Studies on Mutational Resistance to Nalidixic Acid in
Escherichia coli.
Thesis submitted to the University of London by James Hinton for the degree of Doctor of Philosophy October 2000.
Microbiology Section, Department of Pharmaceutics, The School of Pharmacy, Brunswick Square, London. ProQuest Number: 10105159
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest.
ProQuest 10105159
Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author.
All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code. Microform Edition © ProQuest LLC.
ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 Acknowledgements
I would like to thank my supervisor Dr. R. J. Pinney for his continued and substantial support and advice throughout this work.
I am also most thankful to Prof. S.G.B. Amyes and Esther Durham, and all in the
Microbiology Department at the University of Edinburgh for all of their assistance with the PCR and Hinjl restriction experiments. I would also like to thank them for the DNA sequencing of my Escherichia coli strains.
I should also like to thank Marianne and Kate for their technical assistance.
Finally, I would like to thank the School of Pharmacy, University of London for their financial support. Abstract
E. coli ABl 157 cultures incubated for ten days on solid media containing 4 or 8 yug mL ' nalidixic acid (NA) (one or two times the MIC respectively, with the latter concentration being bactericidal) showed an increased apparent mutation frequency to
NA resistance. The presence of the mutator plasmids R46 and R446b further increased the number of NA-resistant mutants recovered. Growth ofE. coli ABl 157 in liquid medium containing 2 or 3 yUg mL ' NA (0.5 and 0.75 x MIC) for ten days increased the number of mutants resistant to 16 yug mL ' NA (4 x MIC), compared to control cultures grown in drug-free broth. Again, R46 and R446b greatly enhanced this effect. As a control, no increase in reversion of the unselected allele,thr-1 was detected during ten days growth in broth containing 0,2 or 3 yUg mL ' NA. These results suggest that the presence of NA might induce mutational resistance to itself via error-prone repair. Strains carrying defective recAlS, umuC36, lexA3 or recB21 alleles, deficient in error-prone repair, displayed a reduction in the numbers of NA resistant mutants recovered after four days incubation in broth containing 0.75 x MIC NA. The presence of a defective uvrA gene led to an increase in the number of NA-resistant mutants recovered, suggesting enhanced error-prone repair of NA-induced DNA lesions in the absence of of error-free excision repair.
Clinically important quinolone-resistance mutations usually occur at position Ser-
83 in the gyrA gene, within the highly conserved quinolone resistance-determining region
(QRDR). Sequence analysis of the QRDR of four quinolone-resistant ABl 157 clones revealed only single changes. Three had mutated at the Ser-83 position, and one showed variation at position Asp-87. It would therefore appear that NA-induced mutation to NA resistance selected in vitro may well be clinically significant. Contents
Page
Acknowledgements 2
Abstract 3
Contents 4
Abbreviations and Symbols 14
List of Figures 19
List of Tables 23
INTRODUCTION 26
1. Antibiotic resistance 26
1.1. Intrinsic resistance 26
1.2. Acquired resistance 26
1.3. Mutational antibiotic resistance 27
1.3.1. Rifampicin resistance 27
1.3.2. Fusidic acid resistance 29
1.3.3. Methicillin resistance in Staphylococcus aureus 3 0
1.3.4. Streptomycin resistance 31
1.3.5. Sulphonamide resistance 32
1.3.6. Trimethoprim resistance 33
2. The quinolones 34
2.1. DNA gyrase 37
2.2. DNA topoisomerase IV 38
2.3. Mode of action of quinolone antibacterials 39 2.4. Quinolone lethality 40
2.5. DNA breaks and cell death 41
2.6. Resistance to quinolones 42
2.6.1. Mutations in the gyrA gene 42
2.6.2. Mutations in the gyrB gene 44
2.6.3. Mutations affecting topoisomerase IV 45
2.6.4. Quinolone resistance resulting from drug impermeability 47
2.7. The absence of plasmid-mediated quinolone resistance 49
2.8. Plasmid elimination by quinolones 49
3. DNA damage and repair 50
3.1. Photoreactivation 51
3.2. Nucleotide excision repair 52
3.2.1. uvrA gene and product 52
3.2.2. uvrB gene and product 53
3.2.3. uvrC gene and product 53
3.2.4. Binding of Uvr proteins to DNA lesions 53
3.2.5. Incision of the phosphate backbone 54
3.2.6. Post-incision events 54
3.3. Methyl-directed mismatch repair 55
3.3.1. Strand discrimination between the parent and newly synthesised daughter strand 56
3.3.2. Requirements for the mutH, mutL, mutS and uvrD gene products for methyl-directed mismatch repair 56
3.3.3. Initiation of mismatch repair 58 3.3.4. Excision of the mismatch 58
3.3.5. Repair of daughter strand gaps 59
3.4. Post-replication recombination repair 59
3.4.1. The role of the RecBCD enzyme in PRR repair 59
3.4.2. The RecF repair pathway 60
3.4.3. Model for the RecBCD repair pathway 60
3.4.4. Search for homology and strand exchange 61
3.4.5. RuvAB and RecG 62
3.5. The SOS response 63
3.5.1. The SOS regulon 64
3.5.2. Regulation of the SOS response 64
3.5.3. The SOS-inducing signal and RecA protein activation 67
3.5.4. RecA*-mediated proteolytic cleavage of the LexA repressor 68
3.5.5. SOS processing 68
3.5.6. SOS mutagenesis 70
3.5.7. DNA polymerases 71
3.5.8. Reduced fidelity of DNA polymerase III 71
3.5.9. Induction of the SOS response by the quinolone antibacterials 73
3.5.10. Repair of quinolone-induced DNA lesions 74
4. Plasmids 76
4.1. Plasmid-encoded mutation-enhancing opérons 76
4.2. Relationship of plasmid-encoded mutation-enhancing systems to the E. coli chromosomalumuDC operon 77 4.3. mucAB 78
4.3.1. Processing of the MucA protein 78
4.3.2. MucAB-mediated mutagenesis 79
4.4. Spectrum of activity of plasmid-encoded mutation- enhancing opérons 81
5. Spontaneous mutation and natural selection 83
5.1. The origin of adaptive mutation 84
5.2. The Lac frameshift assay 87
5.3. Adaptive mutation via the excision of mobile genetic elements 90
5.3.1. The Shapiro deletion 91
5.3.2. Excision from thebgl operon 93
5.4. Prerequisites for the generation of adaptive mutants 96
5.5. Genetic requirements for adaptive mutation 97
S.SA.recA 97
5.5.2. lexA and umu genes 99
5.5.3. Mismatch DNA repair 100
5.5.4. Nucleotide excision repair 102
5.5.5. RecG and RuvABC 102
5.5.6. Plasmid transfer functions 103
5.5.7. DNA polymerases 104
5.6. Proposed mechanisms 105
5.6.1. Variant RNAs 105
5.6.2. Variant DNAs 106
5.6.3. The hypermutable state 107 6. Aims of this Thesis 108
MATERIALS 110
1. Bacterial strains and plasmids 110
2. Chemicals and biochemicals 112
3. Solutions for Polymerase Chaim Reaction (PCR) analysis 115
4. Antibacterials 115
5. Media 116
5.1. Complex media 116
5.1.1. Nutrient broth and nutrient agar 116
5.1.2. MacConkey agar 117
5.1.3. Iso-Sensitest agar 117
5.2. Minimal salts medium 117
5.2.1. Davis and Mingioli salts solution 117
5.2.2. DM minimal agar 118
5.2.3. Agar base plates and ‘soft’ agar overlays used to determine UV-induced reversion to amino acid prototrophy 119
6. Apparatus 120
METHODS 122
1. Overnight cultures 122
2. Viable counts 122
3. Plasmid transfers 122
4. Determination of Minimum Inhibitory Concentrations (MICs) 124
4.1. Preparation of drug-containing plates 124 4.2. Preparation of inocula 124
5. Ultraviolet sensitivity testing 125
6. Reversion to amino acid prototrophy 125
7. Forward mutation to nalidixic acid resistance on solid media 126
7.1. Reversion to threonine independence following growth on nalidixic acid-containing medium 127
8. UV-induced mutagenesi s 127
8.1. UV-induced reversion to amino acid prototrophy 127
8.2. UV-induced mutation to nalidixic acid resistance 128
9. Mutation to nalidixic acid or ciprofloxacin resistance in cultures grown in sub-inhibitory concentrations of nalidixic acid or ciprofloxacin 129
10. Polymerase Chain Reaction (PCR) amplification of nalidixic acid- and ciprofloxacin-resistant clones 131
10.1. Gel electrophoresis of the amplified PCR product 132
10.1.1. Preparation of the gel 132
10. 1.2. Loading the gel 133
11. Hinjl restriction analysis of the PCR product 133
11.1. Gel electrophoresis ofHinfl. digested PCR products 134
11.1.1. Preparation of the gel 134
11.1.2. Loading the gel 134
RESULTS 136
1. Preparatory experiments 136
1.1. Plasmid transfer 136
1.2. Post-UV survival 138 1.2.1. The effect of plasmids on post-UV survival of repair- proficientE. coli AB1157 138
1.2.2. The effect of repair deficiencies on post-UV survival 139
1.3. Determination of nalidixic acid MICs 144
1.3.1. The effect of plasmids on MIC to nalidixic acid in E. coli AB 1157 and its plasmid-bearing derivatives 144
1.3.2. The effect of repair deficiencies on MICs of nalidixic acid 146
1.4. Summary and conclusions 148
2. Reversion to amino acid independence 149
2.1. Introduction 149
2.2. Spontaneous and UV-induced reversion to amino acid independence 150
2.3. Selection-induced reversion to amino acid independence 154
2.4. Summary and conclusions 158
3. Selection-induced mutation to nalidixic acid resistance on solid media 159
3.1. Introduction 159
3.2. Spontaneous and UV-induced mutation to nalidixic acid resistance 160
3.3. Selection-induced mutation to nalidixic acid resistance on solid media 164
3.4. Summary and conclusions 168
4. Selection-induced mutation to quinolone resistance in cultures grown in liquid medium containing sub-inhibitory concentrations of drug 170
4.1. Introduction 170
10 4.2. Mutation to nalidixic acid resistance in E. coli cultures grown in liquid medium containing sub-inhibitory concentrations of nalidixic acid 170
4.2.1. The effect of growth in sub-inhibitory concentrations of nalidixic acid on the development of nalidixic acid resistance in plasmid-ffee E. coli AB 1157 171
4.2.2. The effect of mutator plasmids on the development of nalidixic acid resistance 173
4.2.2.1. The effect of mutator plasmids on the development of nalidixic acid resistance in cultures grown in drug-free broth 173
4.2.2.2. The effect of mutator plasmids on the development of nalidixic acid resistance in cultures grown in sub- inhibitory concentrations of nalidixic acid 174
4.3. Reversion to threonine independence following growth in liquid medium containing sub-inhibitory concentrations of nalidixic acid 178
4.4. Summary and conclusions 182
5. Selection-induced mutation to nalidixic acid resistance in repair- deficient strains of E. coli 183
5.1. Introduction 183
5.2. Mutation to nalidixic acid resistance in repair-deficient strains ofE. coli 183
5.3. Summary and conclusions 187
6. Determination of quinolone MICs of quinolone-resistant mutants 188
6.1. Introduction 188
6.2. Mutation to ciprofloxacin resistance in cultures grown in sub-inhibitory concentrations of ciprofloxacin 188
6.3. Determination of quinolone MICs of the nalidixic acid- and ciprofloxacin-resistant mutants 191
11 6.4. Growth of low-level quinolone-resistant clones in nutrient broth containing sub-inhibitory concentrations of nalidixic acid or ciprofloxacin selects for high level quinolone- resistant mutants 198
6.5. Summary and conclusions 204
7. Determination of mutations within QRDRs of quinolone- resistant clones 206
7.1. Introduction 206
7.2. Detection of mutations in the QRDR ofgyrA in quinolone- resistant E. coli 206
7.2.1. Restriction fragment length polymorphism (RFLP) of the QRDR ofgyrA of quinolone-resistantE. coli ABl 157 clones 208
7.2.2. DNA sequencing of the amplified 620 bp region of gyrA incorporating the QRDR ofE. coli AB 1157 211
7.3. Summary and conclusions 213
DISCUSSION 215
1. The effect of mutator plasmids on the post-UV survival ofE. coli AB 1157 and repair-deficient derivatives 217
2. Selection-induced adaptive mutation to amino acid independence 218
3. Selection-induced mutation to nalidixic acid resistance in E. coli AB 1157 cultures plated on medium containing the MIC or supra- inhibitory (lethal) concentrations of nalidixic acid 220
4. Mutagenesis to nalidixic acid resistance in E. coli ABl 157 cultures during incubation in nutrient broth containing sub-inhibitory concentrations of nalidixic acid 223
5. Examination of unselected mutagenesis following growth in liquid medium containing sub-inhibitory concentrations of nalidixic acid 225
6. Dependence of adaptive mutation to nalidixic acid resistance and R46- mediated effects on host DNA repair genotype 226
12 7. MIC determination and genetic analyses of quinolone-resistant clones following growth in sub-inhibitory concentrations of quinolone 229
8. Summary and conclusions 233
REFERENCES 235
13 Abbreviations and symbols
adenine
Ala/ala alanine
Ap ampicillin
Ara/ara arabinose
Arg/arg arginine
Asn/asn asparagine
Asp/asp aspartate
ATP adenosine triphosphate bgl P-glucosidase bp base pair
C cytosine
Cip ciprofloxacin
Cm chloramphenicol
Cone. concentration
-COOH carboxyl group
Cys/cys cysteine
Da daltons
DHFR dihydrofolate reductase
DHPS dihydropteroate synthetase din DNA damage inducible
DM Davis and Mingioli minimal salts solution/agar
DNA deoxyribonucleic acid
EDTA ethylendiaminetetra-acetic acid
14 EF-G elongation factor G
ExoV exonuclease V
G guanine g grammes gal galactose
GDP guanosine diphosphate
Gln/gln glycine
Glu/glu glutamate
Gly/gly glycine
GTP guanosine triphosphate
H2O water
Hg mercuric ions
His/his histidine
Ile/ile isoleucine
Inc incompatibility group
IRR induced replisome reactivation
J m'^ joules per square metre
J s ' joules per square metre per second kb kilobase
KCl potassium chloride kDa kilodalton
Km kanamycin
L litre
Lac/lac lactose
15 Leu/leu leucine
Lys/lys lysine
M mole
Mar multiple-antibiotic-resistance
Met/met methionine mg milligrammes
Mg'" magnesium ion
MgCl^ magnesium chloride mg mL ' milligrammes per millilitre
Mg mL ' microgrammes per millilitre
ML microlitre
MIC minimum inhibitory concentration mL millilitre mM/mmol millimole
MMR methyl-directed mismatch repair
MMS methylmethanesulfonate mRNA messenger ribonucleic acid
MRSA methicillin-resistant Staphylococcus aureus
MS MacConkey salicin mtl maltose
MucA' activated form of MucA
N neat bacterial suspension
NA nalidixic acid
NAD nicotinamide adenine dinucleotide
16 NaOH sodium hydroxide
Nm neomycin nm nanometre
-OH hydroxyl group
Omp outer membrane protein
PBP penicillin binding protein
PCR polymerase chain reaction
Phe/phe phenylalanine pmol picomole pmol yuL'' picomoles per microlitre
Pol polymerase
Pro/pro proline
PRR post-replication recombination psi pounds per square inch
QRDR quinolone resistance-determining region
R plasmid-free
RecA* activated form of RecA
RFLP restriction fragment length polymorphism
RNA ribonucleic acid rpm revolutions per minute rRNA ribosomal ribonucleic acid
Sal salicin
Ser/ser serine
Sm streptomycin
17 Sp spectinomycin
ss single-stranded
SSB single-strand binding protein
Su sulphonamides
T thymine
To tetracycline
Tet^ tetracycline resistant
Thi/thi thiamine
Thr/thr threonine
Trp/trp tryptophan
Tyr/tyr tyrosine
U yuL'' units per microlitre
UmuD' activated form of UmuD
UV ultraviolet light
Val/val valine
VaP valine resistant xyl xylose
°C degrees Celsius
< less than
< less than or equal to
> greater than
> greater than or equal to
% percentage
%^/y percentage weight per volume
18 List of Figures
Figure Title Page
1 Model of SOS regulation. 65
2 DNA sequence of the lacI-lacZ fusion used in the 88 Lac frameshift assay.
3 Schematic configuration of the regions of 91 E. coli MCS2 before and after fusion formation.
4 The effect of plasmids on the post-UVsurvival of 139 E. coli strain ABl 157.
5 Relative UV sensitivities of repair-deficient £. co// 140 strains.
6 Absence of UV-protective effect of R46 and R446b in 141 E. coli AB2463 recAlS.
I Absence of UV-protective effect of R46 and R446b in 142 E. coli AB2494 lexA3.
8 Effect of R46 and R446b on the post-UV survival of 142 E. coli AB2470 recB21.
9 Plasmids R46 and R446b protect E. coli A B l886 143 uvrA6 against UV.
10 Plasmids R46 and R446b protect E. coli TK702 143 umuC36 against UV.
II Plasmids R46 and R446b increase UV-induced 152 mutagenesis to arginine independence in E. coli ABl 157 argE3.
12 Plasmids R46 and R446b increase UV-induced 152 mutagenesis to threonine independence in E. coli ABl 157 thr-1.
13 Plasmids R46 and R446b increase UV-induced 153 mutagenesis to leucine independence in E. coli ABl 157 leuB6.
19 14 Plasmids R46 and R446b increase UV-induced 153 mutagenesis to histidine independence in E. coli ABl 157 hisG4.
15 Accumulation of leucine-independent revertants in 156 cultures ofE. coli ABl 157 leuB6 plated on medium lacking leucine, and the effect of plasmid R46.
16 Accumulation of threonine-independent revertants in 156 cultures ofE, coli ABl 157 thr-1 plated on medium lacking threonine, and the effect of plasmid R46.
17 Accumulation of histine-independent revertants in 157 cultures of E. coli ABl 157 hisG4 plated on medium lacking histidine, and the effect of plasmid R46.
18 Accumulation of arginine-independent revertants in 157 cultures ofE. coli ABl 157 argE3 plated on medium lacking arginine, and the effect of plasmid R46.
19 Effect of R46 and R446b on the UV-induced 162 mutagenesis ofE. coli AB 1157 to nalidixic acid (NA) resistance (4 yUg mL*).
20 Effect of R46 and R446b on the UV-induced 162 mutagenesis ofE. coli ABl 157 to nalidixic acid (NA) resistance (8 yUg mL'*).
21 Effect of R46 and R446b on the UV-induced 163 mutagenesis ofE. coli ABl 157 to nalidixic acid (NA) resistance (16 yL6g mL'*).
22 Accumulation of nalidixic acid-resistant mutants in 167 cultures ofE. coli AB 1157 plated on medium containing 4 yUg mL * nalidixic acid (NA), and the effect of plasmids R46 and R446b.
23 Accumulation of nalidixic acid-resistant mutants in 167 cultures of E. coli ABl 157 plated on medium containing 8 yUg mL * nalidixic acid (NA), and the effect of plasmids R46 and R446b.
24 Accumulation of nalidixic acid-resistant mutants in 168 cultures of E. coli ABl 157 plated on medium containing 16 /^g mL * nalidixic acid (NA), and the effect of plasmids R46 and R446b.
20 25 The effect of growth in nutrient broth containing 172 0, 2 or 3 yug mL * nalidixic acid (NA) on the production of nalidixic acid-resistant (16 /^g mL’*) mutants in cultures ofE. coli ABl 157.
26 The effect of plasmids on spontaneous mutagenesis to 173 nalidixic acid (NA) resistance (16yUg mL’*) in E. coli ABl 157 cultures grown in drug-free nutrient broth.
27 The effect of growth in nutrient broth containing 0, 2 176 or 3 yUg mL * nalidixic acid (NA) on the production of nalidixic acid-resistant (16 //g mL’*) mutants in cultures ofE. coli ABl 157 (RP4).
28 The effect of growth in nutrient broth containing 0, 2 176 or 3 yUg mL * nalidixic acid (NA) on the production of nalidixic acid-resistant (16 yUg mL’*) mutants in cultures ofE. coli ABl 157 (R46).
29 The effect of growth in nutrient broth containing 0, 2 177 or 3 yUg mL * nalidixic acid (NA) on the production of nalidixic acid-resistant (16 yUg mL’*) mutants in cultures ofE. coli ABl 157 (R446b).
30 The effect of plasmids on mutagenesis to nalidixic acid 177 (NA) resistance (16 yUg mL’*) in E. coli ABl 157 cultures grown in nutrient broth containing 3 yt/g mL * nalidixic acid.
31 The effect of plasmids on spontaneous mutagenesis to 180 threonine independence (Thr^) in E. coli ABl 157 thr-1 cultures grown in drug-free nutrient broth.
32 The effect of growth in nutrient broth containing 180 2 yUg mL * nalidixic acid (NA) on the production of threonine-independent (Thr^) revertants in cultures of E. coli ABl 157 thr-1 and plasmid-containing strains.
33 The effect of growth in nutrient broth containing 181 3 yUg mL * nalidixic acid (NA) on the production of threonine-independent (Thr^) revertants in cultures of E. coli ABl 157 thr-1 and plasmid-containing strains.
34 The effect of growth in nutrient broth containing 0 or 190 0.75 X MIC ciprofloxacin (Cip) on the production of ciprofloxacin-resistant (4 x MIC) mutants in cultures of E. coli ABl 157, and the effect of plasmid R46 on the development of such mutants.
21 35 DNA sequence of the Hinfi. restriction endonuclease 209 recognition site within gyrA, which codes for aspartate-82 and serine-83.
36 Restriction fragment length polymorphisms and 210 undigested PCR products in the amplified 620 bp region ofgyrA in selected E. coli ABl 157 strains sensitive or with a decreased susceptibility to nalidixic acid and ciprofloxacin.
37 The DNA and amino acid sequence of a 620 bp region 212 ofE. coli ABl 157 gyrA (amino acids 1 to 207) incorporating the QRDR.
22 List of Tables
Table Title Page
1 Strains ofEscherichia coli used. 111
2 Plasmids used during this work. 112
3 Components for PCR amplification of theE. coli 131 gyrA gene.
4 Summary of plasmid transfer fi*equencies firom E. coli 137 strain J6-2 or E. coli 343/113 to& coli ABl 157 during three hours mating in nutrient broth.
5 Frequencies of transfer of plasmid R46 from E. coli 137 J6-2 to repair-deficient E. coli strains.
6 Frequencies of transfer of plasmid R446b from E. coli 138 343/113 to repair-deficient E. coli strains.
7 Effects of plasmids R46 and R446b on the post-UV 141 survival of the repair-proficientE. coli ABl 157 and repair-deficient derivatives.
8 MIC values of nalidixic acid for E. coli strain AB 1157 145 and its plasmid bearing derivatives.
9 MIC values of nalidixic acid for repair-deficient 147 strains ofE. coli.
10 Effect of plasmids R46 and R446b on the fi-equency 151 of spontaneous reversion to amino acid independence in E. coli ABl 157.
11 Effect of plasmids R46 and R446b on the UV-induced 154 frequencies of reversion to amino acid independence in E. coli ABl 157, following exposure to a UV dose of 32 J m'^.
12 Reversion frequencies ofE. coli strain AB 1157 to 155 amino acid independence after ten days growth on DM agar lacking the amino acid under test, and the effect of plasmid R46.
13 Effect of plasmids R46 and R446b on the spontaneous 161 mutation frequency to nalidixic acid resistance in E. coli ABl 157.
23 14 Effect of plasmids R46 and R446b on the frequency 161 of mutation to nalidixic acid resistance ofE. coli ABl 157, following exposure to a UV dose of 32 J m^.
15 Numbers of nalidixic acid-resistant E. coli AB 1157 165 clones per plate following incubation for one day, and the effect of plasmids R46 and R446b.
16 Numbers of nalidixic acid-resistant E. coli AB 1157 166 clones per plate following incubation for ten days, and the effect of plasmids R46 and R446b.
17 Plasmid-mediated increases in the number of nalidixic 17 5 acid-resistant mutants recovered after ten days incubation in nutrient broth containing 0, 2 or 3 /Ug mL'* nalidixic acid.
18 Number of Thr^ revertants recovered after ten days 181 growth in liquid medium containing 0, 2 or 3 yUg mL'* nalidixic acid, and the effects of plasmids RP4, R46 and R446b.
19 Minimum inhibitory concentrations (MICs) of nalidixic 184 acid (NA) for repair-deficient E. coli strains.
20 Mutation frequencies to nalidixic acid (NA) resistance 185 (4 X MIC) in repair-deficient mutants of E. coli after four days growth in nutrient broth containing 0.75 x MIC nalidixic acid for each respective strain.
21 The effect of carriage of plasmid R46 on mutation 186 frequencies to nalidixic acid (NA) resistance (4 x MIC) in repair-deficient mutants of E. coli after four days growth in nutrient broth containing 0.75 x MIC nalidixic acid for each respective strain.
22 Minimum inhibitory concentrations (MICs) of nalidixic 193 acid for clones selected on nutrient agar containing nalidixic acid (4 x MIC) or ciprofloxacin (4 x MIC).
23 Minimum inhibitory concentrations (MICs) of 195 ciprofloxacin for clones selected on nutrient agar containing nalidixic acid (4 x MIC) or ciprofloxacin (4xMIC).
24 Comparison of the minimum inhibitory concentrations 196 (MICs) to nalidixic acid and ciprofloxacin of the clones described in Tables 22 and 23.
24 25 Plan of experiment to grow clones, previously selected 199 as described in Tables 22 and 23, in the presence of higher concentrations of quinolones.
26 Minimum inhibitory concentrations (MICs) of nalidixic 201 acid for clones selected on nutrient agar containing nalidixic acid (4 x the resistant MIC) or ciprofloxacin (4 X the resistant MIC).
27 Minimum inhibitory concentrations (MICs) of 202 ciprofloxacin for clones selected on nutrient agar containing nalidixic acid (4 x the resistant MIC) or ciprofloxacin (4 x the resistant MIC).
28 Comparison of the minimum inhibitory concentrations 204 (MICs) to nalidixic acid (NA) and ciprofloxacin (Cip) of the clones described in Tables 26 and 27.
29 Minimum inhibitory concentrations (MICs) to nalidixic 207 acid and ciprofloxacin of the investigated E. coli strains.
30 DNA base sequence alterations ofE. coli strains 213 described in Table 29 and Figure 36.
25 INTRODUCTION
1. Antibiotic resistance
The introduction of antibiotics for the treatment of microbial infections was
accompanied by the suggestion that the development of antibiotic resistance during
therapy was unlikely because the frequency of mutation to resistance was too low. Indeed,
studies on the Murray collection of bacteria, preserved from the pre-antibiotic era,
indicate that resistance to antibiotics occurred at very low frequencies prior to the
introduction of such drugs into clinical practice (Datta and Hughes, 1983; Hughes and
Datta, 1983). However, antibiotic resistance developed quickly under the selective
pressure of antibiotic use and has plagued the treatment of infectious diseases, particularly with the emergence of multiple antibiotic-resistant strains.
1.1. Intrinsic resistance
Antibiotic resistance may be intrinsic or acquired. Intrinsic resistance refers to an
inherent feature of the organism that mediates natural resistance to an antibacterial agent
(Bryan, 1988). Intrinsic resistance mechanisms can complicate the treatment of infectious
diseases by restricting the choice of agents available. For example, intrinsic resistance of
Pseudomonas aeruginosa and mycobacteria results from a low permeability of these
species to many antibacterial agents (reviewed by Nikaido, 1994).
1.2. Acquired resistance
Acquired resistance can occur via mutation or through plasmid- or transposon- determined genes. In tlie clinical setting, the development of acquired antibiotic resistance essentially results from the selective pressure exerted on bacteria during the
26 administration of antibiotics, allowing resistant strains to emerge from previously sensitive bacterial populations (Datta, 1984).
Plasmids are the most frequent cause of clinical antibacterial resistance. Four mechanisms exist by which plasmids confer resistance. These are: (1) alteration of the antibiotic target site, so reducing or eliminating the binding of the antibiotic to its target, e.g. erythromycin and lincosamides (Lai and Weisblum, 1971; Lai et al, 1973a; 1973b);
(2) detoxification or inactivation of the antibiotic, e.g. p-lactams (Bush and Sykes, 1984;
Saunders et al, 1986), aminoglycosides (Shannon and Phillips, 1982), chloramphenicol
(Suzuki and Okamoto, 1967; Shaw, 1967; 1971); (3) prevention of drug transport into the cell or its accumulation by the cell, e.g. tetracycline (Tait and Boyer, 1978; Coleman et al, 1983); and (4) bypass mechanisms specifying a resistant enzyme substitute for a sensitive host-specified enzyme, which is the drug target, e.g. sulphonamides (Wise and
Abou-Donia, 1975), trimethoprim (Amyes and Smith, 1974).
1.3. Mutational antibiotic resistance
Mutational resistance, although not as frequent as plasmid-mediated resistance in causing clinical drug resistance, still accounts for the development of resistance to a number of important antibacterial agents, including quinolones (see Introduction section
2.6i).
1.3.1. Rifampicin resistance
The target of rifampicin in bacteria is the RNA polymerase, specifically the P subunit, encoded by the rpoB gene (Ovchinnikov et al, 1981; 1983). The RNA polymerase ofEscherichia coli is a complex multisubunit enzyme that exists in two
27 forms. RNA polymerase holoenzyme (ttjPP'o) initiates transcription at defined promoter
sites, while elongation of RNA is carried out by core RNA polymerase (ttjpP') (Burgess
et al, 1969; Burgess and Travers, 1970). The mechanism of action of rifampicin is to interfere with transcription by binding to the P subunit of RNA polymerase at a region formed by the appropriate complexing of the different RNA polymerase subunits
(Hartmann et al, 1967; Riva and Silvestri, 1972; Johnston and McClure, 1975). This leads to an interruption of transcription by preventing elongation of the nascent RNA chain past a few nucleotides.
RNA polymerase mutations conferring resistance to rifampicin are exclusively located in the p subunit (Rabussay and Zillig, 1969; Heil and Zillig, 1970; Iwakura et al,
1973), reducing the binding of rifampicin to the enzyme. Mutated sites within the E. coli rpoB gene have been assigned into three clusters (clusters I, II, and III), plus one additional site in the N-terminal region (codon 146) (Lisitsyn et al, 1984; Jin and Gross,
1988). Furthermore, the majority of these rifampicin resistance mutations are tightly clustered in a short region, region I of cluster II, near the middle ofrpoB. Such mutations are likely to affect the conformation of the site at which rifampicin binds to the P subunit, and produce high-level resistance to rifampicin (Jin and Gross, 1988).
Rifampicin is very effective against susceptible strains of Mycobacterium tuberculosis (Vall-Spinosa et al, 1970; Mitchison and Nunn, 1986). However, substitution of a small number of amino acids in rpoB results in rifampicin resistance
(Telenti et al, 1993). These amino acid sites are conserved in many bacterial species, and mutation within them also results in rifampicin resistance in, for example,E. coli (Jin and
Gross, 1988). Rifampicin-resistantMycobacterium leprae are also mutant within rpoB
(Honore and Cole, 1993), and as withM tuberculosis, Ser-531 is the most common site
28 of amino acid substitution (Telenti et al, 1993; Honore and Cole, 1993).
1.3.2. Fusidic acid resistance
Translation elongation factor G (EF-G) is an essential bacterial protein that promotes translocation of the ribosome after peptide bond formation (reviewed by Liljas,
1982; Spirin, 1985). Fusidic acid binds to the EF-G-ribosome complex in combination with either GTP or GDP, and stabilises EF-G-GDP on the ribosome (Willie et al, 1975).
This prevents further elongation, and inhibits protein synthesis (reviewed by Cundliffe,
1981).
Mutations conferring resistance to fusidic acid map tofus A, the gene coding for
EF-G (Bemardi and Leder, 1970; Tanaka et al, 1971; Kuwano et al, 1971), although other sites of mutational resistance have been reported (Isaksson and Takata, 1978;
Bennet and Shaw, 1983). Genetic analysis of fusidic acid-resistant mutants of Salmonella typhimurium reveal that fusA mutations are not randomly distributed but instead map to three distinctly separated regions in the primary structure of EF-G (Johanson and Hughes,
1994). The three clusters coincide fairly well with the most conserved regions in EF-G
(Kohno et al, 1986; Cammarano et al, 1992) suggesting that the mutations are in functionally important regions of EF-G, and therefore may have effects other than simply affecting the binding of fusidic acid (Johanson and Hughes, 1994). The conceivable structural effects of mutations conferring fusidic acid resistance are; (1) altered interaction between EF-G and the ribosome; (2) reduced affinity for the antibiotic by modification of the binding site; or (3) interference with conformational structure. It appears likely that most mutations effect conformational change of the EF-G protein
(Johansonet al, 1996).
29 1.3.3. Methicillin resistance in Staphylococcus aureus
Susceptibility to P-lactam antibiotics reflects the combined affects of drug binding to targets, resistance to inactivation by P-lactamases and, in Gram-negative bacteria, outer membrane permeability. The major mechanism of p-lactam resistance is inactivation of the drug by p-lactamases, particularly plasmid-encoded P-lactamases (Philippon et al,
1989; Jacoby and Medeiros, 1991; reviewed by Jacoby, 1994). Methicillin was the first of a series of p-lactams to be introduced into clinical practice to counteract the widespread occurrence of P-lactamases. It lead to a general decline in the prevalence of multiply resistant Staphylococcus aureus during the early 1960s (reviewed by Shanson,
1981).
P-lactams target enzymes involved in cell wall synthesis, termed penicillin- binding proteins (PBPs) (Blumberg and Strominger, 1974; Georgopapadakou and Sykes,
1983; Waxman and Strominger, 1983) and clinical resistance to methicillin is conferred by mutation within genes encoding PBPs (Chambers, 1988; Brumfitt and Hamilton-
Miller, 1989; Franciolliet al, 1991). For example, S. aureus possesses five PBPs: PBPl,
PBP2, PBP3, PBP3' and PBP4 (Suginaka et al, 1972; Georgopapadakou and Liu, 1980;
Reynolds and Chase, 1981), of which PBPs 1, 2, and 3 are essential and have high affinities for P-lactam antibiotics (Georgopapadakou and Liu, 1980; Reynolds, 1988).
Methicillin-resistant S. aureus (MRSA) strains possess an additional 78 kDa PBP,
2a or 2' (Hartman and Tomasz, 1984; 1986; Rossiet al, 1985; Ubukata et al, 1985; Utsui and Yokota, 1985; Gerberding et al, 1991), encoded by the chromsomal gene, mecA, located on a 30 kb DNA segment of unknown source (Matsuhashi et al, 1986). This PBP has a low affinity for p-lactams, allowing peptidoglycan biosynthesis at P-lactam
30 concentrations which inactivate PBPs 1, 2, 3, and 4 in MRS A (Fontana, 1985; Hartman and Tomasz, 1986; Matsuhashi et al, 1986; Reynolds and Fuller, 1986). Transformation of a strain ofS. aureus, susceptible to methicillin, with a plasmid carrying mecA converts it to MRSA (Tesch et al, 1988; Ubukata et al, 1989). However the level of resistance to methicillin does not correlate with cellular levels of PBP 2a (Murakami and Tomasz,
1989).
Additional chromosomal genes are also involved in the expression of methicillin resistance, (/actor essential for methicillin resistance) (Berger-Bachiet al, 1989) encodes a 48 kDa protein, which does not alter the expression of PBP 2a, but by affecting the glycine content of peptidoglycan and cell wall turnover, reduces susceptibility to P- lactams (Gustafson and Wilkinson, 1989; Maidhoffet al, 1991; Gustafsonet al, 1992). femB, which lies downstream and adjacent to femA, also reduces susceptibility to P- lactams, but to a lesser extent than does femA (Berger-Bachi et al, 1989; 1992). Two additional genes, y^mC and femD, unlinked to femA and femB, have also been identified, howeverfemC and femD produce only a low-level basal resistance to methicillin (Berger-
Bachi et al, 1992).
1.3.4. Streptomycin resistance
The aminoglycoside antibiotic, streptomycin, is commonly used in the treatment of tuberculosis. The drug binds to the 16S ribosomal RNA (rRNA) subunit and interferes with translation proofi*eading. This leads to an inhibition of protein synthesis (Galeet al,
1981; Allen and Noller, 1989). Plasmid-determined aminoglycoside-modifying enzymes are the most fi-equent cause of clinical resistance to streptomycin (Benveniste and Davies,
31 1973; Shannon and Phillips, 1982). However, mutant bacteria with altered responses to streptomycin were also discovered soon after this bactericidal antibiotic was used clinically (Miller and Bohnhoff, 1974). Mutations associated with streptomycin resistance have been identified in M tuberculosis genes encoding 16S rRNA {rrs) (Douglass and
Steyn, 1993) and ribosomal protein S12{rpsL) (Funatsu and Wittman, 1972; Finken et al, 1993; Nair et al, 1993; Heym et al, 1994; Honore and Cole, 1994; Kapuret al, 1995;
Morris etal, 1995). Mutations inrrs alter the 16S rRNA structure, disrupting interactions between 16S rRNA and streptomycin, resulting in resistance (De Stasio et al, 1989).
Single base substitutions in the rrs gene have been identified in two key regions around nucleotide 904 (Douglass and Steyn, 1993; Honore and Cole, 1994; Meieret al, 1994;
Morris et al, 1995; Sreevatsan et al, 1996; Katsukawa et al, 1997) and the 530 loop
(Finken et al, 1993; Meier et al, 1994; Morris et al, 1995; Sreevatsan et al, 1996;
Katsukawa et al, 1997). A point mutation at codon 43 of therpsL gene in several bacterial species, including M. tuberculosis, results in the substitution of an arginine for a lysine in the S12 protein, conferring streptomycin resistance (Nairet al, 1993; Cooksey et al, 1996; Sreevatsan et al, 1996; Katsukawa et al, 1997).
1.3.5. Sulphonamide resistance
The sulphonamide antibacterial agents act as competitive inhibitors of the enzyme dihydropteroate synthetase (DHPS), inhibiting folate biosynthesis in the bacterial cell
(Brown, 1962). Sulphonamides are structural analogues of the enzyme’s normal substrate, para-amino-benzoic acid and act as alternative substrates for the enzyme to produce a sulphur-containing pteroate analogue (Rolandet al, 1979; Swedberg et al,
1979).
32 Clinically occurring sulphonamide resistance in Gram-negative enteric bacteria is largely plasmid-mediated, encoding alternative drug-resistant variants of the DHPS enzymes (Akiba et al, 1960; Skold, 1976; Wise and Abou-Donia, 1975; Swedberg and
Skold, 1980). These allow bypass of the chromosomally-encoded sensitive DHPS enzymes, completion of the blocked metabolic pathway and growth in the presence of sulphonamides. Indeed, plasmid-mediated resistance to sulphonamide antibacterials was the first plasmid-bome antibiotic-resistance phenotype to be observed (Nakaya et al,
1960; reviewed by Watanabe, 1963). However, mutations in the chromosomal dhps gene that confer sulphonamide resistance can be isolated in the laboratory (Pato and Brown,
1963; Skold, 1976; Dallas et al, 1992; Swedberg et al, 1993) and sulphonamide resistance in clinical isolates ofNeisseria meningitidis has been demonstrated to be linked to the chromosome (Kristiansenet al, 1990; Râdstrôm et al, 1992). The dhps gene of clinically-isolated sulphonamide-resistant strains oïN. meningitidis was found to differ by 10% from the corresponding gene in sulphonamide-susceptible isolates, suggesting that the resistance did not result from point mutation but rather had been introduced by recombination. The dhps genes of such resistant clinical isolates carry an insertion of two amino acids in a highly conserved region of the enzyme. If these amino acids are removed by site-directed mutagenesis, sulphonamide susceptibility is returned to the strain
(Huovinen e/a/, 1995).
1.3.6. Trimethoprim resistance
Trimethoprim acts at a later stage in folate biosynthesis, competitively inhibiting the enzyme dihydro folate reductase (DHFR). Plasmid-mediated resistance to trimethoprim was first described in 1972 (Fleming et al, 1972) and is produced by the
33 substitution of a plasmid-specified trimethoprim-resistant DHFR enzyme for the inhibited chromosomal DHFR (Amyes and Smith, 1974; Skold and Widh, 1974).
Chromosomal resistance to trimethoprim can occur via one of three mechanisms.
The first results from the chromosomal location of transposon Tn7 (Lichtenstein and
Brenner, 1981; Heikkila, 1991), the second from a mutational loss of the ability of bacteria to methylate deoxyuridylic acid to thymidylic acid, leading to a requirement for external thymine (Masked et al, 1977; King et al, 1983; Hamilton-Miller, 1984), and the third involves mutational change within the chromosomal gene for DHFR.
Clinical isolates ofHaemophilus influenzae showing levels of resistance to trimethoprim 20- to 400-fold higher than susceptible strains have been demonstrated to contain mutational alterations in the DHFR gene (De Grootet al, 1988; 1991a, Powell et al, 1991) leading to overproduction of altered chromosomal DHFR (De Grootet al,
1988, 1991a; 1991b). Chromosomal mutations to trimethoprim resistanceinE. coli and
Streptococcus pneumoniae were first identified in the laboratory in the 1970s (Sirotnak and McCuen, 1973; Breeze et al, 1975). A combination in E. coli of a structurally altered enzyme and overproduction, resulting from chromosomal mutation, seems to be required to produce high MICs to trimethoprim (Flensburg and Skold, 1987).
2. The quinolones
The quinolones differ from the majority of antibacterial agents in that they are synthesised chemically, rather than being isolated from moulds or fungi. Nalidixic acid, the original quinolone, was discovered during the synthesis of chloroquine (Lesheret al,
1962). It possesses only a narrow spectrum of activity, limited to Gram-negative aerobes.
This, together with its high protein binding, which results in only modest in vivo serum
34 and tissue concentrations, restricts the clinical use of nalidixic acid to the treatment of urinary tract (Stamcy et al, 1963) and enteric infections (Moorhead and Parry, 1965;
McCormack, 1983; Parry, 1983; Malcngracau, 1984).
Mutants ofE. coli resistant to nalidixic acid arc readily detected in vitro (Barlow,
1963). However, the incidence of nalidixic acid resistance among clinical uropathogens has remained low, despite extensive use of the drug since the 1960s and the more recent introdution of the fluoroquinolones into clinical use (Kresken and Wiedemann, 1988; reviewed by Crumplin, 1990a).
The quinolone derivatives developed and introduced during the 1960s and 1970s, included oxolininc acid, cinoxacin, piromidic and pipemidic acids, and flumequine.
These first-generation quinolones offered only slight improvements in Gram-negative activity over nalidixic acid, and continued to lack useful activity against Gram-positive cocci, anaerobes orP. aeruginosa. Following oral administration, these first-generation quinolones were generally well absorbed and were excreted in high concentrations in the urine, allowing them a therapeutic use in the treatment of urinary tract infections
(reviewed by Moellering, 1996; Radi, 1996).
A significant step was made in the development of quinolones with the production of norfloxacin (Itoet al, 1980; Kogaet al, 1980). This drug exhibited both improved activity over first-generation quinolones against Gram-negatives and some Gram-positive activity, but was still only indicated for the treatment of urinary tract infections, sexually transmitted diseases, and prostatitis. Norfloxacin was the first of the fluoroquinolones.
Introduction of a fluorine atom at the 6-position facilitated drug entry into the bacterial cell and increased quinolone activity against the target enzyme, DNA gyrase (Domagala
35 et al, 1986). Modification to the norfloxacin structure led to the development of further second-generation quinolones, including pefloxacin, enoxacin, fleroxacin, lomefloxacin, ciprofloxacin, ofloxacin, and rufloxacin, the most important being ciprofloxacin (Wise et al, 1983) and ofloxacin (Satoet al, 1982). Ciprofloxacin displays improved MIC values, over the first-generation quinolones, against both Gram-positive and Gram- negative pathogens (Wolfson and Hooper, 1985) and was the first quinolone to be used in the treatment of infections beyond the urinary tract and sexually transmitted diseases, including lower respiratory tract, skin and joint infections. Ofloxacin has also been used in the treatment of a variety of systemic infections (Monk and Campoli-Richards, 1987).
The increased potency of the second-generation quinolones against Gram-negative organisms is directly related to their increased activity against DNA gyrase (Crumplin,
1990b).
The development of third-generation quinolones attempted to increase drug potency against both Gram-negative and Gram-positive organisms, including anaerobes.
These later drugs include grepafloxacin, sparfloxacin, gatifloxacin, levofloxacin, pazufloxacin, trovafloxacin, tosulfioxacin, clinafloxacin, and DU-6859a. Clinafloxacin,
DU-6859a and the discontinued BAY y 3118 displayed the broadest spectrum of activity of the class, and were the only agents to possess better Gram-negative activity than ciprofloxacin (Gootz and Brighty, 1996; Wiseetal, 1993; Bauemfeind, 1993). All third- generation fluoroquinolones display better activity againstS. aureus than ciprofloxacin
(Barry and Jones, 1989; Wise etal, 1993; Gootz and Brighty, 1996). However, increasing the lipophilicity of the quinolone to improve potency against Gram-positive organisms, often reduced activity against the Enterobacteriaceae, in comparison to ciprofloxacin
(Barry and Jones, 1989; Gootz and Brighty, 1996). With the exception of pazufloxacin,
36 all third-generation quinolones display enhanced activity, over ciprofloxacin, against
Streptococcus pneumoniae (Ito et al, 1992; Wise et al, 1993; Gootz and Brighty, 1996).
Like ciprofloxacin, the majority of third-generation quinolones retain excellent activity against the respiratory pathogensHaemophilus influenzae, and Legionella pneumophilia
(Gaja, 1992; Edelstein et al, 1996; Gootz and Brighty, 1996) and exhibit enhanced activity over ciprofloxacin, against anaerobes, such asBacteroides fragilis (Aldridge,
1994; Gootz and Brighty, 1996).
2.1. DNA gyrase
DNA gyrase belongs to a group of enzymes called topoisomerases, which catalyse alterations in DNA topology. They are required for DNA replication, transcription, and recombination in both prokaryotic and eukaryotic cells (reviewed by: Reece and
Maxwell, 1991; Roca, 1995).E. coli DNA gyrase, a type II topoisomerase, is a tetrameric enzyme (A2B2) composed of A and B subunits, encoded by the gyrA and gyrB genes respectively. DNA gyrase introduces negative supercoils into DNA in the presence of
ATP and is the only topoisomerase able to perform this activity (Gellert, 1981 ; Shen and
Chu, 1996). Both DNA strands are cut at four base pair staggered positions by DNA gyrase (Morrison and Cozzarelli, 1981) producing single-stranded DNA ends. A covalent link is formed between the hydroxyl group of tyrosine-122 of subunit A of DNA gyrase and the 5 -phosphoryl end of the cleaved DNA backbone (Horowitz and Wang, 1987).
Opening of the DNA chains along the four base pair staggered cuts is necessary to allow subsequent ATP-driven DNA strand passage. This introduces negative supercoils into the
DNA helix, thus allowing the prokaryotic chromosome to be packaged inside the cell
(Worcel, 1974; Crumplin and Smith, 1976; Gellert etal, 1976). The GyrB subunits bind
37 ATP and thus provide the energy required for gyrase-catalysed supercoiling of DNA
(Gellert et al, 1976; Sugino et al, 1978). Passage of the second DNA strand through the double-strand break, mediated by the GyrA subunits, is also essential to the mechanism by which DNA gyrase catalyses relaxation, decatenation and the unknotting of circular
DNA (Brown and Cozzarelli, 1979; Mizuuchi et al, 1982).
2.2. DNA topoisomerase IV
The possibility that quinolones targeted a second site, in addition to DNA gyrase, was suggested by the observation that mutations in gyrA only partially block the lethal action of ciprofloxacin (Lewinet al, 1991). The secondary target was identified as another type II topoisomerase, topoisomerase IV (Katoet al, 1990), which is also inhibited by quinolone antibacterials (Peng and Marians, 1993; Hoshino et al, 1994;
Khodursky et al, 1995). Topoisomerase IV, like gyrase, is composed of two pairs of subunits, encoded by the parC and parE genes (Kato et al, 1990). ParC and ParE subunits show significant homology with GyrA and GyrB respectively (Katoet al, 1992).
Although topoisomerase IV is homologous to DNA gyrase in amino acid sequence, it displays the relaxation activity, but not the supercoiling activity of DNA gyrase (Kato et al, 1992). Both gyrase and topoisomerase IV use a double-strand passage mode of action (Roca, 1995), but differ in the fact that gyrase wraps DNA around itself, whereas topoisomerase IV does not (Peng and Marians, 1995). In contrast to DNA gyrase, topoisomerase IV is believed to recognise DNA crossovers (Zechiedrich and
Cozzarelli, 1995), similar to eukaryotic topisomerase II (Zechiedrich and Osheroff, 1990;
Roca and Wang, 1992; 1994), helping to explain its greater decatenating than relaxing activity (Hoshinoet al, 1994). Purified topoisomerase IV removes catenanes produced
38 by bidirectional replication, an activity only weakly associated with gyrase (Peng and
Marians, 1993; Hiasa and Marians, 1994; Hiasa et al, 1994). The latter enzyme only decatenates DNA after completion of a round of replication, whereas topoisomerase IV can decatenate during replication (Zechiedrich and Cozzarelli, 1995).
2.3. Mode of action of quinolone antibacterials
DNA gyrase emerged as the quinolone target soon after discovery of the enzyme
(Gellert et al, 1976). The supercoiling activity of purified DNA gyrase extracted firom wild-type E. coli cells was inhibited by treatment with nalidixic and oxolinic acids, whereas gyrase extracted from gyrA (nalA) mutants was not (Gellert et al, 1977; Sugino et al, 1977). This inhibitory action was suggested to involve the trapping of a gyrase-
DNA complex in which the DNA is broken (Gellert et al, 1977; Sugino et al, 1977;
Snyder and Drlica, 1979) leading to inhibition of DNA replication (Goss et al, 1965;
Deitz et al, 1966; Cook et al, 1966).
In Gram-positive organisms, such asS. aureus, it appears that topoisomerase IV, rather than DNA gyrase, is the primary target of the quinolones. Sequence analysis of clinically-isolated strains of S. aureus moderately resistant to ciprofloxacin reveal mutations inparC, with high-level resistance being conferred by an additional mutation in gyrA (Ferrero et al, 1994). Analysis of laboratory strains challenged with increasing concentrations of quinolone confirmed that a single mutationparC in confers a low level of resistance to quinolones, with mutations in bothparC and gyrA resulting in high-level resistance (Ferrero et al, 1995; Ng et al, 1996). Quinolone resistance inS. aureus due to a mutation in gyrA has only been found in strains carrying a resistant topoisomerase IV
39 (Ng et al, 1996). The same is true for S. pneumoniae (Munoz and de la Campa, 1996; Pan etal, 1996).
Shen initially proposed that quinolones acted by binding directly to the single stranded “DNA gate” produced by DNA gyrase, via hydrogen bonds with unpaired bases on the single-stranded DNA ends (Shen and Pemet, 1985; Shen et al, 1989a). Such binding traps the gyrase enzyme in a complex with the broken DNA, preventing its turnover (Shen et al, 1989b). However, recent experiments have demonstrated that quinolone binding to gyrase-DNA complexes also occurs in the absence of DNA cleavage (Critchlow and Maxwell, 1996), thus the relevance of DNA cleavage to quinolone binding remains unclear. Inhibition of DNA gyrase activity leads to cessation of DNA replication, induction of the SOS response (see Introduction section 3.5) and eventually cell death. The bactericidal effect of quinolones is accompanied by filamentation of the cells followed by lysis (Gosset al, 1964; Deitz et al, 1966).
2.4. Quinolone lethality
Quinolone lethality is often attributed to inhibition of DNA synthesis, however the concentration of quinolone required to kill cells is higher than that required to block
DNA synthesis. In addition, chloramphenicol, which inhibits killing by quinolones, has little effect on inhibition of DNA synthesis (Gosset al, 1965; Chen et al, 1996). A further observation is that inhibition of DNA synthesis leads to DNA degradation mediated by
RecBCD which, if extensive, should be lethal. But such DNA breakdown is not inhibited by treatment with chloramphenicol or rifampin, whereas the lethal effects of quinolones are (Lewin and Smith, 1990). Also, more extensive DNA breakdown occurs in a recA mutant than in a recA recBC mutant (Willetts and Clark, 1969), but if both recombination
40 effects are defective, survival levels are reduced (Lewin et al, 1989). Therefore, the lethal effects of the quinolones cannot be solely due to inhibition of DNA synthesis and DNA breakdown resulting from formation of the quinolone-gyrase-DNA complex.
