Polaina J., MacCabe A.P. (ed.). Industrial Enzymes. Structure, Function and Applications

Подождите немного. Документ загружается.

METALLOPROTEASES 227

Figure 3. The active site of thermolysin. Amino acids involved in catalysis and mentioned in the text

are shown in detail

and its subsequent nucleophilic attack on the carbonyl carbon atom of the scissile

bond. A protonated His residue (His231 in thermolysin), which is bound via a

hydrogen bond to an Asp residue in the active site cleft (Asp226 in thermolysin), is

claimed to stabilise the transition state by forming a hydrogen bond to the carbonyl

oxygen atom. The breakdown of the tetrahedral intermediate in the transition state is

accomplished by protonation of the amide nitrogen of the scissile bond via Glu143.

An alternative proposal, the ‘reverse protonation’ mechanism, derives from the

observed dependence of the catalytic constants on pH in the hydrolysis of arazo-

formyl peptides which are poor substrates for thermolysin. This alternative questions

the role of Glu143 as a general base and instead proposes the assignment of this

function to a His residue (His231 in thermolysin) in the active site (Mock and

Aksamawati, 1994; Mock and Stanford, 1996). In this case, the scissile bond is

activated for hydrolysis by direct coordination of the carbonyl oxygen atom to

the zinc ion as a potent Lewis acid with simultaneous displacement of the water

molecule otherwise bound to the zinc ion. By analogy to serine proteases, His231

along with Asp226 is supposed to enable proton transfer by deprotonation of an

incoming water molecule which is not bound to the zinc ion.

A recent quantum chemical study (Pelmenschikov et al., 2002) has confirmed

the key role of Glu143 in the thermolysin-catalysed hydrolysis of peptides, thus

supporting the initially proposed mechanism of Matthews and co-workers. In the

context of these data, the reverse protonation mechanism seems to be less favoured.

228 MANSFELD

Glu143 is absolutely conserved in all metalloproteases having the HEXXH

motif, whereas His231 is only partly conserved in metalloproteases employing the

same catalytic mechanism. Additional arguments strengthening the essential role of

Glu143 are provided by site-directed mutagenesis studies. The charge-conserving

replacement of Glu143 by Asp in NprM from B. stearothermophilus MK232, an

enzyme identical to thermolysin at the primary structure level, led to complete

loss of activity (Kubo et al., 1992). This makes it rather unlikely that Glu143

acts solely as a negatively charged counter-ion providing electrostatic stabilisation

of the transition state. Replacement of Glu143 in TLP from B. subtilis caused

nearly complete loss of secreted enzyme, whereas His231 (thermolysin numbering)

mutants were secreted and retained a certain degree of activity (Toma et al., 1989).

Replacement of His231 by Ala or Phe in the TLP from B. stearothermophilus

reduced the k

cat

value 430 and 500-fold, respectively (Beaumont et al., 1995).

These mutants showed reduced pH dependence in the alkaline range. Bearing all

this in mind, the essential role of His231 is not supported. Attempts performed

by our group to produce completely inactive TLPs from B. stearothermophilus

showed that the least active enzyme was the Glu143Gln mutant having less than

0.1% residual activity, whereas the His231Ala mutant enzyme restored 1.6% of

wild-type activity. A combination of both mutations was not able to decrease the

activity further (Mansfeld, unpublished results). The enzymes were expressed in

E. coli and renatured as described in Mansfeld et al. (2005).

Based on structural comparisons, a hinge-bending motion leading to the closure

of the active site cleft upon substrate or inhibitor binding has been proposed

(Holland et al., 1992, 1995; Hausrath and Matthews, 2002). The involvement

of conserved glycine residues at positions 78, 135 and 136 has been assumed

and experimentally proven for thermolysin (Holland et al., 1992), a TLP from

B. cereus (Stark et al., 1992) and a TLP from B. stearothermophilus (Veltman

et al., 1998). Recent comparisons of the ‘open’ and the ‘closed’ structures (Hausrath

and Matthews, 2002) have however shown that these two regions cannot account

completely for the observed movement of the domains at the active site cleft. The

concerted movement of a group of side chains is proposed instead.

