WO1990001548A1

WO1990001548A1 - Gene modification

Info

Publication number: WO1990001548A1
Application number: PCT/GB1989/000862
Authority: WO
Inventors: John Brandon Taylor
Original assignee: Cancer Research Campaign Technology Limited
Priority date: 1988-07-29
Filing date: 1989-07-28
Publication date: 1990-02-22
Also published as: GB8818150D0

Abstract

A process according to any preceding claim for producing a double-stranded DNA construct having a nucleotide base sequence which differs at a predetermined locus from the base sequence of a starting fragment which process comprises the steps of (a) digesting and blunt ending a first sample of starting fragments having a protected 5'-terminal so as to produce a population of fragments including at least one upstream fragment; (b) ligating to the 3'-terminals of fragments produced according to step (a), first precursors for a linker; (c) digesting and blunt-ending a second sample of starting fragments having a protected 3'-terminal so as to produce a population of fragments including at least one downstream fragment; (d) ligating to the 5'-terminal of fragments produced according to step (c), second precursors for the linker; (e) ligating fragments produced according to steps (b) and (d) so as to produce a population of fragments each comprising a linker; (f) obtaining from a population of fragments produced according to step (e) at least one intermediate fragment having the structure: 5'....upstream fragment:linker:downstream fragment....3'; by probing for the upstream fragment:linker and linker:downstream fragment junctions; (g) excising at least the entire linker from an intermediate fragment produced according to step (f); and (h) ligating together the residues of upstream fragments and downstream fragments produced according to step (g), optionally with insertion of an intervening oligonucleotide, so as to produce the double-stranded DNA construct required.

Description

TITLE : "GENE MODIFICATION"

The present invention relates to a process for modifying the nucleotide base sequence of genes and other DNA fragments at a predetermined locus and to materials for use in the process.

Recombinant DNA techniques have advanced to the point where those skilled in the art are no longer content simply to express proteins having their native amino acid sequence; now there is a desire to be able to modify the amino acid sequence in a controlled manner in order to improve desirable properties of the protein, to reduce undesirable properties thereof or to investigate the importance of various domains within the protein. In addition, it is desired to be able to modify regulatory regions within or outside the coding sections of genes in order to alter the way in which expression of the genes is controlled, for instance in order to enhance expression of the protein in transformed cells and to modify the codon usage within the gene to adapt it to expression in particular cells.

Presently available techniques of in vitro mutagenesis enable modifications to be made at or near restriction endonuclease recognition sites. However these sites are often far apart and randomly located so there may be no convenient site near the locus to be modified. For modifications consisting of the substitution of a single base or an adjacent pair of bases, M13 mutagenesis may be used. However this technique relies on hybridisation of* single-stranded DNA to an oligonucleotide which is not completely complementary to the DNA; this in itself can introduce errors. The method is also susceptible to errors generated by the host cell's DNA repair enzymes and it relies on copying of single-stranded DNA, which again may be a source of errors. In addition, where more than two bases are to be modified and/or where the sites of modification are not immediately adjacent, the M13 mutagenesis technique must be used several times to complete all the desired alterations. Except in the simplest cases this method is, therefore, extremely laborious. There are also difficulties in that the size of the M13 phage limits the amount of DNA inserted in the genome.

The present invention seeks to overcome these difficulties by providing a process whereby a DNA sequence can readily be modified by deletion, insertion or substitution at any desired locus.

The invention therefore provides a process for producing a double-stranded DNA construct of modified sequence which comprises constructing an intermediate fragment containing a linker at the site for modification, excising at least the entire linker and ligating the remaining fragments to produce the desired construct. The linker is a double-stranded oligonucleotide which includes recognition sites for restriction endonucleases located within the linker such that by treatment of a fragment containing the linker with appropriate endonucleases capable of cutting outside their recognition sequences, at least the entire linker may be excised from the intermediate fragment.

The manipulations according to the present invention are all conducted on double-stranded DNA fragments. In the majority of cases these fragments will contain a coding strand and its complementary non-coding strand though on occasion both strands may be non-coding. For simplicity in the following description, the manipulations will be described in connection with a "plus" strand which will usually, but not always, be a coding strand. Except where the contrary is indicated, references to 5' and 3' flanks, terminals, and to upstream and downstream directions apply to the plus strand. The complementary minus strand (usually non-coding) will, of course, also be present.

The intermediate fragment is constructed such that, in its plus strand, it contains the sequence

5'.... upstream fragment:linker:downstream fragment....3'

where the upstream fragment contains, at its 3' terminal, the sequence immediately 5' of the modification site. The 5' terminal of the downstream fragment contains the sequence immediately 3' of the modification site and the linker therefore occupies the modification site itself.

The upstream and downstream fragments are obtained by digestion of starting fragments having the starting sequence which is to be modified.

In one embodiment of the invention, a sample of starting fragments have their 5'-termini of the plus strands (and 3' termini of the minus strands) protected from digestion and are then digested and blunt ended. Since the 5'-terminal of each fragment cannot be digested, digestion occurs only from the 3'-terminal and continues in an upstream direction. The reaction is halted when sufficient time has elapsed for the digestion to have reached the vicinity of the modification site such that there is now a population of fragments which will include fragments terminating exactly at the 5' flank of the modification site, i.e. the desired upstream fragments. Downstream fragments are produced in similar manner but by protection of the 3'-termini of the plus strands of the starting fragments. The upstream and downstream fragments are then ligated to the linker to form the intermediate fragment.

Digestion of the starting fragments may be achieved by conventional methods. Use of exonuclease III is particularly preferred as this enzyme cleaves off nucleotides at a fairly constant rate and it is thus a simple matter of timing the reaction until the necessary length of starting fragment remains. Of course, there is some variation in the cleavage of individual starting fragments and the enzyme thus produces, after the appropriate delay, a population of digested fragments all of about the desired length.

