-F286L

recombinant FB with F286L mutation

FB-K323E

recombinant FB with K323E mutation

FB-K350N

recombinant FB with K350N mutation

FB-Native

Native FB purified from human serum

FB-WT

recombinant wild-type FB

FB

complement factor B

FD

complement factor D

FH

complement factor H

FI

complement factor I

FP

properdin

HBS-T

HEPES buffered saline with tween-20

HBS

HEPES buffered saline

LP

lectin pathway of complement

MAC

membrane attack complex

MCP

membrane cofactor protein

MIDAS

metal ion-dependent adhesion site

PEG

polyethylene glycol

RCA

regulator of complement activity

SP

serine protease domain

SPR

surface plasmon resonance

vWF

von Willebrand factor domain