Inhibitory proteaz bakteryjnych

Bacterial Protease Inhibitors

Claudiu T. Supuran,

Andrea Scozzafava,

Brian W. Clare

University of Florence, Dipartimento di Chimica, Laboratorio di Chimica Inorganica e Bioinorganica,

Firenze, Italy

Department of Chemistry, The University of Western Australia,

Nedlands W.A., Australia

Abstract: Serine-, cysteine-, and metalloproteases are widely spread in many pathogenic bacteria,
where they play critical functions related to colonization and evasion of host immune defenses,
acquisition of nutrients for growth and proliferation, facilitation of dissemination, or tissue
damage during infection. Since all the antibiotics used clinically at the moment share a common
mechanism of action, acting as inhibitors of the bacterial cell wall biosynthesis or affecting
protein synthesis on ribosomes, resistance to these pharmacological agents represents a serious
medical problem, which might be resolved by using new generation of antibiotics, possessing a
different mechanism of action. Bacterial protease inhibitors constitute an interesting such
possibility, due to the fact that many specific as well as ubiquitous proteases have recently been
characterized in some detail in both gram-positive as well as gram-negative pathogens. Few
potent, specific inhibitors for such bacterial proteases have been reported at this moment except
for some signal peptidase, clostripain, Clostridium histolyticum collagenase, botulinum neuro-
toxin, and tetanus neurotoxin inhibitors. No inhibitors of the critically important and ubiquitous
AAA proteases, degP or sortase have been reported, although such compounds would presumably
constitute a new class of highly effective antibiotics. This review presents the state of the art in the
design of such enzyme inhibitors with potential therapeutic applications, as well as recent
advances in the use of some of these proteases in therapy.

2002 Wiley Periodicals Inc. Med Res Rev, 22,

No. 4, 329–372, 2002; Published online in Wiley InterScience (www.interscience.wiley.com).

DOI 10.1002/med.10007

Key words: AAA protease; antibiotics; anthrax lethal factor; botulinum neurotoxins; Clostridium
collagenase; cysteine protease; degP; metalloprotease; serine protease; sortase; tetanus neurotoxin

1 . I N T R O D U C T I O N

Proteases (PRs)—also denominated proteinases or peptidases—constitute one of the largest
functional groups of proteins, with more than 560 members actually described.

By hydrolyzing one

329

Correspondence to: ClaudiuT. Supuran,University of Florence,Dipartimento di Chimica,Laboratorio di Chimica Inorganica e Bioi-
norganica,Via della Lastruccia 3, Room188, Polo Scientifico, 50019 Sesto Fiorentino, Firenze, Italy.

E-mail: claudiu.supuran@unifi.it

Contract grant sponsor : Italian CNR^Target Project Biotechnology

Medicinal Research Reviews, Vol. 22, No. 4, 329 ^372, 2002
ß

2002 Wiley Periodicals Inc.

of the most important chemical bonds present in biomolecules, i.e., the peptide bond, PRs play
crucial functions in organisms all over the phylogenetic tree, starting from viruses, bacteria,
protozoa, metazoa, or fungi, and ending with plants and animals. Numerous practical applications in
biotechnology of such enzymes, and the understanding that PRs are important targets for the drug
design, ultimately fuelled much research in this field.

Whether much progress has been registered

in the design and clinical applications of viral PR inhibitors (mainly those targeted against the
aspartic protease of HIV),

2–4

not the same situation is true for the bacterial proteases.

5,6

PRs are

widespread in all types of bacteria, where they are involved in critical processes such as colonization
and evasion of host immune defenses, acquisition of nutrients for growth and proliferation,
facilitation of dissemination, or tissue damage during infection.

5,6

Even more subtle roles for

bacterial PRs have recently been evidenced in the interaction between host and the invading micro-
organisms: interruption of cascade activation pathways, disruption of cytokine network, excision of
cell surface receptors, and inactivation of host protease inhibitors (PIs).

5–8

Nevertheless, it is

surprising that there is little or no regulation of bacterial PRs by plasma-derived PIs, such as for
instance the serpins (serine protease inhibitors) which are present in relatively high concentrations
in plasma.

Even worse for the host is that bacteria developed strategies for neutralizing such

plasma-derived PIs, assuring in this way an efficient attack of the invaded organism.

5–8

Taking into

account all these facts, it is obvious that bacterial PRs may represent very attractive targets for the
development of novel types of antibiotics, since inhibition of such critical enzymes would pre-
sumably lead to the death of the invading pathogen.

Till date all the antibiotics used in clinical

practice share a common mechanism of action, acting as inhibitors of the bacterial cell wall bio-
synthesis or affecting protein synthesis on ribosomes and not intervening in more fundamental
metabolic processes of the pathogen. Considering the specific role that bacterial PRs play in such
critical steps for the successful invasion of the host

5,7

and the constant emergence of antibiotic

resistance,

it is crucial to develop bacterial PIs as a novel antibiotic class. Mention should be made,

that no drugs belonging to this class of pharmacological agents are available at present for clinical
use, and this review is also meant as to attract attention to this relatively unexplored and potentially
important field, in which some progress has nonetheless been registered ultimately. Thus, many
possible targets for the drug design will be discussed, together with the recent progress achieved in
understanding the PR types present in pathogenic bacteria, as well as the inhibitors for such
enzymes. In this review, the discussion will be restricted to PRs present in Eubacteriae, since the
Archeobacteriae are generally non-pathogenic. Also, PRs of viral, fungal, protozoan, or other para-
sitic origin will be not considered in this review.

Five catalytic types of PRs have been recognized so far, in which serine, threonine, cysteine, or

aspartic groups as well as metal ions play a primary role in catalysis. All these types of enzymes are
present in bacteria.

The first three types of PRs are catalytically very different from the aspartic and

metallo-PRs, mainly because the nucleophile of the catalytic site is part of an amino acid in the first
case, whereas it is an activated water molecule for the second group of such enzymes. Thus, acyl
enzyme intermediates are formed only in the reactions of the Ser/Thr/Cys PRs, and only these
peptidases can readily act as transferases.

The classification of PRs used in this review is based on that of Rawlings and Barrett,

1,10

which the catalytic type of the protein represents the top level in the hierarchical classification.
According to this rule, the PRs can be divided into clans based on three-dimensional protein folding
and into families based on evolutionary relationships of the primary sequence. It must be noted that
in the fore-cited classification, several enzymes in which threonine rather than serine residue forms
the nucleophile critical for catalysis, have been placed in the serine PR section. The terminology
used in describing the specificity of PRs depends on a model, in which the catalytic site is con-
sidered to be flanked on one or both sides by specificity subsites, each being able to accommodate
the side chain of a single amino acid residue, as originally proposed by Berger and Schechter,

and adopted thereafter by many researchers.

These sites are numbered from the catalytic site, S1,

330

SUPURAN, SCOZZAFAVA, AND CLARE

S2, . . . Sn towards the N-terminus of the substrate, and S1

, S2

, . . . Sn

towards the C-terminus. The

residues they accommodate are numbered P1, P2, . . . Pn, and P1

, P2

, . . . Pn

, respectively, as

follows (the catalytic site of the enzyme is marked ‘‘*’’, and the scissile peptide bond of the substrate
as ‘‘#’’). The same formalism is also valid for enzyme inhibitors that bind to the catalytic site:

Substrate/Inhibitor:

-P3-P2-P1#P1

-P2

-P3

Enzyme:

-S3-S2-S1*S1

-S2

-S3

For each enzyme with potential therapeutic use per se or for the development of inhibitors

useful as antibiotics, the main characteristics will be mentioned, such as the EC classification (when
available), the catalytic and family type (in the classification of Barrett et al.),

the organisms in

which it is present, the preferred scissile bond(s) and eventually the type of in vivo attacked
substrate, relevant for the pathogenic effects of the considered bacterial species. Furthermore, the
latest developments in the design of inhibitors for such PRs will be discussed, as well as possibilities
to consider such inhibitors as leads for the drug design of novel classes of antibiotics.

It must be mentioned that very few aspartic PRs have been isolated up to now in bacteria, and

those described so far seem to be not involved in pathologic processes connected with bacterial
infections.

This is the reason why no chapter dedicated to aspartic PRs is included in this review.

2 . S E R I N E P R O T E A S E S

PRs in which the catalytic mechanism depends upon the hydroxyl group of a serine residue, which
acts as the nucleophile attacking peptide bonds belong to the serine PR (SPR) class of enzymes.

The catalytic mechanism of serine PRs usually involves, in addition to the serine residue that carries
the nucleophilic attack, a proton donor/general base. In many SPRs, the proton donor is a histidine
residue, and there is a catalytic triad, because a third residue is also required, probably for orien-
tation of the imidazolium ring of the histidine, and this residue may be an aspartate, or another
histidine

1,12

(Fig. 1A). For other SPRs, catalysis is achieved by a catalytic dyad in which a lysine

residue plays the role of proton donor, and a third catalytic residue is not required (Fig. 1B). There
are other SPRs that have a Ser/His catalytic dyad. Finally, for several serine endopeptidases the
N-terminal residue itself is the attacking nucleophile, and they include the threonine-type PRs.

1,12

There are about 40 families of serine- and threonine-PRs, among which enzymes extremely

important for the homeostasis and vital functions of higher vertebrates, such as the digestive

Figure 1.

Schematic representation of the serine protease amino acid residues involved in the proteolytic scission.

A: Catalytic

triad (chymotrypsin numbering).

B: Catalytic dyad (a lysine residue activates the hydroxyl group of the serine residue essential for

catalysis).

BACTERIAL PROTEASE INHIBITORS

331

enzymes trypsin, chymotrypsin, and elastase,

the enzymes involved in the blood coagulation

cascade (thrombin, coagulation factors VIIa, IXa, X),

13–16

the kallikreins, involved in blood pres-

sure regulation or smooth muscle relaxation,

or tryptase, an enzyme predominantly present in

mast cells, and involved in various normal and pathologic processes,

etc. Bacteria seem to possess

all the major types of serine PRs mentioned above (Table 1), and those with potential therapeutic
applications will be discussed in this article. Bacterial SPRs, which are not involved in pathogenesis
(such as subtilisins or proteases isolated from different lactobacilli) will be not included in this
discussion, although such enzymes are important as additives for detergents (subtilisins) or for the
cheese production (lactobacilli proteases).

A. Proteases of the S1 and S2 Families

Trypsin-like proteases are widely distributed in streptomycetes (S. griseus, S. erythraeus,
S. exfoliatus, etc.) (Table I), in which they seem to play important roles related to morphological
differentiation.

19,20

Streptomyces trypsins play a major role in the metabolism of mycelial substrate

or nongrowing mycelial proteins for supporting the growth of aerial mycelium on surface cultures,
but the detailed physiological role of these enzymes is not well defined to date.

These enzymes

are inhibited either by typical trypsin inhibitors such as DFP (diisopropylfluorophosphate), Ts-Lys-
CH

Cl, benzylsulfonyl fluoride, or proteinaceous inhibitors such as leupeptin, soybean trypsin

inhibitor, and aprotinin among others.

The X-ray crystal structures of the enzymes isolated from

S. griseus. and S. erythraeus have been reported, being relatively similar to that of the bovine
trypsin, at least for the active site residues involved in the catalysis, but this did not lead to the
development of potent active site directed inhibitors.

19,20

Several other secreted or membrane bound bacterial endopeptidases, namely streptogrisins

(Table I), were isolated from Streptomyces spp. They are chymotrypsin-like PRs with broad
substrate specificity and readily cleaving peptide bonds on the carboxyl side of Phe, Tyr, Leu, and
Met.

21,22

In contrast to a-chymotrypsin, streptogrisin B is not inhibited by Ts-Phe-CH

Cl, whereas

the tripeptidyl chloromethane derivative Ac-Gly-Leu-Phe-CH

Cl acts as a potent inhibitor.

The

enzyme is also inhibited by 4-iodobenzenesulfonyl fluoride, 4-methoxy-3-chloromercuri-benzene-
sulfonyl fluoride, 4-chloromercuribenzenesulfonyl fluoride, Ac-Gly-Leu-Phe-CH

Cl, turkey ovo-

mucoid inhibitor OMTKY3, potato inhibitor I, Streptomyces subtilisin inhibitor and eglin c.

22,23

The

gene of streptogrisin B codes for a prepro-enzyme, the pre- and pro-peptides being 38 and 76 amino
acid residues long, respectively.

21–23

As in many other bacterial SPRs, the pre- and pro-regions are

responsible for the enzyme secretion across the membrane and for its correct folding. After
synthesis, the pre-region is cleaved off by a signal peptidase. The peptide bond connecting the
propeptide to the mature enzyme is cleaved by a self-processing event, when a Leu#Ile bond is
cleaved. A large number of mutant streptogrisins possessing different substrate specificities and
thermostabilities have been reported.

24,25

Some of them act as peptide-coupling enzymes, and show

interesting biotechnological applications.

The naturally occurring proteinaceous protease in-

hibitor chymostatin forms a hemiacetal adduct with the catalytic residue Ser 195 of streptogrisin
A.

The X-ray structure of this adduct has been reported by Delbaere and Brayer,

and was sub-

sequently used for the design of low molecular weight aldehyde inhibitors of the type Z-Arg-Leu-
Phe-H (Z

¼ benzyloxycarbonyl)

or Ac-Pro-Ala-Pro-Phe-H (Ac

¼ acetyl).

None of these inhibi-

tors have application as antibiotics at the moment.

Glutamate-specific SPRs have also been isolated from many bacteria (Table I).

The bacterial

glutamyl endopeptidases may be divided into three groups according to the source organisms and
sequence relationships: a staphylococcal group, a Bacillus group and a Streptomyces group.

29,30

Some forms of glutamyl endopeptidase I, especially protease V8, have found extensive use in
the specific fragmentation of proteins prior to amino acid sequencing.

Boc-Leu-Glu-CH

(Boc

¼ tert-butoxycarbonyl) is a potent inhibitor of glutamyl endopeptidase I, but no protein

332

SUPURAN, SCOZZAFAVA, AND CLARE

Table I. Bacterial Serine and Cysteine Proteases, With Their Preferred Scissile Bond and Organisms From
Which Were Isolated

Preferred scissile bond

Protease

Family

Organism(s)

(P1#P1

)

Streptomyces trypsins

Streptomyces spp.

Arg#Xaa; Lys#Xaa

3.4.21.81

Streptogrisin B

Streptomyces spp.

Phe#Xaa;Tyr#Xaa

3.4.21.80

Streptogrisin A

Streptomyces spp.

Phe#Xaa;Tyr#Xaa

3.4.21.19

Glutamyl endopeptidase I

Bacillus subtilis;

Glu#Phe; Glu#Val

Staphylococcus aureus

3.4.21.82

Glutamyl endopeptidase II

Streptomyces spp.

Glu#Arg; Asp#Arg

Exfoliative toxin A

Staphylococcus aureus

Glu#Xaa

3.4.21.12

a-Lytic protease

Achromobacter lyticus

Ala#Ala

Lysobacter enzymogenes

DegP (protease Do)

Bacillus subtilis; Brucella

Val#Xaa; Ile#Xaa

abortus

Campylobacter jejuni;
Chlamydia trachomatis;

Escherichia coli

Mycobacterium lepre;

M. paratuberculosis

Rickettsia typhi;

R. tsutsugamushi

Salmonella typhimurium;

Yersinia enterocolitica

3.4.21.50

Lysyl endopeptidase

Achromobacter lyticus

Lys#Xaa

3.4.21.72

IgA1-specific serine

Haemophilus influenzae

Pro#Ser; Pro#Thr

endopeptidase

Neisseria gonorrhoeae;

N. meningitidis

C5a peptidase

Streptococcus agalactiae;

His#Lys

S. pyogenes

Dichelobacter (sheep foot-rot)

Dichelobacter nodosus

Non specific PR

basic serine proteinase

Trepolisin

Treponema denticola

Phe#Xaa

Tripeptidyl-peptidases

S8, S33

Streptomyces spp.

AlaProAla#Xaa

Prolyl tripeptidyl-peptidase

Porphyromonas gingivalis

XaaYaaPro#Xaa

3.4.11.5

Prolyl aminopeptidase

S33

Flavobacterium meningosepticum

N-Pro#Xaa

Mycoplasma genitalium;

Escherichia coli

N. gonorrhoeae,

Streptomyces K15

S11

Streptomyces K15

acyl-

-Ala-

-Ala#Xaa

-Ala-

-Ala transpeptidase

3.4.16.4

Streptomyces R61

S12

Streptomyces spp.

acyl-

-Ala#

-Ala

-Ala-

-Ala carboxypeptidase

3.4.21.89

Signal peptidase I

S26

E. coli; H. influenzae;

Ala#Lys

M. tuberculosis

Pseudomonas fluorescens;

S. typhimurium
Streptococcus pneumoniae

3.4.21.87

Omptin

S18

E. coli; Yersinia pestis

Arg#Arg; Arg#Lys

Salmonella typhimurium

3.4.21.92

Endopeptidase Clp

S14

E. coli; B. Substilis, Listeria spp

Unspecific PR

3.4.22.10

Streptopain

C10

Porphyromonas gingivalis

Phe#Tyr

Streptococcus pyogenes

(continued)

BACTERIAL PROTEASE INHIBITORS

333

inhibitors of this enzyme have been detected to date.

Recently, Hamilton et al.

reported potent

synthetic irreversible inhibitors of V8 of Staphylococcus aureus, of the diphenylphosphonate type.
Thus, compounds such as Ac-Asp-P(OPh)

and Ac-Glu-P(OPh)

inhibited V8, but not granzyme B,

a mammalian chymotrypsin-like PR isolated from cytotoxic T lymphocytes, which also hydrolyzes
Asp#Xaa bonds in proteins. Thus, phosphonates of the type mentioned above might constitute
interesting leads for obtaining antibiotics devoid of undesired side effects.

Exfoliative toxin A (ETA or exotoxin A, Table I), a distinctive SPR produced only by S. aureus,

belonging to phage group II, was first purified and identified as the causative agent in staphylococcal
scalded skin syndrome (SSSS) in 1972 by Melish et al.

SSSS is a blistering disorder that involves

sloughing of the skin encompassing one-half or more of the total epidermis. ETA is the causative
agent of this disease

32,33

primarily affecting neonates. The toxin acts specifically at the zona granu-

losa of the epidermis, to produce characteristic exfoliation, by a mechanism poorly understood at the
moment. Despite the availability of classical antibiotics, SSSS carries a significant mortality rate.

Recently, the X-ray structure of ETA has been reported,

being proved that it is a trypsin-like SPR,

but with significant differences at the amino- and carboxy-termini, as well as in the loop regions.
The active site catalytic residues are similar to those of other glutamate-specific SPRs (such as the
glutamyl endopeptidases mentioned above), suggesting a common catalytic mechanism, except that
the oxyanion hole is closed in ETA.

No specific inhibitors for these SPR have been reported to

date, although they might be important clinically for the treatment of SSSS.

a-Lytic protease (a-LP, Table I) is an extracellular SPR secreted by Lysobacter enzymogenes,

a gram-negative Canadian soil bacterium, resulting in the lysis of other common soil organisms,
including bacteria, fungi, and nematodes.

a-LP is an endopeptidase that has a strong preference

for cleaving after small, hydrophobic residues (such as for instance Ala), being strongly inhibited
by peptide aldehydes/ketones, sulfonyl fluorides, and boron/phosphorus based compounds such
as: Boc-Ala-Pro-boroVal; MeOSuc-Ala-Ala-Pro-Val-B(OH)

; MeOSuc-Ala-Ala-Pro-Phe-B(OH)

;

N-[(2S)-2-(phenoxy(1-R-(N-Boc-

-Ala-Pro)-1-amino-2-methyl-propyl)-phosphinyloxy)-propanoyl]-

-Ala-COOMe;

N-[(2S)-2-(phenoxy(1-R-(N-Boc-oxycarbonyl-

-Ala-Pro)-1-amino-2-methylpro-

pyl)-phosphinyl-oxy)-propanyl]-

-Ala-COOMe.

35,36

The X-ray crystallographic structure of this

enzyme has been reported, being shown that it is similar to that of the homologous PR
a-chymotrypsin.

a-LP was also shown to possess potent staphylolytic activity due to cleavage

of N-acetylmuramoyl-

-Ala amide bonds of the peptidoglycan wall of S. aureus, a widespread

pathogen.

Thus, this protease might be useful for the destruction of such highly pathogenic

bacteria.

The widely conserved heat shock protein DegP (also denominated HtrA, Table I) has both

general molecular chaperone and proteolytic activities.

39,40

This widely spread bacterial protease

plays a major role in the degradation of proteins exported beyond the cytoplasm; its main function
being most likely the removal of misfolded membrane and periplasmic proteins or not properly
processed proteins. DegP seems to recognize the non-native states of proteins, since thermally
aggregated endogenous proteins of E. coli that can be isolated from heat-shocked cells are

Table I. (Continued )

Preferred scissile bond

Protease

Family

Organism(s)

(P1#P1

)

3.4.22.8

Clostripain

C11

Clostridium histolyticum

Arg#Xaa; Lys#Xaa

3.4.19.3

Bacterial

C15

Pseudomonas fluorescens

Glp#Xaa*

pyroglutamyl-peptidase

Staphylococcus aureus, S. pyogenes

3.4.22.37

Gingipain R; gingipain K

C25

Porphyromonas gingivalis

Arg#Xaa; Lys#Xaa

Sortase

ubiquitous in gram-positive bacteria

LeuProXThr#Gly

NC, not classified; S, serine PR family; C, cysteine PR family; *Glp,

-pyroglutamic acid.

