Role Of Complement In Combating Viruses

Published Date: 02 Nov 2017

Abstract

Viruses are constantly under the peril of complement assault as they are efficiently recognized and neutralized by the complement system. In addition, the complement system is also known to enhance the virus specific B and T cell responses which play critical role in controlling viral infections. It is, therefore, important that viruses subvert and escape the complement-mediated attacks for their survival. In this chapter, we provide an overview of how the complement system help in controlling viruses and what elusive strategies viruses have developed to evade these defenses.

Introduction

Viruses are amongst the most successful pathogens to have co-existed with the hosts by maintaining a precarious balance between the two for their successful existence (1). Being predatory in nature, viruses are constantly in the pursuit of survival and thus, there exists a constant struggle for endurance between the viruses and their hosts: viruses pursue host for their propagation and the hosts on the other hand defy the viral intrusions owing to their well-developed and interconnected network of innate and adaptive immune defense mechanisms.

The complement system is one of the most phylogenetically ancient and pivotal innate immune defense mechanisms of the host, which emerged at least 1000 million years ago (2). Although the system was discovered as heat labile components of plasma necessary for antibody-mediated killing of bacteria, over the years it has emerged as a truly multitasking system that not only functions as an innate arm of the immunity, but also plays an instructive role in generation of adaptive immune responses and participates in clearance of immune complexes and apoptotic cells (3,4). Because complement possesses distinctive ability to recognize known as well as newly emerging non-self structures, it possesses the unprecedented capability to protect against varied pathogens including viruses. This is clearly evident from the system’s ability to neutralize diverse DNA (5,6) as well as RNA (7,8) viruses, including the recently emerged pandemic H1N1 virus (Rattan, A., Pawar, S.D., Mullick, J and Sahu, A., unpublished observation).

In contrary to the significant function of the complement system against viruses, it has gained diminutive importance as an antiviral defense system unlike other soluble innate immune barriers such as interferons and interleukins (9). A likely reason for this is that mostly complement deficiencies have been shown to be associated with bacterial infections and not the viral infections (10). It could be argued that non-association of viral infection with complement deficiencies is largely due to the lack of systematic study on association between complement deficiencies and vulnerability to viral infections. Another reason could be that the system is a vital antiviral defense and therefore viruses have developed mechanisms to subvert the complement assault (11,12). This argument is well supported by many studies from other as well as our laboratory, which provide evidence that viruses have developed diverse strategies to target the complement system (13-15). Here, we highlight the mechanisms of complement-mediated control of viral infections and cover the ingenious elusive strategies devised by viruses to annul the complement-mediated assaults.

I) Role of complement in combating viruses

The complement system is evolved to perform immune surveillance in the host, which raises the question: how does the complement system recognize viruses? It is now evident from a significant body of knowledge that complement possesses the ability to recognize viruses with as well as without the help of pattern recognition molecules, which then culminates into complement activation and neutralization of virus particles via three major pathways – the classical (CP), alternative (AP) and lectin (LP) pathways (Fig. 1). Based on the current literature on initiation of complement activation, it can be said that virus recognition by molecules such as antibodies (IgM, IgG3 and IgG1), C-reactive protein (CRP) (16), serum amyloid P (SAP), specific intracellular adhesion molecule-grabbing nonintegrin (SIGN-R1) (17) and C1q (18,19) can lead to activation of the CP, while recognition by lectins such as mannose binding lectin (MBL) and ficolins (L, M and H) can lead to activation of the LP. Among these, recognition of viruses by antibodies and C1q has been shown to be neutralized by the CP (5,6,18,20), while recognition of viruses by MBL has been shown to be neutralized by the LP (21-23).

Although the above repertoire of molecules allows the complement system to recognize a large variety of viruses it does not ensure the recognition of all of them. The system therefore is also equipped with a fail-safe mechanism, which allows labeling of viral particles as non-self without the help of a recognition molecule. In this mode of virus recognition, C3b is deposited onto viruses owing to continuous low level C3 activation in fluid phase by the AP initial C3-convertase {C3(H2O),Bb,P}. Once labeled with C3b, viruses are prone to complement-mediated neutralization as a result of activation of the AP on their surface. A prerequisite for such type of activation however is that, the viral surface should be devoid of complement regulators. A comprehensive list of viruses neutralized by various pathways is provided in Table 1.

Complement-mediated neutralization of viruses

It is clear from the above discussion that recognition of viruses by the complement system results in their neutralization. Thus the next obvious question is, which mechanisms complement utilize to neutralize viruses? Up until now, four mechanisms have been identified for complement-mediated neutralization of viruses: i) aggregation, ii) opsonization, iii) phagocytosis and iv) virolysis.

Aggregation: Viruses are known to be neutralized by antibodies as a result of aggregation, which occurs owing to decrease in the total number of infectious virus units. A similar mode of neutralization has also been shown to occur due to complement activation. Studies on polyoma virus (24) and more recently on influenza (25) and simian virus 5 (26) have shown that complement markedly enhances the aggregation of these viruses when they are coated with antibodies. Interestingly, this aggregation was shown to be dependent on the presence of complement components up to C3. Because C3b is monovalent in nature, it is not clear what induces C3b-dependent crosslinking of these viruses. It is however likely that C3b-binding polymeric molecule(s) present in serum produce this aggregation.

