Decay Accelerating Activity of Complement Receptor Type 1 (CD35)

The goal of this study was to identify the site(s) in CR1 that mediate the dissociation of the C3 and C5 convertases. To that end, truncated derivatives of CR1 whose extracellular part is composed of 30 tandem repeating modules, termed complement control protein repeats (CCPs), were generated. Site 1 (CCPs 1–3) alone mediated the decay acceleration of the classical and alternative pathway C3 convertases. Site 2 (CCPs 8–10 or the nearly identical CCPs 15–17) had one-fifth the activity of site 1. In contrast, for the C5 convertase, site 1 had only 0.5% of the decay accelerating activity, while site 2 had no detectable activity. Efficient C5 decay accelerating activity was detected in recombinants that carried both site 1 and site 2. The activity was reduced if the intervening repeats between site 1 and site 2 were deleted. The results indicate that, for the C5 convertases, decay accelerating activity is mediated primarily by site 1. A properly spaced site 2 has an important auxiliary role, which may involve its C3b binding capacity. Moreover, using homologous substitution mutagenesis, residues important in site 1 for dissociating activity were identified. Based on these results, we generated proteins one-fourth the size of CR1 but with enhanced decay accelerating activity for the C3 convertases.

Recently, CR1 was shown to be the receptor involved in rosette formation between malaria-infected and -uninfected erythrocytes (a phenomenon that correlates with cerebral malaria and a high mortality) (16). This interaction involves CR1 on the uninfected cells and varies with expression of CR1related blood group antigens.
The complement cascade requires stringent control. The C3 and C5 convertases are inhibited by proteins encoded by the regulators of complement activation (RCA) gene family (17). Factor H and C4b binding protein are fluid phase regulators present in plasma, whereas membrane cofactor protein (CD46) and decay accelerating factor (DAF; CD55) are ubiquitously expressed membrane inhibitors. RCA proteins accelerate the dissociation of C3 and C5 convertases (decay accelerating activity (DAA)) and/or serve as cofactors for the factor I-mediated cleavage of C3b and C4b (cofactor activity (CA)). CR1 is the most versatile member of this group. It possesses DAA and CA and inhibits C3 and C5 convertases of both the classical and alternative pathways (CP and AP, respectively). Moreover, it is unique among the cofactor proteins in that through its cofactor activity it generates C3d, which serves as a ligand for complement receptor type 2. A recombinant, soluble form of CR1 (sCR1) inhibits self-destructive C activation in immune complex-mediated syndromes, ischemia/reperfusion injury (18), and hyperacute xenograft rejection (19). It is presently in clinical trials as a complement inhibitor (20).
Like all members of the RCA family, CR1 is composed of the ϳ60-amino acid-long repeating units called complement control protein repeats (CCPs) or short consensus repeats. Of the 30 CCPs in CR1, the first 28 can be organized, based on internal homology, into four long homologous repeats (LHRs), A-D, each composed of seven CCPs (21,22). Analyses of CR1 derivatives carrying a single LHR established that the LHRs A, B, and C contain C3b/C4b binding sites (23,24). LHR B and its structural and functional duplicate LHR C bind C3b and C4b and possess CA for their cleavage (25)(26)(27). LHR A binds C4b with a similar affinity as LHR B or C but binds C3b weakly and has barely detectable CA (27).
The initial three CCPs of LHRs A, B, and C form a functional unit, called site 1 in LHR A and site 2 in LHR B or C (21,27). The amino acid sequences of CCPs 1 and 2 are 39% different from their counterparts in site 2 (CCPs 8 and 9 or CCPs 15 and 16). On the other hand, the amino acid sequences of the third CCP in each site (CCPs 3, 10, and 17) are nearly identical. Consequently, variation in the initial two CCPs determines the functional differences between the sites. Using homologous substitution mutagenesis, amino acids important for ligand binding and for CA in site 1 and in site 2 were identified (24,27,28). In this investigation, our goal was to localize site(s) for DAA and to identify peptides and amino acids critical for this regulatory activity.

General Design
All constructs were expressed as secreted proteins and quantified by enzyme-linked immunosorbent assay (27). They were then assessed for iC3/C4b binding ability and for DAA.

Construction of CR1 Derivatives
Polymerase chain reaction (PCR) products were subcloned in vector pCR™II (Invitrogen, Carlsbad, CA), sequenced, and cut with the appropriate enzyme for cloning in the expression vector to generate the constructs shown in Fig. 1. Amino acid substitutions were made by site-directed mutagenesis (QuikChange™, Stratagene, La Jolla, CA) in the expression vector pSG5 (Stratagene) without subcloning.
