Journal of Antimicrobial Chemotherapy (1977) 3 (Suppl. C), 7-17
Thoughts on the origins of resistance plasmids
Julian Davies, Patrice Courvalin and Douglas Berg
Introduction
The presence of a plasmid can confer enormous advantages to a bacterium. The plasmid can provide additional catabolic potential, often for exotic substrates; resistance to a wide variety of antibiotics, bacteriophages and other cell predators; and in some cases the selective advantage of a higher spontaneous mutation rate (Falkow, 1975). In all probability plasmids code for many more useful characters—at present we can account for only a small proportion of the genetic functions of a plasmid (as little as 20%); thus there are hundreds of genes, carried around on plasmids, that remain to be identified and must presumably have positive survival value for bacteria that possess them. These genes must be 'picked up' during the passage of plasmids from host to host acting as a collection agency for anything that may be useful in its next home. We know very little about where the genes on plasmids come from, how they are picked up, and even less about why they are picked up, e.g., what selective forces are involved. We have some reasonable notions of the origins and mechanisms of construction of plasmids, and these are the topics we would like to discuss. Plasmid-determined antibiotic resistance
The study of plasmid-determined antibiotic resistance has been profitable with respect to ideas about the origins of plasmid genes and their collection and assortment. In particular, studies of resistance to the aminoglycoside antibiotics, possessing the widest range of associated resistance determinants, have been most instructive. Resistance to the aminoglycoside antibiotics is determined by the presence of plasmid-coded enzymes (Table I) that modify the antibiotic in the outer confines of the cell, thereby preventing its transport to its target of action, the ribosome (Haas & Dowding, 1975). At present, kanamycin is subject to at least six different modifications (Figure 1). It was thought, formerly, that resistance to aminoglycoside antibiotics was simply enzyme-catalyzed inactivation, as is the case with the penicillins; we now know that only a small proportion of the aminoglycoside in the medium has to be modified by the resistance enzyme in order to allow the full expression of resistance.
Downloaded from http://jac.oxfordjournals.org/ at Université Laval on July 14, 2015
Department of Biochemistry, University of Wisconsin, Madison, Wisconsin, U.S.A.
J. Davies, P. Courvalin and D. Berg
8
How can we describe the role of the modifying enzymes in the resistance mechanism? We believe that these enzymes are present on the inner membrane of the cell and (probably) closely associated with the transport system for the antibiotic; in some way modification of the entering antibiotic blocks this transport system as is indicated in Figure 2. We have no evidence for this model other than that the modifying enzyme is Table I. The plasmid-determined modifying enzymes associated with resistance to aminoglycoside antibiotics Enzymes that modify aminoglycoside-aminocyclitol antibiotics Enzyme Typical substrates neomycin, kanamycin streptomycin ribostamycin gentamicin
2' '-O-adenylyltransferase 4'-O-adenylyltransferase 3 "-O-adenylyltransferase 6'-O-adenylyltransferase
gentamicin, tobramycin amikacin, tobramycin streptomycin, spectinomycin streptomycin
6'-A'-acetyltransferase 2 '-A'-acetyltransferase 3 '-A'-acetyltransferase
amikacin, tobramycin gentamicin, tobramycin gentamicin, tobramycin
CH 2 NH 2 ~*
acetylation - 0
odenylylation —*- HO
NH 2 -«
acetylation
phosphorylation
adenylylatlon phosphorylation
Kanamycin Figure 1. The sites of enzymatic modification of kanamycin A by resistant strains.
