POLYMERASE
; present invention relates to DNA polymerases. In particular the invention relates to a thod for the generation of DNA polymerases which exhibit a relaxed substrate cificity. Uses of engineered polymerases produced using the methods of the invention also described.
Background:
curate DNA replication is of fundamental importance to all life ensuring the intenance and transmission of the genome and limiting tumorigcncsis in higher

anisms. High-fidelity DNA polymerases perform aa astonishing feat of molecular Dgnition, incorporating the correct micleotide triphosphate (dNTP) substrate molecules
as
specified by the template base with TniTiimal error rates. For example, even without nucleolytic proofreading, the replicative DNA polymerase HI from E.coli on average y makes one error in ~105 base pairs (Schaaper JBC1993).
As energetic differences between correctly and mispaired nucleotides per se are m ch too small to give rise to a 10s fold discrimination, the structure of the polymerase acjtave site in high-fidelity polymerases has evolved to enhance those differences. Recent st ictural studies of the A-family (Pol I-like) DNA polymerases from Thermits aquaticus 0 iq) (Li 98), phage 17 (Ellenberger) and B. stearothermophilus (Bst) (Beese) in ps ticular have revealed how confbrmational changes during the catalytic cycle may explude non-cognate base-pairing geometries because of steric clashes within the closed active site. As a result of these tight steric constraints, not only are mismatched mfcleotides excluded but catalysis becomes exquisitely sensitive to even slight distortions in! the primer-template duplex. This precludes or greatly diminishes the replication of m dified or damaged DNA templates, the incorporation of modified or unnatural
de >xinucleotide triphosphates (dNTP) and the extension of mismatched or unnatural 3' tejrnini
While desirable in nature, such stringent substrate discrimination is limiting for miny applications in biotechnology. Specifically, it restricts the use of unnatural or m dified nucleotide bases and the applications they enable. It also precludes the efficient P( R amplification of damaged DNA templates.
Some other naturally occurring polymerases are less stringent with regard to their su istratc specificity. For example, viral reverse transcriptases like HIV-1 reverse to iscriptase or AMY reverse transcriptase and polymerases capable of translesion sy ithesis such as polY-family polymerases, pol X (Vaisman et al, 2001, JBC) or pol X (V ashington (2002), PNAS; or the unusual polB-family polymerase pol X (Johnson, N< ture), all extend 3' mismatches with elevated efficiency compared to high fidelity pc ymerases. The disadvantage of the use of translesion synthesis polymerases for
bitechnological uses is that they depend on cellular processivity factors for their activity,
su h as PCNA, Moreover such polymerases are not stable at the temperatures at which ce tain biotechnological techniques are performed, such, as PCR. Furthermore most Ti nslesion synthesis polymerases have a much reduced fidelity, which would severely co apromise their utility for cloning.
ha by
Using another approach, the availability of high-resolution structures has guided efforts to rationally alter the substrate specificity of high fidelity DNA polymerases by Sit -directed mutagenesis e.g. to increase acceptance of dideoxi- (ddNTPs) (Li 99) or ril mucleotides (rNTPs) (Astafke 98). In vivo complementation followed by screening
also yielded polymerase variants with increased rNTP incorporation and limited ass of template lesions (Patel 01). Recently, two different in vitro strategies for
se *tion of polymerase activity have been described (Jestin 00, Ghadessy 01, Xia 02). Oi 3 is based on the proximal attachent of polymerase and template-primer duplex on the sai ie phage particle and has allowed the isolation mutants of Taq polymerase, which inc arporate fNTPs and dNTPs with comparable efficiency (Xia 02). However, such rat hods are complex, prone to error and are laborious.
Recently, the technique of compartmentalized self-replication (CSR) (Ghadessy , which is based on the self-replication of polymerase genes by the encoded ymerases within discrete, non-communicating compartments has allowed the selection mutants of Taq polymerase with increased thermostability and/or resistance to the ent inhibitor heparin (Ghadessy et al 01).
H< wever, there still remains a need in the art for an efficient and simple method for re] ixing the substrate specificity of high fidelity DNA polymerases whilst maintaining hi h catalytic turnover and processivity of DNA fragments up to several tens of kb. Such ymerases will be of particular use in applications such as PCR amplification and se uencing of damaged DNA templates, for the incorporation of unnatural base an ilogues into DNA (such as is required for sequencing or array labelling) and as a sfc ting point for the creation of novel polymerase activities using compartmentalised self re lication or other methods.
Si mmary of the invention
Ti s present inventors modified the principles of directed evolution, (in particular co apartmentalised self replication) described in GB97143002, 986063936 and GB 01) 275643 in the name of the present inventors, to relax the steric control of high fidelity A polymerases and consequently to expand the substrate range of such polymerases. Al of the documents listed above are herein incorporated by reference.
Ti ey surprisingly found that by performing the technique of compartmentalised self re lication referenced above, using repertoires of randomly mutated Taq genes, and fli iking primers bearing the mismatches A*G and C*C at their 3' terminus/end, then m tants were generated which not only exhibited the ability to extend the A*G and C*C to [version mismatches used in the CSR selection, but also surprisingly exhibited a ge leric ability to extend roispaired 3' termini. This finding is especially significant since polymerase is not able to extend 3' mismatches (Kwok wt al, (1990), Huang (1992).
! mutant polymerases generated also exhibit high catalytic turnover, concomitant -with high fidelity polymerases and are capable of efficient amplification of DNA lents up to 26kb.
in a first aspect the present invention provides a method for the generation of an engineered DNA polymerase with an expanded substrate range which comprises the step of preparing and expressing nucleic acid encoding an engineered DNA polymerase uti teing template nucleic acid and flanking primers which bear one or more distorting 3'
As! herein defined 'flanking primers which bear a 3' distorting terminus/end" refer to the :e primers which possess at their 3' ends one or more group/s, preferably mi leotide group/s which deviate from cognate base-pairing geometry. Such deviations fro a cognate base-pairing geometry includes but is not limited to: nucleotide mi matches, base lesions (i.e. modified or damaged bases) or entirely unnatural,

ad' 3'
syi thetic base substitutes.. According to the above aspects of the invention,
antageously, the flanking primer/s bear one or more nucleotide mismatches at their nd/terminus.
['antageously, according to the above aspects of the invention the flanking primers have one, two, three, four, or five or more nucleotide mismatches at the 3' primer
end More advantageously, me one or more nucleotide mismatches are consecutive

mifnatcb.es. More advantageously, according to the above aspects of the invention, the

:eotide mismatch at their 3' primer end.
nut
flanking primers have one or two nucleotide mismatches at me 3 primer end. Most bly according to the above aspects of the invention, the flanking primers have one
Mo e specificallythe term 'distorting 3* termini/ends' includes within its scopethe
phe
lomenon f whereby, for example, either the 3' terminal base ( 1-mismatch) or the 3
ten inal and upstream base (2-mismatch, 3-mismatch, 4-mismatch and so on) are not COB plementary to me template base. Preferably mismatches are transversion mismatches apposing purioes with purines and pyrimidines with pyrimidines. Preferably
tn aversion mismatches are G.A and C.C. This type of primer terminus distortion is ared to herein as 'primer mismatch distortion.
addition, and as eluded to above, the term 'flanking primers bearing distorting 3* tea aini/ends' includes within its scope flanking primers beating one or more unatural
e analogues at the 3' termini/end of the one or more flanking primers so that distortion (he cognate DNA duplex geometry is created.
s method of the invention may be used to expand the substrate range of any DNA ymerase which lacks an intrinsic 3-5' exonuclease proofreading activity or where a 3-exonuclease proofreading activity has been disabled, e.g, through mutation . Suitable A polyracrases include polA, polB (see e,g. Pairel & Loeb, Nature Struc Biol 2001) C, polD, polY, polX and reverse transcriptases (RT) but preferably are processive, i-fidelitypolymerases.
Ivantageously, an engineered DNA polymerase with an expanded substrate range according to the invention is generated from a pol A-family DNA polymerase.
vantageously, the DNA polymerase is generated from a repertoire of pol A DNA pc ymerase nucleic acid as template nucleic acid. Preferably the pol A polymerase is Taq p< ymerase and the flanking primers used in the generation of the polymerase are one or m re of those primers selected from the group consisting of the following: S'-CAG GAA
A GCT ATG ACA AAA ATC TAG ATA ACG AGO GA-3';A^G mismatch); 5'GTA A&A CGA CGG CCA GTA CCA CCG AAC TGC GGG TGA CGC CAA GCC-3' C*C
latch

M

re advantageously, according to the above aspect of the invention, the nucleic acid

-oding the engineered polymerase according to the invention is generated using PCR using one or more flanking primers listed herein.

A(
CO

rentageously, the method of the present invention involves the use of tpartmentalised self replication, and consists of the steps listed below:at
(a) preparing nucleic acid encoding a engineered DNA polymerase, wherein the
polymerase is generated using a repertoire of nucleic acid molecules encoding
one or more DNA polymerases and flanking primers which bears a 3'distorting
end.
(b) compartmentalising the nucleic acid of step (a) into microcapsules;
(c) expressing the nucleic acid to produce their respective DNA polymerase
within the microcapsules;
(d) sorting the nucleic acid encoding the engineered DNA polymerase which
exhibits an expanded substrate range; and
(e) expressing the engineered DNA polymerase which exhibits an expanded
substrate range.
st advantageously, the method of the invention comprises the use of one or more A polymerases and flanking primers which bears one or more nwleotide mismatches 'primer ends.

Ac wording to the above aspects of the invention, the term 'engineered DNA po ymerase' refers to a DNA polymerase which has a nucleic acid sequence which is no 100% identical at the nucleic acid level to the one or more DNA polymerase/s or frs jrnents thereof, from which it is derived, and which is synthetic. According to the in' ention, an engineered DNA polymerase may belong to any family of DNA po ymerase. Advantageously, an engineered DNA polymerase according to the i« eation is a pol A DNA polymerase, As referred to above the term 'engineered DNA po ymerase' also includes within its scope fragments, derivatives and homologues of an engineered DNA polymerase' as herein defined so long as it exhibits the requisite
pn >erty of possessing an expanded substrate range as defined herein, in addition, it is an essential feature of the present invention that an engineered DNA polymerase act ording to the invention does not include a polymerase with a 3-5' exonuclease act vity under the conditions used for the polymerisation reaction. (This definition inc udes polymerases ia which the 3-5' exonuclease is not part of the polymerase po: /peptide chain but is associated non-covalentty with the active polymerase). Such a prqpfreading activity would remove any 3' mismatches incorporated according to the

m sthod of the invention, and thus would prevent a. polymerase according to the ration possessing an expanded substrate range as defined hereto.
defined herein the term 'expanded substrate range (of an engineered DNA lymerase) means that substrate range of an engineered DNA polymerase according the present invention is broader than that of the one or more DNA polymerases, or ft gments thereof from which it is derived. The term 'a broader substrate range' refers the ability of an engineered polymerase according to the present invention to extend e or more 3'distorting ends, advantageously transversion mismatches (E nine*purine, pvrirnidmepyrimidine) for example AA, CC, GG, TT and GA, \v rich the one or more polymerase/s from which it is derived cannot extend. That is, ei ;entially, a DNA polymerase which exhibits a relaxed substrate range as herein d fined has the ability not only to extend the 3' distorting endsused in its generation, those of the flanking primers) but also exhibits a. generic ability to extend 3 dfttorting ends (for example AG, AA, GG mismatches). Preferably, 'expanded si t»strate range' (of an engineered DNA polymerase) includes a wider spectrum of u natural nucleotide substrates including ocS dNTPs, dye-labelled nucleotides, aaged DNA templates and so on. More details are given in the Examples.
A wording to the above aspect of the invention advantageously the DNA polymerase g aerated using CSR technology is a pol A polymerase and it is generated using fl along primers selected from the group consisting of the following: S'-CAG GAA
A :A GOT ATG ACA AAA ATC TAG ATA ACG AGG GA-3';&£ mismatch); 5' JTA AAA CGA CGG CCA GTA CCA CCG AAC TGC GGG TGA CGC CAA

G

C-3' C*C mismatch.

ic skilled in the art will appreciate mat in essence, any DNA polymerase flanking pi) met which incorporates a 3' mismatch will work with any suitable repertoire. The pi >cess of mismatch extension will vary in characteristics from polymerase to p( lymerase, and will also vary according to the experimental conditions. For example,

A and C*C are the most disfavoured mismatches for extension by Taq polymerase (K lang et al, 92). Other mismatches are favoured for extension by other polymerases an I this can be routinely determined by the skilled person.
e skilled in the art wfll also appreciate that it is an essential feature of the present ention that the methods described herein will only work far polymerases which are /old of 3-5' exonuclease activity proofreading under the conditions used for the ymerisation reaction, as such activity would result in the removal of the orporated mismatches.
U ing the method of the invention, the present inventors generated a number of pol A pc ymerase mutants. Two of me mutants named Ml and M4 not only exhibit the al lity to extend the G*A and C*C traasversion mismatches used in the CSR selection, bi also surprisingly exhibit a generically enhanced ability to extend 3' mismatched te nini.
Tl us in a further aspect the present invention provides an engineered DNA polymerase w ich exhibits an expanded substrate range. Preferably such an engineered polymerase is ibtainable using one or more method/s of the present invention.
x>rding to the above aspect of the invention, preferably the DNA polymerase is a poj A polymerase.
ording to the above aspect of the invention, preferably the engineered DNA pc ymerase is obtained using the method of the invention.
In a further aspect still, the present invention provides a pol A DNA polymerase with an ex landed substrate range, or the nucleic acid encoding it, wherein the DNA polymerase is lesignated Ml or M4 as shown in fig 1 and fig 2 respectively and depicted as SEQ No 1 i ad SEQ No 2 respectively.
Ac x>rding to the above aspect of the invention, preferably the engineefed DNA ymerase as herein defined is that polymerase designated Ml in fig 1 and depicted SI 3 No 1.
In ret a farther aspect the invention provides a pol A DNA polymerase with an expanded su strate range, wherein the polymerase exhibits at least 95% identity to one or more of th amino acid sequences designated Ml and M4 as shown in fig 1 and fig 2 respectively an depicted SEQ No 1 and SEQ No 2 respectively and which, comprises any one or m re of me following mutations: E520G, D144G, L254P, E520G, E524G, N583S, 1.1-D 44G, L254P, E520G, E524G, N5S3S, VI 131, A129V, L245R, E315K, G364D, G 03R, E432D. P4S1A, I614M, R704W, D144G, G370D, E742G, K56E, I63T, KJPTR, M3171; Q680R, R343G, G370D, E520G, G12A, A109T, D251E, P387L,
08V, R617K, D655E, T710N, E742G, A109T, D144G, V155I, P298L, G370D, 16 4M, E694K, R795G, E39K, R343G, G370D, E520G, T539A, M747V, K767R, Gi 4A, D144G, K314R, E520G, F598L, A608V, E742G, DS8G, R74P, A109T, L245R,
43G, G370D, E520G, N583S, E694K, A743P.
vantageously, Hie invention provides a pol A DNA polymerase with an expanded su istrate range, or the nucleic acid encoding it, wherein the polymerase exhibits at least

'o identity to one or more of me amino acid sequences designated Ml and M4 as
sb wn in fig 1 and fig 2 respectively and depicted SEQ 1 and 2 respectively and which co uprises any one or more of the following mutations: G84A, D144G, K314R, E520G, FS >8L, A608V, E742G, D58G, R74P, A109T, L245R, R343G, G370D, E520G, N583S, EO4K.A743P.
M st advantageously, the invention provides a pol A DNA polymerase with an expanded su strate range, or the nucleic acid encoding it, wherein the polymerase exhibits at least 95 i identity to one or more of the ammo acid sequences designated Ml and M4 as sh wn in fig 1 and fig 2 respectively and depicted SEQ 1 and 2 respectively and which co iprises any one or more of the following mutations: G84A, D144G, K314R» E520G, F5>8L,A608V,E742G.

cording to the above aspect of the invention the mutation 'E520G' describes a DNA p< tymerase according to the invention in which glycjne is present at position 520 of the ai iao acid sequence. The present inventors were surprised to find that E520, which is lo ;ated at the tip of the thumb domain at a distance 20A from the 3'OH of the
smatched primer terminus, would be involved in mismatch recognition or extension .
Tl e mutation of E520 to CS20 is clearly important in such roles however as the present in 'enters have demonstrated. This aspect of the invention is described further in the dt tailed description of the invention.
T e present inventors consider that the method of the invention is applicable to the g< icration of 'blends' of engineered DNA polymerases with an, expanded substrate ra ige, According to the present invention the term a 'blend' of more than one polymerasc re «rs to a mixture of 2 or more, 3 or more 4 or more, S or more enginerred polyaaerases. P jferably the term 'blends' refers to a mixture of 6, 7, 8, 9 or 10 or more 'engineered ymerases'.
It is important to note that the extension of mismatched 3' primer termini is a feature of n« totally occurring polymerases. Viral reverse transcriptases (RT) like HTV-l RT or RT and polymerases capable of translesion synthesis (TLS) such as the polY-fe nily polymerases pol i (Vaisman 2001JBC) or pol x (Washington 2002 PNAS) or the ui usual polB-family polymerase pol£ (Johnson Nature), all extend 3' mismatches with el vated efficiency compared to high-fidelity polymerases. Thus, the mutant polA p lymerases according to the present invention share significant functional similarities w th other polymerases found in nature but so far represent, the only known member of
thpolA-femily polymerases that are proficient in mismatch extension (ME) and
tn oslesion synthesis (TLS) In
contrast to TLS polymerases, which are distributive and depend on cellular pi >cessivity factors such as PCNA, Ml and M4 combine mismatch extension (ME) and tn nslesion synthesis (TLS) with high processivity and in the case of Ml are capable of efficient amplification of DNA fragments of up to 26kb.

