β-Lactam Antibiotics

β-Lactam Antibiotics

-Lactam Antibiotics P Liras and J F Martı´n, Universidad de Leo´n, Leo´n, Spain ª 2009 Elsevier Inc. All rights reserved. Defining Statement Classic...

1MB Sizes 2 Downloads 56 Views

-Lactam Antibiotics P Liras and J F Martı´n, Universidad de Leo´n, Leo´n, Spain ª 2009 Elsevier Inc. All rights reserved.

Defining Statement Classical and Novel -Lactam Families Clavams Carbapenems

Glossary actinomycete Gram-positive filamentous bacteria that produce many important pharmacologically active secondary metabolites. antibiotic Small molecular weight compound produced by microorganisms that inhibit or kill other microorganisms at low concentrations. cephem nucleus Bicyclic chemical structure formed by a -lactam ring and a six-membered dihydrothiazine ring containing a sulfur atom and a double bond. clavams Compounds containing a bicyclic oxazolidinic nucleus, with 5S stereochemistry, lacking -lactamase inhibitory properties. clavulanic acid -Lactam compound with -lactamase inhibitory properties, containing a bicyclic oxazolidinic nucleus and a (3R, 5R) stereochemistry.

Abbreviations ACV ACVS ATP CEA CAS DAOC DAC DGPC GSPG

-(L--aminoadipyl)-L-cysteinyl-D-valine ACV synthetase adenosine triphosphate N2-(2-carboxyethyl)-arginine clavaminate synthase deacetoxycephalosporin C deacetylcephalosporin C deoxyguanidinoproclavaminic acid culture medium containing glycerol glutamic acid, and proline

Monocyclic -Lactams Resistance Genes in Bacterial -Lactam Clusters Further Reading

nonribosomal peptide synthetases (NRPs) High molecular weight enzymes able to activate amino acids with adenosine triphosphate (ATP) to form aminoacyladenylates and to link them together to form peptides. penam nucleus Bicyclic ring formed by a -lactam ring and a five-membered thiazolidine ring containing a sulfur atom. synthases and synthetases Enzymes involved in the biosynthesis of secondary metabolites and other biological compounds. The synthetases require ATP, in contrast to the synthases, which do not require it. Synthases utilize the energy of activated substrates. -Lactam Compound containing a four-membered -lactam ring closed by an amide bond.

GCAS IPN IPNS L-pHPG LAT NRPs ORFs OHHL PBPs PCD TMS

N-glycyl-clavaminic acid synthetase isopenicillin N IPN synthase L-p-hydroxyphenylglycine lysine-6-aminotransferase nonribosomal peptide synthetases open reading frames N(3-oxohexamoyl)-L-homoserine lactone penicillin-binding proteins piperideine-6-carboxylate 12-transmembrane domains

Defining Statement

Classical and Novel -Lactam Families

The structure of classical and novel -lactam antibiotics is compared. Intermediates, enzymes, gene organization, and regulation are studied to understand the different pathways, highlighting the similarities and differences between the different groups.

-Lactams, like many other secondary metabolites, have highly unusual chemical structures when compared to those of primary metabolites. All -lactams contain a fourmembered -lactam ring closed by an amide (-lactam) bond (Figure 1). Penicillins consist of a bicyclic penam

274

Pathogenesis | -Lactam Antibiotics Conventional β-lactam antibiotics

Non-conventional β-lactams O

H N

Penicillins O Benzylpenicillin O Penicillium chrysogenum

H2N

COOH

O

O

Clavams Clavulanic acid Streptomyces clavuligerus

S

S N

O

O

COOH

O

O

Cephabacins Cephabacin F1 Lysobacter lactamgenus

COOH

H

N

O

O

O

Nocardicins Nocardicin A Nocardia uniformis

NH2

O

N

O

COOH

O NH

NH2

O-CH3

O

SO3H

OH

CH

NH

N

O

H2 N

NH2

O

NH

NH

OH NH

H COOH

Monobactams Sulfacezin Pseudomonas acidophila

NH

OH O

N

O

H2N

COOH

OH

NH

H2N

H COH S N

D

NH2

N-OH

O

HOOC

H2N

N

O

COOH

Cephamycins Cephamycin C Streptomyces clavuligerus

S

O

HOOC

COOH

COOH

OH H

Carbapenems Tienamycin Streptomyces cattleya

H O-CH3 S N

D

N

O

COOH

Cephalosporins Cephalosporin C Acremonium chrysogenum

H2N

OH

N

H N

D

275

CH HOOC

Monobactams Tabtoxin Pseudomonas syringae

O

O

NH

Figure 1 Structure of conventional -lactam antibiotics (left panel) and novel nonconventional -lactams (right panel). The generic names of the families, specific names of the structures, and producer microorganisms are indicated. Modified from Liras P and Martı´n JF (2006) Gene clusters for -lactam antibiotics and control of thier expression: Why have clusters evolved, and from where did they originate? International Microbiology 9(1): 9.

nucleus formed by a -lactam and a thiazolidine ring containing a sulfur atom and an acyl side chain bound to the amino group at C-6. They are produced by a few Penicillium and Aspergillus species. Recently, penicillins have been found to be produced by a few other ascomycete fungi. A second -lactam compound, cephalosporin C, produced by the fungus Acremonium chrysogenum and some ascomycetes, contains the cephem nucleus in which the five-membered thiazolidine ring of penicillin is substituted by a six-membered dihydrothiazine ring (Figure 1). Cephalosporin C has an -aminoadipyl side chain attached at the C-7 amino group, which is identical to that of hydrophylic penicillin N but different from those of hydrophobic penicillins. Both classical -lactam compounds, penicillins and cephalosporins, are of great clinical interest as inhibitors of peptidoglycan biosynthesis in bacteria. In the cephamycins, which are produced by various actinomycetes species, the cephem nucleus contains, in addition to the -aminoadipyl side chain, a methoxyl group at

C-7; this group renders the cephamycin structure insensitive to hydrolysis by most -lactamases. In the cephabacins a formyl group is frequently present at C-7, and different peptides are attached to the C-3 carbon of the dihydrothiazine ring in the cephem nucleus; these compounds are produced by some Gram-negative bacteria. All these compounds (penicillins, cephalosporins, cephamycins, and cephabacins) share a common mode of action and are synthesized from similar precursors by pathways with some steps in common. We refer to them as classical -lactams. In addition to these classical -lactams, many nonconventional -lactam compounds have been discovered and characterized since 1970. Some of the new -lactams also inhibit peptidoglycan biosynthesis (carbapenems, nocardicins) but others, such as clavulanic acid, are weak antibiotics but potent -lactamase inhibitors; some clavams have antifungal activity. These novel classes of compounds contain a -lactam ring and generally have a distinct bicyclic structure. The second ring in the

276 Pathogenesis | -Lactam Antibiotics

Penicillin, Cephalosporin, and Cephamycin Biosynthetic Pathways

molecule of clavulanic acid and the clavams is an oxazoline ring, which includes oxygen instead of a sulfur atom, in contrast to the classical -lactams (Figure 1). Members of the carbapenem and olivanic acid families have a carbapenem ring containing a carbon atom instead of a sulfur atom. Thienamycin is the model structure of this family. Finally, there are many compounds with monocyclic structure (monobactams), which contain only the -lactam ring and different side chains. Some of them, such as the nocardicins, are produced by actinomycetes, but other monobactams, such as sulfacezin, are produced by some proteobacteria.

A brief description of the biochemical pathways leading to classical -lactam antibiotic biosynthesis is made here. Several previous reviews give more detailed information on the specific steps. The formation of the -lactam compounds proceeds through a series of sequential reactions including the socalled early biosynthetic steps (formation of a linear peptide intermediate and cyclization to form the nucleus), ‘intermediate steps’ (nucleus modification reactions), and ‘late (decorating) steps’ (Figure 2).

