Protein, and Associates with Lipoprotein Activity, Binds Heparin

Human Factor H-Related Protein 5 Has Cofactor Activity, Inhibits C3 Convertase Activity, Binds Heparin and C-Reactive Protein, and Associates with Lipoprotein This information is current as of October 1, 2021. Jennifer L. McRae, Thomas G. Duthy, Kim M. Griggs, Rebecca J. Ormsby, Peter J. Cowan, Brett A. Cromer, William J. McKinstry, Michael W. Parker, Brendan F. Murphy and David L. Gordon

J Immunol 2005; 174:6250-6256; ; Downloaded from doi: 10.4049/jimmunol.174.10.6250 http://www.jimmunol.org/content/174/10/6250

References This article cites 48 articles, 30 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/174/10/6250.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on October 1, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Human Factor H-Related Protein 5 Has Cofactor Activity, Inhibits C3 Convertase Activity, Binds Heparin and C-Reactive Protein, and Associates with Lipoprotein1

Jennifer L. McRae,2*†‡Thomas G. Duthy,§ Kim M. Griggs,§ Rebecca J. Ormsby,§ Peter J. Cowan,†‡ Brett A. Cromer,¶ William J. McKinstry,¶ Michael W. Parker,¶ Brendan F. Murphy,*†‡ and David L. Gordon§

Factor H-related protein 5 (FHR-5) is a recently discovered member of the factor H (fH)-related protein family. FHR proteins are structurally similar to the complement regulator fH, but their biological functions remain poorly deﬁned. FHR-5 is synthesized in the liver and consists of 9 short consensus repeats (SCRs), which display various degrees of homology to those of fH and the

other FHR proteins. FHR-5 colocalizes with complement deposits in vivo and binds C3b in vitro, suggesting a role in complement Downloaded from regulation or localization. The current study examined whether rFHR-5 exhibits properties similar to those of fH, including heparin binding, CRP binding, cofactor activity for the factor I-mediated degradation of C3b and decay acceleration of the C3 convertase. rFHR-5 bound heparin-BSA and heparin-agarose and a deﬁned series of truncations expressed in Pichia pastoris localized the heparin-binding region to within SCRs 5–7. rFHR-5 bound CRP, and this binding was also localized to SCRs 5–7. FHR-5 inhibited alternative pathway C3 convertase activity in a ﬂuid phase assay; however, dissociation of the convertase was not

observed in a solid phase assay. rFHR-5 displayed factor I-dependent cofactor activity for C3b cleavage, although it was appar- http://www.jimmunol.org/ ently less effective than fH. In addition, we demonstrate association of FHR-5 with high density lipid lipoprotein complexes in human plasma. These results demonstrate that FHR-5 shares properties of heparin and CRP binding and lipoprotein association with one or more of the other FHRs but is unique among this family of proteins in possessing independent complement-regulatory activity. The Journal of Immunology, 2005, 174: 6250–6256.

actor H-related protein 5 (FHR-5)3 is a member of the Regions of fH responsible for binding glycosaminoglycans such as human factor H (fH) protein family (1). This family of heparin have previously been mapped to SCRs 7, 20, and possibly structurally and immunologically related proteins includes 12–14 (3–6), whereas C-reactive protein (CRP)-binding sites have F by guest on October 1, 2021 fH, fH-like protein 1, and five other FHR proteins denoted FHR-1, been localized to SCRs 7 and 8–11 (7). C3b-binding sites are -2, -3, -4A, and -4B. fH is a crucial negative regulator of the within SCRs 2–3, 6–10, and/or 12–14 and 19–20 (8, 9), and com- alternative pathway (AP) of complement activation and targets the plement-regulatory activities reside within SCRs 1–4 (10–12). C3 (C3bBb) and C5 (C3bBbC3b) convertases of the AP in three Little is known about the function of FHR-1 and FHR-2, which ways. It acts as a competitor for factor B binding to C3b to form contain 5 and 4 SCR domains respectively, but both are present in the convertases, is an essential cofactor for factor I-mediated deg- lipoprotein complexes (13–15). FHR-3 and FHR-4B are highly radation of C3b to iC3b, and it accelerates the decay of Bb from related and consist of five SCRs. FHR-4A contains nine SCRs and the C3/C5 convertases (for review, see Ref. 2). fH is a 150-kDa is also closely associated with human triglyceride-rich lipoproteins serum protein consisting of 20 short consensus repeats (SCRs) (16, 17). Despite their sequence similarity, only FHR-3 possesses each composed of ϳ60 aa with 2 overlapping disulfide bonds. heparin-binding activity. This is the likely consequence of FHR-3 possessing an SCR with homology to SCR 7 of fH (3). FHR-3 also *Department of Nephrology, †Immunology Research Centre, and ‡University of Mel- binds streptococcal M protein, an interaction also mediated via bourne Department of Medicine, St. Vincent’s Hospital Melbourne, Fitzroy, Victoria, SCR 7 of fH (18). Both FHR-3 and FHR-4B bind to the C3d region Australia; §Department of Microbiology and Infectious Diseases, Flinders Medical Centre, Bedford Park, Australia; and ¶St. Vincent’s Institute, Fitzroy, Australia of C3b and have been reported to have weak factor I-mediated Received for publication January 14, 2004. Accepted for publication March 14, 2005. cofactor activity, although this activity is not independent of fH The costs of publication of this article were defrayed in part by the payment of page and remains controversial (19). charges. This article must therefore be hereby marked advertisement in accordance FHR-5 is a 65-kDa protein first identified as a component of with 18 U.S.C. Section 1734 solely to indicate this fact. pathological human glomerular preparations. Unlike the other 1 This work was supported by grants from the National Health and Medical Research Council of Australia. FHR proteins, which contain 4 or 5 SCRs, FHR-5 consists of 9 SCR domains. SCRs 1 and 2 are homologous to corresponding 2 Address correspondence and reprint requests to Corresponding author: Jennifer L. McRae, Immunology Research Centre, St. Vincent’s Hospital Melbourne, PO Box SCRs from FHR-1 and FHR-2, whereas SCRs 3–7 and 8–9 share 2900, Fitzroy, Victoria, 3065, Australia., Ph: 61 3 9288 3140. Fax: 61 3 9288 3151. significant homology with SCRs 10–14 and 19–20 of fH (1). The e-mail: [email protected] gene encoding FHR-5 has been localized to human chromosome 3 Abbreviations used in this paper: FHR-5, factor H-related protein 5; fH, factor H; FHR-1, fH-related protein 1; FHR-2, fH-related protein 2; FHR-3, fH-related protein 1q32 within the regulators of complement activation (RCA) gene 3; FHR-4, fH-related protein 4; AP, alternative pathway; SCR, short consensus repeat; cluster. It is closely linked to the other fH family genes and is CRP, C-reactive protein; RCA, regulators of complement activation; RT, room temperature; GVB, gelatin-Veronal buffer; LDL, low density lipoprotein; VLDL, very situated between FHR-2 and factor XIIIb. Each SCR of FHR-5 is low density lipoprotein; HDL, high density lipoprotein. encoded by an individual exon (20).