Thermolysin and related proteases preferably cleave substrates having bulky

hydrophobic residues (Leu, Phe) in the P

' position and smaller amino acids

in position P

(nomenclature according to Schechter and Berger, 1967). Four

major substrate binding pockets S

 S

' S

' on the enzyme have been

identified (Hangauer et al., 1984). The hydrophobic substrate binding pocket S

'of

thermolysin, mainly formed by Phe130, Leu133, Val139 and Leu202, is considered

to be the main determinant of substrate specificity and preferably binds hydrophobic

residues such as Leu (e.g. Hangauer et al., 1984; Matthews, 1988). Mutagenesis

studies on the TLP of B. stearothermophilus have shown that the preference of

this protease for Phe at P

' can be altered toward that of thermolysin by changing

Phe133 to Leu, the latter residue being present in thermolysin at that position (de

Kreij et al., 2000, 2001). Enlargement of the binding pocket by replacement of

Leu202 by smaller amino acids (Val, Ala, Gly) resulted in higher efficiency toward

METALLOPROTEASES 229

substrates with Phe at P

'. Unexpectedly, reduction of its size by substitution of

Leu202 by Phe or Tyr also caused a large increase in activity toward substrates

with Phe at P

The role of other substrate binding pockets has recently been highlighted by the

adaptation of vimelysin (Vibrio sp.) substrate specificity in the P

' position to that

of thermolysin upon exchange of Arg215 in the S

' binding pocket for Asp215

present in thermolysin (Oda et al., 2005). Vimelysin has 35% sequence identity

with thermolysin and is characterised by high stability in organic solvents. It shows

a strong preference for Phe over Leu at position P

' and also a preference for neutral

or acidic amino acids in the P

' position in contrast to thermolysin in which basic

amino acids are preferred at position C.

Enhancements of the activity of thermolysin have been achieved by mutation

of Tyr110 and Phe114 in the S

subsite, as well as Tyr211 (Kubo et al., 1992),

Gln119 (Kidokoro et al., 1995) and Leu155 (Matsumiya et al., 2004). Some of

the effects were additive (Kidokoro, 1998; de Kreij et al., 2002). However, the

effects of mutations on activity were in most cases smaller for large proteinaceous

substrates compared to those observed for short peptides probably as a consequence

of the different possible productive binding modes for the former. The contribution

of binding to catalysis is expected to be much less in the case of short peptides.

Thermolysin activity is considerably enhanced by the addition of neutral salts

(Inouye, 1992; Inouye et al., 1998a; Bedell et al., 1998). A further increase in

activity is observed when the catalytic zinc ion is substituted by cobalt ions (Kuzuya

and Inouye, 2001). Both effects are independent of each other. Activation by NaCl

has been shown to be caused by an increase in k

cat

values (Inouye et al., 1996).

The addition of neutral salts also has beneficial effects on thermolysin solubility

and thermal stability (Inouye et al., 1998a and b). Preliminary studies of crystals

soaked in 4 M NaCl did not show significant changes in the space group (Kamo

et al., 2005). Positive effects on thermolysin activity have also been described for

sugars (e.g. sucrose, trehalose) and other polyols (Mejri et al., 1998), and pressure

up to 2.5 kbar (Kudryashova et al., 1998).

Thermolysin is inhibited by zinc-chelating agents (e.g. 1,10-phenanthroline,

EDTA) (Holmquist and Vallee, 1974). Other more specific inhibitors for

thermolysin have been developed over the years, and their number is still increasing

(reviewed in van den Burg and Eijsink, 2004).

3.2. Stability

Apart from high activity, the stability of an enzyme at higher temperatures and

toward other denaturing influences is one of the most important criteria for the

application of enzymes in biocatalysis. To meet the demands of industry, high

operational stability of enzymes is required as this significantly contributes to cost

reduction.

In order to obtain highly stable enzymes several strategies are used. One is

focused on the isolation of new enzymes from extremophilic organisms (reviewed

230 MANSFELD

by Vieille and Zeikus, 2001). Other strategies are based on: the stabilisation of

available enzymes by rational design of enzyme variants (Eijsink et al., 2004), high-

throughput screening of randomly generated mutant libraries (Eijsink et al., 2005)

in combination with recently described ‘semi-rational’ approaches for a guided

design of these libraries (Patrick and Firth, 2005), chemical modification including

immobilisation (Ulbrich-Hofmann et al., 1999) and de novo design of catalysts

(Kaplan and DeGrado, 2004).

