WO2009149915A1

WO2009149915A1 - Anti-bacterial ligase inhibitors

Info

Publication number: WO2009149915A1
Application number: PCT/EP2009/004176
Authority: WO
Inventors: Mark Odell
Original assignee: University Of Westminster
Priority date: 2008-06-10
Filing date: 2009-06-10
Publication date: 2009-12-17
Also published as: GB0810573D0

Abstract

A method of identifying an agent that inhibits or modulates the activity of a ligase, particularly an ATP-dependent ligase in which a polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group are ligated and a complementary polynucleotide template to both the first and second polynucleotides is present and bound to a surface for detection using surface plasmon resonance. Surface plasmon resonance allows identification of candidate agents that act to inhibit or modulate ligase activity through the binding or lack of binding of the enzyme and allows discrimination from candidate agents that act by inhibiting or modulating the catalytic steps of the enzyme during ligation.

Description

ANTI-BACTERIAL LIGASE INHIBITORS

This invention relates to anti-bacterial ligase inhibitors, an assay for identifying agents acting as ligase inhibitors, particularly a quantitative assay and the use of identified agents in treatment therapies.

Polynucleotide ligases are enzymes which create a covalent bond between discontinuous polynucleotides. This joining of discontinuous polynucleotides by polynucleotide ligases plays a central role in a number of natural biological processes, including chromosome replication, genetic recombination and cellular repair of environmental genetic damage (e.g., X-ray damage). DNA ligases catalyse the formation of phosphodiester bonds at breaks in DNA. They require the presence of a high energy co-factor (either adenosine triphosphate (ATP) or nicotinamide adenine dinucleotide (NAD+)), a 5'PO4 and 3'OH on adjacent strands and a divalent cation normallyMg²⁺. DNA nick repair is a three-step reaction.

DNA ligases catalyse the formation of phosphodiester bonds at breaks in DNA. They require the presence of a high energy co-factor (either ATP or NAD⁺), a 5'PO₄ and 3'OH on adjacent DNA strands and a divalent cation, normally Mg²⁺. DNA nick repair is a three-step reaction (see Scheme 1 1). All eubacteria possess a single, essential NAD⁺ dependent DNA ligase. Cellular organisms have a variable complement of ATP- dependent DNA ligase enzymes.

Scheme 1: General catalytic mechanism of DNA ligase enzymes Step 1 Enzyme + ATP or NAD⁺ --> Enzyme-pA + PPi energy cofactor Stable adenylated enzyme covalent intermediate

Step 2 Enzyme-pA + _PDNA --> Ap_pDNA-Enzyme DNA recognition and binding Stable adenylated DNA intermediate with enzyme bound

Step 3 DNA_0H + Ap_pDNA-Enzyme --> DNApDNA + AMP + Enzyme

Crystal structures of enzymes in complex with DNA provide a static picture and fail to show the dynamics of enzyme-DNA interactions. Gel retardation assays have been used to assess more dynamic aspects of proteins binding to specific sites on DNA. This technique has been applied to analyse DNA ligase-substrate interactions and the essential features of a DNA substrate that are required for stable DNA interaction by Paramecium bursaria Chlorella virus-1 (PBCV-1) ligase (Odell and Shuman, 1999). The limitation of gel retardation analysis as applied to ligase analysis is that it is a binary technique; either the protein binds or it does not. Only protein-DNA interactions that result in stable binding may be analysed. According to this invention, to provide a means of measuring ligase affinity for a DNA molecule, surface plasmon resonance (SPR) can be employed. SPR allows the direct measurement of real-time biomolecular interactions, such as those of proteins with DNA, even if those interactions do not result in the formation of stable complexes and 'real-time' binding kinetics can be measured.

Polynucleotide ligases from viruses, prokaryotic cells and eukaryotic cells have been described and characterized. In particular, NAD+ dependent polynucleotide ligases of the eubacteria and the ATP-dependent polynucleotide ligases of eukaryotes are well characterized and serve essential functions in such cells.

Compounds which modulate ligase activity can act as cytotoxic agents by disrupting cellular processes in normal cell division processes and DNA repair processes. Such modulators with adequate specificity can serve as antibacterial, antifungal, or antineoplastic agents. Similarly, as polynucleotide ligases are also virally encoded (or, in some cases, induced by viral infection), modulators of viral ligase activity can act as antiviral agents.

Polynucleotide ligase enzymes are vital for any manipulation of polynucleotide sequences that leaves a break in a polynucleotide sequence or unjoined polynucleotide sequences, for example DNA replication and repair. All eubacteria possess a single

NAD+ -dependent DNA ligase that is responsible for replication and repair. Inhibitors of the NAD+ -dependent enzymes are known for example from a paper entitled Specific and Potent Inhibition of NAD+-dependent DNA Ligase by Pyridochromanones to Brotz- Oesterhelt et al published in 2003 and US 2007/0082920 A1 relating to pyrido[2,3- lpyrimidines NAD+-dependent DNA ligase inhibitors, published in 2007.

WO2006/016146 describes crystals of DNA ligase A (LigA) and computer-assisted methods for screening, identifying, and designing inhibitors and allosteric modulators of LigA. US-A-5705344 relates to a high-throughput screening assay for inhibitors of nucleic acid helicases

Anti-bacterials are being developed based on the premise that a single NAD+ inhibitor will be selectively toxic for all eubacteria and effective as humans do not possess an NAD+ -dependent ligase. Such anti-bacterials would however also be toxic for desirable bacteria.

Enzymes found in eukaryotes, archaebacteria and bacterial viruses (bacteriophage) derive energy from ATP (ATP-dependent enzymes) and these enzymes all have a common catalytic core. Eukaryotic organisms have a variable complement of ATP- dependent enzymes. An orthologue of DNA ligase I occurs in all eukaryotes and is essential for DNA replication and base excision repair. Eukaryotic ATP-dependent DNA ligases vary greatly in size between about 34 and about 103 kDa. The catalytic, core structure incorporates six conserved amino acid motifs. Paramecium bursaria Chlorella virus DNA ligase ( PBCV-1 ligase) is the smallest characterised eukaryotic DNA ligase (298 amino acids). It has a two domain structure, consisting of an adenylation domain (AdD) and an oligomer binding (OB) fold domain. The crystal structure of the multi- domain Human DNA ligase I (HuLJgI) and PBCV-1 ligase have been solved in complex with nicked DNA substrates. These enzymes encircle the duplex DNA at the site of a nick; HuLigl stabilises the DNA with contacts from the AdD, OB domain and a dedicated DNA binding domain (DBD) (Pascal et al 2004, see reference below). PBCV-1 ligase stabilises the DNA utilising AdD and OB contacts as well as a lysine-rich surface loop or latch module located distal to ligase motif V (Nair et al 2007 - see reference below)..

