The control of expression of the mouse Crry and Cr2 complement receptor genes

The molecular events controlling the expression of the mouse Crry and Cr2 complement receptor genes were investigated. The Crry gene is ubiquitously expressed and encodes a protein product of 65 kilodaltons (kDa). A 1.9 kilobase genomic fragment containing the Crry promoter and other 5' regulatory sequences was fused to a reporter gene. Deletional analysis of 5' sequences revealed an enhancer site. Gel shift and methylation interference assays localized a sequence that was capable of forming a DNA-protein complex. The sequence identified by these assays was able to enhance transcription from a heterologous promoter in a position and orientation independent manner. These experiments identified a Crry gene transcriptional enhancer. The murine Cr2 gene encodes two protein products of 145 kDa and 190 kDa that are restricted in expression to a subset of hematopoietic lineage cells. Sequencing of a genomic fragment of the Cr2 promoter-enhancer region revealed a 49 base pair region that was 70% homologous to an analogous site in the human CR2 promoter-enhancer. Gel shift and methylation interference assays revealed that a single sequence outside the homology area was responsible for the majority of the gel shift complexes. The sequence was similar to the octamer consensus sequence. The size and tissue specificity of the Cr2 gene octamer site complexes is consistent with the size and expression pattern of the Oct-1 and Oct-2 transcriptional control proteins. It was found that bacterial infection decreased both Oct-2 and Cr2 transcription. These experiments indicate that Oct-2 may play a role in B-cell specific transcription of the Cr2 gene. An early report stated that the products of the Cr2 gene were expressed on murine macrophages. It had also been suggested that these cells phagocytized complement coated particles via the 190 kDa Cr2 protein. The transcriptional activity of the Cr2 gene in mouse macrophages was investigated using reverse transcription-rapid polymerase chain reaction and indicated a lack of Cr2 gene expression in macrophages of three different derivations. Protein products of the Cr2 gene were not found on the surface of macrophages. These experiments indicate that the CR1-like activity on mouse macrophages is not due to the presence of Cr2 proteins.

THE CONTROL OF EXPRESSION OF THE MOUSE Crry AND Cr2 COMPLEMENT RECEPTOR GENES by Brian Kelly Martin A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Pathology The University of Utah December 1993 Copyright © Brian Kelly Martin 1993 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a clissertation submitted by Brian Kelly Martin This dissertation has been read by each member of the following supervisory commi ttee and by majority vote has been found to be satisfactory. l { Chair. ohn H. Weis ~~A~ /1- 16 . q.3 David A. Low THE UNIVERSITY OF UTAH GRADUATE SCHOOL FIN AL READING APPROV AL To the Graduate Council of the University of Utah: I have read the dissertation of Brian Kelly Martin in its fmal fonn and have found that (1) its format, citations and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. dlil/L ; ~H.weiS t Supervisory Committee Approved for the Major Department ~(*~1~~7 Carl R. KjeldSfg ~ Chair/Dean Approved for the Graduate Council AnnW.Han Dean of The Graduate School ABSTRACT The molecular events controlling the expression of the mouse Crry and Cr2 complement receptor genes were investigated. The Crry gene is ubiquitously expressed and encodes a protein product of 65 kilodaltons (kDa). A 1.9 kilo base genomic fragment containing the Crry promoter and other 5' regulatory sequences was fused to a reporter gene. Deletional analysis of 5' sequences revealed an enhancer site. Gel shift and methylation interference assays localized a sequence that was capable of forming a DNA­protein complex. The sequence identified by these assays was able to enhance transcription from a heterologous promoter in a position and orientation independent manner. These experiments identified a Crry gene transcriptional enhancer. The murine Cr2 gene encodes two protein products of 145 kDa and 190 kDa that are restricted in expression to a subset of hematopoietic lineage cells. Sequencing of a genomic fragment of the Cr2 promoter-enhancer region revealed a 49 base pair region that was 70% homologous to an analogous site in the human CR2 promoter-enhancer. Gel shift and methylation interference assays revealed that a single sequence outside the homology area was responsible for the majority of the gel shift complexes. The sequence was similar to the octamer consensus sequence. The size and tissue specificity of the Cr2 gene octamer site complexes is consistent with the size and expression pattern of the Oct- 1 and Oct-2 transcriptional control proteins. It was found that bacterial infection decreased both Oct-2 and Cr2 transcription. These experiments indicate that Oct-2 may playa role in B-cell specific transcription of the Cr2 gene. An early report stated that the products of the Cr2 gene were expressed on murine macrophages. It had also been suggested that these cells phagocytized complement- coated particles via the 190 kDa Cr2 protein. The transcriptional activity of the Cr2 gene in mouse macrophages was investigated using reverse transcription-rapid polymerase chain reaction and indicated a lack of Cr2 gene expression in macrophages of three different derivations. Protein products of the Cr2 gene were not found on the surface of macrophages. These experiments indicate that the CR I-like activity on mouse macrophages is not due to the presence of Cr2 proteins. v To Carol T ABLE OF CONTENTS ABSTRACT .................................................. iv LIST OF FIGURES ............................................ ix ACKNOWLEDGMENTS ........................................ xi Chapter I. INTRODUCTION ......................................... 1 The Complement Cascade ..................................... 2 Control of the Complement Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Evolution of Complement Receptors ............................. 34 Mouse Complement Receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Molecular Studies of the Expression of Complement Genes ... . . . . . . . . . . 52 Introduction to Work in This Dissertation ......................... 54 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 II. FUNCTIONAL IDENTIFICATION OF TRANSCRIPTIONAL CONTROL SEQUENCES OF THE MOUSE Cery GENE ...................... 73 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Results ................................................ 76 Discussion .............................................. 83 Acknowledgments ........................................ 85 References .............................................. 85 III. IDENTIFICATION OF SITES FOR DISTINCT DNA BINDING PROTEINS INCLUDING Oct-l AND Oct-2 IN THE Cr2 GENE ......... 87 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Results and Discussion ...................................... 89 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 IV. MURINE MACRO PHAGES LACK EXPRESSION OF THE Cr2-145 (CR2) AND Cr2-190 (CRl) GENE PRODUCTS .......... 96 Introduction ............................................. 97 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Results ................................................ 99 Discussion ............................................. I 0 I References ............................................. 102 V. DISCUSSION .......................................... 103 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Molecular Analysis of Crry Expression . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Molecular Analysis of Cr2 Expression ........................... 118 Analysis of Cr2 Expression in Murine Macrophages .................. 125 Conclusion ............................................. 132 References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , , , 134 viii LIST OF FIGURES Fi~ure ~ 1 . 1 The complement cascade ........................... . . . . . . . , . . 3 1.2 Cleavage fragments of C3 .................................... 16 1.3 Human complement protein linkage groups ........................ 37 2.1 Nucleotide sequence of the Crry promoter/enhancer region .............. 77 2.2 Determination of CITY transcription initiation sites via S 1 nuclease analysis ..................................... 78 2.3 3' deletion analysis of the Crry promoter .......................... 78 2.4 Correct initiation of transcription from the Crry-CA T fusion construct 13 ................................. 79 2.5 5' deletion analysis of the Crry promoter .......................... 79 2.6 Ligation of the terminal 341 bp of CITY sequence into construct 1500 in the reverse orientation restores enhancing activity ............... 80 2.7 Identification of a nuclear proteinlDNA complex within the Crry enhancer sequence .............................. 80 2.8 Gel shift analysis utilizing probes generated by PeR ............. ..... 81 2.9 Determination of the Crry DNA sequences involved in protein binding ................................... 82 2. 10 Gel shift assay using the minimal site defined by methylation interference analysis ............................. 82 2. 11 Gel shift assay competition with normal and mutant annealed oligonucleotides ............................... 83 2. 12 Gel shift analysis using native and mutant probes with nuclear extracts from different tissues and cell lines ................... 83 2. 13 Transcriptional enhancement of a heterologous promoter by the Crry protein binding site ................................ 84 2.14 Binding site comparison of known ets proteins and the binding site in the Crry enhancer element ... . . . . . . . . . . . . . . . . . . . . 85 3.1 DNA sequence of Cr2 5' region ................................ 90 3.2 Identification of murine Cr2 5' noncoding sequences that share homology to analogous regions of human CR2 . . . . . . . . . . . . . .. 91 3.3 Gel shift assay with the 212bp MIH box-containing fragment using B cell nuclear extracts ............................ 91 3.4 Gei shift assay comparing B cell and fibroblast nuclear extracts ........... 91 3.5 Gel shift assay using B cell extracts and subfragments of the conserved region probe ........................ 92 3.6 Gel shift a<;say of Cr2 fragment A with B cell and fibroblast nuclear extracts .............................. 92 3.7 Gel shift assay of Cr2 fragment A with B cell and T cell extracts ....... .. 92 3.8 Identification of the protein binding site within fragment A .............. 93 3.9 Comparison of the octamer sequence motif with the Cr2 octamer protein binding site ............................. 93 3. 10 Competition binding of octamer containing and mutant sequences ......... 93 3. 11 Competition binding of octamer containing and mutant sequences with fragment A .................. . . . . . . . . . . . . 93 3.12 Depression of Oct-2 and Cr2 mRNA after an acute bacterial infection ....... 94 4.1 Diagram of the Cr2-190 and Cr2-145 transcripts and the relative locations of the transcript-specific oligonucleotides .............. 99 4.2 RT-RPCR analysis of transcripts, including Cr2-145 and Cr2-190, in the macrophage cell line J774 .... . . . . . . . . . . . . . . . . . . 99 4.3 RT-RPCR analysis of transcripts, including Cr2-145 and Cr2-190, in bone marrow-derived macrophages (BMM) . . . . . . . . . . . . . . . 100 4.4 RT-RPCR analysis of transcripts, including Cr2-145 and Cr2-190, in thioglycollate-elicited macrophages (TEPM) .............. 100 4.5 RT -RPCR analysis of Cr2 transcripts using an alternate set of oligonucleotides ............................. 101 4.6 FACS analysis of the B cell line 2PK3, T cell line TK-l, the macrophage cell line J774. and TEPM for cell surface proteins including Cr2-145 and Cr2-190 ......................... 101 5. 1 Crry predicted secondary structure ............................. 108 x ACKNOWLEDGMENTS I wish to thank The American Association of Immunologists for their pennission to reprint the articles that appear in Chapter II (Martin, B. K., and J. H. Weis. 1993. Functional identification of transcriptional control sequences of the mouse Crry gene. J. Immunol. 151:857.) and Chapter III (Christensen, S. M., B. K. Martin, S. S. Tan, and 1. H. Weis. Identification of sites for distinct DNA binding proteins including Oct-l and Oct-2 in the Cr2 gene. 1. Immunol. 148:3610.). I would also like to thank VCH Verlagsgesellschaft mbH for their pennission to reprint the article that appears in Chapter N (Martin, B. K., and J. H. Weise 1993. Murine macrophages lack expression of the Cr2-145 (CR2) and Cr2-190 (CR1) gene products. Eur. J. Immunol. 23:3037.) I would like to thank my thesis advisor John H. Weis, for his never-ending support and encouragement. He seemed to always know when to provide a guiding hand and when to stand back and let me run. I would also like to thank the members of the Weis labs, both past and present, for their help over the years. I would like to thank Janis 1. Weis for her encouragement and advice. I would like to thank my parents, Robert and Charlotte Martin. They have always encoumged me to pursue my goals and have been supportive in every aspect of my life. I would especially like to thank my wife Carol and my son Isaac. Isaac has made my life an adventure and a joy. After a day of thesis writing, his arrival at home was always a pleasant excuse to stop for a while. As for Carol, I cannot say enough. She has always been supportive of me. She was always willing to help me with my experiments or with photocopying articles for this thesis. I want to thank her for being my wife, the mother of my son, and., most of all, my best friend. CHAPTER I INTRODUCTION The Complement Cascade The immune system can be functionally divided into two parts based on antigen specificity: the humoral (or specific) and innate systems. Humoral immunity is antigen specific immunity, generally mediated by antibody, while the innate system is antigen nonspecific. The complement system is a major part of the innate immune response. The frrst phenomenon attributed to the complement system was identified in 1898 when Charles Bordet found that fresh serum could mediate bacterial lysis and that this activity is heat labile (1). It was known that antibody mediated effects are not heat sensitive, therefore this system "complemented" the humoral immune system, hence the name, complement. It was subsequently demonstrated that this phenomenon is not due to one protein, but rather a system of proteins. This system is now known to be comprised of more than 25 different molecules. Early work in the complement system involved biochemical assays of function as a method of identifying components. These criteria were used to elucidate the essential factors of the human complement system. The results of these studies pointed to a pathway that involves proteolytic cleavage of serum proteins that then promote cleavage of downstream factors, resulting in a cascade effect. Because of this effect, the system is referred to as the complement cascade. The complement cascade is shown in Figure 1.1. The protein components of the system are listed in Table 1.1. The cascade can be divided into three parts: the alternative pathway, the classical pathway and the membrane attack complex. The alternative pathway The alternative (or properdin) pathway is so named because it was discovered after the classical pathway, but the alternative pathway likely predates the classical pathway in evolution (2,3). Activation of this pathway proceeds via the cleavage of C3 (see 4,5). C3 is a heterodimer composed of two disulfide-linked peptides. Cleavage of C3 results 2 Figure 1.1-The complement cascade. The alternative and classical pathways of complement activation are presented. The classical pathway proceeds via antigen­antibody complexes that activate the protease activity of the C 1 complex. The alternative pathway is initiated by either spontaneous hydrolysis of C3 to C3(H20) or by cleavage of C3 to C3b and attachment to proteins. Both pathways converge in the fonnation of a C5 convertase enzyme. This enzyme leads to fonnation of the membrane attack complex. This figure is adapted from references 6 and 7. 