Division of Mycobacterial Research National Institute for Medical Research, Mill Hill London

RecA expression and DNA damage in mycobacteria

Nicola A Thomas

June 1998

A thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy in the Faculty of Science, University College of London. ProQuest Number: U642358

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 U642358

Published by ProQuest LLC(2015). 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 A bstr a ct

Bacterial responses to DNA damage are highly conserved. One system, the SOS response, involves the co-ordinately induced expression of over 20 unlinked genes. The RecA protein plays a central role in the regulation of this response via its interaction with a repressor protein, LexA. The recA gene of Mycobacterium tuberculosis has previously been cloned and sequenced (Davis et al, 1991) and a LexA homologue has recently been identified (Movahedzadeh, 1996). The intracellular lifestyle of pathogenic mycobacteria constantly exposes them to hostile agents such as hydrogen peroxide. Consequently, the response of mycobacteria to DNA damage is of great interest. In this study the basal levels of RecA protein were shown to be elevated in M. microti and M. smegmatis compared to those found in M. tuberculosis, M. bovis BCG and a wide range of both saprophytic and pathogenic mycobacteria. Expression of this protein was shown to be inducible in both M. smegmatis and M tuberculosis in response to DNA damaging agents, although the kinetics of induction were quite different. A reduction in the levels of LexA following treatment of M. smegmatis, M. tuberculosis and M. microti with a DNA damaging agent, probably as a result of LexA cleavage, provided further corroboratory evidence to suggest that the essential regulatory elements of the SOS response may have been conserved. There was no evidence of RecA induction in M. microti. Sequencing of the M microti recA gene and 700bp of upstream sequence revealed a high degree of homology with equivalent sequences m M tuberculosis', including a putative LexA binding site. M. microti recA was shown to be inducible in M. smegmatis mc^l55 in response to DNA damage using a transcriptional fusion of the putative regulatory elements of this gene to the reporter gene chloramphenicol-acetyl transferase (CAT). Thus it appears that the kinetics of induction are controlled at the level of the host cell; this was further investigated by studying the interaction of purified mycobacterial LexA protein and M. microti cell-free extracts with the putative LexA binding site. T h e T e .\ï p t a t î o n o f S i A n t h o n y Israel van Meckenem

‘Sometimes it felt like this..,” A cknowledgements

I wish to thank Dr. M. Joseph Colston and Dr. Elaine Davis for their invaluable input and guidance over the course of the last four years. They have consistently provided good advice surely no PhD. student could ask for more. I also wish to thank Patricia Brooks for her technical advice as well as Dr. Papavinasasundaram and Dr. Gili Bachrach for their interest and helpful discussions. A special thanks goes to Bob, Vangelis, Colin, Pat, Edith and Fran for their friendship, good humour and support over the past few years.

Beyond academia, I owe a great debt to Mark who has made my existence far more pleasant in many ways.

To my parents T a b l e o f c o n t e n t s

T a b l e o f c o n t e n t s ...... 5

L is t OF F i g u r e s ...... 10

L is t OF T a b l e s ...... 12

L ist o f A bbreviations ...... 13

1.0 I ntroduction ...... 15

1.1 The Mycobacteria ...... 15 1.1.1 Characteristics ...... 15 1.1.2 Mycobacterial pathogens: a re-emerging problem ...... 15 1.2 Molecular biology of mycobacteria: current knowledge and progress ...... 17 1.2.1 Bacterial cloning hosts for expression of mycobacterial genes: gene expression in a heterologous background ...... 17 1.2.2 Mycobacteria as cloning hosts: gene expression in a homologous background ...... 18 1.2.2.1 Mycobacteriophage and shuttle-phasmids ...... 19 1.2.2.2 Mycobacterial plasmids and shuttle vector technology ...... 20 1.2.2.3 Insertion elements and transposons ...... 21 1.2.3 Expression of foreign genes in mycobacteria ...... 22 1.3 Gene regulation ...... 23 1.3.1 Initiation of transcription is catalysed by RNA polymerase: A principal step in the regulation of gene expression ...... 23 1.3.2 Regulatory proteins ...... 24 1.3.3 Gene regulation in mycobacteria: Current knowledge and progress ...... 24 1.4 DNA damage in bacteria: a quick review ...... 26 1.5 DNA damage repair and tolerance systems in prokaryotes ...... 27 1.5.1 DNA repair in prokaryotes ...... 27 1.5.2 DNA damage tolerance systems ...... 30 1.6 The RecA protein: a pivotal component of several cellular mechanisms ...... 30 1.6.1 Homologous recombination ...... 32 1.6.2 Recombinational or post-replication repair ...... 35 1.6.3 Repair of double strand breaks ...... 37 1.6.4 The SOS response: Physiological consequences and regulation ...... 37 1.6.5 SOS Mutagenesis ...... 43 1.7 RecA mediated responses in mycobacteria ...... 43 1.7.1 A novel post-translational modification for the M. tuberculosis RecA protein ...... 44 1.7.2 Homologous recombination in mycobacteria ...... 46 1.8 Aim s ...... 47 1.8.1 Background to the project ...... 47 1.8.2 Specific aims of the project ...... 48

2.0 M a t e r ia l s an d M e t h o d s ...... 49

2.1 Bacterial strains and plasmids; media and growth conditions ...... 49 2.2 Production and purification of recombinant M. tuberculosis RecA protein ...... 51 2.2.1 Over expression of M. tuberculosis recA by heat induction of E. co/z pEJ169 {TG2(pGPl-2)} ...... 51 2.2.2 Production of a crude preparation of recombinant M. tuberculosis RecA protein by differential precipitation ...... 51 2.2.3 Purification of a crude preparation of recombinant M. tuberculosis RecA protein by SDS-polyacrylamide gel electrophoresis ...... 51 2.3 Production of polyclonal antisera raised against recombinant M. tuberculosis RecA protein ...... 52 2.4 Preparation of an affinity chromatography column and purification of polyclonal antisera raised against recombinant M tuberculosis RecA protein...... 52 2.4.1 Production of M. leprae RecA precursor protein ...... 52 2.4.2 Determination of the number of reduced sulfhydryl groups in M. leprae RecA precursor protein ...... 53 2.4.3 Coupling of the M. leprae RecA precursor protein to CNBr activated Sepharose 4B ...... 53 2.4.4 Purification of polyclonal antisera raised against M. tuberculosis RecA protein by affinity chromatography ...... 53 2.5 Production of mycobacterial cell-free extracts...... 54 2.6 Western Blotting ...... 54 2.7 Induction of DNA damage in mycobacteria ...... 55 2.7.1 Exposure to ofloxacin and mitomycin C ...... 55 2.7.2 Ultra-violet irradiation of mycobacterial cultures ...... 55 2.7.3 Determination of the minimum inhibitory concentration (MIC) ...... 56 2.8 Isolation of mycobacterial nucleic acids ...... 56 2.9 Cloning ...... 56 2.9.1 Small scale plasmid preparations ...... 56 2.9.2 Restriction enzyme digestion and alkaline phosphatase treatment ...... 56 2.9.3 Agarose gel electrophoresis ...... 57 2.9.4 Purification of DNA fragments by elution from an agarose gel slice ...... 57 2.9.5 Ligations ...... 57 2.9.6 Transformation of Escherichia coli ...... 57 2.9.7 Transformation of mycobacteria by electroporation ...... 58 2.10 Southern hybridisation ...... 58 2.10.1 Probe labelling ...... 58 2.10.2 DNA transfer and hybridisation ...... 58 2.11 Induction of RecA-CAT fusions in M. smegmatis mc^l 5 5 ...... 59 2.12 Sequencing ...... 59 2.13 Gel Shift Assay ...... 60 2.13.1 Precipitation of oligonucleotides ...... 60 2.13.2 Gel shift assay ...... 60

3.0 P r o d u c t io n a n d purification o f p o l y c l o n a l a n t is e r a r a ise d

AGAINST M. TUBERCULOSIS R E C A ...... 61

3.1 Introduction ...... 61 3.2 The production of polyclonal antisera raised against M. tuberculosis RecA protein 62 3.2.1 Production and purification of recombinant M. tuberculosis RecA protein ...... 62 3.2.2 The immunisation protocol and subsequent monitoring procedures leading to the production of M. tuberculosis RecA antisera 63 3.3 Purification of M. tuberculosis RecA antisera by affinity chromatography using the M. leprae RecA precursor protein covalently linked to a suitable immunoadsorbent ...... 67 3.3.1 Investigations to identify an appropriate coupling g e l ...... 67 3.3.2 Production of A/, leprae RecA precursor protein ...... 68 3.3.3 Preparation of an affinity chromatography column and subsequent purification of polyclonal antisera ...... 70 3.4 Discussion ...... 70

4.0 A QUALITATIVE COMPARISON OF BASAL RECA AND LEXA LEVELS IN

SAPROPHYTIC AND PATHOGENIC MYCOBACTERIA...... 73

4.1 Introduction ...... 73 4.1.1 The importance of the recA genotype ...... 74 4.2 Confirmation that affinity purified M. tuberculosis RecA antibody will recognise RecA precursor protein ...... 76 4.3 A comparison of basal RecA levels in uninduced cell-free extracts of M. smegmatis and members of the M. tuberculosis complex ...... 76 4.3.1 Construction of growth curves for M. smegmatis, M. tuberculosis and M. microti ...... 76 4.3.2 Basal RecA levels in M. smegmatis, M. tuberculosis, M. microti and M. bovis BCG...... 76 4.4 The effect of growth stage on precursor and mature RecA (38kDa) levels in M. tuberculosis and M. microti ...... 78 4.5 A comparison of basal RecA levels in clinical isolates of M. tuberculosis ...... 80 4.6 A comparison of basal RecA levels in a range of saprophytic and pathogenic mycobacteria ...... 80 4.7 The effect of cultivation in alternative media on basal RecA levels in M. smegmatis and the M. tuberculosis complex ...... 83 4.8 An in vitro model that simulates dormancy; the effect on basal RecA levels and the appearance of precursor RecA protein ...... 86 4.9 Confirmation of the presence of a LexA homologue in M. smegmatis and the M. tuberculosis complex and a comparison of basal levels ...... 88 4.10 Discussion ...... 88

5.0 The effect of DNA damaging agents on M . s m e g m a t is ,

M . TUBERCULOSIS AND M . MICROTI...... 95

5.1 Introduction ...... 95 5.2 The optimal conditions for exposure of mycobacteria to DNA damaging agents ...... 96 5.3 The effect of DNA damaging agents on the kinetics of LexA cleavage and RecA induction in mycobacteria ...... 96 5.3.1 The effect of exposure to DNA damaging agents on the kinetics of LexA cleavage and RecA induction in M. smegmatis mc^l55...... 97 5.3.2 The effect of exposure to DNA damaging agents on the kinetics of LexA cleavage and RecA induction in M. tuberculosis H37Rv ...... 97 5.3.3 The effect of growth stage on the kinetics of RecA induction in M. tuberculosis H37Rv ...... 100 5.3.4 The effect of exposure to DNA damaging agents on the kinetics ...... of LexA cleavage and RecA induction in M. microti OV1254...... 100 5.4 The sensitivity of M. tuberculosis, M. smegmatis and M. microti to DNA damaging agents ...... 105 5.4.1 The minimum inhibitory concentration of ofloxacin and mitomycin C in M. tuberculosis, M. smegmatis and M. microti ...... 105 5.4.2 The effect of exposure to ultra-violet (UV) irradiation on the viability of M. tuberculosis, M. smegmatis and M. microti ...... 107 5.5 Discussion ...... I l l

6.0 Determination of whether the M . m ic r o t i r e c A gene possesses

AN INDUCIBLE PROMOTER...... 118

6.1 Introduction ...... 118 6.2 Cloning of the M. microti recA coding region plus regulatory sequences in pBluescript IIK S ...... 119 6.3 Confirmation of the presence of an inducible promoter for the M microti recA gene ...... 119 6.3.1 Preparation of constructs ...... 119 6.3.2 CAT assays of mycobacterial transformants containing DNA from upstream of the recA gene ...... 121 6.4 Sequence determination of XhtM microti recA gene and 700bp of the upstream region ...... 121 6.5 Discussion ...... 125

7.0 Interaction of LexA protein with the M ycobacterium m ic r o t i

r e cA operator region...... 129

7.1 Introduction ...... 129 7.2 Binding of the M. microti LexA homologue to the operator region of the recA g en e ...... 130 7.3 The altered binding observed for M. microti cell-free extracts is due to the specific interaction of a LexA homologue with the conserved Cheo box motif ...... 132 7.4 Investigations to determine if M. microti LexA is binding as a monomer to a half site in the Cheo box motif ...... 132 7.5 Discussion ...... 135

8.0 G e n e r a l D is c u s s io n ...... 142

8.1 The mycobacterial RecA protein ...... 142 8.2 Levels of precursor and spliced RecA protein in untreated mycobacterial cell-free extracts ...... 143 8.3 The effect of exposure to DNA damaging agents on the kinetic profile for RecA induction and LexA cleavage in mycobacteria ...... 147 8.4 Gene regulation in mycobacteria ...... 153 8.5 Expression of recA in M. microti is constitutive ...... 154 8.6 Future work ...... 160 References ...... 163 Appendix I: Media recipes ...... 205 Appendix II: Growth curves ...... 210 Appendix III: Polyacrylamide gel and buffer recipes ...... 213 Appendix IV: Poster abstract ...... 215 L is t o f Fig u r e s

1.1 A general model illustrating the molecular mechanisms of excision repair ...... 28 1.2 The pivotal role of the RecA protein ...... 31 1.3 The molecular mechanism of homologous recombination ...... 34 1.4 Post-replicational repair in E. coli: a model of DNA damage tolerance ...... 36 1.5 The repair of double strand breaks by an Incisional-recombinational repair mechanism ...... 38 1.6 A diagrammatic representation of the mechanism by which DNA damage inducible genes are regulated by the combined action of the products of the recA and lexA genes ...... 39 1.7 Autoproteolytic cleavage of the LexA protein in E. coli ...... 41 1.8 Post-translational modification of the M tuberculosis RecA protein ...... 45 3.1 A schematic illustration of the purification of recombinant M. tuberculosis RecA protein...... 64 3.2 SDS-PAGE analysis of semi-pure and pure recombinant M. tuberculosis RecA...... 65 3.3 Determination of the antigen titre of polyclonal antisera raised against M tuberculosis RecA...... 66 3.4 A schematic illustration of the coupling of M leprae RecA precursor protein to CNBr- activated Sepharose 4B ...... 69 3.5 A comparison of the specificity of M. tuberculosis RecA antisera prior and subsequent to affinity purification ...... 71 4.1 Confirmation that affinity purified M. tuberculosis RecA antibody will recognise RecA precursor protein ...... 77 4.2 A comparison of basal RecA levels in uninduced cell-free extracts ofM smegmatis and members of the M tuberculosis complex ...... 79 4.3 The effect of growth stage on precursor and mature RecA (38kDa) levels in M. tuberculosis andM microti ...... 81 4.4 A comparison of basal RecA levels in clinical isolates of M. tuberculosis ...... 82 4.5 A comparison of basal RecA levels in a range of saprophytic and pathogenic mycobacteria 84 4.6 Basal RecA levels in M. smegmatis and members of the M tuberculosis complex cultivated in either Middlebrook 7H9 or Sautons media ...... 85 4.7 The effect of oxygen starvation on basal RecA levels and the appearance of precursor RecA protein in M tuberculosis ...... 87 4.8 Confirmation of the presence of a LexA homologue in M. smegmatis and the M. tuberculosis complex and a comparison of basal levels ...... 89 5.1 The effect of exposure to ofloxacin on RecA levels in M. smegmatis mc^l 55...... 98 5.2 RecA induction and LexA cleavage in M. smegmatis mc^l55 following exposure to mitomycin C ...... 99 5.3 Western blot monitoring the effect of treatment with a DNA damage-inducing agent on RecA and LexA levels in M. tuberculosis for 5 hours subsequent to exposure ...... 101 5.4 RecA induction and LexA cleavage in M. tuberculosis ...... 102 5.5 The effect of growth stage at the point of treatment with a DNA damage-inducing agent on

10 the kinetic profile for RecA induction in M tuberculosis H37Rv ...... 103 5.6 Western blot monitoring the effect of treatment with a DNA damage-inducing agent on RecA and LexA levels m M microti OV1254 for 7 hours subsequent to exposure ...... 104 5.7 RecA induction and LexA cleavage in M microti ...... 106 5.8 The effect of exposure to ultra-violet irradiation (254nm) on the viability of M. tuberculosis, M. smegmatis and M. microti ...... 110 6.1 The preparation of constructs to determine whether M microti recA possesses an inducible promoter in its upstream sequence ...... 120 6.2 Expression of M. microti recA is inducible inM smegmatis mc^l55...... 122 6.3 Strategy for sequencing the M. microti recA gene and 700bp of upstream sequence ...... 123 6.4 Sequence trace illustrating the two-nucleotide changes found upstream of the M. microti recA gene ...... 124 6.5 Complete nucleotide sequence of 671bp upstream of the M. microti recA gene ...... 126 7.1 Binding of the M. microti LexA homologue to the operator region of the recA gene ...... 131 7.2 Binding of the M. microti LexA homologue to the Cheo box motif is specific ...... 133 7.3 Confirmation that the second, faster migrating complex is due to specific binding of M. microti LexA to the Cheo box motif...... 134 7.4 Mutations in a half site of the Cheo box motif affect binding of M. microti and M. tuberculosis l^QyJ^ipxotém ...... 136 7.5 The affect of excess purified M. tuberculosis LexA on binding to the wild type and mutated Cheo box motif ...... 137 AII.l Growth curve for M smegmatis mc^l55 in Lemco broth ...... 210 AII.2 Growth curve for M. tuberculosis H37Rv in Dubos media ...... 211 AII.3 Growth curve for M. microti OV1254 in Tween-glutamate media ...... 212

11 L is t o f T a b l e s

2.1 Plasmids and constructs used in this study ...... 49 2.2 Mycobacterial species used in this study ...... 50 5.1 The minimum inhibitory concentration of ofloxacin for M. tuberculosis H37Rv, M. smegmatis mc^l55 andM microti OV1254...... 108 5.2 The minimum inhibitory concentration of mitomycin C for M. tuberculosis H37Rv, M. smegmatis mc^l55 andM. microti OV1254...... 109

12 L is t o f A bbreviations

bp- Base pair CAP- Catabolite activator protein CAT- Chloramphenicol acetyltransferase CFUs- Colony forming units DAB- Di-aminobenzidene dATP- Deoxyadenosine 5’-triphosphate dCTP- Deoxycytidine 5’-triphosphate dGTP- Deoxyguanosine 5’-triphosphate DIG- Digoxigenin DMSO- Dimethyl sulphoxide dsDNA- Double stranded deoxyribonucleic acid DTNB- Di-thionitrobenzidene dTTP- Deoxythymidine 5’-triphosphate EDTA- Ethylenediaminetetraacetic acid HR- Homologous recombination IPTG- Isopropyl-P-D-thiogalactopyranoside IR- Illegitimate recombination IS- Insertion elements kb- Kilo-base kDa- Kilo-dalton LPR- Long patch repair MAC- Mycobacterium avium complex MDMR- Methyl directed mismatch repair mRNA- Messenger ribonucleic acid NTM- Non-tuberculous mycobacteria OD- Optical density PBS- Phosphate buffered saline PCR- Polymerase chain reaction PEG- Polyethylene glycol PVDF- Poly-vinylidene-di-fluoride RFLPs- Restriction fragment length polymorphisms SDS- Sodium dodecyl sulphate SOD- Superoxide dismutase SPR- Short patch repair SSB- Single strand binding proteins ssDNA- Single stranded deoxyribonucleic acid TBS- Tris buffered saline

13 TE- Tris-EDTA buffer TEMED- N, N, N ’, N ’-tetramethylethylenediamine Tris- 2-ammo-2-hydroxymethylpropane-1,3-diol TTBS- Tween-tris buffered saline UV- Ultra violet X-gal- 4-Bromo-3-chloro-2-indolyl-beta-galactosidase

14 1.0 Introduction

1.1 The Mycobacteria

Lehmann and Neumann (1896) first established the generic name Mycobacterium to describe the tubercle and leprosy bacilli, both of which are pathogenic to man. It now encompasses over 50 approved species, the majority of which are not deleterious to man.

1.1.1 Characteristics

Mycobacteria are aerobic, acid-alcohol fast actinomycetes that are morphologically small curved or straight rods. They are biologically part of the Gram-positive group but have some distinctive features, notably with respect to the composition of their cell wall. All bacteria possess peptidoglycan, a cross- linked polymer of sugars and amino acids. However, the molecules attached to or associated with this polymer are predominantly lipids in mycobacteria as opposed to the proteins and polysaccharides of other bacteria. In general, the mycobacterial cell wall is composed of galactose, arabinose, meso- diaminopimelic acid and straight chain fatty acids as well as long chain mycolic acids, the size and stmcture of which is characteristic of this genus (Goodfellow and Wayne, 1982; Hinrikson and Pfyffer, 1994; Brennan and Nikaido, 1995). Mycobacteria, in common with other actinomycetes such as Streptomyces, have a high genomic G+C content which varies from 62-70%, with the exception of Mycobacterium leprae which has a G+C content of 56% (Goodfellow and Wayne, 1982; Imaeda et al, 1982; Clark-Curtiss et al, 1985). Genomic size determinations revealed that most mycobacteria have large genomes (3.0-4.5 x 10^ daltons) compared to other prokaryotes {E. coli 2.5 x lO^daltons; Clark- Curtiss et al, 1985; McFadden et al, 1987a). Classification within the genus is dependent upon their growth characteristics under optimal conditions in vitro, with members being either rapid or slow growers. Rapid growers yield visible colonies in 2-7 days; slow growers require 2-6 weeks. Growth of M. leprae in vitro has not been demonstrated. However, the discovery that the nine-banded armadillo {Dasypus novemcinctus Linnaeus) and mouse footpads are susceptible to infection has provided a successful, if slow, method for in vivo propagation of this mycobacterium.

1.1.2 Mycobacterial pathogens: a re-emerging problem

Mycobacteria were among the first bacteria to be associated with human disease. Mycobacterium leprae (Hansens’ bacillus) was identified by Hansen (1874) as the causative agent of leprosy and Mycobacterium tuberculosis was identified by Koch (1882) as that of tuberculosis. Despite major advances in diagnosis, treatment and prevention the magnitude of the TB and leprosy problem is still enormous from a global perspective.

15 The global figure for the prevalence of leprosy is an estimated 5.5 million people, with 20-30% of cases resulting in deformity and subsequent social stigmatisation (Nordeen et al, 1992). Effective chemotherapy exists, but medical supervision and patient compliance are problematical since the disease is common in areas remote from medical supervision. A specific vaccine for leprosy does not yet exist, although preliminary trials with BCG have had limited success.

Archaeological evidence alludes to the existence of tuberculosis in the Western Hemisphere for millenia. Historians have established the existence of TB during the period 4000 to 2000 B.C following the discovery of mummies with tuberculosis of the bone, characterised by vertebral fusion and deformity of the spine, known as Pott’s disease (Bloom and Murray, 1992). It has been postulated that TB occurred as an endemic disease among animals, the causative agent being M. bovis, long before it affected humans. It is thought to have remained an unimportant disease among humans until the early 1600’s when the disease spread with increasing population density and the penchant of Europeans to travel to and colonise distant continents (Steele and Ranney, 1958; Daniel et al, 1994). Determination of the infectious aetiology of TB and, therefore, an effective treatment were slow in coming. However, with the arrival of the twentieth century came the development of an effective vaccine, BCG (Sakula, 1983). The subsequent introduction of antibiotics such as isoniazid, streptomycin and, more recently, rifampicin formed the basis of an effective chemotherapy that dramatically reduced mortality from TB.

The BCG (Bacille Calmette Guerin) vaccine, an attenuated strain of M. bovis, is currently the most widely used vaccine in the world. It is relatively stable and inexpensive, and a single inoculation can provide long lasting sensitisation. It has the potential for being a useful vehicle for delivering protective antigens against multiple pathogens and has powerful adjuvant properties, being able to stimulate both humoral and cell-mediated immunity with minimal serious side effects. In addition, although predominately a vaccine against TB, it has been shown to provide protection against leprosy in trials and observational studies in Asia, Africa and Latin America (Fine and Rodrigues, 1990). Despite the wide usage and the aforementioned advantages of BCG it remains the most controversial of all currently used vaccines, as its protective efficacy has varied in different parts of the world and its impact on tuberculosis world-wide remains unclear (Bloom and Fine, 1994; Horwitz et al, 1995).

In the 1950’s it was recognised that non-tuberculous mycobacteria (NTM) were emerging as pathogens deleterious to man. Opportunistic pathogens such as the M avium-intracellulare complex (MAC) and M. xenopi, which are common within the environment, were shown to be the cause of rare pulmonary disease. Recent interest in these organisms has renewed following the advent of the human immunodeficiency virus (HIV); it has been estimated that they are responsible for 3% of HIV-related deaths (Horsburgh, 1991; Jiva et al, 1997). Furthermore, species such as M. marinum andM. ulcerans cause self-limiting granulomas and progressive malignant ulcers of the skin respectively, and Mycobacterium paratuberculosis, the causative agent of Johne’s disease in cattle and sheep (Chiodini et al, 1984a) has been implicated in the aetiology of Crohn’s disease in man (Chiodini et al, 1984b;

16 McFadden et al, 1987b). However, there is still a great deal of controversy on this subject (Chiodini, 1989; Elsaghier et al, 1992). Few vertebrates are immune to mycobacterial infection but perhaps more worrying is the knowledge that some mycobacteria, such as M. bovis and M. avium, can be transmitted between different host species.

As little as twenty years ago TB was thought to be a disease on the decline. However, an alarming resurgence has occurred. It has been estimated that 1.7 billion people, about one third of the world’s population, are now asymptomatically infected withM tuberculosis. There are approximately 10 million case of tuberculosis world-wide with an annual mortality rate of 3 million (Murray et al, 1990; Kochi, 1991). Many of these infected people reside in developing countries where famine, war and natural disasters create large populations of displaced, malnourished people. Attention has again focused on this ancient disease due to the failure of the BCG vaccine to afford significant global protection (Bloom and Fine, 1994), the spectre of multi-drug resistance strains, and the pernicious combination of infection with M. tuberculosis and the human immunodeficiency virus, HIV (Bloom and Murray, 1992; Narain et al, 1992).

1.2 Molecular biology of mycobacteria: current knowledge and progress

During the past century, major insights into the cellular immunology, physiology and biochemistry of mycobacteria have increased our understanding of the pathology of TB and its impact upon the immune system, though our knowledge is still incomplete (Bloom and Murray, 1994; Horwitz et al. 1995; Ottenhof and Mutis, 1995; Ortalo-Magne et al, 1995). In other bacteria genetic analysis has provided a powerful tool for the elucidation of fundamental biological systems. However, success in using molecular genetics to unravel the complexities of the intricate host-pathogen relationship in mycobacteria has been hampered. The predominant obstacles include the slow generation time of species such as M tuberculosis (18-24 hours) and the hydrophobic nature of mycobacteria, which leaves them prone to clumping during culture. This makes it difficult to obtain individual cells for genetic analysis. The lack of known genetic exchange between mycobacteria, and the small number of identifiable mycobacterial markers that could be exploited for the development of genetic tools have added to the problem (Shinnick et al, 1995). A significant amount of progress has been made in the past ten years despite this delay in the application of molecular biology techniques. It is anticipated that the use of recombinant DNA technology will aid in elucidation of the mechanisms of mycobacterial pathogenesis and immunity as we strive to develop more effective vaccine and treatment regimens.

1.2.1 Bacterial cloning hosts for expression of mycobacterial genes: gene expression in a heterologous background

Due to the lack of systems available for the direct manipulation of mycobacterial genes, Escherichia coli was initially used as a cloning host. Utilisation of this genus was advantageous because of the wealth of

17 genetic knowledge and procedures available for this bacterium. Genomic libraries for M. tuberculosis, M. bovis BCG and M leprae were constructed in E. coli using both plasmid and phage expression vectors, (Clark-Curtiss et al, 1985; Young et al, 1985a; 1985b; Jacobs et al 1986b; Rivoire et al, 1994). The development of recombinant DNA systems for the expression of mycobacterial genes in E. coli represents a significant advance in the study of defined antigenic molecules. Over 50 mycobacterial antigens have been characterised using this methodology, several of which have been identified as proteins with important functions (Young et al, 1991; 1992). However, there are inherent limitations associated with the use of E. coli as a cloning host.

Thole et al (1985) demonstrated that only a small number of M. bovis BCG promoters were active in E. coli. Similarly, Clark-Curtiss et al (1985) demonstrated that mycobacterial DNA could be expressed under the control of an E. coli promoter but mycobacterial promoters were either poorly recognised or not recognised at all by the E. coli transcription machinery. This observation made investigations into the expression and regulation of mycobacterial genes difficult. More recently, additional evidence has accumulated which suggests that some post-translational modifications of mycobacterial proteins do not occur in E. coli, thereby altering the properties of the expressed antigen (Garbe et al, 1993).

An alternative host that would express mycobacterial genes without the need for a foreign promoter was required. Bibb and Cohen (1982) had previously investigated gene expression in Streptomyces, which are also Gram-positive actinomycetes that have a high genomic G+C content (73%). They demonstrated that S. lividans recognised promoters from several bacterial genera including E. coli. Bacillus licheniformis and Serratia marcesans suggesting that Streptomyces were versatile in their ability to recognise heterologous promoters. Furthermore, Westpheling et al (1985) identified two forms of RNA polymerase with alternative sigma factors, which were implicated in the transcription of different promoter classes. Investigations with M tuberculosis, M. bovis BCG and M. leprae DNA (Kieser et al, 1986; Lamb et al, 1986) demonstrated that many genes from these species could be expressed from their own promoters in S. lividans.

1.2.2 Mycobacteria as cloning hosts: gene expression in a homologous background

The limitations inherent with using E. coli as a cloning host for mycobacterial DNA have been discussed above. Although alternative hosts such as Streptomyces remain viable alternatives, the most efficient means of investigating mycobacterial expression and regulation is in a homologous background. The use of non-pathogenic strains such as M. smegmatis and M. bovis BCG as cloning hosts for DNA from virulent strains would minimise the problems inherent in handling pathogenic mycobacteria. However, the full promise of recombinant technology for the genetic analysis of mycobacteria could only be achieved by the development of techniques promoting the stable introduction of recombinant molecules into mycobacterial cells. Despite many attempts over the years by many people it initially remained an elusive goal. Several contributing barriers were hypothesised, namely: the difficulty of introducing DNA

18 through the lipid cell wall, restriction of foreign DNA, unstable replication or maintenance of a plasmid introduced into the mycobacterial environment, and the failure to express a selectable marker gene. Any combination of these could have been responsible for the initial lack of success. The following paragraphs chronicle the progress of recombinant DNA technology in mycobacteria.

1.2.2.1 Mycobacteriophage and shuttle-phasmids

Within a suitable host bacteriophages will selectively redirect the focus of the cell into expressing their proteins, replicating their genome and directing the synthesis of a complex macromolecular virion. Historically they have proved invaluable as experimental tools aiding in elucidation of the nature of bacterial mutation and also in defining systems of recombination, gene expression and regulation. First described in 1947, mycobacteriophages were instrumental in determining whether any of the hypothesised barriers to transformation did indeed contribute to the lack of success. They enabled the process of transformation to be dissected into a number of discrete steps: DNA entry, avoidance of mycobacterial restriction systems, and the stable integration of foreign DNA. This systematic evaluation eventually led to a successful transformation methodology. In recent years they have played key roles in the development of novel cloning vectors enabling the more classical genetic approach to be superseded by recombinant DNA techniques (reviewed by Hatfull and Jacobs, 1994).

Development of shuttle-phasmid technology was a significant breakthrough enabling the introduction of recombinant DNA into fast and slow growing mycobacteria. Phasmids are hybrid molecules that replicate as a plasmid in E. coli and as a bacteriophage in mycobacteria. These modified phage vectors were constructed by insertion of an E. coli cosmid-cloning vector into the mycobacteriophage genome. Plasmid DNA was extracted from E. coli and transfected into M. smegmatis protoplasts, which lack a complete cell wall, where mycobacteriophage genes were expressed and lytic growth initiated. After lysis, mycobacteriophage packaged shuttle-phasmids were efficiently introduced into a broad range of mycobacteria (Jacobs et al, 1987).

Snapper et al (1988) extended the shuttle-phasmid cloning strategy by applying the same methodology to develop vectors based on the temperate mycobacteriophage LI, which is able to form stable lysogens by chromosome integration in M. smegmatis. These authors can also be credited with the expression of the first selectable marker in mycobacteria, achieved by insertion of a cartridge conferring kanamycin resistance into a unique restriction site within the LI shuttle-phasmid vector. Further improvements in the versatility and capacity of the phage based cloning system were achieved by Lee et al (1991; 1993) when the minimum sequences necessary for site specific integration into the mycobacterial chromosome were identified and incorporated into plasmid vectors. Pascopella et al (1994) used these integrating vectors to identify a genomic fragment associated with virulence in M. tuberculosis H37Rv. These vectors have potential as powerful tools for the constmction of recombinant BCG vaccines expressing a novel protective repertoire.

19 1.2.2.2 Mycobacterial plasmids and shuttle vector technology

Progress towards the identification of naturally occurring plasmids in this genus was made only a decade ago, despite the fact that mycobacteria are among the oldest known microorganisms. It is interesting to note that there is no strong evidence for the occurrence of plasmids in M. tuberculosis, M. bovis BCG or M. smegmatis. However, plasmids are common in the M. avium complex (MAC) and have also been found in M scrofulaceum, M. intracellulare andM fortuitum (Crawford and Bates, 1979; 1986; Hull et al, 1984; Franzblau et al, 1986; Meisner and Falkinham, 1986; Goto et al, 1991). The best characterised and possibly the most widely used plasmid replicon to date is that of the M. fortuitum plasmid pAL5000 (Labidi et al, 1984; 1985; 1992), the native function of which, as is true for the majority of mycobacterial plasmids, is unknown. Identification of naturally occurring mycobacterial plasmids was important because of the possibility that they may encode factors involved in drug resistance and virulence as well as their central role in recombinant DNA technology.

Exploitation of these important genetic tools was initially thwarted, specifically in slow growers such as M. tuberculosis, by the lack of an efficient transformation method. The structure of the cell wall and their slow growth combined to present a formidable barrier. Transformation in cell wall deficient M. smegmatis (spheroplasts or protoplasts) could be accomplished with comparative ease using polyethylene glycol (PEG) but the usefulness of this technology was limited due to the failure in regenerating viable bacilli. Crucial achievements in the molecular biology of this genus were the exploitation by Snapper et ûf/ (1988) of electroporation, a technique which vastly improved the transformation of mycobacteria, and the subsequent isolation of a high frequency transformation mutant of M. smegmatis, i.e. M smegmatis mc^l55 (Snapper et al, 1990). Our ability to express genes in mycobacteria is now vastly improved.

A variety of plasmid replicons based on pALSGGO have been combined with E. coli plasmid replicative origins to fashion E. co/i-mycobacteria shuttle vectors. These are capable of extrachromosomal replication in E. coli and mycobacteria (Ranes et al, 199G; Stover et al, 1991; Hinshelwood et al, 1992; Jain et al, 1997). Operon fusions in E. co/i-mycobacteria shuttle vectors have been used to study gene expression in mycobacteria, for example, focusing on “environmentally-active” promoters within macrophages (Dhandayuthapani et al, 1995). This technique has also provided a convenient method for the isolation of new mycobacterial promoters as well as providing insights into the specificity of the RNA polymerases of M. smegmatis and M. bovis BCG (Timm et al, 1994). Other shuttle plasmids not based on M. fortuitum pAL5GGG include pYT72/pYT92, a derivative of the M. scrofulaceum plasmid pMSC262 (Goto et al, 1991), and derivatives of plasmid RSFIGIG (Hermans et al, 1991) and pNG2 (Radford and Hodgson, 1991) which contain non-mycobacterial, broad host range replicons.

A technique for the rapid analysis of plasmid DNA from mycobacterial transformants exploited the inherent ability of shuttle plasmids to replicate in two distantly related bacterial species and the propensity for DNA molecules to be admitted or released from a cell subjected to electroporation.

2G Recombinant mycobacterial colonies mixed with electro-competent E. coli and subjected to a single electric pulse aided in the rapid isolation of recombinant mycobacterial DNA from E. coli in half the time that would otherwise be required for isolation from the host mycobacterial species (Baulard et al, 1992). More recently, further advances have been made toward increasing the transformation efficiency of slow growers such as M. tuberculosis and M. bovis by electroporation at elevated temperatures (Wards and Collins, 1996).

1.2.2.3 Insertion elements and transposons

Insertion (IS) elements and transposons are discrete DNA fragments able to move from one site in a DNA molecule to another independent of host recombination functions. These mobile genetic elements have been isolated in most organisms that have been examined so far and are useful genetic tools for insertional mutagenesis. The first example of an IS element found in mycobacteria was IS900 in M. paratuberculosis, discovered while researchers were looking for strain-specific probes by differential hybridisation (Green et al, 1989). They have since been discovered in several mycobacterial species and include IS6110 and IS 1081, specific to the M. tuberculosis complex (Thierry et al, 1990a; van Soolingen et al, 1992) and IS 902 of M. avium (Moss et al, 1992). The species specificity evident for such elements as IS900 and IS6110 has been exploited for strain differentiation using a novel typing method. An alternative assay method for strain differentiation of mycobacteria in clinical samples has been demonstrated by the combined use of restriction fragment length polymorphism’s (RFLP’s), IS specific probes and the polymerase chain reaction (McFadden et al, 1987; Thierry et al, 1990b; Eisenach et al, 1990;van Soolingen et al, 1992; Groenen et al, 1993).

The development of integrative vectors for insertional mutagenesis was an important goal as it would provide the basic framework for the construction of new vaccines and aid in the investigation of virulence genes as well as defining new drugs able to inhibit their synthesis (McAdam et al, 1994). Transposon mutagenesis is one of the most powerful methods of mutagenesis. It involves using a mobile element to disrupt genes randomly in the chromosome upon transposition. The first transposition events in M. smegmatis were reported using Tn610 (Martin et al, 1990) and transposons engineered from IS900 and IS986 (England et al, 1991; Fomukong and Dale, 1993). Transposition in BCG has been reported using a transposon constructed from IS 1096 (Cirillo et al, 1991; McAdam et al, 1995). However, until recently the only successful methods reported for the successful isolation of auxotrophic mutants of M. tuberculosis were allelic exchange (Balasubramanian et al, 1996) and illegitimate recombination (Kalpana et al, 1991). Efficient transposon mutagenesis in M. tuberculosis has been limited by the inefficiency of the delivery systems. More recently, a successful approach used to deliver transposons into M. smegmatis, which exploited the use of a conditionally replicating vector able to replicate at 30°C but not at 37°C (Guilhot et al, 1994), has been adapted to generate transposon libraries of M tuberculosis mutants. Bardarov et al (1997) constructed a conditionally replicating shuttle phasmid in mycobacteriophages D29 and TM4 using an E. coli cosmid vector, carrying TnlO(kan) or Tn5367, inserted into a non-essential region of the phage genome. Infection of mycobacteria at the non-permissive

21 temperature resulted in highly efficient transposon delivery to the entire population. A second successful approach exploited a replicative delivery vector that is efficiently lost under specific conditions. Pelicic et al (1997) used the counterselective properties of the sacB gene (Pelicic et al, 1996) and a mycobacterial thermosensitive origin of replication to successfully deliver derivatives of IS 1096 into the M. tuberculosis and M. bovis BCG genome.

1.2.3 Expression of foreign genes in mycobacteria

The initial objective for the expression of foreign genes in mycobacteria was the development of a recombinant BCG vaccine delivery system for heterologous antigens. Primary candidates for the construction of expression vectors were the promoters of the stress-regulated proteins. These were chosen to drive the expression of foreign genes in mycobacteria because their expression was observed under all growth conditions and their sequences were among the first available (Young et al 1985; Shinnick et al, 1987; Dellagostin et al, 1993). Other successful examples include the promoter of the alpha antigen of M. kansasii, used to express an epitope of the HIV-1 gag p 17 as a fusion protein, and a promoter sequence isolated from M paratuberculosis (Pan) used to express lacZ in M. bovis BCG (Matsuo et al, 1990; Murray et al, 1992). More recently, sequences encoding the 19kDa-lipoprotein promoter and a signal peptide from M. tuberculosis have been isolated and used by Stover et al (1993) in M. bovis BCG.

Development of the shuttle plasmid technology was a significant step in enabling the expression of foreign genes in mycobacteria, but these plasmids are often unstable in non-selective media. Genes encoding resistance to antibiotics are typically used to select for transformants in other bacteria but the potential for antibiotic selection of mycobacterial transformants is limited. Cell surface glycolipids are thought to be responsible for the impermeability of many commonly used antibiotics, and mycobacteria are naturally resistant to (3-lactams such as ampicillin. Resistance to kanamycin is a commonly used marker in M. tuberculosis, M. smegmatis and M bovis\ M. smegmatis mc^l55 is very sensitive to this antibiotic at low concentrations (Snapper et al, 1988; Ranes et al, 1990; Goto et al, 1991; Hermans et al, 1991). Chloramphenicol is another commonly used marker although its spontaneous mutation rate makes it less attractive for direct selection and it is often used in conjunction with other resistance genes (Snapper et al, 1990). Hygromycin has also been used with both fast and slow growers (Radford and Hodgson, 1991) and has been reported to be more efficient than selection based on kanamycin resistance (Garbe et al, 1994).

Foreign genes have been stably integrated into the chromosome thereby negating the need for continual selection, an obvious advantage in animal model systems. Integration has been achieved by site-specific recombination using the integration system of the mycobacteriophage L5 (Stover et al, 1991; Pascopella et al, 1994), DNA transposition (England et al, 1991) and by homologous recombination, although predominantly in fast growers (Husson et al, 1990). The recalcitrance of slow growers such as M. tuberculosis and M. bovis BCG to undergo homologous recombination (discussed in more detail below) has impeded allelic exchange for mutational analysis in these species. However, current gene replacement

22 technology should be invaluable for investigation into the mechanism of drug action and resistance as well as mycobacterial pathology.

1.3 Gene regulation

Gene expression in prokaryotes has to be controlled in a precise fashion to maximise the metabolic efficiency, this is particularly important for co-ordination of the expression of multi-gene pathways. From DNA to final protein product there is a multi-step pathway incorporating transcription, RNA processing and translation. Regulation of gene expression can be achieved at any stage. Transcription by DNA- dependent RNA polymerases is a cyclic process composed of four steps: promoter binding and activation, RNA chain initiation and promoter escape (Krummel and Chamberlin, 1989), RNA transcript elongation and RNA transcript termination. Transcriptional control is of paramount importance and will be considered further in the context of RNA polymerase activity and the action of positive and negative regulatory proteins. The reader is referred to general molecular biology texts and a review by Uptain et al (1997) for a more comprehensive discussion of gene expression and its regulation.

1.3.1 Initiation of transcription is catalysed by RNA polymerase: A principal step in the regulation of gene expression

Transcription begins with the interaction of RNA polymerase and a site-specific sequence termed the promoter. RNA polymerase consists of a core protein (composed of two a chains, one (3 chain and one (3’ chain) and a fifth dissociable subunit, sigma (a) (Chamberlin, 1982). The core protein has the ability to synthesise RNA on a DNA template but cannot initiate at the promoter in the absence of the sigma (a) subunit. The function of the sigma factor is to ensure that RNA polymerase binds specifically at a promoter site (McClure, 1985; Helmann and Chamberlin, 1988) and, therefore, plays a key role in determining promoter specificity. This is particularly obvious in B. subtilis, a Gram-positive bacterium that possesses nine types of sigma factor used for gene expression at different stages of growth and development (Haldenwang and Losick, 1980; Haldenwang et al, 1981; Losick et al, 1989; Haldenwang, 1995). Although E. coli do not undergo sporulation, it has been reported that they also possess multiple sigma factors: a (a^), (a^), (a^, (ct®) and a'®. The factor mediates promoter recognition under normal conditions (Gralla, 1990; Collado-Vibes et al, 1991) whereas and mediate the heat-shock or stress response (Grossman et al, 1984; Cowing et al, 1985; Strauss et al, 1987; Erickson et al, 1987; 1989; Helmann and Chamberlin, 1988). Some genes are controlled by minor ct factors: the subunit recognises genes involved in nitrogen metabolism (Merrick, 1993), is essential for the expression of some stationary phase genes (Tanaka et al, 1993; 1995; Jishage et al, 1996), mediates the expression of genes for motility and chemotaxis (Amosti et al, 1989; Kundu et al, 1997) and a'® is involved in transcriptional regulation of the genes for extracytoplasmic functions (Angerer et al, 1995).

23 In E. coli, the a ™ promoter has been extensively studied and is characterised by the consensus motifs 5’-TATAAT-3’ at -10 and 5’-TTGACA-3’ at -35 bases respectively from the transcription startsite. Frequently, promoter sequences in other prokaryotes have a different consensus, which can be an obstacle during heterologous gene expression. The sequence at -35 is the site of RNA polymerase recognition and that at -10 the site of formation of a stable polymerase-DNA complex. A comparison of known E. coli a promoters has shown that they all possess a similar, but not necessarily identical, sequence at these two sites and there is generally a good correlation between promoter strength and the degree to which the -10 and -35 elements agree with the consensus. The specificity of promoter recognition, and therefore promoter strength, is affected by both the nature of the promoter and the nucleotide distance between the -10 site, often termed the Pribnow box (Pribnow, 1975a), and the -35 site. In E. coli this distance is usually 17-19 base pairs (Harley and Reynolds, 1987). Other potential regulatory factors affecting initiation of transcription include ions, protein co-factors and DNA supercoiling (Mishra et al, 1990).

1.3.2 Regulatory proteins

Transcription initiation can be subject to either positive or negative regulation by the binding of a regulatory protein at a specific site (the operator) which is adjacent to, or overlaps, the promoter. Examples of negative regulation include repression of the lac operon by the lac repressor (reviewed by Lewin, 1990a) and the action of LexA protein on damage inducible genes in E. coli, where binding of a repressor prevents the expression of more than 20 unlinked genes (Little and Mount, 1982; Walker, 1984). Binding of these proteins to regions of dyad symmetry is a common feature of many prokaryotic repressors (Brent and Ptashne, 1981; Anderson et al, 1985; Jordan and Pabo, 1988; Wolberger et al, 1988; Cheo et al, 1991). An example of positive regulation can be found at the E. coli operon that codes for the enzymes required for arabinose degradation. Arabinose binds to a protein, AraC, inducing a conformational change that allows it to bind to DNA, which in turn stabilises the interaction of RNA polymerase with the promoter. A second clearly defined example of positive regulation is described by catabolite repression in E. coli where the catabolite activator protein (CAP) stimulates transcription of proteins encoded by the lac operon. Binding of cAMP to CAP induces a conformational change in CAP and concomitant binding to a site upstream of the promoter, thereby initiating transcription (de Crombrugghe and Pastan, 1978). The exact mechanism by which this is achieved is not precisely clear although it has been hypothesised that a stronger RNA polymerase binding site is exposed through bending of the DNA (Liu-Johnson et al, 1986; Gartenberg and Crothers, 1988).

1.3.3 Gene regulation in mycobacteria: Current knowledge and progress

Section 1.2.1 briefly discussed the lack of expression of mycobacterial genes from their own promoters in a heterologous system, and this is further exemplified by the lack of complementation of a range of

24 auxotrophic defects in E. coli by a mycobacterial cosmid library (Clark-Curtiss et al, 1985; Jacobs et al, 1986). This defect in expression is believed to be due to the difference in the transcriptional control systems between two groups of organisms having a large variation in the G+C content of their DNA. In his original model for binding of RNA polymerase to the Pribnow box, Pribnow (1975b) hypothesised that the formation of a stable polymerase-DNA complex involved limited unwinding of the DNA around the promoter region and that, thermodynamically, A-T base pairs “melt” more readily than G-C base pairs. The concepts of his original model are still in effect today and may partially account for the lack of expression of mycobacterial promoters within E. coli. Bibb and Cohen (1982) have demonstrated that a comparable situation exists between E. coli and Streptomyces, which, in common with mycobacteria, are Gram-positive actinomycetes that have a high genomic G+C content (73%).

The failure to express mycobacterial promoters in heterologous systems is not absolute, however. Thole et al (1985) demonstrated that a small number of M. bovis BCG promoters were active in E. coli although less efficiently than in their natural host. Further examples of mycobacterial genes expressed from their own promoters in E. coli include the 65kD heat shock protein and the 16S rRNA gene of M leprae (Mehra et al, 1986; Sela and Clark-Curtiss, 1991), M. tuberculosis (Shiimick, 1987; Ji et al, 1994b) and M. bovis BCG (Thole et al 1987; Suzuki et al, 1991). Biotin carrier proteins from several mycobacterial species (Collins et al, 1987) and HPq M tuberculosis 38kD antigen (Anderson et al, 1988) have also been successfully expressed from their own promoters in E. coli.

The nature of mycobacterial promoter sequences has received significantly more attention than other potential gene regulation mechanisms largely due to the ease of computer analysis of DNA sequences and the advent of recombinant techniques such as those described in section 1.2. However, analysis of these studies requires caution because very little is known about the nature of mycobacterial promoters. A positive result using recombinant techniques is not irrefutable evidence that it functions as a promoter in its original context. Similarly, analysis of potential promoter sites by computer algorithm gives no indication that the site concerned possesses promoter activity within the mycobacterial environment. This situation is not helped by the sparseness of information on specific mycobacterial promoters and the fact that few are active within E. coli (Dale and Patki, 1990).

Only a few mycobacterial promoters have been studied. These include: promoters for the 16S rRNA gene of M. smegmatis (Ji et al, 1994b), M. tuberculosis (Ji et al, 1994a; Verma et al, 1994), M bovis BCG (Suzuki et al, 1991) and M. leprae (Sela and Clark-Curtiss, 1991); the bla gene of M. fortuitum (Timm et al, 1994b); the cpn-60 gene of M. tuberculosis (Kong et al, 1993); the 65kDa and 85kDa antigen genes of M. tuberculosis (Shiimick, 1987; Kremer et al, 1995); the rpsL promoter ofM smegmatis, M. tuberculosis, M. bovis and M leprae (Kenney and Churchward, 1996); the Promoter ffomM paratuberculosis transposon IS900 (Murray et al, 1992) and three promoters associated with the transcription of the repressor-like gp71 protein of mycobacteriophage L5 (Nesbitt et al, 1995). A few of these regions contain sequences that resemble the typical E. coli promoter but many belong to specifically regulated genes and differ substantially.

25 With the exception of the heat shock proteins, very little is known of how mycobacterial gene expression may be modified by external conditions at the molecular level. Anderson et al (1990) demonstrated that the 38kDa protein of M tuberculosis may be induced by phosphate starvation and it has been suggested that the 28kDa antigen of M. leprae may be iron regulated but further evidence in this area is not yet available (Dale and Patki, 1990). Analysis of promoter function in vivo, particularly in pathogenic mycobacteria, is of great importance; however, only a few of the nucleotides important in determining promoter function have been identified (Nesbitt et al, 1995; Bashyam et al, 1996; Kenney and Churchward, 1996). A greater understanding of the regulatory sequences involved in the expression of vimlence determinants may aid in the development of new vaccines and a more effective treatment regimen. This may be facilitated by the system developed by Dellagostin et al (1995) enabling the assessment of promoter strength during in vivo growth of both M. leprae and M. tuberculosis.

Information on the nature of mycobacterial RNA polymerases is sparse. Very little information is available regarding their precise nature, although an RNA polymerase and two sigma factors from M. smegmatis have been characterised (Predich et al, 1995). Investigations comparing RNA polymerases in M. smegmatis and M. bovis BCG have suggested that they do not share the same specificity (Timm et al, 1994b), however, this is in direct contradiction to the more recent findings of Bashyam et al (1996). These authors demonstrated that the efficiency and specificity of transcriptional regulation in M. tuberculosis, M. smegmatis and M bovis BCG were conserved. The discordance between these observations may be accounted for by the fact that Timm et al (1994b) were using the promoters of specifically regulated genes. There is a distinct lack of information describing the precise nature and regulation of mycobacterial promoters; it is clear that further intensive investigation is required.

1.4 DNA damage in bacteria: a quick review

DNA damage is an inescapable aspect of life in the biosphere. The vast majority of both prokaryotes and eukaryotes have had to contend with the effects of solar radiation since the beginning of evolution of hfe on this planet. Furthermore, the primary stmcture of DNA is dynamic and subject to constant change as a result of errors introduced during replication and recombination. Consequently, a system that maintains the integrity of DNA is essential for survival in this hostile environment.

DNA damage can be classified into two general classes: alterations to the primary stmcture, and stmctural distortions. The former class affects sequence only and includes a wide range of mutations which, since replication is unaffected, can lead to heritable alterations in DNA. Examples include point mutations, deletions, insertions, duplications and inversions (these can involve large pieces of DNA affecting multiple genes). However, deletions, insertions and some point mutations frequently lead to the complete loss of function of a genetic locus (i.e. a null mutation) and are often eliminated from the gene pool. Conversely stmctural distortions, such as pyrimidine dimers formed following exposure of DNA to radiation at a wavelength approaching its absorption maximum (260nm), present a physical impediment to replication.

26 These dimers are bulky helix distorting lesions whose presence results in the obligatory arrest of replication. (For a detailed review of these mutations the reader is referred to Friedberg et al, 1995.)

1.5 DNA damage repair and tolerance systems in prokaryotes

There are two possible cellular responses to DNA damage: DNA repair or DNA damage tolerance. DNA repair culminates in the restoration of normal base sequence and structure and is achieved by the direct reversal of damage or excision of the damaged sequence. The second cellular response to DNA damage, termed DNA damage tolerance, does not include removal of the primary damage to DNA and therefore, has mutational potential. The latter cellular response is more relevant to the content of this thesis than DNA repair and will be discussed in more detail after a rapid review of basic DNA repair mechanisms in prokaryotes. For more detailed information on the following repair mechanisms both Lewin (1990b) and Friedberg et al (1995) provide good general reviews.

1.5.1 DNA repair in prokaryotes

Direct reversal of DNA damage is rare. The best-characterised example is the photoreactivation of pyrimidine dimers following exposure to ultra-violet (UV) light. A light dependent enzyme, a photolyase, removes the offending intrastrand covalent bonds, thereby restoring structural integrity. Direct reversal of DNA damage may be kinetically advantageous to a cell because it would, presumably, occur more rapidly than multi-step biochemical pathways such as excision repair (see below). Furthermore, the process would tend to be error free because of the high degree of specificity exhibited by such reactions. The types of DNA damage repaired by direct reversal are limited; a more general repair mode is one in which the damaged sequence is excised from the genome and covalent integrity restored by replication. This is termed excision repair. The general nomenclature for excision repair distinguishes between the modes of recognition of the damaged region, namely; base excision, nucleotide excision and mismatch repairs. These mechanisms vary in their specificity but share the same general features, see figure 1.1.

Base excision repair. This is initiated by the action of a specific class of DNA repair enzymes called DNA glycosylases which exclusively recognise a damaged, mispaired or inappropriate base, for example uracil. Incision occurs by a two-step reaction mvolving the action of a DNA glycosylase and an AP endonuclease (either apurinic or apyrimidinic) culminating in the release of the damaged, free base (figure 1.1). The length of the incision is then extended by the action of an exonuclease and new synthesis occurs to replace the missing nucleotides until covalent integrity is restored.

27 Base excision repair General excision repair

q<^qq

q<^qq<^qq^qq^q B a se is rem oved Incision

Endonuclease cuts on both sides of damaged base q^qq^qq^qq^q

Apurinic/Apyrimidinic I endonuclease makes cut I I q^qq^qq

Exonuclease removes q^qq^qq

qqqqqqqqqq<^ q Ligase i se a ls X nick T

q<^qq

Figure 1.1 A general model illustrating the molecular mechanisms of excision repair

Right For nucleotide excision or mismatch repair, damage to tlie genome is recognised and “nicks” introduced into the flanking sequence. The damaged region is removed by the action of an exonuclease and new synthesis initiated to restore covalent integrity. Left The pathway for base excision repair, recognition and removal of an incorrect base is mediated by the action of DNA glycosylases. A specific endonuclease introduces a nick into the site of damage, which is further extended by the action of an exonuclease. Covalent integrity is restored by the initiation of DNA synthesis to fill the gap.

28 Nucleotide excision repair. This is a multi-step process, which culminates in the removal of damaged bases from DNA as intact nucleotides rather than free bases. The general mode of repair is outlined in figure 1.1. Damage is recognised, excised and new DNA synthesis initiated to restore covalent integrity. There are two types of nucleotide excision repair distinguished on the basis of the length of heterogeneous DNA repaired, i.e. short-patch (SPR) and long-patch repair (LPR). Both utilise a damage specific endonuclease (UvrABC). They differ in that SPR is constitutive whereas LPR is induced and has an absolute requirement for induction of the SOS response (see section 1.6.4). In E. coli, LPR correlates with a more rapid recovery and increased survival following DNA damage and is associated with the ability to bypass lesions at or near replication forks. In nucleotide excision repair the genes uvrA, uvrB and uvrC code for the components of an endonuclease that recognises the site of damage and cuts the flanking DNA. An enzyme that possesses 5' to 3' exonuclease activity (in E. coli this is believed to be DNA polymerase I) then removes the intervening sequence. DNA ligase forms a covalent bond to restore genetic integrity following synthesis of new DNA, again by the action of DNA polymerase I, although DNA polymerase II or III can substitute for it.

Mismatch repair. The molecular mechanisms of this repair mode are related to both base and nucleotide excision repair. Mismatched base pairs can arise as a result of the formation of heteroduplex complexes as part of a recombinational process (see section 1.6.1), replication errors or by, for example, deamination of 5- methylcytosine converting a G C pair to a G T pair. The mispaired bases are normal constituents of DNA; therefore enzyme scanning for a lesion or structural impediment is not sufficient for recognition and subsequent removal. For mismatch repair to contribute to the genetic fidelity of DNA the correct base in the mismatch must be distmguished from the incorrect base. In E. coli the best-characterised method for distinguishing between the strands is methyl directed mismatch repair (MDMR). In this bacterium GATC sites are targets for DAM methylase subsequent to replication. This reaction provides a means of targeting repair to the unmethylated donor strand using the methylated strand as a template. If neither strand is methylated then mismatch repair exhibits no preference. The general mechanism of MDMR involves the action of the gene products of mutH, mutL and mutS (also members of the SOS regulatory network) culminating in the generation of a nick in the DNA opposite a GATC hemi-methylated site. Hehcase II, the gene product of uvrD, leads to strand displacement followed by exonuclease activity and ultimately, removal of the mispaired base. MDMR is synonymous with LPR (see nucleotide excision repair) as the signals distinguishing strands are frequently a long way from the lesion, up tolO^bp.

An alternative recognition mechanism, also well characterised in E. coli, involves méthylation of the internal C’s at CCA/TGG sites by DCM methylase and is termed very short patch repair. Interestingly, m mycobacteria there is no detectable méthylation at either GATC (Dam) or CCA/TGG (Dcm) sequences (Hemavathy and Nagaraja, 1995).

There are many diverse types of base damage that have few chemical and structural features in common. Their recognition by the excision repair machinery is not precisely understood. It has, however, been suggested that the presence of many types of base damage that challenged cell survival during biological

29 evolution provided selection for a general excision repair mode which would serve to identify and effect the removal of multiple types of DNA damage.

