Kenneth T. Miyasaki DDS PhD kmiyasak at
Thu Jan 2 19:45:45 EST 1997

Complement (C) is a component of the immune system formed by an
interacting network of plasma glycoproteins and cell receptors. The
complement molecules have a complex origin, many of them arising as
strange mosaics of diverse genes spliced together. Some that appear to
share a common origin aren't encoded on the same chromosome. Others that
share a common origin and share the same chromosomal localization look
completely dissimilar ultrastructurally. 
	In stark contrast to the uniqueness of the complement proteins is their
incredibly non-descriptive nomenclature. This is unfortunate, since it
is rather difficult for students to try to generate interest in a field
in which everything looks the same!
	The complement system (1) provides chemoattractants to recruit
phagocytes to a site of inflammation, (2) provides opsonins for
phagocytes which enables the phagocyte to attach to and ingest the
target microbe, (3) generates kinin-like and anaphylatoxic activities
fundamental to the erythematous and edematous signs associated with
inflammation, (4) confers protection against a number of bacteria by
virtue of its bactericidal and bacteriolytic activity, and (5) serves as
a homeostatic mechanism which mediates the clearance of  antigens from
local and systemic fluid spaces.
	Most complement components are synthesized by hepatic parenchymal cells
(the liver) in a fairly constitutive manner, but many components may
also be synthesized by macrophages and monocytes, including C1, C2, C3,
C4, C5, factor B, C1-INA, Factor D, Factor H, DAF, and several other
complement components. As such, complement clearly is a part of the
immune system. Although complement synthesis does not appear to require
induction, it is also known that certain substances (such as
lipopolysaccharide) can greatly increase the levels of complement mRNA
transcripts by hepatocytes. Some complement components are produced by
vascular endothelium, including C3, factor B, and factor H. Factor D
appears to be made by fat cells (adipocytes). Skin fibroblasts make C1,
C2, C3, C5, factor B, factor H, and C1-INA. Complement accounts for
about 3.5 mg/ml of the total serum protein (normal total serum protein
is about 60-78 mg/ml). 
	The most important thing the complement system does is form C3b from
     Of the complement components in serum, C3 has the highest
concentration (1000-1600 µg/ml). C3 is encoded separately from all the
other complement components on chromosome 19, nevertheless, C3, C4, and
C5 belong to the same family of unusual molecules which possess an
internal thioester  bond. This family also includes a-2-macroglobulin.
The thio-ester bond gives these molecules unique and important biologic
	Reactivity of the thioester group.The thioester of C3 is reactive in
two ways. It can either be hydrolyzed by water via the "Tickover
mechanism" or it can be hydrolyzed by molecules on biological surfaces.
	We will divide the complement system into pathways leading to and
leading away from C3b. Complement exhibits two main pathways leading to
C3b, the alternative and the classical(Loos. 1985. In: Loos (ed.) Curr.
Topics Microbiol. Immunol.   121: 7-18). The alternative pathway is
evolutionarily older, utilizes mainly C3 and factor B, and does not
depend upon antibody. If we were to examine the phylogeny of complement,
we would observe that the alternative pathway was in place in Agnatha
(jawless fish) and today can be found in lampreys and hagfish (Farries
and Atkinson, 1991. Immunol. Today 12: 295-300). In contrast, the
classical pathway of complement activation had to await the development
of the antibody system, and thus did not exist until Chondrichthys (now
represented by cartilaginous fish). Be that as it may, the classical
pathway was discovered before the alternative pathway, and thus, the
German name, “komplement” (rather than the French name, “allexin”) was
coined to emphasize that the system was thought to be no more than a
complementary system of antibody. The classical pathway is usually
activated by antibody (IgG or IgM) complexed with antigen, always
involves C4 and C2, and usually involves C1. Not only can it be said
that the alternative system doesn’t usually depend upon antibody, but
also that there exists one (probably more) antibody-independent
classical pathways (which are undoubtably of some evolutionary
	Pathways away from C3b. After C3b is formed by either the classical or
alternative pathways, subsequent events are dictated by the presence or
absence of various regulators of complement activation (RCA). The major
effects generated in this stage include amplification, anaphylatoxins,
chemoattractants, membrane attack complex formation, inactivation,
opsonization, and B-cell enhancement. The terms “common pathway” or
“terminal pathway” have been used to describe the events leading to the
membrane attack complex (C5b, C6, C7, C8, poly-C9), which reflected the
bias of complementologists to outcomes mediated by molecules which are
not associated with cells. To me, this pathway must be relegated to a
more subordinate position as simply one of the outcomes of the
generation of C3b. However, the student will see such terms in the
literature, and they reflect the notion that fluid phase molecules of
the complement system are distinct from the cell-associated molecules of
the system. 