3 The Complement Cascade Antigen- Antibody + C 1 --.... C2 I \ I Classical Pathway , C2b C2a C4 C4b l C4b2a .,... C4b2a3b C4a C3 C3a ~ C3 ---... cx 1 Factor B Ba C3 activating surface I I I C6 C7 C8 n(C9) + + + + ~ C3b C5 ~5b ~ C5b6 ~ C5b67 ~ C5b678 ~ C5b6789n 1 \ 1 C5a Membrane C3bBb !t _ C3bBb3b Alternative Pathway Attack Complex ~ 5 TABLE 1.1 Protein Components of the Complement Cascade* Protein Size (kD) Number Function of chains Cl 900 20 Initiates the classical pathway Clq 410 18 Immunoglobulin binding domain Clr 85 1 Serine protease, cleaves CIs CIs 85 1 Serine protease, cleaves C4 and C2 C2 110 1 Serine protease, cleaves C3 as part of the classical pathway C3 convertase C3 195 2 Cleaved to form C3a (anaphylatoxin) and C3b (suface binding, convertase and opsonin) C4 210 3 Cleaved to form C4a (anaphylatoxin) and C4b (suface binding and convertase) C5 190 2 Cleaved to form C5a (anaphylatoxin) and C5b (initiates MAC formation) C6 128 1 Component of MAC C7 121 1 Component of MAC C8 155 3 Component of MAC C9 79 1 Component of MAC, polymerization in formation of pores Factor B 93 1 Serine protease, cleaves C3 as part of the alternative pathway C3 convertase Factor D 25 1 Protease, cleaves Factor B * Adaptea from 4,6,7 in the loss of a 77 amino acid peptide, C3a, and a larger fragment, C3b (C3a has important biological functions that will be discussed later). The cleavage can occur spontaneously on certain activating surfaces such as polysaccharides, fungi, bacteria, and viruses. Alternatively, it can be cleaved via complexes whose specific function is to cleave C3 (C3 convertases). This cleavage exposes a reactive thioester bond on the C3b molecule that can covalently attach to surfaces by nucleophilic attack on carboxyl or amino groups. Alternatively, C3 can spontaneously hydrolyze to form C3(H20), which exposes the reactive thioester without loss of the C3a fragment. Evidence suggests that this form of C3 is continuously generated in the fluid phase and attaches randomly to nearby surfaces (4,7). In many cases, this is the initiating event in the complement cascade. If the reactive C3 product does not covalently attach to a surface, it is rapidly inactivated by serum regulatory proteins. Surface bound C3b or C3(H20) serves as a substrate for attachment of the next protein in the cascade, Factor B. After binding, Factor B is cleaved by Factor D to Bb, with release of a smaller peptide, Ba. The complex C3bBb is the alternative pathway C3 convertase and catalyzes additional production of C3b. This is the amplification phase of the alternative pathway. C3b thus produced can be inactivated by regulatory proteins, it can covalently attach to nearby surfaces or it can complex with the convertase to form C3bBb3b, the alternative pathway C5 convertase. The classical pathway The classical pathway is dependent upon the presence of antibody (for a review see 5,8). The classical pathway initiates when the C 1 complex, which is composed of one molecule of Clq and two molecules each of Clr and CIs, binds to the Fc portion of immunoglobulin (Ig). However, it must bind to several antibodies at once, hence activation requires either multimeric Ig (IgM) or antigen-antibody complexes. The process of activating the complement system by antibody is referred to as complement 6 fixation. IgM is the most effective complement fixing antibody isotype, while the IgG isotypes IgG3, IgG I and IgG2 are less effective (lgG3 being the most effective among the group, IgG2 the least) (8-10). Other isotypes are ineffective at complement fixation. The binding of C I to antigen-antibody complexes results in activation of serine protease activity in the Clr subunit, that then cleaves CIs. This cleavage activates the serine protease activity of Cis, that can then cleave both C4 and C2. C4 is highly homologous to C3 and cleavage to C4a and C4b activates a reactive thioester that allows C4b to attach to surface proteins via nucleophilic attack. If the C4b does not attach to a surface, it is inactivated by reaction with water. There are two C4 proteins, termed C4A and C4B (5). C4A preferentially forms bonds with amino groups and C4B forms bonds mainly with hydroxyl groups. C2a that has been cleaved by Cl then associates with bound C4b to form the complex C4b2~ the classical pathway C3 convertase. This complex is able to catalyze the cleavage of C3 to C3b. C3b thus generated can be inactivated by regulatory proteins, it can bind to surface proteins in the vicinity and activate an alternative pathway cascade or it may associate with C4b2b to form C4b2b3b. This complex is the classical pathway C5 convertase. Membrane attack complex Both the alternative and classical pathway C5 convertases, although different in protein constitution, catalyze the cleavage of C5 to C5b and C5a (5), The latter has pleotropic effects that will be discussed later. C5b is loosely bound to the cell surface in association with the C5 convertase (C5 is homologous to C4 and C3, but does not contain the reactive thioester). C5b has a labile binding site specific for C6 and forms C5b6. Incorporation of C7 into the complex changes the affinity of the complex such that it has hydrophobic affinity and inserts into the mernbrane. C8 binds to the membrane inserted complex to form C5b678. This complex has even greater affmity for membrane insertion than C5b67 and is finnly anchored. The next step involves sequential binding 7 of 1 to 18 monomers ofC9 (5), This C5b6789n complex is the membrane attack complex (MAC) and is the tenninal cytotoxic component of the complement cascade. The initial MAC fonTIS a small functional channel in the membrane with an inner diameter of 30 A, but incorporation of larger numbers of C9 monomers increases the pore size (5), The formation of these pores in membranes results in the killing of some microorganisms presumably due to membrane leakage. The gram-negative bacterial genus Neisseria seems to be the group most affected by the MAC (2). Other biological functions of the complement system In addition to the fonnation of the membrane attack complex, the complement cascade has several other very important biological functions. A major function for the system is the fonnation of immune complexes. This process occurs when a foreign cell or particle becomes coated with antibody. If the antibody is of an complement fixing isotype, then the particle will become further coated with complement, a process which is referred to as opsonization. Because of the divalent nature of antibodies, large aggregates of antigen, antibody and complement can fonn. In the absence of antibody, opsonization can still occur if complement is deposited on the surface of the invading organism via activation of the alternative pathway. This activation process occurs because human cells have regulatory proteins capable of inactivating complement, whereas most invading microorganisms have no complement regulatory capacity. An additional reason that complement may be preferentially deposited on infectious cells is that Factor B has a greater affinity for membranes with a low sialic acid content (microbes) than membranes with high sialic acid (self surfaces), hence mernbranes with low sialic acid promote the formation of the alternative pathway convertases (7). The removal of immune complexes is an important process because large complexes may accumulate in the vasculature and activate complement that may then damage nearby tissues and cells. Immune complex removal can be accomplished by two different mechanisms. Fltst, phagocytic cells can 8 ingest the complexes using phagocytic complement receptors or receptors for Ig (Fc receptors). The opsonization of microbes is probably the first line of defense against invading microorganisms. The second way immune complexes can be cleared is by binding complement receptor bearing erythrocytes. The erythrocyte complex travels via the circulation to the liver where the immune complexes are stripped. This process is referred to as irr...mune clearance (11). Another major biological function of the complement cascade is the production of anaphylatoxins. Anaphylatoxins are small molecules that mediate the induction of an inflammatory response. The soluble cleavage products of C3, C4 and CS (C3a, C4a and CSa, respectively) all have these effects. These activities of these molecules were fust recognized when cleavage of C3 and CS in vitro was found to result in the production of anaphylatoxin activity (12,13). The activity of the proteins is the result of binding of the factors to receptors specific for each anaphylatoxin (14). CSa is the most potent anaphylatoxin, while C3a is 20-fold less effective and C4a is 2S00 less potent (7). Due to its relatively low activity, the relevance of C4a production in vivo is suspect. All three peptides share some sequence homology, indicating areas that are probably important for functionality (6). The activity of these anaphylatoxins is modulated by the serum exopeptidase, carboxypeptidase N 1, that catalyzes the removal of the terminal arginine (these products are referred to as "des arg" species, such as CSa des arg) (IS). The removal of this amino acid inactivates some functions of the proteins while only decreasing the activity of others (11). Many functions of CSa des arg appear to be unaffected relative to CSa, apparently because of the presence of an oligosaccharide moiety not present on C3a or C4a. This, however, may also reflect the presence of a serum regulatory protein, cochemotaxin, that enhances the functionality of CSa des arg (16). Because carboxypeptidase N 1 rapidly cleaves the anaphylatoxins in vivo, it has been hypothesized that C5a des arg stabilized by cochemotaxin allows the diffusion of a functional molecule away from the site of complement activation, setting up a gradient 9 that allows the chemoattraction of effector cells ( 17), The biological effects of C3a include the release of 5-hydroxytryptamime from platelets, induction of smooth muscle contraction and degranulation of mast cells and basophils ( 11). The granule products released include vasoactive substances such as histamine. These effects are exerted at C3a concentrations on the order of 10-8 to 10--6 molar, concentrations that are probably biological in the local microenvironment. C5a also causes smooth muscle contraction and degranulation of basophils and mast cells but C5a is greater than two orders of magnitude more potent on a molar basis than C3a (11,18). Functions specific to C5a include triggering oxidative metabolism and production of toxic products in phagocytic cells, induction of chemotactic responses in phagocytic cells and induction of adhesiveness of neutrophils (11). I 0 In addition to mediating the direct biological effects described above, the complement system also influences certain aspects of the humoral immune response. in 1971 it was discovered that IgM, and to a lesser extent IgG, are able stimulate the mouse immune response to sheep erythrocytes (19). These processes seemed to function by concentrating antigen in the spleen. Because IgM is the most effective complement fixing antibody isotype, it seemed a reasonable hypothesis that complement may have some role in this phenomenon. The specific role of complement in the humoral response was investigated by complement depletion in animals. Injection of cobra venom factor transiently reduces the circulating level of C3 to 5% of nonnallevels (20). This treatment reduced the titer and delayed the onset of the primary antibody response, while not affecting the response to thymus-independent antigens. In a subsequent study, it was demonstrated that cobra venon factor and other treatments that decreased complement proteins all suppressed the response to thymus-dependent antigens (21). These reports suggested that complement is required for the presentation of antigen to B cells in the primary antibody response. This theory was supported by the finding that depletion of C3 resulted in lack of antigen localization to splenic follicles and a decrease in the 1 1 antibody memory response (22). Complement depletion had no effect on the B cells that had already been primed. Depletion of C3 led to a decrease in the ability of IgM to mediate a thymus-dependent humoral response (23). Mice that are deficient in C5 did not have a deficient humoral response, demonstrating that it was factors acting before C5 that are important for this effect. Use of an IgM antibody that was mutated such that it was unable to fix complement efficiently showed that the process is dependent upon complement fixation. A problem with depletion of complement via cobra venom factor is that this factor acts by activating C3 very efficiently and depleting the system of complement. This effect is variable, transitory and results in the production of immune complexes and anaphylatoxins that may complicate interpretation of the results. In order to address these caveats, studies were initiated using dogs that are genetically deficient in C3. These animals were shown to have suppressed primary response to thymus-dependent antigens relative to controls and heterozygotes (24). Secondary injection of the antigen resulted in normal antibody levels, but the isotype was predominantly IgM, while in controls the secondary response isotype was mostly IgO. In contrast to C3 depletion studies, these animals also had a suppressed response to thymus-independent antigens. These results demonstrate that C3 is a critical element in the nonnal humoral response. Complement and disease The pathways described above are important in the immune response in several ways. Formation of the MAC is a critical first line of defense in the protection against invading microorganisms. In species whose membranes are resistant to disruption by the MAC, opsonization by complement, either with or without antibodies, allows these microorganisms to be cleared via phagocytosis by complement receptors. The complement system is very important in the prevention of immune complex disease. Lack or proteins responsible for these actions can have serious consequences for the 1 2 health of the individual. Most of the known disease states involving complement are inherited or spontaneous mutations in complement components, resulting in deficiency of the protein(s). Deficiencies in the proteins of the classical pathway are associated with increased susceptibly to immune complex disease, mainly systemic lupus erythematosus (lupus), Patients with these lesions also suffer from an increased susceptibility to infection by pyogenic organisms such as Streptococcus, Staphylococcus and Neisseria. A genetic lesion reSUlting in loss of a functional C I complex is correlated with the most severe disease, while loss of C4 nearly as severe (25). Lack of C2 results in a range a disease phenotypes. Most cases of nonfunctional CI are a result of deficiency in the Clq subunit; lack of Clr or CIs is very rare (25). As stated previously, there are two different genes encoding C4 proteins, therefore each individual has four C4 genes. Only 60% of the population has four functional genes, indicating that lesions at this locus are common (25). Hereditary deficiency in C2 proteins is the most common complement system lesion Caucasian populations (26). There are several types of lesions in the C2 genes. The close linkage of the C4 and C2 genes and the polymorphic nature of the MHC locus in which they are located may contribute to the relatively common deficiencies in both C2 and C4 (25). Deficiencies in factors specific to the alternative pathway are rare. Only a few cases of Factor 0 deficiency have been documented and these patients have increased susceptibility to Neisserial infections (25). Properdin is a positive regulator of the alternative pathway that acts by stabilizing the C3 and C5 convertases. Deficiency in properdin is X chromosome-linked and exhibits a range of phenotypes of which susceptibility to Neisseria infection is the most common result. These patients have no immune complex disease association (25). Lack of C3, which is a component common to both pathways, is also relatively rare in humans and results in recurrent Neisserial infections. 