Effects of base mismatches on joining of short oligodeoxynucleotides by DNA ligases
Effects of base mismatches on joining of short oligodeoxynucleotides by DNA ligasesClare E. Pritchard* and Edwin M. Southern
Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, UK
Received June 16, 1997;Revised and Accepted July 15, 1997
ABSTRACT
The requirement for Watson-Crick base pairing surrounding a nick in duplex DNA to be sealed by DNA ligase is the basis for oligonucleotide ligation assays that distinguish single base mutations in DNA targets. Experiments in a model system demonstrate that the minimum length of oligonucleotide that can be joined differs for different ligases. Thermus thermophilus (Tth) DNA ligase is unable to join any oligonucleotide of length six or less, while T4 DNA ligase and T7 DNA ligase are both able to join hexamers. The rate of oligonucleotide ligation by Tth DNA ligase increases between heptamer and nonamer. Mismatches which cause the duplex to be shortened by fraying, at the end distal to the join, slow the ligation reaction. In the case of Tth DNA ligase, mismatches at the seventh and eighth position 5' to the nick completely inhibit the ligation of octamers. The results are relevant to mechanisms of ligation.
INTRODUCTION
DNA ligases are widely used with oligonucleotides in processes such as oligonucleotide ligation assays (OLA) (1 -4 ), the ligase chain reaction (LCR) (4 ) and to provide primers for DNA sequencing (5 ). The success of these methods depends on ligation of pairs of oligonucleotides which are fully complementary to adjacent regions on the target (template) DNA. The requirement of most DNA ligases for fully base-paired duplex near to the DNA junction has also been exploited to improve the performance of sequencing by hybridisation (6 ).
DNA ligases are found in all organisms; their major functions are to join the Okazaki fragments on the lagging DNA strand during chromosome replication, and to reseal nicks in DNA during repair by the excision/replacement pathway. DNA ligases produce a covalent phosphodiester bond between two nucleic acids, one carrying a 5' phosphate group and the other a 3' hydroxyl group. The ligation process requires Watson-Crick base pairing at the reaction site and it is on this property that in vitro assays depend (7 ,8 ).
In the case of OLA (4 ), where the substrates are typically 20mers, only the exquisite requirement for Watson-Crick base pairing in the sites immediately adjacent to the joining point has been exploited and this fidelity has been studied in detail for Tth ligase (4 ,9 ). We required a DNA ligase which would join short oligonucleotides ranging from hexamers to decamers with high specificity for perfect base pairing in the whole substrate oligonucleotide, for use in a new method of sequence determination under development in this laboratory. Previous studies suggested that the bacterially-derived DNA ligases are more sensitive to the nature of their oligonucleotide targets than viral or eukaryotic DNA ligases (4 ,9 ,10 ), but there have been few studies of the effects of mismatching, and little has been published about the minimum length of oligonucleotide that DNA ligases require as substrate.
Here we show that where the substrates are short oligonucleotides, ligase specificity for fully Watson-Crick base-paired duplex DNA can, in the case of Tth DNA ligase, extend as far as the ninth position 5' to the join and have a significant effect on the rate of ligation.
MATERIALS AND METHODS
Oligonucleotides were synthesised on an ABI 390B DNA synthesiser. Phosphoramidite monomers were from Cruachem. Thermus thermophilus DNA ligase (Tth) was from Advanced Biotechnologies Inc. (Mole Park, Surrey). T4 polynucleotide kinase was from Epicentre Technologies (Cambio, Cambridge, UK). T7 and Bacillus stearothermophilus DNA ligase were gifts from Dale Wigley and Steve Ashford. Acrylamide solution (40%, 19:1 acrylamide:bisacrylamide) from Severn Biotech (Bristol, UK) and Anachem (Luton, UK). All other chemical reagents were purchased from BDH/Merck.
Oligonucleotides were analysed by reversed phase HPLC (Waters 900 series), on C8 columns (Rainin Dynamax 500). Buffer was 0.1 M triethylammonium acetate (pH 7) and oligonucleotides were eluted using gradients of acetonitrile. Composition of the degenerate oligonucleotides (Fig. 4 ) was assessed by separation of the three components in each mixture on a gradient running from 7.2 to 9.6% acetonitrile over 30 min. All mismatch mixtures were equimolar for each component "20%. All oligonucleotides were >90% required product as estimated by integration of HPLC traces.
Phosphorylation of oligonucleotides
Typical phosphorylation reactions contained 20 pmol oligonucleotide, 33 mM Tris-acetate (pH 7.8), 66 mM potassium acetate, 10 mM magnesium acetate, 0.5 mM DTT and 50 mM [[gamma]-33P]- adenosine triphosphate (Amersham International and Dupont). Reactions were stopped and enzyme-killed by addition of 80 [mu]l of TE (10 mM Tris-HCl, 1 mM EDTA) and heating at 70oC for 5 min. Excess salts and ATP were removed using a Sephadex spun column (Pharmacia, Sweden).
DNA ligation
Ligation reaction mixtures containing 20 mM Tris-HCl (pH 8), 50 mM KCl, 10 mM MgCl2, 1 mM EDTA (sodium salt), 1 mM NAD+, 10 mM DTT and 0.5% Triton X-100. ODN1 (250 nM), ODN2 (10 nM) and ODN3 (250 nM) were heated to 95oC for 2 min, allowed to cool, and annealed for 1 h at the reaction temperature. DNA ligase (25 U) was then added and reactions incubated at required temperatures. Samples of the reaction mixtures were removed and the reaction stopped by addition of 50% final volume formamide, 5 mM EDTA. Ligated and unligated oligonucleotides were separated by electrophoresis on 20% denaturing polyacrylamide gels (20% acrylamide, 19:1 acrylamide:bisacrylamide, 7 M urea, 2 mM EDTA, 90 mM Tris-borate). Electrophoresis was carried out in 1* TBE (pH 8.3) at 25 W constant power for 3 h. After fixing and drying, gels were exposed to a storage phosphor plate overnight and scanned using a phosphorimager (Molecular Dynamics 400A). Reactions were quantified using Molecular Dynamics Imagequant software.
RESULTS AND DISCUSSION
Our model substrates consisted of three oligodeoxynucleotides: a template (ODN1) onto which two shorter oligonucleotides (ODN2 and ODN3) could hybridise to produce a duplex containing a single-stranded nick, which could then be sealed by DNA ligases (Fig. 1 ). By keeping ODN1 and ODN2 constant and varying ODN3, we monitored the effects on the reaction of perturbations within the duplex at sites 5' to the nick (Fig. 1 ).
ACKNOWLEDGEMENTS
The authors wish to thank Dale Wigley and Steve Ashford for the gift of T7 and B.stearothermophilus DNA ligases. Also we thank Martin Johnson for technical assistance and Nick Houseby for critical reading of the manuscript and SERC for funding.