In a farther aspect still the present invention provides a nucleic acid construct which is a sable of encoding a pol A DNA polymerase which exhibits an expanded substrate r< ige, wherein said pol A DNA polymerase is depicted in fig 1 and fig 2 as SEQ No 1 or S] Q No 2 and is designated Ml and M4 respectively.
A cording to the above aspect of the invention, preferably the nucleic acid construct ei sodes the Ml pol A polymerase as described herein.
Ir a further aspects the invention provides a pol A DNA polymerase with an expanded si jstrate range, in particular •which is capable of mismatch extension, wherein the polymerase comprises, preferably consists of the amino acid sequence of any e or more of the clones designated herein as 3B5,3BS, 3C12 and 3D1.
tb
yet a further aspect the invention provides a pol A DNA polymerase with an ej panded substrate range, in particular which is capable of abasic site bypass, wherein : DNA polymerase comprises, preferably consists of the amino acid sequence of any e or more of the clones designated herein as 3A10,3B6 and 3811.
a further aspect still the inveation provides a pol A DNA polymerase with an ejj landed substrate range, in particular which is capable of DNA replication involving incorporation of unatural base analogues into the newly replicated DNA, wherein pol A DNA polymerase comprises, preferably consists of the amino acid sequence,
any one or more of the clones designated herein as 4D11 and 5D4.

oi
a further aspect the present invention provides a pol A DNA polymerase with an anded substrate range, wherein the polymerase exhibits at least 95% identity to one more of the amino acid sequences designated 3B5, 3B8, 3C12, 3D1, 3A10, 3B6, 31 1174D11 and 5D4. which comprises any one or more of the mutations (with respect to either of the three parent genes Taq, Tth, Tfl) or gene segments found in clones 31 5, 3B8, 3C12, 3D1, 3A10, 3B6, 3B11, 4D11 and 5D4.12
ID a further aspect still, the present invention provides a vector comprising a nucleic acid cc astroct according to the present invention.
In a further aspect still the present invention provides the use of a DNA polymerase jording to the present invention in any one or more of the following applications ected from the group consisting of the following: PCR amplification, sequencing of
ds naged DNA templates, the incorporation of unnatural base analogues into DNA and creation of novel polymerase activities.
A cording to the above aspect of the invention, preferably the use is of a 'blend' of DNA lymerases according to the invention or selected according to the method of the in] 'entjon. The use of blends of polymerases will be familiar to those skilled in the art and is Jescribed in Barnes, W. M. (1994) Proc. Natl. Acad. Sri. USA 91,2216-2220 which is he ein incorporated by reference,
cording to the above aspect of the invention, preferably the DNA polymerase is a pol DNA polymerase. Advantageously, it is generated using CSR technology using fli nlcing primers bearing one or more 3' mismatch pairs of interest as described herein.
0 icr suitable methods include screening after activity preselection (see Patel & Loeb
1 f and phage display with proximity coupled template-primer duplex substrate (Jestin
01
Xue, 02. CST is also ideally suited as the present inventors have demonstrated.
At cording to the above aspect of the invention, preferably the use of a polymerase ac ording to the invention is in PCR amplification and the polymerase is Ml as herein de cribed.
A( Jording to the above aspect of the invention, advantageously, the creation of novel pc ymerase activities is produced using the technique of compartmentalised self
D initions

term 'engineered DNA polymerase' refers to a DNA polymerase which has a nujleic acid sequence which is not 100% identical at the nucleic acid level to the one or m< re DNA polymerase/s or fragments thereof, from which it is derived, and which has be n generated using one or more biotechnological methods. Advantageously, an en ineered DNA polymerase according to ,the invention is a pol-A family DNA polymerase or a pol-B family DNA polymerase. More advantageously, an engineered DMA polymerase according to the invention is a pol-A family DNA polvmerase.As rejprred to above the term 'engineered DNA polymerase' also includes within its scope fh gmeots, derivatives and homobgues of an 'engineered DNA polymerase' as herein de ined so long as it exhibits the requisite property of possessing an expanded substrate ra ge as defined herein. In addition, it is an essential feature of the present invention that engineered DNA polymerase according to the invention does not include a polymerase h a 3-5' cxonuclease activity under me conditions used for the polymerisation re ction. Such a proofreading activity would remove any 3' mismatches incorporated ac ording to the method of the invention, and thus would prevent a polymerase according
tohe invention possessing an expanded substrate range as defined herein.herein defined 'flanking primers which bear a 3* distorting terminus' refer to se DNA polymerase primers which possess at their 3' ends one or more group/s, ferably nucleotide group/s which deviate from cognate base-pairing geometry. Such
A
th
Pi
d< dations from cognate base-pairing geometry includes but is not limited to:
m ;leotide mismatches, base lesions (i.e. modified or damaged bases) or entirely ut latural, synthetic base substitutes at the 3 end of a flanking primer used according to th methods of the invention. According to the above aspects of the invention,
•antageously, the flanking primer/s bear one or more nucleotide mismatches at their 3'jjend, Advantageously, according to the above aspects of the invention the flanking may have one, two, three, four, or five or more nucleotide mismatches at the 3'jjprirner end. Preferably according to the ahove aspects of the invention, the flanking ;ers have one or two nucleotide mismatches at the 3' primer end. Most preferably
(ording to the above aspects of the invention, me flanking primers have one nvjfcleotide mismatch at their 3' primer end.

to

defined herein the term 'expanded substrate range (of an engineered DNA lymerase) means that substrate range of aa engineered DNA polymerase according the present invention is broader than that of the one or more DNA polymerases, or

gments thereof from which it is derived. The tenn 'a broader substrate range* refers
ined has the ability not only to extend the 3* distorting ends used in its generation, those of the flanking primers) but also exhibits a generic ability to extend 3"
to fthe ability of an engineered polymerase according to the present invention to extend onjfc or more 3'distorting ends, advantageously transversion mismatches (pi rine*purine> pyrimidme*pyrimidijie) for example A*A, C*C, G*G, T*T and G*A, l ich the one or more polymerase/s from which it is derived cannot extend. That is, es sntially, a- DNA polymcrase which exhibits a relaxed substrate range as herein de IB
disforting ends(for example AG, AA, G*G mismatches).

Bi

ef description of thefigures

are 1 shows the Ml nucleic acid (a) and amino acid sequence (b) Fij are 2 shows the M4 nucleic acid (a) amino acid sequence (b)
Fi] are 3 shows the general scheme of mismatch extension CSR selection. Self- ; rep ication of the pol gene by the encoded polymerase requires extension of flanking pit iers bearing G'A and DC 3' mismatches. Polymerases capable of mismatch exl insion (Pol) replicate their own encoding gene (po/*), while Pol" cannot extend mil Hatches and fails to self-replicate. Black bars denote incorporation of the mi match into replication products.
4. Mismatch extension properties of selected polymerases. (A) Polymerase
ana
irity in PCR for matched 3' ends and mismatches. Only mutant polymerases M4
Ml (not shown) generate amplification products using primers with 3'
trai {version mismatches. (B) Mismatch extension PCR assay. Mismatch extension cap bility is expressed as arbitrary mismatch extension units (ratio of polymerase act^ity in PCR with matched vs. mismatched flanking primers). Different

vmerases (black diamonds) and derivatives (open squares, triangles) are shown in se larate columns.

th bt
{. 5. Lesion bypass activity (A) vrtTaq, (B) Ml, (Q M4. Each polymerase was as ayed over time for its ability to extend a radiolabeled primer annealed to either an m lamaged template, or a template containmg an abasie she or a cis-syn cyclobutane
mine-thymine dimer (CPD), Template sequence was identical except for three es located immediately downstream of the primer (Nl-3). The local sequence
context in the Nl-3 region is given on the right haad side of each respective panel abasic site; T-T = CPD.
Fi;. 6, Polymerase activity on unnatural substrates. (A) Polymerase activity in PC R using all oS dNTPs. cxS DNA amplification products of 0.4kb, 0.8kb and 2kb, an obtained with Ml but not with vrtTaq (vrt). $X, Haeffl-digested phage X174 A marker. XH, Hindm-digested phage X DNA marker. (B) Polymerase activity in PdR with complete replacement of dATP with FITO-12-dAT? (teft) or dTTP with

Bi

ti»-16-dUTP (right). Only Ml yields amplification products. M, 1kb DNA ladder

vitrogen). (C) Bypass of a 5-nitroindol template (5NI) base. Polymerase activity assayed over time for its ability to extend a radiolabeled primer annealed to a template containing a 5NI template base.
7. Long range PCR. PCR amplification of fragments of increasing length from a phjlge. DNA template. VftTaq (wt) fails to generate amplification products larger thj i 8.8kb while Ml is able to amplify fragments of > 25kb. AH, HbdHI-digested ph ge A. DNA marker.ext8. Hairpin-ELlSAs to test nucleotide analogue incorporation by mismatch nsion clones.clo?.. Clones 3B5. 3B8,3C12 and 3D1 (where 3 indicates that these are third round es) were able to extend primers containing four mismatches.

110. A list of polymerases selected to extend four mismatches were assayed for ir ability to extend abasic sites in PCR.
{nre 11. Seven polymerases were assayed for their ability -to bypass abasic sites in
rimer extension assay.



ure 12.. Several samples of cave hyena (Crocuta spelaea) were extracted and

ai lysed.

FJ If

we 13, Appropriate primers for use in the method of the invention. See example for details.




ure 14. Polymerases selected for replication of 5NI were tested for activity with a ge of substrates using the hairpin ELISA assay described in example 8. See mple 16 for details.



Fi
raj
tin

ure 15. Polymerases selected for replication of 5NI were tested for activity with a je of substrates . Polymerase 4DU. P is primer, Ch is the chase reaction. Reaction ss in minutes. See example 1 6 for details.



Fi;
rai tin

are 16. Polymerases selected for replication of 5NI were tested for activity with a ;c of substrates Polymerase 5D4. P is primer, Ch is the chase reaction. Reaction :s in minutes. See example 16 for details.

Fi; ure 17. Polymerases selected for replication of 5NI were tested for activity with a rai ge of substrates Polymerase 4D11. P is primer, Ch is the chase reaction. Reaction

tin

$ in minutes. See example 16 for details.

are 18.. Polymerases selected for replication of SNI were tested for activity with a na »e of substrates Polymerase 504. P is primer, Ch is the chase reaction. Reaction tin es in minutes. See example 16 for details.

Fi
cc
m

ore 19. Microarray hybridisations of FITC-Iabelled probes. Microarrays tained 5 replicate features of serial dilutions of Tag, RT and genomic sahnon sperm A target sequences, as indicated. Labelled randomers were used to visualise the Toarray and assess the availability of target sequences for hybridisation, Array co-ridisations were performed with a Cy5-label]ed Taq probe (Cy5^q), as a reference, equivalent uniabelled or FITC-labelled probes



H to

. Single examples from 3 replicate experiments are displayed for each co-
ridisation.



Fi
flu
eai
pr<

ore 20, Figure 21. Microarray signals from FITC-labelled probes. Mean FITC rescence signal of FITC-labelled probes (FITClOTiuj, HTClOMi, FirC100Mi) for i co-hybridisation is plotted against the CyS fluorescence signal of the .reference c (CySraq) for A) Taq, B) RT and C) genomic sahnon sperm DNA target

se< uences, as indicated. D) Microarray background signals from FITC-labelled probes

are
C>
by
fea
co-

Jetcrmined using 3 replicate microarrays for each co-hybridisation experiment of a -labelled Taq probe (CyStiq), as a reference, and unlabeled or FITC-labelled es (FTTCKh,, FITC10M,, HTClOOim). Background inlbrmation -was generated leasuring fluorescence signal from 12 non-feature areas of each microarray. Mean intensities were generated and used to derive a ratiometric value for each non-ire area. A mean of the mean ratio +/- 1 standard deviation is displayed for each ybridisation experiment


ure 22. Fidelity. (A) MutS EUSA. Relative replication fidelity of wtTa?, MI and Mi was determined using mirtS ELISA of two different DNA fragments (either a
cb or 2.5kb region of the cloned Tag gene) obtained by PCR and probed at two dii ereni concentrations. (B) Spectra of nucleotide substitutions observed in PCR fra ;ments amplified with either wtTaq or ML Types of substitutions are given as % of otal substitutions (wtTo?: 48, Ml: 74). Equivalent substitutions on either strand
(e
. G->A, C->T) were added together (GC->AT). Observed-1 delections (wtTaq: 3, 1) are not shown.

Fi

ure 23. Processivity of vrtTaq, Ml and M4 was measured at three different

ymerase concentrations in the absence (A) or presence (B) of trap DNA. The pnj cessivity Jfor nucleotide incorporation at each position was variable but essentially

idi

tical for all three polymerases. For example, the probability of enzyme

du iociation is higher at positions 2-5 compared to positions 6 and 7 for all three

po of

unerases. In the presence of trap DNA (to ensure all primer extension is the result single DNA binding event) 13% of bound wtTaq, 28% of Ml and 15% of M4

exfnded primers to the end of the template. The termination probabilities for

po MA
M

tions 2 mrou^bi 5 varied from 15-25% for vrtlaq and Ml and from 13-35% for while at positions 6 and 7 the termination probability was 5% for wtrcapsules, the contents of each microcapsule must be isolated from the contents of mounding microcapsules, so that there is no or little exchange of the nucleic acids
ene products between the microcapsules over the timescale of the experiment

berwi
possi
prodi :ts of all other nucleic acids. However, even if the meoretically optimal situation
of, ot average, a single nucleic acid or less per microcapsule is not used, a ratio of 5,10,
le, since the gene product of an individual nucleic acid will be isolated from the
Seco id, the method of the present invention requires that there are only a limited uum er of nucleic acids per microcapsule. This ensures that the gene product of an indrv dual nucleic acid will be isolated from other nucleic acids. Thus, coupling berw en nucleic acid and gene product will be highly specific. The enrichment factor is great st with on average one or fewer nucleic acids per microcapsule, the linkage en nucleic acid and the activity of the encoded gene product being as tight as is
50,1 up or 1000 or more nucleic acids per microcapsule may prove beneficial in sorting a large library. Subsequent rounds of sorting, including renewed encapsulation with

21 differing nucleic acid distribution, will permit more stringent sorting of file nucleic acids.

Pr<

:erably, there is a single nucleic acid, or fewer, per microcapsule.

d, the formation and the composition of the microcapsules must not abolish the furl rion of the machinery the expression of the nucleic acids and the activity of the gene

prd

iucts.

Co]
sequentry, any microencapsulation system used must fulfil these three requirements.
appropriate systena(s) may vary depending on the precise nature of the requirements in e ich application of the invention, as will be apparent to the skilled person.
A t ide variety of microencapsulation procedures are available (see Benita, 1996) and ma; be used to create the microcapsules used in accordance with the present invention. Ind ed, more than 200 microencapsulation methods have been identified in the IxtJunre (Finch, 1993).

lipid also
Theie include membrane enveloped aqueous vesicles such as lipid vesicles (lip( somes) (New, 1990) and non-ionic surfactant vesicles (van Hal et al., 1996). The e are closed-membranous capsules of single or multiple bilayers of non- x>valently assembled molecules, with each bilayer separated from its neighbour by 2 i aqueous compartment. In the case of liposomes the membrane is composed of
molecules; these are usually phospholipids but sterols such as cholesterol may >e incorporated into the membranes (New, 1990). A variety of enzyme-catalysed
biocHemical reactions, including KNA and DNA polymerisation, can be performed with i liposomes (Qiafcrabarti et al., 1994; Obetholzer et al., 1995a; Oberholzer et al., 1995 >; Walde et al, 1994; Wick & Laisi, 1996).
With a membrane-enveloped vesicle system much of the aqueous phase is outside the vesic es and is therefore non-compartmentalised. This continuous, aqueous phase shou] 1 be removed or the biological systems in it inhibited or destroyed (for example, by digestion of nucleic acids with DNase or RNase) in order that the reactions are limitdp to the microcapsules (Luisi et al,, 1987),


Enzyme
le-catalysed biochemical reactions have also been demonstrated in mjfsrocapsules generated by a variety of other methods. Many enzymes are active in

re
erse micellar solutions (Bni & Walde, 1991; Bra & Walde, 1993; Creagh et
M d.
19|f3; Haber et al., 1993; Kumar et al^ 1989; Luisi & B., 1987; Mao & Walde, 1991; et at., 1992; Perez et al., 1992; Walde et al, 1994; Walde et al., 1993; Walde et
1988) such as the AOT-isooctane-water system (Monger & Yamada, 1979).