Since thermolysin is produced by a highly thermophilic Bacillus strain it

is relatively thermostable compared to other metalloproteases produced by less

thermophilic strains. As a result of a series of mutational studies a surface-

exposed region between amino acid residues 56 – 69 in the N-terminal part

of the thermolysin-like protease from B. stearothermophilus (TLP-ste) (Takagi

et al., 1985; 85% sequence identity to thermolysin) has been identified that is

extremely sensitive to mutation, whereas the C-terminal part of the protease is only

slightly affected by even dramatic amino acid changes (Vriend and Eijsink, 1993;

Eijsink et al., 1995). Later this region was recognised as being the most labile region

of the protein where local unfolding processes start, resulting in rapid autoprote-

olytic degradation of the protein (Eijsink et al., 1995; Vriend et al., 1998; Fig. 4).

In light of the concept of protein stabilisation developed by Schellenberger and

Ulbrich (1989), this region was consequently called the unfolding region (Mansfeld

et al., 1999). Coherently, stabilisation of this region enabled the construction of

variants displaying considerably enhanced thermostability. The amino acid residues

Figure 4. Homology model of the 3D structure of TLP from B. stearothermophilus (Vriend and

Eijsink, 1993). The large sphere represents the Zn

ion in the active site. The four smaller spheres

represent the Ca

ions bound to the molecule

METALLOPROTEASES 231

selected for replacement were chosen on the one hand on the basis of rational

design strategies, and on the other because they corresponded to residues naturally

occurring in the more thermostable enzyme thermolysin (reviewed in Eijsink

et al., 1995). The introduction of a disulfide bridge between residues 8 and 60

(Mansfeld et al., 1997) was found to strongly stabilise this region against local

unfolding. Detailed studies of this mutant have shown that this effect is mainly

attributable to stabilisation against autoproteolysis rather than global unfolding

(Dürrschmidt et al., 2001). The disulfide bridge was shown to be able to mimic

the stabilising effect of calcium ions in local unfolding processes (Dürrschmidt

et al., 2005). Calcium ions are major determinants of protease stability (Dahlquist

et al., 1976).

These studies culminated in the successful conversion of the moderately stable

TLP from B. stearothermophilus into an extremely stable enzyme (named boilysin)

via a limited number of mutations (van den Burg et al., 1998). The half-life of

the mutant enzyme was 170 min at 100



C in contrast to 1 min for thermolysin.

Compared to naturally occurring enzymes from thermophilic organisms, boilysin

is characterised by wild-type like activity under the usually employed operating

temperatures making this enzyme interesting for industrial application. As temper-

ature is increased, activity is further enhanced. This enzyme has been tested under

extreme conditions for the hydrolysis of substrates that are difficult to digest (van

den Burg et al., 1999; de Kreij et al., 2000), the hydrolysis of prion proteins and

protein removal in nucleic acid purification (van den Burg, personal communi-

cation), and for the synthesis of an aspartame precursor (Kühn et al., 2002).

Taking advantage of knowledge on the enzyme’s unfolding process and the

concept of stabilisation by strengthening the most labile region in a protein (Schel-

lenberger and Ulbrich, 1989; Ulbrich-Hofmann et al., 1999), strong stabilisation

of TLP-ste was achieved by immobilisation in a site-specific manner (Mansfeld

et al., 1999). This was most effective when the protein was fixed to the carrier via

cysteine residues in the unfolding region. Very strong stabilisation has also been

obtained by immobilisation via multiple bonds to a carrier having a high density

of functional groups (Mansfeld and Ulbrich-Hofmann, 2000), suggesting that more

than one labile region might be present in the molecule as has been argued by Vriend

et al. (1998). Rigidification of the enzyme due to the formation of multiple bonds

with the carrier material was found to occur at the expense of activity (Mansfeld

and Ulbrich-Hofmann, 2000).

Another strategy to protect TLPs against autoproteolytic degradation and to

stabilise against thermal inactivation is the identification of primary cleavage sites.

The removal of these autodegradation sites has been described for TLP from

B. subtilis (van den Burg et al., 1990) and thermolysin (Matsumiya et al., 2004).

4. APPLICATIONS

Thermolysin is commercially available at industrial scale as Thermoase PC10F from

Amano Enzyme Inc., formerly Daiwa Kasei Co. Ltd., Japan. Other bacterial metal-

loproteases produced commercially are the TLPs from B. subtilis (Neutrase

from

232 MANSFELD

Novo Nordisk, Denmark and Protin PC10F from Amano Enzyme Inc., Japan). The

highly stable TLP-ste variant Boilysin can be requested from IMEnz (Groningen,

The Netherlands).

