Shigellae are enteropathogenic bacteria that cause the diarrhea-disease dysentery. In the environment Shigella can be found in brackwater but their only natural hosts are primates. Humans take up Shigella from stools or soiled fingers of infected persons or from contaminated food and water. Shigellosis is characterized by mucous diarrhea, fever, nausea and stomach cramps. All Shigellae species are able to cause bacterial dysentery, in which the diarrhea is not only mucopurulent but also bloody (Sansonetti, 1992). Shigella is extremely efficient in invading the hosts’ intestinal epithelium and causing disease: 10-100 bacteria are sufficient to cause shigellosis (DuPont, et al., 1989). Symptoms occur one to two days after exposure to the . The host response usually leads to the resolution of the infection within five to seven days. However, without proper medical treatment the diarrhea can be life threatening to some persons, especially young children and the elderly. This treatment involves rehydration as well as application of electrolytes and antibiotics.

Shigellosis occurs throughout the world with approximately 164.7 million cases per year. Among those cases 99% of them are found in areas of the world with only limited medical support and poor sanitation causing low hygienic standards. Each year 1.1 million people are estimated to die from Shigella infection (Kotloff, et al., 1999). 61% of all fatalities attributable to shigellosis involve children less than 5 years of age as systemic complications of shigellosis occur frequently in children (CDC, 2005;WHO, 2005). These complications include acute renal failure, hemolytic-uremic syndrome, toxic megacolon and neurological ea(Goldfarb, et al., 1982).

The Shigella genus comprises four species: S. dysenterie, S. boydii, S. flexneri and S. sonnei. order reflect severity of symptoms (Mims C., 1998). S. sonnei is classically found in developed countries (Kotloff, et al., 1999). But it seems to become more prevelant in Thailand (71%) as compared to previous years, a phenomenon probably linked to the current development of the country. S. flexneri is predominant in developing countries (60%) and it is the most frequently isolated species worldwide. S. dysenteriae type 1 (Sd1) is the only Shigella species causing epidemic dysentery. Epidemic outbreaks have occurred throughout the world but are often linked to confined populations, e.g. refugee camps. Approximately 5-15% of Sd1 cases are fatal since Sd1 is resistant to many antimicrobials. S. dysenteriae is also the only Shigella species producing the Shiga toxin. into eukaryotic target cells and inhibits protein synthesis (WHO, 2005).

Shigella was first described 1897 by Shiga Kiyoshi in Japan. Shigellae are rod-shaped, non-motile Gram-negative bacteria. They belong to the family of enterobacteriaceaeand do not form spores. Shigella is closely related to Escherichia and is occasionally considered as one strain of the E. coli species. In fact, Shigella shares morphological features with E. coli but it can be easily distinguished biochemically. For example, the cell wall antigens, also known as O-antigens, of Shigella and E.coli are distinct. The O-antigen is the outer polysaccharide portion of the lipopolysaccharide (LPS) and consists of repeating sugar units. Furthermore Shigella is anaerogenic, meaning it does not produce gas from carbohydrates, and it cannot ferment lactose (Levinson W., 2002;School, 1995).

The disease causing properties of Shigella are encoded on a virulence plasmid. Strains cured of this plasmid are non-pathogenic (Sansonetti, et al., 1982). If the Shigella virulence plasmid is transferred into a non-pathogenic E. coli, the plasmid confers invasiveness and cytotoxicity in vitro (Sansonetti, et al., 1983).

After passing the oesopharynx, stomach and small intestine Shigella invades the hosts’ large intestinal epithelium [figure 1.1 and (LaBrec, 1964)]. Shigella traverses the epithelial barrier through specialized membranous epithelial cells, called M-cells (Wassef, et al., 1989). M-cells transport antigens, including enteric pathogens, across the epithelium. They are located in the epithelium covering the gut-associated lymphoid tissue (GALT) (Kraehenbuhl and Neutra, 1992). M-cells are the only port of entry across the epithelium for Shigella, because Shigella cannot invade colonocytes through their apical membrane (Mounier, et al., 1992). Following passage across M-cells, the microorganism interacts with two different host cells: epithelial cells and macrophages. Shigella invades the epithelial cells from the basolateral side. It escapes from the phagosome to the cytoplasm of the cells and replicates there. In order to move intra- and intercellularly, Shigella utilizes the cytoskeleton of the host cells (Makino, et al., 1986). Once infected, the cells secrete the cytokine interleukin-8 (IL-8) to recruit neutrophils (also called polymorphonuclear leukocyte, PMNs) to the site of infection.

