The proteins of the classical and alternative pathways are assigned numbers (e.g. C1, C2). Many of these are zymogens (i.e. pro-enzymes that require proteolytic cleavage to become active). The cleavage products of complement proteins are distinguished from parent molecules by suffix letters (e.g. C3a, C3b). The proteins of the alternative pathway are called ‘factors’ and are identified by single letters (e.g. factor B, which may be abbreviated to fB or just ‘B’). Components are shown in green, conversion steps as white arrows, and activation/cleavage steps as red arrows. The classical pathway is activated by the cleavage of C1r and C1s following association of C1qr₂s₂ with classical pathway activators (see Fig. 4.3), including immune complexes. Activated C1s cleaves C4 and C2 to form the classical pathway C3 convertase . Cleavage of C4 and C2 can also be effected via MASP-2 of the lectin pathway, which is associated with mannan-binding lectin (MBL) or ficolin. The alternative pathway is activated by the cleavage of C3 to C3b, which associates with factor B and is cleaved by factor D to generate the alternative pathway C3 convertase . The initial activation of C3 happens to some extent spontaneously, but this step can also be affected by classical or alternative pathway C3 convertases or a number of other serum or microbial proteases. Note that C3b generated in the alternative pathway can bind more factor B and generate a positive feedback loop to amplify activation on the surface. Note also that the activation pathways are functionally analogous, and the diagram emphasizes these similarities. For example, C3 and C4 are homologous, as are C2 and factor B. MASP-2 is homologous to C1r and C1s. Either the classical or alternative pathway C3 convertases may associate with C3b bound on a cell surface to form C5 convertases, , or , which split C5. The larger fragment C5b associates with C6 and C7, which can then bind to plasma membranes. The complex of C5b67 assembles C8 and a number of molecules of C9 to form a membrane attack complex (MAC), C5b–9.

The classical pathway links to the adaptive immune system

The classical pathway is activated by antibody bound to antigen and requires Ca²⁺

Only surface-bound IgG and IgM antibodies can activate complement, and they do so via the classical pathway. Surface binding is the key:

• IgM is the most efficient activator, but unbound IgM in plasma does not activate complement;

• among IgG subclasses, IgG1 and IgG3 are strong complement activators, whereas IgG4 does not activate because it is unable to bind the first component of the classical pathway.

Q. What occurs when IgM binds to the surface of a bacterium that allows it to activate complement?

A. A transition occurs from a flat planar molecule to a staple form, which exposes binding sites for the first component of the complement system, C1 (see Fig. 3.6).

The first component of the pathway, C1, is a complex molecule comprising a large, 6-headed recognition unit termed C1q and two molecules each of C1r and C1s, the enzymatic units of the complex (Fig. 4.5). Assembly of the C1 complex is Ca²⁺-dependent, and the classical pathway is therefore inactive if Ca²⁺ ions are absent.

Fig. 4.5 Structure of C1

Electron micrograph of a human C1q molecule demonstrates six subunits. Each subunit contains three polypeptide chains, giving 18 in the whole molecule. The receptors for the Fc regions of IgG and IgM are in the globular heads. The connecting stalks contain regions of triple helix and the central core region contains collagen-like triple helix. The lower panel shows a model of intact C1 with two C1r and two C1s pro-enzymes positioned within the ring. The catalytic heads of C1r and C1s are closely apposed and conformational change induced in C1q following binding to complexed immunoglobulin causes mutual activation/cleavage of each C1r unit followed by cleavage of the two C1s units. The cohesion of the entire complex is dependent on Ca²⁺.

(Electron micrograph, reproduced by courtesy of Dr N Hughes-Jones.)

C1 activation occurs only when several of the head groups of C1q are bound to antibody

C1q in the C1 complex binds through its globular head groups to the Fc regions of the immobilized antibody and undergoes changes in shape that trigger autocatalytic activation of the enzymatic unit C1r. Activated C1r then cleaves C1s at a single site in the protein to activate it.

Since C1 activation occurs only when several of the six head groups of C1q are bound to antibody, only surfaces that are densely coated with antibody will trigger the process. This limitation reduces the risk of inappropriate activation on host tissues.

The ability of C4b and C3b to bind surfaces is fundamental to complement function

C3 and C4 are homologous molecules that contain an unusual structural feature – an internal thioester bond between a glutamine and a cysteine residue that, in the intact molecule, is buried within the protein.

When either C3 or C4 is cleaved by the convertase enzyme, a conformational change takes place that exposes the internal thioester bond in C3b and C4b, making it very unstable and highly susceptible to attack by nucleophiles such as hydroxyl groups (-OH) and amine groups (-NH₂) in membrane proteins and carbohydrates. This reaction creates a covalent bond between the complement fragment and the membrane ligand, locking C3b and C4b onto the surface (see Fig. 4.6).