LHR C and LHR D ϩ -The construction of LHR C, which contains CCPs 15-21, (i.e. amino acids 898 -1353) has been described (28). LHR D ϩ (amino acids 1348 -1928) contains LHR D (CCPs 22-28) followed by CCPs 29 and 30. CR1 cDNA in AprM8 obtained from Lloyd Klickstein (Harvard Medical School, Boston, MA) served as a template for PCR amplification of the signal peptide and of LHR D ϩ . The 5Ј-primer for amplification of the signal peptide, ATATACAGATCTATGGGAGCCT-CTTCTCCAAGA, contains a BglII site (underlined) for cloning into pSG5. The 3Ј-primer, CGGATCCCCATGCAACAGGCAGTGCAAGCA-GCACCACAACTGCTAGCAGA, destroys an internal BamHI site by a silent mutation and generates the terminal BamHI site (underlined) by another silent mutation for ligation to LHR D ϩ . Six more silent mutations interrupt GC-rich stretches. The 5Ј-primer for the amplification of LHR D ϩ , ACATAGGATCCGTTCGTGCTGGTC, incorporates codons for six out of eight amino acids between the last Cys of CCP 21 and the first Cys of CCP 22. The BamHI site (underlined) was created by a silent mutation for ligation to the signal peptide. The 3Ј-primer for amplification of LHR D ϩ , TATATAAGATCTTTATGCACGAGAGGTACATTT, replaces His 1929 (the fifth position downstream of the last Cys in CCP 30) with a stop codon followed by a BglII site (underlined) for cloning into pSG5. Each PCR product was cut from pCR™II vector with BamHI and BglII. The signal peptide and LHR D ϩ were then ligated through the BamHI site, and the ligation product was cloned into the BglII site of pSG5. The sequence of LHR D ϩ contained an ATT coding for Ile 1835 instead of the reported ACT coding for Thr 1835 (21). ATT was present in the PCR product and in the original template.
LHR AC (CCP 1-7 followed by  i.e. amino acids 1-449 followed by amino acids 899 -1353) and LHR BC (CCP 8 -21; i.e. amino acids 451-1353)-LHR A in CR1-4 and LHR B in CR1-4 (8,9) are followed by CCP 8 (identical to CCP 15) and the first half of CCP 9 (identical to the homologous part of CCP 16). The latter contains a unique PflMI site. CR1-4 and CR1-4 (8,9) were linearized at this site to generate the amino-terminal portion of LHR AC and LHR BC, respectively, extending to the PflMI site in CCP 16. The remaining part of LHR C, namely to the end of CCP 21, was amplified using CR1 cDNA as a template. The 5Ј-primer, GATCCAGTGAATGGCATGG, includes a PflMI site (underlined) in CCP 16. The 3Ј-primer, CCATTCACTGGT-TAACCAGCACGAACAGAAAGTTC, replaces Gly 1353 (the seventh position downstream of the last Cys of CCP 21) with a stop codon followed by a PflMI site (underlined). The PCR product, cut from pCR TM II with PflMI, was ligated to PflMI-linearized CR1-4 and to CR1-4 (8,9), generating LHR AC and LHR BC, respectively. CCP 1-4,15-18 (amino acids 1-254 followed by amino acids 899 -1153)-CR1-4 was used as a template for the PCR amplification of CCPs 1-4. The 5Ј-primer, ATATAGATCTCAGGTCATTGTTCCA, con-tains a BglII site (underlined) for cloning into pSG5. The 3Ј-primer, ATATATCCTAGGCGTGAGCAGCTTGGTAG, anneals to the end of CCP 4 and includes amino acids 253 and 254 after the last Cys of CCP 4. The AvrII site (underlined), generated by two silent mutations, facilitated ligation to CCP 15. LHR C served as a template for the PCR amplification of CCP 15-18. The 5Ј-primer, CGCTGTGGAATCCTAG-GTCACTGTCAAGCCCCA, includes amino acids 899 -901 preceding the first Cys of CCP 15. The AvrII site (underlined) was created by two silent mutations for ligation to CCP 4. The 3Ј-primer, CTAGTCGTAA-GATCTTTACCTGGAGCAGCTTGG, replaces Val 1155 (the third position after the last Cys in CCP 18) with a stop codon and contains a BglII site (underlined). The two PCR products were cut from pCR™II with BglII and AvrII. CCP 15-18 was then ligated to CCP 1-4 through the AvrII site, generating a cDNA for the construct CCP 1-4,15-18.
CCP 1-4,15-21 (amino acids 1-254 followed by amino acids 899 -1353) was constructed by deleting CCP 5-7 from LHR AC. This was accomplished by replacing the sequence between the two HindIII sites in LHR AC, one in CCP 4 and the other in CCP 18, with the sequence present between the same HindIII sites in CCP 1-4,15-18.