For subsequent use in the process of the invention the digested end of the starting fragments should be blunt-ended. Where a digestion enzyme is selected which leaves a sticky end, the fragments will be subjected either to blunt-ending by digestion of the sticky end, for instance using Si nuclease or mung bean nuclease, or the sticky end may be "filled in" by conventional techniques. Protection of the starting fragments (and subsequent removal of any protecting moieties) may be achieved by conventional methods, the means of protection being selected having regard to the type of digestion to be effected. Often the starting fragments will be produced in a cloning vector and in this case simply by cutting the vector at, say, only the 3'-terminal of the starting fragment vector DNA may be left linked to the 5'-terminal and this vector DNA, if sufficiently long, will afford the necessary protection.

Alternatively the starting fragments may be protected from digestion by the nature of their termini. Thus, for instance if exonuclease III is to be used for digestion, (see, for instance Gene 28, 351-359 (1984), by providing the starting fragment with a 3'-extension at one terminal of the plus strand, the enzyme will be prevented from attacking that terminal whereas the other terminal. being provided with a 5'-extension will be susceptible to digestion. Appropriate termini may be provided by cutting with suitable restriction enzymes, if necessary followed by blunt-ending or filling in.

The present invention permits different starting fragments to be used for the generation of the upstream and downstream fragments. However it will usually be the case that the starting fragments are the same. In this case it is convenient for the starting fragments to be obtained as inserts in a multi-cloning vector and it is particularly preferred to use a mixture where some vectors contain the starting fragments in one orientation and the rest contain the starting fragments in the other orientation. Such multi-cloning vectors with inserts in both orientations may be obtained conventionally.

When these multi-cloning vectors are used to provide the starting fragments, cleavage of the vectors with an appropriate restriction endonuclease having a recognition site near the insert generates two species of linear DNA, the first containing the starting fragment with the vector DNA at its 3' flank, the second containing the starting fragment with vector DNA at its 5' flank. By using an appropriate restriction endonuclease which generates a 3'-extension and/or by use of a sufficiently large vector, the starting fragments are readily produced in protected form ready for digestion. In this case it is not necessary to separate the two types of starting fragments: the digestion can be conducted on the mixture. a sample removed after an appropriate time to form the upstream fragment and another sample removed after an appropriate time to form the downstream fragment. When the modification site is nearer, for example, the 5'-terminal than the 3'-terminal of the plus strand of the starting fragment, the upstream fragment will be shorter, and thus require longer digestion, than the downstream fragment, so that the sample providing the downstream fragment would be taken first.

If it is desired to make modifications at several sites within a starting fragment, a suitable selection of samples are removed from the digestion reaction at appropriate times to form all the necessary upstream and downstream fragments.

Whilst not essential, it is usually much preferred to separate the desired upstream and downstream fragments from the digestion mixture(s) to reduce the amount of extraneous material being incorporated in subsequent steps. DNA fragments may be separated according to size, for instance by HPLC (Dornburg, LC.GC International, 1_^ (3), 71 (1988) and, preferably Merion et al., J^_ Analysis and Purification (Waters) , 1987, pp 60, 61), to provide fractions with a narrow molecular weight distribution and those fractions containing the upstream and downstream fragments can be used in the further steps described below. Where vector DNA is present, for instance when used to protect the starting fragments, this should be removed, for example by treatment with an appropriate restriction endonuclease, before size-dependent separation is conducted.

Preferably deprotection of the digestion fragments is conducted after blunt ending is complete, and most preferably the deprotection step will leave a sticky end ready for later insertion in a cloning vector and to avoid spurious ligation of linker precursors as described below. Where protection is provided by vector DNA, removal using the appropriate restriction enzyme will normally generate the required sticky end.

The next stage in the procedure is ligation of the blunt-ended upstream and downstream fragments with the linker. Whilst this may be achieved directly there is a difficulty in that blunt end ligation reaction rates are generally slow. To enhance the yields of the desired fragments and improve reaction rates it is preferred to use precursors of the linker which are ligated to the upstream and downstream fragments and then ligated to form the complete linker.

In one embodiment the precursors of linkers are part-linkers each having a blunt end and a sticky end. By use of an excess of part-linkers, the blunt end ligation reaction can be driven faster. The ligation affords fragments having sticky ends originating from the part-linkers; the sticky ends of the part-linkers ligated to the upstream fragments may be complementary to those of the part-linkers ligated to the downstream fragments such that annealing and ligating the part-linkers will produce the intermediate fragments. Alternatively the sticky ends of the part-linkers on the upstream and downstream fragments may be ligated via a double-stranded oligonucleotide itself having sticky ends complementary to the sticky ends of the part-linkers.

In this embodiment, in order to avoid the formation of spurious fragments and to avoid sticky-end ligations of the part-linkers, it will be convenient to use part-linkers having sticky ends which are not self-complementary. In this case it will be necessary to separate the upstream fragments from the downstream fragments before ligation to the part-linkers (or to have produced them separately) and to use two separate types of part-linkers for ligation respectively to the upstream and downstream fragments.

In an alternative embodiment, a second type of precursors are used which are protected part-linkers having two blunt ends and which include a restriction endonuclease recognition site, (a "precursor cleaving site") located such that treatment with the enzyme will remove a portion of the precursor to leave a part-linker as described above having a sticky end. In this case the upstream and downstream fragments are each ligated to appropriate precursors; preferably the precursors are present in excess to drive the reaction and although this will tend to result in chains of several precursors being ligated together and to the upstream and downstream fragments, subsequent treatment with the appropriate restriction enzyme will result in cleavage at the precursor cleaving site and removal of all precursor moieties except the desired part-linkers. The upstream and downstream fragments can then be annealed and ligated via the sticky ends of their respective part-linkers to afford the intermediate fragments. This may be achieved directly where the part-linkers on the upstream fragments have complementary sticky ends to those of the part-linkers on the downstream fragments or it may be achieved via a double-stranded oligonucleotide having sticky ends complementary to the sticky ends of the two part linkers.