Opsonization: Activation of the complement system on the viral surface results in coating of the surface with complement components. Notably, such coats are thick enough to be observed under the electron microscope (27,28). It is conceivable that such coating can inhibit the virus attachment process due to steric hindrance. It is however also probable that opsonization hinders post-attachment steps like entry, uncoating, DNA/RNA transport to the nucleus or early gene expression. Examples where complement coating without virolysis has resulted in virus neutralization include complement-mediated neutralization of Newcastle disease virus (NDV) (7), herpes simplex virus-1 (HSV-1) (6,29,30), human T-lymphotropic virus Type I (HTLV-1) (31), human immunodeficiency virus type I (HIV-1) (18,19,32), influenza (25,33), West Nile virus (WNV) (34), Dengue virus (DEN) (23) and glycoprotein C (gc)-null HSV-1 and -2 (29). Of these examples, particularly in HSV-1 and WNV, opsonization has been shown to affect the post-entry steps.

Phagocytosis: Opsonization of pathogens by complement components C3b and C4b, and further inactivation of C3b into iC3b and C3d result in stable labeling of the viral surface with complement fragments. These fragments then can serve as ligands for recognition by various complement receptors on phagocytic cells, e.g., CR1 (CD35), CR2 (CD21), CR3 (CD11b/CD18), CR4 (CD11c/CD18) and CRIg resulting in engulfment of viral particles by phagocytes. Consequently, complement coated viral particles are recognized and engulfed by phagocytes. Such type of neutralization has been shown only in case of HSV (35,36) and Japanese encephalitis virus (JEV) (37) due to lack of focus in this area.

Virolysis: Enveloped viruses are susceptible to complement-mediated virolysis as a result of activation of the terminal complement cascade and insertion of the membrane attack complex (C5b-9) into their envelopes. Many viruses however are known to evade virolysis owing to incorporation of the host complement regulator CD59 into their envelope (discussed later in the review). Viruses that have been shown to be susceptible to complement-mediated virolysis include alphaviruses, herpesviruses, coronaviruses, retroviruses and paramyxoviruses (reviewed in (38)).

Complement-mediated enhancement of acquired immunity

The complement system is known to bridge the innate and acquired immunity (39). The subsequent question therefore is does complement enhance antiviral immunity? It is now well appreciated that the system is capable of enhancing both virus-specific B as well as T cell responses (39,40). Early studies performed in this direction utilized decomplemented animals (obtained by injecting cobra venom factor; CVF) as well as C5-deficient strains of mice to address this question. It was evident from these studies that presence of complement limits viral infection. Infection studies performed using influenza (41), rabies (42), and sindbis viruses (43) in C5-deficient mice and/or CVF-treated animals showed increased viral load in the target organ and higher mortality. Whether lack of virus control was due to direct effect of complement on viruses or due to indirect effect of complement on the acquired immune responses was not clear in these studies. Because it was known that antibodies help neutralize viruses as a result of CP activation, efforts were also made to examine the protective role of complement-fixing antibodies during viral infection. Such antibodies showed protection in mouse models of yellow fever virus (YFV) (44) and DEN-2 (45) infection.

Later studies performed utilizing complement-knockout mice led to a better understanding of the role of complement in boosting antibody and cell-mediated immune responses. It was observed that mice deficient in complement components (C3 and C4) or receptor (CD21/CD35) challenged with HSV produced a reduced IgG response to the virus and this was due to failure of memory B cell generation in these mice (46). The memory B cell generation nonetheless was not altered in C3-/-, C4-/- and CD21/CD35-/- mice when they were challenged with vesicular stomatitis virus (VSV) or lymphocytic choriomeningitis virus (LCMV). The authors therefore proposed that the route of infection and replication capacity of the virus at the site of infection dictates whether complement system is required for enhancing the B cell response (47). Later, Diamond and his colleagues further examined the role of complement in the development of acquired immunity during flavivirus infection. In particular, they showed that mice deficient in C3 or CD21/CD35 generate reduced virus-specific IgM and IgG response and are more susceptible to lethal infection (48). Further, they also observed that different complement pathways participate differently in priming the adaptive immune response, thus AP deficiency results in reduction of T cell responses, while CPl and LP deficiencies result in reduction in B- as well as T cell responses (49).

Because complement deficiencies in mice were associated with exacerbated influenza disease (41), efforts were also made to dissect the effect of complement on T cells. It was shown that C3 deficiency, but not CD21/CD35 deficiency, caused reduced priming of CD4+ and CD8+ cells in lung-draining lymph nodes and as a result reduced the recruitment of virus-specific CD4+ and CD8+ effector T cells in the lung (50). Very recently, it has been shown that defective priming of T cells in influenza-infected C3 deficient mice is due to defect in dendritic cells-mediated transport of viral antigen to the draining lymph node (51). Whether this is also due in part to direct effect of C3a and C5a on T cells is not clear at present. Importance of C3 in induction of virus-specific CD8+ T cells was also observed during LCMV infection though the mechanism remains unresolved (52). In yet another study, it was shown that efficient generation of cytotoxic T lymphocyte response against HIV requires opsonization of virus with complement (53). Importance of complement was also shown during ectromelia virus infection, wherein it was evident that genetic absence of C3, C4 and factor B causes earlier dissemination of virus and increased virus titer in the target organs (54).