Tocapsules can also be generated by interfacial polymerisation and interfacial plexation (Whateley, 1996). Microcapsules of this sort can have rigid, pecmeable membranes, or semipenneable membranes. Semipenneable DSiiles bordered by cellulose nitrate membranes, polyamide membranes and jl-polyamide membranes can all support biochemical reactions, including systems (Chang, 1987; Chang, 1992; Lira, 1984). Alginate/polylysine ssules (Lim & Sun, 1980), which can be fonned under very mild conditions, also proven to be very biocompafi'ble, providing, for example, an effective metjbod of encapsulating living cells and tissues (Chang, 1992; Sun et al., 1992).
Noimembranous mieroencspsttlstion. systems based on phase partitioning of an
aqu ous environment in a colloidal system, such as an emulsion, may also be used.
bly, the microcapsules of the present invention are formed from emulsions; Dgeneous systems of two immiscible liquid phases with one of the phases dispersed in tb| other as droplets of microscopic or colloidal size (Becher, 1957; Sherman, 1968; Lasss rt, 1974; Lissant, 1984).
Emissions
may be produced from any suitable combination of immiscible liquids. Preferably the emulsion of the present invention has water (containing the biochemical pnents) as the phase present in the form of finely divided droplets (the disperse, or discontinuous phase) and a hydrophobic, immiscible liquid (an 'oil') as the

m* rix in which these droplets are suspended (the nondisperse, continuous or external ph se). Such emulsions are termed 'water-in-oiT (W/0). This has the advantage that the enl re aqueous phase containing the biochemical components is compartmentalised in dis reet droplets (the internal phase). The external phase, being a hydrophobic oil, get gaily contains cone of the biochemical components and hence is inert
Th« emulsion may be stabilised by addition of one or more surface-active agents (surfactants). These surfactants are termed emulsifying agents and act at the water/oil inte fece to prevent (or at least delay) separation of the phases. Many oils and many emi sifiers can be used for the generation of water-in-oil emulsions; a recent com Hlation listed over 16,000 surfactants, many of which are used as emulsifying
s
. (Ash and Ash, 1993). Suitable oils include fight white mineral oil and non-ionic (Scbick, 1966) such as sorbitan monooleate (Span™80; Id) and
poly
xyemylenesorbitanmonooleate (Tween™ 80; ICO and Triton-X-100..
i
gene mixtu
The ise of anionic surfactants may also be beneficial Suitable surfactants include sodii n cholate and sodium taurocholate. Particularly preferred is sodium deoxycholate, preferably at a concentration of 0.5% w/v, or below. Inclusion of such surfactants can in so ae cases increase the expression of the nucleic acids and/or the activity of the products. Addition of some anionic surfactants to a non-emulsified reaction
completely abolishes translation. During emulsification, however, the surfactant
is trai jferred from the aqueous phase into the interface and activity is restored. Addition
of an
inionic surfactant to the mixtures to be emulsified ensures that reactions proceed
only a ber compartonentalisation,
Creation of an emulsion generally requires the application of mechanical energy to force the phfses together. There are a variety of ways of doing this which utilise a variety of devices, including stirrers (such as magnetic stir-bars, propeller and turbine | paddle devices and whisks), homogenisers (including rotor-stator homogenisers, p valve homogenisers and jet homogenisers), colloid rpi'lls, ultrasound and '• emulsification1 devices (Becher, 1957; Dickinson, 1994).

Adueous microcapsules formed in water-ia-oil emulsions arc generally stable with little if |ny exchange of nucleic acids or gene products between microcapsules. Additionally, w| have demonstrated that several biochemical reactions proceed in emulsion :apsule$. Moreover, complicated biochemical processes, notably gene transcription and translation are also active in emulsion microcapsules. The technology ex its to create emulsions with volumes all the way up to industrial scales of thousands of fores {Becher, 1957; Sherman, 1968; Lissant, 1974; Lissant, 1984).
Th preferred microcapsule size will vary depending upon the precise requirements of an; individual selection process that is to be performed according to the present inv ntion. In all cases, there will be an optimal balance between gene library size, the enrichment and the required concentration of components in the individual jsules to achieve efficient expression and reactivity of the gene products.
Dei Us of one example of an emulsion used when performing tie method of the present inv otion are given in Example 1.
Exj. •essfon within microcapsules
The(processes of expression must occur within each individual microcapsule provided
11
by I the present invention. Both in vitro transcription and coupled tran caption-translation become less efficient at sub-nanomolar DNA concentrations. Bee use of the requirement for only a limited number of DNA. molecules to be present in ich microcapsule, this therefore sets a practical upper limit on the possible mid jcapsule size. Preferably, the mean volume of the microcapsules is less that 52 x
m3, (corresponding to a spherical microcapsule of diameter less than 10pm, more
prefi tably less than 6.5 x 10'17 m3 (5}Jni), more preferably about 4.2 x 10"18 m3 (2nm)
andieally about 9 x 10'18 m3 (2.6nm).
The effective DNA or RNA concentration in the microcapsules may be artificially I by various methods that will be well-known to those versed in the art These inclute, for example, the addition of volume excluding chemicals such as polyethylene

gl; cols (PEG) and a variety of gene amplification techniques, including transcription us ig RNA polymerases including those from bacteria suck as E. coli (Roberts, 1969; Bl truer and Dahlberg, 1972; Roberts et al., 1975; Rosenberg et al., 1975), eukaryotes e. . (Weil et al., 1979; Manley et al., 1983) and bacteriophage such as T7,13 and SP6 (M dton et aL, 1984); the polymerase chain reaction (PCR) (Saiki et al~, 1988); QJJ tedlicase amplification (Miele et al., 1983; Cahill et al., 1991; Chetverin and Spirin, 19! 5; Katanaev et a!., 1995); the ligase chain reaction (LCR) (Landegren et a/., 1988; Ba my, 1991); and self-sustained sequence replication system (Fahy et aL, 1991) and stn id displacement amplification (Walker et al., 1992), Even gene amplification tec tuques requiting thermal cycling such as PCR and LCR could be used if the em Isions and the in vitro transcription or coupled transcription-translation systems are thermostable (for example, the coupled transcription-translation systems could be made

a thermostable organism such as Thermos aquations).
loci lasing the effective local nucleic acid concentration enables larger microcapsules to be t »ed effectively. This allows a preferred practical upper limit to the microcapsule

volume of about 5.2 x 10"16 m3 (corresponding to a sphere of diameter lOum).

The||microcapsule size must be sufficiently large to accommodate all of the required aents of the biochemical reactions that are needed to occur within the midbcapsule. For example, in vitro, bom transcription reactions and coupled

trans :ription-translation reactions require a total nucleoside triphosphate concentration of at >ut2mM.
For < cample, in order to transcribe a gene to a single short RNA molecule of 500 bases in le gth, this would require a mimmtmr of 500 molecules of nucleoside triphosphate per mien capsule (8.33 x 10"22 moles). In order to constitute a 2mM solution, this number of mole ules must be contained within a microcapsule of volume 417 x 10'19 litres (4.17 x 10"32 m3 which if spherical would have a diameter of 93nm.
Furth|rrnore, particularly io. the case of reactions involving translation, it is to be noted that tjU ribosomes necessary for the translation to occur are themselves approximately

2dbm in diameter. Hence, the preferred lower limit for microcapsules is a diameter of
proximately lOOnm,

me Sp
Thlrefore, the microcapsule volume is preferably of the order of between 52 x 10*22 m3 5.2 x 10"1 m3 corresponding to a sphere of diameter between Q.hnnaad lOum, e preferably of between about 5.2 x 10"19 m3 and 6.5 x 10"17 m3 (lum and Sum).
ere diameters of about 2.6am are most advantageous.
It i| no coincidence mat the preferred dimensions of the compartments (droplets of 2.6um me|n diameter) closely resemble those of bacteria, for example, Eschericfda are 1.1-1.5 x

2.0-15.0 um rods and Azotobacter are 1.5-2.0 um diameter ovoid cells. In its simplest fon , Darwinian evolution is based on a 'one genotype one phenotype' mechanism, The con entran'on of a single compartmentalised gene, or genome, drops from 0.4 nM in a con jartment of 2 um diameter, to 25 pM in a compartment of 5 um diameter. The prol aryotic transition/translation machinery has evolved to operate in compartments

of-
-2 um diameter, where single genes are at approximately nauomolar concentrations.
A si tgle gene, in a compartment of 2.6 um diameter is at a concentration of 02 nM This gea concentration is high enough for efficient translation. Compartoentalisation in such
a volume also ensures that even if only a single molecule of the gene product is formed it
II is pifsent at about 02 nM, which is important if the gene product is to have a modifying
activity of the nucleic acid itself The volume of the microcapsule should thus be selected bearing in mind not only the requirements for transcription and translation of the nucleic atid/jtucleic acid, but also the modifying activity required of the gene product in the bd of the invention.
The
The ize of emulsion microcapsules may be varied simply by tailoring the emulsion
ions used to form the emulsion according to requirements of the selection system, arger the microcapsule size, the larger is the volume that will be required to
encap ;ulate a given nucleic acid/nucleic acid library, since the ultimately limiting factor will bl the size of the microcapsule and thus the number of microcapsules possible per unit vilume.

ft tb si cc
su
s size of the roicrocapsules is selected not only having regard to the requirements of transcription/translation system, but also those of the selection system employed for nucleic acid/nucleic acid construct. Thus, the components of the selection system, h as a chemical modification system, may require reaction volumes and/or reagent icentrations which are not optimal for transociption^anslation. As set forth herein, h requirements may be accommodated by a secondary re-encapsulation step; TDI ceover, they may be accommodated by selecting the microcapsule size in order to timise transcription/translation and selection as a whole. Empirical determination of


nucleic acid/nucleic acid" in accordance with the present invention is as described ve. Preferably, a nucleic acid is a molecule or, construct selected from the group sisting of a DNA molecule, an KNA molecule, a partially or wholly artificial nucleic molecule consisting of exclusively synthetic or a mixture of naturaUy-^xwurring and hetic bases, any one of the foregoing linked to a polypepti.de, and any one of the going linked to any other molecular group or construct Advantageously, the other ecular group or construct may be selected from the group consisting of nucleic acids, merle substances, particularly beads, for example polystyrene beads, magnetic tances such as magnetic beads, labels, such as fluorophores or isotopic labels,che uical reagents, binding agents such as macrocycles and the like.
Th
sue
pro seqnucleic acid portion of the nucleic acid may comprise suitable regulatory sequences, as those required for efficient expression of the gene product, for example aotcrs, enhancers, translatioaal initiation, sequences, polyadenylation ences, splice sites and the like.
Prt luce selection
Dot ils of a preferred method of performing the method of the invention are given in E» Qple 1. However, those skilled in the art will appreciate that the examples given are

L-limiting and methods for product selection are discussed in more general terms be )w.

an
igand or substrate can be connected to the nucleic acid by a variety of means that will be ipparent to those skilled in the art (see, for example, Hennanson, 1996). Any tag will su See that allows for the subsequent selection of the nucleic acid, Sorting can be by
method which allows the preferential separation, amplification or survival of the taj jed nucleic acid. Examples include selection by binding (including techniques based
on
magnetic separation, for example using Dynabeads™), and by resistance to de, radation (for example by nucleases, including restriction endonucleases).
Or > way in which the nucleic acid molecule may be linked to a ligand or substrate is ugh biotinylatiou. This can be done by PCR amplification with a S'-biotinylation
tiff
pri aer such that the biotin and nucleic acid axe covalently linked.

Th
ligand or substrate to be selected can be attached to me modified nucleic acid by a
vai ety of means mat will be apparent to those of skill in the art. A biotinylated nucleic aci may be coupled to a polystyrene microbead (0.035 to 0.2um in diameter) that is co: ted with avidin or streptavidin, that will therefore bind the nucleic acid with very hijj i affinity. This bead can be derivatised with substrate or ligand by any suitable me hod such as by adding biotinylated substrate or by covalent coupling.
Al amatively, a biotinylated nucleic acid may be coupled to avidin or streptavidia coi tplexed to a large protein molecule such as thyroglobulin (669 Kd) or feiritin (440 Kdj This complex can. be dedvatised with substrate or ligand, for example by covalent

coi
ling to the alpha-amino group of lysines or through, a non-covalent interaction such otin-avidin. The substrate may be present in a fona unlinked to the nucleic acid but containing an inactive "tag" that requires a further step to activate it such as
ph,
toactivation (e.g. of a "caged" biotin analogue, (Sundberg et al., 1995; Pinung and
Hu ng, 1996)). The catalyst to be selected then converts the substrate to product The
"ta
' could men be activated and the 'tagged" substrate and/or product bound by a inding molecule (e.g. avidin or streptavidin) complexed with the nucleic acid. The

of substrate to product attached to the nucleic acid via the "tag" will therefore ect the ratio of the substrate and product in solution.
all reactions are stopped and the microcapsules are combined, the nucleic acids active enzymes can be enriched using an antibody or other molecule which birjjis, or reacts specifically with the "tag". Although both substrates and product have the nolecular tag, only the nucleic acids encoding active gene product will co-purify.
Th terms "isolating", "sorting" and "selecting", as well as variations thereof, are used her in. Isolation, according to the present invention, refers to the process of separating

from a heterogeneous population, for example a mixture, such that it is free of >ne substance with winch it was associated before the isolation process. In a embodiment, isolation refers to purification of an entity essentially to city. Sorting of an entity refers to the process of preferentially isolating desired vet undesired entities. In as far as this relates to isolation of the desired entities, "isolating" and "sorting" are equivalent. The method of the present invention peniuts the sorting of desired nucleic acids from pools (libraries or repertoires) of nuc ac acids which contain the desired nucleic acid Selecting is used to refer to the proc ;ss (including the sorting process) of isolating an entity according to a particular prop srty thereof.
miti I selection of a nucleic acid/nucleic acid from a nucleic acid library (for example a mut|at taq library) using the present invention will in most cases require the screening of
i number of variant nucleic acids, Libraries of nucleic acids can be created in a
r of different ways, including the following.
of naturally occurring nucleic acids can be cloned from genomic DNA or cDNA et al^ 1989); for example, mutant Taq libraries or other DNA polymerase |ies, made by PCR amplification repertoires of taq or other DNA polymerase genes /ed very effective sources of DNA polymcrasc fragments. Further details are giverjjinthe examples.
Li nraries of genes can also be made by encoding all (see for example Smith, 1985;
P: jnley and Smith, 1988) or part of genes (see for example Lowman et al, 1991) or
p< ols of genes (see for example Nissan et al. 1994) by a randomised or doped synthetic
ol gonucleotide. Libraries can also be made by introducing mutations into a nucleic acid
pool of nucleic acids 'randomly' by a variety of techniques in vivo, including; using
'n utatar strains*, of bacteria such, as E. coli mutD5 (Liao et al., 1986; Yamagishi et al.,
1! 90; Low et aL, 1996). Random mutations can also be introduced both in vivo and in
vi -o by chemical mutagens, and ionising or UV irradiation (seeFricdberge al.,
11 95), or incorporation of mutagenic base analogues (Freese, 1959; Zaccolo et a/., 1! )6). 'Random' mutations can also be introduced into geaes in vitro during pc ymerisation for example by using error-prone polymerases (Leung et al., 19S9). In a rred embodiment of me method of the invention, the repertoire of nucleic fragments
uspd is a mutant Taq repertoire which has been mutated using error prone PCR. Details art given in Examples 1. According to the method of the invention, the term 'random* m|y be in tenns of random positions with random repertoire of amtno acids at those p( dtions or it may be selected (predetermined) positions with random repertoire of az icto acids at those selected positions.
F rther diversification can be introduced by using homologous recombination either in vi o (see Kowalczylcowski et al., 1994 or in vitro (Stemmer, 1994a; Stemmer, 1994b)).
A crocapsules/sorting
addition to toe nucleic acids described above, the microcapsules according to the in motion will comprise further components required fat the sorting process to take pi ce. Other components of His system will for example comprise those necessary for

scription and/or translation of the nucleic acid. These are selected for the
re [uirements of a specific system from the following; a suitable buffer, an in vitro tr ascription/replication system and/or an in vitro translation system containing all the n< ;essary ingredients, enzymes and cofactors, RNA poiymerase, nucleotides, nucleic ac ds (natural or synthetic), transfer RNAs, ribosomes and amino acids, and the
si >strates of the reaction of interest in order to allow selection of the modified gene pi »duct

s> re, til

suitable buffer will be one in which all of the desired components of the biological tern are active and will therefore depend upon the requirements of each specific ction system. Buffers suitable for biological and/or chemical reactions are known in art and recipes provided in various laboratory texts, such as Sambrook et al.. 1989.

ba rei sui
Ti j in vitro translation system will usually comprise a cell extract, typically from teria (Zubay, 1973; Zubay, 1980; Lesley et al., 1991; Lesley, 1995), rabbit culocytes (Pelham and Jackson, 1976), or wheat germ (Anderson et al., 1983). Many able systems are commercially available (for example from Promega) including so le which will allow coupled transcription/translation (all the bacterial systems and the culocyte and wheat germ TNT™ extract systems from Promega). The mixture of an tno acids used may include synthetic amino acids if desired, to increase the possible nu iber or variety of proteins produced in the library. This can be accomplished by ch rgjng tRNAs with artificial amino acids and using these tKNAs for the in vitro Ira islation of the proteins to be selected (Ellman et al., 1991; Benner, 1994; Mendel et al. 1995).

is
fro
Af sr each round of selection the enrichment of the pool of nucleic acids for those em oding the molecules of interest can be assayed by non-compartmentalised in vitro tra scription/replication or coupled transcription-translation reactions. The selected pool loned into a suitable plasmid vector and RNA or reconibinant protein is produced
i the individual clones for further purification and assay.