Although the exact role of FHR-5 in complement regulation is bled, and the sera were tested by ELISA and Western blot for activity unknown, the protein has been shown to colocalize with C3 in vivo against rFHR-5 and normal human serum. and to bind C3b in vitro in a dose-dependent manner (1). Its dep- Western blot analysis osition in vivo appears widespread in complement-containing glomerular immune deposits and extrarenal sites of complement dep- Recombinant FHR-5 SCR proteins and lipoprotein fractions were separated on 10–12% SDS-PAGE and transferred onto Hybond Cϩ nitrocel- osition (21). A prospective study of 100 renal biopsies showed the lulose (Amersham Biosciences). Native and rFHR-5 proteins were detected pattern of FHR-5 deposition to be similar to that of C3 and C5b-9 using rabbit anti-FHR-5 antiserum (1/2000 v/v) or mouse anti-FHR-5 (22). More recently, FHR-5 and all other FHR proteins have been (K2.254) mAb (Ref. 21; 1/2000 v/v) followed by HRP-conjugated goat detected at high levels in middle ear effusion fluid from patients anti-rabbit IgG (Silenus) or HRP-conjugated sheep anti-mouse IgG (Sile- with otitis media (23). nus) (1/2000 v/v), respectively. Proteins were visualized using ECL and exposure to Hyper Film ECL (Amersham Biosciences). To detect human Given the similarity of FHR-5 with fH and its association with clusterin in lipoprotein complexes, membranes were stripped by washing in complement deposits in vivo, we investigated whether recombi- 62 mM Tris, 2% SDS, 50 mM 2-ME at 42°C for 20 min, washed in PBS, nant FHR-5 (rFHR-5) exhibited functions associated with comple- and reprobed with mouse anti-human clusterin (K2.2G7) mAb (Ref. 21; ment regulation, and shared properties with other FHR proteins 1/2000 v/v) followed by HRP-conjugated sheep anti-mouse IgG as described above. such as circulation with lipoprotein complexes. Silver staining Materials and Methods 12% SDS-PAGE gels were fixed in 30% ethanol, 10% acetic acid solution Bacteria, plasmids, and proteins for 30 min, and rinsed in 20% ethanol and then water for 10 min each. Following sensitization in 0.02% sodium thiosulfate for 1 min and rinsing E. coli strain DH5␣ (Bethesda Research Laboratories, Gaithersburg, MD) Downloaded from ␣ twice in water, 0.2% silver nitrate was added for 30 min. After rinsing, harboring various SCR plasmid constructs in pPICZ A (Invitrogen, Carls- developer (30 g/L sodium carbonate, 10 mg/L sodium thiosulfate, 0.02% bad, CA) was propagated in low salt Luria-Bertani broth containing 25 ␮ formaldehyde) was added, and the reaction was stopped with a 5% Tris, g/ml Zeocin (Invitrogen), with aeration at 37°C. 2.5% acetic acid solution. FH was purified from plasma as previously described (24), with an additional immunoaffinity chromatography step. Functional activity of fH was con- Cofactor assay firmed using factor I-mediated cofactor assays as previously described (25). rFHR-5 was expressed using the baculovirus/insect cell expression sys- The cofactor activity for recombinant FHR-5 vs fH was performed as pre- tem and affinity purified as previously described (1). C3 was isolated from viously described (25). Briefly, 250 ng of C3b, 1 ␮g of factor I (Calbio- http://www.jimmunol.org/ fresh human plasma (26) and C3b was generated by proteolytic cleavage chem-Novabiochem), and varying amounts of fH or rFHR-5 ranging from ␮ with trypsin as described (27). C3i, also known as C3(H2O), was generated 100 ng to 10 g were added in 20 mM phosphate buffer, pH 6.0, to a final by five freeze/thawing cycles of purified C3. volume of 30 ␮l. Samples with or without fH/rFHR-5 or factor I were included as controls. Reactions were incubated at 37oC for 3 h and then Construction of rFHR-5 protein fragments separated under reducing conditions on SDS, 7.5% PAGE and subjected to Western blot analysis using a goat anti-C3c Ab (1/2000 v/v) (Silenus) and Coding regions comprising SCRs 1–4 and 3–7 of FHR-5 were amplified by HRP-conjugated donkey anti-goat IgG (Silenus). PCR and cloned into the yeast expression vector pPICZ␣ A (Invitrogen) downstream of the AOX1 methanol-inducible promoter. The construct de- C3 convertase activity and decay acceleration assays signs were based on the reported SCR intron/exon boundaries of FHR-5 (20). PCR, cloning, and transformation into Escherichia coli were con- The effect of rFHR-5 and fH on the alternative pathway C3 convertase was by guest on October 1, 2021 ducted according to standard protocols (28). SCRs 1–4 were amplified assessed by two methods, a solid phase ELISA-based method developed by using primers FHR-5 SCR1F(EcoRI) 5Ј-GGGAATTCGAAGGAAC Hourcade et al. (31), in which the C3 convertase is generated on a micro- ACTTTGTGATT-3Ј and FHR-5 SCR4R(KpnI) 5Ј-GCGGTACCTCA titer plate surface, and a fluid phase method, in which C3 convertase ac- AACACAAGTGGGTAAAGTTG-3Ј. SCRs 3–7 were amplified using tivity is monitored by C3a generation in solution (32). primers FHR-5 SCR3F(PmlI) 5Ј-GGCACGTGGAAAGGAGAATGT Solid phase method. Microtiter plate wells were coated with 250 CATGTTCC-3Ј and FHR-5 SCR7R(KpnI) 5Ј-GCGGTACCTCAAACA ng of purified C3b. Generation of the C3bBb(Ni2ϩ) complex (al- Ј CAGCGTGGTAATGAT-3 . All PCR-amplified SCR constructs were di- ternative pathway C3 convertase) was achieved by addition of fac- gested with the appropriate restriction enzymes, cloned into the multiple cloning site of pPICZ␣ A, and transformed into E. coli DH5␣. Sequence tor B (400 ng; Calbiochem-Novabiochem) and factor D (25 ng; analysis confirmed each construct had the correct sequence. Calbiochem-Novabiochem) in the presence of 2 mM NiCl2 and 25 mM NaCl in a final volume of 100 ␮l. Mixtures were incubated for Expression in Pichia pastoris 2 h at 37°C. rFHR-5 or fH was then added, and dissociation of the Up to 10 ␮gofSacI-digested pPICZ␣ A DNA containing the various SCR complexes was monitored at further time points during 30 min. constructs was electroporated into P. pastoris strain X33, and transfor- Intact complexes (which had not decayed) on the ELISA plates ␮ mants were selected with 100 g/ml Zeocin. Expression of recombinant were detected using goat anti-human factor B Ab (1/5000 v/v) proteins was induced for 2–4 days by twice daily addition of 2% methanol, according to the manufacturer’s instructions (Invitrogen). (DiaSorin) followed by HRP-conjugated rabbit anti-goat Ab (1/ 2000 v/v) (Dako). Sigma Fast OPD substrate (Sigma-Aldrich) was Purification of recombinant proteins added, and the OD492 was determined. SCR proteins expressed using the P. pastoris system were purified by Fluid phase method. The AP C3 convertase was assembled in immunoaffinity chromatography on CNBr-activated Sepharose (Amersham PBS by the addition of the following purified complement com- Biosciences) coupled with rabbit anti-human FHR-5 IgG, prepared by ca- ponents: 50 ng C3i; 2 ␮g of factor B; 10 ␮g C3; and 12 ␮lof0.1 prylic acid precipitation (29). Briefly, 50–100 ml of P. pastoris supernatant were passed over 5-ml affinity columns several times before extensive M MgCl2. Dilutions of rFHR-5 or fH were added followed by the washing in PBS, pH 7.4, and elution using 3 M potassium thiocyanate. addition of 200 ng of factor D (the enzyme required to activate the Eluates were immediately dialyzed against PBS and concentrated by ul- convertase) to a final volume of 125 ␮l, and the mixture was in- trafiltration, and the concentration of proteins was estimated against an cubated at 37°C for 30 min. Controls of purified complement com- FHR-5 standard curve generated by ELISA. ponents only (positive control), purified complement components Production of anti-FHR-5 polyclonal Abs without factor D (negative control), and purified complement components plus BSA (Sigma-Aldrich; irrelevant protein control) were Rabbits were immunized as previously described (30). In brief, 200 ␮gof purified rFHR-5 in 0.5 ml PBS was emulsified in 1.5 ml CFA and injected included. C3a generation was measured by ELISA using the s.c. at four to five sites. Animals were boosted twice at 3-wk intervals with Quidel C3a Enzyme Immunoassay kit (Quidel). To measure C3a at 100 ␮g of rFHR-5 in IFA. One week after the final boost, rabbits were ear time zero, a sample was taken just before the addition of factor D. 6252 FUNCTIONAL CHARACTERIZATION OF FHR-5

