Functional and Structural Implications of the Complement Factor H Y402H Polymorphism Associated with Age-Related Macular Degeneration

Abstract

purpose. A Tyr-to-His (Y402H) sequence variant in the factor H (FH) and factor H-like protein (FHL-1) gene is strongly associated with an increased susceptibility for age-related macular degeneration (AMD). The purpose of this study was to understand how the Y402H variant in FH/FHL-1 contributes to the pathogenesis of AMD and, in particular, whether interactions mediated by FH/FHL-1, including binding to C-reactive protein (CRP), group A streptococcal M protein (GAS M6), heparin, and retinal pigment epithelial cells (RPE), are affected.

methods. FH was purified from sera of patients homozygous for FH(Y402) or (H402), and recombinant FH fragments representing FHL-1 were generated. Proteins were analyzed for binding to CRP, GAS M6, heparin, and RPE cells.

results. Binding of the FH and FH1 to seven polymorphic variants to CRP and M protein was reduced. The variant did not influence the interaction of FH with heparin but did reduce binding of FHL-1. Binding of the FH and FHL-1 polymorphic variant to RPE cells was not affected.

conclusions. The FH Y402H polymorphism associated with AMD causes a reduction in binding of FH and FHL-1 to CRP and M protein. Both variants show comparable binding to RPE cells, indicating that AMD is unlikely to manifest as a result of impaired host cell-surface recognition. The decreased interaction between FH and CRP, which is essential for the anti-inflammatory function of CRP, provides a possible pathophysiological explanation for the association of the Y402H variant with AMD.

Age-related macular degeneration (AMD) is the primary cause of irreversible blindness in people older than 65 years of age, affecting more than 50 million people in the Western, industrialized world. ¹ It is characterized by the accumulation of extracellular lesions, termed drusen, that form beneath the basal lamina of the retinal pigment epithelium within the Bruch membrane. ² Although drusen are considered precursors to the neovascular changes and progressive degeneration of the retinal pigment epithelium and choroid that lead to central vision loss, the mechanisms by which drusen are associated with neovascularization are unclear. Chronic inflammation and complement activation are important processes in drusen biogenesis, and complement components including C3, C5, C5b-9, factor H (FH), vitronectin, and clusterin have been identified as constituents of drusen. ^{2 3 4 5}  

Genetic variants in the gene encoding complement FH/ factor H-like protein 1 (FHL-1) have been identified as a major risk factor for the development of AMD. ^{6 7 8 9 10} A common missense variant at position 402 (position 384 in the mature polypeptide) of FH, resulting in a Tyr-to-His (Y402H) exchange, increases the risk for AMD 4.6-fold in persons heterozygous for the haplotype and 7.4-fold in persons homozygous for it. ⁶ The Y402H polymorphism is located in exon 9 of the FH/FHL-1 gene, which encodes for the seventh short consensus repeat (SCR). FH is composed of 20 SCR units; each consists of approximately 60 amino acids separated by a short linker region of three to eight amino acids. FHL-1 is a splice variant of the FH gene and consists of the first seven SCRs of FH and an additional four hydrophobic amino acids at the C terminus. ^{11 12 13} FH plays a critical role in regulating the alternative pathway (AP) of complement. The AP is activated by bacteria, parasites, fungi, neoplastic cells, and damaged host cells, potentially labeling them for destruction by lysis or phagocytosis. ¹⁴ However, FH binds to and inactivates C3b bound to host cells, protecting them from complement-mediated damage while allowing complement activation to proceed on nearby foreign surfaces. The ability of FH to selectively protect host cells resides in its capacity to bind to surface-associated sialic acids and polyanions such as glycosaminoglycans (GAGs). Three heparin/polyanion-binding domains have been reported for FH, including one within SCR7. ^{15 16} In addition, FH interacts with group A streptococcal (GAS) M protein through SCR7 and with C-reactive protein (CRP) through binding regions within SCR7 and SCR8–11. ^{17 18 19 20 21} FHL-1 has a number of activities similar to those of FH, including cofactor and decay-accelerating activities ^{22 23} and other interactions mediated by SCR7. Interestingly, GAS M protein binds FHL-1 in preference to FH through SCR7 ¹⁹ and the hydrophobic tail. ²⁴ This enables the bacterium to restrict C3b deposition on its surface, aiding survival within the host. ²⁵ Furthermore, binding of FHL-1 by SCR7 facilitates GAS invasion of epithelial cells. ²⁶ The purpose of this study was to determine the effect of the Y402H polymorphism on the interaction of FH and FHL-1 with its various ligands. 

