The Thermolysin Family (M4)
The thermolysin family of enzymes is classified as the M4 family of metallopeptidases (http://merops.sanger.ac.uk/). At least 200 (in 2007) sequences belong to the M4 family, and they are found in bacteria, fungi, and archaea. Thermolysin from Bacillus thermoproteolyticus is the type-example of the M4 family enzymes. All peptidases in the family bind a single catalytic zinc ion. The zinc ion is tetrahedrally co-ordinated by three amino acid ligands and a water molecule. The water molecule is activated and forms the nucleophile during catalysis. Typically, the M4 peptidases consist of a presequence (signal sequence) that is cleaved off during export, a propeptide sequence with inhibitory and chaperone functions (facilitating folding) and a peptidase unit. The propeptide remains attached until the peptidase is secreted and can be safely activated (1).
Diseases in which peptidases of the family M4 are implicated
Enzymes of the M4 family members are secreted metallopeptidases that degrade extracellular proteins and peptides for bacterial nutrition. They are implicated as key factors in the pathogenesis of various diseases. λ-Toxin, the thermolysin like M4 peptidase from Clostridium perfringens degrades various host proteins and activates the precursors of clostridial potent toxins. λ-Toxin degrades immunoglobulin G, complement C3 component, fibrinogen, fibronectin and α2-macroglobulin, that contribute to innate or adaptive immune defense against infections (2). Another member of the family, coccolysin from Enterococcus faecalis, has been implicated in several opportunistic infections caused by E. faecalis such as soft tissue and urinary tract infections, intra-abdominal abscesses and root canal infections, secondary bacteremia and food poisoning (3). Coccolysin also hydrolyzes casein, gelatin and hemoglobin, and clots milk and inactivates human endothelin-1 (4). The M4 thermolysin like peptidase (hemagglutinin/proteinase) from Helicobacter pylori and Vibrio cholerae is implicated in the pathogenesis of these organisms, and are believed to be causative agents for gastritis, peptic ulcer and gastric carcinoma (5) and cholera, characterized by severe vomiting and watery diarrhea (6–8). The hemagglutinin/proteinase from V. cholerae has also been shown to affect intracellular tight junctions by degrading occluding (9). The M4 thermolysin like peptidase from Legionella cleaves α1-antitrypsin (10), tumor necrosis factor α (11), interleukin 2 and CD4 on human T cell surfaces (12). It has therefore been suggested that it may have a role in the virulence of Legionaire’s disease and pneumonia (13). Pseudolysin, another M4 family member is the extracellullar elastase of Pseudomonas aeruginosa. Pseudomonas aeruginosa is important for the pathogenesis of P. aeruginosa infections (14). The contribution of pseudolysin to disease may be direct, by destructing tissue and damage cell functions, or indirectly by interfering with host-defense mechanisms by impairing the normal function of host proteases. Pseudolysin cleaves casein, elastin, and synthetic peptides (15), human Ig G (16,17), collagen types III and IV (18), serum α1-proteinase inhibitor (15) and human bronchial mucosal proteinase inhibitor (19). It has severe hemorrhagic activity and muscle damaging effects (20), and is connected to lung infections (21–24). Strong evidences suggest that pseudolysin is involved in chronic ulcers by degradation of human wound fluids and human skin proteins (25). Pseudolysin has also been connected to corneal infections, causing corneal liquefaction which can be sight threatening (26). Experimental attempts were made to treat rabbit cornea infections using antibiotics and P. aeruginosa elastase specific inhibitors as adjuncts to antibiotics (27,28), and some inhibitors showed promising effects (29). Based on the current knowledge it is reasonable to believe that enzymes of the M4 family of metallopeptidases are putative targets for therapeutic intervention.