Chen et al (2008) (see reference below) have shown that differential inhibition, of the three human DNA ligases can be achieved by molecules that interfere with the binding step of the individual ligases. Furthermore, the DNA binding step of smaller ATP- dependent ligases, exemplified by the PBCV-1 DNA ligase and commonly found in certain members of the eubacteria, differs from that of the larger human DNA ligases (Pascal et al 2004, Nair et al 2007 - see references below). Molecules that prevent DNA binding by eubacterial ATP-dependent DNA ligases may have utility in combination with NAD-dependent ligase inhibitors that are currently being developed as broad spectrum antibiotics.

Certain bacteria, particularly pathogenic organisms, have two DNA ligases, a single NAD+ and a single ATP-dependent, respectively. Mycobacteria have a number of ATP- dependent enzymes. ATP-dependent polynucleotide ligases are viewed as a single class because of their common catalytic core. Inhibitors to ATP-dependent enzymes for use as antibacterials in mammals, for example humans are not known. There are no commercially available antibacterial ligase inhibitors.

The drawback of employing alone an inhibitor or modulator of the NAD+ -dependent ligases which may attack other beneficial bacteria may be avoided by employing a modulator or inhibitor for a selective ATP-dependent enzyme preferably in combination with an NAD+ dependent ligase. However, inhibitors of ATP-dependent ligases may be harmful insofar as they may inhibit the activity of human ligases.

We have now found that notwithstanding the common core in human and non-human ligases, the ATP-dependent enzymes which do not have a dedicated DNA binding domain ("small" ATP-dependent enzymes) however bind to polynucleotides differently to their human ATP-dependent counterpart enzymes. The chemical steps in ligation, for example adenylation and hydroxy I attack to release the adenylate on sealing the polynucleotides are not believed to be distinct as between human and "small" ATP- dependent enzymes but as the physical or binding step distinguishes between human and "small" enzymes, the invention allows discernment between human and bacterial ATP-dependent enzymes.

The invention provides an assay to determine inhibitors or modulators of ligase activity. By the inclusion of human DNA ligase enzymes in this assay as well as target (and or representative) ligases we will be able to identify inhibitors of ATP-dependent non- mammalian ligase, preferably bacterial ligases which should not be cytotoxic for mammalian cells, at least as regards their ability to inhibit DNA ligase enzymes.

The invention provides an assay, a method and a composition for detecting ligase activity, particularly activity of non-human ATP-dependent ligase and identifying modulators of pathogenic polynucleotide ligase activity.

The invention includes methods of screening for ligase activity by comparing the ligase activity of a candidate modulator with the ligase activity of one or more defined control ligase modulators. The invention also provides a method of screening for a modulator of one or more defined ligase activities by comparing ligase activity in the presence and absence of a candidate modulator. Furthermore the screening method permits interrogation of the individual steps of the DNA ligation reaction. This in turn permits identification of differential modulators that may be employed separately or in concert to act selectively on the individual steps of DNA ligation.

The invention provides a method of identifying an agent that inhibits or modulates the activity of a ligase, particularly an ATP-dependent ligase and human or non-human ATP- dependent ligase, the method comprising: a) providing a polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group, optionally wherein one of these polynucleotides has a labeled group that permits detection and wherein said first and second polynucleotides are selected from DNA and RNA and preferably are both DNA or are both RNA, and a complementary polynucleotide template to both the first and second polynucleotides such that ligation between the said 3' OH and said 5' phosphate groups is possible, the complementary polynucleotide being bound to a surface for detection using surface plasmon resonance, for example a sensor chip comprising a gold surface; b) contacting the said polynucleotides with a ligase and a candidate agent, preferably by passing the ligase and candidate agent over the polynucleotides, the contact being carried out under conditions to promote binding and/or ligation of the first and second polynucleotides; c) detecting any change in the concentration of ligase on the surface preferably using surface plasmon resonance.

The complementary polynucleotide and either the first polynucleotide or second polynucleotide may be part of the same strand and comprise a DNA hairpin such that the complementary templating strand folds back on itself to provide either the first or second polynucleotide.

Preferably, the complementary polynucleotide is bound to the surface by a complex of biotin and avidin or streptavidin.

Surface plasmon resonance is especially advantageous for detection as very small changes on the concentration of ligase on the surface may be detected, for example a change of 10 to 800pg/mm² and preferably less than 10 pg/mm² and desirably less than 1 pg/mm². This assay provides a highly sensitive means to investigate the inhibition or modulation of different ligase elements.

Surface plasmon resonance allows identification of candidate agents that act to inhibit or modulate ligase activity through the binding or lack of binding of the enzyme and allows discrimination from candidate agents that act by inhibiting or modulating the catalytic steps of the enzyme during ligation. Suitably, the candidate agent acts to inhibit transfer of adenosine monophosphate (AMP) to the 5' phosphate group. The inventors have found that this may be determined by attempting to react the polynucleotide having the 5' phosphate group suspected of being adenylated in ligation with a species reactable with an adenylated polynucleotide.

The invention provides a method of determining whether a candidate agent for modulation or inhibition of the activity of an ATP dependent ligase acts to modulate or inhibit such comprising: a) providing a first polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group wherein said first and second polynucleotides are selected from DNA and RNA and preferably are both DNA or are both RNA and suitably a complementary polynucleotide template to both the first and second polynucleotides such that ligation between the said 3' OH and said 5' phosphate groups is possible, the complementary polynucleotide being bound to a surface for detection using surface plasmon resonance, for example a sensor chip comprising a gold surface; b) contacting the said polynucleotides with a ligase and a candidate agent, preferably by passing the ligase and candidate agent over the polynucleotides, the contact being carried out under conditions to promote binding and/or ligation of the first and second polynucleotides; c) contacting the mixture of step b) with a species capable of binding to the first polynucleotide when adenylated under conditions to promote reaction of the species with any adenylated first polynucleotide present; and d) detecting the presence or absence of a reaction product of the species and adenylated first polynucleotide suitably using surface plasmon resonance.

