Beta-Lactamases

b-Lactamases constitute one of the oldest known mechanisms of bacterial antibiotic resistance and one of the most widely distributed. The enzymes hydrolyze b-lactam antibiotics, such as penicillin, rendering them inactive (Fig. 1).Most of these enzymes also react with other cyclic structures, such as g-lactams and isatoic anhydride derivatives (1), and some have low reactivity with simple esters, peptides, and depsipeptides. Richmond and Sykes (2) and Bush (3) classified b-lactamases according to their catalytic activity and apparent substrate specificity, which remains the most useful classification for the clinical microbiologist. Today, the ready availability of b-lactamase amino acid sequences obtained from their genes provides the basis for a structural classification first proposed by Ambler (4). There are two types of b-lactamase: 

1.those with a serine residue at their active site that is transiently acylated by the b-lactam substrate during hydrolysis (Richmond and Sykes groups I and II; Ambler class A, plus the more recently distinguished classes C and D).

2. those with metal ion cofactors, which are Zn(II) in the physiological state, and with apparently no acyl-enzyme intermediate (Richmond and Sykes group III; Ambler class B).

The serine b-lactamases include enzymes that have both narrow and broad substrate specificities and are sensitive, to varying degrees, to b-lactam enzyme-activated inhibitors. This group is the more abundant and currently poses the greater clinical problem. The metallo-b-lactamases are active on a broad spectrum of b-lactam substrates and are much less sensitive to the mechanism-based inhibitors, although they can be inhibited by metal-ion chelating agents. This group has been relatively scarce in nature, but its prevalence is rising because of the increased use of b-lactam antibiotics that resist hydrolysis by serine b-lactamases.

Figure 1. The reaction catalyzed by b-lactamases.

1.Serine b-lactamases

 Sequence comparisons of the active serine enzymes indicate six major groups:

1.The class A b-lactamases: Richmond and Sykes Group II enzymes; Bush groups 2a (penicillinases), 2b (broad spectrum b-lactamases), 2b′ (extended broad spectrum b-lactamases), 2c (carbenicillinases), 2e (cephalasporinases). This group also includes a few proteins identified as D-Ala-D-Ala carboxypeptidases (DAC.(

2.The class C b-lactamases: Richmond and Sykes and Bush Group 1, together with three proteins from Streptomyces, Nocardia, and Bacillus subtilis, identified as carboxypeptidase/transpeptidases.

3.The Class D b-lactamases from gram-negative bacteria (Bush Group 2d oxacillinases), together with b-lactam receptor proteins from gram-positive bacteria.

4.The bifunctional transglycosylase/transpeptidase, class A penicillin-binding proteins (PBP).

5.The monofunctional transpeptidase, class B PBP.

6. D-Ala-D-Ala carboxypeptidases (DAC).

The last three groups are biosynthetic enzymes of the bacterial cell wall that react with D-Ala-D-Ala.

 All six groups have three sets of conserved residues in common, which comprise the active site of the enzyme. The first comprises the sequence Ser-X-X-Lys, where X is any residue and the serine residue is the one that is transiently acylated by substrates and the only residue that is absolutely conserved in all of the proteins. The second has either a serine (class A) or a tyrosine residue (in class C and class D b-lactamase groups), followed after one residue by an asparagine (except in one class D b-lactamase). The third conserved segment is usually Lys-Thr-Gly, but only the glycine residue is absolutely conserved. Sequence analyses have suggested that b-lactamase activity may have arisen several times by a process of evolutionary convergence.