Mi

rocapsule identification

Mi
rocapsules may be identified by virtue of a change induced by the desired gene p« luct which either occurs or manifests itself at the surface of the microcapsule or is

del

ctable from the outside as described in section iii (Microcapsule Sorting). This

chs ige, when identified, is used to trigger the modification of the gene within me
co apartment. In a preferred aspect of the invention, microcapsule identification relies OD a change in the optical properties of the mierocapsule resulting from a reaction le; iing to luminescence, phosphorescence or fluorescence within, the microcapsule. M dification of title gene within the microcapsules would be triggered by identification of luminescence, phosphorescence or fluorescence. For example, identification of Ita linescence, phosphorescence or fluorescence can trigger bombardment of the apartment with photons (or other particles or waves) which leads to modification of
co

th
so
nucleic acid. A similar procedure has been described previously for the rapid ing of cells (KLeij et al, 1994). Modification of the nucleic acid may result, for ex mple, from coupling a molecular "tag", caged by a photokbile protecting group to nucleic acids: bombardment with photons of an appropriate wavelength leads to
ti»

ac,
removal of the cage. Afterwards, all microcapsules are combined and the nucleic is pooled together in one environment. Nucleic acids encoding gene products exhibiting the desired activity can be selected by affinity purification using a molecule

thi

specifically binds to, or reacts specifically with, the "tag".

Mi IS. step procedure

It ill be also be appreciated that according to the present invention, it is not necessary
for
prc
sel
me
all the processes of transcription/replication and/or translation, and selection to seed in one single step, with .all reactions taking place in one microcapsule. The ,tion procedure may comprise two or more steps. Fixst, transcription/replication or translation of each nucleic acid of a nucleic acid library may take place in a first mi|rocapsule. Bach gene product is men linked to the nucleic acid which encoded it (w! ich resides in the same microcapsule). Tile microcapsules are then, broken, and the cut ieic acids attached to their respective gene products optionally purified. All amatively, nucleic acids can be attached to their respective gene products using hods which do not rely on encapsulation. For example phage display (Smith, ,1985), polysome display (Mattaeakkis et al, 1994), RNA-peptide fusion (Roberts
ancSzostak, 1997) or lac rcpressor peptide fusion (Cull, et al, 1992).

tptification
In the second step of the procedure, each purified nucleic acid attached to its gene pr duct is put into a second microcapsule containing components of the reaction to be se scted. This reaction is then initiated. After completion of the reactions, the m aocapsules are again broken and the modified nucleic acids are selected. la the case complicated multistep reactions in which many individual components and reaction ; are involved, one or more intervening steps may be performed between the initial of creation and linking of gene product to nucleic acid, and the final step of ; the selectable change in he nucleic acid.

In
all me above configurations, genetic material comprised in the nucleic acids may be
ax plified and the process repeated in iterative steps. Amplification may be by the lymerase chain reaction (Saiki et al., 1988) or by using one of a variety of other gene plification techniques including; Qp repHcase amplification (Cabill, Foster and 1991; Chetverin and Spain, 1995; Katanaev. Kumasov and Spirin, 1995); the
li|ase chain reaction (LCR) (Landegren et al., 1988; Barany, 1991); the self-sustained aence replication system (Fahy, Kwoh and Gingeras, 1991) and strand displacement plification (Walker et al., 1992),
(lj|) DNA polymerases according to the invention. General
H gh fidelity DNA polymerases such as Pol A(Iike Taq polymerase) and Pol-B family lymerases which lack a 3'-5' exonuclease proofreading capability show a strict b )ckage to the extension of distorted or mismatched 3" primer termini to avoid p >pagation of misincorporations. "While the degree of blockage varies considerably depending on the nature of the mismatch, some transversion (purine-purine / pirimidine-pyrimidme) mismatches are extended up to lO^-fbld less efficiently than
m icbed termini (Huang 92). Likewise, many unnatural base analogues, while in orporated efficiently, act as strong tenninators (Kool, Loakes).
Ti i present inventors have modified the principles described in Ghadessy, F. G et al (2001) Pi c. Nat Acad, Sci, USA, 93, 4552-4557 (compartoaeataHscd self replication) and Gl adessy 2003, and outlined above. Both these documents are herein incorporated by re srence. The present inventors have used these modified techniques to develop a method b] which the substrates specificity of high fidelity DNA polymerases may be expanded in a generic way.
) inventors have exemplified the technique by expanding the substrate specificity 6f the mill-fidelity pol-A family polymerases. In particular, the present inventors created two re ertokes of randomly mutated Taq genes, as described in Ghadessy, F- G et al (2001) re ared to above. Three cycles of mismatch extension CSR was perfozmed using flanking pi aers bearing the mismatches A*G and C*C at their 3* ends. Selected clones were ra ked using a PCR extension assay described herein,
S( ected mutants exhibited the ability to extend the G*A and C*C tranversion mismatches us d in the CSR selection, but also exhibited a genetic ability to extend mispaired 3' tei aim'. These results are surprising, especially since Taq polymerase is unable to extend

su

matches (Kwok et al, (1990); Huang (1992).
Tl is, using mis approach, the inventors have generated DNA polymerases which exhibit a re ixed substrate specificity/expanded substrate range.
An cording to the present invention, the term 'expanded substrate range* (of an en dneered DNA polymerase) means that substrate range of an engineered DNA pc ymerase according to the present invention is broader than that of the one or more A polymerases, or fragments thereof from which it is derived. The term 'a broader su strate range' refers to the ability of an engineered polymerase according to the present in ention to extend one or more 3' mismatches, for example AA, GA, GG, TT, CTJC, which the one or more polymerase/s from which it is derived cannot extend. That
is essentially, a DNA polymerase which exhibits a relaxed substrate range as herein ined has the ability not only to extend the 3 mismatches used in its generation, (JOB th|se of the flanking primers), but also exhibits a generic ability to extend 3* mismatches (fi r example A*G, A*A, G*G).
3 two best mutants Ml (G84A, D144G, K314R, E520G, F598L, A60SV, E742G) and M (D58G, R74P, A109T, L245R, R343G, G370D, E520G, N583S, E694K, A743P) wire chosen for further investigation.
M and M4 not only had greatly increased ability to extend the G-A and C»C tn isversion mismatches used in the CSR selection, but appeared to have acquired a more ge eric ability to extend 3' mispaired termini, including other strongly disfavoured tn isversion mismatches (such as A*G, ArA, G»G) (Fig- IB), which wtTaq polymerase

w

unable to extend, as previously reported (Kwok et al 1990, Huang 92).



c«

Ml and M4 mutants according to the invention.



Ni de
re.

;leio acid sequences encoding Ml and M4 pol A DNA polymerase mutants are icted SEQ No 1 and SEQ Ho 2 respectively and are shown in. Fig 1 and fig 2 >eetively.

D« spite very similar properties, Ml and M4 (and indeed other selected clones) have few m taticms in common, suggesting fhere are multiple molecular solutions to Hie mismatch ex ension phenotype. One exception was ES20G, a mutation that is shared by all but one of he four best clones of the final selection. Curiously, E520 is located at the very tip of m< thumb domain at a distance of 20A from the 3* OH of the mismatched primer tejfninxis and its involvement in mismatch recognition or extension is unclear. However, OG is clearly important for mismatch extension as bacfcrnutation reduces mismatch

ex

nsion in both Ml and M4 to near wt levels (Fig. 2).



10!

only other feature clearly shared by both Ml and M4 are mutations targeting lues, which may be involved ia flipping out the +1 template base. Residue E742
n itated in Ml (E742G) forms a direct contact witii the flipped out +1 base on the t nplate strand (Li et al), while in M4 the adjacent residue A743 is mutated to proline
.743P), which may disrupt interactions by distorting local backbone conformation.
ick mutation of E742G in Ml reduced mismatch extension, but only by ca. 20% ii iicating that it does not contribute decisively to mismatch extension.
rprisingly, mutations in the N-tenninal 5*~3' exonuclease domain (53exoD) also a pear to be contributing to mismatch extension as suggested by the 2-4 fold increased smatch extension ability of chimeras of the 53exoD of Ml, M4 and polD of wtTaq

g. 4). How they promote mismatch extension is unclear but given the apparent distance of the 53exoD fiom the active site (Ute 99, Eom 96) is unlikely to involve direct
ects on extension catalysis. Increased affinity for primer-template duplex could also idbrease mismatch extension (Huang 92) but dissociation constants of wfTaq, Ml and ft 4 for matched and mismatched primer-template duplex were indistinguishable as ji Iged by an equilibrium binding assay (Huang 92) (not shown).
e relationship of Ml and M4 with other naturally occurring DNA polymerases
P P
E tension of mismatched 3' primer termini is a feature of naturally occurring ymerases. Viral reverse transcriptases (RT) like HIV-1 RT or AMV RT and ymerases capable of translesion synthesis (TLS) such as the polY-family polymerases i (Vaisman 2001JBC) or pol K (Washington. 2002 PNAS) or the unusual poffi-iamily ymcrase pol£ (Johnson Nature), all extend 3' mismatches with elevated efficiency
G mpared to high-fidelity polymerases. Thus, the selected polymerases share significant
fifictional similarities with preexisting polymerases but represent, to our knowledge, the y known polA-family polymerases that are proficient in mismatch extension (ME) and
tr|nslesion synthesis (TLS). In contrast to TLS polymerases, which are distributive and >end on cellular processrvity factors such as PCNA (Prakash refs for eta/kappa and
idka). Ml and M4 combine ME and TLS with high processivity and in the case of Ml are
c; pable of efficient amplification of DNA fragments of up to 26kb.


In PI

the case of viral RTs, ME may play a crucial role in allowing error-prone yet cessive replication of a multi-kb viral genome. For TLS polymcrascs, proficient smatch extension is also a necessary prerequisite for their biological function as

>aired and distorted primer termini necessarily occur opposite lesions in the DNA II template strand. The ability of TLS polymerases to traverse replication blocking lesions
mil DNA is thought to arise from a relaxed geometric selection in the active site (( oodman 02). The ability of Ml and M4 to process both bulky mispairs and a distorting D (cys-syn thymidine-thymidine dimer) dimer makes it plausible mat, in analogy to polymerases, they also have acquired a more open active site. Indeed, modelling si wed that a CPD dimer can not be accommodated in the wtTaq polymerase active site w thout mayor steric clashes (TrincaoOl).

(and to a lesser degree M4) also display a much increased ability to incorporate and replicate different types of unnatural nucleotide substrates that deviate to v« ying degrees from the canonical nucleobase structure. Of these the oS substitution is tfc most conservative. However, the sulfur anion is significantly larger than oxygen ai ton and coordinates cations poorly, which may be among the reasons why the wt ei cyme will not tolerate full oS substitution. Fluorescently-labelled nucleotides like aS m clcotidcs retain base-pairing potential but include a bulky and hydrophobic substituent tl (t must be accomodated by the polymerase active site. Steric clashes in the active site at t allievaiedby the presence of a long, flexible linker. Indeed, we find biotin-16-dUTP a] men better substrate for Ml man biotin-11-dUTP, while wtTaq cannot utilize either. T e hydrophobic analogue 5NI represents the most drastic departure from standard a cleotide chemistry we investigated. Comparable in size to a purine base, 5N1 c< mpetely lacks any hydrogen bonding potential but like the natural bases, favours the
ta ti-position with respect to the ribose sugar as judged by NMR (J. Gallego, DX. and
ii PH., unpublished results). Therefore, a 5N1«A or 5NK5 basepair would closely resemble
a nxrine-purine transversion mismatch and may cause similar distortions to the canonical D ^A duplex geometry. Elegant experiments using isosteric non-hydrogen bonding base ai alogues have shown that Watson-Crick hydrogen bonding per se is not required for efficient insertion or replication (reviewed by Kool 02). However, while many non-h|drogen-bonding hydrophobic base analogues are efficiently incorporated, they

sequcntly lead to termination, both at the 3' end aad as a template base (Kool , Re nesberg).
St ictural and biochemical studies have previously .identified regions of the po ymerase structure that are important for mismatch discrimination such as motif A (it 'olved in binding the incoming dNTP), the Ohefcx (motif B) and residues in olved ia. minor groove hydrogen bonding (24, 25). Inspection of the sequence of M and M4 reveals a conspicuous absence of mutations in these regions. Rather mi tations in Ml and M4 implicate regions of fixe polymerase not previously as; rcialed with substrate recognition such as the tip of the thumb subdomain (E520), th< +1 template base-flipping function (E742, A743) in the finger subdomain and the 5-[' exonuclease domain (53exoD),

mi
by als tip tta
Ti s 53exoD is too distant from the active site to have direct effects on mismatch ex snsion. It is, however, thought to be crucial for polymerase processivity and may feus inj neace mismatch extension (24). Indeed, the Stofifel fragment of Taq polymerase (26), wl ich lacks the 53exoD, displays both reduced processivity and more stringent mismatch du Trimination (27). Mutations in the 53exoD of Ml and M4 may therefore contribute to match extension by enhancing polymerase processivity. Together with the ability to ass abasic sites (generated in large DNA fragments during thermocycling) this may contribute to the proficiency of Ml at long PCR (Fig. 5). E520 is located at the very >f the thumb domain at the end of the H2 helix at a distance of 20A from the 3* OH of
mismatched primer terminal base (PI) (2). Mechanistic aspects of the involvement of E520O mutation in mismatch recognition or extension are therefore not obvious eitier. It is worth noting though that adjacent regions, especially the preceding loop co; necting helices HI and H2 and parts of helix I, make extensive contacts with the ter plate-primer duplex between P3-P7 (2). It has previously been observed that mi match incorporation affects extension kinetics up to the P4 position (24). E520G may me liry the structure of these regions to ease passage of mismatches and increase

elc ba:

gation efficiency post incorporation. Base flipping, i.e. rotation of the desigaated
out of the DNA helix axis is a common mechanism among DNA modifying
raies (e.g. glycosylases) but its precise role for polymerase function is less clear. It

been speculated that flipping out of the +1 template base may contribute to-onerase fidelity by preventing out-of-rcgister base-pairing (25) of the 3' nucleotide to
pol
coj late upstream template bases. Interference with this mechanism therefore might pro note apparent mismatch extension but would produce -1 base deletions. However, neither primer extensions nor sequencing of PCR products generated with Ml or M4 usi g primers with 3' GrA and OC mismatches revealed any template slippage but on the confirmed in-register extension of the mismatches (not shown). The utility of base-flipping in the context of the TLS capability of Ml and M4 is easier to understand, especially on the CPD dimer, as the two covalently linked thymine template baz » would be refractory to flipping out Indeed, TLS polymerases, which are naturally ab] ! to bypass CPD dimers, appear to lack a base-flipping function (28).
Ex mination of the extension and incorporation kinetics of the mutant polymerases gests that they have a significantly increased propensity to not oaly extend but also jrate transversion mispairs and consequently should have a significantly increased mijkation rate compared to the wt enzyme. More relaxed geometric selection in the active siti might also be expected to come at the price of significantly reduced fidelity as indeed is |ie case for TLS polymerases (23). However, measurement of the overall mutation rate us ig the MutS assay (not shown) and sequencing of PCR products generated by Ml in< icated only a modest (< 2-fold) increase in the mutation rate (Table 1) mostly due to an increased propensity for transversions. As discussed previously (10), GSR should se «t for optimal self-mutation rates within the error threshold (31). A change in the m tation spectrum towards a more even distribution of transition and transversion mi tations may be an effective solution to accelerate adaptation, while maintaining a he Ithy distance from the error threshold. This may also make Ml a useful tool for pr tein engineering as the bias of Taq (and other DNA polymerases) for transition m tations limits the regions of sequence space mat can be accessed effectively using PC Rmutagenesis
Ts >le 1: Mutation spectrum of wtTaq and Ml in PCR

5 10 1

utatioDS derived from sequencing of 40 clones (SOObp) each.