Preferably the species for reaction with an adenylated polynucleotide comprises a protein, more preferably aprataxin.

Aprataxin is a novel human protein that binds to adenylated polynucleotide and the invention provides for the use of aprataxin in a method for determining whether a candidate agent for modulation or inhibition of DNA ligation acts to modulate or inhibit sealing of DNA after adenylate transfer, thus an inhibitor of the final ligase catalysed step of DNA ligation The ligase assay may also be employed to provide adenylated polynucleotide as a substrate for aprataxin and use the reaction product of aprataxin and adenylated polynucleotide in an assay for a modulator or inhibitor for aprataxin.

Detection of the presence or absence of a reaction product of the species and adenylated first polynucleotide may be carried out using surface Plasmon resonance or by labeling the 5' phosphate polynucleotide.

In another aspect the invention provides a method of identifying an agent that inhibits or modulates the activity of a ligase, particularly an ATP-dependent ligase, the method comprising: a) providing a first polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group, wherein one of these polynucleotides is labeled, a ligase, and a candidate agent, wherein said first and second polynucleotides are selected from DNA and RNA and for example 5'RNA to 3'DNA and 5'DNA to 3'RNA, and a complementary polynucleotide acting as a template for both the first and second polynucleotides such that ligation between the 3' OH and 5' phosphate groups is possible, the complementary polynucleotide being bound to a surface; b) subjecting the mixture to conditions to promote ligation of the first and second polynucleotides to form a third polynucleotide which is distinguishable from the first or second labeled polynucleotide, for example by binding more strongly to an affinity matrix c) subjecting the mixture of b) to a procedure, preferably surface plasmon resonance, to determine whether the labeled third polynucleotide is present and to determine the presence of the first or second labeled polynucleotide .

Upon identification of a suitable agent to inhibit or modulate ligase, the agent may be employed in the method of the invention.

The absence of the labeled third polynucleotide indicates the candidate agent has acted to prevent ligation. A mixture of the labeled third polynucleotide and first or second labeled polynucleotide indicates the agent has acted to partly inhibit ligation. Measuring the relative amounts of the third polynucleotide and the first or second polynucleotide provides a means of quantitatively assessing the ligase inhibiting effect of the candidate agent.

Suitably, the mixture of step b) is contacted with an affinity matrix and the procedure in step c) to determine the presence of the polynucleotide is surface plasmon resonance. DNA ligation typically involves the ligase binding to the DNA, transferring adenylate to the polynucleotide and polynucleotide sealing. Furthermore an agent that affects the third step of the ligation reaction can be distinguished. In this situation the adenylate moiety may be transferred to the 5' phosphate terminated strand however the ligase- catalysed attack of the 3'OH on this adenylated DNA intermediate may be affected. The extent of adenylated nick structure can be revealed by the subsequent incubation with the human protein aprataxin. Aprataxin is capable of reversing this stalled ligation intermediate. The inventor has shown that the creation of an adenylated DNA strand by DNA ligase creates a substrate for aprataxin binding. The assay thus provides a means of identifying inhibitors of three separate events during DNA ligation, namely, DNA binding, transfer of adenylate to the polynucleotide and polynucleotide sealing.

In a preferred embodiment, the invention provides a method of identifying an agent that modulates the activity of a ligase, particularly an ATP-dependent ligase and especially a non-human ATP-dependent ligase, the method comprising: a) providing a first labeled polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group, a ligase, and a candidate agent, wherein said first and second polynucleotides are selected from DNA and RNA, for example 5'RNA to 3'DNA, 5'DNA to 3'RNA, both DNA and both RNA and a complementary polynucleotide linked to both the first and second polynucleotides such that ligation between the 3' OH and 5' phosphate groups is possible, the complementary polynucleotide optionally being bound to a surface; b) subjecting the mixture to conditions to promote ligation of the first and second polynucleotides to form a labeled third polynucleotide which is distinguishable from the first labeled polynucleotide, for example by binding more strongly to an affinity matrix c) subjecting the mixture of b) to a procedure to determine whether the first labeled polynucleotide or the labeled third polynucleotide is present said procedure comprising contacting the said mixture with a solid substrate wherein said labeled third polynucleotide has a greater affinity for said solid substrate than does the first polynucleotide; d) separating any first labeled first polynucleotides to obtain isolated, labeled third polynucleotide; e) measuring the amount of said label retained on the labeled third polynucleotide, wherein the amount of said label is indicative of the degree of modulation of ligase activity by said agent. Suitably, the solid substrate comprises a micro-bead and said greater affinity results from the third polynucleotide being of greater length than said first polynucleotide.

The complementary polynucleotide may comprise a ligand which is bindable to the solid substrate or a component of the solid substrate. Preferably the ligand is biotin and the solid substrate comprises avidin or streptavidin. The ligand may be avidin or streptavidin and the substrate comprises biotin.

The label on the first polynucleotide is suitably a radiolabel, for example radioactive phosphate in the 5' phosphate group. The label may be a known label employed in assays, for example a fluorescent label or biotin. Where biotin is employed as the label, the ligand is something other than biotin.

Steps a) and b) of the assay method may be carried out in conventional apparatus for example a well plate. Multiple procedures may be carried out concurrently to provide a way of screening large numbers of candidate agents simultaneously. The mixture may then be transferred to a separate apparatus and contacted with a polynucleotide binding affinity matrix for example a membrane, such that the first, second and third polynucleotide will bind to the matrix. Suitably liquid components of the mixture are then removed, for example by suction. The bound polynucleotide is then suitably treated so as to differentiate between polynucleotides based on how strongly they are bound by the matrix, for example by contacting with a weak salt solution such that unligated polynucleotide is removed from the matrix. In a preferred embodiment, the matrix is removed and set aside with the bound complementary polynucleotide. The remaining mixture is then suitably contacted with a further affinity matrix so as to bind all the polynucleotide. By measuring the level of bound polynucleotide on the two substrates, for example by radioactive counting, the degree of ligation may be determined quantitatively.