Where the upstream and downstream fragments are separated or generated separately, it is possible but not necessary to ligate different precursors to the upstream fragments and to the downstream fragments. Preferably the precursors ligated to the upstream and downstream fragments are the same.

In this alternative embodiment, it is possible to treat a mixture of the upstream and downstream fragments with the linker precursor but in this case the precursor should be palindromic (i.e. the plus and minus strands in addition to being completely complementary, have identical sequences when respectively read in the 5' to 3' direction) so that it is immaterial which blunt end of the precursor is ligated to the fragments. Preferably the precursor cleaving restriction endonuclease recognition site should be one which yields a self-complementary sticky end on cleavage with the corresponding restriction endonuclease so that the resultant part-linkers on the upstream and downstream fragments may be annealed and ligated directly to form the intermediate fragment.

The precursor cleaving restriction endonuclease recognition site is preferably one for a rare restriction endonuclease such that it is unlikely that use of the enzyme will cause cleavage of the upstream or downstream fragments. More preferably the site is for a restriction endonuclease which is sensitive to modification, such as methylation, of the recognition site, or which requires at least 6 nucleotide base pairs, for instance 7 and preferably 8 or more nucleotide base pairs in its recognition site. Alternatively the precursor is selected to have a cleaving site which does not appear in the sequence of the upstream and downstream fragments (or any associated vector DNA). In a further alternative, cleavage sites in the upstream and/or downstream fragments may be protected for instance by methylation or by in vitro mutagenesis.

Whether part-linkers or protected part-linkers are used, the precursors will be selected in order to provide, where necessary in combination with suitable double-stranded oligonucleotides, the desired linkers.

As previously described, the linkers comprise restriction endonuclease recognition sites (hereafter called "excision recognition sites") located and orientated such that they direct the corresponding restriction enzymes, which cut outside their recognition sequence in at least one strand of the DNA, to excise the entire linker. There are a number of such enzymes known and available to those skilled in the art. Preferred excision recognition sites are for restriction enzymes which cut both strands of DNA outside the recognition sequence and more preferably the excision recognition sites are for enzymes which are sensitive to modification, such as methylation, in the recognition site or which require a recognition site of at least 6 nucleotide base pairs, for instance 7 and preferably 8 or more nucleotide base pairs; this ensures that it is unlikely that the restriction enzymes will damage the upstream and downstream fragments (or any associated vector DNA). Such recognition sites may be palindromic (in which case the plus and minus strands within the site have the same sequences when read respectively in the 5' to 3' direction and, of course, are fully complementary) or non-palindromic. Palindromic recognition sites for enzymes which cut outside their recognition site will allow the enzyme to cut both upstream and downstream of the recognition site so that excision of a linker from an individual intermediate fragment might involve several cleavage events with the enzyme progressively removing portions of the linker. Where the excision enzyme cuts outside its recognition site in only one strand of the DNA, the remnants of the linker may be removed by, for instance, blunt ending.

Non-palindromic sites are preferred since these direct the enzyme to cut only at one flank of the recognition site.

It is conceivable that a single palindromic recognition site would permit one enzyme to excise the entire linker so that only a single excision recognition site is required. However it will be more convenient in most cases to include two excision recognition sites and this is essential when the sites are non-palindromic. The sites may be the same or different and, if they are different, two different restriction endonucleases may be required to complete the excision.

The cutting site of the restriction enzymes used to excise the linkers will always be at or beyond the flank of the linker in at least one strand thereof. Where the enzyme is one which cuts to leave a blunt end, the cutting site must always be at or beyond the 5' or 3' flank of the plus strand of the linker. When the enzyme is one which cuts to leave a sticky end, and this is preferred, the sticky end may be a part of the upstream or downstream fragment, or it may be a part of the linker. In the latter case the remaining linker nucleotides may be removed by blunt ending. When sticky ends originate from the upstream or downstream fragments and it is desired to produce blunt-ended fragments for ligation to form the desired construct, these.may be blunt-ended or filled in as convenient.

Preferred linkers for use in the present invention are at least 13 nucleotide base pairs in length. Particularly preferred linkers according to the invention comprise a third recognition site for a restriction endonuclease; there is no limitation as to the size of the site or to the location thereof, nor even to the position that the restriction enzyme will cut the linker except that the cutting site should be within the linker. It is however preferred that this restriction site is of at least four, more preferably five and yet more preferably six or more nucleotides in length and it is presently preferred that the cutting site is within the recognition site. In these preferred linkers, the third restriction enzyme cutting site is formed by annealing and ligating the part-linkers and the linker oligonucleotide thus comprises, in order from the 5' to the 3' end of the plus strand, a restriction site for cutting 5' of the linker, a restriction site for cutting within the linker and then a restriction site for cutting 3' of the linker. Linkers may be designed with the restriction sites overlapping, adjacent or separated by intervening sequences. Such linkers are also, of course, precursors for part-linkers as previously described. Other preferred linkers for use in the invention include additional restriction endonuclease recognition sites located to facilitate removal of an internal portion of the linker and designed so that the remainder may be ligated directly together or so that a new internal portion may be inserted. Such arrangements facilitate modification-of the linker where this is desirable to avoid the use of restriction enzymes which would damage upstream and downstream fragment DNA (or associated vector DNA) . Thus certain linkers of the invention, in addition to excision recognition sites may contain two, three or even more extra restriction enzyme recognition sites.

Precursors for linkers are either part-linkers or protected part-linkers. The part-linkers each have an excision recognition restriction site (or part thereof) for cutting outside the blunt end of the part-linker and have a sticky end for ligation to another part-linker. Protected part-linkers have at least one restriction site for generation of the sticky end (and this will often also form a restriction site for cutting within the linker) as well as having an excision recognition site or a part thereof.