	The players. The older, alternative pathway depends primarily upon C3
and factor B, which exhibit features that reflect those of the
primordial complement proteins. Factor B is a 90 kdal zymogen
(proenzyme) of a serine protease designated “Bb.” Serum concentrations
of factor B are as high as 225 µg/ml. Factor B is encoded in the major
histocompatibility complex (MHC) gene locus on chromosome 6, and like
the other MHC molecules, is among the most pleomorphic molecules in man
(ie., your factor B is likely to be a bit different, structurally and
functionally, than my factor B).
	Factor D (adipsin). Factor D is a 24 kdal serine protease which unlike
factor B, has no zymogen form in serum, and as a matter of record, is
not synthesized as a zymogen (a rare feature). It has also been called
"adipsin"  by complement experts such as Hans J. Müller-Eberhard (1992.
In: Gallin, Goldstein, and Snyderman (eds) Inflammation: Basic
Principles and Clinical Correlates.  Raven Press. NY. pp 33-61).  This
is based upon the finding that mouse adipsin shares 61% homology with
human factor D (Rosen  et al.,  1989. Science 244: 1483-1487). At
present, the experts believe that human adipsin is the same as human
Factor D. Factor D is responsible for activation of factor B by cleavage
of a single Arg-Lys bond in B to form Ba and Bb (below).
 	Factor P (properdin). At one time, the alternative pathway was called
the "properdin pathway." This is because properdin was once thought to
be the main component of the alternative pathway. Now, factor P has been
relegated to a supportive role. Factor P is a 53 kdal protein (encoded
on the short arm of the X chromosome) which tends to form cyclic dimers,
trimers, tetramers, and larger oligomers. It is an elongated structure,
it is thought to stabilize C3 and C5 convertases.
	Tickover Mechanism. There is a spontaneous, slow hydrolysis
("tickover") of the internal thioester bond of C3, forming C3-H2O.
Hydrolysis of the thioester bond promotes the binding of factor B (in
the presence of Mg++) in the fluid (plasma) phase. The C3-H2OB complex
is bound specifically by factor D. Factor D cleaves a 30 kdal fragment,
Ba, off factor B when it is complexed with  C3-H2O to create the active
enzyme, C3-H2OBb. Fluid phase C3-H2OBb is a fluid phase C3 convertase.
Note that the tickover is random, happens all the time, and does not
require any activator (such as a microbe) to initiate. The fluid phase
C3 convertase catalyzes the cleavage of C3 to C3a and C3b, in the fluid
	Initiation of the alternative pathway. If a suitable surface is near
enough, the alternative pathway may be initiated by the C3b which has
been generated by the “tickover mechanism.” C3b has an exposed
hydrophobic region, which tends to make it associate with membranes.
Fluid phase C3b either binds to a membrane or interacts in the fluid
phase with factor H and is destroyed (discussed below in section on the
RCA). C3b binds to membranes using its exposed internal thioester bond,
which may be hydrolyzed by surface hydroxyls (or amines) and form
metastable covalent ester linkages. Often, the  hydroxyl groups are
present in microbial carbohydrates such as lipopolysaccharide (LPS),
zymosan (cell wall polymer of Saccharomyces cerevisiae, baker's yeast),
inulin (a polyfructan), dextran sulfate (a polymer of glucose with
sulfate groups attached). The C3b thioester can also form covalent amide
linkages with amine groups in proteins. The end result of these
activities is the formation of target-bound C3b. Target bound C3b
exhibits a hydrolyzed thioester, much like the fluid phase C3-H2O. As
such, the binding of C3b to a target signals factor B to then bind to
the target bound C3b. Factor D cleaves Ba off factor B when it is
complexed with target bound C3b, to create the active membrane bound C3
convertase, C3bBb(Mg++). The membrane bound C3 convertase produces more
C3b, which in turn, leads to more membrane bound C3 convertases
(amplification, discussed below).