The other host cells Shigella encounters are resident tissue macrophages. They are situated within lymphoid follicles beneath M-cells (Jarry, et al., 1989;Soesatyo, et al., 1990). Although the macrophages phagocytose Shigella, the bacteria can escape from the phagolysosome to the cytoplasm within minutes (Finlay and Falkow, 1988;Maurelli and Sansonetti, 1988). Unlike in the epithelial cells, Shigella rapidly induces macrophage apoptosis (Zychlinsky, et al., 1992). In an apoptotic process, a cell synthesizes the molecules responsible for its own death (Arends and Wyllie, 1991). Accordingly, macrophages infected with Shigella show the two cardinal signs of apoptosis: specific morphological changes and fragmentation of nuclear DNA into multimers of approximately 200 bp.

Shigella escapes from the dying macrophage and infects further epithelial cells. The apoptotic macrophage releases the pro-inflammatory cytokines interleukin-1β and –18 (IL-1β and IL-18) to recruit neutrophils to the site of infection. Although counterintuitive, the neutrophils first support the bacterial infection. They damage the colonic mucosa by breaking tight junctions to reach bacteria in the intestinal lumen. Thus they the entry for Shigella into the epithelium. But eventually neutrophils resolve the infection. eutrophils engulf Shigella but in contrast to macrophages they prevent the escape of Shigella from the phagolysosome and kill them.

Fig. 1.1: Shigella infection.

Shigella invades the epithelium from the intestinal lumen through M-cells. After reaching the epithelium it invades epithelial cells and is phagocytosed by resident macrophages. Shigella escapes the phagosome of both cells but while Shigella replicates within epithelial cells it induces apoptosis in macrophages probably by activation of caspase-1 (Casp-1). The dying macrophages release the pro-inflammatory cytokines IL-1β and IL-18. Together with IL-8 secreted from the invaded epithelial cells, they signal for PMN (polymorphonuclear leukocyte or neutrophils). The recruited neutrophils eventually clear the infection.

The degree to which pathogenic bacteria are able to cause disease determines their virulence. Virulence depends on the resistance of the host and as well as on the invasiveness and toxicity of the bacteria. Bacterial components and proteins that mediate adhesion, invasion, toxicity, and evasion of host immune cells are termed virulence factors (Mayer-Scholl, et al., 2004). Shigella virulence factors are encoded on a 220 kb virulence plasmid which is essential for Shigella pathogenicity (Sansonetti, et al., 1982). S. flexneri invasion genes are localized in a 31 kb region of the virulence plasmid (Maurelli, et al., 1985). This region encodes the invasion plasmid antigens (ipa) operon, the membrane expression of ipas (mxi) and surface presentation of invasion plasmid antigens (spa) operons, as well as other, independently expressed genes. By transposon insertion and deletion mutagenesis, ipaB, ipaC and ipaD genes were shown to be essential for invasion, vacuolar escape, induction of macrophage apoptosis and virulence in animal models (High, et al., 1992;Menard, et al., 1993;Sasakawa, et al., 1988). IpaB, C and D are secreted by a type III secretion apparatus encoded by the mxi and spa operons (Allaoui, et al., 1992;Allaoui, et al., 1993;Andrews, et al., 1991). The type III secretion apparatus is composed of at least 30 proteins and is conserved among enteropathogenic bacteria (Cornelis and Van Gijsegem, 2000). IpaB and IpaC form a complex (Menard, et al., 1994) that appears to be sufficient to invade epithelial cells (Menard, et al., 1996).

In addition to its role in invasion, several lines of evidence indicate that the Shigella virulence factor IpaB is both necessary and sufficient to induce apoptosis in macrophages. First, mutant strains of Shigella that are invasive but do not express IpaB are not cytotoxic (Zychlinsky, et al., 1994). Second, microinjection of IpaB into the cytosol of macrophages efficiently triggers apoptosis (Chen, et al., 1996). Third, IpaB was shown to bind caspase-1 [also calledIL-1β converting enzyme (ICE)] which plays an important role in Shigella induced apoptosis (Hilbi, et al., 1998;Thirumalai, et al., 1997). Caspase-1 is a proapoptotic and proinflammatorycysteine protease. When activated, it cleaves the pro-inflammatory cytokines IL-1β and IL-18 to their biologicallyactive forms (Dinarello, 1998;Thornberry, 1994). Together these data suggest a model where IpaB is secreted into the macrophage cytosol during infection. IpaB binds to and activates caspase-1, an event that simultaneously induces apoptosis and activates the pro-inflammatory cytokines IL-1β and IL-18.