At laboratory scale, thermolysin can be purchased from different suppliers

(e.g. Sigma, MERCK Biosciences). Thermolysin and TLPs from Bacillus species

are traditionally produced in protease-deficient B. subtilis strains such as DB104

(Kawamura and Doi, 1984) and DB117 (Eijsink et al., 1990). These enzymes

are synthesized as inactive preproenzymes (Takagi et al., 1985; Kubo and

Imanaka, 1988; van den Burg et al., 1991) and processed to the active mature

enzymes via autocatalytic removal of the large propeptides (about 200 amino acids)

(Wetmore et al., 1992; Marie-Claire et al., 1998). The thermolysin propeptide has

been shown to act as a mixed, non-competitive inhibitor of the protease and to facil-

itate the recovery of active enzyme from denatured thermolysin in a stoichiometric

manner (O’Donohue and Beaumont, 1996; Marie-Claire et al., 1999). Recently,

several strategies have been described for the expression of these enzymes in E. coli

with the prosequence in cis or trans (Marie-Claire et al., 1999; Inouye et al., 2005)

and even in the absence of the prosequence (Mansfeld et al., 2005).

The enzymes secreted into the culture broths of B. subtilis and E. coli or renatured

from inclusion bodies formed in E. coli (Mansfeld et al., 2005) can be purified

by affinity chromatography on Bacitracin-silica (van den Burg et al., 1989) or

Gly-D-Phe columns (Walsh et al., 1974).

4.1. Synthesis of Peptides

Large scale synthesis of peptides has become increasingly important for the food

and pharmaceutical industries over the last few decades. The main application of

peptides is their use as low-calorie sweeteners. In addition, several biologically

active peptides have found interest as drugs in the treatment of diseases. Apart

from conventional peptide synthesis, new strategies for production have been tested,

one of which is enzymatic synthesis. The advantages of using enzymes are the

stereospecificity they confer on the reaction, the necessity for only minimal side

chain protection, the mild reaction conditions, and the avoidance of racemisation.

Thermolysin is one of the enzymes that has been studied for its potential in

enzymatic peptide synthesis. Since its first use for this purpose (Isowa et al., 1979) it

has been extensively studied in the laboratory and is used for large-scale production

of N-carbobenzoxy-L-aspartyl-L-phenylalanine methyl ester (Z-Asp-Phe-OMe), the

precursor of the widely used artificial sweetener aspartame (Ooshima et al., 1985;

Murakami et al., 1996). Thermolysin proved to be advantageous in the synthesis of

aspartame due to its low esterolytic activity that results in the preservation of the

methyl ester group which is essential for the sweet taste of the peptide.

As no acyl enzyme is formed in thermolysin catalysis, the reactions cannot be

run in kinetically controlled mode. Therefore, several strategies have been tested

to shift the equilibrium of the thermodynamically controlled enzymatic synthesis

of Z-Asp-Phe-OMe to the desired product. These are based on aqueous systems

METALLOPROTEASES 233

(Inouye, 1992; Murakami et al., 1996), systems with water-miscible organic solvents

(Lee et al., 1992; Kühn et al., 2002), biphasic systems (Hirata et al., 1997; Murakami

and Hirata, 1997; Murakami et al., 1998; Miyanaga et al., 2000b), solid-to-solid

synthesis (Erbeldinger et al., 1998a, b; Erbeldinger et al., 2001) and low-water

solvent systems (Nakanishi et al., 1985). For syntheses in aqueous/organic biphasic

systems (Murakami and Hirata, 1997; Hirata et al., 1997), in low-water solvent

systems with immobilized thermolysin (Nakanishi et al., 1985, 1990) or in

membrane systems (Iacobucci et al., 1994), continuous operation has been used

successfully.

Yields in pure aqueous systems are usually very low. The activity of thermolysin

and, accordingly, the reaction rates in aqueous systems have been found to be

enhanced by the addition of sodium and potassium salts (Inouye, 1992). However,

the pH increase due to salt addition may result in non-enzymatic hydrolysis of the

reactants, a problem which is avoided in reactions at low pH. High yields (95%)

have been achieved by insoluble salt formation between the product and excess Phe-