Lysine lat LAT

NH2

COOH

Piperidein-6-carboxylate

pcd P6C-DH

SH

COOH

NH2 COOH

COOH

α-Aminoadipic acid

Cysteine

H2N

NH

L

COOH

O

O

Valine

ACV synthetase pcbAB

SH

LLD-ACV

D

NH

S

NH O

NH2

H 2N

N

O

NH

cefD1

Isopenicillinyl-CoA

Isopenicillin N

COOH COOH

Penicillin G

IPN cyclase pcbC

S

O

Acyl-transferase penDE H2N

O

N

COOH

O

cefD2

PenN epimerase cefD

S

NH

O

COOH

Penicillin N

N

Penicillinyl-CoA

COOH DAOC synthase cefE, cefEF

H2N

D

COOH

S

NH

O

O

DAOC

N DAC hydroxylase cefF, cefEF

COOH Acetyl-transferase cefG H2N

S

NH

D

H2N

D

S

NH

DAC

COOH

O

O

N

COOH

OCO-CH3

O

O

N

COOH

Carbamoyl-transferase cmcH

OH COOH

Cephalosporin C H2N

D

COOH

H2N

NH

D

COOH Cephabacins

O

O

CHO

S

NH

O

O

N

Methoxyl-transferase cmcl, cmcJ OCO-NH2

CH3 O

COOH

H2N

S

N OCO-peptides COOH

NH

D

COOH

O

Cephamycin C

O

S

N

OCO-NH2 COOH

Figure 2 Biosynthetic pathway of penicillins, cephalosporins, cephamycins, and cephabacins. Note that the conversion of isopenicilllin N into penicillin N by Acremonium chrysogenum occurs in three steps, while in bacteria it is catalyzed by a single epimerase enzyme. Modified from Liras P and Martı´n JF (2006) Gene clusters for -lactam antibiotics and control of thier expression: Why have clusters evolved, and from where did they originate? International Microbiology 9(1): 9.

Pathogenesis | -Lactam Antibiotics

Precursors Three amino acids, L--aminoadipic acid, L-cysteine, and L-valine, are the precursors of the basic structure of all the classical -lactam antibiotics (Figure 2); L-valine and L-cysteine are common amino acids but L--aminoadipic acid is a nonproteinogenic amino acid and must be synthesized by a specific pathway. In fungi, -aminoadipic acid is an intermediate of the lysine biosynthesis pathway. In addition, lysine is catabolized to -aminoadipic acid in Penicillium chrysogenum by (1) an !-aminotransferase, encoded by the oat1 gene, which is induced by lysine, and (2) by a reversal of the lysine biosynthesis pathway catalyzed by the enzymes saccharopine dehydrogenase/ saccharopine reductase. In actinomycetes producing -lactams lysine is converted into -aminoadipic acid semialdehyde by lysine6-aminotransferase (LAT). The lat gene is present only in -lactam-producing microorganisms and is a good probe for detecting novel -lactam producers. The LATreaction product, -aminoadipic semialdehyde, cyclizes ˜ spontaneously to form piperideine-6-carboxylate (PCD) and is later oxidized to -aminoadipic acid by a PCD dehydrogenase, encoded by the pcd gene. Both the lat and the pcd genes are located in the cephamycin gene cluster in Streptomyces clavuligerus.

Early Steps in the Formation of -Lactam Antibiotics Two enzymatic steps are common to all the classical -lactam producers, resulting in the formation of isopenicillin N (IPN), the first compound in the pathway with antibiotic activity. The first enzyme of the pathway is the -(L-aminoadipyl)-L-cysteinyl-D-valine (ACV) synthetase (ACVS), a nonribosomal peptide synthetase. The ACV synthetases are very large multifunctional proteins (Mr in the order of 460 kDa) encoded by intron-free genes of 11 kb named pcbAB, which occur in the fungal and bacterial penicillin and cephalosporin (cephamycin) clusters. This enzyme sequentially activates the three substrates with ATP to form aminoacyl-adenylates, binds them to the enzyme as thioesters, epimerizes the L-valine to D-valine configuration, links together the three amino acids to form the peptide L-(-aminoadipyl)-L-cysteinyl-D-valine, and, finally, releases this peptide from the enzyme by the action of an internal thioesterase activity. The ACV synthetases have three well-conserved domains, specifically to activate each amino acid. The second enzyme in the pathway is the IPN synthase (IPNS), also named ACV cyclase, encoded by the pcbC gene. The IPNSs are intermolecular dioxygenases that require Fe2þ, molecular oxygen, and ascorbate. They remove four hydrogen atoms from the ACV tripeptide forming the

277

bicyclic structure of IPN. The cyclase of P. chrysogenum has been crystallized. The process of direct formation of the bicyclic structure of IPN appears to be different from that of the other nonconventional -lactam antibiotics, which first form the -lactam ring and then, using a different enzyme, cyclize the monocyclic intermediate to form the oxazolidine ring. Monobactam producers lack the ability to close the second ring of the nucleus. However, early studies on cyclization of the ACV tripeptide revealed the formation of a monobactam intermediate before the IPN nucleus is formed, suggesting that perhaps a second P. chrysogenum enzyme helps the IPNS in the cyclization step. In addition to the pcbAB and pcbC genes common to bacteria and filamentous fungi, the producers of penicillin (i.e., Penicillium, Aspergillus ) contain a third gene in the penicillin gene cluster, named penDE of eukaryotic origin (it contains three introns), which encodes an IPN acyltransferase. This enzyme hydrolyzes the -aminoadipic side chain of the IPN and introduces an acyl molecule activated as a CoA derivative to produce benzylpenicillin. This gene is not present in cephalosporin C or cephamycin-producing microorganisms. Intermediate Steps IPN is converted to its D-isomer (penicillin N) in all the cephalosporin and cephamycin producers. The enzyme carrying out this epimerization was purified from actinomycetes, and the gene encoding this activity, cefD, was found to lie in the cephamycin gene cluster. The bacterial cefD-encoded protein is a pyridoxal phosphate-dependent enzyme with an Mr of about 43 kDa. However, attempts to find a homologous cefD gene and a clear epimerase activity in A. chrysogenum failed for many years. Purification of the A. chrysogenum ‘epimerase’ proved to be difficult and unreliable. In 2002 a breakthrough in our understanding of cephalosporin formation occurred when it was reported that the epimerization reaction was different in eukaryotic and prokaryotic microorganisms. The epimerization of IPN in A. chrysogenum is encoded by two linked genes, cefD1–cefD2, located in the early cephalosporin gene cluster. Transcriptional studies on the A. chrysogenum -lactam genes revealed two transcripts in the region downstream of pcbC. Sequencing of the region revealed two open reading frames (ORFs) separated by a bidirectional promoter region. The first, cefD1, has five introns and encodes a 71 kDa protein similar to fatty acid acyl-CoA synthetases. The second, cefD2, has one intron and encodes a protein with high homology to -methyl-CoA racemases of eukaryotic origin. Disruption of either of these ORFs results in a lack of cephalosporin C production, loss of IPN epimerase activity, and accumulation of IPN in the culture. The proposed conversion includes three biochemical steps: CefD1 converts IPN into isopenicillinyl N-CoA; then CefD2 isomerizes the compound into

