METHOD OF DNA SYNTHESIS

Field of the Invention The present invention relates to an in vitro cell-free process for production of deoxyribonucleotides (DNAs) comprising at least one hairpin, corresponding DNA products and uses thereof, and oligonucleotides and kits useful in the process of the invention.

Background of the Invention

In vitro, cell-free processes for amplification of closed linear DNA from a starting template have been described in WO2010/086626 and WO2012/017210. It would be desirable to provide further processes which can synthesise closed linear DNA, and which also allow for synthesis of other types of DNA molecules comprising one or more hairpins. Summary of the Invention

The invention provides a process for synthesis of DNA molecules comprising one or more hairpins which does not require use of any microbiological steps for provision of a starting template. The process also allows for synthesis of circular single-stranded DNA or double- stranded DNA comprising one or more hairpins. These products include DNA molecules which the inventors believe to have structures not previously described in the art, and which have advantages for a wide range of applications.

According to the present invention, synthesis of DNA molecules comprising hairpins is carried out in an in vitro cell-free process starting from an oligonucleotide which can be immobilised to a solid support during all or part of the process. The oligonucleotide is extended enzymatically to synthesise a desired DNA sequence, using short template oligonucleotides which can be synthesised chemically, thus avoiding use of large starting templates encoding the entirety of the desired sequence which would typically need to be propagated in bacteria. Once a desired DNA sequence is synthesised, it can be released from the solid support in a single-stranded or double-stranded form comprising one or more hairpins.

Advantageously, sequences used to provide hairpins in the DNA molecules of the invention also provide a means for enzymatic release of the synthesised DNA from the solid support on completion of DNA synthesis. An oligonucleotide is immobilised on the solid support and extended to incorporate a desired DNA sequence to create a first DNA strand which further comprises a first portion of a protelomerase target sequence proximal to the solid support. A second portion of the target sequence for the protelomerase, complementary to the first portion thereof is also introduced in the extended first strand or on a complementary second strand. The first and second portions of the target sequence for the protelomerase are then used to recreate a complete protelomerase target sequence proximal to the solid support, such that contacting with a protelomerase can then release the synthesised DNA from the solid support in conjunction with generating a hairpin in the DNA molecule released from the solid support.

In further aspects, the invention provides for addition of a second closed end hairpin to a produced DNA molecule by creation of a complete protelomerase target sequence or other sequence capable of forming a hairpin in the distal, extended region of the synthesised DNA strand. Such a sequence capable of forming a hairpin can include neighbouring complementary sequences or two complementary sequences separated by a sequence which is non- complementary.

In further aspects, the invention provides for the addition of a third, fourth, fifth or further closed end hairpin to a produced single stranded DNA molecule by creation of a complete protelomerase target sequence or other sequence capable of forming a hairpin in the synthesised DNA strand.

The DNA molecules synthesised in accordance with the invention may be used for various applications, including medicinal and diagnostic uses, and also as a starting template for further processes of DNA amplification.

In more detail, the invention provides:

In a first aspect, an in vitro cell-free process for production of a deoxyribonucleic acid

(DNA) which comprises a desired DNA sequence, said process comprising:

(a) contacting an oligonucleotide immobilised on a solid support with a series of template oligonucleotides which overlap in sequence, in the presence of at least one DNA polymerase under conditions promoting template-dependent extension of said immobilised oligonucleotide to produce a first DNA strand which comprises the desired DNA sequence and further comprises a first portion of the protelomerase target sequence proximal to the solid support;

(b) introducing a DNA sequence comprising a second portion of the protelomerase target sequence of (a) in the distal part of said first DNA strand produced in (a), or on a second or further strand, such that said first and second portions of the protelomerase target sequence thereby create a complete target sequence for the protelomerase of (a) proximal to said solid support; and

(c) contacting said complete protelomerase target sequence proximal to said solid support with a protelomerase under conditions promoting cleavage and rejoining of said target sequence, to thereby release the produced DNA from immobilisation. In one embodiment, the first and second portions of the protelomerase target sequence are complementary in sequence.

In another aspect, the invention relates to a solid support comprising an immobilised oligonucleotide comprising a first portion of a target sequence for a protelomerase.

In a further aspect, the invention relates to a kit comprising an oligonucleotide comprising a first portion of a target sequence for a protelomerase, a series of template oligonucleotides, and optionally instructions for use in a process of the invention.

In a fourth aspect, the invention relates to a single-stranded circular DNA comprising one or more hairpins, at least one of which comprises a portion of a target sequence for a

protelomerase.

In a fifth aspect, the invention relates to a linear covalently closed double-stranded DNA comprising a first hairpin comprising a portion of a target sequence for a first protelomerase and a second hairpin which has a sequence which is not complementary to the first hairpin. In one embodiment, this is achieved by the second hairpin comprising a portion of a target sequence for a second protelomerase, wherein said first and second protelomerases are different. In an alternative embodiment, this is achieved by the second hairpin comprising a sequence capable of forming a hairpin. Such sequence may include neighbouring or separated complementary sequences, which are capable of annealing to each other.

In a sixth aspect, the invention relates to an in vitro cell-free process for amplification of DNA, comprising contacting a single-stranded circular DNA template comprising a hairpin comprising a portion of a target sequence for a protelomerase, with at least one DNA polymerase in the presence of one or more primers under conditions promoting amplification of said template.

In a seventh aspect, the present invention relates to an in vitro cell-free process for production of a linear covalently closed deoxyribonucleic acid (DNA) comprising:

(a) contacting a linear covalently closed double-stranded DNA comprising a first hairpin comprising a portion of a target sequence for a first protelomerase and a second hairpin comprising a portion of a target sequence for a second protelomerase with a DNA polymerase under conditions promoting DNA amplification, and

(b) contacting the amplified DNA with said first and second protelomerases under conditions promoting production of linear covalently closed DNA,

wherein said first and second protelomerases are different.

In an eighth aspect, the present invention relates to a single-stranded circular DNA comprising at least one hairpin comprising a portion of a target sequence for a protelomerase, for use in therapy or diagnosis, particularly for use in a method for treatment of the human or animal body, or in a diagnostic method practised on the human or animal body.

In a ninth aspect, the invention relates to a linear covalently closed double-stranded DNA comprising at least one hairpin comprising a portion of a target sequence for a first protelomerase and wherein the sequence of the second hairpin is not complementary to the sequence of the first hairpin, for use in therapy or diagnosis, particularly for use in a method for treatment of the human or animal body, or in a diagnostic method practised on the human or animal body. In one embodiment the second hairpin comprises a portion of a target sequence for a second protelomerase, wherein said first and second protelomerases are different. In an alternative embodiment, the second hairpin is provided by neighbouring or separated complementary sequences.

In a tenth aspect, the invention relates to a method of treatment of the human or animal body, comprising administering a therapeutically effective amount of a single-stranded circular DNA comprising a hairpin comprising a portion of a target sequence for a protelomerase to a human or animal in need thereof.

In an eleventh aspect, the invention relates to a method of treatment of the human or animal body, comprising administering a therapeutically effective amount of a linear covalently closed double-stranded DNA comprising at least one hairpin comprising a portion of a target sequence for a first protelomerase and wherein the sequence of the second hairpin is not complementary to the sequence of the first hairpin, to a human or animal in need thereof. In one embodiment the second hairpin comprises a portion of a target sequence for a second protelomerase, wherein said first and second protelomerases are different. In an alternative embodiment, the second hairpin is provided by neighbouring or separated complementary sequences.

Optional features are defined in the dependent claims. Further advantages are described below.

Brief Description of Figures

Figures la to lc: Depicts schematics of examples of various single stranded

oligonucleotides which can act as starting and/or terminating primers and templates. Each is shown in linear format or with the hairpin formed in the portion of the protelomerase target sequence (101 or 107), and has free 3' (102) and 5' (103) ends.

Figure la depicts an oligonucleotide (100) that can be used as a starting primer or as a terminal oligonucleotide template and primer. The oligonucleotide contains a portion of a protelomerase target sequence (101), flanked by a 3' sequence (104) and a 5' sequence (105) that do not form part of the sequence of the protelomerase target sequence. The entire sequence of the oligonucleotide (106) may include a region designed to be complementary to a template oligonucleotide or synthesized strand.

Figure lb depicts a terminal oligonucleotide primer (110). In this example, this contains a portion of a protelomerase target sequence (107) which comprises a different and non- complementary sequence to the one shown in Figure 1(a) as (101). This sequence is flanked by a 3' sequence (109) and a 5' sequence (111) which do not form part of the protelomerase target sequence. The 3' flanking sequence also contains a region (108) which has a sequence which is non-complementary to the synthesized strand, whilst the remaining sequence of the

oligonucleotide primer (110) shown as 112 may be complementary in sequence to a synthesized strand.

Figure lc depicts an oligonucleotide (120) that can act as a starting primer or template, or a terminal oligonucleotide or template. It comprises a hairpin structure (220) that may comprise a portion of a protelomerase target sequence, flanked by a 3' sequence (114) and a 5' sequence (113), which flanking sequences do not form part of the protelomerase target sequence, if such is present. A region of the 3' flanking sequence may comprise a region (115) which is not complementary to the 5' flanking sequence and which is complementary to either a template oligonucleotide or a synthesized strand.

Figure 2 (figures 2a and 2b) depicts various ways of immobilising an oligonucleotide to a solid support.

Figure 2a depicts the use of a spacer molecule (122) which is linked to the solid support (121) and the 5' end (103) of the oligonucleotide (100) via a suitable chemical linker (123 and 124, respectively).

Figure 2b depicts an alternative arrangement where the spacer molecule (122) is linked via a chemical linker to the oligonucleotide (120) 5' flanking sequence (113).

Figure 3 (figures 3a to 3j) depicts an example of some of the steps involved in one method of the invention, in which an oligonucleotide (100a) is immobilised to a solid support (121) via a spacer molecule (122) and a succession of template oligonucleotides (130) are added, together with DNA polymerase, in order to extend the oligonucleotide. The template oligonucleotide is removed after extension is complete and a further template oligonucleotide is added until the desired sequence has been achieved (133). At this point an oligonucleotide (100b) is added which acts as both a template and a terminal primer, and a complementary second strand (136) of DNA is synthesized. Once a double-stranded sequence has been synthesized, the product (140) is released from the solid support using protelomerase (not shown) and leaves a by-product on the solid support (138).

Figure 3a shows the oligonucleotide (100a) comprising a first portion of a protelomerase target sequence (101a) immobilised via a spacer (122) to a solid support (121), in the presence of a template oligonucleotide (130).

Figure 3b shows the template oligonucleotide (130) bound to the immobilised oligonucleotide (100a) via a region of complementary sequence (131). The 5' end of the template oligonucleotide overhangs the 3' end of the immobilised oligonucleotide and can act as a template for extension of the 3' end (102a) of the immobilised oligonucleotide (100a).

Figure 3c shows the extension of the immobilised oligonucleotide (100a) at the 3' end

(102a) once DNA polymerase (not shown) has catalysed extension with a sequence

complementary to the sequence of the template oligonucleotide (130). The oligonucleotide strand, or first strand has an extended sequence (132), which is complementary to a region of the template oligonucleotide (130).

Figure 3d shows the extended sequence (132) once the template oligonucleotide has been removed.

Figure 3e shows the extension of the first strand (133) following a repeated application, extension (using DNA polymerase) and removal of template oligonucleotide. The synthesised sequence of the first strand (133) is complementary to the template oligonucleotides used to construct it, in the order that they are used.

Figure 3f shows the presence of an oligonucleotide (100b), comprising a second portion of a protelomerase target sequence (101b), which acts as a template and a terminal primer.

Figure 3g shows the binding of the oligonucleotide (100b) to the first strand (133) via a region of complementary sequence (134). Both 3' termini (102b and 141) are available for extension by DNA polymerase (not shown).

Figure 3h shows the complementary sequences synthesized by DNA polymerase (not shown) to the terminal primer/template, complementary sequences are shown here in dashed lines. The complementary sequence to the oligonucleotide (100b) is shown as a dashed line (135) and the complementary sequence to the first strand (133) is also shown as a dashed line (136). The complementary sequence is also known as the complementary second strand (136).

Figure 3i shows the first strand (133) and complementary second strand (136). The first and second portions of the protelomerase target sequence (101a and 101b) have been used as templates by the DNA polymerase to form two complete protelomerase target sequences (137a and 137b). Figure 3j shows the result of using a protelomerase on the first strand (133) and complementary second strand (136) of figure 3i. The complete protelomerase target sites (137a and b) are cleaved and ligated, forming a released product (140) which shown here is a closed linear DNA, leaving a by-product linked to the solid support (138) and a further free by-product (139), which is single-stranded DNA with an internal hairpin and free 3' and 5' ends.

Figure 4 (figures 4a to 4f) depicts a further example of a method of the invention for producing covalently closed double stranded or single stranded DNA, via key steps a to f. Other steps are not shown. Double stranded DNA will be produced if the created hairpin structure does not include a loop of single-stranded DNA as shown in Figure 4d. However, if the hairpin (185) in Figure 4d includes a longer intervening single stranded sequence, this may be the major section of the molecule and a minimal amount of double stranded sequence may be present, as shown in Figure 4h

Figure 4a shows an immobilised oligonucleotide (100a) including a first portion of a protelomerase target sequence (101a), immobilised via a spacer molecule (122) attached to a solid-support (121). The immobilised oligonucleotide has been extended via template-dependent extension, to produce a first strand (133). Shown is a template oligonucleotide (130) bound to the first strand (133) near the 3' end (141) including a section of sequence (181). An exemplary sequence (183) of the template oligonucleotide is shown. The template oligonucleotide binds at the 3' end of the first strand (141). The remaining part of the template oligonucleotide overhangs the 3' end of the first strand.

Figure 4b depicts the same structure once DNA polymerase has catalysed the extension of the 3' end (141) of the first strand (133) using the template oligonucleotide (130) to create a complementary sequence (dashed line). This complementary sequence includes a sequence (184) complementary to the exemplary sequence (183) in the template oligonucleotide (130).

Figure 4c shows the same structure as Figure 4b with the template oligonucleotide (130) removed. An earlier template oligonucleotide introduced a sequence (181) in the first strand (133) which is complementary to the sequence (184) introduced by the template oligonucleotide of Figure 4b. The first strand thus includes self-complementary sequences. The distal 3' end of the first strand (133) may loop back and the complementary sequences (181 and 184) may anneal to form a duplex. As shown here, there is no intervening single stranded sequence between the complementary sequences (181 and 184).

Figure 4d depicts the structure of Figure 4c with the two complementary sequences annealed (181 and 184). Figure 4e shows the result of extending the 3' end (141) of the first strand shown in Figure 4d using DNA polymerase. The segment of the first strand (133) between the solid support and the hairpin acts as a template for the second segment of the first strand (186 - dashed line), until the DNA reaches the spacer molecule (122). A second portion of a protelomerase target sequence is synthesised (101b) using the first portion as a template (101a), resulting in the formation of an entire protelomerase target sequence (137).

Figure 4f demonstrates the results of adding a protelomerase to the structure of Figure 4e. A closed linear DNA product (187) is released and a by-product (188) is left immobilised to the solid support. Note that the closed linear DNA is closed at one end with a portion of a protelomerase target sequence at one end (101) and with a template-derived hairpin at the other (189).

Figure 4g depicts an alternative embodiment, wherein the segment of sequence (181) is not present at the 3' end of the first strand (133), and the terminal template oligonucleotide (130) does not anneal thereto. The template oligonucleotide binds to a section of complementary sequence (182).

Figure 4h depicts the structure obtained by extension of the 3' end of the first strand of Figure 4g. An intervening sequence (185) is formed between the complementary sequences (181 and 184), which is looped out as a single strand of DNA, of any particular length.

Figure 5 (Figures 5a to 5e) depicts a further example of the method of the invention. An oligonucleotide (120a) is immobilised and the 3' end (102) extended using rounds of template oligonucleotide dependent extension to form a first strand (133). A final template oligonucleotide (120b) comprising a hairpin (225 is added which acts as a template to introduce sequences in the distal 3' end of the first strand which are capable of forming a hairpin due to self - complementary internal sequences. The synthesis of the complementary second segment of the first strand (211) uses the first segment of the first strand as a template. Once the double stranded sequence is complete, protelomerase (not shown) acts at the complete protelomerase target site (137) to cleave the sequence and re-join the created free ends, releasing a product (212) and a by-product (143).

Figure 5a shows the extended immobilised oligonucleotide (120a) including a first portion of a protelomerase target sequence (220a) forming a first strand (133), to which is bound an oligonucleotide (120b) including a hairpin structure (225), which in this example is not a portion of a protelomerase target sequence.

Figure 5b shows the structure of Figure 5a with the template oligonucleotide (120b) in unfolded, single-stranded form, and the DNA polymerase (not shown) has catalysed extension of the 3' end of the first strand (133), using the template (120b), resulting in synthesis of the sequence shown as a dashed line (210).

Figure 5c shows the structure of Figure 5b with the template oligonucleotide (120b) removed, and the synthesized sequence (210) forming a hairpin due to internal self- complementary sequences (not shown). The 3' end (141) of the extended first strand (133) is thus available for extension.

Figure 5d shows the structure of Figure 5c once DNA polymerase (not shown) has catalysed the extension of the 3' end (141), using the first segment of the first strand (133) as a template, to produce a complementary second segment of the first strand (211). The second segment of the first strand includes a portion of a protelomerase target sequence (220b), thus forming an entire protelomerase target sequence (137) proximal to the solid support.

Figure 5e depicts the results of adding protelomerase (not shown) to the structure of Figure 5d. A product (212), in this case a closed linear DNA is formed, with one end closed with a portion of a protelomerase target sequence (220d) and a hairpin structure at the other (225), together with a by-products (143).

Figure 6 (figures 6a to 6e) depicts the synthesis of a covalently closed single stranded DNA molecule, via key steps a - e. Other steps are not shown.