2.5. DNA breaks and cell death
Double-strand DNA breaks are lethal (Krasin and Hutchinson, 1977) and it has been suggested that it is the liberation of DNA ends from the quinolone-gyrase-DNA complex that accounts for the lethality of the quinolones. The fact that protein synthesis inhibitors block the lethal effects of nalidixic and oxolininc acids (Deitz et al, 1966; Chen et al, 1996) suggests a role for protein in the release of DNA ends from the quinolone- gyrase-DNA complex. This protein does not appear to be a component of the SOS response, since inhibition of the SOS response via alexA mutation has no effect on the rate at which cells are killed by nalidixic acid (Lewin et al, 1989). However, treatment of cells with chloramphenicol or rifampin only led to a partial inhibition of the lethal effects of the more potent fluoroquinolones (Lewinet al, 1991; Chen et al, 1996), suggesting that these drugs have an additional, chloramphenicol-insensitive mechanism of action (Howardet al, 1993a). Based on these observations, Drlica and Zhao (1997) suggested a pathway for quinolone lethality. DNA synthesis and cell growth are inhibited by the formation of a reversible quinolone-gyrase-DNA complex. Cell death then results by one of two mechanisms. The first involves removal of the quinolone-gyrase complex from DNA, revealing lethal double-strand DNA breaks. Such complex removal occurs with all quinolones and is blocked by protein and RNA synthesis inhibitors. The second mode involves the dissociation of DNA gyrase whilst still attached to DNA, resulting in free DNA ends with attached gyrase subunits. This second pathway is proposed to occur
41 only at high fluoroquinolone concentrations, with the resulting lethality being resistant to protein and RNA synthesis inhibition.
2.6. Resistance to quinolones
The absence of plasmid-mediated quinolone resistance leaves chromosomal mutations as the sole mechanism by which bacteria can develop resistance to quinolone antibacterials (Smith, 1984). Resistance results from mutations within genes coding for
DNA gyrase, topoisomerase IV or in genes affecting drug uptake (Hooper and Wolfson,
1991).
Quinolone resistance can be conferred by mutations in either gyrA or gyrB
(Gellert et al, 1977; Relia and Haas, 1982; Robillard and Scarpa, 1988; Nakamura et al,
1989; Satoet al, 1989), resulting in a decreased affinity of the DNA-enzyme complex for quinolones (Willmott and Maxwell, 1993). However, clinical quinolone resistance inE. coli is predominantly conferred by mutations within gyrA, probably because gyrA mutants display high-level resistance, whereas mutations within gyrB confer only a low- level of resistance to quinolones (Nakamuraet al, 1989). Reports of high-level clinical quinolone resistance in many bacterial species, due to mutations ingyrA, have emerged, including Æ influenzae (Setlow et al, 1985), Campylobacter pylori (Glupczynski etal,
1987), P. aeruginosa (Hirai et al, 1987; Inoue et al, 1987; Yoshida et al, 1990),
Citrobacter freundii (Aoyama et al, 1988), Serratia marcescens (Fujimaki et al, 1989), and Campylobacter jejuni (Gootz and Martin, 1991).
2.6.1. Mutations in the gyrA gene
Early studies of quinolone resistance mapped nalidixic acid resistance to a gene
42 that eventually became known as gyrA (Hane and Wood, 1969; Gellert et al, 1977;
Sugino et al, 1977). Sequencing the gyrA genes of four spontaneousgyrA mutants ofE. coli, which were resistant to varying levels of nalidixic or pipemidic acids, showed the mutations to be located in surprisingly close proximity to one another, within a small region near the N-terminus of the GyrA subunit (Yoshida et al, 1988). Several amino acid substitutions were involved: the amino acid serine at position 83 (Ser-83) was substituted with either leucine or tryptophan, alanine (Ala-67) was changed to serine, and glycine
(Gly-106) was replaced by histidine (Yoshida et al, 1988).
Further analysis of ten spontaneous quinolone-resistantgyrA mutants revealed that point mutations occurred within a highly conserved region between amino acids 67 and 106, particularly in the vicinity of the serine residue at position 83 (Yoshida et al,
1990). The site of mutation within this “quinolone resistance-determining region”
(QRDR) correlated with the level of quinolone resistance conferred upon the organism, such that minimum inhibitory concentrations (MICs) decreased in the order of mutations at amino acids 83, 87, 81, 84, 67 and 106 (Yoshida et al, 1990). Replacements of Ser-83 with a hydrophobic amino acid, such as leucine or tryptophan, are the only substitutions found in clinically-isolated quinolone-resistantgyrA mutants (Yoshida et al, 1988; 1990;
Cullen et al, 1989; Fisher et al, 1989; Oram and Fisher, 1991). Alteration of Ser-83 to a hydrophobic amino acid generally confers a higher level of resistance than does mutation at position 87 (Yoshida et al, 1990; Khodursky et al, 1995; Drlica et al, 1996). Such mutation is sufficient to produce a high level of resistance to nalidixic acid (MIC >128 yUg mL'’) (Vila et al, 1994) and low-level resistance to fluoroquinolones (MIC of ciprofloxacin < lyUg mL ') (Yoshidaet al, 1988; Oram and Fisher, 1991).
43 High-level fluoroquinolone resistance, resulting in ciprofloxacin MICs of>2 mL * (Everett et al, 1996), >8 yUg mL* (Vila et al, 1994) or 128jj,g mL* (Heisig et al,
1993), is produced in Ser-83 mutants by the introduction of a second mutation in gyrA replacing aspartate at position 87 with a non-negatively charged amino acid, such as asparagine, glycine or tyrosine.
In addition to substitution of Ser-83 with a hydrophobic amino acid (Yoshida et al, 1988; 1990; Cullen et al, 1989; Fisher et al, 1989; Oram and Fisher, 1991), replacement of the serine residue with alanine also results in a quinolone-resistant GyrA protein (Hallett and Maxwell, 1991). Such change results in the loss of a hydroxyl group indicating that even minor changes at position 83 are sufficient to confer a quinolone- resistance phenotype, although a lower level of quinolone resistance is displayed by the
Ser-83 to Ala substitution, as compared with the Ser-83 to Leu and Ser-83 to Trp mutations (Yoshida et al, 1988; 1990; Cullen et al, 1989; Hallett and Maxwell, 1991).
A Gin-106 to Arg mutation can produce levels of quinolone resistance similar to that of the Ser-83 to Ala mutation (Hallett and Maxwell, 1991). A Gin-106 to His mutation has also been reported at this position (Yoshidaet al, 1988). Both the Gin-106 to Arg and
Gin-106 to His mutations result in the substitution of a polar amino acid with a positively-charged amino acid, suggesting that a positive charge at position 106 of GyrA might be involved in disrupting the quinolone-gyrase interaction, resulting in increased resistance to quinolone antibacterials (Hallett and Maxwell, 1991).
2.6.2. Mutations in the gyrB gene
Mutations in gyrB occur less frequently than gyrA mutants in clinical isolates
(Nakamura et al, 1989) and only give rise to low levels of quinolone resistance
44 (Yamagishi et al, 1986, Yoshida et al, 1991). Two quinolone resistance-determining
sites, Asp-426 and Lys-447, have been identified in the gyrB gene (Yamagishi et al,
1981; 1986). Mutations at these sites, substituting Asp-426 with Asn and Lys-447 with
Glu, affect the overall charge of the GyrB protein and confer nalidixic acid resistance
(Yamagishi et al, 1986; Yoshida et al, 1991). Replacement of Asp-426 with Asn leads to a loss of negative charge and an increase in resistance to all quinolones. Mutation at position 447, from Lys to Glu, changes the charge at the site from positive to negative leading to an increase in resistance to acidic quinolones, such as nalidixic acid, oxolinic acid, cinoxacin, piromidic acid, and flumequine. The same change however, confers a hypersusceptibility to amphoteric quinolones, including pipemidic acid, norfloxacin, enoxacin, ofloxacin, ciprofloxacin, tosulfioxacin, and sparfloxacin (Yoshidaet al, 1991).
2.6.3. Mutations affecting topoisomerase IV
Mutations in genes coding for topoisomerase IV also alter susceptibility to quinolones inE. coli (Heisig, 1996). However these are only detected in gyrA double mutants and at high fluoroquinolone concentrations (Khodurskyet al, 1995; Chen et al,
1996; Kumagai et al, 1996). Two mutation sites have been detected in parC, encoding subunit A of topoisomerase IV. One substitutes the serine at position 80 with a hydrophobic amino acid such as arginine or isoleucine, the other changes the glutamate at position 84 to a positively charged amino acid, such as glycine or lysine (Heisig, 1996;
Vila et al, 1996). Both mutations lead to high-level fluoroquinolone resistance gyrAin mutants (Heisig, 1996; Everett et al, 1996). Mutation in parE, encoding subunit B of topoisomerase IV, also confers fluoroquinolone resistancegyrA in mutants (Breines et al, 1997). The substitution of His for Leu at position 445 of ParE produces a change from
45 a neutral to a positive charge, leading to a two- to four-fold increase in the MIC of norfloxacin. Such an effect was abolished in a strain lacking a resistant GyrA protein, indicating the dependence of resistance conferred by parE on a mutant GyrA (Breines et ah 1997).
In general, a single mutation at Ser-83 of GyrA is associated with a moderate level of quinolone resistance. One or twoparC mutations, in addition to a mutation at Ser-83 of GyrA, produces increased resistance. Three mutations (twogyrA and one parC) are associated with high-level resistance, and four mutations (two gyrA and two parC) produce very high levels of resistance to quinolones (Vilaet al, 1996).
DNA gyrase ofS. aureus is much less susceptible to inhibition by quinolones than is E. coli DNA gyrase (Blanche et al, 1996). S. aureus topoisomerase IV, encoded by the grlA and grlB genes (Ferrero et al, 1994) is the primary target for the quinolones (Ferrero et al, 1994; 1995; Ng et al, 1996). However, gyrA mutations were discovered first in ciprofloxacin-resistant strains ofS. aureus (Sreedharan et al, 1990). These mutations change Ser-84 of GyrA to Leu or Ala, Ser-85 to Pro, or Glu-88 to Lys (Sreedharanet al,
1990; Itoet al, 1994). Quinolone-resistantgyrB mutations also occur, changing Asp-437 to Asn, or Arg-458 to Gin (Itoet al, 1994). Combinations ofgyrA and gyrB mutations lead to high-level quinolone resistance (Itoet al, 1994). Further work demonstrated that mutation in gyrA ofS. aureus does not by itself confer resistance: quinolone-resistant strains also carry a mutation ingrlA (Yamagishi et al, 1996; Ferrero et al, 1995), confirming that the primary target of quinolones inS. aureus is indeed topoisomerase IV
(Ferrero et al, 1995).
46 2.6.4. Quinolone resistance resulting from drug impermeability
Unlike mutations affecting DNA gyrase, which specifically alter susceptibility to quinolones, mutations that reduce quinolone uptake or accumulation may also be associated with decreased sensitivity to unrelated antibacterial agents (Yamagishi et al,
1986; Yoshida 1988; 1990; 1991; Cullene f a / , 1989; Fisher e r a / , 1989; Oram and
Fisher, 1991).
The outer membrane of Gram-negative bacteria represents the major cellular barrier to drug penetration. It has many non-specific, water-filled porin channels through which hydrophilic substances enter the cell. Quinolones together with other antibacterial agents, such as aminoglycosides, p-lactams, chloramphenicol and tetracyclines enter
Gram-negative bacteria through porins by simple diffusion (Hirai et al, 1986a; Hooper et al, 1986; Nikaido, 1989). The isolation of porin-deficient organisms resistant to quinolones supports this hypothesis (Hiraiet al, 1986a; 1986b). Indeed, several species of uptake-deficient Gram-negative bacteria, resistant to nalidixic acid (Relia and Haas,
1982), ciprofloxacin (Sanders etal, 1984), norfloxacin (Hooperet al, 1986) and enoxacin
(Piddock et al, 1987) have been isolated. Such resistance is conferred by inactivation of the ompF gene, either by a mutation within ompF itself, or by post-transcriptional regulation ofompF, produced by mutations within njxB or cficB (Hirai et al, 1986a;
1986b; Hooper gf a/, 1986; 1987; 1989).
Nalidixic acid, and other hydrophobic quinolones, may also enter the bacterial cell through the phospholipid bilayer of the outer membrane (Benbrook and Miller, 1986;
Hirai et al, 1986b). Quinolone resistance may therefore be associated with variation in lipopolysaccharide content of the outer membrane (George and Levy, 1983a; Hiraiet al,
1986b).
47 Resistance to some drugs, such as tetracyclines and fluoroquinolones, may result from increased active drug efflux (George and Levy, 1983a; Cohen et al, 1989). The chromosomal locusmar in E. coli and other members of the Enterobacteriaceae (George and Levy, 1983a; 1983b; Cohen et al, 1988, 1993a) consists of two transcriptional opérons encoded by marC and marRAB. These are divergently expressed from a central operator-promoter region,marO (Cohen et al, 1993b). marRAB is inducible by tetracycline, chloramphenicol, salicylates, and other unrelated substances (Hachler et al,
1991; Cohenet al, 1993c; Seoane and Levy, 1995), and activation in both laboratory mutants and clinical isolates (Maneewannakul and Levy, 1996) produces increased levels of resistance to quinolones as well as to other antibacterials (Cohenet al, 1989; 1993c;
Hachler et al, 1991 ; Ariza et al, 1994). Plating on medium containing low concentrations of tetracycline or chloramphenicol selects multiple-antibiotic-resistant (Mar) mutants of
E. coli that are 6- to 18-fold less susceptible to fluoroquinolones than their wild-type parent. The frequency of emergence of such mutants is at least 1000-fold higher than that of those selected by norfloxacin directly (Cohenet al, 1989).
It has been postulated that activation of the mar locus by mutation within it or in response to an inducing treatment might allow selection of clones carrying mutations at other loci that confer high-level quinolone resistance (Cohenet al, 1989; Maneewannakul and Levy, 1996; Goldman et al, 1996). The Mar phenotype appears to result from a number of mechanisms. Elevated levels of MarA increase micF transcription leading to a decrease in the expression of OmpF (Cohenet al, 1988). However, alone, this is not sufficient to account for the antibiotic resistance phenotypes exhibited by Mar mutants
(Cohen et al, 1989; 1993b). Transcription ofacrAB, which codes for the AcrAB efflux pump is elevated in many marR mutants (Ma et al, 1995) and AcrAB efflux is mainly
48 responsible for the antibiotic resistance phenotype of Mar mutants, including quinolone resistance (Okusu et al, 1996). Similarly, MexAB, a homolog of AcrAB inP. aeruginosa, plays a major role in the intrinsic resistance of this organism to most commonly-used antibiotics (Pooleet al, 1993; Li et al, 1994).
2.7. The absence of plasmid-mediated quinolone resistance
Apparent plasmid-mediated resistance to nalidixic acid was reported in clinical isolates ofShigella dysenteriae type 1 from Kashmir (Panhotraet al, 1985) and from
Bangladesh (Munshi et al, 1987). Neither report has been validated. Ashraf et al{\99\) concluded that although the presence of plasmid pICDOl in the Bangladeshi strain might offer a survival advantage to its host under nalidixic acid stress, a direct relationship between the presence of the plasmid and nalidixic acid resistance could not be demonstrated. Crumplin (1987) suggested that although plasmids do not code for quinolone-resistance genes per se, they could contribute to the development of chromosomal quinolone resistance by conferring a mutator phenotype on their host.
Ambler et al (1993) demonstrated that one of the plasmids, pYDl, harboured by the nalidixic acid-resistant S. dysenteriae from Kashmir (Panhotraet al, 1985) does encode a mutator phenotype. Further in vitro experiments showed that a range of mutator plasmids, including pYDl, increased mutation to nalidixic acid resistance in cultures of
E. coli exposed to lethal concentrations of nalidixic acid in nutrient broth (Ambler and
Pinney, 1995).
2.8. Plasmid elimination by quinolones
Organisms harbouring certain R plasmids are more susceptible to quinolones than
49 plasmid-free strains (Crumplin and Smith, 1981). This enables spontaneously-occurring plasmid-free cells to outgrow plasmid-containing cells and thus bring about an effective
“elimination” of the plasmid (reviewed by Novick, 1969). Quinolones have been shown to cause plasmid elimination in vitro (Michel-Briand et al, 1986; Weisser and
Wiedemann, 1986) and in vivo (Mehtar et al, 1987; Mehtar, 1990), and to inhibit the conjugal transfer of some R plasmids (Gill and Iyer, 1982; Hirai et al, 1984). Active plasmid elimination may occur if replication of the plasmid is more susceptible than chromosomal replication to variation in supercoiling brought about by the action of the quinolone (Uhlin and Nordstrom, 1985).
3. DNA damage and repair
The structural integrity of DNA is vital to living organisms. Any alteration in its organisation will affect the biological activity of the molecule. Many chemical and physical agents and certain biological processes constantly threaten the integrity of DNA.
For this reason a number of DNA repair mechanisms have evolved to counter DNA damage. These may either correct the damage enzymatically in situ, or excise and replace damaged regions with homologous undamaged DNA sequences. If lesions are not corrected or removed, the organism may endure their effects by adjusting its method of
DNA replication or by initiating recombinational events. These may lead to base sequence alterations or genetic rearrangement. Therefore, in addition to their role in maintaining the integrity of DNA, repair mechanisms may also give rise to genetic alteration and, ultimately, mutation.
50 3,1. Photoreactivation
The DNA molecule absorbs light in the wavelength range of 240-300 nm leading to the formation of a variety of lesions within the DNA structure, the most prevalent type being the cyclobutane pyrimidine dimer (Setlow and Setlow, 1972; Gordon and
Haseltine, 1982; Haseltine, 1983). Other photoproducts found less frequently include the pyrimidine-pyrimidone (6-4) photoproduct (Miller, 1985; Glickman et al, 1986). Dimers distort the helical configuration of DNA leading to obstruction of polymerase activity and blockage of DNA replication (Villani et al, 1978). The polymerase bypasses the non coding photoproduct forming a gap in daughter strand DNA. Such gaps represent premutagenic sites.
Photoreactivation repairs pyrimidine dimers by an error-free light-dependent process (Sancar et al, 1985; Brash et al, 1985). DNA photolyase, encoded by thephrA and phrB genes in E. coll (Sancar and Rupert, 1978; Youngs and Smith, 1978;
Sutherland, 1981; Sancar et al, 1983), recognises and binds to a pyrimidine dimer in a light-independent reaction (Sancar et al, 1985; Brash et al, 1985). A single photon of light, of wavelength in the range 300-330nm, is absorbed by the flavin group of the enzyme (Iwatsuki et al, 1980), providing energy to cleave the dimer by electron donation.
This returns the primary structure of the DNA to its original form and allows the enzyme to dissociate (reviewed by Sancar and Sancar, 1988).
DNA photolyase can also enhance survival ofE. coli in the dark (Sancar et al,
1984). Binding of the enzyme to pyrimidine dimers stimulates formation of the UvrABC-
DNA complex (see Introduction section 3.2) by either changing the conformation of the pyrimidine dimer or by protein-protein interactions, which make the dimer more readily located by UvrABC (reviewed by Van Houten, 1990).
51 3.2. Nucleotide excision repair
Nucleotide excision repair is an error-free repair mechanism. It excises bulky, non-coding lesions, which block DNA replication. The process may be divided into five steps: 1) recognition of the non-coding lesion, 2) incision of phosphodiester bonds in the deoxyribose phosphate backbone either side of the lesion, 3) excision of the section of
DNA containing the lesion, 4) DNA synthesis to fill in the resulting gap, 5) sealing the nick to produce an intact DNA strand. The initial three steps in this process require the products of uvrA^ uvrB and uvrC (Howard-Flanders et al, 1965). These function co operatively as a multiprotein enzyme complex, now termed the (A)BC exinuclease, capable of the error-free repair of a broad spectrum of DNA lesions (Levit, 1982; reviewed by Lin and Sancar, 1992).
3.2.1. uvrA gene and product
The uvrA gene has a single promoter located in its 5' region. Expression ofuvrA is damage-inducible (Kacinski and Rupp, 1981; Kenyon and Walker, 1981) and is part of the SOS regulon (see Introduction section 3.5.1) (reviewed by Walker, 1984; 1985).
An SOS box within the promoter region ofuvrA binds LexA protein in vitro (Bertrand-
Burggraf et al, 1987; Sancar, 1987). Following damage to DNA, the level of UvrA increases from 25 to 250 molecules per cell (Selby and Sancar, 1990; reviewed by Van
Houten, 1990). The protein folds into two functional domains, each possessing one zinc finger (Doolittleet al, 1986; Navaratnam et al, 1989), enabling it to bind to DNA (Myles and Sancar, 1991).
52 3.2.2. uvrB gene and product
The uvrB gene has a complex promoter structure: its transcription is under control of both SOS-dependent (Fogliano and Schendel, 1981; Sancar et al, 1982; Sancar and
Rupp, 1983) and SOS-independent promoters (Sancar et al, 1982; Sancar and Rupp,
1983). Such control leads to constitutive expression of UvrB at a level of approximately
250 copies per cell. Following SOS induction, the level of UvrB protein increases to 1000 molecules per cell (Selby and Sancar, 1990).
3.2.3. uvrC gene and product
The level of UvrC is constitutively maintained at between 10 and 20 molecules per cell (Yoakum et al, 1980; Yoakum and Grossman, 1981). Four potential promoters exist within the 5' region ofuvrC (Moolenaaret al, 1987). Sequence analysis of the uvrC gene has indicated three putative LexA binding sites (SOS boxes) (Forster and Strike,
1988; Granger-Schnarr et al, 1986), but none of these bind LexA in vitro, indicating that uvrC is not inducible by DNA damage and is not under SOS control (Forster and Strike,
1988).
3.2.4. Binding of Uvr proteins to DNA lesions
The (A)BC excinuclease does not recognise specific alterations in bases or base- pairing. Rather, the enzyme complex identifies damaged regions of DNA due to strand deformation or helical distortion induced by many physical or chemical agents
(Sedliakova et al, 1987; Lilley, 1992). Such damage recognition allows the (A)BC excinuclease to repair a broad spectrum of DNA lesions (Van Houten, 1990).
Association of the UvrA and UvrB proteins occurs in a 2:1 ratio, producing a
53 complex, UvrAjB (Orren and Sancar, 1989a), which binds to DNA and produces a transient melting of the two strands. The complex travels in the 5'-» 3' direction along the
DNA for twenty to fifty base pairs before dissociating. If a lesion is encountered before the complex dissociates, the UvrB subunit is inserted into the DNA helix at the site of damage. Insertion proceeds in an ATP-dependent manner (Caron and Grossman, 1988a,
1988b; Oh et al, 1989). This leads to formation of an open stable UvrB-DNA complex, with release of the UvrA2 dimer (Orren and Sancar, 1989a; 1989b).
3.2.5. Incision of the phosphate backbone
The UvrC protein interacts with the UvrB-DNA complex in an ATP-independent reaction (Orren and Sancar, 1989a). This leads to cutting of the DNA sugar-phosphate backbone on either side of the DNA lesion (Sancar and Rupp, 1983; Yeung et al, 1983).
3 ' incision is effected by UvrB, and precedes the 5 ' cut catalysed by UvrC (Lin and
Sancar, 1992; Lin et al, 1992). Incision is a relatively slow reaction (Orrenet al, 1992) and is ATP-dependent, but does not require ATP hydrolysis (Sancar and Tang, 1993).
The incision complex hydrolyses the eighth phosphodiester bond 5', and the fourth or fifth phosphodiester bond 3' to the damaged region. The twelve- or thirteen-base oligonucleotide containing the damaged region is then released (Sancar and Rupp, 1983;
Yeung et al, 1983).
3.2.6. Post-incision events
After dual-incision of the sugar-phosphate backbone, the damaged oligonucleotide must be removed, UvrB and UvrC must be released (Caron et al, 1985;
Husain et al, 1985) and DNA repair synthesis must proceed (reviewed by Lin and Sancar,
54 1992). Three enzymes are required: the uvrD gene product (helicase II), DNA polymerase
I (Pol I) (Caronet al, 1985; Husain et al, 1985) and DNA ligase (Youngs and Smith,
1977; reviewed by Sancar and Sancar, 1988). Helicase II turns over UvrC, releasing the damaged oligonucleotide (Kumara et al, 1985; Caronet al, 1985; Husain et al, 1985).
This results in a stable UvrB-gapped DNA complex. Displacement of UvrB requires both helicase II and Pol I. UvrB turnover does not occur following Pol I binding at the 5' incision site, but requires repair synthesis by Pol I to displace UvrB from the gapped
DNA (Orren et al, 1992). Pol I binds to the accessible 3 -OH end of the gap (Van Houten et al, 1988), fills in the gap and displaces UvrB. The nick is then sealed by DNA ligase
(Gellert et al, 1977) resulting in repair patches, nearly all of which are twelve to thirteen nucleotides in length (Sibghat-Ullah et al, 1990; Orren et al, 1992).
Longer patches of repaired DNA occur at low frequencies (Caronet al, 1985).
These maybe greater than 1500 nucleotides (Kuemmerle et al, 1981; Cooper, 1982) and are dependent upon SOS induction (Cooper and Hanawalt, 1972; Hanawalt et al, 1979).
3.3. Methyl-directed mismatch repair
The possibility that E. coli possesses a mismatch repair mechanism was first suggested by White and Fox (1974) and was confirmed in the following two years (White and Fox, 1975; Wildenberg and Meselson, 1975; Wagner and Meselson, 1976).
Mismatch repair permits excision of incorrect mismatches in newly synthesised DNA and resynthesis of the excised region (Lu et al, 1983; 1984).
55 3.3.1. Strand discrimination between the parent and newly synthesised daughter strand
Méthylation of DNA by sequence-specific methylases proceeds after DNA replication (Marinus, 1976; Lyons and Schendel, 1984). Newly synthesised, daughter strands are therefore undermethylated with respect to parental strands. Such undermethylation provides a mechanism for strand discrimination between the correct parental DNA and the newly synthesised, incorrect, daughter strand (Wagner and
Meselson, 1976). This hypothesis was consistent with the observation thatdam mutants oiE. coli, deficient in méthylation at GATC sites (Marinus and Morris, 1975; Marinus et al, 1984), display a spontaneous mutator phenotype.
Direct evidence for the role of méthylation at GATC sequences in strand discrimination came from experiments utilising heteroduplex DNA, in which the two strands differed with respect to their state of flfam^-dependent méthylation (Meselson et al, 1980; huetal, 1983; Pukkila er a/, 1983; Kramer er a/, 1984). If GATC sequences are only methylated on one strand, repair is highly biased toward the unmethylated strand, with the methylated strand serving as template for the correction. If méthylation is absent from both strands, then mismatch repair still proceeds, but occurs equally on both strands.
Very little mismatch correction is observed in heteroduplex DNA in which both strands are highly methylated at GATC sequences.
3.3.2. Requfrement for the mutH, mutL, mutS and uvrD gene products for methyl- directed mismatch repair
Methyl-directed mismatch repair is dependent on functional mutH, mutL and mutS genes (Rydberg, 1978) and the uvrD gene product, DNA helicase II (Nevers and Spatz,
56 1975; Pukkila et al, 1983). muîH, mutL and mutS deficient mutants display a remarkably similar spectrum of spontaneous mutations (Schaaper and Dunn, 1987) with respect to
(1) the frequency at which the mutations occur; (2) the proportion of base substitution mutations to frameshift mutations, with base substitutions predominating; and (3) the predominance of transition mutations over transversion mutations. Such similarities suggested either that the three gene products act as a complex or that their action is very tightly coupled so that essentially the same phenotype results from the loss of any gene function.
mutS encodes a 97-kDa protein (Su and Modrich, 1986), which binds to DNA substrates that contain a mismatch and is critical for mismatch recognition. MutS also possesses a weak ATPase activity (Grilley et al, 1989; Haber and Walker, 1991), with the presence of ATP affecting the behaviour of the MutS protein on mismatch-containing
DNA. MutS binds to mismatches in the absence of ATP, but in its presence, an a-shaped loop structure is formed, stabilised at the DNA junction by bound MutS (Modrich, 1991).
Such loop formation may function in allowing the mismatch and the GATC sites to interact during repair.
The MutH protein ofE. coli exhibits a very weak endonuclease activity, dependent on Mg^^, which cuts unmethylated or hemimethylated GATC sequences in
DNA 5' to the G (Welsh et al, 1987). MutH was suggested to play a role in strand discrimination, since it cuts only the unmethylated strand in hemimethylated sequences
(Welsh etal, 1987).
The mutL gene codes for a 70-kDa protein (Lu et al, 1984), which binds to MutS- heteroduplex complexes (Grilley et al, 1989).
57 3.3.3. Initiation of mismatch repair
Use of purified MutH, MutL, and MutS proteins enabled the detailed mechanism of mismatch repair to be understood (Lahue et al, 1989; Modrich, 1991; Cooperet al,
1993). In addition to these three proteins and DNA helicase II, the system also requires single-strand binding protein (SSB), DNA polymerase III holoenzyme, exonuclease I, exonuclease VII (ûiQxse gene product), the RecJ exonuclease, DNA ligase, ATP, the four deoxynucleoside triphosphates, and NAD , the cosubstrate for E. coli DNA ligase.
The initial step of mismatch repair requires activation of MutH to cleave hemimethylated GATC sites. This requires MutS and MutL, ATP and a divalent cation
(Au et al, 1992). In the presence of MutS and ATP, the formation of an a-loop in the heteroduplex DNA brings together the mismatch and the GATC site, activating MutH to incise the strand containing the mismatch (Modrich, 1991). Excision then proceeds from the site of MutH cleavage in the appropriate direction to remove the mismatch (Grilley etal, 1993).
3.3.4. Excision of the mismatch
DNA helicase II displaces the incised strand, leaving it sensitive to excision
(Modrich, 1991; Grilley et al, 1993). Excision of mismatches occurs in both 3'^5' and
5'^3' directions (Grilley et al, 1993). 3'^5' exonuclease activity is provided by the sbcB gene product, exonuclease I (Lehman and Nussbaum, 1964; Cooper et al, 1993), and 5'
-3 ' activity by either exonuclease VII or the RecJ exonuclease (Chase and Richardson,
1974; Lovett and Kolodner, 1989). Hydrolysis of the excision tract initiates at the GATC site, proceeds beyond the mismatch and terminates at one of a number of discrete sites within a 100-nucleotide region beyond the mismatch (Grilley et al, 1993).
58 3.3.5. Repair of daughter strand gaps
DNA polymerase III resynthesises DNA to fill in the strand gaps left by the excision process (Lahueet al, 1987; Echols and Goodman, 1991). The nick in the newly synthesised DNA is then joined by DNA ligase in a reaction requiring Mg^^ and NAD
(Lehman, 1974).
3.4. Post-replication recombination repair
Post-replication recombination repair (PRR repair) fills gaps in daughter strand
DNA, formed when replicative DNA synthesis is blocked by a non-coding lesion. If left unrepaired, such gaps will lead to DNA degradation via the action of exonucleases.
Stalled polymerases bypass non-coding lesions and resume replication at the next
RNA initiation site, yielding postreplicative gaps in DNA (Johnson and McNeill, 1978).
Such recovery of replication (induced replisome reactivation, IRR) is a function of the
SOS response, requiring recA induction and at least one other gene product, IRR factor
(Khidhir et al, 1985).
3.4.1. The role of the RecBCD enzyme in PRR repair
The recB, recC and recD genes encode the multiprotein RecBCD enzyme, which appears to play two conflicting roles in homologous genetic recombination. Firstly,
RecBCD displays a double-stranded exonuclease activity, exonuclease V (ExoV), which stimulates recombination by interacting with Chi sites (Chaudhury and Smith, 1984).
Secondly, it acts as a helicase, generating regions of single-stranded DNA required by
RecA to promote homologous strand exchange (reviewed by Radding, 1982; Walker,
1984).
59 3.4.2. The RecF repair pathway
An alternative pathway, controlled by recF, accounts for only one percent of recombination in wild-type cells (Horii and Clark, 1973), but provides the only route for
PRR repair in recBC and sbcB mutants (Lloyd and Thomas, 1983).
Eight gene products are involved in RecF repair: recA, recF, recJ, recN, recO, recQ, ruv and uvrD (reviewed by Peterson et al, 1988). The recN, recQ, ruv and uvrD genes are under SOS control, suggesting the involvement of the RecF pathway in SOS
DNA repair (reviewed by: Smith, 1987; Peterson et al, 1988). The demonstration that
RecF activity is increased in DNA-damaged cells seems to confirm this (Lovett and
Clark, 1983). Binding of RecF to the ends of single-stranded, linear DNA has been proposed to protect against degradation, or facilitate DNA strand transfer (Griffin and
Kolodner, 1990).
3.4.3. Model for the RecBCD repair pathway
Together with the RecBCD enzyme, at least five other proteins are involved in recombination repair. These include RecA, SSB, DNA gyrase, Pol I and DNA ligase
(reviewed by: Smith, 1988; Meyer and Laine, 1990). Initially, the RecBCD enzyme, in the form of Exo V, is loaded onto single-stranded ends or gaps in double-stranded DNA
(Rosamond et al, 1979). The bound enzyme unwinds the two strands whilst travelling along the DNA, generating duplex intermediates with long single-stranded tails, and single-stranded DNA fragments (reviewed by Meyer and Laine, 1990; Rosenberg and
Hastings, 1991). The helicase activity of the RecBCD enzyme continues until a Chi site is encountered, a “hot spot” for recombinational activity, of which there are eight hundred or more on theE. coli chromosome (Weinstock, 1987). Here, the endonucleolytic action
60 of the RecBCD enzyme is stimulated by ATP. Cleavage of the DNA strand containing the Chi site occurs four to six nucleotides downstream from the 3' end (Tayloret al,
1985) as the enzyme moves 3'--5' relative to the Chi site (Chaudhury and Smith, 1984;
Ponticelli et al, 1985; Tayloret al, 1985). This cleavage reaction is enhanced by the presence of SSB (reviewed by Meyer and Laine, 1990).
Dissociation of the RecD subunit, the negative effector of the RecBCD enzyme complex (Thaler et al, 1989) then occurs, converting the enzyme from its ExoV form into its helicase form. The RecBC helicase then travels along the double-stranded DNA, unwinding tlie duplex molecule and exposing single-stranded regions. This helicase activity continues until the release of a 3 '-OH single-strand DNA end. It is this single stranded tail which acts as the substrate for presynapsis, requiring RecA protein and SSB
(Smith, 1988; Rosenberg and Hastings, 1991).
3.4.4. Search for homology and strand exchange
RecA protein binds cooperatively to single-stranded DNA in the presence of ATP and SSB forming right-handed nucleoprotein filaments (Flory and Radding, 1982;
Howard-Flanders et al, 1984; Egelman and Stasiak, 1986; reviewed by Meyer and Laine,
1990). A nucleoprotein aggregate is formed by these filaments, which searches for sequence homology in double-stranded DNA, creating many transient, non-specific contacts with the duplex (Chow and Radding, 1985; Gonda and Radding, 1986). Upon locating an area of sequence homology, the nucleoprotein filaments align and pair with the complementary regions of the double-stranded DNA to form a triple-stranded complex, termed a D-loop (Riddles and Lehman, 1985).
Creation of the D-loop initiates strand exchange. One strand of the duplex DNA
61 is unidirectionally displaced by the invading homologous region of single-stranded DNA
(DasGupta etal, 1980; Cox and Lehman, 1981). Exchange of DNA strands advances in the 5 '-^3 ' direction along the single strand, requiring ATP hydrolysis to drive the reaction at a rate of only a few bases per second (Cox and Lehman, 1981; Kahn et al, 1981). The result is an intact, undamaged heteroduplex DNA molecule.
3.4.5. RuvAB and RecG
Holliday junctions (Holliday, 1964) are the classical recombination intermediates formed during reciprocal strand exchange, when a partially single-stranded DNA molecule reacts with a fully duplex molecule that has suitable ends. Such RecA-directed pairing of two homologous DNA molecules requires one of the DNA molecules to be at least partially single-stranded (Roca and Cox, 1990; Kowelczykowski, 1991; Radding,
1991; West, 1992), leading to an exchange of strands that can proceed for several kilobases. Establishment of a Holliday junction allows RecA-promoted branch migration to occur, even if the DNA contains either lesions or considerable regions of nonhomology
(Roca and Cox, 1990; Cox, 1993).
More recently, two further mechanisms of promoting branch migration E.in coli have been discovered, one involving the ruvA and ruvB gene products and the other involving the recG gene product. The ruvAB operon is regulated as part of the SOS response (Shurvinton and Lloyd, 1982; Bensonet al, 1988; Shinagawa et al, 1988). RuvA binds to RuvB and the complex then binds specifically to Holliday junctions (Parsons et al, 1992) producing a RuvAB-Holliday junction complex (Parsons and West, 1993). The principal function of RuvA appears to be to target the RuvB enzyme to the site of the junction, where RuvB promotes branch migration in an ATP-dependent fashion (Iwasaki
62 et al, 1989; Mülleret al, 1993a; 1993b). Such RuvAB-mediated branch migration is much faster and more energy efficient (Tsaneva et al, 1992a) than that promoted by RecA
(Roca and Cox, 1990). In addition, levels of UV-induced lesions present in DNA that inhibit RecA-mediated strand exchange can be bypassed by RuvAB (Tsaneva et al,
1992a).
A third protein, the ruvC gene product, resolves Holliday junctions into recombinant molecules by precise endonucleolytic cleavage across the point of strand exchange (Connelly et al, 1991 ; Dunderdale et al, 1991 ; Iwasaki et al, 1991 ; Bennett et al, 1993). The ruvC gene is found in a separate operon to theruvAB and is not SOS- regulated (Sharpies and Lloyd, 1991; Takahagi et al, 1991). RuvC-mediated cleavage of
Holliday junctions is symmetrical, producing recombinants of either the patch or splice type (Mandai et al, 1993; West, 1994).
The recG gene product is another branch migration-promoting protein, which like
RuvB is a DNA-dependent ATPase that binds to Holliday junctions (Lloyd and Sharpies,
1993a). Under certain conditions, RecG catalyses branch migration in the reverse direction to that of RecA- and RuvAB-driven strand exchange (Whitby et al, 1993).
RuvAB proteins seem to play a more important role than RecG in the processing of damaged DNA, since ruvAB mutants have a greater sensitivity to UV light than recG mutants (Lloyd and Buckman, 1991; Lloyd and Sharpies, 1991).
3.5. The SOS response
Many of the genes expressed in E. coli in response to DNA damage are part of the highly coordinated SOS regulon. SOS induction leads to an increased capacity for both
PRR and excision repair, together with an increase in SOS processing and mutagenesis,
63 inhibition of cell division and respiration, and prophage induction (reviewed by Witkin,
1974; 1976; Little and Mount, 1982; Walker, 1984; 1985).
3.5.1. The SOS regulon
The SOS regulatory network ofE. coli is comprised of at least twenty opérons.
These DNA damage inducible {din) genes are distributed throughout the bacterial chromosome, and are under the control of two proteins, which play opposing roles in the regulation of the SOS response. The LexA protein, the negative regulator of the SOS regulon, represses transcription of alldin genes under normal conditions. LexA-regulated genes include recA (Casaregola et al, 1982a), uvrA (Kenyon and Walker, 1981), uvrB
(Fogliano and Schendel, 1981), uvrD (Siegel, 1983), umuDC (Bagg et al, 1981), recN
(Lloyd et al, 1983), ruvAB (Shurvinton and Lloyd, 1982) and sulA (Huisman and D’Ari,
1981). The RecA protein, the positive regulator, is converted into an activated form
(RecA*) following DNA damage, which facilitates the autoproteolytic cleavage of the
LexA repressor. Therefore, the SOS regulon can exist in two states. In the un-induced state, during normal growth, the LexA protein represses target genes. In the induced state, transcription of SOS genes is greatly increased leading to the expression of the various
SOS phenotypes (Figure 1) (reviewed by Little and Mount, 1982; Walker, 1984; 1985;
Peterson etal, 1988).
3.5.2. Regulation of the SOS response
In the absence of DNA damage, all din genes of the SOS regulon are repressed by the LexA protein, including lexA itself, whose expression is autoregulated by its own gene product (Brent and Ptashne, 1981; Little et al, 1981; Brent, 1982). SOS boxes,
64 The SOS regulon LexA pool
Low level Low level R ecA expression
recA lexA >20 target genes U ninduced State
254nm
e A rp e s r a c m ua e D A damage LexA repressor accum ulatesDN
Inducing signal Decreasing level of RecA
Intermediate State
RecA protein activation Decreased level of signal
LexA repressor cleavage DNA repair
Inducing signal
^ © + © ^
DNA repair and other SOS functions
recA lexA > 20 target genes
Induced State
Figure 1. Model of SOS regulation.
65 regions of DNA approximately twenty base pairs in length, are specific operator sequences, which lie upstream of the repressed genes (Little et al, 1981). The SOS boxes show significant DNA homology with one another and it is to these regions that the LexA protein binds, in the form of dimers (Schnarr et al, 1985; Thliveris et al, 1991). Such
LexA binding blocks the access of RNA polymerase to the promoter regions of the repressed genes, inhibiting their transcription (Little et al, 1981; Brent and Ptashne,
1981). Two SOS boxes, binding LexA with approximately equal affinity, are present in the operator regions of many genes regulated by LexA and RecA, including lexA itself
(Howard-Flanders, 1981). The recA gene is controlled by only one SOS box, to which
LexA binds with greater affinity. These differences in LexA binding affinity may account for the differences in the timing and extent of SOS induction following DNA damage
(Brent and Ptashne, 1981).
Many SOS genes, including recA, are expressed at low, but significant levels whilst in the repressed state (reviewed by Walker, 1984). In the absence of an inducing signal, cellular levels of RecA protein are maintained at about one thousand RecA monomers per cell (Karu and Belk, 1982; Salles and Paoletti, 1983), but can increase as much as twenty-fold upon SOS induction (Karu and Belk, 1982; Weisemann et al, 1984).
Following DNA damage or inhibition of DNA replication, the generation of an inducing signal reversibly activates RecA into RecA* coprotease. The latter facilitates the autoproteolytic cleavage of LexA, leading to a decrease in the pool of LexA repressor protein and increased expression of the various SOS genes. Genes whose operator regions bind LexA relatively weakly are the first to be expressed. On continued induction, more
RecA protein is activated, leading to increased cleavage and inactivation of LexA and expression of genes whose operator regions bind LexA with high affinity. On completion
66 of DNA repair, the inducing signal is no longer produced and RecA returns to its proteolytically inactive state. Levels of LexA protein increase, leading once again to repression of the SOS regulon and return of the cell to the uninduced state (reviewed by:
Little and Mount, 1982; Walker, 1984, 1985; Little, 1991).
3.5.3. The SOS-inducing signal and RecA protein activation
Induction of the SOS response requires the production of an inducing signal, generated by DNA damage, which reversibly activates RecA protein to its coprotease form, RecA* (Casaregola etal, 1982b; Little, 1983). The overproduction of RecA is not sufficient to induce the SOS response (Uhlin and Clark, 1981; Clark, 1982; Quillardet et al, 1982a). The inducing signal has still not been defined clearly in vivo (Little, 1991;
Friedberg et al, 1995), although single-stranded regions of DNA, oligonucleotides and nucleotide cofactors such as dATP or ATP, and the involvement of RecA protein and
DNA replication, have all been suggested (Smith and Oishi, 1978; Craig and Roberts,
1980; Chaudhury and Smith, 1984; Bailone et al, 1985; Devoretet al, 1988; Sassanfar and Roberts, 1990; reviewed by Friedberg et al, 1995). A wide variety of agents damage
DNA and induce the SOS response (Bagg et al, 1981; Quillardet et al, 1982b), the nature of the DNA damage determining whether RecBCD and/or RecF are required for SOS induction (Gudas and Pardee, 1976; Armengod and Blanco, 1978; Karu and Belk, 1982;
Volkert and Hartke, 1984; Thoms and Wackemagel, 1988; Simic et al, 1991; Clark,
1991). DNA unwinding by the RecBCD helicase is necessary for SOS induction following treatment with nalidixic acid (Gudas and Pardee, 1976; Karu and Belk, 1982;
Chaudhury and Smith, 1985), and the recF gene product has been suggested to be required for SOS signal formation following UV-irradiation (Simicet al, 1991).
67 3.5.4. RecA*-mediated proteolytic cleavage of the LexA repressor
Cleavage of LexA is the rate-limiting step in SOS induction (Little, 1984, 1991).
If the inducing signal is weak, only a low rate of LexA cleavage is observed and the cell remains uninduced. Partial induction results in only a proportion of SOS genes being expressed. Where the inducing signal is strong and persistant, the cells are fully induced and express the whole range of SOS responses (Littleet al, 1981; Little, 1984, 1991).
RecA* does not act as a classic protease in the cleavage of the LexA repressor.
It is essential for LexA cleavage at physiological pH, whereas at alkaline pH, LexA cleavage occurs by autodigestion in the absence of RecA*. This suggests that RecA* acts indirectly by facilitating the autodigestion of LexA at physiological pH (Little, 1984;
1991).
Lys-156 and Ser-119 lie close together in the folded LexA protein. Lys-156 is required for the activation of a hydroxyl group on Ser-119 (Slilatyet al, 1986; Slilaty and
Little, 1987; Roland and Little, 1990; Little, 1991), which attacks the Ala-84/Gly-85 bond in the centre of the LexA protein (Little, 1991). RecA* lowers the pk^ of the Lys-156 residue of LexA, to allow cleavage at physiological pH, resulting in two LexA fragments of approximately equal size, possessing no repressor activity (Little, 1991).
3.5.5. SOS processing
The umuDC gene products are required for most chemical and UV mutagenesis in E. coli (Elledge and Walker, 1983; reviewed by Walker, 1984; 1985; Peterson et al,
1988; Battista et al, 1990). Strains carrying mutations in either genes are immutable
(Kato and Shinoura, 1977; Steinbom, 1978; Elledge and Walker, 1983; Shinagawa et al,
1983). SOS-induced mutagenesis also requires RecA protein activity, for which three
68 roles have been proposed (Nohmiet al, 1988; Dutreix et al, 1989; Ennis et al, 1989;
Sweasy et al, 1990; Bates et al, 1991).
UmuD shares sequence homology with the carboxy terminal domain of LexA
(Perry et al, 1985; Battista et al, 1988). Both proteins undergo autodigestion at alkaline pH and form homodimers in solution (Dutreixet al, 1989; Battista et al, 1990). RecA* facilitates autodigestion of both UmuD and LexA at Cys-Gly and Ala-Gly peptide bonds respectively. Cleavage of LexA produces two non-functional fragments leading to derepression of the SOS regulon and induction of the SOS response, whereas UmuD cleavage generates the shorter, carboxy terminal fragment, UmuD' (Nohmi et al, 1988).
The latter is both essential and sufficient for mutagenesis (Nohmi et al, 1988). UmuD cleavage therefore represents a further point of regulation in SOS mutagenesis.
A third role for RecA protein in the regulation of SOS mutagenesis was proposed based upon the findings that mutagenesis is not detected in strains defective for the two known RecA activities. In such strains, LexA cleavage is rendered unnecessary by a
/exT(Def) mutation that inactivates the repressor, and the need for processing of UmuD is bypassed via the carriage of a plasmid carrying a umuDC allele encoding the activated
COOH-terminal fragment, UmuD' (Nohmi etal, 1988; Dutreix et al, 1989; Sweasy et al,
1990).
Mounting evidence suggests the third role of RecA involves direct interaction with Umu mutagenesis proteins. UmuD' forms homodimers and associates with UmuC to form a UmuD'2C complex (Woodgate et al, 1989). Both UmuC and UmuD' interact with RecA* (Freitag and McEntee, 1989; Frank et al, 1993) to target the relatively small number of mutagenically active UmuD'jC complexes at DNA lesions (Sweasy et al,
1990; Bailoneet al, 1991; Dutreix et al, 1992; Frank et al, 1993). UmuD'2C interacts
69 with both free ssDNA and ssDNA that has been precoated with RecA (Bruck et al, 1996).
It has therefore been proposed that UmuD ^C associates with a RecA nucleoprotein filament either through an interaction between UmuC-RecA* (Freitag and McEntee,
1989), or UmuD'-RecA* (Frank et al, 1993). Formation of a RecA* nucleoprotein filament at a lesion in double-stranded DNA (Rosenberg and Echols, 1990), would provide an efficient way to target Umu mutagenesis proteins, via UmuD'C-RecA* interactions, to locations in the cell where they are required to facilitate error-prone translesion DNA synthesis (Frank et al, 1993; Bruck et al, 1996).
3.5.6. SOS mutagenesis
Premutagenic sites opposite UV- or chemically-induced non-coding lesions are generated by interruption of DNA replication at the lesion (Rupp and Howard-Flanders,
1968; Livneh et al, 1993) or by nucleotide excision repair of closely opposed lesions
(Bresler, 1975; Lam and Reynolds, 1987), where the removal of a lesion from one strand results in an excision gap opposite the second lesion. The principle underlying error- prone translesion DNA synthesis is that the replication fork stalls at the non-coding lesion. SOS-induced UmuC, UmuD', and RecA* proteins then interact with DNA polymerase III (Pol III) to shepherd it past the lesion, with the likely production of a base substitution mutation opposite the lesion (Lawrenceet al, 1990).
Effective function of the mutagenic translesion DNA synthesis pathway requires the UmuD'jC complex, activated RecA protein (RecA*) and Pol III (Hagansee et al,
1987; Nohmi et al, 1988; Dutreix et al, 1989; Sweasy et al, 1990; Bailoneet al, 1991).
An initial model proposed for translesion DNA synthesis involved two steps: misincorporation of nucleotides directly opposite the lesion, requiring Pol III and RecA
70 (Bridges, 1988; Sharif and Bridges, 1990), followed by bypass of the lesion involving Pol
III and UmuD'C (Bridges and Woodgate, 1984; Sharif and Bridges, 1990; Napolitano et al, 1997). However, Tang et al (1998) recently came to the conclusion that the efficiency of both base incorporation and strand extension are determined by UmuD' 2C and RecA* since a significant reduction in lesion bypass is demonstrated in the absence of UmuD' 2C.
Single-stranded binding (SSB) protein is also essential for translesion synthesis (Tang et al, 1998). This requirement may be an indirect consequence because of the action of SSB protein in eliminating DNA secondary structures, keeping the p-clamp of Pol HI holoenzyme firom sliding off linear DNA (Bloomet al, 1996) and helping RecA achieve its activated state (Roca and Cox, 1997).
3.5.7. DNA polymerases
Whilst DNA polymerase III (Pol III) is required for SOS mutagenesis (Bridges et al, 1976), the possibilty that DNA polymerase I (Pol I) is involved was ruled out because
Pol I-deficient strains are UV-mutable (Rondo et al, 1970; Witkin, 1970). DNA polymerase II (Pol II) is SOS inducible (Bonner et al, 1988; 1990; Iwasaki et al, 1990).
However,polB mutants show no defect in UV repair or mutagenesis (Rowet al, 1993;
Escarceller et al, 1994). Pol II is therefore not essential for UV repair or mutagenesis, but this does not exclude the possibilty that when present, it can take part in these processes
(Tessman and Rennedy, 1993; Rangarajan et al, 1997).
3.5.8. Reduced fidelity of DNA polymerase III
Pol III holoenzyme consists of ten individual subunits, some of which can also associate with Pol II (Bonneret al, 1988). These subunits are organised into
71 subassemblies, a x dimer, which links two core polymerases; the T clamp-loading complex, composed of the T-, ô-, ô \ and i-subunits; the P-dimer subunit constituting the sliding clamp; and the core polymerase composed of the a-catalytic subunit (Maki and Komberg, 1985), the e-subunit encoding the 3 '^5 ' exonuclease-proofreading activity
(Scheuermann and Echols, 1984), and the 0-subunit of unknown function (Studwell-
Vaughan and O’Donnell, 1993; reviewed by: Kelman and O’Donnell, 1995; Herendeen and Kelly, 1996).
Villani et al (1978) were the first to suggest, later reiterated by others (Echols,
1982; Echols and Goodman, 1991), that an altered, relaxed-fidelity form of DNA polymerase, defective in DNA proofreading, was necessary to enable DNA synthesis past a non-coding lesion. However, recent evidence demonstrates that low levels of bypass occur in the presence of normal proofreading. It would appear, therefore, that inhibition of proofreading is not an absolute requirement for bypass synthesis (Paz-Elizuret al,
1996). It is interesting that overproduction of the e-subunit of Pol III inhibits UV mutagenesis (Foster et al, 1989; Jonczyket al, 1988): an effect attributable to the DNA binding properties of e, rather than its proofreading function (Kanabus et al, 1995).
Efficient translesion synthesis has also been demonstrated in the absence of UmuDC, if the proofreading activity of the e-subunit is also absent (Vandewiele et al, 1998).
There is good evidence that the RecA/UmuD'C protein complex interacts with the e subunit of Pol III (Fersht and Knill-Jones, 1983; Lu et al, 1986; Foster and Sullivan,
1998) and that the 3 '^5 ' exonuclease proofreading activity of Pol III is suppressed during translesion synhesis (Woodgateet al, 1987; Slater and Maurer, 1991; Fijalkowska et al,
1997). Although how the latter is achieved is less clear (Vandewiele et al, 1998). The p- subunit of Pol III also appears to have an effect on translesion DNA synthesis.