On identification of candidate agents which show ligase inhibition activity, particularly for non-human ATP-dependent ligase, the candidate agent is suitably subjected to testing, for example by a DNA binding assay which may employ the surface plasmon resonance assay as described herein or a shifting assay carried out in a gel matrix during electrophoresis, to determine whether the ligase has been inhibited or lack of ligation is due to prevention of normal ligase DNA interaction. Preferably, the separating step by which bound polynucleotide is separated from the rest of the mixture is performed in a tube having a fluid passage comprising a reservoir portion and a evacuation portion a filter extending transversely across said passage and separating said reservoir portion from said evacuation portion.

The first and second polynucleotides may be of any sequence which provides a convenient substrate for the targeted ligase(s). The polynucleotides may be complementary over the entire length of at least one of the polynucleotides or there may be regions of non-complementarity 5' and/or 3' of the complementary region. Generally, synthesized oligonucleotides or conveniently replicated vectors for example phage, or restriction fragments thereof, provide an inexpensive source of the polynucleotides. The assays are generally compatible with the presence of DNA binding proteins.

The first polynucleotide comprises a detectable label, which label is absent from the complementary polynucleotide. A wide variety of directly and/or indirectly detectable labels may be used so long as they are compatible with the assay. Exemplary directly detectable labels include radiolabels, fluorescent labels, and the like. Exemplary indirectly detectable labels include epitope tags, biotin, nucleoside analogs such as digoxigenin and the like.

The pathogenic ligase that is any ligase activity that is harmful or acting harmfully to the host cell or organism is selected based on the target application. Pathogen-selective or - specific ligases may be used to identify pharmacological therapeutics for the treatment of infectious disease. Fungal, viral, bacterial and parasitic ligases, in particular, provide medically urgent targets for identifying inhibitors by the subject methods.

Organisms for which the assay is particularly desirable in seeking inhibitors for pathogenic ligase include Neisseria miningiditis (meningitis), Haemophilus influenzae (pneumonia/meningitis), Vibrio Cholera (cholera), Burkholderia pseudomallei (melioidosis/septicaemia), Trichomonas vaginalis (Trichomoniasis) and Campylobacter jejuni (diarrhoeal disease).

The ligase may be purified from a natural source or may be recombinant and suitably is usually provided in at least a partially-purified form. Only a portion of the native ligase need be used in the assay, the portion being sufficient for ligase activity, preferably not less than an order of magnitude less than that of the full-length ligase. Portions capable of imparting the requisite binding specificity and affinity are readily identified by those skilled in the art. A wide variety of molecular and biochemical methods are available for generating catalytic portions, for example Current Protocols in Molecular Biology (Eds. Aufubel, Brent, Kingston, More, Feidman, Smith and Stuhl, Greene Publ. Assoc, Wiley- Interscience, NY, N. Y., 1992) or that are otherwise known in the art.

The reaction mixture also comprises a candidate agent such as a preselected candidate ligase inhibitor or, especially for high-throughput drug screening, a library-derived candidate agent. Library-derived candidate agents encompass numerous chemical classes, though typically they are organic compounds; preferably small organic compounds. The libraries may comprise synthetic and/or naturally derived compounds. For example, numerous means are available for random and directed synthesis of a wide variety of organic compounds and biomolecules, including expression of randomized oligonucleotides. Alternatively, libraries of natural compounds in the form of bacterial, fungal, plant and animal extracts are available or readily produced. Additionally, natural and synthetically produced libraries and compounds are readily modified through conventional chemical, physical, and biochemical means. In addition, known pharmacological agents may be subject to directed or random chemical modifications, such as acylation, alkylation, esterification, amidification, etc., to produce structural analogs. The agent is provided in standard serial dilutions or in an amount determined by analogy to known modulators.

In addition, the mixture may comprise additional reagents, such as salts, buffers, and the like to facilitate or maximize ligase activity. Also, reagents that reduce non-specific or background denaturation or otherwise improve the efficiency of the assay, such as protease inhibitors, nuclease inhibitors, antimicrobial agents, single-stranded DNA binding protein, or the like may be used.

The complementary polynucleotide is preferentially immobilized on a solid substrate, that is the complementary polynucleotide has a higher affinity for the substrate than does the first, labeled polynucleotide. Preferred solid substrates maximize the disparity in binding affinities and the complementary polynucleotide binding affinity and binding sites, and hence, maximize signal strength and the signal-to-noise ratio. The affinity for the complementary polynucleotide may be direct (substrate-polynucleotide), indirect via a ligand (substrate-ligand-polynucleotide) or ligand receptor complex (substrate-receptor- ligand-polynucleotide). As examples: silica-based bead substrates may be used to size- select the polynucleotides directly; magnetized substrates may be used to distinguish second polynucleotide comprising an iron-based ligand; substrates comprising a surface bound antibody receptor may be used to distinguish second polynucleotide comprising a specific ligand antigen of the antibody receptor. To avoid interference, any selected ligand should not be identical to any selected label or label component.

The methods of the invention are particularly suited to automated high throughput drag screening. In a preferred embodiment, the individual sample incubation volumes are less than about 500 microlitre, preferably less than about 250 microlitre more preferably less than about 100 microlitre. Such small sample volumes minimize the use of often scarce candidate agent, expensive enzymes, and hazardous radioactive waste. Furthermore, the methods provide for automation, especially computerized automation. Accordingly, the method steps are preferably performed by a computer-controlled electromechanical robot.

The invention provides kits for ligase modulator or inhibitor screening which include premeasured amounts of the compositions used in the methods of the invention.

Since ligases are necessary for joining discontinuous polynucleotides, target diseases suitable for treatment with ligase inhibitors or modulators are limited only in that disease or disease progression be subject to inhibition by modulation of the activity of one or more specific ligase. As such, target diseases include viral, bacterial and fungal infections, metabolic disease, genetic disease, cell growth and regulatory disfunction, such as neoplasia, inflammation, hypersensitivity, etc. The target diseases may be afflictions of plants, especially agricultural crops, or animals, especially livestock, house animals and humans.

The invention also provides a screening method for a candidate inhibitor comprising contacting a pathogenic bacterial infection for example Campylobacter jejunii or Burkholderia bacteria with a candidate inhibitor of ATP-dependent ligase and measuring the effect on pathogenicity of the bacteria.