Figure 6a shows an immobilised oligonucleotide (100a) including a first portion of a protelomerase target sequence (101a), immobilised via a spacer molecule (122) attached to a solid-support (121). The immobilised oligonucleotide has been extended via rounds of template- dependent extension, to produce a first strand (133). Shown is a second oligonucleotide (100b) including a second portion of a protelomerase target sequence (101b) bound to the first strand (133) near the 3' end (141). An exemplary sequence in the 3' and 5' flanking regions (104a/b and 105a/b) of the portions of the protelomerase target sequence (101a/ b) is shown for each oligonucleotide (lOOa/b).

Figure 6b depicts the structure of figure 6a after DNA polymerase has catalysed extension of the 3' end (141), to produce a sequence complementary (135 - dashed line) to the

oligonucleotide template (100b). A complete protelomerase target sequence (137) is formed. The 3' end (102b) of the template oligonucleotide (100b) will also be extended by the DNA polymerase unless it is modified to prevent extension or it includes a segment of sequence that is not complementary to the first strand, such as the structure shown in Figure lb.

Figure 6c depicts the first strand of the oligonucleotide (133) as shown in Figure 6b, once the oligonucleotide template (100b) has been removed. The arrow depicts the folding back of the distal 3' end (141) of the first strand (133), bringing the 3' end into close proximity to the solid support (121).

Figure 6d depicts the annealing of the complementary flanking sequences (104a and 105a) at the proximal end of the first strand (133) with the flanking sequences (151 and 152) at the distal end of the first strand (133) to form a double stranded DNA duplex and a complete protelomerase target sequence (137). The distal portion of the protelomerase target sequence forms the second portion of the protelomerase target sequence. Since the sequence of the remaining part of the first strand (133) is not complementary, it remains in single strand form as a looped structure (153).

Figure 6e depicts the result of applying protelomerase (not shown) to the structure of

Figure 6d. The entire protelomerase target sequence is cleaved and the free ends joined, releasing a covalently closed single stranded DNA (155) with a single hairpin (225) and a byproduct (154) on the solid support.

Figure 7 (figures 7a to 7e) depicts an alternative method of producing covalently closed single stranded DNA via key steps a- e. Other steps are not shown.

Figure 7a shows an immobilised oligonucleotide (100a) comprising a first portion of a protelomerase target sequence (101a), immobilised via a spacer molecule (122) attached to a solid-support (121). The immobilised oligonucleotide has been extended via template-dependent extension, to produce a first strand (133). Shown is a template oligonucleotide (130) bound to the first strand (133) near the 3' end (141). An exemplary sequence in the template

oligonucleotide is shown (161). The 3' (160) portion of the template oligonucleotide anneals to the complementary sequence in the first strand (133) and a section of single stranded template (164) overhangs the 3' end of the first strand, including the exemplary sequence (161).

Figure 7b depicts the structure of figure 7a after DNA polymerase has catalysed extension of the 3' end (141) of the first strand (133), to produce a sequence complementary (dashed line) to the oligonucleotide template (130), including a complementary sequence (162) to the exemplary sequence (161)

Figure 7c depicts the first strand of the oligonucleotide (133) as shown in Figure 7b, once the template oligonucleotide has been removed. The arrow depicts the folding back of the 3' end (141) of the first strand (133), bringing the distal 3' end into close proximity to the solid support (121) and the proximal end of the immobilised oligonucleotide (100a).

Figure 7d depicts the annealing of the complementary sequences in the 5' flanking region (104a) of the first portion of the protelomerase target sequence (101a) and the sequence (162) on the distal end of the first strand (133), forming a duplex. Since the sequence of the remaining part of the first strand (133) is not complementary, it remains in single strand form as a looped structure (153).

Figure 7e depicts the structure of figure 7d once DNA polymerase has catalysed extension of the 3' end (141) of the first strand (133) to produce a complementary sequence (166 - dashed lines) using the oligonucleotide (100a) as a template, to produce a second portion of a protelomerase target sequence (101b). An entire protelomerase structure is constructed (137). Note that this structure is identical to that of Figure 6d, and that the structure of the products produced when protelomerase is applied will be the same as shown in figure 6e.

Figure 8 (Figures 8a to 8f) depicts an example of a method of the invention for producing covalently closed single stranded DNA, via key steps a to f. Other steps are not shown.

Figure 8a shows an immobilised oligonucleotide (100a) including a first portion of a

protelomerase target sequence (101a), immobilised via a spacer molecule (122) attached to a solid-support (121). The immobilised oligonucleotide has been extended via template-dependent extension, to produce a first strand (133). Shown is a template oligonucleotide (110) bound to the first strand (133) near the 3' end (141). This includes a first portion of a protelomerase target sequence (107a) which is not complementary to the portion in the immobilised oligonucleotide (101a). Only the flanking sequence (109) of the template oligonucleotide anneals to the complementary sequence in the first strand (133), the remaining part of the 3' section of the template oligonucleotide (108) is not complementary to the first strand (133) and does not anneal. A section of single stranded template remains, including the portion of the protelomerase target sequence (107a) and the 5' flanking sequence (111).

Figure 8b depicts the structure of Figure 8a after DNA polymerase has catalysed extension of the 3' end (141) of the first strand (133), to produce a sequence complementary (171 - dashed line) to the template oligonucleotide (110). A second portion of a protelomerase target sequence (107b) is synthesised, creating an entire protelomerase target sequence (200).

Figure 8c shows the result of applying protelomerase specific for the entire

protelomerase target sequence (200) created in Figure 8b. A by-product (172) is released which is a single stranded piece of DNA that may form a hairpin structure. The first strand (133) now includes a hairpin structure (230). The 3' end of the DNA strand is now the non-complementary sequence section (108) from the oligonucleotide (110), and this forms the start of the non- complementary second segment of the extended first strand. It will be understood that the cognate protelomerase can be applied at any suitable time.

Figure 8d depicts the result of extending the 3' end (141) of the second segment of the extended first strand (173) using template dependent techniques as described previously. Templates are designed such that the terminal sequences (174) of the second segment of the extended first strand (173) are complementary to the immobilised oligonucleotide (100a), and include a second portion of a protelomerase target sequence (101b), a 3' flanking region (175) and a 5' flanking region (176) which are complementary to the immobilised oligonucleotide (100a).

Figure 8e shows the annealing of the sequences (175, 176 and 101b) in the second segment of the extended first strand (173) and sequences (104a, 105a and 101a) in the first segment of the first strand (133). An entire protelomerase target sequence is created (137) as the first portion (101a) and second portion (101b) are juxtaposed. The nucleotide sequences shown are exemplary only.

Figure 8f depicts the result of applying a protelomerase specific for the complete protelomerase target sequence (137). A single stranded circular structure (177) is released. This circular structure may form hairpin structures at two points (179 and 180) around the circle, but may also be present without these folds (not shown). The two strands (133 and 173) separated by the hairpin structures are non-complementary and do not anneal to form a double stranded structure or duplex. A by-product (178) is left immobilised to the solid support.

Figure 9 depicts an exemplary structure of a single-stranded DNA (240) which can be synthesised according to the method of the present invention. This structure is a single strand of DNA, including single stranded segments (241 to 246) separated by hairpin structures (231 to 236). The hairpins can be introduced according to any method of the present invention.

Figure 10 depicts an immobilised oligonucleotide according to one aspect of the invention. The first portion of a protelomerase target sequence (shown here is one strand of the protelomerase target sequence for protelomerase TelN - SEQ. ID No. 17) (101a), flanked by a 3' sequence (104a) and 5' sequence (105a). The sequences of the flanking sequence are irrelevant and shown as X and X' to distinguish them from the portion of the protelomerase target sequence (101a). The 5' end of the 5' flanking sequence (105a) is linked to a spacer molecule (122) immobilised to a solid support (121). The base pairing between the residues forming the portion of the protelomerase target sequence (101a) is shown, although it is postulated that the base pairing at the tip of the hairpin may be disrupted due to structural distortion.

Figure 11a depicts the structure of Figure 10 once template-dependent extension has taken place. The first strand (133) has been extended and the second complementary strand (136) has been synthesised. The extension of the second strand (136) includes using the first portion of the protelomerase target sequence (101a) and the flanking regions (104a and 105a) as a template. This results in the synthesis of complementary sequence in the second strand (136) which includes the 3' and 5' flanking regions (104b and 105b) and a second portion of the protelomerase target sequence (101b). Thus, the entire protelomerase target sequence (137) is formed via the action of DNA polymerase using the immobilised oligonucleotide (100a) as a template.

Figure lib shows the sequence of Figure 11a in a linear format, without the hairpin structures. It can be seen that the first portion of the protelomerase target sequence (101a) has a sequence which is complementary to the second portion of the protelomerase target sequence (101b). At the centre (in this example) of the protelomerase target sequence (190) is the site at which the protelomerase will cleave the sequence, which is in the centre of the telO sequence (201). The complete protelomerase target sequence (137) is composed of Tel L for the enzyme TelN. The sequence forming TelR is shown (203) and TelL (202)

Figure 11c shows what happens to the sequences of Figure lla/llb once protelomerase catalyses the reaction. The sequence is cleaved at the point indicated (190) and the each cleaved ends are re-ligated with the opposing strand to form two separate hairpin structures. In this instance, a by-product (191) is left annealed to the solid support and a product (192) is released, with a closed end, due to the action of the protelomerase.

Figure 12 shows a gel photograph of a 2% agarose gel showing the results from Example 1. The photograph both includes a molecular weight scale on the left hand side (lane 1) and an indication of the product run in the lanes at the bottom. TelN= product of rolling circle amplification (RCA) digested by TelN, TelN/Exolll= RCA product digested by TelN and Exolll

Figure 13 shows the whole native target sequences for a selection of protelomerase enzymes, showing the sequences of both strands of the complementary DNA. Shown are the target sequences: the sequence of SEQ ID NO: 15 (E.coli N15 TelN protelomerase), the sequence of SEQ ID NO: 16 (Klebsiella phage Phi K02 protelomerase), the sequence of SEQ ID NO: 17 (Yersinia phage PY54 protelomerase), the sequence of SEQ ID NO: 18 (Vibrio phage VP882 protelomerase), the sequence of SEQ ID NO: 19 (Borrelia burgdorferi protelomerase) the sequence of SEQ ID NO: 20 (Agrobacterium tumefaciens Te\A protelomerase) and the sequence of SEQ ID NO: 21 (Vibrio parahaemolyticus plasmid Vp58.5 protelomerase). Where the minimum sequence length requirement for the cognate protelomerase is known, this has been indicated by shading the sequence grey, although the enzyme may accept some variation in sequence within this core target sequence. Nucleotides represented in bold and underlined indicate imperfections in the palindrome sequence. The vertical line through the sequences represents the centre of the perfect inverted sequence and the point at which the protelomerase cleaves and joins the target sequence. Figure 14 is a photograph of a gel produced in Example 2. The gel has 5 marked lanes, dsDNA is full length linear DNA, TelN, is the dsDNA treated with TelN, VP58.5 is DNA alternatively treated with VP58.5, whilst both indicates both protelomerase enzymes were applied. T5 represents the sample after addition of the T5 exonuclease. The gel thus shows the progression of the reaction from construction of the dsDNA via template primer extension to cleavage with one or two protelomerases, and final clean up using T5 exonuclease to remove any unwanted open fragments. The photograph is of the gel exposed to blue light at 490nm to show only fluorescently tagged molecules. Arrows indicate the bands depicting dNTPs and fluorescently labelled oligonucleotides that have not been incorporated into full-length product.

Figure 15 is a photograph of the same gel as figure 14, with the exception that it has been stained with GelRed and exposed to UV light at 300nm. This shows all of the oligonucleotides and fragments produced during the reactions of Example 2.

Figure 16 is a photograph of a gel where staining with GelRed has been performed and exposed to UV light at 300nm to show all oligonucleotides and DNA present in the 10% TBE gel. 10 lanes are shown, with molecular ladders included (LR) to allow determination of approximate molecular weight of the DNA/oligonucleotides. The labels are consistent with those used for Figure 14. The production of closed linear DNA (dbDNA) is confirmed by the presence of a band as marked on the gel photograph.

Figures 17a to 17c depicts a schematic representation of the method used in Example 4. An oligonucleotide is attached to a 5' tag (301) and may include a portion of a protelomerase target site (101a), together with two reverse complementary sequences (302 and 303) separated by an intervening sequence (133). This is shown on figure 17a. Figure 17b depicts the looping back and annealing of the complementary sequences (302 and 303), leaving a free 3' end (304) available for extension, using the first strand as a template. Figure 17c shows that an entire protelomerase target site is synthesised by the action of DNA polymerase (137), which allows protelomerase to cleave the molecule at this point and release the 5' tag (301), and leave a circular single stranded DNA structure with one protelomerase-derived hairpin.

Figure 18a is a photograph of a gel obtained using Example 4. This shows DNA oligonucleotides DNA-T and DNA-V, at the varying concentrations indicated, after temperature cycling reactions with DNA polymerase to extend and complete the protelomerase site. The left hand side of the gel (LH) shows the looped extended DNA-T and DNA-V (both 159 nucleotides) which have not been exposed to their cognate protelomerases nor T5 exonuclease. The right hand side (RH) of the gel shows the looped extended DNA-T and DNA-V which have been exposed only to T5 exonuclease. This attacks the DNA structures with free ends and thus no structures are seen.

Figure 18b is a photograph of a gel obtained using Example 4. This shows DNA oligonucleotides DNA-T and DNA-V, at the varying concentrations indicated, after temperature cycling reactions with DNA polymerase to extend and complete the protelomerase site. The left hand side of the gel (LH) shows the looped extended DNA-T and DNA-V which have been treated with the cognate protelomerase (TelN for DNA-T and VP58.5 for DNA-V). This catalyses cleavage/joining resulting in a circular DNA product and by-product as indicated. The right hand side ( H) of the gel shows the looped extended DNA-T and DNA-V which have been exposed to the cognate protelomerase and then to T5 exonuclease.

Detailed Description of the Invention

The process of the present invention utilises an oligonucleotide as a basis for polymerase- mediated extension to produce single-stranded DNA, which may then be converted to double- stranded DNA. The oligonucleotide may be immobilised to a solid support before the polymerase- mediated extension commences, or the oligonucleotide may be immobilised after polymerase- mediated extension in solution. If polymerase-mediated extension takes place in solution prior to immobilisation, it is preferred that 1 to 5 rounds of extension occur, i.e. 1 to 5 separate template oligonucleotides are used in polymerase-mediated extension. The oligonucleotide is immobilised at an appropriate point in the process. The single-stranded DNA or double-stranded DNA is released from the solid support in a closed form comprising at least one hairpin, by contacting with a protelomerase enzyme.

The process of the invention is carried out in an in vitro cell-free environment. Thus, the process is carried out in the absence of a host cell and typically comprises use of purified enzymatic components. Accordingly, the synthesis of DNA and processing by protelomerase is typically carried out by contacting the reaction components in solution in a suitable container. Particular components are provided in immobilisable or immobilised form, which includes the means for attachment to a solid support.

A hairpin is a structure in a polynucleic acid, such as DNA or RNA, formed due to base- pairing between neighbouring complementary sequences of a single strand of the polynucleic acid. The neighbouring complementary sequences may be separated by a few nucleotides, eg. 1- 10 or 1-5 nucleotides. An example of this is depicted in Figure 10. If a loop of non- complementary sequence is included between the two sections of complementary sequence, this forms a hairpin loop. The loop may be of any suitable length. Immobilised oligonucleotide

The immobilised oligonucleotide of the present invention, or starting primer, is capable of being extended. The immobilised oligonucleotide may be chemically synthesised or may be prepared by template-dependent extension of an initial (shorter) primer, for example in solution. An oligonucleotide may thus be immobilised to a solid support to carry out further template- dependent extension, and to allow for completion of production of a desired sequence to be released from the solid support.

Examples of various oligonucleotides are shown in Figure 1, and various points of immobilisation are depicted in Figure 2.

In some embodiments, the immobilised oligonucleotide to be extended on the solid support comprises a first portion of a target sequence for a protelomerase. The first portion of a target sequence for a protelomerase may be derived from any protelomerase target sequence. The first portion of a target sequence for a protelomerase included in the starting primer is designed such that in combination with a second complementary portion of the target sequence for the protelomerase provided as discussed below, a complete protelomerase sequence may be formed. The skilled person is able to divide a protelomerase target sequence into first and second portions which are able to recreate a complete target sequence when juxtaposed together, by reference to the discussion of the characteristics of complete target sequences provided below. The provision of an appropriate pairing of first and second portions of a protelomerase target sequence may also be validated empirically using a suitable assay for protelomerase activity as discussed below. It will be appreciated that the first and second portions of the protelomerase target sequence may each be provided as a single strand of DNA. The first portion may be provided on a different strand of DNA to the second portion, or they may both be provided on the same single strand of DNA, provided that there is sufficient intervening sequence to allow the second portion to be juxtaposed to the first portion. The first and second portions of the protelomerase target sequence may therefore be formed of complementary sequences, allowing them to anneal to one another.

In more detail, a complete protelomerase target sequence cleaved by its cognate protelomerase as described herein is present as a duplex of a first DNA sequence comprising a forward (or sense) portion of a protelomerase target sequence and a complementary second DNA sequence containing the reverse (or antisense) portion of the protelomerase target sequence. The second DNA sequence may comprise the reverse complement of the protelomerase target sequence comprised in the first DNA sequence. In other words, the first portion of the protelomerase target sequence included proximal to the solid support in the first strand forms a complementary duplex with a second portion of the protelomerase target sequence provided at the distal end of the first strand or on a complementary second strand.

