`© 1994 Elsevier Science B.V. All rights reserved. 0378-1119/94/$07.00
`
`1
`
`GENE 08146
`
`DNA polymerase fluorescent substrates with reversible 3' -tags
`
`(Nucleotide sequencing; primer extension; gel; genome; termination; modified nucleotides)
`
`Bruno Canarda and Robert S. Sarfati b
`
`• Facu/te de Medecine 2•me etage. U RA-CN RS 1462, 06107 Nice cedex 2, France; and blnstitut Pasteur, Unite de Chimie Organique. 28, Rue du Dr. Roux,
`75724 PARIS cedex 15, France. Tel. ( 33-1) 4568-8000, ext. 7272
`
`Received by A. Ullmann: 13 January 1994; Revised/Accepted: 21 February/22 February 1994; Received at publishers: 2 June 1994
`
`SUMMARY
`
`We have synthesized 3'-substituted-2'-deoxyribonucleotide-5'-triphosphates corresponding to A, T, G and C. The 3'
`position was esterified by a separate anthranylic derivative ( 3'-tag) giving specific fluorescent properties to each nucleotide
`(nt). These nt acted as substrates with several DNA polymerases leading to chain termination. Upon alkali or enzymatic
`treatment of the terminated DNA chain, free 3'-hydroxyl groups were recovered and found able to undergo chain
`extension when incubated with a mixture of dNTPs and a DNA polymerase. Because each tag has different fluorescent
`properties in itself, i.e., as a free acid, it theoretically is possible, after removal and characterization of the tag, to infer
`which nt has been inserted. Reiteration of the process can then be used to determine a nt sequence with a non-gel(cid:173)
`based method amenable to automation.
`
`INTRODUCTION
`
`DNA sequencing has revolutionized the speed and
`depth of our understanding of complex molecular biology
`processes. Presently classical sequencing techniques were
`introduced around 1977 (Sanger et al., 1977; Maxam and
`Gilbert, 1977). Dideoxy sequencing (Sanger et al., 1977)
`has gained wide acceptance and is now the method of
`choice for determining a nt sequence from a single(cid:173)
`stranded (ss) DNA template. During the four enzymatic
`chain elongations, dideoxy nt are randomly inserted in
`place of the corresponding deoxy nt. Sequencing reac(cid:173)
`tions generate a complex mixture which is subsequently
`resolved by polyacrylamide-gel electrophoresis.
`
`Correspondence to: Dr. B. Canard. Faculte de Medecine 2•me etage,
`URA-CNRS 1462, Avenue de Valombrose, 06107 Nice cedex 2, France.
`Tel. (33) 93377 632; Fax (33) 93533 071;
`e-mail: CANARD@NAXOS.UNICE.FR
`
`Abbreviations: AMY-RT, avian myeloblastosis virus reverse tran(cid:173)
`scriptase; An, anthranyloyl; bp, base pair(s); BSA, bovine serum
`albumine; kb, kilobase(s) or 1000 bp; nt, nucleotide(s); oligo, oligodeoxy(cid:173)
`ribonucleotide; RTa, reversibly tagged; ss, single strand(ed).
`
`SSDI 0378-1119(94)00357-X
`
`We wanted to reconsider the basic features of the
`dideoxy sequencing method, in particular the enzymatic
`reaction itself where most of the complexity of the process
`is generated and analysed during the following steps of
`resolution and data acquisition, i.e., gel electrophoresis
`and sequence reading, respectively. Indeed, we thought
`that we could take advantage of the molecular recogni(cid:173)
`tion performed at the polymerase active site after a single
`nt incorporation. This would considerably reduce pro(cid:173)
`duct diversity generated by an extension reaction in the
`presence of classical deoxy and dideoxy nt. Thus, a
`canonical sequencing reaction would produce a single
`adduct in a controled fashion, allowing identification of
`the added nt, before iteration of the process, leading to
`a stepwise, real-time sequence determination. Protection
`of the 3' -end of the extending DNA molecule would have
`the desired properties for such a reaction, provided that
`a free 3'-hydroxyl group could be eventually regenerated.
