of November 18, 2018.This information is current as

DecayPathway C3 Convertase to Mediate EfficientComponents of the Complement Alternative Decay-Accelerating Factor Must Bind Both

Paul MorganClaire L. Harris, David M. Pettigrew, Susan M. Lea and B.

http://www.jimmunol.org/content/178/1/352doi: 10.4049/jimmunol.178.1.352

2007; 178:352-359; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/178/1/352.full#ref-list-1

, 17 of which you can access for free at: cites 29 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2007 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on N

ovember 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Decay-Accelerating Factor Must Bind Both Components of theComplement Alternative Pathway C3 Convertase to MediateEfficient Decay1

Claire L. Harris,2* David M. Pettigrew,† Susan M. Lea,† and B. Paul Morgan*

Decay-accelerating factor (DAF; CD55) inhibits the complement (C) cascade by dissociating the multimolecular C3 conver-tase enzymes central to amplification. We have previously demonstrated using surface plasmon resonance (Biacore Inter-national) that DAF mediates decay of the alternative pathway C3 convertase, C3bBb, but not of the inactive proenzyme,C3bB, and have shown that the major site of interaction is with the larger cleavage subunit factor B (Bb) subunit. In thisstudy, we dissect these interactions and demonstrate that the second short consensus repeat (SCR) domain of DAF (SCR2)interacts only with Bb, whereas SCR4 interacts with C3b. Despite earlier studies that found SCR3 to be critical to DAFactivity, we find that SCR3 does not directly interact with either subunit. Furthermore, we demonstrate that properdin, apositive regulator of the alternative pathway, does not directly interact with DAF. Extending from studies of binding todecay-accelerating activity, we show that truncated forms of DAF consisting of SCRs 2 and 3 bind the convertase stably viaSCR2-Bb interactions but have little functional activity. In contrast, an SCR34 construct mediates decay acceleration,presumably due to SCR4-C3b interactions demonstrated above, because SCR3 alone has no binding or functional effect. Wepropose that DAF interacts with C3bBb through major sites in SCR2 and SCR4. Binding to Bb via SCR2 increases avidityof binding, concentrating DAF on the active convertase, whereas more transient interactions through SCR4 with C3b directlymediate decay acceleration. These data provide new insights into the mechanisms involved in C3 convertase decay byDAF. The Journal of Immunology, 2007, 178: 352–359.

T he complement system is an essential component of in-nate immunity. Activation by foreign or Ab-targeted sur-faces leads to rapid opsonisation and subsequent phago-

cytosis, inflammation and leukocyte recruitment, and in somecases cell death through formation of a lytic pore. C is activated viaseveral different mechanisms. The alternative pathway (AP)3 of Cactivation turns over continuously at a low level in plasma, prim-ing the system to react immediately when confronted by a patho-gen (1). The AP also functions as an amplification loop for twoother activation pathways, the classical pathway (CP) and thelectin pathway. The CP is activated through various mecha-nisms, the most predominant of which is subsequent to Ab bind-ing to Ag (2). The lectin pathway is activated following bindingof mannan-binding lectin to sugars on bacterial cell walls (3).

The central enzyme of the AP amplification loop is the C3 con-vertase, C3bBb. The proenzyme, C3bB, is formed followingMg2�-dependent binding of factor B (fB) to the activated fragmentof C3, C3b. Bound fB is cleaved by the plasma serine proteasefactor D (fD), resulting in release of Ba (smaller cleavage subunitof fB) and activation of the serine protease domain within Bb(larger cleavage subunit of fB) to form the active convertase (4).The convertase then cleaves many molecules of C3 to C3b, each ofwhich can form a convertase, thus amplifying the pathway. Theenzyme is naturally labile with a half-life of �90 s; association ofthe complex with properdin extends the half-life to �30 min (5).Following cleavage of C3 to nascent C3b, nucleophilic attack onthe internal thioester bond in C3b results in covalent binding to theactivating surface (6). Foreign surfaces are usually devoid of mem-brane regulatory proteins that inactivate the convertase, this lack ofregulation thus leads to amplification. Host cells, however, areprotected from the damaging effects of C by cell surface-associatedconvertase regulatory proteins. In the human, these are decay-ac-celerating factor (DAF; CD55) and membrane cofactor protein(MCP; CD46). These proteins cooperate to prevent opsonisationand damage to self cells at a site of C activation by inactivating theconvertase enzymes (7–9). DAF accelerates decay of the con-vertase, releasing Bb (or C2a in the CP), whereas MCP binds toC3b (or C4b) on the surface and acts as a cofactor for its irre-versible proteolytic inactivation by the plasma protease factor I.DAF and MCP are encoded by genes located in the “regulatorsof C activation” gene cluster and are assembled from similarfunctional domains, termed short consensus repeats (SCRs),found in many C-regulatory proteins that bind C3b and C4b(10). DAF is composed from the amino terminus of four SCRdomains followed by a short “stalk” that is heavily glycosylatedand a GPI anchor (11).