Another virulence factor encoded on the virulence plasmid is IcsA (intracellular spread, also called VirG). IcsAis essential for intra- and inter-cellular movement of Shigella (Bernardini, et al., 1989). Disruption of IcsA leads to loss of bacteria inducedintracellular actin assembly, loss of cell-to-cell spread, andmarkedly reduced virulence in humans and animal models (Bernardini, et al., 1989;Coster, et al., 1999;Lett, et al., 1989;Makino, et al., 1986;Sansonetti, et al., 1991). It is localized to the outer membrane (120 kDa form) and is secreted through a type V system (90 kDa form). IcsA is asymmetrically distributed along the bacterial body (Goldberg, et al., 1993). This is a prerequisite for the polar movement of Shigella in mammalian cells, including bacterial spreading between epithelial cells. The N-terminal part of the IcsA α domain induces the polymerization of actin by interaction with host proteins such as vinculin and neural Wiskott-Aldrich syndrome protein (N-WASP) (Egile, et al., 1999). The non-motile Shigella “hooks on” to these actin tails to move through and in between cells.

Neutrophils are an essential component of the acute inflammatory response and the resolution of microbial infection. Recruitment to inflamed or infected tissue occurs within minutes to hours. Molecules signaling for neutrophil infiltration are predominantly the aminoterminal formylated methionin bacterial peptide (fMLP), IL-8 and TNF-α from macrophages and epithelial cells, and C5a, C3a and C4a from the complement cascade (Burg and Pillinger, 2001). Vasodilatation through TNF-α results in the reduced velocity of blood flow. Physiologically circulating neutrophils in the blood contact the endothelium and transiently interact with it, a phenomenon termed rolling. Molecules mediating this are leucocyte (L), platelet (P) and endothelial (E) selectins, which permit interaction between neutrophils, and neutrophils and endothelial cells (Janeway, et al., 2001). After exposure of circulating neutrophils to chemoattractants (IL-8, fMLP, C5a, LTB4), members of the β2-integrin family mediate the conversion of the rolling state to a state of tight stationary adhesion (Burg and Pillinger, 2001). Neutrophils adhere to the endothelium and secretory vesicles of the neutrophils are mobilized. Neutrophils transmigrate either between or through endothelial cells. During migration through the tissue neutrophils cleave off or shed their selectins and proteases are liberated from the different granule subsets degrading vascular basement membranes and the intercellular matrix (Faurschou and Borregaard, 2003).

The concept of bacterial recognition is based on so-called pathogen associated molecular patterns (PAMPs), which are recognized by pattern recognition receptors (PRRs) (Gordon, 2002). PAMPs are microbial structures, which, upon interaction with elements of the host innate immune system, trigger the initiation of host protective responses accumulating in the clearance of the pathogen by phagocytic cells. PAMPs are ideal targets as they allow distinction between self and microbial non-self. They are found on all microorganisms, which allows a limited number of receptors to recognize PAMPs, and importantly they are essential for microbial survival, therefore no escape mutants can be generated (Mukhopadhyay, et al., 2004).

Neutrophils have the following pattern recognition receptors:

• lectins, e.g. dectin-1, which detect bacterial carbohydrates;

• scavenger receptors, e.g. MARCO, are structurally unrelated membrane molecules, which bind and internalize modified lipoproteins, lipopolysaccharides, and lipoteichoic acid;

• complement receptors, e.g. CR1-4, which recognize complement proteins from the serum that have opsonized the microbes;

• Fc receptors, which recognize IgG opsonized microorganisms;

• Toll-like receptors (TLRs) - TLR 2 and 4 are expressed by neutrophils, they recognize lipoproteins and lipopolysaccharide respectively (Muzio, et al., 2000);

Upon encountering bacteria, neutrophils engulf these microbes into a phagosome, which fuses with intracellular granules to form a phagolysosome (Lee, et al., 2003). In the phagolysosome the bacteria are killed after exposure to enzymes, antimicrobial peptides and reactive oxygen species (ROS). The arsenal of cytotoxic agents has been traditionally divided into either oxygen- independent or -dependent mechanisms (figure 1.2). Both of these systems probably collaborate in killing microbes (Roos and Winterbourn, 2002).