334

SUPURAN, SCOZZAFAVA, AND CLARE

efficiently degraded by this protease in vitro.

DegP is able to cleave aliphatic, b-branched residues,

which are generally buried in the hydrophobic core of most proteins. In model substrates, these
peptide bonds were inaccessible in the native three-dimensional structures.

DegP is a large oligo-

mer (dimer to dodecamer, of 48 kDa each), so that geometric restrictions on access to the active site
may well be involved in the substrate selectivity of this protease.

A major contribution of DegP

and its homologs (isolated in many pathogenic bacteria, see Table I) to pathogenesis of various
bacterial infections has been hypothesized, although no attempts to inhibit this enzyme have been
registered to date.

The gram-negative bacteria Achromobacter lyticus secretes three bacteriolytic proteases, which

were denominated Achromobacter proteases I, II, and III, one of which is a lysyl endopeptidase,
possibly playing an important digestive role.

43,44

No detailed information is available on the

biological functions of this highly active protease, which efficiently lyses Staphylococcus aureus or
Micrococcus luteus cells, primarily by splitting the linkage between the peptide subunit and the
interpeptide in the bacterial cell wall peptidoglycans.

B. Proteases of the S6, S8, S9, and S33 Families

Bacterial immunoglobulin A1 (IgA1) PRs (Table I) that degrade the heavy chain of human IgA1 in
the hinge region have been isolated from many pathogenic species, such as Neisseria spp.,
Haemophilus influenzae, Streptococcus pneumoniae, and Ureaplasma urealyticum among others.

By cleaving human IgA1 at the peptide bond Pro

227

# Thr

228

, these enzymes interfere with the

protective functions of the principal mediator of specific immunity on mucosal surfaces, and
especially in the upper respiratory tract.

Cleavage in the hinge region of IgA1 abolishes the Fc-

mediated secondary effector functions, while the Fab fragments retain antigen-binding capacity and
therefore may mask epitopes on the bacteria.

Thus, IgA1 proteases are critically important for the

ability of bacteria to colonize human mucosal surfaces in the presence of secretory IgA antibodies.
The three leading causes of bacterial meningitis, H. influenzae, N. meningitidis, and S. pneumoniae,
all produce functionally identical IgA1 proteases, suggesting that this activity is the major virulence
factor associated with this particularly invasive disease.

46,47

This SPR might be a very important

target for drug design, but no inhibitors of this enzyme have been reported for the moment.
Furthermore, very recently it has been proved that IgA1 proteases are potent inducers of pro-
inflammatory cytokines, such as tumor necrosis factor alpha (TNF-a), interleukin b (IL-b), IL-6,
and IL-8.

Thus, these proteases contribute substantially to the pathogenesis of bacterial infections

by at least two mechanisms of action.

Group A streptococci are gram-positive pathogens responsible for many diseases such as

pharyngitis, impetigo, rheumatic fever, streptococcal toxic shock syndrome, and acute glomerulo-
nephritis among others. A resurgence of such infections has constantly appeared over the past
years, mainly due to resistance of these pathogens to classical antibiotics.

Virulent strains of

Streptococcus pyogenes weakly activate the alternate complement pathway, an observation that led
to the discovery of S. pyogenes C5a peptidase (SCPA, Table I).

C5a is a product of proteolysis of

the C5 serum complement protein that results from the activation of the complement pathway.

48,49

SCPA is a highly specific endopeptidase that was initially thought, on the basis of carboxypeptidase
sequencing of the N-terminal peptide fragment of C5a, to hydrolyze human C5a at Lys68 in the
primary polymorphonuclear leukocytes (PMN)-binding site. The primary interest in the SCPA
revolves around its role in the virulence of such pathogenic streptococci.

49–51

The scpA gene of

S. pyogenes is linked to and coregulated with genes that encode other surface-associated virulence
factors, being showed that SCPA influences the trafficking of streptococci from skin sites of
infection to lymph nodes and the spleen.

47–49

No inhibitors of this particularly interesting enzyme

(from the drug design point of view) have been reported up to now, although some SCPA-based
vaccines are under investigation.

BACTERIAL PROTEASE INHIBITORS

335

Dichelobacter nodosus (formerly Bacteroides nodusus), an obligate anaerobic gram-negative

bacterium, is the causative pathogen of ovine foot-rot and secretes a family of extracellular pro-
teinases probably responsible for the pathogenesis of the disease.

52–55

Dichelobacter basic pro-

teinase (Table I) contributes to the virulence of the microorganism by promoting tissue invasion;
furthermore, virulence of isolates of D. nodosus correlates with differences in properties of the
extracellular proteinases and these differences have been used in diagnostic tests.

The basic

proteinase has been used as an immunogen in a vaccine formulation and shown to protect sheep
against foot-rot infections.

This unspecific proteinase is completely inhibited by standard SPR

inhibitors, such as DFP, PMSF, and peptidyl chloromethanes with Phe at P1 and Ala or Gly at P2,
but is not inhibited by Tos-Phe-CH

Cl; the chymotrypsin inhibitor, chymostatin, is also only a weak

inhibitor.

Treponema denticola, frequently isolated from the human oral cavity, is a spirochete bacterium

associated with progression of several periodontal diseases, which was shown to secrete a powerful,
phenylalanine specific PR, denominated trepolisin (Table I).

56,57

The proteins degraded by this

chymotrypsin-like enzyme include IgA, IgG, fibrinogen, transferrin, a1-proteinase inhibitor, serum
albumin, casein, fibronectin, laminin, type IV collagen, and gelatin, whereas native type I collagen is
not degraded.

These proteins are in many cases degraded into fragments smaller than 12 kDa, with

Phe the preferred amino acid residue at the cleavage site. Leu, Pro, and Tyr bonds may also be
cleaved at a slower rate.

Trepolisin appears important in migration of the organism through

basement membrane, facilitating thus the invasion by the spirochete.

T. denticola also exerts

several cytopathic effects on epithelial cells, such as loss of cell contacts, degradation of pericellular
fibronectin, cytoskeletal collapse, formation of surface blebs, and intracellular vacuoles, and cell
death and trepolisin probably plays a major part in these effects.

The enzyme is also able to trigger

the activation of latent matrix metalloproteinases (MMPs), which also contribute to tissue
damage.

All in all, trepolisin appears to be a key T. denticola virulence determinant associated

with periodontal disease progression, but no specific inhibitors against this enzyme have been
developed up to now, although such compounds may be clinically relevant in preventing the pro-
gression of the disease.

Four tripeptidyl peptidases have been isolated from Streptomyces spp., which belong to the S8

(tripeptidyl peptidase S) and S33 (tripeptidyl peptidases A, B, and C) family of SPRs, respectively
(Table I).

These enzymes preferentially cleave amino terminal Ala-Pro-Ala-fragments; their

function is less easy to envisage, but may be related to transport of small peptides into the mycelium
or perhaps, more speculatively, in remodeling of the cell wall proteoglycan components.

A prolyl tripeptidyl peptidase was recently isolated from the oral cavity pathogen

Porphyromonas gingivalis,

a gram-negative, asaccharolytic anaerobic bacterium, which is one

of the principal causative agents of periodontal disease (see also Section 3.A dedicated to the
gingipains secreted by this pathogen). This enzyme releases tripeptides from small proteins such as
IL-6 and cystatin C, as well as from peptides, possessing an absolute requirement for Pro in the P1
site. The enzyme is anchored to the membrane through putative signal sequences, thus supporting its
physiological function, which is that of providing nutrients for the growing bacterial cells. The
external localization and uncontrolled activity of this PR contribute significantly to runaway inflam-
mation in the human host, correlated with the pathological degradation of connective tissue during
periodontitis.

It must also be mentioned that this enzyme probably works in concert with the two

cysteine proteases gingipain R and gingipain K secreted by P. gingivalis (which will be discussed
shortly), and this may explain the powerful virulence of this pathogen. No detailed inhibition studies
of this enzyme were reported, although it might be an attractive target for the drug design.

An enzyme that catalyzes the removal of N-terminal proline residues from peptides has been

detected in a variety of organisms, such as human oral cavity bacteria: E. coli, Serratia marcescens,
Xanthomonas campestris, Clostridium difficile, Flavobacterium spp., Mycoplasma genitalium, or
Neisseria spp. among others (Table I).

62,63

Bacterial prolyl aminopeptidases are highly specific for

336

SUPURAN, SCOZZAFAVA, AND CLARE

proline at P1.

Small peptides are easily cleaved by most of these enzymes, and the highest

activities have been observed for di- and tri-peptides. Prolyl aminopeptidase had long been assumed
to be cysteine PRs, but the recent report of the X-ray crystallographic structure for the Serratia

and

Xanthomonas

enzymes, undoubtedly proved that they are serine proteases, with a Ser, His, Asp

catalytic triad for the first PR, and a Ser, Asp, His one for the second enzyme, respectively. Little is
known for the moment on the inhibition of these enzymes, as no specific low molecular weight
inhibitors have been reported.

C. Streptomyces Trans- and Carboxypeptidases

The bacterial cell wall is constituted of peptidoglycans, which are assembled by membrane-bound
transglycosylases/transpeptidases, among which serine-acyl transferases, that characteristically
cleave the C-terminal peptide bond of acyl-

-alanyl-

-alanine peptides.

66–68

These enzymes couple

the breaking of the peptide bond and the transfer of the acyl-

-alanyl moiety to the amino group of

another peptide (

-transpeptidase activity) or to water (

-carboxypeptidase activity) via the

formation of a serine ester-linked acyl-

-alanyl enzyme intermediates.

66,67

With penicillin, they

form a long-lived, serine ester-linked penicilloyl enzyme, and thus behave as penicillin-binding
proteins (PBPs).

66–68

The Streptomyces K15 enzyme (Table I) also catalyzes the cleavage of

peptide, thioester, or ester bonds of carbonyl donors of the type Z-R1-CONH-CHR2-COX-CHR3-
COO

(where X is NH, S, or O) and transfers the electrophilic group Z-R1-CONH-CHR2-CO to

amino acceptors via an acyl-enzyme intermediate.

Penicillins, the most familiar antibiotics, exert

their antibacterial action by inactivating this type of enzymes involved in the peptidoglycan
synthesis, and since penicillins are cyclic analogues of

-Ala-

-Ala-terminated carbonyl donors, the

biosynthetic reaction stops at the level of the serine ester-linked penicilloyl enzyme.

66,68

On the

other hand, resistance to b-lactam antibiotics is also mediated by such transpeptidases (as well as by
b-lactamases, but these enzymes are not discussed here, as they cannot be defined as PRs, although
they catalyze a CONH hydrolytic reaction too), and this explains the considerable interest arisen by
the K15 enzyme, that has recently been crystallized, and its three-dimensional structure determined
by X-ray crystallography.

An enzyme with

-Ala-

-Ala carboxypeptidase activity, which was central to the understanding

of the mode of action of b-lactam antibiotics, has been isolated from Streptomyces strain R61.

The

reactions catalyzed by this enzyme, denominated Streptomyces R61 carboxypeptidase (Table I) are:

R-

-Ala-

-Ala

þ H

! R-

-Ala

-Ala (carboxypeptidase)

R-

-Ala-

-Ala

þ R

-NH

! R-

-Ala-NH-R

-Ala (transpeptidase)

(where R-

-Ala-

-Ala and R

-NH

are the donor and acceptor substrates, respectively).

Antibiotics of the b-lactam family are efficient transient inactivators that acylate the active-site

serine according to the pathway described above for the substrates, but no P1 is formed (due to the
b-lactam structure) and the deacylation step is very slow (10

–10

), so that an inactive acyl

enzyme accumulates.

69,70

Functionalized depsipeptides, derivatives of phenyl phenylacetylglyci-

nates (aryl phenaceturates) with a carboxylate substituent meta to the oxygen of the phenoxide
leaving group and a functionalized methylene group in the ortho- or para-position have been
reported to act as inhibitors of the serine

-peptidase of streptomyces R61.

The biological role of

this enzyme remains mysterious, but it could represent a primitive attempt to protect bacterial cells
against b-lactams produced by organisms living in the same environmental niches.

D. Signal Peptidase I, Omptin, and Clp

Signal peptidases belong to a family of serine proteases widely distributed in prokaryotes and
eukaryotes; these enzymes use a Ser/Lys catalytic dyad for the proteolytic scission.

72,73

Signal

BACTERIAL PROTEASE INHIBITORS

337

peptidases from different bacteria share sequence similarity and have similar substrate specificity.

Ser 90, the catalytic residue in the E. coli signal peptidase 1, is conserved in the whole family.

However, the critical Lys145 residue of the E. coli signal peptidase is conserved only in the bacterial
and the mitochondrial signal peptidases. All bacterial signal peptidases are integral membrane
proteins, residing in the outer leaflet of the cytoplasmic membrane and catalyzing the hydrolytic
cleavage of a specific peptide bond of membrane-imbedded preproteins, their function being that of
removing signal peptides from secreted and membrane proteins, liberating thus the mature proteins
for secretion.

The removal of signal peptides from exported proteins is an essential function for the

bacterial cell, because the uncleaved signal peptide acts as a membrane anchor, and cleavage
releases the secretory protein from the membrane.

72,73

Classical serine protease inhibitors do not

inhibit the type 1 bacterial signal peptidases that are essential enzymes in bacteria, and need to be
investigated as a possible drug target.

Nevertheless, some penem inhibitors of bacterial signal

peptidase have been reported, with affinity in the micromolar range for the E. coli enzyme.

Some

of the best such inhibitors, of type 1–3 are shown in Figure 2.

Such compounds may constitute

valuable lead molecules for obtaining low affinity (nanomolar) inhibitors with potential aplication
as antibiotics.

Omptin (also known as OmpT) is a serine protease present in many gram-negative bacteria. It

is an outer membrane protein (Omp) and the expression of its gene (ompT) is temperature
regulated.

Enzymes of the OmpT family represent a new class of integral membrane PRs with no

sequence homology with other known classes of proteases.

The prototype of the family, the E. coli

K-12 omptin, is an outer membrane endopeptidase with unusual specificity, cleaving the peptide
bond between two basic amino acids (Arg#Arg or Arg#Lys).

75,76

Similarly to signal peptidase

I, omptin utilizes a catalytic dyad in the proteolytic scission, that is constituted in this case by Ser
99 and His 212 (E. coli enzyme numbering).

76,77

Another distinct feature of OmpT is its ability to

function even under extreme denaturing conditions (high concentrations (8 M !) of urea for
example).

Omptin is insensitive to most proteinase inhibitors. Because it remains active under

extreme denaturing conditions, it can cleave recombinant proteins expressed in E. coli as they are
solubilized from inclusion bodies.

75,76

There are a growing number of reports that OmpT and related

proteases are associated with pathogenicity of certain gram-negative bacteria such as Yersinia pestis,
Shigella flexneri, and pathogenic E. coli strains.

No specific inhibitors of this PR have been

reported up to now.

Cytoplasmic bacterial proteases are necessary for the proper cell functioning, but they must be

also strictly regulated. Thus, many bacteria evolved chaperone-like proteases (chaperones are the
protein folding devices)—Clp

, which recognise exposed hydrophobic regions of unfolded/

denatured proteins. Such ATP-dependent proteases may be considered bifunctional enzymes
consisting of a chaperone linked to a protease. The chaperone activity probably functions in the
holoenzyme to unfold or disassemble protein substrates, which are then kinetically partitioned
between refolding and degradation pathways. One of the best studied such enzyme is the E. coli

Figure 2.

Signal peptidase inhibitors

1–3 with their IC

values (in

mM) against the

E. coli enzyme (see Ref. 74).

338

SUPURAN, SCOZZAFAVA, AND CLARE

ClpP (the name Clp was originally intended to suggest the clipping of proteins into a great number
of small peptides without the generation of free amino acids and was derived from the activity of this
enzyme as a caseinolytic protease), but many other chaperone-proteases have been isolated in
such micro-organisms and related bacteria.

78,79

The name might be more usefully interpreted as

chaperone-linked protease too. ClpP has peptidase activity against short ( < 10 amino acids) pep-
tides, whereas degradation of longer polypeptides and proteins requires the complex of ClpP and
ClpA, another chaperone present in these bacteria.

78,79

Cleavage occurs preferentially following

nonpolar residues, but significant rates of cleavage (5–15%) after polar and even charged residues
have been observed with protein substrates.

Although cleavage of the terminal amino acid residue

has been seen with some short peptide substrates, the enzyme does not possess significant exo-
peptidase activity. The enzyme is rapidly inactivated by Suc-Leu-Tyr-CH

Cl, but other, more

specific inhibitors were not found for the moment.

3 . C Y S T E I N E P R O T E A S E S

PRs, in which the nucleophile attacking the scissile peptide bond is the SH moiety of a Cys residue,
are known as cysteine (thiol) proteases (CPR). The catalytic mechanism is very similar to that of the
SPR, in which a nucleophile and a proton donor/general base are required. The proton donor, in all
CPRs, in which it has been identified, is a His residue as in the majority of SPR.

1,80

Although there is

evidence in some families that a third residue is required to orientate the imidazolium ring of the His
(a role analogous to that of the essential aspartate, Asp102, seen in some SPRs, see Fig. 1A), there
are a number of families in which only a catalytic dyad is necessary.

1,80

CPRs have been identified in

many bacteria species (Table I).

A. Proteases of the C10, C11, and C15 Families

A papain-like CPR has been isolated from cultures of S. pyogenes. This enzyme, subsequently
denominated streptopain, is unusual for a bacterial CPR in that it is formed from an inactive
zymogen of 337 amino acid residues, by proteolysis and reduction.

80,81

This enzyme hydrolyzes the

typical trypsin substrates, showing a preference for hydrophobic moieties in P1. It cleaves fibro-
nectin and vitronectin, two extracellular matrix proteins involved in the maintenance of host tissue
integrity, and also human interleukin 1b (IL-1b) precursor, to generate mature IL-1b with full
biological activity, suggesting a role in inflammation and shock.

It was then showed that strep-

topain efficiently releases biologically active kinins from their precursor proteins (H-kininogens).

These peptides act as hypotensors, increase vascular permeability, contract smooth muscle, and
induce fever and pain, and their release by streptopain may represent an important virulence
mechanism of this bacterium.

Streptopain is inhibited by common cysteine proteinase inhibitors,

such as iodoacetate, iodoacetamide, N-ethylmaleimide, p-chloromercuribenzoate, metallic ions
(Cu

, Hg

, and Ag

) and by the azomethane tripeptide Z-Leu-Val-Gly-CHN

80–82

Potentially

important virulence factors from Porphyromonas gingivalis (see later in the text, section 3.B) are
involved in proteolysis and agglutination of erythrocytes (hemagglutination).

The 96–99-kDa

protein derived from the PrtT gene of P. gingivalis has significant similarity in sequence to strep-
topain and its proenzyme.

80,83

Clostripain, a CPR produced and secreted by proliferating cells of the anaerobic bacterium

Clostridium histolyticum, is well known for its selective hydrolysis of arginyl bonds, although lysyl
bonds are also cleaved at a lower rate.

Similarly to the mammalian CPRs caspases, this enzyme

uses a catalytic dyad comprising a histidine residue in addition to the cysteine one.

Potent in-

hibitors of clostripain include oxidizing agents, thiol-blocking agents, and heavy metal ions (Co

BACTERIAL PROTEASE INHIBITORS

339

, Cd

). Tos-Lys-CH

Cl covalently modifies the active site, and is an effective active-site

titrant.

Higher rates of covalent reaction can be obtained with Phe-Ala-Lys-Arg-CH

Cl and

some methylsulfonium salts.

Several aziridine peptide inhibitors of clostripain were recently

reported: derivatives 4–6 for example, act as micromolar inhibitors (Fig. 3).

In contrast to

other CPRs (such as papain, cathepsin B, or cathepsin L), clostripain is preferentially inhibited by
esters 4–6 and not by the corresponding aziridinyl peptide acids.

These compounds act as

reversible inhibitors, whereas they generally irreversibly inhibit the above-mentioned CPRs
(caspases, clostripain). Thus, derivatives such as 4–6 may constitute important leads for the devel-
opment of more effective clostripain inhibitors, as it is generally supposed that this protease
constitutes one of the virulence factors of clostridia.

Clostripain also acts as a transpeptidase, showing maximal activity in the pH range 7.6–

9.0.

84,87,88

Recent studies of Bordusa’s group showed that lysine-containing substrates as well as

modified amino acid containing derivatives can act as effective acyl donors in such peptide synthetic
steps.

87,88

This approach is particularly useful for developing environmentally friendly, efficient,

and selective methods for peptide or peptide isosteres synthesis, and clostripain seems to be one of
the best biocatalysts for such transformations, due to its versatility in accepting non-natural, chemi-
cally modified amino acid residues in its binding pockets. Thus, it is possible to obtain easily peptide
isosteres for the synthesis of pharmacologically active, proteolytically stable derivatives.

87,88

Pyroglutamyl-peptidases, enzymes that remove the aminoterminal pyroglutamate (

-pyroglu-

tamyl, Glp) residue from specific pyroglutamyl substrates, were isolated from mammalian and
bacterial sources.