Methods

Model Building

The FH SCR6–7 (FH6–7) structural models were constructed as described, ²⁷ based on alignment with FH SCR15–16 and Vaccinia virus complement control protein (VCP) SCR3–4, using multiple sequence alignment and homology modeling procedures. ²⁸ The initial model was iteratively refined using in-built molecular dynamics with simulated annealing protocols to improve structural quality, as computed by structural quality analysis software. ²⁹ Electrostatic potentials of protein surfaces were computed and visualized using a graphical analytical tool. ³⁰  

Receptor-Ligand Docking

Docking was performed using the internal coordinate mechanics (ICM) pseudo-Brownian flexible ligand/grid receptor docking algorithm, with the Empiric Conformational Energy Program for Peptides 3 (ECEPP/3) ³¹ force field and Merck Molecular Force Field (MMFF) ³² partial charges for the ligand and a grid map representation of the receptor energy with hydrophobic, electrostatic, hydrogen-bond, and surface potential terms. ^{33 34 35} The initial model was optimized by local energy minimization, incorporating flexibility to receptor side chains within 6 Å of the ligand. ³⁵ VCP in complex with heparin (PDB 1RID) provided coordinates for heparin. ³⁶ H402 was treated as a charged residue in the docking studies. 

Construction, Expression, and Purification of Recombinant FH Constructs

Generation of the recombinant FH fragment FH1–7(Y402) in the Pichia pastoris vector pPICZαA has been described previously. ³⁷ A single nucleotide exchange in SCR7, representing the Y402H sequence variant, was introduced into SCRs 1–7 using the site-directed mutagenesis system (Gene Tailor; Invitrogen, San Diego, CA) according to the manufacturer’s instructions. Recombinant proteins were expressed for 3 days in the presence of 4% methanol and then purified from culture supernatants by immunoaffinity chromatography, as previously described. ³⁸  

Identification of FH(Y402) and (H402) Variants from Serum

Subjects were recruited as part of a study, approved by the joint ethics committee of Flinders University of South Australia and Flinders Medical Centre, of genetic risk factors in AMD. The research adhered to the tenets of the Declaration of Helsinki. Genomic DNA extracted from leukocyte samples was genotyped for the FH Y402H variant with the use of an assay (SNaPshot Multiplex; Applied Biosystems, Foster City, CA). Key domains critical for the interactions to be studied (SCR7 and SCR20) were sequenced to ensure that rare additional polymorphisms that could influence binding were not present. Fresh serum was obtained from subjects identified as homozygous for the FH(Y402) or (H402) variant. FH was purified from serum by immunoaffinity chromatography, as previously described. ³⁸  

Protein Quantification

Serum-purified or recombinant FH proteins were separated on a 5% to 15% gradient minigel in conjunction with 200 to 1000 ng BSA (Pierce, Rockford, IL) to generate a standard curve. After staining with Coomassie blue, protein concentrations were determined using specific software (GS-700 Imaging Densitometer and Molecular Analyst Software 2.1.2; Bio-Rad, Richmond, CA). To confirm that concentrations of the polymorphic variant proteins were equivalent, proteins were serially diluted onto nitrocellulose and incubated with goat antihuman FH antiserum (1:2000 vol/vol) followed by alkaline phosphatase-conjugated donkey antigoat IgG (1:1000 vol/vol; Jackson ImmunoResearch Laboratories, West Grove, PA). Enhanced chemifluorescence substrate (GE Healthcare, Little Chalfont, Buckinghamshire, UK) was added, fluorescence was determined (Typhoon Variable Mode Imager 9410; GE Healthcare), and analysis was performed (Image Quant TL version 2005 software; GE Healthcare). 