As shown in Figure lib, despite the two portions (101a and 101b) of the protelomerase target sequence forming a duplex due to the complementary nature of the sequence of the portions, because of the palindromic nature of the protelomerase target sequence, each portion has the ability to fold into a hairpin due to internal self-complementary sequences within the portion of the target sequence. This is shown in Figure 11a.

Where the immobilised oligonucleotide comprises a first portion of a target sequence for a protelomerase prior to extension, it may further comprise a 5' and/or a 3' flanking region thereto. The 5' and/or 3'flanking regions may have any sequence. Where the oligonucleotide contains 5' and 3' flanking regions, they are preferably not self-complementary. In other words, the 5' and 3' flanking regions are preferably sufficiently non-complementary such that the 3' region of the oligonucleotide remains available for extension under conditions promoting template-dependent extension of the oligonucleotide. In some embodiments, the 3' and/or 5' flanking regions of the oligonucleotide comprises a specific sequence, for example, sequences designed to provide a complementary sequence to sequences synthesised in the first strand or complementary second strand.

The oligonucleotide may comprise natural or modified deoxyribonucleotides and ribonucleic acids or combinations thereof. Chemical modifications may be made to the bases including but not limited to adenine, guanine, thymine, cytosine and uracil, the ribose sugar backbone and the alpha phosphate group. Examples of suitable modified deoxyribonucleotides include, locked nucleic acid (LNA), bridged nucleic acids (BNA), peptide nucleic acid (PNA), unlocked nucleic acids (UNAs) and triazole-linked deoxyribonucleotides. The use and

incorporation of modified deoxyribonucleotides is particularly useful to enhance the binding of the extendable (primer) oligonucleotides to their templates and in conditions where it is desirable for nucleotide sequences at the distal end of a single DNA chain to anneal with complementary sequences at the proximal end immobilised to a surface.

The immobilised oligonucleotide can be of any suitable length. Particularly, the immobilised oligonucleotide can be between 5 and 500 bases, 5 and 400, 5 and 300, 5 and 250 or 5 and 200 bases in length. Particularly, the immobilised oligonucleotide is between 15 and 300 bases in length, more particularly between 15 and 250 bases in length or between 15 and 250, 15 and 200, 15 and 150, 15 and 100, 15 and 75, 15 and 50 bases in length. In some embodiments, the immobilised oligonucleotide is 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90 or 95 bases in length. In alternative embodiments, the oligonucleotide is between 30 and 100, preferably 40 and 90 bases in length.

The immobilised oligonucleotide may include the first portion of a protelomerase target sequence, or this may be included by extension of the immobilised oligonucleotide.

The length of the first or second portion of the protelomerase target sequence is determined by the minimum sequence recognised by the cognate protelomerase in order to bind, cleave and re-join the free ends. Several complete protelomerase target sequences are depicted in Figure 13, and each strand represents a portion of the target sequence for the cognate protelomerase. The length of the first and second portions of the protelomerase target sequence for a cognate protelomerase may be the same or nearly so, since they are capable of annealing to form a duplex. Each portion of a protelomerase target sequence may be 20 to 100 bases in length, more particularly 30 to 100 bases in length.

The first or second portion of the protelomerase target sequence for a particular cognate protelomerase may be flanked by one or more sequences which do not form part of the protelomerase target sequence. These flanking sequences may be of any length, and may include specific sequences. These specific sequences may include a section designed to be

complementary to a first template oligonucleotide, a spacer sequence, a section designed to be complementary to a sequence later included in the extended first strand of DNA or an extended further strand of DNA. The flanking sequences may include sequences which it is desired to include in the DNA product.

The oligonucleotide for immobilisation may include a section of sequence which acts as a spacer adjacent to the solid support. This spacer may be of any suitable length, and may be present to avoid any steric hindrance of the enzymes involved in the process of the invention by the sold support or the linkage entity. Ideally, the spacer is up to 250 bases in length, up to 200, 175, 150, 125, 100, 75, 50, 25, 20, 15, 10 or 5 bases in length. Should the immobilised oligonucleotide include the first portion of a protelomerase target sequence, the spacer may be included as part or all of the 5' flanking sequence (104).

The oligonucleotide for immobilisation may include a section of sequence which acts as a complementary sequence for the oligonucleotide template. Preferably, this section can be of any suitable length. The section can be between 5 and 50 bases, or between 5 and 45, 5 and 40, 5 and 35, 5 and 30, 5 and 30, 3 and 25 or 5 and 20 bases in length. Ideally, the section will be 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 or 19 bases in length. It will be appreciated by those skilled in the art that the length of this section of sequence will be determined by the overlap in complementary sequence designed between the immobilised oligonucleotide and the template oligonucleotide. Furthermore, it will be appreciated that the length of this section of sequence will be determined by the DNA polymerase used and the melting temperature of the

complementary sequence formed between this section of sequence and the template oligonucleotide. The same considerations and sequence length apply for each section of sequence in the extended strand that are to be used as a complementary sequence for the annealing of the template oligonucleotide.

Immobilisation to a solid support

The oligonucleotide may be immobilised by any means to the solid support, provided that the oligonucleotide is capable of being extended from its 3' terminus. The oligonucleotide may be directly or indirectly attached to the solid support. The means of immobilisation to the solid support is selected such that the linkage between the oligonucleotide and solid support remains stable under the conditions used for extension of the primer. Preferably, the linkage is stable under denaturing conditions. The linkage may be reversible or irreversible.

A directly attached oligonucleotide may be covalently or non-covalently bound to the solid support. Any suitable linkage chemistry may be used to covalently bind the oligonucleotide to the solid support. The oligonucleotide may be linked via its 5' terminus to the solid support (Figure 2a). Alternatively, the oligonucleotide may be immobilised to the solid support by a linkage at an internal position (Figure 2b). Typically, where the oligonucleotide to be immobilised includes a first portion of a protelomerase sequence, the linkage is present at a position within the oligonucleotide outside of the target sequence, such as in a spacer sequence. In such an embodiment, the linkage is preferably in the 5' flanking region of the oligonucleotide to the first portion of the target sequence for a protelomerase. However, a linkage in the 3' flanking region of the oligonucleotide to the first portion of the target sequence for a protelomerase is also possible (i.e. Figure 2b). The linker may include a spacer to increase the distance between the solid support and the oligonucleotide. Covalent attachment may occur via conjugation with a coupling agent, via standard phosphoramidite chemistry, reverse amidite chemistry or via a 5'amino linker. Non-covalent attachment encompasses electrostatic interactions, hydrogen bonds and receptor/ligand or antibody/antigen coupling. An example of a non-covalent attachment is the biotin/streptavidin system. The solid support may be coated with streptavidin to attach a biotinylated oligonucleotide.

The spacer may be a molecule comprising for example a carbon chain consisting of multiple methylene bridges or ethylene glycol units. Spacer length in terms of the number of both carbon and oxygen atoms may be between about 3 and about 20, between about 3 and about 15 and optimally between about 5 and about 10, and typically at least 3 atoms. Typically, a plurality of oligonucleotides will be immobilised to the solid support, permitting multiple independent synthesis reactions in parallel on a single solid support. The solid support may comprise a density of at least 10 ³ , at least 10 ^s , at least 10 ⁹ , at least 10 ¹² or at least 10 ¹⁵ immobilised oligonucleotides per square millimetre. The preferred density is between 10 ⁹ and 10 ¹² immobilised oligonucleotides per square millimetre depending on the product to be synthesised. In addition to considerations relating to the density it is also important that the immobilised oligonucleotides are evenly distributed on the surface of the solid support to prevent unwanted intermolecular interactions and steric hindrance which inhibits oligonucleotide synthesis. Methods of ensuring even distribution of oligonucleotides on a solid surface are known in the art.

The oligonucleotides are typically attached to the solid support in the form of an array. The construction of oligonucleotide arrays is well-known in the art. The oligonucleotides may be formed on the solid support by drop deposition or inkjet technology or pre-synthesised oligonucleotides may be deposited on the array and coupled at the desired location.

The solid support according to the invention is any surface to which the oligonucleotide to be extended may be attached, bonded, coupled or tethered. Examples of solid supports include plates, beads, microbeads, hybridisation chips, membranes, crystals and ceramics. Examples of solid support materials include glass, plastics, synthetic polymers, nitrocellulose, nylon, ceramics, metals, resins, gels and membranes.

Extension of the immobilised oligonucleotide

The immobilised oligonucleotide is extended by use of a series of template

oligonucleotides which overlap in sequence to form a sequence complementary to the sequence which is desired to be synthesised. In other words, the template oligonucleotides correspond to the complementary strand to the strand to be extended. The series of template oligonucleotides includes a first template oligonucleotide which comprises a complementary sequence to the immobilised oligonucleotide, typically a complementary sequence to a section in the 3'region of the immobilised oligonucleotide. The first template oligonucleotide anneals to the immobilised oligonucleotide, and includes a sequence which overlaps the 3' end of the immobilised oligonucleotide and provides a template for template-dependent extension of the immobilised oligonucleotide, which thus acts as a primer. Once template-dependent extension of the first template oligonucleotide has been performed, the first template oligonucleotide is removed. A second template oligonucleotide comprises a sequence which overlaps with the sequence of the first template oligonucleotide, and thus which is complementary to the sequence incorporated into the extended oligonucleotide using the first template oligonucleotide as template.

Accordingly, the second template oligonucleotide anneals to the extended oligonucleotide, and again provides a template for further extension of the immobilised oligonucleotide. A sufficient number of further template oligonucleotides are provided so as to incorporate the full sequence desired to be synthesised into the immobilised oligonucleotide to thus produce the first DNA strand. An example of this process is depicted in Figure 3, which shows the template

oligonucleotide (130) annealing to the immobilised oligonucleotide (100a) near the 3' end (102a) at a region of complementary sequence (131). The template oligonucleotide overhangs the 3' end of the immobilised oligonucleotide and is available to act as a template for the polymerase- dependent extension of the 3' end of the immobilised oligonucleotide. The polymerase produces a complementary sequence (132) to the template oligonucleotide. The template oligonucleotide is then removed by appropriate means and further template oligonucleotides are successively added, used to direct extension and removed, leaving an extended first strand (133).

The first portion of the protelomerase target sequence may be incorporated into the first strand by template-dependent extension of the immobilised oligonucleotide with one or more template oligonucleotides. Typically, the first portion of the protelomerase target sequence will be incorporated into the first strand at a position proximal to the solid support. Suitably, the first portion of the protelomerase target sequence maybe located sufficiently distant to the solid support to minimise interference of the linkage to the solid support with cleavage of the complete protelomerase target sequence. In some embodiments, this may mean that no sequence is required between the linkage to the solid support and the first portion of the protelomerase target sequence. For example, a chemical spacer as described above may be utilised. In an alternative embodiment, the first portion of the protelomerase target sequence is separated from the linkage point to the solid support by at least 5 bases. In an alternative embodiment, the first portion of the protelomerase target sequence is separated by 5 to 250 bases, more preferably 5 to 200, 5 to 150, 5 to 100, 5 to 75, 5 to 50 bases. The first portion may be separated from the solid support by 5, 10, 15, 20 or 25 bases or any intervening length.

In some embodiments, an oligonucleotide may be extended by template-dependent extension with one or more template oligonucleotides prior to immobilisation on the solid support for further extension. Initial steps of extension prior to immobilisation may be carried out in solution.

A template oligonucleotide comprises a sequence which overlaps with at least one other template oligonucleotide in the series. The length of overlap is sufficient to allow for annealing to the corresponding complementary sequence in the extended oligonucleotide, and template- dependent extension of said oligonucleotide. The overlapping sequence is typically at least five nucleotides in length, and may be at least 10, at least 12, or more preferably at least 15 nucleotides in length. The overlapping sequence may be about 5 to about 10, about 10 to about 30, about 10 to about 25, about 10 to about 20, or about 15 to about 25 nucleotides in length. Preferably the overlapping sequence is about 5 to about 20 nucleotides in length. A template oligonucleotide is typically at least 30 nucleotides in length, and may be 30-40, 30-50, 30-60, 30- 80, 30-100, 100-130, 130-160, 160-200, 100-200 or 40-60 nucleotides in length. A template oligonucleotide may be up to 200 nucleotides in length. Template oligonucleotides may be about 40 to about 50 nucleotides in length with about 5 to about 20 nucleotides of overlap.

The number of template oligonucleotides is selected according to the length of the sequence desired to be synthesised, and the typical overlaps, total lengths described herein. The length of a desired sequence to be synthesised may range from about 100 to about 150 bases for aptamers, to from about 1 kilobase to about 15 kilobases for vaccines and other therapeutic DNA products. Where the desired DNA sequence comprises one or more aptamer sequences, aptamer lengths may range from about 15 to about 20, about 20 to about 30, about 30 to about 40, about 40 to about 60, about 60 to about 100, about 100 to 150 bases or longer. The length of an aptamer sequence may be in the range of about 15 to about 150 bases. The number of template oligonucleotides required for synthesis in a process of the invention can be estimated for each product from the general formula: 2 + (sequence length/30).

Preferably, one or more of the template oligonucleotides are non-extendable, so as to reduce competition between extension of the immobilised oligonucleotide using the template oligonucleotides as template, and extension of the template oligonucleotides using the immobilised oligonucleotide as template. In some embodiments, all of the template

oligonucleotides are non-extendable. In other embodiments, all template oligonucleotides are extendible. The skilled person is able to vary conditions such as reagent concentrations, DNA polymerase, pH, ionic strength, types of divalent and monovalent ions, temperature, or to incorporate secondary structure destabilisers (e.g. nucleocapsid protein from HIV-1), molecular crowding agents (e.g. trehalose, dextran, DMSO, BSA or polyethylene glycol) or DNA condensing agents (e.g. multivalent cationic charged ligands) to favour extension of the immobilised oligonucleotide where extendable template oligonucleotides are used.

The immobilised oligonucleotide is incubated with the series of template oligonucleotides under conditions promoting template-dependent extension. Such conditions include the presence of at least one DNA polymerase. Any DNA polymerase may be used. Any commercially available DNA polymerase is suitable for use in the process of the invention. Two, three, four, five or more different DNA polymerases may be used, for example one which provides a proof reading function and one or more others which do not. DNA polymerases having different mechanisms may be used e.g. strand displacement type polymerases and DNA polymerases replicating DNA by other methods. A suitable example of a DNA polymerase that does not have strand displacement activity is T4 DNA polymerase.

It is preferred that a DNA polymerase is highly stable, such that its activity is not substantially reduced by prolonged incubation under process conditions. Therefore, the enzyme preferably has a long half-life under a range of process conditions including but not limited to temperature and pH. It is also preferred that a DNA polymerase has one or more characteristics suitable for a manufacturing process. The DNA polymerase may have high fidelity, for example through having proof-reading activity. The DNA polymerase is preferred to have strand-displacing activity, since this may assist in replicating sequences with internal hairpins. Furthermore, it may be preferred that a DNA polymerase displays high processivity. It is preferred that a DNA polymerase does not display non-specific exonuclease activity.

The skilled person can determine whether or not a given DNA polymerase displays characteristics as defined above by comparison with the properties displayed by commercially available DNA polymerases, e.g. phi29, DeepVent®, Bacillus stearothermophilus (Bst) DNA polymerase, and the large fragment of DNA polymerase of Bacillus smithii (BSM) and the large fragment of Bacillus subtilis (BSU) DNA polymerase I. These enzymes are commercially available from several sources including New England Biolabs, Inc., Lucigen and Thermo Scientific.

Where a high processivity is referred to, this typically denotes the average number of nucleotides added by a DNA polymerase enzyme per association/dissociation with the template, i.e. the length of primer extension obtained from a single association event.

Strand displacement-type polymerases are preferred. Strand-displacement polymerases can assist extension through hairpin structures present in protelomerase target sequences. Preferred strand displacement-type polymerases are phi 29, Deep Vent®, BSU, BSM and Bst DNA polymerase I or variants of any thereof. The term "strand displacement" is used herein to describe the ability of a DNA polymerase to displace complementary strands on encountering a region of double stranded DNA during DNA synthesis. It should be understood that strand displacement amplification methods differ from PC -based methods in that cycles of

denaturation are not essential for efficient DNA amplification, as double-stranded DNA is not an obstacle to continued synthesis of new DNA strands. In contrast, PCR methods require cycles of denaturation (i.e. elevating temperature to 94 degrees centigrade or above) during the amplification process to melt double-stranded DNA and provide new single stranded templates.

A strand displacement DNA polymerase used in the method of the invention preferably has a processivity (primer extension length) of at least 20 kb, more preferably, at least 30 kb, at least 50 kb, or at least 70 kb or greater. In particularly preferred embodiments, the strand displacement DNA polymerase has a processivity that is comparable to, or greater than phi29 DNA polymerase.

A preferred type of strand displacement DNA polymerase is a rolling circle amplification (RCA) polymerase, such as phi29 DNA polymerase.

The conditions promoting template-dependent extension of the immobilised

oligonucleotide are suitable to allow for its annealing to a template oligonucleotide, and include a suitable temperature and buffer. Appropriate annealing/hybridisation conditions may be selected empirically. An example of preferred annealing conditions used in the present invention include a buffer comprising 30mM Tris-HCI pH 7.4, 30mM KCI, 7.5mM MgCI ₂ , lOmM (NH ₄ ) ₂ S0 ₄ and ImM DTT. The annealing may be carried out following denaturation by gradual cooling to the desired reaction temperature. Denaturation may assist extension by promoting displacement of a previous template oligonucleotide in favour of binding of a new template oligonucleotide.