Substitution Mutants in CCPs 1 and 2-The strategy for homologous substitution mutagenesis in CCPs 1 and 2 has been described (28) and is summarized in Fig. 4. The new mutants, 13a-d, 9a, 9b, 6c, and 10a11c, are derived from LHR A and are in the expression vector pSG5. The other mutants represented in Fig. 4 have been described (24,27,28).

Preparation of C3b Dimers
C3 was purified from plasma by anion exchange chromatography (29). To generate C3b dimers, C3 (2 mg) in 200 l of phosphate-buffered saline at pH 7 was treated with 20 g of trypsin (Sigma) for 3 min at 37°C. The reaction was stopped with 200 g of soybean trypsin inhibitor (Sigma). Cross-linking through the free sulfhydryl group generated by breaking the thioester bond was then performed for 3 days at 4°C using 15 l of 0.34 mM bismaleimidohexane (Pierce) dissolved in methanol. With this procedure, the yield was over 50%. If, however, as recommended by the manufacturer, bismaleimidohexane was dissolved in Me 2 SO, the yield of C3b dimers did not exceed 10%. The dimers were purified by size exclusion high performance liquid chromatography using a TSK 4000SW column (TosoHaas, Montgomeryville, PA). The dimers were 95% pure based on a Coomassie Blue-stained gel.

Expression of CR1 Derivatives in COS 7 Cells
Plasmids containing cDNAs encoding CR1 derivatives were transfected into COS 7 cells using LipofectAMINE (Life Technologies, Inc.) according to the manufacturer's instructions. One day after transfection, the cells were washed and incubated in Opti-MEM I medium in the absence of serum to avoid possible DAA from the C4b binding protein and factor H in bovine serum. After a 2-or 3-day incubation, supernatants were collected, aliquoted, and stored at Ϫ70°C. Protein levels were estimated by enzyme-linked immunosorbent assay (27) using two monoclonal anti-CR1 antibodies, 3D9 (4) and E11 (30). Because proteins CCP 1-3 and CCP 1-4,15-18 are not recognized by E11, a polyclonal antibody (25) was used for their quantification.

Ligand Binding
iC3 and C4b binding experiments were performed as described (27,28), but the data are presented differently to be consistent with the decay accelerating results. Instead of reporting the fraction of a CR1 derivative that bound to iC3-or C4b-Sepharose, binding of the mutant proteins was expressed relative to the parental protein LHR A. A 100% value was assigned to the fraction of LHR A that bound to C4b-or iC3-Sepharose.

Assays for Decay Accelerating Activity
Alternative pathway factors B, D, and P; classical complement pathway components C1, C2, C4, and C3; terminal components C5-C9; and antibody-sensitized sheep erythrocytes (EA) were obtained from Advanced Research Technologies (San Diego, CA).

Hemolytic Assay for the Classical Pathway DAA
The assay was performed in isotonic veronal-buffered saline containing 2.5 mM veronal (pH 7.35), 71 mM NaCl, 139 mM dextrose, 0.15 mM CaCl 2 , 1 mM MgCl 2 , and 0.1% gelatin (DGVB 2ϩ ) (31). Three ml of washed EA (2.5 ϫ 10 8 /ml) were incubated with C1 (6 g diluted in 3 ml) at 30°C for 15 min. After centrifugation, cells were resuspended in 3 ml of DGVB 2ϩ , and 10 g of C4 in 1.5 ml of buffer was added. After a 15-min incubation at 30°C, the cells were centrifuged and resuspended to 3 ml. One ml of these EAC14 cells was incubated with 0.5 g of C2 diluted in 1 ml for 4 min at room temperature. To prepare the C5 convertase, EAC14 cells were incubated with 0.5 g of C2 and with 10 g of C3 at room temperature for 4 min. The resulting EAC142 or EAC1423 cells were centrifuged, washed, resuspended in 1 ml, and used to assess DAA for the C3 convertase or the C5 convertase, respectively. Specifically, EAC142 cells (50 l) were incubated for 10 min at 30°C with 50 l of an inhibitor or with buffer and transferred to ice, and then 0.5 ml of CЈ-EDTA (a 20-fold dilution of guinea pig serum (Colorado Serum, Denver, CO) in 40 mM EDTA containing veronal buffered saline) was added. In the modified assay, CЈ-EDTA was replaced with 0.5 ml of buffer containing 500 ng of each terminal component (C5-C9), with or without C3. After a 30-min incubation at 37°C and centrifugation, the OD of the supernatant was read at 414 nm.