OMe or unreacted enantiomer D-Phe-OMe followed by subsequent removal of the

precipitate from the aqueous solution (Ager et al., 1998). This method is used in the

commercial aspartame precursor production process of TOSOH (Japan) performed

at Holland Sweetener (The Netherlands) (Fig. 5). Addition of a water-immiscible

solvent like toluene or 4-methylpentan-2-one after the start of formation of the

precipitate was found to permit the process to be run continuously. In the presence

of water-miscible organic solvents (e.g. dimethylsulfoxide) reaction rates were also

enhanced by the addition of salts (Kühn et al., 2002), though yields decreased with

increasing salt concentrations. Yields could be markedly improved by the addition of

alcohols (methanol, 2-propanol) to aqueous systems even though reaction rates were

reduced due to inhibitory effects on thermolysin (Kühn et al., 2002). In biphasic

organic solvent systems, ethyl acetate, tert.-amyl alcohol (Miyanaga et al., 2000b),

n-butyl acetate (Murakami and Hirata, 1997), tributylphosphate and 1-butanol (in

the synthesis of N-formyl-Asp-Phe-OMe – Murakami et al., 2000a) and ionic liquids

(e.g. 1-butyl-3-methylimidazolium hexafluorophosphate – Erbeldinger et al., 2000)

have all been used as solvents. In the solid-to-solid system the pH adjusted by basic

inorganic salt addition played an important role (Erbeldinger et al., 2001). In low-

water solvent systems the water is usually provided by the carrier materials that are

used for adsorptive binding of the enzyme. Polyacrylic ester resins such as XAD-7

(ICN Biomedicals Inc., USA) in ethyl acetate and tert.-amyl alcohol (Miyanaga

et al., 2000a, b), Celite R-640 (FLUKA) in combination with toluene as solvent

(de Martin et al., 2001) or molecularly imprinted polymers (methacrylate-ethylene

glycol dimethacrylate-copolymers) in ethyl acetate (Ye et al., 1999) have all been

used as carrier materials. A considerable increase in thermolysin activity in non-

aqueous media has been achieved by lyophilisation in the presence of KCl or other

inorganic salts (Bedell et al., 1998). Activity could be further improved by the use

of molecular imprinting in combination with activation by salts (Rich et al., 2002).

Cross-linked enzyme crystals (CLECs) of thermolysin which have been used

234 MANSFELD

Figure 5. Principle of commercial aspartame synthesis of TOSOH at Holland Sweetener (The

Netherlands)

successfully for the synthesis of the aspartame precursor (Persichetti et al., 1995)

represent an interesting tool for organic chemists due to their high specific activity

and increased resistance to inactivation by organic solvents, elevated temperatures

and proteolysis.

In all these systems the reaction velocities and yields obtained result from the

interplay between the reaction medium (i.e. buffer, organic solvent, salt concen-

tration), the types and ratios of reactants, and the type of product, as well as

the activity and stability of thermolysin in the corresponding reaction systems.

A compromise always has to be found between initial reaction rates and final yields.

To broaden the scope for enzymatic peptide synthesis new enzymes have been

searched for. A metalloprotease called vimelysin from Vibrio sp. proved to be

superior to thermolysin for the synthesis of aspartame at lower temperatures

and higher solvent concentrations. Like vibriolysin from the Antarctic bacterium

METALLOPROTEASES 235

strain 643 (Adekoya et al., 2006), it might be an interesting alternative for

thermolabile substrates (Kunugi et al., 1997). Due to its broader substrate speci-

ficity, pseudolysin from Pseudomonas aeruginosa might also be interesting for

synthetic purposes (Rival et al., 2000). At laboratory scale, free or immobilized

thermolysin has also been used for the synthesis of other peptides: Z-Gln-Leu-

, Z-Phe-Leu-NH

, various dipeptide fragments of cholecystokinin, and peptides

containing non-proteogenic amino acids (Wayne and Fruton, 1983; Erbeldinger

et al., 1998a, b, 1999; Calvet et al., 1996; Krix et al., 1997). Neutrase

, a TLP

from B. subtilis, with a slightly different substrate specificity to that of thermolysin,

was also applied in peptide synthesis (either as free or immobilized enzyme (on

Celite-545 (Fluka, Germany) or Polyamide-PA6 (Akzo)) (Clapes et al., 1995, 1997).

A scale-up of the suspension-to-suspension approach using mainly undissolved

substrates was performed by Eichhorn et al. (1997). The possibility of using the

extractive reaction in aqueous/organic biphasic systems for the continuous synthesis

of Z-Gly-Phe-OMe at larger scale and in high yield was reported by Murakami

et al. (2000b). Unexpectedly, thermolysin was also shown to catalyse the acylation

of paclitaxel with divinyl adipate (Khmelnitsky et al., 1997) and an acylation of

sucrose with vinyl laurate (Pedersen et al., 2002).