II) Complement evasion by viruses: diverse strategies

The last few decades have seen identification of many diverse strategies adopted by viruses to counteract the host complement. These include molecular mimicry of the host complement regulators, molecular piracy of the host complement regulators and utilization of complement receptors for cellular entry. In this section, we detail how these strategies help viruses to thwart the complement attack.

Molecular mimicry as evasion strategy

The central step in complement activation is the cleavage of C3 by C3-convertases. This results in labeling of pathogens by C3b, which are then treated as "foreign" by the host complement system. On the host cells, this step is tightly regulated by a series of complement regulators belonging to a family of proteins termed as regulators of complement activation (RCA) (55-57). Large DNA viruses, such as pox and herpes viruses, have been shown to encode homologs of RCA proteins (11,12,58). It is believed that these viral regulators have been integrated into the viral genome during co-evolution as a result of horizontal gene transfer (59). Sequence variations in the viral RCA (vRCA) proteins point towards diversification of these homologs after their acquisition. In addition to these, viruses also encode homologs of non-RCA complement regulators as well as novel complement regulators with no structural similarity to human complement regulators to subvert the host complement attack.

The human RCA family members are composed of characteristic structures known as short consensus repeats (SCRs) or complement control protein (CCP) domains (56). These elongated modules are arranged in tandem and vary in number from 4 to 59. Typically, CCP domain consists of approximately 60 amino acids (aa) that are organized into eight or less antiparallel -strands wherein the structure is stabilized by four invariant cysteines which link up through two disulfide bonds in 1-3 and 2-4 arrangement. The domain also contains an invariant tryptophan which is buried inside the small hydrophobic core (60). Interestingly, this arrangement allows exposure of most side chains to the solvent resulting in a larger surface area compared to the proteins of similar molecular weight (15). Functionally, the RCA proteins regulate C3-convertases by accelerating their irreversible decay (decay-accelerating activity; DAA) as well as by acting as cofactor in factor I-mediated inactivation of C3-convertase subunits C3b and C4b (cofactor activity; CFA). The vRCA proteins mimic the human RCA proteins both structurally and functionally. They however vary in length from 2-8 CCPs, though 2 CCP containing viral regulators have not yet been functionally characterized.

RCA homologs of poxviruses: The only subfamily that has been found to encode orthologs of complement regulatory proteins within Poxviridae is the Chordopoxvirinae, and the viruses belong to genera Orthopoxvirus (e.g., vaccinia, variola, monkeypox, ectromelia, cowpox, camelpox and buffalopox), Capripoxvirus (e.g., goatpox, sheeppox and lumpy skin disease virus), Leporipoxvirus (e.g., myxoma and rabbit fibroma virus), Suipoxvirus (e.g., swinepox), Yatapoxvirus (e.g., tanapox, yaba-like disease virus and yaba monkey tumor virus) and Cervidpoxvirus (e.g., deerpox virus). Among these, functional complement regulators are known to be encoded only by orthopoxviruses (Table 2). Interestingly they exhibit greater than 90% sequence similarity to each other. Amongst these, vaccinia virus complement control protein (VCP), smallpox inhibitor of complement enzyme (SPICE), monkeypox inhibitor of complement enzymes (MOPICE), and ectromelia virus inhibitor of complement enzyme (EMICE) have been studied by various groups including ours, which has led to significant insight into pathogenesis of these viruses.

The first vRCA to be discovered was VCP. It was identified when an attenuated strain of the vaccinia virus was found to have a major deletion towards the left end of the genome and one of the ORF encoded by the deleted region matched with human RCA proteins (61). Later, it became clear that it was able to abrogate the antibody-dependent complement-mediated neutralization of the virus and play a role in vaccinia virus pathogenesis (5,62); these results were reconfirmed in recent independent studies (63,64). Analysis of VCP ORF (C21L) showed that it is a four CCP module containing protein, which are connected by short linkers of 4 aa that provide flexibility to the protein (61). Structural studies showed that each of its module fold to form a 6--strand structure (65,66).

Infection of cells with vaccinia virus showed that it is a major secretory protein of the virus (61), which is produced at the late stage of infection (67). Its functional characterization demonstrated that the secreted VCP is capable of regulating complement activity (62) owing to its decay-accelerating activity (DAA) against CP and AP C3 convertases (68,69). Later, it was recombinantly expressed in Pichia pastoris and further functional characterization revealed that it also possesses CFA against C3b as well as C4b (70). Being a soluble protein, it was thought that VCP was able to inhibit complement only in solution, but subsequent studies demonstrated the ability of VCP to bind to heparin sulfate proteoglycans (71,72) as well as the viral protein A56 (73,74) which permits anchorage on the cell surface, thus providing a basis that VCP can additionally protect infected cells from the complement attack.

