Search

The structure, organization and control of the expression of complement genes is now being studied at the RNA and DNA levels. This work was prompted by recognized genetic polymorphisms and complement deficiencies in humans and animals. Here Harvey Colten discusses progress towards an understanding of the molecular basis of these variants and abnormalities.

Factor B (Bf), a constituent of the alternative pathway of complement activation, is encoded by a single gene that is located within the MHC. In murine kidney and intestine, two Bf transcripts (Bf short and Bf long), generated from distinct transcriptional initiation sites, are expressed in approximately equal amounts. In the liver, > 95% of Bf mRNA is the short transcript. To ascertain the biologic consequences of this tissue-specific mRNA polymorphism, we quantitated the effect of structural differences between the two transcripts on net Bf protein synthesis. Cell-free translation of Bf mRNA in vitro revealed that the rate of translation of Bf short is about twice that of Bf long. The 5' extension of Bf long includes four short open reading frames upstream of the authentic translational initiation codons. Mutation of all four upstream AUGs generates a Bf long transcript with a translational rate about equal to that of Bf short. This effect was primarily accounted for by mutation of the second AUG from the 5' end. Similar studies of Bf expression in vivo showed an approximately twofold difference in translational rate between Bf long and the transcript mutated at the upstream AUGs. Because systemic and local inflammation in kidney alters the ratio of Bf long and short, net production of complement protein Bf in extrahepatic tissues is regulated by both transcriptional and translational control mechanisms.

The complement system consists of about two dozen plasma and cell membrane proteins which function as cofactors in defense against pathogenic microbes and in the generation of many immunopathogenic disorders. The complexity of this system and its role in other biological functions has been appreciated within the last two decades. Recognition of genetic deficiencies of the complement proteins and their phenotypic expression has provided additional insights into the physiological role of the complement system. Complementary DNA (cDNA) clones for most of the complement components are now available, and the gene structures for many have been elucidated. Application of molecular biological methods to studies of the complement system and its deficiencies has permitted both the determination of primary structure and chromosomal localization of complement genes (Figure 1) and the capacity to elucidate the molecular basis of complement deficiency disorders.

Proteins of the complement system are important effectors and modulators of inflammation. The complement cascade is triggered by microbes, tissue debris, and specific antibodies. Serum complement proteins are derived primarily from liver, but extrahepatic complement synthesis is important in homeostasis and in local host defenses. Tissue-specific regulation of expression of complement genes is governed by mechanisms similar to those that regulate other “acute phase reactants.” That is, tissue injury or infection elicit changes in expression of these acute phase proteins, which, although variable in kinetics, magnitude, and direction, are a consequence of an elaborate system of cell-to-cell communication. This communication is mediated via a complex network of cytokines, including the interferons, interleukins, several growth factors, and sex hormones. The cell biological and molecular biological details of these mechanisms are now under active investigation. An understanding in molecular terms of the balance between proinflammatory and counterregulatory forces on complement gene expression should provide new insight into the functions of complement and the design of novel therapies for disorders of inflammation.

The biological activities of the more than 30 proteins that comprise the complement system have been elucidated in parallel lines of investigation that resulted in the purification of the proteins and studies of their function in vitro. Twenty years ago, the first complement cDNA clones were generated. Subsequently the structure and chromosomal localization of the complement genes and the primary sequences of their gene products were revealed. For some, even their higher order structure was solved. This work, coupled with studies of complement gene expression, biosynthesis, post-synthetic processing and secretion, contributed to an analysis of the relatively rare naturally occurring genetic deficiencies of complement proteins discovered fortuitously in humans sand experimental animals. Not until the past 5 years, with the application of methods for manipulating genes in vivo (targeted deletion and overexpression), has it been possible to definitively assign specific functions to complement proteins and to assess their importance in the intact organism. These relatively recent studies have confirmed the in vitro work or revealed unexpected roles for complement effector and regulatory proteins in host defenses, specific immunity, immunopathology, metabolism and reproductive biology This work is reviewed and the implications for understanding human diseases and the design of novel pharmaceutical agents are discussed. The promise of this line of investigation is certain but the context imposed by gender, developmental stage, other genes and environment must be taken into account before the practical implications of this deeper understanding of complement biology are fully realized.

The complement protein C4, encoded by two genes (C4A and C4B) on chromosome 6p, is the most polymorphic among the MHC III gene products. We investigated the molecular basis of C4 deficiency in a Finnish woman with systemic lupus erythematosus. C4-specific mRNA was present at low concentrations in C4-deficient (C4D) patient fibroblasts, but no pro-C4 protein was detected. This defect in C4 expression was specific in that synthesis of two other complement proteins was normal. Analysis of genomic DNA showed that the proposita had both deleted and nonexpressed C4 genes. Each of her nonexpressed genes, a C4A null gene inherited from the mother, a C4A null gene, and a C4B null gene inherited from the father, all contained an identical 2-bp insertion (TC) after nucleotide 5880 in exon 29, providing the first confirmatory proof of the C4B pseudogene. This mutation has been previously found only in C4A null genes. Although the exon 29/30 junction is spliced accurately, this frameshift mutation generates a premature stop at codon 3 in exon 30. These truncated C4A and C4B gene products were confirmed through RT-PCR and sequence analysis. Among the possible genetic mechanisms that produce identical mutations is both genes, the most likely is a mutation in C4A followed by a gene conversion to generate the mutated C4B allele.

Inflammation is a programmed response to tissue injury and microbial agents. In primitive multicellular organisms, humoral and cellular constituents of host defenses are focused by the presence of a foreign body or at a site of tissue injury. The evolution in higher organisms of a discrete circulatory system introduced complexity that imposed a need for mechanisms regulating changes in permeability of the circulatory system and for margination/directed migration of circulating cellular elements. For instance, an array of regulated genes expressed in endothelium, blood leukocytes, and in extra- vascular cells direct the accumulation of white cells to sites of inflammation. Proteins within the selectin family (McEver 1991) displayed on the majority of lymphocytes, neutrophils, monocytes, and the corresponding cytokine- regulated proteins on endothelium, are critical for the recruitment of white cells to an inflammatory site. Similar general functions are served by adhesion molecules of the integrin and immunoglobulin (Ig) families (Springer 1990; Kishimoto et al. 1989).Small molecular species such as nitric oxide (Kubes et al. 1991), arachadonic acid metabolites (Smith et al. 1980), and reactive oxygen radicals (Patel 1991) generated at inflammatory sites, play modulating and direct roles in leukocyte adherence, migration, activation. This process involves elaborate control mechanisms and is highly redundant even when considering only the intravascular elements. In addition, separate vascular and peripheral tissue sources of the critical constituents evolved. This has resulted in separate regulatory mechanisms that govern availability of the intravascular and extravascular components of the inflammatory response.