	The classical pathway is best known for tying together specific
antibody responses and complement function. Let us examine how the
classical pathway forms C3b. The main players of the classical pathway
are the C1 complex, C4, and C2. Other elements include C1-inactivator
and mannose-binding protein.
	The C1 complex is a 410 kdal structure which traditionally is said to
consist of five protein subunits: C1q, 2 C1r, and 2 C1s, stabilized by
interaction with Ca++. C1q is a hexamer, consisting of 6 globular
carboxy terminal heads and 6 collagen-like stems. Similar to collagen,
C1q interacts with Ca++, which is important since this permits the
association of C1q with C1r and C1s. The requirement for Ca++ helps
distinguish the classical and alternative pathways. C1q classically
binds to IgG and IgM, via its globular heads. These globular heads may
bind other, as yet undefined structures, and binds C-reactive protein by
a site distinct from that involved in the binding of immunoglobulin.
	C1r and C1s. Two units of C1r and two units of C1s are associated with
C1q, apparently associated with the collagenous regions of C1q. C1r and
C1s are similar, bilobed proteins, thought to have arisen from the same
ancestral protein. Both are serine esterases with the unique, important
property of not being inhibitable by the serum regulatory protein,
alpha-1-antitrypsin. Both are 85 kdal in MW and share a great deal of
amino acid sequences. Both are initially present as zymogens.
	 C1q belongs to a family of similar molecules which includes
mannose-binding protein, conglutinin, and lung surfactant proteins (SP)
A and B. Like C1q, these molecules feature a collagen-like triple helix
and a carboxy-terminal globular head  (Sim and Reid. 1991. Immunol.
Today  12: 307-311). The globular heads of Man-BP, conglutinin, and SP-A
are lectins. SP-A is a hexameric molecule which looks nearly identical
to C1q by electron microscopy, but Man-BP appears to form a mixture of
trimers, tetramers, pentamers, and hexamers. Conglutinin forms only
tetramers and is somewhat longer in appearance. Importantly, Man-BP
activates C1r and C1s in a Ca++-dependent manner completely analogous to
C1q, except antibodies are not required. In this case, Man-BP can
interact with high mannose structures, possibly the mannans found on
fungi; although interestingly, it would appear that Man-BP is misnamed
and has a higher affinity for glucosamine. Neither conglutinin nor SP-A
are able to activate the classical pathway.
	C4 is a 200 kdal heterotrimeric molecule consisting of three subunits,
a (93 kdal), b (75 kdal), and g (33 kdal). Cleavage of the a chain
results in the formation of C4a (9 kdal) and C4b (190 kdal). C4 exists
in two isotypes, C4A and C4B, and both are encoded in the MHC gene
complex on chromosome 6. C2 is a 102 kdal zymogen with 39% structural
homology to factor B. C2b is homologous to Factor Bb, and contains the
active enzyme site. C2 is also encoded in the MHC. 
	C1-INA. C1 inactivator (C1-INA, C1 inhibitor) is a 109 kdal
glycoprotein encoded on chromosome 11 which specifically binds C1r or
C1s. Binding of C1-INA to C1r blocks any possible C4 activation because
C1s remains in the zymogen (inactive) state. C1-INA is a member of the
"serpin" superfamily which includes a-1-antitrypsin and
	The classical pathway is triggered by the binding of IgG or IgM to a
specific target, which exposes a binding area on the antibody which
interacts with the globular heads of C1q. Usually, C1q binds to
antigen-bound antibody of the IgG or IgM subclass; however, classical
pathway activation can occur in the absence of bound IgG or IgM as
discussed below. In order to activate C1q, its binding to antibody must
be divalent or multivalent. 
	The binding of C1q to antigen-antibody results in a conformational
change in C1q which is transmitted to one of the C1r  molecules. The
nature of this conformational change is unknown, however it is believed
that one C1r is pulled farther away from C1s. C1s apparently provides an
inhibitory contact for C1r. This transitional state has been referred to
as the “open complex” state. The open-complex-state C1r which becomes an
active enzyme that acts to cleave the remaining C1r to form a second
reactive enzyme, which reciprocally, cleaves the open-complex-state C1r.
Both of these enzymes are specific for C1s, and cleave that zymogen to
form an active serine protease which is highly specific for  either C4
or C2. The activated C1s has also been referred to as a C4bC2b
	C4 is highly sensitive to the proteolytic activity of activated C1s,
and is cleaved preferentially over C2. The cleavage products are C4a and
C4b. C4a is a 6000 dalton peptide which exerts biologic activity
described later. Removal of the C4a fragment from C4b results in the
exposure of a labile, hydrophobic zone which leads to the association of
C4b to hydrophobic areas (membranes). After this initial attraction, C4b
may form a metastable complex with hydroxyls or amides via its thioester
bond, much like C3b. If C4b cannot find a suitable area to bind, it is
broken down to iC4b, terminating the complement cascade. The activated
C1s will continue to produce C4b molecules, and eventually, a number of
C4b molecules may be bound to the membrane in the vicinity of the C1
complex. It has been estimated that each activated C1s can cleave about
35 C4 molecules of before being inactivated by C1-INA. A substantial
number of the C4b molecules formed remain in the fluid phase, and their
thioester bond hydrolyzes against water, and only a small percentage of
the fluid phase C4b molecules ever become covalently associated with a

	As in the association of C3b with the surface, C4b uses its reactive
thioester bond to associate with surface hydroxyls or amines. The
hydrolysis of the thioester bond signals factor C2 to bind to C4b in the
presence of the divalent cation, Mg++ (totally analogous to the
association of C3b and factor B). Bound C2 is sensitive to proteolytic
enzymes, especially, C1s. C2 is cleaved to C2a (35 kdal) and C2b (75
kdal). C2a dissociates into the fluid phase, leaving a trimolecular
complex, C4bC2bMg++, associated with the target membrane. The C4bC2bMg++
complex possesses enzymatic activity and is referred to as a membrane C3
	C4bC2bMg++ is a C3 convertase analogous to C3bBbMg++ of the alternative
pathway. However, unlike that enzyme, C4bC2bMg++ does not amplify itself
and does not appear to be as efficient as the alternative pathway C3
convertase (this is more than compensated by the much greater activation
efficiency of C1 compared to the tickover mechanism). Additional
classical pathway C3 convertases must be made via C1, as described
above; however, this does not preclude amplification via formation of
	 C1-INA. The flexible C1 complex has a tendency to activate itself
spontaneously. Spontaneous activation of C1 is prevented by C1-INA,
which binds stoichiometrically to the C1r procatalytic sites, thus
preventing the conformational changes which would activate these sites.
C1-INA does not block C1 activation that has been initiated by antibody.

The Regulators of 
Complement Activation

	Once C3b is formed, the complement system must be given guidance. It
needs to know what to do with the C3b. This role is given to the 
products of the regulators of complement activation (RCA) gene region on
the long arm of chromosome 1 and to a protease known as factor I. (C3b4b
INA). Factor I is an 88 kdal heterodimeric neutral serine protease
encoded on chromosome 4 with high specificity for C3b and C4b. Factor I
is the enzyme responsible for modifying C3b. Factor I requires a 
cofactor to tell it what to do. This is where the RCA gene region
products come into play. 
    The RCA gene region products are also known as the "C3b binding
protein family." Actually, some bind C3b and others bind C4b.
Regulators  which  to interact with factor I promote the further
metabolism of C3b (or C4b) along appropriate pathways. Either factor H,
CR1, or MCP direct Factor I to proteolytically remove a 2-3 kdal
fragment (C3f) from C3b thereby forming iC3b. From iC3b, Factor I can
form C3dg, in which factor H (or the B-cell CR1) serves as the cofactor
of choice. Factor I also requires C4 binding protein (C4Bp) or CR1 to
hydrolyze C4b. The ability to bind C3b and C4b evolved long ago (600
million years), and seems to be associated with the presence of a
conserved sequence of 60-70 amino acids called a "short consensus repeat
(SCR)." Besides the RCA gene region products, many other molecules in
the complement system bind to the "C3b, C4b, C5b family" of molecules
and possess SCR (Farries and Atkinson, 1991). For example, factor B, C2,
and C6 bind to C3b, C4b, and C5b, respectively, and also possess the SCR
elements. This indicates that SCR are quite old and fundamental to the
complement system.
	Now remember, both the classical and the alternative pathway result in
the formation of C3b. After this point, the precise fate of the C3b
molecule is dictated by the presence or absence of regulator proteins
(CR1, CR2, Factor H, DAF, MCP, and C4bp). Let us consider the various
options now available once C3b is formed.