Understanding the functioning of any protein requires detailed understanding of its interaction with the target protein(s). For VCP, this was achieved by utilizing a multi-prong approach that included SPR studies, monoclonal antibodies (mAbs), deletion mutagenesis and chimeric proteins. The SPR study revealed that unlike human complement regulators, interaction of VCP with human complement proteins C3b and C4b followed a simple 1:1 binding model and possessed very fast on- and off-rates for the target proteins (75). These results suggested that fast recycling of the viral regulator increases its effectiveness as inactivator. Further, it was also established that like human regulators, these interactions were also highly dependent on ionic strength (75). In another study, binding of truncation mutants of VCP to C3b and C4b suggested that the first three CCP domains were sufficient for binding, but optimal binding was rendered by all the four CCPs (69). The truncated mutants also became handy to map the functional domains in VCP. The results pointed out that the CFA was facilitated by the first three CCPs, while the minimum region required for the CP DAA was CCP1 and 2, but CCPs 3 and 4 were necessary for the complete activity (69). Though these studies (69,76,77) identified the minimum domains required for the functional activity, they failed to identify the domains critical for interacting with factor I during CFA, and that required for dissociation of the catalytic subunit from C3 convertases during DAA. These queries were subsequently answered by examining the activities of domain swap mutants wherein CCP modules of VCP were swapped with homologous modules of the human regulators DAF and MCP as they possess only decay and cofactor activity, respectively. It was explicit from the results that CCP1 of VCP imparts decay of the catalytic subunit from the C3 convertase and CCPs 2 and 3 recruit factor I for cofactor activity (78).

Smallpox inhibitor of complement enzymes (SPICE) is another important poxviral homolog encoded by the highly virulent variola virus (79). Because of its potent activity against human complement, it can serve as an excellent model of structural craftsmanship for complement inhibition. The protein was studied by multiple groups to comprehend the molecular basis of complement inactivation. Initial study by Rosengard et al. (79) demonstrated SPICE to be 100- and 6-fold more efficient than VCP in inactivating human C3b and C4b, respectively. This was followed by a study by Sfyroera et al (80) wherein SPICE was shown to be 75- and 1000-fold more potent than VCP in inhibiting CP and AP, respectively. Subsequently, in a study by Liszewski et al. (81), SPICE was demonstrated to contain higher decay activity against the CP C3 convertases than VCP. It was therefore clear that SPICE possesses enhanced complement regulatory activity against human complement than VCP. Thus, the obvious question to be asked was what dictates this robust activity of SPICE? Since SPICE differs from VCP by only 11 aa, any one or more of these 11 residues in SPICE were liable for this enhanced function. Utilizing an electrostatic modeling approach and mutagenesis experiments, Sfyroera et al. (80) predicted and showed the influence of a two residue substitution in VCP (E108K and E120K) in enhancing its C3b binding and C3b CFA. What remained missing in the study was determining the contributions of these and other residues in C4b CFA and DAA. A systematic analysis of the contribution of each of the 11 variant amino acids of SPICE towards its complement regulatory activities indicated that substitution of four residues (H98Y/S103Y/E108K/E120K) are enough to make VCP as potent as SPICE and this was primarily due to interaction of these residues with factor I (82). Notably, these four residues were also found to dictate the human specificity of SPICE (82). Intriguingly, consistent with the species tropism of vaccinia virus, in our recent report we showed that VCP is a potent inhibitor of the bovine complement (83). Especially, we also demonstrated that species selectivity in VCP and SPICE is primarily dictated by the presence of oppositely charged residues in their middle domains (83).

Other poxvirus RCA homologs that have been studied in detail are the monkeypox inhibitor of complement enzymes (MOPICE) encoded by monkeypox viruses (81,84) and ectromelia virus inhibitor of complement enzyme (EMICE) encoded by ectromelia virus (85). MOPICE is composed of three CCP domains and a truncated fourth domain. Despite truncation, it binds to both C3b and C4b and also possesses CFA against C3b and C4b. It however lacks both the CP and AP DAA. Interestingly, the presence of MOPICE is seen only in the more virulent strains of Congo basin and not in the less virulent strains of West Africa (84). Recent in vivo studies however do not consider MOPICE as the only factor responsible for augmented virulence of the Congo basin strain, though it does play a role in immune modulation (67,86). Unlike MOPICE, EMICE is a 4 CCP module containing protein and shares high homology with all the poxviral inhibitors of complement: it closely resembles VCP with 18 aa variation and 2 aa deletion, MOPICE with 19 aa difference and 1 aa deletion, SPICE with 26 aa difference and 2 aa deletion (85). Amongst all the poxviral inhibitors of complement enzymes, EMICE exhibits greatest divergence in CCP1. Functionally, recombinant EMICE regulates complement activation by factor-I mediated inactivation of C3b and C4b as well as decay of the CP C3-convertase (85). Because inhibitory activities of SPICE and VCP are consistent with the species tropism of variola and vaccinia, it would be interesting to determine whether MOPICE and EMICE also function in species specific manner.

RCA homologs of herpesviruses: Among herpesviruses, only members of subfamily Gammaherpesvirinae encode RCA homologs. Unlike poxviral homologs, herpesviral homologs are more diverse from each other with sequence similarity varying from 43% to 89%. Members that have been characterized functionally include herpesvirus saimiri (HVS), murine γ-herpesvirus 68 (γHV68/MHV68), Kaposi’s sarcoma-associated herpesvirus (KSHV/HHV-8) and rhesus rhadinovirus (RRV).