278 Pathogenesis | -Lactam Antibiotics

penicillinyl N-CoA, which is probably released from the enzyme by the third enzyme, a thioesterase (see Figure 2). The following step in the cephalosporin/cephamycin pathway is the enzymatic expansion of the five-membered thiazolidine ring of penicillin N to a six-membered dihydrothiazine ring. The enzyme responsible for this important conversion is the deacetoxycephalosporin C (DAOC) synthase commonly known as expandase. This protein is an intermolecular dioxygenase very similar to the IPN synthase. It requires Fe2þ, molecular oxygen, and -ketoglutarate to form DAOC and succinic acid. The expandase does not recognize (or does so very poorly) the isomer IPN, penicillin G, or the deacylated 6-aminopenicillanic acid (6-APA) as substrates. The expandase of S. clavuligerus has been crystallized, and the gene cefE has been introduced into P. chrysogenum, leading to the biosynthesis of adipyl-7aminodeacetoxycephalosporanic acid (adipyl-7-ADCA) and adipyl-7-ACA, compounds that can be transformed into economically relevant semisynthetic cephalosporins. Interestingly, the expandase from A. chrysogenum is also able to catalyze the next step of the pathway, that is, the hydroxylation at C-3 that produces deacetylcephalosporin C (DAC). However, in the cephamycin- and cephabacinproducing organisms, two separate but related genes, cefE and cefF, encode enzymes that carry out these two sequential steps. cefE and cefF encode proteins with about 70% identity in amino acids, which are 60% identical to the protein encoded by cefEF in fungi. Both genes probably originated by gene duplication and then specialized in their different functions (expandase and hydroxylase, respectively) with related molecular mechanisms. This is a very interesting example of enzyme ‘specialization’ to perform different, although mechanistically related, reactions. Late Steps The final step in cephalosporin C biosynthesis is the conversion of DAC to cephalosporin C by the DACacetyltransferase, which uses acetyl-CoA as the donor of the acetyl group. This enzyme encoded by the cefG gene has an Mr of 49 kDa and is evolutively similar to O-acetylhomoserine acetyl transferases. The cefG gene contains two introns and is linked to the cefEF gene, but in the opposite orientation. Bioinformatic analysis of cefG reveals a 55% identical amino acid sequence with the met2 gene of Aspergillus fumigatus and Ascobolus immersus. The met2 genes encode O-acetylhomoserine acetyl transferase activity involved in sulfur amino acid biosynthesis. The weak nature of the cefG promoter causes a bottleneck of this enzymatic step in the cephalosporin pathway. When the cefG promoter was substituted by the glyceraldehyde-3phosphate dehydrogenase promoter (gpd) of Aspergillus nidulans or the glutamate dehydrogenase (gdh) promoter of P. chrysogenum a two- to threefold increase of cephalosporin C production was achieved.

In cephamycin-producing actinomycetes, two sequential reactions, a hydroxylation at C-7 and the transfer of a methyl group from S-adenosyl–methionine to the hydroxyl group introduced at C-7, are carried out by a complex of CmcI- and CmcJ that copurify by inmunoaffinity chromatography. This complex binds S-adenosylmethionine and DAC, in contrast to the isolated CmcI- and CmcJ-purified proteins. In addition, a carbamoyl transfer reaction is carried out by the product of the cmcH gene. The final steps of cephabacin biosynthesis have been deduced from the presence of several genes in the cephabacin gene cluster encoding nonribosomal peptide synthetases (NRPs), probably involved in the formation of the lateral peptide chain at C-3.

Regulation of Classical -Lactam Biosynthesis Transcriptional analysis of the penicillin gene cluster indicates that the pcbC and pcbAB genes are expressed from a bidirectional promoter region of 1013 bp. The two genes, pcbC and pcbAB, show a similar pattern of temporal expression and regulation. The third penicillin gene (penDE) carries its own promoter. These promoter regions are under the control of a variety of regulatory mechanisms. In P. chrysogenum, alkaline pH results in a positive regulation of the three promoters. Multiple regulatory sequences have been described in the pcbAB–pcbC bidirectional promoter region. A total of seven Pac sequences recognized by the pH transcriptional regulator PacC, six CreA sites for binding of the general carbon catabolite regulatory protein CreA, six NRE sequences putatively involved in nitrogen repression, and six CCAAT boxes for the binding of the wide-domain trimeric regulator AnCF are present in the promoter region. By coupling the bidirectional pcbAB–pcbC promoter-to-reporter genes and sequential deletion of the promoter, three boxes (A, B, and C) have been defined as essential for optimal expression of the reporter genes. Boxes A and B formed clear protein–DNA complexes as shown by electrophoretic mobility shift assays. Deletion of box A caused a decrease of 41% in the transcriptional activity of the promoter. A palindromic heptanucleotide sequence TTAGTAA is the binding site for a transcriptional activator, named PTA1. Box A also contains four putative CreA binding sites with SYGGRG consensus sequence. Box B contains a CreA site, three putative recognition sites for PacC, and two NRE consensus sequences. However, little is known at present on the nature of the modulator proteins – such as PTA1 – controlling the expression of this promoter. Regulation in prokaryotic -lactam producers is completely different. In Streptomyces, a cascade of regulatory genes control antibiotic production and differentiation. Small diffussible molecules of the -butyrolactone family, such as A-factor, are at the top level of the regulatory

Pathogenesis | -Lactam Antibiotics

cascade since they control morphological differentiation, secondary metabolite biosynthesis, and resistance. At the bottom level of the cascade are the pathwayspecific transcriptional regulators that are found in most actinomycete gene clusters for antibiotic biosynthesis. The pathway-specific transcriptional regulator for cephamycin C and clavulanic acid is encoded by the ccaR gene, located in the cephamycin C cluster. CcaR is a protein (Mr 28 kDa) with amino acid identities in the range of 26–29% with a Streptomyces coelicolor regulatory protein (AfsR) and several antibiotic-specific regulatory proteins (DnrI, RedD, and ActII-ORF4) of the SARP family (Streptomyces antibiotic regulatory proteins). Disruption of ccaR results in lack of production of cephamycin C, clavulanic acid, or other clavams. The wild-type phenotype is restored in the disrupted mutant by complementation with ccaR. Amplification of ccaR results in a 200% increase in the production of cephamycin C and clavulanic acid. Expression of CcaR depends on a putative anti-anti-sigma factor encoded by bldG as shown in bldG-null mutants. Although CcaR contains a TTA codon (Leu35), that in Streptomyces species results in translational control mediated by a rare leucine-tRNA encoded by the bldA gene, this typical regulation by bldA does not occur in CcaR translation. Mobility shift experiments indicate that CcaR is a positive autoregulatory protein; it binds to its own promoter and to the bidirectional cefD–cmcI promoter. Binding of CcaR to the cefD–cmcI bidirectional promoter controls early, intermediate, and late steps of cephamycin biosynthesis and increases the expression of the genes involved. As detected by immunoassay, the lack of expression of ccaR results in the absence of enzymes responsible for the early and middle steps of cephamycin biosynthesis, that is, LAT, IPNS, and DAOC.

279

Like other SARP-like proteins, CcaR contains all the putative DNA-binding domains and a number of highly conserved amino acids corresponding to helical structures. However, the putative heptamer consensus sequences 59-TCGAGCG/C -space- 59-TCGAGCG present in the target DNA of SARP proteins have not been found in the promoters of the biosynthetic genes of clavulanic acid or cephamycin C, and footprinting studies are required to precisely identify the CcaR binding sequences. In addition to the region for CcaR binding, the promoter of the ccaR contains a 26 bp ARE sequence located 890 pb upstream of the ATG start codon. ARE sequences have been reported to be sites of binding for Brp proteins (for butyrolactones receptor proteins). The ARE sequence of S. clavuligerus is functional and binds a Brp protein expressed in Escherichia coli from the S. clavuligerus brp gene. The role of butyrolactones in antibiotic production by Streptomyces has been reported for streptomycin, pristinamycin, and tylosin, but up to now, no butyrolactones have been described to be synthesized in S. clavuligerus or other cephamycin C producers. Besides Brp, the sequence is a binding site for additional proteins, as shown by gel mobility assays using cell extracts from a brp-null mutant of S. clavuligerus. Pleiotropic effectors such as ppGpp and the proteins encoded by relA and relC also play an important role in regulating cephamycin C and clavulanic acid biosynthesis in S. clavuligerus, but their molecular mechanisms are still obscure.