The invention also provides for agents for inhibiting ATP-dependent ligase for use in the treatment of a condition derived from bacteria having an ATP-dependent ligase and compositions containing such agents, especially where the agent has been identified using a method of the invention.

We have further found that ligase inhibitors or modulators are suitably employed in combination with NAD+ inhibitors. An organism treated with a NAD+ inhibitor may survive treatment if the organism has ATP-dependent ligase and survival may enrich the remaining enzymes to resistance to the NAD+ -dependent ligase inhibitor. Treatment in using a combination of ATP-dependent ligase inhibitor and NAD+ -dependent ligase inhibitor reduces this risk.

Without wishing to be bound by theory, it is believed some of the more dangerous pathogens that are likely to be inherently resistant to an NAD+ -inhibitor, include Neisseria miningiditis (meningitis), Haemophilus influenzae (pneumonia/meningitis), Vibrio Cholera (cholera), Burkholderia pseudomallei (melioidosis/septicaemia) and Campylobacter jejuni (diarrhoeal disease). The combination treatment will also provide a means of treating those organisms that survives an initial anti-NAD+ treatment. There are a number of bacterial species that have ATP-dependent ligases and these may be a source of natural resistance to NAD inhibitors. As such even if is possible to eradicate a target organism with an NAD+ inhibitor alone it may be preferential to treat also with a combination of NAD+ /ATP inhibitor mixture to reduce the risk and ideally preclude the arising of and transfer of resistance to NAD+ inhibitors as a result of gene transfer from naturally resistant organisms. The invention is illustrated by the following non-limiting examples.

Examples

Immobilisation of DNA and optimisation of Biacore reaction conditions

All surface plasmon resonance (SPR) analysis was performed on a Biacore 2000 automated system (Biacore AB; Uppsala, Sweden). Prior to substrate immobilisation the Biacore system was desorbed with BIAdesorb solutions 1 (0.5 % w/v SDS in distilled water) and 2 (50 mM glycine-NaOH, pH 9.5; Biacore AB; Uppsala, Sweden) and sanitised with BIAdisinfectant solution (1.0 % sodium hyperchlorite v/v) in distilled water; Biacore AB; Uppsala, Sweden). The system was then primed with degassed and filtered (0.22 Dm filter) Biacore buffer 1 : comprising 50 mM Tris-HCI pH 7.5, 50 mM NaCI, 5.0 mM DTT, 5.0 mM EDTA pH 7.5 and 0.005 % Surfactant p20 (Biacore AB; Uppsala, Sweden). The SA sensor chip was docked and the priming protocol was repeated, the chip was then treated with a minimum of three, one minute pulses of 1.0 M NaCI in 50 mM NaOH at 20 μl/min, to pre-condition the surface and normalised with BIAnormalising solution (70 % glycerol w/w in distilled water; Biacore AB; Uppsala, Sweden). All experiments were performed at 2O⁰C. HPLC-purified and desalted DNA substrates were obtained from Sigma-Aldrich (Poole) and used without any further purification. The biotinylated 21-mer oligonucleotide that forms the bottom strand of all DNA substrates analysed was annealed to various oligonucleotides (details given in Figure 1) to form three different double stranded DNA substrates (see Figure 2).

Figure 1 shows sequence information of DNA substrates immobilised for Biacore analysis

¹The molecular weight of each oligonucleotide was calculated from the nucleotide ratio where A = 15.4; T = 8.8; C = 7.3 and G = 11.7

The DNA substrates were heated at 65⁰C for five minutes in annealing buffer (50 mM Tris pH 7.5, 200 mM NaCI₁ 5.0 mM DTT, pH 7.5) containing 5.0 mM EDTA, before being mixed as shown in figure 1 and then slowly cooled to room temperature and incubated on ice. The biotinylated double-strand DNA substrates were diluted appropriately in Biacore buffer 1 (50 mM Tris-HCI, pH 7.5, 50 mM NaCI, 5.0 mM DTT, 5.0 mM EDTA and 0.005 % Surfactant p20) and immobilised on one of four flow cells of a streptavidin (SA)- coated sensor chip (Biacore AB; Uppsala, Sweden) in a Biacore 2000 (Biacore AB; Uppsala, Sweden). Immobilised DNA surfaces were equilibrated with 20 μl/min Biacore buffer 1. Approximately 200 response units (RU) of each DNA substrate was immobilised by injecting the appropriate oligonucleotide duplex (see Table 1 and Figure 2). The first flow cell was left underivatised to compensate for any effects of refractive index changes with buffer, instrument drift and non-specific binding of the ligase. Figure 2 shows the sequence of the double stranded DNA substrates used to analyse ligase binding and their attachment to a streptavidin-coated SA sensor chip Biotinylated dsDNA substrates were immobilised onto a streptavidin-coated SA chip. Panel A shows a nicked DNA substrate (Nick; top strand composed of the 10mer + phosphorylated 11 mer), panel B shows the double-stranded duplex DNA substrate (Duplex; top strand 21mer) and panel C shows the substrate with a single nucleotide gap (Gap; top strand composed of the 10mer + phosphorylated 10mer). The biotin group is indicated by the red circle attached to the 5'terminus of the bottom strand, the streptavidin groups bound to the surface of the chip are indicated by the green lines.

Following the immobilisation of approximately 200 RU of each substrate a buffer wash was carried out and the baseline was monitored, to ensure substrate stability. Three DNA substrates were utilised: dsDNA with a centrally placed nick (Nick), dsDNA with a one nucleotide gap (Gap) un-nicked duplex DNA (Duplex).

These substrates have been used previously to analyse PBCV-1 ligase DNA binding via gel shift analysis (Odell and Shuman, 1999). The nick represents the productive substrate for DNA ligase binding, whereas the duplex represents the product of a ligation event. The gap is a duplex DNA substrate with 5'PO₄ and 3'OH terminated ends, however sealing across a gap would create a frameshift mutation, a highly deleterious event.

DNA ligase enzymes were injected over the different combinations of the three DNA substrates described and the binding monitored. The flow rate of ligase and buffer was maintained at 30 Dl/min, to limit any effects of mass transport that occurs when the rate at which the analyte binds to the ligand is greater than the rate at which it is delivered to the surface of the chip. Prior to passing over multiple dilutions of each protein, a regeneration buffer (0.7 M NaCI for PBCV-1 ligase) was applied to remove bound ligase enzyme and regenerate the surface of the chip. Following the regeneration protocol buffer was flowed over the chip for three minutes to ensure the baseline was stable.