In
M

iummary DNA polymcrases according to the present invention, in particular Ml and respectively as depicted m SEQ No 1 and SBQ No 2 possess the following

pn jetties:

(1) (2) (3)

DNA Translesion synthesis
A generic ability to incorporate unnatural base analogues into DNA.
Ml has the ability to efficiently amplify DNA targets up to 26kb.

$ of DNA polymerases according to the invention.
Di scted evolution towards extension of distorting transversion mismatches like G-A or CM ! by CSR yields novel, "unfussy" polymerases with an ability to perform not only efi cient mismatch extension and TLS but also accept a range of unnatural nucleot'de sal strates. The present inventors have shown that the evolution of TLS from a high-

fid
ity, polA-family, pol B family or other polymerases requires but few mutations,
suj gestuxg mat TLS and relaxed substrate recognition are functionally connected and ma r represent a default state of polymerase function rather than a specialization.
Th unusual properties of the DNA polymerases according to the present invention, in pax icular Ml and M4 may have immediate uses for example for the unproved inc irporation of dye-modified rmcleotides in sequencing and array labelling and/or the am lification of ultra-long DNA targets. They may prove useful in the amplification of dai aged DNA templates in foreasics or paelobiology, may permit an expansion of the chc oical repertoire of aptamers or deoxi-ribozymes (Benner, Barbas, ribozyme review)

an
mi the
CO
M

may aid efforts to expand the genetic alphabet (Benner, Schultz). The altered Cation spectrum of Ml may make a useful tool in random mutagenesis experiments as
i
strong bias of Taq and other polymerases towards (A->G, T->C) transitions limits the ibinatorial diversity accessible through PCR mutagenesis. Furthermore, the ability of & M4 to extend 3' ends in which the last base is mismatched with the template strand the ability of H10 (see example 6) to extend 3' ends in which the last two bases are matched with the template strand may extend the scope of DNA shuffling methods

(St mmer) by allowing to recombine more distantly related sequences.

In
as
util tha

addition, DNA polymerases according to the invention, in particular pol A unerases, for example Ml and M4 pol A polymerases as herein described may serve L useful framework for mutagenesis and evolution towards polymerases capable of zing an ever wider array of modified nucleotide substrates. The inventors anticipate directed evolution may ultimately permit modification of polymerase chemistry

ito f, allowing the creation of amplifiable DNA-like polymers of defined sequence thus ext aiding molecular evolution to material science.

Th
Urn

invention will now be described by Hie following examples which are in no way
ting of the invention claimed herein.

Ex mplel
Ge eral Methods
Lis of primers:
1: S'-CAG GAA ACA GCT ATG ACA AAA ATC TAG ATA ACG AGG GA-31;
M terials and Methods
DI A manipolatioa and protein expression. Expression of Taq clones for screening CSR selection was as described (10). For kinetic measurements and gel extension as: tys, polymerases were purified as described (32) using a BiorexTO ion exchange res n (BioRad). All PCR and primer extensions were performed in Ix Taq buffer (5( mM KCl/lOmM Tris-HCl (pH 9.0) / 0.1% Triton X-100 / l.SmM MgCfc), with

d£ IPs (0.25mM (Araersham Phannaoia Biotecll, NJ)) and appropriate primers unless sp< cified otherwise. Primer sequences are provided in Supplementary information.
ier extension reactions were terminated by addition of 95% fonnamide/ 10 mM
TA and analysed on 20 % polyacrylamide / 7 M Urea gels.
deselection. Activity preselected libraries Ll and L2 (10) were combined and 3 rounds of CSR selection carried out as described (10) except using primers 1: (A*G

m
match) and 2: (C'C mismatch) and IS cycles of (94°C 1 ™"\ 55°C 1 min, 72°C 8
mi i). Round 2 clones were recombined by staggered extension process (StEP) PCR sh filing (33) as described . For round 3, CSR cycles were reduced to 10 and am ealing times to 30 sec.

PC R. A PCR assay was used to screen and rank clones. Briefly, clones were no malized for activity in PCR with matched primers 3, 4 and activity with mi matched primers 1 and 2 (IpM each) determined at minimal cycle number (15-25 cy les). Extension capability for different mismatches was determined by the same as: iy using mismatch primers 2 (C«C mismatch), 5 (A«A mismatch), 6 (G*G mi match), 7 (G»A mismatch) with matched primer 3 or primer 1 (A*G mismatch) wi i matched primer 4. Incorporation of unnatural substrates in 50 cycle PCR was 20 cai ied out using standard conditions and 50uM ccS dNTPs (Promega) or 50uM FTfC-12-dATP (Perkta-Btoer), Rhodamine-5-dUTP (Perkin-Elmer) or Biotin-16-dtJ HP (Roche) with equivalent amounts of the other 3 dNTPs (all 50uM). Long PCR wa carried out using a two-step cycling protocol as described (22) 94°C for 2 mi .utes, followed by 20 cycles of (94°C 15 sec, 68°C SOmin) using 5ng of phage X D} A. (New England Biolabs) template and either primers 9,10,11 with primer 12 or pri uer 13 with primers 10,14.
nucleotide incorporation/extension kinetics. Kinetic parameters were .using a gel-based assay essentially as described (16). Primers 15,16, 17
(3'||base = G, C, A respectively) were "HP-labeled and annealed to one of template strjads 18, 19, 20 (template base = C, G, A respectively) or 21 (template base C

ferent context). Duplex substrates were used at 50nM final concentration in 1 X buffer with various concentrations of enzyme and dNTP. Reactions were carried at 60°C for times whereby <20% of primer-template was utilized at the highest concentration of dNTP.
mplate affinity assays. An equilibrium binding assay (12) was used to determine relbtive affinity of polymerases for the mismatched primer-tecaplates used in the ki etics assays. Polymerases were preincubated at 60°C in IxTaq buffer with SOnM
"^-labeled matched primer-template and SOnM unlabeled mismatched competitor pi mer-templates. Reactions were initiated by simultaneous addition of dCTP (200uM »d trap DNA (Xbal 1.Sk/I-restricted sheared sahnon sperm DNA, 4.5 mg/ml). Prior >eriments demonstrated trap-effectiveness over the time period used (15 seconds).
mslesf on Replication Assay. Template primers 22 (undamaged) or 23 (containing

Template primer- 24 (containing a single cis-syn thymine dimer), was
mesized as described (34). Primer 25 was P-labeled and annealed to one of the
templates 22,23,24 (at a primer template ratio of molar 1:1.5) and extended in
aM Tris'HCl at pH 8.0, 5mM MgCfc, lOOjiM of each dNTP, lOmM DTT, 250
mi BSA, 2.5% glycerol, 10 nM primer-template DNA and 0.1 Unit of polymerase 0°C for various times.



27
dri

replication assay. Primer 26 was ^-labeled and annealed to template primer containing a single 5-nitroindolc) in IxTaq bufier, 0.1 or 0.5U of the polymerase added and reactions incubated at 60°C for 15 rains, after which 40jiM of each ? were added and incubation at 60°C continued for various times.
Fi elity assays. Mutation rates were determined using the mutS ELISA assay (G snecheck, R. Collins, CO) or by perfonaing 2x50 cycles of PCR on three different ten iplatcs and sequencing the cloned products.
mi
Kifietic analysis. Extension and incorporation kinetics of Ml and M4 for a selection of matches were measured using a gel-based steady-state kinetic assay (Goodman)
CT
bles 1 & 2). Ml and M4 respectively extend a OC mispair 390 and 75-fold more
Jciently than wtTaq. Examination of the other most disfavored mismatches (G*A, A*G, , G»G) reveals generic, although less pronounced., increases of extension efficiencies, as suggested by me PCR assay (Fig. 4, fig 5). The gain in extension efficiency derives from increased relative Vmax values for the mutant polymerases, while i for nucleotidc substrates remains largely unchanged. For most DNA polymerases the relative efficiency of extending a given mispair (fDext) is similar to the relative efficiency , it (fine) (Goodman 1993, Goodman 1990, Washington 2001).Indeed, Ml and respectively incorporate dCTP opposite template base C 206- and 29-fold more Iciently than wtTaq (Table 2).

(Table Removed)

emplate base: 3 primer base; Incorporated base is dCTP enzyme efficiency=Vm^ / Km ^/(mismatched 3'terminus) //(matched terminus) a. (mutant polymerase) I fas. (wtTaq)

(wtTaq) Eximple 3
Tr nslesion synthesis. Transversion mispairs represent distorting deviations from the coj oate duplex structure. We therefore investigated if Ml and M4 were capable of prc Jessing other deviations of the DNA structure such as lesions in the template strand. Us ig a gel-extension assay we investigated their ability to traverse an abasic site and a. ds syn thymine pyrmidine dimer (CPD) template strand lesion. In control assays using an mdamaged template, wtTaq, Ml and M4 efficiently and rapidly extended primers to the end of the template (Fig. 5). On the template containing an abasic site, wtTaq eff aently inserts a base Opposite the lesion but, farther extension i$ largely aboh'shed. In bom Ml and M4 are able to extend past the lesion and to the end of the plate. The size of the final product is similar to that observed on the undamaged plate indicating that bypass occurred without deletions. Perhaps the most striking pie of the proficiency of Ml and M4 to bypass template lesions is observed on the intaining template (Fig. 5). Undo- the assay conditions, wtTaq utilizes a fraction of lie available template and is only able to insert a base opposite the 3'T of the diraei aft|r prolonged reaction conditions. In contrast, both Ml and M4 are able to readily all of the primer to the 3'T of the dimer. In addition, there is also considerable extension of these primers to the 5'T of the CPD. As with die abasic template, a cant traction of these primers are subsequently fully extended to the end of (he

te iplate in an error-free manner without deletions. We estimate that trans-lesion esis (TLS) by Ml and M4 may only be 2-5 fold less efficient than that observed th a naturally occurring TLS polymerase, Dpo4 from S. soljhtaricus (Boudsocq et al (2j|)01), Nucleic Acid Res, 29,46072001) on the same template,
ample 4

U natural substrates. We reasoned that relaxed geometric selection might also aid the in orporation of unnatural base analogues, some of which inhibit or arrest polymerase ac ivity due to poor geometric fit or lack of interaction with either polymerase or te iplate strand, A first, conservative example are phosphothioate nucleotide tri )hosphates (oS dNTPs), in which one of the oxygen atoms in the a phosphate group is re laced by sulfor. As part of a dNTP mixture, oS dNTPs are generally well accepted as s\j >strates by DNA polymerases but when we, replaced all four dNTPs with their aS cc inteiparts in PCR wtTaq failed to generate any amplification products, while Ml (and to lesser extent M4) were able to generate PCR products of up to 2kbp, indicating mat th y could utilize oS dNTPs with much increased efficiency compared to the wt enzyme (F g. 6). As expected, the resulting ccS DNA was completely resistant to cleavage by
A endonucleases (not shown). Nucleotides bearing bulky adducts such as fluorescent d) s are widely used in applications such as dye terminator sequencing or array labelling. A hough generally well tolerated they are incorporated considerably less efficiently man th natural dNTP substrates and can cause permature termination mhomopolvmeric runs,
h> fto by
sumably because of steric crowding due to the bulky dye molecules. In PCR the effect is otentiated because bom template and product strands are labelled. When we replaced dllTP with Biotin-16-dUTP or dATP with FrTC-12-dATP in PCR, wtTaq was unable to generate any amplification products, while Ml was able to generate 2.7 kb amplification >ducts fully labelled with Biotm-16-dUTP or a 0.4kb fully labelled with FTTC-12-d/iTP (Fig. 6). Recently, there has been significant interest in hydrophobicr non-
rogen bonding base-analogues and the applications they may enable. One of these is candiate "universal base" 5-mtroindole (5NJ) (Loakes & Brown 96), which, like other rophobic, strongly stacking base analogues , is readily accepted as a substrate, but

oru B incorporated, acts as a strong terminator both at the 3' end and as a template base. In coi Crast, M4 and in particular Ml efficiently bypass template strand 5NI (Fig. 6) and to a

les

er degree, extend 5NI at the 3' end (not shown).

Ex imple 5
Lo ig PCR. Amplification product size with wtTaq is generally limited to fragments a

fir
kb long but can be extended to much longer targets by inclusion of a proofreading
po ymerase (Barnes 92). We found that the selected polymerases, in particular Ml was
abl ? to efficiently amplify of targets up to 26kb (Fig. 7), using standard PCR conditions
in he absence of auxiliary polymerases or other processivity factors. Under the same
001 ditions wtTaq enzyme failed to amplify targets >9kb. The molecular basis for the
pn duct size limitation in the wt enzyme is thought to be premature termination due to an
ins rility to extend mismatches following nucleotide misincorporation. These are thought
to >e removed by the proofreading polymerase allowing extension to resume. Our results
flu; a generic mismatch extension ability to results in a similarly extended amplification
rai ge supports this concept. '
Esfcrnple 6: Libraries of polymerase chimeras
Libraries of cbimeric polymerase gene variants were constructed using a gene shuffling technique called Staggered extension protocol (StEP, (Zhao, Giver et al. 19 >8)). This technique allows two or more genes of interest from different species to be andomly recombined to produce chimeras, the sequence of which contains parts of,

th<

original input parent genes.

Thermus aquaticus (Tag) "wild type and T8 (a previously selected 11 fold more thi nnostable Taq variant (Ghadessy, Qng et al. 2001)), Thermos thermophiliis (Hth) an Thermus flavus (Tfl) polymerases had previously been amplified from genomic Dl FA and cloned into pASK75 (Sfcerra 1994) and tested for activity. These genes were thi a shuffled using the staggered extension protocol (StEP) as described (Zhao, Giver

aJL 1998) with (CAG GAA ACA GCT ATG ACA AAA ATC TAG ATA ACG 3G GCA A and GTA AAA CGA CG G CCA GTA CCA CCG AAC TGC GGG 3A CGC CAA GCG), recloned into pASK75 and transformed into E. colt TGI. The rary size was scored by dilution assays and detennining the ratio of clones c< qtajqjung insert using PCR screening and was approximately 108. A diagnostic re rtrictioa digest of 20 clones produced 20 unique restriction patterns, indicating that tb j library was diverse. Subsequent sequencing of selected chimeras showed an ai erage of 4 to 6 crossovers per gene.
E ample 7: Selection of two mismatch extension polymeroe.
CSR emulsificatkm and selection was performed on the StEP Taq, Tth and Tfl
II
lifflrary essentially as described (Ghadessy, Ong et al. 2001). Mismatch primers with tv o mismatches at their 3' end (5'-GTA AAA CGA CGG CCA GTT TAT TAA CCA C X5 AAC TGC-3', S'-CAG GAA ACA GCT ATG ACT CGA CAA AAA TCT A 3A TAA CGA CC-3') were in the emulsion as the source of selective pressure. The ac icous phase was ether extracted, PCR purified (Qiagen, Chatsworth, CA) with an ac litional 35% GnHCl, digested with Dpnl to remove methylated plasmid DNA, tn atod with ExoSAP (USB) to remove residual primers, reamplified with outnested pi raers and recloned and transformed into E. coli as above.
The resultant clones were screened and ranked by PCR assay. Briefly, 2 ul of . cells were added to 20 ul of PCR mix with the relevant mismatch primers, that produced a band were then subjected to further analysis and the most acpve clones were sequenced.
In particular, clone HIO has significant activity on the primers with two Smatches. HIO is a chimera of 2*. aquaticus wild type (residues 4 to 20 and 221 to
& )), TS (residues 1 to 3 and 641 to 834) and T. thermophUus (residues 21 to 220). HIO ha five detectable crossover sites and 13 point mutations, of which 4 are silent (F74-»X,
O-»L, P300-+S, T387-*A, A44I-»V, A519-»V Q536-
E ample 8: Selecting for a 4 mismatch extension polymerase.
C R emulsificatioa and selection was perfbizned on the StEP Taq, Tth and Tfl library es entially as described (Ghadessy, Ong et al. 2001). The library had previously been cl nod into pASK75 (see example €). The aqueous phase was ether extracted and re lication products were purified using a PCR purification kit (Qiagen, Chatsworth, C. ) including a -wash with an 35% GnHCl. 7 ul of purified replication products (from 41 were digested with lul DpriL (20 Units) to remove plasmid DNA and treated with 2 il ExoSAP (USB) to remove residual primers for lit at 37°C and reamplified with 01 nested primers (GTAAAACGACGGCCAGT and CAGGAAACAGCTATGAC, 94 °C 2 minutes, and then 30cycles of 94 °C 30 seconds, 50 °C for 30 seconds and 72 °C fo 5 minutes with a final 65 °C for 10 minutes). ReampKfication products were di ested with Xbal and Sail, recloned into pASK75 and transformed into E. coli as ah >ve.
In parallel an alternative selection approach was used: the induced library was en ulsified as above with the additional presence of biotmylated dUTP and incubated at 94°C 5 minutes, 50 °C 1 minute and 72 °C 1 minute. The aqueous phase was ether ex ractod, the DNA in the aqaeous phase was precipitated by addition of 1/10 volume of 3M NaAc, 1 ul glycogen and 2.5 volumes of 100% ethanol. This was then in ubated for 1 hour at -20 °C, spun for at BOOOrpm for 30 minutes in a benchtop m ^centrifuge, washed with 70% ethanol and resuspended in 50 ul buffer EB (<; agen). 20ul of Dynabeads (DynaL Biotech) were washed twice and resuspended in 2( il of bead buffer (lOmM Tris pH 7.5, ImM EDTA, 0.2M NaCl) The washed beads wi « uien mixed with the selection in a total volume of 0.5 col bead buffer and then in ubated overnight under constant agitation at room temperature to capture bi tinylated products. Beads were washed twice in bead buffer, twice in buffer EB and fir illy resuspended in 50 ul bead butler. The resuspended beads were reamplified with ou nested primers (sequences and programme as above) and recloned and transformed

int

& coli as above.