FIGURE 1. FHR-5 has cofactor activity. C3b (250 ng) was incubated with 1 ␮g of factor I (fI) and varying concentrations (as indicated above the lanes) of fH and rFHR-5. C3b cleavage was analyzed by Western blot. Reactions with or without factor I, fH, or FHR-5 were included as controls, and their presence is indicated by ؉. Cofactor activity is demonstrated by a loss of or reduced intensity of the ␣Ј-chain of C3b and/or appearance of the ␣Ј-chain cleavage fragments at 67 and 43 kDa. Downloaded from

Heparin-BSA ELISA A heparin-binding ELISA was performed by coating 1 ␮g of puriﬁed heparin-BSA (Sigma-Aldrich) or BSA only on Maxisorb ELISA microtiter plate wells (Nunc) overnight at 4°C in the presence of 100 mM bicarbonate http://www.jimmunol.org/ buffer, pH 9.5. The samples were blocked with 5% (w/v) skim milk in PBS for1hatroom temperature (RT) and washed three times in 50 mM phosphate buffer, pH 7.4. Samples (100 ␮l) of rFHR-5 fragments, rFHR-5 fragments, or negative control protein (BSA) were then added at 10 ␮g/ml in phosphate buffer and incubated at RT for 1 h. Samples were washed then incubated sequentially for 1 h with mouse anti-FHR-5 (1/2000 v/v) and HRP-conjugated sheep anti-mouse IgG (Silenus) at 1:2000 v/v in 1% skim milk-PBS for 1 h RT. Following washing, Sigma Fast OPD substrate (Sigma-