Accordingly, the process of the invention may comprise one or more steps of incubating an immobilised oligonucleotide bound to a template oligonucleotide under denaturing conditions subsequent to template-dependent extension. The process of the invention may comprise a step of incubation under denaturing conditions after each template-dependent extension of an immobilised oligonucleotide with a template oligonucleotide. Suitable denaturation conditions include chemical denaturation, such as by adjustment of pH, thermal denaturation by increased temperature, and change in ionic strength, such as by removal of cations for example by incubation in deionised water. Thermal denaturation is suitably employed in combination with a thermostable immobilisation of the oligonucleotide to be extended. A suitable pH for denaturation is pH 11, then adjusted to pH 7.5 to subsequently permit extension. The process of the invention may also comprise one or more, such as 1 to 5, preferably at least 2 steps of washing the extended primer to remove a previous template oligonucleotide prior to addition of a new template oligonucleotide. The process of the invention may comprise a washing step to remove template oligonucleotide after each template-dependent extension of a primer with a template oligonucleotide. The process of the invention may also comprise contacting the primer with each member of the series of template oligonucleotides separately or sequentially.

Exemplary denaturation, washing and extension conditions are described in more detail below. The conditions promoting template-dependent extension also comprise conditions promoting DNA polymerase activity. The conditions comprise use of any temperature allowing for DNA polymerase activity, commonly in the range of 20 to 90 degrees centigrade. A preferred temperature range may be about 20 to about 40 or about 25 to about 35 degrees centigrade.

Typically, an appropriate temperature is selected based on the temperature at which a specific DNA polymerase has optimal activity. This information is commonly available and forms part of the general knowledge of the skilled person. For example, where phi29 DNA polymerase is used, a suitable temperature range would be about 25 to about 35 degrees centigrade, preferably about 30 degrees centigrade. The skilled person would routinely be able to identify a suitable temperature for efficient amplification according to the process of the invention. For example, the process could be carried out at a range of temperatures, and yields of amplified DNA could be monitored to identify an optimal temperature range for a given DNA polymerase.

Other conditions promoting template-dependent extension of the immobilised oligonucleotide include the presence of all four dNTPs, ATP, TTP, CTP and GTP, suitable buffering agents/pH and other factors which are required for enzyme performance or stability. Suitable conditions include any conditions used to provide for activity of DNA polymerase enzymes known in the art.

For example, the pH may be within the range of 3 to 10, preferably 5 to 8 or about 7, such as about 7.5. pH may be maintained in this range by use of one or more buffering agents. Such buffers include, but are not restricted to MES, Bis-Tris, ADA, ACES, PIPES, MOBS, MOPS, MOPSO, Bis-Tris Propane, BES, TES, HEPES, DIPSO, TAPSO, Trizma, HEPPSO, POPSO, TEA, EPPS, Tricine, Gly- Gly, Bicine, HEPBS, TAPS, AM PD, TABS, AM PSO, CHES, CAPSO, AMP, CAPS, CABS, phosphate, citric acid-sodium hydrogen phosphate, citric acid-sodium citrate, sodium acetate-acetic acid, imidazole and sodium carbonate-sodium bicarbonate. The reaction may also comprise salts of divalent metals such as but not limited to salts of magnesium (Mg ²⁺ ) and manganese (Mn ²⁺ ), including chlorides, acetates and sulphates. Salts of monovalent metals may also be included, such as sodium salts and potassium salts, for example potassium chloride. Other salts that may be included are ammonium salts, in particular ammonium sulphate.

Detergents may also be included. Examples of suitable detergents include Triton X-100, Tween 20 and derivatives of either thereof. Stabilising agents may also be included in the reaction. Any suitable stabilising agent may be used, in particular, bovine serum albumin (BSA) and other stabilising proteins. Reaction conditions may also be improved by adding agents that relax DNA and make template denaturation easier. Such agents include, for example, dimethyl sulphoxide (DMSO), formamide, glycerol and betaine. It should be understood that the skilled person is able to modify and optimise

amplification and incubation conditions for the process of the invention on the basis of their general knowledge. Likewise the specific concentrations of particular agents may be selected on the basis of previous examples in the art and further optimised on the basis of general knowledge. As an example, a suitable reaction buffer used in CA-based methods in the art is 50mM Tris-HCI, pH 7.4, lOmM MgCI ₂ , 20mM (NH ₄ ) ₂ S0 ₄ , 5% glycerol, 0.2mM BSA, ImM dNTPs. A preferred reaction buffer used in the RCA amplification of the invention is 30mM Tris-HCI, pH 7.4, 30mM KCI, 7.5 mM MgCI ₂ , lOmM (NH ₄ ) ₂ S0 ₄ , ImM DTT, 2mM dNTPs. This buffer is particularly suitable for use with phi29 RCA polymerase.

The reaction conditions may also comprise use of one or more additional proteins. The

DNA template may be amplified in the presence of at least one pyrophosphatase, such as Yeast Inorganic pyrophosphatase. Two, three, four, five or more different pyrophosphatases may be used. These enzymes are able to degrade pyrophosphate generated by the DNA polymerase from dNTPs during strand replication. Build-up of pyrophosphate in the reaction can cause inhibition of DNA polymerases and reduce speed and efficiency of DNA amplification. Pyrophosphatases can break down pyrophosphate into non-inhibitory phosphate. An example of a suitable

pyrophosphatase for use in the process of the present invention is Saccharomyces cerevisiae pyrophosphatase, available commercially from New England Biolabs Inc.

Any single-stranded binding protein (SSBP) may be used in the process of the invention, to stabilise single-stranded DNA. SSBPs are essential components of living cells and participate in all processes that involve ssDNA, such as DNA replication, repair and recombination. In these processes, SSBPs bind to transiently formed ssDNA and may help stabilise ssDNA structure. An example of a suitable SSBP for use in the process of the present invention is T4 gene 32 protein, available commercially from New England Biolabs, Inc.

The washing conditions may be any suitable conditions that allow for the template oligonucleotide to be removed from the immobilised oligonucleotide and the reaction mixture. Ideally, the process of washing includes or causes a denaturation step. A suitable pH for denaturation is pH 11, such as the pH provided by washing with a solution of alkali such as sodium hydroxide (i.e. lOmM), and then it is desirable to then adjust the conditions to pH 7.4 to subsequently permit extension, for example by using reaction buffer as a final wash solution. Any suitable washing conditions may be used.

Production of double-stranded DNA on the solid support The process of the invention may be adapted to produce and release a double-stranded DNA from the solid support. The double-stranded DNA produced is typically a linear covalently closed double-stranded DNA. The double-stranded DNA thus is closed at both ends by hairpins where at least one hairpin comprises a portion of a target sequence for a protelomerase.

Figures 3a - 3j, 4a to 4f, 4g, 4h and 5a to 5e show examples of processes suitable for producing double-stranded DNA.

For production of double-stranded DNA, in one embodiment, an extended first strand comprising a desired DNA sequence is synthesised as discussed above, and this extended first strand is used as a template for extension of a complementary second strand using a primer complementary to the extended first strand (terminal template oligonucleotide or reverse primer). Conditions suitable for extension of the complementary second strand may be selected from those described above for extension of the first strand. The complementary second strand is extended to include a second complementary portion of the target sequence for a protelomerase located proximal to the solid support, which can thus pair or form a duplex with the

corresponding first portion of the target sequence, to create a complete target sequence for a protelomerase proximal to the solid support.

Typically, a distal protelomerase target sequence is included in the double-stranded DNA to allow for formation of a second closed end. Accordingly, the process for production of double- stranded DNA from a solid support typically comprises template-dependent extension of the first strand with a template oligonucleotide comprising a first portion of a target sequence for a protelomerase, typically as a final extension of the first strand. The extended first strand thus comprises a first portion of a target sequence for a protelomerase distal to said solid support. The thus extended first strand can then be used as a template for extension of a complementary second strand, typically using a reverse primer comprising a second portion of the target sequence for a protelomerase located at the distal end of the first strand. Accordingly, a complete, distal, protelomerase target sequence is incorporated in the double-stranded DNA.

The proximal and distal protelomerase target sequences may be target sequences for the same or different protelomerase target sequences, and may each be selected from any of the protelomerase target sequences discussed below. The proximal and distal protelomerase target sequences may be selected from: target sequences for bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof, target sequences for Agrobacterium tumefaciens TelA of SEQ ID NO: 12 or a variant thereof, or target sequences for bacteriophage Vp58.5 gp40 of SEQ ID NO: 14 or a variant thereof. The proximal and distal target protelomerase target sequences may comprise one target sequence for bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof and/or one target sequence of Agro bade um tumefaciens TelA of SEQ ID NO: 12 or a variant thereof and/or a target sequence for bacteriophage Vp58.5 gp40 of SEQ ID NO: 14 or a variant thereof.

Figures 3a to 3j depict an example of such a process. The immobilised oligonucleotide (100a) is extended as previously described (figures 3a to 3e), using a succession of template oligonucleotides (130). In figures 3f and 3g a final template oligonucleotide (or reverse primer) (100b) is added, and this anneals near to the 3' end of the first DNA strand (133), in a region of complementary sequence (134). The final template oligonucleotide comprises a portion of a protelomerase target sequence (101b). DNA polymerase extends both the 3' end of the extended first strand (141) and the 3' end of the final template oligonucleotide (102b), creating a complete protelomerase target sequence (137b) at the distal, extended end of the first strand (133), and also creating a complementary second strand (136). The complementary second strand uses the first strand (133) as a template and thus a complete protelomerase target sequence (137a) is created proximal to the solid-support. These complete protelomerase target sequences may have the same or different sequences, using the same or different cognate protelomerase enzymes. The target sequences can be cleaved and ligated using the cognate protelomerases, releasing a double stranded, closed linear DNA from the solid support (140).

Figures 5a to 5e depict a similar process using an oligonucleotide immobilised at an internal nucleotide.

In an alternative embodiment, a double stranded closed linear DNA may be produced by including a suitable sequence capable of forming a hairpin in the first DNA strand. The double stranded DNA can be formed from a single strand of DNA. Such a process is depicted in Figures 4a to 4f and 4g/4h. For example, two sequences which are complementary to one another (181 and 184) may be incorporated at or near the distal end of the first strand in such a manner that they anneal together within the same strand to form a hairpin. Suitably, the two self-complementary sequences are neighbouring (as shown on Figures 4a to 4f) resulting in the formation of a hairpin. Alternatively, the two self-complementary sequences are separated by an intervening sequence region (185) which may be looped out on formation of the hairpin (as shown on Figures 4g and 4h). In order to produce double-stranded DNA, the 3' end of the hairpin is extended using the segment of the first strand between the solid support and the hairpin as a template, to form a complementary second segment (186) of the first strand (133). The complementary second segment (186) includes a second portion of a protelomerase target sequence (101b), thus forming a complete protelomerase target sequence (137) proximal to the solid support. This may be cleaved by the cognate protelomerase, releasing a double-stranded closed ended DNA molecule (187). In this embodiment, one closed end is formed by the action of a protelomerase on its cognate target sequence and the other closed end is formed by including self-complementary sequences in the sequence of the first strand. The other closed end may be formed by a hairpin loop, with the loop being single stranded DNA.

In an alternative embodiment, an oligonucleotide template comprising a hairpin can be used as the terminal template oligonucleotide for the extension of the first strand. The hairpin in the template oligonucleotide can be a portion of a protelomerase target sequence or can be composed of neighbouring complementary sequences. The neighbouring complementary sequences may comprise sequences other than those of protelomerase target sequences. The template oligonucleotide anneals to the extended first strand, as shown in Figure 5a. The 3' end of the first strand is further extended using the template oligonucleotide (Figure 5b). The template oligonucleotide is then removed, and a hairpin is formed in the first strand due to the introduction of complementary sequences (Figure 5c). The 3' end is then further extended using the first strand as a template, including the first portion of the protelomerase target sequence. This generates a second portion of a protelomerase target sequence, and thus a complete protelomerase target sequence proximal to the solid support (Figure 5d). Once the cognate protelomerase is added, a closed linear DNA is produced (Figure 5e).

Production of single-stranded DNA on the solid support

The process of the invention may be adapted to produce and release a single-stranded DNA from the solid support. The single-stranded DNA produced is typically a single-stranded circular DNA comprising a hairpin comprising a portion of a target sequence for a protelomerase. This single- stranded circular DNA is also described herein as a pinched single-stranded circular DNA (such Figure 6e, 155 and Figure 9, 240). Once the immobilised oligonucleotide has been extended to produce a first strand comprising the desired DNA sequence, this strand may be released as single-stranded DNA in various ways. Examples of processes to synthesize single-stranded DNA include those depicted in Figures 4a to 4g, 6a - 6e, 7a- 7e and 8a to 8f.

In one embodiment, the extended first strand is incubated under conditions promoting annealing of a sequence at its distal 3' end with a complementary sequence in the same strand located proximal to the solid support. This may be achieved by using any of the sequences in the oligonucleotide proximal to the solid support, such as the first portion of the protelomerase target sequence, the 3' flanking sequence and/or the 5' flanking sequence as sequences which are complementary to those introduced into the distal, 3' end of the first strand. An example of this is depicted in Figures 6a to 6e. Promoting the annealing of the distal 3' end of the extended first DNA strand (133) with its proximal end to allow for creation of a complete protelomerase target sequence proximal to the solid support may be achieved in a number of ways.

Firstly at the distal end of the extended first DNA strand, a complementary sequence to a sequence in the same strand proximal to the solid support may be covalently linked to a biotin molecule. Biotin molecules have high affinity binding to streptavidin. Thus, template-dependent extension may be used to incorporate a biotinylated nucleotide sequence at the distal end of the first strand. Where streptavidin, which has multiple biotin binding sites, is used to immobilise the oligonucleotide to the solid support, it can also be used to attract the binding of the high affinity biotin attached to the distal end of the extended primer.

Secondly, magnetic particles may be provided proximal to the solid support and also linked to a sequence the distal end of the extended DNA strand which is complementary to a sequence at the proximal end. In this case, an appropriate application of a magnetic field is used to draw together the proximal and distal ends of the first strand to allow for creation of a complete protelomerase target sequence. The annealing of complementary sequences creates a DNA loop comprising the desired DNA sequence previously incorporated in the extended first strand. Sequence(s) that are complementary to sequence(s) in the first strand proximal to the solid support are thus incorporated in the 3' distal end of the extended first strand by template- dependent extension. In this embodiment, the process of the invention may comprise template- dependent extension of the first DNA strand with a template oligonucleotide comprising a sequence corresponding to a sequence in the 3' flanking region to the first portion of the target sequence of the protelomerase located proximal to the solid support. Thus, template-dependent extension is used to incorporate at the distal 3' end of the first DNA strand a complementary sequence to the 3' flanking region to the first portion of the target sequence of the

protelomerase. Alternatively, or additionally, the process of the invention may comprise template-dependent extension of the distal end of the first DNA strand to incorporate therein the second portion of the target sequence of the protelomerase which is complementary to the first portion thereof located proximal to the solid support. Optionally, a complementary sequence to the 5' flanking region to the first portion of the target sequence of the protelomerase may also be introduced at the distal end of the first DNA strand by template-dependent extension. The sequences complementary to the 3' and/or 5' flanking regions, and/or the second complementary portion of the target sequence for a protelomerase located proximal to the solid support may be introduced through one or more template oligonucleotides. The above template-dependent extension(s) may be the final step(s) of template-dependent extension of the first DNA strand. The introduction of the above complementary sequences at the distal end of the extended first strand may produce a distal end which creates a complete target sequence for a protelomerase proximal to the solid support on annealing of the distal and proximal ends of the first strand. Alternatively, a complete target sequence for a protelomerase may be created proximal to the solid support by annealing of the distal end of the first strand to its proximal end, and subsequent template-dependent extension of the distal end using the proximal end, including the first portion of the protelomerase target sequence, as a template.

The above-defined method of creating a complete protelomerase target sequence proximal to the solid support by annealing the distal 3' end with a complementary sequence may be used for any single-stranded DNA created, including those where further hairpins may have been introduced into the first strand.

In another embodiment, the extended first strand is used to create a single-stranded circular DNA comprising two or more hairpins, at least one of which comprises a portion of a target sequence for a protelomerase. The hairpins may be located at any point around the circle, but if two are present, they are typically located on opposing sides of the single stranded circular DNA (Figure 8f, 177). In order to introduce one or more hairpins on the extended DNA strand, a hairpin or the means to make a hairpin, i.e. a complete protelomerase target sequence, may be incorporated into the distal end of the first strand, i.e. at the 3' end of the first strand. Once this hairpin, or the means to make the hairpin, has been introduced, extension of the first DNA strand may continue. The hairpin at the distal end of the first strand may be formed by the

incorporation of a complete protelomerase target sequence, as depicted in Figure 8b, or from another suitable sequence capable of forming a hairpin, as depicted in Figure 4c. A mixture of both techniques may be used to make a closed single stranded DNA with multiple hairpins.

If a hairpin is introduced using internal self-complementary sequences, an intervening section of single stranded DNA can be created. This is depicted in Figure 4h. For example, two sequences which are complementary to one another (181 and 184) may be incorporated at the distal end of the first strand in such a manner that they anneal together within the same strand to form a hairpin. Suitably, the two self-complementary sequences are separated by an intervening sequence region (185) which is looped out on formation of the hairpin, forming the single- stranded DNA section. This hairpin loop may comprise a desired sequence such as an aptamer or sequence for expression. Thus, the loop may be of any suitable length. It is preferred that the loop is up to 500 bases, or up to 400 bases, up to 300 bases or up to 250 bases in length. Ideally, for applications such as shorter aptamers, the loop may be between 10 and 100 bases, i.e.

between 10 and 90, 10 and 80, 10 and 70, 10 and 60 or 10 and 50 bases. The loop may be 10, 15, 20, 25, 30, 35, 40, 45, 50, 55 or 60 bases in length. Alternatively, the loop (185) may contain a minimal amount of nucleic acid residues. The 3' end of the hairpin can be extended as a second segment of the first DNA strand using the first segment of the first strand as a template or by using of a series of template oligonucleotides forming a sequence that is non-complementary to at least a portion of the first segment of the first strand.

Alternatively, a complete protelomerase target sequence can be introduced at the distal end of the first DNA strand. A complete protelomerase target sequence may be introduced as discussed previously, using a template oligonucleotide comprising a portion of the protelomerase target sequence, such as those depicted in Figures la and lc (100 and 120).