The oxygen-independent mechanisms encompass the contents of the three neutrophil granule subsets: the azurophil, specific and gelatinase granules, which contain characteristic proteases, antimicrobial proteins and peptides enzymes (Borregaard and Cowland, 1997).

Fig. 1.2: Schematic presentation of the oxygen-dependent and oxygen-independent mechanisms during neutrophil phagocytosis of bacteria.

The oxygen-independent mechanisms encompass the contents of the three neutrophil granule subsets: the azurophil, specific and gelatinase granules, which contain characteristic proteases, antimicrobial proteins and peptides, and enzymes. Lysozyme, for instance, disrupts anionic bacterial surfaces, rendering the bacteria more permeable, whereas NE degrades virulence factors. The oxygen-dependent mechanism relies on the NADPH oxidase complex that assembles at the phagosomal membrane and produces O₂ ⁻, which is rapidly converted to hydrogen peroxide. In turn, a constituent of the azurophilic granules, myeloperoxidase, generates hypochlorous acid (HOCl) from hydrogen peroxide. This presentation is taken from (Mayer-Scholl, et al., 2004)

Antimicrobial proteins such as defensins, bactericidal/permeability-increasing protein (BPI) and the enzyme lysozyme, predominantly function by disrupting anionic bacterial surfaces, probably rendering the bacteria more permeable (Kagan, et al., 1990). Proteases, such as neutrophil elastase (NE), degrade bacterial proteins, including virulence factors (Weinrauch, et al., 2002). Other proteases e.g. cathepsin G (CG) have antimicrobial activity independent of their enzymatic activity(Shafer, et al., 2002).

The importance of the oxygen-independent mechanism in is in two very rare inherited diseases, the Chediak-Higashi syndrome (Introne, et al., 1999) and Specific Granule Deficiency (Gombart and Koeffler, 2002). Both disorders are characterized by recurrent infections and shortened life expectancy. In the Chediak-Higashi syndrome, neutrophils contain giant granules resulting from specific and azurophil granule fusion. Specific Granule Deficiency is characterized by the absence of specific granules and defensins. The severity of the symptoms in these diseases underlines the fundamental role of granule proteins in host defense.

The second mechanism of neutrophil killing is oxygen-dependent (Roos, et al., 2003). Phagocytosing neutrophils undergo an ‘oxidative burst’ during which the NADPH oxidase complex assembles at the phagosomal membrane and produces O₂-, which is rapidly converted to hydrogen peroxide by the enzyme superoxide dismutase. In turn, a constituent of the azurophil granules, myeloperoxidase, generates hypochlorous acid (HOCl) from hydrogen peroxide. How the bacteria are actually killed is not known. Hydrogen peroxide is bactericidal only at high concentrations, therefore a variety of secondary oxidants have been proposed to account for the destructive capacity of the neutrophils (Hampton, et al., 1998). The importance of ROS for antimicrobial activity is validated by the susceptibility to infections of patients suffering from chronic granulomatous disease, a condition where the NADPH oxidase complex is inactive (Dinauer, et al., 2000).

In the past, studies often focused on the effects of either the oxygen-dependent or oxygen-independent mechanisms. However, a ROS function might also be to recruit K⁺ to the phagolysosome, allowing granule proteins to go from a highly organized intra-granule structure into solution (Reeves, et al., 2002). The relative contribution of ROS to these two different mechanisms is very intriguing, yet it seems premature to draw conclusions as to whether ROS contribute directly to microbial killing or only serve as activators of granule proteins (Roos and Winterbourn, 2002). Besides killing bacteria inside the phagolysosomes, neutrophils can also degranulate and release antimicrobial factors into the extracellular space (Faurschou and Borregaard, 2003). The cells can also generate neutrophil extracellular traps (NETs), which are composed of granule and nuclear constituents that kill bacteria extracellularly (Brinkmann, et al., 2004).