1 3 Lack of components of the MAC are relatively common in some populations. Nearly all of these lesions result in increased susceptibility to Neisserial infections. Lack of C5 would be assumed to have consequences in inflammation because of a lack of the C5a anaphylatoxin. Although a limited nUlnber of patients have been identified with this defect, they seem to display no inflammatory dysfunction (25). There have been identified several disease states that result from lesions in complement receptors, but these will be discussed in subsequent sections. Control of the Complement Cascade The prevention of complement-mediated damage to self cells and tissues (referred to as homologous lysis) is an important aspect of the complement system. An uncontrolled complement cascade has several effects in addition to the cellular damage created by fonnation of MAC on self cells. C4b and C3b can fonn covalent bonds to various surfaces, resulting in the production of largei mmune complex aggregates. The complexes can become large enough to result in blockage of the vasculature and become so large that they cannot be cleared. The results of the overproduction of anaphylatoxins from C3, C4, and C5 are the stimulation of the inappropriate release of granule contents and vascular leakage. As the cascade proceeds in an uncontrolled fashion, the complement component in the serum would be depleted, leading systemic lack of these protective molecules. It is therefore imperative that the pathway is carefully regulated to avoid these problems. There exist several regulatory factors that control complement deposition on self cells via the complement cascade and a list of these proteins is presented in Table 1.2. These factors can be classified as being either serum proteins or cell surface proteins. The liver is the major site of synthesis for most of the serum regulatory factors. Many of these proteins are acute phase proteins; proteins that are upregulated during an acute phase response. In addition to the proteins that regulate deposition of complement, there are cell surface receptors for complement fragments. 14 TABLE 1.2 Human Complement Regulatory Proteins Protein Serum Proteins C 1 inhibitor Properdin Factor I C4 binding protein S-protein Anaphylatoxin inactivator Sp40/40 Cell Surface Proteins DAF MCP CRI CD59 C8bp Complement Receptors CRt CR2 CR3 CR4 C3a, C4a receptors C5a receptor Adapted from 6,7 Size (in kD) 104 220 88 550 83 310 80 70 45-70 160-250 18 65 160-250 145 165,95 150,95 ? 45 Function Inhibits the serine protease activity of e1r and CIs Stabalizes the altemtive pathway C3 convertase Cleaves C3b to iC3b and C3dg inactivating it for fonnation of the convertase Decay accelerating and cofactor acti vity for C4b complexes Inhibition of MAC Cleaves the terminal Arg of the anaphyl­atoxin, inactivating them Decreases the activity of MAC Decay accelerating function for classical and alternative pathway convertases Cofactor activity for cleavage of C3b and C4b Decay accelerating and cofactor acti vities for both pathways Prevents binding of C9 to C5b-8 Prevents C8 binding to C5b-7 C3b receptor, immune clearance~ binding and phagocytic receptor, C3d receptor, uptake of antigen? iC3b receptor, binding and phagocytic receptor, adhesion iC3b receptor, binding and phagocytic receptor, adhesion Smooth muscle contraction, mast cell granule release Smooth muscle contraction, mast cell granule release, increased vascular permeability, chemotaxis These receptors are important in the control of the immune response via complement. These proteins are also presented in Table 1.2. Human serum complement regulatory proteins 1 5 Cl inhibitor. Cl inhibitor (CI-Inh) is a 110 kilodalton (kDa) protein that is a member of the serine protease inhibitor family. C 1-Inh binds to the C 1 r and Cis components of the C 1 complex with a very high affinity. The affinity of C 1-Inh is slightly higher for CIs than for Clr (27). The CI-CI-Inh complex is so stable that SDS is unable to dissociate the components (8). This binding results in the loss of enzymatic activity of both C lr and C Is. Because of the mode of inhibition, the molar ratio of C 1- Inh to CI must be 1:1 for maximumal efficiency (8). CI-Inh also binds to inactive Cl in the serum and prevents its spontaneous activation (7). Properdin. Properdin is a 220 kDa serum protein that is a positive regulator of the alternative pathway. It binds to the C3bBb convertase and stabilizes the complex to­fold, thus counteracting the activity of Factor H (4). Although this stabilization results in increased alternative pathway activity, properdin is not required for activation via the alternative pathway. Factor I. Factor I is a 88 kDa serine esterase that is involved in the proteolytic cleavage of both C4b and C3b (5,17). In order for Factor I to promote cleavage, it must have cofactors, which include C4 binding protein, complement receptor type 1, Factor H and membrane cofactor protein. The production of these cleaved fragments is a very important process in the control of the complement cascade. The cleavage ofC3b is well understood and is presented in Figure 1.2. The fIrSt cleavage results in a fragment. iC3b. that has all of the C3b sequence, but it inactive for formation of the convertase enzymes due to this single cleavage. The second cleavage, that is also mediated by Factor I, results in two fragments, the C3d,g peptide, that remains surface bound, and C3c, a soluble product. C3d,g is further cleaved by serum peptidases to C3d (the bound Figure 1.2 Cleavage fragments of C3. The structures of various cleavage fragments of C3 are presented. Products that are surface bound after cleavage are labeled as such. 1 6 The fIrSt cleavages are mediated by activating surfaces or by C3 convertases. The next two cleavage is mediated by Factor I in association with cofactor. The final cleavages are accomplished by unidentified serum proteases. This figure is adapted from references 5 and 17. C3 convertase cleav,age site I "s I Cleavage Fragments of C3 o II S - C.... Reactive thioester bond I I Surface bound R I I s I ~ o Factor I I / cleavage site H - S c=o I I R I Surface bound Factor I 0 cleavage site " H-S CI =O I I I S I iC3b I I s Surface bound I C3d,g I H -S I I R I o Tryptic I cleavage C -<?~ 1.JIr Surface bound I C3d I H -s I R I 0 I c=o I 17 I C3g1 1 8 product) and C3g. These cleavage products are ligands for various cellular receptors and mediate important biological effects (discussed in subsequent sections). C4 binding protein (C4bp). C4bp is a serum glycoprotein of 550 kDa that is synthesized in the liver. Its acts as an inhibitor of the classical pathway by binding to C4b and competitively inhibiting the subsequent binding of C2. Is also serves to dissociate the proteins forming the classical pathway C3 convertase. This activity is referred to as decay acceleration. Decay accelerating proteins generally function by binding to C3b or C4b to competitively inhibit the binding of the next factor and they also actively dissociate the prefonned convertase enzymes (7). C4bp has cofactor activity for the Factor I mediated cleavage of C4b. C4bp is an acute phase protein (28). Factor H. Factor H is a 150 kDa serum protein that has been shown to have both cofactor and decay accelerating activities for C4b and C3b. In addition to its synthesis in the liver, Factor H production has also been observed in monocytes and transformed cell lines (29). It is also a product present in platelet granules (30). Factor H is transcriptionally heterogeneous, producing three distinct transcripts in the liver (29). The largest transcript encodes the full length protein while the second largest transcript is a splice variant from the same gene that produces a truncated protein. The smallest transcript is encoded by a separate, but highly homologous gene and encodes two polypeptides (29). The functional significance of the heterogeneity of Factor H proteins is unknown. A receptor for Factor H has been demonstrated on the surface of granulocytes, monocytes and lymphocytes (31). S.protein. S-protein is also called vitronectin. It was at frrst assumed that S­protein was a normal component of the MAC, but biochemical analysis indicated that only a small fraction of MACs contained S-protein (32). It is now known that it is a 83 kDa protein that functions by association with the C5b67 complex, inhibiting the insertion of the complex into the membrane (7). This also prohibits the binding of C8 and C9 (33). 1 9 Anaphylatoxin inactivator. This enzyme binds to the soluble anaphylatoxins~ C3a. C4a and C5a. and cleaves the teoninal arginine, functionally reducing their activity (lI,I5). Sp40/40. This protein is a 80 kDa heterodimer that modulates the activity of the MAC, decreasing its ability to damage cells (7). Human membrane anchored C regulatory proteins Decay accelerating factor (DAF, CD55). This is an ubiquitously expressed membmne protein that has decay accelerating activity for both alternative and classical pathway C3 convertases. However, it has no cofactor activity. OAF is a 70 kDa protein that has four SCR regions (34). Rather than being anchored by a traditional tmnsmembrane domain, OAF is attached to the membrane via a glycosylphosphatidylinositol (GPI) linkage (35). These anchors are a common membmne linkage for cell surface proteins that do not have tmnsmembmne domains. OAF is one of the proteins known to be deficient on the surface of erythrocytes of patients with paroxysmal nocturnal hemoglobinuria (PNH). A portion of these patient's erythrocytes are abnormally sensitive to homologous lysis (36) These erythrocytes were shown to lack several GPI-anchored proteins, including OAF, and it was believed that a clonal mutation that resulted in a defect in G PI linkage was responsible for the PNH condition. This hypothesis was supported by the finding that a genetic mutation in some PNH patients results in a defective enzyme unable to form the GPI anchor (37-39) The lack of DAF on the cell surface of PNH erythrocytes is partially causal for the homologous lysis, however other GPI-inked complement regulatory proteins are also deficient in these patients. The GPI anchor is not required for functionality since chimeric OAF proteins with engineered tmnsmembmne domains are fully functional (40). Membrane cofactor protein (MCP, CD46). Due to its heterogeneous size. MCP was fust identified as gp45-70, a leukocyte membrane protein that bound on a C3 20 affinity column (41). It was found to function as a cofactor for both C4b and C3b cleavage, but it was much more efficient cofactor for C3b (42-44). It was shown to possess no decay accelerating activity. MCP has a diverse expression pattern, being expressed on platelets, leukocytes, endothelial cells, epithelial cells and fibroblasts (45- 47). The notable exception to its wide expression pattern is a lack of the protein on erythrocytes and its absence may make these cells more susceptible to lysis in PNH. The gene for MCP was cloned by screening with redundant oligonucleotides, the sequences of which were deduced by protein sequencing of the amino-terminal portion of the protein (48). The gene contains four SCRs and, unlike DAF, the MCP gene encodes a typical transmembrane domain. The size heterogeneity in the protein has been found to be due to the lack or inclusion of additional exons encoding serine, threonine, and proline rich domains that are sites for O-linked glycosylation (49). Functionality studies have shown that MCP acts primarily by preventing C3 deposition via the alternative pathway (50). Complement receptor type 1 (CRt, CD3S). CRI is a membrane protein that has several functions. It is found on B cells, erythrocytes, monocytes/macrophages. neutrophils, dendritic cells, Langerhans cells and kidney podocytes (51). CRI possesses both decay acceleration and cofactor activities for the alternative and classical pathways. It is able to bind to C3b and C4b which is bound to the cell surface and inhibit the subsequent binding of Factor B or C2, thus inhibiting the formation of the alternative and classical pathway convertases, respectively. CR 1 is a receptor for C3b and has several functions in this capacity. These receptor functions and the molecular biology of CR I will be discussed in the subsequent section. A full length CR 1 molecule has been found as a serum protein (52). This protein is fully as active as membrane bound CRl and has been hypothesized to be secreted into the serum as an additionally serum regulatory protein. However, it has not been demonstrated whether this protein is specifically secreted into the serum or that its presence in the serum is simply due to erythrocyte lysis. 