CRP ELISA

Purified CRP (2.5 μg/mL in PBS; generously provided by Carolyn Mold, University of New Mexico) was coated onto the wells of 96-well ELISA plates (Maxisorb; Nunc, Roskilde, Denmark) and incubated at 4°C overnight. Negative control wells were incubated with gelatin. After blocking with 1% BSA in gelatin veronal buffer (GVB; 1% gelatin, 5 mM veronal, 145 mM NaCl, pH 7.43) for 2 hours at 22°C, increasing concentrations of FH or FH1–7 proteins were added to wells containing CRP or gelatin and incubated for 1 hour. Wells were washed and incubated sequentially with goat antiserum against human FH (1:2000 vol/vol) and peroxidase conjugated rabbit antigoat IgG (1:30,000 vol/vol). Substrate (100 μL ο-phenylenediamine; Sigma-Aldrich, St. Louis, MO) was added, and OD490 nm was determined. Experiments were performed in duplicate and repeated at least three times. To enable comparison between experiments and to determine statistical significance, the highest concentration of the (Y402) variant in each experiment was considered to represent 100% binding, and OD readings were calculated as a percentage of this value. 

Heparin-Sepharose Affinity Chromatography

Heparin-Sepharose binding experiments were conducted using a 1-mL heparin HP column (HiTrap; GE Healthcare) connected to a fast protein liquid chromatography system (FPLC; ÁKTAprime; GE Healthcare). Experiments were conducted at 25°C in PBS using a flow rate of 1 mL/min. Serum-purified or recombinant FH proteins (90 μg) were injected onto the column, washed extensively with PBS, and eluted with a linear salt gradient from 150 to 500 mM NaCl. Five hundred-microliter fractions were collected and analyzed by SDS-PAGE and Western blotting. Each experiment was performed three times. 

Isolation and Culture of Human Retinal Pigment Epithelial Cells

Postmortem eyes without known ophthalmic disease were obtained from the Lions Eye Bank, Flinders Medical Centre (South Australia) after corneal trephination. The research was approved by Flinders Medical Centre Research Ethics Committee. Studies were conducted adhering to the tenets of the Declaration of Helsinki. Eyes from donors ranging in age from 55 to 76 years were enucleated up to 12 hours after death. The anterior segment, vitreous, and neural retina were removed, and the exposed retinal pigment epithelial (RPE) layer was washed twice with sterile PBS. PBS was added to the eyecup, and the RPE cells were removed by gentle scraping. Cells were washed in media (DMEM/Hams F12 [1:1 vol/vol]; Sigma-Aldrich) supplemented with 2 mM l-glutamine, 20% FCS, and antibiotics), resuspended in fresh media in 25-cm² tissue culture flasks, and incubated in 5% CO₂ at 37°C. After 3 days, nonadherent cells were removed, and fresh media containing 10% FCS were added. Cultures were maintained with biweekly feedings and were grown to 50% to 80% confluence. 

Flow Cytometry

Cells were washed in PBS to remove endogenous or FCS-derived FH and were returned to FCS-free media for 1 hour. Cells were detached by incubation in 20 mM EDTA, washed in PBS, and blocked in 3% BSA, 10 mM Tris-HCl, 175 mM PBS on ice for 30 minutes After washing with PBS, 5 × 10⁴ cells were incubated with 10 μg purified FH or FH1–7 diluted in PBS at 37°C for 1 hour. Cells were washed and incubated with goat antiserum against FH (1:160 vol/vol) followed by FITC-conjugated rabbit antigoat IgG (1:50 vol/vol; Chemicon, Temecula, CA). Cells were analyzed using a FACScan (BD Biosciences, San Jose, CA). Experiments were performed on a minimum of three cell populations from different donors. 

GAS M6 ELISA

Purified M6 protein (5 μg/mL), generously provided by Vince Fischetti (Rockefeller University, New York, NY), was coated onto the wells of 96-well ELISA plates (Maxisorb; Nunc) and incubated at 4°C overnight. Negative control wells were incubated with BSA. Nonspecific sites were blocked with 5% skim milk for 1 hour at 22°C. Increasing concentrations of serum-purified FH or recombinant FH1–7 proteins, diluted in PBS, were incubated for 3 hours. Wells were washed with PBS, and binding was assessed as described for the CRP ELISA. Experiments were performed in duplicate and repeated at least three times. Data were analyzed as described for the CRP ELISA. 