It is particularly preferred that such protected part-linkers are palindromic so that on cutting each precursor affords two part-linkers and in this case they will preferably have three restriction sites, two of which are identical excision sites for cutting at or outside the flanks of the part-linkers. The 5' to 3' sequence of the plus and minus strands of these palindromic precursors will, of course, be identical. These particular precursors thus also have the same features as linkers and may be regarded as being linkers as well as precursors therefor.

The linkers may be non-palindromic or, with careful design, may even be palindromic whether the excision sites themselves are non-palindromic or palindromic. Regardless of whether the linkers are palindromic or not, it is preferred that the 5' flank of the plus strand has the same sequence as the 5' flank of the minus strand (when read 5' to 3' respectively) and that these "common sequences" are at least 13 nucleotides long. Thus particularly preferred linkers are at least 13 nucleotides in length; the reasons for this will be explained below in connection with the probes of the invention.

The design of linkers and precursors and part-linkers according to the above requirements is within the ability of those skilled in the art. Particular linkers, precursors and part-linkers are illustrated below and further information assisting their design may be obtained for instance from Gene, _j47_» 1-153 (1986) and the New England Biolabε Catalogue 1988, particularly page 141 giving a summary of restriction enzymes which recognise rare and internal sites and generate complementary ends. Another useful list is that on page 102 of the Catalogue showing the frequency of BspMI in commonly used vectors (page 102) which would need to be protected if using BspMI for instance to excise the linker.

Because the digestion processes available for producing upstream and downstream fragments inevitably produce mixtures of fragments and as the available physical separation techniques are unlikely to be sufficiently discriminating to exclude all but the desired fragments, the intermediate fragments themselves will unavoidably be mixed with unwanted similar fragments, many of which will also contain linkers.

There are two general routes to the formation of the desired construct from these mixtures. The more preferred is the use of probes in identifying and separating the intermediate fragment, and this will be described later. The other general route involves the use of specifically synthesised oligonucleotides and will be described now.

In one embodiment of the process the mixture containing the intermediate fragments is treated with appropriate enzymes to excise the linker and afford sticky ends on the upstream and downstream fragments. The sequence of these sticky ends will be known having been preselected when the site for modification was identified. A specifically designed oligonucleotide having corresponding sticky ends is used to ligate the upstream and downstream fragments by conventional methods. Ideally this is performed using upstream and downstream fragments joined by vector DNA via the 5' flank of the upstream fragment and the 3' flank of the downstream fragment so that ligation of the upstream and downstream fragments via the oligonucleotide results in circularising ligation. By conducting the ligation under high dilution, unimolecular reactions are favoured so that spurious circulariεations are minimised. The DNA is then used in transformation of cells and of course only the circular DNA will result in replication of the vector.

In conducting this variant of the process it is preferable to design the linkers and select the linker-excising enzymes so as to leave a long sticky end to maximise the accuracy of ligation via the oligonucleotide. The oligonucleotide itself is designed not only to provide means for ligation of the upstream and downstream fragments but also to provide the required new sequence at the modification site. Ligation via an oligonucleotide is described in further detail below. The preferred method for forming the desired construct from the mixture containing the intermediate fragment is by identification and separation of the desired intermediate fragments using probes which have sequences capable of identifying the linker when attached to either or both of the upstream and downstream fragments.

In one variant of this process, identification is achieved using two probes. The sequence of the first probe corresponds with the sequence of bases including the junction of the 3'-flank of the upstream fragment and 5'-flank of the linker whereas the second probe corresponds with the sequence of bases including the junction of the 3'-flank of the linker and the 5'-flank of the downstream fragment. The first probe will identify those fragments having the correct upstream fragment and the linker ligated thereto, the second probe will identify those containing the correct downstream fragment and the linker ligated thereto. Thus only the desired intermediate fragment will hybridise with both probes.

In another variant, which is convenient only with short linkers, a single probe is used which has a sequence corresponding with that of the linker and has overlaps at the 5' and 3' flank of this sequence corresponding respectively to the 3' flank of the upstream fragment and the 5' flank of the downstream fragment.

In order to distinguish a unique DNA sequence, statistics indicate that a probe should be at least 17 nucleotides in length. The present probes are preferably this length or longer and more preferably the probes used in the present invention comprise a sequence of at least 3 bases,for instance 5 or more but most preferably 4 bases corresponding with the appropriate flank of the upstream or downstream fragment, the balance (preferably a sequence of at least 13 nucleotide bases) corresponding with the sequence of the linker (thus the linker should itself be at least 13 base pairs in length).

Probes for detecting the linkers according to the present invention may be single or double stranded. Single stranded probes contain a portion of DNA having a nucleotide base sequence identical to the sequence of at least a part of either the plus or the minus strand of a linker according to the present invention. Double stranded probes contain a portion of DNA having a nucleotide base pair sequence identical to at least a part of the nucleotide base pair sequence of a linker according to the present invention. For use in identifying the intermediate fragments, probes according to the present invention will comprise, in addition to a portion of DNA having a sequence corresponding to the sequence of the linker, a portion of DNA having a sequence corresponding to the 3' flank of the upstream fragment or the 5' flank of the downstream fragment. For accuracy in identifying only the intermediate fragments, it is preferred that the probe comprises at least 17 nucleotide bases (or base pairs if a double stranded probe is used) so that the chance of a false positive result is minimised. It is usually convenient to manipulate fragments of DNA having a length of approximately 1 kb and statistics indicate that any 4-base sequence is likely to be occur about four times in such a fragment. It is, therefore, usually sufficient for the portion of the probe having a sequence corresponding to the 3' flank of the upstream fragment, or the 5' flank of the downstream fragment, to consist of 4 nucleotide bases (or base pairs).