72 Overproduction of the P-subunit reduces UV mutagenesis, which can be relieved by providing excess UmuD'C-like proteins (Tadmor et al, 1992). Tadmor et al (1992) proposed that under conditions of SOS induction, an altered Pol III enzyme is formed which contains only a low concentration of the p-subunit. Interaction of this modified Pol
III with RecA, UmuC and UmuD' allows DNA lesions to be bypassed. If the proofreading activity of the e-subunit is also suppressed or inactive during translesion synthesis (Woodgate et al, 1987; Slater and Maurer, 1991; Fijalkowska et al, 1997), then the introduction of incorrect bases opposite and beyond the lesion result in an increased frequency of mutation (reviewed by: Echols and Goodman, 1991; Walker, 1995).
3.5.9. Induction of the SOS response by the quinolone antibacterials
Quinolone antibacterial agents are potent inducers of the SOS regulon (Gudas and
Pardee, 1976; Phillips etal, 1987; Piddock and Wise, 1987), requiring the participation of three proteins: RecA, RecBCD, and LexA. Maximal SOS induction occurs at the most bactericidal concentrations of quinolones, which led Phillipset a/ (1987) to suggest a role for the SOS response in the mechanism of quinolone action. For SOS induction to occur, an inducing signal must be generated (see Introduction section 3.5.3). This may consist of short oligonucleotides and/or single-stranded DNA. Short oligonucleotides induce the
SOS response when introduced into permeabilised cells (Oishi et al, 1979; Irbe and
Oishi, 1980) and single-stranded DNA may be shown to induce the SOS response in vitro
(Craig and Roberts, 1980). Both of these inducing signals could be produced by the
RecBCD enzyme, acting on quinolone-damaged DNA, through its ability to unwind and degrade DNA. RecBCD possesses four activities: unwinding of duplex DNA, exonucleolytic degradation of duplex DNA, exonucleolytic degradation of single
73 stranded DNA, and endonucleolytic degradation of single-stranded DNA (reviewed by
Muskavitch and Linn, 1981; Myers and Stahl, 1994). However, the double-stranded exonuclease activity of RecBCD does not appear to be involved in nalidixic acid- mediated induction of the SOS response, since nalidixic acid treatment of recBCD mutants lacking this activity still leads to SOS induction (Bailone et al, 1985; Chaudhury and Smith, 1985). Neither do the single-strand-specific nuclease activities of RecBCD appear to be required, because other recBC mutations that block SOS induction have no effect on the single-strand-specific nuclease (Kushner, 1974; Karu and Belk, 1982). It is, therefore, likely that the DNA unwinding activity of the RecBCD enzyme is necessary for signal induction (Bailoneet al, 1985; Chaudhury and Smith, 1985). Quinolones are
SOS-dependent mutagens (McDaniel et al, 1978; Phillips et al, 1987; Power and Phillips,
1993). However, they appear non-mutagenic when tested using conventional (excision- repair deficient) Ames strains of Salmonella typhimurium (Schlüter, 1986), although high-level umuC expression may be demonstrated (Ysem et al, 1990). This apparent anomaly was resolved when quinolones were shown to be positive mutagens in strains ofS. typhimurium proficient for excision repair (Gocke, 1991).
3.5.10. Repair of quinolone-induced DNA lesions
recA and recBC mutants are more sensitive to the lethal effects of quinolones than are wild-type cells (McDaniel et al, 1978), with a recAlS recBIl double mutation displaying a greater sensitivity to nalidixic acid than either mutant alone (Lewin et al,
1989). A similar result was observed with recF and recBC mutants (McDaniel et al,
1978) suggesting that several recombinational pathways may contribute to the repair of quinolone-induced DNA damage.
74 RecA and RecBCD are necessary for both PRR and for SOS induction by quinolones. Much work has, therefore centred on determining the relative contributions of recombination and the SOS response to cell survival after quinolone damage.
Recombination appears to be the major mechanism, sincelexA3 mutants, which are defective for SOS induction but not for recombination, are only slightly more sensitive to nalidixic acid than lexA^ strains (McDaniel et al, 1978; Karu and Belk, 1982; Lewin et al, 1989). However, strains carrying mutations that block SOS induction but not recombination (recA430 and lexA3), display an increased susceptibility to ciprofloxacin and norfloxacin, suggesting that SOS repair plays a role in the repair of fluoroquinolone- induced DNA damage (Urios et al\99\', Howard et al, 1993a). The mechanism by which quinolone-induced DNA damage is repaired appears to vary according to the quinolone being examined. Howard et al (1993a) proposed that different repair pathways may be responsible for the repair of quinolone-induced DNA damage. Whereas damage resulting from nalidixic acid exposure is repaired only by recombination repair, DNA damage caused by ciprofloxacin, ofloxacin, and norfloxacin may be subject to SOS repair, in addition to recombination repair (Howard et al, 1993a).
A feature common to both ciprofloxacin and norfloxacin is their interaction with topoisomerase IV. The susceptibility of quinolone-resistantgyrA mutants to ciprofloxacin is increased by recA430 and lexA3 mutations (Urios et al, 1991; Chen et al, 1996), suggesting a role for the SOS response in the recovery of these resistantgyrA mutants from fluoroquinolone-induced DNA damage. ArecA null mutation also renders resistant gyrA mutants more sensitive to ciprofloxacin than thelexA3 mutation (Drlica and Zhao,
1997), indicating that both SOS induction and -dependent recombination participate
75 in topoisomerase IV-mediated killing.
4. Plasmids
Plasmids are circular, double stranded DNA molecules, distributed throughout a large range of bacterial species, and also some eukaryotes. The size of plasmids varies from a few kilobases (kb) up to several hundred kilobases (reviewed by Kües and Stahl,
1989). Plasmids confer upon the host a number of beneficial, but inessential, phenotypes, such as resistance to a variety of antibacterial agents and the ability to produce toxins and virulence factors (reviewed by Krawiec and Riley, 1990). Larger plasmids carry genes required for their own maintenance and for their transfer between organisms via cell-to- cell contact (Nordstrom, 1985; Timmisetal, 1986).
4.1. Plasmid-encoded mutation-enhancing opérons
In addition to the chromosomalumuDC opérons ofE. coli (Kitagawa et al, 1985;
Perry et al, 1985) and S. typhimurium (Smith et al, 1990; Thomas et al, 1990), many plasmid-bome mutation-enhancing opérons have been identified. These include: mucAB from pKMlOl (Perryet al, 1985), impCAB from the Incl plasmid TPllO (Lodwick et al,
1990), samAB from the cryptic plasmid ofS. typhimurium LT2 (Nohmi et al, 1991), rumAB from R391 (Kulaeva et al, 1995), mucAB-\\kQ homologs from R471a
{mucAB^^^^^ and R446b {mucAB^^f^^) (Kulaeva et al, 1995; 1998), and rulAB from plasmids of the plant pathogen Pseudomonas syringae (Sundin et al, 1996). More recently, an impCAB operon has been found on the large, 220 kb virulence plasmid pVP of Shigella flexneri (Runyen-Janecky et al, 1999). Such plasmid-encoded mutation- enhancing opérons are widely distributed throughout the range of plasmid incompatibility
76 groups (Molinaet al, 1979; Pinney, 1980; Upton and Finney, 1983; reviewed by Strike and Lodwick, 1987).
4.2. Relationship of plasmid-encoded mutation-enhancing systems to the E, coli chromosomal umuDC operon
All of the plasmid-encoded mutation-enhancing opérons so far identified are structurally and functionally homologous to the chromosomally-encodedumuDC operon ofE. coli. All show homology at the nucleotide level and, with the exception of the impCAB operon, are di-cistronic, encoding a small UmuD-like protein upstream of a larger UmuC-like protein. In addition, sequence analysis reveals significant amino acid homology between the MucA, Imp A, SamA, RumA, MucAR^^^b, MucA^^yia and RulA proteins, and between the MucB, ImpB, SamB, RumB, MucB^^^^b, MucB^^yig and RulB proteins. Such similarities reveal conservation of both a LexA binding site and a cleavage site in MucA, ImpA, SamA, RumA, MucA^^^^b, MucA^^yi^ and RulA, indicating that these proteins are under SOS control and undergo RecA*-mediated autodigestion to the mutagenically active form (Perry et al, 1985; Kulaeva et al, 1995; 1998; Lodwick et al,
1990; Nohmiet al, 1991; Sundin et al, 1996).
ImpC, a small protein of molecular weight 9,491 Da, has no equivalent in any other mutation-enhancing operon and is not absolutely required for the DNA repair and mutation phenotype conferred by the impCAB operon (Lodwick et al, 1990). The nucleotide sequence of the impCAB operon fromS. flexneri is 96% identical to that of the impCAB operon on the conjugative plasmidTPllO (Lodwick et al, 1990; Runyen-
Janecky et al, 1999). ImpC is homologous to Tum, encoded by prophage 186 inE. coli
(Brumby et al, 1996) and to the E. coli DinI enzyme (Yasuda et al, 1996; Runyen-
77 Janecky et al, 1999). Tum is an antirepressor that causes induction of prophage 186 by directly interfering with the ability of the phage repressor to bind DNA (Shearwin et al,
1998). DinI inhibits RecA*-mediated self cleavage of LexA and UmuD (Yasuda et al,
1998). Thus, ImpC may have a similar role in the regulation of ImpA autodigestion to
ImpA' (Runyen-Janecky et al, 1999).
4.3. mucAB
mucAB is the most studied of all plasmid-encoded mutation-enhancing opérons.
It suppresses umuDC mutants (Walker and Dobson, 1979; Perry and Walker, 1982) in a recA^ /ex4^-dependent fashion, indicating transcriptional control by the LexA repressor
(Walker, 1977; Elledge and Walker, 1983). Indeed, a site has been identified in the 5'- flanking region of themucAB that is homologous to the LexA recognition sites of other SOS loci (Perryet al, 1985).
4.3.1. Processing of the MucA protein
'The mucAB operon encodes two proteins of molecular weights 16,371 Da (MucA) and 46,362 Da (MucB) (Perry et al, 1985), similar in size to the UmuD (15,064 Da) and
UmuC (47,681 Da) proteins respectively (Kitagawa et al, 1985). Significant amino acid sequence homology also exists between UmuD and MucA (41%), and UmuC and MucB
(55%) (Perry et al, 1985). In addition, the Ala25-Gly26 bond of MucA and the Cys24-
Gly25 bond of UmuD are in positions corresponding to the Ala-Gly proteolytic cleavage sites of LexA and the À repressor (Perry et al, 1985). Like UmuD (see Introduction section 3.5.5), the intact 16 kDa MucA protein is processed in a RecA*-dependent fashion to a 13 kDa protein, MucA' (Shiba et al, 1990). MucA cleavage occurs at the
78 Ala25-Gly26 bond, analogous to the Cys24-Gly25 bond cleaved in UmuD (Woodgate et a/, 1989). HoweverrecA430 strains, which cannot process MucA, are nearly as proficient as recA^ strains in MucAB-directed UV mutagenesis (Blanco et al, 1986; Nohmi et al,
1988), whereas UmuDC is not expressed in recA430 strains (Ennis et al, 1985; Blanco et al, 1986). This suggests that intact MucA is nearly as active as the processed protein in UV mutagenesis (Shiba et al, 1990).
Other phenotypes also distinguish the mucAB operon fi*om chromosomalumuDC. mucAB enhances UV-protection and mutagenesis to a greater extent thanumuDC, and the basal levels of DNA repair and mutagenesis conferred by mucAB in cells in which the
SOS response has not been induced are higher than those conferred by umuDC (Walker,
1977; Blanco et al, 1986). Higher mucAB gene expression or differential regulation by
LexA of mucAB and umuDC do not account for these differences (Blancoet al, 1986). mucAB also restores inducible mutagenesis to normally non-mutable lexA(}n&) strains
(Monti-Bragadin et al, 1976; Waleh and Stocker, 1979; McNally et al, 1990) and recA433 and recA435 strains (Hauser et al, 1992) ofE. coli. However, such enhanced effects of mucAB are not attributable to the high activity of unprocessed MucA, as proposed by Shiba et al (1990), but rather to a higher efficiency of processing of MucA to the mutagenically active form MucA' (Hauseret al, 1992).
4.3.2. MucAB-mediated mutagenesis
The active E. coli chromosomal mutagenesis complex, composed of two molecules of UmuD' and one of UmuC (Woodgateet al, 1989; Bruck et al, 1996), is required for error-prone translesion DNA synthesis by DNA polymerase III (Bridges and
79 Woodgate, 1985; Echols and Goodman, 1990; Rajagapolanet al, 1992). High-level expression of the UmiiD ^C complex also prevents recombinational repair of UV- damaged DNA (Sommer ar a/, 1993; Szpilewoska er a/, 1995; Boudsocqet al, 1997). The physical interaction between RecA polymer and the UmuD'jC complex (Witkin et al,
1987; Dutreix et al, 1989; Freitag and McEntee, 1989; Sweasy et al, 1990; Bailoneet al,
1991; Frank et al, 1993; Sommeret al, 1993; 1998) can account for inhibition of both homologous recombination and mutagenesis (Sommeret al, 1993; Boudsocqet al, 1997).
Translesion synthesis involves targetting of UmuD ^C to a DNA lesion by RecA via binding of the complex to the tip of a RecA filament (Sommeret al, 1993; Boudsocqet al, 1997; Sommer et al, 1998) and interaction with a stalled DNA Pol III enzyme
(Jonczyk et al, 1988; Nowicka et al, 1994; Skaliter et al, 1996; Foster and Sullivan,
1998). This has been confirmed in vitro , where translesion synthesis by Pol III has been shown to require the presence of UmuD', UmuC, and RecA (Rajagopalanet al, 1992).
MucA and MucA' interact with themselves and with each other, whereas MucB shows no self-interaction but does interact with both MucA and MucA' (Sarov-Blat and
Livneh, 1998). These interactions parallel those of the homologous UmuD and UmuD' proteins with UmuC (Woodgateet al, 1989; Battista et al, 1990).
Complexes are also formed between MucA' protein and RecA. However, unlike
UmuD and UmuD', which complex on RecA-coated DNA, MucA' can interact with
RecA in a DNA-independent manner. However, in the presence of DNA, the MucA -
RecA complex is specifically targetted to DNA (Frank et al, 1993). In addition, like
UmuD'jC, MucA'B interacts with subunits of Pol III (Tadmor et al, 1992), which may stimulate translesion synthesis. MucB also interacts with single-stranded binding protein
80 (SSB), greatly changing the structure of the SSB-ssDNA complex, which may be another role for MucB in SOS-regulated mutagenesis (Sarov-Blat and Livneh, 1998). Recently, it has been demonstrated that, like its chromosomal homolog, the MucA B complex also inhibits homologous recombination (Venderbureet al, 1999). However, although both
UmuD ^C and MucA'B are targetted to DNA lesions through interaction with RecA polymers (Bailoneet al, 1991; Frank et al, 1993), the two mutagnenic complexes appear to function at distinct sites, since they act synergistically to inhibit recombination
(Venderbure et al, 1999).
MucAB proteins are more proficient in mutagenesis than UmuDC (Upton and
Pinney, 1983; Blancoet al, 1986; Kulaeva et al, 1995), which was initially believed to result from the more efficient processing of MucA to its active form (Hauseret al, 1992).
However, it now seems that such enhanced mutagenic potential is determined by the increased efficiency of MucA'B to aid synthesis across a large range of DNA lesions
(Lawrence et al, 1996; Venderbure et al, 1999).
4.4. Spectrum of activity of plasmid-encoded mutation-enhancing opérons
'The MucAB complex is a stronger promotor of translesion synthesis across abasic lesions than UmuDC. The abasic lesion is produced by a variety of mutagens (Loeb and
Preston, 1986), and therefore has been used as a model for a number of different mutagenic lesions (Strauss, 1991). Enhanced translesion synthesis was suggested to depend equally on an inherently greater capacity to promote this process and on an increased susceptibility of the MucA protein to proteolytic processing (Lawrenceet al,
1996). The enhanced mutagenic potential of mucAB is exploited in the Ames S.
81 typhimurium mutagen tester strains (McCann et al, 1975).
MucAB also alters the relative frequencies of the two major classes of mutations induced byT-T cyclobutane dimers. About five-fold more 3' A than 3' T-C mutations are found in umuDC strains, whereas this ratio is reversed in mucAB and rumAB strains, although overall mutation frequencies are similar for all strains (Szekereset al, 1996).
mucAB, but not umuDC, promote ciprofloxacin-induced mutagenesis in either chromosomal or plasmidichisG428 targets in S. typhimurium. A single copy ofmucAB is sufficient to promote UV mutagenesis, whereas several copies are required to observe ciprofloxacin-induced mutagenesis, suggesting that MucAB-facilitated mutagenic repair of quinolone-induced damage is not very efficient (Clerch et al, 1996).
In most backgrounds, the rumAB operon promotes higher levels of inducible mutagenesis than mucAB. The exception is in a recA430 strain where mucAB is more active than rumAB. This is thought to result from inefficient posttranslational activation of RumA in the recA430 background (Kulaeva et al, 1995). Similarly, although RumA appears more susceptible than UmuD to proteolytic processing, the inherent capacity of
RumAB to promote translesion synthesis at abasic lesions is no greater than UmuDC
(Lawrence et al, 1996).
SamAB proteins, unlike MucAB, can only promote UV mutagenesis when expressed at high-levels (Nohmi et al, 1992; 1995a; Gruz et al, 1996) and since samAB deletion actually increases UV mutability in S. typhimurium about two-fold, it has been suggested that samAB might inhibit UV mutagenesis promoted by the chromosomal umuDC^j operon (Nohmiet al, 1995b). However, under conditions of elevated expression, SamA'B promotes chemically-induced frameshifl; mutagenesis by
82 furylfuramide, aflatoxin Bl, 1-nitropyrene, and 1,8-dinitropyrene with efficiencies comparable to, or even better than, MucA'B (Gruz et al, 1998).
The impCAB operon, like mucAB, produces a spectrum of UV-induced base substitutions similar to that produced by umuDC, the majority of mutations being AT -
GC transitions, with a lower level of GC - AT transitions (Doyle and Strike, 1995).
However, after methylmethanesulfonate (MMS) treatment, theumuDC operon produces almost exclusively AT - GC transitions, whereas impCAB mediates a mutagenic spectrum dominated by transversions, primarily AT^TA. However, the total number of mutations following MMS treatment was not greatly increased byimpCAB (or mucAB), compared to umuDC (Doyle and Strike, 1995).
5. Spontaneous mutation and natural selection
In 1859, Charles Darwin revolutionised the debate on evolution. Before the publication ofOn the Origin o f Species (Darwin, 1859), it was popularly held that species were immutable, and that each was an individual creation. Darwin proposed that species are not immutable, but that variation occurs within species. Some of these variations may confer upon their host an advantage in the current environment. Natural selection then allows the survival of those organisms best suited to the prevailing conditions. Darwin proposed that such variants occur rarely and spontaneously, without regard for their immediate usefulness. However, as better adapted variants gain dominance, less useful variants, and the original parent species, diminish in number and eventually become extinct. Eventually a new variant will become sufficiently diverse from its parent to be regarded as a separate species.
83 The belief that variants occur spontaneously without due regard to their immediate usefulness was generally held true for over a century, strengthened by the demonstration by Luria and Delbrück (1943) that phage-resistant mutants were present in cultures of& coli prior to their selection by phage exposure. Lederberg and Lederberg
(1952), using the technique of replica plating, confirmed that clones of phage- or streptomycin-resistant E. coli mutants existed on control master plates, which had not been exposed to the selective agents.
Although these classic experiments proved that bacterial mutants occur spontaneously in the absence of selection pressure, they did not rule out the possibility that other mutations may be non-random and directed by prevailing conditions. For example, Ryan and Wainwright (1954) observed that His^ revertants continued to arise for a period often days after cultures of histidine-requiring (His) E. coli were inoculated into medium without histidine. Subsequent control experiments showed that the delay in appearance of the His^ mutants did not result from their slow growth, or from phenotypic lag or cell turnover (Ryan, 1955; 1959; Ryanet al, 1963). This led Ryan to conclude that a small amount of DNA must be synthesized, but via some process that increased the error rate (Ryan et al, 1961). However, Ryan did not investigate whether neutral mutations were also being produced during stationary phase, nor if late-arising His^ revertants occurred in the absence of selection.
5.1. The origin of adaptive mutation
In 1988 John Cairns initiated a revolution in evolutionary theory (Cairnset al,
1988). He used a strain of E. coli in which the lactose operon of the bacterial
84 chromosome had been deleted and which harboured an F' plasmid containing the lactose operon but with an amber mutation in lacZ {E. coli bdac F ' ta c t lacZ^^ lacY^). The strain could not utilise lactose unless the amber mutation in the plasmid-bome lacZ gene was reverted. A derivative bearing a deletion in the uvrB-bio region of the chromosome, which abolished nucleotide excision repair, was also used. When overnight cultures of both strains were grown in M9 medium plus 0.05% yeast extract and then plated on M9- minimal plates containing lactose as sole carbon source, both strains exhibited a similar rate of reversion to Lac^, as calculated from the number of visible colonies after 48 hours incubation. On continued incubation, the uvrB strain produced approximately 3 to 5 times as many late-arising Lac^ mutants compared to the uvF strain. Cairns et al (1988) hypothesized that the late-arising mutant Lac^ colonies resulted from late events that occurred on the plates in the presence of lactose when the bacteria were under strong Lac^ selection pressure.
Since the uvrB strain produced a greater number of Lac^ revertants, it was used to examine the role of lactose in the development of late-arising Lac^ mutants. Aliquots were plated onto M9 plates without lactose. Immediate overlay of these plates with top agar containing lactose led to the production of an average of 20 Lac^ colonies per plate after 48 hours. On continued incubation, the lactose-overlayed plates accumulated an additional 8 colonies per day. If the addition of lactose to the plates was delayed by one or three days, the appearance of the Lac^ mutants was delayed by the same period of time.
In the absence of lactose, no Lac^ colonies were observed. This led Cairnset al (1988) to propose that the accumulation of late Lac^ mutants occurred only in the presence of lactose.
As a control to determine if exposure to lactose was specific to increasing the
85 number of Lac^ mutants, Cairns et al (1988) examined the development of unselected valine-resistant mutants in cultures producing late Lac^ revertants. E. coli K-12 strains are sensitive to exogenous valine because of feedback inhibition of acetohydroxy acid synthase enzymes by valine, which creates an artificial auxotrophy for isoleucine
(Umbarger and Brown, 1955; Guardiola et al, 1974). Such valine-induced growth inhibition is overcome by mutations at several loci that confer a valine resistance (VaP) phenotype. Cairns et al (1988) overlayed their lactose plates with agar containing glucose and valine, but could demonstrate no increase in the number of VaP colonies. Therefore, exposure to lactose specifically increases Lac^ reversion: unselected mutations, such as
Val^ do not accumulate. These experiments led the authors to suggest that “populations of bacteria, in stationary phase, have some way of producing, or selectively retaining, only the most appropriate mutations” (Cairnset al, 1988).
Since 1988 and the publication of “The Origin of Mutants” (Cairns et al, 1988), the phenomenon of directed (Cairns et al, 1988), “Caimsian” (Hall, 1990) or adaptive mutagenesis (Hall, 1991) has been reported in many other genes ofE. coli, including the bgl operon (Hall, 1988; 1994),metBl (Hall, 1989), trpA and trpB (Hall, 1990; 1991;
1993; 1995a), cysB (Hall, 1990), tyrA (Bridges, 1994), ebgR and ebgA (Hall, 1995b), mutation to VaP (Jayaraman, 1992) and mutation to nitrofurantoin resistance inVibrio cholerae (Basak et al, 1992). Adaptive mutation has also been demonstrated in the eukaryotic yeast, Saccharomyces cerevisiae (Hall, 1992; Steele and Jinks-Robertson,
1992). Adaptive double mutations have also been shown to occur at higher frequencies than would be predicted if the two mutations were the result of independent events (Hall,
1988; 1991; 1993; 1995b).
86 Although such mutation has been termed adaptive, since it is detected only in genes whose function is selected (Cairns et al, 1988; Hall, 1988; 1989; 1992; 1995a,
Foster and Caims, 1992, Jayaraman, 1992; Steele and Jinks-Robertson, 1992; Bridges,
1994), Hall (1990) observed that approximately 1% of selection-induced Trp^ revertants carried additional mutations for unidentified auxotrophies.
5.2. The Lac frameshift assay
Much of the work that revealed the mysteries of adaptive mutation utilised the
Lac frameshift assay (Caims and Foster, 1991 ; Foster and Caims, 1992; Escarcelleret al,
1994; Rosenberg et al, 1994; Foster and Trimarchi, 1994; 1995; Harrisetal, 1994; 1996;
Galitski and Roth, 1995; Longerich et al, 1995; Radicella et al, 1995; reviewed by
Rosenberg et al, 1996). E. coli strain FC40 is phenotypically Lac due to a chromosomal deletion eliminating the lac operon. Into this background, an F' plasmid has been introduced, which has la d fused to lacZ, This fusion eliminates the coding sequence for the last four residues of lad, all of lacP and lacO, and the first 23 residues of lacZ
(Müller-Hill and Kania, 1974). The fusion protein is transcribed from the mutant lacP promoter, but a frameshift mutation of CCC to CCCC at position 1039 (Calos and Miller,
1981) creates a stop codon at base pair 1359 making the cells Lac (Figure 2). Restoration of the reading frame by deletion or addition anywhere between the two out-of-frame stop codons (boxed in Figure 2) produces reversion to Lac^.
Caims and Foster (1991) observed that liquid cultures of FC40, grown to saturation and plated onto minimal-lactose medium, produced increasing numbers of Lac^ revertants for a period of up to two weeks, until the plate was covered in colonies. During the first five days, the number of Lac^ colonies increased almost linearly from near zero
87 926 ATATCCCGCCGTTAACCACCATCAAACAGGATTTTCGCCTGCT la d
GGGGCAAACCAGCGTGGACCGCTTGCTGCAACTCTCTCAGGG la d
1019 1039* CCAGGCGG TGA AGGGCAATCAGCTGTTGCCCCGTCTCACTG la d
1059 1064 1071 1075 gtgaaaagaaaaaccaccct GG c g CCCaatacgcaaacc la d
1096 1109 g c c t c tCCCCg c g c g t t g gCCgattcattaatgcagctggc la d
1146 1 3 5 9______ACGACAGGTTTCCCGA[A213 bp fusion lad::Z\C^A^CGCCTI ladr.lacZ
1414 GCAGCACATCCCCCTTTCGCCAGCTGGCGTAATAGCGAAGAGGC lacZ
Figure 2. DNA sequence (Rosenberg et al, 1994; Longerichet al, 1995) of thelad-lacZ fusion used in the Lac frameshift assay (Caims and Foster, 1991). la d is fused to lacZ, such that the frameshift mutation of CCC to CCCC at position 1039 inla d (*) is polar on lacZ, making the cells Lac'. Restoration of the reading frame by deletion or addition anywhere between the two out-of-frame stop codons (boxed) reverts the phenotype to Lac^. Underlined bases indicate the sites where reversion mutations have been found to occur in adaptive mutants. The additional cytosine that causes the +1 frameshift is not numbered. to more than 40 per plate. This suggested that the revertants arose after plating as the result of some process that was time-dependent but not growth-dependent. The mutant clones that arose later after plating could not be attributed to growth on the plates, since the population of FC40 cells underwent less than one doubling in the first five days after plating on minimal-lactose medium. Nor could it be attributed to cross-feeding from the first colonies to be formed on the plates, since an excess of non-revertible Lac scavenger cells were also added to the plates. These scavengers remove all contaminating sources of energy from the medium, which might allow growth of Lac' FC40 cells prior to mutation to Lac^. Controls in which cells were plated in the absence of lactose, showed that Lac^ revertants were not produced until plates were overlaid with agar containing lactose (Caims and Foster, 1991). Using FC82 (Lac Trp ), Caims and Foster (1991) also demonstrated that reversion to Lac^ did not occur in the presence of lactose when there was an additional growth requirement absent from the medium, in this case tryptophan.
The addition of tryptophan to minimal lactose plates spread with FC82, but lacking tryptophan, led to the production of Lac^ revertants with the accumulation of almost 100
Lac^ colonies per plate, four days after tryptophan addition. These results led Caims and
Foster (1991) to suggest that mutations that occur during stationary phase are preserved only if they quickly lead to the resumption of growth.
A 276-base pair region of plasmid DNA, which spanned the frameshift mutation, was sequenced from Lac^ isolates. All contained single base deletions in regions of small mononucleotide repeats (underlined in Figure 2) (Rosenberg et al, 1994; Foster and
Trimarchi, 1994). Spontaneous growth-dependent reversion mutations of the samelac frameshift mutation were shown to be much more heterogenous, including duplications, insertions and deletions. This suggested that growth-dependent mutations could be produced by multiple mechanisms, whereas adaptive mutation resulted from a unique mechanism (Rosenberg et al, 1994; Foster and Trimarchi, 1994).
Whilst a high level of adaptive reversion to Lac^ occurs inE. coli FC40 when the
Lac" allele is on an F' plasmid and conjugal functions are expressed (Foster and
Trimarchi, 1995; Galitski and Roth, 1995; Radicella et al, 1995), mutations in a non selected chromosomal gene, rpoB, do not also appear in the Lac' population during selection (Foster, 1994). However, it was not known if nonselected mutations also accumulate on the plasmid. To investigate this, Foster (1997) used a derivative of FC40 carrying a mutation of the plasmidic TnlO element, which reverts the strain to
89 tetracycline sensitivity, to examine the accumulation of tetracycline-resistant (Tet*^) mutants during lactose selection. An increase in the number of Tet^ mutants was observed in the Lac' population during lactose selection, even though resistance to tetracycline should confer no growth or survival advantage (Foster, 1997). In addition, the Lac Tet^ mutants that arose during lactose selection appeared to occur via the same mechanism as the Lac^ mutations of FC40. The conjugal and recombination functions required for efficient “adaptive” reversion to Lac^ in FC40 (Caims and Foster, 1991;
Foster and Trimarchi, 1995; Fosteret al, 1996; Galitski and Roth, 1995; Harris et al,
1994; 1996) were also required for the appearance of Lac' Tet^ mutants of FC722 during lactose selection, suggesting that the two mutations do not arise independently (Foster,
1997). However, that unselected mutations, from tetracycline sensitivity to tetracycline resistance, occur on the plasmid during lactose selection in the Lac fi*ameshift assay does not rule out the possibility that other cases of selection-induced mutation are indeed adaptive (Foster, 1997).
5.3. Adaptive mutation via the excision of mobile genetic elements
Adaptive reversion to Lac^ in the Lac frameshift assay results from single base deletions in regions of small mononucleotide repeats (Rosenberget al, 1994; Foster and
Trimarchi, 1994) (Figure 2). However, other adaptive mutants may involve base substitutions, positive and negative frameshifts, and even excision of mobile elements from within coding regions of DNA (Hall, 1988; 1994).
90 5.3.1. The Shapiro deletion
Shapiro (1984) used E. coli strain MCS2 to study the excision of Mu bacteriophage in response to selection.E. coli MCS2 is both Lac’ and Ara’ due to Mu integration between araB (the positive regulatory component controlling the arabinose operon) and the carboxy-terminal region of thelacZ gene (coding for p-galactosidase).
The promoter forlacZ transcription is absent in MCS2, which also contains a U118ochre triplet at codon 17 of lacZ (Figure 3). Transcription oflacZ can therefore, only occur following fusion to an upstream promoter sequence and removal or reversion of U 118 ochre. Shapiro (1984) showed that appropriate Mu excision events could remove all blocks to transcription between araB and the region that contains the sequence for the catalytically significant domain of P-galactosidase, downstream of codon 17 inlacZ.
Such excision resulted in a fusion strain of MCS2 that was phenotypically Ara-Lac'; i.e, could utilise lactose only if arabinose was present as inducer. Mu excision is often
U 118 ochre araB Mu at codon 17 Pre-fusion strain
Fusion strain (shown I with short Mu linker of <30 bp)
Figure 3. Schematic configuration of thearaB-lacZ regions ofE. coli MCS2 before and after fusion formation.
91 incomplete, leaving oligonucleotide segments derived from Mu termini as linkers between the amino terminal domain of the hybrid cistron and the lacZ sequence (Figure
3; Shapiro and Leach, 1990).
After plating cultures of MCS2 on selective arabinose-lactose medium, Shapiro
(1984) observed a delay of between four and nineteen days before the first Ara-Lac^ colony appeared. The kinetics of colony appearance then showed a rapid increase over the next two to four weeks, followed by a decline (Shapiro, 1984). Production of Ara-
Lac^ mutants by deletion of Mu did not occur spontaneously during prior growth in unselective medium, but apparently arose after plating. This was confirmed when it was shown that Mu deletion revertants appeared much later than control colonies which grew when plates were seeded with a few pre-existing Ara-Lac^ cells (Shapiro, 1984).
Caims et al (1988) confirmed the findings of Shapiro (1984), concluding that; (1) large numbers of Ara-Lac^ cells start to appear after three to four days in rich media containing both lactose and arabinose, (2) Ara-Lac^ cells do not accumulate at a detectable rate in rich media in the absence of lactose (or in the absence of arabinose), and (3) when lactose is added to a stationary culture supplemented with arabinose but not lactose, Ara-Lac^ cells promptly start to accumulate. Caimset al (1988) therefore suggested that excision of Mu from between araB and lacZ was another mechanism for the generation of mutations in response to selection.
Claims that Mu excision could generate selection-induced mutations (Shapiro,
1984; Caims et al, 1988) were examined by Mittler and Lenski ( 1990). When populations of MCS2 were starved for nine days in glucose-limited minimal medium, the frequency of Ara-Lac^ colony-forming cells increased to approximately 10'^ (Mittler and Lenski,
1990), as compared to a frequency of Lac(Ara)^ cells of approximately 10 '® in cultures
92 of MCS2 grown under optimal conditions in fresh medium (Shapiro, 1984; Caimset al,
1988). Mittler and Lenski (1990) concluded that Mu excision was accelerated in response to the non-specific physiological stress of starvation, rather than being selectively favoured by the presence of lactose and arabinose in the medium.
Foster (1993) criticised the conclusions of Mittler and Lenski (1990) because the existence of Ara-Lac^ mutants was not confirmed prior to plating on the lactose- arabinose selective medium. However, Foster and Caims (1994), using fluctuation tests,
Maenhaut-Michel and Shapiro (1994), using sib-selection, and Sniegowski (1995), using replica-plating, eventually confirmed that Mu deletions arise in stationary phase under the stress of starvation, in the absence of selection for the Ara-Lac^ phenotype. It is, therefore, the stress associated with starvation, either in the presence or absence of lactose and arabinose, that is responsible for the development of Ara-Lac^ mutants. However, the spectmm of frisions found after direct selection with lactose and arabinose differs from that produced by non-specific starvation. Sequence analysis revealed the former mostly yielded a single class of “standard” fusions, while indirect sib-selection led to a broader spectmm of fusion stmctures (Maenhaut-Michel and Shapiro, 1994).
5.3.2. Excision from the bgl operon
E. coli is unable to utilise salicin because the bgl (p-glucosidase) operon is cryptic, requiring a mutation within a short region designated bglR to activate the operon and allow its expression (Prasad and Schaefier, 1974). The wild-type allele of bglR
{bglR°) can be converted to its active bglR^ state, either via point mutations (Reynolds et al, 1986; Parker and Hall, 1988) or by insertion of the mobile genetic elements ISl or
IS5 into the bglR site (Reynolds et al, 1981; 1986; Schnetz et al, 1987). This allows
93 utilisation of the p-glucoside sugars, salicin and arbutin.
E. coli strain %342LD harbours a 1.4 kb foreign DNA fragment (IS 103) inserted into, and inactivating, the bglF gene (Parker et al, 1988). The latter specifies a p- glucoside transport protein that both transports and phosphorylates its substrates (Prasad and Schaefier, 1974). Strain %342LD, therefore, requires both excision of IS 103 from within bglF and activation of bglR to allow growth on salicin, and Hall (1988) demonstrated that 60% of %342LD colonies generated Sar papillae in less than two weeks growth on MacConkey medium containing salicin as sole carbon source. Analysis of these SaP mutants indicated that their ability to grow on MacConkey salicin (MS) medium resulted from mutations within the bgl operon producingbglR^ bglF^ double mutants. The frequency at which the SaP double mutants appeared was about twelve orders of magnitude higher than expected on the basis of the observed single mutation rates. Excision of IS 103 from withinbglF did not occur in the absence of salicin (Hall,
1988). Hall (1988) confirmed that the conditions prevailing in aged colonies on MS plates were specific to the production of mutations within thebgl operon because the rate of mutation to valine resistance remained unaffected by conditions on the MS plates.
Using the same strain of E. coli, %342LD, Mittler and Lenski (1992) also investigated the excision of IS 103 (now called IS 150) from withinbglF. After confirming the findings of Hall (1988), that SaP double mutants were produced at a high frequency on MS agar, they showed that single excision mutants could utilise the salicin in MS agar, albeit producing papillae later than SaP double mutants. These results contradicted Hall
(1988), who claimed that single excision mutants cannot utilise salicin. Mittler and
Lenski (1992) also contradicted Hall (1988) by showing that single excision mutations
94 frequently occur in starved populations not exposed to salicin.E. coli %342LD was inoculated into L-broth in the absence of salicin and incubated in this medium for 15 days, without any additional nutrients. After 15 days, they were grown overnight in fresh
L-broth, still without salicin, before plating on MS agar. Control populations, grown overnight in fresh medium from single-colony isolates were also plated on MS agar. The median time for first papilla appearance was three days for the starved populations and ten days for the controls. Mittler and Lenski (1992) concluded, in contrast to Hall (1988), that excision-like mutations from withinbglF do occur spontaneously in the absence of salicin and that many of these intermediate mutants can grow on MS agar, but at a slower rate than SaL double mutants. Since the papillae produced by mutant intermediates on
MS agar contain millions of cells, an increased frequency of SaL double mutants may be expected, via a second mutation in bglR (Mittler and Lenski, 1992).
Hall (1994) dismissed the conclusions made by Mittler and Lenski (1992) by contending that IS 150 excision was indeed consistent with other selection-induced mutations, in that it arose only when of immediate, rather than potential, advantage. Hall
( 1994) failed to detect any SaL excision mutants during prolonged incubation on non- selective medium in the absence of salicin, and demonstrated that excision mutations did not occur at a detectable level in old saturated L-broth cultures, as was claimed by Mittler and Lenski (1992). His results did not rule out the possibility that IS 150 excision could occur in the absence of salicin, but by showing that excisions occur much more frequently in the presence of salicin than in its absence, he demonstrated that such excision events can truly be called “selection-induced” (Hall, 1994).
95 5.4. Prerequisites for the generation of adaptive mutants
A number of requirements need to be fulfilled for adaptive mutation to take place.
This suggests that adaptive mutation differs from spontaneous mutation, which occurs in all growing cultures. Adaptive mutations only arise after exposure to a nonlethal genetic selection (Caims et al, 1988; Hall, 1988; 1989; Foster and Caims, 1992;
Jayaraman, 1992; Steele and Jinks-Robertson, 1992; reviewed by Foster, 1993). For example, in experiments examining the reversion of a lac frameshift mutation in E. coli
(Caims and Foster, 1991; Foster and Caims, 1992; Escarcelleret al, 1994; Rosenberg et al, 1994; Foster and Trimarchi, 1994; 1995; Galitski and Roth, 1995; Longerichet al,
1995; Radicella 1995; Harris er a/, 1994; 1996; reviewed by Rosenberg er a/, 1996), cultures were plated in the presence of lactose as sole carbon source. Here, the selective agent, lactose, is not lethal to the cells. This allows the organisms, defective for lactose utilisation, to survive on the plates in a non-growing state. Only if a mutation occurs at the correct position, conferring a Lac^ phenotype, can the organisms exit stationary phase and grow using lactose as carbon source. In their classic experiments, Luria and Delbrück
(1943) and Lederberg and Lederberg (1952) used lethal selection, and therefore could not detect the appearance of adaptive mutants. Only spontaneous mutants, already present in the bacterial culture, could be detected, since those organisms not resistant to phage
(Luria and Delbrück, 1943; Lederberg and Lederberg, 1952) or streptomycin (Lederberg and Lederberg, 1952) were immediately killed.
Adaptive mutations occur in the apparent absence of cell division (Foster, 1994).
In the Lac frameshift assay, cultures of E. coli FC40 plated onto minimal medium containing lactose as sole carbon source do not proliferate, as determined by viable count,
96 showing that no net cell growth or death occurs in the plated cultures during non-lethal selection (Caims and Foster, 1991; Foster, 1994). Foster (1993) examined mutation rates at a number of targets in E. coli, including the lacIv.lacZ frameshift, (Caims and Foster,
1991), lacZ amber (Caims et al, 1988), trpE frameshift (Caims and Foster, 1991), lacZ missense (Hall, 1991), trpA missense and trpB missense (Hall, 1990), his' (Ryan, 1955), and the lysl frameshift mutation in Saccharomyces cerevisiae (Steele and Jinks-
Robertson, 1992), both before and during selection. She concluded that the cell growth required to generate the observed level of mutation by spontaneous events was so great that the mutations could not be attributed to gross population increases (Foster, 1993).
The possibility that cryptic growth (the tumover of some of the population at the expense of the majority) was responsible for the observed effects, could not be mled out as easily.
Using penicillin, which is only bactericidal to dividing cells, Ryan (1959), Ryan et al
(1963) and Hall (1990) could demonstrate no increase in the rate at which organisms were killed by penicillin. Since viable cell populations remained essentially constant, this implied that cell division was not occurring. In addition. Hall (1990) estimated the maximum possible amount of cell tumover and found it to be insufficient to account for the number of Trp^ mutants that appeared.
5.5. Genetic requirements for adaptive mutation
5.5.1. recA
Adaptive mutagenesis requires the presence or absence of certain gene functions.
For example, genes of the RecA-RecBCD system are required for one class of adaptive mutation, suggesting a role for recombination in the formation of such mutations
(Galitski and Roth, 1995; Longerich et al, 1995). In contrast, there is no requirement for
97 these gene products during spontaneous mutation in growing cells (Caims and Foster,
1991; Foster and Trimarchi, 1995; Harris 2
(Foster and Trimarchi, 1994; Hall, 1992,1993). The only system in whichrecA, and recB and recC, are clearly implicated is the lac frameshift assay of FC40 (Caims and Foster,
1991; Harris et al, 1994; reviewed by Foster 1993; Rosenberg, 1994; Rosenberg et al,
1996). However, when the same lacBS allele is on the chromosome and not on the F ' plasmid, adaptive mutation to Lac^ occurs at a much lower frequency, and is not dependent on recA (reviewed by Rosenberg et al, 1996; Foster, 1998). Strains harbouring null mutations in recD, which functions as a negative effector of the recombination activity of the RecBCD enzyme (see Introduction section 3.4.3), are hyper- recombinogenic and adaptively hypermutable (Harris et al, 1994). Loading of the
RecBCD enzyme onto DNA requires double-strand breaks, implying an involvement of these molecular intermediates in the generation of adaptive mutants in the Lac frameshift assay (Rosenberg, 1994).
The fact that adaptive mutation has also been demonstrated in the absence of functional RecA (Basak etal, 1992, Bridges, 1993; Hall, 1995a), suggests that more than one mechanism may operate (Foster, 1993; Foster and Trimarchi, 1995). Recombination is required for selection-induced frameshifting, as in the E. coli lac frameshift model
(Caims and Foster, 1991; Harris et al, 1994), but not in the generation of selection- induced base substitutions, which occur in the trpA46 allele of E. coli K-12 in the absence of RecA function (Hall, 1995a).
98 5.5.2. lexA and umu genes
Caims and Foster (1991), examining the lac frameshift reversion assay inE. coli, demonstrated that such adaptive mutation requires at least one of the many genes controlled directly or indirectly by LexA, such as recA or the umuDC operon. UmuDC functions are not involved in the generation of adaptive Lac^ fi*ameshift revertants, since the reversion rate is unaffected by a mutant umuC allele, or by mutants ofrecA that are deficient in UmuD, but not LexA, processing (Caims and Foster, 1991 ; see Introduction section 3.5.5). In addition, when the lexAlOl allele, deficient for induction of the SOS response, or theumuC122 allele, which specifically blocks the error-prone SOS pathway, were present in E. coli WP2, the rate of mutation to tryptophan independence in tryptophan-starved cultures was not affected (Bridges, 1993).
SOS induction has been reported in stationary phase bacteria, as indicated by
RecA*-mediated cleavage of the Cl repressor of phage X in resting bacterial populations.
This led to the proposal that both Cl cleavage and adaptive mutation to Lac^ inE. coli strain FC40 depend on high levels of RecA*. The latter is required at least to induce recA for expression of adaptive mutation (Taddei et al, 1995). A role for the SOS response in adaptive mutagenesis is suggested by the similar mutational spectra produced after adaptive mutation of FC40 or when the SOS response is induced in the absence of lesions. In both cases, the relative number of frameshift mutations at mononucleotide repeats is enhanced (Yatagai et al, 1991; Foster and Trimarchi, 1994; Rosenberg et al,
1994). Basak et al (1992) concluded that, except for the recombination functions ofrecA in the lac frameshift assay, induced expression of SOS genes and error-prone repair are unnecessary for adaptive mutation. However, when the SOS regulon is constitutively expressed due to the presence of recA441 and lexA (Def) alleles in E. coli WP2, an
99 increased frequency of selection-induced mutation to tryptophan independence is observed (Bridges, 1993). Therefore, while SOS processing and error-prone repair appear unnecessary to initiate adaptive mutation, such processes may contribute to an increased frequency of adaptive mutagenesis when it is expressed (reviewed by Bridges, 1998).
5.5.3. Mismatch DNA repair
Strains deficient for mutS show an increase in the frequency of adaptive mutation, but a decrease in cell viability during stationary phase (Boe, 1990; Foster and Cairns,
1992; Reddy and Gowrishanker, 1997). These results suggest that DNA synthesis is occurring in stationary phase cells (since methyl-directed mismatch repair (MMR) normally corrects errors only in newly synthesised DNA), and that mismatch repair activity is required to prevent some lethal outcome (reviewed by Foster, 1993). Indeed, the mutations that give rise to adaptive Lac^ revertants of FC40 are produced by DNA polymerase III (Fosteret al, 1995), and consist mainly of -1 base pair frameshifis in runs of iterated bases (Foster and Trimarchi, 1994; Rosenberg et al, 1994). Such polymerase errors are typically corrected by MMR, which led to the hypothesis that MMR might be limiting in stationary phase and nutritionally-deprived cells, leading to the production of adaptive mutations (Foster and Trimarchi, 1994; Rosenberg et al, 1994). Reports that levels of the MMR proteins, MutS, MutH, and MutL, appear to fall during stationary phase and in nutritionally-deprived cells (Feng et al, 1996; Harris et al, 1997a) supported this hypothesis. Also, cells deficient in mutL, mutS, ordam genes, mutations that disable
MMR, produce a spectrum of mutation during exponential growth, which is indistinguishable from that found during adaptive mutation (Longerich et al, 1995).
However, more recent work has demonstrated that a loss of MutL results in an
100 approximately 100-foId increase in the growth-dependent frequency of mutation to Lac^ in E. coli FC722, a close derivative of FC40. A similar 100-fold increase in the frequency of adaptive mutation to Lac^ is also observed in the absence of MutL, suggesting that
MMR is no less effective in correcting errors during lactose selection than during exponential growth (Foster, 1999). In addition, overexpression of MutL or MutS reduces adaptive mutation approximately two-fold, and combined overexpression of these two proteins produces a six-fold decrease in adaptive mutation to Lac^ in FC722 (Foster,
1999). Overproduction of MutL and MutS, either separately or together, also reduces the number of growth-dependent Lac^ mutations, demonstrating that MMR activity can be improved in both growing cells and nutritionally-deprived cells (Foster, 1999). Such results suggest that even if the levels of MutS and MutH decline in stationary phase cells
(Feng et al, 1996), the remaining MMR activity is sufficient for the amount of DNA synthesis that occurs (Foster, 1999).
Jayaraman (1992) proposed that there could be two pathways of adaptive mutagenesis in E. coli, one under mutS, and the other under mutL control. These genes negatively regulate the pathways that produce adaptive mutants. The mutS controlled pathway is dependant on LexA cleavage, but only to produce elevated levels of RecA, as derepression of other SOS genes does not seem to be required (Cairns and Foster,
1991; Jayaraman, 1992). However, the expression of adaptive mutations inmutL strains ofE. coli is independent of recA (Jayaraman, 1992). Hall (1995a), after discovering that a recA deletion did not reduce selection-induced reversion of trpA46 also hinted that adaptive mutations may occur via more than one mechanism: each different class of lesion that initiates selection-induced adaptive mutation could be repaired via a different molecular mechanism.
101 5.5.4. Nucleotide excision repair
Elimination of nucleotide excision repair (see Introduction section 3.2) by mutations in uvrA or uvrB, either has no effect, or increases the frequency of adaptive mutation (Cairns et al, 1988; Foster, 1992; 1993; Foster and Cairns, 1992; Bridges,
1993). The UvrABC pathway is, therefore, not required to produce adaptive mutations, but it may prevent some classes of mutations from occurring. Hall (1995a) observed that lesions in uvrA, uvrB, uvrC, and uvrD increased the selection-induced reversion of trpA46 to trpA^ up to 300-fold during prolonged tryptophan starvation. The spontaneous reversion of cells growing in the presence of tryptophan was unaffected by such defects in nucleotide excision repair. This suggests that during prolonged selection, some form of DNA damage occurs, which is not normally found in cells growing under non- selective conditions (Hall, 1995a).
However, in contrast to the findings of Hall (1995a), Bridges and Timms (1997) observed a comparable frequency of adaptive mutation to tryptophan independence in trpA23 bacteria deficient in nucleotide excision repair, and in trpA23 cells proficient for this repair process. Therefore, the specific function of nucleotide excision repair in selection-induced mutation still remains to be determined.
5.5.5. RecG and RuvABC
In E. coli, two systems operate at Holliday junctions to resolve strand exchange intermediates of conjugational recombination (reviewed by West, 1994; Introduction section 3.4.5). RecG resolution (Lloyd, 1991; Lloyd and Buckman, 1991; Lloyd and
Sharpies, 1993a; 1993b) inhibits Lac^ frameshifr adaptive mutation, whereas the
RuvABC system is necessary for Lac^ adaptive reversion, even in the presence of
102 functional RecG (Connollyet al, 1991; Benson et al, 1991; Dunderdale et al, 1991;
Iwasaki et al, 1991; 1992; Parsonset al, 1992; Tsaneva et al, 1992a; 1992b; Parsons and
West, 1993). This suggests that strand exchange intermediates may also be intermediates
in Lac^ adaptive mutation. Such promotion has been suggested to result from DNA
synthesis at exposed 3' ends of strand exchange intermediates, during which polymerase
errors may occur (Harriset al, 1996; Foster et al, 1996). The opposite effects of the
branch migration enzymes on adaptive mutation are accomodated by assuming that RecG
and RuvAB promote migration of the Holliday junction toward and away from the
replication fork respectively (Foster, 1998).
5.5.6. Plasmid transfer functions
The reverting lacBSv.lacZ allele, which forms the basis of the E. coli Lac
frameshifr assay, is carried on an F ' plasmid. Although conjugation is not required and does not take place during adaptive reversion (Foster and Trimarchi, 1995; Galitski and
Roth, 1995; Radicella et al, 1995), several authors have suggested that aspects of F'
transfer are required, most likely the conjugal form of DNA synthesis (Peters and Benson,
1995, Radicella et al, 1995; Galitski and Roth, 1995; Foster and Trimarchi, 1995). This is demonstrated by the fact that mutation in either traD or traQ reduces Lac^ reversion to 1/10*’’ that of the wild-type FC40 strain (Foster and Trimarchi, 1995; Radicella et al,
1995). In an F' cell, location of the samelacI33::lacZ allele on the chromosome led to a large reduction in the post-plating reversion rate to Lac^, being approximately 1/100* that of cells with the allele on the F' plasmid (Foster and Trimarchi, 1995; Galitski and Roth,
1995; Radicella et al, 1995). Provision of conjugal functions, by the presence of a Tra^
F plasmid or an F ' carrying heterologous genes, to strains harbouring the chromosomal
103 lacI33::lacZ allele did not enhance adaptive Lac^ reversion (Foster and Trimarchi, 1995), demonstrating the importance of the plasmid location of thelacI33::lacZ allele and the expression of conjugal functions for adaptive reversion.
5.5.7. DNA polymerases
The rate at which mutations arise in resting cells appears to be much greater than that expected if the fidelity of their residual DNA synthesis is as high as it is in growing cells (Ryan et al, 1961 ; Foster, 1993). Boe (1990) reported 5% turnover per genome per day in cells entering (glucose-limited) stationary phase, a level which rapidly declined to
0.5%. Foster (1993) concluded that the rate of DNA synthesis in resting cells lies in the range of 0.005 to 0.05 genomes per cell per day. However, the continuation of DNA synthesis under conditions of selection is implied by the observations that defects in mutHLS mismatch correction lead to large increases in the frequency of adaptive mutation (Boe, 1990; Foster and Cairns, 1992; Jayaraman, 1992; Hall, 1995b). It has been suggested that this DNA synthesis may not be detectable when using exogenous labelled triphosphate precursors because the cells use endogenously generated, unlabelled, precursors derived fi*om the breakdown of RNA or DNA (Bridges, 1996).