The exposed thioester remains reactive for only a few milliseconds because it is susceptible to hydrolysis by water. This lability restricts the binding of C3b and C4b to the immediate vicinity of the activating enzyme and prevents damage to surrounding structures.

The alternative and lectin pathways provide antibody-independent ‘innate’ immunity

The lectin pathway is activated by microbial carbohydrates

The lectin pathway differs from the classical pathway only in the initial recognition and activation steps. Indeed, it can be argued that the lectin pathway should not be considered a separate pathway, but rather a route for classical pathway activation that bypasses the need for antibody.

The C1 complex is replaced by a structurally similar multimolecular complex, comprising a C1q-like recognition unit, either mannan-binding lectin (MBL) or ficolin (actually a family of three proteins in man), and several MBL-associated serine proteases (MASPs). MASP-2 provides enzymatic activity. As in the classical pathway, assembly of this initiating complex is Ca²⁺-dependent.

C1q and MBL are members of the collectin family of proteins characterized by globular head regions with binding activities and long collagenous tail regions with diverse roles (see Fig. 6.w3). Ficolins are structurally similar but the head regions comprise fibrinogen-like domains.

MBL binds the simple carbohydrates mannose and N-acetyl glucosamine, while ficolins bind acetylated sugars and other molecules; these ligands are abundant in the cell walls of diverse pathogens, including bacteria, yeast, fungi, and viruses, making them targets for lectin pathway activation. Binding induces shape changes in MBL and ficolins that in turn induce autocatalytic activation of MASP-2. This enzyme can then cleave C4 and C2 to continue activation exactly as in the classical pathway.

The lectin pathway is not the only means of activating the classical pathway in the absence of antibody. Apoptotic cells, released DNA, mitochondria and other products of cell damage can directly bind C1q, triggering complement activation and aiding the clearance of the dead and dying tissue.

Alternative pathway activation is accelerated by microbial surfaces and requires Mg²⁺

The alternative pathway of complement activation also provides antibody-independent activation of complement on pathogen surfaces. This pathway is in a constant state of low-level activation (termed ‘tickover’).

C3 is hydrolyzed at a slow rate in plasma and the product, C3(H₂O), has many of the properties of C3b, including the capacity to bind a plasma protein factor B (fB), which is a close relative of the classical pathway protein C2. Formation of the complex between C3b (or C3(H₂O)) and fB is Mg²⁺-dependent, so the alternative pathway is inactive in the absence of Mg²⁺ ions. (The differences in the ion requirements of the classical and alternative pathway are exploited in laboratory tests for complement activity.)

The complex is the C3 convertase of the alternative pathway

Once bound to C3(H₂O) or C3b, fB becomes a substrate for an intrinsically active plasma enzyme termed factor D (fD). fD cuts fB in the C3bB complex:

• releasing a fragment, Ba;

• while the residual portion, Bb, becomes an active protease.

The complex is the C3 cleaving enzyme (C3 convertase) of the alternative pathway. C3b generated by this convertase can be fed back into the pathway to create more C3 convertases, thus forming a positive feedback amplification loop (Fig. 4.7). Activation may occur in plasma or, more efficiently, on surfaces.

Fig. 4.7 Regulation of the amplification loop

Alternative pathway activation depends on the presence of activator surfaces. C3b bound to an activator surface recruits factor B, which is cleaved by factor D to produce the alternative pathway C3 convertase , which drives the amplification loop, by cleaving more C3. However, on self surfaces the binding of factor H is favored and C3b is inactivated by factor I. Thus the binding of factor B or factor H controls the development of the alternative pathway reactions. In addition, proteins such as membrane cofactor protein (MCP) and decay accelerating factor (DAF) also limit complement activation on self cell membranes (see Fig. 4.9).

Q. What physiological advantages and problems can you see in a system with a positive feedback loop (i.e. where the presence of C3b leads to the production of an enzyme C3bBb that generates more C3b)?

A. The positive feedback amplification and ‘always on’ features of the alternative pathway are well suited to pathogen surveillance. For example, a small initial stimulus could produce the deposition of large amounts of C3b on a pathogen surface, thereby facilitating its phagocytosis. If unregulated, however, the system will continue to activate until all available C3 has been consumed. Self cells would also be subjected to complement activation and could be damaged or destroyed.

Specific features of host cell surfaces, including their surface carbohydrates and the presence of complement regulators (see below), act to protect the host cell from alternative pathway activation and risk of being destroyed, and such surfaces are termed non-activating.

On an activating surface such as a bacterial membrane, amplification will occur unimpeded and the surface will rapidly become coated with C3b (see Fig. 4.7). In a manner analogous to that seen in the classical pathway, C3b molecules binding to the C3 convertase will change the substrate specificity of the complex, creating a C5 cleaving enzyme, .