Expression and purification of proteins for DNA binding analysis

The bacterial plasmid derived from pET28a encoding Human DNA Ligase I (residues 232 - 919, hereafter known as: HuLigl-D232) was kindly provided by Professor Tom Ellenberger (Washington University School of Medicine, Washington, USA).

All PBCV-1 DNA ligase proteins were expressed in E. coli BL21 (DE3) cells grown at 37⁰C in LB broth. When the culture absorbance A_595nm reached 0.4 - 0.5, cells were incubated on ice for 30 minutes, adjusted to 0.4 mM IPTG and incubated, with shaking, at 20⁰C for 20 hours. The cells were harvested and soluble protein extracted by sequential treatment with lysozyme (1 mg/ml final) and triton X-100 (0.1% v/v final). Soluble proteins were applied to Ni-NTA agarose and eluted with imidazole. DNA ligase protein peaks were diluted to 50 mM NaCI with Buffer A (50 mM Tris-HCI pH 7.5 and 5.0 mM DTT), then applied to S-Sepharose resin and eluted using an NaCI gradient. Peak ligase activity fractions were pooled and assayed for ligation activity on Hind\\\ digested D DNA fragments.

Human DNA ligase, HuLigl-D232, was similarly expressed in E. coli BL21(DE3) cells and induced with IPTG. Following Ni-NTA agarose chromatography of the soluble proteins, the eluted human ligase peak fractions were combined, diluted to 25 mM NaCI with Buffer A (50 mM Tris-HCI pH 7.5 and 5.0 mM DTT) and applied to a Q-sepharose column pre-equilibrated with 50 mM Tris-HCI pH 7.5, 5.0 mM DTT and 25 mM NaCI. The column was washed with Buffer A containing 50 mM NaCI and step eluted with NaCI.

The purity of all the proteins was monitored by SDS-PAGE.

Results

Ligase substrate binding and discrimination revealed by SPR

To demonstrate that stable interaction of wild type (wt) PBCV-1 DNA ligase can be monitored by SPR a single cell with nicked DNA substrate bound (as shown in Figure 2) was incubated with the ligase under conditions described above. Nicked DNA represents the substrate in vivo that the enzyme would require to bind and seal as part of maintaining DNA integrity during replication, repair or recombination. Figure 3 shows the binding profile of the PBCV-1 ligase. On injection of the enzyme a deflection of the baseline indicates DNA binding of enzyme. The bound ligase remains apparently stably bound over a subsequent three minute wash period before the addition of 0.7 M NaCI to regenerate the substrate. It is taken that these data reflect the binding phase of the DNA ligase enzyme taking place after step 1 catalysis but before step 2 (the transfer of AMP

*to 5'phosphate moiety of the 5'phosphate terminated strand at the nick site. Step 2 cannot proceed because of the absence of the essential metal cofactor (magnesium) in the buffer system employed. Following buffer wash the nicked DNA was capable of re- binding the ligase (data not shown).

Figure 3 shows the binding of wt PBCV-1 DNA ligase to Nick substrate Wt PBCV-1 ligase (0.25 DM) was injected for 3 minutes over the derivatised SA chip.

Figure 4 shows SPR analysis of wt PBCV-1 DNA ligase binding to DNA reveals discrimination between productive and non-productive substrates.

Wt PBCV-1 ligase (0.5 DM) was injected for 3 minutes at a flow rate of 30 πl/min over the DNA substrates, (described above and immobilised to 200 RU), duplicate runs are shown. Biotinylated dsDNA substrates were immobilised onto a streptavid in-coated SA chip (200 RU). Nick (top strand: 10mer + phosphorylated 11 mer), Duplex (top strand

21mer) and Gap (top strand: 10mer + phosphorylated 10mer). The biotin group is shown as a red circle and streptavidin groups are indicated by the green lines. Binding to the Gap DNA substrate is represented by dark blue and magenta, Duplex DNA by yellow and cyan and the Nick by dark purple and brown. The values from the underivatised cell were subtracted and the baseline zeroed. The start and end of the ligase injection is indicated by black arrows.

The injection profile over the Nick substrate (top strand: 10mer + phosphorylated 11 mer - approximately 200 RU) is shown (black arrows indicate the start and end of the ligase injection). Following a 3 minute dissociation phase, the bound protein was removed from the substrate by a regeneration step (60 second injection of 0.7 M NaCI; start and end of the regeneration injection are highlighted by red arrows).

To demonstrate that the in vitro binding reflects true DNA binding by PBCV-1 DNA ligase two further experiments were performed. Firstly the ligase was shown to be capable of DNA substrate discrimination (Figure 4) and secondly the kinetics of nick sealing (ligase catalytic steps 2 and 3) were analysed (Figure 5). Binding of PBCV-1 DNA ligase to DNA was analysed over 3 substrates: double-stranded duplex DNA with a centrally placed nick (Nick), un-nicked duplex DNA (Duplex), and duplex DNA with a 1 nucleotide gap (Gap), Figure 4. Analysis of the DNA binding profile of wt PBCV-1 ligase reveals that the enzyme has a fast on-rate, evidenced by a rapid increase in RU following the injection of the protein for all three substrates (Figure 4). PBCV-1 ligase formed a stable complex that exhibits slow dissociation from a nicked DNA (a viable substrate) when compared with the Duplex substrate, the sealed reaction product (Figure 4). Twice as much PBCV- 1 protein however binds to the Duplex substrate suggesting more than one ligase molecule is able to bind, non-specifically. Binding to the Gap substrate where ligation is an unfavourable biological outcome was approximately half that shown on nicked DNA.

The enzyme rapidly reached binding equilibrium over the three substrates with an on- rate apparently indistinguishable between all three. However, the PBCV-1 ligase off-rate showed significant variation between each substrate (see Figure 4). On cessation of the ligase injection in the liquid phase it rapidly dissociates from the Gap substrate, 75% dissociates in 20 seconds. The Duplex substrate (equivalent to the ligase reaction product) retains ligase for longer than gapped DNA but the binding profile is markedly different from the binding to the Nick where some 60% of the PBCV-1 ligase remains bound after 100 seconds. Observation of ligase chemistry by SPR To demonstrate that the ligase binding profiled by SPR is consistent with normal DNA binding of the enzyme a further experiment is shown (Figure 5). Figure 5 shows Ligation of nicked substrate by wt PBCV-1 DNA ligase following addition of magnesium.

