WO2001005800A2

WO2001005800A2 - A method for the enrichment of heteroduplexes and its use in mutation detection

Info

Publication number: WO2001005800A2
Application number: PCT/EP2000/006829
Authority: WO
Inventors: Borries Kemper; Stefan Golz
Original assignee: Borries Kemper; Stefan Golz
Priority date: 1999-07-18
Filing date: 2000-07-18
Publication date: 2001-01-25
Also published as: SE9902741D0; WO2001005800A3; AU6694800A

Abstract

A method for selectively enriching heteroduplexes from a mixture of hetero- and homoduplexes comprising binding selectively said heteroduplexes to a heteroduplex binding protein thereby forming a complex between the heteroduplex and the protein. The method is characterized in that the protein is a modified form of a wild-type resolvase, which modified form in comparison to the wild type resolvase retains an ability to selectively bind to heteroduplexes, and has a modified ability to cleave heteroduplexes. In case one of the nucleic strands in the heteroduplexes derives from a sample and the other is a standard nucleic acid strand, the method may be used for detecting mismatches in the sample nucleic acid relative the standard nucleic acid.

Description

C L A I M S

1. A method for selectively enriching heteroduplexes from a mixture of hetero- and homoduplexes comprising the steps of binding selectively said heteroduplexes to a heteroduplex binding protein thereby forming a complex between the heteroduplex and the protein, characterised in that said protein is a modified form of a wild-type resolvase, which modified form in comparison to the wild typ^e resolvase retains an ability to selectively bind to heteroduplexes, and has a modified ability to cleave heteroduplexes .

2. The method of claim 2 , characterised in that said modified form has a reduced ability to cleave heteroduplexes, for instance below 10% compared to the wild type resolvase.

3. The method of any one of claims 1-2, characterised in that said modified form is a wild type resolvase that has been mutated to the modified ability to cleave heteroduplexes .

4. The method of any one of claims 1-3, characterised in that said wild type resolvase is T4 endo VII.

5. The method of any one of claims 1-4, characterised in that said wild type resolvase recognizes all kind of mutations, such as replacements, insertions and deletions.

6. The method of any one of claims 1-5, characterised in that said heteroduplex or said modified form of the resolvase is insoluble or insolubilizable so that said complex will be formed in an insoluble form or transformed to an insoluble form after complex formation, and in that said insoluble form is separated from the reaction medium after its formation.

7. A method of detecting the presence or absence of a modification in a base of a nucleic acid strand (I) in relation to a standard nucleic acid strand (II) , said method comprising the steps of:

(i) forming a duplex comprising nucleic acid I and the standard nucleic acid II; (ii) reacting the duplex from step (i) with a protein selectively binding heteroduplexes thereby forming a complex between the protein and the duplex; and (iii) detecting and taking the formation of said complex as an indication that said nucleic acid (I) contains the modification, characterised in that said protein is a modified form of a wild type resolvase, said modified form in comparison to the wild type resolvase having a different composition of organic constituents, retaining an ability to selectively bind to heteroduplexes, and having a modified ability to cleave heteroduplexes .

8. The method of claim 7, characterised in that said modified form has a reduced ability to cleave heteroduplexes, for instance below 10% compared to the wild type resolvase.

9. The method of any one of claims 7-8, characterised that said modified form is a wild type resolvase that has been mutated to a modified ability to cleave heteroduplexes .

10. The method of any one of claims 7-9, characterised in that said wild type resolvase is T4 endo VII. 20

11. The method of any one of claims 7-10, characterised in that said wild type resolvase is T4 endo VII that is mutated in codon 62, in particular N62D (asparagine to aspartic acid) .

12. The method of any one of claims 7-11, characterised in that said wild type resolvase recognizes all kinds of mutations, such as replacements, insertions and deletions .

13. The method of any one of claims 7-12, characterised in that said wild type resolvase recognizes one or more of the mismatches G:A, C:T, C:C, G:G, A:A, T:T, C:A, and G:T.