Deficiency of the complement protein C2 (C2D), one of the most common genetic deficiencies of the complement system, is associated with rheumatological disorders and increased susceptibility to infection. Two types of C2D have been recognized, each in the context of specific major histocompatibility complex (MHC) haplotypes; type I, a deletion, frameshift and premature stop codon resulting in absence of detectable C2 protein synthesis, and type II, missense mutations resulting in a block in secretion of C2 proteins. Analysis of C2 expression in a child with C2 deficiency, a MHC haplotype different from those associated with type I or II C2D, and recurrent infections revealed additional molecular heterogeneity among C2 deficient patients. No detectable C2 protein was synthesized in the child's fibroblasts under conditions supporting C2 synthesis and secretion in normals and the child's C2 mRNA was reduced to 42% of normal. Nucleotide sequencing of RT-PCR fibroblast mRNA and genomic DNA revealed a type I C2 deficiency (28 base-pair deletion) on one allele and a previously unrecognized two base-pair deletion in exon 2 on the other. Expression of the closely linked factor B gene was markedly decreased (Bf mRNA 25% of normal), though Bf was up-regulated appropriately by interferon-gamma and the flanking sequence containing the Bf promoter was normal in this C2-deficient patient. Moreover, the concentration of Bf protein was normal in the patient's plasma.

The synthesis and secretion of factor H, a regulatory protein of the complement system, were studied in skin fibroblasts from an H-deficient child who has chronic hypocomplementemic renal disease. In normal fibroblasts, factor H transcripts of 4.3 and 1.8 kilobase pairs (kb) encode a 155-kDa protein containing short consensus repeat (SCR) domains 1-20 and a 45-kDa protein which contains SCRs 1-7, respectively. The patient's fibroblasts expressed normal amounts of the 4.3- and 1.8-kb messages constitutively and after tumor necrosis factor-alpha/interferon-gamma stimulation. Lysates of [35S]methionine-labeled fibroblasts from the patient contained the 155- and 45-kDa H polypeptides, but secretion of the 155-kDa protein was blocked; the 45-kDa protein was secreted with normal kinetics. The patient's plasma lacked the 155-kDa protein but contained the small form of H. Moreover, in fibroblasts the retained 155-kDa factor H protein was not degraded, even after 12 h. Immunoflourescent staining and confocal microscopic imaging of the patient's fibroblasts indicated that factor H was retained in the endoplasmic reticulum. Sequence analysis of reverse transcription-polymerase chain reaction products (the entire coding region) and genomic DNA revealed a T1679C substitution on one allele and a G2949A substitution on the other (C518R mutation in SCR 9 and C991Y mutation in SCR 16, respectively). Both mutations affect conserved cysteine residues characteristic of SCR modules and therefore predict profound changes in the higher order structure of the 155-kDa factor H protein. These data provide the first description of a molecular mechanism for factor H deficiency and yield important insights into the normal secretory pathway for this and other plasma proteins with SCR motifs.

Type II complement protein C2 deficiency is characterized by a selective block in C2 secretion. The Type II C2 null allele (C2Q0) is linked to two major histocompatibility haplotypes (MHC) that differ from the MHC of the more common Type I C2 deficiency. To determine the molecular basis of Type II deficiency the two Type II C2Q0 genes were isolated and transfected separately into L-cells. Subsequent molecular biology, biosynthetic, and immunofluorescence studies demonstrated that C2 secretion is impaired in Type II C2 deficiency because of different missense mutations at highly conserved residues in each of the C2Q0 alleles. One is in exon 5 (nucleotide C566 --T; Ser189 --Phe) of the C2Q0 gene linked to the MHC haplotype A11,B35,DRw1,BFS, C4A0B1. The other is in exon 11 (G1930 --A; Gly444 --Arg) of the C2Q0 gene linked to the MHC haplotype A2,B5, DRw4,BFS,C4A3B1. Each mutant C2 gene product is retained early in the secretory pathway. These mutants provide models for elucidating the C2 secretory pathway.

Factor B (Bf), a constituent of the alternative pathway of complement activation, is encoded by a single gene that is located within the MHC. In murine kidney and intestine, two Bf transcripts (Bf short and Bf long), generated from distinct transcriptional initiation sites, are expressed in approximately equal amounts. In the liver,95% of Bf mRNA is the short transcript. To ascertain the biologic consequences of this tissue-specific mRNA polymorphism, we quantitated the effect of structural differences between the two transcripts on net Bf protein synthesis. Cell-free translation of Bf mRNA in vitro revealed that the rate of translation of Bf short is about twice that of Bf long. The 5' extension of Bf long includes four short open reading frames upstream of the authentic translational initiation codons. Mutation of all four upstream AUGs generates a Bf long transcript with a translational rate about equal to that of Bf short. This effect was primarily accounted for by mutation of the second AUG from the 5' end. Similar studies of Bf expression in vivo showed an approximately twofold difference in translational rate between Bf long and the transcript mutated at the upstream AUGs. Because systemic and local inflammation in kidney alters the ratio of Bf long and short, net production of complement protein Bf in extrahepatic tissues is regulated by both transcriptional and translational control mechanisms.