Aldrich) was added, and the OD492 was determined. FIGURE 2. FHR-5 inhibits C3 convertase activity in the fluid phase. A, Heparin-agarose chromatography Fluid phase assay. Fluid phase alternative pathway C3 convertase was gen- by guest on October 1, 2021 erated by addition of purified C3, C3i, factor B, and 0.1 M MgCl . Dilu- Heparin-agarose chromatography was conducted as previously described 2 tions of rFHR-5 or fH were added followed by the addition of factor D to (3) with some modifications. Approximately 7.5 ␮g of rFHR-5, rFHR-5 ␮ ϩ constructs or fH in 1 ml of 50 mM phosphate buffer, pH 7.4, were applied a final volume of 125 l. Purified components only ( control), purified to a 1-ml heparin-agarose column (Pierce), and the flowthrough was col- complement components without factor D (Ϫ control), and purified com- lected and reapplied five times. The column was then washed extensively plement components plus BSA were included as controls. Inhibition of C3 in phosphate buffer, and bound protein was eluted using a linear salt gra- convertase was determined by C3a generation after 30 min of incubation dient to 1 M NaCl (1 ml fractions). Initial, flowthrough, wash, and elution and measured by a C3a ELISA. Shown is the effect of increasing doses of samples were analyzed by SDS-PAGE and Western blots using mouse FH and rFHR-5 on C3a generation (nanograms per milliliter). Reduction in anti-FHR-5, rabbit anti-FHR-5, or sheep anti-fH. C3a generation correlates with decreased C3 convertase activity. Experi- CRP-binding ELISA ments were repeated twice in duplicate, with the means, SDs, and statistical p Ͻ 0.05 vs the BSA protein control) indicated. B, Solid ,ء) significance One microgram of purified CRP (kindly provided by Dr. Carolyn Mold, phase assay. Microtiter plate wells were coated with purified C3b. Gener- University of New Mexico) or gelatin (May and Baker) was coated on ation of the C3bBb(Ni2ϩ) complex (alternative pathway C3 convertase) Maxisorb ELISA microtiter plates) overnight at 4°C in the presence of was achieved by addition of factor B, factor D, and properdin in the pres- gelatin-Veronal buffer (GVB; 5 mM barbitone sodium, 145 mM NaCl, ence of NiCl and NaCl after 2 h incubation. Recombinant FHR-5 or fH 0.1% gelatin). The wells were washed twice with GVB containing 0.02% 2 was then added, and dissociation of the complexes was monitored at 10- v/v Tween 20 and then blocked for2hatRTwith 1% BSA in GVB. Wells were washed three times, and samples of rFHR-5 or rFHR-5 fragments min intervals during 30 min. Intact complexes (which had not been de- (100 ␮lat0–20␮g/ml) in GVB were added for1hatRT.Samples were cayed) were detected using goat anti-human factor B Ab followed by an washed and incubated sequentially for 1 h with rabbit anti-FHR-5 IgG HRP-conjugated rabbit anti-goat Ab. OD readings show the amount of (1/2000 v/v) and HRP-conjugated sheep anti-rabbit IgG (1/2000 v/v) intact AP C3 convertase remaining on the ELISA plate during a 30-min (Chemicon International). Following washing, substrate was added, and the incubation. Experiments were repeated twice in duplicate with the means OD492 was determined. Experiments were repeated twice in duplicate, and and SDs indicated. the mean and SDs were determined. Preparation and analysis of human lipoprotein complexes band in the midsection of the tube), VLDL (faint band at the top of the Low density lipoprotein (LDL), very low density lipoprotein (VLDL), high tube) and HDL and plasma proteins (the lower most fraction) were re- density lipoprotein (HDL), and plasma proteins were isolated from normal moved by needle aspiration. The HDL and plasma protein fraction was human plasma of a fasting individual. LDL and VLDL were isolated by then adjusted to 1.21 g/ml and spun in a Ti 90 rotor at 90,000 rpm for 5 h. single-spin ultracentrifugation. The plasma was adjusted to a density of HDL was aspirated from the top of the tube, and plasma proteins were 1.21 g/ml with KBr, underlaid below density buffer (0.4 mM EDTA, 0.01% collected from the bottom. Lipoproteins were run on a native 1% agarose

NaN3, 1 mM benzamidine-HCl (pH 7.4), and KBr to give a final density of gel (Paragon Electrophoresis System; Beckman Instruments) and lipid 1.019 g/ml) and centrifuged at 50,000 rpm in a VTi 50 rotor (Beckman stained per the manufacturer’s instructions. Association of FHR-5 or hu- Instruments) for2hat4°C. After centrifugation, LDL (a distinct single man clusterin (as a positive control protein) with lipoprotein complexes The Journal of Immunology 6253 was analyzed following separation of 10% whole plasma, 10% plasma mate the relative affinities of rFHR-5 and fH for heparin. A mix- protein fraction, and ϳ50 ␮g of each lipoprotein fraction (delipidated (32) ture of rFHR-5 and fH was applied to a heparin-agarose column, and nondelipidated) by SDS-PAGE under nonreducing conditions and by and the flowthrough, washes, and gradient salt elutions were col- Western blot using the anti-FHR-5 mAb or anti-human clusterin mAb as described above. lected and analyzed by Western blot (Fig. 3B). Both rFHR-5 and fH bound to the heparin agarose column, with neither protein de- Results tected in the flowthrough or wash fractions. However, rFHR-5 was FHR-5 has cofactor activity for the factor I-mediated cleavage eluted later at a higher salt concentration (300 mM), suggesting a of C3b higher affinity for heparin than fH, which eluted in 150 mM NaCl. Like fH, rFHR-5 exhibited cofactor activity for factor I-dependent Expression of rFHR-5 protein fragments in P. pastoris cleavage of C3b to iC3b, as evident by the cleavage of the C3b To localize domains responsible for particular functions, con- ␣Ј-chain into two fragments of ϳ67 kDa and ϳ43 kDa (Fig. 1). structs containing SCR 1–4 or SCR 3–7 of FHR-5 (Fig. 4A) were Cofactor activity was dose dependent; and at the highest concen- produced and expressed using P. pastoris. Affinity-purified tration (10 ␮g) of rFHR-5, a further cleavage product of ϳ40 kDa rFHR-5 protein fragments were subjected to SDS-PAGE and (probably part of C3c) was observed. Western blotting using mouse anti-FHR-5 mAb (K2.254) (Fig. FHR-5 inhibits C3 convertase activity 4B). The mAb recognized both SCR1–4 and SCR3–7 fragments, indicating that the epitope recognized by the anti-FHR-5 mAb is The effect of rFHR-5 on alternative pathway C3 convertase activ- located within SCR3–4. The corresponding silver stain shows high ity was determined in fluid and solid phase assays and compared purity of both SCR fragments. with factor H activity. In the fluid phase, rFHR-5 exhibited sig- Downloaded from nificant inhibition of C3 convertase activity as shown by the dose- The heparin-binding site of FHR-5 resides within SCR 5–7 dependent reduction in C3a generation. Potency of rFHR-5 was To determine the SCR domains of FHR-5 responsible for heparin lower than fH on a weight-for-weight basis (Fig. 2A). However, no binding, SCRs 1–4 and 3–7 were tested by heparin-agarose chro- dissociation of the convertase was detected during 30 min in the matography. SCR 1–4 did not bind heparin, with all protein being solid phase assay following addition of up to 10 ␮g of FHR-5, 2ϩ detected in the flowthrough and wash fractions (Fig. 5A). How- whereas significant decay of C3bBb(Ni ) was observed follow- http://www.jimmunol.org/ ever, SCR 3–7 bound to the heparin column and eluted from 250 ing the addition of 100 ng of fH (Fig. 2B). mM NaCl (Fig. 5B). Combined, these data indicate that there is a FHR-5 binds heparin region within SCRs 5–7 that is essential for heparin binding. The involvement of SCRs 3 and 4 in heparin binding cannot be ruled The binding of rFHR-5 to heparin was examined in two assay out, but alone they are not sufficient for heparin binding. systems. A heparin-BSA ELISA demonstrated binding of rFHR-5 to heparin-BSA but not to BSA alone (Fig. 3A). Heparin-agarose chromatography was used to confirm this observation and to esti- by guest on October 1, 2021

