Binding of Cystatin C to C4

Home , Cystatin

Proc. Natl. Acad. Sci. USA Vol. 87, pp. 1288-1291, February 1990 Biochemistry Binding of cystatin C to C4: The importance of sense-antisense peptides in their interaction (cysteine-proteinase inhibitor/complement/peptide-peptide interaction) JORGE GHISO*, ESTER SABALLt, JULIANA LEONI*, AGUEDA ROSTAGNO*, AND BLAS FRANGIONE* *Department of Pathology, New York University Medical Center, New York, NY 10016; and tCatedra de Inmunobiologia, Facultad de Cs. Bioquimicas y Farmaceuticas, Universidad Nacional de Rosario, Argentina Communicated by Michael Heidelberger, December 6, 1989

ABSTRACT Hydropathic anticomplementarity of amino In addition to the inhibitory activity, the above-mentioned acids indicates that peptides derived from complementary DNA residues (almost identical in all members of the cystatin strands may form amphiphilic structures and bind one an- superfamily) may also be involved in protein-protein inter- other. By using this concept, we have found that the antisense actions. It is known that there is a tendency in the genetic peptide Ser-Tyr-Asp-Leu complementary to the segment Gln- code for codons of hydrophilic amino acids to be comple- Ile-Val-Ala-Gly (residues 55-59) in cystatin C (an inhibitor of mented by codons of hydrophobic amino acids, resulting in cysteine proteases) is located at positions 611-614 ofthe 18 chain peptide structures that might interact specifically through of human C4, the fourth component of complement. Here we amphiphilic conformations (12). It was shown that peptides describe and characterize the specific interaction between generated from the noncoding strand of DNA specifically cystatin C and C4 by ligand affinity chromatography and recognize segments coded by the complementary strand (13). ELISA. Interaction between the two native proteins was mim- This concept was successfully used in the demonstration of icked on replacement of one of them with the corresponding the interaction between different peptide pairs (14) as well as sense-antisense peptide coupled to a carrier protein, and the in the purification of several receptors, among them corti- binding was inhibited by these synthetic peptides in solution. cotropin (ACTH) (13), fibronectin (15), and angiotensin II Through the interaction with C4, cystatin C may play a (16). Practical application of this approach is restricted to regulatory role in complement activation that might be of systems in which DNA sequence information is available. We particular importance at tissue sites where both proteins are have taken advantage of our recent report on the nucleotide produced by macrophages. sequence of the gene coding for cystatin C (17); from the sequence coding for the segment Gln-Ile-Val-Ala-Gly (residues 55-59), we have deduced the complementary DNA Human cystatin C (formerly y trace) (1) is a basic protein of strand and, hence, the amino acid sequence of the antisense known primary structure (2) fully distributed in body fluids in peptide. We now present evidence of the specificity and a wide concentration range. It has been localized immuno- saturability of the interaction between cystatin C and the cytochemically in some cortical neurons, LH cells of the complementary deduced peptide Ser-Tyr-Asp-Leu and dem- adenohypophysis, pancreatic A and thyroid C cells, and onstrate that a protein containing the antisense peptide adrenal medulla (3). It is secreted into tissue culture medium (human C4, the fourth component of complement) interacts by monocytes and other cells, and the down-regulation of its with cystatin C and that this binding is inhibited by the secretion may play a role in inflammation (4). antisense peptide. Structural and genetic studies indicate that cystatin C is part of a superfamily that comprises three families, types I, II, and III (5, 6). Type I cystatins (also called stefins) are MATERIALS AND METHODS proteins of "100 residues with no disulfide bonds. Type II Proteins, Synthetic Peptides, and Antibodies. Human cys- cystatins (family that includes cystatin C) are molecules of tatin C was isolated from urine as described (18). C4 was 115-120 amino acids with two disulfide loops near the C purified from human plasma according to Gresham et al. (19) terminus. Type III cystatins (kininogens) are high molecular with modifications: 20 ml of whole plasma containing 10 mM weight proteins (Mr 68,000-110,000) composed of three cys- EDTA, 10 mM benzamidine hydrochloride, 5% L-lysine (free tatin type II-like domains (about 360 residues), the bradykinin base), and 1 mM phenylmethylsulfonyl fluoride was precip- moiety (9 amino acids), and a C-terminal polypeptide of itated with PEG 6000 (final concentration, 5%) for 1 hr at 4°C. variable length. After centrifugation at 40,000 x g for 30 min, the pellet was Cystatins possess inhibitory activity against a broad spec- discarded and the supernatant was made 21% in PEG 6000 to trum of cysteine proteinases of plant origin (papain, chymo- precipitate C3 and C4. The pellet was redisolved in 3 mM papain, ficin, and actinidin) as well as mammalian lysosomal phosphate buffer (pH 7.3) containing 6.5 mM EDTA, 6.5 mM proteases such as cathepsins B, H, and L (for review, see ref. benzamidine hydrochloride, 33 mM E-amino caproic acid, 7). The active site of inhibitory activity remains unknown, and 61.5 mM NaCl and applied to a 60-ml DEAE-Sephacel although some evidence suggests involvement of glycine at (Pharmacia) column equilibrated with the same buffer. After position 11 (8) and the segment 55-59 (cystatin C numbering) washing off the unbound material, the column was sequen- (6, 9, 10), both highly conserved in all known cystatins (7). tially eluted with 200 ml of a linear NaCI gradient (100 ml of Moreover, peptide Leu-Val-Gly (positions 9-11) inhibits 3 mM phosphate/65 mM NaCI, pH 7.3, and 100 ml of 3 mM growth of many bacteria, especially all group A streptococci, phosphate/133 mM NaCI, pH 7.3), washed with 100 ml of 3 apparently due to inhibition of a cysteine protease produced mM phosphate/133 mM NaCl, pH 7.3, and eluted again with by the bacteria (11). 200 ml of a linear NaCl gradient (100 ml of 3 mM phosphate/ 133 mM NaCl, pH 7.3, and 100 ml of 3 mM phosphate/300 mM NaCl, pH 7.3). As a final purification step, C4 was The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: BSA, bovine serum albumin. 1288 Downloaded by guest on September 24, 2021 Biochemistry: Ghiso et al. Proc. Natl. Acad. Sci. USA 87 (1990) 1289