Alternative splicing of the primary transcript for human complement protein C2 generates templates for translation of a secreted (C2 long) protein and an intracellular (C2 short) form in liver, bronchoalveolar macrophages, and fibroblasts. The approximate ratio of C2 long to C2 short mRNA is 2:1. The C2 short mRNA does not contain the 396-base pair encompassed by exons 2 and 3 of the full-length C2 long and thus lacks codons for the 5 carboxyl-terminal residues of the signal peptide. Synthesis of C2 in cells transfected with full-length RNA corresponding to each of the transcripts show that C2 long is secreted within a half-time of approximately 1 h and that C2 short is not secreted. Cell-free biosynthesis in the presence of microsomes demonstrate that this intracellular C2 protein (70 kDa) is apparently capable of traversing the membrane of the endoplasmic reticulum. Though the function of the intracellular C2 protein is unknown, it is abundant in all cell types that express the C2 gene.

Congenital alveolar proteinosis (CAP), a cause of respiratory failure in fill-term newborns, often leads to death in infancy despite medical therapy. We recently described an inherited deficiency of surfactant protein B (SP-B) (N. Engl. J. Med. 1993; 328:406-410) in two siblings with CAP. The SP-B deficiency was accompanied by marked abnormalities, both quantitative (increase) and qualitative (distribution), of SP-A and SP-C in the lungs of the affected infants. Ultrastructural studies of the lung of one of these infants and of a third affected sibling born in the index family showed abundant alveolar concentric multilamellated structures and membranous vesicles but no typical tubular myelin. In addition, membranous vesicles from type II cells and immunogold labeled SP-A and SP-C were found between type II cells and their basement membrane despite intact interepithelial cell junctions. These findings suggest an important role for SP-B in the directionality of surfactant secretion and in the formation of tubular myelin.

The expression of the complement protein C3 in extrahepatic tissues is highly regulated during the course of inflammation. Hence, systemic acute phase stimuli such as bacterial lipopolysaccharide (LPS) and autoimmune nephritis in aged 'lupus mice' (MRL-lpr/lpr and NZB x NZW F1) both lead to increased C3 mRNA expression in whole kidney. In situ hybridization was used to determine the intrarenal cell type(s) capable of constitutive and regulated C3 mRNA expression. Normal mice injected with Escherichia coli LPS show a marked increase in whole kidney C3 mRNA over control (saline-injected) animals. The renal C3 mRNA in LPS-stimulated mice was found in cortical tubular epithelium. By contrast, in aged (18 week) MRL-lpr/lpr mice, which develop lupus nephritis, the increased intrarenal C3 messenger RNA was localized to perivascular inflammatory cells surrounding medium-sized arteries. Similar perivascular infiltrates were seen in the lungs of the MRL-lpr/lpr mice, and focal inflammatory cell infiltrates were also found in the myocardium. Leucocytes in these infiltrates accounted for the increased C3 expression in these tissues. These findings suggest cell as well as tissue specificity of the response to inflammatory stimuli in the local extrahepatic production of the third component of complement.

Complement protein biosynthesis has been a subject of study for over 100 years. In the 1890s, Metchinoff suggested that leukocytes were a source of complement1. Interest in the liver as a site of complement production was first generated by the experiments of Ehrlich and Morgenroth2, who demonstrated that complement, but not antibody, was depressed in sera of animals poisoned with phosphorus. Shortly thereafter, Muller found that perfusion of liver with serum led to an increase in complement activity in the effluate, and concluded that the liver was the site of complement synthesis3. In the late 1960s, Alper and coworkers4‘5 confirmed this observation that the liver was a major site of serum complement synthesis by studying a patient following liver transplantation; the electrophoretic form of C3 shifted from the recipient FS type to the donor SS type. The complete shift of C3 from recipient to donor type indicated that the liver synthesized at least 95% of the C3 found in serum4,5. Also in the 1960s, Thorbecke and her colleagues6-8 suggested that cell types of endodermal, mesodermal and ectodermal origin in several organs were capable of synthesizing proteins of the complement system. Application of modern molecular biological methods has confirmed that complement proteins are synthesized at multiple sites. In addition, studies of complement metabolism and mRNA expression in vivo9~n and in tissue culture12,13 indicate a role for complement produced at extrahepatic sites, especially in the immunopathogenesis of autoimmune injury. The functional importance of this extrahepatic complement production is supported by the elaborate tissue-specific and developmentally regulated controls of complement gene expression now recognized in several species.

Not since the first decade following the discovery of complement, approximately 100 years ago, has the accumulation of knowledge about the complement system been as rapid, broad and deep. Table 1 lists the complement proteins and receptors and references in recent reviews (Colten and Gitlin, in press; Reid and Day 1989) summarize many of these advances. This remarkable conceptual progress has been primarily due to the application of molecular cloning methods to studies of complement genes and gene products. This symposium highlights a few areas in which these advances have already yielded practical as well as fundamental information.

Murine genomic and cDNA clones were isolated to ascertain the mechanisms accounting for previously recognized multiple forms of complement C2 and factor B mRNA and to analyze structural similarities with the corresponding human gene products (C2, 74% and B, 85%, amino acid identity). Like the human Bf gene, murine Bf and C2 each consist of 18 exons with similar intron-exon organizations. The murine C2 gene (20 kilobases) is more than three times the size of Bf (6 kilobases) due to the presence of large intronic segments separating the exons encoding the NH2-terminal binding and central (von Willebrand factor) domains. Evidence from cDNA clones shows that the two C2 transcripts are generated by an alternative splice at the donor site of exon 14 producing long and short C2 mRNA species that differ by 21 base pairs encoding a region within the binding pocket (amino acids 636-642 (GSTCKDH)) of the serine proteinase domain. Tissue-specific multiple factor B transcripts are generated by alternative transcriptional initiation. From the structure of these transcripts the predicted regulation of expression and rates of translation of B programmed by the short and long mRNA species may differ, but the polypeptides are identical. These data indicate that different molecular mechanisms account for the multiple forms of C2 and factor B mRNA and that the structure of the different transcripts predicts differences in function and expression of the respective gene products.

To study the role of local production of complement proteins during the evolution of a naturally occurring immune complex disease, C3, C4, C2 and Factor B mRNA expression was assessed in several tissues of the inbred mouse strains NZB and (NZB x W) F1 hybrid. In the NZB/W F1 hybrid strain, coincident with the development of glomerulonephritis a marked increase in kidney C3 and C4 mRNA was observed; Factor B mRNA, which is expressed as a doublet in kidney and intestine, showed an increase in expression of the smaller transcript. This alteration of kidney C3, C4 and Factor B mRNA is identical to that noted in association with lupus nephritis in the MRL lpr/lpr strain and following in vivo administration of endotoxin to the BALB/c strain. The development of systemic lupus erythematosis (SLE) in the NZB/W F1 was not associated with a marked change in hepatic complement gene expression. These findings support the hypothesis that local production of complement may play a role in the pathogenesis of glomerulonephritis and other tissue injury in SLE.

