Serpin

Serpins are a group of proteins with similar structures that were first identified as a set of proteins able to inhibit proteases. The acronym serpin was originally coined because many serpins inhibit chymotrypsin-like serine proteases (serine protease inhibitors).

The first members of the serpin superfamily to be extensively studied were the human plasma proteins antithrombin and antitrypsin, which play key roles in controlling blood coagulation (e.g. Figure 1) and inflammation, respectively. Initially, research focused upon their role in human disease: antithrombin deficiency results in thrombosis and antitrypsin deficiency causes emphysema. In 1980 Hunt and Dayhoff made the surprising discovery that both these molecules share significant amino acid sequence similarity to the major protein in chicken egg white, ovalbumin, and they proposed a new protein superfamily. Over 1000 serpins have now been identified, these include 36 human proteins, as well as molecules in plants, fungi, bacteria, archaea and certain viruses. Serpins are thus the largest and most diverse family of protease inhibitors.

While most serpins control proteolytic cascades, certain serpins do not inhibit enzymes, but instead perform diverse functions such as storage (ovalbumin, in egg white), hormone carriage proteins (thyroxine-binding globulin, cortisol-binding globulin) and tumor suppressor genes (maspin). The term serpin is used to describe these latter members as well, despite their noninhibitory function.

As serpins control processes such as coagulation and inflammation, these proteins are the target of medical research. However, serpins are also of particular interest to the structural biology and protein folding communities, because they undergo a unique and dramatic change in shape (or conformational change) when they inhibit target proteases. This is unusual - most classical protease inhibitors function as simple "lock and key" molecules that bind to and block access to the protease active site (see, for example, bovine pancreatic trypsin inhibitor). While the serpin mechanism of protease inhibition confers certain advantages, it also has drawbacks, and serpins are vulnerable to mutations that result in protein misfolding and the formation of inactive long-chain polymers (serpinopathies). Serpin polymerisation reduces the amount of active inhibitor, as well as accumulation of serpin polymers, causing cell death and organ failure. For example, the serpin antitrypsin is primarily produced in the liver, and antitrypsin polymerisation causes liver cirrhosis. Understanding serpinopathies also provides insights on protein misfolding in general, a process common to many human diseases, such as Alzheimer’s and CJD.

Cross-class inhibitors


Most inhibitory serpins target chymotrypsin-like serine proteases (see Table 1 and Figure 2). These enzymes are defined by the presence of a nucleophilic serine residue in their catalytic site. Examples include thrombin, trypsin, and human neutrophil elastase.

Some serpins inhibit other classes of protease and are termed "cross-class inhibitors". A number of such serpins have been shown to target cysteine proteases. These enzymes differ from serine proteases in that they are defined by the presence of a nucleophilic cysteine residue, rather than a serine residue, in their catalytic site. Nonetheless, the enzymatic chemistry is similar, and serpins most likely inhibit both classes of enzyme in a similar fashion.

Examples of cross-class inhibitory serpins include squamous cell carcinoma antigen 1 (SCCA-1) and the avian serpin myeloid and erythroid nuclear termination stage-specific protein (MENT) both inhibit papain-like cysteine proteases

The viral serpin crmA is a suppressor of the inflammatory response through inhibition of IL-1 and IL-18 processing by the cysteine protease caspase-1. In eukaryotes, a plant serpin has been shown to inhibit metacaspases and a papain-like cysteine protease. It is presently unclear whether any mammalian serpins function to inhibit caspases in vivo.

Localization and roles


Approximately two-thirds of human serpins perform extracellular roles. For example, extracellular serpins regulate the proteolytic cascades central to blood clotting (antithrombin), the inflammatory response (antitrypsin, antichymotrypsin, and C1 inhibitor) and tissue remodeling (PAI-1). Non-inhibitory extracellular serpins also perform important roles. Thyroxine-binding globulin and cortisol-binding globulin transport the sterol hormones thyroxine and cortisol, respectively (Figure 3). The protease renin cleaves off a ten-amino acid N-terminal peptide from angiotensinogen to produce the peptide hormone angiotensin I. Table 1 provides a brief summary of human serpin function, as well as some of the diseases that result from serpin deficiency.

The first Intracellular members of the serpin superfamily were identified in the early 1990s. As all nine serpins in Caenorhabditis elegans lack signal sequences, they are probably intracellular. Based upon these data it seems likely that the ancestral serpin to human serpins was an intracellular molecule.