Hemolytic Assay for the Alternative Pathway DAA
The convertase was assembled using purified components (32). EAC14 cells (0.5 ml), prepared as above but at a concentration of 1.5 ϫ 10 9 /ml, were incubated with 0.225 g of C2 and 50 g of C3 for 30 min at 30°C. After washing and resuspending in veronal-buffered saline containing 128 mM NaCl, 4.5 mM veronal (pH 7.35), 0.1% gelatin, and 10 mM Na 2 -EDTA, cells were incubated for 2 h at 37°C to allow dissociation of C1 and C2. The resulting EAC43 cells were washed twice with the above buffer; washed twice with buffer (Mg 2ϩ -EGTA) containing 10 mM Na 2 -EGTA, 7 mM MgCl 2 , 115 mM dextrose, 0.83% gelatin, 59 mM NaCl, and 4.2 mM veronal (pH 7.35); and then resuspended in 6 ml of the same buffer.
EAC43 cells (100 l) were incubated for 30 min at 30°C with 5 ng of factor D, 45 ng of properdin, 1.5 ng of factor B, and 75 l of inhibitor or Mg 2ϩ -EGTA. After adding 0.3 ml of CЈ-EDTA, hemolysis was estimated as for the classical pathway. In the modified assay, CЈ-EDTA was replaced with 300 l of buffer containing 300 ng of each terminal component (C5-C9) with or without C3.

Microtiter Plate Assay for the Alternative Pathway DAA
For the C3 convertase decay accelerating assay (33), microtiter plates were coated overnight with 5 g/ml C3b (Advanced Research Technologies) in phosphate-buffered saline. Plates were blocked for 2 h at 37°C with phosphate-buffered saline containing 1% bovine serum albumin and 0.1% Tween 20. Plates were then incubated for 15 min at 37°C with 10 ng of factor B, 1 ng of factor D, and 0.8 mM NiCl 2 in 2.5 mM veronal buffer, pH 7.4, containing 71 mM NaCl and 0.05% Tween 20. Using this same buffer, sequential 1-h incubations were performed with 1-10 ng of a CR1 derivative, 0.129 g of goat anti-human factor B antibody (Incstar, Stillwater, MN), and 100 l of a 1:15,000 dilution of anti-goat antibody conjugated to horseradish peroxidase (Jackson Immunoresearch Laboratories, West Grove, PA). Color was developed with Ophenylenediamine. In this assay, DAF and factor H behave as expected as mediators of decay accelerating activity (see Ref. 33), and we have detected C3a release using the Amersham Pharmacia Biotech C3a Based on the degree of identity, the first 28 CCPs form LHRs A, B, C, and D, which arose through duplication of a seven-CCPs unit. There are two distinct functional sites, each composed of three CCPs. Site 1 was localized to LHR A, and two nearly identical copies of site 2 were localized to LHRs B and C (23-28). The first two CCPs in site 1 (CCPs 1 and 2) are 39% different from the first two CCPs in site 2 (CCPs 8 and 9) as well as from CCPs 15 and 16, and they are marked by different shading. The third CCP in site 1 is nearly identical to the third CCP (10 or 17) in site 2. CCPs 3, 10, and 17 are represented by boxes with horizontal lines. In addition to LHR D, LHR D ϩ includes CCPs 29 and 30. 12 Ϯ3 LHR C 8 Ϯ 5 LHR D ϩ 0 a A dose-response curve was performed with sCR1 and with each derivative. 100% inhibition was assigned to sCR1. Percentage of inhibition is expressed as a ratio of concentration of sCR1, which reduced hemolysis by 50%, to the concentration of a derivative, which reduced hemolysis by the same amount. Data represent mean Ϯ S.D. of three experiments.
a A dose-response curve was performed with sCR1 and with each derivative. 100% inhibition was assigned to sCR1. Percentage of inhibition is expressed as a ratio of concentration of sCR1, which reduced hemolysis by 50%, to the concentration of a derivative, which reduced hemolysis by the same amount. Data represent mean Ϯ S.D. of three experiments.
des-Arg RIA kit. 2 For the C5 convertase assay, microtiter wells were coated (as above) with C3b dimers at 5 g/ml. The assay was performed under conditions similar to those used for the C3 convertase except that the amounts of factor B and D were 20 and 2 ng, respectively.

RESULTS
Characterization of the structural requirements for DAA proceeded in three steps: first, identification of the required LHR(s); second, localization of the necessary CCPs within an LHR; and, third, determination of critical peptides and individual amino acids within CCPs. DAA was analyzed for the CP and AP C3 and C5 convertases, and the activity of the constructs (illustrated in Fig. 1) was compared.
LHRs Required for DAA-A hemolytic assay was employed in which the CP C3 convertase was formed using purified components, and the lytic sites were developed with CЈ-EDTA. The main site of DAA for the CP C3 convertase was in LHR A (Table I), which had 60% of the activity of sCR1. LHR B or C had approximately 1 ⁄10 the activity of sCR1. LHR D ϩ lacked detectable DAA.