Despite their opposing specificity towards Shigella virulence factors NE and CG are neutral serine proteases of the same subfamily. The super-family of serine proteases contains peptidases with very diverse functions such as digestion, degradation, blood clotting, cellular and humoral immunity, fibrinolysis, fertilization, embryonic development, protein processing and tissue remodelling (Rawlings and Barrett, 1993). The common feature of serine proteases is the occurrence of a highly reactive serine residue in their catalytic center. Apart from that, serine proteases are very diverse and divided into evolutionary unrelated clans (Barrett and Rawlings, 1995). The clans differ in their overall structure and the succession of the catalytic residues in their primary sequences (Krem and Di Cera, 2001). Clan members are subdivided into families based on sequence homology (Barrett and Rawlings, 1995;Rawlings and Barrett, 1993). NE and CG are members of one subfamily of the chymotrypsin-like clan (Lesk and Fordham, 1996).

Chymotrypsin-like serine proteases combine a large group of diverse proteases, including chymotrypsin, trypsin, NE, and CG. The common features among chymotrypsin-like serine proteases are the strictly conserved geometry in their catalytic triad and their overall structure. The identical fold is composed of two asymmetricβ-barrel domains and a C-terminal α-helix (figure 1.3). Enzyme-substrate interactions involve both β-barrel domains. Each barrel consists of six antiparallel β-sheets. In some family members, e.g. , the barrels are connected by an extra α-helix. The barrels form a cleft in which the catalytic triad and the substrate-binding sites are located (Perona and Craik, 1997).

Fig. 1.3: Structure of bovine chymotrypsin.

Bovine chymotrypsin is composed of two β-barrel domains and a C-terminal α-helix. Its structure is a classical example of the common fold of chymotrypsin-like serine proteases. The catalytic triad and the substrate-binding sites are located within the cleft formed by the two β-barrel domains, which consist of six antiparallel ß-sheets each (flat arrows). The residues is shown as a cartoon. The respresentation was generated using PYMOL (DeLano, 2002) and is based on the crystallization of chymotrypsin by (Pjura, et al., 2000).

Catalytic triads cleave a peptide bond between the carbonyl group of the N-terminal amino acid and the amide group of the C-terminal amino acid. Residues adjacent to the scissile bond are termed P1 and P1’ respectively (figure 1.4). The catalytic triad consists of serine, histidine and aspartate: Ser¹⁹⁵, His⁵⁷, and Asp¹⁰² according to the chymotrypsinogen numbering of (Hartley B.S., 1971). The serine side chain forms a hydrogen bond with the imidazole ring of histidine and this histidine shares a hydrogen with aspartate. The resulting geometric arrangement allows the three amino acids to act together in a nucleophilic attack and to cleave the peptide bond of the substrate (see also figure 1.6). The cleavage of the substrate is a coordinated multi-step process. Upon binding of the substrate, the histidine ring positions the serine side chain and polarizes the hydroxyl group of this side chain by transiently binding its hydrogen. The serine hydroxyl group is now more nucleophilic and can attack the carbonyl group of the scissile bond. Aspartate supports the orientation of the histidine side chain and improves its proton acceptor qualities by electrostatic interactions.

Fig. 1.4: Nomenclature of substrate residues at the scissile bond.

The scissile peptide bond is shown in red. The residues N-terminal of the bond are termed P1-Pn, whereas C-terminal residues are called P1’-Pn’ (Berg JM, 2003)

The nucleophilic attack leads to conformational change in the vicinity of the carbonyl carbon atom of the substrate. The previous trigonal planar structure changes to a tetrahedral one. This instable formation is stabilized by the amides of serine 195 and glycine 193. The resulting formation is called the oxyanion hole or oxyanion pocket. The peptide bond is cleaved and the carbonyl group of P1 is transiently attached to the serine side chain, whereas the amide group of P1’ is bound to the histidine imidazole ring. The histidine transfers the proton of the serine to the amide group of P1’ and the former C-terminal part of the substrate can dissociate from the enzyme. To release the N-terminal part of the substrate from the serine side chain, a water molecule is needed and the same steps are repeated as described before. The histidine polarizes the water molecule, which attacks the carbonyl group of P1. The carbonyl group binds the hydroxl group of the water and the P1-serine ester bond is cleaved. The N-terminal substrate dissociates, the histidine-serine hydrogen bond is reestablished and the enzyme is prepared for a new catalytic cycle (Berg JM, 2003).

Fig. 1.5: Representation of protease-substrate interactions.