Probes for use in the present invention may be detected by a variety of techniques known in themselves to those skilled in the art. Preferably the probes bear labels which may be detected directly or indirectly, for instance radio-labels, fluorescent labels, chromophores, enzyme labels and. specific binding agent labels such as antigens which may be detected by corresponding binding partners such as antibodies. The invention also extends to unlabelled probes which may be supplied in a form suitable for labelling shortly before use; this applies particularly where probes are to be used with radio-labels which have a relatively short half-life and cannot therefore be stored in labelled condition for any great length of time.

As previously mentioned, it will usually require at least two probes to identify the desired intermediate fragment. It may be convenient for the linker to be designed to have "common" portions of identical sequences at the 5' flank of its plus strand and the 5' flank of its minus strand. In this particular case, the intermediate fragment may be identified using a pair of probes, each of which comprises a "common" portion having an identical sequence to the sequence of the common portion of the linker. One probe also comprises a portion of at least four nucleotide bases (or base pairs) corresponding to the 3' flank of the upstream fragment, the second probe comprises a portion of at least four nucleotide bases (or base pairs) having a sequence corresponding to the 5' flank of the downstream fragment. This principle enables a complete set of probes to be produced for use in conjunction with a particular linker, all the probes having a common portion and a 'variable' portion of four nucleotide bases for use in identifying any possible upstream and downstream fragments.

As there are only four nucleotide bases (A, C, G and T) there are a total of 4 «256 possible combinations of four nucleotide bases, a complete set of probes could comprise 256 different probe species. Probes for only 3 bases of the upstream or downstream fragment are unlikely to be sufficiently specific whereas a set of probes for probing for five bases of the upstream or downstream fragments would contain 1024 different probes and so would be inconveniently unwieldy. Irrespective of the starting sequence to be modified, such a set of probes would always include the two probes which would be necessary for identifying the particular intermediate fragment produced in accordance with the process of the present invention.

The probing and sub-cloning procedures for separation and identification of the intermediate fragments are conducted in conventional manner. Preferably the probing is conducted under high stringency conditions. It is particularly preferred that the probes are selected such that the full length of each probe will hybridise with the intermediate fragment and that the probing is carried out under conditions such that only fragments which hybridise to the full length of the probe will be selected, for instance as described by Wood e_t al, Proc. Natl. Acad. Sci. USA., 82, 1585-1588). Once the intermediate fragments have been separated, the next stage is removal of the linker followed by ligation of the upstream and downstream fragments. The linker is excised by appropriate restriction enzymes taking advantage of the excision recognition sites which direct restriction endonucleases to cut at or beyond the flanks of the linker.

The restriction enzymes used to excise the linker will normally provide sticky ends which may be blunt-ended, either by filling in or by removal of the single strand extension and the blunt-ended upstream and downstream fragments ligated together to afford a desired final construct. The direct ligation of the upstream and downstream fragments will usually only be used when the modification is by deletion in which case the blunt ends of the upstream and downstream fragments correspond respectively with the bases 5' and 3' of the deletion locus. Alternatively the sticky ends may be used to anneal to corresponding sticky ends on an oligonucleotide for insertion between the upstream and downstream fragments.

When the modification is by substitution or insertion it will almost always be necessary to provide the new base or bases via an oligonucleotide insert and indeed it may be convenient (having regard to the relative ease of sticky end ligation as opposed to blunt end ligation) to make deletion modifications by this method. When ligation is via an oligonucleotide insert, the oligonucleotide will contain the desired modified sequence and have sticky ends corresponding with the sticky ends generated on the upstream and downstream fragments by excision of the linker. The oligonucleotide may and, when used for a deletion mutation, will contain sequences corresponding to the starting sequence at the upstream and downstream flanks of the mutation locus in addition to the sticky ends which themselves will also be the same as, or complementary to, sequences at the upstream and downstream flanks of the modification site.

The oligonucleotide may, in a particular embodiment of the invention, be degenerate in one or more positions not involved in the ligation (i.e. not sticky end extensions). This enables a whole series of modifications to be made simultaneously; for instance in an experiment to optimise codon usage, an oligonucleotide degenerate in one or more codon positions could be used, forming a series of final desired constructs which can be used to transform cells and cells expressing the protein in the most advantageous manner can then be selected.

The use of an oligonucleotide in ligating the upstream and downstream fragments has a particular advantage in that it enables selection of convenient termini on these fragments so that the εequence(ε) to be probed for will be unique within the intermediate fragment. It is also contemplated that ligation may be achieved via a series of two or more oligonucleotideε and that, for convenience in further manipulations, the ultimate ligation^" to form the desired final construct may be preceded by insertion and removal of additional linkers according to the invention.

In a particular aspect the present invention provides a process for producing a double-stranded DNA construct having a nucleotide base sequence which differs at a predetermined locus from the base sequence of a starting fragment which process comprises the steps of

(a) digesting and blunt ending a first sample of starting fragments having a protected 5'-terminal so as to produce a population of fragments including at least one upstream fragment;

(b) ligating to the 3'-terminals of fragments produced according to step (a), first precursors for a linker;

(c) digesting and blunt-ending a second sample of starting fragments having a protected 3'-terminal so as to produce a population of fragments including at least one downstream fragment;

(d) ligating to the 5'-terminal of fragments produced according to step (c), second precursors for the linker;

(e) ligating fragments produced according to steps (b) and (d) so as to produce a population of fragments each comprising a linker; (f) obtaining from a population of fragments produced according to step (e) at least one intermediate fragment having the structure

5' ....upstream fragment:linker:downstream fragment....3' ;

by probing for the upstream fragment:linker and linker: downstream fragment junctions;

(g) excising at least the entire linker from an intermediate fragment produced according to step (f); and

(h) ligating together the residues of upstream fragments and downstream fragments produced according to step (g), optionally with insertion of an intervening oligonucleotide, so as to produce the double-stranded DNA construct required.