14. The method of any one of claims 7-13, characterised in

• that step (i) utilizes nucleic acid strand I or nucleic acid strand II in insoluble or insolubilisable form so that said duplex is formed in an insoluble form or in a soluble form that subsequently is transformed to an insoluble form, and

• that step (ii) comprises reacting said insoluble form of the duplex, if being a heteroduplex, with said modified resolvase form in a soluble form so that said complex will be formed in an insoluble form; and

• that step (iii) comprises detecting the complex in the insoluble form containing the resolvase .

15. The method of any one of claims 7-14, characterised in that step (i) comprises using nucleic acid (II) and nucleic acid (I) in soluble forms, and that step (ii) comprises reacting the modified form of the resolvase in an insoluble form or in a soluble form that is transformed to an insoluble form subsequent to complex formation so that said complex will be formed in an 21 insoluble form, and that step (iii) comprises detecting the complex in the insoluble form containing the resolvase .

16. The method of any one of claims 7-15, characterised in that said complex is detected by detecting said modified form of the resolvase complexed to said heteroduplex, preferably using nucleic acid I or nucleic acid II in insoluble or insolubilisable form.

17. The method of claim 16, characterised in that said resolvase is detected by the use of an analytically detectable reagent (a labelled reagent) having biospecific affinity to the modified form of said resolvase .

18. The method of claim 16, characterised in that said modified form of said resolvase is labelled with an analytically detectable group.

19. The method of any one of claims 7-15, characterised in that said complex is detected by detecting either nucleic acid strand I and/or nucleic acid strand II as part of said complex, preferably with said modified form of said resolvase being in insoluble or insolubilizable form.

20. The method of claim 19, characterised in that either one or both of nucleic acid strand I and nucleic acid strand II is/are labelled with a respective analytically detectable group which may be the same or different for each of the two nucleic acids and which is/are used for detecting nucleic acid I and/or nucleic acid II in said duplex.

21. The method of any one of claims 1-20, characterised in 22 that the modified form has a different composition of organic constituents.

A METHOD FOR THE ENRICHMENT OF HETERODUPLEXES AND ITS USE IN

MUTATION DETECTION. Technical field

The invention concerns a method for selectively enriching heteroduplex DNA from a mixture of hetero- and homoduplex DNA by selectively binding heteroduplex DNA to a protein that selectively binds to heteroduplex DNA. The method can be used in the detection of modifications in bases of nucleic acids. By heteroduplex DNA or heteroduplexes is meant double stranded DNA in which there is a nucleotide in one strand that does not pair through Watson-Crick base pairing and π- stacking interactions with an nucleotide in an opposing complementary strand. This type of disturbance is often caused by a change in one or more oligonucleotides .

Disturbances or modifications of this kind are recognized by heteroduplex-binding proteins. Exemplary base modifications are replacements, deletions, insertions of bases, presence of chemically altered bases, apurinic and apyrimidinic sites etc.

The above-mentioned enrichments have for about 10 years been used for detecting the presence or the absence of the kind of modification referred above in sample nucleic acids. See for instance US 5,698,400; WO 9320233; US 5,217,863; and WO 9623903, which are hereby incorporated by reference. These detection methods comprise the steps of : i) forming a duplex between a sample nucleic acid strand comprising the base modification (nucleic acid I) and a standard nucleic acid strand (nucleic acid II) ; ii) reacting the duplex from step (i) with a heteroduplex- binding protein under conditions which promote the formation of a complex between the heteroduplex-binding protein and the duplex; and iii) detecting and taking the formation of said complex as an indication that the complex formed in step (ii) is a 2 heteroduplex and that nucleic acid (I) contains this kind of modification.

A number of different heteroduplex binding proteins have been used or suggested for selectively binding heteroduplexes in the above-mentioned methods. Some of the proteins have and some have not had the ability to enzymatically cleave the heteroduplexes. Enzymes that specifically cleave heteroduplexes are often called resolvases . One of the most prominent examples is T4 endonuclease VII (T4 Endo VII) which is an enzyme binding to and cleaving all possible mismatches in heteroduplexes. Another example is Mut S that does not have the same broad binding specificity and in addition lacks the ability to cleave heteroduplexes. When T4 Endo VII and other resolvases have been used, the cleavage products obtained have been detected and taken as an indication of the presence of a heteroduplex in the complex. The presence of cleavage products thus also has meant an indication of a modified base in nucleic acid I. By using labelled forms of either or both of nucleic acid I or II, the detection has been facilitated. By analysing the cleavage products (fragments) , the positions of disturbances have been determined.