Cleavage of C5 is the last proteolytic step in the alternative pathway and the C5b fragment remains associated with the convertase.

The alternative pathway is linked to the classical pathway

The alternative pathway is inexorably linked to the classical pathway in that C3b generated through the classical pathway will feed into the alternative pathway to amplify activation. It therefore does not matter whether the initial C3b is generated by the classical, lectin, or alternative pathway – the amplification loop can ratchet up the reactions if they take place near an activator surface.

Complement protection systems

Control of the complement system is required to prevent the consumption of components through unregulated amplification and to protect the host. Complement activation poses a potential threat to host cells, because it could lead to cell opsonization or even lysis. To defend against this threat a family of regulators has evolved alongside the complement system to prevent uncontrolled activation and protect cells from attack.

C1 inhibitor controls the classical and lectin pathways

In the activation pathways, the regulators target the enzymes that drive amplification:

• activated C1 is controlled by a plasma serine protease inhibitor (serpin) termed C1 inhibitor (C1inh), which removes C1r and C1s from the complex, switching off classical pathway activation;

• C1inh also regulates the lectin pathway in a similar manner, removing the MASP-2 enzyme from the MBL or ficolin complex to switch off activation.

C3 and C5 convertase activity are controlled by decay and enzymatic degradation

The C3 and C5 convertase enzymes are heavily policed with plasma and cell membrane inhibitors to control activation. In the plasma:

• factor H (fH) and a related protein factor H-like 1 (fHL-1) destroy the convertase enzymes of the alternative pathway;

• C4 binding protein (C4bp) performs the same task in the classical pathway.

On membranes, two proteins, membrane cofactor protein (MCP) and decay accelerating factor (DAF), collaborate to destroy the convertases of both pathways (Fig. 4.8).

Fig. 4.8 C3 and C5 convertase regulators

The five proteins listed are widely distributed and control aspects of C3b and C4b dissociation or breakdown. Each of these proteins contains a number of short consensus repeat (SCR) domains. They act either by enhancing the dissociation of C3 and C5 convertases or by acting as cofactors for the action of factor I on C3b or C4b.

The regulators of the C3 and C5 convertases are structurally related molecules that have arisen by gene duplication in evolution. These duplicated genes are tightly linked in a cluster on chromosome 1, termed the regulators of complement activation (RCA) locus. This locus also encodes several of the complement receptors (see below).

Control of the convertases is mediated in two complementary ways

Decay acceleration –

The convertase complexes are labile, with a propensity to dissociate within a few minutes of creation. The regulators fH, fHL-1, and C4bp from the fluid phase and DAF on cell membranes bind the convertase complex and markedly accelerate decay, displacing:

• C2a from the classical pathway convertase; and

• Bb from the alternative pathway enzyme (Fig. 4.9).

Fig. 4.9 The two processes by which C3 convertase regulators inactivate the enzymes

DAF binds the enzyme complex, displacing the enzymatically active component (C2a or Bb). Membrane cofactor protein (MCP) binds the C3b/C4b unit released after decay and acts as a cofactor for factor I (fI) cleavage of C3b or C4b, resulting in the irreversible inactivation of the convertase.

Cofactor activity –

Factor I (fI) is a fluid-phase enzyme that, in the presence of an appropriate cofactor, can cleave and irreversibly inactivate C4b and C3b (see Fig. 4.9). MCP is a cofactor for fI cleavage of both C4b and C3b, whereas:

• C4bp specifically catalyzes cleavage of C4b;

• fH/fHL-1 catalyzes cleavage of C3b.

It is interesting to note that, whereas the plasma regulators contain both activities in a single molecule, the two membrane regulators each contain only one activity.

Efficient regulation of the convertases on membranes therefore requires the concerted action of:

• DAF to dissociate the complex; and

• MCP to irreversibly inactivate it by catalyzing cleavage of the central component.

The alternative pathway also has a unique positive regulator, properdin, which stabilizes the C3 convertase and markedly increases its life span. Recent evidence has suggested an additional role of properdin as a catalyst for alternative pathway activation.

.Complement receptor 1 (CR1) is often included in the list of membrane regulators of C3 convertase activity, and, indeed, CR1 is a powerful regulator with both decay accelerating and cofactor activities in both pathways. Nevertheless, it is excluded from the above discussion because CR1 is primarily a receptor for complement-coated particles and does not have a role in protecting the host cell.

The membrane attack pathway

Activation of the pathway results in the formation of a transmembrane pore

The terminal or membrane attack pathway involves a distinctive set of events whereby a group of five globular plasma proteins associate with one another and, in the process, acquire membrane-binding and pore-forming capacity to form the membrane attack complex (MAC).