Three-dimensional structures of class A enzymes from Staphylococcal aureus (5), Bacillus cereus (6) , B. licheniformis (7), Streptomyces albus (8), and Escherichia coli TEM-1 (9, 10) have been published, which include only two of the five activity groups of class A enzymes. Structures of the class C b-lactamases from Citrobacter freundii (11) and Enterobacter cloacae (12) have been published, and this group is extended by the homologous structure of the transpeptidase domain of the Streptomyces carboxypeptidase/transpeptidase (13). The overall structures of all of these enzymes and a DAC are similar. The core of each molecule is a five-stranded, antiparallel b-sheet flanked on one side by three a-helices and on the other by a larger, more diverse a/b domain. The active site lies between the two domains and is bound by one edge of the b-sheet. As shown in Figure 2, the conserved residues listed previously make up the active site regions and occupy very similar positions in all of the structures (Figure 2). Even the alternative serine and tyrosine residues in the second conserved segment of class A (DAC) and class C (DAC,( respectively, have their hydroxyl groups in the same positions; see Ser 130 and Tyr150 in Fig. 2 a and b, respectively. Major differences between the two classes of enzymes that are believed to have functional consequences occur in the area that serves as a recognition pocket for the side chain of the substrate. The class A enzymes have a loop (the W loop) that forms the base of the acyl amino side-chain binding pocket. This loop includes residues Glu166 and Asn170, which localize and activate a water molecule for attack on the ester bond of the substrate (14, 15). The projection of this loop into the active site also limits the scope for binding bulky acyl side chains, which was exploited in developing b-lactamase-resistant antibiotics, such as methicillin.

Figure 2. Comparison of the similar active-site regions of b-lactamases of class A (a) and class C (b). The residues are identified with the one-letter code. The lack spheres are water molecules observed crystallographically that are probably hydrogen-bonded (dashed lines) to various groups of the protein. The active-site serine residue that is acylated in the catalytic reaction is S70 in (a) and S64 in (b).

Evolutionary accumulation of peripheral mutations that modify the conformation of the W loop and of neighboring surface loops that define the edges of the side-chain recognition pocket has led to the appearance of enzymes that belong to groups 2b′, 2c, and 2e that have extended substrate recognition profiles (16). In the class C enzymes, the water molecule that attacks the ester of the acyl intermediate is located on the opposite side of the plane of the ester and is activated by direct interaction with components of a hydrogen-bonded relay formed by the conserved residues. Identification of the individual residues involved in activating of the water has not yet been achieved, although Tyr150 has been implicated (11). The class C enzymes have a deeper, more hydrophobic side-chain binding pocket, lined by Tyr122, that enables them to bind substrates with large 7-acyl amino side chains.

Conformational changes during the b-lactamase reaction have been proposed for the class A, C, and D b-lactamases on a variety of grounds. With simple substrates, NMR and circular dichroism spectroscopy have suggested changes in the structure accompanying acyl-enzyme formation (17,18). A number of substrates show nonstoichiometric bursts of hydrolysis (19), and a conformation change leading to a substrate-induced inactivation has been invoked (1, 20). These observations in solution contrast with the finding of very few differences in the structures of the free enzyme and the acyl-enzyme complex determined by X-ray crystallography (9, 11). An unambiguous description of the reaction mechanism must wait until these observations can be reconciled.

 2.Metallo-b-lactamases

Sequence comparisons suggest that there are four groups of this type of b-lactamase that are not closely related, but share similarities with a protein from the actinorhodin biosynthetic gene cluster of Streptomyces. A few residues involved in binding the metal ion cofactors are absolutely conserved within this diverse group. Three structures are available, that representing two of the major subgroups of metallo-b-lactamases (Fig. 3). All three structures are similar overall to each other, but not to the serine b-lactamases, and the core of each molecule is formed from two antiparallel b-sheets, each with additional parallel strands and flanked by a-helices. The structure of Bacillus cereus b-lactamase II was solved with Cd(II) in place of the natural Zn(II) metal ion (21). One Cd(II(ion is bound tightly, with a dissociation constant of about 1 µM, by the side chains of three histidine and one cysteine residue (22-24). In the homologous enzyme from Bacteroides fragilis, one Zn(II(ion is similarly coordinated by three histidine side chains, as in the B. cereus enzyme, but the fourth ligand is a water molecule that is shared with a second Zn(II) ion (25). The second zinc ion is further coordinated by three other side chains, those of His, Cys, and Asp residues. Two more Cd(II) ions are located in the B. cereus structure, but the coordination of one of them is much weaker (23), so it may not be physiological. Early stoichiometries determined for the binding of Zn(II) and Cd(II(suggested that two ions per molecule are bound (22). The second Cd(II) ion is bound by two Asp residues and lies only 10 Å from the tightly bound ion (Fig. 3), where it might modulate the catalytic activity. It is believed that the water molecule complexed with the two Zn(II) ions in the B. fragilis enzyme is activated to form hydroxide for attack on the b-lactam ring, and a corresponding water molecule is probably activated by the tightly bound ion in the B. cereus enzyme.