1.5.2 DNA damage tolerance systems

Tolerance systems, synonymous with both post-replication and recombination repair in many texts, play an important role in allowing cells to deal with unrepaired damage to their genome. In higher eukaryotic cells tolerance systems are thought to play a significant role since a large genome makes complete repair unlikely. In simplistic terms they allow damaged sequences to be replicated in the presence of a significant proportion of errors. Prokaryotes have evolved a mechanism of dealing with both single strand (ss) and double strand (ds) breaks in DNA utilising proteins that have an important role in the homologous recombination (HR) of undamaged DNA (section 1.6.1). Additionally they possess a mechanism enabling them to process damaged DNA, involving polymerisation across lesions, generally termed trans-lesion synthesis (section 1.6.5). In contrast to homologous recombination, trans-lesion synthesis is a highly mutagenic process whose regulation is closely linked to the SOS regulatory network (see below). In E.coli, tolerance systems have been extensively analysed and are intimately related to the RecA protein which also regulates the complex SOS regulatory network, whose gene products have been implicated in both nucleotide excision and mismatch repair.

1.6 The RecA protein; a pivotal component of several cellular mechanisms

The RecA protein is ubiquitous among bacteria and exhibits dual activities, having both a regulatory and mechanistic role. To date the sequences of more than sixty prokaryotic recA genes are available and have been used as the basis for the constmction of a phylogenetic tree (Miller and Kokjohn, 1990; Karlin et al, 1995; 1996). RecA is a key component of general homologous recombination and has also been implicated in post-replication repair, mutagenesis and derepression of the proteins involved in both the SOS response and induction of 1 prophage into the lytic cycle, see figure 1.2 (Roca & Cox, 1990; Miller & Kokjohn, 1990; Cox, 1993; Kowalczykowski et al, 1994). It has been suggested that the occurrence of RecA depends, to a degree, upon environment (Roca & Cox, 1990).

Among spirochetes it has been demonstrated that many free living species possess an inducible RecA while some virulent parasitic species, which inhabit environments where they encounter less oxygen and UV light, are among examples that lack detectable RecA. It is possible that the challenge to cell survival posed by DNA damage and the ability of microbes to tolerate such damage and exploit certain ecosystems might provide the strong selective pressure required to develop and maintain such a complex molecule (Cox, 1993; Marias et al, 1996).

30 L'l'l-i 4i"i i' SOS response

'Activated DNA repair recA RecA RecA Cell survival gen e 1 Recombination / ^ protein ^ Cell division Mutagenesis

Figure 1.2 The pivotiil role of the RecA protein

31 A RecA protein monomer has a molecular weight of approximately 38kD and an alternating (3 strand-a helix structure. In E. coli, the RecA protein monomer and polymer have been extensively analysed and the functional domains associated with DNA binding and interaction with the LexA repressor have been described (Cox et at, 1987; Miller and Kokjohn, 1990; Story et al, 1992; Karlin et al, 1996; Rehrauer and Kowalczykowski, 1996; Stasiak, 1996; Takahashi et al, 1996; Voloshin et al, 1996). Under physiological conditions both single stranded DNA (ssDNA) and double stranded DNA (dsDNA) provide a scaffold for the formation of a nucleoprotein filament. Filament formation is a rapid process, initially occurring on ssDNA in the 5' to 3' direction. The presence of single-strand binding (SSB) proteins aids the formation of the nucleoprotein filament by removing the secondary structure of DNA. However, if bound to ssDNA prior to RecA they can inhibit the initial “nucléation” event, which is the rate-limiting step of this reaction (Cohen et al, 1983; Kowalczykowski, 1987; Cox, 1991). A major factor accounting for the low rate of nucléation on dsDNA is the absolute requirement for underwound DNA. (Cox, 1991) A single stranded gap provides an optimal nucléation site with extension occurring rapidly to incorporate adjoining duplex DNA. In the following paragraphs the functions attributed to RecA are described; E. coli is used as the primary paradigm because molecular mechanisms in this model have been extensively analysed and are generally the best characterised.

1.6.1 Homologous recombination

The genetic exchange of homologous regions of DNA between two DNA molecules is called general or homologous recombination (HR). The RecA protein is essential for this reaction, catalysing strand transfer between homologous DNA molecules in an ATP-dependent manner via a complex, multi-step pathway. There are other examples of recombination such as site-specific recombination, exemplified by the integration of bacteriophage genes into the host genome (Snapper et al, 1988; Lee et al, 1991; 1993), and transposition (Martin et al, 1990; Cirillo et al, 1991; England et al, 1991; McAdam et al, 1994; Bardarov et al, 1997; Pelicic et al, 1997). The critical difference between these processes and HR is that the enzymes involved act only on specific target sequences with a limited homology or, as with transposition, no homology at all. For its mechanistic role in HR, RecA has an absolute requirement for a nucleotide triphosphate and ssDNA, or a firee 3' end located within a region of homology. In its most simplistic form HR involves binding of RecA protein to ssDNA, pairing of the RecAzssDNA complex to dsDNA, alignment and pairing of homologous sequences on the genome, and finally strand exchange, which can proceed for several kilobases.

The precise molecular mechanisms are not understood with the same clarity as nucleotide excision repair, although identification of the DNA binding domains of E. coli RecA has led to speculation about their role in the search for homology (Baliga et al, 1995; Morimatsu and Horii, 1995; Podyminogin et al, 1995; Yancey-Wrona and Camerini-Otreo, 1995; Kurumizaka etal, 1996; Stasiak, 1996; Takahashi et al, 1996; Voloshin et al, 1996; Wittung et al, 1996; 1997). The active species in RecA mediated strand exchange is the nucleoprotein filament (Howard-Flanders et al, 1984; Cox, 1993). Following formation, this promotes the

32 search of a sister duplex for homologous contacts via the formation of a 3-stranded intermediate. Non- covalent base pairing, via the formation of hydrogen bonds, then serves to ahgn homologous sequences after which strand transfer proceeds in the 5' to 3' direction relative to the ssDNA. This process is termed synapsis and culminates in the formation of a Holhday junction, see figure 1.3 (Kowalczykowski, 1987).

The RecF, RecO, RecR and single-strand binding (SSB) proteins are believed to facilitate the loading of RecA onto ssDNA (Sawitzke and Stahl; 1992; Umezu et al, 1993) as well as participating at presynaptic steps in recombination (Lloyd et al, 1988; Clark, 1991; Wang et al, 1993; Kowalczykowski et al, 1994). The RecF, RecO and RecR proteins are thought to assist in the cleavage of LexA by aiding RecA activation (Volkert and Hartke, 1984; Volkert et al, 1984; Walker, 1985), possibly by promoting the preferential interaction of RecF with dsDNA containing ssDNA regions (i.e. gapped DNA; Griffin and Kolodner, 1990; Hedge et al, 1996; Courcelle et al, 1997). In addition to their reported role in recombination, the RecO and RecR proteins have been shown to alleviate the inhibitory effects of SSB proteins on RecA catalysed joint molecule formation (Umezu et al, 1993; Umezu and Kolodner, 1994). Interestingly, the SSB proteins have been shown to have a positive role in the activation of RecA in addition to their inhibitory effect (Vales et al, 1980; Whittier and Chase, 1981; Chase et al, 1983; Lieberman and Witkin, 1983). The precise nature of this role remains unclear (Johnson, 1977; Sevastopoulos et al, 1977; Meyer et al, 1982; Carlini and Porter, 1997) although SSB proteins have been shown to participate in at least two steps of genetic recombination, i.e. strand assimilation and branch migration (McEntee et al, 1980; West et al, 1982; Lieberman and Witkin, 1983). Subsequent to Holliday junction formation, branch migration proceeds aided by the nucleolytic action of RuvA and RuvB. RuvB has been identified as a DNA dependent ATPase, with RuvA serving to target RuvB to the Holliday junction (Shiba et al, 1991; Iwasaki et al, 1992; Parsons et al, 1992; Mandai et al, 1993; West, 1994; Hiom and West, 1995). Branch migration mediated by a combination of these proteins with RecA is more energy efficient than that promoted by RecA alone. At physiological temperatures spontaneous branch migration is slow, will occur only over short distances and is inhibited by regions of non-homology. In contrast RecA mediated branch migration (in conjunction with the RuvA, RuvB and SSB proteins) is rapid and can occur over several kilobases as well as through DNA lesions and substantial regions of non-homology (Kowalczykowski, 1987; Adams and West, 1995). Resolution can take place by reverse branch migration, mediated by the action of RuvA and RuvB or RecG (Smith, 1994; Adams and West, 1995). Alternatively, resolution can occur via cleavage of the Holliday junction, which is mediated by the endonuclease activity of RuvC (Mandai et al, 1993; West, 1994). Branch migration and the subsequent resolution of the Holliday junction are termed post-synapsis.

Branch migration is an isoenergetic process that involves no net loss or gains of hydrogen bonds; the precise role of ATP hydrolysis remains unclear since binding of ATP is essential for RecA mediated strand transfer but ATP hydrolysis is not (Menetski et al, 1989; Rosselli and Stasiak, 1990). One possibility is that it is required for RecA dissociation from the DNA substrates (Lindsley and Cox, 1990). Alternatively, it may allow branch migration to proceed through regions of non-homology or DNA lesions (Menetski et al, 1989; Rosselli and Stasiak, 1990; 1991; Shan et al, 1996) or render strand

33 z. T ■ z

Binding of RecA protein to ssDNA and initiation of strand exchange (RecO, RecR, RecF, SSB)

z

SYNAPSIS Holliday junction formation RecA (RecO, RecR, RecF)

z

POSTSYNAPSIS Branch migration RecA (RuvA, RuvB, SSB)

B Resolution by cleavage of Holliday junction (RuvC)

Resolution by cleavage at A Resolution by cleavage at B z z z A V

Figure 1.3 The molecular mech;mism of homologous recombination

34 exchange unidirectional (Roca and Cox, 1990; Rosselli and Stasiak, 1990; Cox, 1991; Jain et al, 1994; Shan et al, 1996). The mechanistic role of RecA in homologous recombination has been extensively characterised. Its interaction with, and the role of other rec genes in recombination has been the subject of further investigation (Smith, 1989). For example, the recE and recT genes code for exoVIII and RecT respectively. These two proteins promote homologous pairing and strand exchange in reactions containing linear duplex DNA and homologous, circular ssDNA as substrates (Kolodner et al, 1994).

Homologous recombination serves diverse biological roles: it allows the independent elimination of unfavourable mutations, and reshuffling of favourable mutations without any deleterious effect on the alleles with which they may be associated, an important factor in the generation of diversity. In the laboratory, homologous recombination is an important genetic tool that can be used to create and characterise specific mutations using the "gene knockout" technique, where a wild type gene is replaced by a mutated, non­ functional gene in order to study gene function (Capecchi, 1989).

1.6.2 Recombinational or post-replication repair

In E. coli, exposure to DNA damaging agents leads to a transient inhibition of DNA synthesis, predominantly due to the effect of lesions on replication forks. Although the precise mechanistic basis of re­ initiation of replication has not been clearly defined there is an absolute requirement for RecA synthesis. Two alternative molecular mechanisms that form the basis of re-initiation of replication on templates containing blocks will be discussed: post-replicational repair and trans-lesion synthesis (see section 1.6.5).

Post-replicational repair does not instigate the removal of the DNA lesion and, therefore, does not constitute “repair”; it is a mechanism for tolerance of damage to the genome. In this process, illustrated in figure 1.4, DNA rephcation resumes downstream of the damage by an unknown mechanism leaving a gap in the daughter strand opposite the lesion. RecA polymerises on the ssDNA in the gap to form a nucleoprotein filament. The formation of such a filament is thought to activate RecA, leading to induction of the SOS response (see below), as well as promoting homologous pairing and strand exchange with an undamaged sister DNA molecule. Strand transfer past the lesion closes the gap in the recipient molecule, resulting in the formation of a recombination intermediate (the Holliday junction).

Endonuclease cleavage across the point of strand exchange resolves the connection between the two sister strands leaving DNA polymerases to fill the gap left in the parental strand. The mechanistic basis of this post-replicational repair system, which efficiently deals with a lesion in "daughter” dsDNA, is identical to that previously described for homologous recombination, illustrated in figure 1.3.

35 Z. Z IReplication of damaged DNA z

T Binding of RecA protein to ssDNA and initiation of strand exchange (RecO. RecR. RecF. SSB) Z — 7 A

SYNAPSIS Holliday junction formation RecA (RecO. RecR. RecF)

POSTSYNAPSIS Branch migration RecA (RuvA. RuvB. SSB) I and synthesis of new DNA

\TT B Resolution by cleavage of Holliday junction (RuvC)

Resolution by cleavage at A Resolution by cleavage at B z Ml Ml 7

Figure 1.4 Post-replicatioiiid repair in E. colt a model of DNA damage tolerance

36 1.6.3 Repair of double strand breaks

The repair of double strand breaks demonstrates an absolute requirement for RecA, RecBCD and a homologous DNA duplex, see figure 1.5. RecBCD nuclease loads onto flush DNA ends and proceeds to unwind the duplex molecule leading to the formation of a ssDNA loop that enlarges as it travels along the DNA.

On reaching a correctly orientated Chi recombinational hotspot (5’-GCTGGTGG-3’, Taylor and Smith, 1995), RecBCD cleaves the strand with this sequence producing a long ssDNA region which acts as a substrate for RecA (Smith, 1989; Cox and Lehman, 1987; Cox, 1991). Following strand exchange, repair synthesis enlarges the “D-loop” from the donor dsDNA molecule. The subsequent cleavage of this displaced strand allows it to pair with the gap in the recipient molecule leading to the formation of a molecule with two recombinant joints. Resolution can take place either by reverse branch migration, mediated by the action of RuvA and RuvB or RecG (Smith, 1994; Adams and West, 1995), or by cleavage of the Holliday junction which is mediated by the endonuclease activity of RuvC (Mandai et al, 1993; Smith, 1994; West, 1994).

1.6.4 The SOS response: Physiological consequences and regulation

The SOS response is characterised by a defined set of physiological responses induced as a result of exposure to agents or conditions that damage DNA or perturb replication. Among these responses are an enhanced capacity for DNA repair, chromosome and phage mutagenesis, inhibition of cell division (filamentation) and activation of a co-protease function of RecA. This results in the increased expression of genes that form the basis of the extensively analysed SOS regulatory network (Gudas and Pardee, 1975; Little and Mount, 1982; Walker, 1984; 1995). The status of the SOS response can generally be defined by one of two extremes; fully repressed or fully induced, although in practice an intermediate state exists. In the fully repressed state LexA acts as an upstream repressor of more than 20 genes, including recA and lexA, by binding to a specific DNA sequence (the operator) and repressing transcription (Brent and Ptashne, 1980; 1981; Little and Mount, 1982). It is apparent that many of these genes are expressed at low levels even in this “repressed” state, which, in the case of recA, is sufficient for its role in homologous recombination. When the genome is damaged or replication impaired, an intracellular signal is generated and the SOS response induced. The precise nature of this signal is unknown although possible candidates include ssDNA (D’Ari, 1985; Sassanfar and Roberts, 1990; Higashitani et al, 1995; Morel-Deville et al, 1996; Shinagawa, 1996), ATP (Barbe et al, 1983; Guerrero et al, 1984; Thresher et al, 1988) and deoxyribonucleotides (Phizicky and Roberts, 1981; Menetski and Kowalczykowski, 1989). It is believed that binding of the RecA protein to ssDNA reversibly activates a co-protease function of RecA leading to the autocatalytic cleavage of the LexA protein (Little, 1984) and concomitant de-repression of the genes at which it binds (figure 1.6). The regions of ssDNA disappear as new DNA synthesis restores covalent integrity to the genome. As a consequence, RecA returns to its non-activated state and LexA once again represses expression of the genes of the SOS network.

37 z z. T

Double strand break provides substrate for RecBCD to enter, unwind and expose ssDNA Iregions z z_

Binding of RecA to ssDNA and initiation of I strand exchange (RecO, RecR, RecF, SSB) A Displaced strand forms D-Loop

Repair synthesis extends strand and enlarges duplex D-loop. Donor D-loop pairs with recipient 3' end irA a A

Repair synthesis converts donor D-loop to duplex. Branch migration creates molecule with two recombinant joints RecA (RuvA, RuvB, y SSB)

B i ^

1 z B

Resolution by cleavage (RuvC)

Non-crossover cleavage at A Crossover cleavage at B

^ ^ ------

Figure 1.5 The repjiir of double strand breidis by

38 recA gene SOS response g en es LexA repressor ------► = t RecA Protein SOS response Proteins / \ Repressorf accumulates DNA dam age Loss of RecA co-protease activity Inducing signal i Removal of inducing signal RecA co-protease activated

DNA repaired LexA repressor/ cleaved V X recA gene, . SOS response LexA g en es ^ repressor—► 11— RecA Protein SOS response'I Proteins Activated ------^ ^ RecA Protein DNA repair

Figure 1.6 A diagrammatic representation of the mechanism by which DNA damage inducible genes are regulated by the combined action of the products of the recA and lexA genes

39 The sequence and structure of the E. coli LexA protein have been extensively analysed, it is a 202 amino acid monomer (Horii et al, 1981b; Markham et al, 1981) containing two structurally defined domains linked by a hinge region. The N-terminal domain (amino acids 1-84) is the DNA binding domain (Little and Hill, 1985; Hurstel et al, 1986; Schnarr et al, 1991; Thliveris and Mount, 1992), while the C-terminal domain is called the oligomerization domain (Schnarr et al, 1985; 1988). The 80S response is activated upon proteolytic cleavage of the repressor into its two separate domains (Little, 1984) resulting in a reduction in the specific affinity of LexA repressor for the operator (Bertrand-Burggraf et al, 1987; Kim and Little, 1992). In addition to RecA-mediated cleavage in vivo, LexA cleavage has also been demonstrated in vitro at physiological pH in the presence of RecA and two types of cofactor; ssDNA and a nucleotide triphosphate such as dATP (Little et al, 1994), and at alkaline pH via a RecA independent mechanism (Little, 1984). Cleavage of the E. coli LexA occurs at a specific peptide bond between Ala-84 and Gly-85, see figure 1.7 (Horii et al, 1981a). Further mechanistic studies have led to the identification of two residues (Ser-119 and Lys-156) that are essential to the chemistry of cleavage (Slilaty and Little, 1987; Roland and Little, 1990). These residues and the flanking amino acids have been conserved in all LexA proteins sequenced to date for example; Salmonella typhimurium, Erwinia carotovora, Psuedomonas aeruginosa, Psuedomonas putida (Garriga et al, 1992), Providencia rettgeri, and Aeromonas hydrophila (Riera and Barbe, 1993; 1995), Bacillus subtilis (Raymond-Denise and Guillen, 1991; 1992) and M. tuberculosis (Movahadzadeh et al, 1997a). Both the X prophage repressor and UmuD (see section 1.6.5; Walker, 1995; Harmon et al, 1996; Peat et al, 1996; Rehrauer et al, 1996) are structurally related to LexA and interact with RecA by a similar proteolytic mechanism leading to cleavage at a similar bond.

Sequencing of the E. coli lexA gene (Horrii et al 1981a; 1981b; Markham et al, 1981) led to the identification of an imperfect palindromic upstream of the transcription start site. The general consensus TACTGT-fAT\-ACAGTA was also found upstream of most Gram-negative recA and all E. coli damage inducible genes; the underlined bases were shown to be essential for LexA binding (Brent and Ptashne, 1981; Wertman and Mount, 1985; Ottleben et al, 1991; Schnarr et al, 1991; Umelo et al, 1996). The SOS response in E. coli and B. subtilis are remarkably similar from both a phenotypic and regulatory standpoint (Walker, 1984; Lovett et al, 1988; Sassanfar and Roberts, 1990; Wojciechowski et al, 1991; Raymond- Denise and Guillen, 1991; 1992; Lovett etal, 1993; 1994). To elucidate the mechanisms of regulation in 5. subtilis Cheo e? a/ (1991) cloned the promoter regions of several DNA damage inducible loci.

Comparison of the nucleotide sequences with the promoters of other genes associated with DNA repair identified a different consensus sequence GAAC-N^-GTTC (termed a Cheo box) as the operator site and deletion analysis confirmed the importance of this sequence in the regulation of DNA damage inducible genes (Cheo et al, 1993). More recently, this motif has been identified in the recA promoter regions of Leptospira biflexa (Stamm et al, 1995) and the Gram-positive bacteria: M. tuberculosis, M. leprae, M. smegmatis. Staphylococcus aureus, Lactococcus lactis, Streptomyces lividans, Streptomyces ambofaciens, Clostridium perfringens and Streptococcus pneumoniae. (Davis et al, 1991; 1994; Duwat et al, 1992; Bay les et al, 1994; Nu|3baumer and Wohlleben, 1994; Pearce et al, 1995; Aigle et al, 1997; Johnston et al, 1997;

40 o II N-Terminal — ALA — C GLY Domain 84 A

LYS 156

SER 119

— OOC j

N-Terminal — ALA — C GLY Domain 84 i

— LYS 156

X- SER f 119 i N-Terminal Domain

C-Terminal Domain

Figure 1.7 Autoproteolytic cleavage of the LexA protein in E. coli

41 Papavinasasundaram et al, 1997). Interestingly, no SOS-like or Cheo box motif has been identified in the Gram-negative Helicobacter pylori (Thompson and Blaser, 1995) and a third consensus motif TTG-N,,- CAA has been identified upstream of the recA gene of three members of the Rhizobiaceae, namely: Rhizobium etli. Rhizobium meliloti ziad Agrobacterium tumefaciens (Tapias et al, 1997). Furthermore the sequence TTGT-N, ^-ACAA has been shown to be essential for DNA-damage mediated induction of the Rhodobacter capsulatas recA gene (Fernandez de Henestrosa et al, 1997).

Differential regulation of the SOS response enables the cell to utilise specific aspects of this inducible network, for example nucleotide excision repair (specifically LPR), without committing to the fully-fledged response. The extent to which a gene is expressed following induction of the SOS response is partially dependent upon the degree to which the pools of LexA repressor have been decreased. Other important factors affecting the degree of repression are operator strength, localisation of the operator relative to the promoter and promoter strength (Schnarr et al, 1991). Examples of genes that bind LexA weakly are uvrA, uvrB, uvrC (a damage specific endonuclease in excision repair) and lexA. These are de-repressed in response to a weak inducing signal, whereas sulA, umuC and umuD require a stronger iuducing signal. The gene product of sulA is an inhibitor of septation; expression leads to filamentation, which is a characteristic physiological response to SOS induction in E. coli that can be fatal (Shinagawa, 1996). The gene products of umuC and umuD have a role in trans-lesion synthesis, a highly mutagenic process that is de-repressed as a “last-resort” (see below). Other factors critical to the degree of repression of the SOS response are the ability of the LexA protein to dimerize, although it is not clear whether this is prior or subsequent to association with the operator site (Thliveris and Mount, 1991; Kim and Little, 1992), and the number of operator sites upstream of the damage inducible gene. The concentration range over which LexA will act as a repressor is further reduced by the presence of multiple co-operative binding sites. For example, the E. coli lexA gene has two operator sites and requires a smaller alteration in the concentration of LexA for de-repression than recA, which has only a single operator site (Schnarr et al, 1991). Repression of lexA by LexA has several advantages; it allows differential regulation of the SOS response and ensures a rapid return to the repressed state once the inducing signal is abolished (Brent, 1982). Furthermore, the weak affinity of LexA for its own promoter compared to that of recA provides a buffering capacity by limiting the inducing signal.

The inducible response of prokaryotes to DNA damage has been studied extensively. It is apparent that most bacteria possess an SOS regulatory network which has a significant number of factors in common with that characterised fox E. coli (Little and Mount, 1982; Walker, 1984). Examples iuclude; Salmonella typhimurium, Proteus mirabilis, Haemophilus influenzae, Caulobacter crescentus. Pseudomonas aeruginosa, Aeromonas salmonicida, and Gram-positive bacteria Bacillus subtilis, Streptomyces ambofaciens. Staphylococcus aureus. Streptococcus pyrogenes and Clostridium perfringens (Miller and Kokjohn, 1990; Wojciechowski et al, 1991; Raymond-Denise and Guillen, 1992; Bayles et al, 1994; Tao et al, 1995; Umelo et al, 1996; Aigle et al, 1997; Johnston et al, 1997). However, in Bacillus subtilis filamentation is a recA independent phenomenon and induction of the SOS response is subject to additional controls relating to the development of natural competence in this genus (Raymond-Denise and Guillen, 1992).

42 1.6.5 SOS Mutagenesis

Bacteria such as E. coli are equipped with an array of DNA repair enzymes enabling them to deal with a variety of lesions that may occur within the genome as a result of exposure to DNA damaging agents or conditions that perturb DNA replication. However, a situation can arise whereby damage fails to be processed by an error-free repair pathway and is instead processed by an error-prone pathway. This pathway is called SOS mutagenesis and culminates in trans-lesion synthesis where rephcation passes through a lesion in a continuous mode. This process is tightly regulated such that the “mutagenic” machinery is rapidly disassembled, enduring long enough to permit continued DNA replication and survival but not to support gratuitous mutagenesis (Walker, 1995; Woodgate and Levine, 1996).

The precise molecular mechanism of mutagenesis is unknown although the proteins involved have been defined. Replication across the site of damage is achieved by the combined action of UmuD'C, RecA and DNA polymerase III and results in the insertion of any base opposite the lesion (Rajagopalan et al, 1992). The biochemical properties of the Umu proteins that permit trans-lesion synthesis are not known, but then- elaborate regulatory mechanisms have been extensively investigated. Regulation occurs at the transcriptional level where the umu locus is very tightly repressed by LexA and consequently requires a large inducing signal for its de-repression. Regulation also occurs at the post-translational level. UmuD, which is structurally related to LexA, is functionally inactive until it undergoes RecA mediated cleavage to give UmuD' which can then combine with UmuC to catalyse trans-lesion synthesis (Walker, 1995; Frank et al, 1996; Peat et al, 1996; Woodgate and Levine, 1996). A further level of regulation is provided by the preferential dimerization of UmuD’ with UmuD, which is inactive, over the dimerization of UmuD’ with UmuD’, which is the species which interacts with UmuC (Battista et al, 1990; Jonczyk and Nowicka, 1996; Woodgate and Levine, 1996).

1.7 RecA mediated responses in mycobacteria

Mycobacterial pathogens such as M. tuberculosis survive and rephcate within macrophages, which are an integral part of the host defence mechanism. This intracellular lifestyle constantly exposes them to hostile agents such as hydrogen peroxide and reactive nitrogen intermediates. Consequently, there is great interest in determining the response of mycobacteria to DNA damage and its relevance to pathogenesis. This interest has increased since the identification by Davis e? a/ (1991) of the M. tuberculosis recA gene.

43 1.7.1 A novel post-translational modification for theM. tuberculosis RecA protein

RecA is ubiquitous among prokaryotes and is one of the most conserved proteins found in bacteria (Roca & Cox, 1990; Miller & Kokjohn, 1990; Cox, 1993; Kowalczykowski etal, 1994; Karlin etal, 1995). The stmcture of the M. tuberculosis recA gene is unusual in that it codes for a protein approximately twice the length of that found in other bacteria. Davis et al (1992) demonstrated that the 85kDa product of the M tuberculosis recA gene undergoes a novel post-translational modification to yield a 38kDa RecA protein and a 47kDa "intern" by eliinination of the intervening sequence and joining of the N- and C- terminal domains, see figure 1.8. At the time only one other reported example of this unusual mechanism of protein splicing had been documented and was found in the Saccharomyces cerevisiae VMAl gene, a vacuolar pump ATPase (Hirata et al, 1990; Kane et al, 1990). Further examples have since been discovered in the VMAl gene of Candida tropicalis (Gu et al, 1993) and the DNA pol genes of Thermococcus litoralis (Perler et al, 1992; Hodges et al, 1992) and Pyrococcus sp. (Xu et al, 1993). Analysis of a further 14 species of mycobacteria by Davis et al (1994) revealed that inteins were found only in the recA loci of the pathogenic mycobacterium M. leprae and members of the M. tuberculosis complex. Given the rarity of such elements among bacteria, their presence in different locations in the same gene of two related species is suggestive of a positive selection for splicing in the maturation of RecA and raises questions on the possible imphcations to pathogenicity and vimlence (Davis et al, 1994). Interestingly, an intein has recently been located in a second mycobacterial gene: the gyrh gene of M leprae, M.flavescens, M. gordonae, M. kansasii (Fsihi et al, 1996) and M. xenopi (Telenti et al, 1997). It is perhaps noteworthy that the majority of these inteins appear to be associated with genes that code for proteins involved in DNA metabolism. More recently, fourteen genes in the Methonococcus jannaschii genome have been shown to contain 18 inteins. Within this species the highest identity observed was 33% for a 380bp portion of two inteins, this suggests that the large number of inteins are not a result of this genetic element moving within the genome (Bult et al, 1996).

Sequence analysis of many of the inteins described above has shown them to be distinct from each other, although they do share some common features; notably a conserved hexapeptide motif at the intein-C-extein sphce junction, which suggests that there is a common mechanism for excision (Xu et al, 1993; Clarke, 1994; Colston & Davis, 1994). There is far less homology at the intein-N-extein splice junction. In addition to the conserved motif at the intein-C-extein splice junction, a second pair of conserved motifs located within the intein sequence has been identified. These motifs are characteristic of a group of endonucleases which include the HO endonuclease of yeast (Nakagawa et al, 1992) and the homing endonucleases of group I mobile RNA introns (Biniszkiewicz et al, 1994). These endonucleases mediate gene conversion by initiating recombination at specific sites in the genome resulting in the insertion of their own coding sequence. The invasion of an intron-less gene by an intron is called “homing”. Both homing and endonuclease activity has been demonstrated for the intein in the S. cerevisiae VMAl gene and endonuclease activity has been demonstrated in one of the two inteins in T. litoralis (Bremer et al, 1992; Gimble and Thomer, 1992). However, the mycobacterial recA and gyrA genes have not been shown to possess either homing or endonuclease activity (Fsihi et al, 1996). Apart from these conserved motifs, the inteins in the recA loci of M. tuberculosis and M. leprae bear little or no homology to each other, but the evidence that they are

44 DNA

RNA

Precursor Protein

N-exteIn C-exteIn Intein

Figure 1.8 Post-triUîsi;itioniil modification of the M. tuberculosis RecA protein

45 removed by post-translational splicing is compelling. Of particular significance was the demonstration that conservation of the amino acid sequence is essential for splicing whereas mutations that alter only nucleic acid sequence have little effect. Furthermore, substantial deletions were shown to decrease the abihty of inteins to splice indicating that structural integrity was essential for removal from the precursor (Davis et al, 1992).

1.7.2 Homologous recombination in mycobacteria

Homologous recombination (HR) in mycobacteria was first demonstrated by Husson et al (1990) when they achieved targeted allelic replacement in the M smegmatis pyrF gene. The approach they adopted utilised a ‘suicide-vector’, which favours the selection of integration events. Since then a similar methodology has been used by a number of other groups to achieve allelic exchange by homologous recombination in M. smegmatis (Kalpana et al, 1991; Mustafa et al, 1995; Gordhan et al, 1996). An alternative approach was developed by Pelicic et al (1996). These authors used sacB (the B. subtilis gene encoding levansucrose, which is lethal to mycobacteria in the presence of 10% sucrose) as a marker for the positive selection of gene replacement events in the M. smegmatis pyrF gene. This technique provided a powerful tool for the genetic characterisation of mycobacteria as well as a starting point for the development of a method promoting homologous recombination in M tuberculosis (see below).

Kalpana er a/ (1991) utihsed the suicide vector approach to attempt allelic exchange in the M. tuberculosis met gene. However, Southern blot analysis of transformants demonstrated that the endogenous met gene was intact and that the vector had randomly integrated into the chromosome. Illegitimate recombination (IR) was also the predominant event in M bo\is BCG transformed with linearised DNA fragments carrying an inactivated copy of the uraA gene (Aldovini et al, 1993). Despite the high frequency of IR these authors did demonstrate a low rate of allelic exchange at homologous loci. The factor distinguishing this study from that of Kalpana et al{\99\) was the utilisation of mycobacteria propagated under conditions to maximise the percentage of cells undergoing DNA synthesis. Interestingly, the recalcitrance of slow-growers to undergo HR is not universal. Marklund et al (1995) successfully demonstrated allelic exchange inM intracellulare using the suicide vector approach.

To achieve homologous recombination in the pathogenic mycobacteria a new approach had to be considered. Norman et al (1995) suggested that the lack of detectable homologous recombination events in the M. tuberculosis complex could be due to the severe limitation imposed upon recombination when non- repUcating DNA is used. They demonstrated successful homologous recombination and gene replacement in M. bovis BCG using a rephcating plasmid although, due to the deleterious effect instigated by intermption of the target gene, they were unable to propagate transformants successfully. Baulard et al (1996) adopted a similar approach to dissociate recombination events from transformation efficiencies by the use of a replicative plasmid and Reyrat et al (1995) demonstrated that the isolation of HR events is greatly facilitated by the availability of a phenotypic marker for identification of mutants. They successfully targeted the M.

46 bovis BCG urease gene ureC, this gene was chosen mainly because simple techniques for the rapid isolation of urease-negative colonies are universally available. On the basis of the successes and failures of previous attempts to demonstrate HR in the M. tuberculosis complex Balasubramanian et al (1996) concluded that the use of long linear DNA fragments as donors would increase the probability of detecting homologous recombination events. Transformation of M. tuberculosis with long linear DNA fragments containing an inactivated copy of the leucine biosynthetic gene leuD resulted in a relatively high recovery of integrants. Watt et al (1996) have since demonstrated, in E. coli, that the probability of an HR event is proportional to the length of homologous DNA taking part in the exchange. Another successful method for achieving allelic exchange in M. tuberculosis has recently been reported by Pelicic et al (1997). These authors have adapted the technique they developed to successfully achieve allelic exchange by homologous recombination in the M smegmatis pyrF gene (Pehcic et al, 1996) to M. tuberculosis. Using sacB (discussed above) and a thermosensitive mycobacterial origin of replication (derived from PAL5000) efficient allehc exchange in the purC gene of M tuberculosis has been demonstrated.

Over the past decade, tuberculosis has been one of the most intensively researched infectious diseases due to its resurgence coinciding with that of the AIDS epidemic. Consequently, the ability to achieve a high frequency of homologous recombination would facilitate the understanding of pathogenesis and permit the constmction of defined mutants of M. tuberculosis.

1.8 Aims

1.8.1 Background to the project

At the onset of this project homologous recombination had been demonstrated in M smegmatis but not in any member of the M. tuberculosis complex. Recombination systems in mycobacteria were thought to be broadly similar to those in E. coli on the basis of the limited data available. But why were slow growers so recalcitrant to HR? The presence of an intein in the recA gene of members of the M. tuberculosis complex and M leprae is provocative. It is suggestive of a positive selection for sphcing in the maturation of RecA — perhaps as a mechanism of regulation — and may have implications with respect to pathogenicity and vimlence.

Some mycobacteria are intracellular pathogens that reside within macrophages. It is likely that they will be exposed to many of the mechanisms involved in macrophage killing and consequently, to DNA damaging agents such as hydrogen peroxide and nitric oxide. The importance of the rec A genotype to microbial survival and vimlence has already been established in other bacterial species. Carlsson and Carpenter (1980) demonstrated that resistance to hydrogen peroxide in E. coli correlates with the recA genotype while 80S induction in Erwinia carotovora increases vimlence (McEvoy et al, 1990). More recently, Buchmeier et al

47 (1995) have shown that the ability to repair DNA damage in Salmonella typhimurium is more important to microbial survival than the catalase genotype. Although the thick, lipid-rich mycobacterial cell wall provides a formidable first line of protection, it is conceivable that an analogous requirement for DNA repair is an important survival mechanism of the pathogenic mycobacteria.

1.8.2 Specific aims of the project a) To investigate whether the splicing of the RecA precursor protein of M tuberculosis is conditionally regulated. b) If splicing is regulated, to determine the effect of regulation on RecA activity. c) To compare basal levels of RecA and LexA in different species of mycobacteria with particular emphasis on members of the M. tuberculosis complex. d) To investigate the kinetics of LexA cleavage and RecA induction following exposure to DNA damaging agents and to investigate the basis for any difference.

48 2.0 M aterials and M ethods

2.1 Bacterial strains and plasmids; media and growth conditions

Tabic 2.1 describes the plasmids and constructs used in this work. Escherichia coli strains TG2, DH5 and

DH5a (Sambrook et ah 1989) and Mycobacterium smegmatis mc“155 (Snapper c/ui/, 1990) were used for the cloning of mycobacterial DNA. Unless othei-wise stated E. coli cultures were grown in Luria-Bertani broth

(L-broth; appendix 1) with antibiotic selection at 37°C in a rotaiy shaker (250ij)m).

Table 2.1 Plasmids and constructs used in this study

PL.ASM ID CONSTRUCTION. SOURCE OR REFERENCE

pBluescriptll KS Stratagene pEJ 135 Intact M. tuberculosis recA locus, a 2.7kb /At/ll fragment cloned in pTZl 8R; Davis et at, 1991 pEJ 160 ArwApartial Xmnl deletion of pEJ 135; Daviset al, 1992 pEJ 169 EcoR\-Bci\ deletion of upstream sequences from pEJ 135, Daviset al, 1991 pEJ 170 Non-splicing mutant M. tuberculosis recA gene; Davis et al, 1992 pEJ257 CAT promoter probe vector; Movahedzadehet al, 1997b pFM7 0.7kb fragment of DNA comprising 667bp of upstream sequence and part of the M. tulm culosLs recA gene in pEJ257; Movahedzadeh el a i 1997b pFM 17 0.35kb fragment of DNA comprising 3 lObp of upstream sequence and part of the M. lulwculosis rccA gene in pEJ257, Movahedzadeh et al, 1997b pGPl-2 7.2kb plasmid carrying the T7 RNA polymerase gene and a temperature sensitive (30°C) origin of replication. Appligcne pK233 Whole ofM. leprae recA gene coding for 79kDa precureor protein. Dr. D Emery ^ pNT2 0.7kb Not\-Hind\\\ fragment of pNTmic2 ligated into pEJ257, digested with Notl-HindiW pNT12 0.35kb fm ll-ffrW lI fragment of pNTmic2 ligated into pEJ257, digested with fcoRV-f/mrfrll pNTmicl 3.1 kb Not\-Nae\ fragment of genomic M. microti DNA containing recA ligated into pBluescriptll KS, digested with Aotl-E'coRV pNTmic2 0.7kb Sacï-Xho\ fragment of pNTmic 1 ligated into pUC 19, digested with Sacl-Sall pTZ18R Mead CM/, 1986 pUC19 Yanisch-Perronc/ü/, 1985

National Institute for Medical Research. Mill Hill, London, NW7 1AA.

T able 2.2 lists the mycobacterial species used in this study. Cultures of M. smegmatis were routinely grown in Lemco broth in a rotary shaker (250rpm) and M. m icroti in Tween-glutamate media using a magnetic stirrer. Cultures ofM. tuberculosis H37Rv and all clinical isolates were cultivated in Du bos media in stationary flasks agitated on a daily basis and stationaiy cultures ofM bovis BCG were maintained in

49 Table 2.2 Mycobacterial species used in this study

SPECIES AND STRAIN SOURCE OR REFERENCE

M avium NCTC 08559 M. Im is NCTC 10772 M. IwvLs BCG Denmark Dr. I. Estrada*Garcia M. chetonei subspecies ahscessus ColmdaleNCTC. 10882 M. ckelonei subspecies chelonei Colindale NCTC 946

M fonuitim ColindaleNCTC 10394 M. fmfyana h M. intracellulare NCTC 10425 M. kansasii E239 Prof. E.C. Bottger*" M kansasii E242 Prof. E.C. Bottgcr" M kansasii E250 Prof E.C. Bottger^ M. malmonese M223/94 Prof. E.C. Bottger M. marinum NCTC 02275 M. microti OV1254 M. phlei M. scTofulaceum NCTC 10803 M. smegmatis mc'155 Snapper

National Institute for Medical Research, Mill Hill, London NW7 1AA

" Institute for Medical Microbiology, Hanover;

Dept. Medical Microbiology, St George’s Hospital Medical School, London SW17 ORE

Middlcbrook 7H9 broth. The remaining mycobacterial species were cultured in Dubos media or Lemco broth.

Unless otherwise stated all cultures were grown at 37°C to the optical density (OD^Kjnm) specified for each investigation. Media recipes are given in appendix 1.

50 2.2 Production and purification of recombinantM, tuberculosis RecA protein

2.2.1 Over expression of M. tuberculosis recA by heat induction ofE. coli pEJ169 {TG2(pGPl-2)}

All cultures of E. coli pEJ169{TG2(pGPl-2)} were grown at 32°C in L-broth (appendix I) containing kanamycin, ampicillin and methicillin each at a final concentration of 50|ig/ml. Ten aliquots of 200ml L- broth plus antibiotics were inoculated at 1/20 with an overnight culture of E. coli pEJ 169 (TG2(pGP 1 -2)} and returned to the same 32°C incubator for 3 hours. Fresh L-broth was heated to 62°C and 100ml added to each 200ml volume of culture. After incubation in a static water bath at 42°C for 40 minutes, cultures were returned to an incubator at 37°C for 3 hours. The bacteria were harvested by centrifugation at 5,000rpm in a Sorvall GSA rotor for 15 minutes.

2.2.2 Production of a crude preparation of recombinantM. tuberculosis RecA protein by differential precipitation

The bacterial pellet was washed in R-buffer (0-lM Tris, ImM dithiothrietol, O lmM EDTA, 10% glycerol) then resuspended in a 35ml volume of this buffer. After sonication (Dawe Instruments Ltd., Type 7532A) at 100 watts for 3 x 30seconds the suspension was centrifuged (13,000rpm in a Sorvall SS34 rotor for 35 minutes) and the supernatant retained. Polyethylenimine was added to a final concentration of 0 05% and the sample left on ice for 40 minutes. The precipitate was harvested by centrifugation (9,000rpm in a Sorvall SS34 rotor for 15 minutes), resuspended in 12ml R-buffer then (NHJ^SO^ added (150mM final concentration) and the sample incubated on ice for 45 minutes. After centrifugation (9,000rpm in a Sorvall SS34 rotor for 15 minutes) the supernatant was retained, the precipitate resuspended in 4ml R-buffer and the (NHJ2SO4 precipitation repeated. After centrifugation (9,000rpm in a Sorvall SS34 rotor for 10 minutes) the supernatants were combined and incubated with 20%w/v (NHJ 2SO4 over night at 4°C. Precipitated protein was harvested by centrifugation (13,000rpm in a Sorvall SS34 rotor for 30 minutes) and resuspended in a minimum volume of R-buffer.

2.2.3 Purification of a crude preparation of recombinant M. tuberculosis RecA protein by SDS-poIyacrylamide gel electrophoresis

The protein preparation was denatured in sample buffer at 90°C for 2-3 minutes (4x sample buffer: 30% glycerol, 0-35M SDS, 0-6M dithiothrietol, 0-012% w/v bromophenol blue). After loading onto a 10% denaturing SDS-polyacrylamide gel (appendix III) it was run to completion either overnight at 15V or for 6 hours at 40mA. The RecA protein band was localised by reversible copper staining (Lee et al, 1987), the relevant band excised and the gel shce immersed in destain (0-25M EDTA, 0-25M Tris-Cl) for 15 minutes. This latter procedure was repeated a further 5 times before transfer of the excised gel to dialysis tubing (visking size 2-18/32", Medicell International Ltd., London) containing elution buffer (20mM Tris, 150mM

51 glycine and 0 01% SDS). Following electroelution overnight at 15V the sample was concentrated to a volume of 200-500fil using a Centricon-10 concentrator (Amicon) according to the manufacturer’s instructions. Elution buffer was replaced with phosphate buffered saline (PBS: 1 9mM NaH2?04, 8- ImM NagHPO^, 154mM NaCl, pH 7 4) by buffer exchange during concentration.

2.3 Production of polyclonal antisera raised against recombinant Af. tuberculosis RecA protein

Purified recombinant M. tuberculosis RecA protein (3mg) was emulsified with an equal volume of incomplete Freud’s adjuvant. Each of three rabbits received approximately Img total protein subcutaneously into six sites (contracted to Harlan). Two further immunisations were administered at 21 and 42 days and the progress of antibody levels in a test bleed determined using a standard ELISA technique (Current Protocols in Immunology; Unit 2.1, basic protocol). Exsanguination followed 7 days after a booster immunisation of 200|ig/rabbit of purified RecA protein in aqueous solution. Serum was obtained from whole blood and stored at -20°C until required.

2.4 Preparation of an affinity chromatography column and purification of polyclonal antisera raised against recombinant M. tuberculosis RecA protein

2.4.1 Production of Af.leprae RecA precursor protein

Four 750ml volumes of Terrific broth (appendix I) containing 50pg/ml ampicillin were inoculated 1/60 with an overnight culture of E. coli harbouring plasmid pK233. After 3 hours at 37°C, cultures were induced with ImM isopropyl-P-D-thiogalactopyranoside (IPTG) and returned to the incubator overnight. Bacteria were harvested by centrifugation (5,000rpm in a Sorvall GSA rotor for 15 minutes) and washed in PBS-Tween (l -9mM NaHzPO^, 8-ImM NazHPO^, 154mM NaCl, 0 2% Tween-80, pH 7 4). The bacteria were resuspended in 30ml sonication buffer (50mM Tris-Cl at pH8.5, 7mM 2-mercaptoethanol, ImM Pefabloc — a serine protease inhibitor) and cell membranes dismpted by sonication at 100 watts for 3 x 30seconds (Dawe Instruments Ltd., Type 7532A). Following centrifugation (13,000rpm in a Sorvall SS34 rotor for 35 minutes) the pellet was retained and RecA precursor protein dissolved by continual agitation in 4ml buffer (6 M guanidium-HCl, 50mM Tris-Cl at pH8.5,7mM 2-mercaptoethanol, ImM Pefabloc). After centrifugation (15,000 rpm in a Sorvall SS34 rotor for 10 minutes) the supernatant was retained and stored at -20°C until required.

52 2.4.2 Determination of the number of reduced sulfhydryl groups inM. leprae RecA precursor protein

The number of reduced sulfhydryl groups present in M. leprae RecA precursor protein was estimated using an adaptation of the method described by Elhnan (1959). A solution of lOmM di-thionitrobenzidene (DTNB) in 6 M guanidium-HCl (pH7.2) was prepared immediately prior to use. A 600|il volume of 6 M guanidium-HCl (pH7.2) was placed in a 1ml cuvette and used to zero a spectrophotometer at Following the addition of lOOpl of DTNB solution the increase in was noted. To this cuvette 40|il of

M. leprae RecA precursor protein in 6 M guanidium-HCl (section 2.4.1) was added and the OD^iznm monitored for 10 minutes. A rapid increase in is characteristic of the presence of reduced sulfhydryl groups in the protein.

2.4.3 Coupling of theM. leprae RecA precursor protein to CNBr-activated Sepharose 4B

A 5ml volume ofM leprae RecA precursor protein in 6 M guanidium-HCl (section 2.4.1) was precipitated with 1/9 volume of trichloro-acetic acid. The precipitate was harvested by centrifugation (2,500rpm in a Sorvall HS4 rotor for 10 minutes), washed in coupling buffer (0 IM NaHCO^, 0-5M NaCl, pH8-3) and centrifuged as before. The pellet was dissolved in 1ml di-methyl sulphoxide (DMSO) and the volume increased to 5ml with coupling buffer. The protein content was determined using the BCA micro-titre plate assay (Pierce) according to the manufacturer’s instructions.

The coupling procedure used for CNBr-activated Sepharose 4B and M. leprae RecA precursor protein is based upon the method described by Pharmacia (1986/8). CNBr-activated Sepharose 4B (0 75g) was washed and re-swollen with ImM HCl (200ml/g), allowed to settle under gravity and the majority of HCl removed. The gel suspension was mixed with lOmg M leprae RecA precursor protein (section 2.4.1) in 20ml coupling buffer and left on a rotary mixer overnight at 4°C. The coupling buffer was removed under gravity and all remaining active sites blocked by incubation of the gel suspension with blocking buffer ( 20ml; IM ethanolamine, pH 8 0) for 2 hours at room temperature. Approximately 2 5ml of the gel suspension was transferred to a disposable chromatography column and excess adsorbed protein removed by washing alternately with 10ml acetate buffer (0 IM CHjCOjNa, 0-5M NaCl, pH4 0) then 10ml coupling buffer. These washing stages were repeated once more before removal of excess blocking buffer by washing twice with coupling buffer (10ml per wash). The column was stored at 4°C with 01% sodium azide in PBS until required.

2.4.4 Purification of polyclonal antisera raised against M. tuberculosis RecA protein by affinity chromatography

Polyclonal antisera raised against M. tuberculosis RecA protein was purified by a modification of the method described by Hudson and Hay (1976). The prepared column was washed with 20ml PBS then 5ml of polyclonal antisera applied slowly at 0 5ml / minute. Unbound antisera was removed by passing PBS through the column until the percentage absorbance of the eluate at 260/280nm was <10. The antigen-

53 antibody complex was dissociated by applying 0 IM glycine-HCl (pH2-5) slowly to the column at 0 5ml / minute. The eluate from this stage was collected and titrated to pH8-5 with IM Tris base to prevent protein dénaturation. When the percentage absorbance of the eluate at 260/280nm was <10 the dissociation conditions were altered to maximise the removal of the bound antibody by the addition of 10% dioxane to the dissociation buffer (Anderson et al, 1979). Collection and titration of the eluate to pH8 5 with IM Tris base continued until the absorbance at 260/280nm was <1 0. The pH adjusted eluate was concentrated to T5-2 5ml using a centricon-10 concentrator (Amicon) according to the manufacturer’s instructions and stored in lOOpl aliquots at -20°C until required.

2.5 Production of mycobacterial cell-free extracts

Mycobacterial cultures were grown to the required optical density (ODgoonm) ^nd checked for contamination by the Ziehl Neelsen or Kinyoun staining protocol (Silverton and Anderson, 1961). Bacteria were harvested by centrifugation (10,000rpm in a Sorvall GSA rotor for 10 minutes) and washed with PBS-Tween (1 9mM

NaHgPO^, 8 ImM Na2HP 04, 154mM NaCl, 0 05% Tween-80, pH 7 4). The cell pellet was resuspended in 100-500pl of PBS with Pefabloc (a serine protease inhibitor) and a half volume of glass beads (0 1mm), the cell membranes were then dismpted using a bead beater (2x30 seconds at 4,000 rpm, Biospec Products). Cell-free extracts were harvested by centrifugation (13,000rpm in a microfuge for 5 minutes) and their protein content estimated using the BCA micro-titre plate assay (Pierce) according to the manufacturer’s instmctions.

2.6 Western Blotting

Cell-free extracts were denatured by boiling with sample buffer (4x sample buffer; 30% glycerol, 0-35M SDS, 0-6M dithiothrietol, 0*012% w/v bromophenol blue) and proteins separated on a denaturing SDS- polyacrylamide gel (appendix III). Following incubation for 15 minutes in transfer buffer (25mM Tris, 194mM glycine and 15% methanol) the gel was laid on three sheets of 3MM filter paper pre-cut to the size of the gel. The uncovered side of the gel was overlaid with pre-cut poly-vinylidene-di-fluoride membrane (PVDF Immobilon-P, Millipore) and three further sheets of 3MM filter paper. Proteins were transferred using a standard Western blot protocol (Current Protocols in Immunology: Unit 8.10, Protein blotting with tank transfer systems). Membranes were blocked overnight at 4°C (or at room temperature for 3 hours) in blocking buffer (10% milk powder in TBS: 20mM Tris-Cl, 0-5M NaCl and pH7*5), then incubated for 2 hours at room temperature with polyclonal antisera diluted 1:250-l :500 in 3% blocking buffer. This step can be extended by overnight incubation at 4°C. Following 3 washes, each of 10 minutes duration, in TBS- Tween (TBS with 0.5% Tween-20) the membrane was incubated for 1 hour with an appropriate conjugate at 1:1000 in 3% blocking buffer. The choice of conjugate was dictated by the source of the primary antibody, i.e. peroxidase conjugated swine immunoglobulins to rabbit immunoglobulins or peroxidase conjugated affinity-isolated goat immunoglobulins to mouse immunoglobulins (Dako). After 5 washes in TBS-Tween and a final wash in TBS, membranes were developed using a freshly prepared colorimetric visualisation solution. This solution was prepared by dissolving 4mg/ml di-aminobenzidene (DAB), 8mg/ml NiClj and 1 5pl 30%v/v H 2O2, in 10ml of 015M citrate-phosphate buffer (0 IM citric acid, 0-2M Na2HP 04 titrated to pH5*0). Blots were agitated for 1-5 minutes at room temperature, washed extensively in distilled water and

54 air-dried.

2.7 Induction of DNA damage in mycobacteria

2.7.1 Exposure to ofloxacin and mitomycin C

The DNA-damaging agents used in this study were: i) A 1 mg/ml solution of mitomycin C (Sigma) dissolved in sterile distilled water and used at a final concentration of 2 5pg/ml and, ii) a 1 mg/ml solution of ofloxacin (Sigma) dissolved in sterile 10% ethanol and used at a final concentration of 1 |Lig/ml. All solutions were prepared immediately prior to use and kept on ice. Mycobacterial cultures grown as described in section 2.1 to the specified for each investigation. Following treatment with a DNA damaging agent 50 - 100ml aliquots were removed at designated time intervals and cell-free extracts prepared (section 2.5).

2.7.2 Ultra-violet irradiation of mycobacterial cultures

Mycobacterial cultures were grown as described (section 2.1) to an ODgqonm of 0 6 for M. smegmatis and M. tuberculosis and 0 35 for M microti. Bacteria were harvested by centrifugation (2,500 rpm in a Sorvall GSA rotor for 10 minutes) and resuspended in sterile lOmM MgSO^. In a darkened room 1ml aliquots in a sterile Petri dish (total area 9 6cm^) were exposed to ultra-violet irradiation (254nm, 0 094Jm”V , Jencons) for 0-15 minutes. After a 10”^-10"^-fold dilution in sterile lOmM MgSO^ lOOpl volumes were spread on 7H11 agar (appendix I). Plates were foil wrapped to exclude light and incubated at 37°C for: M. smegmatis 3-4 days, M. tuberculosis 10-15 days andM microti 5-6 weeks. After this time the number of visible colony forming units (CPUs) was determined for each UV dose.

55 2.7.3 Determination of the minimum inhibitory concentration (MIC)

Stock solutions of ofloxacin (lOOpg/ml) and mitomycin C (lOOpg/ml) were prepared using the appropriate solvent (section 2.7.1). Doubling dilutions of ofloxacin (0.04-10pg/ml) or mitomycin C (0.00315-50pg/ml) were made in Sterilins containing 10 ml of sterile media. All tubes were infected with approximately 10^ colony-forming units (CPUs) of M. tuberculosis, M microti or M smegmatis and incubated at 37°C for 10 days, 5 days or 24 hours respectively. The turbidity of each aliquot was visually assessed by comparison with a control culture grown in the absence of DNA damaging agent using the following scale; (-) no growth, (+) poor growth, (++) reasonable growth or (+++) excellent growth.

2.8 Isolation of mycobacterial nucleic acids

Mycobacteria grown to mid/late log phase (appendix II) were harvested by centrifugation (10,000rpm in a Sorvall GSA rotor for 10 minutes) and washed in 15 - 30ml SET buffer (0-3M sucrose, lOmM EDTA, 50mM Tris-Cl, pH8 0). After centrifugation (10,000rpm in a Sorvall GSA rotor for 10 minutes) the pellet was resuspended in SET (1ml for 0 5g-wet weight cells, 2ml for 0.5 - Ig) containing 2mg/ml each of lipase and lysozyme, and then incubated at 37°C for Ihour. Four volumes of GSE buffer (6M guanidium chloride, 1% sarkosyl, 20mM EDTA) were added and incubation continued for a further 2 hours. After a single chloroform extraction, DNA was precipitated overnight at -20°C with 2.5 volumes of absolute ethanol. Following centrifugation (10,000rpm in a Sorvall SS34 rotor for 30 minutes) the precipitate was resuspended m 400pl TE (25mM Tris, lOmM EDTA, pH 8 0) containing 2000U RNAse, 20pl proteinase K (lOmg/ml) and 0 5% SDS, then left at 37°C until dissolved. Following a single phenol/chloroform extraction the suspension was ethanol precipitated overnight at -20°C and the resultant precipitate dissolved m a minimum volume of TE.

2.9 Cloning

2.9.1 Small scale plasmid preparations

Minipreps were prepared by the boiling method (Holmes and Quigley, 1981) or using a QIAGEN miniprep kit according to the manufacturer’s instructions.

2.9.2 Restriction enzyme digestion and alkaline phosphatase treatment

Restriction enzyme digestion of DNA was carried out using a ten-fold excess of the enzyme, in the buffer specified by the manufacturer (New England Biolabs), for 2-3 hours at 37°C. Vector DNA for subcloning was then incubated for 30 minutes with alkaline phosphatase (Boehringer Mannheim) to remove terminal phosphate groups thereby preventing vector re-ligation in the absence of an insert (Sambrook et al, 1989).

56 2.9.3 Agarose gel electrophoresis

Electrophoresis in agarose gels was used to separate, identify and purify DNA fragments according to the methods described in Sambrook et al (1989). Various concentrations of agarose were used depending on the size of the DNA fragments to be resolved; gel thickness was 4 - 8mm. Ethidium bromide was included in the gel at a final concentration of 0 5pg/ml to visuahse DNA fragments. Prior to electrophoresis 1/5 volume of loading buffer (0.25% Bromophenol blue, 0.25% xylene cyanol FF, 30% v/v glycerol) was added to each sample. Electrophoresis was carried out in a Tris-acetate buffer system (TAE Ix solution: 40mM Tris- acetate, ImM EDTA at pH7-7) at 80 - lOOV. Resolved DNA fragments were visuahsed on an ultraviolet transilluminator (300nm), photographed using a Foto/Analyst camera (Fotodyne) and printed from a videocopy processor (Mitsubishi).

2.9.4 Purification of DNA fragments by elution from an agarose gel slice

DNA was extracted from an excised agarose gel slice using either the glass milk binding method (BIO 101 Geneclean II kit) or Centriprep-10 columns (Amicon) according to the manufacturer’s instmctions.

2.9.5 Ligations

Ligations were carried out using a Rapid DNA ligation kit (Boehringer Mannheim) according to the manufacturer’s instmctions. Reactions were set up in a total volume of 21pl containing 50-100ng of vector DNA, 100-200ng of msert DNA, lOpl hgation buffer (2x), 5U of T4 DNA ligase and DNA dilution buffer to the required volume.

2.9.6 Transformation of Escherichia coli

Competent E. coli cells were prepared using an adaptation of the method described by Cohen et al. (1972). A single colony of E. coli (maintained on minimal media for strains carrying an F') was inoculated into 5ml of L-broth (appendix I) and grown overnight at 37°C. Bacteria were subcultured at 1/100 into 50ml warm L- broth and incubated in a rotary shaker (250rpm) at 37°C. At an OD^oonm of 0.3-0.4 cells were chilled on ice for 10 minutes prior to harvesting by centrifugation (2,500rpm in a Sorvall HS4 rotor for 10 minutes at 4°C). Bacteria were resuspended in 20ml ice cold 0 IM CaClj and incubated on ice for 25 minutes. Following a final centrifugation (2,500rpm in a Sorvall HS4 rotor for 10 minutes at 4°C) cells were gently resuspended in 2ml chilled 0 IM CaClj and left on ice until required. Between 10 - lOOng of DNA was added to a lOOpl aliquot of competent cells, left on ice for 30 minutes then heat shocked at 42°C for 90 seconds. The volume was increased to 1ml with L-broth containing glucose at a final concentration of lOmM and the bacteria incubated at 37°C for 1 hour. After a 20-second pulse in a microfuge 900pl of media was decanted, the bacteria resuspended in the remaining L-broth then plated on selective media.