Examination of selection-induced mutation of a number of auxotrophic alleles, which are ‘leaky’ to different extents, has led to the conclusion that little or no mutation under selection occurs in ‘tight’ alleles whereas ‘leaky’ alleles show selection-induced mutation. The rate of such mutation is not directly related to the degree of ‘leakiness’
(Jayaraman, 1995; Galitski and Roth, 1996). DNA synthesis resulting from ‘leakiness’ may not be essential for selection-induced mutation, but does occur and does enhance such mutation (reviewed by Bridges, 1998).
104 Whereas, during non-selective growth, the rate of mutation to lactose utilisation in E. coli FC40 was unaffected by a Pol II deletion (polBAl), this same mutation led to a three-fold increase in adaptive reversion of the lacI33::lacZ allele (Escarceller et al,
1994). Similarly, a deficiency in PolB proofreading exonuclease activity (polBexl) enhanced reversion oflacI33::lacZ four- to six-fold. This led to the suggestion that the exonuclease-defective Pol II synthesizes DNA in non-dividing cells (Foster et al, 1995).
However, presence of the antimutator allele ofdnaE (dnaE915), which encodes the main subunit of the Pol III holoenzyme, decreases adaptive Lac^ reversion approximately three fold, implying that Pol III DNA synthesis can also result in recombination-dependent adaptive mutations (Foster et al, 1995; Harris et al, 1997b). It therefore remains to be determined which polymerase(s) are involved in the generation of adaptive mutations.
5.6. Proposed mechanisms
A number of mechanisms have been proposed to explain the production of adaptive mutations in bacteria. However, no single proposal accounts for all examples so far described.
5.6.1. Variant RNAs
Cairns et al (1988) suggested that cells under selection are engaged in some reversible process of trial and error. Such cells might produce a set of highly variable mRNA molecules, and that reverse transcription would be stimulated if one of these happened to code for a fimctional protein. This model demanded a direct connection for the reverse flow of information from environment (lactose selection) to protein (P- galactosidase) through mRNA, with the mutation finally immortalised in chromosomal
105 DNA by reverse transcription (Cairns et al, 1988). Such direct flow of information from the environment to DNA is most unlikely (Foster and Cairns, 1994). The enzyme, reverse transcriptase has not been demonstrated in E. coli K-12 strains (Inouye and Inouye,
1991), in which the majority of these experiments were performed (Foster, 1992).
5.6.2. Variant DNAs
Stahl (1988) and Boe (1990) suggested the slow-repair hypothesis to explain adaptive mutation. They proposed that non-dividing bacteria may replicate or repair their
DNA irregularly and that polymerase-catalysed DNA synthesis is many times more error- prone under conditions of selection, than mutation rates observed in exponentially growing cells. In a nutritionally deprived cell, the error-correction enzymes that monitor newly synthesised DNA might be slow to act. This would lead to a higher error frequency in the newly replicated DNA. If one of the resulting mutant genes encoded a useful protein, cell division and replicative DNA synthesis would resume, immortalising the mutation. If a useful protein is not encoded by the mutant gene, then the error is eventually corrected in favour of the template DNA strand. As already stated, exponential cultures of cells defective in mismatch DNA repair produce a spectrum of mutations indistinguishable from those seen during adaptive mutation (Foster and Trimarchi, 1995;
Rosenberg et al, 1996) and during lactose selection, the absence of mismatch and alkylation repair pathways increases Lac^ reversion (Foster and Cairns, 1994). However, this model does not explain why more mutants are not found in repair-deficient stationary phase cells in the absence of selection (Foster and Cairns, 1994), as might be expected
(Foster, 1992).
Davis (1989) suggested that selection conditions could impose a bias on the
106 “random” process of mutation. Transcription would be directed to the selected gene or genes. This, he suggests, would result in regions of single-stranded DNA within the transcribed, gene. Single-stranded DNA is more prone to damage, leading to erroneous replication or to error-prone repair. Therefore a biased rate of mutation would occur in the selected gene. This would lead to a higher probability of a successful variant being produced, allowing the organism to resume growth under the selective conditions. This model does not explain how selection pressure actually directs the mutation to the appropriate gene (Foster, 1992).
5.6.3. The hypermutable state
Another model, proposed by Hall (1990), depends upon the “hypermutable” state.
Under selective, and therefore stressful conditions, most cells in a population do not mutate, but a minority transiently experience a high mutation rate (“hypermutation”). In this hypermutable state, extensive DNA damage and resulting error-prone repair synthesis occurs. This leads to an increased incorporation of errors into DNA. If a mutation overcomes the cause of growth inhibition, then growth resumes under the selective conditions, and the organism exits the hypermutable state (Hall, 1990; Foster, 1992).
A specific prediction of the hypermutable state model is that non-selected mutations occur at a higher frequency among cells that bear adaptive mutations than among cells that do not (Hall, 1990). This prediction has been confirmed in four situations (Hall, 1990; Boe, 1990; Foster, 1997; Torkelsonet al, 1997). Rosche et al
(1999) calculated that approximately 0.06% of a population under selection are hypermutators, and that their mutation rate is about 200-fold higher than that of the rest of the population. However, whilst such hypermutators are responsible for nearly all
107 multiple mutations, only about 10% of adaptive Lac^ mutations occur in hypermutating
cells (Rosche et al, 1999). This led these authors to propose that the mechanism which
generates Lac^ mutations is the same in all the cells, but that a deficiency in MMR allows more to be retained in the hypermutators (Roscheet al, 1999).
Foster and Cairns (1992) suggested the cell might retain a pristine master copy of its DNA, but produce extra copies of genes that contain sequence changes. They proposed that such gene duplication was RecA-dependent and that amplification led to the accumulation of mutations at higher rates than in single-copy sequences. If one of the duplicated sequences contained a useful mutation, the amplified region would be resolved by a RecA-dependant process, which retained the useful DNA copy, allowing growth to resume. In the absence of success, the amplified region would also eventually be resolved, but because most copies retained the original DNA sequence, the cell would remain unmutated (Foster and Cairns, 1992).
6. Aims of this Thesis
No single model encompasses all examples of adaptive mutation that have been observed. The precise molecular mechanism(s) of adaptive mutation remain to be discovered. However, it is reasonable to propose that as a result of the DNA damaging effects of nalidixic acid, error-prone repair might be induced in E. coli cells exposed to non-lethal concentrations of nalidixic acid. Such error-prone repair might then act to increase mutation frequencies within the exposed population, when resulting increases in mutation to nalidixic acid resistance might be considered “adaptive”. Therefore, this thesis examines the development of nalidixic acid resistance in nalidixic acid-exposed
108 cultures, and the selective nature of this repair. It also tests the hypothesis that if selection-induced mutation to nalidixic acid resistance were the result of error-prone repair, then the carriage of a mutator plasmid might be expected to increase the frequency at which cells mutate to nalidixic acid resistance.
109 MATERIALS
1. Bacterial strains and plasmids
The strains ofEscherichia coli used throughout this work are listed in Table 1.
Cultures in use were stored on nutrient agar slopes kept at 4°C. Long term storage of bacterial cultures was as nutrient broth cultures frozen and stored in liquid nitrogen.
Plasmids used are listed in Table 2. All bacterial cultures and plasmids were kindly provided from the collection of Dr. R.J. Finney. Plasmid R46, previously known as R-
Brighton (Drabble and Stocker, 1968), has also been referred to as R1818 and TP 120
(reviewed by Strike and Lodwick, 1987).
110 Table 1. Strains of Escherichia coli used.
Strain Genotype Description Source Reference thr-1 leuB6 argES hisG4 proA2 thi-1 X' galK2 Wild-type Dr. R.J. Pinney Howard-Flanders et al AB1157 lacYl mtl-1 ara-14 xyl-5 repair-proficient (1964) rpsL31 tsx-33 supE44 F AB1886 as AB1157 but uvrA6 Excision repair deficient Dr. R.J. Pinney Howard-Flanders et al (1965) AB2463 as AB 1157 but recA13 Recombination deficient, Dr. R.J. Pinney Howard-Flanders and SOS induction deficient Theriot (1966) Recombination deficient, Howard-Flanders (1968); AB2470 as ABl 157 but recB21 lacks exonuclease V, SOS Dr. R.J. Pinney Gudas and Pardee (1975); induction deficient (by McPartland et al (1980) quinolones)
AB2494 as AB 1157 but lexA3 SOS induction deficient, Dr. R.J. Pinney Howard-Flanders (1968) recombination proficient
thi-1 proA2 hisG4 lacYl TK702 galK2 xyl-5 mtl-1 supE44 SOS repair deficient Dr. R.J. Pinney Kato and Shinoura (1977) umuC36 Table 2. Plasmids used during this work.
Plasmid Incompatibility Resistance Reference Group determinance* R46 IncN Ap Sm Su Tc UV Anderson and Datta (1965) R446b IncM Sm T c UV Pinney (1980) RP4 IncP Ap Km Nm Tc Pinney (1980)
* Key for drug resistance
Ap ampicillin Sm streptomycin
Cm chloramphenicol Sp spectinomycin
Hg mercuric ions Su sulphonamides
Km kanamycin Tc tetracycline
Nm neomycin UV ultra-violet protection
2. Chemicals and biochemicals
The chemicals and biochemicals used throughout this work are listed below, together with the name and address of the manufacturer.
L-Amino acids:
L-arginine Sigma-Aldrich Co. Ltd., Poole, Dorset.
L-histidine BDH Chemicals Ltd., Poole, Dorset.
L-leucine BDH Chemicals Ltd.
L-proline BDH Chemicals Ltd.
L-threonine Sigma-Aldrich Co. Ltd.
112 D-glucose BDH Chemicals Ltd.
Vitamin (thiamine HCl) BDH Chemicals Ltd.
Sodium hydroxide (NaOH) BDH Chemicals Ltd.
Hydrochloric acid (Hcl) BDH Chemicals Ltd.
Di-potassium hydrogen orthophosphate (K2HPO4) BDH Chemicals Ltd.
Potassium dihydrogen orthophosphate (KH2PO4) BDH Chmicals Ltd.
Tri-sodium citrate (Na3C6 Hg0 2 .2 H2 0 ) BDH Chemicals Ltd.
Magnesium sulphate heptahydrate (MgS 0 4 .7 H2 0 ) BDH Chemicals Ltd.
Ammonium sulphate ((NH4)2S0 4 ) BDH Chemicals Ltd.
Glycerol Sigma-Aldrich Co. Ltd.
Bromophenol Blue Sigma-Aldrich Co. Ltd.
Agarose Gibco BRL, Life Technologies, Paisley,
Scotland.
Ethidium bromide Sigma-Aldrich Co. Ltd.
Ethylendiaminetetra-acetic acid (EDTA) Sigma-Aldrich Co. Ltd.
Potassium chloride (KCl) Sigma-Aldrich Co. Ltd.
Trizma base (TRIS-HCl) Sigma-Aldrich Co. Ltd.
Sterile mineral oil Sigma-Aldrich Co. Ltd.
Magnesium chloride (MgCl2) Sigma-Aldrich Co. Ltd.
Taq DNA polymerase (5U/mL) Advanced Biotechnologies Ltd., Epsom,
Surrey.
113 Reaction Buffer IV (lOx) {Taq buffer) Advanced Biotecnologies Ltd.
E. coli gyrA Primer A (48 pmol yuL'^) Oswell DNA Services, University of
Southampton, Southampton.
E. coli gyrA Primer B (38 pmoljAS^) Oswell DNA Services. dATP (100 mmol) Advanced Biotechnologies Ltd. dCTP (100 mmol) Advanced Biotechnologies Ltd. dGTP (100 mmol) Advanced Biotechnologies Ltd. dTTP (100 mmol) Advanced Biotechnologies Ltd. lOObp DNA ladder (lyUg mL'^) Gibco BRL.
Hinfl restriction endonuclease (lOU/y^L) Promega, Madison, USA.
Buffer B (lOx) Promega.
All inorganic chemicals purchased from BDH Ltd. were ‘Analar’ grade.
Chemicals and biochemicals purchased from Sigma-Aldrich Co. Ltd. were Sigma grade.
Stock solutions of L-amino acids were prepared at a concentration of 2.5 mg mL ' in sterile distilled water and adjusted to pH 7.2, using a digital pH meter, with sterile O.IM sodium hydroxide (NaOH) solution. Amino acid solutions were sterilised by immersing in a boiling water bath for 20 minutes. Thiamine (vitamin B,) was dissolved in sterile distilled water at a concentration of 0.5 mg mL ' and sterilised by immersion in a boiling water bath for 20 minutes. Glucose solutions were prepared at a final concentration of
40%'^/y and sterilised by autoclaving at 115°C (10 psi) for 20 minutes. Vitamin, amino acid and glucose stock solutions were stored at 4°C.
114 3. Solutions for Polymerase Chain Reaction (PCR) analysis
20 yuL of a solution containing lUjAS^ Taq polymerase was prepared on the day of use by mixing 13 yuL sterile distilled water, 2 yuL Reaction Buffer IV (lOx) {Taq buffer), 1 yuL MgClj (50 mM), and 4 yuL Taq DNA polymerase solution (5U ywL'').
A 100 y^L stock solution of 4mmol dNTP solution containing 1 mmol of each of dATP, dCTP, dGTP and dTTP was prepared using 76y^L sterile distilled water, 20 yuL lOx Taq buffer and 1 jTL of each 100 mmol solution of dNTP.
Primer A and Primer B used for PCR analysis were diluted on the day of use to a concentration of 10 pmol in sterile distilled water and used immediately.
4. Antibacterials
The antibacterial agents used during this work are listed below together with the manufacturer who supplied them.
Ciprofloxacin Bayer pic., Newbury, Berks.
Cloramphenicol Sigma-Aldrich Co. Ltd.
Kanamycin sulphate Sigma-Aldrich Co. Ltd.
Nalidixic acid Sigma-Aldrich Co. Ltd.
Sodium ampicillin (‘Penbritin’) Beechams Research Labs., Brentford, Essex.
Streptomycin sulphate Glaxo Laboratories Ltd.,
Greenford, Middlesex.
Sulphadiazine sodium May and Baker Ltd., Dagenham,
Essex.
115 Tetracycline hydrochloride (‘Achromycin’) Cynamid of Great Britain Ltd.,
Gosport, Hampshire.
All antibacterials were weighed aseptically into sterile containers and dissolved in sterile distilled water to the required concentration. The exception was nalidixic acid, which was initially dissolved in a small quantity of 0.1 M NaOH (0.02 mL per mg of nalidixic acid) and made up to a final volume with sterile distilled water.
5. Media
5.1. Complex media
5.1.1. Nutrient broth and nutrient agar
Nutrient Broth No. 2 (CM67) was obtained from Oxoid (Unipath Ltd.,
Basingstoke, Hampshire), and dissolved in distilled water according to the manufacturer’s instructions. Volumes of 4.5 mL and 9.9 mL were dispensed into universal glass bottles and sterilised by autoclaving at 115 °C for 20 minutes. Sterilised nutrient broth was stored at room temperature.
Nutrient agar plates were prepared by adding 1.5g Lab M agar No. 1 (The
Microbiological Supply Co., Toddington, Beds) to each 100ml nutrient broth, and sterilised by autoclaving at 115°C for 20 minutes. Bottles of molten nutrient agar were then cooled to 55 °C in a water bath. Approximately 20 mL volumes of molten agar were poured into sterile disposable Petri dishes and allowed to set. Once the plates had set, they were overdried in an inverted position without their lids at 44°C for 45 minutes.
After overdrying, the lids were replaced, and if not used immediately, the plates were stored inverted at 4°C.
116 5.1.2. MacConkey agar
MacConkey agar powder (CM7) was obtained from Oxoid and made up
according to the manufacturers instructions. The agar was sterilised, poured into Petri
dishes, overdried and stored as described for nutrient agar plates.
5.1.3. Iso-Sensitest agar
Iso-Sensitest agar (Oxoid, CM471) plates were prepared by the addition of 3.1 g
Iso-Sensitest agar to each 100ml nutrient broth according to the manufacturers
instructions. The agar was sterilised, poured into Petri dishes, overdried and stored as
described for nutrient agar plates.
Nutrient and Iso-Sensitest agars were used for the preparation of antibacterial-
containing agar for selectivity and sensitivity testing. Drug solutions were added to molten cooled agar at 55°C, before pouring, to give appropriate final concentrations of drug in the medium.
5.2. Minimal salts medium
5.2.1. Davis and Mingioli salts solution
Davis and Mingioli (DM) minimal salts solution was used as the minimal medium throughout these experiments. Double-strength Davis and Mingioli salts solution was prepared as described by Davis and Mingioli (1950), using the following salts:
117 K2HPO4 14.00 g
KH2PO4 6 . 0 0 g
Na3C6H507.2H20 0.94 g
MgS04.7H20 0 . 2 0 g
(NH4>2S0 4 2 . 0 0 g
Each constituent was sequentially dissolved in 1 litre of distilled water. 50 mL volumes of double-strength DM solution were distributed into 150 mL glass bottles.
Single-strength DM solution was prepared using the above formula but in a final volume of 2 litres of distilled water. Single-strength DM solution in volumes of 4.5 mL and 9.9 mL was dispensed into universal glass bottles. DM solution was sterilised by autoclaving at 115°C (10 psi) for 20 minutes. DM salts solution was stored at room temperature.
5.2.2. DM minimal agar
DM minimal agar was prepared by the aseptic addition to 50 mL double-strength
DM solution of 1 mL of each of the required sterile amino acid and vitamin solutions to give final concentrations of 25 yug mL’* and 5yUg mL’* respectively. 0.7 mL of sterile 40% glucose solution was added to give a final concentration of 0.28%. Sterile distilled water was added to give a final volume of 60 mL. This solution was added to 40 mL of sterilised hot molten agar (1.5 g Lab M agar in 40 ml distilled water), mixed thoroughly and poured into five sterile disposable Petri dishes. Once set the plates were overdried at 44°C for 45 minutes.
118 5.2.3. Agar base plates and ‘soft’ agar overlays used to determine UV-induced reversion to amino acid prototrophy
Agar base plates were prepared by aseptically adding 0.7ml sterile 40% glucose solution to 50 mL double-strength DM solution. This solution was added to 50 ml sterile molten agar (1.5 g Lab M in 50 mL distilled water) in a 150 mL glass bottle, mixed thoroughly and poured into five sterile disposable Petri dishes. The plates were allowed to set before being overdried at 44°C for 45 minutes.
‘Soft’ agar overlays were prepared in 100 mL volumes by the addition to 50 mL double-strength DM solution of the following:
Lab M agar 0.6 g
Thiamine 5 mg
Amino acid 100 mg
Sterile distilled water to a final volume of 100 mL
The amino acid under test was omitted fi-om the overlay. The solution was heated at 100°C for 5 minutes to melt the agar. 2 mL volumes of the molten ‘soft’ agar were distributed into universal glass bottles and sterilised by autoclaving at 115°C (10 psi) for
5 minutes. ‘Soft’ agar overlays were stored at room temperature.
119 6. Apparatus
All apparatus used throughout this work are listed below.
Centrifuges:
High speed 18 refrigerated centrifuge MSE Scientific Instruments,
Crawley, West Sussex.
Bench top centrifuge Centaur 2 MSE Scientific Instruments.
Microcentaur centrifuge MSE Scientific Instruments.
Digital pH meter (PTI-15) Data Scientific Ltd.
Low pressure mercury lamp (model 12) Hanovia Lamps Ltd., Slough, Berks.
Blak-Ray UV meter model J-225 Ultraviolet Products Inc., San
Gabriel, California, USA.
Multipoint inoculator Denley Ltd.
Incubators:
Static incubator Baird and Tatlock Ltd., Chadwell
Heath, London.
Orbital incubator A. Gallenkamp and Co. Ltd.,
London.
Cyclogene Dri-Block Thermal Cycler Techne Ltd., Cambridge. and HL-1 Heated Lid
Quick Screening Horizontal Gel International Biotechnologies Inc.,
Electrophoresis Unit (Mini-gel) New Haven, Connecticut, USA.
Bethesda Research Laboratories Life Technologies, Petersburg,
Horizon 20.25 gel tank (Large gel) Florida, USA.
UVP Ultraviolet Transluminator Ultraviolet Products, Upland,
120 Model TM-40E California, USA.
Gel Electrophoreisis Photosystem Model QSP International Biotechnologies Inc.
Polaroid Black and White Print Film type 665 Sigma-Aldrich Co. Ltd.
Sterile 9 cm disposable Petri dishes Bibby Sterilin Ltd.
1.5 mL Eppendorf tubes Fisher Scientific UK, Loughborough,
Leics.
Eppendorf tubes for PCR thermal cycler Elkay Laboratory Products UK Ltd.,
Basingstoke, Hants.
121 METHODS
1. Overnight cultures
Overnight cultures were prepared by inoculating colonies from a nutrient agar
slope, or a nutrient agar plate containing appropriate antibiotics for R plasmid-carrying
strains, into 4.5 mL nutrient broth. The inoculated broth was thoroughly mixed using a
vortex mixer and incubated overnight in a static incubator at 37°C.
2. Viable counts
Bacterial cultures were serially diluted in nutrient broth or DM salts solution. 10 ’
dilutions were prepared by adding 0.5 mL of culture to 4.5 mL of diluent, and 10'^
dilutions were produced by the addition of 0.1 mL of culture to 9.9 mL of diluent. These
diluted bacterial suspensions were mixed well and used to make further dilutions. The viability of cultures was determined by either spreading 0.1 mL volumes of appropriate
dilutions onto overdried nutrient agar plates using a sterile glass spreader, or by using the drop inoculum technique (Miles and Misra, 1938): 20 /uL volumes of each dilution were dropped in triplicate using a micropipette (Gilson Ltd.) onto prewarmed nutrient agar plates, scored into segments. The drop method allows triplicate counts over a range of dilutions to be made on a single plate. Once the inocula had soaked into the agar, the plates were inverted and incubated at 37°C.
3. Plasmid Transfers
Prior to conjugation experiments, donor strains were screened for the presence of plasmid-encoded drug-resistance phenotypes by streaking cultures onto appropriate drug- containing media.
122 Cultures of donor and recipient were grown overnight in 4.5 mL of nutrient broth at 37°C. 0.1 mL of the plasmid-carrying donor and 1.0 mL of the recipient culture were added to 4.5 mL of prewarmed nutrient broth and mixed thoroughly. The mating mixture was incubated for 3 hours at 37 °C, after which 9.9 mL of prewarmed DM base was added. The mixture was centrifuged for 20 minutes at 4,000 rpm at room temperature, and the pellet resuspended in 9.9 mL of prewarmed DM base before being centrifuged a second time. The washed pellet was resuspended in 5.6 mL of DM base and serially diluted in DM base. 0.1 mL volumes of each dilution were spread onto selective DM agar containing the growth supplements of the recipient strain and containing an antibiotic to which the plasmid confers resistance. Washed donor and recipient cultures were also spread onto the selective medium to ensure that the selection plates were specific for the growth of transconjugants. 0.1 mL volumes of suitable dilutions of the mating mixture were also spread onto MacConkey agar to determine viable counts. The viable count of the donor culture was also determined by spreading appropriate dilutions of the culture onto MacConkey agar.
Isolated clones recovered from selective plates were streaked onto MacConkey agar and also onto DM minimal or Iso-Sensitest agar containing all antibiotics to which the transferred plasmid confers resistance, to confirm complete plasmid transfer.
MacConkey and Iso-Sensitest agar plates were incubated at 37°C overnight before examination. DM minimal plates were incubated for two nights before being examined for growth. Successful plasmid-carrying recipients were selected from antibiotic- containing plates and grown overnight at 37°C in nutrient broth before being streaked onto nutrient agar slopes for short-term storage, or transferred to liquid nitrogen for long term storage.
123 4. Determination of Minimum Inhibitory Concentrations (MICs)
4.1. Preparation of drug-containing plates
Plates were prepared from nutrient agar containing a range of concentrations of nalidixic acid. Nutrient Broth No. 2 powder and Lab M agar No. 1 sufficient to make 100 mL of nutrient agar were weighed into 150 mL glass bottles. Appropriate volumes of distilled water were added, allowing for the later addition of nalidixic acid solution to a final volume of 100 mL. The agar was sterilised at 115 °C (10 psi) for 20 minutes and the molten agar allowed to cool in a water bath at 55 °C. Once sufficiently cooled, suitable volumes of a fi*eshly prepared nalidixic acid solution were added to the molten agar, giving a final volume of 100 mL. The molten agar was thoroughly, but gently mixed and poured into five disposable sterile Petri dishes. Once set the plates were overdried at
44 °C for 45 minutes. Control drug-free nutrient agar plates were also prepared.
4.2. Preparation of inocula
Overnight cultures were serially diluted in nutrient broth and their viable counts determined on drug-free nutrient agar using the method of Miles and Misra (1938). To determine the MIC for each strain, drug-containing plates were inoculated with 1 jÆ volumes of neat and of 10'^ and 10"^ dilutions of each culture using a multi-point inoculator. Once the inocula had sunk into the agar, the plates were inverted and incubated overnight at 37°C. The MIC value was taken as the lowest concentration of drug to inhibit bacterial growth for all dilutions of culture.
124 5. Ultraviolet sensitivity testing
Cultures were grown overnight at 37°C in 4.5 mL nutrient broth and serially
diluted in nutrient broth. 20 /^L volumes of each dilution were plated, in triplicate, onto
nutrient agar plates. As soon as the inocula had absorbed into the agar, the plates were
irradiated, with the lids removed, under a low pressure mercury lamp which emitted light
at 254 nm. The height of the lamp wes set so that a UV fluence of 1 J m^ s ' was provided
for E. coli strains AB1157 and TK702, and 0.5 J m^ s ' for E. coli strains AB1886,
AB2463, AB2470 and AB2494. Exposure to UV light was carried out in a darkened
room to minimise photoreactivation (Setlow, 1966). Immediately following UV
irradiation, the lids were replaced and the plates incubated in the dark at 37 °C overnight.
Unirradiated control plates were incubated immediately after the drops of bacterial
suspension had sunk into the agar. The fraction of cells that survived following UV
irradiation was calculated as a percentage:
Number of viable UV-irradiated cells per mL x 100% Number of viable non-irradiated cells per mL
6. Reversion to amino acid prototrophy
100 mL volumes of nutrient broth in 1 L conical flasks (plugged with cotton wool
and muslin plugs) were inoculated and incubated overnight at 37°C. The cultures were
transferred to a shaking incubator, pre-heated to 37°C, and incubated for a further two
hours. The cultures were then centrifuged at 6,000 rpm, at 10°C for 20 minutes and the
pellet resuspended in 100 mL of DM base. The suspensions were centrifuged a second
time and the pellet resuspended in 1 mL of DM base producing a concentrated, washed bacterial suspension (lOON). This was serially diluted in DM base and 0.1 mL volumes
of appropriate dilutions were spread onto minimal DM agar containing all but one
125 required amino acid supplement plus all other growth requirements in excess. Once the
inocula had sunk into the agar, the plates were inverted and incubated for up to 14 days
at 37°C. Plates were examined for growth every 24 or 48 hours, and any visible colonies
scored. The viability of the initial bacterial suspensions was determined by plating 0.1 mL volumes of suitable dilutions onto fully-supplemented DM agar. Fully-supplemented DM plates used for viable count determination were incubated at 37°C for two days before being examined.
7. Forward mutation to nalidixic acid resistance on solid media
Two 10 mL overnight nutrient broth cultures of each strain were combines, washed twice in DM base, and resuspended in 2 mL of DM base to give a ION washed suspension. Serial dilutions were performed in DM base and 0.1 mL volumes of appropriate dilutions spread in triplicate onto nutrient agar plates containing different concentrations of nalidixic acid. The plates were inverted and incubated for ten days and the number of visible colonies on each plate scored at 24 or 48 hour intervals. The concentrations of nalidixic acid used were 4,8, 16 and 32 /^g mL’’ (1-, 2-, 4- and 8-times the MIC). Total viable counts of the plated cultures were determined by plating 0.1 mL volumes of suitable dilutions in triplicate onto drug-free nutrient agar. Drug-free control plates were incubated at 37 °C and examined after two days incubation. Mutation frequencies to nalidixic acid resistance were expressed as ratios of the number of visible nalidixic acid-resistant clones at each time point to the total number of viable cells plated.
126 7.1. Reversion to threonine independence following growth on nalidixic acid-
containing medium
After ten days incubation on plates containing nalidixic acid, a selection of
nalidixic acid-resistant clones were scraped off the plates using a sterile loop, and
resuspended in 1 mL of DM base. Control colonies not exposed to nalidixic acid were
selected from drug-free plates, used to determine viable counts, and resuspended in 1 mL
of DM base. The bacterial suspensions were serially diluted in DM base and 0.1 mL
volumes of suitable dilutions were spread, in triplicate, onto DM minimal agar containing
all growth requirements except the amino acid threonine. The plates were incubated at
37 °C and examined after two days. Viable counts of the resuspended colonies were
determined on nutrient agar and examined after overnight incubation.
8. UV-induced mutagenesis
8.1. UV-induced reversion to amino acid prototrophy
Overnight cultures grown in 10 mL nutrient broth were centrifuged at 4,000 rpm
for 20 minutes and resuspended in 10 mL DM base. The centrifugation was repeated and
the pellet resuspended in 10 mL DM base. DM overlays containing all growth requirements except the amino acid under test were kept molten in a water bath at 46 °C.
0.1 mL volumes of the undiluted washed cultures were added to each overlay in triplicate, mixed rapidly by rolling in the hands and poured onto prewarmed minimal DM base plates. Once the ‘soft’ agar had set, the plates were irradiated (lids removed) with UV in the dark before immediately being placed in the incubator at 37°C. The UV doses used were 1,2, 4, 8, 16 and 32 J m'^. Total viable counts were determined by adding 0.1 mL volumes of suitable dilutions to overlays, mixing and pouring onto fully-supplemented
127 base plates. Control, non-irradiated, fully-supplemented plates were incubated
immediately. To determine the viable counts after exposure to each UV dose, 0.1 mL
volumes of suitable dilutions were added to overlays, mixed thoroughly and poured over
fully-supplemented base plates. As soon as the ‘soft’ agar had set, these plates were then
exposed to each UV dose and incubated at 37°C. Plates were examined after two days
incubation and the number of colonies on each plate scored.
UV-induced reversion frequencies were expressed as the number of revertants per
10^ surviving bacteria, corrected for the number of spotaneous revertants.
8.2. UV-induced mutation to nalidixic acid resistance
10 mL overnight nutrient broth cultures were centrifuged twice at 4,000 rpm for
20 minutes and the pellet resuspended in 10 mL DM base. Fully supplemented overlays
containing all amino acids and growth requirements were heated at 100°C for 5 minutes
to melt the agar and cooled to 46 °C. Base plates were nutrient agar containing sufficient
nalidixic acid to give final concentrations of 4,8, 16 and 32 yug mL ' once the overlays
had been poured on. 0.1 mL volumes of the neat washed bacterial suspensions were
added to overlays, in triplicate, mixed and poured over the prewarmed nalidixic acid
containing plates. As soon as the ‘soft’ agar had set, the plates were UV-irradiated (lids
removed) in the dark before being incubated at 37 °C for up to eight days. The UV doses
used were 2,4,8,16 and 32 J m^ s '. Viable counts were determined by plating suitable
dilutions in overlays onto drug-free nutrient agar and incubating for 2 days. Viable counts
following UV exposure were determined by plating 0.1 mL volumes of each washed bacterial suspension in overlays onto drug-free nutrient agar and irradiating with UV
light. These plates were incubated for two days before being counted.
128 UV-induced mutation frequencies to nalidixic acid resistance were expressed as
the number of nalidixic acid resistant mutants per 10^ surviving bacteria, corrected for
the number of spontaneous resistant mutants.
9. Mutation to nalidixic acid or ciprofloxacin resistance in cultures grown in sub- inhibitory concentrations of nalidixic acid or ciprofloxacin
Strains were grown overnight in 4.5 mL nutrient broth. 1.5 mL of each overnight culture was diluted into 150 mL nutrient broth in a 500mL conical flask, containing 0,
0.5 X or 0.75 x MIC of nalidixic acid or ciprofloxacin. The flasks were incubated at 37 °C for ten days. At 24 or 48 hour intervals, 5 mL samples were removed from each flask, centrifuged at 4,000 rpm for 20 minutes, and the pellets resuspended in 5 mL DM base.
The resuspended cultures were centrifuged for a second time before being resuspended in 5 mL DM base. The washed bacterial suspensions were then serially diluted in DM base. 20/uL drops of the neat bacterial suspensions, or suitable dilutions were plated in triplicate onto nutrient agar containing 4 x MIC nalidixic acid or ciprofloxacin for that strain. Once the drops had absorbed into the agar, the plates were inverted and incubated for two days at 37°C. Viable counts of each suspension were determined by plating 20 ywL drops of suitable dilutions in triplicate onto drug-free nutrient agar. These drug-free plates were incubated at 37°C for one day before being counted.
In addition to plating on drug-containing nutrient agar at each time interval, 0.1 mL volumes of the neat washed suspensions were spread in triplicate onto DM minimal agar lacking the amino acid threonine, to test for unselected reversion to threonine prototrophy. Plates were incubated at 37°C for two days and the number of threonine-
129 independent revertants determined.
Results were expressed as the number of nalidixic acid resistant or threonine revertants per mL.
The minimum inhibitory concentrations (MICs) of nalidixic acid and ciprofloxacin were determined for a selection of drug-resistant clones selected from cultures grown in the presence of 0.75 x MIC of nalidixic acid or ciprofloxacin for four or eight days and plated onto medium containing 4 x MIC of nalidixic acid or ciprofloxacin. Selected colonies were transferred to 4.5 mL nutrient broth and grown overnight at 37°C. The overnight cultures were diluted to 10'^ and 10"^ in nutrient broth.
1 yuL drops of the neat culture and the dilutions were plated using a multipoint inoculator onto nutrient agar containing different concentrations of nalidixic acid or ciprofloxain.
Nalidixic acid was used at 2, 4, 8, 16, 32, 64, 128, 256, 512 and 1024 yUg mL ', and ciprofloxacin at 0.001,0.0025, 0.005,0.01,0.02,0.04, 0.08, 0.16, 0.32, 0.64 and 1.28 yUg mL '. Plates were incubated overnight before being examined for growth. The MIC values were taken as the lowest concentration to inhibit bacterial growth for all dilutions.
A selection of clones selected after growth for ten days in the presence of 0.75 x
MIC of nalidixic acid, and whose MICs were known, were grown for a second period of ten days in the presence of 0.75 x their resistant nalidixic acid or ciprofloxacin MICs.
Samples were removed at 24 or 48 hour intervals and plated onto nutrient agar containing
4 X their resistant MIC of nalidixic acid or ciprofloxacin, as described above. The plates were incubated for two days before being examined. Viable counts of each sample were performed on drug-free nutrient agar and counted after incubation for one day. The MICs
130 of a selection of clones growing on 4 x their resistant MIC of nalidixic acid or ciprofloxacin were then determined, as described above.
10. Polymerase Chain Reaction (PCR) amplification of the gyrA genes of nalidixic acid- and ciprofloxacin-resistant clones
PCR reaction mixtures (total volume 99juL) were set up in Eppendorf tubes as shown in Table 3, in the order given. The primers used allow the amplification of the +1 to +620 base pair region of the R coli gyrA gene, including the QRDR.
Table 3. Components for PCR amplification of theE. coli gyrA gene.
Component Cone, of stock Final amount in Volume of stock PCR added
Primer A 10 pmol yuL' 10 pmol 1 yUL
Primer B 10 pmol yuL' 10 pmol 1 yUL
dNTPs 4 mmol 200 yumol 5 lÆ
Taq Buffer (lOx) 200 mmol Tris/ 20 mmol Tris/ 10 yUL 500 mmol KCl 50 mmol Kcl
MgCls 50 mmol 2.5 mmol 5 y^L
Sterile HjO 75 yuL
DNA Small loopful from colony
Taq polymerase 1 U yWL ’ 2U 2iuL Primer A (5'-ATGAGCGACCTTGCGAGAGAAATTACACCG)
Primer B (3 -CTTCTGTAGTCGTAACTTCCCGACTACCTT)
131 30ywL of sterile light mineral oil was added to the surface of each tube to prevent evaporation during cycle amplification. The tubes were loaded into the thermal cycler and amplified according to the schedule shown below:
96 °C for 30 seconds
50 °C for 60 seconds
70 °C for 90 seconds followed by 24 rounds of:
96 °C for 15 seconds
50°C for 30 seconds
70 °C for 90 seconds and a final extension step of:
70 °C for 5 minutes
To ensure that PCR amplification of the DNA was successful, a small sample of the PCR reaction was checked by running on a 2% agarose gel.
10.1. Gel electrophoresis of amplified PCR product
10.1.1. Preparation of the gel
A 2% agarose gel was prepared by adding 0.4 g of agarose to 20 mL TAE buffer
(40 mmol Tris, 2 mmol EDTA, pH 8.0 with acetic acid) and dissolved by heating. The molten agarose was allowed to cool, but not set. Once cooled, 200 yuL of ethidium bromide (50 yUg mL ') was added and mixed thoroughly before pouring into a mini gel tank (Quick Screening Horizontal Gel Electrophoresis Unit) and allowing to set. Once set, 500 mL of TAE buffer was added to the tank.
132 10.1.2. Loading the gel
2 ywL of loading dye (0.1% bromophenol blue, 20% w/v glycerol) was added to
8 ywL of each PCR product and the 10 yuL samples were loaded into the wells of the gel.
2 yuL of loading dye was also mixed with 2 ywL of a 100 bp DNA ladder and this 4 yuL
sample loaded into the gel. The gel was run at 100 volts for one to two hours, until the
dye had almost reached the end of the gel. The gel was then viewed on a UV
transluminator. If PCR amplification of the selected region of DNA had been successful,
a single band of approximately 620bp was present in each lane of the gel.
11. Hinfl restriction analysis of PCR product
The amplified PCR products were subjected to Hinfl restriction endonuclease
treatment. The most common mutation conferring resistance to quinolone antibacterials
occurs at codon 83 of thegyrA gene, in the QRDR. The corresponding base change in gyrA results in the loss of aHinfl. restriction endonuclease recognition site.
The Hinfl digestion mixtures were set up as shown below to a final volume of
50 yUL.
Sterile HjO 4 yuL
PCR product 40 yuL
10 X Buffer B 5 yuL
Hinfl restriction enzyme (lOU ywL ’) 1 yuL
The digestion mixtures were centrifuged for 5 seconds in a microcentrifiige to ensure that the contents were thouroughly mixed. The mixtures were then incubated for
133 2 hours at 37°C in a water bath. The tubes were removed from the water bath and centrifuged for a further 5 seconds in the microcentrifuge. 25 of the mixture containing the digested DNA was removed and transferred to a fresh sterile PCR
Eppendorf tube. 4 fA. of loading dye was added to each tube and the tubes centrifuged for 5 seconds. 10 /.^L of the original, undigested PCR products were also transferred to clean, sterile PCR Eppendorf tubes and 2 yuL of loading dye added.
11.1. Gel electrophoresis of Hinfl digested PCR products
11.1.1. Preparation of the gel
A 4% agarose gel was prepared for the separation ofHinfl restriction products by weighing 8g of agarose into a 500 mL Duram bottle. 200 mL of TAE buffer was added and the mixture heated until all of the agarose had dissolved. The agarose was allowed to cool and 200jA. of ethidium bromide (0.5 yug mL'*) added. The molten gel was mixed gently and poured into an appropriate casting tray. Once set the gel was put into the electrophoresis tank (Bethesda Research Laboratories Horizon 20.25 Gel Tank) and 2 L of TAE buffer added.
11.1.2. Loading the gel
29 yt^L volumes of each digested DNA sample were loaded into wells. 10 yuL samples (2 yuL of loading dye mixed with 8 ywL of undigested PCR product) of the undigested PCR products were also loaded onto the gel. Two control samples were loaded onto the gel: both were subjected to PCR amplification as described above, but one contained no DNA and the other contained no primers. Therefore neither should reveal any bands following electrophoresis. 4 yuL of a marker (2 yuL of loading dye mixed
134 with 2 yuL of the 100 bp DNA ladder) were also loaded onto the gel, which was run at 100 volts for approximately 3 hours.
The gel was viewed on a UV transluminator and if appropriate was photographed using a Gel Electrophoresis Photosystem camera onto Polaroid film (type 665).
135 RESULTS
1. Preparatory experiments
1.1. Plasmid Transfer
Plasmds R46 and RP4 were introduced into the lactose-negative recipient E. coli
AB1157, from E. coli J6-2 donor strains, and plasmid R446b from an E. coli 343/113 donor strain. Plasmids R46 and R446b were introduced into repair-deficient mutants from
E. coli J6-2 and E. coli 343/113 donor strains respectively. The AB1157 and repair- deficient recipient strains bave different nutritional requirements from the donor strains, allowing the recipients, donors and transconjugants to be distinguished following mating.
The transconjugants were, therefore, selected on DM agar containing all growth supplements of the recipient strain plus an antibiotic to which the transferred plasmid confers resistance.
For example, when selecting forE. coli ABl 157 arg his leu pro thr thi (R46) transconjugants, the washed mating mixture was plated onto DM medium containing the amino acids arginine, histidine, leucine, proline and threonine, plus thiamine, and glucose as carbon source. On such medium, growth of the donor strain, J6-2pro his trp was inhibited since tryptophan was absent from the medium. Additionally, recipient ABl 157 cells that had not received the plasmid could not grow due to the presence of an antibiotic to which plasmid R46 confers resistance.
Clones ofE. coli ABl 157 transconjugants were isolated from selective plates and purified by streaking onto DM agar containing all nutritional requirements of the recipient plus one antibiotic to which the plasmid confers resistance. All antibiotic resistance traits were tested for to ensure complete transfer of the plasmid. Isolated colonies that exhibited the expected antibiotic resistance profile were also wired onto
136 MacConkey agar to verify their lactose-fermenting phenotype and stored on nutrient agar slopes or, in the long-term, as nutrient broth cultures in liquid nitrogen.
Plasmid transfer was expressed as the frequency per input donor cell, during a mating period of three hours (Table 4). The frequencies of transfer of plasmids R46 and
R446b into repair-deficient strains are shown in Tables 5 and 6 respectively. All the plasmids transferred at relatively high frequencies of between 10'^ and 10"^.
Table 4. Summary of plasmid transfer frequencies from E. coli strain J6-2 or E. coli 343/113 to& coli ABl 157 during three hours mating in nutrient broth.
Plasmid Resistance Selecting Plasmid transfer transferred phenotype of antibiotic in frequency per plasmid medium input donor cell
R46 Ap Tc Sm Su Ap 10/igmL‘‘ 2.1 X 10-^
R446b Tc Sm Tc 20 /^g m L' 2.0 X 10'^
RP4 Ap Km Nm Tc Km 10 //6g m L' 1.3 X 10-"
Table 5. Frequencies of transfer of plasmid R46 from E. coli J6-2 to repair-deficient E. coli strains.
Recipient strain Plasmid transfer frequency per input donor cell
A B l886 uvrA6 1.9 X 10-2 AB2463 recAlS 2.6 X 10 '
AB2470 recBll 4.5 X 10-^
AB2494 lexAS 1.6 X 10'^
TK702 umuC36 1.7 X 10-"
137 Table 6. Frequencies of transfer of plasmid R446b from E. coli 343/113 to repair- deficient E. coli strains.
Recipient strain Plasmid transfer frequency per input donor cell A B l886 uvrA6 4.6 X 10'^ AB2463 recAlS 2.3 X 10-2
AB2470 recBll 5.7 X 10-^ AB2494 lexA3 3.3 X 10-2
TK702 umuC36 1.1 X 10-^
1.2. Post-UV survival
1.2.1. The effect of plasmids on post-UV survival of repair-proficient Æ". coli AB1157
The post-UV survival properties of the plasmid-bearing AB1157 strains were also determined to confirm that plasmid-mediated UV protection was expressed (Pinney,
1980). The UV sensitivity of the repair-deficient mutants, and the effect of plasmid carriage on their UV sensitivity, were also studied.
Overnight nutrient broth cultures were serially diluted and each dilution plated onto nutrient agar. The plates were exposed to various UV doses, and the fraction of the culture surviving following UV irradiation was expressed as a percentage of the initial viable count.
The non-mutator plasmid RP4 (Pinney, 1980), as expected, had no effect on the post-UV survival of E. coli ABl 157 (Figure 4). Plasmids R46 and R446b enhanced post-
UV survival. After a dose of 100 J m'^, plasmids R46 and R446b conferred similar levels of protection on the host, producing 6.2- and 5.4-fold increases in the surviving fraction respectively. These results verify that plasmids R46 and R446b confer a UV protection phenotype upon their host, whereas plasmid RP4 does not (Pinney, 1980).
138 100
R446b
RP4
0 40 80 120 160 UV Dose (J m"^) Figure 4. The effect of plasmids on the post-UV survival ofE. coli strain ABl 157.
1.2.2. The effect of repair deficiencies on post-UV survival
As expected, the repair-deficient mutants showed increased sensitivities to UV irradiation (Figure 5), the order of sensitivity of the repair-deficient mutants being similar to that previously determined by Upton and Pinney (1983). The effects of the mutator plasmids R46 and R446b on post-UV survival of the repair-deficient strains are reported in Table 7. R46 and R446b produced no significant increase in the surviving fraction of the recAlS (Figure 6) or the lexA3 (Figure 7) mutants (Table 7), confirming the dependence of the plasmid-bome mucAB genes on hostrecA^ and lexA^ genotypes for expression. The Muc^ phenotype of plasmids R46 and R446b is not dependent on functional hostrecB^, uwA^ and genes and was therefore expressed in such repair-
139 1 0 0 ■OAB1157
10 umuC36
>
1 lex A3 uvrA6 recB21
recA 13
0 10 20 30 4 0 UV Dose (J m"^) Figure 5. Relative UV sensitivities of repair-deficient E. coli strains.
deficient backgrounds. In strain AB2470 recB21, the presence of R46 and R446b led to
5.4- and 1.3-fold increases respectively, in the fi’action of surviving cells following a UV dose of 15 J m'^ (Figure 8, Table 7). Similarly, following a UV dose of 6 J m'^, R46 and
R446b increased the post-UV survival of strain A B l886 uvrA6 16.5- and 8.8-fold respectively (Figure 9, Table 7). When transferred into TK702 umuC36, plasmids R46 and R446b increased the fi-action of cells surviving a UV dose of 80 J m"^ by 37.6- and
35-fold respectively (Figure 10, Table 7).
140 Table 7. Effects of plasmids R46 and R446b on the post-UV survival of the repair- proficient Æ coli ABl 157 and repair-deficient derivatives. For each strain, the UV doses selected gave approximately 1% survival of the R strain.
Strain UV Dose Post-UV survival (RVR ) (Jm-2) R46 R446b
ABl 157 120 11.7 8.4 AB2463 recAlS 1.5 0.4 0.3 AB2494 lexAS 10 2.1 0.6
AB2470 recBll 15 5.4 1.3
A B l886 uvrA6 6 16.5 8.8
TK702 umuC36 80 37.6 35.0
00
(O > > 3 (fi
R46 R~ R446b
0 2 4 6 UV Dose (J m"^) Figure 6. Absence of UV-protective effect of R46 and R446b in E. coli AB2463 recAlS.
141 100
<0 > > 3 (/)
R46 1 R446b
0 5 10 15 20 UV Dose (J m"^) Figure 7. Absence of UV-protective effect of R46 and R446b in E. coli AB2494 lexAS.
00
3 CO R46
R446b
0 5 10 15 20 UV Dose (J m"^) Figure 8. Effect of R46 and R446b on the post-UV survival of E. coli AB2470 recB21.
142 100
10 R46 R446b CO 1
0 2 4 6 8 UV Dose (J m“^) Figure 9. Plasmids R46 and R446b protect E. coli A B l886 uvrA6 against UV.
00
(0 > ’> R46 CO R446b
0 40 80 120 UV Dose (J m”^) Figure 10. Plasmids R46 and R446b protect E. coli TK702 umuC36 against UV.
143 1.3. Determination of nalidixic acid MICs
In order to examine frequencies of mutation to nalidixic acid resistance, it was first necessary to determine the MICs of nalidixic acid against the repair-proficient E. coli
ABl 157 strain and its repair-deficient mutants. The effect of plasmid carriage on these
MICs was also examined.
1.3.1. The effect of plasmids on MIC to nalidixic acid in E, coli ABl 157 and its plasmid-bearing derivatives
Neat, 10'^ and ICT^ dilutions of overnight nutrient broth cultures were plated in
I yuL volumes onto nutrient agar plates containing increasing concentrations of nalidixic acid, and the plates examined for growth after overnight incubation at 37°C. The MIC value was taken as the lowest concentration of drug to inhibit bacterial growth at all dilutions. Concentrations of nalidixic acid incorporated into the nutrient agar plates were
0, 0.5, 0.75, 1.0, 1.5, 2.0, 3.0, 4.0 and 5.0 yUgmL’’.
The MIC for the plasmid-free ABl 157 strain was recorded as 4 yUg mL ', since this was the lowest concentration of drug which inhibited growth of the strain at all dilutions (Table 8). A similar result was obtained for the plasmid-harbouring strains
(Table 8). Therefore, plasmids R46 and R446b and RP4 appear to have no effect on the nalidixic acid MIC of E. coli strain ABl 157.
144 Table 8. MIC values of nalidixic acid fbr& coli strain ABl 157 and its plasmid-bearing derivatives. Figures indicate numbers of isolated colonies observed.
Cone, of Growth (+/-) ofE. coli ABl 157 and plasmid-bearing derivatives nalidixic acid (Aig mL'*) R- R46 R446b RP4 N 10-^ 10-^ N 10*^ 10"^ N 10-2 10-^ N 10-2 10-^
0 + + + -k + + + + + -h + +
0.5 4 - + + + + + -t- -f- +
0.75 + + + + + + + + + LA 1.0 + + + + + + + + +
1.5 + -h + -H + + + + + - k -H
2.0 + -t- T -t- + T + + 18 -t- + T
3.0 8 - + 28 - + 12 - T - -
4.0 ------
5.0 ------
+ = confluent growth T = faint trace = no growth 1.3.2. The effect of repair deficiencies on MICs of nalidixic acid
MICs of nalidixic acid were also determined for the repair-deficient E. coli strains
ABl886 uvrA6, AB2463 recA13, AB2470 recB21, AB2494 lexA3 and TK702 umuC36.
1 yuL volumes of neat, 10'^ and 10"^ dilutions were plated onto nutrient agar containing the following concentrations of nalidixic acid; 0, 0.5, 1.0, 1.5, 2.0, 3.0, 4.0 and 5.0jj,% m L'.
Nalidixic acid MICs varied, according to the repair-deficient strain (Table 9).
Strains AB2463 recA13 and AB2470 recBll were more sensitive than their wild-type
ABl 157 parent, with MIC values of 2 yUg mL"'. However, the MIC values for strains
A B l886 uvrA6 and TK702 umuC36 were the same as ABl 157, with strain AB2494 lexA3 giving a higher MIC of 5 yUg mL ' (Table 9).
MIC values were also determined for repair-deficient strains harbouring plasmids
R46 and R446b. It was found that, as with the repair-proficient ABl 157 strain, the presence of either plasmid did not affect the MIC of any strain (data not shown).
146 Table 9. MIC values of nalidixic acid for repair-deficient strains of E. coli. Figures indicate numbers of isolated colonies observed.
Cone, of Growth (+/-) ofE. coli repair-deficient strains nalidixic acid AB 2494 le: kA3 TK7 02 umu C36 (yUgmL'*) AB .886 UV rA6 AB2463 rec A13 AB2470 rec B21 N 10-2 10"* N 10-2 10"* N 10-2 10"* N 10-2 10"* N 10-2 lo-'* 0 + + + + + + + + + + + + + + +
0.5 + + + + + + + + + + + + + + + 1.0 + + + + + 18 + + 4 + + + + + + 1.5 + + + + T - 10 T - + + + + + +
2.0 + + 13 ------+ + • + + + + 3.0 + 18 T ------+ + + + 12 T
4.0 ------+ 13 T ---
5.0 ------
+ = confluent growth T = faint trace nogrowth 1.4. Summary and conclusions
Mutator and control plasmids from different incompatibility groups were readily transferred into E. coli ABl 157 (Table 4) and its repair-deficient mutants (Tables 5 and
6). Complete plasmid transfer was confirmed by determining the antibiotic resistance profiles and UV sensitivities of the new hosts. Carriage of the known mutator plasmids
R46 and R446b (Pinney, 1980; Upton and Pinney, 1983) significantly increased the post-
UV survival of the repair-proficient strain AB 1157 (Figure 4) but did not affect the level of resistance to nalidixic acid in this strain (Table 8), Nalidixic acid MICs of the repair- deficient mutants were not affected by plasmid carriage. Plasmid-mediated UV resistance was expressed in strains A B l886 uvrA6, AB2470 recB21 and TK702 umuC36, but not in AB2463 recAlS or AB2494 lexA3 (Figures 6 to 10).