'o sets of mismatch primers with four mismatches at their 3* end (underlined) (51-CJ iG GAA ACA GCT ATG ACA AAA GTG AAA TGA ATA GTT CGA CTTTT-3' aB|(i 5'-GTA AAA CGA CGG CCA GTC TTC ACA GGT CAA GCT TAT TAA

Gj
' as the fiist set and 5'-CAG GAA ACA GCT ATG ACC ATT GAT AGA
GfT ATT TTA CCA CAGGG-3' and 5'-GTA AAA CGA CGG CCA GTC TTC pA GGT CAA GCT TAT TAA GGTG-3' as>e second set) were used in the

eri

ulsion as two separate sources source of selective pressure.

Bij
HJ|J resultant clones from both CSR and CST were screened and ranked by PCR assay, efly, 2 ul of induced cells were added to 20 ul of PCR mix with the relevant 4

ani
3' C;
T/
po
match primers. Clones that produced a band were then subjected to further analysis their activity on single, double and quadruple mismatch primers (single mismatch prijners: S'-CAG GAA ACA GCT ATG ACA AAA ATC TAG ATA ACG AGO GA-nd S'-GTA AAA CGA CGG CCA GTA CCA CCG AAC TGC GGG TGA CGC A GC£ 3'; double mismatch primers; GAG GAA ACA GCT ATG ACT CGA A AAA TCT AGA TAA CGA CC and GTA AAA CGA CGG CCA GTT TAT A. CCA CCG AAC TGC; four mismatch primers above.) was investigated, ymerases that could extend all of these mismatches were found, though many unerases could do only one of the mismatches and none could do all.

Thl
plasnaid DNA of the ten best clones was then purified and shuffled as described
ab( re (StEP, (Zhao, Giver et aL 1998)). This was then purified, cut and cloned and the resultant library was subjected to another round of CSR as described (Ghadessy, On ; et aL 2001). The same two sets of mismatch primers with four mismatches at their 3' < ad were used in the emulsion as two separate sources source of selective pressure.

Thi by

} was men dealt with as above and the resultant clones were screened and ranked PCR assay (as above). Once again, polymerases that could extend all of these

the disi
mismatches were found (see Table), though many polymerases could do only one of
mismatches and none could do all. There was a notable increase in clones laving mismatch activity over the first round.

The bejft clones from the second round were combined with the best clones from the first roind on a. 96 well plate and were subjected to further screening.

The following table is a summary of the results.

Al Is Tfe polymerase; A2 Tfl; A3 Taq; A4 "Ml; A5 M4; A6 H10 (see previous
I exafiple. 1A7 to 1D12 are first round clones (where 1 indicates that these are first
L clones), 2E1 to 2H12 are second round clones (where 2 indicates that these are I round clones) '
Thf best first and second round closes were shuffled as described above and subjected
I to Knottier round of CSJL The same two sets of mismatch primers with four
milmatches at their 3' end were used in me emulsion as two separate sources of

se^ctive pressure. This was then dealt with as above and the resultant clones were led and ranked by PCR assay (as above). Once again, polymerases that could 1 all of these mismatches were found. In particular, clones 3B5. 3B8,3C12 and
31} 1 (where 3 indicates that these are third round clones) were able to extend primers four mismatches. See figure 9

Some jlromising clones were sequenced. All of the polymerases displayed a similar
I
competition: the first part of the protein, roughly conesponding to the 5-3 exonuclease
of the polymerase, was derived from Tth, whilst the remaining part of the
proteii was derived from Taq. Four point mutations (L33-»P, E78-»K, D145-»G and
1 E$22f»K) re-occurred in the majority of sequenced mutants and one (BIO) had
an extra 16 araino acids at its C terminus through a frame shift at position 2499| Tfl was highly underreprcsected, although some of its sequence was present.
Example 9: Hairpin ELISA to measure polymerase activity.

Thepelow protocol is a sensitive metfaod to measure pqlymerase activity both for the incofporation. of unnatural nucleotide substrates (added to the reaction mixture) or the sion or replication of unnatural nucleotide substrates (incorporated as part of the L oligo).
assay comprises a hairpin oligonucleotide which constitutes both primer and slate in one. In contains as part of the hairpin a biotinylated dU residue, which ve capture of the hairpin oligonucieotide on streptavidin-coated surfaces.
oligonucleotide folds up into a hairpin with a 5' overhang, which serves as the blate strand for the polymerase (typical sequence: 5'- AGC TAC CAT GCC TGC A«I G CAG TCG GCA TCC GTC GCG ACC ACG TT5 TTC GTG GTC GCG ACG lr_GGC_G-3% bases involved in hairpin formation are underlined, 3' base is in ld,5=dU'biotin).
pension reactions are carried out in the presence of small amounts of a labelled leotide typically DIG-16-dUTP. Product is captured ( for example on a streptavidin fated BLISA plate ) and incorporation of labelled nucleotide into the product strand

is measured, (using for example an anti-DIG antibody) and taken as a measure of polym|rase activity.
i
- Metht
reactions are earned out in Ix Taq buffer including 1-lOOnM of hairpin primel and lOOuM dNXP mixture (comprisixxg 0.3-30% dUTP-DIG), rypically incubftcd at 94°C for l-5min, followed by incubation at 50°C for l-5min, followed by ion at 72°C for l-Smia. (I-10jil) Reaction products are added to Streptavidin coatej ELISA plates (Streptawell, Roche) in 200ul PBS, 0.2% Tween20 (PBST) and : at room temperature for lOmin to Ib, ELISA plates are washed 3x in PBST and |00jd of anti-DIG-POD Fab2 fragment (Roche) diluted 1/2000 in PBST is added and me plate is incubated at room temperature for lOmin to Ih, The plate is washed 3-4x iiJPBST and developed with an appropriate POD substrate.
te 10: misjbatoh extension clones
i previously selected for their ability to extend from a 4 basepair mismatch were . for their ability to incorporate a variety of nucleotide analogues.
i were grown at 30°C overnight in 200ul 2XTY + ampicillin (lOOug/ml).
ST + ampiciHin lOOng/ml) overday culture was started from the ovenught grown for 3 hours at 37°C After 3 hours protein expression was induced by the Ition of SOfil of 2XTY + anhydrous tetracycline (Sngtel) to the culture which was . allowed to grow for a further 3h at 31C. The cells were pelleted at 2254xg for 5 : and the growth medium removed by aspiration after which the cell pellet was in lOOjtl IxTaq buffer (lOmM Tris-HO, pH 9.0, LSmM MgCl2, 50mM 0.1% Triton X-100, 0.01% (wA^) stabiHser, HT Biotechnology Ltd), suspended cells were lysed by incubation at 85?C for 10 minutes and the cell debris ; pelleted at 2254xg for 5 minutes.

ELISA protocol: Exteniion reaction.
Reactions were performed in a final volume of 12.5}*1 comprising:
Ix T4 buffer (lOxnM Tris-HO, pH 9.0, l-5mM Mgd2) 50mM KC1, 0.1% Triton X-00, 0.01% (w/v) stabiliser; HT Biotechnology Ltd).
50 pdjoles of primer.

25 uHl of each dNTP (minus the nucleotide analogue) of which 10% (2.5fiM) of the
is digoxigenin-1 1-dUTP and 90% (22,5/jM) is dTTP.

25 ulk the nucleotide analogue. 2.5^|ofcelllysate.
The Reaction conditions were:
9Se<|5 minutes; 50 °C 5 minutes; 72°C 5 minutes.
Detction reaction:
5 i of the extension reaction was added to 200^1 of PBS-Tween (Ix PBS; 0.2% L 20) in StreptaWell high bind plates (Roche) and allowed to bind for 30 minutes at r|om temperature.The plate was washed 3X in PBS-Tween after which was added 20o||il PBS-Tween + anti-digioxigenin-POD Fab fragments (antibody diluted 1/2000; lie). The antibody was allowed to bind for 30 minutes at room temperature.
[plate was washed 3X in PBS-Tween and 200^1 of the substrate added (per ml lOOlil of 1M NaAc pH 6,0, lOfd of DAB, Ipl of HaOz, the reaction was allowed to p after which it was stopped by adding 1 OOjul o
Edberimentl. ELISA with Flnoresceia 12-dAXP:
: ability of clones selected for 4-mismatch extension to incorporate Fluorescein 12-Perfcin Elmer) was assayed using fee primer FITC4. Ite lysates used were loentrated 4-fold.

I!
Expeilmept IL ELISA with Biotin 11-dATP:
The alSility of clones selected for 4-mismatch extension to incorporate Biotin 11-dATP (Perkijji Elmer) was assayed using the primer FUCIO. The lysates used were concejlitrated 4-fold.
I
ExpeKmept HI. ELISA with CyDve 5-dCTP;
II The ability of clones selected for 4-mismatch extension to incorporate CyS-dCTP
(Amareham Biosciences) was assayed using 1he primer ELISAC4P. The lysates used werell concentrated 4-fold,
lent IV. EMSA with CvPve 3-dUTP:
The Ibility of clones selected for 4-mismatch extension to incorporate CyDye 3-dUTP (Amjsrsham Biosciences) was assayed using the primer ELISAT3P. The lysates used concentrated 4-fold. The DIG labelled dUTP in the extension reaction was repl|ced with Fluorescein 12-dATP and the incorporation of Fluorescein 12-dATP was ted by anti-Fluoresccin-POD Fab fragments (Roche).
lent V. Abasic site ELISA...
ability of clones selected for 4-mismatch extension to bypass abasic sites was kyed using the primer PscreenlAbas (AGC TAG CAT GCC TGC ACG CAG 103 GCA TCC GTC GCG ACC ACG TT5 TTC GTG GTC GCG ACG GAT
II
G(|C G, 1 = abasic site
5=| dU biotin). The lysates used were concentrated 4-fold,
selected for 4-mismatch extension were assayed for activity with different bstrates using an ELISA assay.
s Tth Wild-type 'I?? Wild-type •• Tag Wild-type 1 Taq mutant Ml

58
A5= Tkq mutant M4 mutant HI 0
Rows -D Clones isolated after 1 round of 4-mismateh selection Rows E-H Clones isolated after 2 rounds of 4-mismatch selection
The r|sults are shown in Figure 8.
Expdfment V. Abasic site and 5-hydroxyhydantoin bypass
Polyrferases 3A10 and 3D1 were investigated further for their abflity to bypass abasic
sites fnd 5-hydroxy hydantoins, which are both known to exist in damaged DNA such
as fofmd in ancient samples , using the ELISA based activity screen as described
abovi. Both polymerases were more proficient at lesion bypass than wild type Taq by
up tqltwo orders of magnitude.
ii Tbebydantion phosphoramidite was synthesised by standard procedures starting from
the f ydantoin free base. Glycosylation of the silylated hydantoin base in the presence of tlnCtV) chloride with the ditoluoyl (alpha) chlorosugar gave rise to two N-
asylated products which were separated and characterised by 2D-NMR explriments. ITie tolyl groups were removed with ammonia to yield tbe free nucfeoside which was dimethoxytritylated and phosphytylated in the usual manner.
hairpin primer to assay hydantoin bypass was: 5'-AGC TAG CAT GCC TGC ACID CAG XCG GCA TCC GTC GCG ACC ACG TTY TTC GTG GTC GCG AQB GAT GCC G-3', X=hydantoin, Y=Biotin-dU.
ThI sequences of the clones referred to in Examples are shown below; For the avfidance of any doubt, the first sequence provided in each section is the nucleic acid sequence. The second sequence provided is the corresponding ammo acid sequence of
T
1 clone.
CGTrc^
^^ rACACCGCGGX3GACXKKm:aAaXCCX3AC<^^

75
Sevea jolymerases were assayed for their ability to bypass abasic sites in a primer assay (Figure 11).
Primen extension assays were essentially as described in (Ghadessy et al., 2004). Brieflf, undamaged oligonucleotides and a Slmer containing a synthetic abasic site were Hsyntttesized by Lofstrand Laboratories (Gaithersburg, MD) losing standard ques and were gel purified prior to use. A 20mer primer (LESJ20P) with the seqJLce S'-CGTGGTCGCGACGGATGCCG-S' was S'-labeled with [3*P]ATP (5000

Ci/ramole; 1 Ci=37 GBq) (Pharmacia ) using T4 polynucleotide kinase (Invitrogen,
Carl|bad CA). Radiolabeled primer-template DNAs were prepared by annealing the
5'[ f 1 labeled 20mer primer to one of the two following Slmer templates (at a primer ] Jlatc ratio of molar 1:1.5). 1) undamaged DNA (UNDT51T); S'-AGC TAG CAT
GCJC TGC ACG AAT TCG GCA TCC GTC GCG ACC ACG GTC GCA GCG-3'; 2)

an Jligo (LAB AS IT) containing a synthetic abasia site (indicated as an X in bold
forj); 5'-AGC TAG CAT GCC TGC ACG ACA XCG GCA TCC GTC GCG ACC ACJG GTC GCA GCG-31. Standard replication reactions of 10 pi contained 40mM •HC1 at pH 8.0, 5mM MgCli, lOOpM of each ultrapure dKTP (Amersham ia Biotech, NJ), lOmM DTT, 250 ng/ml BSA, 2.5% glycerol, 10 rM S'[32P] ler-template DNA and 0.1 Unit of polymerase. After incubation at 60°C for various tidies reactions were terminated by the addition of 10 jol of 95% formamide/ 10 mM t>TA and the samples heated to 100eC for 5 min. Reaction mixtures (5 fil) were bjected to 20 % polyacrylamide/7 M Urea gel electrophoresis and repEcation p|oducts visuaKzed by Phosphorlmager analysis.

Polymfrases A10 was the most active and was chosen for further analysis (Figure
• nomenclature) on abasic sites and cyclobutane thyroine-thyrtune diraers
(CPD} A10 was clearly better at both abasic site and CPD extension and bypass than both ^lld type and Ml.
14: Error rate iayestigatippr oJL mismatch extension clones as ed b MutS JELISA-
ReI«jbdTspecificitvnMgfatbe expected to be actdevedat the cost of lower fidelity. We

used 1 MutS ELISa to investigate this possibility.

Mutsl is an E. coh derived mismatch binding protein that binds single base pair
mismatches or small (1-4 base) additions or deletions. It can be used to monitor PCR
I fidelity in an ELISA based assay pebble et aL, 1 997).
fobilised Mismatch Binding protein plates (Gcnecheck, Ft Collins, USA) were for fidelity measurements as per manufacturer's instructions, essentially as described in (Debbie et aL, 1997).
Tb| mutation rate of Dl was compared that of wtTaq and Ml Ml was already known to Live a modestly increased mutation rate (approximately 2 fold) (Ghadessy et al., 2dp4). The data presented here suggests that Dl has a 2 fold increased error rate ccjfnpared to Ml and a four fold increased error rate compared to wtTa$. This sonds approximately to a 1 in 2500 error ratio and is sufficiently low to not be bhlematic for many applications.