The enzyme was formerly classified as an aminopeptidase (EC 3.4.11.8), but is

more accurately regarded as an omega peptidase, because the substrate contains no free amino-
terminal NH

group. Pyrases are widely distributed in living organisms, where they play important

roles in the activation and inactivation of Glp-terminated peptides.

These enzymes appear to have

a broad substrate specificity, since most polypeptides with an N-terminal Glp are recognized and
hydrolyzed, the enzyme being thus involved in the regulation of the cellular pool of free Glp. It is
noteworthy that free Glp is known to be pharmacologically important, since many hormones (such
as TRH, thyrotropin-releasing hormone; TSH, thyroid-stimulating hormone; LHRH, luteinizing
hormone-releasing hormone among others) possess this modified amino acid residue in the amino
terminal position.

The specific source of the free Glp that is associated with some diseases remains

unknown, but the involvement of pyrase remains a possibility. In bacteria, this enzyme may have a
nutritional function, or may be involved in intracellular protein metabolism. It was also suggested
that bacterial pyroglutamyl-peptidases are involved in detoxification, since peptides with the amino
terminal Glp residue abnormally acidify the bacterial cell cytoplasm.

An interesting application of bacterial pyroglutamyl-peptidases may be in the diagnosis of

bacterial infections, particularly in the identification of group A streptococci and enterococci,
facilitated by the discovery of a convenient colorimetric procedures.

Activators and inhibitors of

bacterial and mammalian pyrases might constitute valuable therapeutic agents. Thus, the protection

Figure 3.

Aziridinyl peptide inhibitors

4–6 of clostripain. K

values (in

mM) given in the brackets (see Ref. 86).

340

SUPURAN, SCOZZAFAVA, AND CLARE

of the pyroglutamyl residue of peptides from attack by pyrase may improve the delivery of such
peptides in therapeutics.

Pyroglutamyl derivatives of several anticancer drugs have been proposed

as potential prodrugs, designed to be cleaved in situ to the pharmacologically active cytotoxic agent
in the presence of pyrase.

For targeting the tumor site, a pyrase chemically linked to a monoclonal

antibody may be useful.

B. Porphyromonas gingivalis Cysteine Proteases

Periodontal disease is characterized by a group of infections leading to chronic inflammation of the
gingivae, destruction of the periodontal tissue, and loss of alveolar bone, frequently followed by
tooth loss as the final and undesired event.

93,94

Porphyromonas gingivalis, a black-pigmenting

anaerobe and well-established oral pathogen, has been implicated as the major etiologic agent of
this disease, since this pathogen secretes substantial quantities of two thiol-dependent PRs of
molecular weight between 44–50 kDa, that cleave synthetic substrates with arginine and/or lysine
residues at the P1 position, which were thus denominated gingipains R and K, respectively.

5,93,94

addition to protein substrates (such as for instance azocasein), both gingipains are reported to
degrade many protein components of human connective tissue and plasma, including immuno-
globulins, proteinase inhibitors and collagens, explaining thus the powerful tissue destruction
produced by the pathogen.

5,93–95

In most P. gingivalis strains the majority of gingipains are

associated with either cell surface outer membrane or the vesicles that are the blebs of this mem-
brane, indicating that, as a part of the secretory pathway, gingipains translocate through the cyto-
plasmic (inner) membrane and become inserted, at least transiently, into the outer membrane.

The

following secretory events correlated with the pathogenicity of the two gingipains were suggested:
(i) crossing of the cytoplasmic membrane into the periplasmic space, directed by the cleavable
amino terminal signal peptide; (ii) the carboxy-terminal part of the gingipain precursor integrates
with the outer membrane and translocates the remaining part of the protein; (iii) at the cell sur-
face the enzyme acquires its active conformation after proteolytic removal of a pro-fragment; and
(iv) it undergoes further proteolytic processing. Except for the signal peptide removal, the major
post-translational proteolytic processing of gingipain precursors occurs by cleavage at Arg-Xaa
peptide bonds.

Gingipains contribute to the development and/or maintenance of a pathological inflamma-

tory state by at least three mechanisms: (i) activation of the kallikrein-kinin cascade; (ii) release of
neutrophil chemotactic activity from the C5 protein of the complement pathway; (iii) activation
of the coagulation enzymes factor X and prothrombin, which leads to the uncontrolled generation of
thrombin, an enzyme with multiple functions in creating a strong proinflammatory state.

5,61,95,96

Thus, considerable effort has been spent in purifying antigenic complexes and raising anti-

bodies against this pathogen, which might lead to the development of vaccines effective against
periodontitis. The cell surface proteins PrtR and PrtK, as well an antigenic complex including the
two proteins have been reported as effective in raising antibodies against P. gingivalis.

The above-

mentioned complex consists of a 160 kDa Arg-specific gingipain with c-terminal adhesin domains
(designated PtrR) associated with a 163 kDa Lys-specific gingipain also with C-terminal adhesin
domains (designated PrtK) and was administered as a mouth-wash or as a dentifrice, being shown to
induce an immune response against the pathogen.

Although the X-ray structure of gingipain R has recently been reported,

few specific and

potent active-site-directed inhibitors of these enzymes were reported up to now. Like many
CPRs, gingipains are inhibited by thiol-blocking reagents, including iodoacetamide, iodoacetic
acid, and N-ethylmaleimide, but only at rather high concentration (10 mM), and after relatively
long times of incubation (20–30 min).

Recently, Potempa and Travis’ groups showed that

chloromethane derivatives inhibit both these enzymes to varying degree, depending on the peptidyl
components of the inhibitor.

Thus, compounds containing a basic residue at P1 rapidly inactivated

BACTERIAL PROTEASE INHIBITORS

341

the gingipains, some specificity being seen for the P2 residue too. The (acyloxy)methane inhibitor,
Cbz-Phe-Lys-CH

OCO-2,4,6-Me

, was very specific in its rapid inhibition of gingipain K over

the gingipain R. This inhibitor, together with the peptidyl chloromethanes,

-Phe-Pro-Arg-CH

and

-Phe-Phe-Arg-CH

Cl, which reacted most rapidly with the Arg-specific proteinases, has been

used to active site titrate purified forms of the enzymes as well as enzymes found in crude fractions
such as intact P. gingivalis cells, vesicles or membrane fractions. From such experiments, it was
found that gingipains R were always in an about 3-fold excess over gingipain K, and that the two
enzymes as a whole made up 85% of the proteolytic activity associated with this oral pathogen.

conclusion, these two proteases constitute very interesting targets for the drug design of agents
useful in the fight against the periodontal disease.

C. Sortase

A recently described cysteine protease critical for the interaction of gram-positive bacteria with their
animal hosts (and which is ubiquitous in these much feared pathogens) is sortase.

100–106

Cell wall

sorting is the covalent attachment of surface proteins to the peptidoglycan via a C-terminal sorting
signal that contains a consensus LeuProXThrGly (LPXTG—X may be any of the 20 natural amino
acids) sequence.

100–107

Sortase participates in the pathways involved in secretion and anchoring of

cell wall proteins, by a mechanism conserved in almost the entire class of the gram-positive bacteria.
Sortase catalyzes the scission of the Thr # Gly bond of the above-mentioned motif, liberating the
carboxyl moiety of the threonine, which subsequently forms an amide bond with the amino groups
of some amino acid residues present in the bacterial cell wall peptidoglycan, which in turn are cross-
linked to the E-amino group of lysine residues (for S. aureus it is the amino moiety of a pentaglycine
motif that cross-bridge with the COOH moiety of the above-mentioned Thr residue; in S. pyogenes
this motif is constituted by two alanines, whereas Listeria monocytogenes has a meso-diamino-
pimelic acid moiety in place of the lysine residue of the peptidoglycan, and no amino acid in the
bridge).

100–107

Sortase is a 206-amino acid protein with a potential amino-terminal signal peptide that may act

as membrane anchor. The conserved Cys 184 is considered critical for catalysis, since the enzyme
is sensitive to agents that modify SH groups.

103

Sortase possesses both protease as well as trans-

peptidase activities,

102,104

similarly to Streptomyces K15

-Ala-

-Ala transpeptidase and Strepto-

myces R61

-Ala-

-Ala carboxypeptidase previously discussed (see section 2).

It has recently been proved that several LPXTG proteins of staphylococci (such as protein A,

clumping factor, fibronectin-binding proteins, etc.)

105,106

contribute to the virulence of S. aureus,

and that sortase defective strains of these bacteria are non-pathogenic, probably due to incorrect
surface presentation.

105,106

These data clearly showed that cell wall anchoring of surface proteins

is a critical parameter for the development of staphylococcal infection, and that sortase is a very
promising target for the development of novel types of antibiotics, effective against gram-posi-
tive bacteria.

100–107

For the moment, few inhibitors of this enzyme have been reported, maybe also due to the fact

that the target has only recently been purified and demonstrated as critically important for the
life cycle of gram-positive bacteria. Thus, Schneewind’s group

103

showed that sulfhydryl reducing

agents (dithiothreitol, mercaptoethanol, etc.), phenylmethylsulfonyl fluoride or EDTA do not act
as sortase inhibitors, whereas methanethiosulfonates (such as [2-(trimethylammonium)ethyl]-
methanethiosulfonate; (2-sulfonatoethyl)methanethiosulfonate) or organo-mercurials (such as
p-hydroxymercuribenzoic acid) displayed inhibitory effects. Penicillin G, a transpeptidation in-
hibitor was not a sortase inhibitor, whereas vancomycin and moenomycin did slow the sorting
reaction.

103

Thus, it would be stringently important to design specific and powerful inhibitors of this

enzyme, which might have very important applications as new generation antibiotics, but for the
moment such compounds are not available.

342

SUPURAN, SCOZZAFAVA, AND CLARE

4 . M E T A L L O P R O T E A S E S

Metalloproteases (MPRs) are the hydrolases in which the nucleophilic attack on the scissile peptide
bond is mediated by a water molecule coordinated to a divalent metal ion (usually Zn(II) but
sometimes cobalt or manganese, may also activate the water molecule) or bridged to a dimetalic
center (two Zn(II) ions, or one Zn(II) and one Co(II)/Mn(II) ions, etc).

The catalytical metal ion is

coordinated by amino acid moieties present within the active site, usually three in number (the most
frequent being His, Glu, and Asp), the fourth ligand being as mentioned above a water molecule/
hydroxide ion.

108

Thus, MPRs are divided into two main groups depending on the number of metal

ions required for catalysis: in the majority of the described MPRs, only one metal ion is required, but
in some families, there are two metal ions that act co-catalytically. All the metallopeptidases in
which cobalt or manganese is essential for activity, require two metal ions, but there are also
families of zinc-dependent MPR in which two zinc ions are co-catalytic.

1,108

The best studied

dinuclear zinc-PRs are several exopeptidases such as the Aeromonas proteolytica aminopeptidase,
the Streptomyces griseus aminopeptidase, carboxypeptidase G2 of pseudomonads, as well as
methionyl aminopeptidase of E. coli.

1,108

In such MPRs with binuclear metal centers, five amino

acids residues act as ligands (predominantly by means of carboxylate moieties), and one of them
ligates both metal ions, acting as bridging ligand. All MPRs with two catalytic metal ions so far
described are exopeptidases, whereas MPRs with one catalytic metal ion may be exo- or endo-
peptidases.

1,108

MPRs are widely spread in all types of bacteria (Table II), being critical virulence factors, and

playing various pathogenic roles in infection.

6,109,110

Thus, in local bacterial infections, such as

keratitis, dermatitis and pneumonia, MPRs function as decisive virulence determinants, being
generated at the site of infection and causing necrotic or hemorrhagic tissue damage through
digestion of structural tissular components.

6,109

Furthermore, such in situ generated PRs also

enhance vascular permeability, leading to the formation of oedematous lesions through generation
of inflammation mediators such as histamine, bradykinins, kallikreins, or thrombin, allowing further
dissemination of the infection through the systemic circulation.

6,109,110

In the case of systemic

infections, such as septicemia, MPRs act as a synergistic virulence factor, provoking a disordered
proteolysis of various plasma proteins, causing imbalances of the proteinase-proteinase inhibitor
equilibrium, disturbing thus the physiological homeostasis and eliciting an immunocompromised
state of the host.

6,110

MPRs also trigger the disintegration of the iron-carrier proteins of the host,

which has as a consequence an enhanced uptake of iron, an essential element for the growth of
bacteria. Some bacterial toxins, such as some enterotoxins (cholera toxin) are activated by MPRs,
whereas other toxins, such as the botulinum (BoNT) and tetanus neurotoxins (TeNT) are metal
proteases themselves.

111,112

It is thus clear that such MPRs constitute very important potential

targets for the drug design of novel types of antibiotics.

A. Metalloproteases of the Thermolysin Family (M4)

Thermolysin, an unspecific protease secreted by the gram-positive thermophilic bacterium Bacillus
thermoproteolyticus is an extracellular 34.6 kDa metalloendopeptidase, without pharmacological
applications, but which has been much investigated (it was the first MPR for which the X-ray
structure was available

108

) and used to generate active-site models for the design of inhibitors for

mammalian zinc PRs (such as neprilysin for example).

113

This and related M4 family proteases

contain a Zn(II) ion coordinated by two histidine residues belonging to a His-Glu-Xaa-Xaa-His
(HEXXH) motif, whereas the third zinc ligand is a glutamate residue at least 14 residues C-terminal
to the last His of this motif. The zinc-coordinated water molecule also establishes a hydrogen bond
with the carboxylate moiety of the glutamate belonging to the HEXXH motif, which strongly
enhances the nucleophilicity of this water molecule.

1,108,113

BACTERIAL PROTEASE INHIBITORS

343

Table II. Bacterial Metalloproteases, With Their Preferred Scissile Bond and Organisms From Which Were
Isolated

Preferred scissile bond

Protease

Family

Organism(s)

(P1#P1

Listeria metalloprotease (Mpl)

Listeria monocytogenes

Unspecific PR
(thermolysin-like)

3.4.24.30

Coccolysin

Enterococcus faecalis

Phe#Phe; Gly#Phe;
Gly#Leu; Pro#Phe

Hemagglutinin/protease

Vibrio cholerae

Gly#Phe; Gly#Leu;

Helicobacter pylori

Ser#Met; Ser#Ser

3.4.24.26

Pseudolysin

Pseudomonas aeruginosa

Phe#Xaa; Gly#Leu

Legionella

Legionella pneumophila

Unknown

metalloendopeptidase

3.4.24.3

Vibrio collagenase

Vibrio alginolyticus;

Xaa#Gly

V. parahaemolyticus

3.4.24.3

Clostridium collagenases

Clostridium histolyticum;

Xaa#Gly

C. perfringens

Non-haemolytic enterotoxin

M9?

Bacillus cereus

Xaa#Gly

3.4.24.40

Serralysin

M10

Serratia spp.

Xaa#Gly; Xaa#Ala

Pseudomonas aeruginosa;

Erwinia chrysanthemi

3.4.24.40

Aeruginolysin

M10

Pseudomonas aeruginosa

Leu#Gly; Gly#Gly

Mirabilysin

M10

Proteus mirabilis

Leu#Gly (in IgA)

3.4.24.74

Fragilysin

M10

Bacteroides fragilis

Leu#Gly; Gly#Leu

Flavastacin

M12

Flavobacterium

Xaa#Asp

meningosepticum

3.4.17.18

Carboxypeptidase T

M14

Streptomyces spp.

Xaa#Xaa-COOH

3.4.11.10

Leucyl aminopeptidase

M17

Chlamydia trachomatis

N-Leu#Xaa

E. coli; Haemophilus influenzae;
Mycobacterium tuberculosis;

Mycoplasma genitalium

Rickettsia prowazekii, S. typhimurium

3..4.11.18

Methionyl aminopeptidase I

M24

Bacillus spp.; Clostridium

N-Met#Xaa

perfringens

E. coli; Haemophilus influenzae;
Klebsiella pneumoniaea;

Mycoplasma spp.;

Salmonella typhimurium;

Synechocystis spp.

3.4.17.11

Glutamate carboxypeptidase

M20

Pseudomonas spp.;

Xaa#GluCOOH

Flavobacterium spp.

Acinetobacter spp.

VanX

-dipeptidase

M19

Enterococcus spp., Synechocystis

-Ala#

-Ala*

spp.

3.4.24.57

O-Sialoglycoprotein

M22

H. influenzae; M. leprae

Arg#Asp**

endopeptidase

Mycoplasma genitalium;

Pasteurella haemolytica

3.4.24.32

b-Lytic metalloendopeptidase

M23

Lysobacter enzymogenes

N-acetylmuramoyl#Ala
Gly#(

"-amino)Lys

Staphylolysin

M23

Aeromonas hydrophila

Gly#Gly***

Pseudomonas aeruginosa

3.4.24.13

IgA-specific

M26

Streptococcus spp.

Pro#Thr; Pro#Ser

metalloendopeptidase

Neisseria spp.; Haemophilus spp.;
Ureaplasma spp.; Clostridium spp.;
Capnocytophaga spp.; Prevotella spp

3.4.24.68

Tentoxilysin

M27

Clostridium tetani

Gln#Phe

(tetanus neurotoxin)

(continued)

344

SUPURAN, SCOZZAFAVA, AND CLARE

A potent thermolysin inhibitor has recently been reported by Vilcinskas group, who isolated the

first metalloprotease inhibitor from invertebrtates (IMPI, for insect metalloprotease inhibitor).

114

This 8.4 kDa protein was shown to strongly inhibit thermolysin in vitro and in vivo, being detected
only in larvae of the greater wax moth Galleria mellonella that have been injected with bacterial,
fungal, or viral provocators.

114

It was thus suggested that this protease inhibitor may play an

important function in humoral immune response and may possess biomedical applications also
in human therapeutics.

114

The mechansim of thermolysin inhibition by IMPI is unknown for the

moment.

114

Several MPRs belonging to the M4 family were isolated in different bacteria, which might have

important applications for the drug design of new antibiotics (Table II). Thus, Listeria mono-
cytogenes is a gram-positive non spore-forming, facultative intracellular rod-shaped bacterium,
which is capable of causing sepsis and CNS infections in humans and animals.

115

The L. mono-

cytogenes Mpl protease contains 510 amino acids and has a predicted molecular mass of 57.4 kDa.
Following cleavage of the amino-terminal 24 amino acid signal sequence, the 55 kDa inactive
zymogen is secreted to the external medium, whereas mature active protease possessing a molecular
mass of 36 kDa, is formed after further processing of the N-terminal 180 amino acids.

115

The

enzyme is inhibited by the metal chelators of the polyamino-polycarboxylic acid type (EDTA,
EGTA), by 1,10-phenanthroline and the thermolysin-type protease inhibitor phosphoramidon. Its
activity is stimulated by low concentrations (0.1 mM) of zinc chloride, but inhibited by high
concentrations (0.5 mM). The specificity of cleavage by the Mpl protease has not been defined, and
no specific inhibitors were designed, although a model of its active site has recently been reported on
the basis of its homology with thermolysin.

115

Enterococcus faecalis is frequently identified as the etiologic agent of several opportunistic

infections (such as soft tissue and urinary tract infections, intra-abdominal abscesses, and root canal
infections), and as the causative agent in several cases of endocarditis, secondary bacteremia, food
poisoning. It has been proposed, based on the specificity profile of the MPR isolated from this
pathogen—denominated coccolysin—that the extracellular production of this protease is associated
with the above-mentioned clinical conditions.

116

Coccolysin inactivates human endothelin-1 by

hydrolyzing this peptide primarily at the Ser5#Leu6 and the His16#Leu17 bonds, and the human big
endothelin at several bonds involving hydrophobic amino acid residues (similarly with thermolysis
or Mpl, this is also a relatively unspecific protease).

116

The degradation of endothelin by cocco-

lysin resembles the peptidolytic processing of endothelin by thermolysin. Because E. faecalis
is associated with a large number of infectious diseases, it is probable that the manifestation of

Table II. (Continued )

Preferred scissile bond

Protease

Family

Organism(s)

(P1#P1

3.4.24.69

Bontoxilysin

M27

Clostridium botulinum

Gln#Phe; Gln#Arg

(botulinum neurotoxin)

C. barati; C. butyricum

Lys#Ala, Arg#Ile

Anthrax toxin lethal factor

M34

Bacillus anthracis

Unknown

3.4.24.75

Lysostaphin

M37

Staphylococcus staphylolyticus

Gly#Gly***

S. simulans

3.4.24.29

Aureolysin

Staphylococcus aureus

Xaa#Leu

AAA proteases

M41

ubiquitous

(FtsH)

NC, not classified; M, metallo-PR family.

*The dipeptide is the enzyme substrate.

**From

O-sialoglycoproteins.

***From pentaglycine cross-linking peptides of

S. aureus peptidoglycan.

BACTERIAL PROTEASE INHIBITORS

345

inflammatory conditions in the presence of this organism is related to the coccolysin-catalyzed
inactivation of endothelin, but no inhibitors directed against this protease have been reported for the
moment.

116

Cholera is an infectious disease caused by the bacterium Vibrio cholerae and characterized by

severe vomiting and watery diarrhea.

117

These bacteria produce a cytolytic toxin named El Tor

cytolysin/hemolysin, which is encoded by the hlyA gene. The cytolysin is produced as a 79-kDa
precursor form (pro-HlyA) after cleavage of the signal peptide of prepro-HlyA. The pro-HlyA is
then processed to a 65-kDa mature cytolysin (mature HlyA) after cleavage of the 15-kDa amino-
terminal peptide (pro region) of the 79-kDa precursor, usually at the bond between Ala-157 and
Asn-158. The hemagglutinin/protease, a major thermolysin-like protease of V. cholerae, processes
the pro-HlyA to the 65-kDa mature form of the protein. Along with this, the protease-processed
HlyA drastically increases hemolytic activity (the N-terminal amino acid of the mature form of
cytolysin generated by HA/protease was Phe-151).