Investigation of the basis for increased susceptibility to infection with pyogenic organisms in an eight-year-old boy revealed a defect in receptor coupled polymorphonuclear-leukocyte func...

Complementary DNA clones corresponding to the human serum amyloid P component (SAP) were isolated, and the complete nucleotide and derived amino acid sequence of preSAP was determined. PreSAP biosynthesis is directed by a 1.1-kilobase mRNA. Synthesis and postsynthetic processing of preSAP in Xenopus oocytes result in secretion of a protein with mobility similar to native purified SAP when analyzed by sodium dodecyl sulfate gel electrophoresis. The human SAP gene is on chromosome 1, probably closely linked to the gene for C-reactive protein which encodes the related acute phase reactant found in human plasma.

The net production of the complement protein C2, C4, and factor B differ among mononuclear phagocytes from peripheral blood and different tissues in experimental animals and in humans. To examine the mechanisms that regulate these differences in humans, the proportions of C2-producing cells, the average single-cell production rate of C2, the posttranslational glycosylation and kinetics of secretion of C2 and factor B, and the amounts of C2 and factor B mRNA were examined in freshly isolated peripheral blood monocytes, monocytes maintained up to 2 wk in culture, and freshly isolated tissue macrophages from breast milk and bronchoalveolar lavage. In addition, the biosynthesis of two other proteins synthesized and secreted by mononuclear phagocytes, C3 and lysozyme, were examined. We report that despite comparable rates of C3 and lysozyme synthesis and similar processing and kinetics of secretion of C2 and factor B, the freshly isolated tissue macrophage differs from the monocyte-derived macrophage in the proportion of C2-producing cells, in the average single-cell production rate of complement, and in the amounts of specific C2 and factor B mRNA. These differences are tissue specific, because C2-specific mRNA content in bronchoalveolar macrophages is considerably greater than in breast milk macrophages, although the amounts of factor B mRNA are comparable. These data suggest that tissue-specific regulation of complement production in human mononuclear phagocytes occurs at a pretranslational level. These studies now provide a basis for investigation of the molecular effects of agents that modulate the biologic functions of monocytes and macrophages.

Inflammation or acute tissue injury results in a programmed change in the concentration of several plasma proteins. Among these proteins, two--C-reactive protein (CRP) and serum amyloid A protein (SAA)--increase up to 1000-fold after an acute-phase stimulus in humans and rabbits. To determine the mechanism for regulation of acute-phase gene expression, we examined changes in the rates of transcription and specific hepatic mRNA content for rabbit CRP, SAA, and some complement protein mRNA during an acute-phase response. Induction of a sterile inflammatory reaction with intramuscular injection of turpentine resulted in an increase in the hepatocellular content of CRP, SAA, C3, and factor B mRNA and the transcription of CRP, SAA, and C3 genes. These data suggest that the increase in CRP, SAA, and C3 serum concentrations observed during an acute-phase reaction is due to an increase in biosynthesis and is, at least in part, under transcriptional control.

The interaction of C3b, the major cleavage product of C3, with its receptor on human granulocytes results in important biological functions, including phagocytosis, superoxide generation, and release of a variety of enzymes, including histaminase. We have determined the binding kinetics and isotherm of trypsin-generated fluid-phase dimeric C3b at both 0 degrees C and 37 degrees C using human polymorphonuclear leukocytes (PMN). At 37 degrees C the apparent number of receptors per cell was threefold greater than number at 0 degrees C (66,000 vs 21,000), and the affinity (Ka) was slightly less (3.5 X 10(7) M-1 vs 6 x 10(7) M-1). C3b dimer binding was specifically inhibited by a F(ab,)2 anti-C3b receptor antibody. C3b dimer induced histaminase release in a dose-dependent fashion at a concentration of 1 to 4 X 10(7) molecules/cell, whereas at lower concentrations a dose-dependent inhibition of opsonized zymosan-induced release was noted. These effects were independent of IgG. In contrast, C3b monomer failed to demonstrate measurable direct binding or to induce histaminase release. Histaminase release, however, did occur after incubation of PMN with affinity-linked monomer (anti-C3 F(ab')2-C3b complexes). Monovalent complexes did not affect release. Monomeric C3, C3c, and C3d were without effect; however, affinity-linked C3 and C3c did cause release.

Mo1 is a glycoprotein heterodimer found on the surface of phagocytic cells. By use of monoclonal antibodies directed against various epitopes of the 155 kd alpha-chain of Mo1, two distinct functions of this glycoprotein have been identified. Mo1 serves as the receptor (CR3) for iC3b, one of the breakdown products of the third component of complement, and in addition appears to be involved in promoting cell-cell adhesion of granulocytes. In studies performed with a subline of the acute myelogenous leukemia tumor cell line KG1, we found cells from this subline (KG1m) incapable of iC3b-mediated binding despite these cells having surface Mo1. Five distinct epitopes on Mo1 identifiable on normal cells by a panel of anti-Mo1 alpha-chain monoclonal antibodies (including OKM10, thought to be identical to or close to the iC3b binding site) were equally present on KG1m by indirect immunofluorescence. The electrophoretic mobilities of both the alpha- and beta-chains of Mo1 derived from KG1m were identical to those isolated from normal granulocytes. To determine whether altered receptor mobility or decreased surface density of Mo1 was responsible for the lack of Mo1-related functions, binding to EiC3b of isolated Mo1 derived from KG1m lysate was measured. KG1m-derived Mo1 did not bind to EiC3b when compared with normal granulocyte-derived Mo1, suggesting that the lack of iC3b binding is not secondary to decreased surface receptor number or mobility. This subline of KG1 provides an excellent model for the study of the relationship between surface receptor structure and function.