Examile 15: Investigation of mismatch extension dones for the amplification of damaged PNA such as is found ia ancient samples.
DNA recovered from ancient samples is invariably damaged, limiting the information
it cadi yield. Polymerases that can bypass damage (such as abasic site or hydantoins)
11 might therefore be useful in increasing the information that can be recovered from
auciejpt samples of DNA.
ent 1: A mismatch extending polymerase can amplify previously tm-cave hyena DNA
Several samples of cave hyena (Crocuta spelaea) were extracted and analysed. Of thole, seven samples (see Figure 12 for the list) failed to ever produce an ampli fication projpuct.
samples were chosen to test the efficacy of the expanded substrate spectrum poUymerases.
has a slightly reduced kcat/Km, 14% of Taq wild type, and is hence slightly less icient in PCR. Therefore, Ml was blended with a commercial preparation of Taq rTaq (HT biotechnology Ltd)) in a ratio of 1 unit to 10 and compared to Taq in t absence of ML It was hoped that if Ml could bypass the blocking lesions, then the type Taq would amplify the resulting translesion synthesis product. On two

separate occasions, the Ml/SuperTaq mix was able to produce an amplification

produdt whereas SuperTaq alone did not (see figure 12 for one example)

The DJNA was cloned and sequence, and found to differ in two positions (A71-K3,

77A rG) from fbe expected sequence. This could either be a miscoding lesion
I resulllng from a deamination of C or a population variant sequence not seen previously
l
in aJJjNA, Indeed, both mutations exist in modem spotted hyena (Croatia crocuta), argiujpg for the second interpretation. Of the 10 sequences obtained from the same successful PCR, two each had a further unique single mutation, an A to G in different places. These are most likely errors incurred during amplification. Such errors are seen in aDHA PCR and are one reason why multiple sequences need to be obt^|ned from the same PCR product.
jination problems prevented an exhaustive analysis of the benefits of Ml
poljfmerase. However, this result strongly suggested that a suitable altered polymerase
1 could be usefully applied to aDNA.
periment 2: A blend of mismatch extending polymerase needs less ancient DNA for a successful PCR.
Pved ineffective (data not shown). A new two step nested PCR strategy was used. In , first step, the aDNA is amplified over 28 cycles with either SuperTaq or the blend, the second step, the first PCR is diluted 20 fold in a secondary clean room and

amplijled with SuperTaq vising in-nested primers. This is the approach subsequently used tb compare SuperTaq and the blend
ii
BrieflK', 2 jil of ancient sample were added to a 20 nl PCR in SuperTaq buffer (HT
I Biotefch) wifli 1 uM of the appropriate primers (see Figure 13), 2[iM of each
I deoxjoibonucleoside triphosphate (dNTP) as well as 0.1 ul of SuperTaq or an equal
volume of mutant polymerases and amplified for 28 cycles. This PCR was set up in a clean room following precautions appropriate for aDNA.. The first step PCR was then diluted 1 in 20 in a secondary clean room aad thermocycled for a further 32 cycles witathe same buffer and dNTPs conditions, using in-nested primers and SuperTaq. No temf late controls were used to test for contamination.
A tfro fold dilution series of aDNA with equal volumes of SuperTaq and the blend
(ant therefore approximately equal activities, with the blend slightly less active) was
I
performed and repeated this four times
experiment showed that the blend was more likely to produce a band at a lower atration of aDNA than SuperTaq. This therefore represented the second beriment that indicated that the mismatch extension polymerases were more prjbficient at amplifying aDNA than wild-type Taq.
periment 3: The mismatch extension polymerases perform consistently better in lent DNA PCR.

Sampll heterogeneity and the inherent stochasticity of aDNA analysis make the in.terpjtta.tion of a single positive or negative PCR problematic. To address this,
multrdle PCRs of a same sample and count the number of successful PCR
1
amplifications at a limiting sample dilution were performed. Comparison of SuperTaq
with me blend would allowed a statistical analysis. As the amount of aDNA required
for this type of approach is large, samples previously shown to be of high quality were
chosan and tested at limiting dilutions to increase the amount of material available for analysis. A short target sequence was chosen to allow maximal dilutions.
has the additional advantage thai at a sufficiently high dilution, the undamaged DNA. wiH have been diluted out, leaving only damaged template. In such conditions, the Ipifierence between a polymerase mat can bypass blocking lesions and one that caniot should become clearly apparent
A total of nine experiments at limiting amounts of aDNA, where the ?CR would only be stochastically successful (Figures 14 and 15 ) were performed. In eight out of nine experiments, the blend resulted in more successful PCRs than SuperTaq. The
ibability of this occurring by chance is 1.76"%, as determined by binomial distribution analysis. It is commonly accepted that chance can be dismissed as an
slanation when an event is expected to occur at 5% probability or less.
w
fe can therefore state that this effect is not due to chance and that the blend is aeatedly performing better than SuperTaq in the conditions of the experiment. This

proves! beyond reasonable doubt that the mismatch extension polymerases are a more sensitive tool for the recovery of ancient DNA sequences.
16: Selection of a polymerases capable of replicating the unnatural base analogue 5-nitromdol (5NI)
We selected for extension and bypass of 5NI directly from the polymerase chimera library described in example 8 using an analogous strategy to the mismatch selection usmd flanking primers (5'-CAG GAA ACA GCT ATG ACA AAA ATC TAG ATA ACq AGG GCA 5NI-3', 5*-GTA AAA CGA CGG CCA GTA CCA CCG AAC TGC TGA CGC CAA GC5NI-3') comprising 5NI (or a. derivative) at their 3' ends. Aftef round 3 , we used flanking primers (5'-CAG GAA ACA GCT ATG ACA AAA ATQ TAG ATA 5NICG AGG GCA 5NI-3', 5'-GTA AAA CGA CGG CCA GTA CCA C5NIG AAC TGC GGG TGA CGC CAA GC5NI-3') comprising internal 5NI (or | derivative) as well as 3' terminal 5NI (or a derivative) to increase selection i for 5NI replication.
Fivl rounds of selection yielded a number of clones with greatly increased ability to kcate 5NL Among the best clones were round 4 clone 4D11 and round 5 clone 51
TCCG|cAAGGGTTAQ-3'
jl
5D4 lamino acid sequence;
LAWLEVEVGIGEDWIiSAKG*
aple 17: Expanded spectrum of polymerases selected for replication of 5NI
15 polymerases selected for replication of 5NI were tested for activity with a range of substrates using the hairpin ELISA assay described in example 8. tUTP and ceATP were gifts fiom the laboratory of P. Herdewijin, Rega Institute, Kafliolieke Universiteit , Belgium. Results are shown in Figure 14

The amlity of round 5 clones selected for 5NI replication extension to sequentially incoipjrate 2 or 3 of the TNA UTP derivative (3', 2>beta-L-ttoreonyl-UTP was assaydh using the hairpin primers (ELISAT2p: S'-TAG CTC GGT AA CGC CGG CTT ICG TCG CGA CCA CGT TX TTC GTG GTC GCG ACG GAA GCC G-3' , ELJsJpp: 5'-TAG CTC GGT AAA CGC CGG CTT CCG TCG CGA CCA CGT TX TTC GTG GTC GCG ACG GAA GCC G-3' (X=dU-biotin (Glen research)). The lysate| used were concentrated 4-fold. ELISA protocol was a described except that The DIG labelled dUTP in the extension reaction was replaced with Huorescein 12-dATa (Perkin-Elmer) (at 3% of dATP) and the incorporation of Fluorescein 12-dATP was detected by anti-Fluorescein-POD Fab fragments (Roche).
2.EilSAwithceATP:
The ipbiUty of round 5 clones selected for 5NI replication extension to sequentially the cyclohexenyl ATP derivative ceATP was assayed using the hairpin (ELISA2p: 5'-TAG CTC GGA TTTT CGC CGG CTT CCG TCG CGA CCA CGt -TX TTC GTG GTC GCG ACG GAA GCC G-3' , (X=dU-biotin (Glen resejferch)). The lysates used were concentrated 4-fold.
3. flLISA with CvPye S-dCTP and CvDve 3-dCTP:
Th| ability of round 5 clones selected for 5N1 replication extension to sequentially incorporate the fluorescent dye-labelled nucleotides Cy5-dCTP and Cy3-dCTP (Ajhersham Biosciences) was assayed using the hairpin primers (ELISA2p: 5'- TAG CCA GGG CTC CGG CTT CCG TCG CGA CCA CGT TXT TCG TGG TCG
CCjA CGG AAG CCG -3' , (X«dU-biotin (Glen research)). The lysates used were
| concentrated 4-fold.
4.[{Ba$ic site bypass ELISA jl
Tjfe ability of round 5 clones selected for 5NI replication extension to bypass an abasic
was assayed using the hairpin primer (PSoreenlabas: 5'-AGC TAG CAT GCC C ACG CAG YCG GCA TCC GTC GCG ACC ACG TTX TTC GTG GTC GCG

ACG DAT GCC G -3' , (X*dU-biotin, Y=abasic site (Glen research)). The lysates used were concentrated 4-fold,
Example 18: Primer extension reaction with polymerases 4D11 and 5D4 1: Ejiension opposite 5-nitroindole.

?riic|r5'- TAATACGACTCACTATAGGGAGA
Temjlate 31- ATTATGCTGAGTGATATCCCTCT5ATCGAT
5 - 5-Nitroindole

Priijler extension reactions were carried out as follows:
lOl of 32P-labelled primer and lOOpmol of template in a volume of 44^ were .ed in IX Taq buffer. 4D11 or 5D4 polymerase as cell lysate (6ul) was added reactions were incubated at 50°C for 15 minutes followed by addition of one in total volume of SOjil, final dNTP concentration 40uM). 8^1 samples we* taken at various time points and added to 8ul stop solution (7M urea, lOOmM 'A containing xylene cyanol F). At the end of the time course the remaining 3 s were added (final concentration each dKr? 40uM) and reactions incubated at 50HC for a farther 30 minutes. Reaction samples were electrophoretically separated usijig 20% polyacrylamide gels at 25W for 4 hours. The resultant gels were dried and scanned using a phosphorimager (Molecular Dynamics). Data was processed using the fgram ImageQuant (Molecular Dynamics). Results are shown in Figures 35, 36:

SimilaJ reactions using Taq, Tth and Tfl wild-type polymerases under identical conditions leads to almost undetectable extension reactions (data not shown).
2. Incorporation and extension of 5-nitroladole-5'-tripbosphate (SMTP).
Prim* 5'- TAATACGACTCACTATAGGGAGA
Temfate 3'- ATTATGCTGAGTGATATCCCTCTXGTCA
A,T,C,G
• extension reactions were carried out as follows:
SOpaiol of 32P-labelled primer and lOOpmol of template in a volume of 44u.l were
I anndfeled in IX Taq buffer. 4D11 or 5T>4 polymerase as cell lysate (6^1) was added
andlteactions were incubated at 50eC for 15 minutes followed by addition of dSNTTP (lu|in total volume of 50(11, final dNTP concentration 40pM). 8pJ samples were taken at Canons time points and added to 8^1 stop solution (7M urea, lOOmM EDTA con|aining xylene cyanol F). At the end of the time course the 4 native dNTPs were (final concentration each dNTP 40uM) and reactions incubated at SO°C for a 30 minutes. Reaction samples were electrophoretically separated using 20% po^acrylamide gels at 25W for 4 hours. The resultant gels were dried and scanned ; a phosphorimager (Molecular Dynamics). Data was processed using the program Juant (Molecular Dynamics). Results are shown in Figures 17, 18) :
NI-NI self-pair is also formed exceptionally well, though further extension is (data, not shown). Similar reactions using Taq, Tth and Tfl wild-type lymeraees under identical conditions leads to almost undetectable extension ictions (data not shown).

Example 19: Array manufacture and hybridization using Ml.
Targeli were prepared by PCR amplification of 2.5kb Tag gene using primers 29, 28
I
or 2ko| of the HIV pel gene using primers 30, 31. Salmon sperm DNA (Lrvitrogen)
was prepared at 100ag/ul in 50% DMSO. FTTC and Cy5 probes were prepared by
I PCR* implificatioia of 0.4kb fragment of Taq using primers 8, 28 with either 100%
(FTTCJlOOMi) or 10% ofdATP (FITClOMi, FTTClOr,) replaced by FtTC-12-dATP or 10% Jf dCTP replaced by Cy5~dCTP (Cy5Taq). Cy5 and Cy3 random 20mers (MWG) werej|used at 250nM- Targets were purified using PCR purification kit (Qiagen) and in 50% DMSO and spotted onto GAPSH aminosilane-coated glass slides (Corfing) using a MicroGrid (BioRobotics). Array hybridizations were performed ling to standard protocols :
:d slides were baked for 2hr at 80°C, incubated with agitation for SOmin at 42°C in 5x |SC/0.1% BSA Fraction V (Roche)/0.1% SDS, boiled for 2min in ultrapure water, Led 20x in ultrapure water at room temperature (RT), rinsed in propan-2-ol and dried
in aj clean airstream. 50ng of FTTC- and Cy5-labelled probes were prepared in 20|al of
I hybfidization buffer (ImM Tris-BCl pH7.4, 50mM tetiasodium pyrophosphate, Ix
its solution, 40% deioaised formamide, 0.1% SDS, lOOjag/ml sheared salmon
spefm DNA). Each sample was heated to 95QC for 5min, centrifiiged for 2mia, applied to
I
isurface of an array and covered with a 22x22 mm HybnSlip (Sigma). Hybridizations
performed at 48°C for 16hr in a hybridization chamber (Coining). Arrays were wished once with 2xSSC/0.1% SDS at 65°C for 5min once with 0,2xSSC at RT for 5mui i twice with O.OSxSSC at RT for 5mm, Slides were dried in a clean airstream, scanned
with analys
Comp the p subst

89
AnayWoRx autoloader (Applied Precision Instruments) and the array images pd using SoftWoRx tracker (Molecularware).
,ete substitution of natural nucleotides with their unnatural counterparts altered operties of the resulting amplification products. For example, fully alphaS uted DNA was completely resistant to nuclease digestion (not shown).

The with
tosf
rang
leve
peri
gen
and
(Q tar; FF of
Ukb fragment, in -which all adenines (dA) on both strands had been replaced w4lLi||FrrC-12-dAMP (FTTClOOMi), displayed extremely bright fluorescence. The frequency of fluorophore incorporation per 1000 nucleotides (FOI) is commonly used scify the fluorescence intensity of a probe. FOIs of microarray probes commonly from 10-50, while FirC100Mi has an FOI of 295. To investigate if such a bigii of jEluorophore substitution would affect hybridisation characteristics we >rmed a series of microarray experiments. We compared the fluorescent signal srated by FITClOOMi with equivalent probes generated using either vttTaq or Ml replacing only 10% of dAMP with FIT(M2-dAMP (FITClOtaq, HTC10M! 30)), In competitive co-hybridisation with a standard Cy5-labelled probe aq), FITClOOMi hybridised specifically only with its cognate Taq polymerase ;et sequence and not with any non-cognate control DNA. Hybridisation of 'ClOOMi generated an up to 20-fold higher specific signal than equimolar amounts foe FTTC10 probes (Fig. 20) without showing increased background binding (Figs.

19

21).

Example 20 : Mutation rates & spectra of selected polymerases Ml and M4.
Mutation rates were determined using the mutS EUSA assay26 (Genecheck, Ft.
Colliik CO) according to manufacturers instructions. Alternatively, amplification proddfcts derived from 2x50 cycles of PCR of 2 targets with different GC content (38% GC), Taq (68% GC)) were cloned, 40 clones (SOObp each) were sequdhced and mutations (wtTag (51), Ml (75)) analyzed.
Promiscuous mismatch extension might be expected to come at the price of reduced fidelfy, as misincorporation no longer leads to termination. Measurement of the overall mutation rate using both the MutS assay (Fig 22A) and direct sequencing of

amp

fication products, however, indicated an only modestly (1.6 fold) increased

mute ion rate in Ml (or M4). However, Ml displays a significantly altered mutation . compared to wiTaq, with a clearly increased propensity for transversions, in partfular G/C->C/G transversious (Fig 22B).
Example 21: Processivity
1 Naturally occurring translesion polymerases are mostly poorly processive. We therefore
inve stigated, if processivity of Ml and M4 was similarly reduced but found that, even at

the

owest enzryme concentrations, primer extension and termination probabilities by Ml

and|M4 closely matched those of wtTaq (Fig.23), indicating that botfx Ml and M4 exhibit pro|essivity equal (or higher) than wtTog. This is also reflected ia the striking proficiency of Ml in long-range PCR (see example 6).