In another variant the conditions during binding have been selected to make the resolvase inactive. Golz et al (Nucleic Acids Research 26 (1998) 1132-1133) , for instance, have described a variant in which solid phase bound wild type T4 Endo VII (T4 Endo VII^WT) is allowed to bind radioactively 5' -labelled heteroduplexes in the absence of Mg²⁺ . The apparent inactivity or low activity of the enzyme used may depend on a change in the substrate and/or in the enzyme.

In particular this latter variant has several disadvantages. For example, reproducibility was modest due to i) uncontrollable differences in the efficiency of binding of the enzyme to the surface of different microtitre plates, and ii) uncontrolled loss of already bound 3 heteroduplexes due to residual cleavage activity of T4 Endo

VII^WT even under high EDTA concentrations. Furthermore, an enrichment of heteroduplexes from DNA mixtures with a low content of molecules with mismatches was not practical since binding cycles could not be repeated in T4 endo VII seeded microtiter plates.

Golz et al (Eur. J. Biochem 245 (1997) 573-580) have reported about a mutated form of T4 endo VII that binds to heteroduplexes but has a significantly reduced ability to enzymatically cleave them.

Lilley et al (WO 9709434) and Dean (Nature Genetics 9 (1995) 103-14) have speculated that a mutated form of T4 Endo VII that binds but does not cleave heteroduplexes in combination with an antibody might be useful for detecting mismatches in microtiter wells. However, no enabling support for this proposal was given.

Objectives of the invention The primary objective of the invention is to provide improvements in relation to the disadvantages mentioned above .

The invention We have now discovered that improvements can be accomplished if the heteroduplex-binding protein is a modified form of a wild type resolvase. Compared to wild type resolvase this modified form:

• has a modified ability to cleave heteroduplexes, • retains an ability to selectively bind to heteroduplexes.

The broadest aspect of the invention therefore is a method for selectively enriching heteroduplexes from a mixture of hetero- and homoduplexes. The method comprises the step of binding selectively heteroduplexes to a heteroduplex-binding protein thereby forming a complex between the heteroduplex 4 and the protein. The method is characterized in that heteroduplex-binding protein is the modified form of a resolvase as defined in the preceding paragraph.

Unless otherwise specified, the sole term resolvase will mean a wild type resolvase.

Resolvases and modified forms thereof.

In the context of the invention, resolvases are enzymes that are able to selectively resolve native and/or synthetic X- forms of nucleic acid (Kemper B, (1997) Branched DNA resolving enzymes (X-solvases) in DNA Damage and Repair. Biochemistry, Genetics and Cell Biology, eds . J.A. Nickoloff and M. Hoekstra, (Totowa: Humana Press), pp. 179-204). One of the most important resolvases used so far in the field of the invention is T4 Endo VII that is known to bind to and cleave in principle any heteroduplex. Another potent enzyme is Cel 1 from celery (US 5,869,245). Still other potent resolvases are Saccharomyces cerevisiae Endo XI, Endo X2 , Endo X3, and CCE1 (Jensch et al . , EMBO J. 8 (1989) 4325-; Kupfer and Kemper, Eur. J. Biochem. 238 (1995) 77-), T7 endonuclease 1, E. coli MutY (Wu et al . , Proc . Natl . Acad. Sci. USA 89 (1992) 8779-8783), mammalian thymine glycosylase (Wiebauer et al . , Proc. Natl. Acad. Sci. USA 87 (1990) 5842- 5845), topoisomerase I from human thymus (Yeh et al . , J. Biol. Chem. 266 (1991) 6480-6484; and Yeh et al . , J. Biol . Chem. 269 (1994) 15498-15504), and deoxyinosine 3'- endonuclease (Yao and Kow, J. Biol. Chem. 269 (1994) 31390- 31396) .

Depending upon kind, a resolvase will recognize one or more of the mismatches G:A, C:T, C:C, G:G, A:A, T:T, C:A, and G:T.