We have previously used surface plasmon resonance (SPR) toinvestigate formation and regulation of the AP C3 convertase and

*Department of Medical Biochemistry and Immunology, School of Medicine, CardiffUniversity, Cardiff, United Kingdom; and †Laboratory of Molecular Biophysics, De-partment of Biochemistry, University of Oxford, Oxford, United Kingdom

Received for publication August 25, 2006. Accepted for publication October10, 2006.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by the Wellcome Trust Grant 068823.2 Address correspondence and reprint requests to Dr. Claire L. Harris, ComplementBiology Group, Department of Medical Biochemistry and Immunology, School ofMedicine, Cardiff University, Henry Wellcome Building, Heath Park, Cardiff CF144XN, U.K. E-mail address: HarrisCL@cardiff.ac.uk3 Abbreviations used in this paper: AP, alternative pathway; CP, classical pathway;fB, factor B; fD, factor D; Ba, smaller cleavage subunit of fB; Bb, larger cleavagesubunit of fB; DAF, decay-accelerating factor; MCP, membrane cofactor protein;SCR, short consensus repeat; SPR, surface plasmon resonance; vWFA, von Wille-brand factor type A; CR1, C receptor 1; E, erythrocyte; RU, resonance unit; Rmax,maximum binding capacity of ligand.

Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$2.00

The Journal of Immunology

www.jimmunol.org

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

have demonstrated that kinetics of formation of the proenzyme andthe activated convertase were quite different; activation of theproenzyme produced a more stable enzyme with a slower rate ofdecay (12). The two complexes also differed in their susceptibilityto decay using EDTA and soluble DAF. The proenzyme was rap-idly broken apart by Mg2� chelation with EDTA whereas the ac-tivated enzyme was not, presumably reflecting differences in pro-tein conformation and coordination around the Mg2� ion. Only theactivated enzyme was subject to decay acceleration mediated bysoluble DAF. Using SPR, we analyzed the way in which DAFinteracted with individual components of the AP convertase en-zyme (12). DAF bound weakly to C3b (KD � 10�5 M), veryweakly to fB (KD � 4 � 10�5 M), but much more tightly to thecleaved fragment Bb (KD � 10�6 M). DAF bound Bb via the vonWillebrand factor type A (vWFA) domain rather than the serineprotease domain. Functional studies with recombinant C2 and fBmutated within this domain indicated that DAF mediated its reg-ulatory action via interactions with vWFA, and our kinetic analy-ses demonstrated that the Bb fragment and the vWF domain bindDAF with the same kinetic profiles and affinities (12–14). Thespecificity of DAF for the active convertase may be a consequenceof its higher affinity for Bb compared with fB, and suggests that theregulatory function of DAF on the cell membrane is focused en-tirely on inhibition of activated enzymes rather than on binding toinactive proenzymes.

The mechanism of decay acceleration is still incompletely un-derstood. It is clear from work described above that DAF interactswith both subunits of the active C3 convertase and that spacing ofDAF above the membrane is crucial to function (12, 15, 16). Struc-ture/function studies of human DAF using SCR-deletion mutantsindicated that the first SCR was not required for regulatory func-tion, whereas SCRs 2 and 3 were required for inhibition of eitherthe CP or AP (15, 17). Deletion of SCR4 had a critical effect onregulation of the AP. In support of a key role for SCRs 2 and 3,Ab-blocking studies demonstrated that Abs directed against theseSCRs blocked DAF function, with Abs to SCR3 having a partic-ularly profound effect (15, 18). To further define which domains ofDAF are involved in the interaction with components of the con-vertase, we have generated soluble recombinant mutants of DAFlacking specific SCR domains. These were used for SPR analysisto assess the affinity of interactions between specific domains ofDAF and individual components of the AP convertase. Functionalassays, including real-time analyses, were used to determine whichdomains of DAF were required for decay acceleration of the intactenzyme. These data shed new light on the mechanisms involved indecay acceleration of the AP C3 convertase.

Materials and MethodsMaterials

C3, properdin, and fB were prepared by classical column chromatographyusing established methods (19). fD was purchased from Complement Tech-nology. C3b and Bb were prepared from these reagents by activation of theAP as described previously (12). Soluble recombinant C receptor 1 (CR1)was a gift from T Cell Sciences (Needham, MA).

Expression and purification of soluble DAF constructs

Recombinant human DAF comprising the four SCRs (DAF1234) was iso-lated and refolded from Escherichia coli as described previously (20, 21).The structure and function of DAF1234 has been previously studied, andthe purified, refolded, protein is known to consist of four correctly foldedSCR domains and demonstrates the full range of C-regulating and patho-gen-binding activities associated with DAF purified from human erythro-cytes (E). Similarly, N-terminally hexahistidine-tagged recombinant SCR-deletion mutants of DAF consisting of SCR domains 2–3 (DAF23), 3–4(DAF34), and 3 (DAF3) were expressed in E. coli and purified from in-clusion bodies using an identical protocol to that of DAF1234. These mu-

tants have been described previously and have been shown to be functionalwith respect to echovirus binding (22).