required and total inactivation of the antibiotic does not occur; in fact we cannot detect any modified antibiotic in the culture medium. At present, little is known of the transport systems for antibiotics such as aminoglycosides although Bryan and his collaborators have made significant advances in this respect (Bryan et al., 1975, 1976; Bryan & van den Elzen, 1975). Because aminoglycoside antibiotics are, in a medical sense, in vogue—they are used extensively in the treatment of a variety of hospital-acquired Gram-negative and Grampositive infections—new drugs of this type have been introduced in number in the past
Downloaded from http://jac.oxfordjournals.org/ at Université Laval on July 14, 2015
3' -O-phosphotransferase 3' '-O-phosphotransferase 5' '-O-phosphotransferase 2' '-O-phosphotransferase
Origins of resistance plasmids Sensitive
Outer membrane
I D
mJmbVane
Cell cytoplasm
Culture medium
Energy - requiring transport system
Ab -
rlbosome
Cell cytoplasm
Resistant ( R + ) Culture medium
\
Figure 2. A proposal for a mechanism by which enzymatic modification of aminoglycoside antibiotics could lead to a block in antibiotic transport into the cell.
Table II. Substrate ranges for APH-2" (gentamicin phosphotransferase) and AAD-4' (amikacin adenylyltransferase), the 'newest' aminoglycosidemodifying enzymes Substrate Gentamicin d . Amikacin (BBK-8) Kanamycin A Tobramycin Neomycin Butirosin Ribostamycin 4 '-deoxybutirosin 2"-deoxygentamicin C« epi-(2") gentamicin Ci Netilmicin
Activity* with: APH-2" AAD-4' 100 35 25 15 0 0 0 0 0 3 100
0 100 113 95 31 124 113 0 0 0 0
' Expressed as percentage of gentamicin C,, or amikacin as substrate, respectively.
few years and it has been possible to watch the development of resistance in hospital bacteria subsequent to the introduction of each novel aminoglycoside. For example, two new aminoglycoside-modifying enzymes have been described in the past year in clinical isolates of Staphylococcus aureus from France, Great Britain, Switzerland, and the U.S. (Le Gofficetal., 1976; Santanam & Kayser, 1977; Davies & Hoffman, 1977). The substrate ranges of these enzymes indicate clearly the point of attack of the enzyme and also give some idea of their substrate specificity (Table II). For example, although
Downloaded from http://jac.oxfordjournals.org/ at Université Laval on July 14, 2015
Ab-
10
J. Davies, P. Courvalin and D. Berg
gentamicin phosphotransferase (APH-2") clearly distinguishes 4,6-substituted deoxystreptamine antibiotics (gentamicin, kanamycin) from 4,5-substituted deoxystreptamines (neomycin, butirosin), the enzyme also delineates between molecules with kanosamine and garosamine substituents. The structural features responsible for this difference in activity are not known; it is presumably a difference in rate (Km) that is involved. The rate of modification of an aminoglycoside is of great importance. Williams & Northrop (1975) who described the first complete purification of an aminoglycoside-modifying enzyme, gentamicin acetyltransferase, have shown that the enzyme can acetylate both gentamicin and tobramycin in vitro. However, strains possessing the enzyme were resistant to gentamicin and sensitive to tobramycin! The answer to this dilemma appears to lie in the fact that the Km for gentamicin is 10 times lower than that for tobramycin (Williams & Northrop, personal communication); we can conclude that in contrast to gentamicin, tobramycin is not inactivated fast enough to set up a permeability block to prevent the drug entering the cell (Figure 2).