Results

Molecular Modeling

With the use of homology modeling ²⁷ and site-directed mutagenesis, ³⁷ we have previously identified the residues in FH SCR6–7 responsible for interactions with heparin, CRP, and GAS M6 protein. This model, which contains the Tyr residue at position 402, is shown in Figure 1A . A homology model of FH SCR6–7 containing the Y402H variant was constructed (Fig. 1B) , and the electrostatic surface charges of the two variants were compared. FH 6–7(Y402) appears to have two separate areas of positive charge (shown in blue) in contrast to FH 6–7(H402), which shows an extended but evenly distributed surface charge. 

Using the crystal structure of heparin, ³⁶ we assessed the docking of both variants with heparin (Fig. 2) . The heparin and FH 6–7 complex of both variants is electrostatically anchored by R387, K388, R404, K405, and K410, which is in agreement with our previous site-directed mutagenesis studies showing that these residues are implicated in the interaction of FH 6–7 with heparin. ³⁷ In addition, the docking model indicates that heparin also forms electrostatic contacts with Q408 and Y402/H402. However, whereas Y402 forms two hydrogen bonds with heparin, H402 only forms a single bond. 

Identification and Purification of FH(Y402) and FH(H402) from Serum

After identification and purification of FH(Y402) and FH(H402) from sera, proteins were analyzed by SDS-PAGE (Fig. 3A)and Western blotting. Both fragments had an identical apparent molecular mass of approximately 130 kDa and were detected by antiserum against human FH. 

Expression of Recombinant FH1–7(Y402) and FH1–7(H402)

To investigate the effect(s) of the Y402H variant on the interactions and activities of FH/FHL-1, we constructed and expressed two truncated recombinant FH fragments comprising SCR1–7 with either a Tyr or a His residue at position 402. Construct integrity of each polymorphic variant was verified by DNA sequencing. Both fragments had an identical apparent molecular mass of approximately 37 kDa when visualized by SDS-PAGE (Fig. 3B) . FH and FH1–7 proteins were detected by antiserum against human FH and demonstrated similar fluid-phase cofactor activity (data not shown). 

Interaction of FH and FH1–7(Y402) and FH1–7(H402) Variants with CRP

To determine the effect of the Y402H polymorphism on CRP binding, we analyzed the binding of serum-purified and recombinant FH proteins to CRP by ELISA. A 5- to 10-fold molar excess of FH protein relative to the ligand was used to ensure saturation was reached. FH and FH1–7 bound to CRP in a dose-dependent and saturable manner. Polymorphic variants of FH and FH1–7 showed similar low levels of binding (4%–19% of positive readings) to the negative control wells. Figure 4demonstrates that FH(H402) and, in particular, FH1–7(H402) has reduced binding to CRP relative to the Tyr variants. The (H402) variant appears to have less of an effect on the binding of FH than the binding of FH1–7, presumably because of the presence of an additional CRP-binding site in FH within SCR domains 8 to 11. 

Heparin-Binding Properties of FH and FH1–7(Y402) and FH1–7(H402) Variants

We have previously shown that SCR7 of FH contains a heparin/polyanion-binding domain. ^{15 37} To determine whether the AMD-associated polymorphism in SCR7 affects the interaction of FH/FHL-1 with heparin, we subjected purified FH(Y402), FH(H402), and recombinant FH1–7(Y402) and FH1–7(H402) variants to FPLC-controlled heparin-affinity chromatography. Comparison of the elution profiles of the FH polymorphic variants showed that both variants eluted at similar salt concentrations (Fig. 5A) , indicating that the presence of the His residue has no apparent effect on the interaction of FH with heparin. However, the recombinant FH1–7(H402) construct consistently eluted slightly earlier than the FH1–7(Y402) variant (Fig. 5B) . Elution of FH1–7(H402) began at approximately 174 mM NaCl compared with the(Y402) variant, which started eluting at approximately 195 mM NaCl. This is comparable to the decrease observed by Clark et al. ³⁹  

Binding of FH and FH1–7 Y402H Variants to RPE Cells

Immunohistochemistry studies have demonstrated the deposition of FH at the retinal pigment epithelium–choroid interface of AMD patients. ⁸ To determine whether the Y402H variant influences the binding of FH to RPE cells, we examined the binding of FH and FH1–7 to primary RPE cells by FACS. Figure 6shows that both FH (A) and FH1–7 (B) bind to RPE cells; however, this interaction is not affected by the presence of the Y402H polymorphic variant in SCR7. 