Proce|sivity was measured using a primer extension assay the presence and absence of trail DNA. Termination probabilities were calculated according to the method of Kokoska et al
OUgoJuicleotide primer 32 (5'-GCG GTG TAG AGA CGA GTG CGG AG-35) was 32P-bjbelled and annealed to (he template 33 (5'-CTC TCA CAA GCA GCC AGO
CAAjGCT CCG CAC TCG TCT CTA CAC CGC TCC GC-3') (at a primer/template
I ratio |>f molar 1/1.5). wtTaq (0.0025nM; 0.025nM; 0.25nM), M/(0.05nM; 0.5nM;
and M4 (O.OSnM; 0-5nM; 5nM) were preincubated with the primer-template ON/)! substrates (lOnM) in lOmM Tris-HCl at pH 9,0, 5mM MgCl2,50mM KC1,0.1% Tritefi X100 at 25°C for ISmin. Reactions were initiated by addition of lOOuM dNTPs with pr without trap DNA (1000-fold excess of unlabeled primer-templates). Reactions werd performed at 60°C for 2min. Preincubation of poiymerases with the trap DNA subslrate and labelled primer-template before the addition of dNTPs completely abo] shed primer extension (not shown) demonstrating trap effectiveness. Thus, in Hie pres oce of trap DNA/ all DNA synthesis resulted from a single DNA binding event.
Gel band intensities were calculated using a Phosphoimager and ImageQuant (both
I
Mojfecular Dynamics) software. Percentage of polymerase molecules, which extended
(in
info [isity of the band at position 1, 2... Termination probabilities (t) were calculated rding to the method of Kokoska et al1, whereby T at a particular template position waf calculated as the intensity of the band at this position divided by the sum of the intfnsity of this band and me band intensities of all longer products.
pritfers to the end of the template was calculated using the formula: tot x 100% / i, where la is the intensity of the band at position 22 or 23; II, 12... is the

AB publications mentioned in the above specification are herein incorporated by reference. Various modifications and variations of the described methods and system of thejpresent invention will be apparent to those sldlled in the art without departing from ihe scope and spirit of the present invention. Although the present invention has
II
been Ipescribed in connection with specific preferred embodiments, it should be ptood that the invention as claimed should not be unduly limited to such, specific embodiments. Indeed, various modifications of the described modes for carrying out the invention which are obvious to those skilled in biochemistry, molecular biology and biotechnology or related fields are intended to be within the scope of the folloflving claims.
reuces
1. I Schaaper, R.M. (1993) J.BtoLdiem. 268,23762-23765.
2. Li, Y., Korolev, S. & Waksman, G. (1998) Embo J. 17,7514-7525.
3. I Doublie, S,, Tabor, S., Long, A. M., Richardson, C- C. & EUenberger, T.
(1998) Nature 391,251-258. ;
4. I Johnsoa, S. I., Taylor, I S. & Beese, L, S. (2003) Proc. NatL Acad. Sci. USA
100,38895-38900.
5. Li, Y., Mitaxov, V. & Waksman, G. (1999) Proc. NatL Acad. Sci. USA 96,
9491-9496.
6. Astatke, M., Ng, K., Grindley, N. D. & Joyce, C. M. (1998) Proc. NatL Acad.
Sci. USA 95,3402-3407.
7. Patel, P. H, & Loeb, L. A. (2000) J. Biol Ckem. 275,40266-40272.


8. 9. 10.
11. 12.
13.
14
15. 16.
17. 18.
19. 20. 21-
22. 23. 24. 25.
26,
27.

Jestin, J. L, Kristensen, P. & Winter, G. (1999) Angen Chan. Int. Ed. 38,
1124-1127.
Xia, G., Chea, L., Sera, T., Fa, M., Sdwltz, P. G. & Romesberg, F. E. (2002)
Proc. Nad. Acad. Sd. USA 99,6597-6602.
Ghadessy, F. J., Ong, I L. & Holliger, P. (2001) Proc. mi Acad. Sd. USA 98,
4552-4557.
Tawfik, D. S. & Griffiths, A- D. (1998) Nature Biotechnol 16,652-656. Huang, M.-M., Aroheim, N. & Goodman, M. F. (1992) Nucleic Acids Res. 20,
4567-4573.
Kool, E. T. (2000) Curr. Op. Chem. Biol 4,602-608.
Kwok, S., Kellogg, D. R, McKinney, N., Spasic, D., Goda, L., Levenson, C. &
Sninsky, J. J. (1990) Nucleic Acids Res 18,999-1005.
Eom, S. H, Wang, J. & Steitz, T. A. (1996) Nature 382,278-281.
Creighton, S., Bloom, L. B. & Goodman, M. F. (1995) Meth. Enzymol 262,
232-56.
Mendelman, L. V., Petruska, J. & Goodman, M. F. (1990) J. Biol Chem. 265,
2338-2346.
Boudsocq, F., Iwai, $., Hanaoka, F. & Woodgate, R. (2001) Nucleic Adds Res
29,4607-4616.
Venua, S. & Eckstein, F. (1998) Annu. Rev. Biochem. 67, 99-134.
Loakes, D. (2001) Nucleic Acids Res 29, 2437-2447.
Berger, M., Wn, Y, Ogawa, A. K, McMinn, D. L, Schultz, P. G. &
Romesberg, F. E. (2000) Nucleic Acids Res, 291 1-2914.
Barnes, W. M. (1994) Proc. Natl. Acad. ScL USA 91, 2216-2220.
Goodman, M. F. (2002) Anna. Rev. Biochem. 71, 17-50.
Konkd. T. A. & Bebenek, K. (2000) Annu. Rev. Biochem. 69, 497-529.
Patel, P. a, Suzuki, M., Adman, E., Shinkai, A. & Loeb, L. A. (2001) /. MoL
Biol 18, 823-837.
, F. C., Stoffel, S, SaM, R. K., Chang, S. Y, Landre, P. A., Abramson,
R. D. & Gelfand, D. H. (1993) PCR Meth. Appl. 2, 275-87.
Tada, M., Omata, M., Kawai, S-, Saisho, H., Onto, M, Saiki, R. K. & Sninsky,
J. J. (1993) Cancer Res 53, 2472-2474.

28. 29.
30.
31. 32.
33.
34.

Ling, H., Boudsocq, F.3 Woodgate, It & Yang, W. (2001) Cell 107,91-102.
Trincao, J., Johnson, K E., Escalante, C. R., Prakash, S., Prakash, L. &
Aggarwal, A. K. (2001)Mol Cells, 417-426.
Cho, Y. S., Zhu, F. C,, Luxon, B. A. & Gorenstedo, D. G. JBiomol Struct Dyn
11,685-702.
Eigen, M. (1971) Naturwissenschaften 58,465-523.
Engelke, D. R., Krikos, A,, Brack, M- E. & Ginsburg, D. (1990)
Biochem. 191,396-400.
Zhao, H., Giver, L., Shao, Z-, Afitiolter, J. A. & Arnold, F. H. (1998) Nature
Biotechnol. 16,258-61.
Mxucata, T., Iwai, S. & Ohtsuka, E. (1990) Nudeic Acids Res 18,7279-7286.

CLAIMS
1. A [method for the generation of an engineered DNA polymerase with an expanded jstrate range which comprises the step of preparing and expressing nucleic acid ej coding an engineered DNA polymerase wherein the method comprises the use of te iplate nucleic acid and primers which bear one or more distorting 3' termini.
2. A
icthod for the generation of a engineered DNA polymerase with an expanded
subsapte range which comprises the steps of:
(a) preparing nucleic acid encoding a engineered DNA polymerase, wherein me
polymerase is generated using a repertoire of nucleic acid molecules encoding
one or more DNA polymerases and primers which bear distorting 3 termini.
(b) compartmentalising the nucleic acid of step (a) into microcapsules;
(c) expressing the nucleic acid to produce their respective DNA polyraerase
within the microcapsules;
(d) sorting the nucleic acid encoding the engineered DNA polymerase which
exhibits an expanded substrate range; and
(e) expressing the engineered DNA polymerase which exhibits an expanded
substrate range.
3. A lethod according to claim 1 or claim 2 wherein the distorting 3* terminus of the prime s is effected by any one or more of the following techniques: the presence of one i ucleotide mismatch bases at the 3'end of the one or more primers (primer roism itch disortion); the presence of two nucleotide mismatch bases at the 3'end of the one < r more flanking primers (primer mismatch disortion); the presence of three nucle itide mismatch bases at the 3'end of the one or more flanking primers (primer mism itch disortion); the presence of four nucleotide mismatch bases at the 3'end of the 01 e or more flanking primers (primer mismatch disortion) and the presence of one or more unnatural bases at the 3' end of one or more flanking primers or combinations there
4. A i letnod according to claim 3 wherein the distorting 3' terminus of the primers is effect d by the presence of one or more unnatural bases at the 3' end of one or more fianki ig primers; wherein the at least one unnatural base is Snitroindole triphosphate (5NT3 >).

5. A

lethod according to claim 4 wherein the flanking primers bear one or more

nuclec tide mismatches at their 3' end.
6 A method according to aay preceding claim wherein the engineered DNA polyn arase is a pol A DNA polymerase.
7. A method according to claim 6 wherein the pol A polymerase is generated from one 01 more repertoires of randomly mutated Taq genes.
8. Anethod according to claim 6 wherein the pol A polymerase is generated from
rej ertoires generated by recombining related pol A genes.
9 A.I
repert
ethod according to claim 7 wherean the pol A polymerase is generated from ires generated by recombining one or more polA genes selected from the group
consis ing of the following: Thennus aquaticus (Taq), Thcarmus thermophilus (Tth) and Them is flavus (Tfl).
gen en thefol GA-31 GGG'
10 A method according to any of claims. 3 to 9 wherein the pol A polymerase is genera ed from repertoires of Taq genes and wherein the pol A DNA polymerase is
ed using 3' mismatch flanking primers selected from the group consisting of wing; 5'-CAG GAA ACA GCT ATG ACA AAA ATC TAG ATA ACG AGO A*G mismatcb);5'GTA AAA CGA CGG CCA GTA CCA CCG AAC TGC GA CGC CAA GCC-31 CC mismatch.

11. A 3 method according to any of claims 3 to 9 wherein the pol A polymerase is general d from repertoires of Taq genes or from repertoires of chimeras of Taq, Tta and Tfl jenes wherein the pol A DNA polymerase is generated using flanking primers with fo r mismatches at their 3' end (underlined) selected from the group consisting of the i dlowing: 5'-CAG GAA ACA GCT ATG ACA AAA GTG AAA TGA ATA GTT C JA CTTTT-3': 5'~GTA AAA CGA CGG CCA GTC TTC ACA GGT CAA GCT T IT TAA GOTGrS* ; S'-CAG GAA ACA GCT ATG ACC ATT GAT AGA GTT A T TTA CCA CAGGG-3'; 5M3TA AAA CGA CGG CCA GTC TTC ACA

GGTC

^A GCT TAT TAA GGTG-3'

ethod according to any of claims 3 to 9 wherein the pol A polymerase is d from repertoires of Taq genes or from repertoires of chimeras of Taq, Tth
12. A general
with co
and Tfl jenes wherein the pol A DNA polymerase is generated using flanking primers taining unnatural base analogues at their 3' end (X), with X for example being 5-nitroi tdole or derivatives selected from the group consisting of the following: 5'-CAG C \A ACA GCT ATG ACA AAA ATC TAG ATA ACG AGG GCA X-3', 5'-GTA / AA CGA CGG CCA GTA CCA CCG AAC TGC GGG TGA CGC CAA GCX-3 ) ot primers with containing unnatural base analogues at their 3' ends and intema y (X), with X for example being 5-nitroindole or derivatives, selected from the group c insisting of the following: (S'-CAG GAA ACA GCT ATG ACA AAA ATC TAG / TA XCG AGG GCA X-3% 5'-GTA AAA CGA CGG CCA GTA CCA CXG AAC 1 5C GGG TGA CGC CAA GCX-3').
13. A i ethod for the generation of au engineered DNA polymerase with au expanded
substra s range according to any of claims 1 to 12 which comprises the step of
prepari ,g and expressing nucleic acid encoding a blend of engineered polymerases.
14. An snguieered DNA polymerase which exhibits an expanded substrate range.
15 At engineered DNA polymeme which exhibits an expanded substrate range obtains >le using the method of any preceding claim.

16. A

sngineered DNA polymerase which exhibits an expanded substrate range which

is obta ied using the method of claim 12.
17. Ai
engines
18. A
encodu
and2r
. 19. A polyme design? and SB mu.ta.tic E520G, P481A, Q680R. D655E, R795G, K314R, G370D,
20. A p capable consists as 3B5,
21. A p capable consists as3AU
engineered DNA polymerase according to claim 15 or claim 16 which is an red pol A polymerase.
>ol A DNA polymerase with an expanded substrate range, or the nucleic acid g it, wherein the DNA polymerase is designated Ml or M4 as shown in fig 1 jpectively and depicted as SEQ No 1 and SEQ No 2 respectively.
pol A DNA polymerase with an expanded substrate range, wherein the ase exhibits at least 95% identity to one or more of the amino acid sequences ed Ml and M4 as shown in fig 1 and fig 2 respectively and depicted SEQ No 1 I No 2 respectively and which comprises any one or more of the following is: E520G, D144G, U54P, E520G, E524G, N583S, 1.1- D144G, L254P, E524G, N583S, V113I, A129V, L245R, E315K, G364D, G403R, E432D, I614M, R704W, D144G, G370D, E742G, K56E, K3T, K127R, M317I, R343G, G370D, E520G, G12A, A109T, D251E, P387L, A608V, R6I7K, T710N, E742G, A109T, D144G, V155I, P298L, G370D, I614M, E694K, E39K, R343G, G370D, E520G, T539A, M747Y, K767R, G84A, D144G, E520G, F598L, A608V, E742G, D58G, R74P, A109T, L245R, R343G, E520G, N583S, E694K, A743P.
A DNA polymerase with an expanded substrate range, in particular which is f mismatch extension, wherein the DNA polymerase comprises, preferably
3f the amino acid sequence of any one or more of the clones designated herein
B8,3C12and3Dl.
i A DNA polymerase with an expanded substrate range, in particular which is of abasic site bypass, wherein the DNA polymerase comprises, preferably of the amino acid sequence of any one or more of the clones designated herein 3B6and3BH.

22. A |>ol A DNA polymerase with an expanded substrate range, in particular which is
capable of DNA replication involving the incorporation of unatural base analogues into die newly replicated DNA, wherein the pol A DNA polymerase comprises, prefer4>ly consists of the amino acid sequence of any one or more of the clones
design
4.
23. A pol A DNA polymerase with an expanded substrate range, wherein the polymdrase exhibits at least 95% identity to one or more of the amino acid sequences designs ed 3B5,3B8,3C12, 3D1, 3A10, 3B6, 3B11,4D11 and 5D4. which comprises any oni or more of the mutations (with respect to either of the three parent genes Taq, Tth, Tt ) or gene segments found in clones 3B5, 3B8, 3C12, 3DI, 3A10, 3B6, 3B11, 4Dlla:d5D4.
24. A claimsucleic acid construct encoding an engineered poJymerase according to any of 5 to 23.25. A i exhibitsucleic acid construct encoding an engineered pol A DNA polymerase which an expanded substrate range, wherein said pol A DNA polymerase is depicted
in fig 11 nd fig 2 respectively as SEQ No I or SEQ No 2 and is designated Ml and M4
respectrely.26.Avctor comprising a nucleic acid construct according to claim 24 or claim 25.,27. The i
se of an engineered DNA polymerase according to any of claims 15 to 23 in any
one or nkore of the following applications selected from the group consisting of the followinj: PCR amplification, sequencing of damaged DNA templates, the incorporation
ofunatui
il base analogues into DNA and the creation of novel polymerase activities.
28. The v ;e of a blend of engineered polymerases according to claim 27.

29. Ti
derive*
30.TJii is Ml respect
• use according to claim 27 or claim 28 wherein the engineered polymerase is from at least Tag polymerase.
use according to claim 27 or claim 28 wherein the engineered pol A polymerase r M4 as depicted in fig 1 and fig 2 and designated SEQ No 1 and SEQ No 2 respectively.



31. The use according to claim 27 or claim 28of Ml DNA polymerase as depicted in fig signaled SEQ No 1.
32. use according to claim 27 or claim 28 of any one or more pol A polymerases t consisting of the following: 3B5, 3B8, 3C12, 3JD1, 3A10, 3B6, 3B11, 4D11