Studies of complement (C) secretion by single cells indicate that only a subset of a guinea pig macrophage population is capable of secreting the second (C2) and fourth (C4) components of C. Moreover, cell-surface bound C4 antigen is also found on a proportion of guinea pig peritoneal macrophages. The availability of methods for the isolation of macrophages bearing surface membrane C4 antigen and for the detection of the secretion of C by single cells made it possible to ascertain the relationship between these two subsets. Approximately 25% of the freshly isolated peritoneal macrophages were surface C4 antigen positive. When incubated in conditioned medium containing C4 or incubated in small volumes to increase the concentration of fluid phase C4, the proportion of macrophages bearing surface C4 increased to approximately 80%. The surface C4 antigen was adsorbed from the medium and was predominantly native C4. The proportion of peritoneal macrophages secreting functionally active C4 or C2 was approximately 45% as measured by a hemolytic plaque assay technique. The proportion of C4-secreting cells decreased to 5% after incubation in conditioned medium containing preformed C4, whereas the proportion of C2-producing macrophages was unchanged, i.e., it remained at about 45%. The removal of secreted C4 with F(ab')2 anti-C4 or effectively decreasing C4 concentration by increasing the volume of the culture medium abrogated the decrease in the proportion of C4-secreting cells. Conditioned medium derived from cells genetically deficient in C4 had no effect on the proportion of C4- or C2-producing macrophages. The macrophage population bearing surface membrane C4, isolated by fluorescence-activated cell sorting, contained more than 90% of the C4-producing cells. Cells producing C2 were distributed equally in both subpopulations. Maintenance of the surface C4-positive, C4-producing cells in culture for 12 hr resulted in a decrease in the proportion of C4-secreting cells. Conversely, isolated macrophages initially surface C4 negative and not producing C4, developed the capacity to produce C4 in culture. C4 production in the isolated surface C4-negative population was inhibited by incubation in medium containing preformed C4. These results suggest the presence of a negative feedback effect on C4 secretion that is mediated by extracellular C4.(ABSTRACT TRUNCATED AT 400 WORDS)

gamma-Interferon (IFN-gamma) is a well characterized lymphokine known to regulate many mononuclear phagocyte functions, including expression of class I and class II major histocompatibility complex genes. The second component of complement (C2) and factor B are major histocompatibility complex class III gene products synthesized in mononuclear phagocytes. Recombinant IFN-gamma increased the synthesis of C2 and factor B in primary cultures of human mononuclear phagocytes and in murine fibroblasts transfected with cosmid DNA bearing the human C2 and factor B genes. In both cell types the increases in C2 and factor B protein synthesis were detected at concentrations of IFN-gamma less than 1 unit/ml and the regulation of each was pretranslational. The IFN-gamma-induced increases in C2 and factor B mRNA did not require new protein synthesis. In primary cultures of human monocytes, the kinetics of induction of C2 and factor B synthesis differed, but in the transfected L-cells the kinetics were similar, suggesting differences in transduction of the IFN-gamma signal, transcriptional, and/or post-transcriptional events in the two cell types. The small size of the factor B 5' flanking region, which is bounded by the 3' terminus of the IFN-gamma-regulated C2 gene, provides a well defined region to probe the structural basis for IFN-gamma regulation of gene expression.

In order to characterize a genetic deficiency of C2 in guinea pigs, production of C2 by peritoneal macrophage cultures derived from four normal, four heterozygous deficient, and four homozygous deficient animals was measured functionally and immunochemically after metabolic labeling with 35S-methionine. Macrophage monolayers from homozygous deficient animals failed to secrete hemolytically detectable C2 up to 74 hr in culture. A single cell hemolytic plaque assay also failed to demonstrate any functional C2 production by cells from homozygous deficient animals. No C2 protein was detected in media from three of the four homozygous deficient animals, but in one, apparent C2 fragments were present. In contrast, intracellular C2 protein was identified in all four homozygous deficient cell cultures. Its mobility on SDS-PAGE was slightly faster than normal. Much less abnormal intracellular C2 protein was recovered from homozygous deficient macrophage monolayers than intracellular C2 protein from normal macrophage monolayers. Monolayers from heterozygous animals produced functional and immunochemical C2 at approximately 30% of the normal rate. Normal rates of biosynthesis and secretion of two other MHC-linked class III antigens, C4 and factor B, were detected in macrophage cultures from homozygous and heterozygous deficient animals. These data suggest that a specific defect, i.e. a structural abnormality in C2 protein, underlies C2 deficiency in guinea pigs.

The effect of inflammation on C (C2 and C4) biosynthesis by bronchoalveolar and peritoneal cells was studied with methods that detect changes in synthesis rates and qualitative changes in cell populations. Adherent bronchoalveolar macrophages produced less C2 and C4 than adherent peritoneal macrophages. However, the subset of cells capable of producing C was more than 20-fold greater in the peritoneal cell population than in the population of bronchoalveolar cells. The rate of synthesis per C producing bronchoalveolar macrophage was 5 to 10 times the rate by C producing peritoneal macrophages. In contrast, the effect of an inflammatory stimulus on C production was the result of a change in rate of synthesis per cell, not a change in the proportion of C producing cells. This stimulatory effect was exerted locally, not on a cell population harvested from a distant site. At least two mechanisms for local control of C levels in tissues have been identified--that is, a change in synthesis rate and a qualitative change in macrophage cell population. The capacity to alter C levels at a site of inflammation may be important for C dependent functions in host defenses.

Human factor I is a two-chain plasma glycoprotein composed of disulfide-linked 50,000- and 38,000-dalton subunits. Analysis of its biosynthesis and postsynthetic processing demonstrated that factor I is synthesized as a single chain precursor (pro-I) that undergoes glycosylation and limited proteolysis to generate the native protein. One of three human hepatoma cell lines, HepG2 , secreted factor I predominantly (70-90%) in a single chain pro-I form. The other cell lines secrete factor I predominantly in its two chain native form. The defect in conversion of pro-I to I in HepG2 was protein specific since other multichain proteins, derived from single chain precursors, the third, fourth, and fifth components of complement were processed normally. Further analysis of the inefficient pro-I to I conversion by HepG2 revealed that Xenopus oocytes injected with HepG2 mRNA secreted factor I in a predominantly two-chain form. In addition, the apparent sizes of native factor I, transferrin, and alpha-1-antitrypsin secreted by the three hepatoma lines differed due to differences in postsynthetic processing.