2 1 with loss of a small amount of membrane CR 1. A physiological role for serum CR I has not been demonstrated. CDS9 (membrane inhibitor of reactive lysis; homologous restriction factor, 20 kDa; MAC inhibitory factor; protectin). Early reports indicated that, although OAF and MCP are important proteins in the protection of cells at the level of the C3 convertases, there are other proteins acting in the inhibition of the formation of the MAC (53). In 1989 a monoclonal antibody was identified whose binding to erythrocytes enhances homologous lysis (54). The antibody precipitated a 20 kDa protein that was found to be OPI-linked, thus it was absent from the erythrocytes of PNH patients. The protein was purified and protein sequencing followed by redundant oligonucleotide screening of a cDNA library identified a gene that did not have a high degree of homology to other complement regulatory genes (54). These data indicated the protein to be 103 amino acids and confinned that it contains no transmembrane domain. Biochemical analysis of the protein indicated that incorporation of protein from wild type erythrocytes into the membranes of PNH erythrocytes, resulted in a concentration dependent increase in protection from homologous lysis (55). The protein was found to bind C8 and C9 in the newly forming MAC, preventing the binding ofC9 molecules (36). There are reports of PNH patients who do not lack the GPI-linked proteins and one of these patients was found to have a genetic defect within both CD59 genes (56). The C059 protein is widely distributed, being expressed on at least one cell type in every tissue examined (36). These data show the importance of the CD59 protein in the protection of cells from homologous lysis. C8 binding protein (C8bp, homologous restriction factor). A 1986 report showed the presence of a 65 kDa protein on the surface of erythrocytes which acted at the level of C8 binding in formation of the MAC (57). It is a OPI-linked protein and is another complement control protein which is deficient on erythrocytes in PNH patients (5). It is expressed on erythrocytes, neutrophils, monocytes and lymphocytes (7). This protein functions by blocking the binding of C9 to the C5b678 complex, preventing MAC formation. There is no infonnation regarding the amino acid sequence of the protein or nucleotide sequence of the gene. Human complement receptors 22 CRt. CR 1 is the major cellular receptor for C3b, although it does have capacity to bind to iC3b ligands, but with lOO-fold less affinity than for C3b ligands (58). It is a glycosylated protein of heterogeneous size. It has multiple roles in complement biology, In addition to CR l's role as a complement regulatory protein, it serves as the immune complex immune complexesclearance receptor. It is the C3b phagocytic receptor on both macro phages and neutrophils where it is also an endocytic receptor for soluble complexes. CR 1 on B cells and dendritic cells may function in the uptake of antigens for presentation in the context of class II major histocapatibility complex (MHC). CR 1 was ftrSt isolated in 1979 from the membranes of hu~an erythrocytes using its decay accelerating activity as the criterion for its purification (59). Human and primate erythrocytes express less than 2000 CR 1 molecules per cell while all other cell types express greater than 20,000 molecules per cell (60-62). Such a small amount of CR 1 per cell on erythrocytes may not be enough to mediate protective effects for the erythrocytes itself. It may also be that the protein is simply too large to mediate the protective effects on the cell to which it is attached. Indeed, part of the reason that it is believed that CR }'s regulatory properties do not act upon the cell to which it is attached is that PNH erythrocytes have nonnallevels of CR 1, yet they undergo lysis (due to lack of OAF, CD59 and/or C8bp). Even though there is so little CR 1 on erythrocytes, they greatly outnumber all other leukocytes, so most of the the circulating CR 1 is present on erythrocytes (60). The main function of CR I on erythrocytes is believed to be immune clearance. The phenomenon of immune adherence was first described in 1952 (63), It was found that human erythrocytes adhered to bacteria which had been exposed to immune serum. This binding reaction could be abrogated by heating the serum, a classical indication of the involvement of the complement cascade. Immune complexes are bound by erythrocytes and transported via the circulation. The destination for this transport is the liver. In order to study this phenomenon, radiolabeled immune complexes and erythrocytes were infused into baboons and rhesus monkeys, and catheters were used to monitor the loss of erythrocytes or immune complexes in various organs (60). It was discovered that both erythrocytes and immune complexes emerged unchanged from all organs but the liver. In the liver, the complexes were stripped and erythrocytes emerged intact. The receptor that strips the complexes in the liver has not been identified, although Kupfer cells are believed to mediate the process. This process removes immune complexes from the circulation where, in the absence of immune clearance, these complexes could accumulate. Decreased immune clearance ability has been implicated in disease states. Some patients with lupus have been found to have decreased numbers of CR 1 molecules on their E when compared to normal patients (62,64). A restriction fragment length polymorphism has been found to correlate to this low expression and has been hypothesized to be a disease susceptibility indicator for lupus (64). However. this theory is a matter of some controversy (26). In addition to simple clearance of immune complexes. CR 1 also cleaves C3b in the complex. that prevents further convertase formation (65). This function prevents the cascade from forming extremely large immune complexes that would be difficult to clear. 23 After the immune adherence phenomenon for erythrocytes was described in 1952. it was discovered that human monocytes could also mediate this effect (63). In addition. however, to simply binding to the complement coated bacteria, the activated cells were then able to phagocytize them. This process was probably mediated by CR 1 and complement receptor type 3. Circulating neutrophils and monocytes have only about 5000 molecules of CR 1 on the cell surface. but stimulation of these cells by various immune reagents (such as anaphylatoxins) greatly increases the surface expression of the 24 protein (62). CRI is able to mediate both endocytosis and phagocytosis, depending on the size of the complex. Binding of soluble C3b bearing ligands results in the clustering of CR 1 to clatharin-coated pits and endocytosis (66,67). This is followed by ligand delivery to lysosonlal vesicles. It had been recognized that C3b receptors on monocytes were able to bind to large immune complexes, but it was believed that these receptors were unable to mediate phagocytosis (62). Because there is a synergistic activity bet-Neen complement-mediated immune complex binding and Fc receptor induced phagocytosis, and because CR I cocapped with Fc receptors when CR 1 bound to C3b ligands, it was believed that a function of CR 1 was to cooperate with these Fc receptors in phagocytosis by neutrophils and macrophages (68-70). However, it was discovered that activation of the phagocytic cell by phorbol esters, fibronectin or serum amyloid P component, resulted in the ability ofCRI to mediate phagocytosis (62,71,72). It appears that phosphorylation of CR 1 plays a role in the activation of the molecule to a phagocytic state, because in resting cells, CRI is not phosphorylated (73). CRI phosphorylation is induced when macrophages are activated but CR l's phosphorylation pattern is unchanged in non-phagocytic cells under the same stimulation. Part of the difficulty in assessing the function of CRI as a phagocytic receptor is that the cofactor activity of the molecule may result in rapid cleavage of C3b to iC3b, the ligand for CR3, and it may be CR3 that mediates the majority of the macrophage/neutrophils phagocytic activity (51 ). However, it is known that CR I can function as a phagocytic receptor in certain experimental systems (74,75). The function of CRI on cell types other than erythrocytes, neutrophils and monocytesl macrophages is poorly understood. Human B cells express relatively large numbers of CRI molecules per cell when compared to erythrocytes, monocytes and macrophages (76). CR 1 is expressed at varying levels throughout the B cell maturation process, with the exception that plasma cells lack CR 1 expression (77). The role of CR I on B cells is unclear, even though several studies have been designed to test this question 25 (62). There is the suggestion that CRI may promote the maturation ofB cells to antibody secreting cells (51). It has also been found that approximately 15% of peripheral T cells express CRI of unknown functionality (78). Follicular dendritic cells express CRI at relatively high levels (79). It is believed that the function of CR I on these cells may be the uptake of soluble immune complexes and subsequent presentation of these antigens in the context of class II MHC. The final major cell type that expresses CRI is the glomerular podocyte, an epithelial cell of the kidney (80). Although the function of CR I on these cells is unknown, it is interesting to note that there is a lack of CR 1 on glomerular podocytes in a lesion of lupus (62). The protein product encoded by the human CR I gene is polymorphic in the human population. The heterogeneity in the size of CR I in different people was first recognized in 1983 by Dykman et al., when it was discovered that individual blood donors had at least two different allelic subtypes of the molecule as assayed by immunoprecipitation (8l). A subsequent report demonstrated that there were four different alleles of CR 1 that differed in size by 30 kDa. The most common allelic form is 190 kDa in size and is designated the F allotype. It has a frequency of 0.83 (51). The other alleles are as follows: S allotype, 220 kDa, O.l6~ C allotype, 160 kDa, 0.01; D allotype, 250 kDa . . 00 1. Differences in glycosylation account for small differences in molecular weight. however the allelic size differences are not due to these differences (82). The determination of the molecular basis for this polymorphism required the cloning of the CRt gene. A partial cDNA clone for CR I was isolated in 1985 using redundant oligonucleotide probes derived from CRI protein tryptic sequence (83). This initial report suggested the possibility that there were domains within the CR 1 gene that shared a high degree of sequence homology. Further genetic analysis of the gene and its allotypes revealed that the CR 1 gene contained large homology sequences in addition to the standard SCR motif (84,85). These homology regions are referred to as long 26 homologous repeats (LHR). Each LHR contains 7 SCRs that each encode 60-65 amino acids. It was found that each allotype differed by the number of LHRs that they encoded. The most common allelic form (the F allotype) encodes a signal sequence for membrane insertion, 4 LHRs, 2 SCRs not within the LHRs, a transmembrane domain and cytoplasmic sequence (84). The LHRs have been given letter designations, such that the amino-terminal LHR is LHR -A, and the next LHR is LHR -B, etc. The other alleles were found to differ by either the inclusion of additional LHRs or lack of one or two LHRs relative to the F allotype form (85). In some humans, there is an additional LHR 5' of the CR 1 coding sequence. It is not known whether this CR I-like sequence is transcriptionally active (85,86). There is a high degree of homology (greater than 90% in many cases) in all LHR sequences, including introns (and including the CRl-like sequence), which indicates that the duplication events within the CR 1 gene have taken place recently in evolutionary history (86). The specific sequences responsible for the C3b and C4b binding activities of CR 1 have been identified. Deletion mapping revealed that the C3b binding site in CR I is in the amino-terminal two SCRs of both LHR-B and LHR-C and that the C4b binding site was located in the amino-teoninal two SCRs of LHR-A (87). Construction of chimeric proteins has shown that the first four SCRs of LHR -B or LHR -C are required for maximal C3b binding (58). Constructs consisting of combinations of three SCRs have markedly decreased binding efficiencies and those with only two SCRs are unable to bind C3b. Site-directed mutagenesis has indicated that there are only a small number of amino acids in the ftrSt two SCRs of LHR-A that are important for C4b binding and both SCRs are required for this function (88). The function of CR 1 is dependent on the presence of N-linked oligosaccharides. Variations in size of the mature CR 1 protein on individual cells and cell lines that are not due to allotype differences are due to differences in N-glycosylation (CRl has no 0- linked oligosaccharides) (42). Lack of glycosylation causes the CRt protein to have a turnover rate twice that of the glycosylated protein. In the absence of glycosylation both ligand binding and membrane insertion of CR 1 are markedly decreased, pointing to the importance of oligosaccharides in the function and expression of CR 1 (42). 27 Complement receptor type 2 (CR2, CD21, Epstein Barr virus [EBV] receptor). CR2 was functionally defined as the receptor for C3d. This receptor could mediate rosette formation by some B cells and monocytes with complexed C3d ligands (89-91). In the early 1970s the EB V receptor was demonstrated to be present on human B cells and was recommended as a reliable B cell marker (92,93). It was then demonstrated via immuno-flourescence that the EBV receptor colocalized with complement receptors on the surface of a transformed B cell line (94). Although this and subsequent reports suggested that the EB V receptor and CR2 may be the same protein, the connection between the two molecules was not demonstrated until 1984. CR2 was found to be identical to the antigen B2, a monoclonal antibody-defined glycoprotein of approximately 140 kDa present on the surface of B cells only at some maturation stages (95,96). Weis et al. demonstrated that a monoclonal antibody, HB-5, that recognizes an antigen with a expression pattern identical to that of the B2 antigen, also recognizes the CR2 protein (97). The HB-5 antibody plus goat-antimouse IgG F(ab')2 was demonstrated to block the ability of erythrocyte bound C3d to form rosettes with a transformed B cell line. The protein was shown to have a molecular weight of 145 kDa, which is the mostly commonly accepted size for human CR2 (97). Use of the HB-5 antibody confirmed earlier reports that CR2 is expressed on immature, mature and activated B cells but is absent from pre-B lymphocytes and plasma cells (98). The identity of CR2 as the EB V receptor was confirmed in a series of experiments by Fingeroth et al. (99). First, using four different transformed cell lines, the staining patterns with both the HB-5 and the anti-B2 antibodies were identical to the staining pattern using fluorescent labeled EBV. Second, treatment of cells with HB-5 followed by goat-antimouse F(ab'h abrogated the binding ofEBV to the cell line. Finally, CR2 bound to Staphylococcus aureus allowed the specific binding ofEBV, demonstrating CR2 to be the EBV receptor (99). A subsequent study identified an antibody that alone could inhibit the binding of EB V to B cells and could also block C3d binding to B cells, demonstrating that CR2 and the EB V receptor were either the same proteins or were in close association (1 (0). 28 The CR2 gene was cloned by two different groups. The frrst group sequenced tryptic peptides of the CR2 protein and found that the amino acid sequence was homologous to CR 1 sequence ( 101). This information indicated that CR2 gene sequences could possibly be identified by screening a tonsillar cDNA library with CR1 cDNA at low stringency. Two clones were identified that hybridized at low stringency and also hybridized with at set of redundant oligonucleotides based on peptide sequence (101). A subsequent report by this group showed that the protein is composed of a signal sequence for membrane insertion, 15 SCRs, a transmembrane domain and a relatively short cytoplasmic sequence (102). One of the cDNAs isolated contained an additional SCR, indicating that the transcript is alternatively spliced. Isolation of genomic sequences indicated areas of high internal homology, supporting the hypothesis that the gene arose by duplication events (102). The gene was independently cloned by Moore et al. using a specific oligonucleotide identified by Weis et al. (102,103). The cDNA sequence was confinned by tryptic peptide sequencing of approximately 10% of the protein. The variability in the splice pattern of CR2 was confirmed using S 1 nuclease analysis (104). In addition to this finding, the report also demonstrated that there are several alleles of CR2 within the human population. The functional significance of the alternative spicing and allelic variation has not been established. The expression pattern of CR2 has been demonstrated largely using monoclonal antibodies to the protein. Several reports demonstrated its presence on B lymphocytes (94-99), Use of these antibodies demonstrated that the expression of the protein is restricted during B cell development, being present only on immature, mature and activated B cells but being absent from pre-B lymphocytes and plasma cells (98), A C3d,g receptor was found to be expressed by in vi fro cultured macrophages, however this protein was subsequently found to be distinct from CR2 (90,105). Simply because CR2 had been demonstrated as the EB V receptor and EB V infected nasophamygeal epithelial cell, it was assumed that CR2 was present on this cell type (106J07). However, the presence of CR2 on these cells was not experimentally demonstrated until it was shown that different CR2 monoclonal antibodies were able to immunoprecipitate the CR2 protein froln the surface of two