precipitated with PEG 6000 (final concentration, 16%), redi- 0.1% BSA/0.05% Tween 20, pH 7.5) were allowed to react solved in phosphate-buffered saline [(PBS) 20 mM phos- for 1 hr at room temperature. Bound ligand was detected with phate/150 mM NaCl, pH 7.0], and stored at -70TC. Identi- the suitable antibody followed by alkaline phosphatase- fication of C4 in the column fractions was achieved by conjugated goat F(ab')2 anti-rabbit IgG. Between all steps in immunodiffusion (20) and purity was assessed by SDS/ the ELISA, wells were washed three times with Tris/ PAGE (21). BSA/Tween. The reaction was developed for 30 min with Synthetic peptides Gln-Ile-Val-Ala-Gly, Ser-Tyr-Asp-Leu, p-nitrophenyl phosphate (1 mg/ml) in diethanolamine buffer Gin-Ile-Val-Ala-Gly-Cys, Ser-Tyr-Asp-Leu-Cys, Thr-Tyr- (Bio-Rad), stopped with 0.4 M NaOH, and quantified in a Lys-Phe-Phe-Glu-Gln-Met-Gln-Asn-Cys, Asn-Trp-Cys- Microplate reader, model MR600 (Dynatech), at 410 nm. Lys-Arg-Gly-Arg-Lys-Gln, and Asp-Glu-Leu-Leu-Gln-Lys- Inhibition Assays. Synthetic peptides Ser-Tyr-Asp-Leu and Glu-Gln-Asn-Tyr-Ser-Asp were synthesized in the Center for Gln-Ile-Val-Ala-Gly as well as unrelated peptides Thr- the Analysis and Synthesis of Macromolecules (State Uni- Tyr-Lys-Phe-Phe-Glu-Gln-Met-Gln-Asn-Cys, Asn-Trp- versity of New York, Stony Brook) by solid-phase tech- Cys-Lys-Arg-Gly-Arg-Lys-Gln, and Asp-Glu-Leu-Leu- niques (22) and further purified by high-performance liquid Gln-Lys-Glu-Gln-Asn-Tyr-Ser-Asp were screened for inhi- chromatography. Their purity was ascertained by amino acid bition of binding by C4 and cystatin C. Different amounts of analysis using a Waters Pico Tag system and by Edman the synthetic peptides (0.1-10 ,ug) dissolved in 50/ul of50 mM degradation analysis on a 477A Applied Biosystems se- Tris (pH 7.4) were added to 50 ng of either cystatin C or C4 quencer. in 50 ,ul of the same buffer and incubated 1 hr at room Peptides Gln-Ile-Val-Ala-Gly-Cys, Ser-Tyr-Asp-Leu-Cys, temperature. Aliquots of each mixture were transferred into and Thr-Tyr-Lys-Phe-Phe-Glu-Gln-Met-Gln-Asn-Cys were Ser-Tyr-Asp-Leu-BSA-, C4-, cystatin C-, Gln-Ile-Val-Ala- coupled to bovine serum albumin (BSA) by way of their Gly-BSA-, unrelated peptide-BSA-, or BSA-coated wells and C-terminal cysteine, by use of the heterobifunctional reagent the bound cystatin C or C4 was measured as above (see figure m-maleimido-benzoyl-N-hydroxysuccinimide ester (23). legends for details). Anti-human cystatin C was produced in rabbits (24); rabbit anti-human C4 was purchased from Dako (Santa Barbara, CA); alkaline phosphatase-labeled goat F(ab')2 anti-rabbit RESULTS AND DISCUSSION IgG was acquired from Tago. The nucleotide and