Factor B (Bf) gene expression is regulated independently in hepatic and extrahepatic tissues. In studying murine factor B expression, we found two mature mRNA species for Bf in kidney and intestine (2.7 and 2.4 kb) but in other tissues, including liver, lung, spleen, heart, brain, and skeletal muscle, only a single Bf transcript (2.4 kb) was observed. These two Bf transcripts in kidney or intestine are identical in structure except for a 300 bp 5' extension found in the 2.7-kb mRNA. The long Bf transcript originates from a tissue-specific alternative site of transcriptional initiation upstream of the initiation site for the 2.4-kb mRNA, as determined by primer extension and nuclease protection assays. The upstream initiation site lacks a typical promoter motif and is only 80 bp downstream from the terminus of the murine C2 gene. After stimulation in vivo with endotoxin, only the 2.4-kb mRNA increases in kidney indicating independent regulation of the two Bf transcripts. The differences in 5' untranslated region generated in these two Bf mRNA species may affect local concentrations of factor B protein in kidney and intestine by mechanisms depending on differential translation rates or stability of mRNA.

The second component of complement (C2), is a class III major histocompatibility complex gene product and a glycoprotein in the classical complement activating system. Synthesis in the human hepatoma-derived cell line HepG2 results in three intracellular forms: an 84-kDa form secreted in 1-2 h; 79-kDa and 70-kDa forms that remain cell-associated for intervals up to 12 h. All three forms are C2 polypeptides as demonstrated by inhibition of immunoprecipitation with unlabeled C2 and the presence of common major peptide fragments following chymotryptic digestion. The cell-associated forms of C2 are not products of proteolysis as demonstrated by experiments with multiple proteinase inhibitors and by observations of the kinetics of synthesis. Inhibition of core glycosylation by tunicamycin and deglycosylation by acid hydrolysis indicate that the three intracellular C2 polypeptides are glycosylated to a similar extent. Although the 84-kDa form of C2 is susceptible to C1s cleavage, the two cell-associated forms are not. Cell-free biosynthesis by mRNA from HepG2 or human liver results in three primary translation products corresponding to the three unglycosylated forms of C2. These results indicate that HepG2 cells synthesize C2 protein in both secreted and cell-associated forms and that each form is derived from a separate primary translation product.

In order to examine one function of mononuclear phagocytes during maturation from bone marrow precursors to tissue macrophages, an improved hemolytic plaque assay for the detection of synthesis of the second (C2) and fourth (C4) components of C by single cells was developed. With this method, production of C2 and C4 was assessed in cell populations derived from bone marrow, blood, lung, peritoneum, and spleen. The proportion of cells producing C2 and C4 in each population varied. Approximately 10% of bone marrow cells produced C4, but not detectable C2 plaque-forming cells (PFC) were detected. Circulating monocytes yielded about 10% PFC each for C2 and C4. The proportion of C2-producing cells in tissue macrophages varied from approximately 2% in bronchoalveolar macrophages to about 45% in peritoneal and splenic macrophage populations, whereas C4 production by macrophages from lung, peritoneum, and spleen were all approximately 45%. These data suggest that differences in C biosynthesis characterize mononuclear phagocytes at different stages of maturation.

Human monocytes synthesized the third component of complement (C3) up to 5 wk in vitro. Evidence for net C3 synthesis was based on (a) incorporation of 14C-labeled amino acids into C3 protein, (b) indentity of the allotype of C3 produced in vitro with that of the doner's serum C3, even in the presence of carrier C3 protein of a different allotype; (c) correspondence of electrophoretic mobility, size, and subunit structure of C3 protein produced in vitro with serum C3; (d) inhibition of C3 production with cycloheximide. Monocytes from two unrelated C3-deficient patients were studied under conditions that supported C3 synthesis by normal monocytes. Serum from each of the patients contained less than 1% of the normal C3 concentration, buth their monocytes produced C3 at approximately equal to 25% of the normal rate when studied after 2 wk in vitro. The C3 produced in vitro by monocytes from one of the patients had the molecular weight of normal serum C3 and dissociated appropriately under reducing conditions. Monocytes from C3-deficient patients could not be distinguished from normals on the basis of morphology, rosetting with C3-coated erythrocytes, or rates of C2, and total protein synthesis.

To assess the potential role of membrane interleukin 1 (mIL-1) in modulating expression of acute phase proteins, we studied the effect of fixed mouse peritoneal macrophages and isolated cell membranes on the synthesis of C3 and albumin in human hepatoma Hep 3B cells. An increase in C3 synthesis and decrease in albumin synthesis were detected after incubation of Hep 3B cells with fixed stimulated macrophages or with membrane preparations. Estimates of hepatocellular C3 and albumin mRNA indicated that the mIL-1 regulation was exerted at a pretranslational level. The changes in C3 and albumin expression induced by mIL-1 were inhibited by addition of anti-interleukin 1 (IL-1) antiserum to the macrophages, supporting the hypothesis that a form of IL-1 is present on the outer cell membrane of macrophages. Moreover, cell-surface iodination of macrophages allowed the detection of a 33-kDa IL-1 molecule, suggesting that an IL-1 species, similar to the intracellular precursor of IL-1 is present at the outer cell surface of macrophages. The expression of mIL-1 was temporally dissociated from IL-1 release, indicating that the cell-surface associated IL-1 activity is not the result of adsorbed soluble IL-1. These studies provide a basis for further investigation of the role of mIL-1 in modulating monocytes and macrophages function via cell-cell interactions.

Genetic deficiency of the second component of complement (C2) is the most common complement-deficiency state among Western Europeans and is frequently associated with autoimmune diseases. To examine the molecular basis of this deficiency, we established cultures of blood monocytes from four families with C2-deficient members. Using a hemolytic-plaque assay, [35S]methionine metabolic labeling of proteins in tissue culture and immunoprecipitation, RNA extraction and Northern blot analysis, and DNA restriction-enzyme digestion and Southern blot analysis, we found that C2 deficiency is not due to a major gene deletion or rearrangement but is the result of a specific and selective pretranslational regulatory defect in C2 gene expression. This leads to a lack of detectable C2 mRNA and a lack of synthesis of C2 protein. The approach used in this study should prove useful in examination of other plasma protein deficiencies, especially those in which the deficient gene is normally expressed in peripheral-blood monocytes or tissue macrophages and in which ethical considerations preclude the use of liver or other tissue for study.