Having successfully transferred R46 and R446b into both repair-proficient and repair-deficient strains, it was possible to investigate how the mutator phenotypes, conferred by these plasmids, might affect reversion to amino acid independence under conditions that favour adaptive mutation.
148 2. Reversion to amino acid independence
2.1. Introduction
Selection-induced or “adaptive” mutations, which occur only when they are of immediate benefit to the organism, have been the source of controversy for many years
(Luria and Delbriick, 1943; Lederberg and Lederberg, 1952; Ryan and Wainwright, 1954;
Ryan, 1955; 1959; Ryan et al, 1961; 1963; Shapiro, 1984). In 1988, Cairns and co workers demonstrated an increased frequency of mutation to lactose utilisation in an initially lactose negative population ofE. coli when plated on media containing lactose as sole carbon source. Such mutations have been termed adaptive, since their formation has only been detected during non-lethal selection, in genes whose function was selected
(Cairns et al, 1988) (see Introduction Section 5.1).
Hall (1990) proposed that stationary phase populations of bacteria experience nutritional stress, when the majority of DNA synthesis is confined to repair. Spontaneous damage to DNA occurs continuously as a result of thermal and hydrolytic degradation, or from macromolecular synthesis. If repair of such damage is error-prone, then a random mutation may lead to the production of a useful protein, allowing the cell to grow and overcome the nutritional stress, switching DNA synthesis from repair to replication and fixing the new mutation within the chromosome (Stahl, 1990).
In order to examine selection-induced mutation, cultures of E. coli ABl 157, which is auxotrophic for the amino acids arginine, histidine, leucine, proline and threonine, were plated on DM agar in the absence of one required amino acid. The frequency at which each allele reverted could then be observed over a ten day incubation
149 period. In addition, if DNA damage and error-prone repair are responsible, at least in part
(Bridges, 1998), for such selection-induced mutation, then the presence of a mutator plasmid in the system would be expected to increase the frequency at which revertants appeared under selection.
2.2. Spontaneous and UV-induced reversion to amino acid independence
Control experiments were initially performed to determine spontaneous and UV- induced reversion frequencies in the presence or absence of a mutator plasmid. The presence of mutator plasmids has previously been demonstrated to increase the frequency of UV-induced reversion to lysine independence in& coli strain 343/113 (Pinney, 1980) and arginine independence in E. coli strain ABl 157 (Upton and Pinney, 1983).
To determine UV-induced reversion to amino acid independence, aliquots of washed overnight cultures oïE. coli ABl 157 and its R46- and R446b-bearing derivatives were plated in molten ‘soft’ agar overlays lacking one essential amino acid supplement onto DM minimal agar base plates. Once set, the plates were exposed to UV light and immediately incubated at 37°C. Viable counts were determined using ftilly supplemented
‘soft’ agar overlays. UV-induced reversion frequencies to amino acid independence were expressed as the number of revotants per 10^ surviving cells at each UV dose.
Even in the absence of exposure to UV, plasmids R46 and R446b increased frequencies of reversion to threonine, leucine, arginine and histidine independence following an incubation period of two days (Table 10). This confirms the spontaneous mutator effect conferred by such plasmids (Mortelmans and Stocker, 1976). In strain
150 Table 10. Effect of plasmids R46 and R446b on the frequency of spontaneous reversion to amino acid independence in E. coli ABl 157. Reversion frequencies are expressed as the number of revertants per 10* viable cells.
Spontaneous Fold increase in Fold increase in Amino reversion spontaneous reversion spontaneous reversion acid under frequency of frequency due to the frequency due to the test E. coli ABl 157 presence of plasmid presence of plasmid R46 R446b Threonine 0.5 62.4 5.2
Leucine 1.6 7.8 4.4
Arginine 0.3 39.3 10.3 Histidine 0.7 37.0 19.9
ABl 157, the frequencies of both spontaneous and UV-induced reversion to proline
independence were very low (<6 x 10^). Proline reversion was not therefore tested in the
subsequent experiments.
As expected, exposure to UV irradiation increased the frequencies of reversion
of all four alleles tested (Figures 11 to 14). A dose of 32 J m'^ increased the reversion
frequencies of the plasmid-free ABl 157 strain to arginine-, threonine-, leucine- and histidine-independence by 4 X 1 0 \ 130-, 10- and 5.7-fold respectively (Table 11; Figures
11 to 14). Plasmids R46 and R446b increased UV-induced reversion (Table 11), but their mutator effects varied. Following exposure to a UV dose of 32 J m'^, R46 and R446b increased the frequency of UV-induced reversion to histidine independence 660- and 59-
fold respectively (Table 11; Figure 14), whereas UV-induced reversion to arginine independence was increased only 17- and 3.2-fold respectively (Table 11; Figure 11).
151 R46
R446b 0> Q.
10 X)
0 10 20 30 40
UV Dose (J m"^) Figure 11. Plasmids R46 and R446b increase UV-induced mutagenesis to arginine independence in E. coli ABl 157 argE3.
R46
10 0 10 20 30 40 UV Dose (J m ) Figure 12. Plasmids R46 and R446b increase UV-induced mutagenesis to threonine independence in E. coli ABl 157 thr-1. 152 0) R46 Q. R446b 10 0> E 1 G 10 20 30 40 UV Dose (J m“^) Figure 13. Plasmids R46 and R446b increase UV-induced mutagenesis to leucine independence in E. coli ABl 157 leuB6. R46 Q> CL R446b 10 0> R- .O E 1 0 10 20 30 40 UV Dose (J m~^) Figure 14. Plasmids R46 and R446b increase UV-induced mutagenesis to histidine independence in E. coli ABl 157 hisG4. 153 Table 11. Effect of plasmids R46 and R446b on the UV-induced frequencies of reversion to amino acid independence in E. coli ABl 157, following exposure to a UV dose of 32 J m'^. Reversion frequencies are expressed as the number of revertants per 10* viable cells. UV-induced Fold increase Fold increase in Fold increase in Amino reversion over UV-induced UV-induced acid under frequency of spontaneous reversion reversion test E. coli reversion frequency due to frequency due to ABl 157 frequency the presence of the presence of plasmid R46 plasmid R446b Threonine 65 130 26.9 5.3 Leucine 17 10.6 64.1 6.1 Arginine 1220 4067 17.3 3.2 Histidine 4 5.7 660 59.3 2.3. Selection-induced reversion to amino acid independence Results presented in the previous section confirmed the revertability of the four alleles in E. coli ABl 157 and the spontaneous and UV-induced mutator effects of plasmids R46 and R446b. The following experiments were designed to demonstrate selection-induced (adaptive) reversion to amino acid independence and to determine whether the presence of a mutator plasmid might increase the effect: plasmid R46 was used because it gave higher reversion frequencies than plasmid R446b (Tables 10 and 11). Washed overnight cultures of the plasmid-free AB1157 argES hisG4 leuB6 proA2 thr-1 strain and its R46-bearing derivative were plated (approximately 10* cells) onto DM agar lacking either arginine, histidine, leucine or threonine. Plates were incubated for ten days at 37°C and the number of revertant colonies scored at 24 hour intervals. Viable counts were determined on fully-supplemented DM agar. 154 Although very tiny, revertant E. coli AB1157 clones could be detected after 24 hours incubation; they needed two days incubation to be reliably quatifiable. During the next eight days, the number of visible revertant clones increased by varying degrees in all test systems (Figures 15 to 18), illustrating that adaptive mutation was, indeed, occurring. A 20.5-fold increase in the number of leucine revertant clones per plate was observed between day 2 and day 10 (Figure 15), with reversion of the thr-1 allele increasing 5.8-fold (Figure 16). Selection-induced reversion of E. coli ABl 157 to histidine or arginine independence was less pronounced. However, increases in the number of revertant clones of 3.5 and 2.0-fold respectively were observed following ten days incubation (Figures 17 and 18). Reversion frequencies to threonine, leucine and arginine independence were further increased 2.1-, 2.9- and 3.9-fold respectively, by the presence of plasmid R46. However, reversion to histidine independence was unaffected by carriage of R46. Reversion frequencies after ten days incubation are presented in Table 12. Table 12. Reversion frequencies ofE. coli strain ABl 157 to amino acid independence after ten days growth on DM agar lacking the amino acid under test. The effect of the mutator plasmid R46 on reversion frequencies is also shown. Reversion frequencies are expressed as the number of revertants per viable cell plated. Amino acid Fold increase in Reversion frequency Fold increase (+), under test number of revertant to amino acid or decrease (-) clones per plate independence after produced by after 10 days 10 days presence of R46 Threonine 5 .8 3 .0 X 10'" + 2 .1 Leucine 2 0 .5 9.8 X 10'^ + 2 .9 Arginine 2 .0 1.7 X 1 0 ’^ + 3 .9 Histidine 3 .5 2 .5 X 10'^ - 1 .2 155 R46 ^ 100 R" 10 1 0 2 4 6 8 10 Incubation time (days) Figure 15. Accumulation of Icucinc-indcpcndcnt revertants in cultures of E. coli AB 1157 leuB6 plated on medium lacking leucine, and the effect of plasmid R46. a> Q. R46 O R" 00 10 q) .o E 1 0 2 4 6 8 10 Incubation time (days) Figure 16. Accumulation of threonine-independent revertants in cultures of E. coli ABl 157 thr-1 plated on medium lacking threonine, and the effect of plasmid R46. 156 i 100 R" R46 10 1 0 2 4 6 8 10 Incubation time (days) Figure 17. Accumulation of histidine-independent revertants in cultures of E. coli ABl 157 hisG4 plated on medium lacking histidine, and the effect of plasmid R46. JO Q. 0> a. 100 R46 10 0) -Q E 1 0 2 4 6 8 10 Incubation time (days) Figure 18. Accumulation of arginine-independent revertants in cultures of E. coli ABl 157 argES plated on medium lacking arginine, and the effect of plasmid R46. 157 2.4. Summary and conclusions E. coli ABl 157 exhibited selection-induced reversion to amino acid independence, with the frequency of reversion depending upon the allele tested. The presence of plasmid R46 doubled, tripled or quadrupled selection-induced reversion to threonine, leucine or arginine independence respectively, but had no effect on hisG4 reversion (Table 12). It therefore appears, that the Muc phenotype (Shanabruch and Walker, 1980), conferred by plasmid R46 is not only induced by UV damage (Table 11), but also during non-lethal selection for adaptive reversion to amino acid independence. This suggests a role for error-prone DNA repair in the production of “adaptive” mutations, and raises the interesting possibility that, since nalidixic acid damages bacterial DNA and induces mutagenic error-prone repair (Phillips et al, 1987; Piddock and Wise, 1987), exposure to the drug might increase the frequency at which nalidixic acid mutants are produced i.e. the drug is inducing mutation to its own resistance. If error-prone repair is indeed involved, then the presence of a mutator plasmid would be expected to increase this effect. Experiments to test this hypothesis are presented in Results sections 3 and 4. Similarly, if the hypothesis is correct, it might be expected that such selection-induced mutation would be reduced in the absence of error-prone repair, and might be heightened in strains in which the effects of this pathway are increased. Experiments to test this are presented in Results section 6. 158 3. Selection-induced mutation to nalidixic acid resistance on solid media 3.1. Introduction Clinical resistance to quinolone antibacterials, such as nalidixic acid, arises via mutation in the bacterial chromosome, rather than by acquisition of resistance genes carried on R plasmids (Grüneberg, 1992). Smith (1986) studied the development of resistance to ten quinolones in&coli KL16, when plated onto media containing five, ten or twenty times their MIC. The development of mutational resistance to nalidixic acid, cinoxacin, oxolinic acid, fiumequine, rosoxacin and pipemidic acid occurred readily, at fi’equencies of 10'* to 10'^, whereas E. coli KL16 mutated to ciprofloxacin, norfloxacin and ofloxacin resistance at much lower frequencies of 10'^^ or lower. He also demonstrated the inverse relationship between the expression of resistance and the drug concentration on which the mutants had been selected: fewer mutants grew on media containing high concentrations of quinolone than on low concentrations of drug (Smith, 1986). More recently, Riesenfeld et al (1997) reported what appeared to be adaptive mutation to ciprofloxacin resistance. Cultures ofE. coli MG 1655 were spread onto medium containing twice the MIC of ciprofloxacin and incubated for seven days. During this period, the fi'equency of mutation to ciprofloxacin resistance increased 100-fold fi"om 2.5 X 10 * to 2.5 X 10"^. It was, therefore, decided to test for selection-induced mutation to nalidixic acid resistance using medium containing either inhibitory or lethal concentrations of drug. A concentration equivalent to twice MIC, as used by Riesenfeld et al (1997), would kill susceptible cells, allowing resistant mutants to grow through, some of which might grow slowly, appearing several days after the cultures were plated. 159 Use of an inhibitory, and therefore non-lethal, concentration of nalidixic acid would fulfill the criteria of Cairns et al (1988) for adaptive mutation: the appearance of mutants in non-growing cultures during non-lethal selection. 3.2. Spontaneous and UV-induced mutation to nalidixic acid resistance Control experiments were performed initially, to determine both spontaneous and UV-induced mutation frequencies to nalidixic acid resistance of E. coli ABl 157 and its plasmid-carrying derivatives. Washed overnight cultures were added to molten ‘soft’ nutrient agar overlays and poured over nutrient agar plates containing sufficient nalidixic acid to give final concentrations of 4, 8 or 16 mL '. Once set, the plates were either incubated immediately or exposed to UV light before being incubated. Total viable counts were determined on drug-free nutrient agar. Plates were examined for growth after two days incubation, and the number of nalidixic acid-resistant clones scored. The frequency of UV-induced mutation to nalidixic acid resistance was expressed as the number of nalidixic acid-resistant mutants per 10^ viable cells plated. Plasmids R46 and R446b increased mutation frequencies to all concentrations of nalidixic acid tested (Table 13), again illustrating the spontaneous mutator effect conferred by the plasmids. In contrast to reversion to amino acid independence, where R46 produced the greater enhancement of reversion frequencies (Results section 2.2), the presence of plasmid R446b led to a greater increase in mutation frequency to nalidixic acid resistance at all concentrations tested (Table 13). 160 As expected, exposure to UV light increased mutation frequencies to nalidixic acid resistance (Figures 19 to 21). Plating on 4, 8 or 16 yWg mL * nalidixic acid showed that a UV dose of 32 J m'^ increased mutation by 11.5-, 26.2- and 21.5-fold respectively (Table 14). Table 13. Effect of plasmids R46 and R446b on the spontaneous mutation frequency to nalidixic acid resistance in E. coli ABl 157. Mutation frequencies are expressed as the number of resistant mutants per 10^ viable cells. Cone, of Spontaneous Fold increase in Fold increase in nalidixic mutation spontaneous mutation spontaneous mutation acid frequency to NA frequency due to the frequency due to the (^g mU*) resistance of presence of plasmid presence of plasmid E.coliABWSl R46 R446b 4 869 1.3 2.3 8 23 2.3 3.7 16 2 2.5 3.0 Table 14. Effect of plasmids R46 and R446b on the frequency of mutation to nalidixic acid resistance of E. coli ABl 157, following exposure to a UV dose of 32 J m'^. Mutation frequencies are expressed as the number of resistant mutants per 10® viable cells. Cone, of UV-induced Fold increase in Fold increase in Fold increase in nalidixic mutation mutation UV-induced UV-induced acid frequency frequency mutation mutation (/^g mL*) produced by frequency due frequency due UV (compare to R46 to R446b Table 13) 4 1.0 X 10^ 11.5 2.1 3.4 8 602 26.2 1.9 4.5 16 43 21.5 2.4 3.1 161 R446b R46 < 10 1 jQ0) E 0 10 20 30 40 UV Dose (J m"^) Figure 19. Effect of R46 and R446b on the UV-induced mutagenesis oiE. coli ABl 157 to nalidixic acid (NA) resistance (4 /^g mL *). 0> CL m R446b « R46 0) < 10 Q> 1 .O 3E 0 10 20 30 40 UV Dose (J m"2) Figure 20. Effect of R46 and R446b on the UV-induced mutagenesis of E. coli ABl 157 to nalidixic acid (NA) resistance (8 />^g mL'*). 162 (U a. R446b : 1 R46 V < 10 0) 1 3E 0 10 20 30 40 UV Dose (J m“^) Figure 21. Effect of R46 and R446b on the UV-induced mutagenesis of E, coli ABl 157 to nalidixic acid (NA) resistance (16 lag mL *). Plasmids R46 and R446b both increased the frequency of UV-induced mutation to nalidixic acid resistance at all nalidixic acid concentrations tested (Figures 19 to 21; Table 14). Similar to the results obtained for spontaneous mutation to nalidixic acid resistance (Table 13) and in contrast to amino acid reversion (Results section 2), R446b had the greater effect (Figures 19 to 21). After irradiation with a UV dose of 32 J m'^, the presence of plasmid R446b increased the frequency of mutation to resistance to 4, 8 and 16 yug mL * nalidixic acid by 3.4-, 4.5- and 3.1-fold respectively (Table 14). The presence of plasmid R46 led to increases of 2.1-, 1.9- and 2.4-fold in UV-induced mutation frequency to resistance to 4, 8 and 16 yug mL * nalidixic acid respectively (Table 14). Mutation frequencies to nalidixic acid resistance in E. coli cultures plated on solid 163 medium containing nalidixic acid are, therefore, increased by exposure to the DNA damaging agent UV light and are also enhanced by the presence of a mutator plasmid, even in undamaged cells. These results suggest the involvement of error-prone repair in the development of resistant mutants. It therefore seemed appropriate to test not only if nalidixic acid would induce mutation to its own resistance, but also whether the presence of a mutator plasmid would increase such selection-induced mutation. 3.3. Selection-induced mutation to nalidixic acid resistance on solid media Preliminary experiments showed that it was impossible to quantify true nalidixic acid-resistant clones growing against the background of confluent growth that occurred when approximately 10^ cells were plated on solid media containing sub-inhibitory concentrations of nalidixic acid. To determine selection-induced mutation to nalidixic acid resistance, aliquots of washed overnight cultures of the plasmid-free ABl 157 strain and its R46- and R446b- bearing derivatives were therefore plated onto nutrient agar containing 4, 8 or 16 //g mU’ nalidixic acid (one, two or four times MIC). It can be argued that since a nalidixic acid concentration of 4 yUg mU* is the MIC for thisE. coli strain (Table 8) this concentration satisfies the criteria of Cairns et al (1988) that selection for adaptive mutation should occur under conditions that inhibit bacterial growth and provide non-lethal selection. Drug-containing plates were incubated at 37 °C for ten days and the number of visible resistant colonies on each plate scored at 24 or 48 hour intervals. The total viable counts of the plated bacterial suspensions were determined after overnight incubation on drug- free nutrient agar. Results are expressed as the number of nalidixic acid-resistant clones appearing on the plates at each time interval. 164 67 resistant clones of the plamid-free ABl 157 strain were visible following incubation for one day on nutrient agar containing 4 ^ag mL * nalidixic acid, (Table 15). In agreement with Smith (1986), increasing the concentration of nalidixic acid in the medium, to 8 or 16 /.^g m L’, greatly reduced the number of nalidixic acid-resistant clones recovered after 24 hours incubation (Table 15). Even after incubation for one day, the effect of the mutator plasmids was evident. At a nalidixic acid concentration of 4 yug mL', R46 and R446b increased the number of nalidixic acid-resistant clones by 2.8- and 9.1-fold respectively (Table 15). At higher concentrations of nalidixic acid, the plasmid effect was reduced. The number of visible nalidixic acid-resistant clones increased during the ten day incubation period for all strains and at all nalidixic acid concentrations tested (Figures 22 to 24). During the ten day incubation period on medium containing 4 /Wg mL ' nalidixic Table 15. Numbers of nalidixic acid-resistant E. coli ABl 157 clones per plate following incubation for one day. The effect of plasmids R46 and R446b is also shown. Cone, of Number of Fold increase in the Fold increase in the nalidixic resistant clones number of NA resistant number of NA resistant acid clones due to R46 clones due to R446b (Aig mL') 4 67 2.8 9.1 8 3 1.3 1.7 16 0.3 1.3 2.0 165 acid (MIC), the number of visible nalidixic acid-resistant ABl 157 clones per plate increased 240-fold to 1.6 x lO'* (Figure 22 and Table 16). Incubation for ten days on medium containing 8 (Figure 23) or 16 yUg mL ' (Figure 24) of nalidixic acid led to 553- and 37-fold increases in resistant ABl 157 clones respectively (Table 16). Throughout the ten day incubation period on nalidixic acid-containing medium, the presence of plasmids R46 and R446b further increased the number of nalidixic acid-resistant mutants recovered (Figures 22 to 24, Table 16). The greatest increases in selection-induced mutation to nalidixic acid resistance were observed with the IncM plasmid R446b. Following ten days incubation on medium containing 4, 8 and 16 yUg mL ' nalidixic acid, the presence of R446b increased the number of visible drug-resistant clones 4.7-, 2.5- and 3.0-fold respectively (Table 16). The effect of plasmid R46 was less pronounced, producing almost identical increases of 2.2-, 2.0- and 2.1-fold in the number of nalidixic acid-resistant clones recovered following ten days incubation on medium containing 4, 8 and 16 yug mL-1 nalidixic acid respectively (Table 16). Table 16. Numbers of nalidixic acid-resistant E. coli ABl 157 clones per plate following incubation for ten days. The effect of plasmids R46 and R446b is also shown. Cone, of Number of Fold increase in Fold increase Fold increase nalidixic resistant resistant due to R46** due to R446b** acid clones clones* (yug mL') 4 1.6 X 10" 240 2 .2 4.7 8 1.7 X 10^ 553 2 .0 2.5 16 11 37 2.1 3.0 * compared to number after one days incubation. ** compared to the plasmid-free strain after ten days incubation. 166 ■5. 1 0 ‘ R446b a> CL R46 R" 10' 10' 10' 10 0 2 4 6 8 10 Incubation time (days) Figure 22. Accumulation of nalidixic acid-resistant mutants in cultures of E. coli AB 1157 plated on medium containing 4 //g mL * nalidixic acid (NA), and the effect of plasmids R46 and R446b. 0) J2 a . « R446b Q. R46 R" 10 0) E 1 0 2 4 6 8 10 Incubation time (days) Figure 23. Accumulation of nalidixic acid-resistant mutants in cultures oiE. coli ABl 157 plated on medium containing 8 /Wg mL * nalidixic acid (NA), and the effect of plasmids R46 and R446b. 167 o J2 Q. 0) Q. R446b R46 0 0) < Xi0> E =3 0 2 4 6 8 10 Incubation time (days) Figure 24. Accumulation of nalidixic acid-resistant mutants in cultures of E. coli AB 1157 plated on medium containing 16 yUg mL'' nalidixic acid (NA), and the effect of plasmids R46 and R446b. 3.4. Summary and conclusions E. coli ABl 157 mutated to nalidixic acid resistance with increasing frequency during selection on media containing inhibitory or lethal concentrations of nalidixic acid for up to ten days (Figures 19 to 21). The presence of the mutator plasmids, R46 and R446b, further increased the frequency at which cells mutated to nalidixic acid resistance, both in the absence and presence of UV-induced mutagenesis (Tables 13 and 14). The enhancement of mutation to nalidixic acid resistance by the presence of the mutator plasmids indicates the involvement of error-prone repair, not only in UV-induced mutation to drug resistance, which was previously demonstrated by Ambler et al (1993), but also in selection-induced 168 mutation to nalidixic acid resistance in the absence of UV exposure. Demonstration of increased mutation frequencies with time in inhibited but not dying populations, suggests that true Caimsian adaptive mutation to nalidixic acid resistance may be induced by inhibitory concentrations of nalidixic acid. It was therefore decided to examine the effects of sub-inhibitory concentrations of nalidixic acid in liquid media on mutation frequencies to nalidixic acid resistance. 169 4. Selection-induced mutation to quinolone resistance in cultures grown in liquid medium containing sub-inhibitory concentrations of drug 4.1. Introduction One of the accepted criteria for adaptive mutation is that it must result from cells exposed to non-lethal selection (Cairnset al, 1988). In the experiments of Caimset al (1988), cells defective for lactose metabolism were plated onto a medium lacking lactose. Such conditions are not immediately lethal, but growth will only ensue when a successful mutation, permitting lactose utilisation, is produced. Riesenfeld et al (1997) have recently claimed adaptive mutation to ciprofloxacin resistance. They reported a 100-fold increase in the frequency of mutation to ciprofloxacin resistance per viable ciprofloxacin sensitive cell during a seven day period of incubation on medium containing twice the MIC of ciprofloxacin, which is bactericidal to sensitive cells. These data are similar to those obtained with 8 and 16 /Wg mL"’ nalidixic acid (Results section 3.3), and reported by Hinton and Pinney (1997a). Non-lethal selection of nalidixic acid-resistant mutants was therefore examined by growing liquid cultures for extended periods in sub-inhibitory concentrations of nalidixic acid. The effect of mutator plasmids on the development of such adaptive mutation was also determined, to give insight into the role played by SOS repair in the adaptive response. As a control, the effect of the non-mutator plasmid, RP4, was also examined. 4.2. Mutation to nalidixic acid resistance in E, coli cultures grown in liquid medium containing sub-inhibitory concentrations of nalidixic acid 1.5 mL volumes of overnight nutrient broth cultures were diluted into 150 mL of 170 nutrient broth containing 0, 2 or 3 yug mL * nalidixic acid and incubated at 37°C for up to ten days. At 24 or 48 hour intervals, 5 mL samples of each culture were removed, spun down and resuspended in 5 mL of DM salts solution. Total viable counts of each culture were determined by plating 20 yuL drops of suitable dilutions onto drug-free nutrient agar. Mutagenesis to nalidixic acid resistance was determined by plating 20 ywL drops of the neat bacterial suspensions onto nutrient agar containing 16 yug mL * nalidixic acid (4 x MIC). A concentration of 16 pg mL * was selected since E. coli cultures plated on medium containing this concentration of drug produced only low numbers of resistant clones following ten days incubation (Results section 3.3). Plates for the determination of viable counts were incubated for one day before examination. Drug-containing plates were incubated at 37 °C for two days. Results were expressed as the number of nalidixic acid-resistant organisms per mL. 4.2.1. The effect of growth in sub-inhibitory concentrations of nalidixic acid on the development of nalidixic acid resistance in plasmid-free E, coli ABl 157 The total viable counts of drug-free cultures (Figure 25) or of cultures grown in the presence of 2 or 3 yUg mL * nalidixic acid (data not shown) reached approximately 1 X 10^ organisms per mL after one days incubation and remained at that level during the remaining nine days of the experiment. Growth in drug-free medium led to an increase of 2.4-fold in the number of nalidixic acid-resistant mutants recovered between days one and two, but the number of resistant mutants then remained essentially constant throughout the remaining eight days of the experiment (Figure 25). After two days incubation in medium containing 2 or 3 pg mL * of drug, the number of nalidixic acid resistant mutants recovered was 3.2- and 47.5-fold greater than from cultures grown in 171 Total viable counts 0) Q. •Ê 1 o>(O 2 1 V < 0> XX pg mL E 10 0 2 4 6 8 10 Incubation time (days) Figure 25. The effect of growth in nutrient broth containing 0, 2 or 3 ywg mL * nalidixic acid (NA) on the production of nalidixic acid-resistant (16 yUg mL *) mutants in cultures ofE. co/z AB1157. drug-free broth. Continued incubation in 2 or 3 /ug mL * nalidixic acid further increased the number of resistant mutants recovered. After ten days incubation in nutrient broth containing 2 jug mL * nalidixic acid, the number of resistant mutants was increased 20- fold compared to cultures incubated in drug-free broth. Similarly, the presence of 3 /ug mL * nalidixic acid led to a 588-fold increase, giving a count of 1.0 x 10"* resistant mutants per mL, after ten days incubation (Figure 25). 172 4.2.2. The effect of mutator plasmids on the development of nalidixic acid resistance 4 2.2.1. The effect of mutator plasmids on the development of nalidixic acid resistance in cultures grown in drug-free broth Even in the absence of nalidixic acid in the growth medium, the presence of a mutator plasmid led to an increase in the number of nalidixic acid-resistant mutants (Figure 26). After two days incubation, cultures of the plasmid-free and the strain carrying the non-mutator plasmid reached 23 and 15 resistant mutants per mL respectively, remaining essentially constant for the remaining eight days. However, the number of drug resistant mutants recovered from cultures containing the mutator plasmids R46 or R446b continued to increase for a further two or four days before reaching a plateau (Figure 26). Following ten days growth in drug-free nutrient broth, the 0) a. D>fO -V R446b C 2 1 R46 V < RP4 0 a> E 0 2 4 6 8 10 Incubation time (days) Figure 26. The effect of plasmids on spontaneous mutagenesis to nalidixic acid (NA) resistance (16 yUg mL"') in E. coli ABl 157 cultures grown in drug-free nutrient broth. 173 number of nalidixic acid-resistant mutants present per mL were 7- and 24-fold higher in strains ABl 157 (R46) and ABl 157 (R446b) respectively, compared with the plasmid- free control (Figure 26 and Table 17). 4.2.2 2. The effect of mutator plasmids on the development of nalidixic acid resistance in cultures grown in sub-inhibitory concentrations of nalidixic acid Presence of the non-mutator plasmid RP4 strain did not increase the number of nalidixic acid-resistant organisms recovered from broth cultures containing sub-inhibitory concentrations of nalidixic acid (Figures 27 and 30; Table 17). However, the presence of a mutator plasmid produced massive increases in the number of nalidixic acid resistant mutants recovered at every time interval, compared to the control, plasmid-free strain. After only two days incubation in broth containing 2ij,g mL ' nalidixic acid the presence of plasmids R46 and R446b increased the number of resistant mutants in the culture 87- and 16-fold respectively (Figures 28 and 29). Much greater increases were observed after two days incubation in broth containing 3 yUg mL ' nalidixic acid. After only two days, plasmid R46 increased the number of nalidixic acid resistant mutants by 204-fold, and plasmid R446b produced an increase of 288-fold after the same period of time (Figures 28 and 29). After ten days incubation in nutrient broth containing 2 yug mL ' nalidixic acid, the ABl 157 (R46) culture contained 2.3 x iC resistant mutants per mL, a 70-fold increase over the plasmid-free strain (Figure 28). The ABl 157 (R446b) culture contained 4.0 X 10^ resistant mutants per mL after the same period, a 1.2 x ICP-fbld increase over the plasmid-free strain (Figure 29). Incubation for ten days in broth containing 3 yug mL ' nalidixic acid led to an even greater proportion of the R46- and R446b-containing populations becoming resistant to nalidixic acid. Following ten days incubation, 8.0 x 10^ 174 ABl 157 (R46) organisms were resistant to nalidixic acid, accounting for 1 in 10 of the entire population and an increase of 8.0 x 10^-fold over the plasmid-free strain (Figure 28). The effect of plasmid R446b on the development of resistance to nalidixic acid in cultures grown in broth containing 3 yUg mL * nalidixic acid was even greater. 1.6 x 10^ ABl 157 (R446b) organisms, approximately 1 in 8 of the population, were resistant to nalidixic acid after ten days, an increase of 1.6 x 10"^-fold compared to the plasmid-free strain (Figure 29). Figure 30 summarises the data obtained from all strains when grown in nutrient broth containing 3 ywg mL * nalidixic acid. Table 17. Plasmid-mediated increases in the number of nalidixic acid-resistant mutants recovered after ten days incubation in nutrient broth containing 0, 2 or 3 yWg mL ' nalidixic acid. Results are expressed as ratios of mutants of the plasmid-harbouring strain compared with the plasmid-free strain. Cone, of nalidixic Plasmid-mediated increase in the number of nalidixic acid- acid in liquid resistant* mutants recovered medium (ywg m L') RP4 R46 R446b 0 1 . 2 7.1 24 2 0.7 70 1 . 2 X 1 0 ^ 3 0 . 6 8.0 X 10^ 1.6 X 10^ * resistant to 16^g mL'^ of nalidixic acid. 175 Total viable counts -o 0) CL E 'c o>fO o (Oc a> < V _ a E 10 0 2 4 6 8 10 Incubation time (days) Figure 27. The effect of growth in nutrient broth containing 0, 2 or 3 yUg mL ' nalidixic acid (NA) on the production of nalidixic acid-resistant (16 ywg m L') mutants in cultures OÏE. co//ABl 157 (RP4). Total viable counts 0) CL o> 2 jjg mL < 0 2 4 6 8 10 Incubation time (days) Figure 28. The effect of growth in nutrient broth containing 0, 2 or 3 yug mL ' nalidixic acid (NA) on the production of nalidixic acid-resistant (16 yUg mL ') mutants in cultures OÏE. co//ABl 157 (R46). 176 Total viable counts 0) Ql o> 0) E 10 0 2 4 6 8 10 Incubation time (days) Figure 29. The effect of growth in nutrient broth containing 0, 2 or 3 mL ’ nalidixic acid (NA) on the production of nalidixic acid-resistant (16 yUg mL"’) mutants in cultures OÏE. coli ABl 157 (R446b). 10® 0) Q. R446b R46 (O CD 0) R- RP4 < 0> _ o E 10 0 2 4 6 8 10 Incubation time (days) Figure 30. The effect of plasmids on mutagenesis to nalidixic acid (NA) resistance (16 yUg mL"’) in E. coli ABl 157 cultures grown in nutrient broth containing 3 yug mL"’ nalidixic acid. 177 4.3. Reversion to threonine independence following growth in liquid medium containing sub-inhibitory concentrations of nalidixic acid To determine if growth in liquid medium containing sub-inhibitory concentrations of nalidixic acid was specific to the induction of mutation to nalidixic acid resistance, reversion to threonine independence in nutrient broth cultures grown in sub-inhibitory concentrations of nalidixic acid was also determined. Adaptive reversion to threonine independence had previously been demonstrated on minimal medium lacking this amino acid (Results section 2.3; Figure 16). After ten days, 250 threonine independent revertants were observed per plate, an increase of 5.8-fold over the number of revertants observed on day two. Since broth-grown cultures contain excess threonine, no adaptive increase in the number of threonine independent revertants would be expected throughout the ten day incubation period in broth containing nalidixic acid. Samples removed from such cultures were, therefore, also tested for reversion to threonine independence by plating 20 yuL drops of the washed bacterial suspensions onto DM minimal medium lacking threonine. Control experiments showed that growth for ten days in drug-free nutrient broth produced no increase in the number of Thr^ revertants recovered and that the control, non-mutator plasmid RP4 did not increase mutation (Figure 31). After two days incubation in drug-free broth, R46 increased the number of Thr^ revertants recovered 1.9- fold, in contrast to an increase produced by R46 of 62.4-fold after two days incubation on DM medium lacking threonine (Table 10). Even after ten days incubation in drug-free broth, plasmid R46 only increased the number of Thr^ revertants 3.3-fold (Table 18). The presence of plasmid R446b led to a 26-fold increase, over the plasmid-free strain, in the 178 number of Thr^ revertants recovered after only two days (Figure 31), compared to only a 5.2-fold increase after two days incubation on DM medium lacking threonine (Table 10). After ten days incubation in drug-free broth the apparent effect of plasmid R446b was increased (Table 18), possibly because of a reduction in the number of Thr^ revertants in the R strain (Figure 31). Growth in medium containing sub-inhibitory concentrations of nalidixic acid increased Thr^ reversion in all strains (Figures 32 and 33), with the presence of a mutator plasmid increasing the effect (Table 18). During ten days growth in medium containing 2 /^g mL ' nalidixic acid, the number of revertants in the plasmid-less strain increased 1.9- fold, with plasmid RP4 having no effect (Figure 32). However, R46 and R446b enhanced reversion 3.6- and 25.4-fold respectively (Figure 32 and Table 18). Growth in broth containing 3 yUg mL ' nalidixic acid led to a 46.7-fold increase in the number of Thr ABl 157 revertants recovered after ten days, again not significantly affected by RP4 (Figure 33). After two days, the presence of R46 and R446b increased the number of Thr^ revertants 17.4- and 3.5-fold respectively (Figure 33), with much greater effects seen after ten days, with the plasmids increasing Thr^ reversion 20.0- and 50.9-fold respectively (Figure 33 and Table 18). 179 Total viable counts 0) Q. V -V R446b E 10 < 8 RP4 0 2 4 6 8 10 Incubation time (days) Figure 31. The effect of plasmids on spontaneous mutagenesis to threonine independence (Thr^) in E. coli ABl 157 thr-1 cultures grown in drug-free nutrient broth. Total viable counts 0) Q. 0> -V R446b . a E a R46 10 ^ R-/RP4 0 2 4 6 8 10 Incubation time (days) Figure 32. The effect of growth in nutrient broth containing 2 yUg mL ' nalidixic acid (NA) on the production of threonine-independent (Thr^) revertants in cultures ofE. coli ABl 157 thr-1 and plasmid-containing strains. 180 Total viable counts a> Q. R446b R46 RP4 Xi 0 2 4 6 8 10 Incubation time (days) Figure 33. The effect of growth in nutrient broth containing 3 ywg mL ’ nalidixic acid (NA) on the production of threonine-independent (Thr^) revertants in cultures ofE. coli ABl 157 thr-1 and plasmid-containing strains. Table 18. Number of Thr^ revertants recovered after ten days growth in liquid medium containing 0, 2 or 3 //g mL ’ nalidixic acid. The plasmid-mediated increases in the number of Thr^ revertants recovered after ten days are also shown. Cone, of nalidixic Number of Plasmid-mediated increase in the number acid in liquid AB1157 Thr^ of Thr^ revertants recovered medium revertants per mL RP4 R46 R446b (//g m L ’) 0 3 2.0 3.3 71.0 2 13 1.0 3.6 25.4 3 140 1.5 20.0 50.9 181 4.4. Summary and conclusions Large increases in the numbers of nalidixic acid-resistant mutants were observed during ten days incubation of E. coli cultures in liquid medium containing sub-inbibitory concentrations of the drug (Figure 25; Hinton and Pinney, 1997b). This effect occurred against a background in which the total viable counts of the cultures remained constant. The presence of a mutator plasmid such as R46, but particularly R446b, greatly increased the effect (Figures 28 and 29; Hinton and Pinney, 1997b). The non-mutator plasmid RP4 did not increase mutation frequencies (Table 17). Using reversion to threonine independence as a non-selected marker in these nalidixic acid-exposed plasmid-free cultures, it was shown that the number of Tbr^ revertants increased 4.3- and 47-fold following ten days growth in nutrient broth containing 2 or 3 yUg mL * nalidixic acid, further enhanced by the presence of a mutator plasmid (Figures 32 and 33; Table 18). The selection-induced increase in mutation to nalidixic acid resistance may be considered to be “adaptive” since mutation occurred in the absence of apparent cell division and no increase in mutagenesis was observed in unselected cultures. The moderate increase in mutagenesis at an unselected locus during nalidixic acid selection might be predicted as a result of the induction of error-prone repair. 182 5. Selection-induced mutation to nalidixic acid resistance in repair-deficient strains OÏE. coli 5.1. Introduction Previous results presented in this thesis have led to the suggestion that error-prone DNA repair might at least contribute to the mechanism of adaptive mutation to nalidixic acid resistance. Strains of E, coli deficient in individual DNA repair processes were therefore examined to test this hypothesis. It was predicted that strains deficient in error- prone repair would show reduced adaptive mutation to nalidixic acid resistance. This should be seen not only in strains deficient in umuDC but also in strains defective for SOS induction. For example, strains carrying defective recA orlexA genes, would be SOS uninducible and expected to show reduced frequencies of mutation to nalidixic acid resistance. A reduced mutation frequency to nalidixic acid resistance might also be expected in a strain carrying a defective recB gene. The latter encodes a subunit of exonuclease V, which is required for SOS induction by nalidixic acid (Gudas and Pardee, 1975; McPartland et al, 1980). As a control, it was predicted that a strain carrying a defective uvrA gene, which is deficient in error-free excision repair, might show increased frequencies of induced mutation since more damage could be processed via the error-prone repair pathway in the absence of excision repair. 5.2. Mutation to nalidixic acid resistance in repair-deficient strains ofE» coli The MICs of nalidixic acid for the DNA repair-deficient strains are shown in Table 19. Since 3 /^g mL * nalidixic acid, the sub-inhibitory concentration used to grow the wild-type repair-proficient ABl 157 strain, was supra-inhibitory for AB2463 recAlS and AB2470 recB21 (Table 19), it was decided to grow the repair-deficient strains in 0.75 183 times their respective MICs before plating on medium containing four times their respective MICs. The effect of plasmid R46 on the development of nalidixic acid resistance was also examined in each repair-deficient background (Table 21). The presence of the plasmid did not affect the MIC of any strain tested (results not shown). Selection-induced mutation frequencies to nalidixic acid resistance in the repair-deficient strains were determined after four days incubation. On continued incubation, the viability of strain AB2463 recA13 decreased, so affecting mutation frequencies and not allowing accurate comparisons to be made at later times. Table 19. Minimum inhibitory concentrations (MICs) of nalidixic acid (NA) for repair- deficient E. coli strains. Strain and relevant DNA repair phenotype NAMIC genotype (//g m L') ABl 157 Wild-type, fully repair-proficient 4 TK702 umuC36 SOS repair deficient 4 A B l886 uvrA6 Excision repair deficient 4 AB2463 recA13 Recombination repair deficient, SOS 2 induction deficient Recombination repair deficient, lacks AB2470 recB21 exonuclease V, SOS induction 2 deficient (by quinolones) AB2494 lexA3 SOS induction deficient, 5 recombination repair proficient E. coli TK702 is deficient in SOS error-prone DNA repair due to a mutation in umuC. Compared with the fully repair-proficient AB 1157 control strain, TK702 exhibited a 1000-fold decrease in mutation frequency to nalidixic acid resistance after growth in 0.75 X MIC of nalidixic acid for four days (Table 20). However, as predicted, the 184 introduction of plasmid R46 mucAB^ into this strain increased mutation 4.5 x 10^-fold, returning it to the level of the wild-type strain (Table 21). As predicted, the 5-fold increase in mutation frequency to nalidixic acid resistance in strain A Bl886 uvrA6 (Table 20) probably results from enhanced error-prone repair of DNA damage in the absence of error-free excision repair. Introduction of R46 into this strain gave an expected increase in mutation frequency of 36-fold, which was higher than that produced by the plasmid in the wild-type strain (Table 21). This increase could result from enhanced error-prone processing of DNA damage by both themucAB gene products Table 20. Mutation frequencies to nalidixic acid (NA) resistance (4 x MIC) in repair- deficient mutants of E. coli after four days growth in nutrient broth containing 0.75 x MIC nalidixic acid for each respective strain. Cone, of NA Cone, of NA Total viable Number of Mutation in growth on which count of NA"^ frequency Strain medium cultures cultures mutants toNA (/^gmL') plated after 4 days after 4 days resistance* (0.75 X MIC) (/^g mL-') incubation incubation (4 X MIC) ABl 157 3 16 1.3 X 10* 6.7 X 10* 5.3 X 10 * TK702 3 16 8.5 X 10’ 4.4 X 10’ 5.2 X 10* umuC36 ABl 886 3 16 2.0x10* 5.3 X 10* 2.7 X 10’ uvrA6 AB2470 1.5 8 8.5 X 10* 1.5 X 10* 1.8 X 10-* recB21 AB2494 3.75 20 4.9 x 10* 1.8 X 10* 3.8 X 10-* lexAS AB2463 1.5 8 6.7 X 10’ 7.1 X 10’ 1.1 X 10-* recAlS *Mutation frequencies are expressed as the number of nalidixic acid-resistant mutants per viable organism. 185 of R46 and the chromosomally-encoded UmuDC proteins. In line with the prediction, mutation frequencies to nalidixic acid resistance were reduced 2.9 x 10^- and 1.4 x 10^-fold respectively in the recB and lexA mutants compared to the repair-proficient strain (Table 20). The presence of therecAlS mutation in E. coli AB2463 led to a 5.0 x 10^-fold reduction in mutation (Table 20). The latter result was expected since this strain is completely defective in inducible SOS functions (Howard- Flanders and Theriot, 1966). Three results with plasmid-containing strains did not agree with the hypothesis that had been predicted. The presence of R46 in the lexA mutant reduced the mutation frequency 5-fold, whereas it should have no effect since the plasmid mucAB genes should Table 21. The effect of carriage of plasmid R46 on mutation frequencies to nalidixic acid (NA) resistance (4 x MIC) in repair-deficient mutants of E. coli after four days growth in nutrient broth containing 0.75 x MIC nalidixic acid for each respective strain. Mutation Mutation Fold increase (+) Strain frequency* to NA frequency* to NA or decrease (-) resistance in R resistance in R46- produced by strain containing strain presence of R46 ABl 157 5.3 X 10'^ 1 . 1 X 1 0 - 2 + 2 . 1 TK702 umuC36 5.2 X 10'^ 2.3 X 10-2 +4.5 X 10^ ABl 8 8 6 uvrA6 2.7 X 10-2 9.7 X 10-' +36 AB2470 recB21 1 . 8 X 1 0 ’^ 1.9 X 10-^ + 1 0 AB2494 lexAS 3.8 X 10'^ 7.9 X 10-2 -4.8 AB2463 recAlS 1 . 1 X 1 0 ' ^ 9.6 X 10-^ +8.7 X 102 *Mutation frequencies are expressed as the number of nalidixic acid-resistant mutants per viable organism. 186 be uninduced. A 10-fold increase in mutation frequency produced by R46 in the recB mutant was also unexpected, since SOS function should not be induced in this recB- deficient mutant. Similarly, the increase in mutation to nalidixic acid resistance in the recA strain of approximately 9 x 10^-fold was unexpected (Table 21). All these exceptions involve the plasmid-encoded mucAB operon, rather than expression of chromosomalumuDC genes, and it is possible that the R46-mediated nalidixic acid resistance-inducing function may bypass, at least to an extent, the requirement for RecA and RecB to initiate error-prone DNA repair. 5.3. Summary and conclusions The results presented in this chapter demonstrate the involvement of error-prone DNA repair on nalidixic acid-directed adaptive mutation to nalidixic acid resistance in E. coli. Mutation frequencies were greatly reduced in strains defective in SOS induction, such as AB2463 recAlS, AB2470 recB21 and AB2494 lexA3, or expression, such as TK702 umuC36 (Table 20). Introduction of plasmid R46, which codes for MucAB proteins that are homologous to UmuDC, into the UmuC-deficient strain returned its mutation frequency to the wild-type level (Table 21). These data are in agreement with the hypothesis that induction of the SOS response may be at least partly responsible for the development of mutations conferring nalidixic acid resistance in drug-exposed cultures. 