57 When using pUC19 or pBluescriptll KS as the cloning vector with E. coli host DH5a, plates contained 50|ng/ml ampicillin and methicillin, 0-2M isopropyl-P-D-thiogalactopyranoside (IPTG) and 40fxg/ml 4- bromo-3-chloro-2-indolyl-beta-galactosidase (X-gal). This combination permits blue-white screening: colonies with recombinant plasmids are white and parental vector transformants blue. For other vectors selection was based on resistance to 50|ig/ml ampicillin and methicillin or 25 - 50ug/ml kanamycin and the presence of recombinant plasmids confirmed by restriction enzyme digests of isolated plasmid DNA.

2.9.7 Transformation of mycobacteria by electroporation

An overnight culture ofM smegmatis mc^l55 grown as described in section 2.1 was used to inoculate Dubos media (appendix I) at 1/250. After incubation at 37°C for 18hr (ODgoonm of 1.0) the culture was left on ice for 1 hour prior to harvesting by centrifugation (10,000rpm in a Sorvall GSA rotor for 10 minutes). Cells were resuspended in 10% glycerol at 1/3 of the culture volume, incubated on ice for 30 minutes, centrifuged (10,000rpm in a Sorvall GSA rotor for 10 minutes) then resuspended in 10% glycerol at 1/10 of the original culture volume. The cells were left on ice for at least 2hours prior to electroporation. Approximately Ipg of DNA and 400pl competent mycobacteria were added to a precooled 0 2cm cuvette and subject to electro- transformation (25pF, 2.5kV, lOOOQ). The reaction volume was increased to 5ml with prewarmed Dubos media, incubated at 37°C for 3 hrs and lOOpl aliquots plated out on 7H11 agar containing the appropriate antibiotic.

2.10 Southern hybridisation

2.10.1 Probe labelling

DNA was labelled by random primer extension with an ECL DNA labelling kit (Amersham) following the instmctions provided by the manufacturer.

2.10.2 DNA transfer and hybridisation

Southern hybridisation was performed using an adaptation of the method described by Sambrook et al (1989). DNA was separated by overnight electrophoresis through an agarose gel at 25V. After dénaturation in an alkali buffer (0-5M NaOFl and 1-5M NaCl) for 40 minutes the gel was incubated for 2x30 minutes in neutrahsation buffer (0 5M Tris-Cl, 1 5M NaCl at pH7 5). DNA was transferred to Hybond N+ membrane (Amersham) by overnight capillary elution in 20 x SSC (3M NaCl, 0-3M tri-sodium citrate and pH7-7) according to the manufacturer’s instmctions. Membranes were fixed by a 1 minute exposure to UV hght (254nm) then pre-hybridised at 60°C for 1 - 4 hours in hybridisation mix (5 x SSC, 0 5% Amersham blocking reagent, 01% SDS, 5% dextran sulphate, lOOmg/ml herring sperm DNA). After addition of a labelled probe, hybridisation proceeded overnight at 60°C. Filters were washed twice (2 x SSC, 0 1% SDS) at 60°C for a duration of 30 minutes. ECL development of filters was carried out according to the

58 manufacturer’s instructions (Amersham).

2.11 Induction of RecA-CAT fusions inM, smegmatis mc^l55

Transformants of M. smegmatis mc^l55 harbouring one of the following plasmids pEJ257, pFM7, pFM17, pNT2 or pNT12 were grown with constant agitation in Lemco broth (appendix I) containing 25pg/ml kanamycin for 36 hrs at 37°C. This culture was used at 1/500 to inoculate warm Dubos media (appendix I) containing 25pg/ml kanamycin which was then returned to the 37°C incubator for approximately 24 hours. When the ODgoonm h&d reached 0 6-0 7 each culture was split and half used as an uninduced control, the other half was treated with a DNA damage-inducing agent as described in section 2.7.1. All cultures were returned to the incubator for 5 hours after which time cell-free extracts were produced according to the protocol given in section 2.5, omitting the Pefabloc in this instance. Protein estimations were made using the BCA micro-titre plate protocol (Pierce) according to the manufacturer’s instructions. Levels of CAT expression were determined using a CAT-ELISA enzyme immuno-assay kit (Boehringer-Mannheim) according to the manufacturer’s instructions.

2.12 Sequencing

Sequencing samples were prepared using an ABI Prism™ Dye terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer) according to the manufacturer’s instmctions. After PCR and ethanol precipitation, both according to the manufacturer’s instmction, pellets were resuspended in 5|u.l loading buffer (a ratio of 5:1 deionized formamide to 25mM EDTA with 50mg/ml Blue dextran) and applied to a prepared ABI Prism 377 acrylamide gel. Sequence traces were analysed using ABI Prism™ DNA sequencing software (version 2.1.1) and a consensus sequence compiled using DNA STAR Seqman.

59 2.13 Gel Shift Assay

2.13.1 Precipitation of oligonucleotides

Oligonucleotides were supplied ready to use in distilled water (OSWELS) or in ammoitia solution (In-house synthesis, NIMR) which required precipitation prior to use. A 360pl volume of crude oUgonucleotide was precipitated overnight at -20°C using 40pl 3M sodium acetate and SOOpl 100% ethanol. After centrifugation in a microfuge for 15 minutes the pellet was washed in 80% ethanol, air-dried and resuspended in 105 pi sterile distilled water. The DNA concentration of a 1/100 dilution was determined by measuring the ratio of optical density at 260/280nm and applying the following formulae:

(OD260/280 X dilution) / e of oligonucleotide = concentration (pmole/ml)

Where e (extinction coefficient) for each base is: dGTP =11 7ml/pmole dCTP = 7 3ml/pmole dATP = 15 4ml/pmole dTTP = 8-8ml/pmole

2.13.2 Gel shift assay

Gel shift assays were carried out using a DIG Gel shift kit (Boehringer-Mannheim). Complementary, single- stranded oligonucleotides were annealed by denaturing together at 95°C for 10 minutes then slowly cooling to room temperature. Double-stranded oligonucleotides were diluted to 3 - 4pmole/pl with TEN buffer (lOmM Tris-HCl, ImM EDTA, 0 IM NaCl and pH8 0) and labelled with digoxigenin-11-ddUTP according to the manufacturer’s instructions. The labelled oligonucleotide was diluted to 15 - 30finole/pl in TEN buffer immediately prior to use and the gel shift reaction set up according to the manufacturer’s instructions. After incubation at room temperature for 15 minutes, the samples were immediately applied to a pre- electrophoresed native 6 - 8% polyacrylamide gel (14 x 12cm, appendix III). Electrophoresis (180V at 4°C) in 0 5 X TEE buffer (IM Tris base, IM Boric acid, 0 05M EDTA and pH8 0) was terminated when Bromophenol blue loaded in a separate track had travelled 3/4 of the gel length.

Transfer of DNA to a positively charged BM nylon membrane (Boehringer-Mannheim), pre-soaked in 0 25

X TBE, was carried out at 4®C using a semi-dry electroblotter (Ancos) at 400mA for 1 hour. The membrane was soaked in 10 x SSC for 30 seconds and cross-linked for 3 minutes at 254nm on a transilluminator. Chemiluminescent detection was performed according to the manufacturer’s instructions. The membrane was then exposed to X-ray film (Hyper film-MP, Amersham) for 30 - 60 minutes at room temperature.

60 3.0 Pr o d u c tio n and purification o f po ly c l o n a l a n tise r a ra ised

AGAINSTM TUBERCULOSIS^CX

3.1 Introduction

The RecA protein of E. coli has both a regulatory and mechanistic role. It is a key component of general homologous recombination and plays a central role in post-replication repair, mutagenesis and expression of the proteins involved in the SOS response (Roca and Cox, 1990; Miller and Kokjohn, 1990; Cox, 1993; Kowalczykowski et al, 1994). Analogous recA genes have been isolated from more than sixty prokaryotic species (Karlin et al, 1995). A striking degree of similarity exists between them although it appears that their mode of regulation may vary (Brent and Ptashne, 1981; Cheo et al, 1991; 1993; Davis et al, 1991; 1994; Duwat et al, 1992; Bayles et al, 1994; Nu^baumer and Wohlleben, 1994; Pearce et al, 1995; Stamm et al, 1995; Aigle et al, 1997; Johnston et al, 1997; Papavinasasundaram et al, 1997; Tapias et al, 1997; Fernandez de Henestrosa et al, 1997). The importance of the recA genotype to microbial survival and its role in virulence has already been demonstrated in prokaryotes such as E. coli, Erwinia carotovora and Salmonella typhimurium (Carlsson and Carpenter, 1980; McEvoy et al, 1990; Buchmeier et al, 1995). It is envisaged that an analogous requirement for DNA repair may also be important for the survival of pathogenic mycobacteria within the intracellular environment. As a consequence there is great interest in characterising the response of mycobacteria to DNA damage as well as determining its impact on pathogenicity. The identification of a recA-Wks gene in M tuberculosis (Davis et al, 1991) provided the ideal opportunity to assess the consequences of genetic insult in this genus.

To enable the specific aims of this project to be fulfilled a rapid, sensitive method for examining the consequences of genetic insult on recA expression was required. Expression of this gene can be assessed indirectly using an operon fusion to the putative promoter region (Kenyon and Walker, 1980) or directly, by determining the fate of the gene product using a standard Western blotting procedure. The work in this chapter describes the development and optimisation of an assay to determine the effect of genetic insult on RecA protein levels by Western blot. The basic principle behind this method is the exploitation of the specificity inherent in antigen-antibody recognition. A pre-requisite for this technique is an antibody raised against RecA protein. To understand why an antibody raised against M. tuberculosis RecA was preferred over its E. coli counterpart a brief explanation of some basic immunological terms is required.

All antibodies have a common core structure of four polypeptide chains comprising of two identical light chains (each about 24kDa) and two identical heavy chains (55 - 70kDa). One light chain is attached to each heavy chain and the two heavy chains are attached to each other. The antigen-binding site is located at the N-terminal of this structure. There are three separate genetic loci encoding these polypeptide chains, the heavy chain locus and those for the k and X light chains. This engenders a remarkable capacity for structural diversity and accounts for the extraordinary specificity of antibodies, because each

61 amino acid difference may produce a difference in antigen binding. Each antibody is potentially multivalent since it possesses two antigen-binding sites (Abbas et al, 1994). The strength of a single antigen-antibody interaction is defined as its affinity. Measurements of affinity relate to equilibrium conditions and indicate the tendency of antibodies to form stable complexes with the antigen. The strength with which an antibody binds antigen is termed avidity (Roitt et al, 1993). The avidity of an antibody for antigen is dependent on the affinities of the individual binding sites for the antigenic determinants; it is greater than the sum of these affinities if both antigen-binding sites combine with the antigen. In addition, antigen-antibody reactions show a high level of specificity. The specificity of an antiserum is a summation of the interactions of the various antibodies in the population, each reacting with a different antigenic determinant. When a different antigen shares some of the determinants of the original antigen then a proportion of the antibodies will cross-react (Roitt et al, 1993).

The M. tuberculosis RecA protein is highly conserved compared with the RecA proteins of M leprae (92% identity; Davis et al, 1994) andM smegmatis (91% identity; Papavinasasundaram et al, 1997). It is also related, although at a lower level of homology, to its E. coli counterpart (59% identity; Davis et al, 1991). A specific aim of this project is to qualitatively compare basal and induced levels of RecA in different species of mycobacteria. To maximise the potential of the assay it is important to ensure that the antiserum demonstrates a high avidity for the protein under investigation, thereby providing a high degree of specificity and sensitivity. Antisera raised against M. tuberculosis RecA is expected to demonstrate these desirable properties and be preferable to antisera raised against E. coli RecA because of the higher level of homology between mycobacterial RecA proteins compared with that of E. coli.

This chapter describes: i) The production and purification of recombinant M. tuberculosis RecA protein ii) The production of a polyclonal antisera raised against recombinant M. tuberculosis RecA iii) The purification of M. tuberculosis RecA antisera by affinity chromatography using M. leprae RecA precursor protein covalently linked to cyanogen bromide (CNBr)-activated Sepharose 4B

3.2 The production of polyclonal antisera raised against M, tuberculosis RecA protein

3.2.1 Production and purification of recombinant Mtuberculosis RecA protein

Construct pEJ169 is an E co^-B cH deletion of upstream sequences from plasmid pEJ135 (Davis et al, 1991), which contains the whole of the M tuberculosis recA locus expressed from the T7 promoter, and pGPl-2 is a temperature sensitive plasmid which codes for T7 RNA polymerase. An E. coli recA mutant harbouring both plasmids {E. coli TG2 (pEJ169[pGP 1 -2]} ; kindly provided by Dr. E. O. Davis, N.I.M.R.)

62 was grown as described (section 2.2.1) and a cell-fee extract prepared. A semi-pure preparation of recombinant M. tuberculosis RecA was produced by differential protein precipitation (section 2.2.2); the progress of this procedure was monitored at each stage (figure 3.1). Pure recombinant M tuberculosis RecA was prepared by separating this preparation on a 10% SDS-polyacrylamide gel, locating the RecA band by reversible copper staining and excising it from the gel. Dr P Jenner has positively identified this band as RecA (N.I.M.R., unpublished) in previous investigations. The gel slice was transferred to visking tubing and the protein eluted into a Tris based buffer (section 2.2.3). The volume of the electro-eluted sample was reduced to approximately SGOpl and the purified protein estimated to be >90% pure by SDS- polyacrylamide gel electrophoresis (figure 3.2). This sample was stored in lOOpl aliquots at -20°C until required.

3.2.2 The immunisation protocol and subsequent monitoring procedures leading to the production ofM. tuberculosis RecA antisera

Three rabbits were immunised with purified M. tuberculosis RecA protein as described in section 2.3. The ability of the RecA protein to stimulate a humoral immune response in this host is unknown. Therefore, the following procedures were incorporated into the immunisation protocol to maximise its potential: i) Emulsification of the antigen with incomplete Freuds adjuvant (a pharmaceutical grade white mineral oil and an emulsifier) to enhance its antigenicity. ii) The administration of three inoculations over a six-week period. This extends the time of exposure to the antigen and results in a hyperimmunised rabbit whose serum contains vastly increased levels of IgG. iii) Subcutaneous injection of the antigen into several sites. This improves distribution of the antigen to the lymph nodes via the lymphatic system.

The progress of antibody levels in rabbit serum was monitored at days 28 and 49 using a standard ELISA protocol (Current Protocols in Immunology: Unit 2.1, Basic protocol). Antisera dilutions ranging 1/2 - 1/20,000 were assayed on micro-titre plates coated with 50ng of semi-pure recombinant M. tuberculosis RecA protein. A positive signal was consistently achieved at a 1/500 dilution; there was no significant difference between successive test bleeds. A final booster immunisation was administered and exsanguination followed 7 days later. Serum was prepared from whole blood.

The antigen titre and specificity of the antisera was checked by Western blot. Semi-pure M. tuberculosis RecA protein (3mg/ml) subject to ten-fold dilutions was separated on a 10% denaturing SDS- polyacrylamide gel and transferred to PVDF membrane (section 2.6). After probing with a 1/500 dilution of antisera a strong band of the expected molecular weight (38kDa) was visualised in samples containing as little as 20ng of total protein (figure 3.3). A significant amount of cross-reactivity with other proteins was evident indicating the necessity for purification of the antisera, which was stored at -20°C.

63 a)

Over-expression M.of tuberculosis RecA by heat induction of E. co//TG2(pEJ169{pGP1-2})(lane 1)

Cell-free extract prepared by sonication (lane2)

b)

Proteins precipitated with polyethylenimine(0 05% f.c.) (lane 3)

Protein precipitate resuspended in R buffer with (NH4)2S0 4 (150mMfc). After incubation on ice RecA the supernatant is retained, the precipitate resuspended in R buffer and (NH^)2SO^ precipitation repeated 12 3 4

Supernatants retained from the (NH4 )g SO4 precipitation steps are combined and incubated over-night with 20% w/v (NH4)2804 at 4°C

Protein precipitate harvested and resuspended in a minimum volume of R buffer (lane 4)

Figure 3.1 A schematic illustration of the purification of recombinant Af. tuberculosis RecA protein

The M. tuberculosis RecA protein was over-expressed by heat induction ofE. coli TG2(pEJ169{pGPl-2}) and a cell-free extract prepared. A semi-pure preparation of M. tuberculosis RecA was produced by differential protein precipitation (a); the progress of which was monitored at each stage by the removal of samples and their analysis on a 10% denaturing SDS-polyacrylamide gel (6): lane I) lysed bacterial cells prior to sonication, lane 2) cell-free extract of E. coli TG2(pEJ169{pGPl-2)), lane 3) polyethylenimine precipitate, lane 4) semi-pure M. tuberculosis RecA protein.

64 101,000 83,000 50,600 RecA 35,500 29,100

Figure 3.2 SDS-PAGE analysis of semi-pure and pure recombinant M. tuberculosis RecA £. coh TG2 (pEJ169{pGPl-2i) was subject to heat induction and a cell-free extract prepared. A semi-pure preparation of recombinant M. tuberculosis RecA (lane /) was prepared by differential protein precipitation. This preparation was further purified by excision of the RecA band from a 10% SDS-polyacrylamide gel ( 14 x 12cm) followed by electro­ elution into a Tris based buffer. The sample was then concentrated while undergoing buffer exchange with PBS to approximately 1/10 of its original volume (lanel). Molecular w^eights are given in kDa.

05 1 2 3 4 5 6 7 101,000 83,000

50,600 RecA 35 ,5 0 0 29 ,1 0 0

2 0 ,9 0 0

Figure 3.3 Determination of the antigen titre of polyclonal antisera raised against M. tuberculosis RecA Ten-fold dilutions of semi-pure Mtuberculosis RecA (3mg/ml) were analyzed by Western blot with polyclonal M. tuberculosis RccA antisera (diluted 1/500). The total protein content in each sample was estimated as lane 7) 21 pg, lane 2) 2 1 pg, lane i) 210ng, lane -7) 21ng, lane 5) 2-lng, lane 6) 210pg and lane 7) 21pg. A strong band of the expected molecular weight (38kDa) was detected in samples containing as little as 2 Ing of total protein. Molecular weights are given in kDa.

66 3.3 Purification ofM. tuberculosis RecA antisera by affinity chromatography using the M. leprae RecA precursor protein covalently linked to a suitable immunoadsorbent

3.3.1 Investigations to identify an appropriate coupling gel

The basic principle of affinity chromatography relies on the successful covalent linkage of a ligand to an insoluble matrix together with the exploitation of the specificity inherent in antigen-antibody interactions. After application of the substance to be purified to a prepared column, the buffer conditions are altered to maximise dissociation firom the immobilised ligand by manipulating the pH and salt concentration (Dean et al, 1985). A successful separation requires that the ligand be capable of covalent attachment to the adsorbent while maintaining its specific binding affinity for the substance to be purified. As a consequence, the groups available in the ligand for coupling to the adsorbent and the nature of its interaction with the sample to be purified dictate the choice of coupling agent. No specific information was available for the M. leprae RecA precursor protein so a trial and error approach was adopted.

Sepharose 4B is a widely used matrix that is stable in the presence of detergents and dissociating agents as well as under conditions of high or low pH. It is a bead formed agarose gel with an open pore stmcture, which makes the interior of the matrix available for ligand attachment thereby ensuring a good binding capacity. This adsorbent exhibits extremely low non-specific interactions, which is an essential characteristic since affinity chromatography relies on specific interactions. Preliminary investigations focused on the selective, covalent attachment of thiol groups to an activated thiol matrix, primarily because the presence of 6M-guanidium chloride in the buffer of the M. leprae RecA precursor protein (see below) does not interfere with the coupling reaction (Pharmacia, 1986/8). An adsorbent that has been specifically developed for this technique is activated Thiol-Sepharose 4B (Sepharose-glutathione-2- pyridyl disulphide) which comprises a mixed disulphide bond formed between 2,2’-dipyridyl disulphide and glutathione coupled to CNBr activated Sepharose 4B; its effectiveness is based upon a reversible reaction with substances containing reduced sulfhydryl groups. Analysis of the M. leprae RecA precursor protein, using Ellmans reagent and established spectrophotometric procedures (section 2.4.2; Ellman, 1959; Riddles et al, 1983), to determine its suitability for attachment to activated Thiol-Sepharose 4B revealed that it may be an unsuitable choice of adsorbent. The above spectrophotometric technique is based upon the change in following the addition of protein to a fresh solution of di- thionitrobenzidine (DTNB) in 6M-guanidium-HCl (pH7.2); a rapid rise in absorbance is indicative of the presence of free sulfhydryl groups. However, monitoring the OD^iznm following the addition ofM leprae RecA precursor protein revealed that the optical density rose only very slowly which suggests that very few free sulfhydryl groups are present in this protein.

An alternative adsorbent, CNBr-activated Sepharose 4B, is a reactive derivative of Sepharose 4B which efficiently couples proteins under mild conditions (pH 8-10) via their primary amino or similar

67 nucleophillic groups. The reaction of cyanogen bromide with the hydroxyl groups on Sepharose results in the formation of imidocarbonate groups (-CH=NH) which react with nucleophiles to form isourea linkages (-{C=NH}-NHR where R represents the amino group). An advantage of this adsorbent is its capacity for the multi-point attachment of proteins thereby increasing stability of the coupled matrix by decreasing the amount of protein lost via hydrolysis from the adsorbent. CNBr-activated Sepharose 4B is a widely used matrix for purifying specific antibodies (Pharmacia, 1986/8). Therefore it was anticipated that it would be an ideal candidate for purification of the M tuberculosis RecA antibody. The only necessary pre-requisite was the removal of 6M-guanidium chloride from the buffer in which the M leprae RecA precursor protein was dissolved since this chemical may interfere with the coupling reaction by covalently linking Sepharose (Pharmacia, 1986/8).

3.3.2 Production ofM. leprae RecA precursor protein

The selection of a ligand for affinity chromatography is influenced by two factors. Firstly, it should exhibit a specific and reversible binding affinity for the substance to be purified. In addition it should have chemically modifiable groups which enable attachment to the matrix without destroying this binding affinity. Ideally the affinity of the immobilised ligand for binding to the substance to be purified should be in the range lO"'^- lO'^M in free solution (Pharmacia, 1986/8). No information was available regarding the binding affinity of the M. tuberculosis RecA antibody for its protein. However, it seems reasonable to assume that the dissociation constant for the protein against which it was raised will be low, therefore, dissociation of the bound antibody is likely to require strongly denaturing conditions. The M. leprae RecA precursor protein was considered to be a viable alternative since amino acid identity between the M. tuberculosis and M. leprae RecA precursor proteins is 92% for the RecA sequence and 27% for the intein (Davis et al, 1994). Hence, it was anticipated that while the M. leprae RecA precursor protein possesses a sufficiently high degree of homology with its counterpart in M tuberculosis to enable specific, reversible binding to the antibody it would prove easier to dissociate the antigen-antibody complex. Furthermore, the availability of a simple, rapid method for its production made using the M. leprae RecA precursor protein an attractive alternative.

Escherichia coli harbouring plasmid pK233 (kindly supplied by Dr D Emery, NIMR) was induced using isopropyl-|3-D -thiogalactopyranoside (IPTG) as described in section 2.4.1. The M. leprae RecA precursor protein was produced as an insoluble precipitate in inclusion bodies and located in the cell debris following lysis. Following solubilisation of this protein in a guanidium-HCl based buffer it was precipitated with 1/9 volume of trichloro-acetic acid. After two washes in coupling buffer (section 2.4.3) the precipitate was thoroughly dissolved in 1ml dimethyl sulphoxide (DMSO), the volume increased to 5ml using coupling buffer and a sample analysed on a 10% SDS-polyacrylamide gel (figure 3.4a). This buffer exchange is essential because the presence of 6M-guanidium chloride may interfere with the coupling reaction.

68 a) b) Freeze dried CNBr-activated Sepharose 48 rehydrated with 1 mM HCl M. leprae RecA precursor protein dissolved in DMSO/ coupling buffer 101,000 83,000 50,600 - M. leprae RecA lOmg protein solution and 0.75g gel suspension mixed over-night at 4°C 35,500

29,100 Active sites blocked with 1 ethanolamine

2 5ml gel suspension transferred to diposable chromatography column

i Column washed alternately with acetate buffer (10ml) then coupling buffer (10ml). Repeat

Column washed twice with coupling buffer

Apply sample

Figure 3.4 A schematic illustration of the coupling of M. leprae RecA precursor protein to CNBr-activated Sepharose 4B

A preparation of M. leprae RecA precursor protein in DMSO/carbonate buffer was analysed on a 10% denaturing SDS-polyacrylamide gel (a). Coupling of lOmg o f this protein to 0-75g of re-hydrated CNBr-activated Sepharose 4B was performed in the presence of carbonate buffer and active sites blocked with ethanolamine. The coupling reaction and successive washes to remove unbound ligand and excess blocking reagent are illustrated in (6). Molecular weights are given in kDa.

69 3.3.3 Preparation of an affinity chromatography column and subsequent purification of polyclonal antisera

A schematic illustration of the coupling of M. leprae RecA precursor protein to CNBr-activated Sepharose 4B is shown in figure 3.4b. The prepared column was washed with PBS and antisera applied at 0 5ml / minute. Following removal of unbound antisera the antigen-antibody complex was dissociated by the application of 0 IM glycine-HCl (pH2.5), this change in pH alters the degree of ionisation at the site of interaction between the antigen-antibody complex (Hudson and Hay, 1976). Further desorption was facilitated by reducing the polarity of the dissociation buffer with dioxane since antigen-antibody interactions are hydrophobic (Anderson et al, 1979). The eluate was immediately titrated to pH8.5 with Tris base to prevent dénaturation then concentrated to 1-5 - 2-5ml using a centricon-10 concentrator (Amicon) according to the manufacturer’s instmctions. Aliquots of lOOpl of affinity purified serum were stored at -20°C until required.

The antigen titre of the affinity-purified antisera was determined as described previously (figure 3.3). There was no significant alteration in titre; a strong band of the expected molecular weight was visualised in samples containing as little as 50ng of total protein (data not shown). In addition, semi-pure M. tuberculosis RecA protein (5fig total protein) was separated on a 10% SDS-polyacrylamide gel, transferred to PVDF and probed using a 1/500 dilution of “raw” or affinity-purified antisera. Examination of the membrane revealed that the cross-reactivity of the antisera has been significantly reduced (figure 3.5) thereby confirming that the purification by affinity chromatography of polyclonal antisera raised against M. tuberculosis RecA was successful.

3.4 Discussion

This chapter describes the successful production and affinity purification of an antibody raised against M. tuberculosis RecA protein. Examination of the antigen titre of this affinity purified antibody revealed that as little as 50ng of semi-pure recombinant M. tuberculosis RecA is required to give a strong clear signal on a Western Blot. In addition, recombinant M tuberculosis RecA will provide a sensitive, positive control that will enable RecA protein in mycobacterial cell-free extracts to be rapidly identified.

An alternative method for obtaining pure antibody that exhibits the desired specificity is to produce monoclonal antibodies from cells in culture. Monoclonal antibody production is based upon the method devised by Milstein and Kohler (1975) and involves the fusion of mouse splenocytes with a homogenous, non-secreting B-cell myeloma line. The creation of an immortal clone that continually manufactures a single antibody of the desired specificity obviates the requirement for continual polyclonal antiserum production and the concomitant purification. A major advantage of working with monoclonal antibodies is the extreme confidence with which claims can be made about such work. However, their production is both labour and time intensive. Furthermore, monoclonal antibodies do not have a greater overall

70 101,000 83,000

50,600

RecA 35,500 29,100

Figure 3.5 A comparison of the specificity of M. tuberculosis RecA antisera prior and subsequent to affinity purification Scnii-purc M. tuberculosis RccA protein (5pg) was analyzed by Western blot using: lane i) M. tuberculosis RecA antisera, prior to purification and lane 2) affinity-purified M. tuberculosis RecA antisera. Purification by affinity chromatography has significantly reduced the amount of eross-reactivity. Molecular weights are given in kDa.

71 specificity than polyclonal antibodies, which recognise the antigen via a number of different epitopes (Roitt et al, 1993). The fact that monoclonal antibodies recognise only a single molecular epitope can be a disadvantage. Monoclonal antibodies have been raised against E. coli RecA. They have been shown to be useful tools for inhibiting various activities of RecA to elucidate its reaction mechanism as well as for locating active sites or functional domains (Kam and Allen, 1982; Krivi et al, 1985; Makino et al, 1985; Ikeda et al, 1990; Shibata et al, 1991). By comparison, polyclonal antibodies are frequently used to detect RecA protein in vivo or in vitro and to identify related proteins (Lovett et al, 1985; Love and Yasbin, 1986; Angulo et al, 1989; Davis et al, 1991). For the purposes of this project there were no distinct advantages to be gained by using monoclonal antibodies, therefore a decision was made to raise and purify polyclonal antisera.

72 4.0 A QUALITATIVE COMPARISON OF BASAL RECA AND LEXA LEVELS IN

SAPROPHYTIC AND PATHOGENIC MYCOBACTERIA

4.1 Introduction

Bacteria such as Legionella pneumophila, Mycobacterium tuberculosis and Salmonella typhimurium are examples of a specific category of prokaryotes that develop and rephcate within the intracellular milieu of host cells where it is presumed they will encounter a powerful array of antimicrobial defences. They use the host cell to create an optimal environment for their growth and as a shield to evade humoral immune defence mechanisms (Ottenhof and Mutis, 1995). It is envisaged that stress responses such as the SOS response to DNA damage play an important role in their survival. The fact that several “stress” proteins are up regulated in S. typhimurium (Buchmeier and Heffron, 1990), L. pneumophila (Kwaik et al, 1993) and M. tuberculosis (Lee and Horwitz, 1995) by intracellular residence does support this. Prokaryotic stress responses have been extensively investigated. They involve the inducible expression of a set of genes that aid the organism in adapting to new environmental conditions. In general a specific response is associated with a particular stress, namely: heat shock (Lindquist, 1986), oxidative damage (Greenberg and Demple, 1989; Farr and Kogoma, 1991), and the SOS response to DNA damage (Little and Mount, 1982; Walker, 1984), although there are genes that respond to multiple stresses.

Heat shock proteins were among the first mycobacterial antigens for which a functional role was identified; examples include the M. tuberculosis 65kDa and 71kDa proteins. Young et al (1991) proposed that in addition to the temperature regulation of their synthesis they were induced during adaptation to survival within host macrophages; an idea that has recently been corroborated by Lee and Horwitz (1995) who reported that 44 M. tuberculosis proteins exhibit a modulated expression within the intracellular environment. These authors demonstrated that many of these alterations also occurred during extracellular growth in response to heat shock, acidification and oxidative damage. Each stress response produced a unique phenotypic profile with proteins being modulated by some stress conditions but not by others. Noteworthy examples include the 65kDa and 71 kDa antigens of M. bovis and M. tuberculosis (identified as GroEL and DnaK respectively) which are induced by heat shock and acidification but unaffected by oxidative damage (Stover et al, 1991; Young et al; 1991; Lee et al, 1995; Sherman et al, 1995). Interestingly in E. coli, heat shock as well as exposure to ultra-violet irradiation or nalidixic acid induced the expression of the GroEL and DnaK genes. This increase in expression resulting from exposure to DNA damaging agents has been proposed to occur by a RecA-independent mechanism (Kreuger and Walker, 1984) which is in direct contrast to the Gram-positive Lactococcus lactis whose recA gene has been implicated in both the oxidative stress and heat shock response (Duwat et al, 1995). It is perhaps noteworthy that a similar link between recA, the heat shock proteins and oxidative damage in mycobacteria has not yet been established.

73 4.1.1 The importance of the recA genotype

Pathogenic microorganisms possess antioxidant defence mechanisms for protection from toxic oxygen metabolites such as hydrogen peroxide, which are generated during the respiratory burst of phagocytic cells. These defence mechanisms include enzymes such as catalase and superoxide dismutase, and DNA repair systems. The response of Gram-negative organisms has been well characterised; pre-treatment of Escherichia coli and Salmonella typhimurium with non-lethal levels of hydrogen peroxide leads to adaptive resistance. This adaptation is accompanied by the modulated expression of over 30 genes of which 9 are under the control of the oxyR gene product (Demple and Halbrook, 1983; Christman et al, 1985). The OxyR protein can exist in one of two states, oxidised or reduced. Storz et al (1990) reported that the oxidised not the reduced form activates transcription of stress inducible genes such as ahpC (the small subunit of alkyl hydroperoxide reductase), katG (catalase/hydroperoxidase I), gorA (glutathione reductase) and the gene coding for superoxide dismutase (Christman et al, 1985).

Exposure of M. smegmatis to hydrogen peroxide elicits an analogous adaptive response, more than a dozen genes exhibit a modulated expression. In contrast, the protective effect of exposure to non-lethal peroxide levels is absent in pathogenic mycobacteria (Sherman et al, 1995) with only a single protein (KatG) in M. tuberculosis complex and three in M. avium (KatG, AhpC, AhpX) exhibiting modest increases in expression. However, pathogenic mycobacteria are more resistant to hydrogen peroxide than optimally pre-treated M. smegmatis, which is indicative of a constitutive defence against oxidative stress. The basis for this difference in the M. tuberculosis complex is the presence of multiple lesions in the oxyR gene, resulting in a non-functional product. It is perhaps noteworthy that M. avium possesses a putatively functional oxyR gene and an inducible ahpC, which suggests that there may be differences in the oxidative stress response among pathogenic mycobacteria (Deretic et al, 1995; Sherman et al, 1995).

The adaptive response of E. coli and S. typhimurium emphasises the importance of enzymes such as catalase and superoxide dismutase (SOD). However, it has been proposed that DNA damage is the predominant lethal action caused by low doses of hydrogen peroxide (<100pM; Imlay and Linn, 1986); which may explain why these enzymes are unnecessary for long (24 hours) or short (2 hours) term survival of Salmonella in macrophages (Buchmeier et al, 1995). Similarly they do not enhance the survival of E. coli in polymorphonuclear neutrophils (Schwartz et al, 1983). These observations corroborate a previous report suggesting that the recA genotype is more important than the catalase genotype to microbial survival (Carlsson and Carpenter, 1980). In addition, Buchmeier et al (1995) reported that S. typhimurium recA mutants demonstrate an attenuated virulence in mice whereas katG mutants do not. Consequently, the contribution of the recA genotype to the survival of pathogenic mycobacteria within the intracellular environment and its possible role in vimlence is of great interest.

The RecA protein is the most thoroughly characterised enzyme of its type. More than sixty sequences are currently available including over thirty Gram-negative and nearly a dozen Gram-positve sequences (Karlin et al, 1995; 1996). In view of the remarkable similarity between recA genes (43 - 100% amino acid

74 identity) it is reasonable to expect a certain degree of evolutionary conservation in the essential control elements of the SOS system among prokaryotes (see section 1.6.4). The fact that E. coli LexA can influence the induction of SOS-like functions in several species of enterobacteriaceae reinforces this view (Sedgwick and Goodwin, 1985). The identification of the M. tuberculosis and M. leprae recA genes (Davis et al, 1991; 1994) fuelled the already growing interest in the possible importance of the SOS response and its regulation in mycobacteria for survival within the intracellular environment.

Despite the phylogenetic conservation in size, stmcture and overall function in nearly 60 recA genes cloned to date (Karlin et al, 1996) the recA gene in members of the M. tuberculosis complex and M. leprae revealed a strikingly unusual structure. The recA locus specifies a functional 38kDa protein from a single open reading frame with the coding potential for an 85kDa or 79kDa product respectively (Davis et al, 1991; 1994). The presence of an “intein” in the same protein of two related species is suggestive of a positive selection for sphcing in the maturation of RecA and raises questions with respect to its possible implications to pathogenicity and virulence (Davis et al, 1994). More recently, the 85kDa-precursor protein has been shown to be totally defective in its DNA-binding, ATP-binding and ATPase activities (Kumar et al, 1996). This raised the intriguing possibility that splicing in the maturation of RecA may regulate the display of the RecA phenotype. To address this question three members of the M tuberculosis complex were screened by Western blot using affinity purified M. tuberculosis RecA antibody for the presence of precursor RecA protein. The possibility that expression of the precursor protein may be conditionally regulated was also investigated.

Homologous recombination has been reported to occur more efficiently in M. smegmatis (Husson et al, 1990; Kalpana et al, 1991; Mustafa, 1995; Gordhan et al, 1996; Pelicic et al, 1996) than M. tuberculosis (Kalpana et al, 1991) and M. bovis BCG (Aldovini et al, 1993) where the predominant event is illegitimate recombination. One possible explanation for the low levels of homologous recombination found in these two members of the M. tuberculosis complex is that an insufficient quantity of RecA is produced. To address this question the basal levels of RecA in M. smegmatis and three members of the M tuberculosis complex were compared by Western blot using affinity purified M. tuberculosis RecA antibody. In addition, a range of both pathogenic and saprophytic mycobacteria were examined to determine whether there was a common distinction between these two groups in their basal RecA levels. The presence of a recA-like gene in M. tuberculosis was the first evidence to suggest that an inducible response to DNA damage, analogous to the SOS response, may exist in this genus. More recently, a LexA homologue has been identified in M tuberculosis (Movahedzadeh, 1996) which suggests that the essential control elements of the SOS response may have been conserved. To provide further corroboratory evidence M. smegmatis and three members of the M. tuberculosis complex were examined by Western blot using a polyclonal M. tuberculosis LexA antibody to monitor basal LexA levels with a view to assessing the effect of treatment with DNA damaging agents in subsequent investigations.

75 4.2 Confirmation that affinity purified M, tuberculosis RecA antibody will recognise RecA precursor protein

To confirm that affinity purified M. tuberculosis RecA antibody will detect RecA precursor protein (i.e. unspliced protein) a cell-free extract was prepared from a culture ofE. coli TG2 (pEJlTO) grown as described (section 2.1) in the presence of methicillin and ampicillin (each at 50pg/ml f.c.). Construct pEJlTO is a double frameshift mutant of pEJ135 (Davis et al, 1991; 1992) producing RecA precursor protein (85kDa) and a small amount of processed RecA protein (38kDa). Ten-fold dilutions of cell-free extract (O OOl-lOOpg total protein) were separated on a 10% SDS-polyacrylamide gel, transferred to PVDF as described (section 2.6) and the membrane developed using affinity purified M. tuberculosis RecA antibody diluted 1/500. Two major bands of predicted size (85kDa and 38kDa) were detected in samples containing as little as lOOng total protein (figure 4.1) confirming that the polyclonal antibody is an efficient, sensitive means of detecting both spliced and precursor RecA protein.

4.3 A comparison of basal RecA levels in uninduced cell-free extracts ofM smegmatis and members of the M, tuberculosis complex

4.3.1 Construction of growth curves forM. smegmatis, M. tuberculosis and M. microti

Cultures of M. smegmatis, M. tuberculosis and M. microti in their stationary phase were used to inoculate fresh media at 1/250. The change in optical density (ODgoonm) with time was monitored for M. smegmatis (32 hours), M. microti (7 days) and M. tuberculosis (28 days) and a log,o graph constructed. Representative examples of the curves obtained are illustrated in appendix II. Prokaryotic growth curves are typically sigmoidal in shape with clearly defined lag and logarithmic growth phases. However, the lag and logarithmic growth phase cannot be readily distinguished in the curves illustrated in appendix II; this observation is in agreement with previously published data (Fisher and Kirchheimer, 1952; Tsukamura, 1990). Equivalent growth stages for M smegmatis, M. microti and M tuberculosis were determined by visual assessment of their respective growth curves and in subsequent investigations M tuberculosis and M. smegmatis were harvested at an OD^oonm of 0.6 andM microti at an OD^oonm of 0.35.

4.3.2 Basal RecA levels in M. smegmatis, M. tuberculosis, M. microti and M. bovis BCG

Cultures of M. tuberculosis, M. microti, M. smegmatis and M. bovis BCG were grown as described (section 2.1). Cell-free extracts were prepared (section 2.5) and samples with 50pg of total protein separated on a 10% SDS-polyacrylamide gel. Proteins were transferred to PVDF and membranes developed using affinity purified M tuberculosis RecA antibody diluted 1/250 - 1/500 (depending on

76 1 2 3 4 5 6

101,000 83,000

50,600

35,500

Figure 4.1 Confirmation that affinity purified M. tuberculosis RecA antibody will recognise RecA precursor protein

E. co//TG2 harbouring plasmid pEJ170, a double frameshift mutant of pEJ135 (Davis eta l, 1991; 1992) producing RecA precursor protein (U, 85kDa) and a small amount of processed RecA protein (S, 38kDa), was grown in L-broth in the presence of methicillin and ampicillin. A cell-free extract was prepared and ten-fold dilutions (lanes l-6\ lOOpg-lng total protein) were analysed by Western blot with affinity purified

M tuberculosis RecA antibody diluted 1/500. Two major bands of predicted size (85kDa and 38kDa) were detected in samples containing as little as lOOng total protein (lane 4) confirming that tlie polyclonal antibody is an efficient, sensitive means of detectmg both spliced and precursor RecA protein. Molecular weights are given in kDa.

77 batch). Cell-free extracts (50ng total protein) prepared from E. coli TG2{pEJ169(pGPl-2)} and E. co/i{TG2(pGP 1-2)} were included as positive and negative controls respectively. A representative filter is illustrated in figure 4.2a. In each case a band was observed migrating at the same position as the positive control (figure 4.2a, Pr). A visual inspection of band intensities revealed that lower RecA levels are present in M. tuberculosis and M. bovis BCG compared to M. microti and M. smegmatis. Interestingly, precursor RecA protein was not detected in any sample from the M. tuberculosis complex, which suggests that splicing occurs as soon as translation is complete.

To confirm that a variation in the stability of RecA (38kDa) was not an important determinant of the basal RecA levels described above, an identical membrane was developed using polyclonal antisera raised against the intern of M. tuberculosis RecA (kindly provided by Dr. E.O. Davis, N.I.M.R.) diluted 1/500 (figure 4.2b). A band of the expected size (47kDa) was present in M. tuberculosis, M. microti and M. bovis BCG, as anticipated no comparable bands were found m M. smegmatis. A visual inspection of band intensity confirmed that basal RecA levels are mirrored by that of the intein; M. microti has a higher basal level compared to M. tuberculosis and M. bovis BCG. Visual confirmation that the variation in basal RecA levels was not a consequence of differential protein loading was achieved by comparing the basal levels of an internal protein standard (figure 4.2c). For this a third membrane using identical cell- free extracts (5pg total protein per sample) was developed using a monoclonal antibody raised against the M. tuberculosis 65kDa-antigen diluted 1/1000 (kindly provided by Dr R Tascon, N.I.M.R.).

4.4 The effect of growth stage on precursor and mature RecA (38kDa) levels in M. tuberculosis and M. microti

The effect of growth stage at the point of harvesting on precursor and mature RecA levels was also of interest. A partial induction of some E. coli SOS genes has been reported to occur spontaneously during the growth cycle of bacteria. This phenomenon has been linked to transient changes in intracellular pH during growth, which affects the DNA binding affinity of LexA (Dri and Moreau, 1993; 1994). It is perhaps noteworthy that mycobacteria do not exhibit the typical sigmoidal growdh curve characteristic of many prokaryotes (see above). Instead of a lag phase they have been shown to possess an induction phase during which there is an increase in cell size concurrent with a period of rapid multiplication and cell death (Tsukamura, 1990). The successive period is considered to correspond to logarithmic growth.

Cultures of M. tuberculosis and M. microti in their stationary phase were used to inoculate fresh media at 1/250. Two 250ml cultures of M tuberculosis were set up and incubated under identical conditions and a 100ml volume removed from alternate cultures at an ODgoonm 0.25, 0.45, 0.6 and 0.9. For M. microti three 250ml cultures were set up and incubated under identical conditions and a 100ml volume removed from alternate cultures at an OD^oonm of 0.1, 0.23, 0.35, 0.45 and 0.9. Cell-free extracts were prepared and samples with 50pg of total protein separated on a 10% SDS-polyacrylamide gel. Proteins were transferred to PVDF (see section 2.6) and membranes developed using affinity purified M. tuberculosis

78 •Î2 ES co co p s I 1 HH.ÇO O ) II CD s l l ! l llico -Q II» I a) ë ë g C)

101,000 101,000 83,000 83,000 50,600 101,000 Intein 83,000 50,600 35,500 RecA 65KDa 35,500 29,100 50,600 29,100 20,900 35,500 20,900 29,100 Figure 4.2 A comparison of basal RecA levels in uninduced cell-free extracts of M. smegmatis and members of the M. tuberculosis complex

Cell-free extracts of M tuberculosis, M. microti, M. bovis BCG and A4, smegmatis (50|ig protein) were analysed by Western blot with affinity purified M tuberculosis RecA antibody (a). Basal

RecA levels were shown to be elevated in M microti and M smegmatis compared to M. tuberculosis and M bovis BCG. No RecA precursor protein was detected (a). Cell-free extracts (50ng protein) of E. coli TG2(pEJ169(pGPl-2)} ( P r ) and E. co//{TG2(pGPl-2)} (Pf^) were included as positive and negative controls respectively. An identical membrane probed using antisera raised against the intein of M tuberculosis RecA revealed that levels o f this genetic element in the M tuberculosis complex mirror those of RecA(b), which suggests that the observed basal RecA levels are not a result of protein instability. Visual confirmation that the variation in basal RecA levels was not a consequence of differential protein loading was achieved by comparing the basal levels of an internal protein standard, theM tuberculosis 65kDa-antigen (c). Molecular weights are given in kDa.

79 RecA antibody diluted 1/250 - 1/500 (depending on batch). Cell-free extracts (50ng total protein) prepared fromÆ co//TG2{pEJ169(pGPl-2)} and& co//{TG2(pGPl-2)} were included as positive and negative controls respectively. Representative filters are illustrated in figure 4.3. The results suggest that the growth stage at which a culture is harvested has no effect upon basal RecA levels for either M. tuberculosis (figure 4.3a) or M. microti (figure 4.3 b). Furthermore, no RecA precursor protein was detected at any growth stage providing further evidence to suggest that splicing of the precursor is a very rapid process, which occurs as soon as translation is complete.

4.5 A comparison of basal RecA levels in clinical isolates of M. tuberculosis

The evidence presented above suggests that there is a low level of RecA in M. tuberculosis. To determine if this observation is specific for strain H37Rv or a more general property of the species nine clinical isolates of M. tuberculosis (kindly provided by Dr P. Butcher, St George’s Hospital Medical School) were examined. Cultures were grown as described (section 2.1) to an ODgoonm of 0.6 and cell-free extracts prepared. Samples with 50pg of total protein were separated on a 10% SDS-polyacrylamide gel, transferred to PVDF and membranes developed using affinity purified M tuberculosis RecA antibody diluted 1/250 - 1/500 (depending on batch). A band that migrated in the same position as the positive control was observed in all samples, representative filters are shown in figure 4.4. A visual assessment of the intensity of this band revealed that basal RecA levels in the majority of the clinical isolates examined in this assay are comparable to that ofM tuberculosis H37Rv, which suggests that low basal RecA levels are a general property of M. tuberculosis. There were two exceptions; clinical isolate M. tuberculosis 160 consistently exhibited significantly increased levels compared to M tuberculosis H37Rv and slightly elevated levels were seen in M. tuberculosis 155. Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the M. tuberculosis 65kDa antigen (kindly provided by Dr R Tascon, N.I.M.R.) to confirm protein-loading equivalence.

4.6 A comparison of basal RecA levels in a range of saprophytic and pathogenic mycobacteria

In addition to the major pathogenic species M. tuberculosis, M. bovis and M. leprae the genus Mycobacterium includes many species found in the environment, the majority of which are not harmful to man. To determine whether low basal RecA levels are a common feature of pathogenic, as opposed to saprophytic, mycobacteria a range of saprophytic (M. chelonei subsp. abscessus, M. chelonei subsp. chelonei, M. fortuitum, M. phlei, M. thermoresistible) and pathogenic (M. avium, M. bovis, M. habana, M. intracellulare, M. kansasii, M. scrofulaceum) mycobacteria were examined. Cell-free extracts were prepared and samples with 50pg total protein separated on a 10% SDS-polyacrylamide gel, transferred to PVDF and membranes developed using affinity purified M. tuberculosis RecA antibody diluted

80 Optical Density Optical Density

a) b)

101,000 83,000 - 101,000 W 83,000 50,600 ^ 50,600 RecA R ecA 35,500 35,500 29,100 m 29,100

Figure 4.3 The effect of growth stage on precursor and mature RecA (38kDa) levels in M. tuberculosis and M. microti

Cultures o fM tuberculosis and M. microti in their stationary phase were used to inoculate fresh media at 1/250 to give 2 x 250ml volumes M tuberculosis and 3 x 250ml volumes M microti.

A 100ml volume was removed from alternate cultures at an ODgoonm of 0.25, 0.45, 0.6 and 0.9 for M tuberculosis and 0.1, 0.23, 0.35, 0.45 and 0.9 forM microti. Cell-free extracts (50|_ig protein) were analysed by Western blot with aflmity purified M. tuberculosis RecA antibody. The stage of growth at which a culture is harvested has no effect upon basal RecA levels for either

M. tuberculosis (a) orM . microti {b) and no RecA precursor protein was detected at any growth stage. Cell-free extracts (50ng protein) o f£ . coli TG2{pEJ169(pGPl-2)} (P r ) and

E. co//{TG2(pGPl-2)} (?m) were included as positive and negative controls respectively. Molecular weights are given in kDa.

81 M. tuberculosis M. tuberculosis clinical isolates clinical isolates

X > > cr DC c\j a> 00 1- IT) o CDCMCM h- in ^ _ CD LO CD ID c d £ 2 cc ^cD in ^ in 22 X C L CL T—T— 1— -1— 1— X C L Û -

101,000 83,000 101,000 83,000 50,600 RecA- 50,600 35,500 RecA- 29,100 35,500 29,100

Figure 4.4 A comparison of basal RecA levels in clinical isolates of M. tuberculosis

Cell-free extracts (50pg protein) prepared from nine clinical isolates of M tuberculosis were analysed by Western blot with affinity purified M tuberculosis RecA antibody. Basal RecA levels in

the majority of clinical isolates examined in this assay are comparable to that o fM tuberculosis H37Rv. There are two exceptions; M tuberculosis 160 has significantly increased levels compared

to M tuberculosis H31Kv and slightly elevated levels are present inM tuberculosis 155. Cell-free extracts (5Ong total protein) o f£ . co/i TG2{pEJ169(pGPl-2)}(P r ) and£. co//{TG2(pGPl-2)}

(?n) were included as positive and negative controls respectively. Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the

M. tuberculosis 65kDa antigen to confirm protein-loading equivalence. Molecular weights are given in kDa.

82 1/250 - 1/500 (depending on batch). Cell-free extracts (50ng total protein) prepared from E. coli TG2{pEJ169(pGPl-2)} andE. coh {TG2(pGP 1 -2)} were included as positive and negative controls respectively. Representative filters are illustrated in figure 4.5. A band with a migration comparable to that of the positve control is present in all samples. A qualitative assessment of the intensity of this band in each sample suggests that basal RecA levels in the majority of mycobacterial species examined are comparable to that of M tuberculosis. There are minor variations, for example, a consistently less intense signal was observed for M. thermoresistible and M. phlei compared to the M. tuberculosis sample on the same membrane. Similarly, M. habana was intermediate between M. tuberculosis and M. microti.

The M. smegmatis RecA protein (Papavinasasundaram et al, 1997) is highly conserved compared with the RecA proteins of M tuberculosis andMleprae (91% identity; Davis et al, 1991; 1994). Considering the homology between these three species it seems reasonable to assume that a similar level of homology at the amino acid level will exist among a wider range of mycobacteria. However, the precise amino acid identity is not known, therefore, it is worth noting that recognition of the RecA protein in other mycobacteria using affinity purified M. tuberculosis RecA antibody may not be so good as a result of differences in affinity and avidity (see section 3.1). This would lead to a lower specificity for antigen recognition by the M. tuberculosis RecA antibody in other mycobacterial species, which may result in a less intense signal. Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the M. tuberculosis 65kDa antigen (kindly provided by Dr R Tascon, N.I.M.R.) to confirm protein-loading equivalence.

4.7 The effect of cultivation in alternative media on basal RecA levels in M, smegmatis and the M. tuberculosis complex

Mycobacteria are a diverse group in terms of their chemical requirements for optimal growth. The requisite nutrients for pathogenic organisms are frequently more complex than for saprophytic ones. Fast growing species such as M. smegmatis are generally cultivated in Lemco broth or Middlebrook 7H9 media with ADC supplement; slow growers in Dubos or Middlebrook 7H9 with OADC supplement (see appendix I) thereby providing optimal growth conditions. To confirm that the difference in basal RecA levels between M. microti, M. smegmatis, M. tuberculosis and M bovis BCG is not an artefact of the different growth media used they were cultivated in Middlebxook 7H9 (plus supplement) and Sautons media (appendix I). Cell-free extracts with 50pg total protein were separated on a 10% SDS- polyacrylamide gel, transferred to PVDF and probed using affinity purified M. tuberculosis RecA antibody diluted 1/250 - 1/500. Representative filters are illustrated in figure 4.6. A band consistent with the migration of the positve control was present in all samples. The basal RecA levels found following growth of these four species in Middlebrook 7H9 (figure 4.6a) mirror those described previously. RecA levels in M. microti and M smegmatis are considerably greater than those found in M. tuberculosis and M. bovis BCG. Cultivation in Sautons media (figure 4.6b) produced similar results although basal RecA levels in M. bovis BCG also appeared to be slightly elevated compared to M. tuberculosis. Prior to use in

83 . I lCü co î co 8 5 g i Iq3 Q) 0) 0) I I Il Q) ■i -c -2 ICD Ice a LCQ Q . ■S 5 lllîl Iill CE ^ ^ g CL , 101,000 101,000 83,000 83,000 i 50,600 A 50,600 R ecA 35,500 35,500 29,100 1 ^ 2 9 , 1 0 0 m ; 20,900 20,900

Figure 4.5 A comparison of basal RecA levels in a range of saprophytic and pathogenic mycobacteria

Cell-free extracts (50pg protein) prepared from cultures of saprophytic and pathogenic mycobacteria were analysed by Western blot with affinity purified M tuberculosis RecA antibody. Basal

RecA levels in the majority of mycobacterial species examined are comparable to that o fM tuberculosis. There are exceptions, a consistently less intense signal was observed for

M thermoresistible andM phlei compared to theM. tuberculosis sample on the same membrane and RecA levels inM habana are intermediate betweenM tuberculosis andM microti. Cell-free extracts (50ng protein) of E. coli TG2{pEJ169(pGPl-2)} (P r ) and E. co//{TG2(pGPl-2)} (P ^ ) were included as positive and negative controls respectively. Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the M. tuberculosis 65kDa antigen to confirm protein-loading equivalence. Molecular weights are given in kDa.

84 ■s 8 I s CQ DQ I I■§ .o ■§ II I 5 I5 a) b) 0 . ^ 0 ?

101,000 101,000 83,000 « H 83,000

50,600 50,600 R ecA . RecA 35,500 35,500 29,100 29,100

20,900 20,900

Figure 4.6 Basal RecA levels in M. smegmatis and members of the M. tuberculosis complex cultivated in either Middlebrook 7H9 or Sautons media

Cell-free extracts (50|ig protein) prepared from M. tuberculosis, M. microti, M. smegmatis and M. bovis BCG grown in either Middlebrook 7H9 or Sautons media were analysed by Western blot with affinity purified M. tuberculosis RecA antibody. The basal RecA levels in M. microti and M. smegmatis grown in Middlebrook 7H9 are elevated compared to those found inM. tuberculosis and M. bovis BCG (a). Similarly basal RecA levels are elevated in M. microti and M. smegmatis compared to M. tuberculosis grown in Sautons media. In addition, RecA levels are also elevated in M. bovis BCG (b). Cell-free extracts (50ng protein) of E. coli TG2{pEJ169(pGPl-2)} (PJ and E. co//{TG2(pGPl-2)} (Pn) were included as positive and negative controls respectively. Protein-loading equivalence was confirmed with the M. tuberculosis 65kDa antibody. Molecular weights are given in kDa.

85 this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the M. tuberculosis 65kDa antigen (kindly provided by Dr R Tascon, N.I.M.R.) to confirm protein-loading equivalence.

4.8 An in vitro model that simulates dormancy: the effect on basal RecA levels and the appearance of precursor RecA protein

Precursor RecA protein is not present at detectable levels under any of the above in vitro conditions. It has been reported that 44 proteins in M. tuberculosis exhibit a modulated expression within the intracellular environment (Lee and Horwitz, 1995). Hence, the effect that intracellular residence may have on basal RecA levels and the appearance of precursor RecA protein is of great interest, particularly if splicing does regulate display of the RecA phenotype in vivo. Cell-free extracts were prepared from M. tuberculosis bacilli harvested from tissue samples (lung and spleen) collected from previously inoculated mice (kindly provided by Mr. E. Stavropolous, N.I.M.R.). Ten-fold dilutions of a lOmg/ml cell-free extract were separated on a 10% SDS-polyacrylamide gel, transferred to PVDF and membranes developed using affinity purified M tuberculosis RecA antibody diluted 1/250-1/500. However, the presence of a significant proportion of eukaryotic proteins made loading a sufficient amount of prokaryotic protein difficult, and developed membranes had a very high background which made interpretation difficult (data not shown). The fact that the M. tuberculosis 65kDa antigen is induced by intracellular residence (Lee and Horwitz, 1995) made it an unsuitable internal standard in this instance.

As an alternative, an in vitro model thought to be representative of the adaptation of tubercle bacilli to growth-suppressive conditions in inflammatory and necrotic tissues provided the opportunity to mimic the in vivo conditions of dormancy. In the United States this latent infection is thought to account for the majority of tuberculosis cases reported annually (Wayne, 1976; Wayne and Sramek, 1994; Wayne, 1994). Abrupt exposure to anaerobic conditions has a lethal effect on actively growing cultures of M. tuberculosis (Wayne and Lin, 1982). However, bacilli that have settled through a self-generated oxygen depletion gradient in unagitated cultures undergo an orderly metabolic shift-down from active replication to dormancy (Wayne, 1976; Wayne and Sramek, 1994). Using identical methodology the effect of a cessation in active DNA replication on the basal level of RecA (38kDa) and precursor protein (85kDa) was investigated. A 250ml culture of M tuberculosis was grown as described (section 2.1) and its growth monitored by measuring the ODgoonm- When it had reached an ODgoonm of 0.6 10ml aliquots were transferred to 30ml screw-capped polypropylene tubes (Elkay) and incubated upright in a 37°C stationary incubator for 28 days; allowing cells to settle and deplete their available oxygen. After this period of time a mixed population of slowly replicating and dormant bacilli (found in the supernatant suspension and settled sediment respectively) were harvested and a cell-free extract produced. Membranes were prepared and developed as described above. Cell-free extracts (5Ong total protein) prepared from E. coli TG2{pEJ169(pGPl-2)} andÆ co/i{TG2(pGPl-2)} were included as positive and negative controls respectively. A representative filter is illustrated in figure 4.7. Examination of membranes revealed that

86 k 1 1 111 1 Si S I si S S Q."