To study the molecular regulation of C4 and factor B synthesis in mouse peritoneal macrophages, mouse C4 cDNA clones isolated from an H-2d haplotype liver cDNA library, and a previously described mouse factor B cDNA clone, pBmB2 (9), were used to assess quantitative and qualitative differences in C4 and factor B mRNA in resident and elicited cells. The C4 clones that were isolated, pBmS2 (1 Kb) and pBmS10 (0.9 Kb), overlap and together span a 1.5 Kb coding region of mouse pro-C4, extending from the alpha-chain through the gamma-chain; four nucleotide substitutions are evident in comparing 316 bp of the sequence of clone pBmS10 to that of a previously described mouse C4 clone, pMLC4/w7-2 (23). By using these probes, Northern blot analysis of total cellular RNA revealed similar C4 mRNA levels in resident peritoneal macrophages from high-C4 (B10.A) and low-C4 (C3HeB) strains. Pulse and pulse-chase studies of C4 and factor B synthesis were performed on resident, starch-elicited, and thioglycollate-elicited peritoneal macrophages at two culture time periods, 0 to 9 and 24 to 33 hr, and total cellular RNA was isolated from each population at 4.5 and 28.5 hr of culture for Northern blot analysis of C4 and factor B mRNA content. The data demonstrate that as previously reported, C4 production decreases in elicited compared with resident macrophages and decreases with time in culture; however, factor B synthesis does not differ among resident and elicited cells and it increases with time in culture. The variations in C4 and factor B production by mouse peritoneal macrophages are not associated with alterations in C4 and factor B protein processing, catabolism, or secretion; rather, they are a function of differences in net amounts of C4 and factor B mRNA. These data provide direct evidence that the regulation of expression of these class III MHC genes in mouse peritoneal macrophages is a pretranslational event.

The interaction of opsonized particles with human granulocytes promotes a number of important biologic functions, including phagocytosis, superoxide generation, and release of a variety of enzymes, including histaminase. We have previously determined that histaminase release occurs via a C3-dependent process. Although fluid-phase C3b dimers can mediate release, the relative effects of particle-bound C3b and C3bi and of IgG have not been examined. In this report we demonstrate that particle-bound C3 deposited on activators of the alternative C pathway effected histaminase release in the absence of IgG. Particle-bound C3bi and C3b were both effective as mediators of histaminase release. The extent of release varied as a function of the activating surface on which C3 was deposited (zymosan C3b was considerably more potent than C3b bound to rabbit erythrocytes, which was slightly more potent than C3b bound to neuraminidase-treated sheep erythrocytes). In contrast, C3b or C3bi deposited on nonactivating surfaces (such as sheep erythrocytes) at inputs of up to 2,000,000 molecules per granulocyte failed to induce histaminase release unless IgG was also present. The ability of C3b bound to particles that serve as activators of the alternative pathway to induce histaminase release is apparently not the result of decreased susceptibility of C3b to proteolysis or to an increased binding affinity to the C3b receptor, but may relate to the interaction of other surface structures on activating particles with the PMN membrane.

We have used available protein sequence data for the anaphylatoxin (C5a) portion of the fifth component of human complement (residues 19-25) to synthesize a mixed-sequence oligonucleotide probe. The labeled oligonucleotide was then used to screen a human liver cDNA library, and a single candidate cDNA clone of 1.85 kilobase pairs was isolated. Hybridization of the mixed-sequence probe to the complementary strand of the plasmid insert and sequence analysis by the dideoxy method predicted the expected protein sequence of C5a (positions 1-12), amino-terminal to the anticipated priming site. The sequence obtained further predicted an arginine-rich sequence (RPRR) immediately upstream of the N-terminal threonine of C5a, indicating that the promolecule form of C5 is synthesized with a beta alpha-chain orientation as previously shown for pro-C3 and pro-C4. The C5 cDNA clone was sheared randomly by sonication, subcloned into M13 mp8, and sequenced at random by the dideoxy technique, thereby generating a contiguous sequence of 1703 base pairs. This clone contained coding sequence for the C-terminal 262 amino acid residues of the beta-chain, the entire C5a fragment, and the N-terminal 98 residues of the alpha'-chain. The 3' end of the clone had a polyadenylated tail preceded by a polyadenylation recognition site, a 3'-untranslated region, and base pairs homologous to the human Alu concensus sequence. Comparison of the derived partial human C5 protein sequence with that previously determined for murine C3 and human alpha 2-macroglobulin has indicated regions of pronounced sequence similarity. Examination of cytoplasmic RNA prepared from human liver and the human hepatoma cell line Hep G2 by Northern transfer has indicated a C5 mRNA species of about 5.2 kilobase pairs.

The amino acid sequence of the NH2-terminal regions of the intracellular precursor of the fourth component of guinea pig complement (pro-C4) and isolated alpha, beta, and gamma chains of native C4, synthesized by peritoneal macrophages in culture, were determined with a microradiosequencing technique. Radiolabeled pro-C4 was immunoprecipitated from cell lysates and native C4 from culture media which contained one of six 3H-amino-acids or [35S]methionine. The purity of these proteins was established and molecular weights were estimated by electrophoresis in sodium dodecyl sulfate. Seventeen of the first 22 residues of pro-C4 were identified. Fifteen of these 17 were nonpolar. The eight amino acids ascertained within the first 9 NH2-terminal residues of guinea pig pro-C4 were identical with those reported for human C4 beta chain, and identity with residues determined for guinea pig C4 beta chain was established. These data indicate that beta chain is the NH2-terminal segment of pro-C4, that no residues are cleaved from the NH2 terminus during the conversion to native C4, and that this segment of the pro-C4 molecule is conserved phylogenetically.