different human nasophamygeal epithelial cell lines (l08). It was also demonstrated that these cell lines express CR2 mRNA. cDNA sequence revealed that the only difference between the sequence reported from B cell 29 CR2 cDNA and epithelial cDNA is that the 5' untranslated region of the epithelial transcripts is 30-50 nucleotides longer than in the B cell transcript (108). Both malignant and nonmalignant dendritic cells have been found to express the CR2 protein (109). The expression of CR2 on antigen presenting cells such as dendritic cells and B cells, has lead to the hypothesis that CR2 may function in the uptake of antigens in an antigen­independent fashion for presentation in the context of Class II MHC (110). However. no experiments have yet been published that specifically address this theory. Although the function of CR2 has been an area of intense study, it has only been recently that the role of the protein in the immune response has become more clear. It was hypothesized in 1973 that bound C3 may serve as a stimulus for the activation of B lymphocytes (111). It was then demonstrated that various C3 fragments, in either cross­linked or soluble fonns, can cause B cell activation (112-115). But the role of CR2 in this activation process was in doubt because cloning of CR2 indicated that the cytoplasmic tail of protein is too small to encode any of the traditional types of signalling domains needed for the cellular activation (88). Numerous reports have, however, implicated CR2 in mediating enhanced B cell differentiation and proliferation. Three reports in 1985 demonstrated that cross linking of CR2 on the surface of B cells can lead 30 to the subsequent proliferation and/or differentiation of those cells (100,116,117). Furthennore, it was demonstrated that antibody to CR2 allows growth of a B lymphocyte line in suboptimal media conditions (113). These experiments led to the question of how the CR2 antibody was able to mediate these effects in the absence of signalling domains located on the molecule itself. One clue as to the possible mechanism of growth regulatory capacity of CR2 came from the demonstration that CR2 specifically interacts with p53 (118). p53 is an anti­oncogene product that has several functions in the regulation of cellular proliferation (for reviews see 119.120). Lack of a functional p53 protein, via various mechanisms, is an important event in cellular transfonnation. The association of CR2 with p53 led to the hypothesis that it was p53 that was mediating the growth regulatory events seen with anti-CR2 antibodies. however, additional information concerning the CR2/p53 complex has not been forthcoming (118). CR2 has been found to associate with additional proteins. It was discovered that antibody to the CR2 protein co-precipitates the CD 19 molecule from a transfonned B cell lines (121). At this time, CD19 became the leading candidate for the molecule that mediates CR2 signalling. CD 19 is a member of the immunoglobulin superfamily that had been characterized as being expressed throughout B cell development (122,123). It was found that ligation of the CD 19 molecule in singly ttansfected cells replicates the ability of CR2 to cause proliferation in these cells (121). It was hypothesized that CR2 is the ligand binding molecule of the CR2/CD 19 complex, and that it is CD 19 that mediates the signalling event. Other proteins in the CR2ICD 19 complex included the target of antiproliferative antibody-l and Leu-13 (124). Ligation of either of these two proteins produces biological effects similar to ligation via antibody to CR2 or CD 19. These studies have led to the hypothesis that the CR2ICD 19 complex serves a growth regulatory role on B lymphocytes. However, since these studies have all been done in vitro, the in vivo relevance of this growth regulatory capacity is not clear. Although CR2 expression has been demonstrated on the surface of dendritic cells, CD 19 was absent ( 109). This brings up the question of how CR2 is able to mediate signalling in dendritic cells, or if it needs to be able to mediate signalling in these cells. 3 I In addition to the CR2/CD 19 complex, a complex of CR l/CR2 was demonstrated to be present on the surface of B lymphocytes (125). This complex was immunoprecipitated from tonsilar B cells and from K562 cells that were cotransfected with expression vectors for the CR2 and CR 1 proteins. This CR l/CR2 association is distinct from the CR2ICD 19 complex (125). The authors theorized that the complex formed to allow ligands cleaved by the cofactor activity of CR 1 to remain cell associated via binding to CR2 in the complex. Cellular signalling could then be mediated by CR 1, the CR2/CDI9 complex andlor immunoglobulin receptor (125). Studies have not been undertaken to address the importance of the CR l/CR2 complex to the immunobiology of the complement system. Although there is unequivocal evidence that CR2 is complexed with different proteins in B cells, the defmition of the exact biological role of these interactions await further experimentation. Ligand specificity studies have indicated that CR2 has highest affinity for C3d ligands, but it does bind to C3dg with similar efficiency (58). It also has low affinity for iC3b. The presence of N-linked oligosaccharides has been found to be important for the stability and membrane expression of the protein, but not for its ligand binding function (126). Recently, CR2 has been demonstrated to be a ligand for the low affinity IgE receptor, CD23 (127). In this role, CR2 is hypothesized to have some regulatory capacity in the production of 19E. The importance of this finding to the immune response is unknown. CR2 has been unequivocally demonstrated to be a receptor for EB V. The precise peptide of EBV that mediates binding to CR2 has been identified (128). The minimal peptide identified is able to block both C3d,g and EBV binding to B cells. In addition, the peptide is able to block the proliferative response of EBV binding (128). Deletion 32 studies of CR2 showed that the first two amino terminal SCRs mediated binding to both EBV and C3d,g (129). Site-specific chimeric studies designed to identify the specific region of CR2 responsible for EB V binding have indicated that this binding site is conferred by a three dimensional conformation of the protein and can not be localized to a specific sequence (130). In addition to serving as a receptor for EBV, CR2 has also been found to mediate infection of CD4- cells by human immunodeficiency virus type 1 (131). This process was found to occur when viral particles are opsonized by complement and these opsons are taken up via CR2. This process is antibody-independent (131). The fact that at least two different viruses gain entry into cells via CR2 demonstrates that, although CR2 is important in the immune response, it also has functions detrimental to the well-being of the immune system. Complement receptor type 3 (CR3, Mac-I, CDllbCDI8). A monoclonal antibody, M 1170, developed by Beller et al. was shown to immunoprecipitate a heterodimer of 170 and 95 kDa from the sutface of macrophages, neutrophils and natural killer cells (132). This protein is the major cellular receptor responsible for the binding of iC3b ligands. In addition to its role in the binding of complement ligands, CR3 also functions as an adhesion receptor, it major ligands being intercellular adhesion molecule (lCAM)-1 and P-selectin (133). CR3 is a mernber of the integrin family of adhesion molecules (for reviews see 134,135). Each integrin consists of specific a and B subunits. It is the combination of subunits that determines the functionality of the protein. It has long been recognized that in some instances the binding of complement ligands by CR3 is not enough to stimulate phagocytic functions (91). The binding of some bacteria and yeast, however, does result in active phagocytosis and an oxidative burst by the phagocyte. Thus, it is believed that CRJ is a major phagocytic receptor for opsonized particles where its binding to ligand result<ii in a conformational change and subsequent activation of the phagocytic cells only under specific circumstances (133), CR3 is also expressed on some follicular dendritic cells. where its function is unknown (7). Deficiencies in CR3 have been described (26,136). Most of the lesions appear to be in the B subunit, either lack of production or production of a dysfunctional molecule unable to associate with the a subunit. These patients have a range of increased susceptibly to pyogenic infections correlating with the severity of the lesion. It is not known whether the disease phenotype is associated with the complement receptor or leukocyte adhesion functions of the CR3 protein. Complement receptor type 4 (CR4, CDllcCD18). This protein was demonstrated to bind to iC3b via affinity chromatography (137). CR4 is a membrane integrin composed of the same B chain as in CR3, but it uses a different a chain in the 33 heterodimer. It is expressed on monocytes, platelets and neutrophils where its function is believed to be similar to that of CR3 (7). C3a, C4a receptors. The ligands for these receptors are implicit in their names. These proteins are found on mast cells where they promotes granule release in response to ligand binding (7). On smooth muscle, the receptors cause contraction when bound to their ligands. However, due to the low efficiency in mediating these responses. the importance of the C4a receptor in the immune response is in question (11). C5a receptor. Once again the ligand is implicit in the name of this receptor. When C5a binds to this receptor n mast cells, degranulation is induced, with release of histamine and other effector proteins. On macrophages and neutrophils, C5a binding induces an oxidative burst and release of toxic mediators (11). In addition, C5a promotes the release of the inflammatory cytokine, IL-l. Its expression on endothelial cells is important for the increase in vascular permeability required to allow migration of effector cells to the site of complement activation (11). To this end, the presence of the C5a receptor on neutrophils and monocytes promotes chemotaxis. The C5a receptor binds to its ligand with extremely high affinity (approximately 10-9 molar) (11). C5a binding is able to induce the release of platelet activating molecules from many different cell types. The C5a receptor has been cloned and shows domains necessary for the interaction with GTP-binding proteins (138). It is a member of the rhodopsin superfamily of receptors. Evolution of Complement Proteins 34 The complement system of humans is better understood than the system in any other animal. However, due to the constraints on experimentation in humans (among other factors), it is advantageous that these systems be studied in other animals. It is known that there are complement-like proteins in most chordates but the regulatory proteins and receptor systems of lower animals had not been nearly so well characterized as those of humans (3,139). Hints about the function of the systems can be obtained by looking at the differences and similarities of the proteins in humans and other animals. The short consensus repeat (SeR) motif Cloning and sequencing of various complement genes revealed a common structural feature (140). Several of the cloned genes encoded a conserved repetitive element referred to as the short consensus repeat (SCR). Each SCR consists of approximately 60-65 amino acids and has conserved amino acid residues in this motif. The most important conserved residues are four cysteines. The cysteines form a one· three, two-four disulfide linkage that results in the formation of a double-loop in each SCR. Proteins that contain the SCR motif are shown in Table 1.3. These genes include members of the complement activation pathway, complement regulatory proteins and proteins not involved in the complement system. Because proteins other than complement system proteins include this motif in their structures, it has been hypothesized that the SCR can function as a structural motif without complement system function in certain proteins (6). When SCR containing proteins are visualized by electron microscopy, the protein has a "beads-on-a-string" appearance (141,142). It is believed that the function of the SCR as a structural element is to ex.tend the protein away from (he TABLE 1.3 Human SCR Containing Proteins Protein C3b- and C4b-binding proteins C4bp CRI CR2 Factor H MCP DAF Other complement proteins Factor B C2 Clr CIs Proteins unrelated to complement* IL-2 receptor ELAM-l B2-glycoprotein I Factor XIII Number of SCRs 8 23-45 15-16 20 4 4 3 3 2 2 2 6 5 10 References 143,144 80 102-104 145 48 34 146 147 148 149 150 151 152 153 *This represents only a small list of all known noncomplement proteins containing SCRs. 35 cell surface. In addition to its role as a structural element, the SCR can also function in binding of complement proteins, specifically C3b and C4b. The precise binding site for this activity has been localized in several proteins. Thus, the evolution of complement proteins has proceeded via the duplication of SCR motifs. The regulators of complement activation (RCA) locus 36 Because many of the complement proteins in humans contain the SCR motif, it was often hypothesized that these proteins were created by duplication of coding blocks of SCRs. This theory was apparently justified when it was discovered that several of thehuman complement proteins are located at a single locus on chromosome 1. An initial report presented evidence for the linkage of CR1 and C4bp (154). The same group demonstrated an additional linkage of these two genes with Factor H (155). Subsequent reports established that the CR2, DAF and MCP genes are included in this locus and this data provided a detailed genetic map of the region (156-161). The genes and their map positions are shown in Figure 1.3. All the genes are located within approximately 750 kilobases of genomic sequence. The hypothesis that complement proteins arose from a common ancestral sequence seemed to be justified as the cDNA sequences of the RCA genes were compared to genomic sequence. It has been a general finding that each SCR is coded for by a single exon and this provides a mechanism in which exons were duplicated in the genome to provide building blocks for complement proteins. Exons thus duplicated could mutate to provide additional functionality or could simply extend the size of the protein (3). There are several exceptions to the one SCR-one exon rule. These exceptions are generally split exons in which two exons encode a single SCR. It is believed that these split exons evolved from a common ancestral gene, because key sequences are conserved in all split exon SCRs. Data that support the contention that the RCA genes evolved from a common ancestral gene include the fact that all RCA genes have at least one split exon 37 Figure 1.3 Human complement protein linkage groups. The maps for the RCA locus on chromosome 1 and the MHC class III locus on chromosome 6 are presented. The relative distances are indicated by the bars. The sizes of the genes are not to scale. This figure is adapted from references 7,160,162,163. 