Panel A shows the repair of the nicked DNA substrate by wt PBCV-1 ligase. Wt PBCV-1 ligase (0.5 DM) was injected over the immobilised nicked substrate (300 RU); the end of the injection is indicated by an arrow. The chip was washed for 90 seconds with Biacore buffer 1 without 5.0 mM EDTA (start of the injection is highlighted by an arrow), followed by a 90 second wash with Biacore buffer 1 without EDTA, containing 10 mM MgCI₂ (the start of the injection is highlighted by an arrow). Wt ligase (0.25 DM) was reapplied to the chip following the magnesium wash (panel B). The values from the underivatised cell were subtracted and the baseline zeroed.

Following PBCV-1 DNA ligase binding to a nicked DNA substrate in the absence of the essential, divalent metal cofactor the previously seen stable interaction phase can be observed (Figure 5, Panel A). When this complex is exposed to buffer containing 10 mM magnesium chloride the ligase rapidly dissociates (Figure 5, Panel A). We take this to be evidence of steps 2 and 3 of ligase chemistry taking place now that metal is supplied and the ligase then dissociating from the duplex product. Reapplication of PBCV-1 DNA ligase to the washed chip surface shows a ligase binding and dissociation profile consistent with binding to intact duplex (as shown in Figure 4) resulting from the sealing of the nicked DNA (Figure 5, Panel B).

This experiment shows that the binding profile of PBCV-1 ligase when asssessed by SPR represents the true association of the enzyme with nicked DNA substrate and that this technique can be employed to follow both DNA binding and ligase reaction chemistry.

SPR DNA binding data is consistent with crystallographic analysis of DNA binding by PBCV-1 DNA ligase PBCV-1 DNA ligase is known to be dependent on a short 20 amino acid 'latch' module for stable DNA interaction (Nair et al 2007). To further confirm that the SPR system provides a technique for analysing DNA interaction by ligases we deleted this latch region, replacing it with a much shorter 5 amino loop to preserve the overall ligase structure.

The over-expressed Dlatch mutant migrated as a smaller species than the wt PBCV-1 ligase, estimated by SDS-PAGE analysis. The deletion of the latch module did not affect the solubility of the enzyme, as the protein was successfully over-expressed and purified, data not shown. The PBCV-1 ligase Dlatch mutant was diluted and injected over an SA chip derivatised with the three DNA substrates, Nick, Duplex and Gap.

Figure 6 shows SPR analysis of the PBCV-1 ligase πlatch binding to DNA

The binding of the PBCV-1 DNA ligase Δlatch mutant (0.5 DM) to the DNA substrates was assessed by duplicate 3 minute injections (30 Dl/min flow rate). Binding to the Gap DNA substrate is represented by dark blue and magenta, Duplex DNA by yellow and cyan and the Nick by dark purple and brown.

In panel A, the major units on the y-axis have been adjusted to provide a direct comparison to the binding profile of wt PBCV-1 DNA ligase (refer to Figure 3 for reference).

Panel B shows the binding profile of the Δlatch mutant diluted to 50 mM NaCI and panel C shows the profile of Δlatch mutant diluted to 150 mM NaCI. The start and end of each injection is indicated by black arrows. The values from the underivatised cell were subtracted and the baseline zeroed.

Deletion of the latch module resulted in a ten-fold decrease in DNA binding over all three substrates (Figure 6, panel B), compared to that of the wt PBCV-1 ligase (Figure 4). The on-rate over all three substrates was rapid and similar to that observed for the wild type enzyme, but the amount of stable binding was significantly reduced. The highest level of binding was observed over the duplex substrate, giving an equilibrium response of approximately 32 RU. At the end of the injection the enzyme rapidly dissociated and the baseline was restored. Enzyme binding to the nick quickly reached equilibrium at approximately 17 RU and, as with the duplex, dissociation from the substrate occurred rapidly at the end of the injection. Interestingly, these data show that despite the reduction in stable DNA binding, the PBCV-1 ligase Dlatch mutant still shows discrimination of the nick compared to duplex DNA (Figure 6, panels B and C). This finding has not been determined by crystallography as the affinity of the Dlatch for DNA precludes the stable formation of enzyme: DNA co-crystals. This analysis suggests that in the absence of the latch, other residues in the nucleotide binding domain and OB-fold provide sufficient substrate discriminatory moieties although a stable ligase-DNA complex is not formed. Application of SPR to analyse DNA binding by cellular lipases

The DNA binding profile of HuLigl-D232 was analysed on the three substrates, Nick, Duplex and Gap. Ellenberger and co-workers have shown that this form of the human DNA ligase I can be stably co-crystallised with a synthetic DNA nick substrate (Pascal et a/ 2004).

Figure 7 shows SPR analysis of HuLigl-D232 binding to DNA.

The binding of HuLigl-D232 (0.5 DM) was assessed by duplicate 3 minute injections (30 αl/min flow rate) over the immobilised substrates. The enzyme was diluted to approximately 50 mM NaCI.

In panel A, the major units on the y-axis have been adjusted to provide a direct comparison to the binding profile of the PBCV-1 DNA ligase (refer to Figure 3 for reference). Binding to the Gap DNA substrate is represented by dark blue and magenta, Duplex DNA by yellow and cyan and the Nick by dark purple and brown. The start and end of the injection are indicated by black arrows. The response from the underivatised cell were subtracted and the baseline zeroed.

The binding profile for the truncated human DNA ligase to the nicked substrate (Figure 7, panel A) is significantly different to that observed for PBCV-1 ligase (Figure 4). At 75 kDa, HuLigl-D232 is twice the size of the PBCV-1 ligase, but shows half the amount of RU binding to the Nick substrate. At 0.5 DM it showed a binding of approximately 100 RU, although equilibrium was not reached (Figure 7, panel A) compared to the 250 RU observed for the PBCV-1 enzyme (Figure 4). Interestingly the on-rate was approximately two orders of magnitude slower than the PBCV-1 ligase. Unlike the PBCV-1 ligase, the truncated human enzyme did not, at any enzyme concentration, reach equilibrium over the nick; instead, the RU continued to increase gradually over the three minute injection. At the end of the injection, there was a gradual dissociation of unbound protein, with approximately 50 RU of HuLigl-Δ232 remaining bound after a three minute dissociation period. HuLigl-Δ232 also showed a different substrate discrimination to the PBCV-1 ligase, having greater affinity for the Nicked DNA over the Duplex. Binding of human DNA ligase I to DNA is facilitated in part by interaction with PCNA. It is not inconceivable that in the absence of such a factor that the human enzyme DNA binding is somewhat compromised.