Wild type resolvases contain one binding site for the substrate and one active site at which the enzymatic reaction takes place. In some resolvases the two sites more or less coincide. In other resolvases these sites can be remote from each other. For instance, resolvases bind to the 5 distorted part (mismatch) of a heteroduplex and may cleave at a predetermined distance of one or more bases therefrom, depending on the particular resolvase concerned. The binding and the cleaving activities will depend on the three- dimensional arrangement in each respective site. By having knowledge about the primary, secondary, tertiary and/or quaternary structure of a resolvase, it will be possible to predict fairly well at which positions and how amino acid residues should be modified in order to give modified resolvase that is useful in the instant invention. A codon coding for an amino acid that is essential for the cleaving reaction but not for binding reaction may by directly mutated. In some resolvases the enzymic activity, but not the heteroduplex-binding activity, is dependent on a metal ion which chelates to certain amino acid residues. Mutation of one or more of the corresponding codons is then likely to give candidates for useful mutants. It is also of potential interest to change an amino acid at a position essential for the cleaving activity to an amino acid that is non-conserved in relation to the original amino acid of the wild type resolvase .

The structural knowledge for a resolvase may be poor. In such a case spontaneous and/or random mutations may be collected and tested. Chemical derivatization can also give useful modified forms of resolvases .

The appropriateness of a particular modification in a resolvase may be tested in the previously known assays for determining efficiency of a modified resolvase. Testing may also take place by using the method of the instant invention.

The useful enzymic activity may be either enhanced or reduced. In the variants of the inventive method preferred at the priority date, forms exhibiting a significantly reduced enzyme activity are used, for instance reduced to below 10 %, such as below 1 %, of the enzymic activity for the wild type resolvase. These values refer to at least one of the substrates used in Golz et al . , Mutation Research Genomics 382 (1998) 85-92, preferably for two or more up to all . The conditions are preferably the optimal ones given by Golz et al . In the alternative the substrates may be the ones presented in the experimental part of this specification.

The modified form of the resolvase to be used in the instant invention retains a significant ability to bind selectively to a heteroduplex.

Resolvases may consist of one, two or more polypeptide chains that may be the same or different. Modifications can be made in one, two or more of the chains. A particularly interesting variant of a wild type resolvase containing two or more polypeptide chains is to have the chains fused together by an oligopeptide linker. The linker may for instance have 1-20 amino acid residues. If necessary this kind of resolvases are also modified specifically to change the cleaving activity as discussed above. The modified form preferably differs from the wild-type resolvase in the sense that its content of organic constituents is different. By this is meant that the change in cleaving activity is caused by a change in amino acid composition or by covalently attaching organic groups to a polypeptide chain deriving from the wild type resolvase. To use modifications in which an inorganic ion (e.g. Mg²⁺ and Zn²⁺) has been added to or removed from the resolvase without replacing or adding covalently incorporated organic constituents are excluded. The relevant modifications for modifying the cleaving activity should primarily be done at internal positions of the polypeptide chains of the resolvases .

The term modified form of a wild type resolvase includes that the resolvase also has been modified for other reasons. Such modifications can take place before or after the enzyme activity is modified. 7 At the priority date, the preferred wild type resolvase was T4 Endo VII . This enzyme is in active form a dimer of two identical polypeptide chains associated to each other in opposite direction. The binding activity is associated with the terminals and the cleaving activity by internal sites. The preferred modification was in coding codon 62 with the particular mutation being N62D. See Golz et al . , Eur. J. Biochem. 245 (1997) 573-580 and Raajimakers et al . , EMBO J. 18(6) (1999) 1447-1458. Further mutations in T4 Endo VII have been described by Lilley et al (WO 9709434) .

Enrichment and solid phases versus insolubilisation.