Clusters of Genes for -Lactam Antibiotic Biosynthesis Genes for classical -lactam biosynthesis are clustered in all the producer strains (Figure 3). With minor

Penicillium chrysogenum (Penicillin) pcbAB

Acremonium chrysogenum (Cephalosporin C) cefT

pcbAB

orf3

pcbC penDE

Chromosome VII

Chromosome I

pcbC

cefEF cefG

cefD2 cefD1

Streptomyces clavuligerus (Cephamycin C) pcbR pcbC

pcbAB

ccaR cmcH cefF cmcJI cefD cefE pcd cmcTpbp74 bla

lat blp

Lysobacter lactamgenus (Cephabacin)

bla cefD cefF cefE

pcbC

pcbAB

T1

T2

NRPS 1,2,3

T3 NRPS4 cmcJ

Figure 3 Cluster of genes for -lactam biosynthesis in the fungi Penicillium chrysogenum and Acremonium chrysogenum, the actinomycete Streptomyces clavuligerus, and the Gram-negative bacteria Lysobacter lactamgenus. Notice that the pcbAB and pcbC genes are in different orientation in the prokaryotic and eukaryotic producers. The NRPS1 to NRPS4 genes present in the L. lactamgenus cluster encode nonribosomal peptide synthetases (NRPs), putatively involved in the synthesis of the side chain at C-3; T1, T2, and T3 are genes for ABC transport proteins. Modified from Liras P and Martı´n JF (2006) Gene clusters for -lactam antibiotics and control of thier expression: Why have clusters evolved, and from where did they originate? International Microbiology 9(1): 9.

280 Pathogenesis | -Lactam Antibiotics

differences between strains, the pcbC–pcbAB genes are always together and located adjacent to the penDE gene in the penicillin producers. In A. chrysogenum, the ‘early’ gene cluster on chromosome VII (4.6 Mb) contains the genes pcbAB and pcbC encoding the enzymes for the first two steps of the pathway, the genes cefD1 and cefD2 responsible for the epimerization of IPN and cefT, encoding a transmembrane protein for putative exportation of the antibiotic. The ‘late’ gene cluster, on chromosome I (2.2 Mb), contains the genes for the last steps of cephalosporin biosynthesis, cefEF and cefG. A similar gene organization, pcbAB–pcbC–cefD2, has been found in a wood-inhabiting marine fungus, Kallichroma tethys, phylogenetically related to A. chrysogenum. The largest -lactam clusters are those of cephamycinand cephabacin-producing bacteria. In S. clavuligerus the cephamycin C gene cluster is adjacent to the clavulanic acid gene cluster. The whole supercluster of cephamycin C–clavulanic acid extends for about 50 kb. This organization of the biosynthetic genes of both antibiotics in a supercluster also occurs in other clavulanic acid-producing strains, such as Streptomyces jumonjinensis and Streptomyces katsurahamanus. Cephamycin C biosynthesis genes in Amycolatopsis lactamdurans and genes for cephabacin biosynthesis in Lysobacter lactamgenus are also clustered, although the organization of the clusters is not identical in the different producers.

Clavulanic Acid Precursors L-Ornithine and L-arginine are incorporated at carbons C-2, C-3, and C-8 to C-10 of the clavulanic acid molecule; the five carbons of the ornithine molecule are incorporated as an intact unit in the clavulanic acid molecule but the direct precursor of the five-carbon unit is L-arginine. Mutants blocked in the arg F and arg G genes (encoding enzymes for the conversion of ornithine to arginine) are unable to incorporate ornithine into clavulanic acid. Labeled glycerol, glycerate, propionate, and -hydroxypropionate are incorporated into carbons C-5 to C-7 of the clavulanic acid molecule. Purification of the enzyme involved in the condensation of arginine with the C-3 unit showed that it uses D-glyceraldehyde-3-phosphate more efficiently as a substrate than any of the other C-3 possible precursors. Both arginine and glycerol strongly increase clavulanic acid formation. Different 13C-ornithine-labeled molecules are incorporated (via arginine) into clavulanic acid and clavam-2carboxylate with similar efficiency and giving the same regiochemical structure. Feeding studies also showed that the 13C-labeled pathway intermediates clavaminic and proclavaminic acids are incorporated with equal efficiency into clavulanic acid and clavam-2-carboxylate. All these results suggest parallel pathways for the biosynthesis of these two clavams with some common early intermediates.

Gene Clusters for Clavulanic Acid and Clavams

Clavams A group of -lactam compounds contains the clavam structure shown in Figure 4(a). The nucleus resembles that of the penicillins but the five-membered (oxazolidine) ring has an oxygen atom instead of sulfur and lacks the C-6 acylamino side chain of penicillins. The most important commercial clavam compound and the best known is the -lactamase inhibitor clavulanic acid produced by S. clavuligerus. This compound has a bicyclic clavam nucleus with 3R, 5R stereochemistry and shows very poor antibacterial activity. Additional compounds with clavam structure have been isolated from S. clavuligerus culture broths, including clavam-2-carboxylate, 2-formyloxymethylclavam, 2-hydroxymethylclavam, hydroxyethylclavam, and alanylclavam (Figure 4(a)). Metabolites with clavam structure have also been isolated from Streptomyces antibioticus TU1718 and Streptomyces hygroscopicus. All the clavam compounds other than clavulanic acid lack the carboxyl group at C-3 and have a (3S, 5S) stereochemistry. They are antifungal agents but alanylclavam and valclavam possess antibacterial properties. Here we focus on the biochemistry and genetics of clavulanic acid.

Progress in the understanding of clavulanic acid biosynthesis has been closely related to the knowledge of the clavulanic acid gene cluster and to the isolation of its biosynthetic intermediates. The following intermediates of the clavulanic acid pathway have been isolated from S. clavuligerus culture broths: clavaminic acid and proclavaminic acid, two guanidino compounds, N2-(2-carboxyethyl)-arginine (CEA) and deoxyguanidinoproclavaminic acid (DGPC), clavaldehyde, and N-glycylclavaminic acid (see Figure 4(b) for structures). An enzyme, clavaminate synthase (CAS), was reported to convert proclavaminic acid into clavaminic acid. Purification of S. clavuligerus CAS to homogeneity showed that the activity was associated with two polypeptides of 46 and 47 kDa, CAS1 and CAS2, respectively, with 82% identity in amino acid sequence; this leads to the cloning of two genes, cas1 and cas2, 85% identical in nucleotide sequence but located at distant positions in the genome. Since S. clavuligerus also produces clavams, the possibility of duplicated clusters for clavulanic acid and clavams was postulated. The situation becomes more complex when a third cluster of genes related to clavulanic acid and 5S-clavam biosynthesis was reported in S. clavuligerus. The structure of the clusters is shown in

Pathogenesis | -Lactam Antibiotics H

H

(a)

9

O 6 5 7

N

8

3

O

COOH Clavulanic acid H

H

O

2-formyloxymethylclavam

NH

(b)

O

Alanylclavam

NH

+

NH

OH

Glyceraldehyde-3-P +

NH COOH

COOH N-acetylglycylclavaminic acid

NH

COOH Deoxyguanidinoproclavaminic acid (DGPC) cas2

CAS2 O2

H O

NH H CAS2

N

NH2

O

O2

cas2

OH O

COOH CAS

NH2

NH N

O

COOH

N 2-(2-Carboxyethyl)-arginine (CEA)

Arginine

N H

O

β-LS bls

COOH

NH2

N

O

NH2 NH

CEAS2 ceas2

H-O-C -

H

O

H N

O

2-hydroxyethylclavam

NH2

P

COOH N-glycylclavaminic acid

OH

N

O

H NH2

O

H O

N

O

N

O

2-Hydroxymethylclavam

COOH

N H2

OH H

H O

H

H N

O

N

O

H

O

N

O

H

Clavam-2-carboxylate

O

H O

COOH

2

N

O

H O

OH

281

NH2

PAH

N

OH NH

pah

O

COOH

Dihydroproclavaminic acid

NH2

N COOH

Guanidinoproclavaminic acid

Proclavaminic acid

O2 cas2 H

H O N

gcas, orf17 O

O COOH Clavaminic acid

O

NH2 GCAS

H N-CO-CH2-NH2

O

N O COOH

N-Glycylclavaminic acid

H

H

N

O car

COOH Clavaldehyde

CAR

O

OH

N

O COOH Clavulanic acid

Figure 4 (a) Structure of clavulanic acid and the clavams detected in Streptomyces clavuligerus culture broths. (b) Clavulanic acid biosynthetic pathway. The names of the intermediates, the enzymes (in capitals) for the different steps, and the genes (in italics) encoding them are indicated. Some steps (e.g., the conversion of N-glycylclavaminic acid to clavaldehyde) are unknown. Modified from Liras P and Rodrı´guez-Garcı´a A (2000) Clavulanic acid, a -lactamase inhibitor: Biosynthesis and molecular genetics. Applied Microbiology and Biotechnology 54: 467.