Specificity of peptidases in the thermolysin family
The substrate specificity of thermolysin has been studied in large detail while information on other members of the family is scarce. The major site for thermolysin cleavage specificity, (the S1′ site), accepts large hydrophobic residues. Thermolysin preferentially cleaves at the N-terminal side of hydrophobic or bulky amino side chains such as Leu, Phe, Ile and Val (Figure 1). Thermolysin also cleaves bonds of Met, His, Tyr, Ala, Asn, Ser, Thr, Gly, Lys, Glu or Asp at the P1 ′site (30,31). The specificity of aureolysin, coccolysin and lambda toxin (λ-Toxin) is similar to that of thermolysin (3,4,32–34) with preference for hydrophobic P1′ residues (Leu, Val, Tyr, Ile and Phe) and Ala. Similar to thermolysin, pseudolysin and griselysin also favor hydrophobic residues at the P1′ position, but have higher preference for aromatic residues than for other hydrophobic residues. Hydrophobic residues are also favored at P1 and P2′ (15,35,36). Specificity of vimelysin is different from that of thermolysin. Vimelysin specifically recognizes Phe at the P1′ position, whereas thermolysin specifically recognizes Phe at the P1 position (37).
Industrial and therapeutic use of peptidases of family M4
Thermolysin is used as a peptide and ester synthetase in the production of the artificial sweetener aspartame (38–49). The biotechnology industry is also using thermolysin as a non-specific proteinase to obtain fragments for peptide sequencing. Furthermore, the discovery of the mechanistic similarities between thermolysin, carboxypeptidase A and angiotensin converting enzyme was a key factor in the development of antihypertensives in present clinical use (50). Enzymes from thermophilic microorganisms (thermozymes) have unique characteristics such as temperature, chemical, and pH stability. Thus, thermozymes have considerable potential in many industrial applications, and several thermozymes of the thermolysin family (including thermolysin) are in industrial use (51–54). Vimelysin from Vibrio str. T1800 showed great potentials in peptide condensation reactions because of high activity in organic solvents. The optimum pH of vimelysin was pH 8.0 and pH 6.5 when casein and N-[3-(2-furyl)acryloyl]glycl-L-leucin amide (FAGLA) were used as substrates, respectively. The optimum temperature of vimelysin was 50 °C when casein was the substrate, but 15 °C with FAGLA as the substrate (6,55). Vibriolysin from Vibrio proteolyticus has several industrial as well as biomedical applications (53,54). It is used to mediate the coupling of N-protected aspartic acid and phenylalanine methyl ester to yield N-protected aspartylphenylalanine methyl ester, a precursor of the sweetener aspartame. Vibriolysin from V. proteolyticus is also used for the removal of necrotic tissue from wounds such as burns or cutaneous ulcers, and is reported to stimulate the healing of partial-thickness burn wounds (56).
Tertiary structures of the family M4 of metallopeptidases
The 3D structures are known for four members of family M4. These structures are thermolysins from Bacillus cereus (1npc) (57) and B. thermoproteolyticus (1kei) (58), pseudolysin from P. aeruginosa (1ezm) (59) and aureolysin from Staphylococcus aureus (1bqb) (60). Thermolysin from B. thermoproteolyticus was the first metalloendopeptidase crystallized and the 3D structure determined (58,61,62). Both thermolysin structures have four calcium ions suggested to be responsible for their thermostability (61,63–65). The aureolysin structure has three calcium ions, while pseudolysin only has one calcium ion and two disulphide bridges. The structures of at least nine mature thermolysins have also been solved at resolutions of 1.7–2.2 Å (58,66–70). The 3D structure consists of a two-domain structure with the active site between the domains. The 3D structures indicate that two of ligands coordinating the zinc are the histidines of the motif: His-Glu-Xaa-Xaa-His (‘HEXXH’). The third ligand is a Glu located 18–72 residues C-terminal to the HEXXH motif, while the fourth ligand is a water molecule. The N-terminal domain includes both α-helices and β-sheets, and carries the HEXXH motif (X is any amino acid residue). The C-terminal domain is predominantly helical and carries the third zinc ligating amino acid. The HEXXH motif and the third amino acid coordinating zinc (Glu) are located in α helices connected by a turn. This turn is required to bring the amino acids together forming the zinc binding site (Figure 2) (71) Furthermore, at least 25 complexes of thermolysin with different ligands have been solved by X-ray crystallography (72–80). In July 2004, the first X-ray structure of pseudolysin (1u4g, unpublished) in complex with a small molecule inhibitor was also deposited in the PDB database.