FIGURE 3. FHR-5 binds heparin. A, Binding of rFHR-5 to heparin as FIGURE 4. Expression of truncated forms of FHR-5. A, Schematic rep- measured by ELISA with solid phase heparin BSA or BSA control. Results resentation of native FHR-5 and the SCR derivatives constructed. Each are the mean Ϯ SD of three experiments performed in triplicate. B, Binding oval represents one SCR domain. B, Silver stain and Western blot analysis of fH and rFHR to heparin agarose. Western blot of samples from the of puriﬁed rFHR-5 fragments SCR 1–4 (lanes 1 and 3) and SCR 3–7 (lanes heparin-agarose column following NaCl gradient elution. The ﬂowthrough, 2 and 4), respectively. Western blot analysis was performed with the se- washes, and NaCl concentrations of each eluate are shown on the top of the quential addition of mouse anti-FHR-5 mAb (K2.254) and HRP-conju- lanes and a mixture of the starting material (fH and rFHR-5 preparations) gated sheep anti-mouse IgG followed by detection via ECL. SCR 1–4 can is in the left hand lane. be seen as a single protein of ϳ28kDa, and SCR 3–7 is a band of ϳ40 kDa. 6254 FUNCTIONAL CHARACTERIZATION OF FHR-5

FIGURE 5. SCRs 3–7 of FHR-5 contain a heparin-binding site. Binding of FHR-5 fragments 1–4 (A) and 3–7 (B) to heparin-agarose. Western blot of samples from the heparin agarose column following NaCl gradient elution. The NaCl concentrations of each eluate are shown on the top of the lanes and molecular mass marker proteins on the left. Downloaded from FHR-5 binds CRP An ELISA was developed to study rFHR-5 binding to CRP. rFHR-5 binding to CRP was dose dependent and saturable (Fig. 6). To determine the region of FHR-5 responsible for CRP binding, rFHR-5 fragments 1–4 and 3–7 were analyzed in the binding as-

say. The SCR 3–7 fragment also bound in a dose-dependent man- http://www.jimmunol.org/ ner, although with a lower affinity than the complete rFHR-5. The SCR 1–4 fragment, however, showed no specific binding. Thus, FHR-5 binds directly to CRP and binding is partly mediated by SCRs 3–7 with SCRs 5–7 likely to contain the major binding re- FIGURE 7. Analysis of FHR-5 in lipoprotein complexes. A, Following gion. This region also mediated heparin binding, suggesting a separation by ultracentrifugation, purity of LDL, VLDL, HDL, and plasma common or closely related binding site for both ligands. This was protein fractions was assessed by electrophoresis on a 1% agarose gel and further supported by the observation that FHR-5 binding to CRP a standard lipid stain. B, Western blot analysis of 10% whole plasma, 10% ϳ ␮ was totally inhibited by 50 IU of heparin (data not shown). plasma protein, and 50 g of HDL, LDL, and VLDL nondelipidated lipoprotein fractions. Samples were electrophoresed on 10% SDS-PAGE by guest on October 1, 2021 FHR-5 is associated with lipoprotein under nonreducing conditions, and FHR-5 was detected using the mouse anti-human FHR-5 mAb and HRP-conjugated sheep anti-mouse Ig. De- To determine whether FHR-5 is associated with lipoprotein com- lipidated lipoprotein samples gave identical results. C, The Western blot plexes, lipoprotein-containing fractions were isolated from normal was stripped and reprobed with anti-human clusterin mAb and HRP-con- human plasma and analyzed by Western blot. Fig. 7A demonstrates jugated sheep anti-mouse Ig. the purity of lipoprotein fractions following separation by ultracentrifugation. Western blot analysis using anti-human FHR-5 mAb HDL-purified preparation. FHR-5 was not detected in the LDL- or (K2.254) (Fig. 7B) showed that like human clusterin (Fig. 7C), FHR-5 VLDL-purified preparations. This suggests that some FHR-5, like was present in whole plasma, the plasma protein fraction, and the FHR-1, -2, and -4, associates with lipoprotein particles in normal plasma, although the majority probably exists as a free protein.