    The generation of C3b may lead to amplification if no cofactors are
present. Amplification refers to the generation of more C3b and more C3
convertases, which can be an end in itself. Amplification occurs as C3b
interacts with fluid phase factor B. This leads to the formation of
C3bB, which is then metabolized to C3bBb (C3 convertase).These
additional C3b and C3 convertases can eventually form a C5 convertase
(as discussed below), an enzyme which splits C5 into C5a and C5b.

Factor H, A Fluid 
Phase Cofactor
	Factor H is a 160 kdal filamentous (28 nm long), flexible protein which
is an important cofactor in the inactivation of C3b by factor I. In this
regard, factor H can serve as a cofactor for two proteolytic cleavages
by factor I which result in the excision of a small piece (C3f) from
C3b. The remaining large portion of C3b then becomes inactivated with
respect to further complement activation, thus this piece is designated
“iC3b.” Although iC3b is inactivated, it maintains covalent association
with the surface to which it is attached. Receptors for factor H are
also found on neutrophils, monocytes, and B-cells. Factor H can
stimulate the release of factor I from monocytes via interaction with
the factor H receptor. As you may notice, factor H can serve as a
cofactor regardless of the presence or absence of a cell (contrast that
with CR1). 

CR1 and MCP, 
Membrane-Bound Cofactors
	Complement receptor type 1 (CR1, CD35) is a transmembrane glycoprotein
found on phagocytes and B-cells at a density of about 3000-30,000
molecules per cell. CR1 occurs in four or more allotypes with molecular
weights between 160-250 kdal. It promotes cleavage of C3b to iC3b by
factor I, as does factor H. CR1 serves  as a cofactor to factor I in
clipping iC3b to form C3dg. CR1 is thus an important cofactor which
enables phagocytes and B-cells to generate an ingestible surface (iC3b
for phagocytes) or enhance antigen-specific B-cell endocytic events
(C3dg for B-cells). Finally, CR1 binds C4b and serves as a cofactor for
Factor I in the destruction of C4b.
	Membrane cofactor protein (MCP, gp45-70, CD46) is a 45-70 kdal host
integral membrane glycoprotein which is widely distributed in tissues.
Unlike DAF (below), MCP shows cofactor activity: it binds C3b and
permits factor I  to cleave C3b not only to iC3b, but subsequently, to
cleave factor iC3b to C3c and C3dg. Both DAF and MCP serve as a means of
protecting the host cell against C3b mediated phagocyte damage and 
common pathway destruction.

DAF, a Membrane-Bound Regulator 
which is not a Cofactor
	Decay accelerating factor (DAF, CD55) is a 75-80 kdal intramembrane
glycoprotein. DAF is widely distributed in host tissues and is anchored
in membranes via a glycophosphatidyl inositol (GPI, aka "PIG" for
phosphatidylinositol glycans) group (as opposed to being a transmembrane
protein). GPI anchorage facilitates more rapid diffusion of the protein
within the plane of the membrane, an important consideration for a
protective molecule. An alternative thought is that GPI does not
transmit signals across the membrane, thereby permitting the binding of
ligand without activation of the cell. The diffusion of the GPI-anchored
proteins among cells can occur; ie., one cell can give another cell it’s
surface molecule (Kooyman et al. 1995. Science 269:89-92). DAF is the
host cells primary protection agains C3 convertase. DAF specifically
binds to C3bBb and C4b2b (the C3 convertases). It causes the convertases
to fall apart, and then, by binding either C3b or C4b, prevents any
further formation of a C3 convertase. Although DAF may be considered a
regulator, it is not a cofactor. 

RCA Gene-Region Product that is 
Neither a Regulator Nor a Cofactor
Complement receptor type 2 (CR2, CD21) is a 140 kdal RCA gene region
product which binds C3dg (and iC3b, to a lesser extent), at a binding
site near that which is bound by factor H. CR2 is found on B-cells and
stratified squamous epithelium. CR2 is perhaps best known as the
receptor for the Epstein-Barr Virus (EBV). EBV is the agent of
infectious mononucleosis and Burkitt's Lymphoma (a B-cell neoplasm). CR2
doesn't function as a regulator or a cofactor but instead, appears to
enhance specific B-cell responses.

Classical Pathway 
Regulator and Cofactor
	C4bp is a 570 kdal homoheptameric cofactor glycoprotein consisting of 7
identical filamentous subunits arranged in a spider shape. This molecule
binds to fluid phase C4b, is a member of the RCA gene region, and
hastens the destruction of C4b by factor I. Thus, C4bp appears to be
designed to regulate the classical pathway. It is likely that C4bp and
DAF became necessary with the evolution of the classical pathway. Even
though the other cofactors can completely block the alternative pathway,
they do not shutdown the classical C3 convertase (remember, the
classical C3 convertase is comprised of C4bC2b Mg++).