Multiple enzymatic binding sites/pockets (in blue) directly contact the P sites (in pink) of the substrate. The nomenclature of the S sites (Sn ..., S2, S1; S1’, S2’, ... Sn’) is concordant with the P sites [(Schechter and Berger, 1968); see also figure 1.4]. The scissile bond is shown in red [adapted from (Berg JM, 2003)].

An important step for the proper function of the catalytic triad is the correct binding and positioning of the substrate. The substrate binding sites (S) of the enzyme play a key role in this process. These sites are composed of different amino acids and form structural pockets, which interact with the amino acid side chains of the substrate (figure 1.5).

The S1 pocket is thought to have an important role in the substrate specificity of the individual chymotrypsin-like serine proteases (Steitz and Shulman, 1982). In general, the S1 pocket consists of three β-sheets from the C-terminal β-barrel domain [residues 189-193, 214-216, and 226-228 chymotrypsinogen numbering (Hartley B.S., 1971)]. The sat position 214 is highly conserved among the chymotrypsin-like serine proteases and contributes to the S1 binding (Perona and Craik, 1995). It aids in creating a polarenvironment for the catalytic a102 (McGrath, et al., 1992). In most of these proteases, including NE and CG, the β-sheets are connected by two surface loops and the disulfide bond Cys¹⁹¹-Cys²²⁰(Perona and Craik, 1997).

Fig. 1.6: Specificity pocket of trypsin.

A schematic representation of trypsin interacting with a peptide substrate is shown. The catalytic residues (His57, Asp 102 and Ser195, yellow) and the enzyme residues that contact substrate residues are shown (blue). The positively charged arginine side chain at position P1 of the substrate is attracted by the negatively charged aspartate 189 located at the bottom of the S1 specificity pocket. This interaction as well as five enzyme-substrate hydrogen bonds at positions P1 and P3 and glycine 193 help to position the scissile peptide bond (red) for the nucleophilic attack by the polarized hydroxyl group of Ser 195 (red arrow). The respresentation is adapted from (Perona and Craik, 1997).

Besides common characteristics, specific residues differ in the S1 pocket of the individual proteases. It is assumed that these varying amino acid compositions account for the different substrate specificities (Krem, et al., 1999). Trypsin, for example, prefers to cleave after an aor lysine at the P1 position of the substrate. This is because a negatively charged aspartate (Asp 189) is located at the bottom of the trypsin S1 pocket. The aspartate can interact with the long and positively charged side chains of arginine and lysine (figure 1.6). In contrast to that, chymotrypsin contains an uncharged serine at the bottom of its S1 pocket. Thus, uncharged, aromatic side chains of phenylalanine, tyrosine and tryptophan fit into this hydrophobic pocket. The S1 pocket of CG, for example, shows similarity to the chymotrypsin pocket (Harper, et al., 1984).

Apart from the residue at the bottom of S1, other amino acids define the characteristics of the pocket through their side chains. In enzymes with trypsin or chymotrypsin specificities, amino acid 216 is a glycine (figure 1.7). Since the glycine side chain consists only of hydrogen, it allows large substrate side chains to the base of the pocket. On the contrary, in elastase-like enzymes the side chains of the amino acids at position 190 and 216 exclude large and bulky side chains to enter the S1 pocket (figure 1.7). This structural observation is in accordance with the preference of NE for valine as P1residue in peptide substrates (Harper, et al., 1984), since valine only has a small alkyl side chain.

Fig. 1.7: Schematic model of S1 pockets from different chymotrypsin-like serine proteases.

The S1 pocket of chymotrypsin allows large, hydrophobic side chains to enter the pocket completely, whereas trypsin prefers long, negatively side chains. Val190 and 216 confine the S1 pocket of elastase to small alkyl side chains [adapted from (Berg JM, 2003)].

Yet, the model that substrate specificities are only determined by the S1 pockets is incomplete. Substitution of individual amino acids in the S1 siteof trypsin with their counterparts in chymotrypsin fails to transferchymotryptic specificity to the mutant enzyme (Graf, et al., 1987;Hedstrom, et al., 1992). Transferof specificity requires the additional exchange of amino acidsin at least two distal segments of the enzyme, none of whichdirectly contacts substrate (Hedstrom, et al., 1992). It is notable that studies on the different S1 sites of the chymotrypsin-like serine proteases have been carried out with peptide but not with full-length protein substrates. The recognition of these protein substrates likely involves other substrate-enzyme interactions outside of the binding pockets (Perona and Craik, 1995).