Since the aminoglycoside-modifying enzymes constitute a large group of sugar-modifying enzymes that are not normally associated with the bacteria that develop resistance to the aminoglycosides, we considered the possibility that these enzymes may have their origins Table HI. Aminoglvcoside-aminocyclitol modifying enzymes found in antibiotic producing strains 3 '-O-phosphotransferase 3' '-O-phosphotransferase 6 '-JV-acetyltransferase 2 '-.N-acetyltransferase 3-JV-acetyltransferase
in the bacteria that produce these antibiotics (Benveniste & Davies, 1973; Dowding & Davies, 1975). A search for these enzymes in Actinomycetes and other bacteria that produce neomycin, gentamicin and related compounds has revealed that JV-acetyltransferases and Ophosphotransferases are common (Table III). Thus far, no O-adenylyltransferases have been found outside of R + strains. The fact that no aminoglycoside
Thoughts on the origins of resistance plasmids
Julian Davies, Patrice Courvalin and Douglas Berg
Introduction
The presence of a plasmid can confer enormous advantages to a bacterium. The plasmid can provide additional catabolic potential, often for exotic substrates; resistance to a wide variety of antibiotics, bacteriophages and other cell predators; and in some cases the selective advantage of a higher spontaneous mutation rate (Falkow, 1975). In all probability plasmids code for many more useful characters—at present we can account for only a small proportion of the genetic functions of a plasmid (as little as 20%); thus there are hundreds of genes, carried around on plasmids, that remain to be identified and must presumably have positive survival value for bacteria that possess them. These genes must be 'picked up' during the passage of plasmids from host to host acting as a collection agency for anything that may be useful in its next home. We know very little about where the genes on plasmids come from, how they are picked up, and even less about why they are picked up, e.g., what selective forces are involved. We have some reasonable notions of the origins and mechanisms of construction of plasmids, and these are the topics we would like to discuss. Plasmid-determined antibiotic resistance
The study of plasmid-determined antibiotic resistance has been profitable with respect to ideas about the origins of plasmid genes and their collection and assortment. In particular, studies of resistance to the aminoglycoside antibiotics, possessing the widest range of associated resistance determinants, have been most instructive. Resistance to the aminoglycoside antibiotics is determined by the presence of plasmid-coded enzymes (Table I) that modify the antibiotic in the outer confines of the cell, thereby preventing its transport to its target of action, the ribosome (Haas & Dowding, 1975). At present, kanamycin is subject to at least six different modifications (Figure 1). It was thought, formerly, that resistance to aminoglycoside antibiotics was simply enzyme-catalyzed inactivation, as is the case with the penicillins; we now know that only a small proportion of the aminoglycoside in the medium has to be modified by the resistance enzyme in order to allow the full expression of resistance.
Downloaded from http://jac.oxfordjournals.org/ at Université Laval on July 14, 2015
Department of Biochemistry, University of Wisconsin, Madison, Wisconsin, U.S.A.
J. Davies, P. Courvalin and D. Berg
8
How can we describe the role of the modifying enzymes in the resistance mechanism? We believe that these enzymes are present on the inner membrane of the cell and (probably) closely associated with the transport system for the antibiotic; in some way modification of the entering antibiotic blocks this transport system as is indicated in Figure 2. We have no evidence for this model other than that the modifying enzyme is Table I. The plasmid-determined modifying enzymes associated with resistance to aminoglycoside antibiotics Enzymes that modify aminoglycoside-aminocyclitol antibiotics Enzyme Typical substrates neomycin, kanamycin streptomycin ribostamycin gentamicin
2' '-O-adenylyltransferase 4'-O-adenylyltransferase 3 "-O-adenylyltransferase 6'-O-adenylyltransferase
gentamicin, tobramycin amikacin, tobramycin streptomycin, spectinomycin streptomycin
6'-A'-acetyltransferase 2 '-A'-acetyltransferase 3 '-A'-acetyltransferase
amikacin, tobramycin gentamicin, tobramycin gentamicin, tobramycin
CH 2 NH 2 ~*
acetylation - 0
odenylylation —*- HO
NH 2 -«
acetylation
phosphorylation
adenylylatlon phosphorylation
Kanamycin Figure 1. The sites of enzymatic modification of kanamycin A by resistant strains.