101,000 83,000 50,600 R ecA 35,500 29,100 20,900

Figure 4.7 The effect of oxygen starvation on basal RecA levels and the appearance of precursor RecA protein in M. tuberculosis

A 250ml culture o f M tuberculosis was grown to an ODgoonm of 0 6. Ten millilitre aliquots were transferred to

30ml screw-capped polypropylene tubes and incubated upright in a 37°C stationary incubator for 28 days; allowing cells to settle and deplete their available oxygen. After this period of time a mixed population of slowly replicating and dormant bacilli (found in the supernatant suspension and settled sediment respectively) were harvested. Cell-free extracts (50pg protein) were prepared and analysed by Western blot with affinity purified

M tuberculosis RecA antibody. Examination of membranes revealed that levels of RecA inM tuberculosis grown under limiting oxygen conditions (dormant cells) are comparable to those present when it is grown under optimal conditionsin vitro. RecA precursor protein was not present at detectable levels. Cell-free extracts (5Ong total protein) of E.coli TG2{pEJ169(pGPl-2)} (P r ) andE. co//{TG2(pGPl-2)} (P ^ ) were included as positive and negative controls respectively. Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the M tuberculosis 65kDa antigen to confirm protein-loading equivalence. Molecular weights are given in kDa.

87 levels of RecA in M tuberculosis grown under limiting oxygen conditions are comparable to those present when it is grown under optimal conditions in vitro. RecA precursor protein was not present at detectable levels. Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the M tuberculosis 65kDa antigen to confirm protein-loading equivalence.

4.9 Confirmation of the presence of a LexA homologue inM, smegmatis and the M. tuberculosis complex and a comparison of basal levels

Cell-free extracts ofM. smegmatis, M tuberculosis, M microti andM. bovis BCG (2*5pg total protein) were separated on a 15% SDS-polyacrylamide gel. After transfer to PVDF, membranes were developed using a mouse polyclonal antibody raised against M tuberculosis LexA (kindly provided by Dr. P Jenner, N.I.M.R) diluted 1/1000. Purified M. tuberculosis LexA (2 5ng) was included as a positive control. A band that migrated in the same position as the positive control was observed in M. tuberculosis, M. microti and M. bovis BCG; representative filters are shown in figure 4.8. Curiously, a single band with a migration consistent with that of a slightly higher molecular weight protein was found in M smegmatis. A visual assessment of the intensity of each band revealed that comparable amounts of LexA are present in M. smegmatis, M. tuberculosis and M. bovis BCG whereas higher levels are found in M microti. Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against the M. tuberculosis 65kDa antigen (kindly provided by Dr R Tascon, N.I.M.R.) to confirm protein-loading equivalence.

4.10 Discussion

Low levels of homologous recombination have been reported in M tuberculosis (Kalpana et al, 1991) and M bovis BCG (Aldovini et al, 1993). One possible explanation to account for this observation is that an insufficient quantity of RecA is produced. To address this question the basal levels of RecA in M smegmatis and three members of the M tuberculosis complex were compared. Surprisingly, although the amount of RecA in uninduced cell-free extracts was similar for M. tuberculosis and M. bovis BCG it was significantly greater in M. microti and M smegmatis. Probing an identical membrane with an antibody raised against the intein of M tuberculosis RecA revealed that levels of this genetic element in M tuberculosis, M. bovis BCG and M. microti mirror those of RecA, which suggests that the observed basal RecA levels are not a result of protein instability. It is possible that that both the RecA protein and intein are equally unstable in M. tuberculosis and M. bovis BCG compared to M. microti. Although this seems unlikely considering the genetic homogeneity reported among members of the M. tuberculosis complex (Imaeda, 1982; Kempsell et al, 1992; Kirschner et al, 1993; Frothingham et al, 1994; Takewaki et al, 1994). Further corroboration of these results is obviously desirable. Preliminary investigations using CO % CD CO CD o CD ICQ I •§ II 3 I1! LU

101,000 101,000 83,000 83,000 50,600 LexA-^ — 35,500 50,600 29,100 35,500 20,900 LexA- 29,100

20,900

Figure 4.8 Confirmation of the presence of a LexA homologue in M. smegmatis and the M. tuberculosis complex and a comparison of basal levels

Cell-free extracts (2-5pg protein) of M smegmatis, M. tuberculosis, M microti and M. bovis BCG were analysed by Western blot with a polyclonal antibody raised against M tuberculosis LexA. A band that migrated in the same position as the positive control (P lexa ) was observed in M microti, M. tuberculosis and M bovis BCG. In M smegmatis there is a single band with a migration consistent with that of a slightly higher molecular weight protein. Basal LexA levels are comparable in M smegmatis, K4. tuberculosis and M bovis BCG whereas they are elevated in M. microti.

Prior to use in this assay all cell-free extracts had been examined by Western blot using a monoclonal antibody raised against th eM tuberculosis 65kDa antigen to confirm protein-loading equivalence. Molecular weights are given in kDa.

89 Northern blots and the RNase protection assay (Mironov et al, 1995) to compare the expression of recA at the mRNA level were unsuccessful, possibly because of a low expression level. Decreasing the hybridisation temperature to reduce stringency had no positive effect. More recently, investigations have been initiated using quantitative RT-PCR; it is anticipated that these will yield positive results at some point in the near future.

Examination of nine clinical M. tuberculosis isolates revealed that low RecA levels are a general property of this species, which provides further evidence to suggests that an insufficient quantity of RecA (38kDa) may account for the low levels of homologous recombination in M. tuberculosis and M. bovis BCG. However, a qualitative comparison in a wider range of mycobacteria revealed that RecA levels in the majority of species are comparable to that of M. tuberculosis. This includes M. intracellulare in which homologous recombination has been successfully demonstrated (Marklund et al, 1995). The obvious implication of is that a low level of RecA per se may not be a factor in the recalcitrance of M tuberculosis and M. bovis BCG to undergo homologous recombination. The possibihty that the presence of an intein in the recA locus may affect this RecA mediated function is certainly an intriguing notion, particularly in view of further evidence which suggests that the presence of precursor protein might have dominant negative effects, possibly by competing for RecA binding sites on target molecules or by polymerising with the 38kDa RecA monomers (Davis et al, 1992). Interestingly, the 85kDa-precursor protein has been shown to be totally defective in its DNA-binding, ATP-binding and ATPase activities. In contrast the spliced 38kDa RecA protein binds ssDNA in the presence of ATP, possesses ATPase activity and is proficient in homologous pairing; although there are significant differences from E. coli RecA in its requirements for binding to DNA and its affinity for ATP (Kumar et al, 1996). By comparison the functions of three other microbial RecA proteins from Bacillus subtilis (Lovett and Roberts, 1985), Proteus mirabilis (West et al, 1983) and Thermus aquaticus (Angov and Camerini-Otero, 1994) are reported to be largely similar to that of E. coli (Sassanfar and Roberts, 1990; Radding, 1991; West, 1992; Kowalczykowski et al, 1994).

Cultivation in Middlebrook 7H9 had no discernible effect on basal RecA levels. Similarly growth in Sautons media had no effect on RecA levels in M microti and M. smegmatis, which were significantly greater than those found in M. tuberculosis', however, levels of RecA in M. bovis BCG were elevated compared to M. tuberculosis. The precise reason behind this difference is unclear but may be linked to an alteration in extracellular pH since Sautons media provides a mildly acidic environment (pH6.2). Several papers have shown that extracellular pH changes have transient effects on internal cellular pH homeostasis resulting in the altered expression of several LexA repressor controlled genes (Schuldiner et al, 1986; Dri and Moreau, 1994). For example, the predominant action of weak acids in E. coli is to increase the DNA binding affinity of LexA to sfiA leading to further repression. Conversely, Relan et al (1997) have demonstrated in vitro that E. coli LexA tightly binds to the recA operator around neutral pH but is weakened as the pH is lowered. Clearly, further clarification regarding the thermodynamics of the interaction of a repressor molecule with the recA operator in M. smegmatis and the M. tuberculosis complex may be necessary before a satisfactory explanation for this difference can be offered.

90 The spontaneous, partial induction of some E. coli SOS genes during the growth cycle of bacteria has also been linked to transient changes in intracellular pH during growth which affects the DNA binding affinity of LexA (Dri and Moreau, 1993, 1994). This phenomenon was particularly interesting in the context of the difference in basal RecA levels between the M. tuberculosis complex. Perhaps more intriguing was the possibility that the presence of precursor protein might also be affected by growth stage, particularly if the appearance of splicing in the production of mature RecA (38kDa) does regulate the display of the RecA phenotype. To address this question uninduced cell-free extracts were prepared from M. tuberculosis and M. microti cultures harvested at different stages of growth. Interestingly no precursor RecA protein was detected at any of the growth stages examined. Taking into consideration the lack of detectable precursor protein in M. tuberculosis and M microti grown under the aforementioned in vitro conditions this latter observation reinforces the suggestion that splicing and translation may be overlapping processes. Basal RecA levels in both M. microti and M. tuberculosis were found to be consistent in very young to stationary cultures. This tends to dismiss the possibility that the high levels of RecA seen in M. microti compared to other members of the M. tuberculosis complex are due to spontaneous induction of the recA gene during growth.

It is possible that the higher levels of RecA found in M. microti and M. smegmatis compared to M. tuberculosis and M. bovis BCG are due to a lack of repression of the recA gene. Examination of cell-free extracts from these four species by Western blot using a polyclonal antibody raised against M. tuberculosis LexA revealed that they all possess a LexA homologue. The demonstration of a higher basal level of this homologue in M. microti compared to M. tuberculosis and M bovis BCG provides further support for the suggestion that the elevated RecA levels in M. microti may be a due to a lack of repression of SOS genes. Interestingly, basal levels of this LexA homologue in M. smegmatis were comparable to that of M. tuberculosis although its migration was consistent with a protein of slightly higher molecular weight. No direct evidence of a role for this LexA homologue in the regulation of the SOS response has been presented here. However, the effect of treatment with a DNA damage-inducing agent on the levels of this protein will be examined in a later chapter.

A second explanation to account for the low basal RecA levels in M. tuberculosis and M. bovis BCG compared to M. smegmatis and M. microti may be due a difference in promoter strength (i.e. the relative frequency of transcription initiation) between these species. Very few mycobacterial promoters have been studied so far; as a consequence progress toward understanding gene regulation and expression in this genus has been slow. Examples include the rpsY, gene of M. smegmatis (Kenney and Churchward, 1996); the M. bovis hsp60 gene and the M. leprae 28kDa and 18kDa genes (Dellagostin et al, 1995); the 16S rRNA genes o fM leprae, M. tuberculosis andM. smegmatis (Sela and Clark-Curtiss, 1991; Ji e? al, 1994b; Verma et al, 1994); the bla gene ofM fortuitum (Timm et al, 1994); the cpn-60 and 85kDa genes of M tuberculosis (Kong et al, 1993; Kremer et al, 1995) and the P^n promoter from M paratuberculosis (Murray etal, 1992).

The M. smegmatis recA gene has recently been cloned and sequenced (Papavinasasundaram et al, 1997);

91 a high degree of homology exists between this and the M tuberculosis recA gene (91% identity). This homology extends upstream of the recA coding region and includes two putative promoter sites identified on the basis of their homology with the consensus sequences of E. coli and E. coli promoters (Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). However, the spacing between the -10 and -35 elements is different to that found in E. coli (9bp as opposed to 17-21 bp, Rosenberg and Court, 1979). This is by no means unique, the -10 and -35 elements of the promoter for the M. tuberculosis 85kDa antigen are separated by 64bp (Kremer et al, 1995). It is not clear whether mycobacterial RNA polymerases are able to recognise these motifs with a different spacing or whether the -35 region possesses an alternative consensus. Bashyam et al (1996) have suggested that the transcriptional signals of mycobacteria have features in common with Streptomyces in that while the -10 consensus sequence shows a high degree of conservation a larger variety of sequences can be accommodated in the -35 region. In Streptomyces this is thought to be due to the presence of multiple sigma factors with different or overlapping specificities for this region (Bibb and Cohen, 1982; Westpheling et al, 1985; Strohl, 1992).

In addition to the -10 and -35 elements of the promoter site a third motif has been shown to affect promoter activity. This motif is found immediately upstream of the -10 hexamer and is described as an extended -10 motif. A specific TG dinucleotide in this region is reported to be essential for normal promoter activity of the M smegmatis rpsh gene in the absence of a functional -35 hexamer (Kenney and Churchward, 1996). The promoters of both M. bovis hsp60 (Stover et al, 1991) and M. leprae 168 rRNA (Sela and Clark-Curtiss, 1991) also possess this extended -10 region. Notably, this motif is absent in the rpsL promoter of M. tuberculosis, M. bovis and M. leprae which possess a 5- of 6- base pair match with the E. coli a^°-35 site. An extended -10 region, together with a conserved TG dinucleotide, has been identified in B. subtilis (Helmann, 1995; Voskuil et al, 1995) and in approximately 16% of E. coli promoters (Harley and Reynolds, 1987; Kumar et al, 1993). It is thought to affect promoter utilisation and has been shown to be essential to weak promoters but of little benefit to strong promoters (Moran et al, 1982; Graves and Rabinowitz, 1986; Helmann, 1995; Tatti and Moran, 1995) which makes the observation that strong promoters occur less frequently in M. tuberculosis than M smegmatis (DasGupta et al, 1993) particularly interesting.

Examination of the putative promoter site with homology to the E. coli promoter in M. smegmatis recA revealed that an extended -10 region with a conserved TG dinucleotide is present immediately upstream of the -10 hexamer, whereas it is absent in M. tuberculosis. This suggests that promoter strength may be a factor in the difference in basal RecA levels between M. tuberculosis and M. smegmatis. However, analysis of this putative promoter site upstream of the recA genes in M. tuberculosis and M. smegmatis using the chloramphenicol-acetyltransferase (CAT) reporter gene revealed that the uninduced expression of recA is similar in both species (Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). This suggests that there is no significant difference in intrinsic promoter strength between M. tuberculosis and M. smegmatis recA. A second putative promoter site with homology to the E. coli promoter site is located further upstream in both M. smegmatis and M. tuberculosis which lacks an

92 extended -10 motif. This finding is not unexpected since the B. subtilis promoter also lacks this motif (Voskuil et al, 1995). However, this second putative promoter site overlaps the Cheo box, which has been implicated in LexA bmding in both M. smegmatis and M. tuberculosis (Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997) and may be of more consequence than the proximal putative promoter site to regulation of this gene. A more rigorous examination of both these putative promoter sites may provide further insight into their role in the regulation of the recA gene.

Precursor RecA protein was not detected in cell-free extracts harvested from M. tuberculosis, M microti or M. bovis BCG grown under optimal conditions in vitro. Furthermore its appearance was not affected by the stage of growth nor by mild acidity. It is tempting to speculate that this lack of detectable precursor protein in M tuberculosis and M. bovis BCG is a reflection of the low levels of RecA per se in these species. However, this seems unlikely since it was also absent in M microti, which has significantly elevated levels compared to other members of the M. tuberculosis complex. Members of the M. tuberculosis complex are facultative intracellular parasites that survive within host macrophages where it is presumed they will encounter a wide range of anti-microbial defences. Intracellular residence of M. tuberculosis in macrophages induces a profound phenotypic change in protein expression (Lee and Horwitz, 1995). If the appearance of splicing regulates display of the RecA phenotype it may be anticipated that intracellular residence affects the appearance of spliced and precursor RecA protein. To address this issue cell-free extracts were prepared from bacilli harvested from tissue samples (lung and spleen) collected from nude mice that had previously been infected with M. tuberculosis H37Rv. Examination of membranes developed using affinity purified M tuberculosis RecA antibody did not provide any conclusive information. No clear signals were discernible in any sample since a high background made interpretation impossible. The main problem encountered with this protocol was the difficulty in separating the bacilli and tissue prior to cellular disruption. This led to a significant amount of eukaryotic protein contamination making loading of a known amount of prokaryotic protein difficult. This latter problem was compounded by the lack of a suitable internal standard since the M. tuberculosis 65kDa antigen is induced by intracellular residence (Lee and Horwitz, 1995). Investigations have recently been initiated using quantitative RT-PCR to assess RecA mRNA levels in vivo and it is anticipated that these will yield positive results soon.

An alternative in vitro model thought to be representative of the adaptation of tubercle bacilli to growth- suppressive conditions in inflammatory and necrotic tissues provided an opportunity to mimic the in vivo conditions of dormancy (Wayne, 1976; Wayne and Sramek, 1994; Wayne, 1994). The effect of a cessation in active DNA replication on the levels of spliced (38kDa) and precursor RecA protein (85kDa) was examined in bacilli that had settled through a self-generated oxygen depletion gradient. These oxygen starvation conditions cause an orderly metabolic shift-down from active replication to dormancy (Wayne, 1976; Wayne and Sramek, 1994). Considering the proposed role of replication in the production of ssDNA regions, which has been reported to be the inducing signal (D’Ari, 1985; Sassanfar and Roberts, 1990; Higashitani et al, 1995; Morel-Deville et al, 1996; Shinagawa, 1996), it was possible that basal RecA levels would decrease. Furthermore, if splicing of RecA does regulate the display of this phenotype then

93 there may be a concomitant increase in precursor RecA levels. However, examination of membranes developed using affinity purified M. tuberculosis RecA antibody revealed that the levels of spliced and precursor RecA protein were unaffected by growth under these conditions. This model does have significant limitations; a slight change in growth rate may upset the equilibrium between replication and settling that is required to maintain the oxygen gradient necessary to support the shift down to dormancy (Wayne, 1976). More recently, an in vitro model permitting the examination of homogeneous populations of bacilli during discrete stages of this adaptation to low oxygen tensions has been developed (Wayne and Hayes, 1996). This may provide further insights into the possible role of splicing in the appearance of mature RecA protein (38kDa).

In summary, the amount of RecA in uninduced cell-free extracts was similar for M tuberculosis and M. bovis BCG but was found to be significantly greater in M. microti and M. smegmatis. Examination of a wider range of species revealed that RecA levels in many species of mycobacteria are comparable to M. tuberculosis. Hence, a low level of RecA is unlikely to be an important factor in the low levels of homologous recombination seen in M. tuberculosis and M. bovis BCG since homologous recombination has been reported in M. intracellulare (Marklund et al, 1995) which has RecA levels comparable to M. tuberculosis. Precursor RecA protein was not detected in M. tuberculosis, M. microti or M. bovis BCG grown under optimal in vitro conditions. This finding remained constant irrespective of the growth stage of the culture at the point of harvesting, providing further evidence to suggest that translation and splicing of the precursor protein may be overlapping processes. An attempt to determine the effect of intracellular residence on spliced and precursor RecA protein levels was unsuccessful although an in vitro model simulating the conditions of dormancy in M tuberculosis suggests that splicing of the precursor protein and basal RecA levels are unaffected under these growth conditions. Investigations assessing in vivo and in vitro RecA levels using quantitative RT-PCR have recently been instigated. It is hoped that corroboratory data will be available in the near future and that the effect of the effect of intracellular residence on RecA levels may be more readily assessed. The M. tuberculosis 65kDa antigen was used as an internal standard to confirm that differential protein loading was not a factor in the results described above. It is perhaps noteworthy that, while some forms of DNA damage induced its E. coli counterpart (GroEL; Kreuger and Walker, 1984), there is no evidence to suggest a similar phenomenon in mycobacteria. Evidence presented to date has shown that it is induced by heat shock and acidification but unaffected by oxidative damage (Stover et al, 1991; Young et al; 1991; Lee et al, 1995).

94 5.0 The effect of DNA damaging agents on RecA and LexA

LEVELS IN M SMEGMATIS, M, TUBERCULOSIS AiND M. MICROTI

5.1 Introduction

Genetic insult represents a serious challenge to the cellular survival of all living organisms. In E. coli, the RecA protein has a central role in maintaining genetic integrity and has been implicated in homologous recombination and post-replication repair (Roca and Cox, 1990; Miller and Kokjohn, 1990; Cox, 1993; Kowalczykowski et al, 1994). In addition, it co-ordinates the cellular response to agents or conditions that damage DNA by inducing the expression of a set of unlinked genes comprising the SOS regulatory network. Initiation of the SOS response occurs when activated RecA promotes the proteolytic cleavage of LexA protein, the repressor of SOS genes. These events result in the co-ordinated expression of a diverse set of cellular phenomena that characterise the SOS response; among which are the enhanced capacities for DNA repair and induction of the RecA protein (Little and Mount, 1982; Walker, 1984).

The SOS response of E. coli has become a paradigm for the study of inducible DNA repair and recombination processes in many different organisms. In view of the remarkable similarity between recA genes it seems reasonable to expect a certain degree of evolutionary conservation in the essential control elements of the SOS system among prokaryotes (Lovett and Roberts, 1985; Karlin et al, 1995: 1996). The fact that E. coli LexA can influence the induction of SOS-like functions in several species of enterobacteriaceae demonstrates that the mechanism of negative regulation of SOS-like responses may also be highly conserved (Sedgwick and Goodwin, 1985). Further compelling evidence has been provided by the SOS system of Bacillus subtilis, which has been shown to parallel that of E. coli in many respects (Yashin, 1977; Lovett et al, 1988; Raymond-Denise and Guillen, 1991; 1992; Wojciechowski et al, 1991; Yashin et al, 1992; Lovett et al, 1993; 1994). There is also evidence that a SOS-like response exists in Vibrio cholerae, Proteus mirabilis and Haemophilus influenzae (Hofemeister et al, 1975; Notani and Setlow, 1980; Ghosh etal, 1985).

At the outset of this project the M. tuberculosis recA gene had been cloned and sequenced (Davis et al, 1991) but nothing was known about the effect of DNA damaging agents upon its expression. Pathogens such as M tuberculosis survive and replicate within macrophages, which are an integral part of the host defence mechanism. Consequently, there is great interest in determining the effect of DNA damaging agents on recA expression in this genus. Of particular interest is whether it is induced following treatment with a DNA damage-inducing agent. If so, how do the kinetics of RecA induction compare among different species of mycobacteria? The RecA protein in members of the M. tuberculosis complex is unusual in that post-translational splicing of a longer gene product is required to produce a 38kDa protein proficient in binding ssDNA and homologous pairing (Davis et al, 1992; Kumar et al, 1996). Is the appearance of splicing in the maturation of RecA in these species a regulatory mechanism? In the

95 previous chapter we were unable to show that splicing could be conditionally regulated in uninduced cells. In this chapter the effect of exposure to DNA damaging agents on the kinetics of RecA induction in M. tuberculosis, M. smegmatis and M microti was examined together with the possibility that splicing of RecA may be conditionally regulated if RecA is induced following treatment with a DNA damaging agent. In addition, evidence presented in the previous chapter clearly demonstrates that M. smegmatis and members of the M. tuberculosis complex possess a LexA homologue. The presence of both of the essential regulatory elements of the SOS response in mycobacteria was intriguing. As a consequence we examined whether treatment with a DNA damaging agent led to cleavage of this protein.

5.2 The optimal conditions for exposure of mycobacteria to DNA damaging agents

The DNA damaging agents used in this study were ofloxacin and mitomycin C. Ofloxacin is a quinolone inhibitor of DNA gyrase. This compound exhibits a higher toxicity to mycobacteria than does nalidixic acid, the more frequently used gyrase inhibitor (Sugino et al, 1977). Mitomycin C introduces cross-links into DNA by alkylation of guanines (Iyer and Szybalski, 1963; Teng et al, 1989; Borowy-Borowski et al, 1990; Kumar et al, 1992). The ability of these compounds to induce the M tuberculosis recA promoter in M. smegmatis was assessed by Movahedzadeh et al (1997b) and the concentration that gave the greatest degree of induction used in further investigations. Movahedzadeh et al (1997b) also determined the optimal growth stage at the point of induction for M. smegmatis. Equivalent growth stages for M. microti and M. tuberculosis were determined by a visual assessment of the growth curves produced (appendix II) when mycobacteria were cultivated as described in section 2.1. For subsequent investigations cultures of M. tuberculosis and M. smegmatis were grown to an ODgoonm of 0.6; cultures of M. microti were grown to an ODgoonm of 0.35. At this point they were exposed to a final concentration of 2-5fxg/ml mitomycin C or Ipg/ml ofloxacin (prepared as described in section 2.7.1) and bacteria harvested at designated time intervals.

5.3 The effect of DNA damaging agents on the kinetics of LexA cleavage and RecA induction in mycobacteria

Mycobacteria were grown to the required ODgoonm and cell-free extracts prepared from aliquots removed at designated intervals following exposure to ofloxacin (Ipg/ml) or mitomycin C (2 5pg/ml). The protein content of each sample was determined using a BCA micro-titre plate assay and samples containing 50pg total protein separated on a 10% SDS-polyacrylamide gel. Proteins were transferred to PVDF (Immobilon-P, Millipore) and membranes probed using a 1/250 - 1/500 dilution of an affinity purified polyclonal antibody raised against recovcAimzxA M tuberculosis RecA (see section 2.6). Semi-pure M. tuberculosis RecA protein prepared as described (section 2.2) and a cell-free extract of E. co/zTG2(pGPl- 2) were included as positive and negative controls respectively (50ng total protein). The effect of

96 exposure to a DNA damaging agent on LexA protein levels was assessed by separating 3pg total protein on a 15% SDS-polyacrylamide gel. Proteins were immobilised by transfer to PVDF and membranes developed using a 1/1000 dilution of a mouse polyclonal antibody raised against M. tuberculosis LexA (kindly provided by Dr. P.J. Jenner, N.I.M.R.). Purified M tuberculosis LexA (2-5ng) was included as a positive control. A third membrane was prepared by separation of 3pg total protein on a 10% SDS- polyacrylamide gel and immobilisation of proteins by transfer to PVDF. Visualisation using a monoclonal antibody raised against the M. tuberculosis 65kDa antigen (diluted 1/1000; kindly provided by Dr R Tascon, N.I.M.R.) provided visual corroboration that the observed kinetic profiles for LexA cleavage and RecA induction are a consequence of exposure to a DNA damaging agent and not of differential protein loading.

5.3.1 The effect of exposure to DNA damaging agents on the kinetics of LexA cleavage and RecA induction in M. smegmatis mc^lSS

An overnight culture of M. smegmatis mc^l55 grown as described in section 2.1 was used to inoculate fresh media at 1/250. These cultures were grown to an ODeoonm of 0 6, treated with ofloxacin or mitomycin C (prepared as described in section 2.7.1) and cell-free extracts prepared from 50ml aliquots removed every 40 - 60 minutes for 6 hours. RecA and LexA protein levels were analysed by Western blot; representative filters are illustrated in figures 5.1 and 5.2. Examination of membranes developed using affinity purified M. tuberculosis RecA antibody revealed that RecA induction occurs within 1 hour of exposure to mitomycin C and 3 hours for ofloxacin, reaching a peak after 5 hours (figures 5.1a and 5.2a). Maintenance of this level in samples collected after 24 hours is probably due to the unremitting effect of DNA damaging agents still present within the culture (figure 5.2a). Examination of membranes developed using polyclonal M tuberculosis LexA antisera revealed that levels of LexA protein decline within 40 minutes of treatment with a DNA damage-inducing agent (figure 5.2b). An increase in the levels of LexA protein was reproducibly observed in M smegmatis at 4 - 5 hours; possibly due to an increase in lexA expression which became apparent because cleavage was slow. No LexA cleavage products were detected, probably due to in vivo instability of the cleaved products. Monitoring levels of the M. tuberculosis 65kDa antigen (figures 5.1b and 5.2c) provided visual corroboration that the observed kinetic profiles for LexA cleavage and RecA induction are a consequence of exposure to a DNA damaging agent and not of differential protein loading.

5.3.2 The effect of exposure to DNA damaging agents on the kinetics of LexA cleavage and RecA induction inM. tuberculosis H37Rv

Representative filters illustrating the kinetics of LexA cleavage and RecA induction in M tuberculosis H37Rv following exposure to ofloxacin or mitomycin C are shown m figures 5.3 and 5.4. Exposure of M. tuberculosis H37Rv over a 5-hour period had no discernible effect on the kinetics of LexA cleavage

97 a) Time (hours) b) Time (hours) 0 1 2 3 4 4V2 5 6%, 01 2 3 4 5 6%,

101,000 101,000 # 83,000 83,000 65KDa 50,600 50,600 RecA 35,500 35,500 29,100 29,100 20,900

Figure 5.1

The effect of exposure to ofloxacin on RecA levels in M. smegmatis mc^l55

Cultures o fM smegmatis mc^l55 were exposed to ofloxacin (Ipg/ml f.c.) and samples collected over a 6-hour period. Cell-free extracts (50|ag protein) were analysed by Western blot with afflnity-purifled M tuberculosis RecA antibody (a). A visual assessment of band intensity revealed that RecA levels increase 2 - 3 hours subsequent to exposure to a DNA damaging agent, reaching a peak after approximately 5 hours. Cell-free extracts (50ng protein) of E. coli TG2{pEJ169(pGPl-2)} (P r ) and E. coli {TG2(pGPl-2)} (P ^ ) were included as positive and negative controls respectively. Monitoring the levels of the Mtuberculosis 65kDa antigen (b) provided visual corroboration that the observed kinetic profile is a consequence of exposure to a DNA damaging agent and not of differential protein loading. Molecular weights are given in kDa.

98 Time (minutes) O OO OOO LOx: a) eg N . 101,000 83,d00 50,600 RecA -► r n ê m m ^ • 35,500 29,100

20,900

3 OOOOOOlT)^ li| b i ^ 83,000 ' 50,600 LexA —►— — ^ 35,500 - 29,100

20,900

o OO OOO m ^ OOCMCDO'^oqC^OO^ q\ O Tj- 1— 1— T— C\J C\j CO CO C\J 101,000 83,000

65KDa -► 50,600 35,500 29,100

20,900 Figure 5.2 RecA induction and LexA cleavage in M. smegmatis mc^lSS following exposure to mitomycin C

Cultures ofM. smegmatis mc^l 55 were exposed to mitomycin C (2.5pg/ml f.c.) and samples collected over a 6-hour period. Cell-free extracts were analysed by Western blot with affinity-purified M tuberculosis RecA antibody (a) or polyclonal M. tuberculosis LexA antisera (b). Examination o f membranes revealed that there is an increase in RecA levels 1 hour subsequent to exposure to a DNA damaging agent (a) and a decrease in LexA protein levels within 40 minutes, probably as a result of LexA cleavage (6). An increase that was reproducibly observed in LexA levels 4 - 5 hours subsequent to exposure may be due to an increase in lexA expression, which became apparent because cleavage was slow. No LexA cleavage products were detected, probably due to in vivo instability of the cleaved products. Positive and negative control lanes for monitoring RecA induction comprise ofcoli £. TG2{pEJ169(pGPl-

2)1 (Pr) and £. co/i{TG2(pGPl-2)} (P^, 5Ong protein). PurifiedM. tuberculosis LexA (Pujca. 2.5ng) was included as a positive control in all LexA filters. Monitoring levels of theM. tuberculosis 65kDa antigen (c) provided visual corroboration that the observed kinetic profiles for LexA cleavage and RecA induction are a consequence of

exposure to a DNA damaging agent and not of differential protein loading.

99 or RecA induction in this species (figure 5.3). However, additional samples collected every 3 hours for 30-hours revealed that there is a reproducible increase in RecA levels 24 hours subsequent to treatment with a DNA damage-inducing agent (figure 5.4a). Precursor RecA protein was not detected in any sample. Examination of LexA levels during this 30-hour period revealed a consistent decrease 6 hours subsequent to treatment with a DNA damaging agent (figure 5.4b). An increase in the levels of LexA protein that is reproducibly observed at 24 hours is possibly due to an increase in lexA expression that becomes apparent because cleavage is slow. This increase in LexA levels may account for the decrease in RecA protein levels at 27 hours (figure 5.4a). Monitoring levels of the M tuberculosis 65kDa antigen (figures 5.4c) provided visual corroboration that the observed kinetic profiles for LexA cleavage and RecA induction are a consequence of exposure to a DNA damaging agent and not of differential protein loading.

5.3.3 The effect of growth stage on the kinetics of RecA induction in M.tuberculosis H37RV

The growth curve for M. tuberculosis in Dubos liquid medium is not the usual sigmoidal curve seen for most prokaryotes; lag and log phase growth cannot be readily distinguished (appendix II). To confirm that the delay in RecA induction is not a consequence of growth stage at the point of induction, cultures with an ODgoomn oil the “logarithmic” portion of the growth curve (ODgoonm 0 2 - 0.4) were exposed to DNA damaging agents prepared as described (section 2.7.1). Interestingly, an increase in RecA levels was reproducibly observed 6 -1 0 hours subsequent to treatment with a DNA damage-inducmg agent (figure 5.5). Although this differs from the results described previously where induction was not detected for 24 hours, it is still significantly slower than that seen in M. smegmatis which implies that there is an intrinsic delay in M tuberculosis following genetic insult. The discordance between the rate of RecA induction in M tuberculosis cultures induced at an ODgoonm of 0 6 and those at 0 2 - 0 3 is probably due to the increased susceptibility of M tuberculosis at low cell density. It is noteworüiy that precursor RecA protein was not detected in any sample, providing further evidence to suggest that splicing does not regulate the appearance of RecA (38kDa) following induction of this protein as a result of DNA damage (see chapter 8 for a more comprehensive discussion).

5.3.4 The effect of exposure to DNA damaging agents on the kinetics of LexA cleavage and RecA induction inM. microti OVI254

Representative filters illustrating the kinetics of LexA cleavage and RecA induction in M. microti following exposure to ofloxacin or mitomycin C are shown in figures 5.6 and 5.7. Samples collected every 30 minutes for 7 hours following treatment with a DNA damage-inducing agent revealed that the levels of RecA and LexA in M. microti OV1254 are unaffected during this period (figure 5.6). Further samples collected over a 30-hour period also failed to demonstrate any alteration in the levels of RecA

100 a) Time (minutes) b) Time (minutes) X O O O O O ^ _ — o o o o o ^ LU ^OOCVICDO'^O CO O 8 N CL a . OTtOOi-T-CvlCVJCO CM

101,000 83,000 101,000 50,600 83,000 35,500 LexA 29,100 ♦ 50,600 20,900 R ecA 35,500 t 29,100

Figure 5.3 Western blot monitoring the effect of treatment with a DNA damage-inducing agent on RecA and LexA levels m M tuberculosis for 5 hours subsequent to exposure Cultures of M tuberculosis H37Rv were treated with a DNA damaging agent and samples collected over a 5-hour period. In addition, a single sample was removed after 28 hours. Cell-free extracts were analysed by Western blot with affmity-purifled M tuberculosis RecA antibody (a) or polyclonal M tuberculosis LexA antisera (b). A visual assessment of band intensities revealed that RecA and LexA levels were unaffected during this initial 5-hour period, although a decrease in LexA levels was reproducibly observed 28 hours subsequent to e^gosure. Positive and negative control lanes for monitoring RecA induction comprise of£. co//TG2{pEJ169(pGPl-2)} (Pg) and E. co//{TG2(pGPl-2)} (?„). A purified preparation of M mècrc«/o5/j LexA (?lexa) was included as a positive control in all LexA filters. Molecular weights are given in kDa.

101 Time (hours)

101,000 I 83,000

g 50,600

RecA -► — — ^ S 35,500 29,100

0 (0 ^ OjR P l EXA 101,000 w 83,000 50,600

35,500 LexA-► - _ jp 29,100

. # 20,900

. (M If) 00 1- TT M O c ) OCOt- i- t-OJOJCVJCO 101,000 83,000

65KDa 50,600 35,500

Figure 5.4 29,100 RecA induction and LexA cleavage in M tuberculosis

Cultures ofM. tuberculosis H37Rv were exposed to DNA damaging agents and samples collected over a 30-hour period. Cell-free extracts were analysed by Western blot with affinity-purified M. tuberculosis RecA antibody (a) or polyclonal M. tuberculosis LexA antisera {b). Examination of membranes revealed that RecA levels increase 24 hours subsequent to exposure(a) and that LexA protein levels decrease after 6 hours (b). An increase in LexA levels that was reproducibly observed at 24 hours may be due to an increase in lexA expression, which became apparent because cleavage was slow. No LexA cleavage products were detected, probably due to m vivo instability of the cleaved products. Positive and negative control lanes for RecA induction comprise £.coli TG2{pEJ169(pGPl-2)}

(Pr) and E. co//{TG2(pGPl-2)} (P„). PurifiedM. tuberculosis LexA (Plexa) was included as a positive control in all LexA filters. Monitoring levels o f the Mtuberculosis 65kDa antigen (c) provided visual corroboration that the observed kinetic profiles for LexA cleavage and RecA induction are a consequence of exposure to a DNA damaging agent and not of differential protein loading. Molecular weights are given in kDa.

102 Time (hours) ^ ^ rr- o c\j lO 1— Tf N O Z Œ OCVJCO t- t- ^ C V J C M

101,000 ■ 83,000 #W l 50,600 RecA 35,500 29,100

Figure 5.5 The effect of growth stage at the point of treatment with a DNA damage-inducing agent on the kinetic profile for RecA induction in M. tuberculosis H37Rv

To confirm that the delay in RecA induction was not a consequence of growth stage at the point o f induction,

M tuberculosis H37Rv cultures with an ODgoo^m the “logarithmic” portion of the growth curve (ODgoonm 0.2 - 0.4) were ejqjosed to DNA damaging agents. Cell-free extracts were analysed by Western blot with affinity- purified M tuberculosis RecA antibody. A visual assessment of band intensities revealed that RecA levels increased 6-10 hours subsequent to exposure. Although this differs from the results described previously where induction was not detected for 24 hours, it is still significantly slower than that seen in M smegmatis, which

implies that there is an intrinsic delay in RecA induction in M tuberculosis following genetic insult. The discordance between the rate of RecA induction in M tuberculosis cultures induced at an ODgoonm 0 6 and those

at 0 2 - 0 3 is probably due to the increased susceptibility o f M tuberculosis at low cell density. Cell-free extracts

o f £.coli TG2{pEJ169(pGPl-2)} (P r ) and E. co//{TG2(pGPl-2)} (? n ) were included as positive and negative

controls respectively. Molecular weights are given in kDa

103 a) Time (minutes) b) Time (minutes) o o o o o o o o o o o o o o OO OCM in O O '^ O CD CVJ OOOOJID CO O CO (N O C O C O < 3 > t- 1- T-C\JCO CO Tf O COCDCJ) i- t- 1- CM CO CO Tf 83,000 50,600 101,000 35,500 83,000 LexA 29,100 50,600 20,900

R ecA 35,500 I 29,100 20,900 é

Figure 5.6 Western blot monitoring the effect of treatment with a DNA damage-inducing agent on RecA and LexA levels inM microti OV1254 for 7 hours subsequent to exposure Cultures of M microti OV1254 were treated with a DNA damaging agent and samples collected over a 7-hour period. Cell-free extracts were analysed by Western blot with affinity-purified M tuberculosis RecA antibody (a) or polyclonal M tuberculosis LexA antisera (b). RecA and LexA levels were unaffected by exposure to DNA damaging agents during the course of this assay.

Positive and negative control lanes for monitoring RecA induction comprise of E. coli TG2{pEJ169(pGPl-2)} (P r ) and E. co//{TG2(pGPl-2)} (?„). Purified M tuberculosis LexA (P lexa ) was included as a positive control in all LexA filters. Molecular weights are given in kDa.

104 protein (figure 5.7a), which suggests that expression of the M microti recA gene is constitutive. Precursor RecA protein was not detected in any sample. Examination of membranes developed using an antibody raised against M. tuberculosis LexA revealed that LexA levels decline 6 - 7 hours subsequent to treatment with a DNA damage-inducing agent; probably as a consequence of LexA cleavage (figure 5.7b). Interestingly, this profile is remarkably similar to that of M tuberculosis (figure 5.4b). Hence, the increase in LexA levels at 22 hours may also be due to an increase in lexA expression, which became apparent because cleavage of the gene product was slow. Monitoring levels of the M tuberculosis 65kDa antigen (figures 5.7c) provided visual corroboration that the observed kinetic profiles for LexA cleavage and RecA induction are a consequence of exposure to a DNA damaging agent and not of differential protein loading.

5.4 The sensitivity ofM, tuberculosis, M. smegmatis and M microti to DNA damaging agents

Since M. tuberculosis and M smegmatis differ significantly in the rate at which the 80S response is induced, it was anticipated that this would be manifested as a difference in their sensitivity to DNA damaging agents. This was tested experimentally by determining the lowest concentration of ofloxacin and mitomycin C (the minimum inhibitory concentration) that would prevent growth of M tuberculosis, M. smegmatis and M. microti. In addition the effect of exposure to increasing doses of ultra-violet irradiation (254nm) on the viability of these three species was investigated.

5.4.1 The minimum inhibitory concentration of ofloxacin and mitomycin C in M tuberculosis, M. smegmatis and M. microti

The minimum inhibitory concentration (MIC) is the lowest drug concentration in vitro that will inhibit bacterial growth. In this study the DNA damaging agents used were ofloxacin and mitomycin C — agents that induce the E. coli 80S response in two distinct ways. Ofloxacin is a DNA gyrase inhibitor that exhibits a higher toxicity to mycobacteria than does nalidixic acid, the more frequently used gyrase inhibitor (8ugino et al, 1977). Inhibition of DNA gyrase introduces double strand breaks into DNA, which are repaired by the combined action of RecBCD and RecA (8nyder and Drlica, 1979; Chaudhury and 8mith, 1985). Mitomycin C introduces cross-links into DNA via the alkylation of guanines. Repair of these lesions is a 808 inducible response and has aspects of both post-replication and uvr excision repair (Iyer and 8zybalski, 1963; Teng et al, 1989; Borowy-Borowski et al, 1990; 8assanfar and Roberts, 1990; Kumar et al, 1992; Kohn et al, 1992).

The stability of both DNA damaging agents over the time course of this assay was checked by their continuing ability to induce RecA in M smegmatis compared to an unexposed control. A 1ml aliquot removed from a stock solution that had been incubated at 37°C for the requisite number of days was

105 Time (hours) a) ocplg ^ œ 101,000 83,000 # 50,600 RecA -► ^ I 35,500 1 29,100

m OOCN Tf M. o p, b) ocDi-'.-c\JCMC\JC0 r^LEXA 03,000 50,600 ^ 35,500 LexA ** ' 29,100

20,900

V m 00 CM N g c) OCO t- i-C\JC\JCs1CO 101,000 83,000 65KDa 50,600 35,500 29,100 Figure 5.7

RecA induction and LexA cleavage in M . m ic ro ti

Cultures ofM. microti OV1254 were exposed to DNA damaging agents and samples collected over a 30-hour period. Cell-free extracts were analysed by Western blot with affinity-purified M. tuberculosis RecA antibody (a) or polyclonal M. tuberculosis LexA antisera (6). A visual assessment of band intensities revealed that RecA levels are unaffected during this period (a) and that LexA protein levels decreased 6 - 7 hours subsequent to exposure (b). An increase in LexA that was reproducibly observed at 22 hours may be due to an increase in lexA expression, which became apparent because cleavage of the gene product was slow. No LexA cleavage products were detected, probably due to in vivo instability of the cleaved products. Positive and negative control lanes for RecA induction comprise E. coli TG2{pEJ169(pGPl-2)} (P,J and E. co//{TG2(pGPl-2)) (Pp,). Purified M. tuberculosis LexA

(Plexa) was included as a positive control in all LexA filters. Monitoring levels of the M. tuberculosis 65kDa antigen (c) provided visual corroboration that the observed kinetic profiles for LexA cleavage and RecA induction are a consequence of exposure to a DNA damaging agent and not of differential protein loading.

106 diluted to a final concentration of l|o,g/ml (ofloxacin) or 2-5|xg/ml (mitomyc^in C) in an M. smegmatis culture grown to an OD^oonm of 0 6. Following incubation for 5 hours at 37°C a cell-free extract was produced and RecA levels analysed by Western blot as described above.

For determination of the MIC’s, 10ml aliquots ofM tuberculosis, M microtii ox M. smegmatis containing approximately 10^ colony forming units (CFUs) were incubated at 37°C with doubling dilutions of ofloxacin (10 - 0 04|Lig/ml) or mitomycin C (50 - 0 00315pg/ml) for 10 dayrs, 5 days or 24 hours respectively. The turbidity of each aliquot was visually assessed by compariison with a control culture grown in the absence of DNA damaging agents. The results illustrated in tables 5.1 and 5.2 are a compilation of at least 4 assays. It was anticipated that the difference in the irate at which RecA is induced in M. tuberculosis and M. smegmatis and the high constitutive expression o f RecA in M. microti would be reflected by their comparative sensitivities to each agent. This proved true for mitomycin C; a final concentration of 3pg/ml prevents growth of M smegmatis and concentrations inhibitory to M. microti are within two dilutions (0-78pg/ml). However, M. tuberculosis is much more sensitive to mitomycin C, a concentration as low as 25ng/ml prevented growth of this species. Curiously the presence of 2 5pg/ml ofloxacin prevented growth of M. smegmatis and M. microti and concentrati ons inhibitory to M. tuberculosis were within one dilution (1 25pg/ml). This latter observation is in agreement with published MIC data for quinolones andM tuberculosis (Fenlon and Cynamon, 1986; Leysen et al, 1989;Garcia- Rodriguez et al, 1993). The reasons for this similarity are not entirely clear although it has been reported that slow-growing mycobacteria are less susceptible to quinolones that fast-growing species (Garcia- Rodriguez et al, 1993).

5.4.2 The effect of exposure to ultra-violet (UV) irradiation on the viability ofM. tuberculosis, M. smegmatis and M. microti

Cultures of M. tuberculosis and M. smegmatis were grown to an ODgoonm of 0-6 and M. microti to an ODgoonm of 0 35 under the optimal conditions described in section 2.1. Bacteria were resuspended in sterile magnesium sulphate and 1ml aliquots exposed to increasing amounts of UV-light from a calibrated light box (254nm). Samples were diluted 10’^ - lO'^-fold in sterile magnesium sulphate and plated on 7H11 agar plates. After incubation in the dark at 37°C for: M. smegmatis 3 — 4 days, M. tuberculosis 10 - 15 days or M. microti 5 - 6 weeks, the number of CFUs at each exposure was determined and expressed as a fraction of the number of CFUs of an unexposed control. Each curve illustrated in figure 5.8 is a mean of at least 3 assays. Increasing the exposure time to UV irradiation (254nm) led to a marked decrease in the number of CFUs for M. tuberculosis compared to M. microti and M smegmatis'. exposure to 40J/m^UV irradiation (254nm) caused a one-log decrease in the number of CFUs compared to an unexposed control. Interestingly both M. tuberculosis and M. microti exhibit a two-log decrease following exposure to approximately 80J/m^ UV irradiation; by comparison the viability of M. smegmatis decreased only one-log over the course of the assay. The increased sensitivity of M. tuberculosis to UV irradiation lends further support to the premise that the difference in the rate at which the SOS response is

107 Ofloxacin (|.ig nil f.c.) H B 9I M tuberculosis --- + + ++ +++ +++

M smegmatis --- + + ++ +++ +++ +++

M. microti + + ++ +++ +++ +++

Table 5.1 The minimum inhibitory concentration of ofloxacin for M. tuberculosis H37Rv, M. smegmatis mc^l55 and M. microti OV1254 Ten ml aliquots ofM. tuherculosis. M microti or M. smegmatis containing approximately 10 ’ colony forming units (CFUs) were incubated at 37°C with doubling dilutions of ofloxacin ( 10-0.04pg/m l f.c.) for 10 days, 5 days or 24 hours respectively. The turbidity of each aliquot was visually assessed by comparison with a control culture grown in the absence of DNA damaging agent and graded as: no growth (-), poor growth (+), good growth (++) or excellent growth (+++). The results shown above are a compilation of at least 4 assays. The presence of 2.5pg'ml (f.c.) ofloxacin prevents growth ofM. smegmatis and M microti and concentrations inhibitory to Mtuberculosis are within one dilution ( 1.25pg'ml f.c.).

108 n m HP ■■mg H i■omi AT tuberculosis ------+ ++ ++

AT. smegmatis ----- + + + ++ ++ +++ +++ +++ +++ +++

AT. microti + + + ++ ++ ++ +++ +++

Table 5.2 The minimum inhibitory concentration of mitomycin C for M. tuherculosis H37Rv, M. smegmatis me" 155 and M. microti OV 1254 Ten ml aliquots of A/, tuberculosis. M microti or AT smegmatis containing approximately 10^ colony forming units (CFUs) were incubated at 37”C with doubling dilutions of mitomycin C (50 — 0.0031 Spg/ml f.c.) for 10 days, 5 days or 24 hours respectively. The turbidity of each aliquot was visually assessed by comparison with a control culture grown in the absence of DNA damaging agent and graded as: no growth (-), poor growth (+), good growth (++) or excellent growth (+++). The results shown above are a compilation of at least 4 assays. The presence of 3.13 pg/ml (f.c.)

mitomycin C was inhibitory to cultures of ATsmegmatis; concentrations inhibitory to ;VT.microti are within two dilutions (0.78pg'ml). By comparison, 25ng/ml mitomycin C has a pronounced effect on the growth of ATtuherculosis.

109 10 M. microti M. smegmatis I M. tut’ercîdosis 1

o

01 5 "O u0) CQ 001 0r M § 0 001 1

0 0001 0 20 4 0 60 80 100

Dose of UV UTadiation

Figure 5.8 The effect of exposure to ultra-violet irradiation (254nm) on the viability of M. tuherculosis, M. smegmatis and M. microti Mycobacterial cultures grown to the required OD^xinm (M. tuherculosis and M smegmatis to an OD(,,x,nm of 0-6; M. microti to 0-35) were haix estcd, resuspcndcd in sterile magnesium sulphate and I ml aliquots exposed to increasing amounts of UV light from a calibrated light box (254nm). These samples were diluted 10'^- KT^-fold and plated on 7H11 agar plates. After incubation in the dark at 37°C for 3-42 days, depending upon species, the number of CFUs at each exposure was determined and expressed as a fraction of the number of CFUs of an unexposcd control. The results shown above are a mean of at least 3 assays. Increasing the exposure time to UV irradiation (254nm) led to a marked decrease in the number of CFUs for M. tuherculosis compared to M. microti and M. smegmatis: exposure to 40J/nr UV irradiation (254nm) caused a one-log decrease compared to an unexposcd control. Interestingly bothM. tuherculosis and M. microti exhibit a two-log decrease following exposure to approximately 80J/nr UV iiradiation; by comparison the viability of AT smegmatis decreased only one-log over the course of the assay.

10 induced in M. tuberculosis and M. smegmatis is manifested as a difference in their sensitivity to DNA damaging agents. It is probable that the high constitutive levels of RecA in M. microti decrease its susceptibility to DNA damage caused by UV irradiation at < 50J/m^. However, at higher UV doses the amount of RecA present is insufficient to deal with the increasing proportion of lesions, which results in an increased sensitivity to UV doses > 50J/m^.

5.5 Discussion

The recA genes of M. tuberculosis and M. smegmatis were induced following exposure to DNA damaging agents that generate the inducing signal in two distinct ways. Ofloxacin is a DNA gyrase inhibitor. Binding of this dmg stabilises the gyrase complex, preventing enzyme turnover and causing DNA cleavage in the form of double strand breaks (Sugino et al, 1977; Snyder and Drlica, 1979; Chaudhury and Smith, 1985). Processing of these lesions by RecBC helicase provides ssDNA, which activates RecA. Mitomycin C is a bifunctional alkylating agent that demonstrates a propensity to bind covalently to DNA in a sequence specific manner. Structural studies have indicated that it preferentially alkylates the position of 5'-CG-3' guanines leading to the formation of DNA monoadducts and cross­ links (Iyer and Szybalski, 1963; 1964; Teng et al, 1989; Borowy-Borowski et al, 1990; Kumar et al, 1992). This primary DNA damage is processed by uvr excision repair, again yielding ssDNA resulting in RecA activation. The time required to detect induction was protracted in M. tuberculosis (24 hours) compared with M. smegmatis (1 -3 hours) suggesting that induction of RecA in M. tuberculosis is intrinsically delayed. Movahedzadeh et al (1997b) and Papavinasasundaram et al (1997) have recently provided confirmation of these observations using transcriptional fusions to the reporter gene chloramphenicol-acetyltransferase (CAT) to investigate the inducibility of the recA promoter regions of both M. tuberculosis and M. smegmatis.

RecA was not induced in M. microti following treatment with a DNA damage-inducing agent. The high RecA levels found in uninduced cell-free extracts of this species remained constant over the time course of the assay which suggests that M. microti recA is expressed constitutively. This implies that there is no repression of the recA gene in the absence of DNA damaging agents (Little and Mount, 1982; Walker, 1984). Alternatively, the cleaved LexA fragments may still be capable of repressing RecA expression, although this is perhaps less likely. The simplest explanation for a lack of repression would be the absence of a repressor molecule. However, LexA protein was detected in M microti. Nevertheless, the presence of a mutation affecting binding of the repressor could also give rise to the observed phenotype. A mutation in the operator region would affect the ability of the repressor to bind. An operator constitutive mutation has been characterised in E. coli; the recAo mutation is a dominant allele causing constitutive recA expression (Clark, 1982; Ginsburg et al, 1982; Little and Mount, 1982). A defect in the repressor molecule affecting its ability to bind to the operator could also explain this result. In E. coli, a mutation has been characterised that results in the constitutive expression of recA and other lexA target

111 genes; termed the lexA{D&ï) mutation this inactivates or prevents the synthesis of LexA (Mount, 1977; Pacelli et al, 1979;Little and Mount, 1982; Krueger et al, 1983).

The demonstration of LexA cleavage in M tuberculosis, M. microti and M. smegmatis provides compelling evidence for the activation of RecA by an inducing signal. Levels of LexA protein decline much more slowly in M. tuberculosis and M. microti compared to M. smegmatis. A visual inspection of band intensities indicates that there is no alteration in basal LexA levels for approximately 6 - 7 hours subsequent to exposure of M. tuberculosis and M. microti to a DNA damaging agent whereas a significant decrease is seen in M. smegmatis within 40 minutes. By comparison, cleavage of LexA in E. coli has been reported to occur 30-90 seconds after exposure to DNA damaging agents and RecA induction within 10 minutes (Little, 1983. Sanssafar and Roberts, 1990). In B. subtilis, DinR repressor cleavage occurs within 10 minutes of treatment with a DNA damage-inducing agent (Wojciechowski et al, 1991 ; Lovett et al, 1994) and RecA induction within 30 minutes (Lovett et al, 1988; 1994; Raymond-Denise and Guillen, 1992; Miller et al, 1996). This observation is consistent with the reduced rate of repressor inactivation reported for B. subtilis compared to its counterpart in E. coli (Lovett et al, 1988; Sassanfar and Roberts, 1990; Lovett et al, 1994). Interestingly the delay observed for RecA induction in M. smegmatis as a result of treatment with a DNA gyrase inhibitor (3 hours) compared to mitomycin (1 hour) has also been reported in E. coli and B. subtilis (Sanssafar and Roberts, 1990; Wojciechowski et al, 1991; Lovett et al, 1994). This may be due to the different mechanism by which each agent generates the inducing signal; the absence of a similar difference in M. tuberculosis may be a reflection of the intrinsic delay seen in this species.

Because M. tuberculosis and M smegmatis differ significantly in the rate at which the SOS response is induced, one might anticipate that this would be manifested as a difference in their sensitivity to DNA damage-inducing agents. This expectation proved tme. Examination of the relative sensitivities of M. tuberculosis, M. microti and M. smegmatis to ultra-violet irradiation (UV), mitomycin C and ofloxacin revealed an increased sensitivity of M. tuberculosis to UV irradiation and mitomycin C compared to M microti and M. smegmatis. Interestingly, there is no significant difference between any of the species examined following exposure to ofloxacin. It has been firmly established that treatments that damage DNA or inhibit replication will generate an inducing signal, which may be an increase in ssDNA regions (Walker, 1984; D ’Ari, 1985; Higashitani et al, 1992), leading to RecA activation and the concomitant expression of SOS genes. Therefore, it is probable that the sensitivity of M. tuberculosis to DNA damaging agents is a consequence of both the mechanism by which DNA damage is produced and the ability of M. tuberculosis to repair the ensuing damage.

Mitomycin C covalently binds DNA leading to inter-strand cross-links which constitute a block to replication (Iyer and Szybalski, 1963; 1964; Teng et al, 1989; Borowy-Borowski et al, 1990; Kumar et al, 1992; Basak, 1996). Repair of these lesions requires both RecA and UvrABC nuclease. The lesion is excised by the action of UvrABC (section 1.5.1; Kohn et al, 1992) and RecA polymerises on the ensuing ssDNA to form a nucleoprotein filament, initiating recombination repair (section 1.6.2). Binding of RecA

112 to ssDNA reversibly activates its coprotease function leading to LexA cleavage and concomitant induction of the SOS response (Sassanfar and Roberts, 1990; Lovett et al, 1994). Similarly, UV irradiation introduces intra-strand lesions into the DNA duplex. Single-stranded DNA regions are generated as a result of discontinuous replication. Binding of RecA to these regions and the concomitant induction of the SOS response lead to repair of these lesions by uvr excision repair (section 1.5.1). The protracted time for RecA induction in M. tuberculosis would make this species inherently more susceptible to these agents compared to M. smegmatis. It is probable that the elevated levels of RecA found in M. microti decrease its susceptibility to DNA damage-inducing treatments.

Ofloxacin is a DNA gyrase inhibitor (Sugino et al, 1977). Inhibition of its supercoiling activity causes double strand breaks (Snyder and Drlica, 1979). The broken ends provide a substrate for RecBCD (Exonuclease V) which moves along the damaged duplex unwinding DNA in the process. On reaching a correctly orientated “chi-site” it cuts one strand producing a stretch of ssDNA leading to the binding of RecA (Walker, 1984; Sassanfar and Roberts, 1990; Taylor and Smith, 1995). The absolute requirement for RecBCD in the repair of these lesions and the concomitant induction of RecA has been demonstrated (Gudas and Pardee, 1975; 1976; Chaudhury and Smith, 1985; Phillips et al, 1987; Piddock and Wise, 1987; Ysem et al, 1990). Considering the significant differences in the response of M. tuberculosis, M. microti and M. smegmatis to DNA damage it is curious that the sensitivities of all three species to this compound are comparable; the reasons for this are not clear. However, it is interesting to note that concentrations bactericidal to mycobacteria in these investigations correlate well with the concentration required to give maximum induction (Movahedzadeh et al, 1997b). Garcia-Rodriguez et al (1993) has reported that slow growers, including M tuberculosis, are less susceptible than fast growing species. The in vitro activity of quinolones against mycobacteria can be affected by a number of factors including the culture medium and incubation time. Mycobacteria for this assay were cultivated under the optimal conditions described in section 2.1. The different types of media used may have some bearing on the observed results since loss of activity due to binding to albumin and other proteins has been reported (Garcia-Rodriguez e/a/, 1993).

The key regulatory elements of the SOS response in M. tuberculosis and M. smegmatis have been conserved. On the basis of the results described so far it seems reasonable to assume that regulation of the SOS response in these species is analogous to that found in E. coli and B. subtilis. However, there appears to be a significant difference in the kinetics of the mycobacterial SOS response. It is not known whether the difference in response time is due to differences in the rate at which RecA is activated, differences in the rate of repressor inactivation or to some as yet unidentified host factor.