The first studied herpesviral RCA homolog was the complement control protein homolog (CCPH) encoded by Herpesvirus saimiri whose natural host is squirrel monkey. Remarkably, HVS was found to harbor two homologs of the complement regulatory proteins in its genome unlike any other virus: (i) a homolog of the RCA encoded by ORF4, and (ii) a homolog of the terminal complement inhibitor CD59 encoded by ORF15 (87). The ORF4 protein (CCPH) exists as a full length protein with 4 CCP domains and a transmembrane (TM) region as well as a spliced product that contains only 4 CCP domains lacking TM region (88). Its first functional characterization showed that when expressed on cells, it inhibited deposition of C3d on the cell surface (89). However, no studies were performed hereafter for almost a decade to further characterize this protein until our group sought to determine the mechanism of complement inhibition. By utilizing the soluble form of CCPH (sCCPH), we demonstrated that it inhibits both the CP and AP-mediated deposition of C3b onto the target cells owing to decay acceleration of CP and AP C3-convertases as well as inactivation of C3b and C4b (90). Its functional comparison with VCP showed that its rate of inactivation of C4b was comparable to that by VCP, but it was much more potent than VCP in inactivating C3b. This enhanced activity was attributed to R118, which is present at comparable position to that of K120 in SPICE. Mapping of the functional domains in HVS CCPH by generating truncated mutants revealed a very interesting aspect. It was observed that a single CCP module (CCP2) was sufficient to impart complement regulatory activities such as CP-DAA and CFA for C3b and C4b (91). This was against the dogma in the complement field as it was believed that multiple CCP domains are required for imparting any function (92). Thus, these results led to a paradigm shift in the field. Very recently, the protein was also examined for functional sites, which indicated that aa R35, R118, K142, F144 and K191 are critical for its function (93). Together the study suggested that ionic interaction forms a major component during interaction of CCPH with C3b/C4b and C3 convertases.

Akin to HVS CCPH, the KSHV ORF4 protein is also characterized to a significant extent. It is comprised of four CCP domains, a dicysteine motif, a serine/threonine (S/T)-rich region and a transmembrane region. Three putative N-glycosylation sites were present within CCP1, 2 and 4 and several O-linked glycosylations in the S/T region with heparin binding sites in the first CCP similar to the poxviral homologs. Functional characterization of the recombinant ORF4 protein, which was named Kaposica, revealed that it has ability to inhibit human complement-mediated lysis of erythrocytes, block cell surface deposition of C3b and act as a cofactor in factor I-mediated inactivation of C3b and C4b (94). The protein was also characterized by another group which showed that it inhibits C3b deposition on the infected cells and also harbors considerable DAA for the CP C3 convertase and poor DAA for the AP C3 convertase and named it KCP (95). Identification of the functional domains of Kaposica indicated that different CCPs are important for its various functional activities, such as CCPs 1-2 for the CP-DAA and CCPs 2-3 for the CFAs of C3b and C4b. Nevertheless, all the four domains are necessary for its optimal activity and poor AP DAA. (96).

While performing the physicochemical analysis of Kaposica and other vRCA proteins it was observed that positive electrostatic potential is highly conserved in CCPs 1 and 4 and the linkers between CCPs 1-2. Intrigued by this finding, Kaposica mutants were created to reduce the positive potential or convert it to negative. Functional analyses of the mutants indicated that positive potential at these domains and linkers influences the functional activity of the protein. The study also proposed a binding model for vRCA which suggests that positive potential on CCP1 helps their initial docking onto C3b and C4b.

Similar to Kaposica, murine -Herpesvirus 68 also harbors a four CCP containing RCA homolog which is expressed as a membrane bound form and a soluble form (97). The protein is capable of regulating both CP as well as AP of murine complement. In addition to its complement regulatory function and role in neurovirulence owing to subversion of complement (98), MHV68 RCA has also been shown to facilitate the virus replication in macrophages (99) and promote infection by activation of the protein kinase Akt (100). Another herpesviral RCA homolog that has been functionally characterized is the homolog harbored by the RRV which is the only in vivo model for KSHV in primates. The RCA homolog of RRV has been named as RRV complement control protein (RCP) and two different strains of RRV (H26-95 and 17577) have been found to encode RCP that considerably differ in structure. One RRV strain (H26-95) encodes a four CCP containing homolog while the strain 17577 encodes an eight CCP comprising molecule (101). Though varying in structure, both the RCP variants supported factor I mediated cofactor activities for C3b and C4b and decay of the CP C3 convertase. Surprisingly, the 17577 RCP variant also possess DAA for the AP C3 convertase (102).

Non-RCA homologs of herpesviruses: Complement regulators encoded by viruses other than RCA homologs include CD59 of HVS, glycoprotein C (gC) of alphaherpesviruses and yet uncharacterized inhibitor of Epstein-Barr virus (103).