Discussion There are currently only limited data on the functions of the human FHR proteins. Given the structural correlation between these SCR- containing proteins, overlaps in function are expected. To date, similarities have been seen in the ability of some FHR molecules to bind C3b, heparin and streptococcal M protein (1, 18, 19), and to associate with lipoprotein complexes (16, 33). However, no fH- independent complement-regulatory activity has yet been reported for the FHR proteins. FHR-5 is unique among the FHR proteins in that it colocalizes in vivo with complexes resulting from complement activation (22). We therefore investigated the possibility that FHR-5 possesses similar or overlapping functions to fH. The cofactor experiments showed that FHR-5, like fH, pos- FIGURE 6. Binding of FHR-5 and FHR-5 fragments to CRP. Increas- sesses fI-dependent cofactor activity cleaving C3b and generating ing amounts of rFHR-5 and the rFHR-5 SCR fragments were incubated ␣Ј with CRP or gelatin control (1 ␮g) bound to microtiter plates. Binding of C3 chain fragments of 67 and 43 kDa. In addition, FHR-5 sig- rFHR-5 and rFHR-5 fragments were detected by the sequential addition of nificantly inhibited the activity of the fluid phase C3 convertase; rabbit anti-FHR-5 IgG, HRP-conjugated sheep anti-rabbit IgG, and sub- however, FHR-5 displayed no decay acceleration activity in solid 2ϩ strate before reading the OD492. Experiments were repeated twice in du- phase experiments using C3bBb(Ni ) convertase complexes con- plicate, with the means and SDs indicated. structed on the surface of microtiter plates. This was unlikely to be The Journal of Immunology 6255 due to differences in assay sensitivity because fH caused similar complement-mediated damage to self-tissues by limiting the in- decay in both systems. Whether the inhibitory effect of FHR-5 on flammatory response (7, 41, 42). We demonstrated that FHR-5 C3 convertase activity in this fluid phase system is directly due to contains at least one CRP binding site within the SCR 5–7 region, decay acceleration of the C3bBb complex requires clarification, which does not correspond to any of the known CRP-binding sites although the similarity in response to that of fH would suggest of fH. Nevertheless, as proposed for fH, CRP may recruit FHR-5 dissociation of the C3 convertase is a likely mechanism. to sites of tissue damage, where it could have a role in regulating It is conceivable that the C3bBb complexes in the fluid and solid complement activation. Recent immunofluorescence analysis of phases are conformationally distinct and that fH with its multiple lupus nephritis biopsies demonstrated consistent glomerular dep- C3b-binding sites is capable of disrupting either form, whereas osition of CRP in the mesangial and peripheral capillary wall (44), FHR-5 is able only to dissociate the fluid phase complex. Alter- similar to the distribution of FHR-5 that we observed in all lupus natively, the solid phase in vitro system may not accurately reflect nephritis cases examined (22). In contrast, CRP was rarely de- the conditions in vivo where the presence of polyanions on the cell tected in immune deposits of nonlupus glomerulonephropathies, surface may influence binding of fH family regulators (34–37). which invariably contained FHR-5. This suggests that although Although the immunohistological data (22) and the results de- CRP may play a role in recruiting FHR-5 to sites of complement scribed here support a complement regulatory role for FHR-5, se- activation and tissue injury in some circumstances, other mecha- quence analysis of its SCRs does not predict the domains involved. nisms may also be involved. For example, in vitro and in vivo To date, all RCA proteins with complement regulatory ability con- studies by Ren et al. (45) have demonstrated that rat FHR mRNA tain SCRs homologous to the four N-terminal SCRs of fH or C4bp (analogous to human FHR-5) is up-regulated in cultured glomer- ␣-chains (38). Our finding that FHR-5 exhibits cofactor and possible ular epithelial cells subjected to complement attack and in models decay accelerating activity despite lacking such domains suggests that of membranous nephropathy. Downloaded from there is a novel complement-regulatory site in FHR-5. Localization of We have established that FHR-5 is present at 3–6 ␮g/ml in these domains requires analysis of additional FHR-5 constructs. normal human serum (J. L. McRae, P. J. Cowan, and B. F. The ability of fH to distinguish between foreign and host cells Murphy, unpublished data), which is relatively low compared with and regulate alternative pathway activation on the appropriate cell fH (46). However, the potential for recruitment and up-regulation surface is influenced by cell surface molecules. Polyanions such as of FHR-5 may allow it to regulate complement at sites of comple- sialic acid and heparin are abundant on host cell membranes and ment-mediated injury. Whether FHR-5 serum levels vary in response http://www.jimmunol.org/ enhance the binding (34–37) and subsequent activity of fH (39, to disease is also of interest and is currently under investigation. 40). As with fH and factor H-like protein 1, the ability of FHR-5 Although no direct link between fH and lipoprotein has been to bind polyanions may assist in positioning it at the cell surface described, FHR proteins -1, -2, and -4A are present in lipoprotein where it can use its complement-regulatory functions. We local- complexes in human plasma. FHR-1 and FHR-2 are components ized a heparin-binding site in FHR-5 to SCRs 5–7, which show of lipoprotein particles, which also contain apoA-I, LPS-binding highest homology to SCRs 12–14 of fH. Of the 9 SCRs in FHR-5, protein, and fibrinogen, and may be involved in facilitating the SCR 2 shows the closest homology (33%) to SCR 7 of fH, but the adhesive response of neutrophils to lipopolysaccharides (33, 47). FHR-5 SCR 1–4 construct was unable to bind to heparin. The pos- FHR-4A associates with the triglyceride-rich lipoproteins VLDL by guest on October 1, 2021 sibility of a heparin-binding site in SCR 9 of FHR-5, which shows and chylomicrons and to a lesser extent HDL and LDL (1, 16, 17). 42% homology with SCR 20 of fH, was not investigated in this study. Consequently, the use of lipoproteins as transport vehicles or a role FH plays an essential role in regulating complement activation of FHR proteins in lipid transport has been suggested (16). De- at sites of tissue damage. This function appears to be mediated in tection of FHR-5 in the HDL lipoprotein fraction further demon- part by specific binding of fH to the acute phase protein CRP (7). strates a relationship between FHR family members and lipopro- The main biological function of CRP is to recognize pathogens and teins. Furthermore, the complement regulators, C4-binding protein damaged host cells and to activate the classical pathway of com- (48), CD59 (49), and clusterin (50) also interact with lipoprotein in- plement, which results in their uptake by phagocytosis (41, 42). A dicating a possible functional interaction between lipoproteins and the rapid increase in CRP serum levels (up to 1000-fold) is observed complement system. With lipoprotein association and complement- within 6 h after exposure to such stimuli (43). CRP can recruit fH, regulatory capability, FHR-5 lends further support to this theory. which in turn is capable of inhibiting the formation of the AP Fig. 8 summarizes the SCRs of fH and FHR-5 known to be convertase and the C5 convertases. Thus, CRP may act to prevent involved in protein binding and complement-regulatory activity.