117

Both hemagglutinating and protease functions

of this protease are inhibited by chelating agents, including 2-(N-hydroxycarboxamido)-4-methyl
pentanoyl-

-Ala-Gly-NH

(Zincov), a hydroxamic acid derivative specifically designed to inhibit

zinc metalloproteases,

but no other specific or more potent inhibitors of this enzyme have been

reported. Since the proteolytic activity of hemagglutinin/protease plays a key role in activation of
toxins of V. cholerae,

118

finding such inhibitors would constitute an interesting alternative to the

treatment of this disease. On the other hand, a Vibrio cholerae strain defective in hemagglutinin/
protease constitutes an anti-cholera vaccine candidate, which has been examined for safety and
immunogenicity in healthy adult volunteers, showing promising activity.

118

Pseudolysin is the current name of the most abundant extracellular endopeptidase of

Pseudomonas aeruginosa an opportunistic pathogen, that may cause life threatening infections in
compromised patients with underlying respiratory disease like bronchiectasis, cystic fibrosis and
diffuse panbronchiolitis (this protease is commonly called Pseudomonas aeruginosa elastase
too).

119

Most strains of P. aeruginosa produce some kind of protease with broad substrate speci-

ficities during the infectious state in the host. P. aeruginosa elastase has a tissue-damaging
proteolytic activity and is capable of degrading plasma proteins such as immunoglobulins, com-
plement factors, and cytokines.

119

Destruction of the arterial elastic laminae in human systemic

P. aeruginosa infections was the first evidence that P. aeruginosa secretes an elastinolytic protease.
Pseudolysin is the major extracellular virulence factor of P. aeruginosa.

119,120

The contribution of

pseudolysin to disease may be direct, causing tissue destruction, damaging some cell functions, or
indirect, promoting virulence by interfering with host defense mechanisms.

Metal chelators, including EDTA, EGTA, 1,10-phenanthroline, and tetraethylene pentamine,

inhibit the activity of pseudolysin. Phosphoramidon, phosphoryl-dipeptides such as phosphoryl-
Leu-Phe and phosphoryl-Leu-Trp, and peptides containing thiol or hydroxamate groups such as
HSCH

CONH-Phe-Leu, HSCH

CH-(CH

)CO-Ala-Gly-NH

, or HONHCOCH(CH

CO-Ala-Gly-NH

, are potent reversible inhibitors.

121

Inhibition by these compounds is not specific

to pseudolysin since other M4 family PRs are also inhibited.

121

Pseudolysin may contribute in many ways to different diseases, so that effective/specific

inhibitors would be important as potential antibiotics. For example, pseudolysin is probably
responsible for the destruction of arterial elastic laminae in the vasculitis observed in cases of
Pseudomonas septicemia.

120,121

This PR may also induce septic shock through activation of the

Hageman factor-dependent kinin system; activation of the host kinin cascade may also be involved
in pathogenesis of skin burns infected by this pathogen, in which case pseudolysin appears to
support growth and invasiveness of the organisms.

121

By rapidly degrading corneal proteoglycans,

pseudolysin causes severe corneal destruction during Pseudomonas keratitis, but it may also affect
corneal damage indirectly, by activating endogenous corneal proteinases.

121

In Pseudomonas

pneumonia pseudolysin may cause lung damage with hemorrhages and necrosis of alveolar
septal cells, destroying alveolar epithelial cell junctions, which increases epithelial permeability

346

SUPURAN, SCOZZAFAVA, AND CLARE

to macromolecules.

121

Pseudolysin seems also to be involved in the chronic disease caused by

P. aeruginosa in cystic fibrosis patients.

121

An epidemic of pneumonia that broke out among attendees at a convention of the American

Legion in Philadelphia in 1976 was ultimately traced to a novel organism, Legionella pneumophila,
isolated in the water-cooling units of the air-conditioning system of the convention hotel.

122

The

organism is a facultative intracellular parasite in alveolar macrophages, and secretes a thermolysin-
like metalloprotease, which is one of the principal pathogenic factors in Legionnaires’ disease due
to its cytotoxic, tissue-destructive, and phagocyte-inhibitory properties.

122,123

The protease is

cytotoxic to both neutrophils and monocyte/macrophages and interferes with the binding of natural
killer cells to their target cells.

122,123

Activity of this enzyme is inhibited by various metal chelators

such as EDTA, EGTA, and 1,10-phenanthroline, but not by DTT (10 mM). Potent inhibition
is obtained with phosphoramidon and the phosphoramidate analog Z-GlyP(==O)Leu-Ala.

122

specific inhibitors for the Legionella metalloendopeptidase have been designed up to now.

B. Metalloproteases of the M9 Family (Vibrio and
Clostridium Collagenases)

Family M9 contains bacterial collagenases from Vibrio and Clostridium, as well as a number of
other collagenolytic bacterial endopeptidases less investigated for the moment.

The Clostridium

collagenase (ChC, the best studied enzyme in this family) is only distantly related to the Vibrio
collagenase, and had previously been considered to be the sole member of the now defunct M31
family.

Both the Vibrio and Clostridium collagenases are secreted enzymes, and are synthesized as

precursors. The metal ion ligands have only recently been determined for ChC

124

: two of the zinc

ligands occur in the His

415

ExxH motif, while the third one is Glu 447. Similarly to the vertebrate

matrix metalloproteinases (MMP, that also degrade collagen), ChC is a multiunit protein, consisting
of four segments, S1, S2a, S2b, and S3, with S1 incorporating the catalytic domain.

124,125

Vibrio alginolyticus chemovar iophagus is a nonpathogenic marine bacterium isolated from

cured hides, which rapidly lyses collagen under aerobic conditions.

126

The 82 kDa collagenase

responsible for this activity (containing one Zn(II) ion/polypeptide chain), cleaves native collagen
much more rapidly than does vertebrate interstitial collagenase (MMP-1).

126

In the first step, this

enzyme acts on collagen in a manner similar to the MMPs, by attacking at a point three-quarters of
the way from the amino-terminus, although the bond preferentially cleaved is different: Xaa#Gly for
Vibrio collagenase, instead of Gly#Leu or Gly#Ile for the MMPs.

126

The use of Vibrio collagenase

in tissue cell dispersions, elastin purification, and selective cleavages of gene products in
biotechnology is similar to the same use of ChC, which will be discussed shortly. However, more
important applications in human therapy are in the removal of necrotic tissues from burns, ulcers,
and decubitus ulcers because of its strong and specific activity against native collagen, a char-
acteristic it also shares with the Clostridium collagenase.

126

Clostridium histolyticum is a pathogenic anaerobe that causes gas gangrene and other infections

such as bacterial corneal keratitis.

127–129

All strains of this bacterium elaborate a collagenase (ChC),

possibly used as a means to invade the host and to degrade its protein for nutritional purposes.

127

There are at least seven collagenase isozymes, with molecular masses ranging from 68–130 kDa,
that have been purified to homogeneity.

128

They are designated as class I (a, b, g, and Z) and class II

(d, ", and x) enzymes, based on a variety of criteria including their relative activities toward collagen
as well as synthetic substrates.

128

In fact, the crude homogenate of Clostridium histolyticum, is one

of the most efficient systems known for the degradation of connective tissue, being able to hydrolyze
triple helical regions of collagen under physiological conditions, as well as a multitude of synthetic
peptide substrates.

127–129

The initial proteolytic events in this collagenase-mediated hydrolysis of

type I, II, and III collagens have been delineated: the enzymes initially attack all three collagens
at distinct hyper-reactive sites, cleaving between Yaa#Gly bonds in the repeating Gly-Xaa-Yaa

BACTERIAL PROTEASE INHIBITORS

347

collagen sequence.

127

The hyper-reactivity of these cleavage sites appears to be more related to local

conformational features of the collagen fold than to the surrounding sequence. Thus, in contrast to
the vertebrate collagenases, ChC degrades collagen into small peptides.

127,130

ChC could not be crystallized for the moment, and its three-dimensional structure is thus not

available. Only recently some electronic spectroscopic studies of Co(II)-substituted ChC have been
reported,

131

offering thus interesting information regarding the binding of inhibitors within the

active site of this bacterial protease. The Co(II)-substituted ChC retains catalytic activity, similarly
to the native zinc enzyme,

127

and possesses a pH-dependent electronic absorption spectrum, with a

maximum centered at 585 nm and a shoulder at 530 nm, this spectrum being relatively similar to that
of Co(II)-substituted carboxypeptidase A or thermolysin

132

two enzymes in which the Zn(II) ion is

coordinated—such as in ChC—by two histidines and a glutamate.

133

In the presence of hydro-

xamate inhibitors (of the type 7–10, see later in the text), major changes of this spectrum were
evidenced: three absorption maxima instead of the two mentioned above appeared, at 501–505,
562–563, and at 597–598 nm, respectively. The spectra were of low intensity (molar absorbances
around 80–120 M

for the first two maxima, and of around 11–15 M

for the last

one), being characteristic of Co(II) in pentacoordinated geometry.

131

It was thus concluded that the

sulfonylated amino acid hydroxamates, the most potent class of ChC inhibitors, bind to the metal
ion within the enzyme active site, leading to pentacoordinated Co(II) ions

131

(Fig. 4).

Thus, an entire range of sulfonylated amino acid hydroxamate ChC inhibitors of types 7–10

were recently reported, considering the MMP inhibitors

134

of the same type as lead molecules for

obtaining high affinity ligands for this bacterial protease.

131,135–140

These compounds inhibit both

bacterial as well as vertebrate collagenases (Fig. 5).

These studies showed that the sulfonylated amino acid hydroxamates represent a very potent

class of ChC inhibitors, but some structurally related arylsulfonylureas-, arylureas or sulfenamido-
4-nitrobenzyl-Gly derivatives were also proved to inhibit this enzyme.

131,135–140

These studies

showed that the S1

-binding moiety of the arylsulfonamide type, previously investigated for

obtaining non-peptide MMP inhibitors

134

can be efficiently substituted by related moieties such as

alkylsulfonyl-; arylsulfenyl-; arylsulfonylureido-; arylureido; or benzoyl-thioureido, without loss of
the MMP/ChC inhibitory properties. In the large series of alkyl/arylsulfonamido derivatives
investigated as MPR inhibitors, best ChC inhibitory properties were correlated with the presence
of perfluoroalkylsulfonyl-, perfluorophenylsulfonyl-, 3-trifluoromethylphenylsulfonyl-, 3-chloro-4-
nitro-phenylsulfonyl-, 3-/4-protected-amino-phenylsulfonyl-, 3-/4-carboxy-phenylsulfonyl-moi-
eties as S1

anchoring group. Such derivatives possessed inhibition constants in the range of

5–10 nM against ChC. These data indicated that ChC is similar to a short-pocket MMP, eventually
possessing a slightly wider neck than MMP-1.

31,135–141

ChC inhibitors of the above-mentioned type might be very useful for the treatment of bacterial

keratitis.

135–141

Thus, it was reported that collagen shields applied to the corneas of patients affected

Figure 4.

Proposed schematic binding of a sulfonylated amino acid inhibitor within the active site of ChC (see Ref.131).

348

SUPURAN, SCOZZAFAVA, AND CLARE

by bacterial keratitis degrade rapidly, often within a few hours.

142

Once treatment brings the

infection under control, subsequently applied collagen shields degrade more slowly, being shown
that the rate of collagen shield degradation may be a clinically useful index of collagenase activity
on the ocular surface. Ultrastructural studies of collagen shields from patients with acute bacterial
keratitis revealed irregular degradation of shield matrix with no evidence of adherence of micro-
organisms or inflammatory cells.

142

Co-incubation of deepithelialized rabbit corneas and collagen

shields resulted in inhibition of the digestion of the rabbit corneas. Collagen shields may inhibit
corneal collagen degradation in infectious ulceration and melting disorders by effectively com-
peting for collagenase on the ocular surface.

142

Furthermore, combining this inhibitory effect with

the one of an endogenous collagenase inhibitor would highly facilitate the healing in this very
serious eye disease.

135–141

Corneal collagen shields as a drug delivery device were also investigated

for the treatment of bacterial keratitis.

142

Thus, the effectiveness of topical antibiotic treatment, with

and without the use of corneal collagen shields, in a rabbit model of Pseudomonas keratitis, showed
that collagen shields hydrated in tobramycin and supplemented with topical tobramycin were highly
effective in the sustained treatment of experimental Pseudomonas keratitis.

142

Treatment of keratitis

with antibiotic-impregnated collagen shields may reduce the need for very frequent application of
topical drops, but may be more effective with topical drop supplementation to increase the amount
of drug available over the course of therapy.

142,143

Another important application of Clostridium (and Vibrio) collagenases is in dermatology, in

the debridement of dermal ulcers, in patients with chronic non-healing wounds, such as pressure
sores, venous leg ulcers and diabetic ulcers among others.

144,145

In such cases, treatment of the

Figure 5.

Sulfonylated amino acid hydroxamates

7–10 and their inhibition data against the collagen degrading enzymes form ver-

tebrates (MMPs) and

Clostridium hystolyticum (ChC)

(see Refs.131,134^141).

BACTERIAL PROTEASE INHIBITORS

349

wound with bacterial proteases has a debriding effect due to proteolysis of the connective tissue by
the MPR, whereas some proteolysis products may promote healing themselves by a mechanism
little understood for the moment, probably involving the migration and stimulation of activity of
important cells such as wound macrophages, fibroblasts, and keratinocytes.

144,145

The collagen

cleavage products after treatment with bacterial collagenases show also chemoattractant properties
for diverse cells involved in wound healing processes. These enzymes may also be used for the
treatment of burn wounds.

144,145

Recently, a collagenolytic and gelatinolytic MPR has been isolated in Bacillus cereus, this

protein having some sequence homology with the C. histolyticum and C. perfringens collagenases.
This 105 kDa protease presumably belongs to the M9 family, but little is known for the moment on
its role in infection, inhibition, etc.

146

C. Serralysin and Related M10 Proteases; Proteases
of the M12 Family

MPRs belonging to the clan MB have two of the three zinc ligands His residues in an HEXXH motif,
but unlike clan MA discussed above, their third zinc ligand is again a His (or an Asp in rarer cases),
in the extended motif HEXXHXXGXX(H/D) (the zinc ligands are evidenced in bold characters).

The Glu next to the first histidine is predicted to have the same role in catalysis as Glu 142 of
thermolysin and other MA metalloproteases discussed above (see section 4.A), and the conserved
Gly allows the formation of a b turn that brings the zinc ligands together.

The endopeptidases from

clan MB are also known as met-zincins, because there is a conserved Met (Met 145 in astacin) in a
turn that underlies the active site. This clan includes very important enzymes, such as the MMPs as
well as several bacterial MPRs that will be discussed shortly.

Serralysin, a single-chain MPR with a molecular mass of about 55 kDa isolated from Serratia

as well as some other bacteria (Pseudomonas or Erwinia chrysanthemi), shows a broad specificity
with a preference for small- to medium-sized and hydrophobic residues in P1

position (notably Gly

and Ala).

147

Serralysin is considered as one of the virulence factors produced during Serratia or

Pseudomonas infection, though its importance seems to be less than that of other toxins, such as for
instance aeruginolysin from P. aeruginosa.

148

Activity of this MPR is inhibited by EDTA, tetramethylenepentamine and 1,10-phenanthroline,

but can be regained by addition of divalent metal ions. The serralysins are not inhibited by the
classical phosphoramidons, which are thermolysin inhibitors.

148

The physiological function of

serralysins is not clear, but presumably, they play a role in nutrient digestion/uptake by the bacteria.
Thus, some potent and specific inhibitors of this enzyme would be beneficial both for better un-
derstanding the contribution of the protease as virulence factor, as well as for the development of
antibiotics against these pathogens.

Another protease, which has been isolated from various strains of P. aeruginosa such as IFO

3080, IFO 3455, T 30 or from 18 strains isolated from infected patients were at first called
Pseudomonas aeruginosa alkaline proteinases (because of their pH optimum, in the alkaline
domain), and later as aeruginolysin.

149

Aeruginolysin can cause a wide range of pathogenic effects

in hosts infected with P. aeruginosa, including tissue degradation, spreading of infection and
septicemia, inactivation of defense-oriented proteins including immunoglobulins, lysozyme, and
transferrin, being thus an important virulence factor.

149

Aeruginolysin is inhibited by metal ion

chelators such as EDTA and 1,10-phenanthroline as well as by peptidyl-mercaptoanilides such as
Bz-Phe-Arg-SH (K

¼ 7.5 mM), but is insensitive to phosphoramidon and Zincov, which inhibit

many metalloproteinases of clan MA.

149

No potent inhibitors of this protease were reported for

the moment.

Among the virulence components known to be expressed by the urinary tract pathogen Proteus

mirabilis is a MPR of 55 kDA, belonging to the M10 family, referred to as mirabilysin. This is an

350

SUPURAN, SCOZZAFAVA, AND CLARE

extracellular protease that can be isolated, when P. mirabilis is grown on media containing a suitable
substrate such as skim milk agar.

150

Thus, the protease produced during Proteus infections degrades

among other proteins urinary tract IgA. Recombinant mirabilysin also degrades human IgA1 and
IgG in a time-dependent manner resulting in complete digestion of the IgA1 substrate into numerous
smaller fragments.

150

Fragilysin, another metalloproteinase belonging to the metzincin clan has been isolated as the

Bacteroides fragilis enterotoxin.

151

Bacteroides fragilis is an anaerobic bacterium that is a part of

the normal flora found in the large intestines of humans and most other mammals at about the same
concentration as Escherichia coli, but similarly to E. coli, some strains of B. fragilis can produce
toxin. Such enterotoxigenic strains, were shown to produce a 20-kDa zinc metalloprotease toxin,
which has been associated with diarrheal disease in animals and young children.

151

Despite more

than a decade of research by several laboratories, the mechanism of action of the enterotoxin
remained unknown, until it was showed that the enterotoxin gene codes for a signature zinc-binding
motif belonging to the metzincin clan.

152

Fragilysin is cytopathic to intestinal epithelial cells and

induces fluid secretion and tissue damage in ligated intestinal loops. This is the first reported
example of a protease causing diarrhea. Fragilysin causes proteolytic degradation of the tight
junctions and basement membranes (the extracellular matrix) of tissue-cultured epithelial cells, and
this is probably the major mechanism by which the intestinal epithelium is altered in vivo.

153

Fragilysin also may activate host proteases and inactivate protease inhibitors, leading to exa-
cerbation of the destruction.

152

Fragilysin hydrolyzes a number of proteins in the extracellular

matrix of epithelial cells, which is probably how this enzyme causes its pathological effects in the
intestine.

151–153

It has been shown in tissue culture experiments that the enzyme contributes to

the invasion of intracellular bacterial species into epithelial cells by the proteolytic degradation of
the extracellular matrix. Since the intestinal tract has to be resistant to the action of proteases
in general, because it is constantly bathed in proteases, it is not at all clear how this protease can
have such an unusual effect on this tissue. Enzymatic activity of fragilysin was inhibited by metal
chelators (EDTA, DTPA) but not by inhibitors of other classes of proteases. Additionally, cytotoxic
activity of the enterotoxin on human carcinoma HT-29 cells was inhibited by acetoxymethyl ester
EDTA.

154

No potent/selective fragilysin inhibitors were reported, although such compounds would

be very desirable as potential antibiotics. Not very much is known regarding the inhibition of the
M12 family protease flavastacin, isolated from the pathogen Flavobacterium meningosepticum,
which cleaves peptide bonds of the type Xaa#Asp.

D. Bacterial Metallo-Exopeptidases

Metallo-carboxypeptidases were purified from Streptomyces griseus and Thermoactinomyces
vulgaris, a bacterium that shares a number of biochemical traits with bacilli.

155

Although the

physiological role of carboxypeptidase T (CPT)—the enzyme isolated from these bacteria—and
related bacterial enzymes remains to be elucidated, it seems plausible that these exopeptidases are
involved in degradation of protein substrates, which is consistent with their unusually broad
specificity.

155

The crystal structure of CPT from T. vulgaris has been reported, being shown that this

enzyme is a remote homologue of mammalian zinc-carboxypeptidases.

155

In spite of the low degree

of amino acid sequence identity, the three-dimensional structure of CPT is very similar to that of
pancreatic carboxypeptidases A and B, with the active site being located at the C-edge of the cen-
tral part of the beta sheet. Amino acid residues directly involved in catalysis and binding of the
C-terminal carboxyl of the substrate are strictly conserved in CPT, CPA, and CPB, suggesting that
the catalytic mechanism of the three carboxypeptidases is similar. Little is known regarding inhibi-
tion of CPT at this moment.

155

Enzymes similar with the mammalian leucyl aminopeptidase have been isolated in many

bacteria (Table II) and they are referred to as bacterial leucyl aminopeptidases (LAPs) or PepA

BACTERIAL PROTEASE INHIBITORS

351

aminopeptidases. The LAP of E. coli and S. typhimurium are the best characterized of such bacterial
enzymes and exhibit peptidase activities similar to their mammalian counterparts.