In mammals, CD59 targets the terminal pathway of complement by inhibiting C5b-9 mediated cytolysis; it binds C5b-8 and inhibits C9 incorporation into the cell membrane. Because HVS ORF15 showed significant homology to human CD59 (48%), the protein was expressed and examined for complement inhibitory function by two independent groups. Both reported that similar to human CD59, the viral homolog is functional and is capable of protecting the target cells from complement-mediated cytolysis (104,105).

Glycoprotein C of alphaherpesviruses does not show structural similarity to any known complement regulators and thus inhibit complement by a novel mechanism. Binding of gCs to C3 has been shown in case of herpes simplex virus type 1 and 2 (HSV-1 and HSV-2), pseudorabies virus, bovine herpesvirus type-1, equine herpesvirus -1 and -2, and simian herpes B (106-108). Amongst these, functioning of gC molecules of HSV-1 (gC-1) and -2 (gC-2) has been studied in detail. Both gC-1 and gC-2 are homologous to each other and possess several O-glycosylations and nine and seven N-linked glycosylations, respectively. Disulphide bonds are formed by the presence of eight cysteines in these molecules which are imperative for the stability of these proteins (109). gC-1 has been shown to interact with native C3 and C3b, but not with C4b, although its interaction with native C3 does not result in inhibition of C3 activation. More studies on the underlying mechanism have revealed that gC-1 blocks the binding of properdin and C5 to C3b (108) and specifically hastens the decay of the alternate pathway C3 convertase, but not the decay of the CP C3 convertase (110). Consistent with the inhibitory function of gC, HSV-infected cells have been shown to be protected from complement-mediated cytolysis (111,112). The role of gC-1 in pathogenesis has also been established using two different models of infection; gC-null virus was approximately 100-fold less virulent than the wild-type virus and this was reversed when infection studies were performed in C3 null mice (113).

Molecular piracy as evasion strategy

Surface anchored complement regulators such as MCP (CD46), DAF (CD55) and CD59 (protectin) and soluble complement regulators such as factor H and C4-binding protein (C4BP) are critical for protection of host cells from the autologous complement-mediated damage. Many enveloped viruses acquire one or more of these regulators during the process of budding to avert the complement attack (Table 2). It is noteworthy to mention that acquisition of complement regulators is seen in both RNA and DNA viruses as compared to the strategy of molecular mimicry that is restricted only to DNA viruses.

Vaccinia virus is known to exist in two distinct infectious forms, intracellular mature virion (IMV) and extracellular enveloped virus (EEV). Among the two forms of the viruses, EEV was found to be resistant to neutralization by complement, while IMV was sensitive to the same (114). Examination of EEV for the presence of complement regulators by immunoblot as well as immunoelectron microscopy revealed the presence MCP, DAF as well as CD59. Retroviruses are no exception to this strategy as the presence of DAF, MCP and CD59 has also been demonstrated on HIV-1 (32,115). In addition to the general acquisition of DAF, MCP and CD59, specific acquisition of MCP has been shown for Simian virus 5 (SV5) and mumps virus (MuV) (26), and CD59 has been shown for influenza (116).

Among the herpesviruses, protection of pseudorabies virus is believed to be due to recruitment of host regulators (117). Interestingly, some viruses of the family (HCMV, MCMV and HHV-7) apart from acquiring the regulators also upregulate the expression of the host complement regulators on infected cells, which then protect the virus-infected cells from complement attack (118,119).

A recent addition to the list of viruses evading by molecular piracy is hepatitis C virus (HCV) and Newcastle disease virus (NDV). HCV showed the selective incorporation of CD59 on its envelope and HCV produced on CD59 knockdown cells were susceptible to the complement-mediated lysis. These results were further corroborated by detection of high levels of CD59 compared to MCP and DAF in the patient derived HCV isolates (120). Like mammalian viruses, NDV, an avian virus, also demonstrated the presence of DAF and MCP that curtailed complement-mediated lysis of the virus (121).

Examples discussed above illustrate viruses utilizing membrane bound regulators. Nevertheless, there are viruses that recruit factor H and C4BP on their surface. HIV-1 has been shown to recruit factor H which interact with gp41 and gp120 and protect the virions from complement mediated neutralization (32,122,123). Similarly, interaction of factor H with non-structural protein-1 (NS1) of the West Nile virus also protects the virus from complement attack (124). In yet another example, recruitment of factor H on sindbis virus blocks the AP on its surface. However, unlike other viruses where recruitment is mediated by viral proteins, recruitment on sindbis surface is via sialic acid (125,126). Interestingly, the NS1 protein of flaviviruses which interacts with factor H, is also known to interact with C4BP, suggesting that the protein also helps subvert the CP and LP (127).

Exploitation of complement receptors and regulators for cellular entry

Besides subversion of complement, escape is another means by which viruses evade the host immune mechanisms. Consistent with this premise, several viruses exploit complement receptors (CRs) that are widely expressed on the host cells for cellular entry. Overall, viruses belonging to six different families are known to utilize the CRs either directly or indirectly facilitating entry into the host cells (Table 2).