The thermolysin family of enzymes are classified by the MEROPS database, in the subclan MA(E) known as ‘Glu-zincins’. The term ‘Glu-zincins’ is based on that the third zinc coordinating ligand is a Glutamic acid (E) located 18–72 residues C-terminal to the HEXXH motif as opposed to the subclan MA (M) known as the ‘Met-zincins’. In ‘Met-zincins’ the third zinc ligand is a histidine or an aspartate located in the extended motif HEXXHXXGXX(H/D), (X is any amino acid residue). The region C-terminal to the third histidine forms a loop that ends at the catalytic zinc in a unique turn. The turn contains a strictly conserved methionine and is designated the ‘Met-turn’ (81) (Figure 3). In subclan, MA(E) the spacing between the third zinc ligand and an aspartic acid is similar within the subclan (four amino acids). This Glu-(Xaa)3-Asp motif provides a means in addition to the HEXXH motif for finding other peptidases that may share the common thermolysin fold (71).
Catalytic mechanism of the thermolysin family M4
The catalytic mechanism of thermolysin (type-example of the family) is probably the best studied of all metallopeptidases, but all details are not completely understood. It has been suggested that the enzymes in the family undergo hinge-bending motion during catalysis (58,70,82,83). Two mechanisms of action have been proposed, one in which the glutamate of the HEXXH motif acts as a proton acceptor during catalysis (84,85) and a second in which an active-site histidine acts as the general base instead of the glutamate (86,87). As previously mentioned, the catalytic zinc in native thermolysin is tetrahedrally coordinated by His142, His146 of the HEXXH motif, Glu166 of the Glu-(Xaa)3-Asp motif, while the fourth coordination is a water molecule. In the first suggested mechanism, the catalytic cleavage of peptide bonds proceeds by a general base type mechanism with the attack of a water molecule or hydroxide ion on the carbonyl carbon of the scissile bond. The incoming substrate is presumed to displace the water molecule towards Glu143 of the HEXXH motif, such that both hydrogen atoms of the water molecules are hydrogen bonded to Glu143, while the oxygen still ligates the zinc ion (Figure 4). In that way the nucleophilicity of the water molecule is enhanced, and the remaining lone pair of the water oxygen is directed towards the substrate being aligned for nucleophilic attack. The water attacks the carbonyl carbon and forms an intermediate (Figure 4B). The proton accepted by Glu143 is transferred to the leaving nitrogen. The side chain oxygen of Asn112 and the backbone carbonyl of Ala113 accepts hydrogen bonds from the doubly protonated tetrahydral nitrogen of the scissile bond (Figure 4C). A second proton transfer via Glu143 to the leaving nitrogen forms the product (Figure 4D). In the second suggested mechanism three other amino acids are suggested to be important for catalysis besides the activated water molecule, the zinc ion and the zinc coordinating amino acids. These amino acids are Glu143, which acts as an electrophile, Asp226 and His231. Asp226 orientates the imidazolium ring of His231, while His231 acts as proton donor and general base (Figure 4).
Temperature adaptation of the thermolysin family M4
T50 is defined as the temperature where an enzyme looses 50% of the activity in a 30-min period of incubation. van den Burg et al. reported T50 of 82 °C, 76.7 °C and 68.5 °C for the thermolysin like enzymes from B. thermoproteolyticus, B. caldolyticus and B. stearothermophilus CU21, respectively (88). Mutations of six amino acids of the thermolysin like enzyme from B. stearothermophilus CU21, situated near the binding site of the third calcium ion, led to the production of an enzyme with a T50 of 96.9 °C (89). Thermolysins from Alicyclobacillus acidocaldarius, Bacillus sp. Strain EA1, Bacillus megaterium have T50 of 82 °C, 85 °C and 58 °C, respectively (90–92). Bacillolysin from B. cereus is less stable than thermolysin (93). It seems to contain fewer hydrogen bonds than thermolysin even though the four calcium ions are still present (57,82). The additional thermal stability of thermolysin might be due to a combination of rigidification by proline residues, hydrogen bonding or salt-bridges (57). In their 1974 article, Mattews et al. stated (94): ‘it is suggested that the enhanced stability of thermostable proteins relative to thermolabile ones cannot be attributed to a common determinant such as metal ion or hydrophobic stabilization, but in a given instance may be due to rather subtle differences in the hydrophobic character, metal binding, hydrogen bonding, ionic interactions, or a combination of all of these’. In other words, thermostable proteins have similar three-dimensional structures to their less stable counterparts, and the differences in stability are due to a combination of different effects which can be more or less important from protein to protein. A metalloprotease from Listeria monocytogenes is active at temperatures up to 80 °C, and with a T50 of 88 °C it is the most heat-stable natural thermolysin-like peptidase known (95). Vimelysin from Vibrio str. T1800 has an optimum temperature of 50 °C, and high activity in the presence of organic solvents (6,55). The peptidase from Serratia marcescens has a T50 of 50 °C (96).