Interaction of FH and FH1–7(Y402) and FH1–7(H402) Variants with GAS M6 Protein

To determine whether the AMD-associated polymorphism also influences the interaction of FH/FH1–7 with GAS M6, we examined the binding of FH and recombinant polymorphic variants to immobilized M6 protein by ELISA. The FH and FH1–7 polymorphic variants showed comparable low levels of binding (1%–12% of positive readings) to the negative control wells. The (H402) variant of both FH and FH1–7, demonstrated reduced binding compared with the Tyr variant (Fig. 7) , even when a 30-fold molar excess of FH was used (data not shown), indicating that the His residue reduces the affinity of the interaction. 

Discussion

Complement FH plays a critical role in regulating AP complement activation by binding to host cell surfaces or by interacting with surface-attached CRP. This enables FH to protect host cell surfaces from complement-mediated injury while allowing activation to proceed on foreign surfaces. In the maculae of AMD patients, AP regulation is perturbed with increased AP complement activation, localized increase in CRP levels, and reduced clearance of immune complexes. 

In this study, we determined that the Y402H variant causes a reduction in binding of FH and FHL-1 to multiple ligands, including heparin, CRP, and GAS M6 protein. The Y402H variant did not, however, affect binding of either FH or FH1–7 to primary RPE cells. With the exception of GAS M6, the interaction of FH with these ligands is fundamental to the control of AP activation. FH has an intrinsically low affinity for surface-associated C3b and only effectively inactivates C3b or dissociates the C3/C5 convertases after binding to host cells. FH binds by sialic acid-containing GAGs ^{40 41} or by binding to surface-attached CRP, ^{42 43} thereby increasing the affinity of FH for C3b and thus enhancing complement regulation. 

Our modeling studies show that there is minimal structural change associated with the Y402H mutation. This finding is in accord with the nuclear magnetic resonance spectroscopic studies by Herbert et al. ⁴⁴ However, our predicted heparin-binding site, ²⁷ validated by site-directed mutagenesis, ³⁷ comprises the five charged residues R387, K388, R404, K405, K410, with only three of these (K388, R404, K405) predicted by Clark et al. ³⁹ Our docking studies show that all five basic residues, as well as Y/H402 and Q408, interact with heparin. Comparison of the electrostatic surface potential of FH SCR6–7 indicates that substitution of the bulkier Tyr residue with the smaller His residue alters the distribution of the positively charged residues responsible for binding heparin. The loss of a hydrogen bond as a result of the substitution suggests that the Y402H variant should result in a reduction in heparin binding. Heparin-binding studies, showing that FH1–7(H402) elutes at a lower salt concentration, confirmed this prediction. This finding suggests that the complement regulatory functions of FHL-1 in vivo could be compromised. The observation that FH(H402) binding to heparin is not affected is likely to be the result of the presence of additional heparin-binding domains in FH in SCR9 and SCR20. ^{16 18}  

Clark et al. ³⁹ observed that the relative binding of the (H402) and (Y402) variants was dependent on the size and degree of sulfation of the heparin examined. Because heparin is generally used as an in vitro model of polyanion/GAG binding and does not necessarily accurately represent interactions occurring in vivo, we considered it of greater relevance to assess binding of the variants to the polyanion or GAG thought to be interacting with FH in vivo. 

For this reason we analyzed the interaction of FH with RPE cells, which express high levels of heparan sulfate-containing proteoglycans on their surfaces. Initially, using the RPE cell line ARPE-19, we observed that the FH(H402) variant had greater cell surface binding than FH(Y402) (data not shown). This was in contrast to Skerka et al., ⁴⁵ who observed reduced binding of FH(H402). However, the gene and protein expression profiles of this spontaneously immortalized cell line are significantly altered from primary RPE cells, ^{46 47} thus bringing into question its reliability as a physiologically relevant cell line. Consequently, the study was repeated using primary RPE cells. This is the first study to demonstrate that there is no significant difference in the binding of the two variants present in FH and FH1–7 to primary RPE cells, indicating that AMD is unlikely to manifest as a result of impaired FH cell surface recognition through SCR7 and, hence, complement regulation at the retinal pigment epithelium–choroid interface. 