FIGURE 8. SCRs of fH and FHR-5 involved in protein binding and complement activity. SCRs of the individual proteins are represented by ovals. Re- lated SCRs are shaded and vertically aligned, and gaps are indicated by a dot- ted line. Proposed functional and binding domains are represented by hori- ,ء .zontal bars and labeled accordingly C3b binding, decay accelerating (DA), and cofactor activity have been demonstrated for FHR-5 but have not yet been localized to particular SCRs. 6256 FUNCTIONAL CHARACTERIZATION OF FHR-5

FHR-5 is unique among the FHR proteins in its ability to bind C3b, 22. Murphy, B., T. Georgiou, D. Machet, P. Hill, and J. McRae. 2002. Factor H- heparin and CRP, and its possession of cofactor and likely fluid related protein-5: a novel component of human glomerular immune deposits. Am. J. Kidney Dis. 39: 24–27. phase decay-accelerating activity. These results add to our previ- 23. Narkio-Makela, M., J. Hellwage, O. Tahkokallio, and S. Meri. 2001. Comple- ous finding of a direct association between FHR-5 and complement-regulator factor H and related proteins in otitis media with effusion. Clin. ment activation in vivo (22), support a role for FHR-5 in comple- Immunol. 100: 118–126. 24. Hammer, C. H., G. H. Wirtz, L. Renfer, H. D. Gresham, and B. F. Tack. 1981. ment regulation in vivo, and define the first clear complement- Large scale isolation of functionally active components of the human comple- regulatory function among FHR proteins. ment system. J. Biol. Chem. 256: 3995–4006. 25. Giannakis, E., T. S. Jokiranta, R. J. Ormsby, T. G. Duthy, D. A. Male, D. Christiansen, V. A. Fischetti, C. Bagley, B. E. Loveland, and D. L. Gordon. Acknowledgments 2002. Identification of the streptococcal M protein binding site on membrane We thank Dr. Eleni Giannakis for the valuable discussions held regarding cofactor protein (CD46). J. Immunol. 168: 4585–4592. Pichia expression, Dr. Carolyn Mold for her gift of CRP used in these exper- 26. Tack, B. F., J. Janatova, M. L. Thomas, R. A. Harrison, and C. H. Hammer. 1981. iments, and Dr. Steve Christov for his assistance in lipoprotein purification. Third, fourth and fifth components of human complement: isolation and bio- chemical properties. In Methods in Enzymology, Vol. 80. L. Lorand, ed. Aca- demic Press, New York, pp 64–74. Disclosures 27. Avery, V. M., and D. L. Gordon. 1993. Characterization of factor H binding to The authors have no financial conflict of interest. human polymorphonuclear leukocytes. J. Immunol. 151: 5545–5553. 28. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: a Lab- oratory Manual, Vol. 1. Cold Spring Harbor Laboratory Press, New York, pp. References 7.37–7.52. 1. McRae, J. L., P. J. Cowan, D. A. Power, K. I. Mitchelhill, B. E. Kemp, 29. Steinbuch, M., and R. Audran. 1969. The isolation of IgG from mammalian sera B. P. Morgan, and B. F. Murphy. 2001. Human factor H-related protein 5 (FHR- with the aid of caprylic acid. Arch. Biochem. Biophys. 134: 279–284. 5): a new complement associated protein. J. Biol. Chem. 276: 6747–6754. 30. Goding, J. W. 1996. Generation of conventional antibodies. In Monoclonal An-