The E. coli LexA repressor is inactivated both in vivo and in vitro at physiological pH by a specific cleavage reaction requiring the presence of RecA, ssDNA and a nucleotide triphosphate (Little et al, 1984; Walker, 1984). RecA independent cleavage also occurs in vitro at alkaline pH (autodigestion), yielding fragments identical to those reported for RecA-mediated cleavage (Little, 1984). Both reactions cleave a specific peptide bond between Ala-84 and Gly-85 via the action of residues Ser-119 and Lys-156

113 (Horii et al, 1981c; Slilaty and Little, 1987). In principle, the rate of such self-processing reactions can be high due to the presence of the required reactants at high, local concentrations. However, this is not always the case. In the example cited here the rate of LexA cleavage is generally slow until stimulated by an external agent; which implies that the rate of this self-processing reaction is slow but capable of significant increases. Mutational studies using a repressor with an increased rate of in vivo cleavage revealed that LexA cleavage is not an intrinsically slow reaction but rather its rate has evolved to an optimum value (Smith et al, 1991; Roland et al, 1992). Too rapid a rate would prevent repressor accumulation, particularly in slow-growing or stationary phase cells. Conversely, too slow a rate would make prompt inactivation of the repressor upon exposure to an inducing treatment difficult, thereby delaying or preventing an effective response. Miller et al (1996) proposed that the delay in induction of the SOS response in B. subtilis compared to E. coli is due to a difference in the rate of repressor inactivation. It is possible that a similar difference may account for the intrinsic delay in induction of the SOS response in mycobacteria. The M. tuberculosis lexA gene has recently been sequenced and the residues identified as being important for cleavage found to be conserved (Movahedzadeh et al, 1997a), which suggests that there is a common mechanism of proteolytic cleavage. No conclusive information is available on the kinetics of M. tuberculosis LexA cleavage in vitro. However, preliminary investigations suggest that the rate of autodigestion is slow compared to E. coli LexA (Dr. P. Jenner, N.I.M.R. unpublished results) which tends to support the suggestion that a slow rate of repressor cleavage leads to an intrinsic delay in induction of the SOS response.

A second explanation that may account for the delay in induction of the SOS response in mycobacteria is the presence of a defect in an accessory protein; possible candidates include RecF, RecO, RecR and single-strand binding proteins (SSB). In E. coli, the RecA protein promotes homologous pairing and strand exchange in the presence of ssDNA and a nucleotide triphosphate. Umezu et al (1993) reported that the RecF, RecO, RecR and SSB proteins facilitate the loading of RecA onto the initiating single strand. Further genetic evidence has shown that RecF, RecO and RecR act at presynaptic steps in recombination (Lloyd et al, 1988; Wang et al, 1993; Kowalczykowski et al, 1994) and that mutations in the recF gene cause defects in recombinational repair and a delay in induction of the SOS response (Thoms and Wackemagel, 1987; Madiraju and Clark, 1991; Simic et al, 1991; Clark and Sandler, 1994), an intriguing phenotype considering the intrinsic delay observed in induction of the SOS response in M tuberculosis and its recalcitrance to undergo homologous recombination (Kalpana et al, 1991). It has been suggested that RecF assists in the cleavage of LexA by aiding RecA activation (Volkert and Hartke, 1984; Volkert et al, 1984; Walker, 1985). More recent studies have revealed that RecF, in conjunction with RecR, preferentially interacts with dsDNA containing ssDNA regions (i.e. gapped DNA; Griffin and Kolodner, 1990; Hedge et al, 1996; Courcelle et al, 1997). In addition, there are several lines of evidence which show that RecO and RecR are also involved in RecA activation via interaction with the recF gene product (Lloyd et al, 1988; Sawitzke and Stahl; 1992; Wang et al, 1993). Mutational studies have implicated RecO and RecR in recombinational repair (Clark, 1991) and have shown that defects in these genes can delay induction of the SOS response (Hedge et al, 1995; Whitby and Lloyd, 1995). In addition.

114 RecR and RecO have been shown to alleviate the inhibitory effects of SSB proteins on RecA catalysed joint molecule formation in vitro (Umezu et al, 1993; 1994; Umezu and Kolodner, 1994). During normal DNA replication RecA cannot assemble its filament on SSB-bound ssDNA at replication forks. This behaviour is paralleled in vitro at low magnesium concentrations for RecA in the presence of SSB proteins and ssDNA (Kowalczykowski and Krupp, 1987; Lavery and Kowalczykowski, 1990). Conversely, in vitro observations indicate that RecA filament formation on SSB-bound ssDNA does occur at elevated magnesium and ATP concentrations (Barbe et al, 1983; Guerrero et al, 1985; Thresher et al, 1988); inhibition of DNA synthesis in vivo may produce a comparable situation. It has been proposed that modification of the SSB proteins or an increased tendency for RecA filament formation at an elevated ionic concentration may account for RecA polymerisation on SSB-bound ssDNA (Kuzminov, 1995). In addition to the inhibitory effect of SSB proteins, mutational studies have shown that they also have a positive role in RecA activation (Vales et al, 1980; Whittier and Chase, 1981; Lieberman and Witkin, 1983; Chase et al, 1983). Although the precise nature of this involvement remains unclear ssb mutants that fail to induce normal levels of RecA after treatment with an inducing agent have been characterised. Interestingly a similar phenotype exists in cells which overproduce wild type SSB proteins (Johnson, 1977; Sevastopoulos et al, 1977; Meyer et al, 1982; Carlini and Porter, 1997). The majority of ssb mutations, for example 55Z>-1, are attributed to their relative ssDNA-binding affinity, an exception is ssb-l 13 which is thought to involve altered protein-protein associations (Carlini and Porter, 1997). Additional evidence has revealed that SSB proteins also participate in at least two steps of genetic recombination, i.e. strand assimilation and branch migration (McBntee et al, 1980; West et al, 1982; Lieberman and Witkin, 1983). A stoichiometric interaction between RecA, SSB and ssDNA that promotes pairing with dsDNA has been demonstrated (Flory and Radding, 1982) which raises the possibility that when RecA is present in limiting amounts helix-destabilising proteins such as SSB may play an auxiliary role in the pairing reaction (McEntee et al, 1980; Shibata et al, 1980). Basal RecA levels in M. tuberculosis are low but it is not clear whether this is a rate limiting factor. However, the potential influence of SSB proteins on the rate of SOS induction in this genus is certainly intriguing, particularly since the partial complementation of M. tuberculosis RecA in E. coli recA mutants may be due to its inability to displace SSB proteins (Davis et al, 1991; Kumar et al, 1996).

A third explanation to account for the slow induction of the SOS response in mycobacteria is related to the presence of the thick, lipid rich cell wall. The fact that M. tuberculosis was significantly more sensitive to mitomycin C than M. smegmatis demonstrates that this agent does enter the cell of M. tuberculosis. This suggests that a lack of penetration cannot be the explanation for the slow induction seen in M tuberculosis although the mycobacterial envelope may provide a formidable first line of defence. The increased resistance to deleterious agents such as acids and alkalis does support this concept (Goren and Brerman, 1979; Breiman and Nikaido, 1995). However, examination of available data reveals that the contribution that the mycobacterial envelope may make should not be accepted at face value. The mycobacterial envelope consists of the plasma membrane surrounded by a complex wall of carbohydrate and lipid. A capsule composed of polysaccharide and protein with relatively small quantities of lipid surrounds this. The basic structure of the plasma membrane does not differ significantly from that of

115 other bacteria (Goldman, 1970; Kumar et al, 1979; Minnikin, 1982; Silva and Macedo, 1984). The composition of the cell wall is equally well defined; stmctures such as arabinogalactan are common to both fast and slow growers but some inter-species differences between fast and slow growers have been reported (Asselineau and Lederer, 1950; Kanetsuna, 1968). For example, type II or a ' mycolates are predominantly found in fast growers (Davidson et al, 1982; Miimikin et al, 1984; Hinrikson and Pfyffer, 1994; Brennan and Nikaido, 1995). At first glance a difference such as this appears to support the suggestion that the mycobacterial envelope plays a role in delaying entry of DNA damaging agents into the cell. However, the presence of a substantial proportion of a' mycolates decreases fluidity, and hence permeability of the cell wall. It is anticipated that this decrease in permeability would affect the entry of ofloxacin and mitomycin C into the cell since these agents are lipophilic and presumed to enter the cell by traversing the lipid bilayer. Hence, it would be anticipated that induction of the SOS response in M. smegmatis would be delayed compared to M. tuberculosis', this is not the case. One component of the mycobacterial envelope that may play a significant role in decreasing the permeability of the mycobacterial envelope is the capsule. This has been identified as an electron transparent zone surrounding mycobacteria grown both in vitro and in vivo although its precise stmcture and fimction is not clear (Hanks, 1961; Draper and Rees, 1973; Rastogi et al, 1986). The role of the mycobacterial envelope as a permeability barrier which influences the rate at which DNA damaging agents gain access to the cell and hence induce the SOS response remains unclear. Determination of the precise point at which DNA damage occurs following treatment with an inducing agent may provide further insight. Another interesting possibility is that the rate of DNA replication may play a key role in determining the kinetics of the SOS response. A fundamental taxonomic division of the genus Mycobacterium is tied to their growth rates under optimal conditions in vitro. Rapid growers yield visible colonies within 7 days whereas slow growers require 2-6 weeks. Slow growth rate is thought to be a reflection of a retarded rate of macromolecular synthesis; the rate of RNA elongation in M. tuberculosis is 10-fold slower than that for E. coli (Harshey and Ramakrishnan, 1977). Similarly, the rate of DNA chain elongation in M. tuberculosis is approximately 11-fold slower than that of M. smegmatis and 13- to 18-fold slower than E. coli (Hiriyarma and Ramakrishnan, 1986). When DNA replication is perturbed in E. coli an inducing signal is generated (Casaregola et al, 1982b). The precise nature of this signal is unknown, possible candidates include ssDNA (D’Ari, 1985; Sassanfar and Roberts, 1990; Higashitani et al, 1995; Morel- Deville et al, 1996; Shinagawa, 1996), ATP (Barbe et al, 1983; Guerrero et al, 1984; Thresher et al, 1988) and deoxyribonucleotides (Phizicky and Roberts, 1981; Menetski and Kowalczykowski, 1989). Preliminary investigations using a wider spectrum of mycobacteria suggest that an intrinsic delay in RecA induction may also occur in the slow growers M. kansasii and M avium (data not shown). Although these observations are far from conclusive at this stage it does seem likely that the rate of DNA replication may play a key role in determining the kinetics of the SOS response. Perhaps by determining the time taken for DNA damage to manifest as ssDNA or as a result of an accumulation of unused deoxyribonucleotides.

Alternatively, the intrinsic delay in RecA induction in M. tuberculosis may be a result of a difference in promoter strength between M. tuberculosis and M. smegmatis. Such a mutation has been characterised in

116 E. coli (recA453; Clark and Volkert, 1978; Casaregola et al, 1982a) which results in a defect in both recombination and induction of the SOS response. However, treatment of transcriptional fusions of the M tuberculosis and M. smegmatis recA promoter regions to the chloramphenicol-acetyltransferase (CAT) reporter gene with DNA damage-inducing agents resulted in a 2 5- fold increase in expression from both putative promoter sites (Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). This suggests that there is no significant difference in intrinsic promoter strength between M tuberculosis and M smegmatis recA. Clearly a more rigorous examination of these putative promoter sites is required to confirm their role in expression of the recA gene. This may provide further insights into the role that the regulation of recA expression may play in the intrinsic delay observed for RecA induction in M tuberculosis (see chapter 8 for a more detailed discussion).

The essential regulatory elements of the SOS response have been conserved. The evidence presented in this chapter suggests that treatment with a DNA damaging agent generates an inducing signal which activates RecA leading to LexA cleavage and RecA induction in both M tuberculosis and M. smegmatis. However, there is a significant difference in the response time between these two species as well as in comparison to both E. coli and B. subtilis. At present it is not known whether this intrinsic delay is due to differences in the rate at which RecA is activated, a difference in the rate of repressor inactivation or to some as yet unidentified factor. Curiously the high levels of RecA found in uninduced cell-free extracts of M microti are unaffected by treatment with a DNA damaging agent, although it is clear that RecA is activated by an inducing signal since the levels of LexA decline following exposure. Interestingly, the kinetics of this cleavage reaction are very similar to those of M. tuberculosis. This may have been anticipated since both are members of the M. tuberculosis complex and have been reported to share a high degree of genetic homogeneity (Imaeda, 1982; Kempsell et al, 1992; Frothingham et al, 1994; Takewaki et al, 1994). The evidence suggests that the M. microti recA gene is expressed constitutively, which implies that there is no repression of the recA gene in the absence of DNA damaging agents (Little and Mount, 1982; Walker, 1984). Since M microti clearly possesses a LexA homologue the presence of a mutation - either in the operator region or in the protein itself - affecting binding of the repressor will be investigated in subsequent chapters. Cell-free extracts of both M tuberculosis andM microti were examined for the presence of precursor RecA protein. Interestingly, it was not detected in any sample following treatment with a DNA damage-inducmg agent, even in the presence of induced RecA levels. This suggests that the appearance of splicing in the maturation of RecA is not a means of regulating its expression, this will be discussed further in chapter 8.

117 6.0 Determination of whether the M m ic r o t i r e cA g e n e

POSSESSES AN INDUCIBLE PROMOTER

6.1 Introduction

RecA was not induced in M. microti following treatment with a DNA damage-inducing agent. This implies that there is no repression of the recA gene in the absence of DNA damaging agents (Little and Mount, 1982; Walker, 1984) or possibly that the cleaved LexA fragments may still be capable of repressing recA expression. The simplest explanation for a lack of repression would be the absence of a repressor molecule. However, examination ofM microti cell-free extracts by Western blot clearly show that a LexA homologue is present. Moreover this protein exhibits a marked decrease following exposure to DNA damaging agents, which suggests that RecA is activated by an inducing signal and possesses a functional co-protease activity. Nevertheless, the presence of a mutation in the repressor molecule affecting binding to the operator region could also give rise to the observed phenotype. For example, E. coli /ex(Def) mutants have been characterised which result in the constitutive expression of LexA repressed genes (Mount, 1977; Pacelli et al, 1979; Little and Mount, 1982; Krueger et al, 1983; Thliveris and Mount, 1992; Haijema et al, 1996). Alternatively there may be a mutation in the operator region; E. coli recAo mutants have been characterised which constitutively express the recA gene (Clark, 1982; Ginsburg et al, 1982; Little and Mount, 1982; Thliveris and Mount, 1992; Winterling et al, 1997).

To assess whether some feature of the sequence upstream of the M. microti recA gene is responsible for its constitutive phenotype the recA gene was cloned from genomic DNA together with 700bp of upstream sequence. Transcriptional fusions of the putative M microti recA operator / promoter region to the reporter gene chloramphenicol acetyltransferase (CAT) were prepared and expression of CAT assayed in M. smegmatis following DNA damage. Concurrently, the M microti recA gene and 700bp of flanking upstream DNA were sequenced to confirm the presence or absence of a SOS-like box in the putative promoter region.

The work in this chapter describes: i) Cloning of the M. microti recA gene ii) The preparation of constructs to determine whether M. microti recA possesses an inducible promoter in its upstream sequence iii) Investigations to determine if expression of M. microti recA is inducible in M. smegmatis iv) Sequence determination of the M microti recA gene and 700bp of upstream sequence.

118 6.2 Cloning of theM. microti recA coding region plus regulatory sequences in pBluescript IIKS

Southern hybridisation with probes derived from M. tuberculosis recA (section 2.10) demonstrated that many M. microti restriction fragments containing recA are common to both species. Based on knowledge of the M. tuberculosis recA gene and flanking nucleotide sequence (Davis et al, 1991; Movahedzadeh, 1996) a 3.1kb Not\-Nae\ digest, expected to include the whole of the recA coding region plus the regulatory sequence, was selected for cloning. Size-fractionated, Not\-Nae\ digested genomic M. microti DNA was purified using a Geneclean kit, ligated into the Notl-EcoRV sites of pBluescript II KS and transformed in E. coli DH5a as described (section 2.9). A single positive clone was identified (pNTmicl) from twenty-two recombinant plasmids screened by Southern hybridisation using a probe derived from M. tuberculosis recA. The identity of this clone was verified by random sequencing of the insert using primer NTTBl (5’-AGG CGC TGC GGA AAA TGA-3’), the M13 reverse primer (5’- GGA AAC AGC TAT GAC CAT G -3 ’) and the M l3 -20 primer (5 - GTA AAA CGA CGG CCA GT - 3’). Primer NTTBl is based upon sequence 521bp downstream of the transcription start site of M. tuberculosis recA and the M13 reverse and -20 primers are located either side of the 3.1 kb insert. The clone was positively identified by alignment with equivalent sequences in M. tuberculosis.

6.3 Confirmation of the presence of an inducible promoter for theM microti recA gene

In order to test for the inducibility of the M. microti recA gene, the relevant fragment was cloned in a plasmid containing a promoterless CAT gene (Movahedzadeh et al, 1997b).

6.3.1 Preparation of constructs

A 0.7kb Sacl-Xhol fragment of pNTmicl was purified as described (section 2.9.4), ligated into the Sacl-

SaE sites of pUC19 (to give pNTmic2) and transformed into E. coli DH5a. Recombinant plasmids were selected on the basis of blue-white colony screening (section 2.9.6). Plasmid DNA was harvested using a QIAGEN miniprep kit and the following constructs prepared; a 0.7kb Notl-Hindlll fragment ligated into the Notl-HindlU sites of pEJ257 (Movahedzadeh et al, 1997b) to give pNT2 and a 0.35kb PvuW-HindiW fragment ligated into the EcoK^-Hindlll sites of pEJ257 to give pNT12 (figure 6.1). After transformation into E. coli DH5 a plasmid DNA was prepared using a QIAGEN miniprep kit and electroporated into M. smegmatis mc^l55 (see section 2.9.7). Constmcts pFM7, pFM17 and pEJ257 (Movahedzadeh et al, 1997b) were electroporated concurrently with pNT2 and pNT12 for use as positive and negative controls.

119 (kb scale) > a I I pNTmid

U I RecA (kb scale) Ii II pNTmic2

pNT2 CAT

pNT12 CAT

Upstream sequence

g recA gene

pBluescriptll KS (multi cloning site)

I I pUC19 (polylinker cloning site)

CAT reporter gene

Figure 6.1 The preparation of constructs to determine whether M. microti recA possesses an inducihle promoter in its upstream sequence

120 6.3.2 CAT assays of mycobacterial transformants containing DNA from upstream of the recA gene

The effect of DNA damaging agents on the expression of constructs pNT2 and pNT12 was examined as described in section 2.11. Cell-free extracts were prepared ffomM smegmatis mc^l55 harbouring the parental vector (pEJ257) or recombinant clones (pNT2, pNT12, pFM7 and pFM17) and both uninduced and induced cultures assayed for CAT expression and total protein content. The results shown in figure 6.2 are the means of 4 - 10 replicate cultures and clearly show that the M. microti recA gene possesses an inducible promoter. By comparison the level of CAT present in cell-free extracts prepared from the parental strain M smegmatis mc^l55 harbouring plasmid pEJ257 was very low in both the induced and uninduced cultures (< 1 5ng CAT / mg total protein).

Following treatment of M. smegmatis harbouring plasmid pNT2 or pNT12 with a DNA-damage inducing agent the level of CAT in cell-free extracts (ng / mg total protein) is significantly increased (matched- pairs test at a 90% confidence interval). Expression from the recA promoter in plasmid pNT12 (0-35kb insert) was increased 4.8-fold following exposure to mitomycin C and 2-fold for ofloxacin. The increase in expression from the recA promoter in construct pNT2 (0 7kb insert) was approximately 2-3-fold for both DNA damage-inducing treatments. Treatment of M smegmatis harbouring equivalent M. tuberculosis constructs (pFM7 and pFM17, see section 2.1) resulted in a 3-4- to 3- 6 -fold increase in expression following exposure to mitomycin C and a 1-8- to 1-9-fold increase for ofloxacin. There was no difference in induction with respect to construct length in M. tuberculosis.

6.4 Sequence determination of the M microti recA gene and 700bp of the upstream region

Since M. microti clearly possesses a LexA homologue and an inducible recA gene, we examined whether some feature of the upstream sequence could be responsible for its constitutive phenotype. The 3.1 kb 7/o//-A/he/fragment ligated into pBluescript II KS (pNTmicl) was sequenced to determine whether or not a SOS type box was present upstream of the M microti recA gene. The initial sequencing reactions were carried out using the M13 reverse and -20 primers located either side of the 3.1kb insert (S’- GGA AAC AGC TAT GAC CAT G -3 ’ and S’- GTA AAA CGA CGG CCA G T -3 ’ respectively). Subsequent primers were designed based on the sequence obtained enabling determination of the sequence on both strands of the entire fragment (figure 6.3).

V oqM microti and M tuberculosis recA genes have 100% DNA-DNA relatedness. Homology at the DNA level extends 700bp upstream of the recv4-coding region; only two nucleotide changes were evident, both are insertions in M. microti (one S26bp and the other 478bp upstream of the recA gene, see figure 6.4). A motif homologous to the B. subtilis Cheo box (GAAC-N^-GTTC) and identical to that of the

121 Ofloxacin: uninduced CAT level Ofloxacin: induced CAT level Mitomycin C: uninduced CAT level fvditomycin C. induced CAT level

I l

îtu E 1

pEJ257 plTT12 pFl/117

Length of upstream sequence

Figure 6.2 Expression of M. microti recA is inducible in M. smegmatis me 155 M smegmatis nic' 155 harbouring the parental vector (pEJ257) or recombinant clones (pNT2, pNT 12, pFM7 and pFM 17) were grown as described (section 2.11 ) and induced with either ofloxacin ( 1 pg/nil f.c.) or mitomycin C

(2-5pguiil f.c). Cell-free extracts were prepared and both uninduced and induced cultures assayed for CAT expression and total protein content using a CAT-ELISA enzyme immunoassay kit (Bochringcr-Mannheim) and the BCA rnicro- titre plate assay {Pierce) respectively. Each value (ng CAT i mg total protein) is the mean of between 4 - 1 0 replicate cultures; the error bars represent standard deviations calculated using a 90% confidence interval.

122 (kb scale)

pNTm id RecA

Upstream sequence

recA gene

Figure 6.3 Strategy for sequencing the M. microti recA gene and 700bp of upstream sequence

Prelimmary sequencing reactions of construct pNTmicl, wliich contains tlie complete M microti recA gene and

700bp upstream DNA, were performed using the MI 3 reverse and -20 primers located either side of the insert in vector pBluescript H KS. Subsequent primers were designed (position and direction indicated by tlie black arrows) based on the sequence obtained enabling determination of the sequence on both strands of the entire fragment.

123 i ' f t !i \i 'i 1

UM.M

140 150 160 170 . I... - I___ TCGGGGCGG CTTATGGCTCGTCGG TGCTGCTCGATC ACGTG TCGGGGCGG - TT ATGGCT CGT CGG TGCTGCT CG ATC ACGTG

180 190 200 210 _J ___ I__ CTGACCGGCTTCGACGGCCGTAGCGCCGCCCACtGCCA M. microti CTGACCGGCTTCGACG -CCGTAGCGCCGCCCAGGCCA tuberculosis

Figure 6.4 Sequence trace illustrating the two-nucleotide changes found upstream of the M. microti recA gene The M. microti and M tuberculosis recA genes have 100% DNA-DNA relatedness. Homology at the DNA level extends 700bp upstream of the / «vl-coding region. Only two nucleotide changes were found, one 526bp and the other 478bp upstream of the rccA gene ( indicated by arrows on the trace and highlighted in the sequence), both insertions in M. microti.

124 M. tuberculosis recA gene is located 120bp upstream of the M. microti recA gene (figure 6.5).

6.5 Discussion

The results described in this chapter demonstrate that the M. microti recA gene is inducible in M. smegmatis. The maximum level of induction from the putative promoter site in the smallest clone (0-35kb insert) was 4- 8-fold following exposure to mitomycin C and 2-fold for ofloxacin. The induction ratio of the larger construct (0-7kb insert) was approximately 2-3-fold for both DNA damage-inducing agents. There was no apparent difference in induction ratio with respect to the size of the insert for equivalent M. tuberculosis constructs. However, the induction ratio for mitomycin C (3-4- to 3- 6 -fold) was double that of ofloxacin (1-8- to 1-9-fold).

Evidence presented in chapter 5 suggests that mduction of RecA in M. smegmatis is more rapid following exposure to mitomycin C (1 hour) than ofloxacin (3 hours). The variation in the induction ratios reported above may be a reflection of this initial difference in accumulation of RecA since the CAT assay measures absolute levels of CAT expression. Movahedzadeh et al (1997b) reported a difference in induction ratio with respect to construct length (2-5 and 4-3 for the larger and shorter construct respectively) but no difference with respect to inducing agent. The reason for the variation between these two investigations is not clear. It is perhaps worth comparing these induction ratios with those of other bacteria. In E. coli, induction ratios for DNA damage inducible genes are 5- to 7-fold for recA-lacZ fusions (Weisemann et al, 1984; Fernandez de Henestrosa et al, 1991) and 8-5, 6-5, 4-7 and 3-6 respectively for fusions of recA,polB, dinG and Jm//with galactokinase (galK; Lewis etal, 1992). Fusions of B. subtilis recA with a xylE reporter gene gave induction ratios of 5-5- to 10-fold (Gassel and Alonso, 1989; Raymond-Denise and Guillen, 1992) whereas Lovett et al (1988) demonstrated a 5-fold induction by a densitometer scan of immmunoblots. More recently, an induction ration of 2-5- to 2-8-fold has been reported for M smegmatis recA using both the CAT and xylE reporter genes (Papavinasasundaram et al, 1997; Durbach et al, 1997).

Attempts to resolve the apparent paradox between the lack of inducibility of the M. microti recA gene in M. microti and that observed for the M microti recA gene in M smegmatis have not been successful. It has proved difficult to efficiently transform M. microti with plasmids containing transcriptional fusions of the recA promoter regions from either M. smegmatis or M. microti to the CAT gene. More recently the effect of electroporation temperature, biochemical pre-treatment of cells and growth stage of the culture on the efficiency of transformation in slow-growing mycobacteria has been investigated (Wards and Collins, 1996). These authors reported that the efficiency of transformation in slow growers such as M. tuberculosis and M. intracellulare increased at temperatures up to 37°C. Conversely, the efficiency of transformation in fast-growers such as M. smegmatis decreased with increasing temperatures. So far it has proved difficult to efficiently transform M. microti at any temperature.

125 1 GCG GCC GC c AG CGTCGC CG GTGC GGC CG ATCC GGTG CGG GTCTG CCACGACGCGC TDGC G æ

61 TTGGC GGG GTAGC GC GGTG GGTATGTTG GC TGGC GTGC GGC TG AC GG AG TTCC ACGAGCG 120

121 GGTAGCCCTGCATTTCGGGGCGg Q tTATGGCTCGTCGGTGCTGCTCGATCACGTGCTGAC 180

181 CGG CTTC G AC CtQ cCGTAGC GC CGCCCAGG CC ATC GAG G AC GGTGTCGAGC C CC GCG ACG T 240

241 CTGG CG GGC ATTGTGC GCC GACTTCG ACG TTCC CCACGATC GGTG GTGAGGTTGC AC ATT 300

201 CGTCAAAGCGGGCAATCGCCGCACTTGCCTCCGGCGGGTACCCGGTAGAACAGGCAGCAG 360

Chect-box like 361 CTGCGACGCCGAA AGGTCAGATCCGGGCCGGTGA GCAC 3CCG 3ATC CGGCCAGGCTAGCG 420

421 GTGTTG AGCAG ATC GTC GGTG ATCC GG AC C AG CC GCG CACGCAAGTC GGGC CG C AC CG CC 480

481 GCC AGGGCGTTCG ACGCGCCGACGAGCGCGGACGCGATCtTTGCC ACACGCGGCG IY3TC AC 540

oZl -35 Cheo-box j32 ~10 a?0 -35 341 ACTTGA^srC GAAC AGGr T p n ^GGCTACTG TG GTG ATC0TC GG AG C AG CC G AC ITGTC A 600

fj70-10 æ 1 TGGC TGTClrCTAGTb TC AC GG C(3©iCC G AC CGATAC CGG TC A ATC GAAC ACC GAC C AC AG 660

M T Q T P D R E K A L E L A V A Q 661 G AG AG GC AC C ATG ACGCAGACCCCCGATCGGGAAAAGGCGCTCGAGCTGGCAGTGGCCCA 720

I E K S Y G K G S V M R L G D E A R Q 721 GATCG AG A AG AG TTACGG CAAAG GTTCG GTGATGC GC CTC GGC G AC G AGG CGC GTC AGC C 780

Figure 6.5 Complete nucleotide sequence of 671 bp upstream of the M. microti recA gene The A/, microfi recA gene and TOObp of upstream sequence have almost 100% DNA-DNA relatedness with equivalent sequences in M. tuberculosis. Only two bases differ; both insertions inM. microti (O)- Putative transcription start sites based on homology with the M.tuberculosis recA gene (Movahedzadeh et al, 1997b) are circled (Q)- The putative promoter elements w ith homology toE. coli promoters identified near these transcription start sites, the Cheo-box and the Cheo-box like motif are boxed and labelled.

126 There is almost 100% DNA-DNA relatedness between the M. microti recA gene and VOObp of upstream DNA with equivalent sequences in M tuberculosis. Only two base changes were evident, both insertions in the DNA upstream of M microti recA. Homology between these two species includes a motif (120bp upstream) with an exact match to the SOS box (GAAC-N 4-GTTC) identified by Cheo e? a/ (1991) and a second motif 275bp upstream possessing 6/8 matches (GCAC-N 4-GATC). Imaeda (1985) demonstrated that members of the M. tuberculosis complex (M tuberculosis, M. bovis, M. microti and M. africanum) are genetically very similar (85-100% homology) but differ in their epidemiology. Restriction enzyme analysis revealed that the distribution of DNA fragments produced by a single enzyme, for example £'coRI, EcoKV or Kpn\, are indistinguishable among representative strains of the M tuberculosis complex. Moreover, they have highly conserved DNA sequences in several regions. The M tuberculosis 16S rRNA sequence differs in only one position fromM bovis and M microti (Kempsell et al, 1992) and both the dnaJ genes (Takewaki et al, 1994) and the 16S to 23S rDNA internal transcribed spacer are completely conserved (Kirschner et al, 1993; Frothingham et al, 1994). Similarly, both M. tuberculosis and M. microti are indistinguishable by PCR for insertion sequences IS 1081 and IS6110 and the gene MPB70 (Liebana et al, 1996). It is perhaps noteworthy that these observations are not a general property of all mycobacterial complexes. The M. avium complex dnaJ genes have only 94% similarity (Takewaki et al, 1994) and the 16S to 23S rDNA internal transcribed spacer differs by 21 nucleotides (Frothingham et al, 1994).

Several hypotheses have been offered to explain this high degree of sequence conservation. The intracellular nature of members of the M. tuberculosis complex may result in a lower total number of organisms being present resulting in fewer opportunities for mutation. In addition the opportunity for recombination with other species or strains would be limited and may result in genetic homogeneity. It is assumed that members of the M. tuberculosis complex are well adapted to the host environment, which presumably remains relatively constant; mutations would only be conserved if they provided a selective advantage (Frothingham et al, 1994). Another intriguing possibility is that M. tuberculosis, M. microti, M. bovis BCG and M. africanum have diverged as separate species only very recently allowing little time for divergent mutations to accumulate.

Evidence presented here and in chapter 5 demonstrates that the M. microti recA gene is inducible in M. smegmatis but not in its native host. A comparative sequence analysis between M microti recA and equivalent sequences in M. tuberculosis has identified a Cheo box motif (GAAC-N 4-GTTC) identical to that upstream of recA genes inM tuberculosis, M. smegmatis andM leprae. Binding of LexA to this consensus has been demonstrated in these species by gel mobility assays and DNase I footprinting but not to the second motif 275bp upstream possessing 6/8 matches (GCAC-N 4-GATC, Durbach et al, 1997; Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). The results described so far suggest that the basic regulatory mechanism of the SOS response has been conserved and that a defect affecting the ability of M. microti LexA to bind to the SOS box may be responsible for its constitutive phenotype.

The physiological consequences of a lack of repression by LexA (such as filamentation and a reduced

127 viability or growth rate) are not obviously manifest in M microti. Therefore, other plausible explanations to account for the constitutive phenotype must be considered. One possibility is that a functional LexA homologue is present but unable to bind efficiently. It has been reported that the intracellular environment plays a role in fine tuning repression of the SOS system by LexA; binding affinity and cooperativity have been demonstrated to be dependent on both pH and the specific anion / cation composition and concentration (Shaner and Gaissarian, 1996; Relan et al, 1997). Alternatively, M. microti LexA may be preferentially displaced from the operator site; Preobrajenskaya et al (1994) have shown that protein HU can displace LexA from its binding sites on the lexA, recA and sulA genes in E. coli. Although this seems unlikely considering the high degree of homology between members of the M tuberculosis complex (Imaeda, 1982; Kempsell et al, 1992; Kirschner et al, 1993; Frothingham et al, 1994; Takewaki et al, 1994) and the absence of evidence for a similar phenomenon in any other member. A third possibility is that LexA does not regulate the M. microti recA gene. Not all bacteria that possess an E. co/i-like SOS system have the same genes under LexA control. For example the uvrB gene of E. coli is DNA damage inducible (Fogliano and Schendel, 1981) but that of Psuedomonas aeruginosa is not (Rivera et al, 1996).

Another intriguing possibility is that the constitutive phenotype is due to the presence of a host factor that affects LexA binding or acts as a positive regulator. Evidence provided by lacZ fusions with Rhodobacter capsulatas recA suggests that the protein controlling this gene is a positive regulator. The level of recA expression in cells with a doubly mutated SOS-like box is equivalent to that of untreated wild type cells and not that of treated wild type cells, which would be expected if the controlling protein were a repressor (Fernandez de Henestrosa et al, 1997). A few more examples of damage inducible genes under positive control are known, among them the Erwinia carotovora pnlA and ctv genes encoding pectin lyase and carotovoricin respectively (McEvoy et al, 1992). The product of rdgB stimulates transcription of these genes. Expression of this gene is blocked by rdgA\ a repressor that has been reported to be cleaved by RecA. A similar situation exists for the transcriptional regulation of the various types of pyocins by Psuedomonas aeruginosa (Matsui et al, 1993). However, it is worth noting that neither the pectin lyase nor the pyocin genes possesses a LexA binding site in their promoter regions. In chapter 7 the function of the consensus GAAC-N 4-GTTC as an operator in M. microti recA expression will be examined by gel retardation analysis. It is anticipated that these investigations will provide further insight into the binding properties of the M. microti LexA protein.

128 7.0 Interaction of LexA protein with the M ycobacterium m icroti

r e c A o p e r a t o r r e g i o n

7.1 Introduction

The E. coli LexA repressor comprises of 202 amino acids (Horii et al, 1981b; Markham et al, 1981) arranged in two structurally defined domains linked by a hinge region. The N-terminal domain (residues 1-84) is the DNA binding domain (Little and Hill, 1985; Hurstel et al, 1986; Schnarr et al, 1991; Thliveris and Mount, 1992), while the C-terminal domain (residues 85-202) is called the oligomerization domain (Schnarr et al, 1985; 1988). The 80S response (see section 1.6.4) is activated upon proteolytic cleavage of the repressor into its two separate domains (Little, 1984) resulting in a reduction in the specific affinity of LexA for the operator (Bertrand-Burggraf et al, 1987; Kim and Little, 1992).

In previous chapters M. microti has been shown to possess a LexA homologue that is cleaved following exposure to DNA damaging agents. However, the lack of RecA induction in this species following DNA damage and the high basal levels of RecA suggests that there is no repression of the recA gene. Sequencing of the M. microti recA gene and TOObp of upstream DNA revealed almost 100% DNA-DNA relatedness with equivalent sequences in M. tuberculosis. This homology includes the putative regulatory elements of the M. microti recA gene, which were shown to be inducible in M. smegmatis following treatment with a DNA damage-inducing agent. The evidence presented so far suggests that the basic regulatory elements of the SOS response in M. microti have been conserved and that a defect in the sequence upstream of the recA gene is not responsible for the constitutive RecA phenotype in M. microti.

In M tuberculosis, M. smegmatis and M. leprae binding of LexA to the recA operator (GAAC-N 4-GTTC) has been demonstrated by gel mobility assays and DNase I footprinting (Durbach et al, 1997; Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). The following investigations will focus on the interaction of M. microti LexA with the operator (GAAC- N 4-GTTC) identified upstream of M. microti recA to determine whether the repressor is able to bind and therefore regulate expression of this gene. To directly assess whether the M. microti LexA homologue binds to the Cheo box motif (GAAC-

N4-GTTC) gel mobility assays will be performed. This technique is based on the differing electrophoretic mobilities of free DNA and DNA-protein complexes, thereby facilitating the study of their interactions.

The work in this chapter describes: i) Binding of M. microti LexA to the Cheo box motif ii) Evidence that the altered binding properties are due to specific interactions between M. microti LexA and the Cheo box motif iii) Investigations to determine whether LexA will bind to a half-site.

129 7.2 Binding of the M. microti LexA homologue to the operator region of therecA gene

Two 32bp oligonucleotides were designed comprising the Cheo box motif (bold) identified 120bp upstream of M. microti recA plus lObp each of upstream and downstream flanking sequence of the putative operator site: Gel shift 1 (5’-GAG TTG AAT GGA ACA GGT GTT CGG GTA GTG TG-3’) Gel shift 2 (5’-GAG AGT AGG GGA ACA GGT GTT GGA TTG AAG TG-3’). A double stranded oligonucleotide (referred to as wild type in future text) was prepared from these by denaturing them together at 95°G then cooling slowly to room temperature to allow them to anneal. A second set of 32bp oligonucleotides with a mutation in the Gheo box were designed (altered bases are shown in bold italics) to confirm that binding to the Gheo box motif was specific; Mutated gel shift 1 (5’ GAG TTG AAT CTC CAA GGT TGGAGG GTA GTG TG 3’) Mutated gel shift 2 (5’ GAG AGT AGG CTC CAK GGT TGG AG A TTG AAG TG 3’). After preparation of a double stranded oligonucleotide (referred to as mutated in future text) both the wild type and mutated oligonucleotides were labelled with digoxigenin as described (section 2.13). A native polyacrylamide gel ( 8% in 0.25 x TBE; appendix IV) was prepared 12 hours in advance and pre­ run for 2 hours at 150V to ensure complete gel polymerisation and decomposition of ammonium persulphate. The labelled DNA fragments were diluted to 30fmole/pl in TEN buffer and the binding reaction set up on ice using 30fmole labelled DNA, 5-lOpg M. microti cell-free extract and Ipl of poly[d(I-G)] competitor DNA (included to prevent non-specific binding). Purified M. tuberculosis LexA (01 fig; Movahedzadeh et al, 1997a) and M. tuberculosis cell-free extract (5-1 Opg) were used in control reactions. After incubation for 15 minutes at room temperature, samples were immediately applied to the pre-electrophoresed gel. Electrophoresis and transfer of the protein-DNA complex was carried out as described (section 2.13) and chemiluminescent detection performed according to the manufacturer’s instructions.

A retarded complex of comparable migration to that seen with purified M. tuberculosis LexA and wild type DNA (figure 7.1, lane 2) was present for both M. tuberculosis (figure 7.1, lanes 6 and 7) and M. microti cell-free extracts (figure 7.1, lanes 9 and 10). In addition, the interaction of M. microti LexA with the wild type oligonucleotide produced a second, stronger signal for a retarded complex that migrated further into the gel. The ratio of intensity between these two bands was variable between experiments. No retarded fragments were seen when mutated DNA altered only in the Gheo box motif was used (figure 7.1, lanes 3, 4 and 11 through 16).

130 wt m wt m Cheo box motif M. tuberculosis LexA - + - + M. tuberculosis cell-free extract ----- M. microti cell-free extract ------

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Figure 7.1 Binding of the M. microti LexA homologue to the operator region of the recA gene

Two32bp double stranded oligonucleotides were designed; one spanning the Cheo box motif (wild type {wt},lanes

!,2 and 5 through !0\ 32fmoles) GAG TTG AAT GGA ACA GGT GTT CGG GTA GTG TG and an identical sequence except for mutations introduced into the Gheo box motif (mutated {m},lanes 3, 4 and 11 through 16: 32fmoles) GAG TTG AAT GTC CAA GGT TGG AGG GTA GTG TG. Each of these was analysed with the following: no protein added, 0-1 pg pureM. tuberculosis LexA, 5 or lOpg of either M. tuberculosis or M. microti cell-free extract. The protein added to each reaction is indicated above the autoradiograph. A retarded complex of comparable migration to that seen with purified M. tuberculosis LexA and wild type DNA {lane 2) was present for both M. tuberculosis (lanes 6 and 7) and M. microti cell-free extracts (lanes 9 and 10). In addition, the interaction of M. microti LexA with the wild type oligonucleotide produced a second, stronger signal for a retarded complex that migrated further into the gel. The ratio of intensity between these two bands was variable between experiments. No retarded fragments were seen when the mutated oligonucleotide altered only in the Gheo box motif was used (lanes 3. 4 and II through Iff).

131 7.3 The altered binding observed for M, microti cell-free extracts is due to the specific interaction of a LexA homologue with the conserved Cheo box motif

The results described above clearly show that labelled, wild type oligonucleotide is retarded in the presence of purified M. tuberculosis LexA andM microti or M. tuberculosis cell-free extracts. To confirm that the second, faster migrating complex was due to specific binding of M microti LexA to the Cheo box motif, competition experiments were performed.

To assess the specificity of the DNA sequence in the interaction the reaction was performed in the presence of excess unlabelled oligonucleotides. Incubation with increasing amounts of unlabelled wild type DNA (16-10,000fmoles) reduced the proportion of retarded DNA with M tuberculosis LexA (figure 7.2a, lanes 1 through 5) and of both complexes with M. microti cell-free extract (figure 7.2b, lanes 7 through 11). In comparison, incubation of M. microti cell-free extract with excess unlabelled mutated oligonucleotide had no effect upon the proportion of retarded DNA (figure 7.2b, lanes 12 through 16). These results indicate that both the retarded DNA fragments observed with M. microti cell-free extract are due to specific interactions with the Cheo box motif.

To investigate the specificity of the protein in the interaction the reaction was performed following preincubation of the cell-free extract with an antibody raised against M. tuberculosis LexA (kindly donated by Dr. P. Jenner, N.I.M.R.). This treatment drastically reduced the proportion of retarded DNA with purified M. tuberculosis LexA (figure 7.3, lanes 3 through 5) and both M. tuberculosis (figure 7.3, lanes 9 through 11) andM. microti cell-free extracts (figure 7.3, lanes 14 through 16). WithM. microti cell-free extract both the retarded bands were equally affected. This analysis suggests that both of the retarded DNA complexes are due to the specific binding of an M. microti LexA homologue to the Cheo box motif.

7.4 Investigations to determine if M. microti LexA is binding as a monomer to a half site in the Cheo box motif

Mutagenesis assays (Thhveris and Mount, 1991) in combination with DNase I and hydroxy radical footprinting have provided compelling evidence which suggests that the operator bound form of E. coli LexA is a dimer (Brent and Ptashne, 1981; Little et al, 1981; Hurstel et al, 1986; Kim and Little, 1992; Miller et al, 1996). The dyad symmetry of the operator site is common to several repressor-operator complexes; in each case one protein monomer contacts each half of the palindrome (Anderson et al, 1985; Jordan and Pabo, 1988; Wolberger et al, 1988). Therefore, it seems reasonable to expect that the mycobacterial LexA would bind as a dimer. Thus the faster-migrating DNA-protein complex seen with M microti LexA could be due to LexA binding as a monomer, which would not retard the complex as much as a dimer.

132 a) b) m wt Cheo box motif wt M. tuberculosis LexA - + - + Cheo box motif ------' *------M. microti cell-free extract + 4- + + + + + + + + + M. tuberculosis LexA + + + + + + - Unlabelled wild type Cheo box

Uniabelled wild type Cheo box — ------~ I - - Unlabelled mutated Cheo box 1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Figure 7.2

Binding of the M. m icroti LexA homologue to the Cheo box motif is specific

Purified M. tuberculosis LexA (0-1 pg) plus labelled wild type DNA {wt} (CAC TTG AAT CCA ACA GGT GTT CGG CTA CTG TG; 32fmoles, a lanes 1 through 7) was incubated with increasing amounts of unlabelled wild type sequence as competitor {a lanes I through 5; using 5- fold increments from 16-I0,000fmoles). The proportion of retarded complex was reduced as the concentration of competitor increased. Control reactions comprise wild type DNA with and without O lpg purifiedM. tuberculosis LexA (a, lanes 6 and 7). Incubation ofAf. microti cell-free extract (lOpg) with wild type DNA in the presence of excess unlabelled wild type sequence (b lanes 7 through II) reduced the proportion of both retarded complexes. Whereas, the presence of unlabelled mutated DNA as competitor(b lanes 12 through 16) had no affect upon the formation of either retarded protein-DNA complex. Control reactions comprising labelled mutated DNA {m}{b lanes I and 2) or wild type

DNA {b lanes 3 and 4) with and without purified M. tuberculosis LexA are shown. The protein added to each reaction is indicated above the autoradiograph.

133 wt Cheo box motif

Pre-incubated with a LexA antibody

M. tuberculosis LexA - + + + +

M. tuberculosis cell-free extract + - ++ +

M. microti cell-free extract 4- 4-4-4-

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Figure 7.3 Confirmation that the second, faster migrating complex is due to specific binding of M. microti LexA to the Cheo box motif

Labelled wild type DNA {wt| (CAC TTG AAT CCA ACA GGT GTT CGG CTA CTG TG; 32fmoles in lanes I through I6\ 32fmoles) was analysed with each of the following: no protein added, Olpg M. tuberculosis LexA, 5pg of either M. microti or M. tuberculosis cell-free extract. Prior to addition to the gel shift reaction samples of cell-free extract were pre-incubated with a polyclonal antibody raised against M tuberculosis LexA (undiluted lanes 3, 9 and 14\ 1/10 dilution lanes 4, 10 and 75; 1/100 dilution lanes 5. II and 16). The protein added to each reaction is indicated above the autoradiograph. The following control reactions were included: wild type DNA only{lanes I, 6 and 12) and wild type DNA plus antibody in the absence of protein {lane 8). The proportion of retarded DNA with purified M. tuberculosis LexA {lanes 3 through 5) and both M. tuberculosis {lanes 9 through 11) and M. microti cell- free extracts {lanes 14 through 16) was drastically reduced. With M. microti cell-free extract both retarded bands were equally affected.

134 To determine whether the protein-DNA complex which migrates further into the gel is indeed due to binding of a LexA monomer, two 32bp double stranded oligonucleotides with mutations introduced into each half site were prepared (altered bases are in bold italics): i) Mutated gel shift 3/4 (5’-CAC TTG AAT CTC CAA GGT GTT CGG CTA CTG TG-3’) and its complement, ii) Mutated gel shift 5/6 (5 ’-CAC TTG AAT CGA ACA GGT TGG AGG CTA CTG TG-3 ’) and its complement. These were used in gel shift assays as described above. As anticipated, the interaction of M. microti cell- free extract with the wild type oligonucleotide resulted in the formation of two protein-DNA complexes (figure 7.4, lanes 13 and 14). No comparable bands were observed with either the mutated oligonucleotide (figure 7.4, lane 12) or those with mutations in a half-site (figure 7.4, lanes 15, 16, 21 and 22) although non-specific binding of variable intensity was apparent in some reactions. Similarly, the interaction of purified M. tuberculosis LexA or M tuberculosis cell-free extract with the wild type oligonucleotide resulted in the formation of a retarded protein-DNA complex (figure 7.4). No comparable complexes were observed with any mutated oligonucleotide.

Binding of LexA to an intact operator has been reported to have a 1000-fold greater affinity than binding to a half-site (Kim and Little, 1992). To assess whether binding of purified M. tuberculosis LexA has a lower affinity for binding to a half-site as opposed to an intact operator, the wild type and mutated oligonucleotides were incubated with an increasing concentration of purified M. tuberculosis LexA protein (0-001— lOpg, figure 7.5). It should be noted that both the concentration of the M. tuberculosis LexA stock solution and the maximum volume of each gel shift reaction are factors limiting the maximum LexA concentration used in these investigations.

An increasing proportion of DNA was retarded for the interaction of labelled wild type DNA with an increasing concentration of purified M. tuberculosis LexA (figure 7.5, lanes 2 through 6 ). However, no retarded fragments of the anticipated migration were present in equivalent reactions using labelled mutated DNA (figure 7.5, lanes 8 through 12) or DNA mutated in a half-site (figure 7.5, lanes 14 through 18 and 21 through 25) even with a 100-fold excess of the concentration sufficient to retard all wild type sequence. Curiously a retarded fragment that migrated only partially into the gel is seen following the interaction of wild type DNA (figure 7.5, lanes 5 and 6 ) or DNA mutated in a half-site (figure 7.5, lanes 17, 18, 24 and 25) m the presence of a 10-fold excess (>lpg) of purified M tuberculosis LexA. A protein-DNA complex with a comparable migration is formed following the interaction of a 100-fold excess of purified M tuberculosis LexA (lOpg) with labelled mutated DNA (figure 7.5, lane 12).

7.5 Discussion

The work in this chapter describes the specific binding of M. microti LexA to the conserved Cheo box motif (GAAC-N 4-GTTC) found 120bp upstream of the M. microti recA gene. A retarded DNA fragment

135 m wt wt3/4 m wt m3/4 m wt m3/4 m5/6 Cheo box motif r— ^ r— ^ ^ ' ' ' M. tuberculosis LexA _ + _ + - + - - - - M. tuberculosis cell-free extract ______- - M. microti cell-free extract - - - -

1 23456789 1011 1213141516 17 18 19 20 21 22

Figure 7.4 Mutations in a half-site of the Cheo box motif affect binding of M. microti and M. tuberculosis LexA protein

To determine whether the protein-DNA complex which migrates further into the gel is due to binding of a LexA monomer four 32bp double stranded oligonucleotides were designed: wild type {wt}, CAC TTG AAT CGA ACA

GGT GTT CGG CTA CTG TG (32fmoles, lanes 3. 4, 8. 9. 13 and /4); mutated {m |, CAC TTG AAT CTC CAA GGT TGG AGG CTA CTG TG (32fmoles, lanes I. 2. 7 and /2); mutated gel shift 3/4 {m3/4}, 5'-CAC TTG AAT CTC C A k GGT GTT CGG CTA CTG TG-3’ (32fmoles, lanes 5. 6. 10. 11. 15 and 16), mutated gel shift 5/6 {m5/6}, 5'-CAC TTG AAT CGA ACA GGT TGGAGG CTA CTG TG-3’ (32fmoles, lanes 17 through 22). Each of these was analysed with no protein added, 0 Ipg M. tuberculosis LexA, 5 or lOpg of either M. microti or M. tuberculosis cell-free extract. The protein added to each reaction is indicated above the autoradiograph. The interaction ofM. microti cell-free extract with the wild type oligonucleotide resulted in the formation of two protein- DNA complexes {lanes 13 and 14). No comparable bands were observed with the mutated DNA {lane 12) or DNA mutated in a half-site {lanes 15, 16, 21 and 22), although non-specific binding of variable intensity was apparent in some reactions. The interaction of purifiedM. tuberculosis LexA or M. tuberculosis cell-free extract with the wild type oligonucleotide resulted in the formation of a retarded protein-DNA complex. No comparable complexes were observed with any mutated oligonucleotide.

136 wt m3/4 wt m5/6 Cheo box motif

M. microti cell-free extract

M. tubercuiosis LexA

8 9 10 11 12 13 14 15 16 17 18 19 20 21 21 22 23 24 25

Figure 7.5 The affect of excess purified M. tuberculosis LexA on binding to the wild type and mutated Cheo box motif

To assess whether purified M. tuberculosis LexA can bind as a monomer at a higher concentration the following double stranded oligonucleotides were used: wild type {wt}, CAC TTG AAT CGA ACA GGT GTT CGG CTA

CTG TG (32fmoles, lanes I through (5); mutated {m}, CAC TTG AAT CTC CAA GGT TGG AGG CTA CTG TG (32fmoles, lanes 7 through 72); mutated gel shift 3/4 {m3/4}, 5'-CAC TTG AAT CTC CAA GGT GTT CGG CTA

CTG TG-3’ (32fmoles, lanes 12 through /5); mutated gel shift 5/6 {mS/6}, 5'-CAC TTG AAT CGA ACA GGT TGG AGG CTA CTG TG-3’ (32fmoles, lanes 20 through 25). Each of these was analysed with no protein added or 10-fold increments of 0 001 - lOpg purifiedM. tuberculosis LexA. The protein added to each reaction is indicated above the autoradiograph. An increasing proportion of DNA was retarded for the interaction of labelled wild type DNA with an increasing protein concentration (Janes 2 through 6). No retarded fragments of the anticipated migration were present in equivalent reactions using mutated DNA (lanes 8 through 12) or DNA mutated in a half- site (lanes 14 through 18 and 21 through 25) even with a 100-fold excess of the concentration sufficient to retard all the wild type sequence. The interaction o f wild type DNA (lanes 5 and 6) or DNA mutated in a half-site (lanes 17. 18. 24 and 25) with a 10-fold excess (> 1 pg) of purified M. tuberculosis LexA resulted in the formation of a retarded fragment that migrated only partially into the gel. A protein-DNA complex with a comparable migration was formed following the interaction of a 100-fold excess of purifiedM. tuberculosis LexA (lOpg) with labelled mutated DNA (lane 12). The interaction ofM. microti cell-free extract (5pg) with wild type DNA is shown in lane 19.

137 observed with M. microti and M. tuberculosis cell-free extracts is comparable to that produced with purified M. tuberculosis LexA. In addition, a second more intense protein-DNA complex that migrates further into the gel is consistently observed with M. microti cell-free extracts. Neither retarded fragment was observed when an oligonucleotide mutated only in the Cheo box motif was used. Furthermore, increasing amounts of unlabelled wild type oligonucleotide as competitor reduced the proportion of both retarded, labelled DNA complexes whereas equivalent amounts of unlabelled mutated oligonucleotide did not. These observations suggest that proteins present in M. microti cell-free extract bind specifically to the conserved Cheo box motif. Pre-incubation of M. microti cell-free extract with an antibody raised against M. tuberculosis LexA drastically reduced the proportion of retarded, labelled DNA in both complexes indicating that they are due to binding of an M. microti LexA homologue. The results suggest that M microti LexA does bind to the conserved Cheo box motif but in a manner distinct from that of M. tuberculosis.

The E. coli LexA repressor is a two-domain stmcture comprising the DNA-binding domain (N- terminal residues 1-84; Little and Hill, 1985; Hurstel et al, 1986; Schnarr et al, 1991; Thliveris and Mount, 1992) and the oligomerization domain (C- terminal residues 85-202). The structure of the N- terminal domain has been determined by nuclear magnetic resonance (Fogh et al, 1994) revealing that LexA is a member of a family of proteins which recognise DNA through a helix-tum-helix motif (Harrison and Aggerwal, 1990). Mutagenesis assays have identified several E. coli LexA mutants defective in their DNA binding ability (Thliveris and Mount, 1992). In addition, E. coli spr derivatives have been identified that have a high rate of constitutive expression of SOS genes; although these are thought to make little functional LexA {lexA Def mutants; Mount, 1977; Pacelli et al, 1979).

Evidence presented in this chapter demonstrates that M microti LexA does bind to the Cheo box motif, although with altered DNA binding properties. This suggests that mutations at positions in the DNA- binding domain that form no direct protein-DNA contacts may affect “occupancy” of the operator. One possibility is a defect affecting the stability or strength with which the repressor binds the operator; the variation in the comparative intensity of the two retarded complexes formed with M microti LexA in successive experiments provides evidence for this hypothesis. Defective repressors that cannot interact with the operator sequence tightly have been identified in the E. coli lac repressor (Kleina and Miller, 1990). These are predominantly due to incorrect folding of the protein and are usually a result of an alteration in the N-terminal (DNA-binding) domain. Mutations in this domain have been identified that affect dimerization of the E. coli LexA repressor, for example Knegtel et al (1995) identified several residues in the DNA binding domain that enhance binding by stabilising the dimerization interface.

Mutations in the C- terminal domain are also of consequence. Kleina and Miller (1990) have characterised several in the E. coli lac repressor that lead to a constitutive phenotype. In particular mutations affecting subunit contacts can alter protein function; for example, defects that affect folding of the C-terminal domain and hence oligomerization. Golemis and Brent (1992) have postulated that protein-protein interactions of E. coli LexA are of equal significance to protein-DNA interactions. These

138 authors suggested that the carboxy-terminal domain of a LexA monomer promotes a spatially precise association with a second monomer attached to the operator site; LexA monomers lacking this interaction are more likely to be displaced from the DNA. Based on mutational and biochemical evidence a model for the sequential, co-operative binding of two LexA monomers in a two-step process has been proposed (Schnarr etal, 1985; 1988; Thliveris and Mount, 1991; Golemis and Brent, 1992; Kim and Little, 1992). Binding of a second monomer to the operator involves both protein-DNA and protein-protein contacts (Kim and Little, 1992). It is generally believed that the co-operative binding observed for the second monomer arises from protem-protein interactions. Moreover that the structure of the protein may change upon binding to DNA (Brennan et al, 1990; Sauer, 1990) allowing additional protein-protein contacts to form or stabilising a conformation not favoured in solution. However, it is equally plausible that the first bound monomer confers a conformational change on the DNA-binding site that increases the affinity of this second site for the second monomer (Kim and Little, 1992).

The DNA-bmding properties of E. coli LexA have been extensively analysed and specific bases important for the interaction between the N-terminal (DNA-binding) domain and the conserved motif of the SOS box (CTGT) identified (Ottleben et al, 1991; Schnarr et al, 1991; Thliveris and Mount, 1992). Furthermore non-specific protein-DNA contacts of the repressor with the intervening AT rich sequence have been shown to modulate repressor affinity (Wertman and Mount, 1985; Knegtel et al, 1995). To date no evidence is available describing the precise interaction of a mycobacterial repressor with specific bases in the operator site. In addition the importance of non-specific protein-DNA contacts with the nucleotide sequence flanking the Cheo box motif and the effect of mutations on protein-DNA and protein-protein contacts are unknown. However, the importance of the conserved motif (GAAC-N 4- GTTC) has been established in B. subtilis, M. tuberculosis and M. smegmatis (Haijema et al, 1996; Miller et al, 1996; Winterling et al, 1997; Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). Binding to a motif with one mismatch has been demonstrated (Movahedzadeh et al, 1997a) whereas LexA will not bind to a second Cheo box with 6/8 matches (GCAC-N 4-GATC) located 275bp upstream of the M tuberculosis recA gene (Movahedzadeh, 1996) nor to that in M. microti (data not shown).

Elucidation of the M. microti lexA sequence was not feasible within the time-scale of this project; as a consequence no specific mutations have been identified in M microti LexA. However, investigations were initiated to investigate the possibility that the altered binding properties of M. microti LexA are due to monomer as opposed to dimer binding. Like many prokaryotic repressors (Anderson et al, 1981; Anderson et al, 1985; Jordan and Pabo, 1988; Wolberger et al, 1988) LexA binds as a dimer to an operator site that possesses dyad symmetry (Brent and Ptashne, 1981; Wertman and Mount, 1985; Cheo et al, 1991; 1993; Fernandez de Henestrosa et al, 1997; Tapias et al, 1997). Investigation of these sites in E. coli, B. subtilis and M. tuberculosis by DNase I and hydroxy radical footprinting have shown that both half sites are bound by LexA in a pattern which appears symmetrical (Brent and Ptashne, 1981 ; Little et al, 1981 ; Hurstel et al, 1986; Kim and Little, 1992; Miller et al, 1996; Movahedzadeh et al, 1997b). In addition, E. coli mutagenesis assays have provided compelling evidence for the interaction of at least one LexA monomer with each half site of the recA operator (Thliveris and Mount, 1991).