In this means of generating an introduced hairpin at the distal end of the first strand, a first portion of a protelomerase target sequence is introduced in the extended first strand distal to the solid support, typically in a template-dependent extension of the first strand. The thus extended first strand is then annealed to a template oligonucleotide comprising a portion of a target sequence for a protelomerase located at the distal end of the extended first strand, to thereby create a complete target sequence for a protelomerase once polymerase-mediated extension has taken place. The template oligonucleotide comprising the portion of the target sequence for a protelomerase is then extended as a second segment of the first DNA strand. This complete protelomerase target sequence at the distal end or 3' end of the first strand may have the same or different cognate protelomerase as the complete protelomerase target sequence proximal to the solid support.

It is preferred that the template oligonucleotide used to introduce a complete protelomerase target sequence at a distal or 3' end of the first strand includes a portion of the 3' flanking sequence that is not complementary to the first segment of the first strand. The first segment of the first strand is the sequence between the solid support and the complete protelomerase target sequence. Such a template oligonucleotide (120) is shown in Figures lb and 8a. Should the 3' end of this oligonucleotide be extended, the first segment of the first strand is not used as a template, but template oligonucleotides are required for polymerase-mediated extension as described previously.

Once the first hairpin or means to generate a hairpin has been included in the first strand, further extension of the 3' end of the hairpin generates a second segment of the first strand. A further hairpin or means to generate a hairpin may be introduced at the distal end of the second segment of the first strand, further extension of the 3' end of that hairpin results in extension of the third segment of the first strand, and so on. Thus, once a hairpin or means to generate a hairpin has been introduced at the distal end of the first strand, extension of the 3' end of the hairpin results in further extension of a new segment of that strand.

This process can be repeated to insert as many hairpins or means to generate the same as required.

The complete protelomerase target sequence(s) introduced into the first strand may be recognised by the same or different cognate protelomerases.

Once the complete protelomerase target sequence has been included in the first strand, it may be processed using its cognate protelomerase immediately, or at any other appropriate point in the process, including after release of the single-stranded DNA.

In an embodiment, the complete protelomerase target sequence used to create an introduced hairpin is distinct from the complete protelomerase target sequence used proximal to the solid support. In a further embodiment, each of the complete protelomerase target sequences introduced into the first strand are recognised by different cognate protelomerase enzymes.

Alternatively, the first and second portions of the protelomerase target sequence may both be incorporated into the distal end of the strand such that the complementary portions anneal together within the same strand and can form a hairpin at the distal end once cleaved by the cognate protelomerase.

In any of the above embodiments providing for further extension of the first strand via different segments, each separated by a hairpin, the extension may be carried out with a series of template oligonucleotides which overlap in sequence to form a non-complementary sequence to the earlier segment(s) of the first strand. Advantageously, this allows for inclusion of further single-stranded DNA sequences of interest in the synthesised DNA in addition to the DNA sequence of interest included in the earlier segment(s) of the first strand. The further extension of the first strand may use template oligonucleotides, extension conditions, and washing and denaturation steps as described above.

In order to form a circular single-stranded structure and release the DNA from the solid support, a complete protelomerase target sequence is required proximal to the solid support. This may be introduced using any of the means previously discussed. Advantageously, the further template-dependent extension of the first strand further comprises introduction of

complementary sequence(s) in the extended first strand to sequence(s) in the first segment of the first strand proximal to the solid support, annealing of these complementary sequence(s), and creation of a complete target sequence for a protelomerase proximal to the solid support. In this way, the first extended strand creates a non-complementary single-stranded region flanked by a proximal complete protelomerase sequence at the proximal end and one or more complete protelomerase target sequence(s) or other sequence(s) capable of forming a hairpin within the extended first strand. In other words, the extended first strand may contain a complete protelomerase target sequence proximal to the solid support (the proximal complete

protelomerase target sequence), which allows for release from the solid support, and also further complete protelomerase target sequences or other sequences capable of forming a hairpin, such as regions with neighbouring complementary sequences. This enables the released sequence to contain a plurality of hairpins, whilst being composed of a single strand of DNA. Due to the action of the protelomerase on the proximal complete protelomerase target sequence, the single stranded DNA has no free ends, and if the hairpins are denatured, the single stranded DNA is circular.

Where proximal and one or more distal protelomerase target sequences are used, they may be target sequences for the same or different protelomerases, and may each be selected from any of the protelomerase target sequences discussed below or shown on Figure 13. The proximal and distal protelomerase target sequences may both be target sequences for bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof, both be target sequences for Agrobacterium tumefaciens TelA of SEQ ID NO: 12 or a variant thereof or both be target sequences for bacteriophage Vp58.5 gp40 of SEQ ID NO. 14 or a variant thereof. The proximal and distal target protelomerase target sequences may comprise sequences selected from: a target sequence for bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof, a target sequence for Agrobacterium tumefaciens TelA of SEQ ID NO: 12 or a variant thereof, or a target sequence for bacteriophage Vp58.5 gp40 of SEQ ID NO: 14 or a variant thereof.

As described above, sequence(s) complementary to sequences in the first strand proximal to the solid support that may be used to facilitate creation of a complete protelomerase target sequence may comprise a sequence complementary to the 3' flanking region to the first portion of a target sequence for a protelomerase located proximal to the solid support, a second complementary portion of the proximal protelomerase target sequence, and optionally a sequence complementary to the 5' flanking region to the first portion of the proximal protelomerase target sequence. The above complementary sequences may be introduced on the distal end of the extended first strand, using one or more template oligonucleotides, or a partially complementary sequence may be annealed and then extended using the first strand as template to create a complete protelomerase target sequence proximal to the solid support.

The complete protelomerase target sequences may be contacted with the cognate protelomerase at any appropriate point of the process. Figure 8b and 8c depicts that once the complete protelomerase target sequence is introduced in the distal end of the first strand, it is contacted with the cognate protelomerase before further extension of the strand commences. However, this could be delayed until the first strand has been further extended, as discussed below.

Following strand extension, the complete protelomerase target sequences distal (where employed, alternatively substituted for another hairpin-forming sequence) and proximal to the solid support are each contacted with a cognate protelomerase, thereby generating a single- stranded circular DNA comprising two or more hairpins, at least one of which comprises a portion of a target sequence for a protelomerase. The number of hairpins depends upon the number of complete protelomerase target sequences introduced into the extended strand and/or inclusion of other hairpin-forming sequences. The hairpins are located between the desired DNA sequence(s) introduced in the first strand, and thus the first strand is split into segments of single stranded DNA, each of which is flanked by hairpins.

In all embodiments for production of single-stranded DNA discussed above, the single- stranded DNA molecule is released from the solid support by contacting the complete

protelomerase target sequence proximal to the solid support with a protelomerase, as described below.

Protelomerase target sequence

A protelomerase target sequence is used in accordance with the invention as a substrate for a protelomerase enzyme, to provide for release of synthesised DNA from the solid support, and/or to generate a closed end for a synthesised DNA. A protelomerase target sequence used in the invention is created from first and second portions of the target sequence which are synthesised in separate steps, either on the same DNA strand or on separate DNA strands. The process of the invention may comprise creation of a single protelomerase target sequence proximal to the solid support or creation of protelomerase target sequences both proximal and distal to said solid support.

A complete protelomerase target sequence is any DNA sequence whose presence in a DNA template provides for cleavage and religation of the template by the enzymatic activity of protelomerase to form at least one hairpin. Examples of native complete protelomerase target sequences are given in Figure 13. A complete protelomerase target sequence may be the minimal sequence required for the action of the cognate protelomerase and may not represent the entire native recognition sequence. Where a template is double-stranded, a complete protelomerase sequence allows for its conversion into a closed linear DNA. In other words, a complete protelomerase target sequence is required for the cleavage and religation of double stranded DNA by protelomerase to form covalently closed linear DNA. A complete protelomerase target sequence thus contains the minimum amount of sequence required for target recognition, cleavage and religation of the sequence.

Typically, a protelomerase target sequence comprises any perfect palindromic sequence i.e. any double-stranded DNA sequence having two-fold rotational symmetry, also described herein as a perfect inverted repeat. As shown in Figure 13, the protelomerase target sequences from various mesophilic bacteriophages and bacterial plasmids all share the common feature of comprising a perfect inverted repeat. The length of the perfect inverted repeat differs depending on the specific organism. In Borrelia burgdorferi, the perfect inverted repeat is 14 base pairs in length. In various mesophilic bacteriophages, the perfect inverted repeat is 22 base pairs or greater in length. Also, in some cases, e.g. E. coli N15, the central perfect inverted palindrome is flanked by inverted repeat sequences, i.e. forming part of a larger imperfect inverted palindrome.

A complete protelomerase target sequence as used in the invention preferably comprises a double stranded palindromic (perfect inverted repeat) sequence of at least 14 base pairs in length, and thus each portion of the protelomerase target sequence comprises at least 14 bases in length. As shown in Figure 13, base pairs of the perfect inverted repeat are conserved at certain positions between different bacteriophages, while flexibility in sequence is possible at other positions. An example of a perfect inverted repeat from a protelomerase target sequence is SEQ ID NO: 22, particularly preferred for use with Agrobacterium tumefaciens TelA, which also is the minimum sequence required for the protelomerase TelA to bind, cleave and religate the open ends. This is shown on Figure 13 in grey.

The perfect inverted repeat may be flanked by additional inverted repeat sequences. The flanking inverted repeats may be perfect or imperfect repeats i.e. may be completely symmetrical or partially symmetrical. The flanking inverted repeats may be contiguous with or non-contiguous with the central palindrome. The protelomerase target sequence may comprise an imperfect inverted repeat sequence which comprises a perfect inverted repeat sequence of at least 14 base pairs in length. The imperfect inverted repeat sequence may comprise a perfect inverted repeat sequence of at least 22 base pairs in length. Particularly preferred protelomerase target sequences comprise the sequences of SEQ ID NOs: 15 to 21 or variants thereof.

The sequences of SEQ ID NOs: 15 to 21 comprise perfect inverted repeat sequences, and additionally comprise flanking sequences from the relevant organisms. A protelomerase target sequence comprising the sequence of SEQ ID NO: 15 or a variant thereof is preferred for use in combination with E.coli N15 TelN protelomerase of SEQ ID NO: 10 and variants thereof. A protelomerase target sequence comprising the sequence of SEQ ID NO: 16 or a variant thereof is preferred for use in combination with Klebsiella phage Phi K02 protelomerase of SEQ ID NO: 12 and variants thereof. A protelomerase target sequence comprising the sequence of SEQ ID NO: 17 or a variant thereof is preferred for use in combination with Yersinia phage PY54

protelomerase of SEQ ID NO: 4 and variants thereof. A protelomerase target sequence comprising the sequence of SEQ ID NO: 18 or a variant thereof is preferred for use in combination with Vibrio phage VP882 protelomerase of SEQ ID NO: 8 and variants thereof. A protelomerase target sequence comprising the sequence of SEQ ID NO: 19 or a variant thereof is preferred for use in combination with a Borrelia burgdorferi protelomerase. A protelomerase target sequence comprising the sequence of SEQ ID NO: 20 or a variant thereof is preferred for use in combination with Agrobacterium tumefaciens TelA of SEQ ID NO: 12 and variants thereof. A protelomerase target sequence comprising the sequence of SEQ ID NO: 21 or a variant thereof is preferred for use in combination with Vibrio parahaemolyticus plasmid Vp58.5 of SEQ ID NO: 14.

Due to the presence of a central section of perfect inverted repeat section, which may be surrounded by imperfect repeat sections, the protelomerase target site may not be symmetrical. If this is the case, the site may be seen as two halves, such as the TelL and Tel sections of the protelomerase target sequence shown in Figure 11 for protelomerase TelN. The protelomerase will still recognise the site if it is made entirely symmetrical, i.e. TelN will recognise a TelL/TelL site and a Tel /Tel site.

Variants of any of the palindrome or protelomerase target sequences described above may also be used in the invention, including homologues or mutants thereof. Mutants include truncations, substitutions or deletions with respect to the native sequence, and can thus also include fragments of the sequence. A variant sequence is any sequence whose presence in a DNA template allows for its cleavage and re-ligation to form at least one hairpin by the enzymatic activity of protelomerase, or to convert the synthesised DNA into a closed linear DNA in the case of a double-stranded DNA. This can readily be determined by use of an appropriate assay for cleavage and re-ligation of the template or for the formation of closed linear DNA. Any suitable assay described in the art may be used. An example of a suitable assay is described in Deneke et al, PNAS (2000) 97, 7721-7726. An example of a suitable assay for protelomerase activity in the process of the invention is monitoring for release of the synthesised DNA from the solid support. Preferably, a variant sequence allows for protelomerase binding and activity that is comparable to that observed with the native sequence. Examples of preferred variants of palindrome sequences described herein include truncated palindrome sequences that preserve the perfect repeat structure, and remain capable of allowing for cleavage and re-ligation of a template to form a hairpin or to allow for formation of closed linear DNA. However, variant protelomerase target sequences may be modified such that they no longer preserve a perfect palindrome, provided that they are able to act as substrates for protelomerase activity.

It should be understood that the skilled person would readily be able to identify additional suitable protelomerase target sequences for use in the invention on the basis of the structural principles outlined above. Candidate protelomerase target sequences can be screened for their ability to promote cleavage and re-ligation of a template to form a hairpin or formation of closed linear DNA using the assays for protelomerase activity described above.

The hairpin formed by the action of a protelomerase does not generally include a region of non-complementary sequence, i.e. the sequence is normally entirely complementary. With reference to Figure 11c showing the hairpins created by TelN on the Tel L site, the entire sequence within the hairpin is complementary, and there is no loop structure at the end of the hairpin composed of non-complementary sequence. However, some structural distortion may create some strain on the pairing of bases at the tip of the hairpin, meaning that these are not available for base-pairing, despite their complementary nature. It is preferred that the hairpins formed by the protelomerase do not include any non-complementary loop section at the tip. Some "wobbles" of non-complementary bases within the length of a hairpin may not affect the structure. It is, however, preferred that the hairpin is entirely self-complementary.

Complementarity describes how the bases of each polynucleotide in a sequence (5' to 3') are in a hydrogen-bonded pair with a complementary base, A to T (or U) and C to G on the anti-parallel (3' to 5') strand, which may be the same strand (internal complementary sequences) or on a different strand. It is preferred that the sequences in the hairpin are 90% complementary, preferably 91%, 92%, 93%, 94%, 95%, 96%, 98%, 99% or 100% complementary. Formation of hairpins and release from solid support

The DNA synthesised from a solid support in accordance with the invention comprises one or more hairpins comprising a portion of a protelomerase target sequence. These may act to close the ends of the DNA in the case of a double-stranded closed linear DNA. The above hairpins are generated by contacting with a protelomerase enzyme, which also acts to release the synthesised DNA from the solid support.

A protelomerase used in the invention is any polypeptide capable of cleaving and rejoining a template comprising a cognate protelomerase target sequence in order to form a hairpin or to produce covalently closed DNA. Thus, the protelomerase has DNA cleavage and ligation functions. Enzymes having protelomerase-type activity have also been described as telomere resolvases (for example in Borrelia burgdorferi). If a DNA comprises a protelomerase target sequence, the enzyme can cut the DNA at this sequence and ligate the ends to create hairpins which may covalently close the DNA. Where the protelomerase target sequence in a template is proximal to a solid support, one of the resulting hairpins will remain bound to the solid support, and the template will also be released from the solid support in a form comprising a hairpin comprising a portion of a protelomerase target sequence, such as a covalently closed DNA. The requirements for protelomerase target sequences are discussed above. As also outlined above, the ability of a given polypeptide to cleave and rejoin a protelomerase target sequence can be determined using any suitable assay described in the art.

Protelomerase enzymes have been described in bacteriophages. In some lysogenic bacteria, bacteriophages exist as extrachromosomal DNA comprising linear double strands with covalently closed ends. The replication of this DNA and the maintenance of the covalently closed ends (or telomeric ends) are dependent on the activity of the enzyme, protelomerase. An example of this catalytic activity is provided by the enzyme, TelN, from the bacteriophage N15 that infects Escherichia coli. TelN recognises a specific nucleotide sequence; a slightly imperfect inverted palindromic structure termed Tel L comprising two halves, TelR and TelL, flanking a 22 base pair inverted perfect repeat (TelO) (see Figure lib). TelR and TelL comprise the closed ends of the DNA once the protelomerase has acted on the TelRL target site (Figure 11c).

The process of the invention requires use of at least one protelomerase. The process of the invention may comprise use of more than one protelomerase, such as two, three, four, five or more different protelomerases. Examples of suitable protelomerases include those from bacteriophages such as phiHAP-1 from Halomonas aquamarina (SEQ ID NO: 2), PY54 from Yersinia enterolytica (SEQ ID NO:4), phiK02 from Klebsiella oxytoca (SEQ ID NO:6) and VP882 from Vibrio sp. (SEQ ID NO: 8), N15 from Escherichia coli (SEQ ID NO:10), Agrobacterium tumefaciens TelA (SEQ ID NO:12), Vp58.5 from Vibrio parahaemolyticus (SEQ ID NO:14) or variants of any thereof. Use of E. coli bacteriophage N15 protelomerase (SEQ ID NO: 10) or a variant thereof, Vibrio parahaemolyticus bacteriophage Vp58.5 gp40 (SEQ ID NO: 14) or a variant thereof and/or Agrobacterium tumefaciens TelA (SEQ ID NO: 12) or a variant thereof is particularly preferred.

Variants of SEQ ID NOs: 2, 4, 6, 8, 10, 12 and 14 include homologues or mutants thereof. Mutants include truncations, substitutions or deletions with respect to the native sequence. A variant must cleave and religate (forming a hairpin) or produce closed linear DNA from a template comprising a protelomerase target sequence as described above.