4.2. Production of Protein Hydrolysates

Another very important industrial application of metalloproteases (mostly in combi-

nation with other proteases) is the production of hydrolysed food proteins and

flavour-enhancing peptides to replace the chemical methods of synthesis. Soy and

wheat hydrolysates are used in flavour-enhancement of soups and sauces, and milk

protein hydrolysates are preferred for the refinement of cheese products. Meat

hydrolysates find application in the enhancement of the flavour of meat products,

soups, sauces and other instant products. An advantage of enzymatic processes

is the minor formation of unwanted by-products of negative impact on health;

the disadvantageous formation of bitter tasting peptides in enzymatically produced

protein hydrolysates can be overcome by simultaneous treatment with exopeptidases

(reviews in Saha and Hayashi, 2001; Raksakulthai and Haard, 2003). Bitter peptides

are characterised by a high content of hydrophobic amino acids. Hydrolysates of

meat and fish proteins or gelatin develop less bitter taste than hydrolysates of maize

protein, casein or haemoglobin. Flavour development by a cocktail of proteases,

including Neutrase

has been used to accelerate the ripening of dry fermented

sausages (e.g. Fernandez et al., 2000). The development of high-value functional

foods and nutraceuticals has made a major impact on dairy protein hydrolysate

production because of the latter’s probiotic, antimicrobial and digestive effects.

Another beneficial effect of protein hydrolysates on health might be their antiox-

idant effects (Hernandez-Ledesma et al., 2005) and their inhibition of angiotensin-

converting enzyme (Vercruysse et al., 2005). Low-molecular weight hydrolysis

products of protamine have been tested successfully as a delivery system for DNA

in gene therapy (Park et al., 2003).

236 MANSFELD

4.3. Other Applications

Further commercial applications of metalloproteases can be found in the brewing

industry (improved filtration of beer, reduced calorie content), in the leather industry

(bating and dehairing), the processing of slaughter waste, improvement of the

baking characteristics of flour, and in the film industry (recovery of waste silver).

An interesting application of immobilized thermolysin is the removal of protein

coatings from the surface of old documents and art work (Moeschel et al., 2003).

In protein science, thermolysin is an important tool for limited proteolysis to

determine primary structures and gain first insights into the conformation of proteins

whose crystal structures are not yet known (reviewed in Fontana et al., 2004).

It is also used to analyse confined local fluctuations and global unfolding events

in proteins and to determine their stabilities (Arnold et al., 2005; Park and

Marqusee, 2005), or to isolate protein fragments that can fold autonomously and

therefore be considered as domains which might be useful in crystallisation and

high-throughput applications (Gao et al., 2005).

Inhibitors of metalloproteases involved in diseases are of potential therapeutic

use. In this respect, thermolysin has been used as a template for the creation

of homology models of the active sites of medically relevant mammalian metal-

loproteases. Details of these important classes of metalloproteases can be found

in Barrett et al. (2004). Interesting targets are: neprilysin (Roques et al., 1993);

angiotensin-converting enzyme, being responsible for degradation of biologically

active peptides such as enkephalins; endothelin-converting enzymes, which liberate

endothelin (a potent vasoconstrictor) from its precursor; highly potent neuro-

toxins like bontoxilysin and tentoxilysin from Clostridium species which block the

release of acetylcholine at neuromuscular junctions and cause motor paralysis in

tetanus and botulism; anthrax lethal factor, which acts by disrupting intracellular

signalling by cleaving mitogen-activated protein kinase kinases and causes multiple

haemorrhagic lesions; matrix metalloproteases or matrixins (like collagenase,

elastase, stromelysin, matrilysin, gelatinase) which are involved in the degradation

of extracellular matrix proteins including collagen and are required for tissue

repair and remodelling but are also involved in pathological processes (arthritis,

atherosclerosis, tumour growth and metastasis); ADAM17 (tumour necrosis factor

-converting enzyme) and ADAM10 (myelin-associated metalloendopeptidase),

pappalysins 1 and 2 which cleave insulin-like growth factor 1 binding protein-4

and liberate the growth factor.

5. ACKNOWLEDGEMENT

I would like to thank Prof. Dr. R. Ulbrich-Hofmann for critical reading of the

manuscript, Dr. S. Gebauer’s group of Molecular Modelling at our department for

help in preparing the Figures, Prof. G. Vriend, Dr. B. van den Burg, Prof. V.

Eijsink, and Prof. G. Venema for getting me started with the metalloprotease work

and fruitful cooperation.