amino acid sequences of the segment Affinity Chromatography Assays. Cystatin C and BSA were 55-59 in cystatin C (the most conserved in all cystatins), the coupled to CNBr-activated Sepharose 4B (Pharmacia) at 5 nucleic acid sequence ofcomplementary DNA strand, and the mg/ml of beads according to manufacturer's instructions. amino acid sequence deduced for its protein translation prod- Two sets of experiments were performed using cystatin uct are shown in Fig. 1. By use ofthe same reading frame, the C-Sepharose and control beads of BSA-Sepharose (columns antisense peptide to the sequence Gln-Ile-Val-Ala-Gly proved of 1.5 ml each). In the first set, different amounts of C4 to be Pro-Ser-Tyr-Asp-Leu. A homology search in the Protein (0.5-2.5 mg) were allowed to interact with the affinity ma- Sequence Database [National Biomedical Research Founda- trices equilibrated in 10 mM phosphate (pH 7.4) containing tion, Georgetown, VA (26)] indicated that the peptide Ser- either 25 or 150 mM NaCl. In the second set, 1.0 mg ofC4 was Tyr-Asp-Leu was part of human C4, the fourth component of subjected to cystatin C affinity chromatography under dif- complement (positions 611-614 of the A3 chain). ferent salt concentrations (0-150 mM NaCI). In all cases the The interaction between cystatin C and C4 was studied by columns were washed with the equilibrating buffer until no means ofaffinity chromatography and ELISA. As a first step, further material absorbing at 280 nm was detected. Bound cystatin C and C4 were purified as in Materials and Methods; protein was eluted with 1 M NaCl in the starting buffer. In C4 was recovered from the DEAE-Sephacel columns at 270 both experiments, the amount of C4 was calculated using an mM NaCl. After precipitation with PEG, only one band ofMr extinction coefficient, el% , at 280 nm of 8.2 (25). 210,000 was detected in SDS/10% polyacrylamide gels, Enzyme-Linked Immunosorbent Assay (ELISA). Plastic mi- whereas, under reducing conditions, the three characteristic crotiter plates (Immulon 2; Dynatech) were coated (100 ,ul per bands of C4 (Mr 90,000, Mr 80,000, and Mr 30,000) were well, overnight at 4°C) with the different proteins in 0.1 M found. Purified cystatin C was immobilized onto Sepharose NaHCO3 buffer at pH 9.6. Coating was terminated by two 4B (1.5 mg/ml of beads) and allowed to react with purified washes of TBS (20 mM Tris/150 mM NaCl, pH 7.4) followed C4. Fig. 2 shows that the interaction follows a dose-response by blocking with 1% BSA in TBS. Serial dilutions of the relationship and that the binding is enhanced by lowering the corresponding ligands in Tris/BSA/Tween (50 mM Tris/ NaCl concentration. The amount of protein bound increased