required and total inactivation of the antibiotic does not occur; in fact we cannot detect any modified antibiotic in the culture medium. At present, little is known of the transport systems for antibiotics such as aminoglycosides although Bryan and his collaborators have made significant advances in this respect (Bryan et al., 1975, 1976; Bryan & van den Elzen, 1975). Because aminoglycoside antibiotics are, in a medical sense, in vogue—they are used extensively in the treatment of a variety of hospital-acquired Gram-negative and Grampositive infections—new drugs of this type have been introduced in number in the past
Downloaded from http://jac.oxfordjournals.org/ at Université Laval on July 14, 2015
3' -O-phosphotransferase 3' '-O-phosphotransferase 5' '-O-phosphotransferase 2' '-O-phosphotransferase
Origins of resistance plasmids Sensitive
Outer membrane
I D
mJmbVane
Cell cytoplasm
Culture medium
Energy - requiring transport system
Ab -
rlbosome
Cell cytoplasm
Resistant ( R + ) Culture medium
\
Figure 2. A proposal for a mechanism by which enzymatic modification of aminoglycoside antibiotics could lead to a block in antibiotic transport into the cell.
Table II. Substrate ranges for APH-2" (gentamicin phosphotransferase) and AAD-4' (amikacin adenylyltransferase), the 'newest' aminoglycosidemodifying enzymes Substrate Gentamicin d . Amikacin (BBK-8) Kanamycin A Tobramycin Neomycin Butirosin Ribostamycin 4 '-deoxybutirosin 2"-deoxygentamicin C« epi-(2") gentamicin Ci Netilmicin
Activity* with: APH-2" AAD-4' 100 35 25 15 0 0 0 0 0 3 100
0 100 113 95 31 124 113 0 0 0 0
' Expressed as percentage of gentamicin C,, or amikacin as substrate, respectively.
few years and it has been possible to watch the development of resistance in hospital bacteria subsequent to the introduction of each novel aminoglycoside. For example, two new aminoglycoside-modifying enzymes have been described in the past year in clinical isolates of Staphylococcus aureus from France, Great Britain, Switzerland, and the U.S. (Le Gofficetal., 1976; Santanam & Kayser, 1977; Davies & Hoffman, 1977). The substrate ranges of these enzymes indicate clearly the point of attack of the enzyme and also give some idea of their substrate specificity (Table II). For example, although
Downloaded from http://jac.oxfordjournals.org/ at Université Laval on July 14, 2015
Ab-
10
J. Davies, P. Courvalin and D. Berg
gentamicin phosphotransferase (APH-2") clearly distinguishes 4,6-substituted deoxystreptamine antibiotics (gentamicin, kanamycin) from 4,5-substituted deoxystreptamines (neomycin, butirosin), the enzyme also delineates between molecules with kanosamine and garosamine substituents. The structural features responsible for this difference in activity are not known; it is presumably a difference in rate (Km) that is involved. The rate of modification of an aminoglycoside is of great importance. Williams & Northrop (1975) who described the first complete purification of an aminoglycoside-modifying enzyme, gentamicin acetyltransferase, have shown that the enzyme can acetylate both gentamicin and tobramycin in vitro. However, strains possessing the enzyme were resistant to gentamicin and sensitive to tobramycin! The answer to this dilemma appears to lie in the fact that the Km for gentamicin is 10 times lower than that for tobramycin (Williams & Northrop, personal communication); we can conclude that in contrast to gentamicin, tobramycin is not inactivated fast enough to set up a permeability block to prevent the drug entering the cell (Figure 2).
Since the aminoglycoside-modifying enzymes constitute a large group of sugar-modifying enzymes that are not normally associated with the bacteria that develop resistance to the aminoglycosides, we considered the possibility that these enzymes may have their origins Table HI. Aminoglvcoside-aminocyclitol modifying enzymes found in antibiotic producing strains 3 '-O-phosphotransferase 3' '-O-phosphotransferase 6 '-JV-acetyltransferase 2 '-.N-acetyltransferase 3-JV-acetyltransferase
in the bacteria that produce these antibiotics (Benveniste & Davies, 1973; Dowding & Davies, 1975). A search for these enzymes in Actinomycetes and other bacteria that produce neomycin, gentamicin and related compounds has revealed that JV-acetyltransferases and Ophosphotransferases are common (Table III). Thus far, no O-adenylyltransferases have been found outside of R + strains. The fact that no aminoglycoside