Any homologues of DNA polymerases or protelomerases mentioned herein are typically a functional homologue and are typically at least 40% homologous to the relevant region of the native protein. Homology can be measured using known methods. For example the UWGCG Package provides the BESTFIT program which can be used to calculate homology (for example used on its default settings) (Devereux et al (1984) Nucleic Acids Research 12, 387-395). The PILEUP and BLAST algorithms can be used to calculate homology or line up sequences (typically on their default settings), for example as described in Altschul S. F. (1993) J Mol Evol 36:290-300; Altschul, S, F et al (1990) J Mol Biol 215:403-10. Software for performing BLAST analyses is publicly available through the National Center for Biotechnology Information

(http://www.ncbi.nlm.nih.gov/).

The BLAST algorithm performs a statistical analysis of the similarity between two sequences; see e.g., Karlin and Altschul (1993) Proc. Natl. Acad. Sci. USA 90: 5873-5787. One measure of similarity provided by the BLAST algorithm is the smallest sum probability (P(N)), which provides an indication of the probability by which a match between two nucleotide or amino acid sequences would occur by chance. For example, a sequence is considered similar to another sequence if the smallest sum probability in comparison of the first sequence to the second sequence is less than about 1, preferably less than about 0.1, more preferably less than about 0.01, and most preferably less than about 0.001.

A variant polypeptide comprises (or consists of) sequence which has at least 40% identity to the native protein. In preferred embodiments, a variant sequence may be at least 55%, 65%, 70%, 75%, 80%, 85%, 90% and more preferably at least 95%, 97% or 99% homologous to a particular region of the native protein over at least 20, preferably at least 30, for instance at least 40, 60, 100, 200, 300, 400 or more contiguous amino acids, or even over the entire sequence of the variant. Alternatively, the variant sequence may be at least 55%, 65%, 70%, 75%, 80%, 85%, 90% and more preferably at least 95%, 97% or 99% homologous to full-length native protein. Typically the variant sequence differs from the relevant region of the native protein by at least, or less than, 2, 5, 10, 20, 40, 50 or 60 mutations (each of which can be substitutions, insertions or deletions). A variant sequence of the invention may have a percentage identity with a particular region of the full-length native protein which is the same as any of the specific percentage homology values (i.e. it may have at least 40%, 55%, 80% or 90% and more preferably at least 95%, 97% or 99% identity) across any of the lengths of sequence mentioned above.

Variants of the native protein also include truncations. Any truncation may be used so long as the variant is still able to produce closed linear DNA as described above. Truncations will typically be made to remove sequences that are non-essential for catalytic activity and/or do not affect conformation of the folded protein, in particular folding of the active site. Truncations may also be selected to improve solubility of the protelomerase polypeptide. Appropriate truncations can routinely be identified by systematic truncation of sequences of varying length from the N- or C-terminus.

Variants of the native protein further include mutants which have one or more, for example, 2, 3, 4, 5 to 10, 10 to 20, 20 to 40 or more, amino acid insertions, substitutions or deletions with respect to a particular region of the native protein. Deletions and insertions are made preferably outside of the catalytic domain. Insertions are typically made at the N- or C- terminal ends of a sequence derived from the native protein, for example for the purposes of recombinant expression. Substitutions are also typically made in regions that are non-essential for catalytic activity and/or do not affect conformation of the folded protein. Such substitutions may be made to improve solubility or other characteristics of the enzyme. Although not generally preferred, substitutions may also be made in the active site or in the second sphere, i.e. residues which affect or contact the position or orientation of one or more of the amino acids in the active site. These substitutions may be made to improve catalytic properties.

Substitutions preferably introduce one or more conservative changes, which replace amino acids with other amino acids of similar chemical structure, similar chemical properties or similar side-chain volume. The amino acids introduced may have similar polarity, hydrophilicity, hydrophobicity, basicity, acidity, neutrality or charge to the amino acids they replace.

Alternatively, the conservative change may introduce another amino acid that is aromatic or aliphatic in the place of a pre-existing aromatic or aliphatic amino acid. Conservative amino acid changes are well known in the art and may be selected in accordance with the properties of the 20 main amino acids as defined in Table A.

Table A - Chemical properties of amino acids

Lys polar, hydrophilic, charged(+) 7rp aromatic, hydrophobic, neutral

Leu aliphatic, hydrophobic, neutral 7yr aromatic, polar, hydrophobic

It is particularly preferred that the variant is able to cleave and religate, release from a solid support or produce covalently closed DNA as described above with an efficiency that is comparable to, or the same as the native protein.

The DNA synthesised on the solid support is incubated with at least one protelomerase under conditions promoting protelomerase activity. In other words, the conditions promote the cleavage and religation of DNA comprising a protelomerase target sequence to form DNA comprising one or more hairpins comprising portions of the protelomerase target sequence, such as a covalently closed DNA with hairpin ends. Conditions promoting protelomerase activity comprise use of any temperature allowing for protelomerase activity, commonly in the range of 20 to 90 degrees centigrade. The temperature may preferably be in a range of 25 to 40 degrees centigrade, such as about 25 to about 35 degrees centigrade, or about 30 degrees centigrade. Appropriate temperatures for a specific protelomerase may be selected according to the principles outlined above in relation to temperature conditions for DNA polymerases. A suitable temperature for use with E.coli bacteriophage TelN protelomerase of SEQ ID NO: 10 is about 25 to about 35 degrees centigrade, such as about 30 degrees centigrade.

Conditions promoting protelomerase activity also comprise the presence of suitable buffering agents/pH and other factors which are required for enzyme performance or stability. Suitable conditions include any conditions used to provide for activity of protelomerase enzymes known in the art. For example, where E.coli bacteriophage TelN protelomerase is used, a suitable buffer may be 30mM Tris-HCI pH 7.4, 30mM KCI, 7.5mM MgCI ₂ , lOmM (NH ₄ ) ₂ S0 ₄ and ImM DTT, which may be diluted 2-8 fold depending on the concentration of the protelomerase used.

Agents and conditions to maintain optimal activity and stability may also be selected from those listed for DNA polymerases.

Where proximal and distal protelomerase target sequences are created which are target sequences for different protelomerases, the distal protelomerase target sequence may be contacted with its cognate protelomerase to form a distal covalently closed DNA end prior to contacting of the protelomerase target sequence proximal to the solid support with its cognate protelomerase. Thus, the distal end of the DNA may be closed by a protelomerase prior to extension of a second non-complementary or complementary strand, or prior to release of the DNA from the solid support. Where necessary, the reaction conditions may be changed to allow for optimal processing of two different protelomerase target sequences by their cognate protelomerases.

In some embodiments, it may be possible to use the same conditions for activity of protelomerase as are used for extension of first and second DNA strands. In other embodiments, it may be necessary to change reaction conditions where conditions used to provide optimal DNA polymerase activity lead to sub-optimal protelomerase activity. Removal of specific agents and change in reaction conditions may be achievable by filtration, dialysis and other methods known in the art. The skilled person would readily be able to identify conditions allowing for optimal DNA polymerase activity and/or protelomerase activity.

In a particularly preferred embodiment, for use in synthesis of DNA by an RCA DNA polymerase, preferably phi29, the DNA synthesis is carried out under buffer conditions substantially identical to or consisting essentially of 30mM Tris-HCI, pH 7.4, 30mM KCI, 7.5mM MgCI ₂ , lOmM (NH ₄ ) ₂ S0 ₄ , ImM DTT, 1 to 4mM dNTPs, such as 2mM dNTPs, at a temperature of 25 to 35 degrees centigrade, such as about 30 degrees centigrade. The processing step with protelomerase may then preferably be carried out with TelN, and/or preferably under buffer conditions substantially identical to or consisting essentially of a 2 to 8 fold dilution of the buffer used for the DNA polymerase, 30mM Tris-HCI, pH 7.4, 30mM KCI, 7.5mM MgCI ₂ , lOmM (NH4) ₂ S0 ₄ and ImM DTT at a temperature of 25 to 35 degrees centigrade, such as about 30 degrees centigrade.

All enzymes and proteins for use in the process of the invention may be produced recombinantly, for example in bacteria. Any means known to the skilled person allowing for recombinant expression may be used. A plasmid or other form of expression vector comprising a nucleic acid sequence encoding the protein of interest may be introduced into bacteria, such that they express the encoded protein. For example, for expression of SEQ. ID NOs: 2, 4, 6, 8, 10, 12 or 14, the vector may comprise the sequence of SEQ ID NOs: 1, 3, 5, 7, 9, 11 or 13, respectively. The expressed protein will then typically be purified, for example by use of an affinity tag, in a sufficient quantity and provided in a form suitable for use in the process of the invention. Such methodology for recombinant protein production is routinely available to the skilled person on the basis of their general knowledge. The above discussion applies to the provision of any protein discussed herein.

Purification and formulation of DNA

Following release of a DNA product comprising one or more hairpins from the solid support by the action of protelomerase, the process of the invention may further comprise a step of purifying the DNA product. The purification referred to above will typically be performed to remove any undesired products. Purification may be carried out by any suitable means known in the art. For example, processing may comprise phenol/chloroform nucleic acid purification or the use of a column which selectively binds nucleic acid, such as those commercially available from Qjagen. The skilled person can routinely identify suitable purification techniques for use in isolation of DNA.

Once the DNA product has been generated and purified in a sufficient quantity, the process may further comprise its formulation as a DNA composition, for example a therapeutic DNA composition. A therapeutic DNA composition will comprise a therapeutic DNA molecule of the type referred to below. Such a composition will comprise a therapeutically effective amount of the DNA in a form suitable for administration by a desired route e.g. an aerosol, an injectable composition or a formulation suitable for oral, mucosal or topical administration.

Formulation of DNA as a conventional pharmaceutical preparation may be done using standard pharmaceutical formulation chemistries and methodologies, which are available to those skilled in the art. Any pharmaceutically acceptable carrier or excipient may be used.

Auxiliary substances, such as wetting or emulsifying agents, pH buffering substances and the like, may be present in the excipient or vehicle. These excipients, vehicles and auxiliary substances are generally pharmaceutical agents which may be administered without undue toxicity and which, in the case of vaccine compositions will not induce an immune response in the individual receiving the composition. A suitable carrier may be a liposome or a DNA nanoparticle. DNA nanoparticles may be created by use of compaction agents such as cationic polymers to condense the DNA into nanoparticles. DNA condensing polymers may be conjugated to peptides that act as nuclear localisation signals (NLS) to overcome intra- and extracellular barriers to DNA delivery.

Pharmaceutically acceptable excipients include, but are not limited to, liquids such as water, saline, polyethylene glycol, hyaluronic acid, glycerol and ethanol. Pharmaceutically acceptable salts can also be included therein, for example, mineral acid salts such as

hydrochlorides, hydrobromides, phosphates, sulfates, and the like; and the salts of organic acids such as acetates, propionates, malonates, benzoates, and the like. It is also preferred, although not required, that the preparation will contain a pharmaceutically acceptable excipient that serves as a stabilizer, particularly for peptide, protein or other like molecules if they are to be included in the composition. Examples of suitable carriers that also act as stabilizers for peptides include, without limitation, pharmaceutical grades of dextrose, sucrose, lactose, trehalose, mannitol, sorbitol, inositol, dextran, and the like. Other suitable carriers include, again without limitation, starch, cellulose, sodium or calcium phosphates, citric acid, tartaric acid, glycine, high molecular weight polyethylene glycols (PEGs), and combination thereof. A thorough discussion of pharmaceutically acceptable excipients, vehicles and auxiliary substances is available in

REM INGTON'S PHARMACEUTICAL SCIENCES (Mack Pub. Co., N.J. 1991), incorporated herein by reference.

Oligonucleotides and kits for synthesis of DNA

The invention also provides an oligonucleotide suitable for use in the process of the invention. The oligonucleotide comprises a first portion of a target sequence for a protelomerase, flanked by 5' and 3' regions. The flanking 5' and 3' regions are preferably non-complementary.

The invention further provides an oligonucleotide of the invention immobilised to a solid support. The solid support may be any solid support as described above. The oligonucleotide may be immobilised to the solid support by any linkage or spacer as described above. The solid support typically comprises a plurality of oligonucleotides as described above.

The invention additionally provides a kit comprising an oligonucleotide of the invention and at least one further component selected from a solid support, a series of template oligonucleotides overlapping in sequence and instructions for use in a process of synthesis of DNA of the invention. The kit may comprise two or all three of the above further components. The kit may further comprise a DNA polymerase and/or a protelomerase, each selected from any of these enzymes described above. The DNA polymerase is preferably phi29 or a variant thereof. The protelomerase is preferably bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof, bacteriophage Vp58.5 gp40 of SEQ ID NO: 14 or a variant thereof, or Agrobacterium tumefaciens TelA of SEQ ID NO: 12 or a variant thereof. The kit may comprise the oligonucleotide of the invention attached to a solid support. DNA molecules comprising hairpins

The invention additionally provides the single-stranded and double-stranded DNA molecules that may be synthesised using the process of the invention as products per se.

Accordingly, the invention provides a single-stranded circular DNA comprising a hairpin comprising a portion of a target sequence for a protelomerase. This product is also described herein as a pinched single-stranded circular DNA and/or IbDNA.

The invention further provides a single-stranded circular DNA comprising (at least) two hairpins, at least one of which comprises a portion of a target sequence for a protelomerase. The hairpins are typically separated from each other by single-stranded segments that are non- complementary with each other. The non-complementary single-stranded segments may each comprise at least one desired sequence, such as an aptamer sequence or sequence for expression. A different desired DNA sequence, such as a different aptamer sequence may be provided in each of the non-complementary single-stranded segments.

Preferably, the hairpin comprising a portion of a target sequence for a protelomerase comprises a portion of a target sequence for bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof, a portion of a target sequence for bacteriophage Vp58.5 gp40 of SEQ ID NO 14 or a variant thereof or a portion of a target sequence for Agrobacterium tumefaciens TelA of SEQ ID NO: 12 or a variant thereof.

Where two or more hairpins comprising a portion of a target sequence for a

protelomerase are present in the single-stranded circular DNA, they may each comprise a portion of a target sequence for the same protelomerase. Alternatively, each hairpin may comprise a portion of a target sequence for a different protelomerase. One hairpin may comprise a portion of a target sequence for bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof, another hairpin a portion of a target sequence for Agrobacterium tumefaciens TelA of SEQ ID NO: 12 or a variant thereof and another hairpin a portion of the target sequence for bacteriophage Vp58.5 gp40 of SEQ ID NO 14 or a variant thereof

Alternatively, the single-stranded circular DNA may comprise a hairpin that is not formed by a portion of a protelomerase target sequence, in addition to a hairpin that comprises a portion of a protelomerase target sequence. The above alternative types of hairpin may be formed by any two complementary sequence regions within the same strand which can anneal together. The loop joining the two arms of the hairpin may comprise a desired DNA sequence, such as an aptamer sequence as described herein, and thus form a hairpin loop. The single-stranded circular DNA may comprise an aptamer sequence within a hairpin and one or more aptamer sequences in non-complementary single-stranded segments either side of a hairpin. Alternatively, the DNA sequences comprising the hairpin are fully complementary and no section of sequence loops at the end of the hairpin.

The invention additionally provides a linear covalently closed double-stranded DNA closed at one end by a hairpin comprising a portion of a target sequence for a first protelomerase and at a second end by a hairpin comprising a portion of a target sequence for a second protelomerase, wherein the first and second protelomerases are different. In other words, the linear covalently closed DNA (or closed linear DNA) comprises a first hairpin comprising a portion of a target sequence for a first protelomerase and a second hairpin comprising a portion of a target sequence for a second protelomerase The first protelomerase may be bacteriophage N15 TelN of SEQ ID NO: 10 or a variant thereof and the second protelomerase Agrobacterium tumefaciens TelA of SEQ ID NO: 12 or a variant thereof. Alternatively, the first and/or second protelomerase may be bacteriophage Vp58.5 gp40 of SEQ ID NO 14 or a variant thereof.

The single-stranded circular and double-stranded closed linear DNA molecules of the invention have particular utility as therapeutic agents i.e. DNA medicines which can be used to express a gene product in vivo. This is because their covalently closed structure prevents attack by enzymes such as exonucleases, leading to enhanced stability and longevity of gene expression as compared to "open" DNA molecules with exposed DNA ends. Linear double stranded open-ended cassettes have been demonstrated to be inefficient with respect to gene expression when introduced into host tissue. This has been attributed to cassette instability due to the action of exonucleases in the extracellular space.

Sequestering DNA ends inside covalently closed structures also has other advantages. The DNA ends are prevented from integrating with genomic DNA and so closed linear DNA molecules are of improved safety. Also, the closed linear structure prevents concatamerisation of DNA molecules inside host cells and thus expression levels of the gene product can be regulated in a more sensitive manner.

In addition, the single-stranded circular DNA molecules of the invention are considered by the inventors to offer advantages over linear covalently closed DNA molecules for expression, since their structure is more open and thus expected to be more readily transcribed.

The DNA molecules of the invention may include a DNA sequence (described above as the desired DNA sequence) which encodes a therapeutic product. The therapeutic product may be a DNA aptamer, a protein, a peptide, or an NA, such as small interfering RNA. Exemplary lengths for DNA aptamers are discussed above. In order to provide for therapeutic utility, such a DNA molecule may comprise an expression cassette comprising one or more promoter or enhancer elements and a gene or other coding sequence which encodes an mRNA or protein of interest. The expression cassette may comprise a eukaryotic promoter operably linked to a sequence encoding a protein of interest, and optionally an enhancer and/or a eukaryotic transcription termination sequence.