38 Human Complement Protein Linkage Groups The RCA Locus-Human Chromosome 1 MCP CRl CR2 DAF C4bp Factor H I • • I ~ 7/~L . £J // -- ~ lOOkb 6.9cM The MHC Class III Locus-Human Chromosome 6 DRa C2 Factor B C4A C4B TNF II)l~ ~ y;;t"11 10 kb 21-hydoxylase 39 and that the intron border is precisely located at glycine 34 (162). No other SCR containing gene has been identified that has a split exon at that particular residue. This infoffiution would indicate that either gene duplications created the RCA genes or that the split exon contains activity that is common to all the RCA genes. In addition to the RCA linkage group, there is another group of complement genes located within the MHC complex on human chromosome 6. The genes in this area are Factor B, C2 and two different genes encoding C4 (C4A and C4B) (7). The map for these genes is also presented in Figure 1.3. These genes are highly polymorphic, perhaps reflecting the polymorphic nature of the MHC region (3). There is yet a thirdlinkage group containing the genes for components of the MAC: C6, C7, C8 and C9 (3). This group is also located on chromosome 1. The fact that the genes in each of these linkage groups share some functional aspect provides strong support for the hypothesis that the complement proteins evolved via duplication of common ancestral sequences. The CRI gene It appears that the evolution of the CR 1 gene is a relatively recent event. The fact that there are several alleles of CR 1 in the human population indicates that stabilizing evolution has yet fix an allele as the dominant fonn of CRI. Alternatively, it is possible that all alleles are equally effective in mediating CRl functionality. Supporting the idea that CR 1 evolution is a recent event is the fact that there is a high degree of sequence homology between LHR sequences, including intron nucleotides (86). This indicates that the LHR sequences have been recently duplicated and have yet to accumulate mutations, both in conservative manner within the coding portion of the gene and within introns. Finally, it is interesting to note that some humans have an LHR coding block 5' of the coding portion of the CRI gene, the function of which is unknown (85,86). ThIS indicates that the duplication events that have led to the heterogeneity in the size of the CR I proteins may have also produced this CR I-like duplication. The manner in which the CR I gene has recently duplicated LHR repeats to produce a larger protein has lead some to propose that there has been a selective pressure for a larger C3b binding protein (164). C3b receptors in primates 40 There has been a body of data recently on C3b receptors of primates other than humans. As stated previously, it is known that the immune clearance receptor of primates is on erythrocytes; however the molecular properties of that protein(s) were not investigated in that report (60). Gorillas have a protein of 220 kDa on the surl'ace of both erythrocytes and neutrophils that cross-reacts with both polyclonal antisera and two monoclonal antibodies against human CR 1 (165). The size of the protein corresponded to the most common allelic variant of human CR 1 and as such this protein was labeled gorilla CR 1. This group was also able to immunoprecipitate a protein believed to be gorilla MCP (165). Orangutans have a CR I-like protein with similar characteristics to human and gorilla CRI. Orangutan erythrocytes also have a'DAF-like protein (165). These data indicate that these primates possess complement receptors not unlike those of the human. Immunoprecipitation of chimpanzee erythrocytes using antibodies against human CR 1 has shown that, in addition to a protein of similar size to human CR I, chimps have a smaller protein of approximately 70 kDa (166). Cloning and sequencing of polymerase chain reaction generated cDN As for these products indicated that both the 220 and 70 kDa products originate from the same gene. The difference in size is due to alternative transcript splicing of the primary transcript (166). These data indicate a basic difference in the C3b-binding proteins of the chimpanzee relative to humans. The profile of C3b-binding proteins on baboon erythrocytes is quite unlike that of all other primates that have been studied. Baboon erythrocytes do react with monoclonal 4 1 antibodies and polyclonal antisera directed against human CR 1 (167). Immunoprecipitation indicated that there is a single molecule that reacts with these antibodies. This protein was 65 kDa, much smaller than any allelic form of CR 1 but similar in size to the smaller C3b product of chimpanzees (116,167). C3b affinity chromatography showed that this 65 kDa protein is the only C3b binding molecule on baboon erythrocytes (167). Relative phylogenetic relationships show that in terms of degree of relatedness to humans, chimpanzees are closest, then gorillas, orangutans and baboons (168). The degree of difference in the profile of C3b-binding proteins on these primates indicates an evolution from a small 65 kDa protein in baboons, to a single large protein of approximately 220 kDa in humans, gorilla and orangutans and an intermediate fonn in chimpanzees, which express both proteins. Investigation of the complement receptors in lower animals could provide additional information on the evolution of these receptors, information that will provide clues as to the functionality of these molecules. Mouse Complement Receptors Functional assays Early studies on immune clearance in rodents indicated that in these animals immune clearance is mediated by platelets, not by erythrocytes (169-172), This result showed that there is a basic difference between the complement systems of primates and rodents. Other information on the presence of mouse complement receptors must be gleaned from literature that addresses the presence of complement receptors based on functional assays. The most often studied mouse cell type for complement receptor function is the macrophage. Initial studies in 1972 identified the presence of complement receptors on the surface of mouse macrophages that mediate attachment of these cells to complement­coated immune complexes, but promote only minimal phagocytosis (173). Investigation of the ligand binding properties of this receptor or receptors indicated that C3d is not a 42 ligand for this binding and that immune complexes bind to murine macrophages via a C3b receptor(s) (174,175). It was found that, although unactivated macrophages bind immune complexes, activated cells both bound and phagocytize via a C3b receptor(s) (176). Other studies indicate that iC3b may be a major ligand for mouse complement­mediated binding of macrophages to these complexes (177). A number of signals have been identified that allow macrophages to become competent to phagocytize immune complexes via complement. Thioglycollate, specific cytokines, lipopolysaccharide. inflammatory products, and macrophage stimulating protein are all able to stimulate the phagocytosis of cells and particles via complement receptors (177-185). Certain bacteria are both bound and phagocytized via complement receptors (186-188). These studies indicate that mouse macrophages possess complement receptors for C3b ligands and that these receptors are able to bind to ligands via complement. They are also able to phagocytize them under some conditions. Ligand binding studies CR3 was identified by a rat anti-mouse monoclonal antibody, Mlno (132). This antibody recognized a integrin heterodimer on the surface of mouse macrophages. Binding of the antibody was found to inhibit macrophage rosetting mediated by iC3b. but not C3b-mediated rosetting (132). Several groups have used the M1I70 antibody in their studies of macrophage phagocytic function. Zymosan uptake by mouse peritoneal macrophages can be inhibited approximately 50% by M1I70 (189). Similarly, 65% of phagocytosis of Leshinulnia promastigotes is inhibitable by MinO (184). In these types of experiments by other groups, the M1I70 percent inhibition of complement-mediated phagocytosis ranged from 50% to 96% (179,187,188,190-192). However, one group found that Cryptococcus neoformnns phagocytosis by mouse macrophages is entirely mediated by a C3b receptor, since the MlnO antibody had no effect on this process (193), These studies indicate that CR3 on mouse macrophages is a major phagocytic receptor. However, the percent of phagocytosis that is inhibitable by M Ino plus the studies that have shown that some phagocytic function mediated by a C3b receptor, suggests the presence of a protein other than CR3. 43 Experiments demonstrating the substrate specificities of mouse cornplement receptors have indicated the presence of proteins other than CR3. As stated previously, initial reports in 1972 indicated that mouse macrophages have receptors for C3b and not C3d (173). Other experiments have shown that macrophages have both C3b and iC3b receptors. Studies using C3b-affinity chromatography of mouse spleen cells revealed the presence of three binding activities (194). The first is a protein of 190 kDa, a size consistent with the most common allelic fonn of human CR 1. Due to the size of the protein and the fact that it has cofactor activity, the protein was assumed to be the mouse homologue of CR 1 (194). The identities of the other proteins were unknown and were not pursued in that report. A subsequent report that also utilized C3b affinity chromatography of spleen cell membrane proteins demonstrated the presence of proteins of 65 and 210 kDa that bind C3b (195). The presence of the 140 kDa protein identified by Kinoshita et ala was not demonstrated. Once again it was assumed that the larger C3b binding protein was murine CRt, while the smaller protein was tenned p65. However, when antisera against human CR 1 was used to immunoprecipitate proteins from the surface of splenocytes, only the 65 kDa protein precipitated, indicating that p65 is more closely related to human CRI than the protein dubbed murine CRI (195). p65 was expressed on every cell examined in the report, including splenocytes, transfonned fibroblasts, thymocytes, lymph node cells, erythrocytes and peritoneal macrophages. It was found that the polyclonal antisera that reacted only with p65 can block C3b-mediated rosette formation by macrophages and spleen cells (195). C3d mediated rosettes fonned by splenocytes are not affected by the anti-human CRI antisera. The fact that the p65 protein was able to mediate rosette fonnation was unexpected because it was assumed that mCR 1 would be the protein that would that fonned rosettes in the mouse, not the 44 smaller p65 protein. Additional studies of these proteins made use of monoclonal antibody technology. Using C3b-affinity chromatography, Kinoshita et al. purified the 190 kDa mCR 1 protein and produced monoclonal antibodies against the molecule (196). One of the three monoclonal antibodies recognizes only the 190 kDa protein while the other two recognize both the 190 kDa protein and a 145 kDa protein. The 190 kDa protein is expressed by all peripheral B lymphocytes, macrophages and stimulated neutrophils (196). T cells, thymocytes, erythrocytes and platelets are negative for expression of the protein. Since it is known that platelets mediate immune clearance in mice, the 190 kDa protein can not be the immune clearance receptor. The relationship between the 145 kDa and 190 kDa proteins was not pursued in this report (196). The 190 and 145 kDa proteins were subsequently demonstrated to be present on the surface of dendritic cells, but absent on the specialized macrophage population, Kupfer cells (197). The basis for the interrelationship of these proteins was not elucidated until the genes for these proteins were cloned. Crry Human CR2 was cloned by screening a cDNA library with CRI sequences at low stringency (101). It was reasoned that the mouse homologues to the human complemenl receptors could be similarly cloned by screening a mouse library with CR 1 and CR2 sequences at low stringency. The approach identified two genes as homologues to the human complement receptor genes. In order to clone the mouse homologue to CRl, human CRI cDNA sequences were used as a probe on a mouse genomic library at low stringency (198). This procedure identified two genomic sequences as having sequences homologous to CR I . These genes were tenned Crry and Crry-ps (the nomenclature of the genes was changed and the terms used here represent present usage) (198). Partial sequence analysis 45 indicated that these genes code for proteins with a typical SCR motif. However, the size of the mRNA indicated that the transcript size is much smaller than transcripts for CR 1 (198). Also, Northern blots using these sequences as probes indicated that the gene has a very diverse expression pattern. Because of the small transcript size and the ubiquitous expression of the gene, it was suggested that these sequences do not represent mouse CR1, but a closely related gene. A subsequent report presented the full eDNA sequence of the Crry gene and indicated that the gene codes for a signal sequence, five SCRs, transmembrane and cytoplasmic domains (164). This product has a high degree of homology with CR 1 and it was suggested that Crry is the evolutionary homologue to CRl. with gene duplications leading to the larger size of the CR1 product (164). No suggestion was made for the mechanism involving the change in expression pattern between Crry and CRI. The Crry gene is located on mouse chromosome 1 (164). Production of polyclonal antisera against an E. coli fusion protein demonstrated that the Crry protein product is approximately 70 kDa in size. When the genomic organization of Crry was analyzed, it was discovered that gene consists of ten exons (199). A single exon codes for the signal sequence and an alternatively spliced domain, the flfSt and fifth SCRs are coded by single exons. The third and fourth SCRs are located on one exon. The second SCR, the transmembrane and cytoplasmic sequences are all coded by split exons (199). Interestingly, the split in the second SCR occurs at glycine 34, the same position in which the split occurs in all split SCRs in the human RCA locus, demonstrating Crry's relationship to that gene family (162). The Crry-ps sequence was demonstrated to be a pseudogene derived from the Crry gene based on several pieces of data (199). First sequence analysis demonstrated an 11 bp deletion relative to Crry sequence that results in a frame shift and premature termination via a stop codon. Second, there are no intronic sequences within the Crry-ps locus and the gene contains a degenerate poly-A tail. Fmally, S1-nuclease protection analysis indicated that there are no transcripts matching Crry-ps sequences (199). These reports demonstrate that Crry is 46 closely related to CR I in sequence and represents the mouse genetic homologue to human CR I. However, the Crry encoded protein has a much smaller size and a very diverse expression pattern. Because the size of the protein identified by Crry antisera is similar to the size of p65, it was suggested that the Crry gene codes for the p65 protein (164). Experimental evidence that the Crry gene encodes the p65 protein has been presented by Molina et aI. (200). Recombinant Crry expressed on the surface of K562 erythroleukemia cell line is specifically recognized by the polyclonal antisera against human CRI that recognized p65. These data positively demonstrate that the Crry gene encodes the p65 protein. Unlike the experiments by Wong and Fearon, the Crry protein expressed on the surface of K562 cells is unable to mediate rosette fonnation (195,200). This indicates that either one group's experiments were flawed or there is an intrinsic difference between Crry protein expression on the surface of macrophages and K562 cells. It was also demonstrates that the Crry protein possesses intrinsic complement regulatory activity, presumably cofactor activity (200). Monoclonal antibodies have been prepared against Crry using the recombinant protein (201). Expression of the protein was shown to coincide with the previously reported mRNA expression pattern and expression ofp65 (164,195,201). Because mRNA expression cannot be defined in erythrocytes and platelets, the use of the monoclonal antibodies was important and indicated that the Crry protein is expressed on the surface of both cell types (201). An interesting correlation can be obtained by comparing the expression patterns of CR 1 and Crry. CRt is