Hemolysis assays

To assess regulation in the CP, sheep E (TCS Microbiology) were sensi-tized by incubating 1 volume of 4% E (v/v) with 1 volume of 1/250 rabbitanti-sheep E (Amboceptor; Behring Diagnostics) for 30 min at 37°C. Sen-sitized cells were washed three times in complement fixation diluent (Ox-oid) supplemented with 0.1% gelatin) and resuspended to 2%. To assessregulation in the AP, rabbit E were washed into AP buffer (5 mM sodiumbarbitone (pH 7.4), 150 mM NaCl, 7 mM MgCl2, 10 mM EGTA, and 0.1%gelatin) and resuspended to 2%. For both assays, a 50 �l aliquot wasincubated with 50 �l of normal human serum and 50 �l of a dilution of testprotein or a control protein. All dilutions were made in complement fixa-tion diluent/gelatin or AP buffer, and serum was previously titered to yield70% lysis in the absence of inhibitor. Cells were incubated at 37°C for 30min, centrifuged to pellet cells, and 50 �l of supernatant was removed formeasurement of absorbance at 415 nm (hemoglobin release). Control in-cubations included cells incubated in buffer only (0%) or in 0.01% TritonX-100 (100%). Percentage of lysis was calculated as follows: percentage oflysis � 100 � (A415 test sample � A415 0% control)/(A415 100% control �A415 0% control). A noninhibitory protein was used as a control, andinhibitory activity of the test protein was calculated: percentage of inhibi-tion � 100 � (percentage of lysis with negative control protein � per-centage of lysis test sample)/percentage of lysis with control protein.

Surface plasmon resonance

All analyses were conducted on a Biacore 3000 (Biacore International)other than that in Fig. 4, which was conducted on a Biacore T100. C3b wascoupled to the chip surface using its native thioester bond as describedpreviously (12, 23). In brief, a very small nidus of C3b (�50 resonance unit(RU)) was coupled to the chip surface using amine coupling. Subsequently,a mix of fB and fD was flowed over the surface to form the convertasefollowed by C3. Cleavage of C3 to C3b close to the chip surface enabledcoupling through the native thioester group. Affinity measurements weretaken on surfaces with different amounts of C3b immobilized. It was notpossible to immobilize Bb on the surface and retain binding activity; there-fore, all Bb interactions were studied with soluble recombinant DAF mu-tants on the chip surface, these were immobilized using amine couplingaccording to the manufacturer’s instructions (amine coupling kit; BiacoreInternational). Properdin was also immobilized on to the surface of a CM5chip using amine coupling as described previously (24). For all kineticanalyses a CM5 chip (carboxymethylated dextran surface; Biacore Inter-national) was used and data collected at 25°C, the flow rate was maintainedat 30 �l/min, and data from a reference cell was subtracted to control forbulk refractive index changes. The Rmax (maximum binding capacity ofligand (covalently coupled to the sensor chip surface) was kept low and theflow rate high to eliminate mass transfer. Proteins were polished by sizeexclusion chromatography to ensure removal of any aggregates before ki-netic analysis. Samples were injected using the KINJECT command toensure accurate association kinetics. Interactions were analyzed in HEPES-buffered saline (HBS; 10 mM HEPES (pH 7.4), 150 mM NaCl, and 0.005%surfactant P20) supplemented with 1 mM MgCl2. Harsh regeneration con-ditions were avoided in all interactions described in this study. Becausebinding was of relatively low affinity, sufficient time was allowed for totaldecay and return to original baseline levels. Data were evaluated usingBiaevaluation software (Biacore International). Concentration of analytes(moiety flowed across the sensor chip surface) was assessed using absor-bance at A280, and molarities were calculated using the following extinc-tion coefficients in milligrams per milliliter: DAF1–4, 1.30; DAF23, 1.16;DAF34, 1.29; DAF3, 0.88; C3b, 1.03; Bb, 1.29.

Decay assays

Surface plasmon resonance. The ability of DAF mutants to decay theAP C3 convertase was assessed in real time using SPR. A total of 3800RU of C3b was deposited on the surface of a CM5 chip as describedabove. Active convertase was formed by flowing 20 �l of a mix of fB(3.8 �M) and fD (0.07 �M) over the C3b surface at 20 �l/min. DAFmutants (20 �l) at specified concentrations were then flowed across theconvertase, and decay was visualized in real time. To regenerate backto baseline C3b, 10 �l of DAF1234 (0.5 �M) was injected over the chipsurface.Plate assay. The ability of DAF mutants to decay the AP C3 convertasewas also assessed using an ELISA-based protocol essentially as describedpreviously (13, 25). All steps were for 1 h at 37°C unless otherwise statedin 2.5 mM sodium barbitone (pH 7.4), 71 mM NaCl, 1 mM NiSO4, and0.1% Tween 20. ELISA plates (Medisorb plates; Nunc) were coated with

353The Journal of Immunology

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

C3b at 3 �g/ml and blocked with buffer containing 1% BSA. Convertasewas formed by incubating plates for 30 min with 0.5 �g/ml fB and 50ng/ml fD. Plates were washed and incubated with DAF, soluble CR1, orbuffer only for 20 min at room temperature. Plates were washed, and re-sidual fB in the convertase was detected using polyclonal sheep anti-fB(Binding Site) followed by HRP-conjugated anti-sheep Ig (Binding Site).Color was developed using o-phenylenediamine dihydrochloride (Dako-Cytomation). Background values were obtained by omitting fB and fD.Absorbance was measured at 490 nm, and percentage of convertase re-maining was calculated as follows: percentage of convertase remaining �100 � (A490inh � A490Bgnd)/(A490buffer-A490bgnd).

ResultsAssessment of C inhibitory activity by hemolysis assay

DAF1234 and deletion mutants DAF3, DAF34, and DAF23 weretested for their ability to prevent CP-mediated lysis of Ab-coatedsheep EA or AP-mediated lysis of rabbit E (Fig. 1). In the CPassay, DAF1234 and DAF23 were able to inhibit lysis, and theactivity of DAF23 was reduced by 4-fold compared withDAF1234. In the AP assay, DAF1234 was the only active reagent.These data are in agreement with previous reports that describe anessential role for SCR4 in the regulation of the AP.