To identify the mechanisms accounting for hemodialysis-induced granulocytopenia, we undertook quantitative kinetic studies of a granulocyte-adhesion-promoting surface glycoprotein (Mo1). In eight patients undergoing maintenance hemodialysis, there was a fivefold increase in the mean cell-surface expression of Mo1 within 15 minutes after the start of dialysis with a new cuprophane membrane. The peak increase in surface Mo1 coincided with the maximal drop in neutrophil count and with the peak rise in the plasma levels of the complement-activation products C5a desArg and C3a desArg. During dialysis on a membrane being reused for the fifth time, no significant complement activation, no increase in Mo1 expression, and no change in neutrophil count were seen. C5a desArg (but not C3a desArg) induced a comparable increase in Mo1 expression on normal granulocytes in vitro at concentrations similar to those measured in vivo. Chemotactic peptide-induced granulocyte aggregation (a reflection of increased cell-to-cell adhesiveness) was specifically blocked by mouse monoclonal antibodies to Mo1 in vitro. These data suggest that the increased expression of Mo1 on granulocytes in vivo is in part mediated by C5a (and C5a desArg). The quantitative increase in granulocyte-surface Mo1 may provide a mechanism for initiating leukoaggregation, sequestration of granulocytes, and neutropenia during hemodialysis.

The structure of a human serum amyloid A (SAA) genomic clone (SAAg9) has been analyzed and the nucleotide sequence of the coding regions is compared with that of the cDNA for apoSAA1. The leader and coding sequences of exons 2 and 3 are identical to SAA1. However, there are 10 nucleotide and 7 derived amino acid substitutions in exon 4. These changes are identical to the amino acid sequence of the amyloid protein associated with familial Mediterranean fever. In particular, the amino acid substitution (Thr to Phe) at residue 69 of SAA1 may have an important role in this type of hereditary amyloidosis. The genomic clone SAAg9 has been transfected into mouse L cells, and constitutive expression of human specific mRNA and protein were observed in stable transfected clones. The expression of both SAA mRNA and protein were increased by incubation of the transfected cells with purified human interleukin-1 (IL-1), both human and mouse recombinant IL-1, and recombinant human tumor necrosis factor alpha. The induction of SAA is pretranslational and is likely to be mediated by protein factor(s) since incubation with cycloheximide diminished IL-1-dependent increase in SAA mRNA.

The interaction of C3b receptor with C3b, the major cleavage product of C3, elicits important biologic functions, such as enhanced phagocytosis and release of cellular enzymes. We determined the binding kinetics and binding isotherm of C3b-receptor interaction by using human cells and fluid phase C3b generated by trypsin cleavage of purified native C3. 125I labeled C3b was separated into 2 molecular species, a dimer and a monomer by column chromatography. We found that dimeric C3b bound to human erythrocyte C3b receptors with an affinity that was more than 25 times that of the monomer. 125I dimeric C3b did not bind to sheep red blood cells, which lack the C3b receptor, nor to trypsinized or 2-mercaptoethanol treated normal human red blood cells, 2 methods for abrogating the immune adherence activity. Binding of 125I fluid phase C3b dimer to the C3b receptor was specific, saturable (about 90 ng of C3b dimer bound maximally per 1 X 10(9) red blood cells), reversible (in the presence of a 100-fold molar excess of unlabeled ligand), and of moderate affinity (Kd about 9.53 nM). The equilibrium binding constants were similar with the various cells tested. Binding was characterized by rapid on and off rates and did not exhibit ligand cooperativity. This specific interaction reached a steady state within 10 to 15 min at 0 degrees C; 50% of specifically bound ligand dissociated from its binding site on human red blood cells in approximately 1 min at 0 degrees C. The density of C3b receptors on human red blood cells, polymorphonuclear leukocytes, monocytes, and B lymphocyte-enriched preparations was 360, 20,000, 30,000, and 21,000 receptors/cell, respectively.

On the cell intermediate EAC142, made with limited C1, excess C4 and excess C2, more than one (up to six) hemolytically effective complexes may be present at a single SAC1, i.e., complexes cluster at a singl hemolytice site. These cells lyse more rapidly, utilize C-EDTA more efficiently and have more C3 convertase activity per hemolytic site than EAC142 made with excess C1 and C4 and limited C2, although both cell types lyse to the same extent with excess C-EDTA. Multiple complexes per hemolytic site were generated with guinea pig and with human C1, C4 and C2.

A C4 cDNA clone has been used to localize the C4 genes within the human major histocompatibility complex (MHC). The clone, pC4AL1, detects a DNA polymorphism of the C4 genes in Southern blot hybridization experiments and established the genotype of a lymphoblastoid cell line, LCL 721, as heterozygous for the C4 DNA patterns. Subclones of LCL 721 that had gamma-ray-induced lesions of the MHC were analyzed with pC4AL1. Loss or retention of chromosome-specific C4 DNA patterns relative to the loss or retention of other MHC genes on the same chromosome was assessed. This permitted the mapping of the C4 genes, and the other MHC-linked complement genes, between the HLA-DR and HLA-B loci.

The NH2-terminal structures of the intracellular precursor of the third component of guinea pig complement (pro-C3), synthesized in macrophage cultures, and beta chain of C3 from guinea pig plasma were determined. Six of the first 16 residues of pro-C3 were identified by a microradiosequencing method. All six are nonpolar and are identical with the residues in beta chain of native plasma C3. This establishes beta chain as the NH2-terminal subunit in pro-C3. A comparison of primary structures of guinea pig and human C3 beta chains revealed identity of at least 8 residues within the NH2-terminal decapeptide. Incubation of plasmin with radiolabeled pro-C3 in macrophage lysate resulted in loss of pro-C3 and the generation of a two-chain molecule electrophoretically indistinguishable from native C3. These data indicate similarities between pro-C3 and the precursor of the fourth complement component, pro-C4, with respect to subunit organization, presence of phylogenetically conserved NH2-terminal regions, and the probable mechanism for generation of their respective native proteins.

A patient with striking susceptibility to infection by pyogenic organisms was found to have 1/1000th or less of the normal serum concentration of C3. Investigation of members of her family revealed many individuals, including her mother and father, with C3 levels approximately half-normal, and it seems that she is homozygous for C3 deficiency.