Vibriolysin from Antarctic bacterium str. 643, VAB, is an M4 peptidase from a cold-adapted organism. However, that VAB is cold adapted has so far not been proven biochemically and the analyses done so far are based on comparative amino acid content analysis and molecular dynamics simulations experiments. Cold-adapted enzymes have attracted attention due to their industrial potentials, and so far the areas of uses include: additives in the food industry (fermentation, cheese manufacturing, baking, meat tenderization), vaccine development, additives to detergents, environmental bioremediations (digesters, in composting, oil or xenobiotic biodegradation), biotransformation and molecular biological applications, heterologous gene expression in psychrophilic hosts to prevent inclusion bodies, and the paper industry (51,97) . A psychrophile is defined as an organism that permanently thrives at temperatures close to the freezing point of water.
VAB as a model to understand cold adaptation
Comparative amino acid sequence analysis of VAB and other family M4 peptidases yielded insight into molecular features that may be connected to cold adaptation of VAB. It is obvious that VAB uses a number of factors for adopt to cold climate. Amino acid sequence analysis of 44 (September 2002) sequences of the thermolysin family of enzymes showed that VAB compared to the other enzymes has: (i). Fewer arginines. (ii) A smaller Arg/(Lys + Arg) ratio. (iii) A lower fraction of large aliphatic side chains, expressed by the (Ile + Leu)/(Ile + Leu + Val) ratio. (iv) More methionines. (v) More serines. (vi) More of the thermolabile amino acid asparagine.
Ideally the X-ray structure of VAB would give tremendous insight into the structural aspects of cold adaptation. However, the X-ray structure is not available to date, and we therefore used a 3D model constructed by homology with pseudolysin to study the structural patterns of VAB compared with other M4 peptidases. The sequence identity to template and the structural validation carried out on the built model indicated that the model was reliable. MD simulations for 3 nanoseconds (ns) of VAB, pseudolysin and thermolysin supported the assumption that cold-adapted enzymes have a more flexible 3D structure than their thermophilic and mesophilic counterparts, especially in some loop regions. The structural analysis of the three enzymes indicated that:
- • VAB has fewer intramolecular cation- π electron interactions and hydrogen bonds than its mesophilic (pseudolysin) and thermophilic (thermolysin) counterparts.
- • Lysine is the dominating cationic amino acids involved in salt bridges in VAB, while arginine is the dominating in thermolysin and pseudolysin.
- • VAB had a greater volume of the inaccessible cavities than pseudolysin and thermolysin.
- • The electrostatic potentials at the surface of the catalytic domain were also more negative for VAB than for thermolysin and pseudolysin (11).
These differences indicate that there are weaker intramolecular interactions leading to greater flexibility in VAB than in thermolysin and pseudolysin. Comparative amino acid sequence analysis suggested that VAB has a slightly lower fraction of large aliphatic amino acids, fewer arginines, more methionines, more serines and more asparagines than other enzymes of the thermolysin family. These characteristics are part of the general mechanisms of cold adaptation. MD simulations of VAB, thermolysin and pseudolysin supported the hypothesis that increased local flexibility is a strategy for cold adaptation. The MD simulations suggested that loop regions and substrate binding regions in VAB have greater structural flexibilities than in the mesophilic and thermophilic counterpart, expecially in two loop regions. These loop regions had at least one glycine residue.