Binding affinity between DAF, C3b, and Bb

We have previously demonstrated that DAF binds components ofthe AP convertase (12). In the presence of Mg2�, DAF bindsweakly to C3b and fB and much more tightly to the Bb fragment.The binding affinities of the DAF SCR-deletion mutants to C3b

and Bb were analyzed in this study using SPR. To analyze thebinding of C3b to DAF, C3b was coupled to the carboxy-methyldextran chip (CM5) by successive cycles of C3 and fB/fD as de-scribed previously (12). This resulted in coupling of C3b in a na-tive conformation and orientation via its internal thioester group.DAF reagents were subjected to size exclusion chromatography toeliminate any aggregates and were passed across the C3b surface.Binding demonstrated low-affinity fast on/fast off kinetics (Fig. 2).Affinity was determined by steady-state analysis from multiple de-terminations at different densities of C3b. DAF1234 and DAF34demonstrated similar binding profiles and bound with the sameaffinity to C3b (Table I). DAF23 and DAF3 did not bind to C3b atconcentrations of 22.3 �M (DAF23) or 61.7 �M (DAF3), even

FIGURE 1. Hemolysis assays to assess inhibition of C by DAF mutants.a, CP: Ab-coated sheep E were attacked with human C, and the ability ofrecombinant soluble DAF to prevent hemolysis was determined. Percent-age of inhibition of lysis was calculated as described in Materials andMethods by comparison with a noninhibitory protein. DAF1234 (�),DAF23 (f), DAF34 (Œ), and DAF3 (x). b, AP: Rabbit E were attackedwith human C and inhibitory activities determined as described for the CP.Results are means of three determinations and vertical bars represent SD.

FIGURE 2. Equilibrium analysis of recombinant soluble DAF mutantsbinding to C3b. C3b was immobilized on the chip surface via its internalthioester group. Interaction with DAF34 (a) or DAF1234 (b) was analyzedin HBS/1 mM Mg2� at various concentrations and sensorgram traces wereobtained. Interaction experiments were performed multiple times, a typicalexample is illustrated here for each form of DAF. The affinity of the in-teraction was analyzed by steady-state analysis (see inset); Req refers to theRU of DAF specifically bound to C3b at equilibrium. c, DAF23 and DAF3were flowed across a surface coated with high levels of C3b, no interactionwas evident.

354 DAF-MEDIATED DECAY OF THE AP

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

when the density of C3b on the chip surface was increased to 2700RU. To analyze binding of the Bb fragment to DAF, the variousDAF mutants were immobilized on to the chip surface and Bb was

flowed across the surface. It was not possible to flow DAF acrossa Bb-immobilized surface because amine coupling of Bb abolishedthe binding interaction, presumably due to the conformational sen-sitivity of Bb. The binding interaction between DAF1234 and Bbdemonstrated a kinetic profile similar to that which we previouslyreported, with a KD of 0.77 �M (Table I) (12). The data generatedduring the association/dissociation phases fit a Langmuir 1:1model tightly; the rapid change in signal at the inject start/stop maysuggest the presence of a small proportion of fB with an alteredconformation in the vWFA domain and very low affinity for DAF.In contrast to the C3b binding described above, DAF34 did notbind Bb, whereas DAF23 bound to Bb with the same affinityand kinetic profile as DAF1234 (Fig. 3). Surface density ofDAF34 was increased such that the theoretical Rmax was 2850RU; however, binding of Bb was not evident. When DAF3 wasimmobilized on the chip surface no interaction with Bb wasevident, even when 1000 RU of protein was immobilized (datanot shown).

Interaction of DAF with properdin

In vivo, the active AP convertase is likely to be associated withproperdin. To investigate whether properdin bound to DAF, pro-perdin was immobilized on to the dextran chip surface andDAF1234 was flowed across it. No binding interaction was evidentat 48.2 �M DAF1234, even when the density of properdin on thesurface was increased to 2900 RU (Fig. 4a). The surface was con-firmed as active by assembly of the C3bB proenzyme and C3bBbconvertase as described previously (Fig. 4b) (24). When C3bBbwas formed on the properdin surface it was sensitive to decay byDAF (4.8 �M DAF, illustrated in this study). The decay profileresembled that reported previously for DAF-mediated decay ofC3bBbP (24); the decrease back to baseline presumably reflects

FIGURE 3. Binding kinetics of Bb to recombinant soluble DAF mu-tants. DAF1234 (a), DAF23 (b), or DAF34 (c) were immobilized on thechip surface via amine groups. Interaction with Bb was analyzed in HBS/1mM Mg2� at various concentrations and sensorgram traces were obtained.Interaction experiments were performed multiple times, a typical exampleis illustrated here for each form of DAF. The affinity of the interaction wasanalyzed by 1:1 Langmuir binding, gray lines are fitted data (these linesunderlie the real data on each trace and are only just visible in places dueto the close nature of the fit).