The process of the invention may include additional cloning steps for instance to increase the amount of material available for use in subsequent steps; such cloning is by conventional methods. Where fragments are cloned, subsequent manipulation may be conducted using fragments excised from cloning vectors or on the cloning vectors still containing the fragments except, in the latter case, where the presence of the cloning vector would interfere with a step in the process.

Usually the linkers will be selected such that their restriction enzyme recognition sites do not also appear elsewhere in the fragments being manipulated. Where this is not possible, or is inconvenient, it is preferred to protect restriction sites which are not to be cut in the present process by modification, for instance by protective methylation using appropriate enzymes or M13 mutagenesis. Protection and deprotection at a later stage is conventional.

In a further aspect the present invention provides double-stranded DNA fragments containing linkers as hereinbefore described.

Intermediate fragments containing linkers as hereinbefore described may be inserted into cloning vectors and produced in bulk and/or stored, and replenished in perpetuity. A bulk supply of a cloned intermediate fragment containing a linker may be used for introduction of any number of substitutions, additions and/or deletions at or near a particular site in the starting fragment and these may all be accomplished by the same mechanism, namely the insertion by forced-cloning of a custom-made, double-stranded oligonucleotide incorporating the sought after modification. Because of this in-built flexibility, whereby substitutions, additions and/or deletions can be generated as optional products of a single intermediate fragment containing a linker at the appropriate site, those wishing to conduct multiple experiments on a particular starting fragment can choose the 3' terminal sequence of the upstream fragment and the 5' terminal sequence of the downstream fragment to suite the selection of probes which they have in stock thus avoiding the need to obtain a complete set of probes as described above. Particularly in this sort of case, probes having a 5^* or 6-base or longer, overlap with the cloned DNA may be chosen to afford additional accuracy in identifying upstream : linker and/or downstream : linker junctions. The variability in selecting the 3' terminal of the upstream fragment and/or the 5' terminal of the downstream fragment also enables the skilled worked to avoid cleavage sites which are resistant to cleavage by a particular restriction enzyme used for excision of the linker.

The invention also provides precursors for linkers as hereinbefore described.

In yet a further aspect the invention provides probes for detecting the linkers as hereinbefore described.

The present invention further provides a kit for DNA modification comprising, separately packaged, precursors for a linker and at least two probes, the probes all having the same common portion of nucleotide base sequence identical to the nucleotide base sequence of the common portion of linker and a variable portion of at least four nucleotide bases in length where each probe has a different variable portion. The variable portion may be either at the 3' flank or the 5' flank of the constant portion. If the variable portion is at the 5' flank of the constant portion, to identify a desired intermediate fragment using single-stranded probes, a probe is selected having a variable portion corresponding to the 3' flank of the plus strand of the upstream fragment and a second probe is selecte having a variable portion corresponding to the 3' flank of the minus strand of the downstream fragment. Double-stranded probes can be selected on a very similar basis as can single and double-stranded probes wherein the variable portion is the 3' flank of the constant portion.

For use with linkers which do not have a common portion, it may be more convenient to use one probe to detect the junction of the upstream fragment and 5' flank of the linker and a second probe to detect the junction of the 3' flank of the linker and the downstream fragment so that, in order to provide a kit for detection of all possible intermediate fragments involving a particular linker, it will be necessary to have two sets of probes each substantially as described above. In one set of probes the constant region would correspond with the 13 nucleotide bases at the 5' flank of the plus strand of the linker and in the second set the constant region would correspond with 13 nucleotide bases at the 3' flank of the plus strand of the linker.

The invention also therefore provides a kit for DNA modification comprising, separately packaged, precursors for a linker and two sets of probes as described above. The kits optionally also contain reagents, buffers, enzymes and substrates for use in DNA modification according to the process of the invention.

The invention will now be described with reference to the Figures of the accompanying Drawings in which:

Fig. 1 shows series of part-linkers according to the invention. In each case the left hand column shows the recognition site (Capitals, A « adenosine, C » cytosine, G - guanosine, T « thymidine) and a suggested sticky end section (lower case, a « adenosine, c ■ cytosine, g = guanosine, t ■= thymidine) together with the cutting site (arrows). "N" indicates any deoxynucleotide and this may be part of the linker or part of the upstream or downstream fragments. The sticky end section may be varied as convenient. The longest possible part-linkers of each series have a sticky end section and the complete recognition and cutting site sections. Shorter part linkers could also be used, these would simply result in upstream or downstream DNA being lost on excision of the linker. The shortest part linkers have a sticky end section and the recognition site section only. The right hand column in each case gives one example of a shorter, or the shortest part-linker in each series. The central column indicates, by conventional abbreviations, the relevant restriction enzyme for each series. Series 18, 19 and 20 have two restriction sites which are indicated by corresponding underlining. Series 14, 15 and 17 have degenerate sites where X = C or G, Y = A or G, Z = T or C. Fig. 2 shows the sequence of the human GSH transferase pi gene. The first underlined section (of 8 bases) is the T.R.E.

The following Examples illustrate the invention but are not intended to limit the scope of the invention in any way.

EXAMPLE 1

To examine the effect upon transcriptional efficiency, of chemically modified bases within cis-acting regulatory sequences upstream of a gene.

12-0-tetradecanoyl phorbol-13-acetate (TPA) is a tumour promoting substance which causes enhanced expression of many genes and especially for the purposes of this particular Example, the human glutathione transferase pi gene (Fig. 2.). This enhanced gene expression is regulated at least in part through a short DNA sequence (TGACTCAG, henceforth called TRE) upstream of the gene (see Fig. 2. residues -66 to -59). This transcriptional regulation can be assessed experimentally by attaching the upstream region of the gene (eg from residue -1000 to +1) to a reporter DNA fragment and, following expression in nuclear extracts, determining the level of the reporter transcript, in this Example a chemically modified base substitution is made within the TRE to examine the effect of the substitution upon transcriptional efficiency. The substitution will be to replace the residue -66 (thymine), which is essential to the functional integrity of the TRE, with 5-hydroxymethyl uracil ("5-hydroxythymine") .