Figure 5(a), and the characteristics of the proteins encoded by the genes in the different clusters are indicated in Figure 5(b). The first gene cluster for clavulanic acid biosynthesis (Figure 5(a, I)), named the clavulanic acid gene cluster, was cloned by complementing clavulanic acid nonproducing mutants. It includes the cas2 gene and is located downstream of the pcbC gene for cephamycin C biosynthesis, forming the so-called cephamycin C–clavulanic acid supercluster. The first fragment cloned (6.6 kb) included the genes bls2, pah2, cas2, oat2, and oppA1. The cluster was enlarged by the inclusion of claR and car, cyp, fd, and orf12 to

orf19. Disruption of some of the genes, that is, cyp, claR, and bls2, resulted in a total lack of clavulanic acid biosynthesis. However, surprisingly, disruption of other genes, some of them known to encode enzymes for clavulanic acid biosynthesis (ceaS2, pah2), blocked clavulanic acid formation only under some growth culture conditions and had no clear effect on clavam biosynthesis, supporting the conclusion that some genes, encoding early enzymes of the pathway, might be duplicated and used independently for clavulanic and 5S-clavam biosynthesis. Using the cas1 gene as a probe, a second cluster of genes was isolated. This cluster contains the cas1 gene

282 Pathogenesis | -Lactam Antibiotics E SN

(a)

B

B

S

B

E

B

N

I pbp54 ceaS

K

N N

bls

B K K

pah

B

cs2

oat2

N N

S

oppA1

claR

N/S N N

car cyp-fd

N

orf12 13 14

NK S N

oppA2

16

gcas pbpA

N Sp

pbp2

N

K

II cvmG cvmP cvmH cvm13

12

3

7

11

N

NN

2

1

N

cas1

cvm4

5

N

6

N

tRNA cvm9

10

N

III ceaS1

bls1 pah1

oat1 cvm6P

cvm7P

(b) I. Gene cluster for clavulanic acid Gene ceaS2 bls2 cas2 pah2 oat2 oppA1 claR car cyp-fd orf12 orf13 orf14 oppA2 orf16 gcas pbpA pbp2

Similarity/Identity Carboxyethylarginine synthetase β-Lactam synthase Clavaminate synthase Proclavaminic acid hydrolase Ornithine acetyl transferase Periplasmic oligopeptide-binding Transcripitional regulator LysR-type Clavaldehyde reductase P450-ferredoxin Unknown Unknown Unknown Periplasmic oligopeptide-binding Unknown N-glycyl-clavaminic synthetase Penicillin-binding protein Penicillin-binding protein

II. Gene cluster for clavams Gene cvmG cvmP cvmH cvm13 cvm12 cvm11 cvm7 cvm3 cvm2 cas1 cvm4 cvm5 cvm6 cvm9 cvm10

Similarity/Identity Secreted protein Arginine deiminase Hydrolase LanU-type Asparaginase Transcripitional regulator Efflux Protein Paralogous to cvm7P Flavin reductase Hypothetical isomerase Clavaminate synthase Homoserine acetyltransferase Oxidoreductase Aminotransferase Transcripitional regulator Protein kinase

III. “Paralogues” gene cluster Gene ceaS1 bls1 pah1 oat1 cvm6P cvm7P

Similarity/Identity Paralogue to ceaS2 Paralogue to bls2 Paralogue to pah2 Paralogue to oat2 Paralogue to cvm6 Transcriptional regulator PimR-type

Figure 5 (a) Cluster of genes for clavulanic acid (I), the clavams (II), and the paralogous cluster (III). Notice in I the gene pcbR (black arrow) that connects the clavulanic acid gene cluster and the cephamycin C gene cluster. (b) The function of the genes deduced from biochemical studies or from gene disruption studies (ceaS2, bls2, cas2, pah2, cas1, oat2, oppA1, oppA2, gacs) is indicated; in other cases the function has been deduced from the sequence of the protein encoded by the gene.

and genes cvm1 to cmv6. Disruption of some of these genes (cvm1, cvm4, and cvm5) resulted in mutants unable to produce any 5S clavams but still able to synthesize clavulanic acid. This cluster for clavam biosynthesis was later extended to include genes cvm7 to cvm13 and cvmG and cvmP (see Figure 5(b, II)). The role of the cvm genes is still poorly known. Mutation of cvm2 results in a strain still producing some alanylclavam and 2-hydroxymethylclavam but completely blocked in clavam-2carboxylate production; however, null mutants in cvm3, 4, 6, 7, 9, 10, or 11 had no appreciable effect on 5S-clavam production. Hybridization of S. clavuligerus DNA with the pah2 gene of the clavulanic acid gene cluster, encoding the

proclavaminic acid hydrolase, allowed the identification of a second copy of the pah gene (named later pah1). Surrounding pah1 a third cluster of genes, named the paralogous gene cluster for clavulanic acid biosynthesis, has been located. This cluster (Figure 5(a, III)) contains genes duplicated from those of the first clavulanic acid cluster that were named ceaS1, bls1, pah1, and oat1. In addition, in this third cluster two other genes, cvm6P and cvm7P, for clavam biosynthesis have been located (Figure 5(a, III)). Clavulanic Acid Biosynthesis The purification and characterization of enzymes by expression of the clavulanic acid genes in E. coli has been

Pathogenesis | -Lactam Antibiotics

of great importance in the understanding of the biosynthetic pathway (Figure 4(a)). CAS purification elucidated the formation of the second ring in the clavulanic acid molecule. The CASs are nonheme iron dioxygenases that, like the IPNS involved in penicillin and cephalosporin biosynthesis, require -ketoglutarate as cosubstrate. However, in contrast to IPNS, CAS2 is a multifunctional enzyme, able to introduce an oxygen atom from molecular oxygen as a hydroxyl group into DGPC, forming guanidinoproclavaminic acid, and then to carry out two additional oxidative steps, bringing about cyclization/desaturation of proclavaminic to clavaminic acid (Figure 4(a)). Disruption of different ORFs in the clavulanic acid cluster facilitated the characterization of additional steps in the pathway. The first enzyme in the pathway is encoded by ceaS2 (also named pyc in a separate study). The CEAS2 (Mr 62 kDa) is a protein with domains recognizing pyruvate and thiamine pyrophosphate and is similar (30% identity) to acetohydroxyacid synthases that condense two molecules of pyruvate to form -acetolactate. The enzyme, therefore, appears to use pyruvate or a related 3-carbon molecule as a substrate. Inactivation of ceaS2 leads to the inability to produce clavulanic acid in complex media but not in a glycerol-supplemented glutamic acid, and proline (GSPG) defined medium, supporting the existence of a second enzyme with similar function (probably CEAS1). Simultaneously, it has been reported that CEAS2 requires ATP and Mg2þ and uses arginine and D-glyceraldehyde-3phosphate (D-G3P) as substrates. This enzyme has the unusual ability to form C–N bonds instead of the C–C bonds associated with thiamine–pyrophosphate-requiring enzymes. Disruption of ORF3 (bls2) in the clavulanic cluster causes the accumulation of large amounts of CEA. Chemical complementation of the bls2 mutant with DGPC restored the ability to produce clavulanic acid. The enzyme catalyzing the conversion of CEA to DGPC has been named -lactam synthetase. It directly forms the four-membered -lactam ring by means of an intramolecular amide bond in the CEA molecule, in the presence of ATP and Mg2þ. This is a new mechanism to form a -lactam ring different from that of the classical -lactams. The -lactam synthetase is similar to asparagine synthetases, which are primary metabolism aminotransferases. Hydrolysis of the guanidine group of guanidinoproclavaminic acid (Figure 4(b)) is carried out by a 33.4 kDa protein monomer, which was accidentally copurified with the ACV synthetase. Probes from the N-terminus of this protein were used to clone pah2. Recombinant PAH2 converts guanidinoproclavaminic acid to proclavaminic acid and urea. PAH2 has been crystallized; it is a hexameric protein of the arginases family that requires Mn2þ but lacks arginase, agmatinase, or -guanidinobutyrate ureidohydrolase activities. It does not hydrolyze the guanidino group from either DL-arginine or N-acetylarginine but uses