FIGURE 4. Interaction of DAF with properdin and decay of properdin-stabilized convertase. High levels of properdin were immobilized on thechip surface and interaction with DAF1234 was assessed (a). The integrityof the surface was confirmed by forming the proenzyme C3bB (gray line)or C3bBb (black line) on immobilized properdin (b). Soluble DAF1234was able to decay the active convertase C3bBbP.

Table I. Kinetic constants for binding of DAF mutants to the ligandsC3b and Bb and ability to decay AP convertasea

DAFConstruct

C3b-Binding KD

(�M)Bb-Binding

KD (�M)

Ability toDecayC3bBb

DAF1234 10.8 � 1.0 (n � 5) 0.77 � 0.18 (n � 2) ��DAF23 No interaction 0.92 � 0.40 (n � 3) �DAF34/C3b 8.7 � 1.1 (n � 4) No interaction �DAF3/C3b No interaction No interaction �

a Values of KD given are mean � SD of multiple experiments (n). Ability to causedecay acceleration of the AP C3 convertase was assessed as follows: ��, powerfuldecay acceleration; �, decay acceleration evident; �, no decay acceleration.

355The Journal of Immunology

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

rapid release of Bb from the convertase followed by a slower re-lease of C3b from properdin.

DAF-mediated decay of C3bBb monitored using SPR

The AP convertase, C3bBb, was assembled on a Biacore chip sur-face by covalently binding C3b to the chip surface and formingconvertase by flowing fB and fD. The ability of DAF to decay theconstituent proteins was analyzed by SPR. DAF1234 was flowedacross the surface at different concentrations. DAF1234 decayedthe convertase very efficiently; at 5 �M and 0.5 �M baseline levelswere reached almost immediately (Fig. 5, a and b), and at 0.05 �M

DAF1234, decay was still evident, albeit less efficient (Fig. 5c). Abinding interaction between the convertase (C3bBb) and DAF1234was not visible because the drop in RU from accelerated decayoccurred too rapidly for DAF interaction to be visualized or quan-titated. At very high concentrations of DAF1234 (5 �M), a resid-ual DAF/C3b binding event characterized by very rapid dissocia-tion kinetics was seen (Fig. 5a).

The ability of the DAF deletion mutants to accelerate decay ofthe convertase was similarly analyzed. When DAF34 was flowedacross the convertase surface at 5 �M, accelerated decay wasevident (Fig. 5a, second inject); and at 0.5 �M DAF34 decay was

FIGURE 5. Decay of AP C3 convertase by DAF and DAF mutants monitored by SPR. C3b was immobilized on the chip surface via its thioester group.A mixture of fB and fD were flowed across the surface to form active convertase, C3bBb, as described in Materials and Methods; the point of inject isindicated in each sensorgram. Formation of the convertase and natural decay is indicated on the left side of each figure. At the indicated point, soluble DAFwas injected across the chip surface at 5 �M (a), 0.5 �M (b), or 0.05 �M (c). Accelerated decay or binding of DAF is indicated for DAF1234 (black trace),DAF34 (gray trace), DAF23 (black dashed trace), or DAF3 (gray dashed trace).

356 DAF-MEDIATED DECAY OF THE AP

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

barely visible (Fig. 5b). These data illustrate that SCRs 3 and 4together contain functional sites that accelerate decay of the con-vertase, although �100-fold more DAF34 was required to decaythe convertase compared with DAF1234. In contrast, DAF23bound to the convertase but accelerated decay was not apparent,even at the highest concentration (Fig. 5a, third inject). Becausedecay was not accelerated, it was possible to see DAF23 bindingto C3bBb; the binding profile resembled the interaction of DAFwith Bb (Fig. 3). DAF3 had no decay-accelerating ability at 5 �M(Fig. 5a, fourth inject). No binding of DAF3 to the convertase wasevident, confirming that the observed binding of DAF23 wasthrough SCR2.

DAF-mediated decay of C3bBb monitored using an ELISA

A second assay was used to monitor the ability of the DAFdeletion mutants to decay the AP convertase. A microtiter platewas coated with C3b, and the AP convertase was assembled onthe surface using fB and fD. In early experiments, plates werecoated with either C3b monomer or dimer to differentiate be-tween functional decay of the C3 and C5 convertase; however,the data were identical (data not shown) and all subsequentassays used C3b monomer on the plate surface. Following as-sembly of the convertase, different concentrations of DAF wereincubated in wells of the plate to accelerate decay of the en-zyme. Residual Bb was detected using a polyclonal anti-fB Ab,and decay was quantitated by comparing to control wells thathad contained just buffer during the decay period. These dataconfirmed the SPR results: DAF1234 was very efficient at de-caying the AP convertase with an activity comparable to solubleCR1, whereas DAF34 could decay the convertase only at highconcentrations (Fig. 6). DAF23 was inactive, confirming theSPR decay data.