In contrast, if DAA for the AP was assessed by a similar hemolytic assay, LHR A had only 0.5% of the activity of sCR1 (Table II), while LHRs B, C, and D ϩ had no detectable activity (even if the molar concentration of these proteins was 100-fold higher than sCR1). Mixing LHR A with LHR B or D ϩ did not result in activity above that of LHR A alone. These results were surprising because it was expected that, as for the CP C3 convertase, a single LHR would be capable of dissociating the AP C3 convertase. To further assess this issue, constructs bearing more than one LHR were evaluated (Table III). While LHR BC had barely detectable activity, LHR AC had DAA equal to 50% of the DAA of sCR1.
These data suggested that LHR A was sufficient for DAA of the CP C3 convertase, while LHR AC was required for the AP C3 convertase. Because the two C3 convertases are so similar and because it was difficult to reconcile results for the AP with those for the CP, a second assay for the AP C3 convertase was employed in which the convertase was assembled on a microtiter plate using purified components. In this assay, LHR A had DAA equivalent to that of sCR1, whereas LHR B or C had 12 and 14%, respectively (Table IV). These results, consistent with the data obtained in the CP C3 convertase assay, indicate that LHR A contains the major site for CP and AP DAA.
To account for the requirement that LHR A be linked to LHR C for efficient decay of the AP convertase on EA, we hypothesized that during the convertase assembly step both the C3 convertase (C3bBb) and the C5 convertase ((C3b) 2 Bb) had been generated. This might occur if C3b dimers were produced by the classical pathway C3 convertase employed to deposit C3b. If so, LHR A would be sufficient for DAA for the C3 convertase, whereas LHR AC might be necessary for efficient dissociation of the C5 convertase. To test this possibility, the hemolytic assay was modified such that instead of developing lytic sites with CЈ-EDTA, they were developed using purified C5-C9, with or without the addition of C3. The rationale was that, if a C5  a A dose-response curve was performed with sCR1 and with each derivative. 100% inhibition was assigned to sCR1. Percentage of inhibition is expressed as a ratio of concentration of sCR1, which reduced hemolysis by 50%, to the concentration of a derivative, which reduced hemolysis by the same amount. Data represent mean Ϯ S.D. of three experiments. a Concentration of the inhibitors was a log lower in the C5 convertase assay than in the C3 convertase assay. b A dose-response curve was performed with sCR1 and with each derivative. 100% inhibition was assigned to sCR1. Percentage of inhibition is expressed as a ratio of concentration of sCR1, which reduced hemolysis by 50%, to the concentration of a derivative, which reduced hemolysis by the same amount. Data represent mean Ϯ S.D. of three experiments. a A dose-response curve was performed with sCR1 and with each derivative. 100% inhibition was assigned to sCR1. Percentage of inhibition is expressed as a ratio of concentration of sCR1, which reduced hemolysis by 50%, to the concentration of a derivative, which reduced hemolysis by the same amount. Data represent mean Ϯ S.D. of three experiments.
convertase were generated during the initial steps of the assay, hemolysis would be independent of additional C3. 60% of erythrocytes were lysed with the addition of C5-C9, and 80% were lysed if C3 was also added. The 60 Ϯ 7% value is similar to the 56 Ϯ 8% lysis if CЈ-EDTA was used to develop lytic sites. Because hemolysis was largely independent of the added C3, we concluded that the AP C5 convertase had been generated on erythrocytes, probably as a result of C3b dimer formation (34).
To further address the question, microtiter wells were precoated with purified C3b dimers and incubated with Ni 2ϩ and factors B and D. The resulting complexes, (C3b) 2 BbNi 2ϩ , were treated with LHR A or with LHR AC. LHR AC was 10 times more effective than LHR A in dissociating the C5 convertase (Fig. 2). This finding was consistent with the results obtained in the hemolytic assay (Table III) in which a C5 convertase was also present.
We next asked if LHR AC is required for efficient DAA of the CP C5 convertase as well. An assay was employed in which a C5 convertase was formed on EA and lytic sites developed by the addition of purified C5-C9. DAA of the C5 convertase by LHR AC was similar to that of sCR1, while LHR A was approximately 200-fold less effective (Table V). As in the case of the AP C5 convertase, LHR BC had little DAA for the CP C5 convertase (Table V). These results establish that optimal DAA of C5 convertases requires two functional sites.  16 Ϯ 4 0 a DAA for the classical pathway was tested in a hemolytic assay and DAA for the alternative pathway was tested in a microtiter plate assay. b A dose-response curve was performed with sCR1 and with each derivative. 100% inhibition was assigned to sCR1. Percentage of inhibition is expressed as a ratio of concentration of sCR1, which reduced hemolysis by 50%, to the concentration of a derivative, which reduced hemolysis by the same amount. Data represent mean Ϯ S.D. of three experiments.