156

These are

enzymes with broad substrate specificity that catalyze the release of N-terminal amino acids from
peptides, especially leucyl and methionyl substrates.

156

Initially, unlike the zinc-dependent mam-

malian enzymes, the bacterial enzymes were considered to contain two Mn

ions, but it was later

shown that they contain two Zn

ions, similarly with the mammalian LAP.

156,157

The active sites

of PepA from E. coli and LAP from bovine lens are isostructural, in both structures, a bicarbonate
anion being bound to an arginine side chain (Arg-356 in PepA and Arg-336 in bovine lens LAP)
very near the two catalytic zinc ions.

157

Peptidase B (PepB) of Salmonella enterica serovar

Typhimurium is one of three broad-specificity aminopeptidases found in this organism, being a 427-
amino-acid (46.36-kDa) protein, which has been assigned to the LAP structural family.

158

The

Aeromonas proteolytica aminopeptidase (AMP), Pseudomonas sp. (RS-16) carboxypeptidase G2
(CPG2), and Streptomyces griseus aminopeptidase (SGAP) are other zinc dependent exopeptidases
with cocatalytic zinc ion centers and a conserved aminopeptidase fold, probably possessing a three-
dimensional structure similar to that of the prostate-specific membrane antigen (PSMA) and the
transferrin receptor (TfR).

159

In E. coli and S. typhimurium PepA functions as an aminopeptidase to hydrolyze exogenous

peptides as a source of amino acids; to promote protein turnover during starvation and probably to
degrade abnormal proteins.

156,157

Thus, inhibitors of such enzymes might be important for the

design of new antibiotics. Indeed, Huntington et al.

160

reported that peptide-derived thiols of the

general structure 11 are potent, slow-binding inhibitors of the aminopeptidase from Aeromonas
proteolytica, with K

values in the range from 2.5 to 57 nM (Fig. 6). To investigate the nature of the

interaction of these thiol-based inhibitors with the dinuclear active site of AAP, the electronic
absorption and EPR spectra of Co(II)Co(II)-, Co(II)Zn(II)-, and Zn(II)Co(II)-AAP in the presence
of the strongest binding inhibitor have been recorded,

160

being shown that both [CoZn(AAP)] and

[ZnCo(AAP)], in the presence of such inhibitors, exhibited an absorption band centered at 320 nm
characteristic of an S

! Co(II) ligand-metal charge-transfer band. It was shown that the inhibitor is

interacting with each active-site metal ion: practically the inhibitor interacts weakly with one of the
metal ions in the dinuclear site and that the crystallographically identified micro-OH(H) bridge,
which has been shown to mediate electronic interaction of the Co(II) ions, is likely broken upon
inhibitor binding. These data suggested that the thiolate moiety of the inhibitor may bind to either of
the metal ions in the dinuclear active site of AAP, but does not bridge the dinuclear cluster.

160

These

compounds constitute interesting lead molecules for developing more potent LAP inhibitors with
potential antibiotic properties.

The methionine aminopeptidases (MetAPs) represent a unique class of cobalt-containing PRs

that have been isolated in eukariotes as well a prokariotes, in which they are involved in the
nonprocessive cleavage of the initiator methionine of protein synthesis.

161

Although virtually all

protein synthesis is initiated with methionine, most proteins do not appear to retain it in their mature
forms, and thus, inhibitors of MetAP offer hope for stopping critical physiological processes,

Figure 6.

Structure of AAP inhibitors of the peptide-thiol type

11.

352

SUPURAN, SCOZZAFAVA, AND CLARE

leading to the development of therapies of cancer, and microbial/fungal infections.

161–163

Proteins

that are transported across membranes generally have an N-terminal signal peptide that is removed
by a specific peptidase, resulting in the loss of the initiator methionine along with several other
N-terminal residues. However, cytoplasmic proteins must rely on specific MetAPs to remove the
initial methionine, and this explains why these enzymes are present throughout the phyloge-
netic tree.

161

The high-resolution crystal structure of E. coli MetAP (eMetAP) has been solved by single

crystal X-ray diffraction.

164

The structure has been described as a ‘‘pita-bread’’ fold, where the two

Co(II) ions are sandwiched between two b sheets, which are surrounded by four a helices yielding a
structure with pseudo-2-fold symmetry.

161,164

Although this is a novel fold for a protease, the

structure is similar to that of Pseudomonas putida creatinase, which does not possess metal ions in
its active site.

161

The two metal ions of eMetAP are only 2.9A

˚ apart and are liganded by two aspartic

acids, two glutamic acids and one histidine, and a bridging OH, which is the nucleophile in the
proteolytic scission.

161–164

Since the removal of the N-terminal Met is a critical step in the maturation of many proteins,

MetAP are critical enzymes required for the proper subcellular localization and eventual degra-
dation of proteins; deletion or inhibition of MetAP activity in many bacteria and yeasts leads to the
death of these organisms.

161

Thus, inhibitors of these enzymes started to be developed in order to

assess them for obtaining pharmacological agents. The natural products fumagillin 12 and ovalicin
13

as well as the synthetic derivative TNP-470 14 possess antibacteriophagic, amoebicidal, and anti-

angiogenesis properties, which are correlated with their inhibitory properties against MetAPs
from different organisms (Fig. 7).

161

Another potent inhibitor of eMetAP (and related bacterial

enzymes) is a bestatin-based compound of type 15, which binds to the enzyme in the substrate-like
manner, but it is not hydrolyzable (IC

against the E. coli enzyme of 5 mM). One must mention

that these compounds inhibit both bacterial as well as eukariotic MetAPs, and that in order to
design novel antibiotics based on such lead molecules, inhibitors selective for the bacterial enzymes
are needed.

A dinuclear, manganese-containing proline-specific aminopeptidase, relatively similar to

eMetAP discussed above has recently been isolated in E. coli and characterized by X-ray crys-
tallography.

165

The dipeptide Pro-Leu is a competitive inhibitor of this enzyme,

165

but no stronger

Figure 7.

Methionyl aminopeptidase inhibitors

12–15.

BACTERIAL PROTEASE INHIBITORS

353

inhibitors for its activity were reported up to now, nor is it known whether its inhibition may have
critical consequences for the pathogenic bacteria.

E. Proteases of the M19, M20, M22, M23, and M26 Families

Proteases able to release the C-terminal glutamate residues from a wide range of N-acylating
moieties, including peptidyl, aminoacyl, benzoyl, benzyloxycarbonyl, folyl, and pteroyl deriva-
tives are denominated glutamate carboxypeptidases (GCPs), and such enzymes were isolated
only in prokariotes (Table II).

166

On the other hand, enzymes with similar activity, that hydrolyze

the g-glutamyl tail of antifolate and folate polyglutamates (named glutamyl hydrolases) were
isolated in higher vertebrates,

167

but both the prokariotic as well as the eukariotic such enzymes

have important pharmacological applications in the treatment of cancer. The prime interest in such
enzymes was due to their ability to cleave the glutamate residue from folate polyglutamates 16 and
more significantly folate analogs, such as the chemotherapeutic agent methotrexate 17 an inhibitor
of dihydrofolate reductase (Fig. 8).

166,167

This property provided the opportunity to assess these

enzymes as antitumor agents through depletion of reduced folates, essential cofactors in DNA
synthesis and as rescue agents against methotrexate toxicity.

166

Progress was limited through

availability of the enzyme until isolation of carboxypeptidase G from Pseudomonas sp. strain
RS-16.

166

Bacterial glutamate carboxypeptidases are characterized as zinc(II)-requiring exopepti-

dases with specificity for glutamate, whereas g-glutamyl hydrolase (EC 3.4.19.9) is a cysteine
protease and glutamate carboxypeptidase II a metallo-protease belonging to the M28 family (but
since these are mammalian enzymes will be not dealt with here).

166,167

Carboxypeptidase G2 is a dimeric protein of 83.6 kDa containing two Zn(II) ions per subunit,

and there are no disulfide bonds in its molecule.

166

In both native and recombinant forms, the

enzyme is located in the periplasmic space, being targeted by a 22 amino acid signal peptide.
Carboxypeptidase G2 has been crystallized and its structure determined at 2.5 A

˚ resolution.

168

Each

subunit of the molecular dimer consists of a large catalytic domain containing two Zn(II) ions at the
active site, and a separate smaller domain, which forms the dimer interface. The two active sites in
the dimer are more than 60 A

˚ apart and are presumably independent; each contains a symmetric

distribution of carboxylate and histidine ligands around the dimetalic center, the two zinc ions being
at a distance of 3.2 A

˚ .

168

Figure 8.

Structure of folate pentaglutamate

16 and methotrexate 17.

354

SUPURAN, SCOZZAFAVA, AND CLARE

The enzyme has been developed for clinical use in two settings: (i) as a rescue agent during high

dose (3–30 g

) methotrexate therapy, against a range of cancers.

166

The principle of ‘‘rescue’’

is well established using leucovorin (5-formyltetrahydrofolate) to rapidly restore cellular levels
of reduced folates depleted by the inhibitory action of methotrexate on dihydrofolate reductase.
Removal of methotrexate then relies on patient hydration and renal function. If the latter is impaired
then high circulating levels of methotrexate can result in bone marrow and other toxicities. GCPs
can thus be used to rapidly eliminate methotrexate from the circulation. Rescue protocols have been
extended to include intrathecal, as opposed to systemic, methotrexate therapy; (ii) by using targeted
carboxypeptidase G2 to remove the glutamate residue from prodrugs, releasing a highly cytotoxic
agent at tumor sites.

166

This method has been given the acronym ADEPT (antibody directed enzyme

prodrug therapy). Targeting is achieved by covalent linkage of carboxypeptidase G2 to tumor-
associated antibodies and monoclonal antibody fragments, including antihuman chorionic gona-
dotrophin, antihuman carcinoembryonic antigen, or antihuman c-erbB2 protooncogene product
among others.

166,169

Once delivered to the tumor site, the enzyme is used to activate prodrugs,

predominantly based on glutamyl derivatives of nitrogen mustard compounds such as 4-[(2-
mesyloxyethyl)(2-chloroethyl)amino] benzoyl glutamic acid and 4-[N,N,-bis(2-iodoethyl)amino]-
phenol linked to glutamic acid.

166

Clinical studies emerging by using such approaches are very

promising.

166

There are at least 17 families of MPRs that could not yet been assigned to clans.

Of these, ten

possess the HEXXH motif that includes two zinc ligands and the catalytic Glu that is found in clan
MA and clan MB, whereas for seven families the HEXXH motif has not been determined, nor any
motifs similar to those in any clans of metallopeptidase.

The bacterial enzymes that will be

discussed in the end of this review all belong to these 17 families of less characterized proteases.

The zinc-containing

-alanyl-

-alanine (

-Ala-

-Ala) dipeptidase VanX has been detected in

both Gram-positive and Gram-negative bacteria, being associated with resistance to the glyco-
peptide antibiotic vancomycin.

170–172

It appears that this enzyme has at least three distinct physio-

logical roles: (i) in pathogenic vancomycin-resistant enterococci (such as for instance Enterococcus
faecium), vanX is part of a five-gene cluster that is switched on to reprogram cell-wall biosynthesis
to produce peptidoglycan chain precursors terminating in the depsipeptide

-Ala-

-lactate rather

than the dipeptide

-Ala-

-Ala. This modified peptidoglycan exhibits a 1,000-fold decrease in

affinity for vancomycin (accounting for the observed phenotypic resistance) due to the loss of
one hydrogen bond between the antibiotic molecule and the amide NH of the dipeptide (now re-
place by the ester bond of the depsipeptide);

171

(ii) in the glycopeptide antibiotic producers, such

as Streptomyces toyocaensis (which secretes a vancomycin-like antibiotic) and Amylocatopsis
orientalis, a vanHAX operon may have co-evolved with antibiotic biosynthesis genes in order to
provide immunity by reprogramming cell-wall termini to

-Ala-

-lactate, as antibiotic biosynthesis

proceeds and the toxic compound accumulates; (iii) in the Gram-negative bacterium E. coli, which
is never challenged by the glycopeptide antibiotics, because such molecules are unable to penetrate
the outer membrane permeability barrier, a vanX homologue (named ddpX) is cotranscribed
with a putative dipeptide transport system (ddpABCDF) in stationary phase by the transcription
factor RpoS (s

171

The combined action of DdpX and the permease would permit hydrolysis of

-Ala-

-Ala transported back into the cytoplasm from the periplasm as cell-wall crosslinks are

refashioned, and

-Ala resulting in hydrolysis could then be oxidized as an energy source for cell

survival under starvation conditions.

171

VanX is a 202 amino acid polypeptide (23.5 kDa) existing in solution as a homodimer, and

it contains 1.0 mol of bound Zn(II) per mol of polypeptide. It is a

-dipeptidase, catalyzing the

hydrolysis of the bacterial cell wall biosynthesis intermediate

-Ala-

-Ala into 2 molar equi-

valents of

-alanine; as mentioned above, VanX does not catalyze the hydrolysis of the depsipeptide

-Ala-

-lactate. The exact mechanism by which VanX preferentially hydrolyzes the amide subs-

trate versus its kinetically and thermodynamically more favorable ester analog has not yet been

BACTERIAL PROTEASE INHIBITORS

355

determined and is the subject of current research efforts. VanX is resistant to the action of b-lactam
antibiotics, but is inhibited by transition-state substrate analogs or by chelating agents. Dithiols
such as DTT and dithioerythreitol are also potent time-dependent micromolar inhibitors whose
mechanism of inactivation is postulated to occur by the formation of an enzyme-metal-dithiol
ternary complex.

170–172

The crystal structure of the E. faecius enzyme has recently been determined: the Zn(II) ligands

are His 116, Asp 123, and His 184, whereas Glu 181 hydrogen-bonds the zinc-coordinated water
molecule.

173

The most important finding of this study was that the active site is small and constricted

(of around 150 A

), which may make rational design of VanX inhibitors a significant medicinal

chemistry challenge.

172,173

Unfortunately, potent, nanomolar inhibitors of VanX have not been

reported for the moment, although such compounds might have an important clinical impact for the
treatment of infections resistant to vancomycin.

Another enzyme potentially important for the development of non-classical antibiotics is

O-sialoglycoprotein endopeptidase, abbreviated as glycoprotease, which derives its name from
its unique specificity for the proteolytic cleavage of glycoproteins. To date, all known substrates
of this enzyme are glycoproteins rich in sialoglycans O-linked to threonine or serine residues.

174

The enzyme was discovered in Pasteurella haemolytica A1 (an organism associated with fibrinous
pneumonia in cattle) due to its ability to hydrolyze human erythrocyte glycophorin A.

174

The high

specificity of the P. haemolytica glycoprotease for cell surface O-sialoglycoproteins has made the
enzyme a useful tool in cell surface antigen studies, since the lack of action of the glycoprotease
against protein substrates, which do not bear O-linked glycans means that cell surface molecules
and their roles can be delineated by its use.

174

However, there is as yet no clear indication of

the role of the enzyme in the pathogenesis of fibrinous pneumonia in cattle, although animals
vaccinated with P. haemolytica culture proteins supplemented with the recombinant glyco-
protease fusion protein rGcp-F show enhanced protection against experimental challenge with
live pathogen.

174

The soil bacterium Lysobacter enzymogenes has a remarkable ability to lyse other bacteria

and some soil nematodes, a property which allowed the isolation of a protease named b-lytic
protease, since this enzyme has the ability to cause lysis of bacterial cell walls and differentiates it
from a-lytic protease. This metalloendopeptidase lyses Arthrobacter globiformis, Micrococcus
luteus as well as Staphylococcus aureus cells, inhibiting thus the growth of sensitive organisms and
potentially serving as an antimicrobial agent.

175

This 19.1 kDa protein (containing 178 amino acid

residues) incorporates two disulfide bonds between residues 65–111 and 155–168 and one zinc ion
per mole, that is essential for activity. It does not contain the consensus zinc-binding site HEXXH,
but has a conserved HXH sequence that is probably implicated in zinc binding.

175

Staphylolysin is the new name suggested for Pseudomonas aeruginosa LasA endopeptidase

that causes lysis of Staphylococcus aureus cells. Recent studies show that staphylolysin is a secreted
endopeptidase that can slowly degrade insoluble elastin, rendering it a better substrate for pseudo-
lysin and other endopeptidases.

176

Staphylolysin cleaves peptide bonds within the pentaglycine

cross-linking peptides of S. aureus peptidoglycan leading thus to cell lysis. It also lyses cell walls of
Micrococcus radiodurans and Gaffkia tetragena, which contain di- and tri-glycine sequences in
their interpeptides, respectively.

176

Staphylolysin is a 19.965 kDa protein, consisting of 182 resi-

dues; its amino acid sequence is

40% identical to those of Achromobacter lyticus and Lysobacter

enzymogenes

b-lytic endopeptidases.

176

Staphylolysin contains four conserved Cys residues (that,

by analogy to the Lysobacter enzyme, are likely to be engaged in disulfide bonds) and one zinc ion.
Although it does not possess the classical zinc-binding motif (HEXXH), a conserved HXH sequence
has been evidenced to be the potential zinc-binding site (where X

¼ Leu 121).

176

Staphylolysin appears to play a role in pathogenesis of several infections, including corneal

infections: in a mouse model of P. aeruginosa corneal infection, a defined lasA mutant causes mild
to no disease following infection; application of 5 mg of purified staphylolysin to scarified mouse

356

SUPURAN, SCOZZAFAVA, AND CLARE

corneas produced an acute toxic reaction.

176

In experimental models of lung infections, bacterial

production of staphylolysin increases the virulence of P. aeruginosa Staphylolysin inhibits the
growth of S. aureus and thus may play a role in acquiring certain niches, such as the cystic fibrosis
lung environment. Its ability to stimulate elastin degradation by other endopeptidases, as demons-
trated in vitro suggests that it may contribute to the tissue destruction associated with P. aeruginosa
infections.

176

No inhibitors of this enzyme were reported, but the measurement of antibodies against

Staphylococcus aureus staphylolysin was used to try to discriminate between complicated and
uncomplicated S. aureus septicaemia.

177

IgA proteinases were recognized by identification of unusual fragments of immunoglobulin A

(IgA) in cell-free fluids of the human digestive tract and since this cleaving activity was also found
in normal human saliva, it appeared to be of bacterial origin, not from the pancreas or intestine. This
was confirmed by finding an IgA-cleaving activity in culture filtrates of many bacteria colonizing or
infecting human beings.

178

IgA proteinases are a group of endopeptidases produced by patho-

genically important bacteria belonging to the genera Streptococcus, Neisseria, Haemophilus,
Ureaplasma, Clostridium, Capnocytophaga, and Prevotella.

178

All these proteinases cleave the

glycosylated hinge region of human IgA, a flexible, peptide stretch that lies within the heavy chain,
between the F

and F

domains.

178

An analysis of 13 immunoglobulin A1 (IgA1) protease genes

(iga) of strains of Streptococcus pneumoniae, Streptococcus oralis, Streptococcus mitis, and
Streptococcus sanguis showed all of them to encode proteins with molecular masses of approxi-
mately 200 kDa, containing the sequence motif HEMTH and an aspartate residue 20 amino acids
downstream, which are characteristic of Zn MPRs.

179

In addition, all had a typical gram-positive

cell wall anchor motif, LPNTG, which, in contrast to such motifs in other known streptococcal and
staphylococcal proteins, was located in their N-terminal parts.

179

The independent evolution of several distinct classes of enzymes with a similar biological

function points to the importance of inactivation of IgA1 for colonization by bacterial mucosal
pathogens (see section 2.B for the serine protease family of enzymes that inactivate IgA). Yet no
specific role for any IgA proteinase in infection has been proven for the moment. IgA proteinase
activity is, in principle, capable of interfering with most functions of IgA, because cleavage
separates the antigen-binding F

from the effector F

domains of the molecule, and this, not only

facilitates bacterial evasion of immunity by establishing an immunodeficiency towards the micro-
organism, but the secretory IgA cleavage products themselves may contribute to immune
evasion.

178

No inhibitors active in the submillimolar range have been successfully made for these

enzymes. Attempts to synthesize small, substrate-based inhibitors for the metallo-type IgA pro-
teinases have been unsuccessful, while millimolar inhibitors have been made for the serine-type
enzymes (see section 2.B).

F. Tetanus and Botulinum Neurotoxins

Clostridium tetani and Clostridium botulinum (but also C. barati and C. butirycum) are strictly
anaerobic pathogens that provoke tetanus and botulism, respectively.

111,112

The first disease has

already been described by Hippocrates of Kos in the fourth century BC, being characterized by an
often fatal spastic paralysis with contraction of skeletal muscles that work one against the other;
botulism on the other hand is a neurologic syndrome of vertebrates, characterized by the loss of
function of peripheral cholinergic synapses.

111,112

This functional loss may be highly variable, from

minimal unnoticed effects up to a generalized flaccid paralysis. Such a manifestation accounts for
the fact that botulism was first described only at the beginning of the nineteenth century and infant
botulism was identified only 20 years ago.

111,112

Both diseases are caused entirely by clostridial

neurotoxins, the tetanus neurotoxin (TeNT) and botulinum neurotoxin (BoNT, of which seven
distinct serotypes are known, BoNT/A-G), respectively. These toxins constitute a group of bacterial
metalloprotease with unique properties activity.