DiscussionTo further investigate the mechanism by which DAF acceleratesdecay of the AP C3 convertase, we have generated recombinantsoluble constructs of DAF with deletions of SCR domains andhave assessed in this study their binding interactions and func-tional activities in the AP. Together, these mutants covered andoverlapped the proposed C-regulatory domains and were centeredon SCR3, a domain suggested from various studies to be importantin regulation of both the AP and CP (15, 26, 27). We used hemo-lysis assays to study C inhibitory function of these DAF mutantsand confirmed previous observations that SCR4 was an essential

requirement for regulation of the AP but not the CP (Fig. 1). Toassess direct binding interactions between the DAF constructs andthe AP C3 convertase components, C3b and Bb, we used SPR. Thedata implicated SCR2 in binding to the Bb portion of the conver-tase, whereas binding to C3b was mediated by SCR4. Neither theDAF23 nor the DAF3 construct bound C3b (Figs. 2 and 3). Thedata demonstrate two primary sites of interaction between DAFand the AP C3 convertase, but do not rule out the possibility ofother weaker interactions, particularly in the context of the wholemolecule where the presence of other domains may have subtleeffects on the overall tertiary structure of the molecule and orien-tation of the domains with respect to each other. It is importantto note, however, that the binding affinity of DAF1234 for C3b(�10 �M) was equal to that of DAF34 (Fig. 2; Table I), andaffinity of DAF1234 for Bb (�0.9 �M) was equal to that ofDAF23 (Fig. 3; Table I). This result demonstrates that these twosites of interaction can function independently and are not com-promised in the smaller constructs. Because both of these in-teractions can take place simultaneously when DAF bindsC3bBb, avidity effects will enhance overall binding affinity tothe active convertase.

To further address the relative importance of the C3b/SCR4 andBb/SCR2 interactions in regulation of the AP convertase, we usedtwo different assays that assessed the ability of constructs to decaythe convertase. Decay of the enzyme was monitored using SPR(Fig. 5). By visualizing binding and decay in real time, we foundthat the construct consisting of SCRs 3 and 4 mediated decay ofC3bBb, albeit less efficiently than the construct containing all fourSCRs. DAF23 bound the convertase (binding affinity for Bb is�0.9 �M) but binding did not mediate accelerated decay, in con-trast to binding of SCR34 that accelerated release of Bb. Theseactivities were confirmed using ELISA-based assays to monitorrelease of Bb from C3b on the plate surface (Fig. 6). One inter-pretation of these data is that decay of C3bBb occurs as a conse-quence of the C3b/SCR4 interaction. A possible sequence ofevents is as follows: DAF binds to the active convertase, C3bBb,through interactions in SCR2 and SCR4. Efficient binding to Bb(but not fB) targets DAF to the active enzyme. Because bindingbetween SCR4 and C3b is likely to be weak (KD of DAF34/C3binteraction is �10�5 M), we suggest that the primary role of theBb/SCR2 interaction in regulation of the AP is to focus DAF onthe active convertase and to hold the regulator in position, therebyaiding the SCR4/C3b interaction. Following binding of SCR4, ac-celerated decay is induced, possibly through a conformational tran-sition in C3b that results in destabilization of the interaction withBb and its release. Because C3b can rebind fB, any transition inC3b must be reversible following dissociation of DAF and Bbfrom C3b. It is noteworthy that the conformation of Bb once re-leased from the convertase is likely to differ from that in theC3bBb complex or in free fB because released Bb does not rebindto C3b. Binding of the mutants DAF34 and DAF23 to the conver-tase (Fig. 5) can be explained as follows: DAF 34 binds to theC3b fragment of the convertase; there is no avidity of bindingthrough Bb; therefore, 100-fold more DAF is required for bind-ing (Fig. 5a, gray trace, compared with Fig. 5c, black trace).Destabilization of the convertase proceeds via the SCR4 inter-action with C3b and Bb is dissociated. In the case of DAF23,the mutant binds to Bb, and this interaction can be visualized bySPR (Fig. 5a, black dashed trace); however, there is no inter-action with C3b, the enzyme does not break apart, and Bb re-mains bound to C3b.

We propose that the role of SCR3 is as a crucial spacing domainenabling correct orientation and presentation of SCR2 to Bb andSCR4 to C3b. The data summarized above contradict the perceived

FIGURE 6. Plate assay to monitor decay of AP C3 convertase byDAF and DAF mutants. C3b-coated wells were incubated with fB andfD to form active convertase C3bBb. Wells were incubated withDAF1234 (f), DAF23 (Œ), DAF34 (x), or soluble CR1 (�), and resid-ual Bb bound to C3b was detected using anti-fB Ab as described inMaterials and Methods. Results are means of three determinations andvertical bars represent SD.

357The Journal of Immunology

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

view of convertase regulation in that they do not reveal any bind-ing interaction between SCR3 and the convertase. SCR3 has beenimplicated in regulatory function in various studies including Ab-blocking and mutation studies. Molecular modeling has high-lighted two regions that might interact with convertases, astretch of positively charged amino acids (KKK125–127) pre-dicted to lie in the groove between SCRs 2 and 3, and a hy-drophobic patch (L147F148) located on the surface of SCR3 (28).Indeed, these amino acids are highly conserved across species.and disruption of either of these patches impaired ability ofDAF to regulate C (29, 30). Point mutation studies have iden-tified numerous amino acids whose substitution affects C reg-ulation, with some of these appearing to be key to function (26,27). Residues whose mutation affects regulation of both path-ways are located in SCRs 2 and 3 and include R69, R96, R100,L171 (SCR3), and two residues in the previously identified re-gions K127 and F148 (SCR3). Other residues are specific toregulation of a particular pathway such as E134 in the CP andK126, F169 (SCR3), and E206 (SCR4) in the AP.