The Events which Occur in 
the Absence of Regulators
	Amplification (the formation of more C3 convertases and C3b) occurs in
the absence of RCA gene region products. C3b is generated and
amplification can occur via the binding of the alternative pathway
component, factor B. The bound factor B is hydrolyzed by factor D
resulting in the formation of target-bound C3bBb, which is a C3
convertase. Additional stabilization of C3bBb is achieved if properdin
(factor P) is bound. Stability of this complex is about 30 min. This
entire complex  is also a C3 convertase.  If a sufficient amount of C3b
clusters about the C3 convertase, a C5 convertase is formed. The diagram
shows the bound C3b initially interacting with Factor B (interestingly,
this initial binding may be mediated by the Ba portion of Factor B) in
the presence of Mg++. Factor B is cleaved by factor D, forming the
active C3 convertase, which rapidly makes more C3b molecules. Once a
membrane bound C3 convertase is formed (by either the classical or the
alternative pathways) the alternate pathway can serve to amplify the
process. In this case, the membrane bound C3 convertase can generate
more C3b which can also bind to the target membrane. These additional
bound C3b molecules can bind to more molecules of factor B and form
additional C3 Convertases. Again, factor P may bind and help stabilize
the complex.

Formation of 
C5 convertase 
	Formation of C5a and C5b. The addition of C3b clusters to the C3
convertase forms a C5 convertase. The alternative pathway forms a C5
convertase with the formula (C3b)2 or moreBbMg++ ± P. The classical
pathway forms a C5 convertase with the formula (C3b)1 or moreC4bC2bMg++.
The C3b does not actually appear to bind to the C3 convertase, but
instead, numerous C3b molecules cluster around the C3 convertase.
Interestingly, rather than altering the specificity of the enzyme, it
appears that the C3b molecules alter the incoming C5 molecule, rendering
it susceptible to conversion by Bb. This may be desirable in that the
formation of the C5 convertase does not shut down the activity of the C3

C5 Convertase 
	The formation of the C5 convertase has two biologically significant
results: (1) the formation of C5a, the major complement chemoattractant
for host leukocytes, and (2) the initiation of the membrane attack
complex by the formation of C5b. The Membrane Attack Complex (MAC)
consists of a dimeric series of complement proteins, C5b, C6, C7, and C8
and a polymer of C9. The major protein in this group on a molar (and
wt/vol) basis in serum is C9. The function of the MAC is to form
transmembrane channels across exposed lipid membranes.

Components of the 
Membrane Attack Complex
	C5 (encoded on chromosome 9q34.1) is a 191 kdal heterodimeric protein
consisting of one115 kdal and one 75 kdal subunit. Cleavage of the
larger subunit by the C5 convertase results in the formation of a 11.2
kdal C5a chemoattractant molecule and 180 kdal C5b. C5 is structurally
related to C3 and C4, and has the internal thioester bond that dominates
the behavior of these molecules.
	The membrane attack complex family consists of a group of related
molecules which are involved with the membrane lytic function of
complement. C6, C7, C8 (there are two isomers of C8: C8a and C8b
isomers), C9, C8-binding protein (discussed below), and T-cell perforin
are members of this family. However, the membrane attack complex
consists of only C5b, C6, C7, C8, and C9. C6, C7, and C9 are single
chain glycoproteins encoded on chromosome 5 (5p13). C6 and C7 have
approximate molecular weights near 110 kdal. C8 and C9 are structurally
related molecules. C8 is a 151 heterotrimer, consisting of two different
64 kdal chains and one 22 kdal chain (designated a, b, and g,
respectively). C9 is a 71 kdal glycoprotein which polymerizes in
membranes. C9 shares substantial structural homology with the cytotoxic
lymphocyte pore-forming protein called T-cell perforin.

attack complex 
	C5b interacts with the target via its thioester group to form a
covalent bond with surface hydroxyls or amines. When you think about it,
this may explain why we need so many other proteins to attack the
membranes. The problem that the MAC attack addresses is one involving
bridging the gap between hydrophilic milieus to the more hydrophobic
cell membrane. The binding of C5b initiates a series of events which
leads to the successive binding of C6, C7, C8, and C9. Addition of C8 to
the complex results in a slight increase in membrane permeability. C8 is
capable of producing limited membrane damage, forming channels
approximately 1 nm in diameter. According to current thought, more
critical function of the C5b-C8 dimer is to induce the polymerization of
C9. This process involves the binding of C9 to C8. The binding of C9 to
C8 results in a conformational change in C9, exposing a previously
sequestered, hydrophobic zone. The hydrophobic zone then penetrates the
membrane and eventually, as polymerization occurs, forms a dough-nut
like polymer in the membrane with a hydrophilic pore. Poly-C9 produces a
large transmembrane channel (10 nm). Polymers of C9 have been estimated
to be between 12-21 C9 units. The transmembrane channels have been shown
to permit the flux of potassium outward and that of  calcium and sodium