2. Zipfel, P. F., T. S. Jokiranta, J. Hellwage, V. Koistinen, and S. Meri. 1999. The tibodies: Principles and Practice. Academic Press, San Diego, pp. 465–479. Downloaded from factor H protein family. Immunopharmacology 42: 53–60. 31. Hourcade, D. E., L. M. Mitchell, and M. E. Medof. 1999. Decay acceleration of the 3. Blackmore, T. K., T. A. Sadlon, H. M. Ward, D. M. Lublin, and D. L. Gordon. complement alternative pathway C3 convertase. Immunopharmacology 42: 167. 1996. Identification of a heparin binding domain in the seventh short consensus 32. Harris, C. L. 2000. Functional assays for complement regulators. In Complement repeat of complement factor H. J. Immunol. 157: 5422–5427. Methods and Protocols. B. P. Morgan, ed. Humana Press, Totowa, NJ, pp. 83– 4. Pangburn, M. K., M. A. Atkinson, and S. Meri. 1991. Localization of the heparin- 101. binding site on complement factor H. J. Biol. Chem. 266: 16847–16853. 5. Blackmore, T. K., J. Hellwage, T. A. Sadlon, N. Higgs, P. F. Zipfel, H. M. Ward, 33. Park, C. T., and S. D. Wright. 1996. Plasma lipopolysaccharide-binding protein and D. L. Gordon. 1998. Identification of the second heparin-binding domain in is found associated with a particle containing apolipoprotein A-I, phospholipid, human complement factor H. J. Immunol. 160: 3342–3348. and factor H-related proteins. J. Biol. Chem. 271: 18054–18060. http://www.jimmunol.org/ 6. Prodinger, W. M., J. Hellwage, M. Spruth, M. P. Dierich, and P. F. Zipfel. 1998. 34. Pangburn, M. K., and H. J. Muller-Eberhard. 1978. Complement C3 convertase: ␤ The C-terminus of factor H: monoclonal antibodies inhibit heparin binding and cell surface restriction of 1H control and generation of restriction on neuramin- identify epitopes common to factor H and factor H-related proteins. Biochem. J. idase-treated cells. Proc. Natl. Acad. Sci. USA 75: 2416–2420. 331: 41–47. 35. Pangburn, M. K., D. C. Morrison, R. D. Schreiber, and H. J. Muller-Eberhard. 7. Jarva, H., T. S. Jokiranta, J. Hellwage, P. F. Zipfel, and S. Meri. 1999. Regulation 1980. Activation of the alternative complement pathway: recognition of surface of complement activation by C-reactive protein: targeting the complement inhib- structures on activators by bound C3b. J. Immunol. 124: 977–982. itory activity of factor H by an interaction with short consensus repeat domains 36. Kazatchkine, M. D., D. T. Fearon, J. E. Silbert, and K. F. Austen. 1979. Surface- 7 and 8–11. J. Immunol. 163: 3957–3962. associated heparin inhibits zymosan-induced activation of the human alternative 8. Jokiranta, T. S., J. Hellwage, V. Koistinen, P. F. Zipfel, and S. Meri. 2000. Each complement pathway by augmenting the regulatory action of the control proteins of the three binding sites on complement factor H interacts with a distinct site on on particle-bound C3b. J. Exp. Med. 150: 1202–1215. C3b. J. Biol. Chem. 275: 27657–27662. 37. Fearon, D. T. 1978. Regulation by membrane sialic acid of ␤1H-dependent de- 9. Sharma, A. K., and M. K. Pangburn. 1996. Identification of three physically and cay-dissociation of amplification C3 convertase of the alternative complement by guest on October 1, 2021 functionally distinct binding sites for C3b in human complement factor H by pathway. Proc. Natl. Acad. Sci. USA 75: 1971–1975. deletion mutagenesis. Proc. Natl. Acad. Sci. USA 93: 10996–11001. 38. Krushkal, J., O. Bat, and I. Gigli. 2000. Evolutionary relationships among pro- 10. Gordon, D. L., R. M. Kaufman, T. K. Blackmore, J. Kwong, and D. M. Lublin. teins encoded by the regulator of complement activation gene cluster. Mol. Biol. 1995. Identification of complement regulatory domains in human factor H. J. Im- Evol. 17: 1718–1730. munol. 155: 348–356. 39. Horstmann, R. D., M. K. Pangburn, and H. J. Muller-Eberhard. 1985. Species 11. Kuhn, S., C. Skerka, and P. F. Zipfel. 1995. Mapping of the complement regu- specificity of recognition by the alternative pathway of complement. J. Immunol. latory domains in the human factor H-like protein 1 and in factor H1. J. Immunol. 134: 1101–1104. 155: 5663–5670. 40. Kazatchkine, M. D., D. T. Fearon, and K. F. Austen. 1979. Human alternative 12. Kuhn, S., and P. F. Zipfel. 1996. Mapping of the domains required for decay complement pathway: membrane-associated sialic acid regulates the competition acceleration activity of the human factor H-like protein 1 and factor H. Eur. between B and ␤1 H for cell-bound C3b. J. Immunol. 122: 75–81. J. Immunol. 26: 2383–2387. 41. Volanakis, J. E. 2001. Human C-reactive protein: expression, structure, and func- 13. Estaller, C., V. Koistinen, W. Schwaeble, M. P. Dierich, and E. H. Weiss. 1991. tion. Mol. Immunol. 38: 189–197. Cloning of the 1.4-kb mRNA species of human complement factor H reveals a 42. Mold, C., H. Gewurz, and T. W. Du Clos. 1999. Regulation of complement novel member of the short consensus repeat family related to the carboxy termi- activation by C-reactive protein. Immunopharmacology 42: 23–30. J. Immunol. nal of the classical 150-kDa molecule. 146: 3190–3196. 43. Pepys, M. B., and M. L. Baltz. 1983. Acute phase proteins with special reference 14. Skerka, C., R. D. Horstmann, and P. F. Zipfel. 1991. Molecular cloning of a to C-reactive protein and related proteins (pentaxins) and serum amyloid A pro- J. Biol. Chem. human serum protein structurally related to complement factor H. tein. Adv. Immunol. 34: 141–212. 266: 12015–12020. 44. Zuniga, R., G. S. Markowitz, T. Arkachaisri, E. A. Imperatore, V. D. D’Agati, 15. Skerka, C., C. Timmann, R. D. Horstmann, and P. F. Zipfel. 1992. Two additional and J. E. Salmon. 2003. Identification of IgG subclasses and C-reactive protein in human serum proteins structurally related to complement factor H: evidence for lupus nephritis: the relationship between the composition of immune deposits and a family of factor H-related genes. J. Immunol. 148: 3313–3318. FC␥ receptor type IIA alleles. Arthritis Rheum. 48: 460–470. 16. Skerka, C., J. Hellwage, W. Weber, A. Tilkorn, F. Buck, T. Marti, E. Kampen, U. Beisiegel, and P. F. Zipfel. 1997. The human factor H-related protein 4 (FHR- 45. Ren, G., M. Doshi, B. K. Hack, J. J. Alexander, and R. J. Quigg. 2002. Isolation 4): a novel short consensus repeat-containing protein is associated with human and characterization of a novel rat factor H-related protein that is up-regulated in triglyceride-rich lipoproteins. J. Biol. Chem. 272: 5627–5634. glomeruli under complement attack. J. Biol. Chem. 277: 48351–48358. 17. Jozsi, M., H. Richter, I. Loschmann, C. Skerka, F. Buck, U. Beisiegel, A. Erdei, 46. Whaley, K., and S. Ruddy. 1976. Modulation of the alternative complement ␤ and P. F. Zipfel. 2005. FHR-4A: a new factor H-related protein is encoded by the pathways by 1H globulin. J. Exp. Med. 144: 1147–1163. human FHR-4 gene. Eur. J. Hum. Genet. 13: 321–329. 47. Park, C. T., and S. D. Wright. 2000. Fibrinogen is a component of a novel 18. Blackmore, T. K., V. A. Fischetti, T. A. Sadlon, H. M. Ward, and D. L. Gordon. lipoprotein particle: factor H-related protein (FHRP)-associated lipoprotein par- 1998. M protein of the group A streptococcus binds to the seventh short consen- ticle (FALP). Blood 95: 198–204. sus repeat of human complement factor H. Infect. Immun. 66: 1427–1428. 48. Xu, N., B. Dahlback, A. K. Ohlin, and A. Nilsson. 1998. Association of vitamin 19. Hellwage, J., T. S. Jokiranta, V. Koistinen, O. Vaarala, S. Meri, and P. F. Zipfel. K-dependent coagulation proteins and C4b binding protein with triglyceride-rich 1999. Functional properties of complement factor H-related proteins FHR-3 and lipoproteins of human plasma. Arterioscler. Thromb. Vasc. Biol. 18: 33–40. FHR-4: binding to the C3d region of C3b and differential regulation by heparin. 49. Vakeva, A., M. Jauhiainen, C. Ehnholm, T. Lehto, and S. Meri. 1994. High- FEBS Lett. 462: 345–352. density lipoproteins can act as carriers of glycophosphoinositol lipid-anchored 20. McRae, J. L., B. E. Murphy, H. J. Eyre, G. R. Sutherland, J. Crawford, and CD59 in human plasma. Immunology 82: 28–33. P. J. Cowan. 2002. Location and structure of the human FHR-5 gene. Genetica 114: 50. Jenne, D. E., B. Lowin, M. C. Peitsch, A. Bottcher, G. Schmitz, and J. Tschopp. 157–161. 1991. Clusterin (complement lysis inhibitor) forms a high density lipoprotein 21. Murphy, B. F., and A. J. d’Apice. 1988. Identification of the components of glomer- complex with apolipoprotein A-I in human plasma. J. Biol. Chem. 266: ular immune deposits using monoclonal antibodies. Pathology 20: 130–136. 11030–11036.