Resolution of the solution structure of pairs of DAF domains(SCR23) and the crystal structure of the entire molecule have en-abled location of residues on the DAF surface and surface chargeto be mapped. Residues highlighted from point mutation studiesare located on both “faces” of DAF, implying that DAF may fitinto a groove on the convertase with contacts on both sides of themolecule. The solution structure of SCR23 indicated a flexiblehinge between the two domains with a positively charged interfaceencompassing the KKK residues highlighted above (31). The crys-tal structure of SCR1234 demonstrated a near linear molecule witha rigid junction between SCR2 and SCR3, a band of positivecharge encircling the upper portion of this junction, and a band ofuncharged residues encircling the amino-terminal end of SCR3,including F148, F169, and L171 (20, 27). This hydrophobic patchwas prominent only in the crystal structure, a consequence of dif-ferences in side-chain orientations between this and the solutionstructure. It was proposed that the energy of binding between DAFand the convertase was provided in part by association betweenthis hydrophobic domain and a similar hydrophobic patch on theconvertases (20). It is clear that SCR3 is indeed key to function;however, we demonstrate no direct binding with the convertase,C3bBb, or with properdin (Fig. 4). The role of SCR3 in main-taining SCR2 and SCR4 in the correct orientation crucial forefficient decay acceleration is supported by the crystal structureshowing the SCR2–3 and SCR3– 4 junctions to be rigid. Alter-ation of residues, such as KKK125–127 at the SCR2–3 join orhydrophobic residues at the amino terminus of SCR3, will im-pact on DAF function by altering this crucial domain orienta-tion. Mutations in SCR2 or its deletion weaken regulatory func-tion due to loss of interaction with Bb. Mutations in SCR3 affectthe relative orientations of SCR2 and SCR4, whereas deletionof SCR3 will destroy the correct spacing between two crucialdomains, effectively removing the cooperative binding betweenSCR2 and SCR4. The observed function blocking effects of Absagainst SCR3 are explained by steric hindrance because thisSCR is located between the two SCR that interface with theconvertase.

In summary, we have used sensitive techniques to study theinteractions of individual domains of DAF with components ofthe AP C3 convertase and have identified SCR2 and SCR4 asprimarily responsible for binding the active enzyme. Further-more, we have interrogated the mechanism by which DAF me-diates decay. The ability of DAF to “lock” in to an active en-zyme, but not to the proenzyme is due to the comparativelytight interaction between SCR2 and the Bb domain. Once held

in place, SCR4 interacts efficiently with C3b and decay accel-eration follows.

Interaction between SCR4 and C3b is prevented in the proen-zyme C3bB because the Ba domain (comprised from three SCRdomains) is still juxtaposed with C3b. The data presented in thisstudy shed new light on the mechanisms underlying the remark-able activity of DAF and show that multiple mechanisms ensurethat the activities of DAF are focused on regulating the activeenzyme.

DisclosuresThe authors have no financial conflict of interest.

References1. Pangburn, M. K., R. D. Schreiber, and H. J. Muller-Eberhard. 1981. Formation of

the initial C3 convertase of the alternative complement pathway: acquisition ofC3b-like activities by spontaneous hydrolysis of the putative thioester in nativeC3. J. Exp. Med. 154: 856–867.

2. Lachmann, P. J., and N. C. Hughes-Jones. 1984. Initiation of complement acti-vation. Springer Semin. Immunopathol. 7: 143–162.

3. Turner, M. W. 2003. The role of mannose-binding lectin in health and disease.Mol. Immunol. 40: 423–429.

4. Volanakis, J. E., and S. V. Narayana. 1996. Complement factor D, a novel serineprotease. Protein Sci. 5: 553–564.

5. Medicus, R. G., O. Gotze, and H. J. Muller-Eberhard. 1976. Alternative pathwayof complement: recruitment of precursor properdin by the labile C3/C5 conver-tase and the potentiation of the pathway. J. Exp. Med. 144: 1076–1093.

6. Law, S. K., and R. P. Levine. 1977. Interaction between the third complementprotein and cell surface macromolecules. Proc. Natl. Acad. Sci. USA 74:2701–2705.

7. Brodbeck, W. G., C. Mold, J. P. Atkinson, and M. E. Medof. 2000. Cooperationbetween decay-accelerating factor and membrane cofactor protein in protectingcells from autologous complement attack. J. Immunol. 165: 3999–4006.

8. Morgan, B. P., and C. L. Harris. 1999. Complement Regulatory Proteins. Aca-demic Press, London.

9. Morgan, B. P., and S. Meri. 1994. Membrane proteins that protect against com-plement lysis. Springer Semin. Immunopathol. 15: 369–396.

10. Reid, K. B., and A. J. Day. 1989. Structure-function relationships of the com-plement components. Immunol. Today 10: 177–180.

11. Nicholson-Weller, A., and C. E. Wang. 1994. Structure and function of decayaccelerating factor CD55. J. Lab. Clin. Med. 123: 485–491.

12. Harris, C. L., R. J. Abbott, R. A. Smith, B. P. Morgan, and S. M. Lea. 2005.Molecular dissection of interactions between components of the alternative path-way of complement and decay accelerating factor (CD55). J. Biol. Chem. 280:2569–2578.

13. Hourcade, D. E., L. Mitchell, L. A. Kuttner-Kondo, J. P. Atkinson, andM. E. Medof. 2002. Decay-accelerating factor (DAF), complement receptor 1(CR1), and factor H dissociate the complement AP C3 convertase (C3bBb) viasites on the type A domain of Bb. J. Biol. Chem. 277: 1107–1112.