Events which occur in 
the Presence of Regulators
	Opsonization (formation of iC3b). Opsonization is the preparation of a
microbial surface for ingestion (ie., coating the microbe with
self-derived molecules) by phagocytes (macrophage and neutrophils).
Although the complement system coats microbes with C3b, phagocytes have
chosen not to ingest particles with C3b bound to the surface. Why not?
Well, perhaps the phagocytes are being polite, waiting for C3b to do
whatever else it has to do on that surface. Or maybe it’s for their own
protection that they inactivate C3b before trying to ingest the particle
that the C3b is coating (analogously, although we like our food hot, we
don't like it on fire). The bottom line is that C3b itself doesn't
signal phagocytosis. Instead, the phagocyte uses one of two cofactor
proteins to further prepare C3b: CR1 or factor H. Either factor H or the
phagocyte complement receptor, CR1, will function to make ready C3b by
assisting factor I in knicking C3b molecule in two spots, thereby
releasing a tiny (2-3 kdal) fragment, C3f, and leaving the bulk of the
molecule covalently attached to the target as an iC3b molecule. Factor H
may do much of this before the phagocyte arrives, but recent studies
show that CR1 is the crucial cofactor for converting C3b to iC3b. It is
probable that most iC3b produced in the presence of factor H is degraded
	On B-cells, CR1 can promote the decay of iC3b to C3dg, but on
phagocytes, this is prevented by two additional complement receptors,
CR3 and CR4 (these are beta 2-integrins). CR3 (Mac-1) and CR4 (p150,95)
bind specifically to iC3b and are the complement receptors which lead to
the ingestion of the iC3b-coated surface. These two receptors associate
with cytoskeletal elements of the phagocyte, and trigger the
intracellular events within the phagocyte which lead to the engulfment
and secretory events associated with phagocytosis. 
 	Notice that opsonization precludes further amplification (in fact, the
binding of either CR1 or factor H  in itself to C3b - in the absence of
factor I - prevents Factor B from binding). Therefore, the student
should remember that opsonization and amplification are exclusive. For
this reason, opsonization also precludes the formation of any additional
membrane attack complexes (by preventing the formation of  C5

	B-Cell “enhancement” (adjuvanticity). B-cells possess CR1 and CR2 but
lack both CR3 and CR4. Therefore, CR1 on B-cells serves as a cofactor
for factor I which permits the formation of iC3b and subsequently, C3dg.
Two proteolytic cleavages are required to obtain iC3b with the release
of C3f, and one additional cleavage results in the release of a fairly
large fragment, C3c. C3dg remains associated with the target surface.
C3dg  is covalently linked to antigen via its thioester, and if it
presents a multivalent complex (many C3dg molecules), then interaction
with CR2 has been shown to promote cell entry into S-phase and can
replace the need for macrophage-derived growth factors (Melchers  et
al.,  1985. Nature 317: 264-267). Monomeric (univalent) forms of C3dg
tends to inhibit B-cell activation. Monomeric C3dg may be the more usual
	CR2 works in association with a cell surface molecule called “CD19.”
CD19 is required for normal antibody responses, and it amplifies BCR
signals via it’s capacity to interact with several intracellular
proteins (including phosphotidylinositol-3-kinase [to be discussed
later]). To elicit the CD19 enhancement, CD19 must be brought into close
association with the BCR (even simple biochemical cross-linkage will
suffice). Thus, it is now believed that CR2 functions as a “coreceptor”
with the BCR for antigen coupled to C3dg. CR2 serves to bring CD19 into
association with the BCR, enabling the enhancing intracellular activity
mentioned above. Enhancement enables the B-cell to respond to antigen
levels several orders of magnitude lower than otherwise (Carter and
Fearon, 1992. Science  256: 105-107) or respond at a much higher level
(1000-10000 fold) to similar amounts of antigens (Dempsey  et al., 
1996.  Science 271: 348-350). In this model, the CR2-CD19 complex
enhances B-cell activation by promoting binding of antigen by both the
BCR and CR2 (Noesel  et al.,  1993. Immunol. Today 14: 8-11). 
	Host protection against C3 convertases. There is nothing special about
the bilamellar structure of the host membrane. The host protects its own
membrane against inadvertant C3 convertase activity using the regulators
of complement activation, DAF and MCP. DAF binds to both C4b and C3b.
DAF can displace Bb from C3b thus inactivating the alternative C3
convertase. Analogously, DAF can displace C2b from C4b, thus
inactivating the classical C3 convertase. DAF subsequently prevents any
other factor B's from binding to C3b and any other C2's from binding to
C4b. Although DAF is a regulatory molecule, it has no cofactor activity.
	MCP has been shown to bind C3b. In a manner analogous to CR1 and Factor
H, MCP is a cofactor for Factor I and generates iC3b from C3b. Thus, the
astute student will notice that DAF appears to be aimed at blocking the
convertase enzymes that make C3b, and MCP appears to be aimed at
destroying the products made by the enzymes (ie., C3b). 