The protease targets of intracellular inhibitory serpins have been more difficult to identify. Characterization is complicated by the observation that many of these molecules appear to perform overlapping roles. Further many human serpins lack precise functional equivalents in model organisms such as the mouse. An important function of intracellular serpins may be to protect against the inappropriate activity of proteases inside the cell. For example, one of the best-characterised human intracellular serpins is SERPINB9, which inhibits the cytotoxic granule protease granzyme B. In doing so, SERPINB9 may protect against inadvertent release of granzyme B and premature or unwanted activation of cell death pathways.

Intracellular serpins also perform roles distinct from protease inhibition. For example, maspin, a non-inhibitory serpin, is important for preventing metastasis in breast and prostate cancers. Another example is the avian nuclear cysteine protease inhibitor MENT, which acts as a chromatin remodelling molecule in avian red blood cells.

Phylogenetic studies show that most intracellular serpins belong to a single clade (see Table 1). Exceptions include the non-inhibitory heat shock serpin HSP47, which is a chaperone essential for proper folding of collagen, and cycles between the cis-Golgi and the endoplasmic reticulum.

Structure


Structural biology has played a central role in the understanding of serpin function and biology. Over eighty serpin structures, in a variety of different conformations (described below), have been determined to date. Although the function of serpins varies widely, these molecules all share a common structure (or fold).

The structure of the non-inhibitory serpin ovalbumin, and the inhibitory serpin antitrypsin, revealed the archetype native serpin fold. All typically have three β-sheets (termed A, B and C) and eight or nine α-helices (hA-hI) (see figure 4). Serpins also possess an exposed region termed the reactive centre loop (RCL) that, in inhibitory molecules, includes the specificity determining region and forms the initial interaction with the target protease. In antitrypsin, the RCL is held at the top of the molecule and is not pre-inserted into the A β-sheet (figure 4, left panel). This conformation commonly exists in dynamic equilibrium with a partially inserted native conformation seen in other inhibitory serpins (see figure 4, right panel).

Conformational change and inhibitory mechanism
Early studies on serpins revealed that the mechanism by which these molecules inhibit target proteases appeared distinct from the lock-and-key-type mechanism utilised by small protease inhibitors such as the Kunitz-type inhibitors (e.g. basic pancreatic trypsin inhibitor). Indeed, serpins form covalent complexes with target proteases. Structural studies on serpins also revealed that inhibitory members of the family undergo an unusual conformational change, termed the Stressed to Relaxed (S to R) transition. During this structural transition the RCL inserts into β-sheet A (in red in figure 4 and 5) and forms an extra (fourth) β-strand. The serpin conformational change is key to the mechanism of inhibition of target proteases.

When attacking a substrate, serine proteases catalyze peptide bond cleavage in a two-step process. Initially, the catalytic serine performs a nucleophilic attack on the peptide bond of the substrate (Figure 5). This releases the new N-terminus and forms an ester-bond between the enzyme and the substrate. This covalent enzyme-substrate complex is called an acyl enzyme intermediate. Subsequent to this, this ester bond is hydrolysed and the new C-terminus is released. The RCL of a serpin acts as a substrate for its cognate protease. However, after the RCL is cleaved, but prior to hydrolysis of the acyl-enzyme intermediate, the serpin rapidly undergoes the S-to-R transition. Since the RCL is still covalently attached to the protease via the ester bond, the S-to-R transition causes the protease to be moved from the top to the bottom of the serpin. At the same time, the protease is distorted into a conformation, where the acyl enzyme intermediate is hydrolysed extremely slowly. The protease thus remains covalently attached to the target protease and is thereby inhibited. Further, since the serpin has to be cleaved to inhibit the target protases, inhibition consumes the serpin as well. Serpins are therefore irreversible enzyme inhibitors. The serpin mechanism of inhibition is illustrated in figures 5 and 6, and several movies illustrating the serpin mechanism can be viewed at this link.

Conformational modulation of serpin activity
The conformational mobility of serpins provides a key advantage over static lock-and-key protease inhibitors. In particular, the function of inhibitory serpins can be readily controlled by specific cofactors. The X-ray crystal structures of antithrombin, heparin cofactor II, MENT and murine antichymotrypsin reveal that these serpins adopt a conformation wherein the first two amino acids of the RCL are inserted into the top of the A β-sheet (see figures 4 and 7). The partially inserted conformation is important because co-factors are able to conformationally switch certain partially inserted serpins into a fully expelled form. This conformational rearrangement makes the serpin a more effective inhibitor.