187 6. Determination of quinolone MICs of quinolone-resistant mutants 6.1. Introduction Experiments described so far in this thesis have been performed with nalidixic acid as the selecting quinolone, with no attempt to determine the MICs of resistant clones. Although nalidixic acid is still used clinically, it would be of interest to investigate adaptive mutational resistance to a more modem quinolone. This chapter, therefore, reports experiments that determine minimum inhibitory concentrations to both nalidixic acid and ciprofloxacin of resistant clones selected after growth in sub-inhibitory concentrations of either quinolone. A representative number of these clones will then be subject to a second period of growth in broth containing concentrations corresponding to 0.75 X their resistant MICs, of either nalidixic acid or ciprofloxacin, and their MICs again determined. This will indicate whether the process of adaptive mutation can function to produce highly resistant and therefore clinically very significant mutants. It is then intended to use PGR amplification and Hinfl restriction of the QRDR of thegyrA gene of representative resistant mutants (see Results section 7) to reveal whether the genetic alterations that have occurred in response to adaptive quinolone-induced mutation are similar to those already identified in clinically-isolated, highly quinolone-resistant strains. 6.2. Mutation to ciprofloxacin resistance in cultures grown in sub-inhibitory concentrations of ciprofloxacin Minimum inhibitory concentrations of nalidixic acid and ciprofloxacin were determined for resistant clones that had been selected after growth in sub-inhibitory concentrations of either dmg. To obtain data for drug resistance levels of ciprofloxacin- 188 grown clones, cultures ofE. coli ABl 157 were grown for up to ten days in broth containing a sub-inhibitory concentration of ciprofloxacin and plated at 24 hour intervals onto medium containing an inhibitory concentration of ciprofloxacin. The MICs of ciprofloxacin againstE. coli strain AB 1157 and AB 1157 (R46) were determined to be 0.005 yUg mL'. 1.5 mL volumes of overnight nutrient broth cultures of ABl 157 or ABl 157 (R46) were diluted into 150 mL of drug-free nutrient broth or nutrient broth containing 0.00375 yug mL ' ciprofloxacin (0.75 x MIC) and incubated at 37°C. At 24 hour intervals, 5 mL samples were removed from each culture, washed and resuspended in 5 mL DM salts solution. Total viable counts were determined by plating 20 yuL drops of suitable dilutions onto drug-free nutrient agar. Ciprofloxacin-resistant mutants were selected by plating 20 yuL drops of the neat bacterial suspensions onto nutrient agar containing 0.02 yUg mL ' ciprofloxacin (4 x MIC). Plates for the determination of viable counts were incubated for one day before examination. Drug- containing plates were incubated at 37 °C for two days. Results were expressed as the number of ciprofloxacin-resistant organisms per mL of culture. After only one day’s incubation, in nutrient broth containing 0.75 x MIC of ciprofloxacin the number of ciprofloxacin-resistant ABl 157 mutants recovered was 46- fold higher, compared to the number recovered from drug-free broth (Figure 34). 189 a> Q. Total viable counts o>ra 0) Q. B- R46 R46 ja E 10 0 2 4 6 8 10 Incubation time (days) Figure 34. The effect of growth in nutrient broth containing 0 (dashed lines) or 0.75 x MIC (solid lines) ciprofloxacin (Cip) on the production of ciprofloxacin-resistant (4 x MIC) mutants in cultures of E. coli ABl 157, and the effect of plasmid R46 on the development of such mutants. However, unlike the results obtained after growth in 0.75 x MIC nalidixic acid (Figure 25), the number of ciprofloxacin-resistant mutants increased only 1.6-fold during the next nine days incubation (Figure 34). The presence of R46 increased the number of resistant clones recovered by 2.3-fold after one day, with, again, no significant increase over the next nine days (Figure 34). This plasmid-mediated effect was also evident in cultures grown in the absence of ciprofloxacin (Figure 34). Resistant clones were then selected from plates containing 4 x MIC ciprofloxacin or 4 X MIC nalidixic acid (see Results section 4.2) and their MICs to both nalidixic acid and ciprofloxacin determined. 190 6.3. Determination of quinolone MICs of the nalidixic acid- and ciprofloxacin- resistant mutants Five quinolone-resistant clones ofE. coli AB 1157 and of AB 1157 (R46), selected from cultures grown for four days in nutrient broth containing 0.75 x MIC of either nalidixic acid or ciprofloxacin and then plated on nutrient agar containing 4 x MIC of each respective drug, were tested to determine their MICs. Another five clones were tested from cultures that had grown for eight days in the presence of 0.75 x MIC of the drugs, and which had been subject to similar selection procedures. Overnight cultures grown in drug-free nutrient broth were plated directly onto nutrient agar containing 4 x MIC of drug to give control colonies for testing. Table 22 shows the MICs of nalidixic acid for the selected colonies. From 20 control clones plated on medium containing nalidixic acid, 11 showed MICs to nalidixic acid of > 512 yUg mL'\ Similarly, from the 20 control clones plated on medium containing ciprofloxacin, 10 had MICs to nalidixic acid of >512 yUg mL \ However, no control clone selected after growth on ciprofloxacin-containing medium exhibited a nalidixic acid MIC of 128 yUg mL'% compared with 4 control clones selected on medium containing nalidixic acid which did record this MIC. Of the 10 more sensitive control clones selected on ciprofloxacin, 6 had nalidixic acid MICs of 64 yUg mL *, and 4 gave MICs of 32 yUg mL *. Only a single control clone selected on medium containing nalidixic acid gave an MIC for this drug of 32 yUg mL'*, with four clones having nalidixic acid MICs of 64 yUg mL * (Table 22). With the exception of colonies selected after growth of the plasmidless ABl 157 strain for eight days in broth containing 0.75 x MIC nalidixic acid, clones selected after 191 growth in the presence of sub-inhibitory concentrations of nalidixic acid gave similar resistance patterns to control strains (Table 22). Four ABl 157 clones selected after eight days growth in nalidixic acid-containing broth had MICs for this drug of only 16 yUg mL'% the same concentration of drug on which the clones were selected. The fifth clone gave a nalidixic acid MIC of 32 ij,g mL"’. Clones of AB 1157 selected after four days in broth containing 0.75 x MIC nalidixic acid gave MICs of 64 ywg mL"’ (3 colonies) and ^512 /^g mL"’ (2 colonies). After four days growth in sub-inhibitory nalidixic acid, selected colonies of strain ABl 157 (R46) gave a similar pattern of nalidixic acid resistance as clones recovered after eight days growth in the same medium (Table 22). All 10 colonies of ABl 157 and ABl 157 (R46) selected after four days growth in a sub-inhibitory concentration of ciprofloxacin and plating on medium containing 4 x MIC ciprofloxacin, gave nalidixic acid MICs of <64 yUg mL"’, whereas 6 colonies selected after eight days growth in the same medium had nalidixic acid MICs of ^512 yUg mL"’, with 4 recording an MIC of 64 yWg mL "’ (Table 22). Table 23 shows the ciprofloxacin MICs of the same colonies whose nalidixic acid resistance levels are reported in Table 22. The clones of ABl 157 R" selected after 8 days growth in nalidixic acid and which gave lower nalidixic acid MICs than control clones, also exhibited lower ciprofloxacin MICs than the controls (Table 23). Control clones selected on nalidixic acid gave a similar pattern of ciprofloxacin resistance as control clones selected on ciprofloxacin. Of the 20 control clones growing on nalidixic acid- containing medium, 4 gave a ciprofloxacin MIC of 0.16 ywg mL"’, 13 an MIC of 0.08 yUg mL"’ and 3 a ciprofloxacin MIC of 0.04 yUg mL"’. Similarly, of the control clones selected on ciprofloxacin, 4 gave a ciprofloxacin MIC of 0.16 yUg mL"’, 10 an MIC of 0.08 yUg mL "’ and 6 an MIC of 0.04 yWg mL"’ (Table 23). Clones grown in a sub-inhibitory concentration 192 Table 22. Minimum inhibitory concentrations (MICs) of nalidixic acid for clones selected on nutrient agar containing nalidixic acid (4 x MIC) or ciprofloxacin (4 x MIC). Cultures ofEscherichia coli ABl 157 either harbouring or not harbouring plasmid R46, were grown for four or eight days in nutrient broth containing 0.75 x MIC of nalidixic acid or ciprofloxacin. Control clones were selected from colonies produced by plating drug-free cultures onto plates containing nalidixic acid (4 x MIC) or ciprofloxacin (4 x MIC). Figures refer to the number of clones with that particular MIC to nalidixic acid. Strain of Grown in Plated on Period of Nalidixic acid MIC (yUg mL ’) ABl 157 growth (days) 16 32 64 128 256 k512 R- - NA 1 2 2 5 R46 - NA 2 2 6 VO U) R NA NA 4 3 2 8 4 1 R46 NA NA 4 1 2 2 8 1 1 3 R - Cip 1 4 5 R46 - Cip 3 2 5 R Cip Cip 4 3 2 8 4 1 R46 Cip Cip 4 4 1 8 5 of ciprofloxacin before selection on ciprofloxacin generally exhibited higher ciprofloxacin resistance than clones selected from nalidixic acid-containing medium. Only one clone selected after nalidixic acid treatment had a ciprofloxacin MIC of 0.16 yUg mL % whereas 5 clones selected from medium containing ciprofloxacin had a ciprofloxacin MIC of 0.16fxg mL'\ No clone was isolated with a ciprofloxacin MIC of greater than 0.16 /^g mL \ No clone selected after exposure to ciprofloxacin exhibited a ciprofloxacin MIC of less than 0.04 /^g mL ', whereas almost half (8 out of 20) of the clones selected from nalidixic acid-containing medium gave ciprofloxacin MICs <0.02 yUgmL ' (Table 23). Table 24 compares the MICs of nalidixic acid and ciprofloxacin for all the clones described in Tables 22 and 23. The results clearly show that colonies exhibiting high nalidixic acid MICs acid also had high ciprofloxacin MICs, and vice versa. Almost half of all the colonies tested (34 from 80) gave nalidixic acid MICs of 512 yUg mL ' or greater (^ 128 x MIC): all 34 had ciprofloxacin MICs of either 0.08 or 0.16 yUg mL ' (16 or 32 X MIC) (Table 24). The data in Table 22 show that clones with high levels of resistance to nalidixic acid (^ 128 X MIC) were readily selected from both control and drug-exposed cultures, regardless of whether nalidixic acid or ciprofloxacin was used for selection. Of the 80 clones reported in Table 22, 34 showed a nalidixic acid MIC of >512 yUg mL ', an increase of 128-fold or greater over the wild-type MIC. These same clones did not exhibit such high levels of resistance to ciprofloxacin, only 14 out of 80 gave ciprofloxacin MICs of 0.16 ywg mL ' (32 x MIC), with none giving greater than this. 194 Table 23. Minimum inhibitory concentrations (MICs) of ciprofloxacin for clones selected on nutrient agar containing nalidixic acid (4 x MIC) or ciprofloxacin (4 x MIC). Cultures ofEscherichia coli ABl 157 either harbouring or not harbouring plasmid R46, were grown for four or eight days in nutrient broth containing 0.75 x MIC of nalidixic acid or ciprofloxacin. Control clones were selected from colonies produced by plating drug-free cultures onto plates containing nalidixic acid (4 x MIC) or ciprofloxacin (4 x MIC). Figures refer to the number of clones with that particular MIC to ciprofloxacin. Strain of Grown in Plated on Period of Ciprofloxacin MIC (yUg mL ') ABl 157 growth (days) 0.005 0.01 0.02 0.04 0.08 0.16 R- - NA 3 6 1 R46 - NA 7 3 VO LA RNANA 4 1 2 2 8 1 1 3 R46 NA NA 4 1 1 3 8 1 1 2 1 R - Cip 4 6 R46 - Cip 2 4 4 R Cip Cip 4 1 3 1 8 4 1 R46 Cip Cip 4 1 4 8 2 3 Table 24. Comparison of the minimum inhibitory concentrations (MICs) to nalidixic acid and ciprofloxacin of the clones described in Tables 22 and 23. Figures refer to numbers of clones. Nalidixic acid MIC (Mg mL ') 16 32 64 128 256 >512 0.005 1 VO o \ 0.01 1 Ciprofloxacin 0.02 3 1 4 1 MIC (Mg m L') 0.04 1 15 1 0.08 9 3 4 23 0.16 1 1 11 These results are in agreement with the currently accepted concept that one mutation in E. coli, commonly in the QRDR of thegyrA gene, is required to produce a high level of resistance to nalidixic acid (MIC >128 /^g m L’) (Vila et al, 1994) and a low level of resistance to fluoroquinolones (ciprofloxacin MIC < 1 yUg mL ’) (Yoshidaet al, 1988; Oram and Fisher, 1991). A second mutation in gyrA is necessary for the development of high-level resistance to fluoroquinolones, resulting in ciprofloxacin MICs of >2 yUg mL ’ (Everett et al, 1996), >8 yUg mL ’ (Vila et al, 1994) or 128 yUg mL ’ (Heisig et al, 1993). The data presented in Tables 22 and 23 therefore suggest that the selected clones carry only one mutation, leading to a high level of resistance to nalidixic acid but only a low-level resistance to ciprofloxacin. The variation in the nalidixic acid resistance levels of the selected clones (Table 22), suggests that the mutational site within the E. coli chromosome may vary. High-level resistance to nalidixic acid (> 128 yUg mL'’) results from a single mutation in the QRDR of thegyrA gene, altering the serine amino acid at position 83 of the GyrA protein (Vilaet al, 1994). Whereas, those clones showing only a moderate increase in nalidixic acid MIC may have a wild-type GyrA protein, but exhibit altered expression of outer membrane porins, such as OmpF (Hooperet al, 1986; 1992), induction of the marRAB operon (Cohenet al, 1989; 1993c), or mutation ingyrB (Yoshida 1991). It was, therefore, predicted that if a low-level ciprofloxacin resistant mutant was given a second period of growth in nutrient broth containing a concentration of quinolone approaching its MIC, the resulting culture would contain high-level ciprofloxacin resistant mutants. 197 6.4. Growth of low-level quinolone resistant clones in nutrient broth containing sub- inhibitory concentrations of nalidixic acid or ciprofloxacin selects for high level quinolone-resistant mutants Table 25 summarises the experimental design. Three clones of ABl 157 that exhibited varying nalidixic acid MICs were subcultured into nutrient broth containing 0.75 X their resistant nalidixic acid MICs. These clones had all been selected on medium containing nalidixic acid. One of the clones had been grown in drug-free broth before plating on the selecting medium, the second clone had been grown for four days and the third clone for eight days in broth containing a sub-inhibitory concentration of nalidixic acid before plating on the selecting medium. Three ABl 157 clones were selected from ciprofloxacin-containing medium according to the pattern described above and subcultured into nutrient broth containing 0.75 x their resistant ciprofloxacin MICs. Samples of all cultures were plated after four and eight days onto nutrient agar containing 4 X the respective resistant nalidixic acid or ciprofloxacin MIC (Table 25). Six resistant clones of ABl 157 (R46) were selected and tested in the same manner (Table 25). 198 Table 25. Plan of experiment to grow clones, previously selected as described in Tables 22 and 23, in the presence of higher concentrations of quinolones. Organism Cone, of drug MIC of clone Grown in Plated on (yUg mL ’) in broth (yUg m L ’) (yUg m L ’) (yUg m L ’) before plating on (0.75 X MIC) (4 X MIC) selecting medium 0 NA 16 NA 12 NA 64 ABl 157 (R-) NA 3 (4 days) NA 64 NA 48 NA 256 NA 3 (8 days) NA128 NA 96 NA512 0 NA128 NA96 N A 5I2 ABl 157 (R46) NA 3 (4 days) NA 128 NA96 NA512 NA 3 (8 days) NA128 NA 96 NA512 0 Cip 0.04 Cip 0.03 Cip 0.16 ABl 157 (R-) Cip 0.005 (4 days) Cip 0.08 Cip 0.06 Cip 0.32 Cip 0.005 (8 days) Cip 0.08 Cip 0.06 Cip 0.32 0 Cip 0.08 Cip 0.06 Cip 0.32 ABl 157 (R46) Cip 0.005 (4 days) Cip 0.08 Cip 0.06 Cip 0.32 Cip 0.005 (8 days) Cip 0.16 Cip 0.12 Cip 0.64 Due to problems associated with the solubility of nalidixic acid, it was not possible to quantify nalidixic acid MICs greater than 1024 yUg mL ’ (256 x wild-type MIC). Clones with only a relatively high MIC of nalidixic acid (128 yUg mL ’) were therefore used to determine how a second exposure to this drug might affect resistance. This allowed medium containing 512 yUg mL ’ nalidixic acid (4 x resistant MIC) to be used to select resistant clones. Five colonies were tested after 4 or 8 days growth on selecting plates from each treatment outlined in Table 25 (a total of 120 colonies) and their MICs determined (Tables 26 and 27). 199 The results in Table 26 clearly show that a second round of growth in a sub- inhibitory concentration of either nalidixic acid or ciprofloxacin selects clones with high- level MICs to nalidixic acid. 101 of 120 colonies, selected after growth of cultures in 0.75 X their respective resistant nalidixic acid or ciprofloxacin MICs, gave MICs to nalidixic acid of 1024 fÀg mL ' or greater. This is an increase of at least 256-fold over the wild-type MIC of 4 jjLg mL '. The initial level of resistance (16-128 mL ' ), the presence of plasmid R46 or the growth period in the sub-inhibitory concentration of quinolone (four or eight days) did not affect the pattern of results (Table 26). Similar data were obtained from clones selected initially on medium containing ciprofloxacin. With the exception of an AB1157 clone with an initial ciprofloxacin MIC of 0.08 yUg mL', which gave a total of 7 clones with nalidixic acid MICs of < 128 yUg mL ', high-level nalidixic acid resistance was exhibited by all clones irrespective of whether their initial selection had been after growth in nalidixic acid or ciprofloxacin. When the same clones described in Table 26 were tested for resistance to ciprofloxacin, it was found that these MICs were also increased by a second period of growth in nutrient broth containing 0.75 x the resistant MIC of quinolone (Table 27). However, a difference can be seen between those colonies previously selected after growth in medium containing a sub-inhibitory concentration of nalidixic acid and plated on nutrient agar containing nalidixic acid and those selected after similar treatment in media containing ciprofloxacin. No colony initially selected fi*om nalidixic acid plates, regrown in 0.75 x their resistant nalidixic acid MIC and selected again on nalidixic acid- containing medium had an MIC to ciprofloxacin of >1.28 yug mL ' (256 x MIC). Whereas, 21 out of 60 colonies selected fi'om ciprofloxacin plates and regrown in ciprofloxacin had ciprofloxacin MICs of >1.28 yUg mL ' (Table 27). The highest MIC 200 Table 26. Minimum inhibitory concentrations (MICs) of nalidixic acid for clones selected on nutrient agar containing nalidixic acid (4 x the resistant MIC) or ciprofloxacin (4 x the resistant MIC). Cultures ofE. coli ABl 157 either harbouring or not harbouring plasmid R46, were grown for four or eight days in nutrient broth containing 0.75 x the resistant MIC of nalidixic acid or ciprofloxacin. Figures refer to the number of clones with that particular MIC to nalidixic acid. Organism Initial Period of Nalidixic acid MIC (/^g mL*’) MIC growth C^g mL’’) (days) 32 64 128 256 512 ^1024 NA 16 4 5 8 5 ABl 157 (R-) NA 64 4 5 8 5 NA 128 4 5 8 5 NA 128 4 5 8 5 ABl 157 (R46) NA 128 4 2 3 8 5 NA 128 4 5 8 5 Cip 0.04 4 5 8 5 ABl 157 (R-) Cip 0.08 8* 1 8 1 Cip 0.08 4 3 2 8 2 1 2 Cip 0.08 4 5 8 5 AB1157 (R46) Cip 0.08 4 5 8 5 Cip 0.16 4 5 8 5 * No colonies were detected after four days growth in 0.75 x resistant MIC of ciprofloxacin. Therefore 10 colonies were selected after eight days growth. 201 Table 27. Minimum inhibitory concentrations (MICs) of ciprofloxacin for clones selected on nutrient agar containing nalidixic acid (4 x the resistant MIC) or ciprofloxacin (4 x the resistant MIC). Cultures ofE. coli ABl 157 either harbouring or not harbouring plasmid R46, were grown for four or eight days in nutrient broth containing 0.75 x the resistant MIC of nalidixic acid or ciprofloxacin. Figures refer to the number of clones with that particular MIC to ciprofloxacin. Organis Initial Period of Ciprofloxacin MIC (/^g mL ') m MIC growth (yUg m L') (days) 0. 0. 0. 0. 0. 0. 1. 2. 5. 02 04 08 16 32 64 28 56 12 NA 4 2 3 16 8 4 1 AB1157 (R-) NA 4 1 2 2 64 8 5 NA 4 5 128 8 2 3 NA 4 5 128 8 5 AB1157 (R46) NA 4 5 128 8 5 NA 4 3 2 128 8 5 Cip 4 3 1 1 0.04 8 3 2 ABl 157 (R-) Cip 4 4 1 0.08 8 1 4 Cip 0.08 8* 10 Cip 4 3 2 0.08 8 1 4 AB1157 (R46) Cip 4 2 3 0.08 8 5 Cip 4 3 2 0.16 8 3 1 1 * No colonies were detected after four days growth in 0.75 x resistant MIC of ciprofloxacin. Therefore 10 colonies were selected after eight days growth. 202 detected was 5.12 /^g mL *, an increase of 1 x 10^-fold over the wild-type MIC of 0.005 jjLg mL * ciprofloxacin. Table 28 compares the MICs of ciprofloxacin and nalidixic acid of the clones described in Tables 26 and 27. These comparisons suggest that the development of high- level resistance to nalidixic acid (>1024 mL *, equivalent to >256 x MIC) does not necessarily correlate with high levels of resistance to ciprofloxacin. 81 of the 101 colonies displaying a nalidixic acid MIC of > 1024 yUg mL *, gave ciprofloxacin MICs of 0.64 ywg mL * (128 X MIC) or less (Table 28). High-level resistance to ciprofloxacin (> 1.28 mL *, equivalent to 256 x MIC) was only observed in clones that had been selected after growth in broth containing 0.75 X the resistant MIC of ciprofloxacin before plating on 4 x the resistant MIC of ciprofloxacin. 21 of the 60 ABl 157 and ABl 157 (R46) colonies initially selected from ciprofloxacin plates showed MICs for ciprofloxacin of > 1.28 ywg mL * (Table 27). All but one of these 21 clones had a nalidixic acid MIC of >1024 yUg mL * (Table 28). Whereas, all 60 colonies selected from nalidixic acid plates had ciprofloxacin MICs of <0.64 ywg mL* (Table 27): 58 of these gave nalidixic acid MICs of >1024 yWg mL* (Table 26). No colony selected on ciprofloxacin plates after growth in the presence of 0.75 x the resistant MIC of ciprofloxacin gave MICs to ciprofloxacin of <0.16 yUg mL * (Table 27), whereas almost half (29 out of 60) of the colonies selected after growth in nalidixic acid containing medium had ciprofloxacin MICs of <0.16 yUg mL * (Table 27). 203 Table 28. Comparison of the minimum inhibitory concentrations (MICs) to nalidixic acid (NA) and ciprofloxacin (Cip) of the clones described in Tables 26 and 27. Figures refer to numbers of clones. NA MIC (yUg mL ') 32 64 128 256 512 >1024 0.02 4 0.04 5 0.08 2 18 0.16 16 Cip MIC 0.32 24 (yUg m L') 0.64 5 1 1 9 14 1.28 1 15 2.56 3 5.12 2 6.5. Summary and conclusions Mutant clones exhibiting high-level nalidixic acid resistance (MIC ^512 yUg mL ', equivalent to ^128 x wild-type MIC) and low level resistance to ciprofloxacin (MIC <0.16 yUg mL ', equivalent to <32 x wild-type MIC) were selected after a single period of growth in sub-inhibitory concentrations of either nalidixic acid or ciprofloxacin (Tables 22 and 23). Such increased resistance can be produced by a single mutation in the QRDR within gyrA o f E. coli (Vila et al, 1994). Table 26 shows that a second period of growth in nutrient broth containing 0.75 x the respective “resistant” MICs of clones selected after a one step treatment led to 101 of 120 selected colonies having an MIC to nalidixic acid of > 1024 yUg mL ', irrespective of whether the initial resistant clone was selected on 64 yUg mL ' or 512 yUg mL ' of nalidixic acid. In these colonies it is possible 204 that a second mutation has occurred within gyrA to increase the level of resistance further (Heisigeta/, 1993). 21 of 60 colonies selected from cultures grown for a second time in sub-inhibitory concentrations of ciprofloxacin gave MICs to ciprofloxacin of ^1.28jug mL'’ (Table 27). A double mutation in gyrA alone has previously been shown to confer a ciprofloxacin MIC of 0.5 jug mL'’ (Truong et al, 1995). However others have shown that a single mutation within gyrA, replacing serine with leucine at position 83, is sufficient to produce ciprofloxacin MICs of ^ I yUg mL ’ (Vilaet al, 1994) or 2 yUg mL ’ (Everettet al, 1996). In the next chapter the QRDRs within the gyrA genes of a selection of the mutants reported in Tables 25 to 27, which are resistant to varying concentrations of nalidixic acid and ciprofloxacin will be determined using the Polymerase Chain Reaction (PCR). This will reveal any mutations that have occurred within the QRDRs, which may explain the observed resistance patterns, and will give insight into the mechanisms of quinolone resistance in E. coli cultures exposed, either singly or repeatedly, to sub-inhibitory quinolone concentrations. 205 7. Determination of mutations within QRDRs of quinolone-resistant clones 7.1. Introduction Mutation to quinolone resistance inE. coli requires only one base-pair change in the QRDR of thegyrA gene, but the mutation must be at one of only a few specific sites (see Introduction section 2.6.1). A change of the amino acid serine at position 83 of GyrA to a hydrophobic amino acid, such as leucine or tryptophan, is sufficient to produce a high level of resistance to nalidixic acid (MIC > 128 yUg mL ’) (Vila et al, 1994) and low-level resistance to fiuoroquinolones, i.e. MIC of ciprofloxacin < 1 yUg L ’ (Yoshidaet al, 1988; Oram and Fisher, 1991). An additional mutation in gyrA causing a change in aspartate at position 87 of the GyrA protein to a non-negatively charged amino acid, such as asparagine, glycine or tyrosine, confers high level fluoroquinolone resistance i.e. MIC of ciprofloxacin ^2 yUg mL ’ (Everettet al, 1996), >8 yUg mL’ (Vila et al, 1994), 128 yUg mL ’ (Heisig et al, 1993). Mutations in genes coding for the secondary quinolone target, topoisomerase IV, also alter susceptibilty, in some cases increasing MICs to high-level fluoroquinolone resistance (see Introduction section 2.6.3). 7.2. Detection of mutations in the QRDR ofgyrA in quinolone-resistant E, coli Eleven clones ofE. coli ABl 157 with varying levels of resistance to both nalidixic acid and ciprofloxacin were selected fi’om those described in Tables 26, 27 and 28. The selection procedures and MICs of each strain to both drugs are summarised in Table 29. The Polymerase Chain Reaction (PCR) was used to amplify the QRDR regions of the gyrA genes of each strain. These were then subjected to Hinfl. restriction and the DNA sequenced to determine if any point mutations had occurred. 206 Table 29. Minimum inhibitory concentrations (MICs) to nalidixic acid and ciprofloxacin of the investigated E. coli strains. The lane numbers (see Figure 36) for each amplified PCR product and each Hinfi. digested PCR product are also shown. Cone, of MICs Lane number drug in Cone, of drug (Adg m L') (see Figure 36) Strain growth in selecting number medium medium PCR Hinfl (Atg m L') (/^g m L') NA Cip product restricted product NAI 0 0 4 0.005 5 6 (wild-type) NA2 0 0 4 0.005 7 8 (wild-type) NA3 NA3 NA 16 16 0.02 9 10 NA4 Cip 0.005 Cip 0.02 16 0.02 11 12 NA5 Cip 0.06 Cip 0.32 32 0.64 13 14 NA6 NA 48 NA256 64 0.04 15 16 NA7 NA 48 NA 256 64 0.04 17 18 NAS NA 12 NA 64 1024 0.02 24 25 NA9 Cip 0.12 Cip 0.64 1024 5.12 26 27 NAIO NA 48 NA 256 1024 0.04 28 29 N A ll Cip 0.12 Cip 0.64 1024 2.56 30 31 The size of the E. coli K-12 gyrA gene is known to be 2.6 kilobases (kb) (Swanberg and Wang, 1987). The primers used (see Methods section 10) allowed amplification of the +1 to +620 bp region of the gene, which contains the QRDR. Fragments produced following PCR amplification of all quinolone-resistant clones corresponded to the expected fragment size (Figure 36). The successful PCR amplification of fragments of the QRDR regions of the 9 quinolone-resistant clones and 2 wild-type controls allowed Hinfl. restriction and DNA sequencing. It was hoped that 207 these genetic analyses would reveal sequence alterations that may have occurred, especially within the QRDR between amino acids 67 and 106. 7.2.1. Restriction fragment length polymorphism (RFLP) of the QRDR ofgyrA of quinolone-resistant E. co/f ABl 157 clones Restriction fragment length polymorphism (RFLP) was determined for the amplified gyrA fragments of the resistant E. coli ABl 157 clones using restriction endonuclease Hinfi. This recognises and cleaves the sequence GANTC (Figure 35), and was used to screen for base substitutions in the amplified gyrA fragments, specifically around the coding sequences for aspartate-82 and serine-83 (5'-GACTCG-3'; Figure 36) and aspartate-115 and serine-116 (5'-GACTCT-3'; Figure 36). A PCR fragment of gyrA from a quinolone-sensitive strain with no base substitution at serine-83 (nucleotide sequence encoding amino acids 82 and 83 being 5'-GACTCG-3') would be cleaved twice by Hinfi resulting in three fragments of length 99 bp, 246 bp and 275 bp. Certain commongyrA mutations in E. coli occurring within a region of the enzyme centered around serine-83 (Fisher et al, 1989), result in the abolition of aHinfi restriction site in gyrA. A gyrA PCR fragment of a quinolone-resistant strain ofE. coli, containing a mutation at serine-83, would therefore only yield two fragments of length 275 bp and 345 bp, following restriction treatment withHinfi, due to the loss of aHinfi recognition site in the nucleotide sequence of amino acids 82 and 83. The amplified PCR and Hinfi restricted products of the eleven selected strains are shown in Figure 36. 208 Figure 35. DNA sequence of the Hinfl. restriction endonuclease recognition site within gyrA, which codes for aspartate-82 and serine-83. The points of cleavage are indicated by the arrows. V 5' GACTCG 3' 3' CTGAGC 5' À The amplified, undigested PCR product can clearly be seen in Figure 36, lanes 5, 7, 9, 11, 13, 15, 17,24, 26,28 and 30 producing a bright band at approximately 620 bp. This confirmed that amplification of the desired region ofgyrA has been successful for all strains examined. Restriction fragments produced by Hinfl digestion of the PCR product from two samples of the wild-type, quinolone-sensitive ABl 157 strain are shown in lanes 6 and 8 (Figure 36). As expected, the PCR product from the wild-type strain was cleaved twice due to the presence of two Hinfl recognition sites, revealing three bands in lanes 6 and 8, corresponding to the expected three fragments of sizes 99 bp, 246 bp and 275 bp. All but three of the quinolone-resistant isolates studied yielded the same pattern of products as the wild-type following restriction withHinfl restriction endonuclease: NA3, NA4, NA5, NA6 and NA7 all yielded three bands of 99 bp, 246 bp and 275 bp, indicating that no mutation had occurred at serine-83 to abolish the Hinfl recognition site. It can be seen from Table 29 that these five isolates exhibited relatively low levels of resistance to nalidixic acid (< 64 /^g mL'*). Strain NAI 0, which gave high-level resistance to nalidixic acid and low-level resistance to ciprofloxacin (MICs of > 1024 and 0.04 yUg mL'*, respectively) also yielded three bands, with lengths of 99 bp, 246 bp and 275 bp (a 209 Figure 36. Restriction fragment length polymorphisms and undigested PCR products in the amplified 620 bp region oigyrA in selected E. coli ABl 157 strains sensitive or with a decreased susceptibility to nalidixic acid and ciprofloxacin. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 a,------— — kk : 300 bp 200 bp 100 bp 200 bp 100 bp 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 A description of each lane is shown in Table 29. In addition, a 100 bp ladder was run in Lanes 2 and 23. very faint band could be seen on the gel in lane 29, corresponding to a fragment length of 99 bp), indicating the absence of mutation at serine-83. However strains NAS, NA 9 and NAI 1 revealed a pattern of restriction which differed from the others. Only two bands were produced following Hinfl restriction of NAS, NA9 and NAI 1 indicating that a mutation had occurred at one of theHinfl recognition sites withingyrA (Figure 36). 210 That the two bands on the gel correspond to fragments of lengths 275 bp and 345 bp indicated that the lost Hinfi recognition site was at serine-83. If the lost restriction recognition site had occurred at aspartate-115 and serine-116, fragments of size 246 bp and 374 bp would have been produced. Since Hinfi restriction of strains NA3, NA4, NA5, NA6 and NA7 suggested no alteration in the region of Ser-83, they were not selected for further sequencing. However, strains NA8, NA9, NAIO and NAI 1 were all high-level resistant mutants to nalidixic acid (MICs ^1024 /ig mL'^), with strains NA9 and N A ll also exhibiting high-level resistance to ciprofloxacin (MICs >2.56 /^g mL’^). Strains NA8, NA9 and N A ll, which had lost the Hinfi restriction site around Ser-83, were sequenced to determine the change in base sequence that had occurred. Strain NAIO still retained the Ser-83 restriction site, but displayed high-level resistance to nalidixic acid (MIC k 1024 yUg mL ') with only low level ciprofloxacin resistance (MIC 0.04 yUg mL '). It was, therefore, sequenced to reveal other possible mutations within its QRDR. 7.2.2. DNA sequencing of the amplified 620 bp region ofgyrA incorporating the QRDR OÎE, coli AB1157 The amplified 620 bp region of gyrA, incorporating the QRDR, of the quinolone- sensitive, wild-type E. coli ABl 157 strain NAI (Table 29) is shown in Figure 37. This sequence agrees with that published by Yoshida et al, 1988. Amplified 620 bp fragments of the four selected quinolone-resistant mutants were identical to the sequence of the control quinolone-sensitive strain, except for the changes shown in Table 30. Three of the four resistant strains examined carried mutations that altered the amino acid at serine-83 (Table 30). Strains NA9 and NAI 1, both of which showed high- 211 Figure 37. The DNA and amino acid sequence of a 620 bp region of E. coli AB 1157 gyrA (amino acids 1 to 207) incorporating the QRDR. The nucleotide sequence of thegyrA gene is presented from the 5' (left) to 3' (right) end. Amino acid positions are numbered. The sequence agrees with that published by Yoshida et al (1988). ATG AGC GAG CTT GCG AGA GAA ATT ACA CCG GTC AAC ATT GAG GAA GAG CTG AAG Met Ser Asp Leu Ala Arg Glu lie Thr Pro Val Asn He Glu Glu Glu Leu Lys 1 5 10 15 AGC TCC TAT CTG GAT TAT GCG ATG TCG GTC ATT GTT GGC CGT GCG CTG CCA GAT Ser Ser Tyr Leu Asp Tyr Ala Met Ser Val He Val Gly Arg Ala Leu Pro Asp 20 25 30 35 GTC CGA GAT GGC CTG AAG CCG GTA CAC CGT CGC GTA CTT TAG GCC ATG AAC GTA Val Arg Asp Gly Leu Lys Pro Val His Arg Arg Val Leu Tyr Ala Met Asn Val 40 45 50 CTA GGC AAT GAG TGG AAC AAA GCC TAT AAA AAA TCT GCC CGT GTC GTT GGT GAG Leu Gly Asn Asp Trp Asn Lys Ala Tyr Lys Lys Ser Ala Arg Val Val Gly Asp 55 60 65 70 GTA ATC GGT AAA TAG CAT CGC CAT GGT GAG TCG GCG GTC TAT GAG ACG ATC GTC Val He Gly Lys Tyr His Pro His Gly Asp Ser Ala Val Tyr Asp Thr He Val 75 80 85 90 CGC ATG GCG GAG CCA TTC TCG CTG CGT TAT ATG CTG GTA GAG GGT GAG GGT AAC Arg Met Ala Gin Pro Phe Ser Leu Arg Tyr Met Leu Val Asp Gly Gin Gly Asn 95 100 105 TTC GGT TCT ATC GAG GGC GAG TCT GCG GCG GCA ATG CGT TAT ACG GAA ATC CGT Phe Gly Ser He Asp Gly Asp Ser Ala Ala Ala Met Arg Tyr Thr Glu He Arg 110 115 120 125 CTG GCG AAA ATT GCC CAT GAA CTG ATG GCC GAT CTC GAA AAA GAG ACG GTC GAT Leu Ala Lys He Ala His Glu Leu Met Ala Asp Leu Glu Lys Glu Thr Val Asp 130 135 140 TTC GTT GAT AAC TAT GAC GGC ACG GAA AAA ATT CCG GAC GTC ATG CCA ACC AAA Phe Val Asp Thr Tyr Asp Gly Thr Glu Lys He Pro Asp Val Met Pro Thr Lys 145 150 155 160 ATT CCT AAC CTG CTG GTG AAC GGT TCT TCC GGT ATC GCC GTA GGT ATG GCA ACC He Pro Asn Leu Leu Val Asn Gly Ser Ser Gly He Ala Val Gly Met Ala Thr 165 170 175 180 AAC ATC CCG CCG CAC AAC CTG ACG GAA GTC ATC AAC GGT TGT CTG GCG TAT ATT Asn He Pro Pro His Asn Leu Thr Glu Val He Asn Gly Cys Leu Ala Tyr He 185 190 195 GAT GAT GAA GAC ATC AGC ATT GAA GGG Asp Asp Glu Asp He Ser He Glu Gly 200 205 212 levels of resistance to both nalidixic acid (MICs of >1024 yUg mL *) and ciprofloxacin (MICs of 5.12 and 2.56 yUg mL * respectively), revealed C - G transversions, which changed serine-83 to tryptophan. Strain NAS was also mutant at position 83, but in this case, a C - T transition mutation altered the amino acid from serine to leucine. Strain NAIO was wild-type at position 83, but revealed a mutation at position 87, an A - G transition changing aspartate to glycine. Since strain NAIO still retained the GACTCG sequence at position 82 and 83, this explained why it yielded three fragments after Hinfl. restriction analysis (Figure 36). Table 30. DNA base sequence alterations of E. coli strains described in Table 29 and Figure 36. Base changes resulting in amino acid substitutions are highlighted in bold and underlined. Strain number MICs mL'*) Nucleotide Amino acid change change Nalidixic acid Ciprofloxacin NAI 4 0.005 NA8 >1024 0.02 TCG - TTG Ser83 - Leu NA9 >1024 5.12 TCG - TGG Ser83 ^ Try NAIO >1024 0.04 GAC -> GGC Asp87 - Gly N A ll >1024 2.56 TCG - TGG Ser83 - Try 7.3. Summary and conclusions Hinfl endonuclease digestion of nine quinolone-resistant isolates revealed that three had lost a Hinfl recognition site in the region coding for aspartate-82 and serine-83. A fourth resistant isolate, (NAIO), revealed a pattern of restriction fragments identical to the wild-type, indicating that no base substitution had occurred at either///«/I recognition 213 site. All four strains exhibited a high level of resistance to nalidixic acid (MIC >1024 yUg mL '). However their MICs to ciprofloxacin ranged from 0.02 yUg mL' (4 x wild-type MIC) to 5.12 yUg mL ' (1024 x wild-type MIC). Sequencing of the QRDRs of these four strains revealed alterations in base sequence (Table 30). Three of the four strains carried an amino acid substitution at serine-83, whereas the fourth, (NAIO), contained a base alteration at aspartate-87. Retention of the wild-type base sequence at seine-83 explains why Hinfl. digestion of strain NAIO resulted in three DNA fragments (Table 30). Mutation at Ser-83 is frequently reported in clinically-isolated quinolone-resistantE. coli strains (Yoshida et al, 1988; 1990; Cullen et al, 1989; Fisher et al, 1989; Oram and Fisher, 1991). 214 DISCUSSION Resistance to nalidixic acid and other quinolone antibacterial agents in E. coli essentially results jfrom mutations within chromosomal genes. Alterations at a number of genetic loci can confer such resistance, including mutations within genes coding for DNA gyrase and topoisomerase IV, or in genes affecting drug uptake (Yoshida et al, 1990; Hooper and Wolfson, 1991; Yamagishiet al, 1996; Vila et al, 1996; Heisig, 1996; Kumagai et al, 1996; Okusu et al, 1996). In E. coli, the gyrA gene, which encodes subunit A of DNA gyrase, is the major target for mutation, specifically within the quinolone resistance-determining region (QRDR) encoding amino acids 67 to 106 (Yoshida et al, 1990). The Ser-83 residue of GyrA is the most frequently mutated site (Cullen et al, 1989; Yoshida etal, 1990; Oram and Fisher, 1991; Maxwell, 1992). Quinolone-resistant clinical isolates ofE. coli and S. typhimurium have been widely reported, displaying similar Ser-83 and/or Asp-87 mutations within the QRDR ( Heisig et al, 1994; 1995; Ruiz et al, 1995; Truonget al, 1995; Griggs et al, 1996; Bazile-Pham-Khac et al, 1996). Such quinolone-resistantE. coli isolates commonly show a Ser-83 to leucine (Yoshida et al, 1988; Oram and Fisher, 1991; Heisig et al, 1993) or Ser-83 to tryptophan change (Yoshida et al, 1988; Cullen et al, 1989; Oram and Fisher, 1991). Earlier reports of plasmid-mediated resistance to nalidixic acid (Panhotra et al, 1985; Munshi et al, 1987) have not been validated, and are postulated to result from carriage of a plasmid-encoded mutator phenotype, which increases frequencies of mutation to quinolone resistance (Crumplin, 1987; 1990a).In vitro work has confirmed that mutator plasmids do increase apparent mutation to nalidixic acid resistance in E. coli cultures incubated in broth containing lethal concentrations of nalidixic acid (Ambler and Pinney, 1995), and that one of the plasmids, pYDl, carried by the nalidixic acid-resistant 215 strain of S. dysenteriae from Kashmir (Panhotraet al, 1985) does indeed encode a mutator phenotype (Ambler et al, 1993). Adaptive mutation has recently, been reported to increase mutation to ciprofloxacin resistance. During seven days incubation ofE. coli cultures on medium containing twice the MIC of ciprofloxacin, mutations to ciprofloxacin resistance continually occurred in non-dividing cells, while mutations to rifampin resistance did not accumulate (Riesenfeld et al, 1997). However, by definition (Cairns et al, 1988), adaptive mutation is said to occur only in selected genes during the application of a non-lethal selection in non-growing cultures. Therefore the use of twice the MIC of ciprofloxacin by Riesenfeld et al (1997), does not fit with the accepted criteria for demonstration of adaptive mutation. Experiments, reported in this thesis, performed to investigate adaptive mutation to quinolone resistance have shown: • The effect of prolonged incubation ofE. coli AB 1157 on minimal media lacking an essential amino acid, on the frequency of mutation to amino acid independence and the effect of mutator plasmids on the rate of mutation. • An increased frequency of mutation to nalidixic acid resistance in E. coli AB 1157 plated on nutrient agar containing inhibitory or lethal concetrations of nalidixic acid, further enhanced by the presence of a mutator plasmid. • The influence of growth in sub-inhibitory concentrations of nalidixic acid on the production of nalidixic acid-resistant E. coli ABl 157 mutants. 216 • Dependence of adaptive mutation to nalidixic acid resistance on the DNA repair capacity of the host. • That the base substitutions within the QRDRs of gyrA of the quinolone-resistant clones are similar to those found in clinically-isolated quinolone-resistant mutants. 1. The effect of mutator plasmids on the post-UV survival of E, coli ABl 157 and repair-deficient derivatives Initial experiments transferred the mutator plasmids R46 and R446b, and the control non-mutator plasmid RP4, by conjugation, into the repair-proficientE. coli strain ABl 157 (Table 4). Plasmids R46 and R446b were also transferred into the repair- deficient & coli strains ABl886 uvrA6^ AB2463 recA13, AB2470 recB21, AB2494 lexA3 and TK702 umuC36 (Tables 5 and 6). Post-UV survival experiments with E. coli AB 1157 and its plasmid-containing derivatives confirmed the UV-protection phenotypes, and by inference the mutator phenotypes, conferred by plasmids R46 and R446b, and that plasmid RP4 does not carry such a phenotype. Plasmid R46 conferred the higher level of UV protection, producing a 30.5-fold increase in the surviving fi*action following exposure to a UV dose of 160 J m'^. R446b increased the surviving fi*action of the repair- proficient ABl 157 strain 5-fold following the same UV dose (Table 7, Figure 4). Such variation in post-UV survival conferred by mutator plasmids has previously been documented (Siccardi, 1969; Molina et al, 1979; Pinney, 1980; Upton and Pinney, 1983) and is believed to result from differential expression of the muc genes encoded by the plasmids. 217 The UV-protection abilities of plasmids R46 and R446b were also examined in repair-deficient strains. Neither R46 nor R446b bad any effect on post-UV survival in recAlS or lexA3 backgrounds (Table 7, Figures 6 and 7), re-iterating the dependence of muc gene expression on hostrecA^ and lexA^ genotypes (Tweats et al, 1976; Upton and Pinney, 1983). Plasmid-mediated UV- protection is not dependent on functional host recB^ oruvrA^ genotypes (Upton and Pinney, 1983; Table 7), in agreement with previous findings. Similarly, plasmids R46 and R446b were shown to increase post-UV survival of the umuC mutant (Table 7, Figure 10; Upton and Pinney, 1983), via substitution of the non-functional chromosomalumuDC operon with the plasmid-encoded mucAB operon (Perry et al, 1985; Kulaeva et al, 1998). 2. Selection-induced adaptive mutation to amino acid independence The phenomenon of adaptive mutation has been reported at many different loci on the E. coli chromosome. Such selection-induced mutation occurs with varying fi*equencies, depending on the mutation involved (Hall, 1988; 1989; 1990; 1991; 1993; 1994; 1995a; 1995b; Jayaraman, 1992; Bridges, 1994). Before examining the possibility of selection-induced mutation to nalidixic acid resistance in E. coli strain ABl 157, the effect of selection on reversion to amino acid independence (leucine, arginine, threonine and histidine) was studied. The effects of R46 and R446b carriage on selection- and UV- induced reversion to amino acid independence were also examined, to determine any role for the plasmid-encoded mutator phenotype. As a control, plasmid enhancement of UV- induced reversion to amino acid independence was, indeed, demonstrated (Table 11, Figures 11 to 14), in agreement with previous reports (Molinaet al, 1979; Pinney, 1980; Upton and Pinney, 1983), R46 conferring greater reversion to amino acid independence 218 at all loci tested (Table 11). Four of the mutations conferring amino acid auxotrophy onE. coli ABl 157 reverted at measurable frequencies during selection (Table 12, Figures 15 to 18). Production of such revertants in bacterial cultures plated onto medium lacking the required amino acid may be considered adaptive, since it fulfills the criteria for adaptive mutation in that mutations arise in non-growing cells under non-lethal selection, in genes whose function was selected (Cairns et al, 1988). The frequencies at which reversions occurred varied depending on the allele selected. Between day two (when spontaneous revertants already present in the culture prior to selection had grown to form visible colonies) and day ten, a 20.5-fold increase in the number of Leu^ revertants was observed (Table 12, Figure 15), whereas during the same period of time, the number of selection- induced Arg^ revertants only doubled (Table 12, Figure 18). Such variation in mutability of chromosomal targets, with no apparent relationship between target size and frequency of mutation (Torkelsonet al, 1997), has lead to the proposal that hot and cold regions exist within the chromosome for stationary-phase “adaptive” mutation (Rosenberget al, 1998). Similar to the plasmid enhancement of UV-induced reversion to amino acid independence, the presence of mutator plasmid R46 was shown to increase selection- induced frequencies of reversion to threonine, leucine and arginine independence (Table 12). However, in contrast to UV-induced reversion to histidine independence (Table 11), plasmid R46 had no significant effect on selection-induced mutation at this locus (Table 12, Figure 17), again illustrating variation in the mutability of different chromosomal sites (Torkelsonet al, 1997; Rosenberg et al, 1998). The findings that the R46 mutator plasmid increases selection-induced reversion 219 frequencies of three out of the four amino acid-encoding loci tested, suggests that the Muc phenotype is expressed during such non-lethal selection. Therefore, it is possible that the mechanism determined by mutator plasmids that enhances spontaneous and UV- induced mutagenesis is also utilised to produce amino acid-independent adaptive mutants. If error-prone DNA repair is involved in adaptive mutation, then, it was predicted, exposure to an inducer of such repair, such as nalidixic acid (Gudas and Pardee, 1976; Phillips et al, 1987; Piddock and Wise, 1987), might trigger adaptive mutation to nalidixic acid resistance, with mutation frequencies enhanced by carriage of a mutator plasmid. 3. Selection-induced mutation to nalidixic acid resistance in E. coli ABl 157 cultures plated on medium containing the MIC or supra-inhibitory (lethal) concentrations of nalidixic acid Initial experiments examined UV-induced mutation to nalidixic acid resistance in E. coli ABl 157 cultures plated on medium containing 4 (inhibitory, but not lethal), 8 or 16 yUg mL ’ (lethal concentrations) nalidixic acid. As expected, similar to UV-induced reversion to amino acid independence (Figures 11 to 14), control experiments utilising increasing doses of UV exposure increased the frequency of mutation to nalidixic acid resistance at all drug concentrations tested. The carriage of a mutator plasmid further increased UV-induced mutation to nalidixic acid resistance (Figures 19 to 21). These data were in agreement with previous findings demonstrating the role of mutator plasmids in increasing the frequency of UV-induced mutation to nalidixic acid resistance in E. coli cultures exposed to lethal concentrations of drug (Ambler et al, 1993; Ambler and Pinney, 1995). 220 Riesenfeld et al (1997) claimed to have demonstrated adaptive mutation to ciprofloxacin resistance in E. coli cells plated on ciprofloxacin-containing medium. However, whether such selection-induced ciprofloxacin-resistant mutants can be considered truely adaptive is open to question since cultures were plated on medium containing twice the MIC of ciprofloxacin for the strain under examination. Such a concentration would kill susceptible cells and allow resistant mutants to grow through. Selection-induced mutation to nalidixic acid resistance was observed in plasmid- ffee E. coli cultures incubated on medium containing 4 yWg mL ' nalidixic acid for ten days (Figure 22). An increase in the number of nalidixic acid-resistant mutants was also demonstrated in E. coli cultures plated on medium containing 8 or 16 /.