14. Kuttner-Kondo, L. A., M. P. Dybvig, L. M. Mitchell, N. Muqim, J. P. Atkinson,M. E. Medof, and D. E. Hourcade. 2003. A corresponding tyrosine residue in theC2/factor B type A domain is a hot spot in the decay acceleration of the com-plement C3 convertases. J. Biol. Chem. 278: 52386–52391.

15. Coyne, K. E., S. E. Hall, S. Thompson, M. A. Arce, T. Kinoshita, T. Fujita,D. J. Anstee, W. Rosse, and D. M. Lublin. 1992. Mapping of epitopes, glyco-sylation sites, and complement regulatory domains in human decay acceleratingfactor. J. Immunol. 149: 2906–2913.

16. Pangburn, M. K. 1986. Differences between the binding sites of the complementregulatory proteins DAF, CR1, and factor H on C3 convertases. J. Immunol. 136:2216–2221.

17. Brodbeck, W. G., D. Liu, J. Sperry, C. Mold, and M. E. Medof. 1996. Localiza-tion of classical and alternative pathway regulatory activity within the decay-accelerating factor. J. Immunol. 156: 2528–2533.

18. Harris, C. L., D. M. Lublin, and B. P. Morgan. 2002. Efficient generation ofmonoclonal antibodies for specific protein domains using recombinant immuno-globulin fusion proteins: pitfalls and solutions. J. Immunol. Methods 268:245–258.

19. Harrison, R. A. 1996. Purification, assay and characterization of complementproteins from plasma. In Weir’s Handbook of Experimental Immunology, Vol-ume II. Cell Surface and Messenger Molecules of the Immune System.L. A. Herzenberg, D. M. Weir, L. A. Herzenberg, and C. Blackwells, eds.Blackwell Science, Cambridge, pp. 75.71–75.50.

20. Lukacik, P., P. Roversi, J. White, D. Esser, G. P. Smith, J. Billington, P. A. Williams,P. M. Rudd, M. R. Wormald, D. J. Harvey, et al. 2004. Complement regulation at themolecular level: the structure of decay-accelerating factor. Proc. Natl. Acad. Sci. USA101: 1279–1284.

358 DAF-MEDIATED DECAY OF THE AP

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

21. White, J., P. Lukacik, D. Esser, M. Steward, N. Giddings, J. R. Bright,S. J. Fritchley, B. P. Morgan, S. M. Lea, G. P. Smith, and R. A. Smith. 2004.Biological activity, membrane-targeting modification, and crystallization of sol-uble human decay accelerating factor expressed in E. coli. Protein Sci. 13:2406–2415.

22. Pettigrew, D. M., D. T. Williams, D. Kerrigan, D. J. Evans, S. M. Lea, andD. Bhella. 2006. Structural and functional insights into the interaction of echo-viruses and decay-accelerating factor. J. Biol. Chem. 281: 5169–5177.

23. Jokiranta, T. S., J. Westin, U. R. Nilsson, B. Nilsson, J. Hellwage, S. Lofas,D. L. Gordon, K. N. Ekdahl, and S. Meri. 2001. Complement C3b interactionsstudied with surface plasmon resonance technique. Int. Immunopharmacol. 1:495–506.

24. Hourcade, D. E. 2006. The role of properdin in the assembly of the alternativepathway C3 convertases of complement. J. Biol. Chem. 281: 2128–2132.

25. Krych-Goldberg, M., R. E. Hauhart, V. B. Subramanian, B. M. Yurcisin 2nd,D. L. Crimmins, D. E. Hourcade, and J. P. Atkinson. 1999. Decay acceleratingactivity of complement receptor type 1 (CD35): two active sites are required fordissociating C5 convertases. J. Biol. Chem. 274: 31160–31168.

26. Kuttner-Kondo, L. A., L. Mitchell, D. E. Hourcade, and M. E. Medof. 2001.Characterization of the active sites in decay-accelerating factor. J. Immunol. 167:2164–2171.

27. Williams, P., Y. Chaudhry, I. G. Goodfellow, J. Billington, R. Powell,O. B. Spiller, D. J. Evans, and S. Lea. 2003. Mapping CD55 function: the struc-ture of two pathogen-binding domains at 1.7 Å. J. Biol. Chem. 278:10691–10696.

28. Kuttner-Kondo, L., M. E. Medof, W. Brodbeck, and M. Shoham. 1996. Molec-ular modeling and mechanism of action of human decay-accelerating factor. Pro-tein Eng. 9: 1143–1149.

29. Brodbeck, W. G., L. Kuttner-Kondo, C. Mold, and M. E. Medof. 2000. Structure/function studies of human decay-accelerating factor. Immunology 101: 104–111.

30. Kuttner-Kondo, L., V. B. Subramanian, J. P. Atkinson, J. Yu, and M. E. Medof.2000. Conservation in decay accelerating factor (DAF) structure among primates.Dev. Comp. Immunol. 24: 815–827.

31. Uhrinova, S., F. Lin, G. Ball, K. Bromek, D. Uhrin, M. E. Medof, andP. N. Barlow. 2003. Solution structure of a functionally active fragment of decay-accelerating factor. Proc. Natl. Acad. Sci. USA 100: 4718–4723.

359The Journal of Immunology

by guest on Novem

ber 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from