PROTOCOL

The double-stranded Pstl-generated DNA fragment corresponding to residues (approx.) -2200 to +495 of the GSH transferase pi gene was cloned, using EcoRI linkers, in both orientations in plasmid pUC-18 at the ECoRl site of the multicloning linker. These constructs have unique restriction sites for BamHI and Pstl present in the remainder of the multicloning linker since the cloned DNA fragment does not contain a BamHI site and the EcoRI linkers do not regenerate the Pstl sites. The BspMI site is protected by methylation at base +37.

(a) A clone in a given orientation* is digested with BamHI and Pstl then treated with exonuclease III progressively to reduce the size of the cloned DNA from one end [this enzyme 1) yields a processive digestion at a uniform rate starting from the BamHI generated 4-base 5'protuding end and 2) fails to initiate digestion at the Pstl generated 4-base 3'protuding end thereby protecting the vector DN (Henikoff, S. 1984, Gene 2 _, 351-359)]. Samples of the digest are taken at suitable time intervals and the reaction stopped. A sample (or a pool of samples if preferred) is then blunt-ended using Sl-nuclease followed by EcoRI cleavage within the remaining portion of the multicloning linker thus generating a mixture of molecules with one blunt end and one single-stranded extension generated by the EcoRI cleavage.

Each sample or pool of samples (max. 50ug) is loaded onto an HPLC and the DNA duplexes are separated according to size using a Waters Gen-Pak FAX column (or equivalent, and a sodium chloride gradient (Merion e_t al. , 1987. J. Analysis & Purification (Waters) 60-61). Suitable sample volumes are collected and the size range within each is determined by agarose gel electrophoresis. [At this point the fractions can be used for sequencing]. Fractions containing suitable DNA lengths are mixed with a linker (I)

GCTAGCAGGTGCGGCCGCACCTGCTAGC (BspMI recognition sites) CGATCGTCCACGCCGGCGTGGACGATCG (Notl recognition site) ^{(I )}

and ligated using phage T4 DNA ligase.

(b) The whole process of step (a) is repeated for the construct containing the cloned DNA in the alternative orientation.

Each ligation is then cut with Notl and the linker fragments separated from the bulk on an Ultrogel AcA34 column. Suitable fractions containing linker fragments are then mixed and ligated followed by digestion with EcoRI. [The last step is necessary since the molecules initially had a complementary single-stranded EcoRI-generated extension which would have annealed and ligated.] The DNA is then mixed with EcoRI-cut, phosphatased plasmid vector, ligated and transformed into a suitable host to generate a clone library containing inserts conforming to the general structure:

-2200...GCTAGCAGGTGCGGCCGCACCTGCTAGC...+495 (BspMI sites) CGATCGTCCACGCCGGCGTGGACGATCG (Notl site)

i.e. the clone library includes a population of cloned molecules containing a linker in a position corresponding to deletions which are localised to an area determined by the original choice of HPLC eluates.

The precise construct desired including the sequence

-74 .

,GCCGGCTAGCAGGTGCGGCCGCACCTGCTAGCCACT.. .CGGOCGATCGTCCACGCCGGCGTGGACGATCGGTGA..

-59

(where the boxed portion arises from the linker and the other residues arise from the cloned gene numbered as indicated) is found using probes

(a) 5' CGGCCGGCTAGCAGGTG 3' for the upstream : linker overlap and (b) 5' AGTGGCTAGCAGGTGCG 3' for the downstream : linker overlap both radiolabelled by kinasing and by the oligonucleotide screening method of Wood e_t al_. (1985, Proc. Natl. Acad. Sci. USA. S2__, 1585 to 1588) whereby, under very stringent conditions, only clones containing sequences which are complementary to the full-length of a probe will be selected. Usually any clone can be identified according to the sequence of four bases immediately adjacent to the flanks of the linker sequence. This example includes the unusual case whereby probe (a) overlaps the cloned DNA by 6 nucleotides since the regulatory region upstream of eukaryotic genes is G/C rich and so has a high probability of more than one construct being selected.

The bacterium containing the desired construct s grown up and the plasmid purified and the position of the linker checked by sequencing. Subsequent digestion with BspMI will remove the linker leaving sticky ends which will allow the forced ligation of the short synthetic duplex oligonucleotide

GCCGGCGCCGTGACTCAG

CGCGGCACTGAGTCGTGA *

Where T is 5-hydroxymethyluracil.

This DNA prepartion is then used in nuclear extracts in order to assay the effect of the base modification upon transcription.

EXAMPLE 2

In an alternative, to Example 1, the part linker (II)

GGCCGCACCTGCTAGC (BspMI recognition site) (II) CGTGGACGATCG (Notl site forms on ligation below) is used in place of the protected part linker (I)

EXAMPLE 3

In Examples 1 and 2 the constructs 2200—linker—2200 and 494-linker—495 can also form. In Example 3 a linker (III) is used which contains a Sfil site (GGCCNNNNNGGCC in which N is any deoxynucleotide and the central 3 residues are non-palindromic) in place of the Notl site.

(Ill) eg GCTAGCAGGTGGCCNGGG/NGGCCACCTGCTAGC (BspMI sites) CGATCGTCCACCGGN/CCCNGGCCTGGACGATCG (Sfil site)

/ « cleavage site)

This reduces the number of clones needed to be screened to find a precise construct. To insert such a linker in the construct to the procedure of Example 2 is followed except that in step (a) a part linker (Ilia) is used and in step (b) a part linker (Illb) is used: GCTAGCAGGTGGCCNGGG (BspMI site) (Ilia) CGATCGTCCACCGGN (Sfil site)

NGGCCACCTGCTAGC (BspMI site) (Hlb) CCCNCCGGTGGACGATCG (Sill site)

EXAMPLE 4

To optimise the efficiency of expression of a protein in any system such as higher eukaryote cells lines, bacteria, yeast or transgenics by varying the nucleotide sequence about the AUG (initiator) codon.