Our data indicate that the Y402H polymorphism causes a small reduction in the binding of FH and FH1–7 to CRP. Others have found a similar slight reduction in binding of the H402 variant to CRP, ^{48 49 50} suggesting that any change in FH/FHL-1 regulatory function would be subtle. This may explain why AMD only manifests later in life and takes up to 20 years to develop. CRP is an acute-phase reactant with both proinflammatory and anti-inflammatory properties. It plays an important role in downregulating AP complement activation ^{43 51} by the recruitment of FH/FHL-1, allowing FH/FHL-1 to regulate C3b deposition and C3 convertase formation by the AP and to inhibit C5a and C5b-9 formation generated by AP or CP. ^{43 51 52} A reduction in binding of FH/FHL-1 by CRP would, therefore, impede the ability of CRP to inhibit AP complement activation. Recent investigations have shown that persons carrying the Y402H polymorphism in addition to CRP haplotypes, which confer high serum CRP levels, are at greater risk for AMD than carriers with normal CRP levels or noncarriers with high CRP levels. ⁵³ This suggests that elevated CRP levels with normal proinflammatory abilities but attenuated anti-inflammatory abilities, because of the presence of the Y402H variant in FH/FHL-1, could lead to uncontrolled and chronic inflammation. The significance of complement activation and regulation in the pathogenesis of AMD has been further illustrated by the recent finding of an association between C3 functional polymorphisms and risk for AMD. ⁵⁴  

It is possible that additional polymorphisms may also contribute to AMD susceptibility. Li et al. ⁵⁵ have identified multiple polymorphisms showing stronger association with AMD risk than the FH/FHL-1 Y402H polymorphism. Although the authors suggest that Y402H is simply in linkage disequilibrium with nearby alleles, our work and that of others demonstrates that the Y402H polymorphism does affect the protein function of FH and FHl-1. Many of the polymorphisms identified are located in intronic regions; thus, it is possible that these polymorphisms influence expression of the FH and FHL-1 genes. 

The Y402H allele is relatively prevalent and encompasses a wide range of ethnic groups. ^{56 57 58 59 60 61} The high frequency of the variant indicates a phenotypic advantage may be responsible for its positive selection. One explanation is that the variant evolved in response to the capacity of a number of pathogens to evade complement attack by recruiting FH/FHL-1. Some of these pathogens include GAS, Treponema denticola, Candida albicans, and Borrelia burgdorfer. Each of these pathogens binds to FH/FHL-1 through the polyanion binding site on SCR7, ^{19 62 63 64 65 66 67} enabling it to restrict C3 convertase formation on its surface. Furthermore, GAS uses FHL-1 to facilitate intracellular invasion of epithelial cells. ²⁶ GAS is responsible for severe invasive infections, including pneumonia, bacteremia, necrotizing fasciitis, and streptococcal toxic shock syndrome. Thus, it is possible that the polymorphism was selected for during evolution to limit immune evasion by potential pathogens. Although it appears that this polymorphism also causes deleterious effects in the form of AMD and myocardial infarction, ⁶⁸ these conditions generally develop after reproductive age. Providing that the deleterious phenotypes have less impact on reproductive success than the pathogen, the polymorphism will be retained in the population. ⁶⁹ A number of heritable chronic human conditions are thought to be the direct result of natural selection for resistance against human pathogens. ^{70 71 72} One example is a polymorphism in the β-globin gene, which protects against malarial parasite infection but causes sickle cell anemia. ⁷⁰ The observation that FH and FHL-1 Y402H variant have significantly reduced binding to GAS M6 protein supports this hypothesis. 