Figure 3. Comparison of the similar active-site regions of metallo-b-lactamases from Bacillus cereus (a) and Bacteriodes fragilis (b). The shaded spheres are the two metal ions, cadmium (II) in (a) and zinc (II) in (b), and the smaller solid spheres are water molecules observed crystallographically.

3.Inhibition

 The important role of b-lactamases in antibiotic resistance has led to the development of specific b-lactamase inhibitors that can be used to protect b-lactamase-sensitive antibiotics. For the serine b-lactamases, several classes of mechanism-based inhibitors have been discovered, and several inhibitor proteins are known (26). In many cases, the mechanism of action of the small-molecule inhibitors is not clear, and only two classes have been used in combination with antibiotics.

Clavulanic acid has been widely used clinical application. It is selective for the class A and class D b-lactamases. Opening of the b-lactam ring of clavulanic acid by the initial attack of the b-lactamase triggers a series of chemical rearrangements in the inhibitor moiety (Fig. 4) that results in an acyl-enzyme that is resistant to attack by water and may even be cross-linked by the inhibitor moiety (27).

Figure 4. Suggested mechanism for the reaction of TEM-2 class A b-lactamases with clavulanic acid (27).

The penam sulfone acids sulbactam and tazobactam are also used clinically as b-lactamase inhibitors. As with clavulanic acid, a series of chemical rearrangements are provoked by the reaction with b-lactamase. Although sulbactam and tazobactam are relatively selective for class A and class D b-lactamases, analogs that have increased activity against class C b-lactamases have been reported (28) .

4. Regulation of Biosynthesis

Expression of chromosomal b-lactamase in Bacillus and of plasmid-encoded b-lactamases in staphylococci is under the control of a typical repressor protein (Blal and Mecl, respectively) and a second regulatory protein (BlaR and MecR, respectively). The BlaR and MecR regulatory proteins are similar. Both are integral membrane proteins that have a b-lactam-binding domain homologous to the class D serine b-lactamases (29). The predicted topology of the polypeptide in the membrane is three to five membrane-spanning a-helices, and the N-terminus in the cytoplasm and the b-lactam-binding domain is on the outer surface of the membrane, where it could act as a receptor for b-lactams.

Expression of chromosomal class C b-lactamases in Enterobacteriaceae from the ampC gene is under the control of the ampD, ampE, ampG, and ampR gene products. AmpR protein is a typical transcription activator that is lost in organisms, such as E. coli, that have constitutive AmpC production. AmpR responds to the binding of 1,6-anhydro-N-acetylmuramyl-[L]-alanyl-[D]-glutamyl-meso-diaminopimelic           acid (MurNAc tripeptide) produced by breakdown of the peptidoglycan of the cell wall (30). The three other proteins are involved in transmembrane signaling for induction of the b-lactamase. Inactivation of AmpD results in massive overproduction of b-lactamase, whereas loss of either AmpE or AmpG activity results in a total block of induction.

AmpD is an amidase that cleaves the tripeptide from the MurNAc tripeptide, thus inactivating it as an inducer. It is thought that AmpG is an integral membrane protein that acts as a transporter for the MurNAc tripeptide, whereas it has been suggested that AmpE, also an integral membrane protein, provides energy for its uptake (31).

References

1.M. G. P. Page (1993) Biochem. J. 295, 295–304. 

2.M. H. Richmond and R. B. Sykes (1973) Adv. Microb. Physiol. 9, 31–88. 

3.K. Bush (1989) Antimicrob. Agents Chemother. 33, 264–270, 271–276.

4.R. P. Ambler (1980) Phil. Trans. R. Soc. Lond. B239, 321–331. 

5.O. Herzberg and J. Moult (1987) Science 236, 694–701. 