EBV is the most popular example of this category. It has been well established that EBV infects B cells using complement receptor 2 (CR2/CD21) for entry, which is predominantly present in mature B cells and follicular dendritic cells. And, the viral protein involved in interaction with CR2 is gp350/220, which is the major enveloped glycoprotein of EBV (128-130). The receptor binding epitope of gp350/220 was identified by using synthetic peptides, having sequence homology with C3d domain of C3 (natural ligand CR2) and deletion mutants of gp350/220. The data from these studies have clearly demonstrated that a short stretch of peptide (EDPGFFNVE) present at the amino terminus of gp350/220 specifically bound to the purified CR2 (130) and the amino-terminal, 470-amino-acid domain is pondered as the key ligand domain of gp350/220 (131). Further, it was also illustrated that the presence of EBV-neutralizing monoclonal antibody 72A1 restricted EBV internalization by binding with an N-terminal epitope and ligand activity of gp350/220 was prohibited in the absence of amino-terminal amino acids 28 and 29 (131). Recent X-ray crystallographic exploration of gp350/220 contributed to the same supposition (132). For more than three decades it was believed that CR2 is the only receptor EBV utilizes for its entry, but recent investigations have unmasked that it also uses CR1 (CD35) as a receptor for its entry into primary B cell (133). It should however be pointed out that though CR1 is adequate for EBV adsorption, its entry and latent infection requires co-expression of HLA II. Importantly, the study unravels how EBV infects cells that are devoid of CD21 including immature B cells and hence offers a better comprehension of EBV viral pathogenesis.

Another complement regulator, membrane cofactor protein (MCP/CD46), has also been utilized to access entry into host cells by the Edmonston strain (vaccine strain) and laboratory adapted isolates of measles virus (MeV), human herpes virus-6 (HHV-6) (134-136), some adenoviruses (Ads) (e.g., Ad3, Ad11 and Ad35 of group B and Ad37 of group D adenovirus) and bovine viral diarrhea virus (137-140). MCP comprises of four CCPs and studies on the interactions between MeV and MCP has revealed that the ectodomain of MeV viral glycoprotein hemagglutinin plays a role in binding to MCP via CCP domain 1 and 2, which is distinct from the binding site of C3b and C4b, the natural ligands of MCP (141-143). Additionally, it was found that all the four isoforms of MCP (C1, C2, BC1 and BC2) were able to interact with MeV (144). In the case of HHV-6, the heterotetramer gH/gL/gQ1/gQ2 complex has been recognized as a CD46 ligand (145). Even though both MeV and HHV-6 utilize MCP as a viral internalization receptor, studies have shown that the interacting domain of HHV-6 to MCP is CCP 2 and 3 which is distinct from MeV-MCP interacting domains. Whether MCP-mediated entry of these viruses also results in modulation of immune response owing to intracellular signaling through its cytoplasmic tail of MCP is not clear at present.

Numerous coxsackieviruses and enteroviruses are well known to exploit another complement regulator, the decay-accelerating factor (DAF/CD55), as a co-receptor for cellular entry (146-150) i.e., DAF by itself does not facilitate the cellular entry of virus. Mapping of interacting domains on DAF for various viruses revealed that different viruses interact with different CCPs of DAF when used as a co-receptor (151-154). In a recent study by Pan et al (155), a single aa change in the Coxsackievirus B3 capsid protein facilitated binding to DAF and provided cellular tropism.

Above examples clearly exemplify how viruses use CRs to gain entry in the host cells. Another mechanism that many viruses have adopted is to use the opsonized components on their surfaces as ligands for binding the CRs to augment infection. The examples include HIV, WNV, HTLV-1 and RRV. It is believed that HIV carefully balances complement activation on its surface to avoid virolysis by MAC formation, but utilized opsonized complement fragments to target CRs (reviewed in detail in (156)). A number of studies have shown that pre-treatment of complement increase the antibody dependent enhancement of HIV infection and opsonized HIV can efficiently interact with different CRs present on host cells (157,158). Further, it has also been shown that opsonized HIV coated with C3 fragments enhance infection in monocytes/macrophages and dendritic cells expressing CR3 and CR4 (159,160). In addition to aiding infection, complement receptors have also been implicated in spreading of the virus (156,161,162) and maintenance of the extracellular HIV reservoir in the germinal centers (163,164).

Other viruses like the WNV, HTLV-1 and RRV have also shown to make use of CRs for augmentation of infection. Utilizing in-vitro infection assays it has been demonstrated that CR3 is involved during antibody-dependent enhancement (ADE) of WNV infection (165). Likewise, HTLV adsorption on host cells was found to upsurge ~5-fold when pre-treated with complement, which could be inhibited by anti-CR2 antibody (166). Recently, in vivo investigations on Ross river virus using CR3-/- mice have revealed that lack of CR3 results in less severe disease without affecting viral replication (167).