References Chen, X., Zhong, S., Zhu, X., Dziegielewska, B., Ellenberger, T., Wilson, G. M., MacKerell, A.D. Jr and Tomkinson, A.E. (2008) Rational design of human DNA ligase inhibitors that target cellular DNA replication and repair. Cancer Res. 68:3169-3177

Nair, P.A., Nandakumar, J., Smith, P., Odell, M., Lima, CD. and ShumanS. (2007) Structural basis for nick recognition by a minimal pluripotent DNA ligase. Nature Structural & Molecular Biology 14, 770-778.

Odell, M. and Shuman, S. (1999) Footprinting of Chlorella virus DNA ligase bound at a nick in duplex DNA. J. Biol. Chem, 274, 14032-14039.

Pascal, J. M., O'Brien, P.J., Tomkinson, A.E., and Ellenberger, T. (2004). Human DNA ligase completely encircles and partially unwinds nicked DNA. Nature 432, 473-478.

Claims

1. A method of identifying an agent that inhibits or modulates the activity of an ATP dependent ligase, the method comprising: a) providing a polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group, and wherein said first and second polynucleotides are selected from DNA and RNA and a complementary polynucleotide template to both the first and second polynucleotides such that ligation between the said 3' OH and said 5' phosphate groups is possible, the complementary polynucleotide being bound to a surface for detection using surface plasmon resonance; b) contacting the said polynucleotides with a ligase and a candidate agent, the contact being carried out under conditions to promote binding and/or ligation of the first and second polynucleotides; and c) detecting any change in the concentration of ligase on the surface.

2. A method according to claim 1 wherein the ligase is a non-human or human ATP- dependent ligase.

3. A method according to claim 1 or claim 2 wherein the first and second polynucleotide are both DNA or are both RNA.

4. A method according to any one of the preceding claims wherein the surface for detection comprises a sensor chip comprising a gold surface.

5. A method according to any one of the preceding claims wherein the complementary polynucleotide and either the first polynucleotide or second polynucleotide are part of the same strand and comprise a DNA hairpin such that the complementary templating strand folds back on itself to provide either the first or second polynucleotide.

6. A method according to any one of the preceding claims wherein the complementary polynucleotide is bound to the surface for detection by a complex of biotin and avidin or streptavidin.

7. A method according to any one of the preceding claims wherein the detection step has a sensitivity to detect a change of less than 10 pg/mm² in the ligase present.

8. A method of identifying an agent that inhibits or modulates the activity of a ligase comprising: a) providing a first polynucleotide having a 5' phosphate group and a second polynucleotide having a 3" OH group, wherein one of these polynucleotides is labeled, a ligase, and a candidate agent, wherein said first and second polynucleotides are selected from DNA and RNA and a complementary polynucleotide acting as a template for both the first and second polynucleotides such that ligation between the 3' OH and 5' phosphate groups is possible, the complementary polynucleotide being bound to a surface; b) subjecting the mixture to conditions to promote ligation of the first and second polynucleotides to form a third polynucleotide which is distinguishable from the first or second labeled polynucleotide; and c) subjecting the mixture of b) to a procedure to determine whether the labeled third polynucleotide is present and to determine the presence of the first or second labeled polynucleotide .

9 A method according to claim 8 wherein the third polynucleotide is distinguishable from the first or second polynucleotide on the basis of different binding strength with an affinity matrix.

10. A method according to claim 8 or claim 9 wherein the mixture of step b) is contacted with an affinity matrix and the procedure in step c) to determine the presence of the polynucleotide is surface plasmon resonance.

11. A screening method for a candidate inhibitor or modulator of ATP-dependent ligase comprising identifying an inhibitor or modulator using a method according to any one of the preceding claims and contacting a pathogenic bacterial infection with a candidate inhibitor of ATP-dependent ligase and measuring the effect on pathogenicity of the bacteria.

12. An agent for inhibiting ligase for use in the treatment of a condition derived from bacteria having an ATP-dependent ligase and compositions containing such agents wherein the agent has been identified using a method according to any one of the preceding claims.

13. A composition for inhibiting or modulating both ATP-dependent ligase and NAD+ dependent ligase comprising a NAD+ dependent ligase inhibitor or modulator and an ATP-dependent ligase inhibitor or modulator.

14. A method of determining whether a candidate agent for modulation or inhibition of activity of an ATP-dependent ligase acts to modulate or inhibit such activity comprising: a) providing a first polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group and a complementary polynucleotide template to both the first and second polynucleotides such that ligation between the said 3' OH and said 5' phosphate groups is possible, the complementary polynucleotide being bound to a surface for detection using surface plasmon resonance;; b) contacting the said polynucleotides with a ligase and a candidate agent under conditions to promote binding and/or ligation of the first and second polynucleotides; c) contacting the mixture of step b) with a species capable of binding to the first polynucleotide when adenylated under conditions to promote reaction of the species with any adenylated first polynucleotide present; and d) detecting the presence or absence of a reaction product of the species and adenylated first polynucleotide using surface plasmon resonance.

15. A method according to claim 14 wherein the species for reaction with an adenylated polynucleotide comprises a protein.

16. A method according to claim 14 or claim 15 wherein the species comprises aprataxin.

17. Use of aprataxin in an assay for a modulator or inhibitor for aprataxin as defined in any one of claims 14 to 16.

18. Use according to claim 17 comprising: a. providing a first polynucleotide having a 5' phosphate group and a second polynucleotide having a 3' OH group wherein said first and second polynucleotides; b. contacting the said polynucleotides with an ATP-dependent ligase and a candidate agent to act as a modulator or inhibitor for aprataxin under conditions to promote adenylation of the first polynucleotide; and c. detecting the presence or absence of a reaction product comprising adenylated polynucleotide and aprataxin.