Identification of CCPs Required for DAA-To determine which CCPs within LHR A are responsible for DAA of the CP and AP C3 convertases, truncated LHR A derivatives, containing a full or partial site 1, were analyzed (Table VI). DAA of CCP 1-3 (site 1) was comparable with that of LHR A or CCP 1-4, indicating that CCPs 4 -7 do not contribute to this regulatory function. We next constructed LHR A derivatives deleted of CCP 1, 2, or 3 (Table VI). DAA was substantially reduced, indicating that each CCP is necessary for DAA as they are for ligand binding (24,28).
Since LHR A linked to LHR C was required for DAA for the C5 convertases, we asked if CCPs 5-7, located between the two active sites, play a role. The construct CCP 1-4,15-18, lacking CCPs 5-7, had 7-and 10-fold lower activity for the CP and AP, respectively, than did LHR AC but much higher activity than LHR A (Fig. 3, Table V). Construct CCP 1-4,15-21 was very similar to CCP 1-4,15-18 (not shown). Because CCPs 5-7, located between the active sites, possess no activity on their own, a requirement for proper spacing between sites 1 and 2 in the decay acceleration of C5 convertases is strongly suggested.
Identification of Peptides and Amino Acids Important for DAA-Modified forms of LHR A, carrying amino acid substitutions in CCPs 1 and 2 derived from the homologous positions of CCPs 8 and 9, respectively (Fig. 4), were assessed for DAA for C3 convertases. For these experiments, a hemolytic assay for the CP and a microtiter assay for the AP were employed. All mutants shown in Fig. 4 were tested, but only mutations that produced a Ϯ20% change are listed in Table VII. The effects of the mutations on DAA and on ligand binding (Refs. 24 and 28; present report) defined the three groups.
Group I represents substitutions, which decrease DAA for one or both convertases but have no effect on ligand binding. Hence, these mutations identify residues likely to be specifically required for DAA. In this group, the most marked effect was caused by F82V, which nearly abrogated DAA for both C3 convertases. To a lesser extent, mutations T103E, T110A, and V111A also reduced DAA for both C3 convertases. Other mutations reduced DAA for one pathway only. For example, W7H almost abrogated DAA for the AP, while mutations K92T and S99H reduced DAA for the CP.
Groups II and III are mutations that caused parallel changes in DAA and in ligand binding. Group II mutations produced a decrease in DAA and binding. For example, mutant G35E had reduced C4b and C3b binding and DAA for the AP and CP C3 convertases. Two substitutions, R64K and Y94H, produced a reduction in C4b binding and DAA for the CP C3 convertase. N65T reduced C3b binding and DAA for the AP C3 convertase only.
Group III mutations increased ligand binding and DAA. Mutant D109N had enhanced DAA for the C3 convertases and ligand binding (Table VII). This mutant and the double mutant D109N,E116K are of special interest because of the marked enhancement in DAA, which is further illustrated in the kinetic analysis shown in Fig. 5. Two other mutations, E6D and N29K, increased DAA for the CP C3 convertase, while mutation S37Y increased DAA for the AP convertase. In addition, mutant 14, with nine amino acid substitutions, showed increased DAA for the AP. No single mutation was responsible for this effect. DISCUSSION DAA for C3 Convertases-LHR A was identified as the main site of DAA for the CP and AP C3 convertases (Refs. 35 and 36; this report). Within LHR A, CCPs 1-3 account for all DAA, being equivalent to CCPs 1-4 or LHR A itself. If CCP 1, 2, or 3 was deleted from LHR A, DAA was markedly reduced. Thus, CCPs 1-3 (site 1) are necessary and sufficient for the dissociating activity. That site 1 was the major region for DAA for both convertases was unexpected because C3b binding by CCPs 1-3 is very low. However, this seemingly paradoxical observation is consistent with the functional profile of DAF, which has a low affinity for C3b or C4b but efficiently dissociates both C3 convertases (37). Transient interaction of an inhibitor with a convertase appears sufficient and probably optimal for deactivation. Like DAF, site 1 of CR1 may have a higher affinity for the convertases than for C3b or C4b alone (37). Previously, we found that most of the cofactor activity for C3b and C4b resides in site 2, which is located in CCPs 8 -10 and duplicated in the nearly identical CCPs 15-17 (27). Here we show that this site has inefficient DAA in its own right. Therefore, site 1 is DAFlike, whereas site 2 is membrane cofactor protein-like.