( Cystatin C, residue* 55 - 59 )

a ... GLN ILE VAL ALA GLY ...

b 5' CAA GCC CAG ATC GTA GCT GGG GTG AAC 3

.T.

C 31 GTT CGG GTC TAG CAT CGA CCC CAC TTG 5' d ... LEU ASP TYR SER P

( Cystatin C, deduced antisense ) FIG. 1. Sense-antisense sequences. (a) Amino acid sequence of segment 55-59 of cystatin C. (b) Corresponding nucleotide sequence from cloned human cystatin C. (c) Nucleotide sequence of complementary DNA strand. (d) Antisense amino acid sequence from translation of complementary DNA strand. Downloaded by guest on September 24, 2021 1290 Biochemistry: Ghiso et al. Proc. Natl. Acad. Sci. USA 87 (1990)

I ~~~~xI _o ' 0 x ~l x

c) E

- . 0.3 0

0.1 Protein, Ag per well 2 3 FIG. 4. Binding of cystatin C to C4 and to a carrier protein mg containing antisense sequence Ser-Tyr-Asp-Leu. Microtiter plates C4 offered, were coated with variable amounts (0.01-1 ,Ag per well) of C4 (o), C-C4 interaction. Ser-Tyr-Asp-Leu-BSA (X), BSA (e), and Thr-Tyr-Lys-Phe- FIG. 2. Dose-response relationship ofcystatin Phe-Glu-Gln-Met-Gln-Asn-BSA (i), followed by sequential addition One-halfto 2.5 mg ofC4 was applied to a 1.5-ml cystatin C-Sepharose alkaline column (5 mg/ml of beads) in 10 mM phosphate (pH 7.4) containing of cystatin C (50 ng per well), anti-cystatin C (1:5000), and 150 mM NaCl (e). Bound protein was eluted with phosphatase-labeled goat F(ab')2 anti-rabbit IgG (1:2000). The reac- 25 mM NaCl (o) or tion was developed as indicated in the text. Data represent the mean 10 mM phosphate/1 M NaCl, pH 7.4. Control for nonspecific binding ±2 SD. (U) was accomplished by use of BSA-Sepharose under identical of four duplicate experiments. Vertical bars indicate conditions as above. Data represent the mean of three independent experiments. Vertical bars indicate ±2 SD. cystatin C, whereas BSA or an unrelated peptide-BSA yielded no binding activity. Ser-Tyr-Asp-Leu was coupled to almost 5-fold when the NaCl molarity was lowered from 150 BSA to avoid nonspecific interactions often observed when mM to 10 mM (Fig. 3), indicating that ionic forces are small peptides are coated onto microtiter plates (27). important in this binding. The reaction C4-cystatin C was inhibited by peptide A series ofELISA experiments was carried out to establish Ser-Tyr-Asp-Leu in solution. When cystatin C was preincu- the role of the peptides Gln-Ile-Val-Ala-Gly and Ser- bated with increasing concentrations of Ser-Tyr-Asp-Leu Tyr-Asp-Leu in the interaction between cystatin C and C4. before the assay, cystatin C binding to C4 decreased, as As shown in Fig. 4, a dose-response relationship was ob- shown in Fig. 5, corroborating the importance of the se- tained when cystatin C was added to increasing amounts of quence Ser-Tyr-Asp-Leu (positions 611-614 of the C4) for C4 or Ser-Tyr-Asp-Leu-BSA bound to the microtiter plates. the binding activity. Unrelated peptides preincubated with Thus, the antisense peptide Ser-Tyr-Asp-Leu contained in C4 cystatin C under identical conditions failed to inhibit the was able to mimic the behavior of C4 in its interaction with interaction.

0.6

c 3 .0 -0

. .