One of the most common purposes for enriching heteroduplexes from a mixture of hetero- and homoduplexes is for processing the heteroduplexes further. A typical example is in the assays described above for detecting base modifications in a sample nucleic acid I . A subordinate purpose is for preparing a heteroduplex-depleted mixture. In both cases it is, for practical reasons, important to physically separate the complex between the heteroduplex and the heteroduplex-binding protein from homoduplexes. This can be accomplished if the heteroduplex-binding protein initially is in insoluble form. The protein is then preferably covalently, biospecifically or physically adsorbed to a solid phase. In an alternative the heteroduplex-binding protein may be insolubilizable, i.e. transformable to an insoluble form after complex formation. Illustrative examples of insolubilizable variants of the protein are modified forms which are linked to a member of a biospecfic affinity pair, for instance biotin or glutathione or gluthathione transferase (GST) or a hapten. After complex formation the complex is contacted with a solid phase carrying the biospecific affinity counterpart to the member. The counterpart on the solid phase may be strepavidin or glutathione transferase or an antibody specific to the hapten, respectively. The conjugate between a member of a biospecific affinity pair and the resolvase may be a fusion protein.

An alternative route for preparing insoluble heteroduplex- protein complexes is to start duplex formation from an insoluble or insolubilisable form of either of the two nucleic acid strands. In the preferred cases one of the nucleic acid strands is covalently or biospecifically or via physical adsorption attached to a solid phase.

Insolubilisable forms are accomplished in the same way as for the heteroduplex-binding protein, except for the nucleic acid strand is attached the member of the biospecific affinity pair. Insolubilization may take place either subsequent to the formation of the heteroduplexes but prior to the complex formation, or subsequent to the complex formation.

The solid phases may have different shapes and forms. They may be in the form of particles such as beads, tube walls, planar surfaces, wells such as in microtiter plates etc. The solid phases may be porous or non-porous. Single surfaces may be divided into distinct part surfaces with different nucleic acids attached to different part surfaces. Such surfaces may comprise from 2 or from 10 up to hundreds or thousands of distinct part surfaces. This type of surfaces may be of potential interest for testing heteroduplex formation of a sample nucleic acid strand (I) against several standard nucleic acid strands (II) in parallel.

The surface of a solid phase is typically hydrophilic in the sense that it carries polar groups containing heteroatoms. Typical groups are carboxy, hydroxy, amido etc. These groups may be present on polymers selected from polyhydroxy polymers such as polysaccharides (dextran, agarose, starch, cellulose etc) and synthetic polymers such as polyvinyl alcohols, poly (hydroxy alkyl) methacrylates or corresponding polyacrylates . The solid phases may also be based on other polymers, for instance polyacryl or polymethacryl amides . 9

Length of nucleic acid strands and conditions for duplex formation.

The nucleic acid strands involved in duplex formation typically comprise at least 20 nucleotides and preferably have between 90 and 50,000 nucleotides, more preferably between 160-6,000 nucleotides. The strands may have been obtained via amplification, for instance by PCR, of a sample containing more or less native nucleic acid. The nucleic acid involved in duplex formation is preferably DNA. The figures are valid for both strands. Nucleic acid strands I and II may be of different length.

Reaction conditions for formation of complex between duplexes and the heteroduplex-binding protein.

The reaction media used for complex formation is typically aqueous. The conditions applied and the specific components in the aqueous medium are commonly known in the field of hybridisation. See for instance Ausebel et al . , Current Protocols in Molecular Biology, John Wiley & Sons, Inc. New York) .

The temperature, pH and constituents of the medium are the same as commonly applied in the field.

Detection of formation of complex between heteroduplex and protein.

Detection is possible by utilizing a detectable form of either or both of the strands forming the heteroduplex or of the heteroduplex-binding protein. The detectable form can be a labelled form of one of the reactants just mentioned with the provision that the label should not be on the insoluble or insolubilizable reactant . Typical labels (analytically detectable groups) are radioactive atoms or groups, fluorescent groups, chemiluminescent groups, enzyme related groups (enzymes, cofactors, substrates, cosubstrates etc), members of biospecific affinity pairs (such as biotin and 10 haptens) etc. The preferred labels are radioactive or otherwise radiation emitting, such as fluorescent or chemiluminescent .

Detection may also be accomplished by using an analytically detectable reagent that reacts with the complex. In this case the complex or its components contains a group that is recognizable by the reagent.