While the major site of DAA for the C3 convertases is in LHR A, several lines of evidence indicate a contribution from site 2. First, LHRs B and C have detectable DAA. Second, compared with sCR1, LHR A has 60% of DAA for the CP C3 convertase, while LHR AC has 100%. These results suggest that site 2 participates in regulating the CP C3 convertase. Regarding the AP C3 convertase, in the microtiter assay system, site 1 accounts for all of the DAA detected with sCR1. In the hemolytic assay, the contribution of site 2 to the decay of the C3 conver- tase could not be evaluated because of the simultaneous presence of the C5 convertase (see below).
DAA for C5 Convertases-In contrast to the C3 convertases, optimal DAA for the C5 convertases requires both site 1 and site 2. LHR B or C, single or linked together as in LHR BC, lacked DAA for the C5 convertases. If LHR A is mixed with LHR B, DAA is not augmented above the level observed with LHR A alone. However, if the two sites are in a single protein as in LHR AC, efficient DAA results.
CCPs 5-7, which are not required for ligand binding and are not formally part of a functional site, are necessary for full CP and AP convertase DAA. This suggests that a function of CCPs 5-7 is to serve as a spacer; i.e. this arrangement of active sites may optimize an interaction with C4bC3b heterodimers and C3bC3b homodimers. These studies therefore provide an explanation for the seven-CCP-long unit, which is repeated four times in CR1. In addition to facilitating the decay of the C5 convertases, binding to dimers and other multiplicatives of ligands is probably optimized by this arrangement. Indeed, efficiency of C3b dimer binding by the four polymorphic size variants of CR1 is enhanced as the copy number of site 2 increases from one to three (38). A cooperative interaction between the two functional sites of CR1 was suggested previously based on greater cofactor activity for C4b if C3b was adjacent (39).
The different requirements for decay of the CP C3 versus the CP C5 convertase were readily demonstrable. In contrast, the interpretation of results for the AP convertases was initially complicated by the simultaneous presence of both C3 and C5 convertases in a hemolytic assay designed to assess decay of the C3 convertase. The presence of the AP C5 convertase in this system was conclusively shown by developing the lytic sites with purified C5-C9 rather than with CЈ-EDTA. Because lysis occurred without the addition of C3, C5 convertases must have been present. Thus, besides monomers, C3b dimers were deposited in this assay system, and the subsequent addition of factors B, D, and P resulted in the formation of the AP C5 convertase. That AP C5 convertase can be generated without an intermediate stage of the C3 convertase has been reported TABLE VII Effect of amino acid substitutions in site 1 on decay accelerating activity for C3 convertases DAA for the CP was tested in a hemolytic assay and for the AP in a microtiter assay. Binding was defined as unchanged if the difference relative to LHR A was less than 20%. All binding data for 9a and 9b, 13a and 13b, and 6c are reported here for the first time as are iC3 binding data for 1a, 1b, 5a, 5b, 8a, and 8b. Binding results for the other proteins were reported in Refs. 24 (34). The requirement for two sites for optimal DAA for the AP C5 convertase was supported by the second assay system, in which microtiter wells were coated with C3b dimers and factor B and D were added to produce a C5 convertase. Consistent with results of the hemolytic assay, LHR AC had greater DAA than LHR A.
DAA of site 1 for the C5 convertases is much reduced compared with LHR AC or sCR1. On the other hand, LHR B or C carrying site 2 possesses no DAA for the C5 convertases. These results indicate that site 1 is the primary site for DAA, whereas the role of site 2 is probably to bind to the third subunit, C3b, of the C5 convertases. Experiments with the construct LHR AA, composed of two tandem copies of LHR A, support this concept, since its DAA for AP and CP C5 convertases is similar to LHR A. 3 Binding to C3b could destabilize the C5 convertase complex directly or facilitate the interaction between the dissociating site of LHR A and its convertase target. We are currently examining the relation between the strength of C3b binding by the carboxyl-terminal LHR and the efficiency of decay of the C5 convertases.
For optimal DAA of C5 convertases, two sites in CR1 and intervening CCPs are required. More than one site may also be required for full DAA in factor H. This is based on the observation that DAA of a factor H-like protein, composed of the initial seven CCPs, is about 100 times lower compared with full-length factor H, with its 20 CCPs and two additional C3b binding sites (40). However, DAF, with only four CCPs, has DAA for the C5 convertases equivalent to CR1. 3 One interpretation of these findings is that there is a difference in the mechanism for deactivation of convertases between DAF and CR1. Alternatively, dimers of DAF may be necessary for C5 convertase decay. A biologically active form of DAF with a molecular weight approximately twice that of DAF was described by Kinoshita et al. (41). Although it constitutes less than 10% of DAF on human erythrocytes, it exceeds 50% on orangutan erythrocytes (42), where it was shown to be a nondisulfide covalently linked homodimer (43). Also, many more C3 than C5 convertases are formed during complement activation, so only a small percentage of DAF may need to be dimerized.