I 0.1 1 10 50 100 150 Inhibitor, ,g NaCI, mM FIG. 5. Inhibition of cystatin C-C4 binding by peptide Ser- 3. Effect of NaCl concentration on cystatin C-C4 interac- Tyr-Asp-Leu. Microtiter plates coated with C4 (100 ng per well) were FIG. room tion. One milligram ofC4 was added to a 1.5-ml cystatin C-Sepharose allowed to react with 50 ng of cystatin C preincubated (1 hr at column (5 mg/ml of beads) in 10 mM phosphate (pH 7.4) containing temperature) with 0.1-10 ,g of Ser-Tyr-Asp-Leu (o). Controls were different NaCl concentration (0-150 mM). Control for nonspecific carried out with unrelated peptides Thr-Tyr-Lys-Phe-Phe- was on a 1.5-ml BSA-Sepharose column (5 mg/ml Glu-Gln-Met-Gln-Asn-Cys, Asn-Trp-Cys-Lys-Arg-Gly-Arg-Lys- binding performed as of beads) at salt concentration as above. o, C4 bound to cystatin Gln, and Asp-Glu-Leu-Leu-Gln-Lys-Glu-Gln-Asn-Tyr-Ser-Asp Bound C was e, C4 bound to BSA-Sepharose. Data represent the inhibitors under identical conditions (e). cystatin C-Sepharose. the mean mean of three independent experiments. Vertical bars indicate ±2 evaluated as described in legend to Fig. 4. Data represent SD. of four duplicate experiments. Vertical bars indicate ±2 SD. Downloaded by guest on September 24, 2021 Biochemistry: Ghiso et al. Proc. Natl. Acad. Sci. USA 87 (1990) 1291 sites where cystatin C and the complement proteins are locally produced by macrophages (29, 30). We thank Fran Hitchcock for help with manuscript preparation. 1.0 - J.L. is the recipient of a fellowship from Consejo Nacional de Investigaciones Cientificas y Tecnicas (CONICET), Argentina. This x~~~~~~~ work was supported by National Institutes of Health Grant AR 01431-33. 0.5- ;_ ; ,- t~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1. Hochwald, G. & Thorbecke, J. (1962) Proc. Soc. Exp. Biol. Med. 109, 91-95. 2. Grubb, A. & Lofberg, H. (1982) Proc. Nail. Acad. Sci. USA 79, 3024-3027. 3. Grubb, A. & Lofberg, H. (1985) Scand. J. Clin. Lab. Invest. 45, 0.01 0.1 1 Suppl. 177, 7-14. 4. Warfel, A. H., Zucker-Franklin, D., Frangione, B. & Ghiso, J. C4, ,ug (1987) J. Exp. Med. 166, 1912-1917. 5. FIG. 6. Binding of C4 to to Muller-Esterl, W., Fritz, H., Kellerman, J., Lottspeich, F., cystatin C and BSA containing sense W. & V. FEBS Lett. 221-226. peptide Gln-Ile-Val-Ala-Gly. Microtiter were coated Machleidt, Turk, (1985) 191, plates with 6. Salversen, G., Parkes, C., Abrahamson, M., Grubb, A. & cystatin C (o), Gln-Ile-Val-Ala-Gly-BSA (X), BSA (o), and Thr- A. J. Biochem. J. 234, 429-434. Tyr-Lys-Phe-Phe-Glu-Gln-Met-Gln-Asn-BSA ng per Barrett, (1986) (a) (500 well) 7. Barrett, A. J. (1987) Trends Biochem. Sci. 12, 193-196. and allowed to react with C4 (0.01-1,g) for 1 hr at room temperature. 8. Plates were sequentially incubated anti-C4 Abrahamson, M., Ritonja, A., Brown, M., Grubb, A., with (1:3000) and alkaline Machleidt, W. & Barrett, A. J. (1987) J. Biol. Chem. 262, phosphatase-labeled goat F(ab')2 anti-rabbit IgG (1:2000). The reac- 9688-9694. tion was developed as indicated in the text. Data represent the mean of three duplicate 9. Teno, N., Tsurboi, S., Otoh, N., Okamoto, H. & Okada, Y. experiments. Vertical bars indicate ±2 SD. (1987) Biochem. Biophys. Res. Commun. 143, 749-752. 10. The peptide sense Machleidt, W., Theils, U., Laber, B., Assfalg-Machleidt, I., Gln-Ile-Val-Ala-Gly (the sequence con- Esterl, A., Wiegand, G., Kos, J., Turk, V. & Bode, W. (1989) tained in cystatin C) coupled to BSA interacted with C4 in a FEBS Lett. 243, 234-238. dose-response relationship almost identical to that of cysta- 11. Bjork, L., Akerson, P., Bohus, M., Trojnar, J., Abrahamson, tin C-C4 (Fig. 6). BSA or an unrelated peptide-BSA was used M., Olafsson, I., & Grubb, A. (1989) Nature (London) 337, as a control for nonspecific interaction. 