111,112,180,181,182

BACTERIAL PROTEASE INHIBITORS

357

BoNT and TeNT are translated as single chain pro-toxins, which are subsequently cleaved

within a disulphide loop situated about one-third from the amino-terminus, generating the fully
active di-chain 150 kDa neurotoxin.

111,112

This is composed of a light (L) chain of about 50 kDa and

a heavy (H) chain of 100 kDa linked by an interchain disulphide bridge and by non-covalent
interactions.

111,112,182

The protease activity is associated to the L-chain, whereas the heavy chain is

important for the membrane translocation of the L-chain into the neuronal cytosol, playing a critical
role in pore formation.

111,112

The carboxyterminal part of the H-chain may also be subdivided in two

subdomains: a lectin-like domain and a b-trefoil fold, and is probably involved in the toxin binding
to presynaptic membranes via a double interaction, most likely with two different molecules
of the nerve terminal.

111,112

The zinc endopeptidase activity of the L-chain is inhibited when

the toxin is intact, being expressed only after reduction of the disulphide bond bridging the L- and
H-chains.

111,112

One Zn(II) ion is present bound to the L-chain, which is critical for the protease

activity of these neurotoxins.

111,180

The Zn(II) of these MPRs is coordinated by two histi-

dine residues (TeNT numbering: His 222 and His 226, belonging to the consensus zinc-binding site
HEXXH), by the carboxylate group of Glu 261, in addition to the water molecule critical for
catalysis.

111,112

As for many other MPRs discussed here, the glutamate residue (Glu 223 in this case)

of the consensus zinc-binding sequence mentioned above, is also critical in catalysis, as it parti-
cipates in hydrogen bonding and activation of the zinc-bound water molecule.

111,112

The catalytic

domain of the BoNT/A contains both a-helix and b-strand secondary structures, having little
similarity with other MPRs, except for the zinc binding motif.

182

The protease active site of these

enzymes is quite deep, accommodating at least 16 amino acid residues, which makes them among
the most ‘‘bulky’’ metallo-proteases studied so far.

111,182

This may have very important con-

sequences for the successful design of active-site directed inhibitors of the protease activity of
BoNT/TeNT.

The only known proteolytic substrate of TeNT is a 120-residue protein anchored to the mem-

brane of cell vesicles, which has been termed VAMP (vesicle-associated membrane protein) but also
synaptobrevin and cellubrevin.

111,112

Similarly, BoNTs/B, D, F, and G cleave specifically VAMP at

different single peptide bonds, as shown in Table II.

111,112,181

The three VAMP isoforms, VAMP-1,

VAMP-2, and cellubrevin, of human and mice are cleaved at the same site by BoNT/B and TeNT,
this site being Gln76#Phe77.

111,112,180,181

VAMPs of other species may carry mutations at the

cleavage/recognition site(s) that render them neurotoxin insensitive, as is the case of the rat and
chicken VAMP-1, which are not cleaved by BoNT/B.

180

BoNTs also possess two other intracellular

targets in addition to VAMP: SNAP-25 (25 kDa synaptosomal-associated protein) and syntaxin,
both of which are cleaved between Gln#Arg (by BoNT/A), Lys#Ala (by BoNT/C), Arg#Ala
(by BoNT/C), or Gln#Lys (different neurotoxins) residues.

111,112

One must mention that syntaxin is

a 35 kDa type II membrane protein located in the neuronal plasmalemma, being important for
neuronal development and survival, possibly modulating calcium entry within the neurons.

111

SNAP-25 is a palmitoylated protein of the CNS, highly conserved in the evolutionarily tree (from
yeasts to humans), being also involved in the calcium-dependent phase of neurotransmitter release
at exocytotic sites.

111,112

Tetanus is acquired by contamination of wounds with spores of toxigenic strains of Clostridium

tetani, which are ubiquitous, but highly enriched in faeces from many animals including farmyard
animals. Under anaerobic conditions, spores germinate, and tetanus neurotoxin is produced and
released upon autolysis. The toxin spreads in the body and binds, to an as yet unidentified receptor of
the presynaptic terminal of the neuromuscular junction and after endocytosis, the toxin migrates
retroaxonally inside the motor neuron and reaches the spinal cord. The toxin is then released in the
intersynaptic space and it enters small synaptic vesicles, when they expose their lumen to the outside
following fusion with the presynaptic membrane and release of neurotransmitter.

111,112

Synaptic

vesicles are then endocytosed and their lumen is acidified by a vacuolar ATPase proton pump to
drive the reuptake of neurotransmitter. Such lumenal acidification is essential for intoxication to

358

SUPURAN, SCOZZAFAVA, AND CLARE

occur because tetanus neurotoxin at low pH changes conformation and becomes hydrophobic.

111

In such a form, the toxin inserts in the lipid bilayer of the vesicle membrane and somehow the
H-domain manages to translocate the L-chain into the cytosol, with the consequent cleavage of
VAMP.

111,112

Such specific proteolysis is sufficient to cause a prolonged blockade of neuro-

exocytosis, and this demonstrates the essential role of VAMP in neuroexocytosis.

111,112

Clostridial spores harboring the genes encoding botulinum neurotoxins are widespread in the

environment and may germinate in anaerobic foods, and produce and release the neurotoxins.
Eating poisoned food causes botulism, following transcytosis of botulinum neurotoxin across the
intestinal epithelial layer and its spread through the body. Botulinum neurotoxins bind the un-
myelinated presynaptic membrane of motor neurons and of other cholinergic synapses. This results
in a flaccid paralysis as well as in alterations of the autonomic nervous system.

111,112

Thus, membrane permeant inhibitors of the protease activity of these enzymes would be helpful

therapeutic agents for the treatment of tetanus and botulism, as at the present time, there is no
effective drug therapy to prevent the progressive evolution of tetanus or botulism following intoxi-
cation or infection.

183

Only recently the first compounds that potently inhibit these zinc-proteases

have been reported.

183,184

One must mention that similarly to many other metallo-proteases, BoNT,

and TeNT are inhibited by chelating agents such as EDTA, 1,10-phenantroline or captopril, gene-
rally in millimolar concentrations, and thus, such compounds do not have practical applications for
the treatment of diseases provoked by these neurotoxins.

111,112

Thus, Martin et al.

183,184

recently showed that various b-aminothiols of type 18–24 act as

micromolar inhibitors (affinity in the 35–250 mM range) against the protease activity of the TeNT
light chain (Fig. 9). The compounds incorporating the aromatic ring, and mainly those containing
the free sulfonamide group were the best inhibitors. This is an extremely interesting result, since
carbonic anhydrase (CA, EC 4.2.1.1), another zinc enzyme possessing CO

hydrase and esterase

(but not protease) activities is strongly inhibited by aliphatic/aromatic sulfonamides, with the in-
hibitor coordinated to the metal ion by means of the ionized sulfonamide nitrogen.

185

The fact that

in the entire series of derivatives reported by the French group,

183,184

the most active were those

containing the SO

moiety, might signify that similarly to the case of inhibiting CA, such

sulfonamides may bind to the Zn(II) ion within the metallo-protease active site (alone, or in chelate
form, together with the thiolate sulphur atom of the amino-thiol molecule). The same group reported
that dipeptidyl-based amino-thiols incorporating aliphatic/aromatic sulfonamide moieties, of types
25–28

show good inhibitory properties against both TeNT as well as BoNT/B.

184

This is the first

example of potent synthetic inhibitors for these metallo-proteases, and open very interesting
potential applications of such protease inhibitors for the treatment of botulism/tetanus (Fig. 10).

The high specificity of the action of BoNT provided the basis for its clinical use the therapy of

a variety of human diseases caused by hyperfunction of cholinergic terminals.

111,186–188

Thus,

Figure 9.

TeNT light chain protease inhibitors of types

18–24 and their inhibition constants (K

values^in

mM in brackets)

(see Ref.183).

BACTERIAL PROTEASE INHIBITORS

359

injection of minute amounts of neurotoxin into the muscle(s) to be paralyzed leads to a depression of
the symptoms for a period of several months, and this is the best available treatment for dystonias,
strabismus,

186

facial wrinkling, brow position,

187

as well as palmar and axillary hyperhidrosis.

188

Mainly BoNT/A is used for such treatments, but some other types of modified toxins are also in-
vestigated clinically.

112

G. Anthrax Toxin Lethal Factor, Lysostaphin, and Aureolysin

Bacillus anthracis secretes three proteins, which associate in binary combinations to form toxic
complexes at the surface of mammalian cells: receptor-bound protective antigen (PA) is pro-
teolytically activated, yielding a 63 kDa fragment, which oligomerizes into heptamers, which bind
oedema factor (EF) or lethal factor (LF) to form the toxic complexes.

189

Anthrax toxin lethal factor

is one of the three components that are collectively termed anthrax toxin. The combination of PA
(82.7 kDa) and LF (90.2 kDa) is designated lethal toxin because it kills certain animals and rapidly
lyses macrophages. The combination of PA and EF (88.8 kDa), termed oedema toxin, inhibits
phagocytes by increasing cAMP concentrations to unphysiologic levels.

189,190

Both LF and EF enter

cells by binding to furin-activated, receptor-bound protective antigen.

189,190

Edema factor is a

calmodulin-dependent adenylate cyclase. LF is a metalloprotease which cleaves specifically
isoforms 1, 2, and 3 of MAPK kinases within their N-terminal tail thus interfering in a major
pathway of signaling from the plasma membrane to the cell nucleus.

191,192,193

Indeed, macrophages,

which are killed by PA

þ LF, are protected by certain hydrophobic peptides such as Leu-NH

(EC

1 mM), Phe-NH

(EC

0.2 mM), Leu-CH

Cl (EC

0.1 mM), and bestatin (EC

0.2 mM).

194

Staphylococcus simulans secretes lysostaphin, a protease that hydrolyzes the peptidoglycan

of all staphylococci that synthesize peptidoglycans incorporating the pentaglycine cross-bridges

Figure 10.

BoNT/B and TeNT inhibitors of type

25–28 (see Ref.184).

360

SUPURAN, SCOZZAFAVA, AND CLARE

(see also the sortase at section 3.C).

107,195

Lysostaphin is synthesized as a preproenzyme of 42 kDa

with a signal peptide that is cleaved intracellularly and an amino-terminal propeptide removed
extracellularly to yield the mature 27 kDa enzyme, which contains one zinc ion, whose coordination
has not been elucidated for the moment.

195

Lysostaphin has been used as an antistaphylococcal

agent, as a reagent for typing staphylococci, and also to lyse staphylococcal cell walls for the
liberation of intracellular enzymes, nucleic acids, cell membrane, and surface components, which is
widely employed for research purposes.

195

Since elastolytic activities of microbial pathogens are

generally associated with virulence, lysostaphin or a lysostaphin-like enzyme may play an impor-
tant role in staphylococcal disease pathogenesis, but no potent inhibitors of this zinc protease have
been reported up to now.

The extracellular metalloproteinase from the pathogenic bacterium Staphylococcus aureus

denominated aureolysin has only recently been characterized as belonging to the thermolysin
family.

196

The enzyme binds one Zn(II) and three Ca(II) ions, comprises one chain of 301 amino

acid residues and the catalytic metal ion is coordinated by His 144, His 148, Glu 168, and a water
molecule, whereas Glu 145 is liganded to the zinc-bound water molecule by means of two hydrogen
bonds.

196

Aureolysin involvement in the pathophysiology of S. aureus diseases is unknown and the

enzyme is not considered as a virulence factor for the moment. In spite of this, there is some
indication that this metalloproteinase can indirectly modulate an inflammatory reaction. The en-
zyme is responsible for the pseudocoagulase activity of coagulase-negative S. aureus strains,
apparently through direct proteolytic activation of prothrombin in human plasma.

197

The recently

reported structure should facilitate anyhow the synthesis of potent, active-site directed inhibitors,
which are not available for the moment.

H. AAA Proteases

AAA proteases (ATP-ases associated with a variety of cellular activities) represent a conserved class
of ATP-dependent zinc-proteases that mediate the degradation of membrane-proteins in bacteria,
mitochondria, and chloroplasts.

198,199

These proteins combine proteolytic and chaperone-like

activities, forming a membrane-integrated ‘‘quality-control’’ system, and their inactivation/inhibi-
tion provokes severe defects in various organisms.

198,199

They are present probably in all bacteria,

being some of the few proteases essential for viability of the microorganisms.

198,199

Such proteins

were first detected in E. coli, being denominated FtsH.

199

FtsH is an ATP- and Zn

-dependent

metalloprotease with a molecular mass of about 70 kDa, being anchored to the cytoplasmic
membrane via two transmembrane regions in such a way that the very short amino- and the long
carboxy-termini are exposed into the cytoplasm.

199

A 200 amino acid residues long module contains

the ATP-binding site, whereas the second transmembrane segment, which neighbors the C-terminal
cytoplasmic region of FtsH, participates in not only its membrane-anchoring, but also in its protease
activity (ascribed to an HEXXH zinc-binding motif) and homo-oligomerization.

200

The specificity

of peptide bond cleavage reaction by this protease has not been elucidated, but it was shown that
FtsH cleaves the C-terminal side of hydrophobic residues and produces a characteristic set of small
peptides ( < 30 kDa) without releasing a large intermediate.

201

Thus, FtsH recognizes the unfolded

structure of proteins and progressively digests them at the expense of ATP.

198–201

In the absence of

substrate proteins, FtsH hydrolyzes ATP with an apparent K

value for ATP of about 80 mM.

198–201

Proteolytic activity of FtsH is inhibited by chelators such as 1,10-phenanthroline and EDTA, as well
as by vanadate, whereas it is insensitive to N-ethylmaleimide, azide, KNO

, and PMSF.

198–201

In E. coli (but presumably in many other bacteria), FtsH (or its homologs) forms a complex with

a two periplasmically exposed membrane proteins, HflK and HflC. FtsH is required for proteolytic
degradation of some unstable proteins, that include both soluble regulatory proteins such as sigma-
32 (heat-shock sigma factor) and phage lambda CII (transcriptional activator), as well as membrane
proteins, including uncomplexed forms of SecY (that forms the translocon together with SecE and

BACTERIAL PROTEASE INHIBITORS

361

SecG) and the a subunit of the F0 complex of the H

-ATPase.

199

Its activity can be modulated by

the HflKC proteins, by another membrane protein designated YccA, which can transiently associate
with both the FtsH and the HflKC proteins, or by small peptides such as CIII encoded by phage
lambda (involved in lysogenization) or SpoVM (needed for sporulation) encoded by Bacillus
subtilis.

195

Besides being a protease, there is circumstantial evidence that FtsH also acts as a mole-

cular chaperone, being designated as a ‘‘charonin’’.

199

Since AAA proteases are essential for the viability of the bacterial cells, developing inhibitors

targeted against them would lead to a new class of efficient antibiotics. Still, such studies have not
been performed for the moment, as this class of proteases has only recently been studied in some
detail, but they are very promising targets for the drug design of antibiotics of the future.

5 . C O N C L U S I O N S

The bacterial proteases and their inhibition constitute important, emerging research fields both for
the drug design of novel therapeutic agents that would avoid drug resistance problems typical of
classical antibiotics as well as for the applications of such enzymes per se in therapy. Important
advances have been registered for the moment only in the second field mentioned above, with the
wide use of clostridial collagenases (and sometimes also Vibrio collagenases) for the debridement
of wounds, removal of necrotic tissues from burns, ulcerations, decubitus ulcers, pressure sores,
venous leg ulcers, and diabetic ulcers among others. In all such conditions, these difficultly to heal
wounds show dramatic improvements when treated with these bacterial proteases, by a mechanism
not totally understood for the moment. These and other metalloproteases (such as BoNT, TeNT) are
also extensively used for the preparation of different vaccines. BoNT/A has important other ap-
plications, mainly in dermatology, for the treatment of dystonias, facial wrinkling, brow position,
palmar and axillary hyperhidrosis as well as for the non-surgical correction of strabismus. A method
for treating mucus hypersecretion (and thus chronic obstructive lung disease, asthma, and other such
debilitating conditions) by inhibition of exocytosis in mucus secreting cells or neurones, mainly by
BoNT/A, has also been recently reported, showing that the clostridial metalloproteases are indeed
versatile therapeutic agents.

Much less progress has been registered on the other hand in the first of the above-mentioned

fields, i.e., in the design of novel antibiotics of the bacterial protease inhibitor type. Although in the
case of several viral proteases, much success has recently been reported, mainly for the design of
inhibitors of HIV proteases, in the case of the bacterial proteases—which as can be seen from this
review are widely distributed in pathogenic bacteria—this approach has rarely been taken into
consideration by the main pharmaceutical companies or academic research. Thus, all antibiotics
available at the moment possess the same mechanism of action, intervening at different points in the
bacterial cell wall biosynthesis, and this explains the wide-spread cross-resistance to these classical
antibiotics, that constantly emerged during the last decades. Design of bacterial protease inhibitors
would thus constitute one of the best alternatives to classical antibiotics, since it was firmly esta-
blished that these proteases constitute major pathogenic factors in the constant ‘‘battle’’ between the
pathogens and the host’s immune system. Unfortunately for us, bacteria are able to evade our
defenses by a variety of mechanisms, many of which involve just the proteases they secrete. As seen
throughout this review, for the largest majority of the discussed bacterial proteases, few potent,
selective inhibitors have been reported up to now, with several exceptions such as the signal
peptidase inhibitors with affinity in the micromolar range for the enzyme, some micromolar clos-
tripain inhibitors, the potent (nanomolar) class of sulfonylated amino acid hydroxamate inhibitors of
Clostridium histolyticum collagenase, some relatively potent methionyl aminopeptidase inhibitors
as well as the recently reported micromolar inhibitors of the clostridial neurotoxins (TeNT and
BoNT), but not one of these compounds arrived in clinical trials as far as we know. One may anyhow

362

SUPURAN, SCOZZAFAVA, AND CLARE

see that this accounts for less than 10 enzymes in the list of at least 55 taken into consideration in this
review!

It must also be stressed that recently some ubiquitous bacterial proteases have been described in

some detail (such as for instance DegP in gram-negative bacteria, or sortase in the gram-positive
ones, or the AAA proteases present in all types of bacteria, and essential for their viability) and their
function in the bacterial cell cycle started to be understod in some depth. These facts, together with
the hope that X-ray (or NMR) 3-D structure should be soon available for some of these proteins
allow a certain basis of optimism that this knowledge may be exploited for the drug design of new
generation of antimicrobials. Such compounds have chances to be effective antimicrobials against a
large number of bacteria, since the above-mentioned proteases are ubiquitous. On the other hand,
the development of specific such agents for a given protease (for instance specific inhibitors of the
clostridial neurotoxins, or of the anthrax lethal factor, etc.) would constitute small, but certainly
important therapeutic advances for the treatment of these conditions. In fact many diseases caused
by several of the pathogens discussed in this review (and due to the attack of some of their proteases)
have no efficacious drug treatment at this moment. Thus, bacterial proteases may indeed be con-
sidered as some of the most important future targets for the drug design.

A C K N O W L E D G M E N T

Research from the author’s laboratory was financed by the Italian CNR—Target Project
Biotechnology.

R E F E R E N C E S

1. Barrett AJ, Rawlings ND, Woessner JF, Jr. (Eds.). Handbook of proteolytic enzymes (CD-ROM).

London: Academic Press, 1998, Chapters 1–569.

2. Leung D, Abbenante G, Fairlie DP. Protease inhibitors: current status and future prospects. J Med Chem

2000;43:305–341.

3. Ogden RC, Flexner CW (Eds.). Protease Inhibitors in AIDS therapy. New York: Marcel Dekker, 2001,

310 p.

4. Wlodawer A, Gustchina A. Structural and biochemical studies of retroviral proteases. Biochim Biophys

Acta 2000;1477:16–34.

5. Travis J, Potempa J. Bacterial proteinases as targets for the development of second generation antibiotics.

Biochim Biophys Acta 2000;1477:35–50.

6. Miyoshi SI, Shinoda S. Bacterial metalloproteases as the toxic factor in infection. J Toxicol Toxin Rev

1997;16:177–194.

7. Maeda H. Role of microbial proteases in pathogenesis. Microbiol Immunol 1996;40:685–699.
8. Rice SA, Givskov M, Steinberg P, Kjelleberg S. Bacterial signals and antagonists: the interaction between

bacteria and higher organisms. J Mol Microbiol Biotechnol 1999;1:23–31.

9. Wright GD. Resisting resistance: new strategies for battling superbugs. Chem Biol 2000;7:R127–R132.

10. Rawlings ND, Barrett AJ. Evolutionary families of peptidases. Biochem J 1993;290:205–218.
11. Berger A, Schechter I. Mapping the active site of papain with the aid of peptide substrates and inhibitors.

Philos Trans R Soc Lond [Biol] 1970;257:249–264.

12. Rawlings ND. Introduction: Serine peptidases and their clans. Handbook of proteolytic enzymes

(CD-ROM). London: Academic Press, 1998, Chapter 1.

13. Supuran CT, Scozzafava A, Briganti F, Clare BW. Protease inhibitors: synthesis and QSAR study of novel

classes of non-basic thrombin inhibitors incorporating sulfonylguanidine and O-methyl-sulfonyl-isourea
moieties at P1. J Med Chem 2000;43:1793–1806.

14. Clare BW, Scozzafava A, Briganti F, Iorga B, Supuran CT. Protease inhibitors. Part 2: weakly basic

thrombin inhibitors incorporating sulfonylaminoguanidine moieties as S1 anchoring groups—Synthesis
and structure-activity correlations. J Enz Inhib 2000;15:235–264.