Human neutrophil cathepsin G (CG; EC 3.4.21.20) and human neutrophil elastase (NE; EC 3.4.21.37) are major constituents of the antimicrobial proteins of neutrophils. At the sites of infection they are mainly released into the phagocytic vacuole and to a lesser extend to the extracellular space to exert their specific an unspecific functions. One of the specific functions of NE is the degradation of virulence factor of Shigella and other Gram-negative enteropathogenic bacteria. It is intriguing that CG does not cleave these effectors since the two proteins share many characteristics.

As mentioned above, CG and NE belong to the same subfamily of chymotrypsin-like serine proteases. Thus they are homologous and their primary amino acid sequence is to 37% identical. The residues of the catalytic triad are the same positions (His⁵⁷, Asp¹⁰² and Ser¹⁹⁵) and their overall fold consists of two ß-barrels. In addition to that, NE and CG show a striking structural similarity beyond their overall fold (figure 1.8). When superimposed, the α-carbons of NE and CG only display a root mean square deviation (RMSD) of 0,9 Å. The RMSD describes the difference in localization of the α-atoms at similar positions in the structure. A RSMD of zero means that the structures are identical in conformation (Maiorov and Crippen, 1994). However, the crystal structures show that the S1 sites of NE and CG are different (Bode, et al., 1989;Hof, et al., 1996;Navia, et al., 1989). The S1 pocket of CG can harbor large and bulky side chains whereas only amino acids with small alkyl side chains fit into the S1 pocket of NE. Using peptide substrates, it was confirmed that CG and NE prefer different amino acids at the P1 position: phenylalanine and lysine for CG versus valine and leucine for NE (Harper, et al., 1984;Tanaka, et al., 1985). In general, NE hydrolyzes peptide substrates much faster than CG (Harper, et al., 1984;Tanaka, et al., 1985).

Human NE and CG are processed and stored in the same manner. They are both synthesized as inactive zymogens in the premyeloid and myeloid stage of neutrophil and monocyte differentiation in the bone marrow (Campbell, et al., 1989;Fouret, et al., 1989). Protein synthesis seems restricted to the differentiating neutrophil, since no transcription, at least of the NE gene (Fouret, et al., 1989), is observed in neutrophils after they have left the bone marrow. Both enzymes are N-glycosylated. NE has two asparagine-linked carbohydrate chains (asparagine 95 and 144, chymotrypsin numbering), whereas CG has one potential glycosylation site [asparagine 60 (Salvesen, et al., 1987;Sinha, et al., 1987)]. The composition of the complex mannose oligosaccharides sugars can differ resulting in different NE and CG isoforms (Kim and Kang, 2000;Lindmark, et al., 1990;Watorek, et al., 1993). How the glycosylated precursors are transferred to the maturing granule compartment is unclear, but obviously it is independent of the mannose-6-phophate receptor (Glickman and Kornfeld, 1993;Hasilik, 1992;Rijnboutt, et al., 1991;Rijnboutt, et al., 1991) which is often involved in targeting proteins to lysosomes (Kornfeld and Mellman, 1989).

In the developing granule, NE and CG are post-translationally processed to mature, active proteins. The N-terminal processing of the preproenzymes involves two cleavage steps: first a signal peptide, then the two-residue activation peptide is cleaved off (Brown, et al., 1993;McGuire, et al., 1993;Salvesen and Enghild, 1990;Urata, et al., 1993). Both mature proteins start with an isoleucine. This N-terminal processing seems essential for activity of the protein, since recombinant expression of full-length NE or mature NE starting with a methionine did not yield active proteins (Li and Horwitz, 2001). In addition, the C-terminus of both enzymes is processed to its individual mature form (Gullberg, et al., 1994;Salvesen and Enghild, 1990). However, the carboxyl prodomains do not seem to be important for proper targeting and enzymatic activity (Gullberg, et al., 1995). Finally, the 267 amino acid NE precursor is processed to its 218 amino acid mature form and the CG 255 amino acid precursor results in a 226 amino acid mature CG protein.