Protection of Host 
Against the MAC Attack 
	As described above, the host cell has several mechanisms of regulating
C3 and C5 convertase activities on its surface. Apparently, this isn't
enough protection. Most cells possess a means of protecting themselves
against the membrane attack complex. Why? Well, as you recall, there are
only three molecules of the complement system which can covalently
associate with the cell via a thioester. The first two, C3 and C4 are
controlled by the RCA gene region products. The third, C5 is not.
Therefore, in a torrid firefight, innocent bystanders can easily become
targets of C5b bullets shot by the C5 convertase. A covalent association
with C5b can have grave consequences for host cells and cell matrices.
The host has several mechanisms of regulating the MAC attack.
Interestingly, the  known regulators are  not "cofactor" molecules which
bind C5b. I don't know why. 
	S protein (vitronectin). The main regulator of MAC is vitronectin,
which binds to C5b-C7. Vitronectin is a connective tissue glue which
promotes cell attachment and binds proteoglycans. It appears to be
multifunctional, and the S-protein also blocks activation of C9
(formation of poly C9).
	Homologous restriction and homologous restriction factors. One of the
earliest observed phenomenon in immunology was that complement from one
species was far more efficient at lysing the red blood cells from a
xenogeneic donor rather than a donor of the same species. This
phenomenon was called “homologous restriction.” Thus, ox complement
lyses guinea pig red blood cells much more efficiently than ox  red
blood cells. Now it is understood that homologous restriction is due to
the specific binding of C8 and C9 by membrane factors called “homologous
restriction factors” and C3 convertases by DAF. The homologous
restriction factors simply “see” homologous (syngeneic or allogeneic) C8
and C9 better than xenogeneic C8 and C9. Similarly, DAF “sees” C3bBb
from the same species better than it sees xenogeneic C3bBb. Protection
against “MAC attack” is performed by a 65 kdal C8-binding protein
(homologous restriction factor, HRF,C8bp) and a 20 kdal HRF called
“CD59” or the “MAC inhibiting factor (MACIF).” CD59 is widely
distributed in host tissues and binds both C8 and C9. Both C8bp and CD59
are GPI-anchored (like DAF). A mechanism for the protective effect of
C8bp has been proposed. C8bp binds very close to the site for C9 binding
by  C8. It thereby prevents C9 from binding to the C9-binding site.
Unable to bind to C8, C9 is unable to unfold and expose its hydrophobic
domain. As such the C9 is unable to insert into the membrane, undergo
polymerization, and form a channel.
	Interestingly, although we normally view the MAC attack on self tissues
with some degree of abhorance, there is mounting evidence that this can
be a very important aspect of our inflammatory responses. For example,
although platelets possess CD59, C5b-C9 appear capable of modulating
intracellular calcium without inducing cell lysis, thereby modifying the
procoagulant activities of the platelets (Sims and Wiedmer. 1991.
Immunol. Today 12:338-342).
	Hyperacute rejection. Another important aspect of homologous
restriction involves pigs. Did you know that pigs have organs (hearts
and so forth) that are considered just about the right size to be
transplanted into humans? Actually, the converse is also probably true.
Nowadays, pig organs are receiving renewed interest as xenografts. In
the past, the problem was that pig DAF had little effect on human C3bBb
and C4b2b. Pig MCP was very ineffective in inactivating human C3b. And
pig CD59 did little to prevent the activity of the human membrane attack
complex. As a result, if you put a pig heart in a human, it would be
rejected within an hour (patients would complain). This rapid rejection,
called hyperacute rejection, is based upon complement, and the inability
of the donor organ to regulate the spontaneous activity of the
recipient’s complement system at the surface of it’s cells. Now,
scientists have begun to develop transgenic pigs which carry the genes
for human DAF, MCP, and CD59. The transgenic pigs have been developed
such that they express high levels of DAF, MCP, and/or CD59 on their
cell surfaces. Organs from such animals transplanted into primates
(whose complement is more like human complement) survive 2-5 days
without immunosuppression (Kingman. 1995. J. NIH Res. 7:30-32). If you
use cyclosporin A or cyclophosphamide to suppress the other, slower form
of transplant rejection mediated by lymphocytes, monkeys can survive
more than two months with pig hearts!

Function of the 
Low Molecular Weight Fragments
	The splitting of complement components C2, C3, C4, and C5 leads to the
generation of both high and low molecular weight fragments. We have
discussed the activities of the high molecular weight fragments, now we
will briefly consider the low molecular weight fragments.
	Kinin-like activity refers to the direct effects that a molecule may
have on the vasculature and smooth muscles. The 35 kdal C2a exhibits
kinin-like activity and causes smooth muscle contraction, pain, and
increase in vascular permeabilization. 
	Anaphylatoxins. The term “anaphylatoxin” was coined to point out C3a,
C4a, and C5a induce a reaction resembling anaphylaxis if injected into
animals. Anaphylatoxic activity is similar in end result to kinin-like
activity, but differs in the involvement of mast cells. The 9 kdal C3a,
9 kdal C4a, and 11 kdal C5a stimulate mast cell degranulation and
arachidonate metabolism (described in the chapter on Phagocytes),
resulting in the release of mast cell mediators histamine and
thromboxane A2, respectively. As such, certain anaphylatoxic effects may
be blocked by antihistamines, and others may be blocked by
cyclooxygenase-pathway inhibitors (like Motrin®). In man, the most
potent anaphylatoxin appears to be C5a. Some classic anaphylatoxic
effects include increased vascular permeability, smooth muscle
contraction, and mast cell and basophil activation. The anaphylatoxins
are regulated by a 300 kdal molecule with carboxypeptidase N activity
called "anaphylatoxin inactivator." 