^g mL ' nalidixic acid (Figures 23 and 24), however, the increase in the number of nalidixic acid-resistant mutants observed here cannot be considered to be truely adaptive, since concentrations of 8 and 16 yUg mL ' nalidixic acid are lethal to the strain. Between day one and day ten, the average number of nalidixic acid-resistant clones that grew on plates containing 4, 8 or 16 //6g mL ' nalidixic acid increased 240-, 553- and 37-fold respectively. At day ten there was an average of 1.6 x 10\ 1.7 x 10^ and 11 nalidixic acid-resistant colonies on plates containing 4, 8 or 16 yWg mL ' nalidixic acid respectively (Table 16). In agreement with Smith (1986), these results re-iterated the fact that fewer mutants grew on media containing higher concentrations of quinolone. In parallel with mutator plasmid enhancement of UV- and selection-induced reversion to amino acid independence (Tables 11 and 12; Pinney, 1980; Upton and Pinney, 1983), the carriage of R46 and R446b further increased the number of nalidixic acid-resistant mutants recovered after ten days incubation on all concentrations of drug tested (Table 16, Figures 22 to 24). However, whereas plasmid R46 had a greater effect 221 on selection-induced reversion to amino acid independence (Tables 11 and 12, Figures 11 to 18; Pinney, 1980; Upton and Pinney, 1983), R446b was the more effective enhancer of UV- and selection-induced mutation to nalidixic acid resistance (Tables 14 and 16, Figures 19 to 24). The capacities of the various mutation-enhancing opérons to promote mutations at different DNA lesions vary (Nohmiet al, 1992; 1995a; 1995b; Doyle and Strike, 1995; Kulaevaera/, 1995; 1998; Lawrenceera/, 1996; Szekeres etal, 1996; Gruz et al, 1996; 1998) suggesting that the mucAB-Xike, operon of R446b is better at promoting mutations to nalidixic acid resistance in drug exposed cultures than the similar operon of R46 (Mortelmans and Stocker, 1979; Kulaeva et al, 1998), whereas R46 enhances reversion to amino acid independence to a greater extent than R446b. That increasing numbers of nalidixic acid-resistant mutants are recovered during incubation of the plasmid-free strain on medium containing the MIC of the drug fits with the requirements for “adaptive” mutation (Cairns et al, 1988). Such mutants are said to occur in non-growing cultures in genes whose function is selected by a non-lethal process: at its MIC of 4 mL'% nalidixic acid is inhibitory, but not lethal, to E. coli ABl 157. Increase in the number of nalidixic acid-resistant mutants during nalidixic acid selection by the presence of a mutator plasmid supports the hypothesis that error-prone DNA repair is involved in adaptive mutation to nalidixic acid resistance. A possible sequence of events could commence with nalidixic acid interfering with DNA replication by inhibition of DNA gyrase. This leads to SOS induction (see Introduction section 3.5.9), resulting in error-prone repair of nalidixic acid-induced DNA lesions. This produces mutations within the chromosome that confer resistance to nalidixic acid. Enhancement of error-prone repair, by carriage of a mutator plasmid, would further 222 increase the frequency of selection-induced mutation to nalidixic acid resistance. 4. Mutagenesis to nalidixic acid resistance in E. coli ABl 157 cultures during incubation in nutrient broth containing sub-inhibitory concentrations of nalidixic acid The proposal that adaptive mutations only occur under the pressure of non-lethal selection (Cairns et al, 1988) has been challenged by demonstration that some mutations arise in the absence of selection, under the non-specific stress of starvation (Mittler and Lenski, 1990; 1992), although a higher frequency of mutation occurs in the presence of specific selection pressure (Hall, 1994). By incubating cultures of E. coli ABl 157 in nutrient broth containing 0, 2 or 3 //g mL ' nalidixic acid for up to ten days, the effect of sub-inhibitory concentrations of drug on the development of nalidixic acid resistance could be examined. After two days incubation in drug-free broth, the number of nalidixic acid- resistant mutants of the plasmid-free ABl 157 strain had increased 2.4-fold. That no further increase in the number of mutants was observed during the next eight days suggests that the resistant clones that had appeared by day two grew from spontaneous mutants already present in the culture. However, the presence of 2 or 3 /ig mL ' nalidixic acid in the broth medium (sub-inhibitory, and therefore, non-lethal concentrations) produced increases of 20- and 588-fold respectively, in the number of drug-resistant mutants recovered after ten days, as compared to cultures incubated in drug-free broth (Figure 25). One of the criteria of Cairnset al (1988) for adaptive mutants is that they only arise in non-growing cultures. Following one day’s incubation in drug-free or drug- containing nutrient broth, the total viable count of the plasmid-free E. coli ABl 157 223 culture reached approximately 1x10^ organisms per mL, and remained essentially constant during the remaining nine days of the experiment (Figure 25). However, this does not rule out the possibility of cryptic growth, where mutations are produced in a proportion of cells in the population able to grow and divide, at the expense of death of other cells. Hall (1990) examined the possibility that cryptic growth was responsible for the production of adaptive Trp^ mutants but could demonstrate no cell division in such populations. He also estimated that the maximum amount of cell turnover during tryptophan selection was insufficient to account for the number of Trp^ mutants that appeared (Hall, 1990). Without examining total counts or DNA synthesis within cultures under nalidixic acid selection, it is impossible to conclude that no cell division is occurring. These results confirm the need for selection pressure in the production of adaptive nalidixic acid-resistant mutants, since in the absence of such selection no increase in the number of nalidixic acid-resistant mutants was observed after day two, when spontaneous mutants would be expected to appear. Also evident fi*om these results is the fact that increasing the sub-lethal concentration of drug in the medium leads to an increase in the number of nalidixic acid-resistant mutants. Exposure to a nalidixic acid concentration of 3 yUg mL ' produces greater numbers of nalidixic acid-resistant mutants than a nalidixic acid concentration of 2 yUg mL '. Similar to results obtained when examining selection-induced mutation to nalidixic acid resistance on solid media, plasmids R46 and R446b, but not the control non-mutator plasmid RP4, further increased the number of nalidixic acid-resistant mutants recovered fi*om broth cultures containing 0,2 or 3 jj,% mL ' nalidixic acid (Table 17). Even when grown in drug-free broth, the spontaneous mutator effect of R46 and R446b can be seen, producing 7- and 24-fold more nalidixic acid-resistant mutants 224 respectively after ten days, than the plasmid-free and RP4-containing strains (Figure 26). Again, differences were observed in the level of enhanced mutagenesis conferred by R46 and R446b. The R46-containing strain produced 70- and 8 x 10^-fold greater numbers of resistant mutants compared to the plasmid-free strain, following ten days incubation in nutrient broth containing 2 or 3 ^ ^ mL ’ nalidixic acid (Figure 28, Table 17). Further increases were observed when the strain harboured plasmid R446b. After ten days incubation in medium containing 2or3^:g m L ’ nalidixic acid, the number of nalidixic acid-resistant mutants recovered was 1.2 x ICP- and 1.6 x 10 '‘-fold greater than the plasmid-free strain (Figure 29, Table 17). Such increases led to massive numbers of nalidixic acid-resistant ABl 157 (R446b) mutants being recovered, such that after ten days incubation in nutrient broth containing 3 /ig mL ' nalidixic acid, 1 in 8 of the entire population was now resistant to nalidixic acid (Figure 29). The increase in adaptive mutation to nalidixic acid resistance observed when the strain harboured a mutator plasmid, and the absence of such an increase conferred by the non-mutator plasmid, confirmed the importance of mutator genes in selection-induced mutation and provides further evidence for error-prone DNA repair being involved in the production of adaptive mutants to nalidixic acid resistance. 5. Examination of unselected mutagenesis following growth in liquid medium containing sub-inhibitory concentrations of nalidixic acid For mutations to be truely adaptive, they should occur only in selected genes and not at unselected loci (Cairns et al, 1988). For this reason, reversion to threonine independence was examined in cultures undergoing selection in broth containing sub- inhibitory concentrations of nalidixic acid (Results section 2.3; Figure 16), i.e. conditions 225 that contained excess threonine. An increase in unselected mutagenesis to threonine independence was observed with more Thr^ revertants recovered from cultures grown in the higher concentration of drug (Table 18, Figures 32 and 33). Carriage of R46, but especially R446b, further increased the number of Thr^ revertants, suggesting that both selected and unselected mutations were produced via the same mechanism of error-prone repair. However, whilst 1.0x10"^ nalidixic acid-resistant AB1157 mutants were recovered per mL after ten days growth in medium containing 3 yUg mL ’ nalidixic acid, only 140 Thr^ revertants were recovered from the same population. R46 and R446 increased the number of Thr^ revertants recovered under the same conditions to 2.8 x 10^ and 7.1 x 10^ Thr^ revertants per mL respectively (Figure 32). This reversion of the unselected marker is very small in comparison to the 8.0 x 10^ ABl 157 (R46) and 1.6 x 10^ ABl 157 (R446b) nalidixic acid-resistant mutants per mL present in the same cultures (Figures 28 and 29). Adaptive mutations have been proposed to only occur in genes under selection. However an increase in unselected mutagenesis might be expected in cultures grown in the presence of nalidixic acid, since quinolones are potent inducers of error-prone DNA repair (Gudas and Pardee, 1976; Piddock and Wise, 1987; reviewed by Hooper et al, 1988; Lewin et al, 1989; Walters et al, 1989; Ysem et al, 1990). The consequence of such induction would be the production of mutations throughout the chromosome, in addition to those specific to the adaptive response, which results from growth in sub- inhibitory concentrations of nalidixic acid. 6. Dependence of adaptive mutation to nalidixic acid resistance and R46-mediated effects on host DNA repair genotype One way of testing the hypothesis propounded in this thesis, that selection- 226 induced adaptive mutation to nalidixic acid resistance involves error-prone repair of nalidixic acid-induced DNA lesions would be to determine if adaptive mutation occurred in the absence of error-prone repair. Induction of SOS error-prone repair by quinolone antibacterials such as nalidixic acid requires the participation of RecA, RecBCD and LexA proteins (Gudas and Pardee, 1976; McPartland et al, 1980; Phillips et al, 1987; Piddock and Wise, 1987), in addition to a functional umuDC-VikQ operon for SOS processing. Error-prone DNA repair has been proposed, on the one hand, to contribute to quinolone-induced cell death (McDaniel et al, 1978; Piddock and Wise, 1987; Piddock et al, 1990) and, on the other hand, to repair DNA damage produced by quinolone action (Ginsburg et al, 1982; Drlica, 1984; Walters et al, 1989). Results presented by Lewin et al (1989) suggest that nalidixic acid-induced DNA damage is not repaired by the inducible mutagenic SOS system, but rather by the error-free recombination pathway. However, DNA damage caused by the fluoroquinolones ciprofloxacin, norfloxacin and ofloxacin, which have additional bactericidal mechanisms, is subject to SOS repair (Howard gr a/, 1993a; 1993b). Selection-induced mutation to nalidixic acid resistance was reduced 4.8 x 1 0 \ 2.9 X 10^-, 1.4 X 10^- and 1.0 x 10^-fold in repair-deficient strains carrying the mutant alleles recAlS, recBll, lexAS and umuC36 respectively (Table 20). Such results add further evidence to the hypothesis that at least a portion of selection-induced mutation to nalidixic acid resistance occurs via error-prone repair following SOS induction by nalidixic acid. Abolition of error-free excision repair by the carriage of the mutant uvrA6 allele, increased selection-induced mutation to nalidixic acid resistance 5-fold (Table 20). Such an increase in mutation frequency may be expected, since in the absence of error- 227 free excision repair, a greater number of lesions could be subject to error-prone repair, increasing the frequency of mutation to nalidixic acid resistance. Carriage of plasmid R46 in strain TK702 umuC36 increased the frequency of mutation to nalidixic acid resistance 4.5 X 10^-fold (Table 21), producing greater numbers of mutants than the plasmid-free repair-proficient strain. This demonstrates the increased mutagenic efficiency of the mucAB operon, compared to chromosomalumuDC genes. Indeed, Clerch et al (1996) have recently reported that mucAB, but not umuDC, promote ciprofloxacin-induced mutagenesis in both chromosomal and plasmidic hisG428 targets in S. typhimurium. This enhanced mutagenic potential conferred by mucAB appears to result from an increased efficiency of MucA'B to allow translesion DNA synthesis across a large range of DNA lesions (Lawrence et al, 1996; Venderbure et al, 1999). Surprisingly, the presence of plasmid R46 increased the frequencies of selection- induced mutation to nalidixic acid resistance in strains AB2463 recAlS and AB2470 recB21 (Table 21), both of which should be uninducible for 80S function by nalidixic acid. However, mucAB has previously been shown to increase mutation in strains in which DNA is undamaged (Walker, 1977; Waleh and Stocker, 1979; Blanco etal, 1986), and the increases in mutation observed may reflect such a background effect. Indeed, when tested in a recA430 strain, which cannot process MucA, MucAB-directed UV mutagenesis was shown to be nearly as proficient as in therecA^ wild-type (Blanco et al, 1986; Nohmi et al, 1988). However, UmuDC was not expressed against the recA430 background (Ennis et al, 1985; Blancoet al, 1986). The higher efficiency of processing of MucA to the mutagenically active form, MucA', has been suggested to account for the enhanced mutagenic effects of mucAB when compared with umuDC (Hauser et al, 1992). Such differential processing may also explain the increases in selection-induced 228 mutagenesis to nalidixic acid resistance in the AB2463 recAlS and AB2470 recBll strains carrying plasmid R46 (Table 21). 7. MIC determination and genetic analyses of quinolone-resistant clones following growth in sub-inhibitory concentrations of quinolone MIC determination of selected nalidixic acid-resistant clones following a single period of growth in nutrient broth containing 3 yUg mL ' nalidixic acid (0.75 x MIC) revealed that clones exhibiting a high level of resistance to nalidixic acid (^ 128 x MIC) were readily selected. Even when nalidixic acid was absent from the growth medium, 11 of the 20 resistant mutants selected showed MICs to nalidixic acid of >512 yUg mL ' (Table 22). The presence of 0.75 x MIC ciprofloxacin in the growth medium was also tested to examine if exposure to this drug led to a different spectrum of quinolone- resistant mutants. Similar to the resistance levels of control clones plated on nalidixic acid-containing medium following growth in drug-free broth, of the 20 control clones plated on medium containing ciprofloxacin, 10 had nalidixic acid MICs of >512 ywg mL ' (Table 22). Growth in broth containing a sub-inhibitory concentration of ciprofloxacin produced mutants with higher ciprofloxacin MICs than clones selected in nalidixic acid- containing medium; five clones selected from medium containing ciprofloxacin had an MIC to this drug of 0.16 yUg mL ', whereas only one clone selected from nalidixic acid- containing medium exhibited the the same MIC to ciprofloxacin (Table 23). Such a single period of growth in sub-inhibitory concentrations of either nalidixic acid or ciprofloxacin readily produced clones with a high level of resistance to nalidixic acid (>512 yUg mL'). However, these same clones, regardless of the selection applied, did not also exhibit high ciprofloxacin MICs, with only 14 out of 80 selected clones giving ciprofloxacin MICs 229 of 0.16 mL"’ (32 x MIC) (Tables 22 and 23). A second priod of growth of quinolone-resistant clones in nutrient broth containing 0.75 x the respective “resistant” MICs of either nalidixic acid or ciprofloxacin led to 101 of 120 selected colonies giving a nalidixic acid MIC of >1024 yUg mL ' (Table 26). High ciprofloxacin MICs were exhibited by clones selected on ciprofloxacin following a second period of growth in 0.75 x their “resistant” ciprofloxacin MIC: 21 of 60 colonies selected on ciprofloxacin gave MICs to ciprofloxain of >1.28 mL'', whereas all 60 colonies selected on nalidixic acid plates had ciprofloxacin MICs of <0.64 A6g mL ' (Table 27). The QRDRs within the gyrA genes of selected resistant mutants were determined using the Polymerase Chain Reaction (PCR) and Hinfi restriction. It was hoped that such sequencing would reveal mutations within this region, that might explain the patterns of resistance observed. Other mutations, for example, within gyrB (Yamagishi et al, 1986; Yoshida et al, 1991) or associated with a decrease in the level of outer membrane porins (Hirai et al, 1986a; 1986b; Cohen et al, 1989; Hooperet al, 1992) or induction of the marRAB operon (Cohenet al, 1989; 1993c; Hachleret al, 1991; Ariza et al, 1994), can also increase quinolone MICs. But since the increases in quinolone resistance they produce are relatively slight, only the QRDRs of the gyrA genes of our high-level quinolone-resistantE. coli clones were examined. Hinfl restriction analysis of eleven selected quinolone-resistant clones revealed that three contained mutations at position Ser-83 (Figure 36), abolishing a Hinfl. recognition site within the quinolone resistance-determining region (QRDR) ofgyrA. Sequence analysis was carried out on these three clones (NA8, NA9 and NAl 1) and on another (NAIO), which exhibited high-level resistance to nalidixic acid (MIC >1024 yUg 230 mL ') but only low-level resistance to ciprofloxacin (MIC 0.04 mL ') (Table 30). Two of the four resistant strains, NA9 and NAl 1, displaying high levels of resistance to both nalidixic acid (MICs >1024 yWg mL ') and ciprofloxacin (MICs of 5.12 and 2.56 /Ug mL ' respectively), revealed a C G transversion mutation changing Ser-83 to tryptophan (Table 30). The third mutant (NAS), with nalidixic acid and ciprofloxacin MICs of kI024 and 0.02 yUg ml ' respectively, revealed a C ^ T transition mutation altering Ser-83 to leucine. Hinfl. restriction of clone NAIO (nalidixic acid MIC >1024 jj,g mL ', ciprofloxacin MIC 0.04lug mL ') revealed that the Hinfl recognition site around Ser-83 was maintained, explaining why it continued to yield three fragments following Hinfl restriction analysis (Figure 36). Further sequence analysis of the QRDR of thegyrA of this latter clone revealed a mutation changing Asp-87 to glycine (Table 30), explaining why it yielded three fragments following Hinfl restriction analysis (Figure 36). Replacement of Ser-83 by a hydrophobic amino acid, such as leucine or tryptophan, increases nalidixic acid MICs to >128 yUg mL ' (Vila et al, 1994) and produces a low level of resistance to fluoroquinolones; i.e. MIC for ciprofloxacin < 1 yUg mL ' (Yoshida et al, 1988; Oram and Fisher, 1991). Similarly, replacement of aspartate at position 87 of GyrA by a non-negatively charged amino acid, such as asparagine, glycine or tyrosine, also results in a decreased susceptibility to quinolone antibacterials (Yoshida et al, 1990; Khodursky et al, 1995; Drlica et al, 1996). Both substitutions are thought to weaken the interaction between DNA gyrase and the quinolone-DNA complex (Oram and Fisher, 1991; Hallett and Maxwell, 1991). The site of mutation within the QRDR of GyrA correlates with the level of quinolone resistance conferred upon the organism. Mutations at position 83 generally confer a greater level of resistance than mutations occurring at position 87, suggesting that changing Ser-83 leads to a weaker 231 interaction between DNA gyrase and the quinolone DNA complex, than alteration of Asp-87 (Yoshida et al, 1990; Khodursky et al, 1995; Drlica et al, 1996). However, this does not explain why replacement of Ser-83 with tryptophan led to a higher level of resistance to ciprofloxacin than the Ser-83 to leucine alteration (Table 30). Mutations have also been reported at other sites within the QRDR of GyrA, including positions 67, 81, 84 and 106 (Yoshida et al, 1990), which may also contribute to quinolone resistance. However sequence analysis of the QRDRs of the four quinolone- resistant mutants revealed no other sequence alterations other than those at positions 83 and 87 (Table 30). In addition to mutations in the QRDR of GyrA, mutations contributing to quinolone resistance can also occur ingyrB (Yamagishi et al, 1996; Yoshida et al, 1991), in genes encoding topoisomerase IV ingyrA mutants (Heisig, 1996; Vila et al, 1996; Everett et al, 1996; Breines et al, 1997), and in genes affecting drug uptake and accumulation (Relia and Haas, 1982; George and Levy, 1983a; Sanders etal, 1984; Hirai et al, 1986a; 1986b; Hooper et al, 1986; Piddock et al, 1987; Cohen et al, 1989). Therefore, mutations at other sites within the genomes of the four resistant clones examined could also contribute to their decrease in quinolone susceptibility. However, clinically-isolated E. coli quinolone-resistantgyrA mutants commonly display either a Ser-83 or Asp-87 mutation (Yoshidaet al, 1988; 1990; Cullenet al, 1989; Fisher et al, 1989; Oram and Fisher, 1991) indicating the importance of these sites and the ease at which they mutate to confer quinolone resistance. Indeed, the quinolone-resistant mutants described here, including NA9 and NAl 1 (ciprofloxacin MICs of 5.12 and 2.56 yUg mL ' respectively) were isolated after only two periods of exposure to either nalidixic acid or ciprofloxacin. It is possible that under conditions of prolonged or repeated exposure to 232 quinolones, further mutations ingyrA, gyrB, topoisomerase IV or loci affecting drug uptake and accumulation, may be produced resulting in cells highly resistant to quinolone antibacterials, giving, for example ciprofloxacin MICs as high as 128jug mL ' as reported by Heisig et al, 1993. 8. Summary and conclusions To summarise: the work presented in this thesis has shown that exposure ofE. coli cultures to sub-lethal or inhibitory concentrations of nalidixic acid in vitro, significantly increases the frequency at which cells mutate to become resistant to this drug. The effect is greatly enhanced by the carriage of a mutator plasmid, so much so that following ten day’s incubation in nutrient broth containing 3^g mL ' nalidixic acid (0.75 X MIC), approximately 1 in 8 of ABl 157 (R446b) cells had mutated to nalidixic acid resistance. The mechanism by which such selection-induced mutations are introduced appears to involve the error-prone repair of nalidixic acid-induced DNA lesions. Although, there are few claims in the literature for the involvement of error-prone repair in adaptive mutation (Yatagai et al, 1991; Bridges, 1993; Foster and Trimarchi, 1994; Rosenberg et al, 1994; reviewed by Bridges, 1998), it is possible that under the condition of strong SOS induction provided by nalidixic acid, error-prone repair could plat at least a part in the adaptive process under these conditions. Strains deficient in error-prone repair show greatly reduced frequencies of selection-induced mutation to nalidixic acid resistance. The observation that the mutator plasmid R46 returns the frequency of mutation to resistance in strain TK702 umuC36 to wild-type levels, confirms the requirement for error-prone DNA repair. Reports of plasmid-mediated resistance to nalidixic acid (Panhotra et al, 1985; Munshi et al, 1987) were later suggested (Crumplin, 233 1987) and then shown (Ambler et al, 1993; Ambler and Pinney, 1995) to result from a plasmid-conferred mutator phenotype. These reports suggested a role for SOS induction and error-prone repair in the development of nalidixic acid resistance in nalidixic acid- exposed cultures, which has been confirmed by the results presented here. Genetic analysis of selection-induced quinolone-resistantE. coli clones revealed mutations that play a major role in contributing to the development of clinical quinolone resistance. Amino acid substitution at serine 83 or aspartate 87, witin the QRDR of gyrA, were detected and are believed to be responsible for the majority of clinical quinolone resistance (Yoshida et al, 1988; 1990; Cullen et al, 1989; Fisheret al, 1989; Oram and Fisher, 1991). The data presented in this thesis therefore suggest that the presence of nalidixic acid induces mutational resistance to itself. This clearly has important implications in the clinical management of bacterial infections, suggesting that inappropriate dosage regimens could contribute to the development of strains resistant to the quinolone antibacterials. 234 REFERENCES Akiba, T., K. Koyama, Y. Ishiki, S. Kimura, and T. Fukushima (1960)Japanese Journal of Microbiology4:219-227. Aldridge, K.E (1994)Antimicrobial Agents and Chemotherapy 38:1671-1674. Allen, P.N., and H.F. NoUer (1989)Journal of Molecular Biology208:457-468. Ambler, J.E., Y.J. Drabu, P.H. Blakemore, and R.J. Pinney(1993) Journal of Antimicrobial Chemotherapy 31:831-839. Ambler, J.E., and R.J. Pinney (1995) Journal of Antimicrobial Chemotherapy 35:603- 609. Amyes, S.G.B., and J.T. Smith (1974) Biochemical and Biophysical Research Communications58:412-418. Anderson, E.S., and N. Datta (1965) Lancet i:407-409. Aoyama, H., K. Sato, T. Fujii, K. Fujimaki, M. Inoue, and S. Mitsuhashi (1988) Antimicrobial Agents and Chemotherapy 32:104-109. Ariza, R.R., S.P. Cohen, N. Bachhawat, S B. Levy, and B. Demple (1994)Journal of Bacteriology176:143-148. Armengod, M E., and M. Blanco (1978)Mutation Research 52:37-47. Ashraf, M.M., Z.U. Ahmed, and D A. Sack (1991)Canadian Journal of Microbiology 37:59-63. An, K.G., K. Welsh, and P. Modrich (1992) Journal of Biological Chemistry 267:12142-12148. Bagg, A., C.J. Kenyon, and G.C. Walker(1981) Proceedings of the National Academy of Sciences of the United States of America 78:5749-5753. Bailone, A., S. Sommer, and R. Devoret (1985)Proceedings of the National Academy of Sciences of the United States of America82:5973-5977. Bailone, A., S. Sommer, J. Knezevic, M. Dutreix, and R. Devoret (1991)Biochimie 73:479-484. Barlow, A.M. (1963) British Medical Journal 2:1308-1310. Barry, A.L., and R.N. Jones (1989)Journal of Antimicrobial Chemotherapy23:527- 535. 235 Basak, J., U. Mukherjee, and S.N. Chatterjee (1992)Environmental and Molecular Mutagenesis 20:53-60. Bates, H., B.A. Bridges, and R. Woodgate (1991) Mutation Research 250:199-204. Battista, J R., T. Nohmi, L. Dodson, and G.C. Walker (1988)In: University College Los Angeles Symposium Molecular Cell Biology “Mechanisms and Consequences of DNA Damage Processing.”, pp. 455-459,Eds. Friedberg, E.D. and P.C. Hanawalt, New York. Battista, J R., T. Ohta, T. Nohmi, W. Sun, and G.C. Walker (1990)Proceedings of the National Academy of Sciences of the United States of America87:7190-7194. Bauernfeind, A (1993) Journal of Antimicrobial Chemotherapy 31:505-522. Bazile-Pham-Khac, S., Q.C. Truong, J.P. Lafont, L. Gutmann, X.Y. Zhou, M. Osman, and N.J. Moreau(1996) Antimicrobial Agents and Chemotherapy 40:1504- 1507. Benbrook, D M., and R.V. Miller (1986) Antimicrobial Agents and Chemotherapy 29:1-6. Bennett, A.D., and W.V. Shaw 1983. Biochemical Journal 215:29-38. Bennett, R.J., H.J. Dunderdale, and S.C. West (1993)Cell 74:1021-1031. Benson, F., S. Collier, and R.G. Lloyd (1991) Molecular and General Genetics 225:266-272. Benson, F.E., G.T. Illing, G.J. Sharpies, and R.G. Lloyd (1988)Nucleic Acids Research 16:1541-1550. Benveniste, R., and J. Davies (1973) Annual Review of Biochemistry 42:471-506. Berger-Bachi, B., L. Barberis-Mario, A. Strassle, and F.H. Kayser (1989)Molecular and General Genetics 219:263-269. Berger-Bachi, B., A. Strassle, J.E. Gustafson, and F.H. Kayser (1992)Antimicrobial Agents and Chemotherapy 36:1367-1373. Bernardi, A., and P. Leder (1970) Journal of Biological Chemistry 245:4263-4268. Bertrand-Burggraf, E.S., S. Hurstel, M. Daune, and M. Schnarr(1987) Journal of Molecular Biology 193:293-302. 236 Blanche, R., B. Cameron, F.-X. Bernard, L. Maton, B. Manse, L. Ferrero, N. Ratet, C. Lecoq, A. Goniot, D. Bisch, and J. Crouzet (1996)Antimicrobial Agents and Chemotherapy 40:2714-2720. Blanco, M. G. Herrera, V. Aleixandre (1986) Molecular and General Genetics 205:234-239. Bloom, L.B., J. Turner, Z. Kelman, J.M. Beecham, M. O’Donnell, and M F. Goodman (1996) Journal of Biological Chemistry 271:30699-30708. Blumberg, P.M., and J.L. Strominger (1974)Bacteriological Reviews38:291-335. Boe, L. (1990) Molecular Microbiology 4:597-601. Bonner, C.A., S. Hays, K. McEntee, and M.F. Goodman (1990)Proceedings of the National Academy of Sciences of the United States of America 87:7663-7667. Bonner, C.A., S.K. Randall, C. Rayssiguier, M. Radman, R. Eritja, B E. Kaplan, K. McEntee, and M.F. Goodman (1988) Journal of Biological Chemistry 263:18946- 18952. Boudsocq, F., M. Campbell, R. Devoret, and A. Bailone (1997)Journal of Molecular Biology270:201-211. Brash, D.E., W.A. Franklin, G.B. Sancar, A. Sancar, and W.A. Haseltine (1985) Journal of Biological Chemistry260:11438-11441. Breeze, A S., P. Sims, and K.A. Tracey (1975)Genetical Research 74:543-556. Breines, D M., S. Ouabdesselam, E.Y. Ng, J. Tankovich, S. Shah, C.J. Soussy, and D C Hooper (1997)Antimicrobial Agents and Chemotherapy 41:175-179. Brent, R. (1982)Biochimie 64:565-569. Brent, R., and M. Ptashne (1981)Proceedings of the National Academy of Sciences of the United States of America 78:4204-4208. Bresler, S.E. (1975) Mutation Research 29:467-472. Bridges, B.A. (1988)Mutation Research 198:343-350. Bridges, B.A. (1993)Mutation Research 302:173-176. Bridges, B.A. (1994) Mutation Research 307:149-156. Bridges, B.A. (1996)Journal of Bacteriology178:2709-2711. 237 Bridges, B.A. (1998) Mutation Research 408:1-9. Bridges, B.A., R.P. Motershead, and S.G. Sedgwick (1976)Molecular and General Genetics 144:53-58. Bridges, B.A., and A.R. Timms (1997) EMBO Journal 16:3349-3356. Bridges, B.A., and R. Woodgate (1984) Molecular and General Genetics 196:364-366. Bridges, B.A., and R. Woodgate (1985)Proceedings of the National Academy of Sciences of the United States of America 82:4193-4197. Brown, G.M. (1962) Journal of Biological Chemistry 237:536-540. Brown, P.O., and N.R. Cozzarelli (1979)Science 206:1081-1083. Bruck, I., R. Woodgate, K. McEntee, and M.F. Goodman (1996) Journal of Biological Chemistry 271:10767-10774. Brumby, A.M., I. Lamont, I B. Dodd, and J.B. Egan (1996)Virology 219:105-114. Brumfitt, W., and J. Hamilton-Miller (1989)New England Journal of Medicine 320:1188-1196. Bryan, L.E. (1988) Journal of Antimicrobial Chemotherapy 22 (Supplement A): 1-15. Bryan, S.K., M. Hagensee, and R E. Moses (1990)Mutation Research 243:313-318. Burgess, R.R., A.A. Travers, J.J. Dunn, and E.K.F. Bautz (1969)Nature 221:43-46. Burgess, R.R., and A.A. Travers (1970) Federation Proceedings 29:1164-1169. Bush, K., and R.B. Sykes (1984) In ‘Antimicrobial Drug Resistance’, pp. 1-31, Ed. Bryan, L.E., Academic Press, Orlando. Cairns, J., and P.L. Foster (1991)Genetics 128:695-701. Cairns, J., J. Overbaugh, andS. Miller (1988) Nature 335:142-145. Calos, M.P., and J.H. Miller (1981)Journal of Molecular Biology153:39-66. Cammarano, P., P. Palm, R. Creti, E. Ceccarelli, A.M. Sanangelantoni, and O. Tiboni (1992) Journal of Molecular Evolution 34:396-405. Caron, P R., and L. Grossman (1988a) Nucleic Acids Research 16:9641-9650. Caron, P R., and L. Grossman (1988b) Nucleic Acids Research 16:9651-9662. 238 Caron, P R., S.R. Kushner, and L. Grossman (1985)Proceedings of the National Academy of Sciences of the United States of America 82:4925-4929. Casaregola, S., R. D’Ari, and O. Huisman(1982a) Molecular and General Genetics 185:430-439. Casaregola, S., R. D’Ari, and O. Huisman(1982b) Molecular and General Genetics 185:440-444. Chambers, H.F. (1988) Clinical Microbiology Reviews 1:173-186. Chase, J.W., and C.C. Richardson (1974)Journal of Biological Chemistry249:4553- 4561. Chaudhury, A.M., and G.R. Smith (1984)Proceedings of the National Academy of Sciences of the United States of America 81:7850-7854. Chaudhury, A.M., and G.R. Smith (1985) Molecular and General Genetics 201:525- 528. Chen, C.-R., M. Malik, M. Snyder, and K. Drlica (1996)Journal of Molecular Biology 258:627-637. Chow, S.A., and C M. Radding (1985) Proceedings of the National Academy of Sciences of the United States of America 82:5646-5650. Clark, A.J. (1982) Biochimie 64:669-675. Clark, A.J. (1991) Biochimie 73:523-532. Clerch, B., J.M. Bravo, and M. Llagostera(1996) Mutation Research 349:201-208. Cohen, S.P., L.M. McMurry, and S B. Levy (1988)Journal of Bacteriology170:5416- 5422. Cohen, S.P., L.M. McMurry, D C. Hooper, J.S. Wolfson, and S B. Levy (1989) Antimicrobial Agents and Chemotherapy 33:1318-1325. Cohen, S.P., W. Yan, and S B. Levy (1993a) Journal of Infectious Diseases 168:484- 488. Cohen, S.P., H. Hachler, and S B. Levy (1993b) Journal of Bacteriology 175:1484- 1492. Cohen, S.P., S B. Levy, J. Foulds, and J.L. Roner (1993c) Journal of Bacteriology 175:7856-7862. 239 Coleman, DC., I. Chopra, S.W. Shales, T.G.B. Howe, and T.J. Foster (1983)Journal of Bacteriology153:921-929. Connolly, B., C.A. Parsons, F.E. Benson, H.J. Dunderdale, G.J. Sharpies, R.G. Lloyd, and S.C. West (1991) Proceedings of the National Academy of Sciences of the United States of America 88:6063-6067. Cook, T.M., K.G. Brown, J.V. Boyle, and W.A. Goss (1966)Journal of Bacteriology 92:1510-1514. Cooksey, R.C., G.P. Morlock, A. McQueen, S.E. Glickmann, and J.T. Crawford (1996) Antimicrobial Agents and Chemotherapy 40:1186-1188. Cooper, D.L., R.S. Lahue, and P. Modrich (1993)Journal of Biological Chemistry 268:11823-11829. Cooper, P.K. (1982) Molecular and General Genetics 185:189-197. Cooper, P.K., and P.C. Hanawalt (1972)Proceedings of the National Academy of Sciences of the United States of America 69:1156-1160. Cox, M.M. (1993) Bioessays 15:617-623. Cox, M.M., and I.R. Lehman (1981)Proceedings of the National Academy of Sciences of the United States of America 78:6018-6022. Craig, N.L., and J.W. Roberts (1980) Nature 283:26-30. Critchlow, S.E., and A. Maxwell (1996) Biochemistry 35:7387-7393. Crumplin, G.C. (1987) Lancet ii:854-855. Crumplin, G.C. (1990a) Reviews in Medical Microbiology 1:67-75. Crumplin, G.C. (1990b) Molecular effects of 4-quinolones upon DNA gyrase: DNA systems. In ‘The 4-Quinolones’, Ed. G.C. Crumplin, pp.53-68, Springer-Verlag, New York. Crumplin, G.C., and J.T. Smith (1976) Nature 260:643-645. Crumplin, G.C., and J.T. Smith (1981)Journal of Antimicrobial Chemotherapy7:379- 388. Cullen, M E., A.W. Wyke, R. Kuroda, and L.M. Fisher (1989)Antimicrobial Agents and Chemotherapy 33:886-894. 240 Cundliffe, E. (1981) In ‘Molecular Basis of Antibiotic Action’, Eds. Gale, E.F., E. Cundliffe, P.E. Reynolds, M. Richmond, and M.J. Waring, pp.405-547, Wiley Interscience, London. Dallas, W.S., J.E. Gowen, P.H. Ray, M.J. Cox, and I.K. Dev (1992)Journal of Bacteriology174:5961-5970. Darwin, C. (1859) “On The Origin of Species by Means of Natural Selection.”, John Murray, London. DasGupta, C., T. Shibata, R.P. Cunningham, and C.M. Radding (1980)Cell 22:437- 446. Datta, N. (1984) British Medical Bulletin 40:1-2. Datta, N., and V.M. Hughes (1983) Nature 306:616-617. Davis, B.D. (1989)Proceedings of the National Academy of Sciences of the United States of America 86:5005-5009. Davis, B.D., and E.S. Mingioli (1950) Journal of Bacteriology 60:17-28. Deitz, W.H., T.M. Cook, and W.A. Goss (1966)Journal of Bacteriology91:768-773. De Groot, R., J. Campos, S.L. Moseley, and A.L. Smith (1988)Antimicrobial Agents and Chemotherapy 32:477-484. De Groot, R., G. Dzoljic-Danilov, B. van Klingeren, W.H. Goessens, and H.J. Neyens (1991a) European Journal of Paediatrics 150: 534-546. De Groot, R., D.O. Chaffin, M. Kuehn, and A.L. Smith (1991b) Biochemical Journal 274:657-662. De Stasio, E.A., D. Moazed, H.F. NoUer, and A.E. Dahlberg (1989)EMBO Journal 8:1213-1216. Devoret, R., A. Bailone, M. Dutriex, P.L. Moreau, S. Sommer, and M. Bagdasarian (1988) In “Mechanisms and Consequences of DNA Damage Processing.”, pp. 437-443, Eds. E.G. Friedberg and P.C. Hanawalt, Alan Liss Inc. Domagala, J.M., L.D. Hanna, C.L. Heifetz, M.P. Hutt, T.F. Mich, J.P. Sanchez, and M. Solomon (1986)Journal of Medicinal Chemistry 29:394-404. Doolittle, R.F., M.S. Johnson, I. Husain, B. Van Houten, D C. Thomas, and A. Sancar (1986)Nature 323:451-453. Douglass, J., and L.M. Steyn (1993)Journal of Infectious Diseases167:1505-1506. 241 Doyle, N., and P. Strike (1995)Molecular and General Genetics 247:735-741. Drabble, W.T., and B A D. Stocker (1968)Journal of General Microbiology53:109- 123. Drlica, K. (1984) Microbiological Reviews 48:273-289. Drlica, K., C. Xu, J.-Y. Wang, R.M. Burger, and M. Malik (1996)Antimicrobial Agents and Chemotherapy 40:1594-1599. Drlica, K., and X. Zhao (1997)Microbiology and Molecular Biology Reviews61:377- 392. Dunderdale, H.J., F.E. Benson, C.A. Parsons, G.J. Sharpies, R.G. Lloyd, and S.C. West (1991) Nature 354:506-510. Dutreix, M., B. Burnett, A. Bailone, CM. Radding, and R. Devoret (1992)Molecular and General Genetics 232:489-497. Dutreix, M., P.L. Moreau, A. Bailone, F. Galibert, J R. Battista, G.C. Walker, and R. Devoret (1989)Journal of Bacteriology171:2415-2423. Echols, H. (1982) Biochimie 64:571-575. Echols, H., and M.F. Goodman (1990)Mutation Research 26:301-311. Echols, H., and M.F. Goodman (1991)Annual Review of Biochemistry60:477-511. Edelstein, P.H., M A C. Edelstein, K.H. Lehr, and J. Ren (1996)Journal of Antimicrobial Chemotherapy 37:117-126. Egelman, E.H., and A. Stasiak(1986) Journal of Molecular Biology 191:677-697. Elledge, S.J., and G.C. Walker (1983)Journal of Bacteriology155:1306-1315. Ennis, D.G., B. Fisher, S. Edmiston, and D.W. Mount (1985)Proceedings of the National Academy of Sciences of the United States of America82:3325-3329. Ennis, D.G., N. Ossanna, and D.W. Mount (1989) Journal of Bacteriology 171:2533- 2541. Escarceller, M., J. Hicks, G. Gudmundsson, G. Trump, D. Touati, S. Lovett, P.L. Foster, K. McEntee, and M.F. Goodman (1994)Journal of Bacteriology176:6221- 6228. Everett, M.J., Y.F. Jin, V. Ricci, and L.J.V. Piddock (1996)Antimicrobial Agents and Chemotherapy 40:2380-2386. 242 Feng, G., H.-C.T. Tsui, and M E. Winkler (1996)Journal of Bacteriology178:2388- 2396. Ferrero, L., B. Cameron, B. Manse, D. Lagneaux, J. Crouzet, A. Fameckon, and F. Blanche (1994) Molecular Microbiology 13:641-653. Ferrero, L., B. Cameron, and J. Crouzet (1995) Antimicrobial Agents and Chemotherapy 39:1554-1558. Fersht, A.R., and J.W. Knill-Jones(1983) Journal of Molecular Biology 165:669-682. Fijalkowska, I.J., R.L. Dunn, and R.M. Schaaper (1997)Journal of Bacteriology 179:7435-7445. Finken, M., P. Kirschner, A. Meier, A. Wrede, and E C. Bottger (1993) Molecular Microbiology 9:1239-1246. Fisher, L.M., J.M. Lawrence, I.C. Josty, R. Hopewell, E E C. Margerrison, and M E. Cullen (1989) The American Journal of Medicine 87 (Supplement 5A):2S-8S. Fleming, M.P., N. Datta, and R. Griineberg (1972) British Medical Journal 1:726-728. Flensberg, J., and O. Skold(1987) European Journal of Biochemistry 162:473-476. Flory, J., and C M. Radding (1982) Cell 28:747-756. Fogliano, M., and P.F. Schendel (1981) Nature 289:196-198. Fontana, R. (1985) Journal of Antimicrobial Chemotherapy 16:412-416. Forster, J.W., and P. Strike (1985) Gene 35:71-82. Forster, J.W., and P. Strike (1988)Molecular and General Genetics 211:531-537. Foster, P.L. (1992)Journal of Bacteriology174:1711-1716. Foster, P.L. (1993) Annual Review of Microbiology 47:467-504. Foster, P.L. (1994)Genetics 138:253-261. Foster, P.L. (1997) Journal of Bacteriology 179:1550-1554. Foster, P.L. (1998) Genetics 148:1453-1459. Foster, P.L. (1999) Mutation Research 436:179-184. Foster, P.L., and J. Cairns (1992) Genetics 131:783-789. 243 Foster, P.L., and J. Cairns (1994) EMBO Journal 13:5240-5244. Foster, P.L., G. Gudmundsson, J.M. Trimarchi, H. Cai, and M.F. Goodman (1995) Proceedings of the National Academy of Sciences of the United States of America 92:7951-7955. Foster, P.L., and A.D. Sullivan (1998) Molecular and General Genetics 214:467-473. Foster, P.L., A.D. Sullivan, and S.B. Franklin (1989)Journal of Bacteriology 171:3144-3151. Foster, P.L., and J.M. Trimarchi(1994) Science 265:407-409. Foster, P.L., and J.M. Trimarchi (1995)Proceedings of the National Academy of Sciences of the United States of America 92:5487-5490. Foster, P.L., J.M. Trimarchi, and R.A. Maurer (1996)Genetics 142:25-37. FrancioUi, M., J. Bille, M.P. Glauser, and P. Moreillon (1991)Journal of Infectious Diseases 163:514-523. Frank, E.G., J. Hauser, A S. Levine, and R. Woodgate (1993)Proceedings of the National Academy of Sciences of the United States of America90:8169-8173. Freitag, N., and K. McEntee (1989)Proceedings of the National Academy of Sciences of the United States of America 86:8363-8367. Friedberg, E.C., G.C. Walker, and W. Siede. 1995. DNA repair and mutagenesis. The American Society for Microbiology, Washington, D.C. Fujimaki, K., T. Fujii, H. Aoyama, K-I. Sato, Y. Inoue, M. Inoue, and S. Mitsuhashi (1989) Antimicrobial Agents and Chemotherapy 33:785-787. Funatsu, G., and H.G. Wittman (1972)Journal of Molecular Biology68:547-550. Gaja, M. (1992)Chemotherapy 40:1-10. Gale, E.F., E. Cundliffe, P.E. Reynolds, M.H. Richmond, and M.J. Waring (1981) The molecular basis of antibiotic action. John Wiley and Sons, Inc., New York. Galitski, T., and J R. Roth (1995) Science 268:421-423. Galitski, T., and J R. Roth (1996) Genetics 143:645-659. Gellert, M. (1981) Annual Review of Biochemistry 50:879-910. 244 Gellert, M., K. Mizuuchi, M.H. O’Dea and H.A. Nash (1976)Proceedings of the National Academy of Sciences of the United States of America73:3872-3876. Gellert, M., K. Mizuuchi, M.H. O’Dea, T. Itoh, and J-I. Tomizawa (1977) Proceedings of the National Academy of Sciences of the United States of America 74:4772-4776. George, A.M., and S.B. Levy (1983a) Journal of Bacteriology 155:541-548. George, A.M., and S.B. Levy (1983b) Journal of Bacteriology155:531-540. Georgopapadakou, N.H., and F.Y. Liu (1980) Antimicrobial Agents and Chemotherapy 18:834-836. Georgopapadakou, N.H., and R.B. Sykes (1983)Bacterial enzymes interacting with P-lactam antibiotics. In ‘Antibiotics containing the beta lactam structure II. Handbook of Experimental Pharmacology, Vol67/11’, Eds. Demain, E.L., and N.A. Solomon, pp.1-77. Springer-Verlag, Berlin. Gerberding, J.L., C. Miick, H.H. Liu, and H. Chambers (1991)Antimicrobial Agents and Chemotherapy 35:2574-2579. Gill, S., and V.N. Iyer (1982) Canadian Journal of Microbiology28:256-258. Ginsburg, H., S.H. Edmiston, J. Harper, and D.W. Mount (1982)Molecular and General Genetics 187:4-11. Glickman, B.W., R.M. Schaaper, W.A. Haseltine, R.L. Dunn, and D.E. Brash (1986) Proceedings of the National Academy of Sciences of the United States of America 83:6945-6949. Glupczynski, Y., M. Labbe, A. Burette, M. Delmee, V. Avesani, and C. Bruck (1987) Lancet i:1096. Gocke, E. (1991)Mutation Research 248:135-143. Goldman, J., D.G. White, and S.B. Levy (1996) Antimicrobial Agents and Chemotherapy 40:1266-1269. Gonda, D.K., and C M. Radding (1986)Journal of Biological Chemistry261:13087- 13096. Gootz, T.D., and K.E. Brighty (1996) Medicinal Research Reviews 16:433-486. Gootz, T.D., and B.A. Martin (1991) Antimicrobial Agents and Chemotherapy 35:840- 845. 245 Gorden, L.K., and W.A. Haseltine (1982).Radiation Research 89:99-112. Goss, W.A., W.H. Deitz, and T.M. Cook (1964)Journal of Bacteriology88:1112-1118. Goss, W.A., W.H. Deitz, and T.M. Cook (1965) Journal of Bacteriology 89:1068-1074. Granger-Schnarr, M., M. Schnarr, and C.A. van Sinis (1986)FEBS Letters 198:61- 65. Griffin, T.J. (TV), and R.D. Kolodner (1990) Journal of Bacteriology 172:6291-6299. Griggs, D.J., K. Gensberg, and L.J.V. Piddock (1996)Antimicrobial Agents and Chemotherapy 40:1009-1013. Grilley, M., J. Griffith, and P. Modrich (1993) Journal of Biological Chemistry 268:11830-11837. Grilley, M., K.M. Welsh, S.S. Su, and P. Modrich (1989)Journal of Biological Chemistry 264:1000-1004. Griineberg, R.N. (1992)Journal of Medical Microbiology36:23-25. Gruz, P., K. Matsui, T. Sofuni, and T. Nohmi (1996)Mutation Research 354:157-170. Gruz, P., K. Matsui, T. Sofuni, and T. Nohmi (1998)Mutation Research 398:33-42. Guardiola, J., M. De Felice, and M. laccarino (1974)Journal of Bacteriology120:536- 538. Gudas, L.J., and A.B. Pardee (1975)Proceedings of the National Academy of Sciences of the United States of America 72:2330-2334. Gudas, L.J., and A.B. Pardee(1976) Journal of Molecular Biology 101:459-477. Gustafson, J.E., B. Berger-Bachi, A Srassle, and B. Wilkinson (1992)Antimicrobial Agents and Chemotherapy 36:566-572. Gustafson, J.E., and B. Wilkinson (1989)FEMS Microbiology Letters59:107-112. Haber, L.T., and G.C. Walker (1991) EMBO Journal 10:2707-2715. Hachler, H., S.P. Cohen, and SB. Levy (1991)Journal of Bacteriology173:5532-5538. Hagensee, M E., T.L. Timme, S.K. Bryan, and R E. Moses (1987)Proceedings of the National Academy of Sciences of the United States of America84:4195-4199. Hall, B.G. (1988) Genetics 120:887-897. 246 HaU, B.G. (1989) Genome 31:265-271. Hall, B.G. (1990) Genetics 126:5-16. Hall, B.G. (1991) Proceedings of the National Academy of Sciences of the United States of America 88:5882-5886. Hall, B.G. (1992) Proceedings of the National Academy of Sciences of the United States of America 89:4300-4303. Hall, B.G. (1993)Journal of Bacteriology175:6411-6414. Hall, B.G. (1994)Molecular Biology and Evolution11:159-168. Hall, B.G. (1995a) Journal of Molecular Evolution 40:86-93. Hall, B.G. (1995b) Proceedings of the National Academy of Sciences of the United States of America 92:5669-5673. Hallett, P., and A. Maxwell (1991) Antimicrobial Agents and Chemotherapy 35:335- 340. Hamilton-Miller, J.M.T. (1984) Resistance to antibacterial agents acting on antifolate metabolism, p. 173-190. In L.E. Bryan (ed.), Antimicrobial Drug Resistance. Academic Press Inc., New York. Hanawalt, P.C., P.K. Cooper, A.K. Ganeson, and C.A. Smith(1979) Annual Review of Biochemistry 48:783-836. Hane, M.W., and T.H. Wood (1969) Journal of Bacteriology 99:238-241. Harris, R.S., G.Feng, K.J. Ross, R. Sidhu, C. Thulin, S. Longerich, S.K. Szigety, M.E. Winkler, and S.M. Rosenberg (1997a) Genes and Development 11:2426-2437. Harris, R.S., H.J. Bull, and S.M. Rosenberg(1997b) Mutation Research 375:19-24. Harris, R.S., S. Longerich, and S.M. Rosenberg (1994) Science 264:258-260. Harris, R.S., K.J. Ross, and S.M. Rosenberg(1996) Genetics 142:681-691. Hartman, B.J., andA. Tomasz (1984) Journal of Bacteriology 158:513-516. Hartman, B.J., and A. Tomasz(1986) Antimicrobial Agents and Chemotherapy 29:85- 92. Hartmann, G., R.O. Honikel, R. Kniisel, and J. Niiesch (1967)Biochimica et Biophysica Acta. 145:843-844. 247 Haseltine, W.A. (1983) Cell 33:13-17. Hauser, J,, A.S. Levine, D.G. Ennis, K.M. Chumakov, and R. Woodgate (1992) Journal of Bacteriology174:6844-6851. Heikkila, E. (1991)Ann. Univ. Turkuensis. Ser. D 79:1-68. Heil, A., and W. Zillig (1970)FEBS Letters 11:165-168. Heisig, P (1996)Antimicrobial Agents and Chemotherapy 40:879-885. Heisig, P., B. Kratz, E. Haile, Y. Graser, M. Altwegg, W. Rabsch, and J.-P. Faber (1995) Microbial Drug Resistance 1:211-218. Heisig, P., H. Schedletzky, and H. Falkenstein-Paul (1993) Antimicrobial Agents and Chemotherapy 37:696-701. Heisig, P., L. Tzouvelekis, A. Vatopoulos, and N. Legakis (1994)Journal of Antimicrobial Chemotherapy 33:887-888. Herendeen, D R., and T.J. Kelly (1996) Cell 84:5-8. Heym, B., N. Honore, C. Truffot-Pernot, A. Bannerjee, C. Schurra, W.R. Jacobs, Jr., J.D.A. von Embden, J.H. Grosset, and S.T. Cole(1994) Lancet 344:293-298. Hiasa, H., R. DiGate, and K. Marians (1994)Journal of Biological Chemistry 269:2093-2099. Hiasa, H., and K. Marians (1994) Journal of Biological Chemistry 269:16371-16375. Hinton, J., and R.J. Pinney (1997a) Journal of Pharmacy and Pharmacology 49 (Supplement 4): 13. Hinton, J., and R.J. Pinney (1997b) Journal of Pharmacy and Pharmacology 49 (Supplement 4): 13. Hirai, K., H. Aoyama, S. Suzue, T. Irikura, S. lyobe, and S. Mitsuhashi (1986a) Antimicrobial Agents and Chemotherapy 30:248-253. Hirai, K., H. Aoyama, T. Irikura, S. loybe, and S. Matsuhashi (1986b) Antimicrobial Agents and Chemotherapy 29:535-538. Hirai, K., B. Irikura, S. lyobe, and S. Mitsuhashi (1984)Chemotherapy 32:471-476. Hirai, K., S. Suzue, T. Irikura, S. lyobe, and S. Mitsuhashi (1987)Antimicrobial Agents and Chemotherapy 31:582-586. 