Psychrophilic enzymes produced by cold-adapted microorganisms display a high catalytic efficiency over a temperature range of roughly 0–30 °C and are most often associated with high thermosensitivity. Increased flexibility has been proposed to be the main structural feature of cold-adapted enzymes. Their high specific activity has been related to their possible higher flexibility leading to rather high instability. Several studies have indicated that adaptation of enzymes to temperature is achieved by minor adjustments of the flexibility in regions outside the active site, and these localized increases in conformational flexibility constitute an essential element in cold adaptation (98–101). Increased local flexibility of regions involved in conformational changes occurring during catalysis and rigidity in all areas not directly related to the catalytic process are adaptive traits observed in several cold-adapted enzymes (40,100,102–104). However, more experimental evidence is necessary to clearly conclude that higher structural flexibility is a consequence of cold adaption of enzymes. Lack of experimental procedures for the study of global or local protein flexibility of enzymes has led to rare direct experimental evidence supporting the temperature-dependent adaptation of flexibility in proteins. Studies that compare thermal factors (B-factor) of crystal structures of homologous protein have been used to give an indication about the structural flexibility of psychrophilic, mesophilic, and thermophilic proteins. However, other factors such as resolution, crystal packing, solvent content and experimental temperature do affect B-factors making such comparison faulty. Other methods used in accessing protein flexibilities include amide hydrogen-deuterium (H/D) exchange by Fourier-transformed infrared (FTIR) spectroscopy and the accessibility of intrinsic tryptophan residues by monitoring the quenching of protein intrinsic fluorescence by small quencher molecules such as acrylamide (105,106). In the former, the disappearance of the amide N-H band by FTIR spectroscopy is used for studying protein H/D exchange (98).
Cold-adapted enzymes have great potentials for biotechnological applications but a major draw back is that they have to be characterized adequately and screened for useful purposes. A thorough knowledge of the relationship between function and stability is needed. This will allow redesigning known enzymes to function at lower temperatures (107).
Inhibition of enzymes of the thermolysin family
Enzymes of the family M4 are important drug targets. The knowledge about the mode of inhibition of thermolysin has been used in the design of antihypertensive in clinical use. The increase in resistance to antibiotics is giving a greater need for development of anti-infective agents that can be used alongside current antibiotics. High affinity inhibitors of the M4 family of proteinases have immerse therapeutic as well as industrial values.
The 3D structures of thermolysin in complex with competitive or irreversible inhibitors (108), showed that the inhibitors replace the zinc coordinated water molecule with a carboxylate, hydroxamate or phosphoramidate group, providing structural analogues for the tetrahedral intermediates formed during catalysis. Naturally occurring inhibitors of M4 proteinases are phosphoramidon from Streptomyces tanashiensis, the proteinaceous nous inhibitors Streptomyces metalloproteinase inhibitor (SMPI) and the inducible metalloproteinase inhibitor (IMPI) from the moth Galleria mellonella caterpillar. IMPI was identified in the larvae of the greater wax moth, G. mellonella, and shares no sequence similarity with known vertebrate or invertebrate proteins or other natural inhibitors of metalloproteinases (109). IMPI is not active against matrix metalloproteinases but inhibits microbial metalloproteinases, such as bacterial thermolysin. The 3D structure of IMPI has not been determined, and its molecular mass is 8.6 kDa (110). The IMPI contains 10 cysteine residues which are likely to form five disulfide bridges. The pronounced stability of the IMPI against heat or trichloroaceteic acid (TCA) is attributed to the five intermolecular disulfide bonds (111). Low sequence similarity between IMPI and known vertebrate or invertebrate proteins or other natural inhibitors of metalloproteinases precludes the construction of 3D homology models of IMPI. Certainly, the understanding of the inhibition interactions between IMPI and the peptidases in the thermolysin family will be of utmost importance in the development of effective anti-infectives suitable for clinical use.
Streptomyces metalloproteinase inhibitor from Streptomyces nigrescens TK-23, was the first known proteinaceous inhibitor of metalloproteinases of the gluzincin family. The NMR structure of SMPI (112) shows that the protein contains two disulphide bridges. The inhibitor function is connected to Cys64-Val65 segment in the Arg60-Ala73 loop. This loop also contains a disulphide bridge between Cys64 and Cys69, resulting in a rigid loop structure (6). The several hundred folds stronger derived binding affinity of SMPI compared with presently available inhibitors indicates a huge potential in designing M4 proteinase inhibitors that mimic the peptidase binding region of SMPI. However, for a rational design of such inhibitors, it is important to distinguish between the ‘structural epitope’ and the ‘functional epitope’ of SMPI (113), and to identify residues along the protein-protein interfaces that yield significant contributions to the free energy of binding. The X-ray structures of the complexes of SMPI with M4 peptidases would ideally guide a comprehensive functional survey of residues present at the interaction interface. However, the X-ray structures of such complexes are not available. Experimental work was done to probe the effect of mutation of the residues at the reactive site loop of SMPI. These studies shed light onto the roles these amino acids for the interactions of SMPI with thermolysin, pseudolysin and vimelysin (6). Mutational analysis of the SMPI-thermolysin complex showed that Arg60, Arg61 and Arg66 within the rigid active site loop are important for thermolysin inhibition. The double mutant, R60/61A, had 5 times weaker inhibition than the wild type, while the single mutant R66A showed 100 times reduced inhibition, whereas triple mutant R60/61/66A showed 200 reduced inhibition compared to the wild type (6).