The active proteins are predominantly stored in the azurophilic granules of neutrophils in fairly high concentrations [1-2 pg/cell, (Wiedow, et al., 1996)]. NE and CG are very basic with a PI of > 9 and ~12, respectively, and through their arginine residues they are probably anchored to the negatively charged heparin and chondroitin sulfate proteoglycan matrix of the granule. Upon activation, neutrophils discharge their granule contents into the bacteria containing vacuole or to the extracellular space. A minimal proportion of the active proteins is also localized on the surface of unstimulated neutrophils (Owen, et al., 1995). This surface expression is increased when the neutrophils are exposed to chemoattractants such as fMLP (Owen, et al., 1995). The correct distribution of NE in the granules and on the surface seems important to prevent neutrophil deficiencies (neutropenia). Hereditary neutropenia is rare but predisposes people to infections. In cyclic neutropenia (CN) the neutrophil number in the blood oscillates from zero to normal (Lange, 1983;Morley, et al., 1967) whereas in severe congenital neutropenia (SCN) the total neutrophil number is drastically reduced (Ancliff, et al., 2001). All cases of CN and 75% of SCN cases are caused by mutations in the NE gene (Horwitz, et al., 1999). These mutations are thought to interfere with the correct targeting of NE from the trans-golgi network to the granules and the plasma membrane (Horwitz, et al., 2004).

Fig. 1.8: Superimposition of the NE and CG crystal structures.

190 α-carbons of the crystal structures of NE (yellow) and CG (grey) were superimposed with a root mean square deviation (rmsd) of 0,9 Å. The catalytic triad consisting of histidine (position 57), aspartate (102) and serine (195) is shown in purple. The superimposition was achieved using the SWISS-PDB-VIEWER (Guex, 1997).

Despite its low activity towards peptide substrates, CG, like NE, unspecifically degrades extracellular matrix components such as proteoglycan, collagen, laminin, fibronectin and even elastin (Baggiolini, et al., 1979;Caughey, 1994;Roughley, 1977;Roughley and Barrett, 1977). These proteolytic characteristics may be helpful for the neutrophil egress from the bloodstream and their subsequent transmigration through tissue. But the role of NE and CG in these processes remains controversial (Shapiro, 2002).

At sites of inflammation, neutrophils degranulate and discharge NE and CG mainly into the phagocytic vacuole and to a lesser extent into the extracellular space (Faurschou and Borregaard, 2003). When exposed to the extracellular environment, the enzymes are either membrane-bound, part of neutrophil extracellular traps (NETs) (Brinkmann, et al., 2004), or freely released. To avoid digestion of healthy tissue NE and CG levels are tightly controlled by their physiological inhibitors, α1-antitrypsin and α1-antichymotrypsin, respectively (Beatty, et al., 1980). These plasma-derived inhibitors belong to the family of serpins (serine protease inhibitors) and are highly abundant extracellular proteins (Shapiro, 2002). They can inhibit free but not membrane-bound NE and CG (Owen, et al., 1995). Hereditary deficiency in α-antitrypsin causes chronic destruction of alveolar walls and leads to pulmonary emphysema development (Eriksson, 1984;Gadek, et al., 1981). Deficiencies in α1-antichymotrypsin predispose patients for lung diseases (Faber, et al., 1993). In addition, NE is present in cystic fibrosis airways (Delacourt, et al., 2002) and induces excess mucus production in chronic obstructive pulmonary disease (COPD) (Shapiro, 2002).

Apart from their degradative properties, NE and CG are thought to be important regulatory tools in inflammatory processes (Wiedow and Meyer-Hoffert, 2005). For example, NE induces IL-8 expression via toll-like receptor 4 in vitro (Devaney, et al., 2003) and CG activates PAR-4 (protease-activated receptor) to initiate thrombocyte aggregation (Sambrano, et al., 2000).

The importance of NE and CG in immunity is proven by various studies. Mice deficient in NE and / or CG are susceptible to fungal infections despite normal neutrophil development and recruitment (Tkalcevic, et al., 2000). The same study indicates that NE and CG act as effectors in the endotoxic shock cascade downstream of TNFα. In accordance with the observed specificity of NE for virulence factors of Gram-negative bacteria, mice deficient in NE have impaired survival following infections with gram-negative pathogens. In contrast, mice deficient in CG are more susceptible to Gram-positive bacteria (Belaaouaj, et al., 1998;Reeves, et al., 2002). Interestingly, CG shows an antimicrobial activity that is independent of its enzymatic function (Bangalore, et al., 1990).