The presence of a heteroduplex in the complex may be detected by directly measuring the presence of the label in the complex, for instance on a solid phase. The presence may also be measured indirectly after complex formation by measuring the amount of added labelled reagent not being bound to the complex.

The labelled reactant may be either or both of the nucleic acid strands. In case the modified form of the resolvase is enzymatically active, labelled nucleic acid cleavage products released from the complex can be measured and taken as an indication of the presence of heteroduplexes in the complex. As an alternative an active resolvase that may be a wild-type variant or an active modification thereof may be added to the complex. The added resolvase may, for instance, be the variant described above containing fused polypeptide chains. Another potentially useful alternative is chemical cleavage at the disturbances in heteroduplexes (US 5,217,863; and Roberts et al . , Nucleic acid Research 25 (1997) 3377-3378) . By collecting the nucleic acid cleavage fragments released, and analysing them with respect to size and base composition, for instance by gel electrophoresis, further information with respect to type and position of base modifications can be gained.

The inventive method is useful in detecting mutations and polymorphisms in nucleic acids, for instance associated with mammalian diseases and in the context of forensic medicine. The method is also useful for typing and identifying bacteria, viruses, fungis, and other nucleic acid containing 11 uni- or multicellular organisms.

The invention will now be illustrated by the experimental part .

E X P E R I M E N T A L P A R T MATERIALS AND METHODS

Buffers, chemicals and radiochemicals : T4 Endo VII -buffer contained 50 mM Tris/HCl pH 8.0 , 12 mM MgCl₂ and 1 mM 2 - mercaptoethanol. TBE-buffer contained 45 mM Tris-borate (pH 8.0) and 1 mM EDTA. Phosphate buffer contained appropriate mixtures of K₂HP0₄ and KH₂P0₄. Acrylamide/Bisacrylamide Premix (29:1) for denaturing polyacrylamide gel electrophoresis (PAGE) was purchased from Serva

32 (Heidelberg) . γ- P-dATP (specific activity >3000 Ci/mmole) was purchased from Amersham (Braunschweig) . Glutathione Sepharose 4B was from Amersham Pharmacia Biotech (Freiburg) . All other chemicals were purchased from Merck (Darmstadt) .

Proteins: Wild type T4 Endo VII was purified as previously described (Golz et al . , DNA Res., 2 (1998) 277-284).

Cleavage deficient GST- fusion protein T4 Endo VII-N62D^GST was purified as previously described (Christoph et al . , J. Mol . Biol. 277 (1998) 529-540).

Oligonucleotides and DNA: Synthetic oligonucleotides were purchased from Pharmacia (Freiburg) , purified, radioactively 5^Λ -end-labelled and assembled into hybrid DNA substrates as described (Golz et al . , Nucleic Acids Res . 26 (1998) 1132 - 1133 ) . The nucleotide sequences of the substrates used in this study were the same as those given by Golz & Kemper (.Enzymatic mutation detection: Enrichment of heteroduplexes from hybrid DNA mixtures by cleavage-deficient GST- tagged endonuclease VII, Nucleic Acids Research 27 (15) (1999 ) (under the heading: Oligonucleotides and DNA) . 12

Mismatch substrates were assembled from the oligos by mixing adequate quantities plus and minus-strands . For example, the C/C-mismatch in position 21 of a 41bp long heteroduplex substrate was made by mixing oligonucleotides MM41C²¹⁺ and MM41C^21-. The homoduplex was assembled by mixing MM41C²¹⁺ and MM41G^21". The 8nt-insertion was assembled from oligos MM41CCATCCAG²¹⁺ and MM40[0]^21~ a minus-strand oligo with one nucleotide deleted in position 21 (marked by brackets) opposite the 8nt insertion in the plus-strand oligo.

T4 Endo VII digests and PAGE. A total volume of 10 μl T4

Endo VII reaction buffer contained 25 fmol of radioactively labelled DNA. In reactions with phosphate buffer, MgCl₂ was added shortly before the reaction was started to avoid precipitation. After the addition of T4 Endo VII the reaction mixtures were incubated for 15 minutes at the temperatures indicated. Reactions were terminated by ethanol precipitation followed by resuspension in M&G stopmix.