expressed at only about 1000 molecules per erythrocyte were it functions as the immune clearance receptor (62). In other cells, CRt expression is at least an order of magnitude greater than that on erythrocytes. Similarly, Crry expression on platelets is at least an order of magnitude less than on other cell types (201). This could indicate a different functionality of the Crry protein on platelets, perhaps as the immune clearance receptor in mice. The authors of these reports stress their contention that Crry is not a complement receptor, since it does not mediate rosette fonnation (200-202). Their point 47 is that CITY functions as a complement regulator analogous to the human DAF and MCP proteins. However, there is a report that indicates that under some circumstance the Crry protein may function as a rosetting receptor (195). Regardless, experiments have not been carried out to detennine if the Crry protein has other complement receptor properties such as immune clearance and endocytic or phagocytic capacity. Cr2 The mouse homologue to human CR2 was isolated by techniques similar to the method used for isolating Crry gene sequences. Screening of a mouse cDNA library with human CR2 cDNA sequences at low stringency yielded murine sequences related to CR2 (203,204). One report showed that polyclonal antisera to human CR2 immunoprecipitated a single protein from the surface a transfonned murine B cell line (203). Isolation of cDNA sequences indicated a high degree of relatedness with the human CR2 gene and Northern experiments demonstrated a transcript of similar size to the CR2 transcript. Experiments by Kurtz et al. similarly demonstrated murine sequences closely related to CR2 gene sequences (204). However, Northern analysis of splenic mRNA revealed the presence of four transcripts of 3,5,9 and 11 kilobases. The suggestion was made that the differences in size of transcripts is due to alternative transcript splicing (204). The gene was dubbed Cr2. Both groups subsequently reported full length cDNA sequences. Kurtz et al. showed that the Cr2 coded for two proteins via alternative transcript splicing (205). The cDNA encoding the first protein contained a signal sequence, 15 SCRs, transmembrane and cytoplasmic domains. The predicted amino acid sequence and protein size were found to be extremely homologous to human CR2 (205). The second cDNA sequence encoded all sequences contained in the 15 SCR protein, but after the signal sequence. the protein contained an additional six SCRs. These six SCRs were found to be homologous to CRI in sequence and organization (205). Using sequences specific for the larger 48 transcript, they were able to demonstrate that the 5 and 11 kilobase mRNA species coded for the larger protein, and by inference the three and nine kilobase transcripts encoded the smaller protein. The 9 and 11 kilobase species were suggested to be incompletely spliced forms (205). Fingeroth found only a single cDNA species in the transformed B cell line, AJ9, that was used for the cDNA library from which CR2-related sequences were cloned (206). It is possible that the transformation process of the B cell line used by Fingeroth resulted in this cell line lacking proper splice forms of the Cr2 gene. Kurtz et al. suggested that the product of the larger transcript corresponded to the previously designated mCR 1 protein, but this was not supported by experimental data (205). The exons encoding the six amino-terminal SCRs that are present in the larger Cr2 transcript, have been located within human CR2 genomic sequences (207). The failure of the human CR2 gene to produce a 190 kDa product is the apparent result of mutations within these six SCRs that inhibit proper splicing. Support for the hypothesis that the larger CR2 gene product encodes the mCR 1 protein involved characterization of the recombinant protein with the previously defined monoclonal antibodies. This group cloned the Cr2 gene by screening a mouse cDNA library with human CR2 cDNA sequences at low stringency (208). As did Kurtz et aI., they were able to show that the Cr2 gene encodes two proteins (205,208). Molina et al. developed a fusion protein and showed that cross immunoprecipitation with the previously defined monoclonal antibody 7E9 (which recognized both the 190 and 145 kDa proteins) indicated that the Cr2 gene sequences code for the proteins that had been previously defined as mCRI and mCR2 (208). Additionally, Molina et al. have shown that the protein encoded by the larger transcript when expressed on the surface of K562 cells, is specifically recognized by all monoclonal antibodies against the 190 kDa protein (200). This recombinant protein functions as a rosetting receptor. These experiments unequivocally demonstrate that the Cr2 gene encodes the 190 and 145 kDa proteins. Many reports have referred to the 190 and 145 kDa proteins as mCR 1 and mCR2. 49 respectively (196,200,208,). However, this nomenclature does not take into account the genetic relationships among the proteins, and other manuscripts have referred to the proteins as Cr2-190 and Cr2-145 (205,209). This thesis will use the latter convention. The designation of the 190 kDa protein as mCR 1 suggests that it is derived from CR 1 related sequences~ but it is c1ear that it is the Crry gene which is most c10sely related to CR 1. Should the cluster designation nonlenclature be assigned to these proteins, it is clear that the products of the Cr2 gene should be referred to as C021 a and C021 b, while the mouse protein should be referred to as CD35. Cluster designations refer to the ability of different antibodies to precipitate a given protein. Antisera against CR 1 precipitates the Crry/p65 protein from murine cells (195,200). An immunoprecipitation has not been performed with antisera to CR2 with cells expressing both Cr2 proteins, but it does precipitate Cr2-145 from the surface of a cell line that expresses only that protein (203). Since all antibodies prepared against Cr2-145 recognize Cr2-190, human CR2 antisera should precipitate both proteins from mouse spleen cells. This dissertation makes the suggestion that the mouse Crry/p65 protein be given the designation C035, whereas the products of the Cr2-145 and Cr2-190 proteins be designated C021 a and C021 b, respectively. The expression pattern of the Cr2 proteins has been delineated by antibody staining and molecular techniques. As mentioned previously, studies from the laboratory of Kinoshita have indicated that the Cr2 proteins are expressed on B cells, dendritic cells. activated neutrophils and peritoneal macrophages (196,197). No Cr2 expression was observed on T cells, thymocytes, platelets, erythrocytes and Kupfer cells. In no case has there been observed expression of one protein in the absence of the other, but this is somewhat difficult to access due to the cross reactivity of all of the antibodies that react with Cr2-145, with Cr2-190. These reports indicate that Cr2 gene expression in the mouse most closely resembles that of CR 1 in the human. Molecular analysis using a polymerase chain reaction (PCR) approach that is able to differentiate the two forms at the 50 transcript level has conclusively shown that there is no expression of one transcript in the absence of the other (210). Also, this analysis has indicated high dose microbial challenge results in a decrease of Cr2 transcription in splenocytes (209). Because the Cr2-190 protein contains all of the sequence of the Cr2-145 protein, it was hypothesized that the larger protein may contain all of the ligand-binding capacity of the smaller protein, plus ligand-binding capacity specific to its unique sequence. The theory was confmned by studies by Pramoonjago et ale (211). Cr2-190 was found to bind to both C3b and C3d with similar efficiencies, while binding to iC3b with 5 fold less affinity. Cr2-145 bound to C3d and iC3b with less affinity, but did not bind to C3b (211). These studies demonstrate that the products of the Cr2 gene have similar ligand specificities for C3d and iC3b, while the C3b-binding activity is localized to the amino­tenninal6 SCRs in Cr2-190. Other than the ligand-binding function, functionality of the Cr2 proteins has been demonstrated in only two reports. As stated previously, the Cr2-145 protein functions as a C3d-specific rosetting receptor while the Cr2-190 protein functions in both C3b- and C3d-mediated rosette formation (200). However, what function this rosetting may have in the immunobiology of complement is unclear. It may be the case that the binding complement ligands allows phagocytosis via a second signal or it may be mediated by another receptor (such as an Fc receptor). Neither the Cr2-145 nor Cr2-190 proteins had the ability to protect the cells on which they were located from complement mediated damage (200). Demonstration of in vivo functionality made use of intravenous injection of three Cr2 protein monoclonal antibodies (212). Injection of two monoclonal antibodies that do not block the C3d-binding ability of the Cr2 proteins had little effect on the primary response to antigen injection 24 hours after infusion of the monoclonals. Injection of the monoclonal antibody that recognized both Cr2-145 and Cr2-190 and blocks C3d-binding, was able to block C3d mediated rosette formation (212). This resulted in nearly total inhibition of the primary antibody response. These data suggest 5 1 that the C3d binding capacity of the Cr2 proteins may be important for the initial uptake of antigen in antigen presenting cells. The cells then present the antigen to T cells. The T lymphocytes produce the necessary cytokines that stimulate appropriate B cells, leading to the antibody production characteristic of the primary response. The RCA locus in the mouse Several mouse proteins related to the proteins of the human RCA locus have been characterized. Screening of a mouse liver cDNA library with human Factor H sequences resulted in the isolation of the murine homologue (213). The mouse gene contains a signal sequence and 20 SCRs (the same number as are in the human gene), The murine C4bp gene was similarly isolated (214). This gene contains a signal and six SCRs and has extensive homology between the mouse and human sequences. As discussed previously, the Crry gene represents the genetic homologue to human the CRI gene while the Cr2 gene is the homologue of the CR2 gene. A study was initiated in order detennine the genetic relationship between the RCA locus on human chromosome 1 and the mouse genes. Results indicated that mouse chromosome 1 encodes all four identified mouse proteins, but in a very different arrangement from that in the human (163). Identification of mouse markers homologous to human markers indicated that the mouse chromosome 1 has undergone a translocation or inversion with the breakpoint in the middle of the RCA locus . .\10st markers lined up in appropriate positions relative to human chromosome 1, but the Crry and Cr2 genes have been removed to a distal position relative to C4bp and Factor H (163). This event is not specific for inbred strains of mice, since the M. spretus species has the same genetic arrangement of this locus. This proposed translocation or inversion seems to have had another effect. Despite numerous attempts, researchers have been unable to locate mouse sequence homologous to the DAP and MCP genes (202). It is possible that translocation or inversion actually resulted in the deletion of the mouse genetic homologues to DAP and MCP. There has been a report 52 on a mouse protein with OAF functionality, but information such as protein andlor gene sequences have not been forthcoming (215). The lack of these regulatory proteins has been part of the reason for referring to Crry as a complement regulator, with functions analogous to OAF and MCP (200,201). The lack of OAF and MCP may have placed a selective pressure on the Crry gene to become a ubiquitously expressed protein and provide a regulatory role analogous to the two deleted proteins. Alternatively, it is possible that the DAF and MCP genes did not evolve until after the evolutionary split between mice and humans. This would suggest that the Crry gene has had a diverse expression pattern throughout evolution and it changed to a restricted pattern when the sequences were duplicated in creating the CRI gene. However, this is an unlikely possibility because if this were the case, then OAF and MCP genes should be more closely related to the progenitor gene, CR 1, than the Crry gene. Sequence data show that the Crry and CRI genes are much more closely related than CRI to OAF and MCP, indicating that the first hypothesis is more plausible. In order to distinguish between these possibilities, the Crry, OAF and MCP related sequences in species closely related to the mouse should be cloned and sequences for comparison. Molecular Studies of the Expression of Complement Genes The molecular events controlling the expression of five complement related genes have been studied. The 5' untranslated region of the C4bp a gene has consensus regulatory sequences, but the functionality of these elements was not tested (216). Factor H has been shown to be responsive to retinoic acid and the specific nucleotides in the promoter of the gene that mediate this responsiveness have been delineated (217). The promoter of the CR2 gene has been characterized by two different groups. Rayhel et al. demonstrated the presence of sequences homologous to known regulatory elements but did not demonstrate functionality of these sequences (218). Yang et al. did these same analyses, but included deletional analysis of the CR2 promoter (219). It was 53 demonstrated that relatively short sequences of the promoter were necessary for transcription, with subtle regulatory sequences upstream. However, these short constructs were expressed equally well in cell lines that did not transcribe CR2, indicating that these constructs lack tissue-specific regulatory elements. Cloning of the 5' genomic sequences of the DAF gene revealed no consensus TATA or CAAT elements (220). This is characteristic of the organization of housekeeping genes. The fusion of these upstream region to a reporter gene and subsequent deletion analysis indicated that the region from 2.8 to 0.8 kilobases contained no regulatory elements for the cell lines tested. A single enhancer element active only in the EBV cell line was identified in the -794 to -206 region. The remaining sequence was found to contain the majority of transcriptional regulatory activity. This area is GC rich, again a feature of housekeeping genes. Another report on the OAF promoter suggested that 99% of the transcriptional activity in COS cells resides within the -200 to -77 region (221). The reason for the discrepancy may be the fact that this report used the COS cell line to assay the human OAF promoter and this promoter may have species restriction. The 5' region of the MCP gene also has no consensus TAT A motif and is GC rich (222). This region was fused to a reporter gene and assayed for expression. The sequence required for promoter activity was localized to within 624 nucleotides upstream of the initiation site and an inhibitory sequence was identified in the region from -1204 to -624. In addition to this analysis of the MCP gene, the 5' region of an MCP-like sequence was analyzed and found to have similar promoter activity. These reports indicate that the ubiquitously expressed proteins OAF and MCP both have promoter characteristic of housekeeping genes and this arrangement may contribute to their diverse expression patterns. Introduction to Work in This Dissertation The work presented in this dissertation involves the characterization of the molecular events controlling the expression of the mouse complement receptor genes, Crry and Cr2. The Crry gene has a diverse expression pattern that may reflect its function a~ a complement regulatory protein. Elucidation of the mechanism involved in the transcriptional control of Crry may give clues as to the possible mechanism for the change in the expression of Crry versus CRl. Similarly, investigation of CR2 expression may provide evidence for the change, if any, in the transcriptional control sequences of the mouse versus human genes. The ability to modulate the expression of Crry and Cr2 could provide clues for possible modulation of CR 1 expression in the human, that may aid in controlling immune complex disease. 