The archetypal example of this situation is antithrombin, which circulates in plasma in a partially inserted relatively inactive state. The primary specificity determining residue (the P1 Arginine) points toward the body of the serpin and is unavailable to the protease (Figure 7). Upon binding a high-affinity heparin pentasaccharide sequence within long-chain heparin, antithrombin undergoes a conformational change, RCL expulsion, and exposure of the P1 Arginine. The heparin pentasaccharide-bound form of antithrombin is, thus, a more effective inhibitor of thrombin and factor Xa (figure 7). Furthermore, both of these coagulation proteases contain binding sites (called exosites) for heparin. Heparin, therefore, also acts as a template for binding of both protease and serpin, further dramatically accelerating the interaction between the two parties (Figure 7). After the initial interaction, the final serpin complex is formed and the heparin moiety is released. This interaction is physiologically important. For example, after injury to the blood vessel wall, heparin is exposed, and antithrombin is activated to control the clotting response. The understanding of the molecular basis of this interaction formed the basis of the development of Fondaparinux, a synthetic form of Heparin pentasaccharide used as an anti-clotting drug.



Certain serpins spontaneously undergo the S-to-R transition as part of their function, to form a conformation termed the latent state (Figure 8). In latent serpins, the first strand of the C-sheet has to peel off to allow full RCL insertion. Latent serpins are unable to interact with proteases and are not protease inhibitors. The transition to latency represents a control mechanism for the serpin PAI-1. PAI-1 is released in the inhibitory conformation, however, undergoes conformational change to the latent state unless it is bound to the cofactor vitronectin. Thus PAI-1 contains an "auto-inactivation" mechanism. Similarly, antithrombin can also spontaneously convert to the latent state as part of its normal function. Finally, the N-terminus of tengpin (see pdbs 2PEE and   2PEF), a serpin from Thermoanaerobacter tengcongensis, is required to lock the molecule in the native inhibitory state. Disruption of interactions made by the N-terminal region results in spontaneous conformational change of this serpin to the latent conformation.



Serpin receptor interactions
In humans, extracellular serpin-enzyme complexes are rapidly cleared from circulation. In mammals, one mechanism by which this occurs is via the low-density lipoprotein receptor-related protein (LRP receptor), which binds to inhibitory complexes made by antithrombin, PA1-1, and neuroserpin, causing uptake and subsequent signaling events. Thus, as a consequence of the conformational change during serpin-enzyme complex formation, serpins may act as signaling molecules that alert cells to the presence of protease activity. The fate of intracellular serpin-enzyme complexes remains to be characterised.

Recently, it has been shown that the Drosophila serpin necrotic is taken up via the Lipophorin Receptor-1 (LpR1), which is related to the mammalian LDL receptor family. Trafficking studies reveal that following uptake by LpR1, necrotic is delivered to lysosomes where it is targeted for degradation.

Conformational change and non-inhibitory function
Certain non-inhibitory serpins also use the serpin conformational change as part of their function. For example, the native (S) form of thyroxine-binding globulin has high affinity for thyroxine, whereas the cleaved (R) form has low affinity. In similar manner, native (S) Cortisol-Binding Globulin (CBG) has higher affinity for cortisol than its cleaved (R) counterpart (Figure 3). Thus, in these serpins, RCL cleavage and the S to R transition has been commandeered to allow for ligand release, rather than protease inhibition.

Serpins, serpinopathies and human disease


Serpins are vulnerable to inactivating disease-causing mutations that result in the formation of misfolded polymers or protein aggregates ("serpinopathies"). Well-characterised serpinopathies include alpha 1-antitrypsin deficiency (alpha-1), which may cause familial emphysema and sometimes liver cirrhosis, certain familial forms of thrombosis related to antithrombin deficiency, types 1 and 2 hereditary angioedema (HAE) related to deficiency of C1-inhibitor, and familial encephalopathy with neuroserpin inclusion bodies (FENIB; a rare type of dementia caused by neuroserpin polymerisation). Serpins thus belong to a large group of molecules such as the prion proteins and the glutamine repeat containing proteins that cause proteopathies or conformational diseases.

Serpin polymerisation causes disease in two ways. First, the lack of active serpin results in uncontrolled protease activity and tissue destruction; this is seen in the case of antitrypsin deficiency. Second, the polymers themselves clog up the endoplasmic reticulum of cells that synthesize serpins, eventually resulting in cell death and tissue damage. In the case of antitrypsin deficiency, antitrypsin polymers cause the death of liver cells, sometimes resulting in liver damage and cirrhosis. Within the cell, serpin polymers are removed via endoplasmic reticulum associated degradation. However, the mechanism by which serpin polymers cause cell death remains to be fully understood.