283

N-acetyl-(L)-arginine and (3R)hydroxy-N-acetylarginine as substrates. Disruption of orf15 and orf16 in the clavulanic acid cluster leads to the accumulation of new clavam compounds characterized as N-acetylglycyl-clavaminic acid and N-glycyl-clavaminic acid. These results facilitated the characterization of orf17. It encodes an N-glycyl-clavaminic acid synthetase (GCAS) belonging to the ATP-grasp fold superfamily protein, which forms in vitro N-glycyl-clavaminic acid from clavaminic acid, ATP, and glycine. The mechanism by which the orf15- and orf16-encoded proteins regulate the synthesis of GCAS is unclear. The C-9 oxygen in the clavulanic acid molecule (occupying the position of the C-9 amino group of clavaminic acid) derives from molecular oxygen as shown by incorporation of labeled dioxygen. Clavaldehyde, an intermediate of clavulanic acid biosynthesis, is accumulated by randomly obtained mutants of S. clavuligerus. If N-glycyl-clavaminic acid is an intermediate between clavaminic acid and clavaldehyde, it seems that the N-glycyl group of NGCA is removed by oxidative cleavage to yield clavaldehyde, either directly or indirectly. However, the steps between N-glycyl-clavaminic acid and clavaldehyde are still unknown (Figure 4(b)). Alternatively N-glycyl-clavaminic acid may be a side product, and the conversion of clavaminic acid to clavaldehyde may be a direct process involving an oxidative step with removal of the C-9 amino group by a mechanism similar to those of deaminating omega-aminotransferases (which convert ornithine to glutamic-5semialdehyde). Clavaldehyde possesses -lactamase inhibitory properties, which correlates well with its 3R, 5R stereochemistry, but it is more unstable than clavulanic acid. The conversion of clavaldehyde into clavulanic acid is carried out by an NADP-dependent, 28 kDa monomer, clavaldehyde reductase. The N-terminal amino acid sequence of CAR confirmed that it is encoded by a gene, car, located about 4 kb downstream of cas2. In vivo, CAR appears to be a tetramer, as is usual in short-chain dehydrogenases, and acts reversibly on clavaldehyde and clavulanic acid. The precise role of other genes in the clavulanic acid cluster is unknown. Genes cyp-fd encode a P450 cytochrome– ferredoxin essential for clavulanic acid biosynthesis. This type of protein is usually involved in hydroxylation, and the P450 cytochrome might be an in vivo electron carrier protein for the CAS; alternatively, it might play a role in the final introduction of a hydroxyl group at C-9 into the clavaldehyde formation. Gene oat2 encodes an N-acetyl transferase similar to ArgJ in the arginine biosynthesis pathway. Expression of oat2 is negatively controlled by ArgR, the arginine biosynthesis repressor protein. Disruption of this gene only affects clavulanic acid production.

284 Pathogenesis | -Lactam Antibiotics

Regulatory Genes in the Clavam Clusters In the regulatory cascades controlling antibiotic production and differentiation, some signal molecules, identified as peptides, have been shown to be involved in restoration of aerial mycelium in S. coelicolor. Peptide mixtures at micromolar concentration accelerate the onset and increase the level of the antibiotics produced by S. clavuligerus. Two genes in the clavulanic acid gene cluster, oppA1 and oppA2, which do not complement each other, encode proteins with 48% amino acid identity to periplasmic oligopeptide-binding proteins. Null mutants in oppA1 or oppA2 produce normal levels of cephamycin C but do not produce any clavulanic acid; these mutants show a higher resistance to the toxic tripeptide bialaphos, confirming the role of oppA1 and oppA2 in peptide transport. As reported previously, CcaR positively controls clavulanic acid and clavam production, but the promoters in the clavulanic acid cluster to which CcaR binds are unclear. A second gene, claR, encoding a different regulatory protein is present in the clavulanic acid cluster. This gene encodes a polypeptide of 47 kDa with a significant degree of identity to many transcriptional activators of the LysR family. The protein possesses two characteristic helixturn-helix motifs, corresponding to the DNA-binding regions of transcriptional activators, which are located at the N- and C-carboxy termini of the protein. Disruption of claR results in mutants producing cephamycin C but unable to produce clavulanic acid. This null mutant accumulates clavaminic acid, indicating that claR controls only the late genes involved in the conversion of clavaminic acid to clavulanic acid. Accordingly, the transcripts of car, ORF9, and ORF10 in the clavulanic acid cluster are not expressed in claR-disrupted strains. Initial experiments showed a lack of transcription of claR, as detected by Northern hybridization, in ccaR-disrupted mutants. However, RT-PCR experiments show a low expression of the clavulanic acid genes (ceaS2, cas2, pah2, bls2, car) in the S. clavuligerus claR mutant. Three genes, cvm7, cvm9, and cvm12, encoding different types of regulatory proteins are present in the clavam gene cluster but their disruption does not affect clavulanic acid or clavam biosynthesis. Only the disruption of cvm7P, encoding a PimR-type regulator, results in a complete lack of 5S-clavam production but this mutant retains normal clavulanic acid biosynthesis.

Carbapenems The carbapenems have the -lactam ring fused to a fivemembered ring containing a carbon atom instead of the sulfur atom characteristic of the penicillins or the oxygen present in the clavams (Figure 1). Several research groups found carbapenem compounds produced by Streptomyces species in

screenings for cell-wall inhibitors – for example, thienamycins, carpetimycins, asparenomycins, PS antibiotics, and olivanic acids. Thienamycin, produced by Streptomyces cattleya, is the most potent broad-spectrum antibacterial compound of all of them, being resistant to -lactamases, but it is chemically unstable and only chemically modified derivatives are used clinically. The best-studied carbapenem is the carbapenem-2-em-3-carboxylic acid produced by unicellular Gram-negative bacteria, such as Serratia and Erwinia. Precursors of thienamycin have been identified by using radioactive and stable isotope-labeled substrates. Cysteine, glutamate, acetate, and the methyl group of methionine are incorporated into the thienamycin molecule (Figure 6(a)). A cluster of 22 genes (thn A to V) has been located using a cosmid library in a 32 kb DNA region of the S. cattleya genome (Figure 6(b)). Independent insertional inactivation of thn N, thn L, or thn O gave rise to thienamycin nonproducing mutants. Based on the deduced function of the genes, they have been separated into possible structural genes, including genes for putative cysteine transferases (thn T,V), methyl transferases (thn K,L), and a -lactam synthetase (thn M), genes involved in thienamycin export or resistance, regulatory genes (thnU) and genes of unknown function. A biosynthetic pathway of at least ten steps has been proposed for thienamycin biosynthesis, but requires experimental confirmation. The precursors of the carbapenem produced by Erwinia are proline and malonyl-CoA. The pathway, confirmed by the purification and crystallization of the three main enzymes, is much simpler than in other -lactam producers. It is a novel -lactam biosynthetic pathway consisting of four steps (Figure 7(a)). In Erwinia carotovora, a plant pathogen, the cluster contains merely nine genes (car R–car A to H). Genes car ABC are required for carbapenem biosynthesis (Figure 7(b)). Genes car DE encode putative enzymes involved in the transformation of proline (or glutamate) to L-glutamate semialdehyde. This compound is fused to malonyl-CoA by CarB, a trimeric carboxymethylproline synthase, similar to crotonases, able to form carbon–carbon bonds. CarB is 37% identical in amino acid sequence to the deduced S. cattleya protein ThnE for thienamycin biosynthesis. The (5S-carboxymethyl)-S-proline intermediate formed by CarB already contains the five-membered ring of the carbapenem molecule and is additionally cyclized by CarA, the enzyme that forms the intracellular amide bond of the -lactam ring. CarA is an ATP/Mg2þdependent enzyme homologous to the -LS enzyme forming the -lactam ring in clavulanic acid biosynthesis and to ThnM, for thienamycin biosynthesis. A third enzyme, CarC, the carbapenem synthase, oxydizes the (3S, 5S) carbapenam intermediate to the final product (5R)-carbapenem. This enzyme is an -ketoglutarate, Fe2þdependent hexameric dioxygenase with a 23% amino acid