BACTERIAL PROTEASE INHIBITORS

363

15. Scozzafava A, Briganti F, Supuran CT. Protease inhibitors. Part 3: synthesis of non-basic thrombin

inhibitors incorporating pyridinium-sulfanilylguanidine moieties at the P1 site. Eur J Med Chem 1999;34:
939–952.

16. Supuran CT, Briganti F, Scozzafava A, Ilies MA. Protease inhibitors. Part 4: synthesis of weakly basic

thrombin inhibitors incorporating pyridinium-sulfanilylaminoguanidine moieties. J Enz Inhib 2000;15:
335–356.

17. Barrett AJ. Introduction: tissue kallikrein and its relatives. Handbook of proteolytic enzymes (CD-ROM).

London: Academic Press, 1998, Chapter 28.

18. Johnson DA. Tryptase. Handbook of proteolytic enzymes (CD-ROM). London: Academic Press, 1998,

Chapter 19.

19. Kim IS, Kang SG, Lee KJ. Physiological importance of trypsin like protease during morphological

differentiation of Streptomyces spp. J Microbiol 1995;33:315–321.

20. Yamane T, Iwasaki A, Suzuki A, Ashida T, Kawata Y. Crystal structure of Sterptomyces erythraeus

trypsin at 1.9 A resolution. J Biochem (Tokyo) 1995;118:882–894.

21. Baardsnes J, Sidhu S, MacLeod A, Elliott J, Morden D, Watson J, Borgford T. Streptomyces

griseus protease B: secretion correlates with the length of the peptide. J Bacteriol 1998;180:3241–
3244.

22. Qasim MA. Streptogrisin B. Handbook of proteolytic enzymes (CD-ROM). London: Academic Press,

1998, Chapter 77.

23. Laskowski M, Jr., Qasim MA. What can the structure of enzyme-inhibitor complexes tell us about the

structures of enzyme substrate complexes? Biochim Biophys Acta 2000;1477:324–337.

24. Sidhu SS, Borgford TJ. Selection of Streptomyces griseus protease B mutants with desired alterations in

primary specificity using a library screening strategy. J Mol Biol 1996;257:233–245.

25. Elliott J, Bennet AJ, Braun CA, MacLeod A, Borgford TJ. Active-site variants of Streptomyces griseus

protease B with peptide-ligation activity. Chem Biol 2000;7:163–171.

26. Delbaere LT, Brayer GD. The 1.8 A structure of the complex between chymostatin and Streptomyces

griseus protease A. A model for serine protease catalytic tetrahedral intermediates. J Mol Biol 1985;183:
89–103.

27. Tomkinson NP, Galpin IJ, Beynon RJ. Synthetic analogues of chymostatin. Inhibition of chymotrypsin

and Streptomyces griseus proteinase A. Biochem J 1992;286:475–480.

28. Birktoft JJ, Breddam K. Glutamyl endopeptidases. Methods Enzymol 1994;244:114–126.
29. Stennicke HR, Breddam K. Glutamyl endopeptidase I. Handbook of proteolytic enzymes (CD-ROM).

London: Academic Press, 1998, Chapter 79.

30. Stennicke HR, Breddam K. Glutamyl endopeptidase II. Handbook of proteolytic enzymes (CD-ROM).

London: Academic Press, 1998, Chapter 80.

31. Hamilton R, Walker B, Walker BJ. Synthesis and proteinase inhibitory properties of diphenyl phso-

phonate analogues of aspartic and glutamic acids. Bioorg Med Chem Lett 1998;8:1655–1660.

32. Melish MF, Glasgow LA, Turner MD. The staphylococcal scalded-skin syndrome: isolation and partial

characterization of the exfoliative toxin. J Infect Dis 1972;125:129–140.

33. Ladhani S, Joannou CL, Lochrie DP, Evans RW, Poston SM. Clinical, microbial, and biochemical aspects

of the exfoliative toxins causing staphylococcal scalded-skin syndrome. Clin Microbiol Rev 1999;12:
224–242.

34. Papageorgiou AC, Plano LR, Collins CM, Acharya KR. Structural similarities and differences in

Staphylococcus aureus exfoliative toxins A and B as revealed by their crystal structures. Protein Sci
2000;9:610–618.

35. Rader SD, Agard DA. a-Lytic protease. Handbook of proteolytic enzymes (CD-ROM). London: Aca-

demic Press, 1998, Chapter 82.

36. Davis JH, Agard DA. Relationship between enzyme specificity and the backbone dynamics of free and

inhibited a-lytic protease. Biochemistry 1998;37:7696–7707.

37. Fujinaga M, Delbaere LT, Brayer GD, James MN. Refined structure of a-lytic protease at 1.7 A resolu-

tion. Analysis of hydrogen bonding and solvent structure. J Mol Biol 1985;184:479–502.

38. Li S, Norioka S, Sakiyama F. Purification, staphylolytic activity, and cleavage sites of a-lytic protease

from Achromobacter lyticus. J Biochem (Tokyo) 1997;122:772–778.

39. Pallen MJ, Wren BW. The HtrA family of serine proteases. Mol Microbiol 1997;26:209–221.

364

SUPURAN, SCOZZAFAVA, AND CLARE

40. Spiess C, Beil A, Ehrmann M. A temperature-dependent switch from chaperone to protease in a widely

conserved heat shock protein. Cell 1999;97:339–347.

41. Laskowska E, Kuczynska-Wisnik D, Sko´rko-Glonek J, Tylor A. Degradation by proteases Lon, Clp, and

HtrA, of Escherichia coli proteins aggregated in vivo by heat shock; HtrA protease action in vivo and
in vitro. Mol Microbiol 1997;22:555–571

42. Kolmar H. DegP or protease Do. Handbook of proteolytic enzymes (CD-ROM). London: Academic

Press, 1998, Chapter 83.

43. Sakiyama F, Masaki T. Lysyl endopeptidase. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, 1998, Chapter 85.

44. Li S, Norioka S, Masaki T. Bacteriolytic activity and specificity of Achromobacter beta-lytic protease.

J Biochem (Tokyo) 1998;124:332–339.

45. Poulsen K, Reinholdt J, Jespersgaard C, Boye K, Brown TA, Hauhe M. A comprehensive genetic study of

strptococcal immunoglobulin A1 proteases: evidence for recombination within and between species.
Infect Immun 1998;66:181–190.

46. Poulsen K, Kilian M. IgA1-specific serine endopeptidase. Handbook of proteolytic enzymes (CD-ROM).

London: Academic Press, 1998, Chapter 87.

47. Lorenzen DR, Dux F, Wolk U, Tsirpouchtsidis A, Haas G, Meyer TF. Immunoglobulin A1 protease, an

exoenzyme of pathogenic Neisseriae, is a potent inducer of proinflammatory cytokines. J Exp Med 1999;
190:1049–1058.

48. Cunningham MW. Pathogenesis of group A streptococcal infections. Clin Microbiol Rev 2000;13:

470–511.

49. Efstratiou A. Group A streptococci in the 1990s. J Antimicrob Chemother 2000;45 (Suppl): 3–12.
50. Ji Y, Schnitzler N, deMaster E, Cleary P. Impact of M49, Mrp, Enn, and C5a peptidase proteins on

colonization of the mouose oral mucosa by Streptococcus pyogenes. Infect Immun 1998;66:5399–
5405.

51. Stafslien DK, Cleary PP. Characterization of the streptococcal C5a peptidase using a C5a-green fluor-

escent fusion protein substrate. J Bacteriol 2000;182:3254–3258.

52. Kortt AA, Stewart DJ. Dichelobacter (sheep foot-rot) basic serine proteinase. Handbook of proteolytic

enzymes (CD-ROM). London: Academic Press, 1998, Chapter 101.

53. Kortt AA, Stewart DJ. Properties of the extracellular acidic proteases of Dichelobacter nodosus. Stability

and specificity of peptide bond cleavage. Biochem Mol Biol Int 1994;34:1167–1176.

54. Kortt AA, Burns JE, Vaughan JA, Stewart DJ. Purification of the extracellular acidic proteases of

Dichelobacter nodosus. Biochem Mol Biol Int 1994;34:1157–1166.

55. Liu D, Yong WK. Improved laboratory diagnosis of ovine footrot: an update. Vet J 1997;153:99–105.
56. Ishihara K, Okuda K. Molecular pathogenesis of the cell surface proteins and lipids from Treponema

denticola. FEMS Microbiol Lett 1999;181:199–204.

57. Rosen G, Naor R, Sela MN. Multiple forms of the major phenylalanine specific protease in Treponema

denticola. J Periodontal Res 1999;34:269–276.

58. Uitto VJ. Trepolysin. Handbook of proteolytic enzymes (CD-ROM). London: Academic Press, 1998,

Chapter 102.

59. Supuran CT, Scozzafava A. Matrix metalloproteinases (MMPs). In: Smith HJ, Simons C (Eds.). Pro-

teinase and peptidase inhibition: recent potential targets for drug development. London: Taylor & Francis,
2001, pp 35–52.

60. Butler MJ. Tripeptidyl-peptidases A, B, and C. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, 1998, Chapter 140.

61. Banbula A, Mak P, Bugno M, Silberring J, Dubin A, Nelson D, Potempa J. Prolyl tripeptidyl peptidase

from Porphyromonas gingivalis: a novel enzyme with possible pathological implications for the
development of periodontitis. J Biol Chem 1999;274:9246–9252.

62. Kitazono A, Kabashima T, Huang HS, Ito K, Yoshimoto T. Prolyl aminopeptidase gene from Flavo-

bacterium meningosepticum: cloning, purification of the expressed enzyme, and analysis of its sequence.
Arch Biochem Biophys 1996;336:35–41.

63. Fedorko DP, Williams EC. Use of cycloserine-cefoxitin-fructose agar and

-proline-aminopepti-

dase (PRO Discs) in the rapid identification of Clostridium difficile. J Clin Microbiol 1997;35:1258–
1259.

BACTERIAL PROTEASE INHIBITORS

365

64. Yoshimoto T, Kabashima T, Uchikawa K, Inoue T, Tanaka N, Nakamura KT. Crystal structure of prolyl

aminopeptidase from Serratia marcescens. J Biochem (Tokyo) 1999;126:559–565.

65. Medrano FJ, Alonso J, Garcia JL, Romero A, Bode W, Gomis-Ruth FX. Structure of proline im-

inopeptidase from Xanthomonas campestris pv. citri: a prototype for the prolyl oligopeptidase family.
EMBO J 1998;17:1–9.

66. Ghuysen JM. Streptomyces K15

-Ala-

-Ala transpeptidase. Handbook of proteolytic enzymes (CD-

ROM). London: Academic Press, Chapter 144.

67. Grandchamps J, Nguyen-Disteche M, Damblon C, Frere JM, Ghuysen JM. Streptomyces K15 active-site

serine

-transpeptidase: specificity profile for peptide, thiol ester, and ester carbonyl donors and

pathways of the transfer reactions. Biochem J 1995;307:335–339.

68. Fonze E, Vermeire M, Nguyen-Disteche M, Brasseur R, Charlier P. The crystal structure of a penicilloyl-

serine transferase of intermediate penicillin sensitivity. The

-transpeptidase of streptomyces K15.

J Biol Chem 1999;274:21853–21860.

69. Fre`re JM. Streptomyces R61

-Ala-

-Ala carboxypeptidase. Handbook of proteolytic enzymes (CD-

ROM). London: Academic Press, Chapter 145.

70. Adediran SA, Pratt RF. Beta-secondary and solvent deuterium kinetic isotope effects on catalysis by the

Streptomyces R61

-peptidase: comparisons with a structurally similar class C beta-lactamase. Bio-

chemistry 1999;38:1469–1477.

71. Cabaret D, Liu J, Wakselman M, Pratt RF, Xu Y. Functionalized depsipeptides, substrates, and inhibitors

of beta-lactamases and

-peptidases. Bioorg Med Chem 1994;2:757–771.

72. Dalbey RE. Signal peptidase I. Handbook of proteolytic enzymes (CD-ROM). London: Academic Press,

Chapter 153.

73. Stein RL, Barbosa MD, Bruckner R. Kinetic and mechanistic studies of signal peptidase I from

Escherichia coli. Biochemistry 2000;39:7973–7983.

74. Allsop AE, Brooks G, Bruton G, Coulton S, Edwards PD, Hatton IK, Kaura AC. Penem inhibitors of

bacterial signal peptidase. Bioorg Med Chem Lett 1995;5:443–448.

75. Stathopoulos C. Structural features, physiological roles, and biotechnological applications of the mem-

brane proteases of the OmpT bacterial endopeptidase family: a micro-review. Membr Cell Biol 1998;
12:1–8.

76. Kramer RA, Dekker N, Egmond MR. Identification of active site serine and histidine residues in

Escherichia coli outer membrane protease OmpT. FEBS Lett 2000;468:220–224.

77. Kramer RA, Zandwijken D, Egmond MR, Dekker N. In vitro folding, purification and characterization of

Escherichia coli outer membrane protease ompT. Eur J Biochem 2000;267:885–893.

78. Maurizi MR. Endopeptidase Clp. Handbook of proteolytic enzymes (CD-ROM). London: Academic

Press, Chapter 177.

79. Herman C, D’Ari R. Proteolysis and chaperones: the destruction/reconstruction dilemma. Curr Opin

Microbiol 1998;1:204–209.

80. Otto HH, Schirmeister T. Cysteine proteases and their inhibitors. Chem Rev 1997;97:133–171.
81. Watts AB, Brocklehurst K. Streptopain. Handbook of proteolytic enzymes (CD-ROM). London: Acade-

mic Press, Chapter 221.

82. Herwald H, Collin M, Muller-Esterl W, Bjorck L. Streptococcal cysteine proteinase releases kinins:

a novel virulence mechanism. J Exp Med 1996;184:665–673.

83. Sugawara S, Nemoto E, Tada H, Miyake K, Imamura T, Takada H. Proteolysis of human monocyte CD14

by cysteine proteinases (gingipains) from Porphyromonas gingivalis leading to lipopolysaccharide
hyporesponsiveness. J Immunol 2000;165:411–418.

84. Ulmann D, Bordusa F. Clostripain. Handbook of proteolytic enzymes (CD-ROM). London: Academic

Press, Chapter 257.

85. Chen JM, Rawlings ND, Stevens RA, Barrett AJ. Identification of the active site of legumain links it to

caspases, clostripain, and gingipains in a new clan of cysteine endopeptidases. FEBS Lett 1998;441:361–
365.

86. Schirmeister T, Peric M. Aziridinyl peptides as inhibitors of cysteine proteases: effect of a free carboxylic

acid function on inhibition. Bioorg Med Chem 2000;8:1281–1291.

87. Gunther R, Stein A, Bordusa F. Investigations on the enzyme specificity of clostripain: a new efficient

biocatalyst for the synthesis of peptide isosteres. J Org Chem 2000;65:1672–1679.

366

SUPURAN, SCOZZAFAVA, AND CLARE

88. Bordusa F. Nonconventional amide bond formation catalysis: programming enzyme specificity with

substrate mimetics. Braz J Med Biol Res 2000;33:469–485.

89. Cummins PM, O’Connor B. Pyroglutamyl peptidase: an overview of the three known enzymatic forms.

Biochim Biophys Acta 1998;1429:1–17.

90. Awade´ A, Cleuziat P, Gonzale`s T, Robert-Baudouy J. Pyrrolidone carboxyl peptidase (Pcp): an enzyme

that removes pyroglutamic acid (pGlu) from pGlu-peptides and pGlu-proteins. Proteins Struct Funct
Genet 1994;20:34–51.

91. Bundgaard H, Moss J. Prodrugs of peptides. IV. Bioreversible derivatization of the pyroglutamyl group by

N-acylation and N-aminoethylation to effect protection against pyroglutamyl aminopeptidase. J Pharm
Sci 1989;78:122–126.

92. Cheung HTA, Dong Z, Escoffer L, Smal MA, Tattersall MHN. Activation by peptidases and cytotoxicity

of 2-(

-a-aminoacyl) prodrugs of methotrexate. In: Ayling JE, Nair MG, & Baugh CM (eds.). Chemistry

and biology of pteridines and folates. New York: Plenum Press, 1993, pp 457–460.

93. Genco CA, Potempa J, Mikolajczyk-Pawlinska J, Travis J. Role of gingipains R in the pathogenesis of

Porphyromonas gingivalis-mediated periodontal disease. Clin Infect Dis 1999;28:456–465.

94. Graves D, Jiang Y, Genco CA. Periodontal disease: bacterial virulence factors, host response, and impact

on systemic health. Curr Opin Infect Dis 2000;13:227–232.

95. Potempa J, Travis J. Gingipain R and gingipains K. Handbook of proteolytic enzymes (CD-ROM).

London: Academic Press, Chapter 258.

96. Shi Y, Ratnayake DB, Okamoto K, Abe N, Yamamoto K, Nakayama K. Genetic analyses of proteolysis,

hemoglobin binding, and hemagglutination of Porphyromonas gingivalis. Construction of mutants with a
combination of rgpA, rgpB, kgp, and hagA. J Biol Chem 1999;274:17955–17960.

97. Genco CA, Odusanya BM, Potempa J, Mikolajczyk-Pawlinska J, Travis J. A peptide domain on gingipain

R, which confers immunity against Porphyromonas gingivalis infection in mice. Infect Immun 1998;66:
4108–4114.

98. Eichinger A, Beisel HG, Jacob U, Huber R, Medrano FJ, Banbula A, Potempa J, Travis J, Bode W. Crystal

structure of gingipain R: an Arg-specific bacterial cysteine proteinase with a caspase-like fold. EMBO
J 1999;18:5453–5462.

99. Potempa J, Pike R, Travis J. Titration and mapping of the active site of cysteine proteinases from

Porphyromonas gingivalis (gingipains) using peptidyl chloromethanes. Biol Chem 1997;378:223–230.

100. Jones CH, Hruby DE. New targets for antibiotic development: biogenesis of surface adherence structures.

Drug Dev Today 1998;3:495–504.

101. Mazmanian SK, Liu G, Ton-That H, Schneewind O. Staphylococcus aureus sortase, an enzyme that

anchors surface proteins to the cell wall. Science 1999;285:760–763.

102. Ton-That H, Liu G, Mazmanian SK, Faull KF, Schneewind O. Purification and characterization of sortase,

the transpeptidase that cleaves surface proteins of Staphylococcus aureus, at the LPXTG motif. Proc Natl
Acad Sci USA 1999;96:12424–12429.

103. Ton-That H, Schneewind O. Anchor structure of staphylococcal surface proteins. IV. Inhibitors of the cell

wall sorting. J Biol Chem 1999;274:24316–24320.

104. Ton-That H, Mazmanian SK, Faull KF, Schneewind O. Anchoring of surface proteins to the cell wall of

Staphylococcus aureus. J Biol Chem 2000;275:9876–9881.

105. Cossart P, Jonquie`res R. Sortase, a universal target for therapeutic agents against gram-positive bacteria?

Proc Natl Acad Sci USA 2000;97:5013–5015.

106. Mazmanian SK, Liu G, Lenoy E, Schneewind O. Staphylococcus aureus sortase mutants deffective in the

display of surface proteins and in the pathogenesis of animal infection. Proc Natl Acad Sci USA 2000;97:
5510–5515.

107. Navarre WW, Schneewind O. Surface proteins of gram-positive bacteria and mechanisms of their tar-

geting to the cell wall envelope. Microbiol Mol Biol Rev 1999;63:174–229.

108. Coleman JE. Zinc enzymes. Curr Opin Chem Biol 1998;2:222–234.
109. Shinoda S, Miyoshi SI, Wakae H, Rahman M, Tomochika KI. Bacterial proteases as pathogenic factors,

with special emphasis on vibrio proteases. J Toxicol Toxin Rev 1996;15:327–339.

110. Vollmer P, Walev I, Rose-John S, Bhakdi S. Novel pathogenic mechanism of microbial metalloproteinase:

liberation of membrane-anchored molecules in biologically active form exemplified by studies with the
human interleukin-6 receptor. Infect Immun 1996;64:3646–3651.

BACTERIAL PROTEASE INHIBITORS

367

111. Schiavo G, Matteoli M, Montecucco C. Neurotoxins affecting neuroexocytosis. Physiol Rev 2000;80:

717–766.

112. Humeau Y, Doussau F, Grant NJ, Poulain B. How botulinum and tetanus neurotoxins block neuro-

transmitter release. Biochimie 2000;82:427–446.

113. Beynon RJ, Beaumont A. Thermolysin. Handbook of proteolytic enzymes (CD-ROM). London: Acade-

mic Press, Chapter 351.

114. Wedde M, Weise C, Kopacek P, Franke P, Vilcinskas A. Purification and characterization of an inducible

metalloprotease inhibitor from the hemolymph of greater wax moth larvae, Galleria mellonella. Eur
J Biochem 1998;255:535–543.

115. Coffey A, van den Burg B, Veltman R, Abee T. Characteristics of the biologically active 35-kDa

metalloprotease virulence factor from Listeria monocytogenes. J Appl Microbiol 2000;88:132–141.

116. Makinen PL, Makinen KK. The enterococcus faecalis extracellular metalloendopeptidase (EC 3.4.24.30;

coccolysin) inactivates human endothelin at bonds involving hydrophobic amino acid residues. Biochem
Biophys Res Commun 1994;200:981–985.