Docking and molecular dynamics (MD) simulation studies have been used to study the interactions of SMPI with thermolysin and pseudolysin (114). Computational alanine scanning of the residues at the interaction interfaces has predicted residues that appear to be responsible for the experimentally determined binding affinities. These predictions are good starting point for further experimental studies and/or de novo design of novel inhibitors that might be of therapeutic and/or industrial value (114). The MD simulation studies also revealed interactions that have not previously been known between M4 peptidases and SMPI. All the SMPI-thermolysin and SMPI-pseudolysin complexes studied had auxiliary binding sites away from the active site, stabilizing the protein-protein interactions. These auxiliary binding sites were also subjected to computational alanine scanning which predicted hot spot residues that might contribute to the high free energy of binding (114). However, there is no available experimental studies performed to confirm the role and or existence of such auxiliary binding sites. For example, the docking and MD simulations indicated that there are altogether almost 50 residues at the Pseudolysin-SMPI interfaces contributing to the complex building. For each of the pseudolysin-SMPI complexes there are 23–30 at the pseudolysin side, and about 20–22 at the SMPI side. These residues are the so-called ‘hot-spots’ of binding energy, and identification of the hot spot residues at the interaction interfaces is important. Molecular mimics of ‘hot-spots’ may functions as lead compounds for further rational and combinatorial design of new molecules with improved affinities (115).
The SMPI’s interactions’ with pseudolysin and thermolysin
Experimental binding studies have indicated that pseudolysin binds its inhibitors 10–100 fold stronger than thermolysin (6). The experimentally determined inhibition constants (Ki) of SMPI is 2.54 × 10−12 M for pseudolysin, and 1.14 × 10−10 m for thermolysin. Thermolysin is irreversibly inhibited by ClCH2CO-HOLeu-OCH3 (HOLeu: N-hydroleucine), while pseudolysin is not, whereas ClCH2CO-HOLeu-Ala-Gly-NH2 can irreversibly inhibit both enzymes (116). It has therefore been postulated that the binding pocket of pseudolysin is bigger than the binding pocket of thermolysin, since an increase in the length of the ligand by two amino acids (Ala and Gly) results in inhibition of both pseudolysin and thermolysin. Most of the amino acids involved in active site interactions and ligand binding are highly conserved between thermolysin and pseudolysin. Previous studies of thermolysin have indicated the following amino acids are important for substrate/inhibitor binding or catalyses (85) (the corresponding amino acids in pseudolysin are given in brackets): Asn112 (Asn112), Ala113 (Ala113), Trp115 (Trp115), Arg203 (Arg198), Tyr157 (Tyr155), Phe130 (Phe129), Leu133 (Leu132), Val139 (Val137), Glu143 (Glu141), Ile188 (Ile186), Val192 (Ile190), Leu202 (Leu197) and His231 (His223). Our MD simulation studies have however indicated that some amino acids not conserved between pseudolysin and thermolysin were also important for SMPI binding to pseudolysin. These amino acids were Tyr114, Asp206 and His224. The corresponding amino acids of thermolysin are Phe114, Tyr206 and Ile224, respectively. These differences are believed to contribute to the stronger affinities of SMPI for pseudolysin than for thermolysin (114, 117). A major problem in the theoretical calculations was the overestimation of the binding affinity compared with the expeirmenental results, most probably due to inappropriate parameterization of the zinc ion. The fact that the zinc ion is in the active site complicates the calculations. This problem need urgent attention due to the importance of the roles metal ions play in metalloproteinases (116).