Pathogenesis | -Lactam Antibiotics (a)

Thienamycin

OH COOH H2N

COOH

8

S

8 9

9

5 6 7

1 4

N

S

2

SH

NH2

3

O

Methionine

285

NH2 Cysteine

COOH

O 4 6

7

NH2

S-CoA

Acetyl-CoA

3

2 5

COOH

COOH Glutamate

(b)

thnA

B

C

D

E

F

G H

I

J

K

L

M N

O P

Q

R S

T

U V

Figure 6 (a) Precursors of the thienamycin molecule. (b) Cluster of genes for thienamycin. The function of some of these genes is indicated in the text. Modified from Nu´n˜ez LE, Me´ndez C, Bran˜a AF, Blanco G and Salas JA (2003) The biosynthetic gene cluster for the -lactam carbapenem thienamycin in Streptomyces cattleya. Chemistry and Biology 10: 301.

sequence identity to CAS. CarC epimerizes and desaturates the carbapenem structure. Control of the expression of the car cluster is exerted by a car R/carI system (similar to the luxR/I control system of Vibrio). The CarR protein acts as a positive regulator binding to the car A promoter region but is not required to express the car FG genes, responsible for carbapenem resistance, which are expressed from a promoter region internal to carD (Figure 7(b)). In addition, carbapenem biosynthesis in E. carotovora is regulated by a quorum sensing control, in which N(3-oxohexanoyl)-Lhomoserine lactone (OHHL) is the signal molecule. OHHL binds the CarR protein and induces the formation of multimeric aggregates of CarR that bind strongly to the carA promoter, increasing the expression of the car cluster.

Monocyclic -Lactams This group of compounds contains only the -lactam ring bound to a side chain, but their structures lack the thiazolidine, oxazolidine, or dihydrothiazine ring characteristics of the other -lactam groups. The first monobactams described were the -lactamaseresistant nocardicins (Figure 1) discovered as a group of compounds moderately active against Gram-negative bacteria. These compounds are produced by the actinomycete Nocardia uniformis var tsuyamanensis as a family of seven related compounds (nocardicin A to G), with nocardicin A being the most active. Structural differences among the nocardicins are

due to a homoserine side chain in nocardicins A, B, C, and D, which is absent in nocardicins E, F, G, and the presence and orientation of an oxime group at C-3 of the monobactam ring in nocardicins A, B, E, and F. Nocardicins were later isolated from Actinosynnema mirum, Nocardiopsis atra, and Microtetraspora caesia. A number of unicellular bacteria, both Gram-positive (Staphylococcus and Micrococcus sp.) and Gram-negative (species of Pseudomonas, Chromobacter, and Agrobacterium), produce monobactams with a simple structure (Figure 1). They contain a sulfonic acid group bound to the N-atom of the amide group at the -lactam ring, acyl substituents at C-3, and, in some cases, a methoxyl group at C-3. All these monobactams show weak antibacterial activity, especially against Gram-negative bacteria. In addition some plant pathogen bacteria, such as Pseudomonas syringae, produce very simple -lactam compounds, such as tabtoxin, a cyclic peptide, that acts as an inhibitor of plant glutamine synthetases. Serine is known to be the precursor of the -lactam ring in most of the monocyclic -lactams studied. The acyl group of the unicellular bacterial monobactams is derived from D-glutamyl-D (or L)-alanine. Most biosynthetic studies have been performed on the nocardicins. Precursor incorporation into nocardicin A indicates that the compound is formed from L-serine, L-homoserine, and L-tyrosine, which is metabolized to the direct precursor, a nonproteinogenic amino acid L-p-hydroxyphenylglycine (L-pHPG) (Figure 8(a)). Nocardicins G and E were found to be precursors of nocardicin C and A, respectively. An

286 Pathogenesis | -Lactam Antibiotics

(a) Proline carD, E CHO HO HN

H2N COOH

COOH

L-Glutamic Semialdehyde

Malonyl-CoA

CarB

H

O HO

HN COOH

5S-(Carboxymethyl)-S-Proline ATP

CarA

AMP + PPi

N O

COOH

(3S, 5S)- Carbapenam 2 KG + O2 N

CarC

O

COOH

(3S, 5S)- Carbapenam

Succinate + CO2 + HO2

6

5

7

N

O8

1 2

4

3

COOH

(5R)-Carbapenem

(b)

carI

carR

carA

carB

carC carD carE carF carG carH

Figure 7 (a) Pathway of carbapenem biosynthesis in Erwinia carotovora. The precursors (proline and malonyl-CoA) and the enzymes of the pathway are indicated. Modified from Clifton IJ et al. (2003) Crystal structure of carbapenem synthase (CarC). Journal of Biology and Chemistry 278: 20843. (b) Organization of the carbapenem gene cluster of E. carotovora. The transcriptional units are indicated with discontinuous lines, and the promoters with dots. Note that the regulatory gene carI is not located in the cluster. Modified from McGowan SJ et al. (1997) Analysis of the carbapenem gene cluster of Erwinia carotovora: definition of the antibiotic biosynthetic genes and evidence for a novel -lactam resistance mechanism. Molecular Microbiology 26 (Part B) 26: 545–546.

Pathogenesis | -Lactam Antibiotics (a)

NH2 HO

(b)

CH2 CH

NH2

COOH

HO

COOH

Nocardicin C

N-OH NH H

H2N

CH2 COOH

O

O

HOOC

CH

N-OH

N-OH

HOOC

HO

C-CO-NH-CH-CH2 CO-N-CH COOH

L-Homoserine

N H

Nocardicin A

NH

CH-CH2-CH2-O

OH

COOH

nat SAM

COOH

H2N

O

Epimerase

NH2 HO

N H

O

CH2 COOH

nocG p-OH phenylglycine transaminase

L-pHPG

OH

nocL P450

nocF p-OH mandelate synthetase HO

O

O

HOOC

CH2 CO COOH

H

NH

H2N

nocN p-OH mandelate oxidase HO

287

OH

O

O

H

OH N H COOH

Nocardicin E

L-Serine

(from SAM)

Figure 8 (a) Biosynthesis of L-p-hydroxyphenylglycine (L-pHPG) by Nocardia uniformis and condensation with other precursors, L-homoserine and L-serine, in the molecule of nocardicin A. Modified from Demain AL and Solomon NA (eds.) (2004) Antibiotics containing the -Lactam Structutre I. Berlin: Springer-Verlag, 1983. (b) Biosynthetic steps for the formation of the oxime group in nocardicin A using nocardicin C as substrate (upper part) and incorporation of the L-homoserine precursor (from SAM) using nocardicin E as substrate (lower part). Modified from Kelly WL and Townsend CA. (2005). Mutational analysis of nocK and nocL in the nocardicin A producer Nocardia uniformis. Journal of Bacteriology 187: 739.

S-adenosylmethionine-dependent enzyme catalyzes the addition of the homoseryl side chain to a phenolic group present in nocardicins E, F, and G (Figure 8(b)). This enzyme (Mr 33.2 kDa) is partially homologous to methyl transferases. It was purified to homogeneity by a combination of affinity chromatography steps to S-adenosylhomocysteine and nocardicin A-agarose. Internal peptides were used to clone the nat gene from the N. uniformis genome. The cluster of nocardicin biosynthetic genes was located around the nat gene (Figure 9). Fourteen ORFs encode hypothetical genes for resistance and export of nocardicins (nocD, encoding an N-acetyltransferase, nocH for a membrane transport protein), for the biosynthesis of the nonproteinogenic precursor L-pHPG (nocFGN), for regulation (nocR, encoding a SARP type of regulator similar to AfsR and to the putative regulator of the

calcium-dependent CDA antibiotic of S. coelicolor), and for the structural genes (nat, nocA, nocB, and nocL). Upstream of nat are the structural genes nocA and nocB, spanning 17 kb in the noc cluster, which encodes two type C NRPs of 3692 and 1925 amino acids, respectively. These proteins are required for nocardicin A formation. Based on the alignment of the domains of the encoded proteins, some modules have been predicted to activate L-pHPG, L-serine, and L-N5-hydroxyornithine. In addition, genes encoding tailoring enzymes involved in the modifications of nocardicins are present in the cluster. nocL encodes a cytochrome P450 catalyzing the formation of the syn-configurated oxime moiety of nocardicin A. Disruption of nocL results in the lack of formation of metabolites containing the oxime moiety, and their production was restored by transformation of the null mutant with nocL.