Aliquots were loaded on 12 % denaturing polyacrylamide gels containing 7 M urea. Reaction products were visualised by autoradiography and quantitated by phosphorimaging on a Fuji BAS1000.

RESULTS AND DISCUSSION

The T4 Endo VII-N62D^GST based binding assay was tested for all possible mismatches in synthetic heteroduplex DNAs. For formation of protein-DNA complexes, T4 Endo VII-N62D^GST and DNA-substrates were reacted before the protein was bound to Gluthatione-Sepharose . This improved the final yield of bound heteroduplex DNA over homoduplex DNA considerably. In particular, routine reactions were performed in a total volume of 500 μl containing 1.4 μg T4 Endo VII-N62D^GST and 2 fmole of labelled sample DNA (about 10,000 cp ) . 100 μg of Gluthatione-Sepharose were added after the incubation period, and the mixture was transferred to a spin column or dot-blot apparatus to separate solid from liquid phases. Separation conditions were chosen such that no extra washing or drying steps were necessary to determine the amount of bound DNA by either short term phosphoimaging or autoradiography over night . It should be noted that the 5 binding characteristics of T4 Endo VII N62D^GST to heteroduplex DNA were the same as for non-tagged mutant T4 Endo VII N62D or wild type enzyme T4 Endo VII^T described previously (Golz et al . , Nucleic Acids Res. 26 (1998) 1132- 1133) .

10

Reaction conditions. The binding of hetero- and homoduplex DNAs to T4 Endo VII-N62D^GST depends on the reaction conditions, and temperature, time, salt concentration and different buffers were tested. An optimal temperature of

15 16°C (figure 1 a) and an optimal incubation time of 15 minutes were determined (figure 1 b) . Best results were obtained with phosphate buffer at pH 6.5 (P0₄-buffer) as described before (Golz et al . , Mutat Res. 382 (1998) 85-92). When P0₄-buffers of 125 mM or 150 mM were tested in

20 combination with 50 mM, 100 mM or 200 mM KCl, an optimal KCl concentration of 50 mM was determined for detection of all substrates in either of the two concentrations of phosphate buffer. A KCl concentration of 100 mM and higher is selective for branched DNAs like cruciform DNA, only (figure

25 1 c) . In conclusion, for mismatch detection a 125 mM P0₄- buffer at pH 6.5 and 50 mM KCl were the optimal conditions. They were used here for the following experiments.

As shown in figure 1 e, all mismatches were detectable under 30 the optimal binding conditions. Characteristic differences in the relative efficiencies of binding were observed for individual mismatches. The affinity of T4 Endo VII-N62D^GST was the highest for C/C- and G/G-mismatches, the lowest for A/A- and T/T-mismatches and intermediate for A/G- , A/C-, 35 T/G- and T/C-mismatches (mismatches T/G and T/C are not included in figure e) . All experiments had been repeated at least 5 times giving identical results, thus reflecting a high degree of reproducibility of the method.

Locating mutations in heteroduplex DNA. To locate a mutation in the heteroduplex-DNA, spin columns were used for trapping, washing and draining the GST-Sepharose . After closure of the columns, 500 μl T4 Endo VII reaction buffer containing 100 U of wild type (cleavage active) T4 Endo VII^WT was added. After incubation for 15 minutes at 37 °C 25 μl of SDS (20 %) was added and the DNA was removed by centrifugation. Samples were EtOH precipitated and analyzed for diagnostic fragments on 12 % PAA gels (see figure 2 c) .

Enriching heteroduplex DNA. For the enrichment of heteroduplex-DNAs from samples with a low content of heteroduplexes the spin column procedure was used successfully. After trapping Gluthatione-Sepharose from the binding reaction in a spin-column and the first centrifugation, another sample of DNA was added in 500 μl binding buffer, incubated for 5 minutes at 16 °C and then centrifuged. This step was repeated up to five times. The amount of bound DNA was determined after each round in separate experiments and the amount of trapped heteroduplex DNAs was determined by enzymatic cleavage. The total amount of DNA bound to the column did not markedly increase with repetitions (compare lanes 3 and 10 in figure 2 c) . However, the content of heteroduplex DNAs among the total DNA bound, increased from about 10 % to 45 % after the third cycle. No further increase was observed with additional cycles.