248 Holliday, R. (1964) Genetics Research 5:282-304. Honore, N., and S.T. Cole (1993) Antimicrobial Agents and Chemotherapy 37:414-418. Honore, N., and S.T. Cole (1994) Antimicrobial Agents and Chemotherapy 38:238-242. Hooper, D C., C.C. Sanders, and J.S. Wolfson(1988) Reviews of Infectious Diseases 10:516-527. Hooper, D C., and J.S. Wolfson (1989)Reviews of Infectious Diseases11 (Supplement 5):5902-5911. Hooper, D C., and J.S. Wolfson (1991) European Journal of Clinical Microbiology and Infectious Diseases 10:223-231. Hooper, D C., J.S. Wolfson, M.A. Bozza, and E.Y. Ng (1992)Antimicrobial Agents and Chemotherapy 36:1151-1154. Hooper, D C., J.S. Wolfson, E.Y. Ng, and M. Swartz (1987) American Journal of Medicine 82 (Supplement 4A): 12-20. Hooper, D C., J.S. Wolfson, K.S. Souza, E.Y. Ng, G.L. McHugh, and M.N. Swartz (1989) Antimicrobial Agents and Chemotherapy 33:283-290. Hooper, D C., J.S. Wolfson, K.S. Souza, C. Tung, G.L. McHugh, and M.N. Swartz (1986) Antimicrobial Agents and Chemotherapy 29:639-644. Horii, Z.I., and A.J. Clark (1973)Journal of Molecular Biology80:327-344. Horowitz, D.S., and J.C. Wang (1987)Journal of Biological Chemistry262:5339-5344. Hoshino, K., A. Kitamura, I. Morrissey, K. Sato, J-I. Kato, and H. Ikeda (1994) Antimicrobial Agents and Chemotherapy 38:2623-2627. Howard, B.M., R.J. Pinney, and J.T. Smith (1993a) Arzneimittel-Forschung Drug Research 43:1125-1129. Howard, B.M., R.J. Finney, and J.T. Smith (1993b)Journal of Pharmacy and Pharmacology 45:658-662. Howard-Flanders, P. (1968) Advances in Biological and Medical Physics 12:299-317. Howard-Flanders, P. (1981) Scientific American 245:72-80. Howard-Flanders, P., R.P. Boyce, and L. Theriot (1965)Genetics 53:1119-1136. Howard-Flanders, P., E. Simpson, and L. Theriot (1964)Genetics 49:237-246 249 Howard-Flanders, P., and L. Theriot (1966)Genetics 53:1137-1150. Howard-Flanders, P., S.C. West, and A. Stasiak (1984)Nature 309:215-219. Hughes, V.M., and N. Datta (1983)Nature 302:725-726. Huisman, O., and R. D’Ari (1981) Nature 290:797-799. Huovinen, P., L. Sundstrom, G. Swedberg, and O. Skold (1995)Antimicrobial Agents and Chemotherapy 39:279-289. Husain, L, B. Van Houten, D C. Thomas, M. Abdel-Monem, and A. Sancar (1985) Proceedings of the National Academy of Sciences of the United States of America 82:6774-6778. Inouye, M., and S. Inouye (1991) Annual Review of Microbiology45:163-186. Inoue, Y., K. Sato, T. Fujii, D. Hirai, M. Inoue, S. lyobe, and S. Mitsuhashi (1987) Journal of Bacteriology169:2322-2325. Irbe, R.M., and M. Oishi (1980)Journal of Bacteriology144:1061-1067. Isaksson, L.A., and R. Takata (1978) Molecular and General Genetics 161:9-14. Ito, A., K. Hirai, M. Inoue, H. Koga, S. Suzue, T. Irikura, and S. Mitsuhashi (1980) Antimicrobial Agents and Chemotherapy 17:103-108. Ito, H., H. Yoshida, M. Bogaki-Shonai, T. Niga, H. Haffori, and S. Nakamura(1994) Antimicrobial Agents and Chemotherapy 38:2014-2023. Ito, T., M. Otsuki, and T. Nishino (1992) Antimicrobial Agents and Chemotherapy 36:1708-1714. Iwakura, Y., A. Ishikama, and T. Yura (1973) Molecular and General Genetics 121:181-196. Iwasaki, H., A. Nakata, G.C. Walker, and H. Shinagawa (1990)Journal of Bacteriology172:6268-6273. Iwasaki, H., T. Shiba, K. Makino, A. Nakata, and H. Shinagawa (1989)Journal of Bacteriology171:5276-5280. Iwasaki, H., M. Takahagi, A. Nakata, and H. Shinagawa (1992)Genes and Development 6:2214-2220. Iwasaki, H., M. Takahagi, T. Shiba, A. Nakata, and H. Shinagawa (1991)EMBO Journal 10:4381-4389. 250 Iwatsukî, N., C.O. Joe, and H. Werbin (1980) Biochemistry 19:1172-1176. Jacoby, G.A. (1994) Trends in Microbiology 2:357-360. Jacoby, G.A., and A.A. Medeiros (1991) Antimicrobial Agents and Chemotherapy 35:1697-1704. Jayaraman, R. (1992)Journal of Genetics71:23-41. Jayaraman, R. (1995) Journal of Genetics 74:85-97. Jin, D.J., and C.A. Gross (1988)Journal of Molecular Biology202:45-58. Johanson, U., A. Ævarsson, A. Lilias, and D. Hughes (1996)Journal of Molecular Biology 258:420-432. Johanson, V., and D. Hughes(1994) Gene 143:55-59. Johnson, R.C., and W.F. McNeill (1978)Electron microscopy of UV-induced postreplication repair of daughter strand gaps, p. 95-99.In PC. Hanawalt, E.C. Friedberg, and C.F. Fox (ed.), DNA Repair Mechanisms. Academic Press, Inc., New York. Johnston, D.E., and W.R. McClure (1975) Abortive initiation of in vitro RNA synthesis on bacteriophage lambda DNA polymerase. In ‘RNA Polymerase’, Eds. Losick, R., and M. Chamberlain, pp. 413-428. Cold Spring Harbour Laboratory, Cold Spring Harbour, New York. Jonczyk, P., I. Fualkowska, and Z. Ciesla (1988) Proceedings of the National Academy of Sciences of the United States of America 85:9124-9127. Kacinski, B.M., and W.D. Rupp (1981) Nature 294:480-481. Kahn, R., P R. Cunningham, C. DasGupta, and C M. Radding (1981)Proceedings of the National Academy of Sciences of the United States of America 78:4786-4790. Kanabus, M., A. Nowicka, E. Sledziewska-Gojska, P. Jonczyk, and Z. Ciesla(1995) Molecular and General Genetics 247:216-221. Kapur, V., L.-L. Li, M R. Hanu*ick, B.B. Plikaytis, T.M. Shinnick, A. Telenti, W.R. Jacobs, Jr., A. Banner]ee, S. Cole, K.Y. Yuen, J.E. Clarridge HI, B.N. Kreiswirth, and J.M. Musser (1995) Archives of Pathology and Laboratory Medicine 119:131-138. Karu, A.E., and E.D. Belk (1982) Molecular and General Genetics 185:275-282. Kato, J-I., Y. Nishimura, R. Imamura, H. Niki, S. Hiraga, and H. Suzuki(1990) Cell 63:393-404. 251 Kato, J-I., H. Suzuki, and H. Ikeda (1992)Journal of Biology Chemistry267:25676- 25684. Kato, T., and Y. Shinoura (1977) Molecular and General Genetics 156:121-131. Katsukawa, C., A. Tamaru, Y. Miyata, C. Abe, M. Makino, and Y. Suzuki (1997) Journal of Applied Microbiology 83:634-640. Kelman, Z., and M. O’Donnell (1995) Annual Review of Biochemistry 64:171-200. Kenyon, C.J., and G.C. Walker (1981) Nature 289:808-810. Khidhir, M.A., S. Casaregola, and I.E. Holland (1985)Molecular and General Genetics 199:133-140. Khodursky, A.B., E.L. Zechiedrich, and N.R. Cozzarelli (1995)Proceedings of the National Academy of Sciences of the United States of America92:11801-11805. King, CH., DM . Shlaes, and M.J. Dul (1983)Journal of Clinical Microbiology18:79- 83. Kitagawa, Y., E. Akaboshi, H. Shinagawa, T. Horii, H. Ogawa, and T. Kato (1985) Proceedings of the National Academy of Sciences of the United States of America 82:4336-4340. Koga, H., A. Itoh, S. Murayama, S. Suzue, and T. Irikura (1980)Journal of Medicinal Chemistry 23:1358-1363. Kohno, K., T. Uchida, H. Ohkubo, S. Nakanishi, T. Nakanishi, T. Fukui, E. Ohtsuka, M. Ikehara, and Y. Okada (1986)Proceedings of the National Academy of Sciences of the United States of America 83:4978-4982. Kondo, S., H. Ichikawa, K. Iwo, and T. Kato (1970) Genetics 66:187-217. Kow, Y.W., G. Faundez, S. Hays, C.A. Bonner, M.F. Goodman, and S.S. Wallace (1993) Journal of Bacteriology175:561-564. Kowaiczykowski, S.C. (1991) Biochimie 73:289-304. Kramer, B., W. Kramer, and H.J. Fritz (1984)Cell 38:879-887. Krasin, F., and F. Hutchinson (1977)Journal of Molecular Biology116:81-98. Krawiec, S., and M. Riley (1990)Microbiological Reviews54:502-539. Kresken, M., and B. Wiedemann (1988)Antimicrobial Agents and Chemotherapy 32:1285-1288. 252 Kristiansen, B.-E., P.Râdstrôm, A. Jenkins, E. Ask, B. Facinelli, and O. Skold (1990) Antimicrobial Agents and Chemotherapy 34:2277-2279. Kuemmerle, N., R. Ley, and W. Masker (1981) Journal of Bacteriology 147:333-339. Kiies, U., and U. Stahl (1989) Microbiological Reviews 53:491-516. Kulaeva, O.I., E.V. Koonin, J.C. Wooton, A S. Levine, and R. Woodgate (1998) Mutation Research 397:247-262. Kulaeva, O.I., J.C. Wooton, A S. Levine, and R. Woodgate (1995)Journal of Bacteriology 177:2737-2743. Kumagai, Y., J.I. Kato, K. Hoshino, T. Akasaka, K. Sato, and H. Ikeda (1996) Antimicrobial Agents and Chemotherapy 40:710-714. Kumura, K., M. Seguchi, A.L. Steinum, and E. Seeberg (1985)Nucleic Acids Research 13:1483-1492. Kushner, S.R. (1974)Journal of Bacteriology120:1219-1222. Kuwano, M., D. Schlesinger, G. Rinaldi, L. Felicetti, and G.P. Tocchini (1971) Biochemical and Biophysical Research Communications42:441-444. Lahue, R.S., K.G. An, and P. Modrich (1989)Science 245:160-164. Lahue, R.S., S.S. Su, and P. Modrich (1987)Proceedings of the National Academy of Sciences of the United States of America 84:1482-1486. Lai, C.J., and B. Weisblum (1971)Proceedings of the National Academy of Sciences of the United States of America 68:856-860. Lai, C.J., J.E. Dahlberg, and B. Weisblum (1973a)Biochemistry 12:457-460. Lai, C.J., B. Weisblum, S.R. Fahnestock, and M. Namur a (1973b)Journal of Molecular Biology74:67-72. Lam, L.H., and R.J. Reynolds (1987) Mutation Research 178:167-176. Lawrence, C.W., A. Borden, S.K. Bannerjee, and J.E. LeClerc (1990)Nucleic Acids Research 18:2153-2157. Lawrence, C.W., A. Borden, and R. Woodgate (1996)Molecular and General Genetics 251:493-498. Lederberg, J., and E.M. Lederberg (1952)Journal of Bacteriology63:399-406. 253 Lehman, I.R. (1974) Science 186:790-797. Lehman, I.R., and A L. Nussbaum (1964) Journal of Biological Chemistry 239:2628- 2636. Lesher, G.Y., E.J. Froelich, M.D. Gruet, J.H. Bailey, and R.P. Brundage (1962) Journal of Medicinal and Pharmaceutical Chemistry 5:1063-1068. Levit, M (1982)Cold Spring Harbour Symposium Quantitative Biology47:251-262. Lewin, C.S., B.M.A. Howard, N.T. Ratcliffe, and J.T. Smith (1989)Journal of Medical Microbiology 29:139-144. Lewin, C., B. Howard, and J.T. Smith (1991)Journal of Medical Microbiology34:19- 22. Lewin, C., and J.T. Smith (1990)Journal of Medical Microbiology31:65-70. Li, X.-Z., D.M. Livermore, and H. Nikaido (1994)Antimicrobial Agents and Chemotherapy 38:1732-1741. Lichtenstein, C., and S. Brenner (1981) Molecular and General Genetics 183:380-387. Liljas, A. (1982) Progress in Biophysics and Molecular Biology 40:161-228. Lilley, D.M.J. (1992)Nature 357:282-283. Lin, J-J., A.M. Phillips, J.E. Hearst and A. Sancar(1992) Journal of Biological Chemistry 267:17693-17700. Lin, J-J., and A. Sancar(1992) Molecular Microbiology 6:2219-2224. Lisitsyn, N.A., E.D. Sverdlov, E.P. Moiseyeva, O.N. Danilevskaya, and V.G. Nikifurov (1984)Molecular and General Genetics 196:173-174. Little, J.W. (1983) Journal of Molecular Biology 167:791-808. Little, J.W. (1984)Proceedings of the National Academy of Sciences of the United States of America 81:1375-1379. Little, J.W. (1991)Biochimie 73:411-422. Little, J.W., and D.W. Mount (1982) Cell 29:11-22. Little, J.W., D.W. Mount, and C.R. Yanisch-Perron (1981)Proceedings of the National Academy of Sciences of the United States of America78:4199-4203. 254 Livneh, Z., O. Cohen-Fix, R. Skaliter, and T. Elizur (1993) Critical Reviews in Biochemistry and Molecular Biology28:465-513. Lloyd, R.G. (1991) Journal of Bacteriology173:5414-5418. Lloyd, R.G., and C. Buckman (1991) Journal of Bacteriology173:1004-1011. Lloyd, R.G., S.M. Picksley, and C. Prescott (1983) Molecular and General Genetics 189:157-161. Lloyd, R.G., and G.J. Sharpies (1991) Journal of Bacteriology 173:6837-6843. Lloyd, R.G., and G.J. Sharpies (1993a) Nucleic Acids Research 21:1719-1725. Lloyd, R.G., and G.J. Sharpies (1993b) EMBO Journal 12:17-22. Lloyd, R.G., and A. Thomas (1983) Molecular and General Genetics 190:156-161. Lodwick, D., D. Owen, and P. Strike (1990)Nucleic Acids Research 18:5045-5050. Loeb, L.A., and B.D. Preston (1986) Annual Review of Genetics 20:201-230. Longerich, S., A.M. Galloway, R.S. Harris, C. Wong, and S.M. Rosenberg (1995) Proceedings of the National Academy of Sciences of the United States of America 92:12017-12020. Lovett, ST., and A.J. Clark(1983) Journal of Bacteriology 153:1471-1478. Lovett, S T., and R.D. Kolodner (1989)Proceedings of the National Academy of Sciences of the United States of America 86:2627-2631. Lu, A-L., S. Clark, and P. Modrich (1983) Proceedings of the National Academy of Sciences of the United States of America 80:4639-4643. Lu, A-L., K. Welsh, S. Clark, S.S. Su, and P. Modrich (1984) Cold Spring Harbour Symposium Quantitative Biology 49:589-596. Luria, S.E., and M. Delbrück (1943) Genetics 28:491-513. Lyons, S.M., and P.P. Schendel (1984) Journal of Bacteriology 159:421-423. Ma, D., D.N. Cook, M. Alberti, N.G. Pon, H. Nikaido, and J.E. Hearst (1995) Molecular Microbiology16:45-55. Maenhaut-Michel, G., and J.A. Shapiro(1994) EMBO Journal 13:5229-5239. 255 Maidhoff, H., B Reinecke, P. Blumel, B. Berger-Bâchi, and H. Labischinski (1991) Journal of Bacteriology173:3507-3513. Maki, H., and A. Kornberg (1985) Journal of Biological Chemistry 260:12987-12992. Malengreau, M. (1984) The Lancet ii:172. Mandai, T.N., A.A. Mahdi, G.J. Sharpies, and R.G. Lloyd (1993)Journal of Bacteriology175:4325-4334. Maneewannakul, K., and S B. Levy (1996)Antimicrobial Agents and Chemotherapy 40:1695-1698. Marinus, M.G. (1976)Journal of Bacteriology128:853-854. Marinus, M.G., and N.R. Morris (1975)Mutation Research 28:15-26. Marinus, M.G., A. Poteete, and J.A. Arraj (1984)Gene 28:123-125. Maskeil, R., O.A. Okubadejo, R.H. Payne, and L. Pead (1977) Journal of Medical Microbiology 11:33-45. Matsuhashi, M., M.D. Song, F. Ishino, M. Wachi, M. Doi, M. Inoue, K. Ubukata, N. Yamashita, and M. Konno (1986) Journal of Bacteriology 167:975-980. Maxwell, A (1992)Journal of Antimicrobial Chemotherapy30:409-414. McCann, J., N.E. Springarn, J. Kobori, and B.N. Ames (1975)Proceedings of the National Academy of Sciences of the United States of America72:979-983. McCormack, J.G (1983) The Lancet ii:1091. McDaniel, L., L.H. Rogers, and W.E. Hill (1978)Journal of Bacteriology134:1195- 1198. McNally, K.P., N.E. Freitag, and G.C. Walker (1990)Journal of Bacteriology 172:6223-6231. McPartland, A., L. Green, and H. Echols (1980)Cell 20:731-737. Mehtar, S. (1990)Quinolones Bulletin6:21-23. Mehtar, S., P H. Blakemore, and K. Ellis (1987) American Journal of Medicine 82 (Supplement 4A):55-57. Meier, A., P. Kirschner, F.-C. Bnge, U. Vogel, and E.C. Bottger (1994)Antimicrobial Agents and Chemotherapy 38:228-233. 256 Meselson, M., P. Fukkîla, M. Rykowski, J. Peterson, M. Radman, R. Wagner, G. Herman, and P. Modrich (1980) Journal of Supramolecular Structure 4 (Supplement): 311. Meyer, R.R., and P.S. Laine (1990)Microbiological Reviews54:342-380. Michel-Briand, Y., V. Uccelli, J-M. Laporte, and P. Plesiat (1986)Journal of Antimicrobial Chemotherapy 18:667-674. Miles, A.A., and S.S. Misra (1938)Journal of Hygiene38:732-748. Miller, C.P., and M. Bohnhoff (1974)Journal of Bacteriology54:467-475. Miller, J.H. (1985) Journal of Molecular Biology 182:45-65. Mitchison, D.A., and A.J. Nunn (1986)American Review of Respiratory Disease133: 423-430. Mittler, J.E., and R.E. Lenski (1990)Nature 344:173-175. Mittler, J.E., and R E Lenski (1992) Nature 356:446-448. Mizuuchi, K., M. Mizuuchi, and M. Gellert (1982)Journal of Molecular Biology 156:229-243. Modrich, P. (1991) Annual Review of Genetics 25:229-253. Moellering, R.C., Jr. (1996)Chemotherapy 42 (Supplement 1):54-61. Molina, A.M., N. Babudri, M. Tamaro, S. Venkerini, and C. Monti-Bragadin (1979) FEMS Microbiology Letters5:33-37. Monk, J.P., and D.M. Campoli-Richards (1987) Drugs 33:346-391. Monti-Bragadin, C., N. Babudri, and L.S. Samer (1976)Molecular and General Genetics 145:303-306. Moolenaar, G.E., C.A. van Sluis, C. Backendorf, and P. van den Putte (1987)Nucleic Acids Research 15:4273-4289. Moorhead, P.J., and HE. Parry (1965) British Medical Journal 2:913-915. Moreau, N.J., H. Robaux, L. Baron, and X, Tabary (1990) Antimicrobial Agents and Chemotherapy 34:1955-1960. Morris, S., G.H. Bai, P. Suffys, L. Portillo-Gomez, M. Fairchok, and D. Rouse (1995) Journal of Infectious Diseases171:954-960. 257 Morrison, A., and N.R. Cozzarelli (1981)Proceedings of the National Academy of Sciences of the United States of America 78:1416-1420. Mortelmans, K.E., and B.A.D. Stocker (1976) Journal of Bacteriology 128:271-282. Mortelmans, K.E., and B.A.D. Stocker (1979) Molecular and General Genetics 167:317-327. Millier, B., I.R. Tsaneva, and S.C. West (1993a) Journal of Biological Chemistry 268:17179-17184. Müller, B., I.R. Tsaneva, and S.C. West (1993b) Journal of Biological Chemistry 268:17185-17189. MUller-Hill, B., and J. Kania(1974) Nature 249:561-563. Munoz, R., and A.G. de la Campa (1996) Antimicrobial Agents and Chemotherapy 40:2252-2257. Munshi, M.H., D A. Sack, K. Harder, Z.U. Ahmed, M.M. Rahaman, and M.G. Morshed (1987) Lancet ii:419-421. Murakami, K., and A. Tomasz (1989) Journal of Bacteriology 171:874-879. Muskavitch, K.M.T., and S. Linn (1981) Enzymes 14:233-250. Myers, R.S., and F.W. Stahl (1994) Annual Review of Genetics 28:49-70. Myles, G.M., and A. Sancar (1991) Biochemistry 30:3834-3840. Nair, J., D A. Rouse, G.H. Bai, and S.L. Morris (1993)Molecular Microbiology 10:521-527. Nakamura, S., M. Nakamura, T. Kojima, and H. Yoshida (1989)Antimicrobial Agents and Chemotherapy 33:254-255. Nakaya, R., A. Nakamura, and Y. Murata (1960)Biochemical and Biophysical Research Communications3:654-659. Napolitano, R.L., I B. Lambert, and R.P.P. Fuchs (1997)Proceedings of the National Academy of Sciences of the United States of America 94:5733-5738. Navaratnam, S., G.M. Myles, R.W. Strange, and A. Sancar (1989)Journal of Biological Chemistry 264:16067-16071. Nevers, P., and H.C. Spatz (1975) Molecular and General Genetics 139:233-243. 258 Ng, E.Y., M. Trucksis, and D C. Hooper (1996) Antimicrobial Agents and Chemotherapy 40:1881-1888. Nikaido, H. (1989) Antimicrobial Agents and Chemotherapy 33:1831-1836. Nikaido, H. (1994) Science 264:382-393. Nohmi, T., J.R. Battista, L.A. Dodson, and G.C. Walker (1988)Proceedings of the National Academy of Sciences of the United States of America85:1816-1820. Nohmi, T., A. Hakura, Y. Nakai, M. Watanabe, S.Y. Murayama, and T. Sofuni. (1991) Journal of Bacteriology 173:1051-1063. Nohmi, T., M. Yamada, M. Matsui, K. Matsui, M. Watanabe, and T. Sofuni(1995a) Molecular and General Genetics 247:7-16. Nohmi, T., M. Yamada, K. Matsui, M. Watanabe, and T. Sofuni(1995b) Mutation Research 329:1-9. Nohmi, T., M. Yamuda, M. Watanabe, S.Y. Murayama, and T. Sofuni (1992) Journal of Bacteriology 174:6948-6955. Nordstrom, K. (1985) In ‘Plasmids in Bacteria.’, pp. 119-123, Eds. Helinski, D.R., S.N. Cohen, D.B. Clewell, D.A. Jackson, A. Hollaender, L. Hagar, S. Kaplan, J. Konisky, and C M. Wilson, Plenem Press, New York and London. Novick, R.P. (1969) Bacteriological Reviews 33:210-235. Nowicka, A., M. Kanabus, E. Sledziewska-Gojska, and Z. Ciesla (1994)Molecular and General Genetics 243:584-592. Oh, E.Y., L. Classen, S. Thiagalingam, S. Mazur, and L. Grossman (1989)Nucleic Acids Research 17:4145-4159. Oishi, M., C.L. Smith, and B. Friefeld (1979)Cold Spring Harbour Symposium Quantitative Biology43:897-907. Okusu, H., D. Ma, and H. Nikaido (1996)Journal of Bacteriology178:306-308. Oram, M., and L.M. Fisher (1991) Antimicrobial Agents and Chemotherapy 35:387- 389. Orren, D.K., and A. Sancar (1989a) University College Los Angeles Symposium of Molecular Cell Biology 83:87-94. Orren, D.K., and A. Sancar (1989b) Proceedings of the National Academy of Sciences of the United States of America 85:8785-8789. 259 Orren, D.K., C.P. Selby, J.E. Hearst, and A. Sancar (1992)Journal of Biological Chemistry 267:780-788. Ovchinnikov, Y.A., G.S. Monastyrskaya, V.V. Gubanov, V.M. Lipkin, E.D. Sverdlov, I.E. Kiver, I.A. Bass, S.Z. Mindlin, O.N. Danilevskaya, and R.B. Khesin (1981) Molecular and General Genetics 184:536-538. Ovchinnikov, Y.A., G.S. Monastyrskaya, SO. Guriev, N.F. Kalinina, ED. Sverdlov, A.I. Gragerov, I.A. Bass, I.E. Kiver, E.P. Moiseyeva, V.N. Igumnov, S.Z. Mindlin, V.G. Nikiforov, and R.B. Kesin (1983) Molecular and General Genetics 190:344-348. Pan, X.-S., J. Ambler, S. Mehtar, and L.M. Eisher (1996)Antimicrobial Agents and Chemotherapy 40:2321-2326. Panhotra, B.R., B. Desai, and P.L. Sharma(1985) Lancet i:763. Parker, L.L., P.W. Betts, and B.G. Hall (1988)Journal of Bacteriology170:218-222. Parker, L.L., and B.G. Hall (1988)Genetics 119:485-490. Parry, H E. (1983) The Lancet ii:1206. Parsons, C.A., I. Tsaneva, R.G. Lloyd, and S.C. West (1992)Proceedings of the National Academy of Sciences of the United States of America89:5452-5456. Parsons, C.A., and S.C. West (1993) Journal of Molecular Biology 232:397-405. Pato, M L., and G.. Brown (1963) Archives of Biochemistryand Biophysics 103:443- 448. Paz-Elizur, T., M. Takeshita, M. Goodman, M. O Donnell, and Z. Livneh (1996) Journal of Biological Chemistry271:24662-24669. Peng, H., and K.J. Marians (1993)Journal of Biological Chemistry268:24481-24490. Peng, H., and K.J. Marians(1995) Journal of Biological Chemistry 270:25286-25290. Perry, K.L., S.J. Elledge, B.B. Mitchell, L. Marsh, and G.C. Walker (1985) Proceedings of the National Academy of Sciences of the United States of America 82:4331-4335. Perry, K.L., and G.C. Walker (1982) Nature 300:278-281. Peters, J.E., and S.A. Benson (1995)Journal of Bacteriology177:847-850. Peterson, K.R., N. Ossanna, A T. Thliveris, D.G. Ennis, and D.W. Mount (1988) Journal of Bacteriology170:1-4. 260 Philippon, A., R. Labia, and G. Jacoby (1989)Antimicrobial Agents and Chemotherapy 33:1131-1136. Phillips, I., E. Culebras, F. Moreno, and P. Baquero (1987)Journal of Antimicrobial Chemotherapy 20:631 -638. Piddock, L.J.V., R.N. Walters, and J.M. Diver (1990)Antimicrobial Agents and Chemotherapy 34:2331-2336. Piddock, L.J.V., W.J.A. Wijnands, and R. Wise(1987) Lancet ii:907. Piddock, L.J.V., and R.N. Wise (1987) FEMS Microbiology Letters 41:289-294. Pinney, R.J. (1980) Mutation Research 72:155-159. Ponticelli, A.S., D.W. Schultz, A.F. Taylor, and G.R. Smith (1985) Cell 41:145-151. Poole, K., K. Krebes, C. McNally, and S. Neshat (1993) Journal of Bacteriology 175:7363-7372. Powell, M., Y. Hu, and D.M. Livermore (1991)Infection 19:174-177. Power, E., and I. Phillips (1993)FEMS Microbiology Letters112:251-254. Prasad, I., and S. Schaefler (1974)Journal of Bacteriology120:638-650. Pukkila, P.J., J. Peterson, G. Herman, P. Modrich, and M. Meselson(1983) Genetics 104:571-582. Quillardet, P., P.L. Moreau, H. Ginsburg, D.W. Mount, and R. Devoret (1982a) Molecular and General Genetics 188:37-43. Quillardet, P., O. Huisman, R. D’Ari, and M. Hofnung (1982b)Proceedings of the National Academy of Sciences of the United States of America79:5971-5975. Rabussay, D., and W. Zillig (1969)FEBS Letters 5:104-106. Radding, C M. (1982) Annual Review of Genetics 16:405-437. Radding, C M. (1991) Journal of Biological Chemistry 266:5355-5358 Radicella, J.P., P.U. Park, and M.S. Fox (1995)Science 268:418-420. Radi, S. (1996) Archiv der Pharmazie 329:115-119. Radstrom, P., C. Fermer, B.-E. Kristiansen, A. Jenkins, O. Skold, and G. Swedberg (1992) Journal of Bacteriology174:6386-6393. 261 Rajagopalan, M., C. Lu, R. Woodgate, M. O’Donnell, M F. Goodman, and H. Echols (1992) Proceedings of the National Academy of Sciences of the United States of America 89:10777-10781. Rangarajan, S., G. Gudmundsson, Z, Qiu, P.L. Foster, and M.F. Goodman (1997) Proceedings of the National Academy of Sciences of the United States of America 94:946-951. Reddy, M., and J. Gowrishanker (1997) Genetics 147:991-1001. Reece, R., and A. Maxwell (1991) Critical Reviews in Biochemistry and Molecular Biology 26:335-375. Relia, M., and D. Haas (1982) Antimicrobial Agents and Chemotherapy 22:242-249. Reynolds, A.E., J. Felton, and A. Wright (1981) Nature 293:625-629. Reynolds, A.E., S. Mahadevan, S.F. LeGrice, and A. Wright (1986) Journal of Molecular Biology 191:85-95. Reynolds, P.E. (1988) The essential nature of staphylococcal penicillin-binding proteins. In ‘Antibiotic inhibition of bacterial cell surface assembly and function’, Eds. Actor, P., L. Daneo-Moore, M.L. Higgins, M.R.J. Salton, and G.D. Shockman, pp.343-351. American Society for Microbiology, Washington, D.C. Reynolds, P.E., and H. Chase (1981)p-lactam-binding proteins: identification as lethal targets and probes of P-lactam accessibility. In ‘P-Lactam antibiotics: mode of action, new developments, and future prospects’, Eds. Salton, M.R.J., and G.D. Shockman, pp. 153-168. Academic Press, Inc., New York. Reynolds, P.E., and C. Fuller (1986) FEMS Microbiology Letters 33:251-254. Riddles, P.W., and I.R. Lehman (1985) Journal of Biological Chemistry 260:165-169. Riesenfeld, C., M. Everett, L.J.V. Piddock, and B.G. Hall (1997)Antimicrobial Agents and Chemotherapy 41:2059-2060. Riva, S., and L.G. Silvestri (1972)Annual Review of Microbiology26:199-224. Robillard, N.J., and A.L. Scarpa (1988) Antimicrobial Agents and Chemotherapy 32:535-539. Roca, A.I., and M.M. Cox (1990) Critical Reviews in Biochemistry and Molecular Biology25:415-456. Roca, A.I., and M.M. Cox (1997)Progress in Nucleic Acid Research and Molecular Biology56:129-223. 262 Roca, J. (1995) Trends in Biochemical Sciences 20:156-160. Roca, J., and J.C. Wang(1992) Cell 71:833-840. Roca, J., and J.C. Wang(1994) Cell 77:609-616. Roland, K., and J.W. Little(1990) Journal of Biological Chemistry 265:12828-12835. Roland, S., R. Erone, R.J. Harvey, V.L. Styles, and R.W. Morrison (1979)Journal of Biological Chemistry254:10337-10345. Rosamond, J., K.M. Telander, and S. Linn (1979)Journal of Biological Chemistry 254:8646-8652. Rosche, W.A., P.L. Foster, and J. Cairns (1999)Proceedings of the National Academy of Sciences of the United States of America96:6862-6897. Rosenberg, S.M. (1994) Genome 37:893-899. Rosenberg, S.M., and H. Echols (1990)Journal of Biological Chemistry265:20641- 20645. Rosenberg, S.M., R.S. Harris, S. Longerich, and A.M. Galloway (1996) Mutation Research 350:69-76. Rosenberg, S.M., and P.J. Hastings(1991) Biochimie 73:385-397. Rosenberg, S.M., S. Longerich, P. Gee, and R.S. Harris (1994) Science 265:405-407. Rosenberg, S.M., C. Thulin, and R.S. Harris (1998) Genetics 148:1559-1566. Rossi, L., E. Tonin, Y.R. Cheng. And R. Fontana (1985) Antimicrobial Agents and Chemotherapy 27:828-831. Ruiz, J., F. Marco, P. Goni, F. Gallardo, J. Mensa, A. Trilla, T. Jimenez de Anta, and J. Vila (1995) Journal of Antimicrobial Chemotherapy 36:737-738. Runyen-Janecky, L.J., M. Hong, and S.M. Payne (1999) Infection and Immunity 67:1415-1423. Rupp, W.D., and P. Howard-Flanders (1968) Journal of Molecular Biology 31:291- 304. Ryan, F.J. (1955) Genetics 40:726-738. Ryan, F.J. (1959) Journal of General Microbiology 21:530-549. 263 Ryan, F.J., D. Nakada, and M.J. Schneider (1961)Zeitschrifl fur Vererbungslehre 92:38-41. Ryan, F.J., T. Okada, and T. Nagata (1963)Journal of General Microbiology30:193- 199. Ryan, F.J., and L.K. Wainwright (1954)Journal of General Microbiology11:364-.379 Rydberg, B. (1978)Mutation Research 52:11-24. Salles, B., and C. Paoletti (1983)Proceedings of the National Academy of Sciences of the United States of America 80:65-69. Sancar, A. (1987) Photobiochemistry and Photobiophysics 11:301-315. Sancar, A., K.A. Franklin, and G.B. Sancar (1984)Proceedings of the National Academy of Sciences of the United States of America 81:7397-7401. Sancar, A., and C.S. Rupert (1978)Gene 4:295-308. Sancar, A., and W.D. Rupp (1983) Cell 33:249-260. Sancar, A., and G.B. Sancar (1988) Annual Review of Biochemistry 57:29-67. Sancar, A., G.B. Sancar and W.D. Rupp (1982) Nature 289:196-198 Sancar, G.B., F.W. Smith, and A. Sancar (1983)Nucleic Acids Research 11:6667- 6678. Sancar, G.B., F.W. Smith, and A. Sancar (1985) Biochemistry 24:1849-1855. Sancar, A., and M.S. Tang (1993) Photochemistry and Photobiology 57:905-921. Sanders, C.C., W.E. Sanders, Jr., R.V. Goering, and V. Wemer (1984)Antimicrobial Agents and Chemotherapy 26:797-801. Sarov-Blat, L., and Z. Livneh (1998) Journal Biological Chemistry 273:5520-5527. Sassanfar, M., and J.W. Roberts(1990) Journal of Molecular Biology 212:79-96. Sato, K., K. Hoshino, T. Une, and Y. Osada (1989) Reviews of Infectious Diseases 11 (Supplement 5):S915-8916. Sato, K., V. Matsuura, M. Inoue, T. Une, Y. Osada, H. Ogawa, and S. Mitsuhashi (1982) Antimicrobial Agents and Chemotherapy 22:548-553. 264 Saunders, J R., C.A. Hart, and V.A. Saunders (1986) Journal of Antimicrobial Chemotherapy 18 (Supplement C):57-66. Schaaper, R.M., and R.L. Dunn (1987) Proceedings of the National Academy of Sciences of the United States of America 84:6220-6224. Scheuermann, R., and H. Echols (1984) Proceedings of the National Academy of Sciences of the United States of America 81:7747-7751. Schliiter, G. (1986)Current Clinical Practice 34:64-67. Schnarr, M., J. Pouyet, M. Granger-Schnarr, and M. Daune (1985)Biochemistry 24:2812-2818. Schnetz, K., C. Toloczyki, and B. Rak (1987)Journal of Bacteriology169:2579-2590. Sedliakova, M., K. Kieibl, and F. Masek (1987)Mutation Research 191:13-16. Selby, C.P., and A. Sancar (1990)Mutation Research 236:203-211. Seoane, A., and S B. Levy (1995) Journal of Bacteriology 177:3414-3419. Setlow, J.K., E. Caberra-Juarez, W.L. Albritton, D. Spikey, and A. Mutschler (1985) Journal of Bacteriology 164:525-534. Setlow, R.B. (1966) Science 153:379-386. Setlow, R.B., and J.K Setlow (1972) Annual Review of Biophysics and Biomolecular Structure 1:293-346. Shanabruch, W.G., and G.C. Walker(1980) Molecular and General Genetics 179:289- 297. Shannon, K., and I. Phillips (1982)Journal of Antimicrobial Chemotherapy9:91-102. Shanson, D.C. (1981)Journal of Hospital Infection2:11-36. Sharif, F., and B.A. Bridges (1990)Mutagenesis 5:31-34. Shapiro, J.A. (1984) Molecular and General Genetics 194:79-90. Shapiro, J.A., and D. Leach(1990) Genetics 126:293-299. Sharpies, G.J., and R.G. Lloyd (1991)Journal of Bacteriology173:7711-7715. Shavitt, O., and Z. Livneh (1989)Journal of Biological Chemistry264:11275-11281. 265 Shaw, W.V. (1967) Journal of Biological Chemistry 242:687-693. Shaw, W.V. (1971) Annuals of the New York Academy of Sciences 182:234-242. Shearwin, K.E., A.M. Brumby, and J.B. Egan (1998)Journal of Biological Chemistry 273:5708-5715. Shen, L.L., J. Baranowski, andA.G. Pernet (1989a) Biochemistry 28:3879-3885. Shen, L.L., and D.T.W. Chu (1996) Current Pharmaceutical Design 2:195-208. Shen, L.L., L A. Mitscher, P.N. Sharma, T.J. O’Donnell, D.W.T. Chu, C.S. Cooper, T. Rosen, and A.G. Pernet (1989b) Biochemistry 28:3886-3894. Shen, L.L., and A.G. Pernet (1985)Proceedings of the National Academy of Sciences of the United States of America 82:307-311. Shiba, T., H. Iwasaki, A. Nakata, and H. Shinagawa (1990)Molecular and General Genetics 224:169-176. Shinagawa, H., H. Iwasaki, T. Kato, and A. Nakata (1988)Proceedings of the National Academy of Sciences of the United States of America85:1806-1810. Shinagawa, H., T. Kato, T. Ise, K. Makino, and A. Nakata (1983)Gene 23:167-174. Shurvinton, C.E., and R.G. Lloyd (1982) Molecular and General Genetics 185:352- 355. Sibghat-UUah, A. Sancar, and J.E. Hearst(1990) Nucleic Acids Research 18:5051- 5053. Siccardi, A.G. (1969)Journal of Bacteriology100:337-346. Siegel, E.C. (1983) Molecular and General Genetics 191:397-400. Simic, D., B. Vukovic-Gacic, A. Ajanovic, and J. Knezevic-Vukcevic(1991) Mutation Research 254:255-262. Sirotnak, P.M., and R.W. McCuen (1973)Genetics 74:543-556. Skaliter, R., T. Paz-Elizur, and Z. Livneh (1996)Journal of Biological Chemistry 271:2478-2481. Skold, O. (1976) Antimicrobial Agents and Chemotherapy 9:49-54. Skold, O., and A. Widh. (1974)Journal of Biological Chemistry249:4324-4325. 266 Slater, S.C., and R. Maurer (1991) Proceedings of the National Academy of Sciences of the United States of America 88:1251-1255. Slilaty, S.N., and J.W. Little (1987) Proceedings of the National Academy of Sciences of the United States of America 84:3987-3991. Slilaty, S.N., J.A. Rupley, and J.W. Little (1986) Biochemistry 25:6866-6875. Smith, C.L., and M. Oishi (1978)Proceedings of the National Academy of Sciences of the United States of America 75:1657-1661. Smith, CM., W.H. Koch, S.B. Franklin, P.L. Foster, T.A. Cebula, and E. Eisenstadt (1990) Journal of Bacteriology 172:4964-4978. Smith, G.R. (1987) Annual Review of Genetics 21:179-201. Smith, G.R. (1988) Microbiological Reviews 52:1-28. Smith, J.T. (1984) Phannaceutical Journal 233:299-305. Smith, J.T. (1986) Scandinavian Journal of Infectious Diseases 49 (Supplement): 115- 123. Sniegowski, P.D. (1995) Journal of Bacteriology 177:1119-1120. Snyder, M., and K. Drlica (1979)Journal of Molecular Biology131:287-302. Sommer, S., A. Bailone, and R. Devoret (1993) Molecular Microbiology 10:963-971. Sommer, S., F. Boudsocq, R. Devoret, and A. Bailone (1998)Molecular Microbiology 28:281-291. Soussy, C.J., J.S. Wolfson, E.Y. Ng, and D.C. Hooper (1993)Antimicrobial Agents and Chemotherapy 37:2588-2592. Spirin, A S. (1985) Progress in Nucleic Acid Research and Molecular Biology 32:75- 114. Sreedharan, S., M. Oram, B. Jensen, L.R. Peterson, and L.M. Fisher (1990)Journal of Bacteriology172:7260-7262. Sreetvatsan, S., X. Pan, K.E. Stockbauer, D.L. Williams, B.N. Kreiswirth, and J.M. Musser (1996) Antimicrobial Agents and Chemotherapy 40:1024-1026. Stahl, F.W. (1988) Nature 335:112-113. Stahl, F.W. (1990) Nature 346:791. 267 Stamey, T A., N.J. Neymoy, and M. Higgins (1963) Investigative Urology6:582-591. Steele, D.F., and S. Jinks-Robertson (1992)Genetics 132:9-21. Steinborn, G. (1978) Molecular and General Genetics 165:87-93. Strauss, B.S. (1991) Bioessays 13:79-84. Strike, P., and D. Lodwick (1987) Journal of Cell Science 6 (Supplement):303-321. Studwell-Vaughan, P.S., and M. O’Donnell (1993)Journal of Biological Chemistry 268:11785-11791. Su, S.S., and P. Modrich (1986) Proceedings of the National Academy of Sciences of the United States of America 83:5057-5061. Suginaka, H., P.M. Blumberg, and J.L. Strominger (1972)Journal of Biological Chemistry 247:5279-5288. Sugino, A., N.P. Higgins, P.O. Brown, C.L. Peebles, and N.R. Cozzarelli (1978) Proceedings of the National Academy of Sciences of the United States of America 75:4838-4842. Sugino, A., C. Peebles, K. Kruezer, and N. Cozzarelli (1977)Proceedings of the National Academy of Sciences of the United States of America74:4767-4771. Sundin, G.W., S.P. Kidambi, M. Ullrich, and C.L. Bender (1996)Gene 177:77-81. Sutherland, B.M. (1981) In “Enzymes.” pp. 482-515, Ed. Boyer, P.O., Academic Press, New York. Suzuki, ¥., and S. Okamoto(1967) Journal of Biological Chemistry 242:4722-4730. Swanberg, S.L., and J.C. Wang(1987) Journal of Molecular Biology 197:729-736. Sweasy, J.B., E.M. Witkin, N. Sinha, and V. Roegner-Maniscalco (1990)Journal of Bacteriology172:3030-3036. Swedberg, G., S. Castensson, and O. Skold (1979) Journal of Bacteriology 137:129- 136. Swedberg, G., C. Fermer, and O. Skold (1993) Point mutations in the dihydropteroate synthase gene causing sulfonamide resistance, p. 555-558. In J.E. Ayling (ed.). Chemistry and biology of pteridines and folates. Plenum Press, New York. Swedberg, G., and O. Skold (1980)Journal of Bacteriology142:1-7. 268 Szekeres, E.S., R. Woodgate, and C.W. Lawrence (1996)Journal of Bacteriology 178:2559-2563. Szpilewska, H., P. Bertrand, A. Bailone, and M. Dutreix (1995)Biochimie 77:848- 853. Taddei, F., I. Matic, and M. Radman (1995)Proceedings of the National Academy of Sciences of the United States of America 92:11736-11740. Tadmor, Y., R. Ascarelii-Goell, R. Skaliter, and Z. Livneh (1992)Journal of Bacteriology174:2517-2524. Tait, R.C., and H.W. Boyer (1978) Cell 13:73-81. Takahagi, M., H. Iwasaki, A. Nakata, and H. Shinagawa (1991)Journal of Bacteriology173:5747-5753. Tanaka, N., G. Kawano, and T. Kinoshita (1971)Biochemical and Biophysical Research Communications42:564-567. Tang, M., I. Bruck, R. Entja, J. Turner, E.G. Frank, R. Woodgate, M. O’Donnell, and M.F. Goodman (1998)Proceedings of the National Academy of Sciences of the United States of America 95:9755-9760. Taylor, A.F., D.W. Schultz, A S. Ponticelli, and G.R. Smith (1985)Cell 41:153-163. Telenti, A., P. Imboden, F. Marchesi, D. Lowrie, S. Cole, M.J. Colston, L. Matter, K. Schopfer, and T. Bodimer (1993)Lancet 341:647-650. Tesch, W., A. Strassle, B. Berger-Bachi, D. O’Hara, P. Reynolds, and F.H. Kayser (1988) Antimicrobial Agents and Chemotherapy 32:1494-1499. Tessman, I., and M.A. Kennedy (1993) Genetics 136:439-448. Thaler, D.S., E. Sampson, I. Siddiqi, and S.M. Rosenberg (1989)Genome 31:53-67. Thliveris, A T., J.W. Little, and D.W. Mount (1991)Biochimie 73:449-456. Thomas, S.M., H.M. Crowne, S.C. Pidsley, and S.G. Sedgwick (1990) Journal of Bacteriology 172:4979-4989. Thoms, B., and W. Wackernagel (1988) Journal of Bacteriology 170:3675-3681. Timmis, K.N., M l. Gonzales-Carrero, T. Sekizaki, and F. Rojo (1986) Journal of Antimicrobial Chemotherapy 18 (Supplement C):l-12. 269 Torkelson, J., R.S. Harris, M.-J. Lombardo, J. Nagendran, C. Thulin, and S.M. Rosneberg (1997) EMBO Journal 16:3303-3311. Truong, Q.C., S. Ouabdesselam, D.C. Hooper, N.J. Moreau, and C.J. Soussy (1995) Journal of Antimicrobial Chemotherapy36:1055-1059. Tsaneva, I.R., B. Müller, and S.C. West (1992a)Cell 69:1171-1180. Tsaneva, I.R., G.T. lUing, R.G. Lloyd, and S.C. West (1992b)Molecular and General Genetics 235:1-10. Tweats, D.J., M.J. Thompson, R.J. Pinney, and J.T. Smith (1976)Journal of General Microbiology93:103-110. Ubukata, K., N. Yamashita, and M. Konno (1985)Antimicrobial Agents and Chemotherapy 27:851 -857. Ubukata, K., R. Nonoguchi, M. Mitsuhashi, and M. Konno (1989)Journal of Bacteriology171:2882-2885. Uhlin, B.E., and A.J. Clark(1981) Journal of Bacteriology 148:386-390. Uhlin, B.E., and K. Nordstrom (1985)Journal of Bacteriology162:855-857. Umbarger, H E., and B. Brown (1955)Journal of Bacteriology70:241-248. Upton, C., and R.J. Pinney (1983) Mutation Research 112:261-273. Urios, A., G. Herrera, V. Aleixandre, and M. Blanco (1991)Biochimie 73:519-521. Utsui, ¥., and T. Yokota (1985)Antimicrobial Agents and Chemotherapy 35:397-403. Vall-Spinosa, A., W. Lester, T. Moulding, P.T. Davidson, and J.K. McClatchy (1970)New England Journal of Medicine 283:616-621. Van Houten, B. (1990) Microbiological Reviews 54:18-51. Van Houten, B., H. Gamper, J.E. Hearst, and A. Sancar (1988)Journal of Biological Chemistry 263:16553-16560. Vandewiele, D., A. Borden, P.I. O’Grady, R. Woodgate, and C.W. Lawrence (1998) Proceedings of the National Academy of Sciences of the United States of America 95:15519-15524. Venderbure, C., A. Chastanet, F. Boudsocq, S. Sommer, and A. Bailone (1999) Journal of Bacteriology181:1249-1255. 270 Vila, J., J. Ruiz, F. Marco, A. Barcelo, P. Goni, E. Giralt, and T. Jimenez De Anta (1994) Antimicrobial Agents and Chemotherapy 38:2477-2479. Vila, J., J. Ruiz, P. Gomi, and M.T.J.D. Anta (1996)Antimicrobial Agents and Chemotherapy 40:491 -493. Villani, G., S. Boiteux, and M. Radman (1978) Proceedings of the National Academy of Sciences of the United States of America 75:3037-3041. Volkert, M.R., and M.A. Hartke (1984) Journal of Bacteriology 157:498-506. Wagner, R.Jr., and M. Meselson (1976) Proceedings of the National Academy of Sciences of the United States of America 73:4135-4139. Waleh, N.S., and B A D. Stocker (1979) Journal of Bacteriology 137:830-838. Walker, G.C. (1977) Molecular and General Genetics 152:93-103. Walker, G.C. (1984) Microbiological Reviews 48:60-93. Walker, G.C. (1985) Annual Review of Biochemistry 54:425-457. Walker, G.C. (1995) Trends in Biochemical Sciences 10:416-420. Walker, G.C., and P.P. Dobson (1979) Molecular and General Genetics 172:17-24. Walters, R.N., L.J.V. Piddock, and R. Wise (1989)Journal of Antimicrobial Chemotherapy 24:863-873. Watanabe, T. (1963) Bacteriological Reviews 27:87-115. Waxman, D.J., and J.L. Strominger(1983) Annual Review of Biochemistry 52:825- 869. Weinstock, G.M. (1987) In coli and S. typhimurium: Cellular and Molecular Microbiology.”, pp. 1034-1043, Eds. Neidhardt, F.C., J.L. Ingraham, K.B. Low, B. Magasanik, M. Schaechter and H.E. Umbarger, ASM, Washington. Weisemann, J.M., C. Funk, and G.M. Weinstock (1984)Journal of Bacteriology 132:197-208. Weisser, J., and B. Wiedemann (1986)Journal of Antimicrobial Chemotherapy18:575- 583. Welsh, K.M., A C. Lu, S. Clark, and P. Modrich (1987)Journal of Biological Chemistry 262:15624-15629. 271 West, S.C. (1992) Annual Review of Biochemistry 61:603-640. West, S.C. (1994) Cell 76:9-15. Whitby, M.C., L. Ryder, and R.G. Lloyd (1993) Cell 75:341-350. White, R.L., and M.S. Fox (1974)Proceedings of the National Academy of Sciences of the United States of America 71:1544-1548. White, R.L., and M.S. Fox (1975)Genetics 81:33-50. Wildenberg, J., and M. Meselson (1975) Proceedings of the National Academy of Sciences of the United States of America 72:2202-2206. Willetts, N., and A.J. Clark(1969) Journal of Bacteriology 100:231-239. Willie, G.R., N. Richman, W.O. Godtfredsen, and J.W. Bodley (1975) Biochemistry 14:1713-1718. Willmott, C.J.R., and A. Maxwell (1993)Antimicrobial Agents and Chemotherapy 37:126-127. Wise, E.M., and M.M. Abou-Donia (1975) Proceedings of the National Academy of Sciences of the United States of America 72:2621-2625. Wise, R., J.M. Andrews, and L.J. Edwards (1983)Antimicrobial Agents and Chemotherapy 23:559-564. Wise, R., J.M. Andrews, and N. Brenwald (1993) Journal of Antimicrobial Chemotherapy 31:73-80. Witkin, E.M. (1970) Nature 229:81-82. Witkin, E.M. (1974) Proceedings of the National Academy of Sciences of the United States of America 71:1930-1934. Witkin, E.M. (1976) Bacteriological Reviews 40:869-907. Witkin, E.M., V. Roegner-Maniscalco, J.B. Sweasy, and J.O. McCall (1987) Proceedings of the National Academy of Sciences of the United States of America 84:6805-6809. Wolfson, J.S., and D.C. Hooper (1985)Antimicrobial Agents and Chemotherapy 28:581-586. Woodgate, R., B.A. Bridges, G. Herrera, and M. Blanco (1987)Mutation Research 183:31-37. 272 Woodgate, R., M. Rajagopalan, C. Lu, and H. Echols (1989)Proceedings of the National Academy of Sciences of the United States of America86:7301-7305. Worcel, A. (1974) ‘Studies on the folded chromosome ofEscherichia colV. In ‘Mechanisms and Regulation of DNA Replication’, pp. 201-224, Eds. Kolberg, A.R., and M. Kohiyama, Plenum Press, New York and London. Yamagishi, J-., Y. Furutani, S. Inoue, T. Ohue, S. Nakamura, and M. Shimizu (1981)Journal of Bacteriology148:450-458. Yamagishi, J-I., H. Yoshida, M. Yamayoshi, and S. Nakamura (1986)Molecular and General Genetics 204:367-373. Yamagishi, J., T. Kojima, Y. Oyamada, K. Fujimoto, H. Hattori, S. Nakamura, and M. Inoue (1996)Antimicrobial Agents and Chemotherapy 40:1157-1163. Yasuda, T., K. Morimatsu, T. Horii, T. Nagata, and H. Ohmori(1998) EMBO Journal 17:3207-3216. Yasuda, T., T. Nagata, and H. Ohmori (1996)Journal of Bacteriology178:3854-3859. Yatagai, F., J.A. Halliday, and B.W. Glickman(1991) Molecular and General Genetics 230:75-80. Yeung, A.T., W.B. Mattes, E.Y. Oh, and L. Grossman (1983)Proceedings of the National Academy of Sciences of the United States of America80:6157-6161. Yoakum, G.H., and L. Grossman (1981) Nature 292:171-173. Yoakum, G.H., S.R. Kushner, and L. Grossman (1980) Gene 12:243-248. Yoshida, H., M. Bogaki, M. Nakamura, L.M. Yamanaka, and S. Nakamura (1991) Antimicrobial Agents and Chemotherapy 35:1647-1650. Yoshida, H., T. Kojima, J-I. Yamagishi, and S. Nakamura (1988)Molecular and General Genetics 211:1-7. Yoshida, H., M. Nakamura, B. Bogaki, and S. Nakamura (1990)Antimicrobial Agents and Chemotherapy 34:1273-1275. Youngs, D.A., and K.C. Smith (1977) Molecular and General Genetics 152:37-41. Youngs, D.A., and K.C. Smith (1978)Mutation Research 51:139-143. Ysern, P., B. Clerch, M. Castano, I. Gibert, J. Barbé, and M. Llagostera (1990) Mutagenesis 5:63-66. 273 Zechiedrich, E L., and N.R. Cozzarelli (1995) Genes and Development 9:2859-2869. Zechiedrich, E L., and N. Osheroff (1990) EMBO Journal 9:4555-4562. 274 INCREASE IN MUTATION FREQUENCY TO NALIDIXIC ACID RESISTANCE INDUCED BY GROWTH ON SOLID MEDIUM CONTAINING INHIBITORY OR LETHAL CONCENTRATIONS OF DRUG J. Hinton and R.J. Pinney, Microbiology Section, Department of Pharmaceutics, The School of Pharmacy, University of London, Brunswick Square, London WCIN lAX, UK Increased frequency of mutation to lactose utilisation has been demonstrated Fig. 1. Effect of plasmid carriage on Fig. 2. Effect of plasmid carriage on when strains ofEscherichia coli that cannot metabolise lactose were plated on mutation to resistance to 4 pig mL ' NA mutation to resistance to 8pig mL ' NA media containing lactose as sole carbon source (Caimset al, 1988). "Adaptive R446b mutations" of this type are said to occur only in selected genes after the R446b application of non-lethal selection in non-growing cultures. The antibacterial R- drug, nalidixic acid (NA) damages DNA and induces mutagenic error-prone | - 4 R" DNA repair (Piddock & Wise, 1987). This study reports the generation of, what might be considered “adaptive” NA resistant mutants in E. coli ABl 157 by growth on solid medium containing NA, and the effect that mutator plasmid 1 -7 R446b (Pinney, 1980) has on the frequency at which mutants appear. 0.1 mL volumes of overnight nutrient-broth grown cultures ofE. coli r® ABl 157 and ABl 157(R446b) were plated onto nutrient agar containing NA at 4 or 8 pig mL'^ (one and two times MIC, with the latter concentration being bactericidal). These plates were incubated for 10 days at 37°C and the number 0 2 4 6 8 10 0 2 4 6 8 10 of visible colonies on each plate scored at 24 or 48 hour intervals. Total viable Incubation time (days) Incubation time (days) counts of the plated cultures were also determined on drug-free nutrient agar. restreaking, but did not grow on higher concentrations of NA. Colonies of both Mutation frequencies were expressed as ratios of the number of visible NA strains grown in the presence of 8 pig mL ' NA, grew when streaked onto resistant clones to the number of cells plated. nutrient agar containing 4,8, 16 or 32 pig mL ' NA. The number of visible NA resistant colonies of strain ABl 157 These results suggest that the presence of NA induces mutational increased with time on media containing both 4 and 8 p^g mL'^ NA (Figs. 1 and resistance to itself, with a mutator plasmid enhancing the effect. Since plasmid- 2), giving 250- and 500-fold increases respectively in apparent mutation mediated resistance per se to quinolones is unknown, the data may have frequencies to NA resistance. The strain carrying plasmid R446b gave similar important implications on our understanding of how mutational resistance to results, with the plasmid-mediated mutator effect increasing mutation frequency these clinically important antibacterial drugs develops. 5- and 2.5-fold respectively after 10 days. The NA resistance phenotype was confirmed by streaking onto nutrient agar containing 4,8, 16 or 32 p&g mL'\ Caims, J. et al (1988) Nature 335:142-145 Colonies of strains ABl 157 and ABl 157(R446b) growing on medium Piddock, L.J.V., Wise, R. (1987) FEMS Microbiol. Lett. 41:289-294 containing 4 p^g mL‘* NA, retained resistance to this concentration on Pinney, R.J. (1980) Mutat. Res. 72:155-159 INCREASE IN MUTATION FREQUENCY TO NALIDIXIC ACID RESISTANCE INDUCED BY GROWTH IN LIQUID MEDIUM CONTAINING SUB-LETHAL CONCENTRATIONS OF DRUG J. Hinton and R.J. Pinney, Microbiology Section, Department of Pharmaceutics, The School of Pharmacy, University of London, Brunswick Square, London WCIN lAX, UK The antibacterial drug, nalidixic acid (NA), the archetypal 4-quinolone, Fig. 1. Effect of growth in 0 (#), 2 (□), Fig. 2. Effect of plasmid carriage on damages bacterial DNA and induces mutagenic error-prone DNA repair or 3 (V) ;ug mL * NA on NA resistance NA resistance in E. coli grown m (Piddock & Wise, 1987). This study reports the effects of exposure to sub- in E, coli (R446b) drug-free broth inhibitory concentrations of NA in liquid growth medium on the occurrence of NA resistance inEscherichia coli. The effect of mutator plasmids (Ambler Total viable counts & Pinney, 1995) on the development of resistance has also been investigated. 10* 1.5 mL of overnight nutrient broth (NB) cultures ofE. coli ABl 157 3pg mL |i R446b and derivatives of this strain carrying the mutator plasmids R46 or R446b (Pinney 1980) were diluted in 150 mL NB containing 0, 2 or 3 /^g mL'^ NA. R46 Cultures were incubated at 37°C for ten days. Samples were removed at 24 or "o 48 hour intervals, diluted in NB and titered for nalidixic acid resistant mutants R" S) 10 by diluting and plating on nutrient agar containing 16 /^g mL * NA (four times 10 the minimum inhibitory concentration). Plates were incubated for 2 days at 37°C. Total viable counts were also determined on drug-free nutrient agar. Growth ofE. coli ABl 157 in NB containing 2or3^g mL * NA led 0 2 4 6 8 10 0 2 4 6 a 10 Incubation time (days) Incubation time (days) to 17- and 500-fold increases respectively in the number of NA resistant mutants recovered after 10 days incubation, compared to control cultures These results show that growth in the presence of sub-inhibitory concentrations grown in drug-free broth. Plasmid R446b enhanced this effect. The numbers of NA significantly increases the frequency of bacterial resistance to this of NA resistant cells present in cultures of AB1157(R446b) after 10 days antibacterial drug. They may therefore have important implications in the incubation in 2 or 3 //g mL * NA were 1 x 10^ - and 4x10^ -fold greater clinical management of bacterial infections, suggesting that inappropriate respectively than in control cultures grown in drug-free broth (Fig. 1). Plasmid dosage regimens could lead to the development of resistant strains. The R46 gave similar results (data not shown). Carriage of R46 and R446b also presence of mutator plasmids further increases the frequency at which bacteria increased the number of NA resistant mutants after growth in drug-free broth: become resistant. 6-fold by R46 and 20-fold by R446b (Fig. 2). Ambler, I.E., Pinney, R.J. (1995) J. Antimicrob. Chemother. 35:603-609 Piddock, L.J.V., Wise, R. (1987) FEMS Microbiol. Lett. 41:289-294 Pinney, R.J. (1980) Mutat. Res. 72:155-159