385-386. Our results indicate that there is specific binding between 12. Blalock, J. E. & Smith, E. M. (1984) Biochem. Biophys. Res. cystatin C and C4, which contain in their sequences peptides Commun. 121, 203-207. encoded by DNA 13. Bost, K. L., Smith, E. M. & Blalock, J. E. (1985) Proc. Natl. complementary strands. These sense and Acad. Sci. USA 82, 1372-1375. antisense peptides are able to mimic or inhibit the interaction 14. Blalock, J. E., Elton, T. & Oparil, S. (1989) Biochem. J. 261, between both proteins, indicating that they play a preeminent 311-312. role in the binding activity. It will be important to establish 15. Brentani, R., Ribeiro, S., Potocnjak, P., Pasqualini, B., Loper, the effect of certain amino acid substitutions on the interac- J. D. & Nakaie, C. R. (1988) Proc. Natl. Acad. Sci. USA 85, tion, since the recognition signal could require a special 364-367. arrangement of charges and/or specific three-dimensional 16. Elton, T., Dion, L., Bost, K., Oparil, S. & Blalock, J. E. (1988) structure. Interestingly, very similar sequences can be found Proc. Natl. Acad. Sci. USA 85, 2518-2522. among proteins that are members of different interactive 17. Levy, E., Lopez-Otin, C., Ghiso, J., Geltner, D. & Frangione, B. J. Med. pathways. In this light, the residues Ser-His-Asp-Leu are (1989) Exp. 169, 1771-1778. 18. Ghiso, J., Pons-Estel, B. & Frangione, B. (1986) Biochem. present in prekallikrein and kallikrein, whereas kininogens of Biophys. Res. Commun. 136, 548-554. high and low molecular weight, also members of the cystatin 19. Gresham, H., Matthews, D. & Griffin, F. (1986) Anal. Bio- superfamily, contain the sequence Gln-Val-Val-Ala-Gly, chem. 154, 454-459. closely related to the sense peptide in cystatin C. 20. Ouchterlony, 0. (1948) Acta Pathol. Microbiol. Scand. 25, Although further studies are necessary to establish the 186-191. biological relevance of these findings, we have evidence that 21. Laemmli, U. K. (1970) Nature (London) 227, 680-685. the cystatin C-C4 interaction does indeed occur under phys- 22. Steward, J. M. & Young, J. D., eds. (1984) SolidPhase Peptide iological conditions. Preliminary studies indicate that cysta- Synthesis (Pierce, Rockford, IL). tin C significantly inhibits the lytic activity of the classical 23. Lin, F. T., Zinnecker, M., Hamaoka, T. & Katz, D. H. (1979) complement Biochemistry 18, 690-697. pathway, preventing the erythrocyte/antibody/ 24. Ghiso, J., Jensson, 0. & Frangione, B. (1986) Proc. Natl. Acad. complement 1-4 complex to bind C2 and form the C3 Sci. USA 83, 2974-2978. convertase (unpublished observations). It is widely known 25. Isenman, D. E. & Kells, D. I. (1982) Biochemistry 21, 1109- that complement activation is normally a local event occur- 1117. ring on cell membranes and antigen-antibody complexes. 26. Lipman, D. J. & Pearson, W. R. (1985) Science 227, 1435- The biologically active peptides produced during the activa- 1441. tion sequence are able to increase vascular permeability, 27. Eberle, A. N., Drozdz, R., Baumann, J. B. & Girard, J. (1989) attract and immobilize leukocytes at sites of inflammation, Peptide Res. 2, 213-220. 28. stimulate phagocytosis, and promote the formation of a Whaley, K. (1987) in Complement in Health and Disease, ed. Whaley, K. (MTP Press, Lancaster, cytolytic complex. This very powerful system is U.K.), pp. 1-36. regulated by 29. Hochwald, G. M., Pepe, A. & Thorbecke, G. J. (1967) Proc. several plasma proteins (28). The interaction described here Soc. Exp. Biol. Med. 124, 961-966. could constitute a different regulatory mechanism for com- 30. Stecher, V., Morse, H. & Thorbecke, G. J. (1967) Proc. Soc. plement activation that may be ofparticular interest at tissue Exp. Biol. Med. 124, 433-438. Downloaded by guest on September 24, 2021