The MAC is a transmembrane pore (Fig. 4.10). Cleavage of C5 creates the nidus for MAC assembly to begin. While still attached to the convertase enzyme, C5b binds first C6 then C7 from the plasma. Conformational changes occurring during assembly of this trimolecular C5b67 complex:

• cause release from the convertase; and

• expose a labile hydrophobic site.

Fig. 4.10 The membrane attack pathway

(1) C5b, while still attached to the C5 convertase, binds C6 and C7 from the fluid phase. The trimolecular C5b–7 complex dissociates from the convertase and binds the target cell membrane. Binding of C8 and multiple copies of C9 generates a rigid protein-lined transmembrane channel, the membrane attack complex (MAC). (2 and 3) Electron micrographs of the MAC. The complex consists of a cylindrical pore in which the walls of the cylinder, formed by C9, traverse the cell membrane. In these micrographs the human C5b–9 complex has been incorporated into a lecithin liposomal membrane. × 234 000.

(Courtesy of Professor J Tranum-Jensen and Dr S Bhakdi.)

The complex can stably associate with a membrane through the labile hydrophobic site, though the process is inefficient and most of the C5b67 formed is inactivated in the fluid phase.

Membrane-bound C5b67 recruits C8 from the plasma, and, finally, multiple copies of C9 are incorporated in the complex to form the MAC.

The latter stages of assembly are accompanied by major conformational changes in the components with globular hydrophilic plasma proteins unfolding to reveal amphipathic regions that penetrate into and through the lipid bilayer.

The fully formed MAC creates a rigid pore in the membrane, the walls of which are formed from multiple copies of C9 (up to 12), arranged like barrel staves around a central cavity.

The MAC is clearly visible in electron microscopic images of complement lysed cells as doughnut-shaped protein-lined pores, first observed by Humphrey and Dourmashkin 40 years ago (see Fig. 4.10). The pore has an inner diameter approaching 10 nm:

• allowing the free flow of solutes and electrolytes across the cell membrane; and

• because of the high internal osmotic pressure, causing the cell to swell and sometimes burst.

Metabolically inert targets such as aged erythrocytes are readily lysed by even a small number of MAC lesions, whereas viable nucleated cells resist killing through a combination of ion pump activities and recovery processes that remove MAC lesions and plug membrane leaks.

Even in the absence of cell killing, MAC lesions may severely compromise cell function or cause cell activation.

Regulation of the membrane attack pathway reduces the risk of ‘bystander’ damage to adjacent cells

Although regulation in the activation pathways is the major way in which complement is controlled, there are further failsafe mechanisms to protect self cells from MAC damage and lysis.

First, the membrane binding site in C5b67 is labile. If the complex does not encounter a membrane within a fraction of a second after release from the convertase, the site is lost through:

• hydrolysis; or

• binding of one of the fluid-phase regulators of the terminal pathway – S protein (also termed vitronectin) or clusterin – both of which are multifunctional plasma proteins with diverse roles in homeostasis.

C8, an essential component of the MAC, also behaves as a regulator in that binding of C8 to C5b–7 in the fluid phase blocks the membrane binding site and prevents MAC formation.

The net effect of all these plasma controls is to limit MAC deposition to membranes in the immediate vicinity of the site of complement activation, so reducing risk of ‘bystander’ damage to adjacent cells.

CD59 protects host cells from complement-mediated damage

Complexes that do bind are further regulated on host cells by CD59, a membrane protein that locks into the MAC as it assembles and inhibits binding of C9, thereby preventing pore formation (Fig. 4.11). CD59 is a broadly expressed, glycolipid-anchored protein that is structurally unrelated to the complement regulators described above.

Fig. 4.11 Role of CD59 in protecting host cells from complement damage

The upper diagram models assembly of the MAC in the absence of the regulator CD59 – C9 binds C5b–8, unwinds, and traverses the membrane and recruits further C9 molecules, which in turn unfold and insert to form the MAC. In the lower diagram, CD59 binds the C5b–8 complex and prevents the unfolding and insertion of C9, which is essential for the initiation of MAC pore formation.

The importance of CD59 in protecting host cells from complement damage is well illustrated in the hemolytic disorder paroxysmal nocturnal hemoglobinuria (PNH), in which erythrocytes and other circulating cells are unable to make glycolipid anchors and as a consequence lack CD59 (and also DAF). The ‘tickover’ complement activation that occurs on all cells without much consequence is then sufficient in the absence of CD59 to cause low-grade hemolysis and hemolytic crises.

Membrane receptors for complement products

Receptors for fragments of C3 are widely distributed on different leukocyte populations

Many cells express one or more membrane receptors for complement products (Fig. 4.12). An understanding of the receptors is essential because the majority of the effects of complement are mediated through these molecules. The best characterized of the complement receptors are those binding fragments of C3.