Identification of Amino Acids Important for DAA for the C3 Convertases-Several amino acids including Phe 82 , Trp 7 , Ile 112 , Lys 92 , and Ser 99 are specifically required for DAA (Ref. 44; this report). Their substitution reduced DAA for one or both C3 convertase(s) without altering ligand binding. The Phe 82 substitution almost abrogated DAA for both AP and CP C3 convertases. Therefore, Phe 82 , which is not a conserved amino acid in a CCP, is a critical residue for DAA for both convertases. The Trp 7 and Ile 112 substitutions affected DAA for the AP, while Ser 99 and Lys 92 substitutions led to a reduction of DAA for the CP C3 convertase. These results indicate distinct structural requirements for regulatory activity versus ligand binding and for dissociation of the CP versus the AP.
The effects of several other mutations indicate that DAA is also a function of ligand binding. For example, substitution of Gly 35 caused a parallel reduction in DAA for both C3 convertases and for C3b and C4b binding. If Arg 64 or Tyr 94 were altered, C4b binding was reduced, as was DAA for the CP C3 convertase. Additional evidence that ligand binding is important for DAA comes from the mutant D109N, in which increased C3b and C4b binding is associated with enhanced DAA for both C3 convertases. Another example is mutant 14 with increased C3b binding and DAA for the AP C3 convertase.
In a previous report (27), the peptide sequence in site 2 homologous to regions 10 and 11 in site 1 was shown to be critical for ligand binding and cofactor activity. Here we demonstrate that peptides 10 and 11 are important for the DAA of site 1. Reduced DAA was observed after replacement of Thr 110 and Val 111 . The importance of the region 10 and 11 is further underscored by the increase in DAA for C3 convertases, above the level of sCR1, as a result of the substitution of one amino acid, D109N, or two amino acids, D109N and E116K. In addition to increased DAA, these proteins have increased C3b and C4b binding and cofactor activity (28). They are candidates for a new generation of complement inhibitors because 1) their DAA for C3 convertases not only exceeds DAA of their parental protein LHR A but also that of sCR1, 2) they are one-fourth the size of sCR1, and 3) they can be further truncated to include only CCPs 1-3. The potential also exists for improving the efficiency of DAA for the C5 convertases by modification of LHR A in LHR AC. Last, as previously suggested (24) and further emphasized by these data, homologous peptides in regions 10 and 11 are likely to be functionally important in DAF, membrane cofactor protein, C4b binding protein, and factor H.
Some of the amino acids important for DAA in site 1 are present in the homologous positions of the active site (CCPs 2-4) of DAF (45) (Fig. 6). One example is Phe 82 , whose counterpart in DAF, Phe 148 , is especially noteworthy because its substitution, along with the preceding Leu 147 , caused a substantial decrease in DAA by DAF for convertases of both pathways (46). These residues were predicted to be involved in DAA of DAF, based on a computer model in which they are a part of the groove between CCP 2 and 3 (47). Other residues important for DAA in site 1, Gly 35 , Asn 65 , Tyr 94 , Ser 99 , and Val 111 , are also found in the homologous positions of the active site of DAF, and one other, Thr 110 , is replaced with a conservative Ser. The prediction is that these amino acids are also essential for the function of DAF, serving as convertase contact points or as essential structural elements.
In conclusion, DAA for the CP and AP C3 convertases is mediated primarily by CCPs 1-3 (site 1) of LHR A. Optimal DAA for C5 convertases requires two contiguous LHRs, one carrying site 1 for DAA and the other carrying site 2 for C3b binding. The spacing between these two sites is important. Within site 1, amino acids were identified that are important solely for DAA, and others were identified that alter DAA and ligand binding in parallel. Proteins were generated in which 3 M. Krych-Goldberg, R. E. Hauhart, and J. P. Atkinson, unpublished results.
FIG. 6. Amino acids important for DAA in CR1 and conserved in DAF. Alignment of the amino acids of CCP 1 of CR1 with those of CCP 2 of DAF and of the amino acids of CCP 2 of CR1 with those of CCP 3 of DAF. The numbered amino acids represent residues in site 1 of CR1 important for DAA that are identical or conserved in the active site of DAF but not in site 2 of CR1. DAA for only one or the other C3 convertase is reduced as well as proteins with increased DAA for C3 convertases but onefourth the size of CR1. These results increase our understanding of the structure-function relationships in CR1 including a reason for its unusual highly homologous repetitive structure. They provide insights into the structural requirements for DAA of CR1 as well as of related proteins. Together with our prior analysis of ligand binding and cofactor activity in CR1, these data should facilitate the analysis of the three-dimensional structure of active sites of proteins bearing CCPs.