288 Pathogenesis | -Lactam Antibiotics 2 kb P450 N

R

K J IH G

Module 1 A1

nocA

T1

A2

nat D

F

Module 2 C2

nocB

Module 3 T2

C3

A3

Module 4 T3

C4

E

Module 5 A4

T4

C5

A5

T5 TE

E3 Figure 9 Cluster of genes for nocardicin A in Nocardia uniformis. Genes nocL (P450) and nat were the starting point to locate the cluster. The nocA and nocB genes encode nonribosomal peptide synthetases (NRPs). The different modules of the synthetases are shown below. Some of them have been postulated to be nonfunctional. Modified from Gunsior M, Breazeale SD, Lind AJ, Ravel J, Janc JW and Townsend CA (2004) The biosynthetic gene cluster for a monocyclic -lactam antibiotic, nocardicin A. Chemistry & Biology 11: 927.

No biosynthetic or genetic studies on the monobactams produced by unicellular bacteria have been reported. The biosynthetic precursors of tabtoxin are L-threonine, L-aspartate, pyruvic acid, and the methyl group of L-methionine. A biosynthetic gene cluster for tabtoxin has been described, which contains genes for precursors of the compounds (a tetrahydropicolinate N-succinyl transferase encoded by tabB), putative genes for tabtoxin resistance (tblD), and genes that relate this compound to other -lactams, such as a gene encoding a -lactam synthetase (tblS) and a gene (tblC) encoding a protein similar to clavaminic acid synthetase, both of them required for tabtoxin biosynthesis.

Resistance Genes in Bacterial -Lactam Clusters Most bacteria are sensitive to -lactams, and the producer bacteria might be killed by their own antibiotics. However, the -lactam-producing bacteria are somehow less sensitive to -lactams than nonproducer ones. Frequently, the -lactam clusters include genes for -lactamases and penicillin-binding proteins (PBPs). The class A -lactamase present in the cephamycin cluster of S. clavuligerus is a typical penicillinase, inactive against cephalosporins or cephamycins. S. clavuligerus is a naturally cephalosporin-resistant strain (up to 1 mg ml1 cephalosporin) but is sensitive to about 1 mg ml1 of penicillin G. The -lactamase present in the cephamycin cluster protects the producer strain against any intermediate penicillins formed in the pathway and is released by secretion or by the lysis of the cells. Disruption of the -lactamase gene alters the sensitivity of the producer strain to penicillins. Mutants of A. lactamdurans defective in the bla gene showed a modified morphology in a solid medium and produced increased levels of cephamycin C.

Also, the cluster of cephabacin biosynthesis genes in L. lactamgenus contains a class A -lactamase gene, but in this case the enzyme resembles typical cephalosporinases. In contrast, there are no -lactamase genes in the penicillin cluster of P. chrysogenum or the cephalosporin cluster of A. crhysogenum. These fungi are intrinsically resistant to -lactams. In the S. clavuligerus cephamycin cluster, there are four genes encoding putative PBPs (pcbR, pbp74, pbpA, and pbp2). In addition, two genes have been found encoding proteins with -lactamase inhibitory properties (bli and blp). When pcbR, encoding a PBP protein, was disrupted, the null mutant strains were found to have a decreased resistance to benzylpenicillin and cephalosporin. pbpA and pbp-2 have been expressed heterologously in E. coli , where only pbpA appears to confer some resistance to penicillin. In the carbapenems a gene for a -lactamase (thnS) has been found in the thienamycin cluster of S. cattleya and two genes (carFG) that, when disrupted, produce carbapenemsensitive phenotypes, in E. carotovora. In addition, the cluster of genes for -lactam biosynthesis contains genes encoding proteins that appear to be involved in secretion of the antibiotic from the producer cells. They are membrane proteins of the Mayor Facilitator Family (MFS). A. chrysogenum contains a gene, cefT1, located downstream of pcbAB in the opposite orientation encoding a protein with 12-transmembrane domains (TMS) with the characteristic motifs of Drug:Hþ antiporters of the 12-TMS class. Amplification of cefT1 results in a 100% increase in cephalosporin C production, supporting the importance of these genes in secretion of the antibiotic or regulation of its biosynthesis. Similar genes, named cmcT, are also present in the cephamycin C clusters of S. clavuligerus and A. lactamdurans. Both encode 14-TMS proteins of the MFS family, which are 73% identical between themselves. Hybridization studies using probes internal to cmcT gave positive

Pathogenesis | -Lactam Antibiotics

hybridization with total DNA from all the -lactam producers tested but not with Streptomyces lividans JI1326, a nonproducer strain. See also: Antibiotic Production; Industrial Biotechnology, (overview)

Further Reading Aharonowitz Y, Cohen G, and Martı´n JF (1992) Penicillin and cephalosporin biosynthetic genes: Structure, organization, regulation, and evolution. Annual Review of Microbiology 46: 461–495. Baggaley KH, Brown AG, and Schofield CJ (1997) Chemistry and biosynthesis of clavulanic acid and other clavams. Natural Product Reports 140: 309–333. Bibb M (1996) The regulation of antibiotic production in Streptomyces coelicolor A3(2). Microbiology 142: 1335–1344. Brakhage AA (1998) Molecular regulation of beta-lactam biosynthesis in filamentous fungi. Microbiology and Molecular Biology Reviews 62: 547–585. Coulthurst SJ, Barnard AM, and Salmond GP (2005) Regulation and biosynthesis of carbapenem antibiotics in bacteria. Nature Reviews Microbiology 3: 295–306. Dı´ez B, Mellado E, Rodrı´guez M, Fouces R, and Barredo JL (1997) Recombinant microorganisms for industrial

289

production of antibiotics. Biotechnology and Bioengineering 55: 216–226. Liras P (1999) Biosynthesis and molecular genetics of cephamycins. Cephamycins produced by actinomycetes. Antonie Van Leeuwenhoek 75: 109–124. Liras P and Rodrı´guez-Garcı´a A (2000) Clavulanic acid, a betalactamase inhibitor: Biosynthesis and molecular genetics. Applied Microbiology and Biotechnology 54: 467–475. Martı´n JF (1998) New aspects of gene and enzymes for -lactam antibiotic biosynthesis. Applied Microbiology and Biotechnology 50: 1–15. Martı´n JF (2000a) Alpha-aminoadipyl-cysteinyl-valine synthetases in beta-lactam producing organisms. From Abraham’s discoveries to novel concepts of non-ribosomal peptide synthesis. The Journal of Antibiotics 53: 1008–1021. Martı´n JF (2000b) Molecular control of expression of penicillin biosynthesis genes in fungi: Regulatory proteins interact with a bidirectional promoter region. Journal of Bacteriology 182: 2355–2362. Martı´n JF, Casqueiro J, and Liras P (2005) Secretion systems for secondary metabolites: How producer cells send out messages of intercellular communication. Current Opinion in Microbiology 8: 282–293. Martı´n JF, Gutie´rrez S, and Aparicio JF (2000) Secondary metabolites. In: Lederberg J (ed.) Encyclopedia of Microbiology, 2nd edn., vol. 4, pp. 213–236. San Diego, CA: Academic Press. Martı´n JF, Ulla´n RV, and Casqueiro FJ (2004) Novel genes involved in cephalosporin biosynthesis: The three-component isopenicillin N epimerase system. In: Brakhage A (ed.) Advances in Biochemical Engineering-Biotechnology, pp. 91–109. Berlin: Springer-Verlag.