139 Binding of LexA to an intact operator has been reported to have a 1000-fold greater affinity than binding to a half-site (Kim and Little, 1992). Consequently, reactions were performed between labelled wild type DNA, mutated DNA or that mutated in a half-site and M. tuberculosis LexA, which was present at concentrations 100-fold greater than that required for retardation of all wild type DNA. It should be noted that both the concentration of the M. tuberculosis LexA stock solution and the maximum volume of each gel shift reaction were limiting factors with respect to the maximum LexA concentration in these reactions. As anticipated, the proportion of retarded wild type DNA increased with increasing LexA concentration. However, no retarded complexes with a migration comparable to that seen with either M. tuberculosis or M. microti LexA were formed in the presence of mutated DNA or that mutated in a half-site. Interestingly, a strong band that migrated only a small distance into the gel was seen following the interaction of wild type DNA and that mutated in a half-site in the presence of a 10-fold excess of purified M. tuberculosis LexA. A 100-fold excess of purified M. tuberculosis LexA was required to produce a comparable observation with the fully mutated ohgonucleotide. It is probable that these retarded protein-DNA complexes are due to the formation of a non-specific complex between DNA and a multimeric form of LexA, for example a tetramer (Schnarr et al, 1985; 1988). The E. coli LexA repressor dimerises only weakly in the absence of DNA and has a dissociation constant of 15 - 50|liM (Schnarr et al, 1985; 1988; Kim and Little, 1992). Cellular concentrations of E. coli LexA have been estimated at 1 - 2 pM of which approximately 20% is unbound (Moreau, 1987; Sassanfar and Roberts, 1990). This implies that the most of the unbound cellular repressor is monomeric. Assuming a similar dissociation constant in the presence of DNA it is plausible that as LexA concentration increases the tendency of LexA to form a multimeric complex increases. Specific information regarding the cellular concentration and kinetics of dissociation is not available for a mycobacterial LexA. However, it is plausible that a similar mechanism may account for this high molecular weight protein-DNA complex described above.

The evidence presented above suggests that M. microti LexA does not bind to a half-site. However, these experiments are far from conclusive and further investigation is recommended. DNase I and hydroxy radical footprinting (Lin and Shiuan, 1995; Miller et al, 1996; Movahedzadeh et al, 1997) may provide a more precise indication of the bases important for repressor binding. Moreover further investigation would confirm whether or not all the sequence necessary for binding is present in the wild type oligonucleotide. Haijema et al (1996) reported a partially retarded fragment for a construct of B. subtilis difiR lacking part of the sequence necessary for binding of DinR; a band that migrated further into the gel was observed on gel mobility assays. A second explanation for the lack of induction of M. microti RecA following treatment with a DNA damaging agent is that the cleaved LexA product is still able to repress expression of the recA gene. Examination of cell-free extracts prepared from M. microti following treatment with a DNA damage-inducing agent by DNase I footprinting may provide further insight.

In E. coli, there is evidence demonstrating that pH and monovalent anion identity and concentration have a profound effect upon operator binding (Dri and Moreau, 1994; Shaner and Gaissarian, 1996; Relan et al, 1997). Moreover, the binding affinity of LexA for its operator as a function of LexA monomer concentration is dependent upon the exact solution conditions. Examining the thermodynamics of LexA

140 binding to the Cheo box using gel mobility assays may provide further insight into the binding of M. microti LexA to the Cheo box motif. Other factors that may give rise to a constitutive phenotype include the presence of a host factor that acts as a positive regulator (McEvoy et al, 1992; Matsui et al, 1993; Fernandez de Henestrosa et al, 1997) or a molecule that displaces LexA from its binding site (Preobrajenskaya et al, 1994). These possibilities have aheady been discussed in a previous chapter.

141 8.0 G en er a l D isc u ssio n

Mycobacteria such as M. tuberculosis, M. bovis and M. leprae are examples of a specific category of prokaryotes that develop and replicate within the intracellular milieu of host cells where it is presumed that they will encounter a powerful array of antimicrobial agents such as peroxide and nitric oxide. It is envisaged that stress responses such as the SOS response to DNA damage are an integral part of their survival mechanism. Such responses involve the induction of a set of genes that help protect and adapt the organism in new environmental conditions; as evidenced by the modulated expression of several “stress” proteins in M tuberculosis resident within macrophages (Lee and Horwitz, 1995). While part of the stress response appears to be specialised for a particular stress, namely, heat shock (Lindquist, 1986) oxidative damage (Greenberg and Demple, 1989; Farr and Kogoma, 1991) and DNA damage (Little and Mount, 1982; Walker, 1984), there are genes that respond to multiple stresses. Examples include the groEL and dnaK genes of E. coli, which are induced by UV irradiation and nalidixic acid via a recA independent mechanism (Krueger and Walker, 1984) as well as by heat-shock. In the Gram-positive Lactococcus lactis the recA gene has been implicated in both the oxidative stress response and the heat shock response. (Duwat et al, 1995). In E. coli and other bacteria, the heat shock proteins have been implicated in DNA repair (Ogata et al, 1993; Petit et al, 1994) and the oxidative stress response (Morgan et al, 1986; Farr and Kogoma, 1991; Lindquist, 1992). By comparison the 65kDa and 71kDa antigens of M. bovis and M. tuberculosis (identified as GroEL and DnaK respectively) are induced by heat shock and acidification but unaffected by oxidative damage (Stover et al, 1991; Young et al; 1991; Lee et al, 1995; Sherman et al, 1995). To date no link between recA, the heat shock proteins and oxidative damage has been reported for mycobacteria.

8.1 The mycobacterial RecA protein

Bacterial responses to DNA damage are highly conserved. One system, the SOS response to DNA damage, involves the co-ordinately induced expression of over 20 genes through a common regulatory mechanism (Little and Mount, 1982; Walker, 1984). The RecA protein is ubiquitous among bacteria and plays a central role in the regulation of this response. In addition to its regulatory function it also mediates homologous recombination, post-replication repair, induction of phage X lysogens and mutagenesis (Roca and Cox, 1990; Miller and Kokjohn, 1990; Kowalczykowski et al, 1994). These processes have been extensively characterised in E. coli but despite major insights into the cellular immunology, physiology and biochemistry of mycobacteria over the past century they are poorly understood in this genus.

In E. coli, resistance to hydrogen peroxide correlates with the recA genotype of the cell rather than the level of the protecting enzymes catalase, peroxidase or superoxide dismutase (Carlsson and Carpenter, 1980). Furthermore, the protective enzymes catalase and superoxide dismutase were shown to be unnecessary for short (2 hours) and long (24 hours) term survival of Salmonella in macrophages

142 (Buchmeier et al, 1995) and they do not enhance the survival of E. coli in polymorphonuclear neutrophils (Schwartz et al, 1983). These observations correlate well with the proposal that the predominant lethal action caused by low doses of hydrogen peroxide (lOOpM) is DNA damage (Imlay and Linn, 1986). In addition, the recA gene has been linked to virulence: Salmonella typhimurium recA mutants demonstrate an attenuated virulence in mice (Buchmeier et al, 1995), and 80S induction in Erwinia carotovora increases virulence by increasing the synthesis of pectin lyase and carotovoricin (McEvoy et al, 1990). Consequently, the contribution of recA to the survival of M tuberculosis within the intracellular environment and the possibility that it may affect vimlence is of great interest.

Despite the phylogenetic conservation in size, structure and overall function of more than sixty prokaryotic recA gene sequences (Karlin etal, 1995; 1996) the recA gene of M tuberculosis revealed a strikingly unusual structure. The recA locus was found to specify a functional 38kDa protein from a single open reading frame with the coding potential for an 85kDa product (Davis et al, 1991). The gene coding sequence is interrupted by an open reading frame encoding an intein; a genetic element that is post-translationally spliced out of the mature protein (Davis et al, 1992). Examination of recA gene structure in other mycobacteria revealed a normal recA gene with the coding potential for a 38kDa product in all species with the exception of the M tuberculosis complex and M. leprae (Davis et al, 1994). Interestingly the recA gene of M. leprae is interrupted by an unrelated intein inserted at a different site to that of M. tuberculosis. The independent acquisition of an intein in the recA locus of the pathogenic mycobacteria was intriguing and raised the question of whether its presence affects RecA function, particularly since low levels of homologous recombination have been reported for M. tuberculosis (Kalpana et al, 1991) and M. bovis BCG (Aldovini et al, 1993). The possibility that the appearance of splicing in the recA locus of the M. tuberculosis complex and M. leprae regulates the display of the RecA phenotype was intriguing. The investigations reported here focused on characterisation of the role of RecA in the response of mycobacteria to treatment with a DNA damage inducing agent and determination of whether splicing of the M tuberculosis precursor RecA protein could be conditionally regulated.

8.2 Levels of precursor and spliced RecA protein in untreated mycobacterial cell-free extracts

Low levels of homologous recombination have been reported in M. tuberculosis (Kalpana et al, 1991) and M. bovis BCG (Aldovini et al, 1993) compared to M. smegmatis (Husson et al, 1990; Kalpana et al, 1991; Mustafa, 1995; Gordhan et al, 1996; Pelicic et al, 1996). One possible explanation to account for this observation is that an insufficient quantity of RecA is produced. To address this question the basal levels of RecA in M. smegmatis and three members of the M. tuberculosis complex were compared. Surprisingly, although the amount of RecA in uninduced cell-free extracts was similar for M. tuberculosis and M bovis BCG it was significantly greater in M. microti and M. smegmatis. Probing an identical membrane with an antibody raised against the intein of M tuberculosis RecA revealed that the basal

143 levels of this genetic element in M. tuberculosis, M. bovis BCG and M. microti mirror those of RecA. This suggests that RecA protein instability is not a significant factor in the difference in basal RecA levels. It is possible that both the RecA protein and intein are equally unstable in M. tuberculosis and M. bovis BCG, although this seems unlikely considering the high basal RecA levels seen in M microti and the genetic homogeneity reported among members of the M. tuberculosis complex (Imaeda, 1982; Kempsell et al, 1992; Kirschner et al, 1993; Frothingham et al, 1994; Takewaki et al, 1994).

Examination of nine clinical isolates of M. tuberculosis revealed that low RecA levels are a general property of this species, which suggests that an insufficient quantity of RecA (38kDa) may be a factor in the low levels of homologous recombination in M. tuberculosis and M. bovis BCG. However, a qualitative comparison of basal RecA levels in a wider range of mycobacteria revealed that RecA levels in the majority of species are comparable to that of M. tuberculosis including M. intracellulare', a species in which efficient homologous recombination has been demonstrated (Marklund etal, 1995). The level of amino acid identity between M. tuberculosis RecA and that found in other species of mycobacteria is not known. However, considering the homology between the M. smegmatis RecA protein (Papavinasasundaram et al, 1997) and its counterpart in M tuberculosis dxAM leprae (91% identity; Davis et al, 1991; 1994) it seems reasonable to assume that a significant level of homology exists among a wider range of mycobacteria. Therefore, the weak signals seen on Western blots examining RecA levels in a wider range of mycobacterial species are probably not due to significant differences in affinity and avidity (see section 3.1), which would lead to poor antigen-antibody recognition resulting in a less intense signal. Consequently, analysis of the evidence described above suggests that a low level of RecA perse may not be a significant factor in the recalcitrance of M tuberculosis and M. bovis BCG to undergo homologous recombination. This suggests that the presence of an intein in the recA locus may affect this RecA mediated function. This is clearly an intriguing possibility that gains further support in light of the following in vivo and in vitro evidence. Davis et al (1992) reported that the expression of a mutant form of the M. tuberculosis precursor protein that was inefficiently spliced to yield both precursor and spliced protein products failed to complement E. coli recA mutants. They postulated that the presence of precursor protein might have dominant negative effects, possibly by competing for RecA binding sites on target molecules or by polymerising with the 38kDa RecA monomers. Further compelling evidence was provided by Kumar et al (1996) who demonstrated that the 85kDa-precursor protein is totally defective in its DNA-binding, ATP-binding and ATPase activities. In contrast the spliced 38kDa RecA protein was shown to bind ssDNA in the presence of ATP, possess ATPase activity and be proficient in homologous pairing; although there were significant differences compared to E. coli RecA in its requirements for binding to DNA and its affinity for ATP. By comparison, the function of three other microbial RecA proteins from Bacillus subtilis (Lovett and Roberts, 1985), Proteus mirabilis (West et al, 1983) and Thermus aquaticus (Angov and Camerini-Otero, 1994) are reported to be largely similar to that of E. coli (Sassanfar and Roberts, 1990; Radding, 1991; West, 1992; Kowalczykowski et al, 1994).

Unspliced precursor RecA protein was not detected in the untreated cell-free extracts of M. tuberculosis, M. microti or M bovis BCG grown under optimal conditions in vitro. It is possible that the lack of detectable precursor protein in both M. tuberculosis and M. bovis BCG is a reflection of the low levels of

144 RecA present per se. However, this seems unlikely since it was also absent in M. microti, which has significantly elevated RecA levels compared to other members of the M. tuberculosis complex. Isolation of precursor protein has proved difficult. Kumar et al (1996) successfully over produced this protein in an E. coli recA mutant strain harbouring plasmid pEJ135 (which contains an intact M. tuberculosis recA gene under the control of a lacZ promoter) only when cells were induced and harvested during exponential growth. Investigations revealed that growth stage (ODgoonm of 0.2-1.0) had no effect on the appearance of precursor protein in neither M. tuberculosis nor M. microti. Interestingly treatment of M. tuberculosis with a DNA damage inducing agent in its “exponential” phase of growth (ODgoonm of 0.2- 0.3; see appendix II) resulted in an increase in the levels of spliced RecA (after 6-10 hours) but had no effect on the appearance of unspliced precursor RecA protein. (The effect of treatment with an inducing agent on RecA levels in M. smegmatis and members of the M. tuberculosis complex is discussed further below.) This lack of precursor protein in cells overproducing the spliced RecA product suggests that splicing in its native host is a rapid process that is not conditionally regulated in vitro following an increase in recA expression as a result of DNA damage.

The effect of growth stage at the point of harvesting on basal RecA levels was also of interest. A partial induction of some E. coli SOS genes has been reported to occur spontaneously during the growth cycle of bacteria. This phenomenon has been linked to transient changes in intracellular pH during growth, which affects the DNA binding affinity of LexA (Dri and Moreau, 1993, 1994). Several papers have shown that extracellular pH changes have transient effects on internal cellular pH homeostasis resulting in the altered expression of several LexA repressed genes (Schuldiner et al, 1986; Dri and Moreau, 1994). For example, the predominant action of weak acids in E. coli is to increase the DNA binding affinity of LexA to sfiA leading to further repression (Dri and Moreau, 1994). Conversely, Relan et al (1997) have demonstrated in vitro that E. coli LexA tightly binds to the recA operator around neutral pH but is weakened as the pH is lowered. Interestingly, basal RecA levels in both M. tuberculosis and M microti were unaffected by the growth stage at the point of harvesting. The obvious inference is that the elevated levels of RecA in M. microti are unlikely to be a result of similar transient changes in intracellular pH during growth. Cultivation in Sautons media provides a mildly acidic environment (pH6.2). Consistent with previous observations basal RecA levels were elevated in M. microti and M. smegmatis compared to M. tuberculosis. Curiously, levels of RecA inM bovis BCG were elevated following growth in this media. The precise reason behind this difference is unclear but may be due to an altered binding affinity of LexA as a result of an alteration in the internal cellular pH homeostasis due to an extracellular pH change.

Members of the M. tuberculosis complex are facultative intracellular parasites that survive and replicate within host macrophages. It is presumed that they will encounter a wide range of anti-microbial defences and that their response to DNA damage may be an integral part of survival within this environment. The effect of intracellular residence upon recA expression and the possible role of splicing in the maturation of RecA in vivo as a mechanism by which the RecA phenotype is regulated are of great interest. To address whether intracellular residence affects the appearance of spliced and precursor RecA protein cell-

145 free extracts were prepared from bacilli harvested from lung and spleen tissue collected from nude mice that had previously been infected with M. tuberculosis. However, there were intrinsic difficulties associated with this protocol. Perhaps the most pertinent was associated with the difficulty in separating bacilli and eukaryotic tissue prior to cellular disruption. This resulted in high backgrounds and a difficulty in assessing the precise amount of prokaryotic protein present in each sample. This latter problem was compounded by the lack of a suitable internal standard since M. tuberculosis hsp65 is induced by intracellular residence (Lee and Horwitz, 1995).

An alternative in vitro model that simulates the in vivo conditions of dormancy was used to examine the effect of a cessation in active DNA replication as a result of oxygen starvation, which may occur within macrophages under growth suppressive conditions in inflammatory and necrotic tissues. Reactivation of latent disease is estimated to be the cause of the majority of tuberculosis cases reported annually in the United States (Wayne, 1976; 1994; Wayne and Sramek, 1994). Based on the model described by Wayne (1976, 1994), M. tuberculosis bacilli were allowed to settle through a self generated oxygen depletion gradient and uninduced cell-free extracts examined for spliced and precursor RecA protein. It has been proposed that DNA replication plays a role in the manifestation of ssDNA regions, which has been postulated to be the inducing signal for induction of the SOS response (D’Ari, 1985; Sassanfar and Roberts, 1990; Higashitani et al, 1995; Morel-Deville et al, 1996). It was anticipated that a cessation of DNA synthesis would be reflected in the levels of precursor protein and RecA detected, perhaps by a decrease in spliced RecA levels and a concomitant increase in unspliced precursor protein; particularly if the appearance of splicing does regulate display of the RecA phenotype. However, examination of cell-free extracts using affinity purified M tuberculosis RecA antibody revealed that basal RecA levels were unaffected and that precursor protein was absent. This model does have significant limitations. It only provides the opportunity to examine bacilli at a specific stage of this adaptation when the majority of cells have undergone the shift-down to dormancy. It was not possible to examine discrete changes in recA expression that may occur during this process. In addition, a slight change in growth rate may upset the equilibrium between replication and settling that is required to maintain the oxygen gradient supporting the shift down to dormancy (Wayne, 1976). Recently, an in vitro model permitting the examination of homogeneous populations of bacilli during discrete stages of this adaptation to low oxygen tensions has been described (Wayne and Hayes, 1996). This would enable examination of precursor and spliced RecA levels at specific stages of the shift-down from active replication to dormancy and may provide fiurther insight as to whether splicing in the maturation of RecA in M. tuberculosis has a role in vivo.

No compelling evidence is presented here to suggest that the appearance of splicing in the maturation of RecA in pathogenic mycobacteria regulates the display of the RecA phenotype, although its role in vivo is still unclear. The lack of detectable precursor protein in members of the M. tuberculosis complex suggests that splicing and translation may occur concurrently. Interestingly, this observation was unaffected by the growth stage of the culture at the point of harvesting, mild acidity or a cessation in active DNA replication by depleting available oxygen levels. A low basal level of RecA (38kDa) appears to be a common denominator of many mycobacterial species, the elevated levels seen in M. microti and

146 M. smegmatis are therefore intriguing; possible causes of this phenomenon will be discussed further below. In light of the difference in basal RecA levels between M microti and other members of the M. tuberculosis complex the response of mycobacteria to treatment with a DNA damage-inducing agent was of great interest.

8.3 The effect of exposure to DNA damaging agents on the kinetic profile for RecA induction and LexA cleavage in mycobacteria

The SOS response in both E. coli and B. subtilis is activated by physical and chemical agents that produce, either directly of indirectly, a variety of DNA lesions. I chose to investigate the effects of mitomycin C and ofloxacin — agents that generate the SOS inducing signal in two distinct ways. Mitomycin C introduces cross-links into DNA by the alkylation of guanines (Iyer and Szybalski, 1963; Borowski et al, 1990; Kumar et al, 1992) thereby blocking replication. These lesions are repaired by UvrABC endonuclease, which makes a cut either side of the cross-link creating a substrate for RecA- mediated recombinational repair (Sancar and Sancar, 1988; Sassanfar and Roberts, 1990; Kohn et al, 1992). Ofloxacin is a DNA gyrase inhibitor that blocks DNA replication and causes double strand breaks in DNA. Repair of these lesions requires the unwinding activity of RecBCD nuclease. On reaching a correctly orientated “Chi-site” this enzyme cuts one stand producing a stretch of ssDNA leading to the binding and subsequent activation of RecA (Snyder and Drlica, 1979; Phillips et al, 1987; Piddock and Wise, 1987; Ysem etal, 1990).

The recA genes of M. tuberculosis and M smegmatis were induced by both treatments. The time required to detect induction was protracted in M. tuberculosis (24 hours) compared to M. smegmatis (1-3 hours) suggesting that induction of RecA in M. tuberculosis is intrinsically delayed. This has recently been confirmed using a transcriptional fusion to the chloramphenicol-acetyltransferase reporter gene (CAT; Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997), evidence which suggests that host factors may play a significant role. Interestingly, treatment of M. tuberculosis with DNA damaging agents at an earlier stage of growth (ODgoonm of 0 2 - 0 3) resulted in an increase in RecA after 6-10 hours. The discordance between the rate of RecA induction in M. tuberculosis cultures induced at an OD^oonm of 0 6 and those at 0 2 - 0 3 is probably due to the increased susceptibility of M. tuberculosis at low cell density. Alternatively, it may be a result of an alteration in cell wall permeability and an increase in the percentage of free water within the cell that is frequently observed in prokaryotes at an early stage of growth. Although this observation differs from that described previously, where induction was not seen until 24 hours, it is still significantly slower than that seen in M. smegmatis and implies that that there is an intrinsic delay in RecA induction in M. tuberculosis following genetic insult. Curiously, treatment of M. microti with DNA damage-inducing agents did not result in an increase in RecA levels; the high RecA levels in this species remained unchanged over a 30 hour period. This suggests that there is no repression of the recA gene in the absence of DNA damaging agents, which may account for the high basal levels of RecA seen in this species. Alternatively, the cleaved LexA product may still be capable of repressing

147 recA expression. The simplest explanation for a lack of repression is the absence of a repressor molecule. However, a LexA homologue was identified in this species. Nevertheless a mutation in either the repressor molecule or the operator region affecting binding of the repressor would also result in a constitutive phenotype (see below).

In E. coli and B. subtilis the SOS response is co-ordinately regulated by the action of RecA and LexA (Little and Mount. 1982; Lovett et al, 1988; Wojciechowski et al, 1991). Examination of cell-free extracts by Western blot using a polyclonal antibody raised against M. tuberculosis LexA clearly demonstrates that a LexA homologue is present in M. tuberculosis, M. microti, M. bovis BCG and M smegmatis. This suggests that the basic regulatory elements of the SOS response in mycobacteria have been conserved. It is interesting to note that the basal level of this homologue among members of the M. tuberculosis complex mirrored those of RecA; elevated levels appear to be present in M. microti compared to M tuberculosis and M. bovis BCG. This provides further evidence to suggest that there is a lack of repression of SOS genes by LexA in M. microti and would tend to imply that there is a mutation affecting binding of the repressor to the operator site; this will be discussed further below. Examination of LexA levels in M. smegmatis revealed that they are comparable to that of M. tuberculosis and M. bovis BCG. Curiously, the LexA homologue present in M. smegmatis appears to be of a slightly higher molecular weight than that of the M. tuberculosis complex. Papavinasasundaram et al (1997) reported that a LexA homologue binds to the SOS-like box found upstream of the M. smegmatis recA gene. These observations argue against the suggestion that the high level of RecA in this species is due to a lack of repression. Alternative explanations will be discussed further below.

Treatment of M. smegmatis with DNA damage-inducing agents causes a decrease in LexA levels within 40 minutes. In both M. tuberculosis and M. microti levels of the LexA homologue remained constant for 6-7 hours subsequent to exposure before a decrease was observed. It is assumed that this decrease is a result of LexA cleavage, providing evidence for the activation of RecA by an inducing signal. An increase in LexA levels was reproducibly observed 4-5 hours following exposure in M. smegmatis and after 22-24 hours in M. tuberculosis and M. microti and is probably due to an increase in lexA expression which became apparent because cleavage was slow. The lack of detectable cleavage products may be a reflection of their instability in vivo. By comparison, cleavage of LexA in E. coli has been reported to occur within 90 seconds of exposure to DNA damaging agents and RecA induction within 10 minutes (Little, 1983; Sassanfar and Roberts, 1990). In B. subtilis, DinR repressor cleavage has been demonstrated within 10 minutes of exposure to DNA damaging agents (Wojciechowski et al, 1991; Lovett et al, 1994). B. subtilis has been reported to have a reduced rate of repressor inactivation compared to E. coli which is consistent with the observation that induction of RecA is slower in B. subtilis; occurring within 30 minutes (Lovett et al, 1988; 1994; Raymond-Denise and Guillen, 1992; Miller et al, 1996). In mycobacteria there is an identical correlation between the kinetic profiles of these two proteins; LexA cleavage precedes RecA induction. Interestingly the delay observed for RecA induction in M. smegmatis as a result of treatment with a DNA gyrase inhibitor (3 hours) compared to mitomycin (1 hour) has also been reported in E. coli and B. subtilis (Sanssafar and Roberts, 1990; Wojciechowski et al, 1991; Lovett et al, 1994). This may be

148 due to the different mechanisms by which each agent generates the inducing signal. The absence of a similar difference in M. tuberculosis may be a reflection of the intrinsic delay seen in this species.

Because M. tuberculosis and M. smegmatis differ significantly in the rate at which the SOS response is induced, one might anticipate that this would be manifested in a difference in their sensitivity to DNA damage-inducing agents. This expectation proved true. Examination of the relative sensitivities of M tuberculosis, M. microti and M. smegmatis to ultra-violet irradiation (UV), mitomycin C and ofloxacin revealed an increased sensitivity of M. tuberculosis to UV irradiation and mitomycin C compared to M. microti and M. smegmatis. Interestingly, there was no significant difference between any of the species examined following exposure to ofloxacin. It has been firmly established that treatments that perturb replication induce the SOS response possibly by the manifestation of ssDNA regions, which has been postulated to be the inducing signal (Walker, 1984; D’Ari, 1985; Higashitani et al, 1995). The overall rate of SOS induction has been reported to depend upon the rate of RecA activation and concomitant repressor cleavage, which are essentially determined by the density of lesions in the genome (Little, 1983; Lovett et al, 1988; Sassanfar and Roberts, 1990). Therefore, the difference in the rate of RecA induction together with mode of action of each of the DNA damage inducing treatments and the mechanisms by which the ensuing lesions are repaired may play a role in the relative sensitivities described above.

Mitomycin C covalently binds DNA forming inter-strand cross-links which constitute a complete block to replication (Iyer and Szybalski, 1963; 1964; Teng et al, 1989; Borowy-Borowski et al, 1990; Kumar et al, 1992; Basak, 1996). As discussed previously, the UvrABC endonuclease makes a cut either side of the cross-link, thereby creating a substrate for RecA-mediated recombinational repair (Sancar and Sancar, 1988; Sassanfar and Roberts, 1990; Kohn et al, 1992). Binding of RecA to ssDNA reversibly activates its coprotease fimction leading to repressor cleavage and concomitant induction of the SOS response (Sassanfar and Roberts, 1990; Lovett et al, 1994). Similarly, exposure to UV irradiation leads to the formation of intra-strand cross-links resulting in discontinuous replication and the production of ssDNA regions. RecA binds to these regions resulting in its activation and leading to induction of the SOS response and uvr mediated repair (see section 1.5.1). The protracted time for RecA induction in M. tuberculosis (24 hours) compared to M. smegmatis (1-3 hours) would lead to the former species exhibiting an inherently increased sensitivity. The comparative resistance of M. microti is probably a result of the elevated basal RecA levels observed in this species; if the recA gene is not repressed then expression is constitutively high (see below).

The comparable sensitivity of M. tuberculosis, M. microti and M. smegmatis to ofloxacin is paradoxical. Ofloxacin is a DNA gyrase inhibitor that leads to double strand breaks in DNA which are repaired by the combined action of RecBCD and RecA (Snyder and Drlica, 1979; (Taylor and Smith, 1980; 1995). Phillips et al, 1987; Piddock and Wise, 1987; Ysem et al, 1990). The precise reason behind the comparable sensitivity of these three species is unclear although it is perhaps noteworthy that concentrations bactericidal to mycobacteria in these investigations (1 25-2 5pg/ml) correlate well with

149 the concentration required to give maximum induction (lp,g/ml; Movahedzadeh et al, 1997b). Garcia- Rodriguez et al (1993) has reported that slow growers, including M. tuberculosis, are less susceptible than fast growing species to quinolones, the in vitro activity of which can be affected by a number of factors including the culture medium and incubation time. Mycobacteria for use in this assay were cultivated under optimal conditions (see section 2.1). It is possible that the different types of media used may have some bearing on the observed results since loss of quinolone activity due to binding to albumin and other proteins has been reported (Garcia-Rodriguez et al, 1993).

The work described here confirms that the key regulatory elements of the SOS response in M. tuberculosis and M smegmatis have been conserved (Davis et al, 1991 ; Movahedzadeh et al, 1997a; 1997b; Papavinasasundaram et al, 1997). Despite the intrinsic delay for induction of the SOS response in mycobacteria, particularly M tuberculosis, it seems reasonable to assume that SOS regulation in these species is analogous to that observed in E. coli and B. subtilis. There are several possibilities to account for this delay in induction of the SOS response; these include differences in the rate at which RecA is activated or in the rate of repressor inactivation. It is equally possible that it is a result of some as yet unidentified host factor.

The E. coli LexA repressor is inactivated by a specific cleavage reaction that requires activated RecA protein in vivo (Walker, 1984). LexA cleavage also occurs in vitro at physiological pH in the presence of RecA and two types of cofactor; ssDNA and a nucleotide triphosphate such as dATP (Little et al, 1984). RecA independent cleavage is also observed in vitro at alkaline pH, yielding fragments identical to those seen following RecA-mediated cleavage (Little, 1984). Both reactions cleave a specific peptide bond between Ala-84 and Gly-85 of LexA (Horii et al, 1981). Mechanistic considerations have led to the identification of two residues (Ser-119 and Lys-156) that are essential to the chemistry of cleavage (Slilaty and Little, 1987). In principle, the rate of such self-processing reactions can be very high due to the presence of all reactants at high, local concentrations. However, in some instances, for example LexA, the rates are slow and greatly stimulated by an external agent implying that these self-processing molecules undergo a slow intrinsic reaction but are capable of significant rate increases. Mutational studies utilising a repressor with an increased rate of in vivo cleavage revealed that LexA cleavage is not an intrinsically slow reaction but rather its rate has evolved to an optimum value (Smith et al, 1991 ; Roland et al, 1992). Too rapid a rate would prevent repressor accumulation, particularly in slow growing or stationary phase cells. Conversely, too slow a rate would make prompt inactivation of the repressor upon exposure to an inducing treatment difficult, thereby delaying or preventing an effective response. Induction of the 80S response in B. subtilis (Lovett et al, 1988; Raymond-Denise and Guillen, 1992; Miller et al, 1996) is delayed compared to E. coli (Little, 1983; Sassanfar and Roberts, 1990). This delay has been reported to be due to a difference in the rate of repressor inactivation (Miller et al, 1996). It is possible that a similar difference may account for the comparative delay in SOS induction observed in mycobacteria. The M. tuberculosis lexA gene has recently been cloned and sequenced (Movahedzadeh et al, 1997a); the residues identified as being important for cleavage have been conserved which suggests that there is a common mechanism of proteolytic cleavage. There is no conclusive information available

150 on the kinetics of M. tuberculosis LexA cleavage in vitro. However, preliminary investigations suggest that the rate of autodigestion is slow compared to E. coli LexA (Dr. P. Tenner, N.I.M.R. unpublished results).

A second possible explanation for the slow induction of the SOS response in mycobacteria is that the thick, lipid rich cell wall provides a formidable first line of defence. At this point it is worth remembering that the increased sensitivity of M. tuberculosis to mitomycin C compared to M. smegmatis demonstrates that a lack of penetration is not the cause of the slow induction seen in M. tuberculosis, since mitomycin C clearly enters the cell. However, the increased resistance of mycobacteria to deleterious agents such as acids and alkalis does support the suggestion that this stmcture provides some protection (Goren and Brennan, 1979; Brennan and Nikaido, 1995). The mycobacterial envelope comprises of a plasma membrane enclosed by a complex wall of covalently linked carbohydrates and lipids within a capsule composed of polysaccharides and proteins. The basic stmcture of the plasma membrane is similar to that of other bacteria and the stmcture of the cell wall is equally well defined. Stmctures such as arabinogalactan are common to the cell wall of both fast and slow growers whereas other stmctures exhibit inter-species differences (Asselineau and Lederer, 1950; Kanetsuna, 1968), for example type II or cl' mycolates are predominantly found in fast growers (Davidson et al, 1982; Minnikin et al, 1984; Hinrikson and Pfyffer, 1994; Brennan and Nikaido, 1995). Difference such as this need to be evaluated individually to assess their role in delaying entry of DNA damaging agents into the cell. In the example cited here the presence of a substantial proportion of a' mycolates in M smegmatis will decrease fluidity, and hence permeability of the cell wall. This decrease in permeability would affect entry into the cell of lipophilic agents such as ofloxacin and mitomycin C, which is difficult to reconcile with the delayed induction seen in M. tuberculosis compared to M. smegmatis. One component of the mycobacterial envelope that may play a significant role in decreasing its permeability is the capsule. This stmcture has been identified as an electron transparent zone surrounding mycobacteria grown both in vitro and in vivo although its precise stmcture and function is not clear (Hanks, 1961; Draper and Rees, 1973; Rastogi et al; 1986). Investigations addressing precisely when DNA damage occurs following treatment with a DNA damage-inducing agent may provide further insight as to whether permeability of the mycobacterial envelope influences the time taken for RecA induction in M. tuberculosis compared to M. smegmatis.

Alternatively, the delay in SOS induction may arise as a result of a defect in an accessory protein. Possible candidates include RecF, RecO, RecR and single-strand binding proteins (SSB). Following the production of ssDNA, the RecA protein promotes homologous pairing and strand exchange leading to the formation of recombination intermediates. The RecF, RecO, RecR and SSB proteins are believed to facilitate the loading of RecA onto ssDNA (Sawitzke and Stahl; 1992; Umezu et al, 1993) as well as participating at presynaptic steps in recombination (Lloyd et al, 1988; Clark, 1991; Wang et al, 1993; Kowalczykowski et al, 1994). It has been reported that RecF and RecR have a role in the resumption of replication at a stalled replication fork (Courcelle et al, 1997). Furthermore, the RecF protein is thought to assist in the cleavage of LexA by aiding RecA activation (Volkert and Hartke, 1984; Volkert et al, 1984; Walker, 1985), possibly by promoting the preferential interaction of RecF with dsDNA containing

151 ssDNA regions (i.e. gapped DNA; Griffin and Kolodner, 1990; Hedge et al, 1996). In addition to the roles of RecR and RecO described above these proteins have been shown to alleviate the inhibitory effects of SSB proteins on RecA catalysed joint molecule formation in vitro (Umezu et al, 1993; 1994; Umezu and Kolodner, 1994).

During normal DNA replication RecA carmot assemble its filament on SSB-bound ssDNA at replication forks. This behaviour is paralleled in vitro at low magnesium concentrations for RecA in the presence of SSB proteins and ssDNA (Kowalczykowski and Krupp, 1987; Lavery and Kowalczykowski, 1990). Conversely, in vitro observations indicate that RecA filament formation on SSB-bound ssDNA does occur at elevated magnesium and ATP concentrations (Barbe et al, 1983; Guerrero et al, 1985; Thresher et al, 1988). It is possible that inhibition of DNA synthesis in vivo produces a comparable situation. It has been proposed that modification of the SSB proteins or an increased tendency for RecA filament formation at an elevated ionic concentration may account for RecA polymerisation on SSB-bound ssDNA (Kuzminov, 1995). In addition to the inhibitory effect of SSB proteins mutational studies have provided evidence for a positive role in the activation of RecA (Vales et al, 1980; Whittier and Chase, 1981; Lieberman and Witkin, 1983; Chase et al, 1993). Although the precise nature of this involvement remains unclear ssb mutants that fail to induce normal levels of RecA after treatment with an inducing agent have been characterised. Interestingly a similar phenotype exists in cells which overproduce wild type SSB proteins (Johnson, 1977; Sevastopoulos et al, 1977; Meyer et al, 1982; Carlini and Porter, 1997). The majority of ssb mutations, for example 5jfe-l, are attributed to alterations in their ssDNA- binding affinity; an exception is 5sft-l 13 which is thought to involve altered protein-protein associations (Carlini and Porter, 1997). Additional evidence has revealed that SSB proteins also participate in at least two steps of genetic recombination, i.e. strand assimilation and branch migration (McEntee et al, 1980; West et al, 1982; Lieberman and Witkin, 1983). Interestingly, when RecA is present in limiting amounts helix-destabilising proteins such as SSB may play an auxiliary role in the pairing reaction (McEntee et al, 1980; Shibata et al, 1980); a stoichiometric interaction between RecA, SSB and ssDNA that promotes pairing with dsDNA has been demonstrated (Flory and Radding, 1982). RecA levels in M tuberculosis are low; it is not known whether this is a rate limiting factor, but the influence that SSB proteins may have on the rate of SOS induction in this genus is certainly intriguing. This is particularly so in view of the evidence which suggests that the partial complementation of M tuberculosis RecA in E. coli (Davis et al, 1991) may be due to an inability to displace SSB proteins (Kumar et al, 1996).

Another interesting possibility is that the rate of DNA replication may play a key role in determining the kinetics of the SOS response. A fimdamental taxonomic division of the genus Mycobacterium is tied to their growth rate under optimal conditions in vitro. Rapid growers yield visible colonies within 7 days whereas slow growers require 2-6 weeks. It has been reported that the slow growth rate reflects a retarded rate of macromolecular synthesis; the rate of RNA elongation in M. tuberculosis is 10-fold slower than that for E. coli (Harshey and Ramakrishnan, 1977). Similarly, the rate of DNA chain elongation in M. tuberculosis is approximately 11-fold slower than that of M. smegmatis and 13- to 18- fold slower than E. coli (Hiriyarma and Ramakrishnan, 1986). When DNA replication is perturbed in E.

152 coli the SOS response is induced (Casaregola et al, 1982b). The precise nature of the inducing signal is unknown, possible candidates include ssDNA (D’Ari, 1985; Sassanfar and Roberts, 1990; Higashitani et al, 1995; Morel-Deville et al, 1996; Shinagawa, 1996), ATP (Barbe et al, 1983; Guerrero et al, 1984; Thresher et al, 1988) and deoxyribonucleotides (Phizicky and Roberts, 1981; Menetski and Kowalczykowski, 1989). Preliminary investigations initiated using a wider spectrum of mycobacteria suggest that there may also be an intrinsic delay in RecA induction in the slow growers M. kansasii and M. avium (data not shown). Although these observations are far from conclusive at this stage it does seem likely that the rate of DNA replication may play a key role in determining the kinetics of the SOS response, perhaps by determining the time taken for DNA damage to manifest as ssDNA or as a result of an accumulation of unused deoxyribonucleotides.

8.4 Gene regulation in mycobacteria

One possible explanation to account for both the low basal RecA levels in M. tuberculosis and M. bovis BCG and for the intrinsic delay in RecA induction in M. tuberculosis may be related to promoter strength. Progress towards understanding gene regulation and expression in mycobacteria has been slow, as very few mycobacterial promoters have been studied so far. Examples include the rpsL gene of M. smegmatis (Kenney and Churchward, 1996); the M. bovis hsp60 gene and the M. leprae 28kDa and 18kDa genes (Dellagostin et al, 1995); the 16S rRNA genes of M leprae, M. smegmatis and M. tuberculosis and (Sela and Clark-Curtiss, 1991; Ji et al, 1994; Verma et al, 1994); the bla gene of M. fortuitum (Timm et al, 1994); the cpn-60 and 85kDa genes of M tuberculosis (Kong et al, 1993; Kremer et al, 1995) and the P^n promoter from M. paratuberculosis (Murray et al, 1992).

A high degree of homology has been shown to exist between the recA genes of M tuberculosis and M. smegmatis (91% identity; Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). This homology extends upstream of the recA coding region and includes two putative promoter sites identified on the basis of their homology with the consensus sequences of E. coli and E. coli promoters. The -10 and -35 elements of the a^°-like promoter are found at a different spacing (9bp) at these sites in mycobacteria compared to E. coli (17-21 bp, Rosenberg and Court, 1979). A similar observation was made for the M. tuberculosis 85kDa antigen a™-like promoter whose -10 and -35 sites are separated by 64bp (Kremer et al, 1995). It is not yet known whether mycobacterial RNA polymerases are able to recognise these motifs with a different spacing or whether the -35 region possesses an alternative consensus. However, examination of the available evidence suggests that the transcriptional signals of mycobacteria have features in common with Streptomyces (Bashyam et al, 1996). This genus is able to tolerate a larger variety of sequences at the -35 region due to the presence of multiple sigma factors that possess different or overlapping specificities for this region (Bibb and Cohen, 1982; Westpheling, 1985; Strohl, 1992).

153 In addition to the -10 and -35 sites, a third site slightly upstream of the -10 hexamer (described as an extended -10 motif) has been reported to affect promoter activity. Kenney and Churchward (1996) reported that, in the absence of a functional -35 hexamer, the M. smegmatis rpsL promoter is dependent on a specific TG dinucleotide found in this extended -10 region for normal promoter activity. Interestingly, an extended -10 region is also found in the promoters of M. bovis hsp60 (Stover et al, 1991) and M. leprae 16S rRNA (Sela and Clark-Curtiss, 1991) whereas it is absent in the promoter of the rpsL genes in M. tuberculosis, M. bovis and M. leprae, which possess a 5- of 6 -base pair match with the E. coli ct’°-35 site. The TG dinucleotide in an extended -10 region has been identified in B. subtilis (Helmann, 1995; Voskuil et al, 1995) and in approximately 16% of E. coli promoters (Harley and Reynolds, 1987; Kumar et al, 1993). It is thought to affect promoter utilisation and has been shown to be essential to weak promoters but of little benefit to strong promoters (Moran et al, 1982; Graves and Rabinowitz, 1986; Helmann, 1995; Tatti and Moran, 1995), which is particularly interesting since strong promoters have been reported to occur less frequently in M tuberculosis than M. smegmatis (DasGupta et al, 1993).

Examination of the putative promoter site with homology to the E. coli a™ promoter in M. smegmatis recA revealed that an extended -10 region with a conserved TG dinucleotide is present immediately upstream of the -10 hexamer, whereas it is absent in M tuberculosis. This suggests that promoter strength may be an important determinant of the difference in basal RecA levels and rate of induction of the SOS response between M. tuberculosis and M. smegmatis. However, analysis of this putative promoter site upstream of the recA genes in M. tuberculosis and M. smegmatis using the chloramphenicol- acetyltransferase (CAT) reporter gene revealed that the uninduced expression of recA is similar in both species. Moreover, treatment with a DNA damage inducing agent led to a 2.5-fold increase in expression of both recA fusions (Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). This suggests that there is no significant difference in intrinsic promoter strength between M. tuberculosis and M. smegmatis recA. A second putative promoter site with homology to the E. coli promoter site is located further upstream in both M smegmatis and M tuberculosis which lacks an extended -10 motif. This finding is not unexpected since the B. subtilis promoter also lacks this motif (Voskuil et al, 1995). However, this second putative promoter site overlaps the Cheo box, which has been implicated in LexA binding in both M. smegmatis and M. tuberculosis (Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997) and may be of more consequence than the proximal putative promoter site to regulation of this gene. A more rigorous examination of both these putative promoter sites may provide further insight into their role in the regulation of the recA gene.

8.5 Expression ofrecA in M. microti is constitutive

The high basal levels of RecA in M microti are unaffected by treatment with a DNA damage-inducing agent, which suggests that that the recA gene is not repressed in the absence of DNA damage. Alternatively, the cleaved LexA product may sill be capable of repressing recA expression. Normally, the repressor protein LexA binds to a specific sequence (the SOS box) upstream of the genes it regulates to

154 prevent transcription (Brent and Ptashne, 1980; 1981). A constitutive phenotype might be expected to arise if (i) the repressor molecule is absent, (ii) the SOS box is altered to prevent binding of the repressor, or (iii) the repressor molecule is defective in binding to the SOS box. Since M. microti clearly possess a LexA homologue that exhibits a marked decrease following treatment with a DNA damage inducing agent the first of these possibilities was rapidly dismissed. In E. coli, /ex(Def) mutants have been characterised, these produce little or no functional LexA and lead to the constitutive expression of LexA- repressed genes (Mount, 1977; Pacelli et al, 1979; Little and Mount, 1982; Krueger et al, 1983; Thliveris and Mount, 1992; Haijema et al, 1996). Mutations in the operator region leading to a constitutive phenotype have also been characterised (recAo; Clark, 1982; Ginsburg et al, 1982; Little and Mount, 1982; Thliveris and Mount, 1992; Winterling et al, 1997).

The promoter upstream of the M microti recA gene is inducible in M. smegmatis following treatment with DNA damage inducing agents. Examination of this region using a plasmid containing the relevant fragment upstream of a promoterless CAT (chloramphenicol-acetyltransferase) gene revealed a 2- to 4-8- fold increase in expression. The level of induction varied with size of the insert and inducing treatment applied; the shorter 350bp insert exhibited a 4-8-fold induction following treatment with mitomycin C and 2-fold for ofloxacin. Treatment of the plasmid carrying the larger 700bp insert with either ofloxacin or mitomycin C resulted in a 2-3-fold increase in expression. These induction ratios are comparable to some isolated E. coli DNA damage inducible promoters fused to galactokinase (3-6- to 8-5-fold; Lewis et al, 1992) or lacZ (5- to 7-fold) in Gram-negative bacteria treated with an inducing agent (Weisemann et al, 1984; Fernandez de Henestrosa et al, 1991). Similarly, fusions of B. subtilis recA with a xylE reporter gene gave induction ratios of 5-5- to 10-fold (Gassel and Alonso, 1989; Raymond-Denise and Guillen, 1992) and Lovett et al (1988) demonstrated a 5-fold induction of RecA by a densitometer scan of immmunoblots. More recently, an induction ration of 2-5- to 2- 8-fold has been demonstrated for M. smegmatis recA using both CAT SiVidxylE reporter genes (Durbach et al, 1997; Papavinasasundaram et al, 1997).

Examination of equivalent constructs with the relevant fragments from M. tuberculosis revealed that while there was no difference with respect to the size of the insert, the induction ratio following treatment with mitomycin C (3-4- to 3-6-fold) was also approximately double that of ofloxacin (1-8- to i-9-fold). This variation in induction ratios with respect to inducing treatment may be a reflection of the delay observed in RecA induction in M. smegmatis between ofloxacin (3 hours) and mitomycin C (1 hour), which results in a difference in the rate of RecA accumulation. Curiously, Movahedzadeh et al (1997b) reported a 2-5-fold increase in expression from the plasmid with the 700bp insert compared to a 4-3-fold increase for the 350bp insert; there was no difference with respect to the inducing treatment applied. The reason for the variation between these two investigations is not clear.

Attempts to resolve the apparent paradox between the lack of inducibility of the M. microti recA gene in M. microti and that observed for the M. microti recA gene in M. smegmatis have not been successful. It has proved difficult to efficiently transform M. microti with plasmids containing transcriptional fusions of

155 the recA promoter regions from either M. smegmatis or M. microti to the CAT gene. More recently the effect of electroporation temperature, biochemical pre-treatment of cells and growth stage of the culture on the efficiency of transformation in slow-growing mycobacteria has been investigated (Wards and Collins, 1996). These authors reported that the efficiency of transformation in slow growers such as M. tuberculosis and M. intracellulare increased at temperatures up to 37°C. Conversely, the efficiency of transformation in fast-growers such as M. smegmatis decreased with increasing temperatures. So far it has proved impossible to efficiently transform M microti at any temperature.

Examination of the M. microti recA gene and 700bp of upstream sequence revealed almost 100% DNA- DNA relatedness with equivalent sequences in M. tuberculosis. Only two base changes were evident, both insertions in the upstream sequence (526bp and 478bp upstream of the recA gene). There are precedents for this high degree of homology between members of the M. tuberculosis complex. Imaeda (1982) demonstrated that members of the M. tuberculosis complex are genetically very similar (85-100% homology); restriction enzyme analysis revealed that the distribution of DNA fragments produced by a single enzyme, for example EcoRI, EcoRV or Kpnl, are indistinguishable among representative strains. Furthermore, they have highly conserved DNA sequences in several regions. The M. tuberculosis 16S rRNA sequence differs in only one position from M. bovis and M. microti (Kempsell et al, 1992) and both the dnaJ genes (Takewaki et al, 1994) and the 16S to 23S rDNA internal transcribed spacer are completely conserved (Kirschner et al, 1993; Frothingham et al, 1994). Similarly, both M. tuberculosis and M. microti are indistinguishable by PCR for insertion sequences IS 1081 and IS6110 and the gene MPB70 (Liebana et al, 1996). It is perhaps noteworthy that these observations are not a general property of all mycobacterial complexes. The M. avium complex dnaJ genes have only 94% similarity (Takewaki et al, 1994) and the 16S to 23S rDNA internal transcribed spacer differs by 21 nucleotides (Frothingham et al, 1994).

Frothingham et al (1994) have offered several explanations for this high degree of sequence conservation. The intracellular nature of these species may lead to a lower total number of organisms being present resulting in fewer opportunities for mutation. In addition, the opportunity for recombination with other species or strains would be limited and may result in genetic homogeneity. It is assumed that members of the M. tuberculosis complex are well adapted to the host environment, which presumably remains relatively constant; mutations would only be conserved if they provided a selective advantage. Another intriguing possibility is that M tuberculosis, M. microti, M. bovis and M. africanum have diverged as separate species only very recently allowing little time for divergent mutations to accumulate.

Homology between these two species includes a motif (120bp upstream) with an exact match to the SOS box (GAAC-N 4-GTTC) identified in Bacillus subtilis by Cheo ei a/ (1991) and a second motif 275bp upstream possessing 6/8 matches (GCAC-N 4-GATC). The lack of inducibility of the M. microti recA gene in its native host is intriguing, particularly since the evidence presented here clearly demonstrates that it possesses an inducible promoter and a putatively functional Cheo box motif (GAAC-N^-GTTC) in its upstream sequence. An identical sequence is located upstream of the recA genes in M. tuberculosis, M.

156 smegmatis and M. leprae and binding of a LexA homologue to this consensus has been demonstrated in all three species by gel mobility assays and DNase I footprinting (Durbach et al, 1997; Movahedzadeh et al, 1997a; 1997b; Papavinasasundaram et al, 1997). Interpretation of the evidence so far clearly supports the suggestion that the basic regulatory elements of the SOS response in M microti have been conserved. Interestingly, gel mobility assays performed to directly assess whether the M microti LexA homologue identified by Western blot binds to the Cheo box motif (GAAC-N 4-GTTC) identified upstream of its recA gene revealed that this protein has unusual binding properties. A retarded DNA fragment comparable to that produced with purified M. tuberculosis LexA was present with cell-free extracts of both M. microti and M. tuberculosis. In addition, a second more intense DNA fragment that migrates further into the gel was consistently observed with M microti cell-free extracts. Binding to the Cheo box motif is specific; no retarded fragments were seen when an oligonucleotide mutated only in the Cheo box motif was used. Furthermore, utilisation of increasing amounts of unlabelled wild type oligonucleotide as competitor reduced the proportion of retarded, labelled DNA whereas equivalent amounts of unlabelled mutated oligonucleotide did not. Further investigation revealed that the retarded fragments were a result of the specific binding of a LexA homologue to the Cheo box motif; pre-incubation of M. microti cell-free extract with an antibody raised against M. tuberculosis LexA drastically reduced the proportion of retarded, labelled DNA.

The M. microti LexA homologue appears to have altered DNA binding properties compared to its counterpart in M. tuberculosis and M. smegmatis (Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). In E. coli the LexA repressor is a two-domain structure comprising the DNA-binding domain (N- terminal residues 1-84; Little and Hill, 1985; Hurstel et al, 1986; Schnarr et al, 1991; Thliveris and Mount, 1992) and the oligomerization domain (C-terminal residues 85-202). The structure of the N- terminal domain has been determined by nuclear magnetic resonance (Fogh et al, 1994) revealing that LexA is a member of a family of proteins which recognise DNA through a helix-tum-helix motif (Harrison and Aggerwal, 1990). Mutagenesis assays have identified several E. coli LexA mutants defective in their DNA binding ability (Thliveris and Mount, 1992).

One explanation for the unusual binding properties of M. microti LexA is that mutations at positions in the DNA-binding domain that form no direct protein-DNA contacts affect “occupancy” of the operator. There may be a defect affecting the stability or strength with which the repressor binds the operator, which may account for the variation in the comparative intensity of the two retarded fragments formed with M. microti LexA in successive experiments. Defective repressors that cannot interact with the operator sequence tightly have been identified in the E. coli lac repressor (Kleina and Miller, 1990). These are mainly due to incorrect folding of the protein and are usually a result of an alteration in the N- terminal (DNA-binding) domain. Mutations in the DNA binding of the E. coli LexA repressor that affect dimerization have also been identified; Knegtel et al (1995) have identified several residues in the DNA binding domain that enhance binding by stabilising the dimerization interface.

157 Mutations in the C-terminal domain are also of consequence. Kleina and Miller (1990) have characterised several mutations in the E. coli lac repressor that lead to a constitutive phenotype. In particular mutations affecting subunit contacts can alter protein function; for example a defect affecting folding of the C- terminal domain and hence oligomerization. Golemis and Brent (1992) have postulated that protein- protein interactions of E. coli LexA are of equal significance to protein-DNA interactions. They suggested that the carboxy-terminal domain of a LexA monomer promotes a spatially precise association with a second monomer attached to the operator site; LexA monomers lacking this interaction are more likely to be displaced from the DNA. Like many prokaryotic repressors (Anderson et al, 1985; Jordan and Pabo, 1988; Wolberger et al, 1988) LexA binds as a dimer to an operator site that possesses dyad symmetry (Brent and Ptashne, 1981; Wertman and Mount, 1985; Cheo et al, 1991; 1993; Fernandez de Henestrosa et al, 1997; Tapias et al, 1997). Examination of these operator sites in E. coli, B. subtilis and M. tuberculosis by DNase I and hydroxy radical footprinting have shown that both half sites are bound by LexA in a pattern which appears symmetrical (Brent and Ptashne, 1981; Little et al, 1981; Hurstel et al, 1986; Kim and Little, 1992; Miller et al, 1996; Movahedzadeh et al, 1997b). In addition, mutagenesis assays in E. coli have provided compelling evidence for the interaction of at least one LexA monomer with each half site of the recA operator (Thliveris and Mount, 1991). Based on mutational and biochemical evidence a model for the sequential, co-operative binding of two LexA monomers in a two-step process has been proposed for E. coli (Schnarr etal, 1985; 1988; Thliveris and Mount, 1991; Golemis and Brent, 1992; Kim and Little, 1992). Binding of a second monomer to the operator involves both protein-DNA and protein-protein contacts (Kim and Little, 1992). It is generally believed that the co-operative binding observed for the second monomer arises from protein-protein interactions and that the structure of the protein may change upon binding to DNA (Brennan et al, 1990; Sauer, 1990) allowing additional protein-protein contacts to form or stabilising a conformation not favoured in solution. Alternatively, the first bound monomer may confer a conformational change on the DNA-binding site that increases the affinity of this second site for the second monomer (Kim and Little, 1992).

The possibility that the unusual binding properties of the M. microti LexA homologue were due to monomer as opposed to dimer binding of the repressor to the operator site was examined. Kim and Little (1992) have reported that binding of E. coli LexA to an intact operator has a 1000-fold greater affinity than binding to a half-site. We performed similar investigations using gel mobility assays with the relevant labelled oligonucleotides and a 100-fold excess of the LexA concentration required to completely retard the wild type oligonucleotide. Both the concentration of the M tuberculosis LexA stock solution and the maximum volume of each gel shift reaction were limiting factors with respect to the maximum LexA concentration used in these reactions. The presence of a mutation in a half-site of the Cheo box motif did not result in the appearance of retarded DNA-protein complexes comparable to those observed with M. microti cell-free extract. A retarded fragment that migrates only a small distance into the gel was present in reactions containing the wild type or half-site mutated operators together with a 10-fold excess of purified M. tuberculosis LexA. An identical band was identified in reactions containing a completely mutated operator in the presence of a 100-fold excess of purified M. tuberculosis LexA. It is probable that these retarded DNA fragments are due to the formation of a non-specific complex between DNA and a multimeric form of LexA,

158 for example a tetramer. (Schnarr et al, 1985; 1988). The E. coli LexA repressor dimerises only weakly in solution in the absence of DNA with a dissociation constant of 15-50fiM (Schnarr et al, 1985; 1988; Kim and Little, 1992). Cellular concentrations of E. coli LexA have been estimated at 1-2 pM with approximately 20% present as unbound protein (Moreau, 1987; Sassanfar and Roberts, 1990). This implies the most of the unbound cellular repressor is monomeric. Assuming a similar dissociation constant in the presence of DNA it is plausible that as LexA concentration increases the tendency of LexA to form a multimeric complex increases. No specific information regarding the cellular concentration and energetics of dissociation is available for a mycobacterial LexA. However, it seems plausible to suggest that a similar mechanism may account for the observations described here.

The evidence presented above suggests that M. microti LexA does not bind to a half-site. However, these experiments are far from conclusive and further investigation is recommended. DNase I and hydroxy radical footprinting (Lin and Shiuan, 1995; Miller et al, 1996; Movahedzadeh et al, 1997) may provide a more precise indication of the bases important for binding of the M microti LexA homologue. Moreover this technique would confirm whether or not all the sequence necessary for binding is present in the wild type oligonucleotide. This is of particular interest since Haijema et al (1996) reported a partially retarded fragment for a construct of B. subtilis dinR lacking part of the sequence necessary for binding of DinR. A second explanation for the lack of induction of M microti RecA following treatment with a DNA damaging agent is that the cleaved LexA product is still able to repress expression of the recA gene. Examination of cell-free extracts prepared from M. microti following treatment with a DNA damage- inducing agent by DNase I footprinting may also provide further evidence to support or dismiss this suggestion.

Because the physiological consequences of a lack of repression by LexA (such as filamentation and a reduced viability or growth rate) are not obviously manifest in M. microti it is worth considering other plausible explanations for the constitutive phenotype. One possibility is that a functional LexA homologue is present but it is unable to bind efficiently. It has been reported that the intracellular environment plays a role in fine tuning repression of the SOS system by LexA; binding affinity and cooperativity have been demonstrated to be dependent on both pH and the specific anion/cation composition and concentration (Shaner and Gaissarian, 1996; Relan et al, 1997). Alternatively, the M. microti LexA homologue may be preferentially displaced from the operator site. This phenomenon has been reported in E. coli-, the protein HU has been shown to displace LexA from its binding sites on the lexA, recA and sulA genes (Preobrajenskaya et al, 1994). However, this is unlikely to account for the unusual binding properties of M microti lexA due to the genetic homogeneity reported for the M. tuberculosis complex (Imaeda, 1982; Kempsell et al, 1992; Kirschner et al, 1993; Frothingham et al, 1994; Takewaki et al, 1994) and the absence of any evidence for a similar phenomenon in any other members. Another possibility is that LexA does not regulate the M. microti recA gene. Not all bacteria that possess an E. co/i-like SOS system have the same genes under LexA control. For example, the uvrB gene of E. coli is damage inducible (Fogliano and Schendel, 1981) but that of Psuedomonas aeruginosa is not (Rivera et al, 1996). A second example is found in the Anabeana variabilis recA gene, this has no

159 apparent E. coli SOS-like box but is induced following exposure to UV irradiation and mitomycin C (Owttrim and Coleman, 1989). More recently, Black et al (1998) have demonstrated that an SOS-like system which is induced in response to DNA damage is absent in Neisseria gonorrhoeae; the recA, uvrA and uvrB genes are not damage inducible and do not possess a LexA binding site (Black et al, 1995; 1997; 1998).