	Phagocyte behavior. C5a is the primary chemotaxin for host phagocytes
and increases the surface expression of CR1 and CR3 on host phagocytes.
It can also stimulate phagocyte secretion (mainly limited in neutrophils
to the "specific granules"). The proteolytic cleavage of C5 by the C5
convertase into C5a and C5b is actually not essential. A more limited
proteolysis by trypsin leaves C5a and C5b associated by disulfide bonds
yet the molecule becomes a  chemoattractant.
	Complement deficiencies afford insight into the role of complement in
maintaining homeostasis and host defense.
	C1, C2, and C4 deficiencies. Systemic lupus erythematosis (SLE) is a
disease with which the dentist will confront (two diagnostic clinical
signs which are accepted by the American Rheumatism Association include
facial erythema [the famous malar rash] and oral or nasopharyngeal
ulceration. For a more complete description, see Ferrell and Tan. 1985.
In: Rose and Mackay (ed) The Autoimmune Diseases. Academic Press,
pp29-57). The association of C1, C2, or C4 deficiency with lupus
suggests some role of C1, C2, or C4  in the clearance of immune
complexes (IC). As you will learn in your future pathology courses,
subepithelial IC deposits are part of the histopathology of SLE. A good
example of the significance of complement in maintaining IC homeostasis
is the fact that fresh serum can dissolve IC, whereas heat-inactivated
serum cannot. Apparently, the activation of C3 (forming C3b and also
C3d), can result in the dissolution of the IC. Another potential problem
of a person with C1, C2, or C4 deficiencies is deep, recurrent pyogenic
infections. In western European populations, homozygous C2 deficiency is
the most common inherited complement deficiency state (afflicting
1:10,000 to 1:30,000); and tellingly, many of these individuals suffer
no apparent disease. However, forty per cent of individuals with this
disorder develop some systemic autoimmune disease or SLE-like disease at
some point. 
	C3 deficiency and Factor I and Factor H deficiency. C3 deficiency is
grave. Patients with C3 deficiency suffer severe disseminated infections
with "pyogenic," serum-resistant bacteria. Luckily, only sixteen cases
have lived long enough to be described in humans. C3 deficiency can
arise either as a primary genetic disorder or as a result of unregulated
depletion of C3. Unregulated depletion of C3 can occur as a result of a
genetic deficiency of either factor H or factor I. In these cases,
severe pyogenic infection can result secondarily from a depletion of C3
(due to unregulated amplification). Negative regulatory mechanisms are
thus important in preventing the consumption of complement components.
	C1-INA (C1 Inhibitor) deficiency. C1-INA deficiency can result in
angioedema. This is a disorder for the doctor to remember, since the
signs include swollen mucosae (most apparent are swollen lips).
Angioedema of the oropharynx can lead to life-threatening airway
obstruction. Angioedema results from the disregulation of C1r (and other
serum proteases which C1-INA also regulates, notably kallikrein and
bradykinin) which normally is prevented from spontaneous activation by
C1-INA. This results in increased complement activation, and the
formation of C4a, which has anaphylatoxin effects and causes mast cell
degranulation (which results in increased vascular permeability), and
perhaps C2a, which has kinin-like activity (among other functions,
kinins increase vascular permeability) and leads to edema. There are two
forms of this disease (inherited and acquired). The inherited form is
autosomal and is a result of a gene defect. Although a person inherits
two genes for C1-INA (one from mom and one from dad), apparently, one
good gene is not sufficient to keep up with the rate of depletion of
C1-INA. Therefore, the disease is autosomal dominant.  Acquired C1-INA
(C1 Inhibitor) deficiency occurs as a result of an intense expenditure
of C1-INA as a result of some increase in C1 activation. The C1
activation can be indicative of autoimmunity, in which the body is
reacting with complement fixing antibodies in some massive fashion
against itself. For example, B cell lymphomas usually lead to acquired
angioedema probably because of an antiidiotype response against the
lymphoma cells.
	C5-8 deficiencies are associated with recurrent Neisserial
(meningococcal and gonococcal) infections. These organisms possess
either a capsular polysaccharide which precludes successful phagocytosis
or can invade cells and propagate in an intracellular manner. In either
case, the phagocyte is inneffective, but happily, these organisms are
often serum-sensitive. Thus, serum killing and the terminal complement
pathway is crucial in controlling these organisms. 
	Rarely, SLE-like diseases have been associated with these deficiencies.
The SLE-like diseases may be secondary to complement depletion resulting
from recurrent or long term infections rather than to the terminal
complement components playing some role in immune complex clearance. 
	The occurrence of the terminal pathway deficiencies show ethnic
variability and also variability as to the incidence. The incidences of
C5 deficiency is rare (27 cases); C6 deficiency is 1:60:000
(Caucasians); and C7 deficiency is less common in Caucasians than C6
deficiency, but has an incidence of 1:25,000 in Japanese. C8 deficiency
has been reported (incidence is not known); however, because C8 is
encoded on three different genes, there appear to be differences in the
type of C8 deficiencies (C8b deficiency has been found in Caucasians,
but C8a or C8g defiencies are more often uncovered in Asian and African
	C9 deficiency may be important. It is common among Japanese (1:1000),
who seem to suffer no ill effects. It is a rare occurrence in
Caucasians; and of the five known cases, three suffered from
meningococcal infection. The reason for this disparity may be related to
the the incidence of these infections within each population rather than
a greater dependence upon C9 among Caucasians (Caucasians in Europe have
a meningococcal infection incidence of 1:10,000 compared to an incidence
of 1:1,000,000 in the Japanese population....Unfortunately, now we have
to explain why those differences occur!).
	GPI protein deficiency. As mentioned above, several protective,
membrane control proteins are GPI-anchored, including DAF, CD59, and
C8bp. Individuals with a lack of GPI anchored proteins are prone to
paroxysmal nocturnal hemoglobinuria (PNH), a disease in characterized
by  intravascular destruction of red blood cells (hemolytic anemia),
paroxysms brought about by oxygen insufficiency, and excretion of
hemoglobin by the kidneys. Thrombotic events also occur as a result of
platelet activation. The lack of DAF, CD59, and C8bp probably
contributes to the disease pathogenesis. More specific deficiencies
provide insight: both DAF and CD59 deficiencies have been documented.
DAF deficiency  leads to no clinical problems. CD59 deficiency has been
associated with a mild PNH type of anemia. These observations suggest
that CD59 is more important than DAF in preventing PNH (You should be
able to explain this based upon what you know about DAF and CD59). C8bp
deficiency has not yet been documented.

More information about the Immuno mailing list