Like cleaved serpins, serpin polymers are hyperstable with respect to heating, and each serpin monomer appears to have undergone the stressed to relaxed transition. Furthermore, serpin polymers are unable to inhibit target proteases, suggesting that the RCL is unavailable and inserted into the A-sheet. In the absence of definitive structural data, it was, therefore, postulated that serpins polymerise via a mechanism known as A-sheet polymerisation. In normal function the RCL inserts into the A β-sheet to form a fourth strand (figure 4). In the A-sheet polymerisation model, it was suggested that the RCL of one serpin molecule spontaneously inserted into the A-sheet of another, to form a long-chain polymer (figure 9). In effect, it was, thus, proposed that polymerization occurred as a consequence of the requirement of the serpin scaffold to accept an additional β-strand.

Serpins were one of the first families for which disease-causing mutations were directly analyzed in reference to the available crystal structures. In support of the A-sheet polymerisation model, it was noted that many serpin mutations that cause polymerisation localise to two distinct regions of the molecule (highlighted in figure 4a) termed the shutter and the breach. The shutter and the breach contain highly conserved residues, underlie the path of RCL insertion, and are proposed to be important for conformational change.

Two structures of cleaved serpin polymers have been solved; both of which reveal RCL / A-sheet sheet linkages similar to those predicted by the A sheet polymerisation mechanism. However, in direct contrast to the known properties of physiological serpin polymers, crystals of cleaved serpin A-sheet polymers readily dissociate into monomeric forms.



A large body of data now suggest that the events associated with serpin polymerisation occur during the folding of the molecule, and that mutations that cause serpinopathies interfere with the ability of the serpin to fold to the metastable native state. In normal serpin folding, the serpin rapidly moves through a key folding intermediate to attain the native state. Many studies have shown that it is the serpin folding intermediate that has the ability to polymerise, hence it is important that this folding species rapidly moves on to adopt native state. It was shown that mutations such as the Z-antitrypsin variant (Glu 342 to Lys) somehow prevented the final stage of seprin folding and caused the accumulation of the folding intermediate. As a result, population of the folding intermediate resulted in polymer formation. Interestingly, it was noted that once folded, the Z-antitrypsin variant closely resembles wild-type material in terms of thermal stability and inhibitory activity.

Together, these data have presented an important challenge to the A-sheet model for serpin polymerisation. On the one hand, the idea that serpin polymer formation essentially takes advantage of the serpin mechanism of conformational change is an attractive one. On the other, the biophyiscal data in particular suggest that it is a folding intermediate (rather than the native form) that polymerises, and it is clear that this intermediate must have different structural properties to the native, folded state.

In 2008, a key serpin crystal structure was determined that strongly suggests that physiological serpin polymers do not form via the A-sheet mechanism and instead form via a more extensive domain swapping event. The first such structure solved was of an antithrombin dimer (figure 10), and revealed that both strands s5A and the RCL can be incorporated into the A-sheet of another serpin molecule. This structure can readily be adapted to form long chain polymers. In 2011, the structure of a domain swapped antitrypsin trimer revealed that in polymers of this serpin the RCL is inserted, and that the C-terminal region of the molecule (comprising strands s1C, s4B and s5B) formed the domain swap (figure 11). In support of the physiological relevance of the latter structure, it was shown that antitrypsin polymers formed via a C-terminal domain swap were recognised by a monocloncal antibody specific for pathogenic antitrypsin polymers.



The new "domain swapped" model for serpin polymerisation begins to reconcile the available biophysical and biochemical data. Together, these data suggest that domain swapping events occur when mutations or environmental factors somehow interfere with the final stages of serpin folding to the native state. These data also reveal that different serpins can apparently polymerise via different types of domain swaps. Finally, while these data shed light on the final polymeric form, it is important to note that the precise toxic species of intermediate and / or polymer that causes cell death in, for example, antitrypsin deficiency, remains to be identified.

Mutations that result in spontaneous formation of latent (or latent-like), inactive conformations


Certain pathogenic mutations in serpins can promote inappropriate transition to the monomeric latent state (see figure 8a for the structure of the latent state). This causes disease because it reduces the amount of active inhibitory serpin. For example, the disease-linked antithrombin variants wibble and wobble, both promote formation of the latent state.