In conclusion, we have demonstrated that the FH/FHL-1 Y402H polymorphism, associated with an increased risk for AMD, causes a reduction in binding of FH and FH1–7 to CRP and GAS M6 but does not affect binding to RPE cells. The impaired interaction with CRP may potentially affect CRP-mediated AP complement regulation, leading to the chronic inflammation associated with AMD. In contrast, the Y402H polymorphism does not affect binding of FH or FH1–7 to RPE cells, suggesting that AMD pathogenesis is not caused by impaired complement regulation on this surface. The reduced binding of FH and FH1–7 to GAS M6 provides a potential explanation for the prevalence of the Y402H genetic variant in diverse populations. Further investigations are required to determine whether the Y402H polymorphism affects binding of FH to drusen or other cell surfaces and to establish the significance of the impaired interaction between CRP and FH in the maculae of AMD patients. 

²  These authors contributed equally to the work presented here and should therefore be regarded as equivalent authors.

 Supported by the National Health and Medical Research Council of Australia and Flinders Medical Centre Foundation.

 Submitted for publication October 5, 2007; revised December 10, 2007, and January 7, 2008; accepted March 4, 2008.

 Disclosure: R.J. Ormsby, None; S. Ranganathan, None; J.C. Tong, None; K.M. Griggs, None; D.P. Dimasi, None; A.W. Hewitt, None; K.P. Burdon, None; J.E. Craig, None; J. Hoh, None; D.L. Gordon, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Comparison of the surface electrostatic potentials of (A) FH6–7 (Y402) and (B) FH6–7 (H402). Residues are shaded from red (< −10 kT) to white (0 kT) to blue (> +10 kT) Amino acids implicated in the binding of heparin are labeled; those involved in CRP or GAS M6 protein binding, or both, are underlined.

Figure 2.

 

LIGPLOT ²⁹ representation of FH6–7 variants docked with heparin. The crystal structure of heparin was docked to (A) FH6–7 (Y402) and (B) FH6–7 (H402), as described in the text, using computational energy minimization. Residues implicated in the interaction of FH6–7 with heparin are identified. Heparin bonds are purple; protein bonds are gold. Carbon atoms are black, oxygen red, nitrogen blue, and sulfur yellow. Hydrogen bonds are shown as dashed green lines, and hydrophobic contacts are depicted as semicircles with lines.

Figure 3.

 

Coomassie staining of (Y402) and (H402) variants of (A) serum-purified FH and (B) recombinant FH1–7. Proteins were separated using 7.5% or 12.5% SDS-PAGE, respectively, and stained with Coomassie blue.

Figure 4.

 

Comparison of CRP-binding properties of the (Y402) and (H402) variants of (A) FH and (B) FH1–7. Microtiter wells were coated with CRP, blocked, and then incubated with increasing concentrations of purified proteins to obtain saturation of the immobilized ligand. Binding was detected using goat antiserum against FH and HRP-conjugated antigoat IgG. Data are represented as the mean ± SE of three to four experiments.

Figure 5.

 

Comparison of heparin-binding properties of the (Y402) and (H402) variants of (A) FH and (B) FH1–7. Serum-purified FH or recombinant FH1–7 proteins were injected onto a 1-mL heparin HP column, washed with PBS, and eluted with a linear salt gradient. Solid lines represent absorbance of the eluent at 280 nm, and dashed lines represent conductivity. The lower panels represent SDS-PAGE and Western blot analysis of fractions collected during the elution gradient. Experiments were performed three times, and a representative chromatogram for each protein is shown.

Figure 6.

 

Binding of the (A) FH and (B) FH1–7 (Y402) and (H402) variants to primary RPE cells. Cells were detached and incubated with 10 μg purified protein. Binding was detected using goat antihuman FH antiserum and FITC-conjugated IgG, followed by flow cytometric analysis. Cells incubated without primary antibody were included as a negative control. Experiments were repeated at least three times, and a representative profile is shown.

Figure 7.

 

Comparison of GAS M6 protein-binding properties of the (Y402) and (H402) variants of (A) FH and (B) FH1–7. Microtiter wells were coated with purified M6 protein, blocked, and incubated with increasing concentrations of purified proteins to obtain saturation of the immobilized ligand. Binding was detected using goat antiserum against FH and HRP-conjugated antigoat IgG. Data are represented as the mean ± SE of three to four experiments.