54 In order to pursue these goals, Crry promoter-enhancer sequences were fused to a reporter gene, chloramphenicol acetyltransferase. These constructs were evaluated for their ability to promote reporter gene expression in fibroblast cells. 5' deletion analysis indicated the presence of a Crry enhancer sequence. Gel shift and methylation interference analyses were used to determine the specific protein-DNA contacts in the enhancer region. These analysis identified a minimal sequence which was able to enhance expression of a heterologous promoter. The core of this sequence contained a GGAA sequence, the common core DNA-binding sequence for the ets family of proto­oncogenes. These data indicate that Crry expression may be controlled, in part, by a protein related to the ets family of transcriptional regulators. The study of expression of the Cr2 gene was prompted by the observation that there was an area of high homology in the promoters of the Cr2 and CR2 genes. This area was subjected to gel shift and methylation interference analysis using B cell nuclear extracts. These experiments demonstrated that the major DNA-protein complex in this region occurred not in the homology sequence, but in an area just 5' of this site. This sequence contained nucleotides corresponding to an octamer motif. This sequence is 55 responsible for the B cell specific transcription of several promoters and enhancers via the transcription factor Oct-2 It was demonstrated that in vivo injection of Listeria monocytogenes resulted in a decrease of Oct -2 transcription and a corresponding decrease in Cr2 transcription. These data indicate that Cr-2 transcription may be controlled, in part, by the Oct-2 transcription factor. As it was previously demonstrated that the Cr2-190 protein was expressed on the surface of peritoneal macrophages, the mechanism of transcription in macrophages was compared to that in B cells. Investigation of transcription of the Cr2 gene products revealed no expression in three different macrophage cell types. Analysis of protein expression on the surface of these macro phages indicated that neither the Cr2-145 nor Cr2-190 proteins were expressed on the cell surface. These data indicate that the previous report on expression of Cr2-190 on the surface of murine macrophages was in error. This suggests that the C3b binding activity on the macrophage cell surface must be due to the presence of another protein(s). References 1. Barrett, J. T. 1988. In Textbook of Immunology. The C. V. Mosby Company. St. Louis, Missouri. pp. 11. 2. Kinoshit~ T. 1991. Biology of the complement system: the overature. Immunol. Today. 12:291. 3. Farries, T. C., and 1. P. Atkinson. 1991. Evolution of the complement system. Immunol. Today. 12:295. 4. Pangburn, M. K. 1983. Activation of complement via the alternative pathway. Federation Proc. 42:139. 5. Muller-Eberhard, H. 1988. Molecular organization and function of the complement system. Ann. Rev. Biochem. 57:321. 6. Campbell, R. D., S. K. A. Law, K. B. M. Reid, and R. B. Simm. 1988. Structure, organization, and regulation of the complement receptor genes. Ann. Rev. Immunol. 6:161. 7. Abbas, A. K., A. H. Lichtman, and J. S. Porter. 1991. In Cellular and Molecular Immunology. W. B. Sanders Company. pp.259-282. 8. Cooper, N. R. 1985. The classical complement pathway: Activation and regulation of the first complement component. Adv. Immunol. 37: 151. 9. Ishizak~ T., K. Ishizaka, T. Borsos, and H. Rapp. 1966. C'1 fixation by human isoagglutinins: Fixation ofC'1 "(-G and "(-M but not by "(-A antibody. 1. Immunol. 97: 716. 10. Augener, W., H. M. Grey, N. R. Cooper, and H. J. Muller-Eberhard. 1971. The reaction of monomeric and aggregated immunoglobulins with C 1. Immunochem. 8:1011. 11. Frank, M. M., and L. F. Fries. The role of complement in inflammation and phagocytosis. Immunol. Today. 12:322. 12. Cochrane, C. G., and H. J. MUller-Eberhard. 1968. The derivation of two distinct anaphylatoxin activities from the third and fifth components of human complement. 1. Exp. Med. 127:371. 13. Gorski, J. P., T. E. Hugle, and H. J. MUller-Eberhard. 1979. C4a: the third anaphylatoxin of the human complement system. Proc. Nat/. Acad. Sci. USA. 76:5399. 14. Glovsky, M. M., T. E. Hugle, T. Ishizaka, L. M. Lichtenstein, and B. W. Erickson. 1979. Anaphylatoxin-induced histamine release with human leukocytes. 1. Clin. Invest. 64:804. 56 15. Ward, P. A., and J. Ozols. 1976. Characterization of the protease activity of the chemotactic factor inactivator. J. Clin. Invest. 58:123. 16. Perez, H. D., I. M. Goldstein, R. O. Webster, and P. M. Henson. 1981. Enhancement of the chemotactic activity of human C5a des arg by an anionic polypeptide (" cochemotaxin If) in normal serum and plasma. J. Immunol. 126:800. 17. Fearon, D. T., and W. W. Wong. 1983. Complement ligand-receptor interactions that mediate biological responses. Ann. Rev. Immunol. /:243. 57 18. Fernandez, H. N., P. M. Henson, A. Otani, and T. E. Hugli. 1978. Chemotactic response to humrul C3a and C5a anaphylatoxins. J. Imntunol. 120: 109. 19. Dennert, G. 1971. The mechanism of antibody-induced stimulation and inhibition of the immune response. J.Immunol. 106:951. 20. Pepys, M. B. t 972. Role of complement in induction of the allergic response. Nature New Bioi. 237: 157. 21. Pepys, M. B. 1974. Role of complement in induction of antibody production in vivo. Effect of cobra factor and other C3-reactive agents on thymus-dependent and thymus-independent antibody responses. J. Exp. Med. 140: 126. 22. Klaus, G. G. B., and 1. H. Humphrey. 1977. The generation of memory cells: the role ofC3 in the generation of memory B cells. Immunology. 33:31. 23. Heyman, B., L. Pilstrom, and M. 1. Shulman. 1988. Complement activation is required for IgM-mediated enhancement of the antibody response. J. Exp. Med. 167:1999. 24. O'Neil, K. M., H. D. Ochs, S. R. Heller, L. C. Cork, 1. M. Morris and 1. A. Winkelstein. 1988. Role of C3 in humoral immunity. Defective antibody production in C3 deficient dogs. J. lmmunol. 140:1939. 25. Morgan, B. P., and M. 1. Walport. 1991. Complement deficiency and disease. lmmunol. Today. 12:301. 26. Colten, H. R., and F. S. Rosen. 1992. Complement Deficiencies. Ann. Rev. lmmunol. 10:809. 27. Reid, K. B. M., and R. R. Porter. 1981. The proteolytic activation systems of complement. Ann. Rev. Biochem. 50:433. 28. Saeki, T., S. Hirose, M. Nukatsuka, Y. Kusunoki, and S. Nagasawa. 1989. Evidence that C4b-binding protein is an acute phase protein. Biochem. Biophys. Res. Comm. 164: 1446. 29. Estaller, C., V. Kiostinen, W. Schwaeble, M. P. Dierich, and E. H. Weiss. 1991. Cloning of the 1.4-kb mRNA species of human complement factor H reveals a novel member of the short consensus repeat family related to the carbo=< y terminal of the classical 150-kOa molecule. J. Immunol. 146:3190. 30. Devine, D. V., and W. F. Ross. 1987. Regulation of the activity of platelet­bound C3 convertase of the alternative pathway of complement by platelet factor H. Proc. Natl. Acad. Sci. USA. 84:5873. 58 31. Schmitt, M., H. Mussel, K. P. Hammann, O. Scheiner, and M. P. Dierich. 1981. Role of 61 H for the binding of C3b-coated particle to human lymphoid and phagocytic cells. Eur. 1. lmmunol. 11: 739. 32. Bhakdi, S., R. Kaflein, T. S. Halstensen, F. Hugo, K. T. Preissner, and T. E. Mollnes. 1988. Complement S-protein (vitronectin) is associated with cytolytic membrane-bound C5b-9 complexes. Clin. Exp. lmmunol. 74:459. 33. Dahlback, B., and E. R. Podack. 1985. Characterization of human S protein, an inhibitor of the membrane attack complex of complement. Demonstration of a free reactive thiol group. Biochem. 24:2368. 34. Caras, I. W., M. A. Davitz, L. Rhee, G. Weddell, D. W. Martin, Jr., and V. Nussenzweig. 1987. cDNA cloning of decay accelerating factor indicates novel use of splicing to generate two protein fonns. Nature. 325:545. 35. Medof, M. E., E. I. Walter, W. L. Roberts, R. Haas, and T. L. Rosenberry. 1986. Decay accelerating factor of complement is anchored to cells by a C­terminal glycolipid. Biochem. 25:6740. 36. Lachmann, P. J. 1991. The control of homologous lysis. lmmunol. Today. 12:3125. 37. Miyata, T., J. Takeda, Y. Hda, N. Yamada, N. Inoue, M. Takahashi, K. Maeda, T. Kitani, and T. Kinoshita. 1993. The cloning of PIG-A, a component in the early step of GPI-anchor biosynthesis. Science. 259:1318. 38. Takahashi, M., J. Takeda, S. Hirose, R. Hyman, N. Inoue, T. Inoue, T. Miyata, E. Ueda, T. Kitani, M. E. Medof, and T. Kinoshita. 1993. Deficient biosynthesis of N-acetylglucosaminyl-phosphatidylinositol, the fust intennediate of glycosyl phosphatidylinositol anchor biosynthesis, in cell lines established from patients with paroxysmal nocturnal hemoglobinuria. 1. Exp. Med. 177:517. 39. Takeda, J., T. Miyata, K. Kawagoe, Y. Iida, Y. Endo, T. Fujita, M. Takahashi, T. Kitani, and T. Kinoshita. 1993. Deficiency of the GPI anchor caused by a somatic mutation of the PIG-A gene in paroxysmal nocturnal hemoglobinuria. Cell. 73:703. 40. Lublin, D. M., and K. E. Coyne. 1991. Phospholipid-anchored and transmembrane versions of either decay-accelemting factor or membrane cofactor protein show equal efficiency in protection from complement-mediated cell damage. 1. Exp. Med. 173:35. 41. Cole, J. L., G. A. Housley, Jr., T. R. Dykman, R. P. MacDermott, and J. P. Atkinson. 1985. Proc. Natl. Acad. Sci. USA. 82:859. 42. Lublin, D. M., R. C. Griffith, and J. P. Atkinson. 1986. Influence of glycosylation on allelic and cell-specific Mr variation, receptor processing, and ligand binding of the human complement C3b/C4b receptor. J. Bioi. Chem. 261:5736. 43. Oglesby, T. J., C. J. Allen, M. K. Liszewski, D. J. G. White, and J. P. Atkinson. 1992. Membrane cofactor protein (CD46) protects cells from complement-mediated attack by an intrinsic mechanism. 1. Exp. Med. 175:1547. 59 44. Seya, T., T. Ham, M. Matsumoto, Y. Sugita, and H. Akedo. 1990. Complement-mediated tumor cell damage induced by antibodies against membrane cofactor protein (MCP, CD46). J. Exp. Med. 172: 1673. 45. McNeary, T., L. Ballard, T. Seya, and 1. P. Atkinson. 1989. Membrane cofactor protein of complement is present on human fibroblast, epithelial, and endothelial cells. J. Clin. Invest. 84:538. 46. Seya, T., J. R. Turner, and 1. P. Atkinson. 1986. Purification and characterization of a membrane protein (gp45-70) that is a cofactor for cleavage of C3b and C4b. J. Exp. Med. 163:837. 47. Seya, T., L. L. Ballard, N. S. Bora, V. Kumar, W. Cur, and 1. P. Atkinson. 1988. Distribution of membrane cofactor protein of complement on human peripheral blood cells. An altered form is found on granulocytes. Eur. J. Immunol. 18:1289. 48. Lublin, D. M., M. K. Liszewski, T. W. Post, M. A. Arce, M. M. Le Beau, M. B. Rebentisch, R. S. Lemons, T. Seya, and J. P. Atkinson. 1988. Molecular cloning and chromosomal localization of human membrane cofactor protein (MCP): evidence for inclusion in the multigene family of complement-regulatory proteins. J. Exp. Med. 168:181. 49. Post, T. W., M. K. Liszewski, E. M. Adams, I. Tedja, E. A. Miller, and J. P. Atkinson. 1991. Membrane cofactor protein of the complement system: Alternative splicing of serinelthreoninelproline-rich exons and cytoplasmic tails produces multiple isoforms that correlate with protein phenotype. J. Exp. Med. 174:93. 50. Kojima, A., K. Iwata, T. Seya, M. Matsumoto, H. Ariga, J. P. Atkinson, and S. Nagasawa. 1993. Membrane cofactor protein (CD46) protects cells predominantly from alternative complement pathway-mediated C3-fragment deposition and cytolysis. J.Immunol. 151:1519. 51. Weiss, L., E. Fischer, N. Haeffner-Cavaillon, M. Jouvin, M. Appay, 1. Bariety, and M. Kazatchkine. 1989. The human C3b receptor (CRt). Adv. Nephrol. 18:249. 52. Yoon, S. H., and D. T. Fearon. 1985. Characterization of a soluble form of the C3b1C4b receptor (CR1) in human plasma. J.Immunol. 134:3332. 53. Shin, M. L., G. Hansch, V. W. Hu, and A. Nicholson-Weller. 1986. Membrane factors responsible for homologous species restriction of complement-mediated lysis: evidence for a factor other than DAP operating at the stage of C8 and C9. J. Immunol. 136: 1777. 54. Davies, A., D. L. Simmons, G. Hale, R. A. Harrison, H. Tighe, P. J. Lachmann, and H. Waldmann. 1989. CD59, an LY-6-like protein expressed in human lymphoid cells, regulates the action of the complement membrane attack complex on homologous cells. J. Exp. Med. 170:637. 55. Holguin, M. H., L. R. Fredick, N. J. Bernshaw, L. A. Wilcox, and C. J. Parker. 1989. Isolation and characterization of a membrane protein from normal human 60 erythrocytes that inhibits reactive lysis of the erythrocytes of paroxysmal nocturnal hemoglobinuria. 1. Clin. Invest. 84:7. 56. Motoyama, N., N. Okada, M. Yamashina, and H. Okada. 1992. Paroxysmal nocturnal hemoglobinuria due to hereditary nucleotide deletion in the HRF20 (CD59) gene. Eur. 1. Immunol. 22:2669. 57. Schonennark, S., E. W. Rauterberg, M. L. Shin, S. Loke, D. Roelcke, and G. Hansch. 1986. Homologous species restriction in lysis of human erythrocytes: a membrane-derived protein with C8-binding capacity functions as an inhibitor. 1. Immunol. 136:1772. 58. Kalli, K. R., J. M. Ahearn, and D. T. Fearon. 1991. Interaction ofiC3b with recombinant isotypic and chimeric fonns ofCR2. 1.lmmunol. 147:590. 59. Fearon, D. T. 1979. Regulation of the amplification C3 convertase of human complement by an inhibitory protein isolated from human erythrocyte membrane. Proc. Natl. A cad. Sci. USA. 76:5867. 60. Comacoff, J. B., L. A. Hebert, W. L. Smead, M. E. VanAman, D. 1. Birmingham, and F. J. Waxman. 1983. Primate erythrocyte-immune complex­clearing mechanism. 1. Clin. Invest. 7 1:236. 61. Fearon, D. T., and K. F. Austen. 1980. The alternative pathway of complement­a system for host resistance to microbial infection. New Eng. 1. Med. 303:259. 62. Fearon, D. T. 1984. Cellular receptors for fragment of the third component of complement. Immunol. Today. 5:105. 63. Nelson, D. S. 1953. The immune-adherence phenomenon, an immunologically specific reaction between microorganisms and erythrocytes leading to enhanced phagocytosis. Science. 118: 733. 64. Wilson,1. G., and D. T. Fearon. 1984. Altered expression of complement receptors as a pathogenetic factor in systemic lupus erythematosus. Arthritis Rheum. 27: 1321. 65. Medof, M. E., K. Iida, C. Mold, and V. Nussenzweig. 1982. Unique role of the complement receptor CR 1 in the degradation of C3b associated with immune complexes. 1. Exp. Med. 156:1739. 66. Abrahamson, D. R., and D. T. Fearon. 1983. Endocytosis of the C3b receptor of complement within coated pits in human polymorphonuclear leukocytes and monocytes. Lab. Invest. 48:162. 67. Fearon, D. T., 1. Kaneko, and G. G. Thomson. 1981. Membrane distribution and adsorptive endocytosis by C3b receptors on human polymorphonuclear leukocytes. 1. Exp. Med. 153:1615. 68. Ehlenberger, A. G., and V. Nussenzweig. 1977. The role ofmerubrane receptors for C3b and C3d in phagocytosis. J. Exp. Med. 145:357. 69. Jack. R. M., and D. T. Fearon. 1984. Altered surface distribution of both C3b and Fc receptors on neutrophils induced by anti-C3b receptor or aggregated IgG. J.Immunol. 132:3028. 70. Wright, S. D., L. S. Craigmyle, and S. C. Silverstein. 1983. Fibronectin and serum amyloid P component stimulate C3b-and C3bi-media