Note that the heteroduplexes are distinguishable from homoduplexes in this experiment because an 8nt insert was chosen in the plus-strand of the construct which makes it distinguishable from the minus-strand on denaturing gels (figure 2 a, lanes 3 - 8 and figure 2 b) .

A size distinction can not be made with an equal number of nucleotides in both strands of a heteroduplex as expected 15 during routine mutation detection analyses. However, if the cleavage reaction is performed after the binding cycles are completed the relative content of bound heteroduplexes becomes visible. Again, a relative increase of heteroduplex DNA is observed with each repetition cycle starting from an initial, barely detectable proportion of about 10 % to a clearly detectable portion of about 40 % after three to four cycles (figure 2 c, lanes 3 - 10 and figure 2 d) .

Detection of mismatches as well as other distortions in hybrid DNAs made from longer PCR-fragments of 80 - 250 bp by solid phase EMD technology have previously been described (Golz et al . , Nucleic Acids Res. 26 (1998) 1132- 1133) .

From the intensity of the signal (marked 'product' in figure 2 c) , one may conclude that mismatch heteroduplexes can be made visible even from samples with lower than 10 % heteroduplexes. This, however, was not the case for unknown reasons, indicating that 10 % heteroduplex content in a sample provides a kind of threshold concentration for the reaction.

In conclusion, these experiments show that mutant cleavage deficient T4 Endo VII can be used for selective binding and enrichment of diagnostic heteroduplex DNAs from DNA mixtures with contents of heteroduplexes too low for detection using regular analytical procedures. Repeated cycles of binding reactions followed by diagnostic cleavage using wild type T4 endo VII, allows reliable detection of heteroduplexes in samples with as low as 10 % heteroduplex content.

LEGENDS TO FIGURES

Figure 1. Binding of heteroduplex DNA by T4 Endo VII-N62D^GST under various assay conditions, a) For determination of temperature dependence, samples were incubated at 4°C, 16°C, 21°C or 37°C respectively, and analysed according to the standard protocol described in the text . Images obtained for different oligonucleotides are shown for the different temperatures as indicated, b) Time dependence of the reactions was determined for 60 min, 15 min or 5 min with an 8nt insertion heteroduplex substrate. Images obtained from each reaction are shown, c) Influence of KCl concentrations was determined in reactions with different DNA samples either with 125mM or 150mM phosphate buffer at pH 6.5. Images obtained for the different oligonucleotides are shown, d) The sensitivity of the binding reaction was determined using DNA-mixtures containing varying relative concentrations of heteroduplex DNA as indicated. Images of each reaction are shown, e) Comparison of the relative binding efficiencies for various mismatch oligonucleotides. Equal amounts of mismatch containing DNAs (1 fmole) were mixed with homoduplex DNA (1 fmole) and analysed following the standard protocol. Images of each sample are shown. Abbreviations are: inp : input DNA in the assay; ins : heteroduplex oligonucleotide with an 8nt insertion; ds : homoduplex control DNA; cf : cruciform DNA CFM13 ; mm: oligonucleotides with mismatches. Mismatch C/C was used in a) , b) , c) and d) as an example.

Figure 2. Enrichment of heteroduplex DNAs from DNA mixtures and their visualisation by cleavage, a) Selection of heteroduplex-DNA with an 8 nt insertion oligo. A DNA mixture with 10 % heteroduplex DNA and 90 % homoduplex DNA was analysed according to the protocol. The incubation was repeated 1 to 5 times and the bound material was analysed on 12 % denat . PAA-gels. b) Quantitation of the results shown in figure 2a) . c) Selection of heteroduplex-DNA with a C/C mismatch from a DNA mixture containing 10 % of heteroduplex DNA and 90 % homoduplex DNA by binding. Binding reactions were repeated 1 to 5 times. Each sample was then exposed to 100 U of wild type T4 Endo VII^WT under standard reaction conditions and samples were then analysed on 12 % denat. PAA-gels. Kl : Untreated homoduplex-DNA; K2 : untreated heteroduplex-DNA. d) Quantitation of the result in figure c.