The immediate nucleotide sequence about the initiator codon is known to have a large effect upon the precision and efficiency of translation during gene expression. However an optimal sequence for expression in mammalian cell lines may not be optimal for expression in other systems such as Escherichia coli or yeast. For example, a feature influencing the precision of translation in higher eukaryotes only is the presence of a purine residue three nucleotides upstream of the initiator AUG (Kozak, M. (1987), Nucleic Acids Res., 15, 8125-8148).

This feature appears to ensure the utilisation of the correct AUG codon in those 10% (approximately) of eukaryotic mRNAs which have additional AUG codons upstream of the known start of protein synthesis. In addition, to optimise translational efficiency it is usually necessary to alter the codon usage of the first few codons of a cloned sequence by trial and error.

The insertion of a linker in place of the bases immediately upstream of the initiator codon and some (e.g. six) of the first six codons allows its replacement by a duplex oligonucleotide which is degenerate at suitable residues. Following ligation, the degenerate constructs are transformed into the appropriate expression system and selection made on a phenotypic basis, eg yield.

Claims

1. A process for producing a double-stranded DNA construct of modified sequence which comprises constructing an intermediate fragment containing a linker at the site for modification, excising at least the entire linker and ligating the remaining fragments to produce the desired construct, wherein the linker is a double-stranded oligonucleotide which includes recognition sites for restriction endonucleases located within the linker such that by treatment of a fragment containing the linker with appropriate endonucleases capable of cutting outside their recognition sequences, at least the entire linker may be excised from the intermediate fragment.

2. A process according to claim 1 wherein the linker contains a palindromic recognition site for a restriction endonuclease capable of cutting outside the recognition site and located so as to permit excision of at least the entire linker from the intermediate fragment.

3. A process according to claim 1 wherein the linker contains two non-palindromic recognition sites for restriction endonucleases capable of cutting outside their recognition sites and located and oriented so as to permit excision of at least the entire linker from the intermediate fragment.

4. A process according to claim 2 or claim 3 wherein the linker further comprises the recognition site for a restriction endonuclease located such that the cutting site of the restriction endonuclease is within the linker sequence.

5. A process according to any one of claims 1 to 4 comprising the steps of: (a) producing a population of digestion products of a starting fragment containing an upstream fragment

(b) producing a population of digestion products of a starting fragment containing a downstream fragment and

(c) ligating the upstream and downstream fragments with a linker to form an intermediate fragment having the structure

5' ...upstream fragment:linker:downstream fragment...3'

6. A process according to claim 5 comprising the steps of probing for the upstream fragmentrlinker junction using a first probe and for the linker:downstream fragment junction using a second probe.

7. A process according to claim 6 wherein the linker comprises common sequences at the 5' and 3' flanks and wherein the first and second probes contain identical common sequences.

8. A process according to claim 5 or claim 6 wherein the probing is conducted under conditions such that only fragments which hybridise with the full length of the probes will be selected.

9. A process according to any preceding claim for producing a double-stranded DNA construct having a nucleotide base sequence which differs at a predetermined locus from the base sequence of a starting fragment which process comprises the steps of

(b) ligating to the 3'-terminals of fragments produced according to step (a), first precursors for a linker; (c) digesting and blunt-ending a second sample of starting fragments having a protected 3'-terminal so as to produce a population of fragments including at least one downstream fragment;

(e) ligating fragments produced according to steps (b) and (d) so as to produce a population of fragments each comprising a linker;

(f) obtaining from a population of fragments produced according to step (e) at least one intermediate fragment having the structure

5... pstream fragmentrlinker:downstream fragment...3' ; by probing for the upstream fragment:linker and linker: downstream fragment junctions;

10. A linker which is a double-stranded oligonucleotide containing a palindromic recognition site for a restriction endonuclease capable of cutting outside its recognition site located such that when the linker is part of a DNA fragment, at least the entire linker may be excised from that fragment.

11. A linker which is a double-stranded oligonucleotide containing two non-palindromic recognition sites for restriction endonucleases capable of cutting outside their recognition sites located and oriented such that, when the linker is part of a DNA fragment, at least the entire linker may be excised from that fragment.

12. A linker according to claim 10 or claim 11 further comprising a recognition site for a restriction endonuclease located such that the linker may be cleaved by the endonuclease.

13. A precursor for a linker according to any one of claims 10 to 12 comprising a restriction endonuclease recognition site located for cleaving the precursor to form a part-linker containing at least a part of a recognition site for a restriction endonuclease capable of cutting outside its recognition site.

14. A part-linker which is a precursor for a linker according to any one of claims 10 to 12 comprising at least a part of a recognition site for a restriction endonuclease capable of cutting outside its recognition site and having a sticky end suitable for ligation to a second part-linker so as to form a linker.

15. A kit for DNA modification according to any one of claims 1 to 9 comprising separately packaged, precursors for a linker having a common nucleotide sequence and at least two probes, the probes all having the same common portion of nucleotide base sequence identical to a common portion of the linker.

16. A kit for DNA modification according to any one of claims 1 to 9 comprising separately packaged, precursors for a linker, a set of probes for detecting upstream fragmentrlinker junctions and a set of probes for detecting linkerrdownstream fragment junctions.

17. A set of probes for DNA modification according to any one of claims 1 to 9 comprising a plurality of oligonucleotide probes each having a different variable portion and an identical common portion of nucleotide sequences.

18. A set according to claim 17 wherein the common portion is at least 13 nucleotides in length and the variable portion is 4 nucleotides in length.

19. A set of 256 probes according to claim 18 wherein each of the probes has a different variable portion.

20. A set of double-stranded probes according to any one of claims 17 to 19.