117. Nagamune K, Yamamoto K, Naka A, Matsuyama J, Miwatani T, Honda T. In vitro proteolytic processing

and activation of the recombinant precursor of El Tor cytolysin/hemolysin (pro-HlyA) of Vibrio cholerae
by soluble hemagglutinin/protease of V. cholerae, trypsin, and other proteases. Infect Immun 1996;64:
4655–4658.

118. Benitez JA, Garcia L, Silva A, Garcia H, Fando R, Cedre B. Preliminary assessment of the safety and

immunogenicity of a new CTXPhi-negative, hemagglutinin/protease-defective El Tor strain as a cholera
vaccine candidate. Infect Immun 1999;67:539–545.

119. Kon Y, Tsukada H, Hasegawa T, Igarashi K, Wada K, Suzuki E, Arakawa M, Gejyo F. The role of

Pseudomonas aeruginosa elastase as a potent inflammatory factor in a rat air pouch inflammation model.
FEMS Immunol Med Microbiol 1999;25:313–321.

120. Pesci EC, Milbank JB, Pearson JP, McKnight S, Kende AS, Greenberg EP, Iglewski BH. Quinolone

signaling in the cell-to-cell communication system of Pseudomonas aeruginosa. Proc Natl Acad Sci USA
1999;96:11229–11234.

121. Kessler E, Ohman DE. Pseudolysin. Handbook of proteolytic enzymes (CD-ROM). London: Academic

Press, Chapter 357.

122. Woessner JF. Legionella metalloendopeptidase. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, Chapter 358.

123. Moffat JF, Edelstein PH, Regula DP, Cirillo JD, Tompkins LS. Effects of an isogenic Zn-metalloprotease-

deficient mutant of Legionella pneumophila in a guinea-pig pneumonia model. Mol Microbiol 1994;12:
693–705.

124. Jung CM, Matsushita O, Katayama S, Minami J, Sakurai J, Okabe A. Identification of metal ligands in the

Clostridium histolyticum ColH collagenase. J Bacteriol 1999;181:2816–2822.

125. Matsushita O, Jung CM, Minami J, Katayama S, Nishi N, Okabe A. A study of the collagen-

binding domain of a 116-kDa Clostridium histolyticum collagenase. J Biol Chem 1998;273:3643–
3648.

126. Fukushima J, Okuda K. Vibrio collagenase. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, Chapter 367.

127. Van Wart HE. Clostridium collagenase. Handbook of proteolytic enzymes (CD-ROM). London: Acade-

mic Press, Chapter 368.

128. Bond MD, Van Wart HE. Purification and separation of individual collagenases of Clostridium

histolyticum using red dye ligand chromatography. Biochemistry 1984;23:3077–3085.

129. Bond MD, Van Wart HE. Characterization of the individual collagenases from Clostridium histolyticum.

Biochemistry 1984;23:3085–3091.

130. Jung W, Winter H. Considerations for the use of clostridial collagenase in clinical practice. Clin Drug

Invest 1998;15:245–252.

131. Scozzafava A, Supuran CT. Protease inhibitors: synthesis of potent bacterial collagenase and matrix

metalloproteinase inhibitors incorporating N-4-nitrobenzylsulfonylglycine hydroxamate moieties. J Med
Chem 2000;43:1858–1865.

132. Valee BL, Auld DS. Zinc coordination, function, and structure of zinc enzymes and other proteins.

Biochemistry 1990;29:5647–5659.

368

SUPURAN, SCOZZAFAVA, AND CLARE

133. Larsen KS, Zhang K, Auld DS.

-Phe complexes of zinc and cobalt carboxypeptidase A. J Inorg Biochem

1996;64:149–162.

134. Scozzafava A, Supuran CT. Carbonic anhydrase and matrix metalloproteinase inhibitors. Sulfonylated

amino acid hydroxamates with MMP inhibitory properties act as efficient inhibitors of carbonic
anhydrase isozymes I, II, and IV, and N-hydroxysulfonamides inhibit both these zinc enzymes. J Med
Chem 2000;43:3677–3687.

135. Scozzafava A, Supuran CT. Protease inhibitors. Part 5. Alkyl/arylsulfonyl- and arylsulfonylureido-/

arylureido-glycine hydroxamate inhibitors of Clostridium histolyticum collagenase. Eur J Med Chem
2000;35:299–307.

136. Supuran CT, Briganti F, Mincione G, Scozzafava A. Protease inhibitors: synthesis of

-alanine hydro-

xamate sulfonylated derivatives as inhibitors of Clostridium histolyticum collagenase. J Enz Inhib 2000;
15:111–128.

137. Supuran CT, Scozzafava A. Protease inhibitors. Part 7. Inhibition of Clostridium histolyticum collagenase

with sulfonylated derivatives of

-valine hydroxamate. Eur J Pharm Sci 2000;10:67–76.

138. Scozzafava A, Supuran CT. Protease inhibitors. Part 8. Synthesis of potent Clostridium histolyticum

collagenase inhibitors incorporating sulfonylated

-alanine hydroxamate moieties. Bioorg Med Chem

2000;8:637–645.

139. Scozzafava A, Supuran CT. Protease inhibitors. Part 9. Synthesis of Clostridium histolyticum collagenase

inhibitors incorporating sulfonyl-

-alanine hydroxamate moieties. Bioorg Med Chem Lett 2000;10:499–

502.

140. Scozzafava A, Ilies MA, Manole G, Supuran CT. Protease inhibitors. Part 12. Synthesis of potent matrix

metalloproteinase and bacterial collagenase inhibitors incorporating sulfonylated N-4-nitrobenzyl-b-
alanine hydroxamate moieties. Eur J Pharm Sci 2000;11:69–79.

141. Clare BW, Scozzafava A, Supuran CT. Protease inhibitors. Synthesis of a series of bacterial collagenase

inhibitors of the sulfonyl amino acyl hydroxamate type. J Med Chem 2001;44:2253–2258.

142. Clinch TE, Hobden JA, Hill JM, O’Callaghan RJ, Engel LS, Kaufmann HE. Collagen shields containing

tobramycin for sustained therapy (24 hours) of experimental Pseudomonas keratitis. CLAO J 1992;18:
245–247.

143. Kuwano M, Horibe Y, Kawashima Y. Effect of collagen cross-linking in collagen corneal shields on

ocular drug delivery. J Ocul Pharmacol Ther 1997;13:31–40.

144. Harding KG, Bale S, Llewellyn M, Baggot J, Robbins K. A pilot study of Clostridium collagenase

(Collagenase ABC

(TM)

) ointment in the debridement of dermal ulcers. Clin Drug Invest 1996;11:139–

144.

145. Jung W, Winter H. Considerations for the use of clostridial collagenase in clinical practice. Clin Drug

Invest 1998;15:245–252.

146. Lund T, Granum PE. The 105-kDa protein component of Bacillus cereu non-hemolytic enterotoxin is a

metalloprotease with gelatinolytic and collagenolytic activity. FEMS Microbiol Lett 1999;178:355–361.

147. Louis D, Bernillon J, Paisse JO, Wallach JM. Use of liquid chromatography-mass spectrometry coupling

for monitoring the serralysin-catalyzed hydrolysis of a peptide library. J Chromatogr B Biomed Sci Appl
1999;732:271–276.

148. Baumann U. Serralysin. Handbook of proteolytic enzymes (CD-ROM). London: Academic Press,

Chapter 386.

149. Morihara K. Aeruginolysin. Handbook of proteolytic enzymes (CD-ROM). London: Academic Press,

Chapter 387.

150. Belas R. Mirabilysin. Handbook of proteolytic enzymes (CD-ROM). London: Academic Press, Chapter

388.

151. Franco AA, Cheng RK, Chung GT, Wu S, Oh HB, Sears CL. Molecular evolution of the pathogenicity

island of enterotoxigenic Bacteroides fragilis strains. J Bacteriol 1999;181:6623–6633.

152. Obiso RJ, Bevan DR, Wilkins TD. Molecular modeling and analysis of the Bacteroides fragilis toxin.

Clin Infect Dis 1997;25 (Suppl 2):S153–S155.

153. Moncrief JS, Duncan AJ, Wright RL, Barroso LA, Wilkins TD. Molecular characterization of the

fragilysin pathogenicity islet of enterotoxigenic Bacteroides fragilis. Infect Immun 1998;66:1735–1739.

154. Moncrief JS, Obiso R, Jr., Barroso LA, Kling JJ, Wright RL, Van Tassell RL, Lyerly DM, Wilkins TD.

The enterotoxin of Bacteroides fragilis is a metalloprotease. Infect Immun 1995;63:175–181.

BACTERIAL PROTEASE INHIBITORS

369

155. Teplyakov A, Polyakov K, Obmolova G. Crystal structure of carboxypeptidase T from Thermo-

actinomyces vulgaris. Eur J Biochem 1992;208:281–288.

156. Strater N, Sherratt DJ, Colloms SD. X-ray structure of aminopeptidase A from Escherichia coli and a

model for the nucleoprotein complex in Xer site-specific recombination. EMBO J 1999;18:4513–4522.

157. Strater N, Sun L, Kantrowitz ER, Lipscomb WN. A bicarbonate ion as a general base in the mechanism of

peptide hydrolysis by dizinc leucine aminopeptidase. Proc Natl Acad Sci USA 1999;96:11151–11155.

158. Mathew Z, Knox TM, Miller CG. Salmonella enterica serovar typhimurium peptidase B is a leucyl

aminopeptidase with specificity for acidic amino acids. J Bacteriol 2000;182:3383–3393.

159. Mahadevan D, Saldanha JW. The extracellular regions of PSMA and the transferrin receptor contain an

aminopeptidase domain: implications for drug design. Protein Sci 1999;8:2546–2549.

160. Huntington KM, Bienvenue DL, Wei Y, Bennett B, Holz RC, Pei D. Slow-binding inhibition of the

aminopeptidase from Aeromonas proteolytica by peptide thiols: synthesis and spectroscopic characteriza-
tion. Biochemistry 1999;38:15587–15596.

161. Lowther WT, Matthews BW. Structure and function of the methionine aminopeptidases. Biochim

Biophys Acta 2000;1477:157–167.

162. Arfin SM, Kendall RL, Hall L, Weaver LH, Stewart AE, Matthews BW, Bradshaw RA. Eukaryotic

methionyl aminopeptidases: two classes of cobalt-dependent enzymes. Proc Natl Acad Sci USA 1995;92:
7714–7718.

163. Li X, Chang YH. Amino-terminal protein processing in Saccharomyces cerevisiae is an essential function

that requires two distinct methionine aminopeptidases. Proc Natl Acad Sci USA 1995;92:12357–12361.

164. Roderick SL, Matthews BW. Structure of the cobalt-dependent methionine aminopeptidase from

Escherichia coli: a new type of proteolytic enzyme. Biochemistry 1993;32:3907–3912.

165. Wilce MCJ, Bond CS, Dixon NE, Freeman HC, Guss JM, Lilley PE, Wilce JA. Structure and mechanism

of proline-specific aminopeptidase from Escherichia coli. Proc Natl Acad Sci USA 1998;95:3472–
3477.

166. Sherwood RF, Melton RG. Glutamate carboxypeptidase. Handbook of proteolytic enzymes (CD-ROM).

London: Academic Press, Chapter 483.

167. Galivan J, Ryan TJ, Chave K, Rhee M, Yao R, Yin D. Glutamyl hydrolase: pharmacological role and

enzymatic characterization. Pharmacol Ther 2000;85:207–215.

168. Rowsell S, Pauptit RA, Tucker AD, Melton RG, Blow DM, Brick P. Crystal structure of carboxypeptidase

G2, a bacterial enzyme with applications in cancer therapy. Structure 1997;5:337–347.

169. Melton RG, Sherwood RF. Antibody-enzyme conjugates for cancer therapy. J Natl Cancer Inst 1996;88:

153–165.

170. Boger DL. Vancomycin, teicoplanin, and ramoplanin: Synthetic and mechanistic studies. Med Res Rev

2001;21:356–381.

171. Lessard IA, Walsh CT. VanX, a bacterial

-alanyl-

-alanine dipeptidase: resistance, immunity, or survival

function? Proc Natl Acad Sci USA 1999;96:11028–11032.

172. Wu Z, Walsh CT. Dithiol compounds: potent, time-dependent inhibitors of VanX, a zinc-dependent

-dipeptidase required for vancomycin resistance in Enterococcus faecium. J Am Chem Soc 1996;118:

1785–1786.

173. Bussiere DE, Pratt SD, Katz L, Severin JM, Holzman T, Park C. The structure of VanX reveals a novel

amino-dipeptidase involved in mediating transposon-based vancomycin resistance. Mol Cell 1998;2:
75–84.

174. Mellors A. O-Sialoglycoprotein endopeptidase. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, Chapter 505.

175. Kessler E. b-Lytic metalloendopeptidase. Handbook of proteolytic enzymes (CD-ROM). London: Acade-

mic Press, Chapter 506.

176. Kessler E, Ohman DE. Staphylolysin. Handbook of proteolytic enzymes (CD-ROM). London: Academic

Press, Chapter 507.

177. Ryding U, Renneberg J, Rollof J, Christensson B. Antibody response to Staphylococcus aureus whole

cell, lipase and staphylolysin in patients with S. aureus infections. FEMS Microbiol Immunol 1992;4:
105–110.

178. Plaut AG. IgA-specific metalloendopeptidase. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, Chapter 508.

370

SUPURAN, SCOZZAFAVA, AND CLARE

179. Poulsen K, Reinholdt J, Jespersgaard C, Boye K, Brown TA, Hauge M, Kilian M. A comprehensive

genetic study of streptococcal immunoglobulin A1 proteases: evidence for recombination within and
between species. Infect Immun 1998;66:181–190.

180. Schiavo G, Benfenati F, Poulain B, Rossetto O, Polverino de Laureto P, DasGupta BR, Montecucco C.

Tetanus and botulinum-B neurotoxins block neurotransmitter release by proteolytic cleavage of synapto-
brevin. Nature 1992;359:832–834.

181. Schiavo G, Rossetto O, Catsicas S, Polverino de Laureto P, DasGupta BR, Benfenati F, Montecucco C.

Identification of the nerve terminal targets of botulinum neurotoxins serotypes a, D, and E. J Biol Chem
1993;268:23784–23787.

182. Lacy DB, Tepp W, Cohen AC, DasGupta BR, Stevens RC. Crystal structure of botulinum neurotoxin type

A and implications for toxicity. Nature Struct Biol 1998;5:898–902.

183. Martin L, Cornille F, Coric P, Roques BP, Fournie´-Zaluski MC: b-Amino-thiols inhibit the zinc metallo-

peptidase activity of tetanus toxin light chain. J Med Chem 1998;41:3450–3460.

184. Martin L, Cornille F, Turcaud S, Meudal H, Roques BP, Fournie´-Zaluski MC. Metallopeptidase inhibitors

of tetanus toxin: a combinatorial approach. J Med Chem 1999;42:515–525.

185. Supuran CT, Scozzafava A. Carbonic anhydrase inhibitors. Curr Med Chem Imm Endoc Metab Agents

2001;1:61–97.

186. Scott AB. Botulinum toxin injection into extraocular muscles as an alternative to strabismus surgery.

Ophthalmology 1980;87:1044–1049.

187. Huang W, Foster JA, Rogachefsky AS. Pharmacology of botulinum toxin. J Am Acad Dermatol 2000;43:

249–259.

188. Solomon BA, Hayman R. Botulinum toxin type A for palmar and digital hyperhidrosis. J Am Acad

Dermatol 2000;42:1026–1029.

189. Elliott JL, Mogridge J, Collier RJ. A quantitative study of the interactions of Bacillus anthracis edema

factor and lethal factor with activated protective antigen. Biochemistry 2000;39:6706–6713.

190. Leppla SH. Anthrax toxin lethal factor. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, Chapter 511.

191. Duesbery NS, Webb CP, Leppla SH, Gordon VM, Klimpel KR, Copeland TD, Ahn NG, Oskarsson MK,

Fukasawa K, Paull KD, Vande Woude GF. Proteolytic inactivation of MAP-kinase-kinase by anthrax
lethal factor. Science 1998;280:734–737.

192. Vitale G, Pellizzari R, Recchi C, Napolitani G, Mock M, Montecucco C. Anthrax lethal factor cleaves the

N-terminus of MAPKKs and induces tyrosine/threonine phosphorylation of MAPKs in cultured macro-
phages. Biochem Biophys Res Commun 1998;248:706–711.

193. Pellizzari R, Guidi-Rontani C, Vitale G, Mock M, Montecucco C. Anthrax lethal factor cleaves MKK3 in

macrophages and inhibits the LPS/IFNgamma-induced release of NO and TNFalpha. FEBS Lett 1999;
462:199–204.

194. Menard A, Papini E, Mock M, Montecucco C. The cytotoxic activity of Bacillus anthracis lethal factor

is inhibited by leukotriene A4 hydrolase and metallopeptidase inhibitors. Biochem J 1996;320:687–
691.

195. Park PW, Mecham RP. Lysostaphin. Handbook of proteolytic enzymes (CD-ROM). London: Academic

Press, Chapter 515.

196. Banbula A, Potempa J, Travis J, Fernandez-Catalan C, Mann K, Huber R, Bode W, Medrano F. Amino-

acid sequence and three-dimensional structure of the Staphylococcus aureus metalloproteinase at 1.72 A
resolution. Structure 1998;6:1185–1193.

197. Potempa J, Dubin A, Travis J. Aureolysin. Handbook of proteolytic enzymes (CD-ROM). London:

Academic Press, Chapter 538.

198. Langer T. AAA proteases: cellular machines for degrading membrane proteins. TIBS 2000;25:247–251.
199. Schumann W. FtsH—a single chain chaperonin? FEMS Microbiol Rev 1999;23:1–11.
200. Makino S, Makino T, Abe K, Hashimoto J, Tatsuta T, Kitagawa M, Mori H, Ogura T, Fujii T, Fushinobu

S, Wakagi T, Matsuzawa H. Second transmembrane segment of FtsH plays a role in its proteolytic activity
and homo-oligomerization. FEBS Lett 1999;460:554–558.

201. Asahara Y, Atsuta K, Motohashi K, Taguchi H, Yohda M, Yoshida M. FtsH recognizes proteins with

unfolded structure and hydrolyzes the carboxyl side of hydrophobic residues. J Biochem (Tokyo)
2000;127:931–937.

BACTERIAL PROTEASE INHIBITORS

371

Claudiu Supuran received his B.Sc. in Chemistry from the Polytechnic University of Bucharest, Roumania
(1987) and Ph.D. in Chemistry at the same university in 1991. In 1990, he became Assistant Professor of
Chemistry at the University of Bucharest. After a stage at the University of Florida in Gainesville, USA, he
returned in Bucharest, where he became Associate Professor in the Department of Organic Chemistry. In 1995,
he moved at the University of Florence, Italy, where he is currently Research Fellow and Contract Professor of
Chemistry. His main research interests include medicinal chemistry, heterocyclic chemistry, design of enzyme
inhibitors and activators, chemistry of sulfonamides, biologically active organo-element derivatives, Quanti-
tive structure-activity relationship (QSAR) studies, X-ray crystallography of metallo-enzymes, metal complexes
with biologically active ligands (metal-based drugs), carbonic anhydrases, serine proteases, matrix
metalloproteinases, bacterial proteases, and amino acid derivatives among others.

Andrea Scozzafava graduated at the University of Florence (1971), where he became Assistant Professor
of Chemistry in 1972 and Associate Professor in 1975, at the Faculty of Pharmacy. In 1980, he became full
Professor at the University of Sassari, Italy and then in 1981, he returned at the University of Florence, where
he is currently Professor of General and Inorganic Chemistry. His main research interests include structure-
function relationships of metallo-enzymes, biotransformation processes in environment; biocatalytic synthesis
for green chemistry, X-ray crystallography of metallo-enzymes, drug design of enzyme inhibitors.

Brian Clare received his B.Sc. from the University of Western Australia in 1969 and his Ph.D. from the same
university in 1973. He worked for Murdoch University as Research Officer on novel hydrometallurgical
processes from 1973 to 1981 and experimental and theoretical studies related to the application of amorphous
silicon to solar energy from 1982 to 1999, and also at UWA on theoretical studies related to boron clusters,
fullerenes, and QSAR of series of compounds of medicinal interest from 1981 to the present. His chief research
interests at present are in computational chemistry and its application to medicinal chemistry.

372

SUPURAN, SCOZZAFAVA, AND CLARE

Wyszukiwarka

Podobne podstrony:
GKS, chemioterapeutyki i inhibitory proteaz we wstrzasie
Inhibitory proteazy HIV
Aktywność inhibitorów proteaz na powierzchni ciała pszczoły miodnej
Inhibitory syntezy białek bakteryjnych (makrolidy, chloramfenikol)
Inhibitory aromatazy w leczeniu uzupeĹ‚niajÄ…cym raka piersi
Bakterie spiralne do druk
choroby wirus i bakter ukł odd Bo
chemioterapia zakazeń bakteryjnychskrócona
Bakterie 2
Szkol Lampy bakteriobójcze
CHEMIOTERAPEUTYKI W CHOROBACH BAKTERYJNYCH
Bakterie w biotech
inhibicja enzymy wykresy
bakteria 2 ppt
Zakażenia bakteryjne skóry STOMATOLOGIA
Bakterie Gram

więcej podobnych podstron