Conclusions

Being a component of the innate immune system, the complement system is required to perform immune surveillance in the host. It is clear from the findings discussed herein that it accomplishes this effectively by directly neutralizing viruses as well as by boosting the acquired immunity. Additionally, because it is located in blood as well as tissues, it has the potential to neutralize viruses even in the immunologically privileged sites. It is therefore necessary that viruses develop evasion strategies to counter the complement assaults and maintain a balance between viral propagation and host complement defense. It could thus be argued that all successful viruses must have developed various stratagems to annul the complement assails. Encoding complement regulators along with other evasion molecules within viral genomes, which does not aid in replication, would require maintenance of a larger genome. This therefore suggests that complement evasion by molecular mimicry is expected to occur only in large DNA viruses like pox and herpesviruses. Nevertheless, molecular piracy and escape mechanisms (including inhibition of complement synthesis (168)) could be utilized by large as well as small genome viruses, which is obvious from the data discussed here. Thus, identification and better understanding of complement evasion and complement modulation mechanisms of various viruses in animal models, and in particular during human infections, would pave way to better manage viral infections.

Up until now, insight into complement regulation came primarily by studying interaction of host regulatory protein with human complement. The viral regulators now provide invaluable tools to dissect novel features (e.g., special distribution of electrostatic potential (83,169)) and control points of complement activation proteins that are central to their inactivation. We therefore believe that detailed characterization of functioning of viral regulators will not only provide new insight into host-virus interplay, but will also help design novel complement therapeutics to control complement-mediated pathology during various disease conditions.

Acknowledgements

Work presented here was supported by a project grant from the Department of Biotechnology, New Delhi, India. The authors also acknowledge financial assistance to AR and RK by the Council of Scientific and Industrial Research, New Delhi and the University Grant Commission, New Delhi, respectively.

Figure legends

Figure 1. Pathways of complement activation. Complement activation occurs through three major pathways namely classical (CP), lectin (LP) and alternative (AP) that result in the formation of C3 convertases (C4b,2a in case of CP/LP, and C3b,Bb in case of AP), which on the host cells are regulated by membrane anchored (DAF, MCP and CR1) and soluble (factor H and C4BP) RCA family members (marked in red). Though all the pathways have different mechanism of initiation, they converge to cleave the central C3 molecule to form C3b, the major opsonin that gets deposited on the viral surface. Deposition of C3b on the viral surface results in virus neutralization and boosting antibody response. Binding of C3b to C3 convertases result in formation of C5 convertases (C4b,2a,3b in case CP/LP and (C3b)2,Bb in case of AP) and further activation of the pathway leading to lysis. Both C3- and C5-convertases in AP are stabilized by properdin (P). Anapylatoxins C3a and C5a generated as a result of C3 and C5 cleavages induce chemotaxis, inflammatory responses and enhancement of T cell responses. Viruses encode structural homologs of RCA (e.g., VCP, SPICE, MOPICE etc) and CD59 (e.g., HVS CD59), and other complement regulators (e.g., gC-1) to regulate complement on their surface; viral regulators are marked in black.

Figure 2. Complement regulation by viral regulators through C3 convertase decay and C3b/C4b inactivation. Viral regulators (vRCA) inactivate alternative pathway (C3b,Bb) and classical/lectin pathway (C4b,2a) C3-convertases by two mechanisms: a) by accelerating the decay of C3-convertases (decay-accelerating activity) and b) by recruiting factor I on C3b/C4b for their inactivation (cofactor activity). (i) Recruitment of viral regulator onto C3 convertases causes dissociation of catalytic subunit of the enzyme, (ii) viral regulator is retained on C3b/C4b, (iii) viral regulator bound to C3b/C4b recruits factor I, (iii) factor I cleaves C3b into iC3b and C3f, and C4b into C4c and C4d.

Table 1: Neutralization of viruses by various complement pathway.

Pathway activated

Viruses

References

Classical pathway

Bacteriophages,

Herpes simplex virus 1 and 2(HSV),

Vaccinia virus

Japanese encephalitis virus (JEV),

Vesicular stomatitis virus (VSV),

Influenza virus

Hantavirus.

Ectromelia virus (mousepox virus),

Venezuelan equine encephalitis virus (VEEV)

Human immunodeficiency virus-2(HIV-2)

Human immunodeficiency virus-1(HIV-1),

Human T-cell leukemia virus type I (HTLV-1)

Polyoma virus

Parainfluenza virus type 3 (PIV3)

Lymphocytic choriomeningitis virus (LCMV)

Newcastle disease virus (NDV)

Yellow fever virus (YFV)

(170)

(6,29,30)

(5,114)

(171)

(8)

(25,33)

(172)

(54)

(173)

(174)

(18-20,32)

(31)

(24)

(175)

(7)

(7)

(44)

Alternate pathway

Epstein–Barr virus (EBV),

Sindbis virus,

Sendai virus,

Simian virus 5,

Vesicular stomatitis virus (VSV),

Mumps virus (MuV)

Measles virus (MeV)

Ectromelia virus (mouse pox virus),

Nipah virus (NiV)

(176)

(177)

(178)

(26,179)

(179)

(180,181)

(182-184)

(54)

(185)

Lectin pathway

Oncolytic viruses,

Hepatitis C viruses (HCV),

Influenza virus,

Herpes Simplex Virus-2 (HSV-2),

Human Immunodeficiency Virus (HIV),

Dengue Virus,

West Nile virus

(186)

(187)

(188)

(21)

(189,190)

(23)

(34)

Table 2: Evasion strategies of various viruses.

Evasion strategy

Virus family

Virus

Proteins encoded/

acquired or complement receptor used

Target proteins/

enzymes