Another intriguing possibility is that the constitutive phenotype is due the presence of a host factor that affects LexA binding or acts as a positive regulator. Evidence provided by lacZ fusions with Rhodobacter capsulatas recA suggests that the protein controlling this gene is a positive regulator. The level of recA expression in cells with a doubly mutated SOS-like box was equivalent to that of untreated wild type cells and not that of treated wild type cells which would be expected if the controlling protein were a repressor (Fernandez de Henestrosa et al, 1997). A few more examples of damage inducible genes under positive control are known, among them the Erwinia carotovora ctv and pnlA genes encoding carotovoricin and pectin lyase respectively (McBvoy et al, 1992). The product of rdgB, whose expression is blocked by rdgA, stimulates transcription. The repressor RdgA is cleaved by activated RecA. A similar situation exists for the transcriptional regulation of the various types of pyocins by Psuedomonas aeruginosa (Matsui et al, 1993). However, it is worth noting that neither the pectin lyase nor the pyocin genes possess a LexA binding site in their promoter regions.

8.6 Future work

The essential regulatory elements of the SOS response appear to have been conserved in mycobacteria. The evidence suggests that activation of RecA by an inducing signal following treatment with a DNA damaging agent results in LexA cleavage and RecA induction, although there appears to be a delay in induction of the SOS response in mycobacteria compared to E. coli and B. subtilis. This is in addition to the intrinsic delay apparent for RecA induction in M tuberculosis (24 hours) compared to M. smegmatis (1-3 hours). It is not yet clear whether these differences in response time are due to differences in the rate at which RecA is activated, differences in the rate of repressor inactivation or to some as yet unidentified host factor. It would be interesting to precisely identify the point at which DNA damage occurs following treatment with a DNA damaging agent. Clarification as to whether the mycobacterial envelope provides a permeability barrier which influences the rate at which DNA damaging agents gain access to the cell may aid in determining whether the rate of RecA activation contributes to the intrinsic delay seen in this genus. In addition, elucidation of the kinetics of RecA-mediated LexA cleavage and autodigestion may provide a clearer indication of whether the rate of repressor inactivation is a significant factor accounting for the intrinsic delay observed in M. tuberculosis compared to M. smegmatis. A further difference in the response of mycobacteria to treatment with a DNA damage-inducing agent has also been revealed by these studies: the recA gene of M. microti appears to be expressed constitutively. The evidence suggests that there is a lack of repression in the absence of DNA damaging agents and that this may be a result of the unusual binding properties of the M. microti LexA homologue. Clearly, the differences highlighted

160 with regard to the kinetics of RecA induction in these three species emphasises the requirement for caution when choosing a non-pathogenic model for the examination of the regulation of M. tuberculosis genes since both M. smegmatis and M. microti are frequently used alternatives.

The role that splicing may play in the maturation of RecA among members of the M. tuberculosis complex and M leprae is still unclear. The in vitro evidence presented here suggests that splicing is a rapid process that plays no role in regulating the display of the RecA phenotype. More recently, the construction and complementation of an M. smegmatis recA mutant with the M. tuberculosis recA gene has revealed that the presence of an intein does not affect UV sensitivity or recombination in vitro in M. smegmatis (Papavmasasundaram et al, 1997b). The role that this genetic element may play in vivo remains elusive.

Basal RecA levels have been qualitatively compared in a wide spectrum of mycobacteria. Extending these investigations to examine the effect of treatment with a DNA damage-inducing agent would determine the impact that the rate of DNA replication may or may not have on the observed kmetics of RecA induction. The use of quantitative RT-PCR may yield more definitive information by providing a quantitative analysis of recA expression in mycobacteria in uninduced and induced cells.

The identification of further DNA damage inducible {din) genes in mycobacteria would enable functionally important features of the operator/promoter sites to be established and may also be particularly illuminating in M. microti with respect to regulation of the SOS response. The fact that filamentation and a high mutagenesis rate are not manifest in this species would suggest that the umu locus and the sulA gene are subject to some form of regulation. Furthermore, elucidation of the sequences involved in gene regulation and characterisation of the transcriptional signals should contribute towards a better understanding of gene regulation and expression generally in mycobacteria.

The DNA-binding properties of E. coli LexA have been extensively analysed and interactions between the N-terminal domain and the conserved motif (CTGT) characterised (Ottleben et al, 1991; Schnarr et al, 1991; Thliveris and Mount, 1992). In addition, non-specific protein-DNA contacts of the repressor with the intervening AT rich sequence have been shown to modulate repressor affinity (Wertman and Mount, 1985; Knegtel et al, 1995). The precise interaction of a mycobacterial LexA homologue with specific bases in the operator site and the effect of mutations in this site on protein-DNA contacts are still unknown. The importance of the conserved motif (GAAC-N 4-GTTC) has been established for B. subtilis, M. tuberculosis and M. smegmatis (Haijema et al, 1996; Miller et al, 1996; Winterling et al, 1997; Movahedzadeh et al, 1997b; Papavinasasundaram et al, 1997). Investigations have shown that the substitution of all eight bases in the Cheo box motif eliminates LexA binding. It is not clear whether all these bases are equally important for recognition by a mycobacterial LexA although it is known that LexA will bind to a Cheo box motif with one mismatch (Movahedzadeh et al, 1997a) but not one with two (Movahedzadeh, 1996). In addition, a clearer understanding of the potential role of non-specific

161 protein-DNA contacts with the nucleotide sequence flanking the Cheo box motif and their impact on repressor binding affinity remains an important goal.

In E. coli, the structure and sequence of the LexA DNA binding domain is highly conserved with other Gram-negative bacteria, for example; Salmonella typhimurium, Erwinia carotovora, Psuedomonas aeruginosa, Psuedomonas putida and H. influenzae (Garriga et al, 1992; Winterling et al, 1997). Considering that M tuberculosis LexA recognises a different sequence (GAAC-N^-GTTC; Davis et al, 1991) to E. coli (CTGT-Ng-ACAG; Wertman and Mount, 1985) the lack of conservation in the N-terminal domain is not unexpected (Movahedzadeh et al, 1997a). However, in view of the homology between the LexA binding sites in mycobacteria (Davis et al, 1994) and B. subtilis (Cheo et al, 1991; 1993) the lack of sequence conservation in the N-terminal domain of their respective repressors is surprising (Movahedzadeh et al, 1997a; Winterling et al, 1997). A comparative sequence analysis of the N-terminal domain of other mycobacterial LexA homologues may yield useful information regarding the identification of bases essential for protein-DNA interactions. Elucidation of the sequence of further Gram-positive LexA homologues may provide further insights into the bases important for both protein- DNA and protein-protein contacts at the repressor-binding site. The M. microti lexA sequence may be particularly useful in this respect since the evidence suggests that there is a mutation in this protein affecting binding to the SOS box.

Further investigation into the response of mycobacteria to DNA damage will enhance our understanding of the survival of pathogenic mycobacteria within the intracellular environment. Studies aimed at identifying other damage inducible genes may provide further insight into their regulation and may be important for a clearer understanding of pathogenicity. The ability to permanently modify specific genes using homologous recombination would make a major contribution to the genetic analysis of mycobacteria and may facilitate the identification of virulence factors. In addition, it would provide a powerful tool facilitating the stable expression of foreign genes of immunological interest in M. bovis BCG, which may aid in the development of effective recombinant vaccines.

162 Re f e r e n c e s

Abbas, A.K., Lichtman, A.H. and Pober, J.S. (1994). Antibodies and antigens. In Cellular and Molecular Immunology, chapter 3, 2"^ edition. W.B. Saunders Company.

Adams, D.E. and West, S.C. (1995). Relaxing and unwinding on Holliday: DNA helicase-mediated branch migration. Mutat. Res. 337, 149-159.

Aigle, B., Holl, A C., Angulo, J.F., Leblond, P. and Decaris, B. (1997). Characterisation of two Streptomyces ambofaciens recA mutants: identification of the RecA protein by Immunoblotting. FEMS Microbiol. Lett. 149, 181-187.

Aldovini, A., Husson, R.N. and Young, R.A. (1993). The uroA locus and homologous recombination in Mycobacterium bovis BCG. Journal o f Bacteriology 175(22), 7282-7289.

Anderson, A.B., Ljungqvist, L. and Olsen, M. (1990). Evidence that protein antigen b of Mycobacterium tuberculosis is involved in phosphate metabolism. Journal Gen. Microbiol. 136, 477-480.

Anderson, A.B., Worsaae, A. and Chaparas, S.D. (1988). Isolation and characterisation of recombinant Lambda g tll bacteriophage expressing eight different antigens of potential immunological relevance. Infect. Immun. 56, 1344-1351.

Anderson, K.K. and Benjamin, Y. (1979). The effects of organic solvents and temperature on desorption of yeast 3-phosphoglycerate kinase from immunoadsorbent. Journal Immunol. Methods 25,375-381.

Anderson, J.E., Ptashne, M. and Harrison, S.C. (1985). A phage repressor-operator complex at 7 resolution. Nature 316, 596-601.

Anderson, W.F., Ohlendorf, D.H., Takeda, Y. and Matthews, B.W. (1981). Structure of the cro repressor from bacteriophage X and its interaction with DNA. Nature 290, 754-758.

Angerer, A., Enz, A., Ochs, M. and Braun, V. (1995). Transcriptional regulation of ferric citrate in Escherichia coli K-12. Feci belongs to a new subfamily of a ’°-type factors that responds to extracytoplasmic stimuli. Molec. Microbiol. 18, 163-174.

Angov, D. and Camerini-Otero, R.D. (1994). The recA gene from thermophile Thermus aquaticus. Journal

163 of Bacterial. 176, 1405-1412.

Angulo, J.F., Moreau, P.L., Maunory, R., Laporte, J., Hill, A.M. Bertolotti, R. and Devoret, R. (1989). KIN, a mammalian nuclear protein immunologically related to Escherichia coli RecA protein. Mutat. Res. 217, 123-134.

Arnosti, D.N. and Chamberlin, M.J. (1889). Secondary a factor controls transcription of flagella and chemotaxis genes in Escherichia coli. Proc. Natl. Acad. Sci. USA 86, 830-834.

Asselineau, J. and Lederer, E. (1950). Structure of the mycolic acids of mycobacteria. Nature 166, 782-783.

Balasubramanian, V., Pavelka. Jr., M S., Bardarov, S.S., Martin, J., Weisbrod, T.R., McAdam, R. A., Bloom, B. R. and Jacobs Jr., W.R., (1996). Allelic exchange in Mycobacterium tuberculosis with long linear recombination substrates. Journal o f Bacterial. 178(1), 273-279.

Baliga, R., Singleton, J.W. and Dervan, P.B. (1995). RecA oligonucleotide filaments bind in the minor groove of double-stranded DNA. Proc. Natl. Acad. Sci. USA 92, 10393-10397.

Barbe, J., Villaverde, A. and Guerrero, R. (1983). Evolution of cellular ATP concentration after UV-mediated induction of SOS system in. Escherichia coli. Biochem. Biophys. Res. Commun. 117, 556-561.

Bardarov, S., Kriakov, J., Carrière, C., Yu, S., Vaamonde, C., McAdam, R.A., Bloom, B.R., Hatfull, G.F. and Jacobs Jr., W.R. (1997). Conditionally replicating mycobacteriophages: A system for transposon delivery to Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 94,10961-10966.

Basak, J. (1996). On the nature of the adaptive response induced by mitomycin C in Vibrio cholerae OGAWA 154 cells. Biochem. Biophys. Comms. 220,509-514.

Bashyam, M.D., Kaushal, D., Dasgupta, S.K. and Tyagi, A.K. (1996). A study of mycobacterial transcriptional apparatus: Identification of novel promoter elements. Journal o f Bacterial. 178(16), 4847-4853.

Battista, J R., Ohta, T., Nohmi, T., Sun, W. and Walker, G.C. (1990). Dominant negative umuD mutations decreasing RecA mediated cleavage suggest roles for intact UmuD in modulation of the SOS response. Proc. Natl Acad. Sci. USA 87, 7190-7194.

Baulard, A., Jourdan C., Mercenier, A. and Locht, C. (1992). Rapid mycobacterial plasmid analysis by electroduction between Afyco&ac/m'MW sp. dnà Escherichia coli. Nucleic Acid. Res. 20(15), 4105.

164 Baulard, A., Kremer, L. and Locht, C. (1996). Efficient homologous recombination in fast growing and slow growing mycobacteria. Journal o f Bacterial. 178(11), 3091-3098.

Bayles, K.W., Brunskill, E.W., landolo, J.J., Hruska, L.L., Huang, S., Pattee, P.A., Smiley, B.K. and Yashin. R.E. (1994). A genetic and molecular characterisation of the recA gene from Staphylococcus aureus. Gene. 147, 13-20.

Bertrand-Burgraff, E., Hurstel, S., Duane, M. and Schnarr, M. (1987). Promoter properties and negative regulation of the uvrA gene by the LexA repressor and its amino terminal DNA binding domain. Journal Molec. Biol. 193, 293-302.

Bibb, M«F. and Cohen, S.N. (1982). Gene expression in Streptomyces: Construction and apphcation of promoter-probe plasmid vectors in Streptomyces lividans. Molec. Gen. Genet. 187,265-277.

Biniszkiewicz, D., Cesnaviciene, E. and Shub, D.A. (1994). Self-splicing group-I intron in cyanobacteria initiator methionine tRNA: Evidence for lateral transfer of introns in bacteria. The EMBO Journal 13(19), 4629- 4635.

Black, C.G., Fyfe, J.A.M. and Davies, J.K. (1995). A promoter associated with the neisserial repeat can be used to transcribe the uvrB gene from. Neisseria gonorrhoeae. Journal Bacterial. 177, 1952-1958.

Black, C.G., Fyfe, J.A.M. and Davies, J.K. (1997). Cloning, nucleotide sequencing and transcriptional analysis of the uvrA gene from Neisseria gonorrhoeae. Molec. Gen. Genet. 254,4769-485.

Black, C.G., Fyfe, J.A.M. and Davies, J.K. (1998). Absence of an SOS-like system in Neisseria gonorrhoeae. Gene 208,61-66.

Bloom, B.R. and Fine, P.E.M. (1994). The BCG experience: Implications for future vaccines against tuberculosis. In Tuberculosis. Edited by Bloom, B.R. American Society for Microbiology Press.

Bloom, B.R. and M urray C.J.L. (1992). Tuberculosis: Commentary on a reemergent küler. Science 257,1055- 1064.

Borowy-Borowski, H., Lipman, R. and Tomasz, M. (1990). Recognition between mitomycin C and specific DNA sequences for cross-link formation. Biochemistry 29,2999-3006.

165 Bremer, M.C.D., Gimble, F.S., Thorner, J. and Smith, C.L. (1992). VDE endonuclease cleaves Saccharomyces cerevisiae genomic DNA at a single site: physical mapping of the VMAl gene. Nucleic Acids Res. 20(20), 5484.

Brennan, P.J. and Nikaido, H. (1995). The envelope of mycobacteria. Ann. Rev. Biochem. 64,29-63.

Brent, R. (1982). Regulation and autoregulation by LexA protein. Biochimie 64, 565-569.

Brent, R. and Ptashne, M. (1980). The lexA gene product represses its own promoter. Proc. Natl. Acad. Sci. USA 77(4), 1932-1936.

Brent, R. and Ptashne, M. (1981). Mechanism of action of the LexA gene product. Proc. Natl. Acad. Sci USA. 78(7), 4204-4208.

Buchmeier, N.A. and Heffron, F. (1990). Induction of Salmonella stress proteins upon infection of macrophages. Science 248, 730-732.

Buchmeier, N.A., Libby, S.J., Xu, Y., Loewen, P.C., Switala, J. and Guiney, D.G. (1995). DNA repair is more important than catalase for Salmonella virulence in mice. Journal o f Clinical Investigations 95, 1047-1053.

Bult, C.J., White, O., Olsen, G.J., Zhou, L., Fleischmann, R.D., Sutton, G.G., Blake, J.A., Fitzgerald, L.M., Clayton, R.A., Gocayne, J.D., Kerlavage, A.R., Dougherty, B.A., Tomb, J.B., Adams, M.D., Reich, C.I., Overbeek, R., Kirkness, E.F., Weinstock, K.G., Merrick, J.M., Glodek, A., Scott, J.L., Geoghagen, S.M., Weidman, J.F., Fuhrmann, J.L., Nguyen, D., Utterback, T.R., Kelley, J.M., Peterson, J.D., Sadow, P.W., Hanna, M.C., Cotton, M.D., Roberts, K.M., Hurst, M.A., Kaine, B.P., Borodovsky, M., Klenk, H-P., Fraser, C M., Smith, H.O., Woese, C.R. and Venter, J.C. (1996). Complete genome sequence of the methanogenic archaeon, Methanococcus jannaschii. Science 273, 1058-1072.

Capecchi, M R. (1989). Altering the genome by homologous recombination. Science. 244, 1288-1292.

Carlini, L.E. and Porter, R.D. (1997). Analysis of ssb mutations in vivo imphcates SSB proteins in two distinct pathways of SOS induction and in recombinational DNA repair. Molec. Microbiol. 24(1), 129-139.

Carlsson, J. and Carpenter, V.S. (1980). The recA gene product is more important than catalase and superoxide dismutase in protecting Escherichia coli against hydrogen peroxide toxicity. Journal o f Bacteriol. 142(1), 319-321.

166 Casaregola, S., D’Ari, R. and Huisman, O. (1982a). Quantitative evaluation of recA gene expression in Escherichia coli. Molec. Gen. Genet. 185,430-439.

Casaregola, S., D’Ari, R. and Huisman, O. (1982b). Role of DNA replication in the induction and turn-off of the SOS response in Escherichia coli. Molec. Gen. Genet. 185,440-444.

Chamberlin, M.J. (1982). Bacterial DNA-dependent RNA polymerases. In: The Enzymes. Edited by Boyer, P.D. 15, 61-86. Academic Press, London.

Chase, J.W., Murphy, J.B., Whittier, R.F., Lorenson, E. and Sninksky, J.J. (1983). Amplification of ssb- 1 mutant single stranded DNA binding protein in Escherichia coli. Journal Molec. Biol. 164, 193-211.

Chaudhury, A.M. and Smith, G.R. (1985). Role of Escherichia coli RecBCD enzyme in SOS induction. Molec. Gen. Genet. 201, 525-528.

Cheo, D.L, Bayles, K.W. and Yashin, R.E. (1991). Cloning and characterisation of DNA damage-inducible promoter regions itom Bacillus subtilis. Journal o f Bacteriol. 173(5), 1696-1703.

Cheo, D.L., Bayles, K.W. and Yashin, R.E. (1993). Elucidation of the elements that control damage induction and competence induction of the Bacillus subtilis SOS system. Journal of Bacteriol. 175(18), 5907- 5915.

Chiodini, R.J. (1989). Crohn’s disease and the Mycobacterioses: a review and comparison of two disease entities. Clinical. Microbiol. Rev. 2(1), 90-117.

Chiodini, R.J., van Kruiningen, H.J., Thayer, W.R., Merkal, R.S. and Couteau, J.A. (1984a). Ruminant paratuberculosis (Johne’s disease): the current status and future prospects. Cornell Vet. 74, 218-262.

Chiodini, R.J., van Kruiningen, H.J., Thayer, W.R., Merkal, R.S. and Couteau, J.A. (1984b). Characteristics of an unclassified Mycobacterium species isolated from patients with Crohn’s disease. Journal Clin. Microbiol. 20, 966-971.

Christman, M.F., Morgan, R.W., Jacobson, F.S. and Ames B.N. (1985). Positive control of a regulon for defences against oxidative stress and some heat shock proteins in Salmonella typhimurium. Cell 41, 753-762.

Cirillo, J.D., Barletta, R.G., Bloom, B.R. and Jacobs Jr., W.R. (1991). A novel transposon trap for mycobacterial isolation and characterisation of IS 1096. Journal o f Bacteriol. 173, 7772-7780.

167 Clark-Curtiss, J E., Jacobs Jr., W.R., Docherty, M.A., Ritchia, L.R. and Curtiss, R. (1985). Molecular analysis of DNA and construction of genomic libraries of Mycobacterium leprae. Journal o f Bacteriol. 161(3), 1093-1102.

Clark, A.J. (1982). recA operator mutations and their usefulness. Biochimie 64, 669-675.

Clark, A.J. (1991). rec genes and homologous recombination proteins in Escherichia coli. Biochimie 73, 523-532.

Clark, A.J and Sandler, S.J. (1994). Homologous recombination: the pieces begin to fall in place. Crit. Rev. Microbiol. 20, 125-142.

Clark, A.J. and Volkert, M R. (1978). A new classification of pathways repairing pyrimidine dimer damage in DNA. In DNA repair mechanisms, pp576-72. Hanawalt, P C. and Fox, C.F. (Ed.)

Clarke, N.D. (1994). A proposed mechanism for the self-splicing of proteins. Proc. Natl. Acad. Sci. USA 91, 11084-11088.

Cohen, S.P., Resnick, J. and Sussman, R. (1983). Interaction of single-strand binding protein and RecA protein at the single strand DNA site. Journal Molec. Biol. 167, 901-909

Collado-Vibes, J., Magasanik, B. and Gralla, J.D. (1991). Control site location and transcription regulation in Escherichia coli. Microbiol Rev. 55, 371-394.

Collins, M E., Moss, M.T., Wall, S. and Dale, J.W. (1987). Expression of biotinylated proteins in mycobacterial gene libraries. FEMS Microbiol. Lett. 43, 53-56.

Colston, M.J. and Davies, E.O. (1994). Homologous recombination, DNA repair and mycobacterial recA genes. In Tuberculosis: Pathogenesis and control. Edited by Barry R Bloom. American Society for Microbiology, Washington, DC 20005.

Colston, M.J. and Davis, E.O. (1994). The ins and outs of protein splicing. Molecular Microbiology 12(3), 359- 363.

Courcelle, J., Carswell-Crumpton, C. and Hanawalt, P C. (1997). RecF and RecR are required for the resumption of replication at DNA replication forks in Escherichia coli. Proc. Natl. Acad. Sci. USA 94, 3714- 3719.

168 Cowing, D.W., Bardwell, J.C.A., Craig, E.A., Woolford, C., Hendrix, R.W. and Gross, C.A. (1985). Consensus sequence for Escherichia coli heat shock gene promoters. Proc. Natl. Acad. Sci. USA 82, 2679- 2683.

Cox, M.M. and Lehmann, I. (1987). Enzymes of general recombination. Ann Rev Biochem. 56, 229-262.

Cox, M.M. (1991). The recA protein as a recombinational repair system. Molecular Microbiology 5(6), 1295- 1299.

Cox, M.M. (1993). Relating biochemistry to biology: How the recombinational repair function of recA protein is manifested in its molecular properties. Bioessays 15(9), 617-623.

Crawford, J.T. and Bates, J.H. (1979). Isolation of plasmids from mycobacteria. Infect. Immun. 24,979- 981.

Crawford, J.T. and Bates, J.H. (1986). Analysis of plasmids in Mycobacterium avium-intracellulare isolates from persons with acquired immunodeficiency virus. Amer. Rev. Resp. Dis. 134, 659-661.

Dale, J.W and A Patki. (1990). Mycobacterial gene expression and regulation. In Molecular Biology of Mycobacteria. Edited by J.J McFadden, Academic Press Limited.

Daniel, T.M., Bates, J.H. and Downes, K.A. (1994). History of tuberculosis. In Tuberculosis. Edited by Bloom, B.R. American Society for Microbiology Press.

D’Ari, R. (1985). The SOS response. Biochimie 67, 343-347.

DasGupta, S.K., Bashyam, M.D. and Tyagi, A.K. (1993). Cloning and assessment of mycobacterial promoters by using a plasmid shuttle vector. Journal o f Bacteriol. 175(16), 5186-5192.

Davidson, L.A., Draper, P. and Minnikin, D.E. (1982). Studies on the mycolic acids from the walls of Mycobacterium microti. Journal Gen. Microbiol. 128, 823-828.

Davis, E.G., Jenner, P.J., Brooks, P.C., Colston, M.J. and Sedgwick, S.G. (1992). Protein splicing in the maturation of Mycobacterium tuberculosis RecA protein: A mechanism for tolerating a novel class of intervening sequence. Ce/Z 71,201-210.

Davis, E.G., Thangaraj, H.S., Brooks, P C. and Colston, M.J. (1994). Evidence of selection for protein introns in the recAs of pathogenic mycobacteria. The EMBO Journal 13(3), 699-703.

169 Davis, E.O., Colston, M.J. and Sedgwick, S.G. (1991). Novel structure of the recA locus of Mycobacterium tuberculosis implies processing of the gene product. Journal o f Bacteriol. 173(18), 5653-5662.

Dean, P.D.G., W.S Johnson and F.A Middle. (1985). Affinity Chromatography: A practical approach. Arlington, VA: IRL press.

De Crombrugghe, B. and Pastan, I. (1978). Cychc AMP, the cychc AMP receptor protein, and their dual control of the galactose operon. In The operon. Eds. Miller. J.H. and Reznikoff, W.S. Cold Spring Harbour, New York, pp303-324.

Dellagostin, O.A., Esposito, G., Eales, L.J., Dale, J.W. and McFadden, J.J. (1995). Activity of mycobacterial promoters during intracellular and extracellular growth. Microbiol. 141, 1785-1792.

Dellagostin, O.A., Wall, S., Norman, E., O'Shaughnessy, T., Dale, J.W. and McFadden J.J. (1993). Construction and use of integrative vectors to express foreign genes in mycobacteria. Molecular Microbiology 10(5), 983-993.

Demple, B. and Halbrook, J. (1983). Inducible repair of oxidative damage in Escherichia coli Nature 304, 466-468.

Deretic, V., Phillips, W., Dhandayuthapani, S., Mudd, M.H., Curcic, R., Garbe, T., Heym, B., Via, L.E. and Cole, S T. (1995). Mycobacterium tuberculosis is a natural mutant with an inactivated oxidative-stress regulatory gene: implications for sensitivity to isoniazid. Molec. Microbiol. 17(5), 889-900.

Dhandayuthapani, S., Via, L.E., Thomas, C.A., Horowitz, P.M., Deretic, D. and Deretic, V. (1995). Green fluorescent protein as a marker for gene expression and cell biology of mycobacterial interactions with macrophages. Molec. Microbiol. 17(5), 901-912.

Draper, P. and Rees, R.J.W. (1973). The nature of the electron-transparent zone that surrounds Mycobacterium lepraemurium inside host cells. Journal Gen. Microbiol. 77,79-87.

Dri, A M. and Moreau, P.L. (1993). Phosphate starvation and low temperature as well as ultraviolet irradiation transcriptionally induce the Escherichia coli Lex-controlled gene sflA. Molec. Microbiol. 8, 697-706.

Dri, A M. and Moreau, P.L. (1994). Control of the LexA regulon by pH: evidence for a reversible inactivation of the LexA repressor during the growth cycle of Escherichia coli. Molec. Microbiol. 12(4), 621-629.

170 Durbach, S I., Andersen, S.J. and Mizrahi, V. (1997). SOS induction in mycobacteria: analysis of the DNA-binding activity of a LexA-like repressor and its role in DNA damage induction of the recA gene from Mycobacterium smegmatis Molec. Microbiol. 26(4), 643-653.

Duwat, P., Ehrlich, S.D. and Gross, A. (1992). Use of degenerate primers for polymerase chain reaction cloning of the Lactococcus lactis subsp lactis recA gene. Appl. Environ. Microbiol. 58,2674-2678.

Duwat, P., Ehrlich, S.D. and Gross, A. (1995). The recA gene of Lactococcus lactis: characterisation and involvement in oxidative and thermal stress. Molec. Microbiol. 17(6), 1121-1131.

Eisenach, K.D., Cave, M.D., Bates, J.H. and Crawford, J.T. (1990). Polymerase chain reaction amplification of a repetitive DNA sequence specific for Mycobacterium tuberculosis. Journal Infect. Dis. 161,977-981.

Eisenach, K.D., Sifford, M.D., Cave, M.D., Bates, J.H. and Crawford, J.T (1991). Detection of Mycobacterium tuberculosis in sputum samples using a polymerase chain reaction. Am. Rev. Resp. Dis. 144, 1160-1163.

Ellman, G.L. (1959). Tissue Sulfhydryl groups. Arch. Biochem. Biophys. 82. 70-77.

Elsaghier, A., Prantera, C., Moreno, C. and Ivanyi, J. (1992). Antibodies to Mycobacterium paratuberculosis-sçtcr^ic protein antigens in Crohn’s disease. Clin. Exp. Immunol. 90, 503-508.

England, P.M., Wall, Q. and McFadden, J.J. (1991). IS900-promoted stable integration of a foreign gene into mycobacteria. Molec. Microbiol. 5,2047-2052.

Erickson, J.W. and Gross, C.A. (1989). Identification of the subunit of Escherichia coli RNA polymerase: a second alternative a factor involved in high-temperature gene expression. Genes Dev. 3,1462-1471.

Erickson, J.W., Vaughn, V., Walter, W.A., Neidhardt, F.C. and Gross, C.A. (1987). Regulation of the promoters and transcripts of rpoR, the Escherichia coli heat shock regulatory gene. Genes Dev. 1,419-432.

Farr, S B. and Kogoma, T. (1991). Oxidative stress responses in Escherichia coli and Salmonella typhimurium. Microbiol. Rev. 55(4), 561-585.

Fenlon, C.H. and Cynamon, M.H. (1986). Comparative in vitro activities of ciprofloxacin and other 4- quinolones against Mycobacterium tuberculosis and Mycobacterium intracellulare. Antimicrobial Agents 29(3), 386-388.

171 Fernandez de Henestrosa, A.R., Labazi, M., Mar Lopez, M. and Barbe, J. (1997). Functional analysis of the recA promoter of Rhodobacter capsulatas. Molec. Gen. Genet. 255,487-494.

Fernandez de Henestrosa, A.R., Calero, S. and Barbe, J. (1991). Expression of the recA gene of Escherichia coli in several species of Gram-Negative bacteria. Molec. Gen. Genet. 226, 503-596.

Fine, P.E.M. and Rodrigues, L.C. (1990). Mycobacterial diseases. The Lancet 335^ 1016-1020.

Fisher, M.W. and Kirchheimer, W.F. (1952). Studies on the growth of mycobacteria; The occurrence of arithmetic linear growth. Amer. Rev. Tuberc. 66, 758-761.

Flory, J. and Radding, C M (1982). Visualisation of RecA protein and its association with DNA: a priming effect of single strand binding proteins. Cell 28, 747-756

Fogh, R.H., Ottleben, G., Ruterjans, H., Schnarr, M., Boelens, R. and Kaptein, R. (1994). Solution structure of the LexA repressor DNA binding domain determined by spectroscopy. EMBO Journal. 13, 3936-3944.

Fogliano, M. and Schendel, P.F. (1981). Evidence for the inducibility of the uvrB operon. Nature 289, 196- 198.

Fomukong, N.G. and Dale, J.W. (1993). Transpositional activity of IS986 in Mycobacterium smegmatis. Gene. 130, 99-105.

Frank, E.G., Gonzalez, M., Ennis, D.G., Levine, A S. and Woodgate, R. (1996). In vivo stability of the Umu mutagenesis proteins: a major role for RecA. Journal o f Bacteriol. 178(12), 3550-3556.

Franzblau, G.S., Takeda, T. and Nakamura, M. (1986). Mycobacterial plasmids: screening and possible relationships to antibiotic resistance in Mycobacterium avium-intracellulare. Microbiol. Immunol. 30, 903- 907.

Friedberg, E C, Walker, G.C and W Siede. (1995) DNA repair and mutagenesis. American Society for Microbiology Press.

Frothingham, R., Hills, H.G. and Wilson, K.R. (1994). Extensive DNA sequence conservation throughout the Mycobacterium tuberculosis complex. Jour. Clin. Microbiol. 32(7), 1639-1643.

172 Fsihi, H., Vincent, V. and Cole, S T. (1996). Homing events in the gyrA gene of some mycobacteria. Proc. Natl Acad. Sci. USA 93, 3410-3415.

Garbe, T.R., Barathi, J., Barnini, S., Zhang, Y., Abou-Zeid, €., Tang, D., Mukherjee, R. and Young, D.B. (1994). Transformation of mycobacterial species using hygromycin resistance as selectable marker. Microbiol. 140, 131-138.

Garbe, T., Harris, D., Vordermeier, M., Lathigra, R., Ivanyi, J. and Young, D. (1993). Expression of the Mycobacterium tuberculosis 19-kilodalton antigen m Mycobacterium smegmatis: Immunological analysis and evidence of glycosylation. Infect, and Immun. 61(1), 260-267.

Garcia-Rodriguez, J.A. and Gomez-Garcia, A C. (1993). In vitro activities of quinolones against mycobacteria. Jour. Antimicrob. Agents and Chemother. 32, 797-808.

Garriga, X., Calero, S. and Barbe, J. (1992). Nucleotide sequence analysis and comparison of the lexA genes from Salmonella typhimurium, Erwinia carotovora, Psuedomonas aeruginosa and Pseudomonas putida. Molec Gen. Genet. 236, 125-134.

Gartenberg, M R. and Crothers, D M. (1988). DNA sequence determination of CAP-induced bending and protein binding affinity. Nature.333, 824-826.

Gassel, M. and Alonso, J.C. (1989). Expression of the recE gene during induction of the SOS response in Bacillus subtilis recombination-deficient strains. Molec. Microbiol. 3(9), 1269-1276.

Ghosh, R.K, Siddiqui, K.A.I., Mukhaopidhyay, G. and Ghosh, A. (1985). Evidence that a similar system to the recA system of Escherichia coli exists in Vibrio cholerae. Molec Gen. Genet. 200,439-441.

Gimble, F.S. and Thorner, J. (1992). Homing of a DNA endonuclease gene by meiotic gene conversion in Saccharomyces cerevisiae. Nature. 357, 301-306.

Ginsburg, H., Edmiston, S.H., Harper, J. and Mount, D.W. (1982). Isolation and characterisation of an operator constitutive mutant of Escherichia coli K-12. Molec. Gen. Genet. 187,4-11.

Goldman, D.S. (1970). Subcellular localisation of individual mannose-containing phospholipids in Mycobacterium tuberculosis. Amer. Rev. Respir. Dis. 102, 543-555

Golemis, E.A. and Brent, R. (1992). Fused protein domains inhibit DNA binding by LexA. Molec. Cell. Biol 12(7), 3006-3014.

173 Goodfellow, M. and Wayne, L.G. (1982). Taxonomy and nomenclature. In The Biology in Mycobacteria. Vol 1 Eds. Ratledge, C. and Stanford, J. Academic press, London. Chapter 10, pp471-521.

Goren, M B and P.J Brennan (1979). Mycobacterial lipids. In Tuberculosis pp69-193. Edited by Youmans, O.P. Philadelphia.

Gordhan, B.G., Andersen, S.J., De Meyer, A.R. and Mizrahi, V. (1996). Construction by homologous recombination and phenotypic characterisation of a DNA polymerase domain polA mutant of Mycobacterium smegmatis. Gene 178, 125-130.

Goto, ¥., Tanuguchi, H., Udou, T., Mizuguchi, Y. and Tokunaga, T. (1991). Development of a new host vector system in mycobacteria. FEMS Microbiol Lett. 83, 277-282.

Gralla, J.D. (1990). Promoter recognition and mRNA initiation by Escherichia coli E-a’°. Methods in Enzymology. 185, 37-54.

Graves, M.C. and Rabinowitz J.C. (1986). In vivo and in vitro transcription of the Clostridium pasteurianum ferredoxin gene: evidence for “extended” promoter elements in Gram-positive organisms. Journal Biol Chem. 261,11409-11415.

Green, E.P., Tizard, M.L.V., Thompson, J., Winterbourne, D.J., McFadden, J.J. and Hermon-Taylor, J. (1989). Sequence and characteristics of 1S900, an insertion element identified in a human Crohn’s disease isolate of Mycobacterium paratuberculosis. Nucleic Acids Res. 17(22), 9063-9073.

Greensburg, J.T. and Demple, B. (1989). A global response induced in Escherichia coli by redox-cycling agents overlaps with that induced by peroxide stress. Journal o f Bacteriol. 171, 3933-3939.

Griffin, T.J. and Kolodner, R.D. (1990). Purification and preliminary characterisation of the Escherichia coli RecF protein. Journal Bacteriol. 172, 6291-6299.

Groenen, P.M.A., Bunschoten, A.E., Soolingen, D.V. and Van Embden, J.D.A. (1993). Nature of DNA polymorphism in the direct repeat cluster of Mycobacterium tuberculosis-, application for strain differentiation by a novel typing method Molecular Microbiology 10(5), 1057-1065.

Grossman, A.D., Erickson, J.W. and Gross, C.A. (1984). The htpR gene product of Escherichia coli is a sigma factor for heat shock promoters. Cell 38, 383-390.

174 Gu, H.H., Xu, J., Gallagher, M. and Dean, G.E. (1993). Peptide splicing in the vacuolar ATPase subunit A from Candida tropicalis. Journal Biol. Chem. 268, 7372-7381.

Gudas, L.J. and Pardee, A.B. (1975). Model for regulation of Escherichia coli DNA repair functions. Proc. Natl. Acad. Sci. USA 72(6), 2330-2334.

Gudas, L.J. and Pardee, A.B. (1976). DNA synthesis and inhibition of protein X in Escherichia coli. Journal Molec. Biol. 101,475-477.

Guerrero, R., Llagostera, M., Villaverde, A. and Barbe, J. (1984). Changes in ATP concentration in Escherichia coli during induction of the SOS system by mitomycin C and bleomycin. Journal Gen. Microbiol. 130, 2247-2251.

Guilhot, C., Otal, I., Van Rompaey, L, Martin, C. and Gicquel, B. (1994). Efficient transposition in mycobacteria: Construction of Mycobacterium smegmatis insertional mutagenesis libraries. Journal ofBacteriol. 176(2), 535-539.

Haijema, B.J, Sinderen, D.V., Winterllng, K., Koolstra, J., Venema, G. and Hamoen, L.W. (1996). Regulated expression of the dinR and recA genes during competence development and SOS induction in Bacillus subtilis. Molec. Microbiol. 22(1), 75-85.

Haldenwang, W.G. (1995). The sigma factors of Bacillus subtilis. Microbiol. Rev. 59, 1-30.

Haldenwang, W.G., Lang, N. and Losick, R. (1981). A sporulation-induced sigma-like regulatory protein from Bacillus subtilis. Cell 23, 615-624.

Haldenwang, W.G. and Losick, R. (1980). Novel RNA polymerase a factor from Bacillus subtilis. Proc. Natl. Acad. Sci. USA 77(12), 7000-7004.

Hanks, J.H. (1961). Significance of capsular components of Mycobacterium leprae and other mycobacteria. Int. Journal Leprosy 29(1), 74-83.

Hansen, G.A. (1874). Unders pgelser ang ende spedalskhedens rsager. Norsk magasin for laegevitenskapen. 4, 1-88.

Harley, C.B. and Reynolds, R.P. (1987). Analysis of Escherichia coli promoter sequences. Nucleic Acid Res. 15, 2343-2361.

175 Harmon, F.G., Rehrauer, W.M. and Kowalczykowski, S.C. (1996). Interaction of Escherichia coli RecA protein with LexA repressor. Journal Biol. Chem. 271(39), 23874-23883.

Harrison, S.C. and Aggerwal, A.K. (1990). DNA recognition by proteins with the helix-tum-helix motif. Annu. Rev. Biochem. 59, 933-969.

Harshey, R.M. and Ramakrishnan, T. (1977). Rate of ribonucleic acid chain growth in Mycobacterium tuberculosis H37Rv. Journal ofBacteriol. 129(2), 616-622.

Hatful!, G.F. and Jacobs, W.R. (1994). Cornerstones of mycobacterial research.. In Tuberculosis. Edited by Bloom, B.R. American Society for Microbiology Press.

Hedge, S.P., Rajagopalan, M. and Madiuaju, M.V.V.S. (1996). Preferential binding of Escherichia coli RecF protein to gapped DNA m the presence of adenosine (y-thio) triphosphate. Journal o f Bacterial. 178(1), 184-190.

Hedge, S., Sandler, S.J., Clark, A.J. and Madiraju, M.V.V.S. (1995). recO and recR mutations delay induction of the SOS response. Molec. Gen. Genet. 246, 254-25.

Helmann, J.D. (1995). Compilation and analysis of Bacillus subtilis o*-dependent promoter sequences: evidence for extended contact between RNA polymerase and upstream promoter DNA. Nucleic. Acid. Res. 23(13), 2351-2360.

Helmann, J.D. and Chamberlin, M.J. (1988). Structure and function of bacterial sigma factors. Ann Rev Biochem 57, 839-872.

Hemavathy, K.C. and V. Nagaraja (1995). DNA méthylation in mycobacteria: Absence of méthylation at GATC (Dam) and CCA/TGG (Dcm) sequences. FEMS Immunol, and Medical Microbiol. 11, 291-296.

Hermans, J., Martin, C., Huyberts, G.N.M., Goosen, T. and de Bont, J.A.M. (1991). Transformation of Mycobacterium aurum and Mycobacterium smegmatis with the broad host range Gram-negative cosmid vector pJRD215. Mo/ec. Microbiol. 5, 1561-1566.

Higashitani, N, Higashitani, A. and Horiuchi, K. (1995). SOS induction in Escherichia coli by single­ stranded DNA of mutant filamentous phage: Monitoring by cleavage of LexA repressor. Journal ofBacteriol. 177(12), 3610-3612.

176 Hinrikson, H P. and Pfyffer, G.E. (1994). Mycobacterial mycolic acids. Med. Microbiol. Lett. 3,49-57.

Hinshelwood, S. and Stoker, N.G. (1992). An Escherichia co/i-mycobacteria shuttle cosmid vector, pMSCl. Gene 110, 115-118.

Hiom, K. and West, S.C. (1995). The mechanism of branch migration in homologous recombination: Assembly of RuvAB-Holliday junction complex in vitro. Cell 80, 787-793.

Hirata, R., Ohsumi, Y., Nakano, A., Kawasaki, K. and Anraku, Y. (1990). Molecular structure of a gene, VMAl, encoding the catalytic subunit of H+-translocating adenosine triphosphate from vacuolar membranes of Saccharomyces cerevisiae. Journal Biol. Chem. 265, 6726-6733.

Hiriyanna, K.T. and Ramakrishnan, T. (1986). Deoxyribonucleic acid replication in Mycobacterium tuberculosis H37Rv. Arch. Microbiol. 144, 105-109.

Hodges, R.A., Perler, F.B., Noren, C.J. and Jack, W. (1992). Protein splicing removes intervening sequences in an archaea DNA polymerase. Nucl. Acids Res. 20, 6153-6157.

Hofemeister, J. and Bohme, H. (1975). DNA repair in Proteus mirabilis. Ill Survival, dimer excision and UV reactivation in comparison to Escherichia coli K-12. Molec Gen. Genet. 141, 147-161.

Holmes, D.S. and Quigley, M. (1981). A rapid boiling method for the preparation of bacterial plasmids. Anal. Biochem. 114, 193-197.

Horii, T., Ogawa, T., Nakatani, T., Hase, T., Matsubara, H. and Ogawa, H. (1981c). Regulation of SOS functions: purification of Escherichia coli LexA protein and determination of its specific site cleaved by the RecA protein. Cell 27, 515-522.

Horii, T., Ogawa, T. and Ogawa, H. (1981a). Regulation of SOS functions: purification of Escherichia coli LexA protein and determination of its specific site cleaved by RecA protein. Cell 27, 515-522.

Horii, T., Ogawa, T. and Ogawa, H. (1981b). Nucleotide sequence of the lexA gene of Escherichia coli. Cell 23, 689-697.

Horsburgh Jr., C.R.. (1991). Mycobacterium avium complex infection in the acquired immunodeficiency syndrome. New Engl. Journal Med. 324, 1332-1337.

177 Honvitz, M.A., Lee, B.W.E., Dillon, B.J. and Harth, G. (1995). Protective immunity against tuberculosis induced by vaccination with major extracellular proteins of Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 92, 1530-1534.

Howard-Flanders, P., West, S.C. and Stasiak, A. (1984). Role of recA protein spiral filaments in genetic recombination. Nature 309,215-219.

Hudson, L. and Hay, F.C. (1976). Practical Immunology, pp 193-197. Blackwell Scientific Publications. Oxford.

Hull. S.I., Wallace, R.J., Bubgy, D.G., Price, K.E., Goodlines, R.A., Siverson, J.M. and Silcox, V.A. (1984). Presence of aminoglycoside acetyl-transferase and plasmids in Mycobacterium fortuitum. Annu. Rev. Resp. Dis. 129, 614-618.

Hurstel, S., Granger-Schnarr, M., Duane, M. and Schnarr, M. (1986). In vitro binding of LexA repressor to DNA: evidence for the involvement of the amino terminal domain. EMBO Journal 5, 793-798.

Husson, R.N., James, B E. and Young, R.A. (1990). Gene replacement and expression of foreign genes in mycobacteria. Journal ofBacteriol. 172(2), 519-523.

Ikeda, M., Makino, O. and Shibata, T. (1990). Epitopes and active sites of the RecA protein. Journal Biol. Chem. 265, 8948-8956.

Imaeda, T., Kirchheimer, W.F. and Barksdale, L. (1982). DNA isolated from Mycobacterium leprae-. Genomic size, base ratio and homology with other related bacteria as determined by optical DNA-DNA reassociation. Journal o f Bacteriol. 150(1), 414-417.

Imlay, J. A. and Linn, S. (1986). Bimodal patterns of killing of DNA repair defective or anoxically grown Escherichia coli by hydrogen peroxide. Journal o f Bacteriol. 166, 519-527.

Iwasaki, H., Takahagi, M., Nakata, A. and Shinagawa, H. (1992). Escherichia coli RuvA and RuvB proteins interact with Holliday junctions and promote branch migration. Genes Dev. 6, 2214-2220.

Iyer, V.N. and Szybalski, W. (1963). A molecular mechanism of mitomycin action: Linking of complementary DNA strands. Microbiology 50, 355-362.

Iyer, V.N. and W. Szybalski (1964). Mitomycin and porfiromycm: Chemical mechanism of activation and cross-linking of DNA. Science 145, 55-68.

178 Jacobs Jr., W.R., Barret, J.F., Clark-Curtiss, J.E. and Curtiss, R. (1986). In vivo repackaging of recombinant cosmid molecules for analysis of Salmonella typhimurium. Streptococcus mutans and mycobacterial genomic libraries. Infect, and Immun. 52, 101-109.

Jacobs Jr., W.R., Docherty, M.A., Curtiss in, R. and Clark-Curtiss, J.E. (1986b). Expression of Mycobacterium leprae genes from a Streptococcus mutans promoter in Escherichia coli K-12. Proc. Natl. Acad. Sci. USA 83, 1926-1930.

Jacobs Jr., W.R, Tuckman, M. and Bloom, B.R. (1987). Introduction of foreign DNA into mycobacteria using a shuttle phasmid. Nature 327, 532-535.

Jain, S.K., Cox, M.M. and Inman, R.B. (1994). On the role of ATP hydrolysis in RecA protein-mediated DNA strand exchange. Journal o f Biol. Chem. 269(32), 20653-20661.

Jain, S.K., Kaushal, D., DasGupta, S.K. and Tyagi, A.K. (1997). Construction of shuttle vectors for genetic manipulation and molecular analysis of mycobacteria. Gene 190, 37-44.

Ji, Y.E., Colston, M.J. and Cox, R.A. (1994a). The ribosomal RNA {rrn) opérons of fast growing mycobacteria: primary and secondary structures and their relation to rm opérons of pathogenic slow growers. Microbiol. 140, 2829-2840.

Ji, Y.E., Kempsell, K.E., Colston, M.J. and Cox, R.A. (1994). Nucleotide sequences of the spacer-1, spacer-2 and trailer regions of the rm opérons and secondary stmctures of precursor 23 S rRNA and precursor 5S rRNA of slow growing mycobacteria. Microbiol. 140, 1763-1773.

Ji, Y.E., Colston, M.J. and Cox, R.A. (1994b). Nucleotide sequence and secondary stmcture of precursor 16S rRNA of slow-growing mycobacteria. Microbiol. 140, 128-132.

Jishage, M., Iwata, A., Ueda, S. and Ishihama, A. (1996). Regulation of RNA polymerase sigma subunit synthesis in Escherichia coli: Intracellular levels of four species of sigma subunit under various growth conditions. Journal o f Bacteriol. 178(18), 5447-5451.

Jiva, T.M., Jacoby, H.M., Weymouth, L.A., Kaminski, D.A. and Portmore, A C. (1997). Mycobacterium xenopi; Innocent bystander or emerging pathogen? Clin. Infect. Dis. 24, 226-232.

Johnson, B.F. (1977). Genetic mapping of the lexC-113 mutation. Molec. Gen. Genet. 157, 91-97.

179 Johnston, J.L., Sloan, J., Fyfe, J.A.M., Davies, J.K. and Rood, J.I. (1997). The recA gene from Clostridium petfringens is induced by methyl methane sulphonate and contains an upstream Cheo box. Microbiology 143, 885-890.

Jonczyk, P. and Nowicka, A. (1996). Specific in vivo protein-protein interactions between Escherichia coli SOS mutagenesis proteins. Journal ofBacteriol. 178(9), 2580-2585.

Jordan, S.R. and Pabo, C O. (1988). Structure of the X complex at 2.5 resolution: details of the repressor- operator interactions. Science 242, 893-899.

Kalpana, G.V., Bloom, B.R. and Jacobs Jr., W.R. (1991). Insertional mutagenesis and illegitimate recombination in mycobacteria. Proc. Natl. Acad. Sci. USA 88,5433-5437.

Kane, P.M., Yamashiro, C.T., Wolczyk, D.F., Neff, N., Goebl, M. and Stevens, T.H. (1990). Protein splicing converts the yeast TFPl gene products to the 69kDa subunit of the vacuolar H+ adenosine triphosphatase. Science. 250, 651-657.

Kanetsuna, F. (1968). Chemical analysis of mycobacterial cell walls. Biochem. Biophys. Acta. 158, 130-143.

Karlin, S., Weinstock, G.M. and Brendel, V. (1995). Bacterial classifications derived from RecA protein sequence comparisons. Journal o f Bacteriol. 177(23), 6881-6893.

Karlin, S. and Brocchieri, L. (1996). Evolutionary conservation of recA genes in relation to protein structure and fimction. Journal ofBacteriol. 178(7), 1881-1894.

Karu, A.E. and Allen, L. (1982). Interaction of hybridoma antibodies with normal and mutant forms of Escherichia coli RecA protein. Biochimie 64, 723-729.

Kenney, T.J. and Churchward, G. (1996). Genetic analysis of the Mycobacterium smegmatis rpsL promoter. Journal ofBacteriol. 178(12), 3564-3571.

Kenyon, C.J. and Walker, G.C. (1980). DNA-damaging agents stimulate gene expression at specific loci in Escherichia coli. Proc. Natl. Acad. Sci. USA 77(5), 2819-2823.

Kempsell, K.E., Ji, Y.E., Estrada-Garcia, I.C.E., Colston, M.J. and Cox, R.A. (1992). The nucleotide sequence of the promoter, 16S rRNA and spacer region of the ribosomal RNA operon of Mycobacterium tuberculosis and comparison with Mycobacterium leprae precursor rRNA. Journal Gen. Microbiol. 138,

180 1717-1727.

Kieser, T., Moss, M.T., Dale, J.W. and Hopwood, D.A. (1986). Cloning and expression of Mycobacterium bovis BCG DNA in Streptomyces lividans. Journal o f Bacteriol. 168(1), 72-80.

Kim, B. and Little, J.W. (1992). Dimerization of a specific DNA-binding protein on the DNA. Science 252, 203-205.

Kirschner, P., Springer, B., Vogel, U., Meier, A., Wrede, A., Kiekenbeck, M. and Bottger, E.G. (1993). Genotypic identification of mycobacteria by nucleic acid sequence determination; report of a 2-year experience in a clinical laboratory. Journal o f Clin. Microbiol. 31, 2882-2889.

Kleina, L.G. and Miller, J.H. (1990). Genetic studies of the lac repressor: XIII Extensive amino acid replacements generated by the use of natural and synthetic nonsense suppressors. Journal Molec. Biol. 212,295- 318.

Knegtel, R.M.A., Fogh, R.H., Ottlegen, G., Rutterjans, H., Dumoulin, P., Schnarr, M., Boelens, R. and Kaptein, R. (1995). A model for the LexA repressor DNA complex. Proteins: Structure, Function and Genetics 21, 226-236.

Koch, R. (1882). Die tiologie der Tuberkulose. Berliner Klinischen wochenschrifi.15, 221-230.

Kochi, A. (1991). The global tuberculosis situation and the new control strategy of the World Health Organisation. Tubercle. 72, 1-6.

Kohn, H., Ven-Shun, L. and Moon-Shong, T. (1992). Recognition of mitomycin C-DNA monoadducts by UvrABC nuclease, y o w m a / Chem. Soc. 114, 5501-5509.

Kolodner, R., Hall, S.D. and Luisi-Deluca, C. (1994). Homologous pairing proteins encoded by the Escherichia coli recE and recT genes. Molecular Microbiology 11(1), 23-30.

Kong, T.H., Coates, A.R.M., Butcher, P.D., Hickman, C.J. and Shinnick, T.M. (1993). Mycobacterium tuberculosis expresses two chaperonin-60 homologues. Proc. Natl. Acad. Sci. USA. 90,2608-2612.

Kowalczykowski, S.C. (1987). Mechanistic aspects of the DNA strand exchange activity of Escherichia coli RecA protein. TIBS 12,141-145.

181 Kowalczykowski, S.C., Dixon, D.A., Eggleston, A.K., Lauder, S.D. and Rehrauer, W.M. (1994). Biochemistry of homologous recombination in Escherichia coli. Microbiol. Rev. 58(3), 401-465.

Kowalczykowski, S.C. and Krupp, R.A. (1987). Effects of Escherichia coli SSB proteins on the single­ stranded DNA dependent ATPase activity of Escherichia coli RecA protein. Evidence that SSB protein facihtates the binding of RecA protein to regions of secondary structure within single-stranded DNA. Journal Molec. Biol. 193,97-113.

Kremer, L., Baulard, A., Estaquier, J., Content, J., Capron, A. and Locht, C. (1995). Analysis of Mycobacterium tuberculosis 85kDa antigen promoter region. Journal ofBacteriol. 177, 642-653.

Krivi, G.G., Bittner, M L., Rowold, E., Wong, E.Y., Glenn, K.C., Rose, K.S. and Tiemeier, D C. (1985). Purification of RecA based fusion proteins by immunoadsorbent chromatography. Characterisation of a major antigenic determinant of Escherichia coli RecA protein. Journal Biol. Chem. 260, 10263-10267.

Krueger, J.H. and Walker, G.C. (1984). The groEL and dnaK genes of Escherichia coli are induced by ultraviolet irradiation and nalidixic acid in an A/p/?^-dependent fashion. Proc. Natl. Acad. Sci. USA 81, 1499- 1503.

Krueger, J.H., Elledge, S.J. and Walker, G.C. (1983). Isolation and characterisation of Tn5 insertion mutations in the lexA gene of Escherichia coli. Journal ofBacteriol. 153, 1368-1378.

Krummel, B. and Chamberlin, M .J. (1989). RNA chain initiation by Escherichia coli RNA polymerase. Structural transitions of the enzyme in early ternary complexes. Biochemistry 28, 7829-7842.

Kumar, A., Malloch, R.A., Fujita, N., Smillie, D.A., Ishihama, A. and Hayward, R.S. (1993). The minus 3 5-recognition sequence of Escherichia coli sigma 70 is inessential for initiation of transcription at an “extended minus 10” promoter. Journal Molec. Biol. 232,406-418.

Kumar, G., Kalra, V.K. and Brodie, A.F. (1979). Asymmetric distribution of phospholipids in membranes from Mycobacterium phlei. Arch. Biochem. Biophys. 198, 22-30.

Kumar, R.A., Vaze, M B., Chandra, N.R., Vijayan, M. and Muniyappa, K. (1996). Functional characterisation of the precursor and spliced forms of RecA protein of Mycobacterium tuberculosis. Biochemistry 25^ 1793-1802.

Kumar, S., Lipman, R. and Tomasz, M. (1992). Recognition of specific sequences by mitomycin C for alkylation. Biochemistry 1399-1407.

182 Kundu, T.K., Kusano, S. and Ishihama, A. (1997). Promoter selectivity of Escherichia coli RNA polymerase o’^holoenzyme involved in transcription of flagellar and chemotaxis genes. Journal of Bacteriol. 179(13), 4264-4269.

Kurumizaaka, H., Aihara, H., Ikawa, S., Kashima, T., Bazemore, L.R., Kawasaki, K., Sarai, A., Radding, C M. and Shibata, T. (1996). A possible role of the C-terminal domain of the RecA protein. A gateway model for double-stranded DNA binding. Journal o f Biol. Chem. 271(52), 33515-33524.

Kuzminov, A. (1995). A mechanism for induction of the SOS response in Escherichia coli: Insights into the regulation of reversible protein polymerisation in vivo. Journal Theor. Biol. 177, 29-43.

Kwaik, Y.A., Einstein, B.I. and Engleberg, N.C. (1993). Phenotype modulation by Legionella pneumophila upon infection of macrophages. Infect. Immun. 61, 1320-1329.

Labidi, A., Dauget, C., Gob, K.S. and David, H.L. (1984). Plasmid profiles of Mycobacterium fortuitum complex isolates. Curr. Microbiol. 11, 235-240.

Labidi, A C., David, H.L. and Roulland-Dussoix, D. (1985). Restriction endonuclease mapping and cloning of Mycobacterium fortuitum var. fortuitum plasmid pAL5000. Ann. Inst. Pasteur Microbiol. 136, 209-215.

Labidi, A., Mardis, E., Roes, B.A.and Wallace, R.J. (1992). Cloning and DNA sequence of the Mycobacterium fortuitum plasmid pAL5000 Plasmid 27, 130-140.

Lamb, F.I. and Colston, M.J. (1986). Cloning of Mycobacterium leprae gene in Streptomyces. Lepr. Rev. 57, 225-229.

Lavery, P.E. and Kowalczykowski, S.C. (1990). Properties of recA441 protein-catalysed DNA strand exchange can be attributed to an enhanced ability to compete with SSB proteins. Journal Bol. Chem. 265,4004-4010.

Lee, B.Y. and Horwitz, M.A. (1995). Identification of macrophage and stress-induced proteins of Mycobacterium tuberculosis. Journal o f Clin. Invest. 96, 245-249.

Lee, C., Levin, A. and Branton, D. (1987). Copper staining; a five-minute protein stain for sodium-dodecyl- sulphate-polyacrylamide gels. Analytical Biochemistry 166,308-312.