The DNA molecules of the invention may be used for production of DNA for expression in a host cell, particularly for production of DNA vaccines. DNA vaccines typically encode a modified form of an infectious organism's DNA. DNA vaccines are administered to a subject where they then express the selected protein of the infectious organism, initiating an immune response against that protein which is typically protective. DNA vaccines may also encode a tumour antigen in a cancer immunotherapy approach. Any DNA vaccine may be used in the DNA molecules of the invention. Also, the process of the invention may produce other types of therapeutic DNA molecules e.g. those used in gene therapy. For example, such DNA molecules can be used to express a functional gene where a subject has a genetic disorder caused by a dysfunctional version of that gene. Examples of such diseases are well known in the art.

The novel structures of the invention may also have non-medical uses including in material science, in data storage and the like.

Medical uses

The products of the invention are particularly preferred for use in medicine, since they for example are considered to provide advantages for stability in vivo. Accordingly, the invention provides a single-stranded circular DNA of the invention, or a linear covalently closed double- stranded DNA of the invention, for use in a method for treatment of the human or animal body, or in a diagnostic method practised on the human or animal body. The invention further provides a method of treatment of the human or animal body, comprising administering a therapeutically effective amount of a single-stranded circular DNA of the invention or a linear covalently closed double-stranded DNA of the invention to a human or animal in need thereof.

In particular therapeutic aspects, the DNA molecules of the invention may be used as DNA vaccines or for gene therapy. The DNA molecules may be used to induce an immune response to an antigen of interest, typically by expression of a DNA sequence encoding the antigen. The invention thus provides a method of inducing an immune response against an antigen in a host, said method comprising administering a single-stranded circular DNA or linear covalently closed double-stranded DNA of the invention that encodes said antigen to said host in such a way that said antigen is expressed in said host and induces an immune response against said antigen. The antigen may be selected from any suitable antigen.

In other embodiments, a DNA molecule of the invention may include one or more aptamer sequences which can be used to specifically recognise target molecule(s) for therapeutic purposes. An aptamer sequence for any target molecule of interest may be included in the DNA molecule.

A DNA molecule of the invention may be formulated for therapeutic purposes as described above, and for example be provided in a suitable carrier such as a liposome or DNA nanoparticle. A pharmaceutical formulation comprising a DNA molecule of the invention can be delivered to a subject in vivo using a variety of known routes and techniques. For example, a pharmaceutical formulation can be provided as an injectable solution, suspension or emulsion and administered via parenteral, subcutaneous, epidermal, intradermal, intramuscular, intra- lymphatic, intra-arterial, intraperitoneal, or intravenous injection using a conventional needle and syringe, a microneedle and syringe or using a liquid jet injection system. The administration may be made using a patch, such as a microtine patch. Pharmaceutical formulations can also be administered topically to skin or mucosal tissue, such as nasally, intratonsillarly, intratracheal^, intestinal, rectally or vaginally, or provided as a finely divided spray suitable for respiratory or pulmonary administration. Other modes of administration include oral administration, suppositories, sublingual administration, and active or passive transdermal delivery techniques.

A suitable amount of the DNA molecule to be administered may be determined empirically. Dosages for administration will depend upon a number of factors including the nature of the DNA molecule and mode of action, pharmaceutical formulation, the route of administration and the schedule and timing of the administration regime. Suitable doses of a DNA molecule described herein may be in the order of microgram ^g), milligram (mg) or up to grams (g). A single administration of the DNA molecule may be sufficient to have a beneficial effect for the patient, but it will be appreciated that it may be beneficial if the DNA molecule is administered more than once, in which case typical administration regimes may be, for example, once or twice a week for 2-4 weeks every 6 months, or once a day for a week every four to six months.

Synthesis of novel structures

The novel structures disclosed herein are preferably synthesised using the methods of the present invention. Alternative methods of synthesising such structures are conceivable, and include template-dependent extension using no immobilisation, or alternatively, attaching the initial oligonucleotide to a 5' tag. The template dependent extension may involve thermal cycling to replace templates in a stepwise fashion, since this allows for denaturation of the "spent" template and annealing of the fresh template to allow extension. Alternatively or additionally, the methods to make the novel structures may involve different enzymes such as helicases, i.e. Escherichia coli Rep and Bacillus stearothermophilus PcrA. Any suitable method to make these novel structures involving at least one protelomerase-derived hairpin may be used.

Downstream amplification of a synthetic template

The novel structures and/or products of the process of the invention carried out on a solid support may be used as a template for further DNA amplification. This provides for a process of large-scale DNA production which can be carried out wholly in an in vitro cell-free environment without a requirement for propagation of a starting template in bacteria. The invention thus provides an in vitro cell-free process for amplification of DNA using a single-stranded circular DNA of the invention or a linear covalently closed double-stranded DNA of the invention as a template. The above process may further comprise initial production of the DNA template in accordance with a process for production of DNA of the invention carried out on a solid support. Prior to amplification, the single stranded circular DNA of the invention may be converted to double-stranded DNA, this process requiring a non-strand displacing DNA polymerase and a ligase enzyme.

The above process using either template comprises contacting the template with at least one DNA polymerase in the presence of one or more primers under conditions promoting amplification of the template. The DNA polymerase may be selected from any described above for use in the process of production of DNA on a solid support of the invention. However, the DNA polymerase is preferably a strand-displacement type polymerase, more preferably a rolling circle amplification (RCA) polymerase. A preferred RCA polymerase is phi29 or a variant thereof.

The conditions promoting amplification of the template are typically selected from those described above for template-dependent extension of the immobilised oligonucleotide on the solid support. Where the template is a linear covalently closed double-stranded DNA, it may be incubated under denaturing conditions to form a single stranded circular DNA before or during conditions promoting amplification of the template DNA.

The primer or primers may be an oligonucleotide that hybridizes to the DNA template and generates a DNA: primer hybrid that primes a DNA synthesis reaction. The primers may be nonspecific (i.e. random in sequence) or may be specific for one or more sequences comprised within the DNA template. It is preferred that the primers are of random sequence so as to allow for nonspecific initiation at any site on the DNA template. This allows for high efficiency of amplification through multiple initiation reactions from each template strand. Examples of random primers are hexamers, heptamers, octamers, nonamers, decamers or sequences greater in length, for example of 12, 15, 18, 20 or 30 nucleotides in length. A random primer may be of 6 to 30, 8 to 30 or 12 to 30 nucleotides in length. Random primers are typically provided as a mix of

oligonucleotides which are representative of all potential combinations of e.g. hexamers, heptamers, octamers or nonamers in the DNA template.

In other embodiments, the primers are specific. This means they have a sequence which is complementary to a sequence in the DNA template from which initiation of amplification is desired. In this embodiment, either a single primer or a pair of primers may be used to specifically amplify the DNA template to produce single stranded or double stranded DNA, respectively. In another embodiment a primer capable of specifically binding to a palindromic sequence within a protelomerase target sequence comprised within the DNA template may be used. Such a primer is capable of binding to each complementary strand of the template and thus priming

amplification on both strands, so only one species of primer molecule is required per template. If the template is single-stranded, only one species of primer will be required. Where the DNA template is pinched single stranded circular DNA, a single species of primer may be used. This primer may bind to a target sequence within the single stranded DNA, or alternatively bind specifically to a portion of the protelomerase target sequence within the hairpin.

Primers may be unlabelled, or may comprise one or more labels, for example

radionuclides or fluorescent dyes. Primers may also comprise chemically modified nucleotides. Primers may also comprise a portion of non-complementary spacer sequence for immobilisation to a solid support. Primer lengths/sequences may typically be selected based on temperature considerations i.e. as being able to bind to the template at the temperature used in the amplification step.

The contacting of the DNA template with the DNA polymerase and one or more primers takes place under conditions promoting annealing of primers to the DNA template. The conditions include the presence of single-stranded DNA allowing for hybridisation of the primers. The conditions also include a temperature and buffer allowing for annealing of the primer to the template. Appropriate annealing/hybridisation conditions may be selected depending on the nature of the primer. An example of preferred annealing conditions used in the present invention include a buffer 30mM Tris-HCI pH 7.5, 20mM KCI, 8mM MgCI ₂ . The annealing may be carried out following denaturation by gradual cooling to the desired reaction temperature.

The above processes for amplification from a DNA template of the invention may result in amplified DNA comprising concatamers comprising repeat units of amplified DNA sequence, typically where RCA amplification is used. If the template is a single stranded DNA in which a primer is chosen from a region outside of the protelomerase target sequence, with at least one hairpin according to the present invention, the concatamer made by amplifying this structure will be multiple repeats of the single stranded DNA (which is complementary in sequence to the template DNA) with intervening portions of protelomerase target sequences. In one embodiment, the single stranded DNA is synthesised as an antisense strand, such that the concatamer forms the sense strand. If a primer for RCA amplification is chosen from within the protelomerase target sequence the resulting product will be double-stranded concatamers, since the initial strand of amplified DNA will be available for priming.

The single stranded DNA concatamers produced by amplification of the novel structures may be resolved into single units of amplified DNA comprising a desired sequence of interest by various methods. The concatamers may be resolved by one or more nucleases, such as endonucleases. In particular embodiments, a single-stranded circular DNA template may comprise endonuclease site(s) flanking (at one or both ends) a sequence of interest such as a DNA aptamer, such that the aptamer sequence can be specifically released in linear form and/or excised from the remainder of the amplified DNA sequence. The double stranded concatamers can be resolved into closed linear DNA by the action of a protelomerase.

In another embodiment, where a linear covalently closed double-stranded DNA of the invention is used as a template, protelomerase target sequences for first and second

protelomerases are present in the template, such that the amplified DNA may be contacted with the first and second protelomerases to release single linear covalently closed DNA units from the amplified DNA. The first and second protelomerases are preferably selected from bacteriophage N15 TelN of SEQ ID NO: 15 or a variant thereof, Vibrio parahaemolyticus bacteriophage Vp58.5 gp40 (SEQ ID NO: 14) or variant thereof and/or Agrobacterium tumefaciens TelA of SEQ ID NO: 31 or a variant thereof.

The amplified DNA resulting from the above processes for DNA amplification may be purified and formulated as described above. The invention further provides a process for making a pharmaceutical composition comprising carrying out a process for production of a DNA or for amplification of a DNA as described above, and formulating the resulting DNA product or amplified DNA with a pharmaceutically acceptable carrier or diluent.

Examples

Example 1

Reagents

Dynabeads MyOne Streptavidin CI (Life Technologies), NxGen Phi20 DNA polymerase (Lucigen), Protelomerase TelN produced in-house (stock concentration 20.5μΜ), Exonuclease III

(Enzymatics), dNTPs Mix (containing lithium salts) (Bioline), Nuclease free water (Sigma Aldrich), NaOH (Fisher Scientific), Qubit dsDNA BR (Broad Range) Assay Kit (Life Technologies), Qubit dsDNA HS (High Sensitivity) Assay Kit (Life Technologies), SafeWhite Nucleic Acid Stain (NBS Biologicals), Agarose (NBS Biologicals), Oligonucleotide SI, SEQ ID No. 23, (Oligo Factory), Low Molecular Weight DNA Ladder (New England Biolabs), lOx TLG buffer reagent: 300mM Tris HCI pH 7.4 (Sigma Aldrich), 300mM KCI (Sigma Aldrich), 75mM MgCI ₂ (Sigma Aldrich), 50mM (NH ₄ ) ₂ S0 ₄ (Sigma Aldrich), nuclease free water (Sigma Aldrich)

IBA Oligonucleotides (5' to 3'): PEGS1: BiotinC6 attached to SEQ ID No. 24, PEGS2: SEQ ID No. 25, PEGS3: SEQ ID No. 26, PEGS4: SEQ ID No. 27, PEGS5: SEQ ID No. 28, PEGS6: SEQ ID No. 29, PEGS7: SEQ ID No. 30.

Oligonucleotide design

Oligonucleotide templates were designed to produce a covalently closed linear DNA construct comprising a 194 bp sequence, which contained the first 108bp of the CMV promoter sequence terminated at each end with a hairpin loop comprising a portion of the protelomerase TelN target sequence (Figure lib). Seqbuilder (http://www.dnastar.com/t-seqbuilder.aspx) was used to design the seven oligonucleotide sequences used to construct the closed linear DNA by template- dependent primer extension.

Oligonucleotide PEGS1 (with a length of 71 bases) encodes first portion of the protelomerase TelN target sequence and has a biotin C6 modification on the 5' end, which enables it to be immobilised onto streptavidin-coated magnetic beads. Oligonucleotides PEGS2 to PEGS6 (lengths of 45bases, 45bases, 35bases, 40bases and 30bases, respectively) collectively encode the CMV promoter sequence and act as sequential templates for extension of the first immobilised oligonucleotide, which acts as a primer. The seventh oligonucleotide (PEGS7, with a length of 74bases) contains the second portion of the protelomerase TelN target sequence). Each oligonucleotide template contains a 15bp overlap at its 3' end with the extending 3'end of the extending immobilized oligonucleotide.

First extension step in solution

The first extension step of oligonucleotide PEGS1, using oligonucleotide PEGS2 as a template, was carried out in solution. The following reagents were combined in a 0.2ml PC tube: 33.5μΙ nuclease free water, 5μΙ lOx TLG buffer pH 7.4, 2.5μΙ lOOmM dNTPs (final concentration of 500μΜ), 2.5μΙ ΙΟΟμΜ oligonucleotide 1 (final concentration of 5μΜ), 5μΙ ΙΟΟμΙ oligonucleotide 2 (final concentration of ΙΟμΜ) and 15 units Phi29 to provide a total reaction of 50μΙ. The reaction was mixed and incubated at 30°C for 30 minutes in an Innova 40 incubator without shaking (New Brunswick Scientific).

Oligonucleotide immobilization and extension

Aliquots of streptavidin-coated magnetic beads (30μΙ per extension reaction) were placed in 1.5ml microcentrifuge tubes and washed three times with 400μΙ lx TLG buffer (pH 7.4). A magnetic microcentrifuge rack was used to enable the buffer to be separated and removed from the beads according to the manufacturer's instructions. The beads were then resuspended in 60μΙ of Ix TLG buffer and mixed with 50μΙ of first extension reaction described above. The beads were incubated at room temperature, with gentle shaking (50rpm) for one hour. After immobilisation, the first extension reaction was removed and the beads were washed in 400μΙ of lOmM NaOH. This results in strand denaturation and allows removal of the second oligonucleotide template from the newly extended first oligonucleotide attached to the beads. Three lOmM NaOH washes were carried out to ensure complete denaturation and full removal of the previous oligonucleotide. The beads were then washed in 400μΙ of lx TLG buffer pH 7.4. After the removal of the buffer wash, the following reagents were added to the beads for a 50μΙ extension reaction: 31μΙ water, 5μΙ lOOmM NaOH, 5μΙ oligonucleotide template 3 (final concentration of ΙΟμΜ), 5μΙ lOx TLG buffer pH 7.4, 2.5μΙ lOOmM dNTPs (500μΜ) and 15 units of Phi29 DNA polymerase, to provide a total reaction volume of 50μΙ. The reaction was mixed and incubated at 30°C for 10 minutes without shaking. The reaction was then removed and the beads were washed three times with 400μΙ lOmM NaOH and once with 400μΙ lx TLG buffer pH 7.4.

The extension reactions were then repeated, as described above, for oligonucleotide templates PEGS4, PEGS5, PEGS6 and PEGS7. When the final oligonucleotide template (PEGS7) had been added and the first strand extension completed, the beads were washed with lOmM NaOH to denature and remove it and allow the incorporated TelN sequence to fold and form a hairpin. The 3' end of the hairpin was used to initiate the synthesis of a complementary strand to the now complete template extended first strand. The following components were added to complete the reaction: 41.5μΙ nuclease free water, 5μΙ ΙΟχ TGL buffer pH 7.4, 2.5μΙ lOOmM dNTPs (for a final concentration of 500μΜ) and 15 units of Phi29, for a final reaction volume of 50μΙ. The reaction was mixed and incubated at 30°C for 10 minutes without shaking.

Protelomerase TelN digestion

Following the final reaction to complete the synthesis of a double stranded extension product, the beads were washed twice with 400μΙ lx TLG buffer pH 7.4. The beads were then resuspended in 98μΙ lx TLG buffer and 2μΙ protelomerase TelN (400nM) was added to cleave the extension product from the bead surface. The reaction was incubated at 30°C for 15 minutes. High sensitivity Qubit readings were taken before and after the TelN was added and the amount of product cleaved from the surface was 1.15μβ/ηιΙ.

Rolling Circle Amplification (RCA)

Rolling circle amplification (RCA) using Phi29 DNA polymerase was used to increase the amount of product cleaved from the bead surface. The RCA reaction was set up as follows: the

TelN sequence-specific oligonucleotide SI (30μΜ) was added to the TelN digest. The reaction was heated to 95°C for 5 minutes and then cooled to 4°C. dNTPs were then added to give final concentration of ImM followed by 20 units of Phi29 DNA polymerase. The reaction was incubated at 30°C overnight. TelN and Exonuclease III digestion

The overnight RCA concatameric DNA product (15.]^g/ml) was digested with 500nM protelomerase TelN at 30°C for 15 minutes. A 25μΙ aliquot was taken from the TelN digest reaction and 20U Exonuclease III was added and incubated at 37°C for 30 minutes. Both the TelN digest product and the TelN/Exolll product (25μΙ of each sample) were heat treated at 75°C for one minute and then run on a 2% agarose gel at 50V with SafeWhite DNA stain.

Figure 12, lane 1 shows a reference DNA ladder. Lane 2 shows bands resulting from

amplification (by RCA) of the covalently closed linear DNA product to produce concatamers which were then digested with protelomerase TelN. Clearly there is a band in the 200bp region corresponding to the size of the expected product. Further treatment with exonuclease III (Lane 3) removes contaminating (open ended) fragments but leaves the product band. This

demonstrates that the template extended product is of the expected size and has covalently closed ends (resistant to exonuclease action).