It is also worth highlighting a structure of a disease-linked human antichymotrypsin variant that further demonstrates the extraordinary flexibility of the serpin scaffold. The structure of antichymotrypsin (Leucine 55 to Proline) revealed a novel "δ" conformation that may represent an intermediate between the native and latent state (Figure 12). In the delta conformation, four residues of the RCL are inserted into the top of β-sheet A. The bottom half of the sheet is filled as a result of one of the α-helices (the F-helix) partially switching to a β-strand conformation, completing the β-sheet hydrogen bonding. It is unclear whether other serpins can adopt this conformer, and whether this conformation has a functional role. However, it is speculated that the δ-conformation may be adopted by Thyroxine-binding globulin during thyroxine release.

Other mechanisms of serpin-related disease
In humans, simple deficiency of many serpins (e.g., through a null mutation) may result in disease (see Table 1).

It is rare that single amino acid changes in the RCL of a serpin alters the specificity of the inhibitor and allow it to target the wrong protease. For example, the Antitrypsin-Pittsburgh mutation (methionine 358 to arginine) allowed the serpin to inhibit thrombin, thus causing a bleeding disorder.

Serpins are suicide inhibitors, the RCL acting as a "bait." Certain disease-linked mutations in the RCL of human serpins permit true substrate-like behaviour and cleavage without complex formation. Such variants are speculated to affect the rate or the extent of RCL insertion into the A-sheet. These mutations, in effect, result in serpin deficiency through a failure to properly control the target protease.

Several non-inhibitory serpins play key roles in important human diseases. For example, maspin functions as a tumour suppressor in breast and prostate cancer. The mechanism of maspin function remains to be fully understood. Murine knockouts of maspin are lethal; these data suggest that maspin plays a key role in development.

Evolution
Serpins were initially believed to be restricted to eukaryote organisms, but have since been found in a number of bacteria and archaea. It remains unclear whether these prokaryote genes are the descendants of an ancestral prokaryotic serpin or the product of lateral gene transfer (genetic transfer between organisms not by evolutionary descent). Rawlings et al. showed that serpins are the most widely distributed and largest family of protease inhibitors.

Human
In 2001, a serpin nomenclature was established (see table 1, below). The naming system is based upon a phylogenetic analysis of ~500 serpins. The human genome encodes 16 serpin clades, termed serpinA through to serpinP, encoding 29 inhibitory and 7 non-inhibitory serpin proteins (see Law et al. (2006) for a recent review). The proteins are named serpinXY where X is the clade of the protein and Y the number of the protein within that clade. Table 1 lists each human serpin, together with brief notes in regards to each molecules function and the consequence (where known) of dysfunction or deficiency.

Specialised Mammalian Serpins
Many mammalian serpins have been identified that share no obvious orthology with a human serpin counterpart. Examples include numerous rodent serpins (particularly some of the murine intracellular serpins) as well as the uterine serpins (discussed below).

Uterine
The term uterine serpin refers to members of the serpin A clade that are encoded by the SERPINA14 gene. Uterine serpins are produced by the uterine endometrium of a restricted group of mammals in the Laurasiatheria clade under the influence of progesterone or estrogen. They are probably not functional proteinase inhibitors and may function during pregnancy to inhibit maternal immune responses against the conceptus or to participate in transplacental transport.

Insect
The Drosophila melanogaster genome contains 29 serpin encoding genes. Amino acid sequence analysis has placed 14 of these serpins in serpin clade Q and 3 in serpin clade K with the remaining 12 serpins classified as orphan serpins not belonging to any clade. The clade classification system is difficult to use for Drosophila serpins and instead a nomenclature system has been adopted that is based on the position of Drosphila serpin genes on the Drosophila chromosomes. 13 of the Drosophila serpins occur as isolated genes in the genome (including Serpin-27A, see below), with the remaining 16 organised into three gene clusters that occur at chromosome positions 28D (2 serpins), 42D (5 serpins), 43A (4 serpins), 77B (3 serpins) and 88E (2 serpins).

Studies on Drosophila serpins reveal that Serpin-27A inhibits the Easter protease (the final protease in the Nudel, Gastrulation Defective, Snake and Easter proteolytic cascade) and thus controls dorsoventral patterning. Easter functions to cleave Spätzle (a chemokine-type ligand), which results in Toll mediated signaling. In addition to its central role in embryonic patterning, Toll signaling is also important for the innate immune response in insects. Accordingly, serpin-27A additionally functions to control the insect immune response. Interestingly, in Tenebrio molitor (a large beetle), a protein (SPN93) comprising two discrete tandem serpin domains functions to regulate the toll proteolytic cascade.