6زB. Samaraoui, B. Sutton, R. Todd, P. Artimyuk, S. G. Waley, and D. Phillips (1986) Nature 320, 378-380.

7.P. C. Moews, J. R. Knox, O. Dideberg, P. Charlier, and J.-M. Frère (1990) Proteins: Struct. Function Genet. 7, 156–171. 

8.O. Dideberg, P. Charlier, J. P. Wery, P. Dehottay, J. Dusart, T. Erpicum, J.-M. Frère, and J.-M. Ghuysen (1987) Biochem. J. 245, 911–913. 

9.N. C. J. Strynadka, H. Adachi, S. E. Jensen, K. Johns, A. Sielecki, C. Betzel. K. Sutoh, and M. N. G. James (1992) Nature 359, 700–705. 

10.C. Jelsch, L. Mourey, J. M. Masson, and J.-P. Samama (1992) Proteins: Struct. Function Genet. 16,364-383.

11.C. Oefner, A. D''Arcy, J. J. Daly, K. Gubernator, R. L. Charnas, I. Heinze, C. Hubschwerlen, and F. K. Winkler (1990) Nature 343, 284–288. 

12.E. Lobkovsky, P. C. Moews, J. Liu, H. Zhao, J.-M. Frère, and J. R. Knox (1993) Proc. Natl. Acad. Sci. USA 90, 11257–11261. 

13.J. A. Kelly and A. P. Kuzin (1995) J. Mol. Biol. 254, 223–236. 

14.R. M. Gibson, H. Christensen, and S. G. Waley (1990) Biochem. J. 272, 613–619. 

15.W. A. Escobar, A. K. Tan, and A. L. Fink (1991) Biochemistry 30, 10783–10787. 

16.G. A. Jacoby and A. A. Medeiros (1991) Antimicrob. Agents Chemother. 35, 1697–1704. 

17.M. Jamin, C. Damblon, A. M. Baudin-Misselyn, F. Durant, G. C. K. Roberts, P. Charlier, G. Llabres, J.-M. Frère. (1994) Biochem. J. 301, 199–203. 

18.P. Taibi-Tonche, I. Massova, S. B. Vakulenko, S. A. Lerner, and S. Mobashery (1996) J. Am. Chem. Soc. 118, 7441–7448. 

19.A. Samuni and N. Citri (1975) Biochem. Biophys. Res. Commun. 62, 7–11. 

20.S. G. Waley (1991) Biochem. J. 279, 87–94. 

21.B. J. Sutton, P. J. Artymiuk, A. E. Cordero-Borboa, C. Little, D. C. Phillips, and S. G. Waley (1987) Biochem. J. 248, 181–188. 

22.R. B. Davies and E. P. Abraham (1974) Biochem. J. 143, 129–135. 

23.G. S. Baldwin, A. Galdes, A. O. Hill, B. E. Smith, S. G. Waley, and E. P. Abraham (1978(Biochem. J. 175, 441–447. 

24.A. Carfi and O. Dideberg (1995) EMBO J. 14, 4919–4921. 

25.N. O. Concha, B. A. Rasmussen, K. Bush, and O. Herzberg (1996) Structure 4, 823–836. 

26.N. C. J. Strynadka, S. E. Jensen, K. Johns, H. Blanchard, M. Page, A. Matagne, J.-M. Frère, and M. N. G. James (1994) Nature 368, 657–660. 

27.R. P. A. Brown, R. T. Aplin, and C. J. Schofield (1996) Biochemistry 35, 12421–12432. 

28.H. G. F. Richter, P. Angehrn, C. Hubschwerlen, M. Kania, M. G. P. Page, J.-L. Specklin, and F. K. Winkler (1996) J. Med. Chem. 39, 3712–3722. 

29. T. Kobayashi, Y. F. Zhu, N. J. Nicholls, and J. O. Lampen (1987) J. Bacteriol. 169, 3873–3878.

30.J. T. Park (1995) Mol. Microbiol. 17, 4521–426. 

31. S. Lindquist, M. Galleni, F. Lindberg, and S. Normark (1989) Mol. Microbiol. 3, 1091–1102.