Lee, M.H. and Hatfull, G.F. (1993). Mycobacteriophage L5 integrase-mediated site-specific integration in vitro. Journal ofBacteriol. 175(21), 6836-6841

183 Lee, M.H., Pascopella, L., Jacobs, W.R. and Hatfull, G.F. (1991). Site specific integration of mycobacteriophage L5; Integration proficient vectors Mycobacterium smegmatis, Mycobacterium bovis BCG and Mycobacterium tuberculosis. Proc Natl Acad. Sci USA 88, 3111-3115

Lehman, K.B. and Neumann, R. (1896). Atlas und Grundis der Backterologie und Lehrbuch der speciellen bakteriologischen Diagnostik. edition. Lehmann, J.F., M»nchen, Germany.

Lewin, B. (1990a). A panapoly of opérons: The lactose paradigm and others. In Genes IV, chapter 13. Oxford University Press, New York and Cell Press, Massachusetts.

Lewin, B. (1990b). Systems that safeguard DNA. In Genes IV, chapter 19. Oxford University Press, New York and Cell Press, Massachusetts.

Lewis, L.K., Jenkins, M.E. and Mount, D.W. (1992). Isolation of DNA damage-inducible promoters in Escherichia coli: Regulation oipolB (dinA), dinO and dinHhy LexA repressor. Journal o f Bacteriol. 174(10), 3377-3385.

Leysen, D C., Haemers, A. and Pattyn, S.R. (1989). Mycobacteria and the new quinolones. Antimicrobial Agents and Chemother. 33, 1-5.

Liebana, E., Aranaz, A., Francis, B. and Cousins, D. (1996). Assessment of genetic markers for species differentiation within the Mycobacterium tuberculosis complex. Journal Clin. Microbiol. 33(4), 933-938.

Liebermann, H. B. and Witkin, E M. (1983). DNA degradation, ultraviolet sensitivity and SOS mediated mutagenesis in strains of Escherichia coli deficient in single strand binding proteins: Effects of mutations and treatments that alter levels of exonuclease V and RecA. Molec. Gen. Genet. 190, 92-100.

Lin, K.C. and Shiuan, D. (1995). A simple method for DNase I footprinting analysis. Journal Biochem Biophys Methods 30, 85-89.

Lindquist, S. (1986). The heat shock response. Ann. Rev. Biochem. 55,1151-1191.

Lindsley, J.E. and Cox, M.M. (1990). Assembly and disassembly of RecA protein filaments occur at opposite filament ends. Journal o f Biol. Chem. 265(16), 9043-9054.

Little, J.W . (1983). The SOS regulatory system: control of its state by the level of RecA protease. Journal o f Molec. Biol. 167, 791-808.

184 Little, J.W. (1984). Autodigestion of LexA and phage X repressors. Proc. Natl Acad. Sci. USA 81, 1375- 1379.

Little, J.W. and Hill, S.A. (1985). Deletions within a hinge region of a specific DNA binding protein. Proc. Natl Acad. Sci USA 82, 2301-2305.

Little, J.W. and Mount, D.W. (1982). The SOS regulatory system of Escherichia coli. Cell 29,11-22.

Little, J.W., Kim, B., Roland, K.L., Smith, M.H., Lin, L.L. and Slilaty, S.N. (1994). Cleavage of LexA repressor. Methods in Enzymology 244, 266-284.

Little, J.W., Mount, D.W. and Yanisch-Perron, C.R. (1981). Purified LexA protein is a repressor of the lexA and recA genes. Proc. Natl Acad. Sci USA. 78,4199-4203.

Liu-Johnson, H.N., Gartenberg, M R. and Crothers, D M. (1986). The DNA binding domain and bending angle of Escherichia coli CAP protein. Cell. 47, 995-1005.

Lloyd, R.G., Porton, M.C. and Beckman, C. (1988). Effect of recF, recJ, recN, recO and ruv mutations on ultraviolet survival and genetic recombination in a recD strain of Escherichia coli K12. Molec. Gen. Ge«er.212, 317-324.

Losick, R., Youngman, P. and Piggot, P.J. (1989). Genetics of endospore formation in Bacillus subtilis. Annual Rev. Genet. 20, 625-669.

Love, P.E. and Yasbin, R.E. (1986). Induction of the Bacillus subtilis SOS-like response by Escherichia coli RecA protein. Proc. Natl Acad. Sci USA. 83, 5204-5208.

Lovett Jr., C M. and Roberts, J.W. (1985). Purification of a RecA protein analogue from Bacillus subtilis. Journal Biol. Chem. 260, 3305-3313.

Lovett Jr., C M., Cho, K.C. and O’Gara, T.M. (1993). Purification of a SOS repressor from Bacillus subtilis. Journal ofBacteriol. 175(21), 6842-6849.

Lovett Jr., C M., Love, P.E., Yasbin, R E. and Roberts, J.W. (1988). SOS-like induction in Bacillus subtilis: Induction of the RecA protein analogue and a damage-inducible operon by DNA damage in recA^ and DNA repair-deficient strains. Journal o f Bacteriol. 170(4), 1467-1474.

185 Lovett Jr., C.M., O’Gara, T.M. and Woodruff, J.N. (1994). Analysis of the SOS inducing signal in Bacillus subtilis using Escherichia coli LexA as a probe. Journal of Bacteriol. 176(16), 4914-4923.

M adiraju, M.V.V.S. and Clark, A.J. (1991). Effect of RecF protein on reactions catalysed by RecA protein. Nucleic Acids Res. 19, 6925-6300

Makino, O., Shibata, Y., Maeda, H., Shibata, T. and Ando, T. (1985). Monoclonal antibodies with specific effects on the partial activities of RecA protein of Escherichia coli. Journal o f Biol. Chem. 260, 15402-15404.

Mandai, T.N., Mahdi, A.A., Sharpies, G.J. and Lloyd, R.G. (1993). Resolution of Holliday intermediates in recombination and DNA repair: indirect suppression of ruvA, ruvB and ruvC mutations. Journal o f Bacteriol. 175, 4325-4334.

Marais, A., Bove, J.M. and Renaudin, J. (1996). Characterisation of the recA gene regions of Spiroplasma citri and Spiroplasma melliferum. Journal o f Bacteriol. 178(23), 7003-7009.

Markham, B.E., Little, J.W. and Mount, D. (1981). Nucleotide sequence of the lexA gene of Escherichia coli K12. Nucleic Acid Res. 9, 4149-4161.

Marklund, B.I., Speert, D.P. and Stokes, R.W. (1995). Gene replacement through homologous rtcorNoinadon in Mycobacterium intracellulare. Journal o f Bacteriol. 177(21), 6100-6105.

Martin, C., Timm, J., Rauzier, J., Gomez-Lus, R., Davies, J. and Gicquel, B. (1990). Transposition of an antibiotic resistance element in mycobacteria. Nature 21,739-743.

Matsui, H., Sano, Y., Ishihara, H. and Sinomiya, T. (1993). Regulation of pyocin genes in Psuedomonas aeruginosa by positive {prtN) and negative {prtR) regulatory genes. Journal of Bacteriol. 175(5), 1257-1263.

Matsuo, K., Yamaguchi, R., Yamazaki, A., Tasaka, H., Terasaka, K., Totsuka, M., Kobayashi, K., Yukitake, H. and Yamada, T. (1990). Establishment of a foreign antigen secretion system in mycobacteria. Infect, and Immun. 58,4049-4054.

McAdam, R.A., Guilhot, C and Gicquel, B. (1994). Transposition in mycobacteria. . In Tuberculosis. Edited by Bloom, B.R. American Society for Microbiology Press.

McAdam, R.A., Weisbrod, T.R., Martin, J., Scuderi, J.D., Brown, A.M., Cirillo, J.D., Bloom, B.R. and Jacobs Jr., W.R. (1995). In vivo growth characteristics of leucine and methionine auxotrophic mutants of

186 Mycobacterium bovis BCG generated by transposition mutagenesis. Infect. Immunol. 63, 1004-1012.

McClure, W.R. (1985) Mechanism and control of transcription in prokaryotes. Ann Rev Biochem. 54, 171- 204.

McEntee, K., Weinstock, G.M. and Lehman, I.R. (1980). RecA protein catalysed strand assimilation; stimulation by Escherichia coli single-strand DNA-binding protein. Proc. Natl. Acad. Sci. USA 77, 857-861.

McEvoy, J.L., Murata, H. and Chatterjee, A.K. (1990). Molecular cloning and characterisation of an Erwinia carotovora subsp. carotovora pectin lyase gene that responds to DNA damaging agents. Journal o f Bacteriol 172, 3284-3289.

McEvoy, J.L., Murata, H. and Chatter jee, A.K. (1992). Genetic evidence for an activator required for induction of pectin lyase in Erwinia carotovora subsp. carotovora by DNA-damaging agents. Journal o f Bacteriol. 174(16), 5471-5474.

McFadden, J.J., Butcher, P.D., Chiodini, R.J. and Hermon-Taylor, J. (1987a). Determination of genomic size and DNA homology between an unclassified mycobacterium species isolated from patients with Crohn’s disease and other mycobacteria. Journal Gen. Microbiol. 133, 211-214.

McFadden, J.J., Butcher, P.D., Chiodini, R.J. and Hermon-Taylor, J. (1987b). Crohn’s disease isolated mycobacteria are identical to Mycobacterium paratuberculosis, as determined by DNA probes that distinguish between mycobacterial species. Journal Clin Microbiol. 25, 796-801.

McNeil, M R. and Brennan, P.J. (1991). Structure, function and biogenesis of the cell envelope of mycobacteria in relation to bacterial physiology, pathogenesis and drug resistance; some thoughts and possibilities arising from recent structural information. Res. Microbiol. 142,451-463.

Mehra, V., Sweetser, D. and Young, R.A. (1986). Efficient mapping of protein antigen determinants. Proc Natl Acad. Sci USA 83, 7013-7017.

Meisner, P.S. and Falkinham, J.O. (1986). Plasmid DNA profiles and epidemiological markers for clinical and environmental isolation of Mycobacterium avium, Mycobacterium intracellulare and Mycobacterium scrofulaceum. Journal o f Infect. Dis. 153, 526-531.

Menetskl, J.P. and Kowalczykowski, S.C. (1989). Enhancement of Escherichia coli RecA protein enzymatic functions by dATP. Biochem. 28, 5871-5881.

187 Merrick, M.J. (1993). In a class of its own — the RNA polymerase factor (a^). Molec. Microbiol. 10(5), 903-909.

Meyer, R.R., Voegele, D.W., Ruben, S.M., Rein, D C. and Trela, J.M. (1982). Influence of single-stranded DNA-binding protein on recA induction in Escherichia coli. Mutat. Res. 94, 299-313.

Milstein, C. and Kohler, G. (1975). Continuous cultures of fused cells secreting antibodies of pre-defmed specificity. Nature 256,475-497.

Miller, M.C., Resnick, J.B., Smith, B.T. and Lovett Jr., C M. (1996). The Bacillus subtilis dinR gene codes for the analogue of Escherichia coli LexA. Journal o f Biol. Chem. 271(52), 33502-33508.

Miller, R.V. and Kokjohn, T.A. (1990). General microbiology of recA\ Environmental and evolutionary significance. Annual Review Microbiology 44,365-394.

Minnikin, D.E. (1982). Lipids: Their chemistry, biosynthesis and roles. In: The biology o f mycobacteria, Vol. 1. Physiology, Identification and Classification. (Edited by Ratledge, C. and Stanford, J.) pp 95-184. Academic press, London.

Minnikin, D.E., Minnikin, S.M., Parlett, J.H., Goodfellow, M. and Magnuson, M. (1984). Mycolic acid patterns of some mycobacteria. Arch. Microbiol. 139,225-231.

Mishra, R.K., Gopal, V. and Chatterji, D. (1990). Correlation between the DNA supercoiling and the initiation of transcription by Escherichia coli RNA polymerase in vitro: the role of sequences upstream of the promoter region. FEBS letters. 260(2), 273-276.

Mironov, V.M., Montagu, M.V. and Inze, D. (1995). High throughput RNase protection assay. Nucleic Acid Res. 23(16), 3359-33360.

Moran Jr., C.P., Lang, N., LeGrice, S.F.J., Lee, G., Stephens, M., Sonenshein, A.L., Pero, J. and Losick, R. (1982). Nucleotide sequences that signal the initiation of transcription and translation in Bacillus subtilis. Journal o f Bacteriol. Molec. Gen. Genet. 186, 339-346.

Moreau, P.L. (1987). Effects of overproduction of single stranded DNA binding protein on RecA protein dependent processes in Escherichia coli. Journal o f Molec. Biol. 194, 621-634.

Morel-Deville, F. and Dusko Ehrlich, S. (1996). Theta-type DNA replication stimulates homologous recombination in the Bacillus subtilis chromosome. Molec. Microbiol. 19(3), 587-598.

188 Morgan, R.W., Christman, M F., Jacobson, F.S., Storz, G. and Ames, B.N. (1986). Hydrogen peroxide inducible proteins in Salmonella typhimurium overlap with heat shock and other stress proteins. Proc. Natl. Acad. Sci. USA. 83, 8059-8063.

Morimatsu, K. and Horii, T. (1995). DNA-binding surface of RecA protein. Europ. J. Biochem. 234, 695- 705.

Moss, M.T., Malik, Z.P., Tizard, M.L.V., Green, E.P., Sanderson, J.D. and Hermon-Taylor, J. (1992). IS902, an insertion element of the chronic-enteritis-causing Mycobacterium avium subsp. silvaticum. Journal Gen. Microbiol. 138, 139-145.

Mount, D.W. (1977). A mutant of Escherichia coli showing constitutive expression of the lysogenic induction and error-prone DNA repair pathways. Proc. Natl. Acad. Sci. USA 74(1), 300-304.

Movahedzadeh, F. (1996). Regulation of Mycobacterium tuberculosis RecA expression. PhD. thesis.

Movahedzadeh, F., Colston, M.J. and Davis, E.O. (1997a). Characterisation of Mycobacterium tuberculosis LexA: recognition of a Cheo {Bacillus-typQ 80S) box. Microbiology 143, 929-936.

Movahedzadeh, F., Colston, M.J. and Davis, E.O. (1997b). Determination of DNA sequences required for regulated Mycobacterium tuberculosis recA expression in response to DNA-damaging agents suggests that two modes of regulation exist. Journal o f Bacteriol. 179(11), 3509-3518.

Murray, A., Winter, N., LaGranderie, M., Hill, D.F., Rauzier, J., Timm, J., LeClerc, C., Moriarty, K.M. and Gicquel, B. (1992). Expression of Escherichia co/i-P-galactosidase in Mycobacterium bovis BCG using an expression system from Mycobacterium paratuberculosis which induced humoral and cellular

immune responses. Molec. Microbiol. 6 , 3331-3342.

Murray, C.J.L., K. Styblo and A. Rouillon (1990). Tuberculosis in developing countries: burden intervention and cost. Bull. Int. Union against Tuberculosis and Lung Dis. 65(1), 6-25.

Mustafa, A.B. (1995). Genetic transformation of mycobacteria by homologous recombination. Supplement to Nutrition 11(5), 670-673.

Nakagawa, K-I., Morishima, N. and Shibata, T. (1992). An endonuclease with multiple cutting sites, Endo. Seel, initiates genetic recombination at its cutting site in yeast mitochondria. The EMBO Journal 11(7), 2707-2715.

189 Narain, J.P., Raviglione, M.C. and Kochi, A. (1992). HIV associated tuberculosis in developing countries: epidemiology and strategies for prevention. Tubercle and Lung Dis. 73, 311-321.

Nesbitt, C.E., Levin, M., Donnelly-Wu, M.K. and Hatfull, G.F. (1995). Transcriptional regulation of repressor synthesis in mycobacteriophage L5. Molec. Microbiol. 17, 1045-1056.

Nordeen, S.K., Lopez-Bravo, L. and Sundaresan, T.K. (1992). Estimated numbers of leprosy cases in the world. Leprosy Rev. 63, 282-287.

Norman, E., Dellagostin, O.A., McFadden, J.J. and Dale, J.W. (1995). Gene replacement by homologous recombination in Mycobacterium bovis BCG. Molec. Microbiol. 16(4), 755-760.

Notani, N.K. and Setlow, J.E. (1980). Inducible repair system in Haemophilus influenzae. Journal of Bacteriol. 143, 516-519.

Nupbaumer, B. and Wohlleben, W. (1994). Identification, isolation and sequencing of the recA gene of Streptomyces lividans TK24. FEMS Micro. Lett. 118, 57-64.

Ortalo-Magne, A., Dupont, M.A., Lemassu, A., Andersen, A.B., Gounon, P. and Daffe, M. (1995). Molecular composition of the outermost capsular material of the tubercle bacillus. Microbiology 141, 1609- 1620.

Ottenhof, T.H.M. and Mutis, T. (1995). Role of cytotoxic cells in the protective immunity against and immunopathology of intracellular infections. Europ. Journal Clin. Invest. 25, 371-377.

Ottleben, G., Messori, L., Rutterjans, H., Kaptein, R., Granger-Schnarr, M. and Schnarr, M. (1991). H- NMR investigation of the interaction of the amino terminal domain of the LexA repressor with a synthetic half operator. Journal ofBiomolec. Struct, and Dynam. 9(3), 447-461.

Owttrim, G.W. and Coleman, J R. (1989). Regulation of expression and nucleotide sequence of Anabeana variabilis recA gene. Journal Bacteriol. 171, 5713-5719.

Pacelli, L.Z., Edmiston, S.H. and Mount, D.W. (1979). Isolation and characterisation of amber mutations in the lexA gene of Escherichia coli K-12. Journal ofBacteriol. 137(3), 568-573.

Papavinasasundaram, K.G., Movahedzadeh, F., Colston, M.J. and Davis, E.O. (1997b). Constmction and complementation of Mycobacterium smegmatis reveals that the intein in Mycobacterium tuberculosis does

190 not affect RecA fimction. Submitted.

Papavinasasundaram, K.G., Movahedzadeh, F., Keer, J.T., Stoker, N.G., Colston, M.J. and Davis, E.O. (1997). Mycobacterial recA is cotranscribed with a potential regulatory gene called recX. Molec. Microbiol. 24(1), 141-153.

Parsons, C.A., Tsaneva, I., Lloyd, R.G. and West, S.C. (1992). Interaction of Escherichia coli RuvA and RuvB proteins with synthetic Holliday junctions. Proc. Natl. Acad. Sci. USA 89, 5452-5456.

Pascopella, L., Collins, F.M., Martin, J.M., Lee, M.H., Hatfull, G.F., Stover, C.K., Bloom, B.R. and Jacobs Jr., W.R. (1994). Use of in vivo complementation in Mycobacterium tuberculosis to identify a genomic fragment associated with virulence. Infect, and Immun. 62(4), 1313-1319.

Pearce, B.J, Naughton, A.M., Campbell, E.A. and Masure, H R. (1995). The rec locus, a competence induced operon in Streptococcus pneumoniae. Journal of Bacteriol. 177, 86-93.

Peat, T.S., Frank, E.G., McDonald, J.P., Levine, A.S., Woodgate, R. and Hendrickson, W.A. (1996). Structure of the UmuD’ protein and its regulation in response to DNA damage. Nature 380,727-730.

Pelicic, V., Jackson, M., Reyrat, J.M., Jacobs Jr., W.R., Gicquel, B. and Guilhot, C. (1997). Efficient allelic exchange and transposon mutagenesis m. Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 94, 10955- 10960.

Pelicic, V., Reyrat, J.M. and Gicquel, B. (1996). Generation of unmarked directed mutations in mycobacteria, using sucrose counter-selectable suicide vectors. Molec. Microbiol. 20(5), 919-925.

Perler, F.B., Comb, D.G., Jack, W.E., Moran, L.S., Qiong, B., Kucera, R.B., Brenner, J., Slatko, BE., Nwankwo, D.O., Hempstead, S.K., Carlow, C.K.S. and Jannasch, H. (1992). Intervening sequences in an archaea DNA polymerase gene. Proc. Natl. Acad. Sci. USA. 89, 5577-5581.

Pharmacia: Laboratory separation division. (1986-8). Affinity Chromatography: principles and methods, pp 6 - 94. Printed by Ljungforetagen (Sweden).

Phillips, I., Culebras. E., Moreno, F. and Baquero, F. (1987). Induction of the SOS response by new 4- quinolones. Journal Antimicrob. Chemother. 20,631-638.

Phizicky, E.M. and Roberts, J.W. (1981). Induction of SOS fimctions: regulation of proteolytic activity of Escherichia coli RecA protein by interaction with DNA and nucleotide triphosphate. Cell 25,259-267.

191 Piddock, L.J.V. and Wise, R. (1987). Induction of the SOS response in Escherichia coli by 4-quinolone antimicrobial agents. FEMS Micro. Lett. 41, 289-294.

Podyminogin, M.A., Meyer, R.B. and Camper, H.B. (1995). Sequence-specific covalent modification of DNA by cross-linking oligonucleotides. Catalysis by RecA and implications for the mechanism of synaptic joint formation. Biochem. 34, 13098-13108.

Predich, M., Doukhan, L., Nair, G. and Smith, I. (1995). Characterisation of RNA polymerase ands two sigma-factors genes from Mycobacterium smegmatis. Molec. Microbiol. 15(2), 355-366.

Preobrajenskaya, O., Boullard, F., Boubrik, F., Schnarr, M. and Rouviere-Yaniv, J. (1994). The protein HU can displace the LexA repressor from its DNA-binding sites. Molec. Microbiol. 13(3), 459-467.

Pribnow, D. (1975a). Nucleotide sequence of an RNA polymerase binding site at an early T7 promoter. Proc. Natl. Acad. Sci. USA 72(3), 784-788.

Pribnow, D. (1975b). Bacteriophage T7 early promoters; Nucleotide sequences of two RNA polymerase binding sites. Journal Molec. Biol. 99,419-443.

QIAGEN. (1995). Standard plasmid mini-preparation, preparation of large plasmids, preparation of low-copy plasmids and cosmids and purification of plasmid DNA prepared by other methods. QIAPREP Plasmid Handbook 03/95.

Radding, C M. (1991). Helical interactions in homologous pairing and strand exchange driven by RecA protein. Journal o f Biol. Chem. 266(9), 5355-5358.

Radford, A.J. and Hodgson, A.L.M. (1991). Construction and characterisation of a mycobacterium- Escherichia coli shuttle vector. Plasmid 25, 149-153.

Rajagopalan, M., Chi, L., Woodgate, R., O’Donnell, M., Goodman, M.F. and Echols, H. (1992). Activity of purified mutagenesis proteins UmuC, UmuD’ and RecA in replicative bypass of an abasic DNA lesion by DNA polymerase III. Proc. Natl. Acad. Sci. USA 89, 10777-10781.

Ranes, M.G., Rauzier, J., LaGranderie, M. and Gicquel, B. (1990). Functional analysis of pAL5000, a plasmid from Mycobacterium fortuitum: Construction of a mini mycobacteria-£sc/ienc/im coli shuttle vector. Journal o f Bacteriol. 172: 2793-2797.

192 Rastogi, N., Frehel, C. and David, H.L. (1986). Triple layered structure of mycobacterial cell wall: Evidence for the existence of a polysaccharide-rich outer layer in 18 mycobacterial species. Current Microbiol. 13, 237-242.

Raymond-Denise, A. and Guillen, N. (1991). Identification of dinR, a DNA damage inducible regulator gene of Bacillus subtilis. Journal o f Bacteriol. 173(22), 7084-7091.

Raymond-Denise, A. and Guillen, N. (1992). Expression of the Bacillus subtilis dinR and recA genes after DNA damage and during competence. Journal o f Bacteriol. 174(10), 3171-3176.

Rehrauer, W.M. and Kowalczykowski, S.C. (1996). The DNA binding sites of the Escherichia coli RecA ^xoX&m. Journal Biol. Chem. 271(20), 11996-12002.

Rehrauer, W.M., Lavery, P.E., Palmer, E.L., Singh, R.N. and Kowalczykowski, S.C. (1996). Interaction of Escherichia coli RecA protein with the LexA repressor. Journal Biol. Chem. 271(39), 23865-23873.

Relan, N.K., Jenuwine, E.S., Gumbs, O H. and Shaner, S.L. (1997). Preferential interactions of the Escherichia coli LexA repressor with anions and protons are coupled to binding the recA operator. Biochemistry 36., 1077-1084.

Reyrat, J.M., Berthet, F.X. and Gicquel, B. (1995). The urease locus of Mycobacterium tuberculosis and its utilisation for the demonstration of allelic exchange m Mycobacterium bovis BCG. Genetics 92, 8768-8772.

Riddles, P.W., Blakeley, R.L. and Zerner, B. (1983). Reassessment of Ellmans reagent. Methods in Enzymology 91, 49-61.

Riera, J. and Barbe, J. (1993). Sequence of the Providencia rettgeri lexA gene and its control region. Nucleic Acid Res. 21, 2256.

Riera, J. and Barbe, J. (1995). Cloning, sequencing and regulation of expression of the lexA gene of Aeromonas hydrophila. Gene 154, 71-75.

Rivera, E., Vila, L. and Barbe, J. (1996). The uvrB gene of Pseudomonas aeruginosa is not DNA damage inducible. Journal ofBacteriol. 178(18), 5550-5554.

Rivoire, B., Pessolani, M. C.V., Bozic, C M., Hunter, S.W., Hefta, S.A., Mehra, V. and Brennan, P.J. (1994). Chemical definition, cloning and expression of the major protein of the leprosy bacillus. Infect, and Immun. 62(6), 2417-2425.

193 Roca, A I. and Cox, M.M. (1990). The RecA protein: Structure and function. Critical Reviews in Biochemistry and Molecular Biology 25(6), 415-456.

Roitt, I., Brostoff, J. and Male, D. (1993). Antibodies and Antigens. In Immunology, S"' edition. Mosby.

Roland, K.L. and Little, J.W. (1990). Reaction of LexA repressor with diisopropyl fluorophosphate. Journal Biol. Chem. 265(22), 12828-12835.

Roland, K.L., Smith, M.H., Rupley, J.A. and Little, J.W. (1992). In vitro analysis of mutant LexA proteins with an increased rate of specific cleavage. Journal Molec. Biol. 228, 395-408.

Rosenberg, M. and Court, D. (1979). Regulatory sequences involved in the promotion and termination of RNA transcription. Annu. Rev. Genet. 13, 319-353.

Rosselli, W. and Stasiak, A. (1990). Energetics of RecA-mediated recombination reactions; without ATP hydrolysis RecA can mediate polar strand exchange but is unable to recycle. Journal Molec. Biol. 216, 335- 352.

Rosselli, W. and Stasiak, A. (1991). The ATPase activity of RecA is needed o push the DNA strand exchange through heterologous regions. EMBO Journal 10(13), 4391-4396.

Sakula, A. (1983). BCG: who were Calmette and Guerin? Thorax 38, 806-812.

Sancar, A. and Sancar, G.B. (1988). DNA repair enzymes. Annu. Rev. Biochem. 57, 29-68.

Sambrook, J., Fritscb, E.F. and Maniatis, T. (1989). Molecular Cloning, a Laboratory Manual. T'^ edition. Cold Spring Harbour laboratory, Cold Spring Harbour, New York.

Sassanfar, M. and Roberts, J.W. (1990). Nature of the SOS inducing signal in Escherichia coli: The involvement of DNA replication. Journal Molec. Biol. 2212, 79-96.

Sauer, R.T. (1990). Scissors and helical forks. Nature 347, 514-515.

Sawitzke, J.A. and Stahl, F.W. (1992). Phage X has an analogue o f Escherichia coli recO, recR and ree f genes. Genetics. 130, 7-16.

Schnarr, M., Pouyet, J., Granger-Schnarr, M. and Duane, M. (1985). Large scale purification, oligomerisation, equilibria and specific interaction of the LexA repressor of Escherichia coli. Biochemistry

194 24, 2812-2818.

Schnarr, M., Granger-Schnarr, M., Hurstel, S. and Pouyet, J. (1988). The carboxy terminal domain of the LexA repressor oligomerizes essentially as the entire protein. Febs. Lett. 234, 56-60.

Schnarr, M., Oertel-Buchheit, P., Kazmaier, M. and Granger-Schnarr, M. (1991). DNA binding properties of the LexA repressor. Biochimie 73,423-431.

Schuldiner, S., Agmon, V., Brandsma, J., Cohen, A., Friedman, E. and Padan, E. (1986). Induction of 80S functions by alkaline intracellular pH in. Escherichia coli. Journal ofBacteriol. 168, 936-939.

Schwartz, C.E., Krall, J., Norton, L., McKay, K., Kay, D. and Lynch, R.E. (1983). Catalase and superoxide dismutase in Escherichia coli. Journal Biol. Chem. 258, 6277-6281.

Sedgwick, S. and Goodwin, P.A. (1985). Interspecies regulation of the SOS response by the Escherichia coli lexA gene. Mutation Res. 145, 103-106.

Sela, S. and Clark-Curtiss, J.E. (1991). Cloning and characterisation of the Mycobacterium leprae putative ribosomal RNA promoter in Escherichia coli. Gene 98, 123-127.

Sevastopoulos, C.G., Wehr, C.T. and Glaser, D.A. (1977). Large scale automated isolation of Escherichia coli mutants with thermosensitive DNA replication. Proc. Natl. Acad. Sci. USA 74, 3485-3489.

Shan, Q., Cox, M.M. and Inman, R.B. (1996). DNA strand exchange promoted by RecA K72R. Journal o f Biol. Chem. 271(10), 5712-5724.

Shaner, S.L. and E.S. Gaissarian (1996). Binding of Escherichia coli LexA repressor to the recA operator. Journal Molec. Recognition 9,468-473.

Sherman, D R., Sabo, P.J., Hickey, M.J., Arain, T.M., Mahairas, G.G., Yuan, ¥., Barry HI, C.E. and Stover, K.C. (1995). Disparate responses to oxidative stress in saprophytic and pathogenic mycobacteria. Proc. Natl. Acad. Sci. USA 92, 6625-6629.

Shiba, T., Iwasaki, H., Nakata, N. and Shinagawa, H. (1991). SOS-inducible DNA repair proteins RuvA and RuvB, of Escherichia coli: Functional interactions between RuvA and RuvB for ATP hydrolysis and renaturation of the cruciform stmcture in supercoiled DNA. Proc. Natl. Acad. Sci. USA 88, 8445-8449.

195 Shibata, T., Cunningham, R.P., DasGupta, C. and Radding, C M. (1980). Homologous pairing in genetic recombination: formation of D-loops by combined action of RecA protein and a helix-destabilising protein. Proc. Natl. Acad. Sci. USA 77, 2606-2610.

Shibata, T., Ikeda, M., Makino, O. and Ikawa, S. (1991). Monoclonal antibodies and mechanistic studies on RecA protein. Biochimie 73, 209-217.

Shinagawa, H. (1996). SOS response as an adaptive response to DNA damage in prokaryotes. In: Stress inducible cellular responses, pp221-235. Edited by Feige, U., Morimoto, R.I., Yahara, I. and Polla, B. Birkhauser Verlag Basel. Switzerland.

Shinnick T.M. (1987). The 65kDa antigen of Mycobacterium tuberculosis. Journal o f Bacteriol. 169, 1080- 1088.

Shinnick, T.M., King, C.H. and Quinn, F.D. (1995). Molecular biology, virulence and pathogenicity of mycobacteria. The Amer. Journal Med. Sci. 309(2), 92-97.

Silva, M.T. and Macedo, P.M. (1984). Structural characterisation of normal and damaged membranes of Mycobacterium leprae and cultivable mycobacteria. Journal Gen. Microbiol. 130, 369-380

Silverton, R.E and Anderson, M.J. (1961). Handbook of medical laboratory formulae, pp9. Butterworths. London.

Simic, D., Vukovic-Gacic, B., Ajanovic, A. and Knezevic-Vukcevic, J. (1991). Activation of RecA protein in recombination deficient strains of Escherichia coli following DNA-damaging agents. Mutat. Res., DNA repair. 254, 255-262.

Slilaty, S.N. and Little, J.W. (1987). Lysine-156 and serine-119 are required for LexA repressor cleavage: a possible mechanism. Proc Natl. Acad. Sci. USA. 84, 3987-3991.

Smith, G.R. (1989). Homologous recombination in Escherichia coll. Multiple pathways for multiple reasons. Ce//58, 807-809.

Smith, G.R. (1994). Hotspots of homologous recombination. Experientia 50, 234-241.

Smith, M.H., Cavenagh, M.M. and Little, J.W. (1991). Mutant LexA proteins with an increased rate of in vivo cleavage. Proc. Natl. Acad. Sci. USA 88, 7356-7360.

196 Snapper, S B., Lugosi, L., Jekkel, A., Melton, R., Keiser, T., Bloom, B.R. and Jacobs, W.R. (1988). Lysogeny and transformation of mycobacteria: stable expression of foreign genes. Proc. Natl. Acad. Sci. USA 85, 6987-6991.

Snapper, S B., Melton, R.E., Mustafa, S., Kieser, T. and Jacobs Jr., W.R. (1990). Isolation and characterisation of efficiency plasmid transformation mutants of Mycobacterium smegmatis. Molec. Microbiol. 4(11), 1911-1919.

Snyder, M. and Drlica, K. (1979). DNA gyrase on the bacterial chromosome: DNA cleavage induced by oxolinic acid. Journal Molec. Biol. 131, 287-302.

Stamm, L.V., Frye, J.G. and Hardham, J.M. (1995). Sequence of the Leptospira biflexa serovar patoc recA gene. Gene 167, 339-340.

Stasiak, A. (1996). Getting down to the core of homologous recombination. Science 272, 828-829.

Steele, J.H. and Ranney, A.F. (1958). Animal tuberculosis. Amer. Rev. Tuberc. 77, 908-922.

Story, R.M., Weber, I.T. and Steitz, T.A. (1992). The stmcture of Escherichia coli RecA protein monomer and polymer. Nature 355, 318-325.

Storz, G., Tartaglia, L A. and Ames, B.N. (1990). Transcriptional regulator of oxidative stress-inducible genes: Direct activation by oxidation. Science 248, 189-194.

Stover, C.K., Bansal, G.P., Hanson, M.S., Burlein, J.E., Palaszynski, S R., Young, J.F., Koenig, S., Young, D.B., Sadziene, A. and Barbour, A.G. (1993). Protective immunity elicited by recombinant Bacilli Calmette-Guerin (BCG) expressing outer surface protein A (OspA) lipoprotein, a candidate for Lyme disease vaccine. Journal Exp. Med. 178, 197-209.

Stover, C.K., de la Cruz, V.F., Fuerst, T.R., Burlein, J.E., Benson, L.A., Bennet, L.T., Bansal, G.P., Young, J.F., Lee, M.H., Hatfull, G.F., Snapper, S B., Barletta, R.G., Jacobs, W.R. and Bloom, B.R. (1991). New use of BCG recombinant vaccines. Nature 351,456-460.

Strauss, D.B., Walter, W.A. and Gross, C.A. (1987). The heat shock response of Escherichia coli is regulated by changes in the concentration of Nature 329, 348-351.

Strohl, W.R. (1992). Compilation and analysis of DNA sequences associated with apparent Streptomyces promoters. Nucleic Acid Res. 20, 961-974.

197 Sugino, A., Peebles, C.L., Kreuzer, K.N. and Cozzarelli, N.R. (1977). Mechanism of action of nalidixic acid: purification of Escherichia coli nalA gene product and its relationship to DNA gyrase and a novel nick closing enzyme. Proc. Natl. Acad. Sci.USA. 74,4767-4771.

Suzuki, Y, A Nagata and T Yamada. (1991). Analysis of the promoter region in the rRNA operon from Mycobacterium bovis BCG. Antonie Leeuwenhoek. 60, 7-11.

Takewaki, S. I., Okuzumi, K., Manabe, I., Tanimura, M., Miyamura, K., Nakahara, K.I., Yazaki, Y., Ohkubo, A. and Nagai, R. (1994). Nucleotide sequence comparison of the mycobacterial dnaJ gene and PCR-restriction fragment length polymorphism analysis for identification of mycobacterial species. InternationalJournal o f Systemic Bacteriol. 44(1), 159-166.

Tao, L., Hollingshead, S.K., Suvorov, A.N., Ferretti, J.J. and McShan, W.M. (1995). Constmction of a Streptococcus pyogenes recA mutant via insertional inactivation, and cloning and sequencing of the complete recA gene. Gene 162, 59-62.

Takahashi, M., Maraboeuf, F. and Norden, B. (1996). Locations of the functional domains in the RecA protein. Overlap of domains and regulation of activities. Europ. Journal Biochem. 242, 20-28.

Tanaka, K., Fujita, N., Ishihama, A. and Takahashi, K. (1993). Heterogeneity of the principal sigma factor in Escherichia coli: the rpoS gene product, is a principal sigma factor of RNA polymerase in stationary phase Escherichia coli. Proc. Natl. Acad. Sci. USA 90, 3511-3515.

Tanaka, K., Kusano, S., Fujita, N., Ishihama, A. and Takahashi, K. (1995). Promoter determinants for Escherichia coli RNA polymerase holoenzyme containing (the rpoS gene product). Nucleic Acids Res. 23, 827-834.

Tapias, A., Fernandez De Henestrosa, A.R. and Barbe, J. (1997). Characterisation of the promoter of the Rhizobium etli recA gene. Journal o f Bacteriol. 179(5), 1573-1579.

Tatti, K.M. and M oran Jr., C.P. (1995). changed to a® specificity by amino acid substitution in its -10 binding region. Journal o f Bacteriol. 177(22), 6506-6509.

Taylor, A.F. and Smith, G.R. (1995). Strand specificity of nicking of DNA at Chi sites by RecBCD enzyme. Journal Biol. Chem. 270(41), 24459-24467.

Telenti, A., Southworth, M., Alcaide, F., Daugelat, S., Jacobs Jr., W.R and Perler, F.B. (1997). The Mycobacterium xenopi GyrA protein splicing element: Characterisation of a minimal intein. Journal o f

198 Bacteriol. 179(20), 6378-6382.

Teng, S.P., Woodson, S.A. and Crothers, D.M. (1989). DNA sequence specificity of mitomycin cross- linking. Biochemistry 28, 3901-3907.

Thierry, D., Caves, M.D., Eisenach, K.D., Crawford, J.T., Bates, J.H., Gicquel, B. and Guesdon, J.L. (1990a). IS6110, an IS like element of Mycobacterium tuberculosis complex. Nucleic Acid Res. 18, 188.

Thierry, D., Brisson-Noel, A., Vincent-Levy-Frebault, V., Nguyen, S., Guesdon, J. L. and Gicquel, B. (1990b). Characterisation of a Mycobacterium tuberculosis insertion sequence, IS6110, and its application in diagnosis. Journal Clin. Microbiol. 28(12), 2668-2673.

Thiiveris, A T. and Mount, D.W. (1992). Genetic identification of the DNA binding domain of Escherichia coli LexA protein. Proc. Natl. Acad. Sci. USA 89,4500-4504.

Thiiveris, A T., Little, J.W. and Mount, D.W. (1991). Repression of the Escherichia coli recA gene requires at least two LexA protein monomers. Biochimie 73,449-455.

Thole J.E., Dauwerse, H.G., Das, P.K., Groothuis, D.G., Schools, L.M. and VanEmbden, J.D.A. (1985). Cloning of Mycobacterium bovis DNA and expression of antigens in Escherichia coli. Infect, and Immun. 50, 800-806.

Thole J.E., Keulen, W.J., Kolk, A.H.J., Groothuis, D.G., Berward, L.G. and VanEmbden, J.D.A. (1987). Characterisation, sequence determination and immunogenicity of a 64kDa protein of Mycobacterium bovis expressed 'm Escherichia coli K12. Infect. Immun. 55, 1466-1475.

Thoms, B. and Wackernagel, W. (1987). Regulatory role of recF in the SOS response of Escherichia coli: impaired induction of SOS genes by UV-irradiation and nalidixic acid in a recF mutant. Journal o f Bacteriol. 157,498-506.

Thompson, S. A. and Blaser, M.J. (1995). Isolation of the Helicobacter pylori recA gene and involvement of the recA region in resistance to low pH. Infect, and Immun. 63(6), 2185-2193.

Thresher, R.J., Christiansen, G. and Griffith, J.D. (1988). Assembly of presynaptic filaments. Factors affecting the assembly of RecA protein onto single-stranded DNA. Journal Molec. Biol. 201, 101-113.

Timm, J., Lim, E.M. and Gicquel, B. (1994a). Escherichia coli- mycobacteria shuttle vectors for operon and gene fusions to lacZ\ the pJEM series. Journal ofBacteriol. 176(21), 6749-6753.

199 Timm, J., Periili, M.G., Duez, C., Trias, J., Orefici, G., Fattorini, L., Amicosante, G., Oratore, A., Joris, B., Frere, J.M., Pugsley, A.P. and Gicquel, B. (1994b). Transcription and expression analysis, using lacZ and phoA gene fusions of Mycobacterium fortuitum (3-lactamase genes cloned from a natural isolate and high level P-lactamase producer. Molec. Microbiol. 12,491-504.

Tsukamura, M. (1990). Growth physiology of mycobacteria in modified Dubos liquid medium. Microbiol. Immunol. 34(12), 995-1003.

Umelo, E., Noonan, B. and Trust, T.J. (1996). Cloning, characterisation and expression of the recA gene of Aeromonas salmonicida. Gene 175, 133-136.

Umezu, K., Chi, N.W. and Kolodner, R.D. (1993). Biochemical interaction of the Escherichia coli RecF, RecO and RecR protein and single strand binding protein, Proc. Natl. Acad. Sci. USA 90, 3875-3879.

Umezu, K. and Kolodner, R.D. (1994). Protein reactions in genetic recombination in coli. Interactions involving RecO and RecR overcome the inhibition of RecA by single strand DNA binding protein. Journal Biol. Chem. 269, 30005-30013.

Uptain, S.M., Kane, C M. and Chamberlin, M.J. (1997). Basic mechanisms of transcript elongation and its regulation. Annu. Rev. Biochem. 66 , 117-172.

Vales, L.D., Chase, J.W. and Murphy, J.B. (1980). Effect of the ssbAl and lexC-113 mutations on lambda prophage induction, phage growth and cell survival. Journal of Bacteriol. 143, 887-896.

Van-Soolingen, D., Hermans, P.W.M., De Haas, P.E.W. and Van Embden, J.D.A. (1992). Insertion element IS 1081-associated restriction fragment length polymorphism’s in Mycobacterium tuberculosis complex species: a reliable tool for recognising Mycobacterium bovis BCG. Journal Clin. Microbiol. 30(7), 1772-1777.

Verma, A, Kinger, A.K. and Tyagi, J.S. (1994). Functional analysis of transcription of Mycobacterium tuberculosis 16S rDNA-encoding gene. Gene 148, 113-118.

Volkert, M R. and Hartke, M.A. (1984) Suppression of Escherichia coli recF mutation by rec.4-linked srfA mutation. Journal o f Bacteriol. 157,498-506.

Volkert, M R., Margossian, L.J. and Clark, A.J. (1984). Two component suppression of recF143 by recA441 in Escherichia coli K12. Journal o f Bacteriol. 160,702-705.

200 Voloshin, O.N., Wang, L. and Camerini-Otero, R.D. (1996). Homologous DNA pairing promoted by a 20- amino acid peptide derived from RecA. Science 272, 868-872.

Voskuil, M.I., Voepel, K. and Chambliss, G.H. (1995). The -16 region, a vital sequence for the utilisation of a promoter in Bacillus subtilis and Escherichia coli. Molec. Microbiol. 17(2), 271-279.

Walker, G.C. (1984). Mutagenesis and inducible responses to DNA damage m Escherichia coli. Microbiological Reviews 48(1), 60-93.

W alker, G.C. (1985). Inducible DNA repair systems. Annu. Rev. Biochem. 54,424-457.

W alker, G.C. (1995). SOS regulated proteins in translesion DNA synthesis and mutagenesis. TIBS 20,416- 420.

Wang, T.C.V., Chang, H.Y. and Hung, J.L. (1993). Cosuppression of recF mutation, recR mutation and recO mutation by mutant recA alleles in Escherichia coli cells. Mutat. Res. 294, 157-166.

Wards, B.J. and Collins, D.M. (1996). Electroporation at elevated temperatures substantially improves transformation efficiency of slow growing mycobacteria. FEMS Microbiol. Lett. 145, 101-105.

Watt, V.M., Ingles, C.J., Urdea, S.M. and Rutter, W.J. (1996). Homology requirements of recombination in Escherichia coli. Proc. Natl. Acad. Sci. USA 82,4768-4772.

Wayne, L.G. (1976). Dynamics of submerged growth of Mycobacterium tuberculosis under aerobic and microaerophillic conditions. Amer. Rev. Respir. 114, 807-811.

Wayne, L.G. (1994). Dormancy of Mycobacterium tuberculosis and latency of disease. Europ. Journal Clin. Microbiol, and Infect. Dis. 13(11), 908-914.

Wayne, L.G. and Hayes, L.G. (1996). An in vitro model for the sequential shift-down of Mycobacterium tuberculosis through two stages of non-rephcating persistence. Infect, and Immun. 64(6), 2062-2069.

Wayne, L.G. and Sramek, H.A. (1994). Metronidazole is bactericidal to dormant cells of Mycobacterium tuberculosis. Antimicrob. Agents and Chemother. 38(9), 2054-2058.

Wayne, L.G. and Lin, K.Y. (1982). Glyoxylate metabolism and adaptation of Mycobacterium tuberculosis to survival under anaerobic conditions. Infect .and Immun. 37,1042-1049.

201 Weisemann, J.M., Funk, C. and Weinstock, G.M. (1984). Measurement of in vivo expression of the recA gene

0Ï Escherichia co/i by using lacZ gene füsions. Journal o f Bacteriol. 160(1), 112-121.

Wertman, K.F and Mount, D.A. (1985). Nucleotide sequence binding specificity of the LexA repressor of Escherichia coli K-12. Journal o f Bacteriol. 163(1), 376-384.

West, S.C. (1992). Enzymes and molecular mechanisms of genetic recombination. Ann. Rev. Biochem. 61,603- 640.

West, S.C. (1994). The processing of recombination intermediates: Mechanistic insights from studies of bacterial proteins. Cell 76,9-15.

West, S.C., Cassuto, E. and Howard-Flanders, P. (1982). Role of the SSB protein in RecA promoted branch migration reactions. Molec. Gen. Genet. 186, 333-338.

West, S.C., Countryman, J.K. and Howard-Flanders, P. (1983). Purification and properties of the RecA protein of Proteus mirabilis. Journal o f Biol. Chem. 258, 4648-4654.

Westpheling, J., Ranes, M. and Losick, R. (1985). RNA polymerase heterogeneity in Streptomyces coelicolor. Nature 313, 22-26.

Whitby, M.C. and Lloyd, R.C. (1995). Altered SOS induction associated with mutations in recF, recO and recR. Molec. Gen. Genet. 246, 174-179.

Whittier, R.F. and Chase, J.W. (1981). DNA repair in Escherichia coli strains deficient in single strand binding proteins. A/o/ec. Gen. Genet. 190, 341-347.

Winterling, K.W., Levine, A.S., Yasbin, R E. and Woodgate, R. (1997). Characterisation of DinR, the Bacillus subtilis SOS repressor. Journal o f Bacteriol. 179(5), 1698-1703.

Wittung, P., Bazemore, L.R., Takahashi, M., Norden, B. and Radding, C. (1996). Second-site RecA- DNA interactions: Lack of identical recognition. Biochem. 35, 15349-15355.

Wittung, P., Ellouze, C., Maraboeuf, F., Takahashi, M. and Norden, B. (1997). Thermochemical and kinetic evidence for nucleotide sequence dependent RecA-DNA interactions. Europ. J. Biochem. 245, 715- 719.

202 Wojciechowski, M.F., Peterson, K.R. and Love, P.E. (1991). Regulation of the SOS response in Bacillus subtilis: Evidence for a LexA repressor homologue. Journal of Bacteriol. 173(20), 6489-6498.

Wolberger. C, Dong, Y.C., Ptashne, M. and Harrison, S.C. (1988). Structure of a phage 434 Cro/DNA complex. Nature. 335, 789-795.

Woodgate, R. and Levine, A S. (1996). Damage inducible mutagenesis: Recent insights into the activities of the Umu family of mutagenesis proteins. Cancer Surveys 28, 117-140.

Xu, M-Q., Sothworth, M.W., Mersha, F.B., Hornstra, L.J. and Perler, F.B. (1993). In vitro protein splicing of purified precursor and the identification of a branched intermediate. Cell. 75, 1371-1377.

Yancey-Wrona, J.E. and Camerini-Otero, R.D. (1995). The search for DNA homology does not limit stable homologous pairing promoted by RecA protein. Current Biology 5(10), 1149-1158.

Yasbin, R E. (1977). DNA repair in Bacillus subtilis. I. The presence of an inducible system. Molec. Gen. Genet. 153,211-218.

Yasbin, R E. (1977). DNA repair in Bacillus subtilis. II. Activation of the inducible system in competent bacteria. A/o/ec. Gen. Genet. 153, 219-225.

Yasbin, R.E., Cbeo, D.L. and Bayles, K.W. (1992). Inducible DNA repair and differentiation in Bacillus subtilis: Interactions between global régulons. Molec. Microbiol. 6(10), 12263-1270.

Young, D. and Garbe, T. (1991). Heat shock proteins and antigens of Mycobacterium tuberculosis. Infect, and Immun. 59(9), 3086-3093.

Young, D., Garbe, T., Latbigra, R., Abou-Zied, C. and Zbang, Y. (1991). Characterisation of prominent protein antigens from mycobacteria. Bull. Int. Union Tuberc. Lung Disease 66,47-51.

Young, D.B., Kauffmann, S.H.E., Hermans, P.W.M. and Tbole, J.E.R. (1992a). Mycobacterial protein antigens: a compilation. Molecular Microbiology 6(2), 133-145

Young, R.A., B.R Bloom, C.M Grosskinsky, J Ivanyi, D Thomas and RW Davis. (1985b). Dissection of Mycobacterium tuberculosis antigens using recombinant DNA. Proc Natl Acad. Sci USA 82, 2583-2587.

Young, R.A., V. Mebra, D Sweetser, T Buchanan, J Clark-Curtiss, R.W Davis and B.R Bloom. (1985). Genes for major protein antigens of leprosy parasite, Mycobacterium leprae. Nature 316,450-452.

203 Ysern, P., Clerch, B., Castano, M., Gilbert, I., Barbe, J. and Llagostera, M. (1990). Induction of SOS genes in Escherichia coli and mutagenesis in Salmonella typhimurium by fluoroquinolones. Mutagenesis 5, 63-66.

204 A p p e n d ix I: M e d ia r e c ip e s

Dubos Medium

Part A

Casamino acids 2-5g Asparagine 03g NazHPO^.UH^O 2 5g

KH2PO4 Ig Sodium citrate l-5g

MgS04.7H20 0 -6 g 10% Tween-80 5ml

Dissolve asparagine in hot water (50ml) and add to remaining constituents. Increase volume to 950ml with distilled water and adjust pH to 7.2 with 2M NaOH. Make up to 1000ml with distilled water, sterilise by autoclaving.

Part B Bovine serum albumin fraction V 5g Distilled water to 1000ml, filter sterilise.

For use mix 240ml part A with 10ml part B

Lemco Broth

Bacto-peptone lOg Bacto-Lab Lemco powder 5g NaCl 5g Distilled water to 1000ml (pH to 7.2 with 2M NaOH)

Sterilise by autoclaving.

Luria-Bertani agar (L-agar)

Bacto agar 15g Bacto-tryptone lOg

205 Bacto-yeast extract 5g NaCl lOg Distilled water to 1000ml

Sterilise by autoclaving.

Luria-Bertani medium (L-broth)

Bacto-tryptone lOg

Bacto-yeast extract 5g NaCl lOg Distilled water to 1000ml (sterilise by autoclaving)

Middlebrook 7H9 broth

Part A Ammonium sulphate 0 5g L-glutamic acid 0-5g Sodium citrate 0 -lg Pyridoxine OOOlg Biotin 00005g Disodium phosphate 2-5g

Monopotassium phosphate Ig Ferric ammonium citrate 004g Magnesium sulphate 005g Calcium chloride 00005g Zinc sulphate OOOlg Copper sulphate OOOlg Distilled water to 1000ml

Sterilise by autoclaving.

P artB

Glycerol 2ml Filter sterilise

206 Part C

OADC (Difco) 20ml (for cultivation of slow growing mycobacteria) or, ADC (Difco) 20 ml (for cultivation of fast growing mycobacteria)

For use mix 920ml part A with one part B and four parts C. Middlebrook 7H11 agar

Part A

Pancreatic digest of casein Ig L-glutamic acid 0-5g Sodium citrate 0-4g Pyridoxine OOOlg Biotin 00005g Ferric ammonium citrate 0 04g Ammonium sulphate 0 5g Disodium phosphate 1-5g Monopotassium phosphate l-5g Magnesium sulphate 005g Bacto agar 15g Malachite green OOOlg Distilled water to 1000ml

Sterilise by autoclaving.

P artB

Glycerol 2ml Filter sterilise

Part C

OADC 20ml (for cultivation of slow growing mycobacteria) or, ADC 20 ml (for cultivation of fast growing mycobacteria)

For use mix 920ml part A with one part B and four parts C.

Sautons media (without glycerol)

Part A

Sodium pyruvate 6 g

207 Sodium glutamate 6 g

Citric acid 2g K2HPO4 l-5g MgSO^.VHzO 0-25g Distilled water to 1000ml (adjust pH to 6.2 with ammonia)

P artB ZnClz 40mg

FeCl3.6H20 200mg CUCI2.2H2O lOmg MnCl2.4H20 lOmg

Na2B407.H20 lOmg

Distilled water to 1000ml

Sterilise part A and B by autoclaving.

Part C Biotin lOOpg Glucose 30g Dissolve biotin in 10ml very dilute NaOH and add to glucose, filter sterilise.

For use mix 990ml part A with 2ml part B and 10ml part C.

Terrific Broth

Bacto-tryptone 12g Bacto yeast extract 24g Glycerol 4ml

KH2PO4 2.31g K2HPO4 12.54g

Distilled water up to 1000ml

Sterilise by autoclaving.

208 Tween-Glutamate media

Part A

L-monosodium glutamate 8g Bacto-casitone Ig Ferric ammonium citrate 0-lg

KH2PO4 Ig Na^HPO^ 2-5g CaClz.ZHzO OOOlg

CUSO4.5H2O 00005g ZnS04.7H20 00005g 10% Tween-80 5ml Distilled water to 1000ml (sterilise by autoclaving)

P artB Bovine serum albumin fraction V 5g Distilled water to 1000ml, filter sterilise.

For use mix 230ml part A with 20ml part B.

209 A ppen d ix II: G r o w t h c u r v es

Cultures of M. smegmatis, M. tuberculosis and M. m icroti in their stationary phase were used to inoculate fresh media at 1 '250 and cultures grown as described (section 2.1 ). The change in optical density

(ODf^onm) with tim e w as m onitored for M. smegmatis (32 hours), M. microti (1 days) and M. tuberculosis

(28 days) and a log,o graph constructed. Prokaryotic growth cuiwes are typically sigmoidal in shape with clearly defined lag and logarithmic growth phases. However, the lag and logarithmic growth phase cannot be readily distinguished in the curves illustrated below; this obseiwation is in agreement with previously published data for mycobacteria (Fisher and Kirchheimer, 1952; Tsukamura, 1990). For the purposes of these investigations cultures of M. smegmatis, M. microti and M. tuberculosis were used at “equivalent growth stages”, which were determined by a visual assessment of their respective growth curves. Unless othei-wise stated M. tuberculosis and M. smegmatis were harvested at an OD^oonm of 0.6 and M. m icroti at an ODf,oonmOf0.35.

Pf

o o

0 01 0 5 10 15 20 25 30 35

Hours

Figure AII.l Growth ciir\ e for A/, smegmatis mc^l55 in Lemco broth

210 10

OC3 O o 0.1 f

0.01 0 5 10 15 2.0 25 30

Days

Figure AU.2 Growth curv e for M. tuberculosis H37Rv in Dubos media

21 1 un 5 û

-3 6 o f

0 01 1 2 3 4 5 6 7 8 9 10

Days

Figure AII.3 (jrowlh curv e for M. microti OV1254 in Tween-glutaiiiate media

212 A p p e n d ix III: P olyacrylamide g e l a n d b u f f e r r e c ip e s

10% Denaturing SDS-Polyacrylamide gel

Protogel (37-5:1 acrylamide to bisacrylamide stabilised solution, National Diagnostics) 8-3ml 1-5M Tris-HCl (pH8-7) 6-25ml Distilled water 10ml 10% SDS 0-25ml 10% Ammonium persulphate (Biorad) 120pl TEMED (Biorad) 25pi

15% Denaturing SDS-Polyacrylamide gel

Protogel (37-5:1 acrylamide to bisacrylamide stabilised solution, National Diagnostics) 12-5ml 1-5M Tris-HCl (pH8-7) 6-25ml

Distilled water 6 ml 10% SDS 0-25ml 10% Ammonium persulphate (Biorad) 120pl TEMED (Biorad) 25pl

4% Stacking Polyacrylamide gel

Protogel (37-5:1 acrylamide to bisacrylamide stabilised solution. National Diagnostics) 1 -3ml

1-5M Tris-HCl (pH 6 -8) 0-83ml Distilled water 7-87ml 10% Ammonium persulphate (Biorad) 25pi TEMED (Biorad) lOpl

8% Native polyacrylamide gel

Protogel (37-5:1 acrylamide to bisacrylamide stabilised solution. National Diagnostics) 19-95ml TBE(lOx) 3-75ml Distilled water 51.3ml 25% Ammonium persulphate (Biorad) 112pl TEMED (Biorad) 112pl

213 For a 6% native polyacrylamide gel use 15ml Protogel and 56 25ml distilled water.

Denaturing SDS-Polyacrylamide gel running buffer (lOx)

Tris-HCl (pH8-3) 30-3g Glycine 144g SDS lOg

For use dilute 1:10 in distilled water.

Running buffer for native polyacrylamide gels: TBE (lOx)

Tris base 121 Ig Boric acid 61 83g EDTA 186g Distilled water up to 1000ml (natural pH)

214 A ppen d ix IV: P o ster A b str a c t

Constitutive and Induced levels of RecA expression in mycobacteria

Nicola A Thomas, Peter Jeimer, Elaine O Davis, Patricia Brooks, Farahnaz Movahedzadeh and M. Joseph Colston.

The pathogenic mycobacteria favour an intracellular environment where they are exposed to a variety of DNA damaging agents as part of the host’s defence mechanism. Several systems exist to maintain the integrity of bacterial DNA; one of the most important of these is the SOS response, which involves the co-ordinated expression of several genes in response to exposure to DNA damaging agents. The RecA protein is a pivotal component of the SOS response; being involved in derepression of the SOS proteins via its interaction with the repressor protein LexA, and also in homologous recombination and mutagenesis repair.

In this study we have been investigating the expression of RecA in mycobacteria in order to understand its role in promoting intracellular survival of M tuberculosis and its involvement in homologous recombination. Using antibodies generated against M tuberculosis RecA, we have found that basal expression of the protein is significantly lower in some mycobacterial species, particularly M tuberculosis, than in others. Exposure to DNA damaging agents does appear to increase expression of RecA, and we are currently comparing the induction levels in different mycobacterial species. The use of anti-RecA antibodies will also enable us to investigate regulation of the unusual post-translational splicing reaction, which is involved in the maturation of RecA in M. tuberculosis.

Poster presented at the Third International Conference on the Pathogenesis of Mycobacterial Infections June 27-30, 1996 Saltsjobaden, Stockholm, Sweden

215