Worm
The genome of the nematode worm C. elegans contains nine serpins, however, only five of these molecules appear to function as protease inhibitors. One of these serpins, SRP-6, has been shown to perform a protective function and guard against stress induced calpain-associated lysosomal disruption. Further SRP-6 functions to inhibit lysosomal cysteine proteases released after lysosomal rupture. Accordingly, worms lacking SRP-6 are sensitive to stress. Most notably, SRP-6 knockout worms die when placed in water (the hypo-osmotic stress lethal phenotype or Osl). Based on these data it is suggested that lysosomes play a general and controllable role in determining cell fate.

Plant
The presence of serpins in plants has long been recognised, indeed, an abundant barley grain serpin (barley Protein Z) is one of the major protein components in beer.

The MEROPS database identifies 18 serpin family members in the Arabidopsis thaliana genome, but only about eight of these are full-length serpin sequences. Plant serpins are potent inhibitors of mammalian chymotrypsin-like serine proteases in vitro, the most well-studied example being barley serpin Zx (BSZx), which is able to inhibit trypsin, chymotrypsin as well as several blood coagulation factors. However, close relatives of chymotrypsin-like serine proteases are absent in plants. Interestingly, the RCL of several serpins from wheat grain and rye contain poly-Q repeat sequences similar to those present in the prolamin storage proteins of the endosperm. It has therefore been suggested that plant serpins may function to inhibit proteases from insects or microbes that cleave grain storage proteins. In support of this hypothesis, specific plant serpins have been identified in the phloem sap of pumpkin (CmPS-1) and cucumber plants. However, while an inverse correlation between up-regulation of CmPS-1 expression and aphid survival was observed, in vitro feeding experiments revealed that recombinant CmPS-1 did not appear to affect insect survival.

Alternative roles and protease targets for plant serpins have been proposed. It has been shown that Arabidopsis AtSerpin1 (At1g47710; 3LE2) inhibits metacaspase-like proteases in vitro. More recently the major in vivo protease target for AtSerpin1 was identified as the papain-like cysteine protease RESPONSIVE TO DESICCATION-21 (RD21). In the same study the X-ray crystal structure of the native, stressed form of AtSerpin1 was shown to have plant-specific features. Two other Arabidopsis serpins, AtSRP2 (At2g14540) and AtSRP3 (At1g64030) appear to be involved in responses to DNA damage caused by plant exposure to methane methylsulfonate (MMS).

Fungal
A single fungal serpin has been characterized to date: celpin from Piromyces sp. strain E2. Piromyces is an anaerobic fungus found in the gut of ruminants and is important for digesting plant material. Celpin is predicted to be an inhibitory molecule and contains two N-terminal dockerin domains in addition to the serpin domain. Dockerins are commonly found in proteins that localise to the fungal cellulosome, a large extracellular mulitprotein complex that breaks down cellulose. It is therefore suggested that celpin protects the cellulosome against plant proteases. Interestingly certain bacterial serpins also localize to the cellulosome.

Prokaryote
Predicted serpin genes are sporadically distributed in prokaryotes. In vitro studies on some of these moelcules have revealed that they are able to inhibit proteases and it is suggested that they function as inhibitors in vivo. Interestingly, several prokaryote serpins are found in extremophiles. Accordingly, and in contrast to mammalian serpins, these molecule possess elevated resistance to heat denaturation. The precise role of most bacterial serpins remains obscure, however,  Clostridium thermocellum serpin localises to the cellulosome. It is suggested that the role of cellulosome-associated serpins may be to prevent unwanted protease activity against the cellulosome.

Viral
Serpins are also expressed by viruses as a way to evade the host's immune defense. In particular, serpins expressed by pox viruses, including cow pox (vaccinia) and rabbit pox (myxoma), are of interest because of their potential use as novel therapeutics for immune and inflammatory disorders as well as transplant therapy. A study on Serp1 reveals this molecule suppresses the Toll-mediated innate immune response and allows indefinite cardiac allograft survival in rats. Studies on Crma and Serp2, reveal both are cross-class inhibitor and targets both serine (Granzyme B; albeit weakly) and cysteine proteases (Caspase 1 and Caspase 8). In comparison to their mammalian counterparts, viral serpins contain significant deletions of elements of secondary structure. Specifically, structural studies on crmA reveals this molecule lacks the D-helix as well as significant portions of the A- and E-helices.