Example 2 - Enzymatic synthesis of covalently closed linear DNA (dbDNA)

Reagents

Phusion ^® High Fidelity DNA polymerase (New England Biolabs), Protelomerase TelN produced in- house (stock concentration 20.5μΜ), Protelomerase VP58.5 produced in house (stock concentration 21.5 μΜ), T5 Exonuclease (New England Biolabs), dNTPs Mix (containing lithium salts) (Bioline), Nuclease free water (Sigma Aldrich), GelRed ss/dsDNA stain (Biotium), 10% TBE precast PAGE gels (ThermoFisher), Low Range (LR) and 1 kb DNA Ladder (New England Biolabs) Isothermal Amplification Buffer ("IAB", New England Biolabs) comprising at lx: 20mM Tris-HCI; lOmM (NH ₄ ) ₂ S0 ₄ ; 50mM KCI; 2mM MgS0 ₄ & 0.1% Tween ^® 20, Buffer 4 (New England Biolabs) comprising at lx: 50mM potassium acetate; 20mM Tris-acetate; lOmM magnesium acetate & ImM DTT, TBE buffer from 20x stock (Thermofisher).

Table 1: Oligonucleotides (Integrated DNA Technologies)

Name Length Description Sequence (5'-3')

1.120 79 nt Initiating oligonucleotide primer - TATC AG C AC AC A ATTG CCC ATT ATAC

Alexa Fluor ^® 488 fluorophore linked GCGCGTATAATGGACTATTGTGTGCT to the 5' end. Encodes TelN target GATATGTACACTTAAGTAGTAATCAA sequence T

1.112 60 nt First template oligonucleotide - 3' TGGGCTATGAACTAATGACCCCGTA end is a dideoxy nucleotide, A TTGA TTACTACTTAAGTGTACA TA T preventing extension by polymerase CAGCACACA

1.113 54 nt Second template oligonucleotide - 3' CCGTAAGTTATGTAACGCGGAACTC end is a dideoxy nucleotide, CAT AT A TGGG CTA TGAA CTAA TGA CC preventing extension by polymerase CCG

1.114 60 nt Third template oligonucleotide - 3' CT AGT AG ATCTG CTAG CCGCCAGGC end is a dideoxy nucleotide, GGGCCATTT ACCGTAAGTTATGTAAC preventing extension by polymerase GCGGAACTC

1.107V 80 nt End capping oligonucleotide A ACCTG C AC AG GTGTAC AT AT AGTCT containing VP58.5 target sequence - AATTAGACTATATGTACACCTGTGCA Alexa Fluor ^® 488 fluorophore linked GGTTACTAGTAGATCTGCTAGCCGCC to the 5' end AG

Oligonucleotide design

Oligonucleotide templates were designed to produce a covalently closed linear DNA construct comprising a 171 base pair sequence, terminated at one end with a hairpin loop comprising a portion of the protelomerase TelN target sequence and at the other, a hairpin loop comprising a portion of the protelomerase VP58.5 target sequence.

Oligonucleotide 1.120 (with a length of 79 nucleotides) encodes the protelomerase TelN target sequence and is modified at the 5' end with an Alexa Fluor ^® 488 fluorophore. Oligonucleotides 1.112, 1.113 and 1.114 (lengths of 60, 54 and 60 nucleotides respectively) collectively encode part of the CMV promoter sequence and act as sequential templates for extension of the first oligonucleotide, 1.120, which acts as a primer. The fifth oligonucleotide (1.107V, with a length of 80 nucleotides) encodes a portion of the protelomerase VP58.5 target sequence. Each oligonucleotide template contains an overlap at its 3' end with the 3'end of the extending oligonucleotide primer; these overlapping sections are marked by italics and underlined in the table. Reaction conditions

The extension reaction in this experiment was performed using a temperature cycling procedure. The reaction volumes were 50μΙ in all cases comprising Isothermal Amplification Buffer (lx concentration), dNTPs mix (800μΜ), 200μΜ of each oligonucleotide (1.120, 1.112, 1.113, 1.114 and 1.107V) and 1 unit of Phusion ^® High Fidelity DNA polymerase (New England Biolabs)

Temperature cycling reactions were performed in a BioRad CIOOO Thermal Cycler under the following conditions: 2 minutes at 95°C; 50 cycles comprising: 1 minute at 95°C, 1 minute at 50°C, 1 minute at 72°C; 10 minutes at 72°C; 4°C END

The reaction produces an open ended double stranded DNA comprising sequences from the 5 oligonucleotides. One end of the molecule incorporates a full TelN protelomerase target sequence while the other incorporates a full VP58.5 target sequence (full sized linear product). The 5' end of each strand is labeled with an Alexa Fluor ^® 488 fluorophore making it visible under blue light at a wavelength of 490 nm. The product of the extension reaction when treated with protelomerases produces covalently closed linear DNA.

Protelomerase reactions were performed in the same buffer as the extension reaction - that is, the protelomerase enzyme was added directly to the cycling reaction upon its completion. Individual protelomerases were added to a final concentration of 1 μΜ but where both TelN and VP58.5 protelomerases were added the concentration of each was reduced to 0.5 μΜ. Reactions were incubated at 37°C for 15-30 minutes, and inactivated at 75°C for a further 10 minutes.

Gel electrophoresis

Gels in this example are all native 10% TBE PAGE gels, run at 180 volts in lx TBE buffer.

Imaging was performed with excitation at wavelength of 490nm for fluorescently tagged molecules or at 300nm after staining with 3 times concentrated GelRed™ DNA stain for 1 hour to image all the DNA/oligonucleotides present. Results are shown as Figures 14 and 15. Lane 1, Figure 14 shows a bright band corresponding to the full sized linear product formed following the templated extension reaction. The lower bright band corresponds to unincorporated fluorescently tagged oligonucleotides, i.e. 1.107V and 1.120. The corresponding lane in Figure 15 confirms the presence of the full sized linear product.

Lanes 2 and 3 (Figure 15) show bands corresponding to full sized linear products cleaved and joined at one end with TelN or VP58.5 protelomerase respectively. The release of the very short hairpin protelomerase sequences from this reaction are also clearly evident in Figure 14, Lanes 2 and 3 and confirm protelomerase activity.

Treatment of the full sized linear product with both protelomerases TelN and VP58.5 produces a covalently closed linear DNA product, dbDNA (Figure 15, Lane 4). Confirmation of this is indicated in Figure 15, Lane 5 which shows a single band only, corresponding to the correct size of the desired dbDNA product. All other open ended linear products have been completely hydrolysed by exonuclease T5 which was added to reactions at a concentration of 0.2 units/μΙ, in lx NEB buffer 4. T5 exonuclease attacks both single and double stranded DNA with blunt or overhanging ends and from both 5' and 3' ends. dbDNA is resistant to exonuclease T5, as it is circular with no free ends (5' or 3') to attack. This is further confirmed in Lane 5, Figure 14 where, as expected, no fluorescent DNA /oligonucleotides are evident and all fluorescence is located in a band at the very bottom of the gel where it would be expected to find dNTPs. The fluorescent tags at each end of the full sized linear product have been cleaved off by the protelomerases leaving a non-fluorescent and therefore non-visible dbDNA when this imaging is used, whereas it can be seen using Gel ed as shown on Figure 15. The fluorescent hairpins released have been hydrolysed by T5 exonuclease to dNTPs.

The data confirms successful synthesis of the closed linear DNA product.

Example 3 - Enzymatic synthesis of covalently closed linear DNA (dbDNA) on a solid surface Reagents

The same reagents were used as indicated in Example 2 plus 5 nm gold nanoparticles

(Cytodiagnostics).

Name Length Description Sequence (5'-3')

3.1 81 nt Initiating oligonucleotide primer - I A I GGAAAAACGO.AGCAACGCGGCU 1 1 1 two thiol groups are linked to the TACGGTTCCTGGCU 1 1 I GU GGCU \ \ \ GC 5' phosphate group to allow TCA CA TGTAGA TCTTGTA CA attachment of the oligo to a

surface

3.2 200 nt First template oligonucleotide of GGGCGGGGGTCGTTGGGCGGTCAGCCAGG the series - 3' end has a 3 CG G G CC ATTTACCGT AAGTT ATGTA ACG CG nucleotide mismatch to its binding GAACTCCATATATGGGCTATGAACTAATGA site, preventing extension by CCCCGTAATTGATTACTACTTAAGTGTACAT polymerase ATC AG C AC AC AATAGTCC ATTATACG CG CG

TATA ATG G G C A ATTGTGTG CTG ATA TGTA C AAGA TCTACA TGTGAGC 1 1 1

3.3 200 nt Second template oligonucleotide of A TAA TGCCAGGCGGGCCA TTTA CCGTC ATT the series - 3' end has a 3 G ACGTC AATAG G G G G CGTACTTG G CAT AT nucleotide mismatch to its binding G ATAC ACTTG ATGT ACTG CC AAGTG G G C AG site, preventing extension by TTTACCGTAAATACTCCACCCATTGACGTCA polymerase ATG G AA AGTCCCT ATTG G CGTTACTATG G G

AACATACGTCATTATTGACGTCAATGGGCG GGGGTCGTTGGGCGGTCTCG

3.4 200 nt Third template oligonucleotide of GCCAAAACAAACTCCCATTGACGTCAATGG the series - 3' end has a 3 GGTGGAGACTTGGAAATCCCCGTGAGTCA nucleotide mismatch to its binding AACCGCTATCCACGCCCATTGATGTACTGCC site, preventing extension by AAAACCGCATCACCATGGTAATAGCGATGA polymerase CTAATACGTAGATGTACTGCCAAGTAGGAA

AGTCCC ATA AG GTC ATGTACTG GGCA TAA T

GCCAGGCGGGCCATTTAGGC

3.5 64 nt End capping oligonucleotide - CC ATTATACG CG CGTATAATG G G C A ATTGT includes sequence for TelN GTG CTG ATAG CCAAAA CAAA CTCCCA TTGA loopback extension CAG

Table 2: Oligonucleotides (Integrated DNA Technologies) encode the eGFP gene sequence flanked by a CMV promoter and poly-A tail, all within protelomerase TelN target sites. Oligonucleotide design

Oligonucleotide templates were designed to produce a covalently closed linear DNA (dbDNA) construct comprising a 525 base pair sequence, terminated at each end with hairpin loops comprising a portion of the protelomerase TelN target sequence.

Oligonucleotide 3.1 (with a length of 81 nucleotides) is modified with two thiol groups attached to the terminal 5' phosphate to allow covalent attachment to a prepared solid surface. Oligonucleotide 3.2 (with a length of 200 nucleotides) encodes the protelomerase TelN target sequence and a portion of the CMV promoter sequence, and oligonucleotides 3.3 and 3.4 (each 200 nucleotides in length) collectively encode a further portion of the CMV promoter sequence. Together, these three oligonucleotides act as sequential templates for extension of the first oligonucleotide, 3.1, which acts as a primer. The fifth oligonucleotide (3.5, with a length of 64 nucleotides) encodes a portion of the protelomerase TelN target sequence. This oligonucleotide does not contain the full protelomerase site, but rather encodes just over half - this will allow the extended strand to 'loop back' on itself to enable extension to form a double stranded oligonucleotide. The loop thus formed is the same sequence as the hairpin loop produced by successful protelomerase TelN cleavage of its target site. Each oligonucleotide template contains an overlap at its 3' end with the 3' end of the extending oligonucleotide primer (sequences underlined and italicised mark these overlapping sections). The last 3 nucleotides at the 3' end of oligonucleotides 3.2 to 3.5 are not complementary to the extending strand to prevent their unwanted extension by DNA polymerase.

Immobilization of oligonucleotide 3.1 on 'gold nanoparticles'

De-protection of the thiolated oligonucleotide, and linkage of that oligonucleotide to gold nanoparticles, was performed as directed by the manufacturer's instructions for the Cytodiagnostics ^® 5nm Oligo EADY Gold Nanoparticle Conjugation Kit. De-protected oligonucleotide was separated from DTT using a G-25 column from GE Healthcare, and linked oligonucleotide was purified by repeated (10 x) washes in ddH ₂ 0 in a 30 kDa spin concentrator column (Millipore) by centrifugation at 4,500g. Reaction conditions

The extension reaction by thermal cycling was carried out as described in Example 2 using the oligonucleotide templates described above. However, in this experiment, the initiating oligonucleotide primer was covalently attached to a gold nanoparticle such that the extension reaction was carried out on a surface. The reaction produces an immobilised dsDNA molecule comprising sequences from the 5 oligonucleotides. The end of the molecule proximal to the gold nanoparticle incorporates a full TelN protelomerase target sequence while the distal end incorporates a hairpin sequence identical to that produced by TelN cleavage of its target site. The product of the extension reaction when treated with TelN protelomerase is therefore a covalently closed linear DNA (dbDNA) cleaved from the gold nanoparticle.

Protelomerase TelN reactions were performed in the same buffer as the extension reaction i.e., the protelomerase enzyme was added directly to the cycling reaction upon its completion. Protelomerase was added to a final concentration of 0.5 μΜ. Reactions were incubated at 37°C for 15-30 minutes, and inactivated at 75°C for a further 10 minutes. Exonuclease T5 was added to reactions at a concentration of 0.2 units/μΙ, in lx NEB buffer 4 (50mM potassium acetate; 20mM Tris-acetate; lOmM magnesium acetate & ImM DTT).

Gel electrophoresis

The illustrated gel in Figure 16 is a native 10% TBE PAGE gel, run at 180 volts in lx TBE buffer. Imaging was performed with excitation at a wavelength of 300nm after staining with 3 times concentrated GelRed™ DNA stain for 1 hour to image all the DNA/oligonucleotides present. As shown in Figure 16, dbDNA can be produced by the method described in Example 2 but with the starting oligonucleotide 3.1 immobilised on a surface of a 5nm diameter. Much of the DNA in the reaction is impeded from passing into the agarose by its fusion to the gold nanoparticle (Lane 3). However, dbDNA, when released by TelN cleavage and cleaned up by treatment with T5, runs as a single exonuclease resistant band to a position expected by its size (Lane 7).

Example 4 - Synthesis of circular single stranded DNA covalently closed by the action of a protelomerase

This example describes the synthesis of a single stranded circular DNA containing a

protelomerase target sequence (referred to as IbDNA). This structure may be exploited in the same way as doubled stranded covalently closed linear DNA but with the potential advantage that the single stranded loop is permanently open to transcription into RNA. In addition, the single stranded loop section may be designed to encode an aptamer sequence with specific molecular binding properties for diagnostic and medicinal applications.

A circular single stranded DNA has been synthesized from a 159 nucleotide long

oligonucleotide encoding an aptamer sequence reported to inhibit metallo- -lactamase enzyme activity in Bacillus cereus (Kim. S-K et al, 2009, Chemical Biology & Drug Design, 74(4), 343-348.) The sequence of the oligonucleotide and process for conversion into a circular single stranded construct is outlined in Table 3 and Figures 17a to c.

Table 3: Sequences of IbDNA precursor oligonucleotides. Reverse complementary sequences are underlined, aptamer sequence is in italics, spacer sequences are in bold. Supplied by Integrated DNA Technologies

A single oligonucleotide may be extended and processed to IbDNA

A 159 nt oligonucleotide consisting of a protelomerase site followed by an aptamer sequence flanked by two reverse complementary sequences was designed. Two variants were prepared, DNA-T and DNA-V, where the incorporated protelomerase target sites were for TelN and VP58.5 respectively (Table 3).

The self-complementary sequences within the oligonucleotide can bind to each other under the right conditions to form a short double stranded sequence, the 3' end of which may be extendable by a DNA polymerase. Extension of the 3' end results in the formation of a full protelomerase site which may be cleaved and joined (by the appropriate protelomerase enzyme) to form the desired circular covalently closed single stranded construct (IbDNA) and a short waste hairpin sequence. Confirmation of a successful reaction is made by treatment with T5 exonuclease as described previously and identification of the surviving band by gel

electrophoresis (see Figures 18a and 18b). Reaction conditions

The 159nt oligonucleotides supplied by IDT had been subjected to standard desalting purification and were diluted to a stock concentration of 200 μΜ in double distilled water. Each oligonucleotide was then diluted to 3 concentrations - 0.1, 0.5 and 1 μΜ - in 50μΙ reaction volumes also containing:

IAB buffer @ lx concentration, dNTP mix @ 100 μΜ, 1 unit of Phusion ^® HF DNA polymerase These reactions were cycled as follows: 2 minutes @ 95°C; 35 cycles comprising: 30 seconds @ 95°C then ramping down at 2°C/sec to 30 seconds @ 55°C, 30 seconds @ 72°C; 10 minutes @ 72°; 4°C END

These reactions were split into 10 μΙ aliquots with NEB buffer 4 added to a final concentration of (lx) and where indicated, protelomerase added to a final concentration of 2.5 μΜ, to ensure it was in excess. All the aliquots were then incubated for 15 minutes at 37°C.

5 units of T5 exonuclease were then added to indicated aliquots in the presence and absence of protelomerase and incubated at 37°C for 30 minutes. The temperature was then raised to 75°C for 15 minutes to heat inactivate the enzymes. Table 4 depicts the enzymes added to each aliquot:

5 μΙ of these reactions were loaded and run on 10% TBE PAGE gels, as shown in Figures 18a and 18b. Figure 18a (LH) shows that in the absence of any enzyme, the oligonucleotides ran at their expected size but in the presence of T5 exonuclease, they were completely hydrolysed (Figure 18a RH).

When exposed to protelomerase, a waste 'cap' or by-product band appeared as well as a corresponding product band (Figure 18b LH). This confirmed the loopback extension and successful formation of the protelomerase target site (Figure 18b LH). Subsequent exposure to exonuclease (Figure 18b RH) revealed a surviving band representing covalently closed circular DNA (IbDNA). This is in contrast to the linear oligonucleotides exposed directly to exonuclease, which were fully hydrolysed (Figure 18a RH).