`As a first step toward such a method, we describe such
`deoxy nt derivatives and their use by different DNA poly(cid:173)
`merases. We show using a end-labelled primer and a gel
`assay that these nt are chain terminators which, once
`incorporated, can be converted to regular, functional
`
`Columbia Ex. 2007
`Illumina, Inc. v. The Trustees
`of Columbia University
`in the City of New York
`IPR2020-01177
`
`
`
`Gene, 148 (1994) 1-6
`© 1994 Elsevier Science B.V. All rights reserved. 0378-1119/94/$07.00
`
`1
`
`GENE 08146
`
`DNA polymerase fluorescent substrates with reversible 3' -tags
`
`(Nucleotide sequencing; primer extension; gel; genome; termination; modified nucleotides)
`
`Bruno Canarda and Robert S. Sarfati b
`
`• Facu/te de Medecine 2•me etage. U RA-CN RS 1462, 06107 Nice cedex 2, France; and blnstitut Pasteur, Unite de Chimie Organique. 28, Rue du Dr. Roux,
`75724 PARIS cedex 15, France. Tel. ( 33-1) 4568-8000, ext. 7272
`
`Received by A. Ullmann: 13 January 1994; Revised/Accepted: 21 February/22 February 1994; Received at publishers: 2 June 1994
`
`SUMMARY
`
`We have synthesized 3'-substituted-2'-deoxyribonucleotide-5'-triphosphates corresponding to A, T, G and C. The 3'
`position was esterified by a separate anthranylic derivative ( 3'-tag) giving specific fluorescent properties to each nucleotide
`(nt). These nt acted as substrates with several DNA polymerases leading to chain termination. Upon alkali or enzymatic
`treatment of the terminated DNA chain, free 3'-hydroxyl groups were recovered and found able to undergo chain
`extension when incubated with a mixture of dNTPs and a DNA polymerase. Because each tag has different fluorescent
`properties in itself, i.e., as a free acid, it theoretically is possible, after removal and characterization of the tag, to infer
`which nt has been inserted. Reiteration of the process can then be used to determine a nt sequence with a non-gel(cid:173)
`based method amenable to automation.
`
`INTRODUCTION
`
`DNA sequencing has revolutionized the speed and
`depth of our understanding of complex molecular biology
`processes. Presently classical sequencing techniques were
`introduced around 1977 (Sanger et al., 1977; Maxam and
`Gilbert, 1977). Dideoxy sequencing (Sanger et al., 1977)
`has gained wide acceptance and is now the method of
`choice for determining a nt sequence from a single(cid:173)
`stranded (ss) DNA template. During the four enzymatic
`chain elongations, dideoxy nt are randomly inserted in
`place of the corresponding deoxy nt. Sequencing reac(cid:173)
`tions generate a complex mixture which is subsequently
`resolved by polyacrylamide-gel electrophoresis.
`
`Correspondence to: Dr. B. Canard. Faculte de Medecine 2•me etage,
`URA-CNRS 1462, Avenue de Valombrose, 06107 Nice cedex 2, France.
`Tel. (33) 93377 632; Fax (33) 93533 071;
`e-mail: CANARD@NAXOS.UNICE.FR
`
`Abbreviations: AMY-RT, avian myeloblastosis virus reverse tran(cid:173)
`scriptase; An, anthranyloyl; bp, base pair(s); BSA, bovine serum
`albumine; kb, kilobase(s) or 1000 bp; nt, nucleotide(s); oligo, oligodeoxy(cid:173)
`ribonucleotide; RTa, reversibly tagged; ss, single strand(ed).
`
`SSDI 0378-1119(94)00357-X
`
`We wanted to reconsider the basic features of the
`dideoxy sequencing method, in particular the enzymatic
`reaction itself where most of the complexity of the process
`is generated and analysed during the following steps of
`resolution and data acquisition, i.e., gel electrophoresis
`and sequence reading, respectively. Indeed, we thought
`that we could take advantage of the molecular recogni(cid:173)
`tion performed at the polymerase active site after a single
`nt incorporation. This would considerably reduce pro(cid:173)
`duct diversity generated by an extension reaction in the
`presence of classical deoxy and dideoxy nt. Thus, a
`canonical sequencing reaction would produce a single
`adduct in a controled fashion, allowing identification of
`the added nt, before iteration of the process, leading to
`a stepwise, real-time sequence determination. Protection
`of the 3' -end of the extending DNA molecule would have
`the desired properties for such a reaction, provided that
`a free 3'-hydroxyl group could be eventually regenerated.
`As a first step toward such a method, we describe such
`deoxy nt derivatives and their use by different DNA poly(cid:173)
`merases. We show using a end-labelled primer and a gel
`assay that these nt are chain terminators which, once
`incorporated, can be converted to regular, functional
`
`
`
`2
`
`3'-ends. We describe the basic steps of a putative cycle
`(namely complete
`incorporation. deprotection and
`re-incorporation). as a contribution towards a new nt
`sequencing method that circumvents gel electrophoresis
`and the use of radio-isotopes.
`
`TAHU: I
`
`Spectral properties or uncuupled lluurnphllres·' and !heir c·ll1Te,pnnd(cid:173)
`ing nt
`
`;\nthranylic derivativeb
`
`Rl
`
`R2
`
`R3
`
`R4
`
`Wavelength
`(111111
`
`<'r,upbl
`l\l
`
`RESULTS AND DISCUSSION
`
`The aim of this work was to design 3'-modified
`2'-deoxynucleotides 5'-triphosphate substrates for DNA
`polymerases, such that the 3'-moiety would be different
`for each base G, A, T or C, be easily identified (e.g ..
`fluorescent), and be removed under conditions compati(cid:173)
`ble with DNA stability
`to restore an unprotected
`3'-hydroxyl end.
`
`(a) Synthesis of nt analogs
`We synthesized 3' -anthranyloyl 2' -deoxy-nucleotide-5' -
`triphosphates derivatives and evaluated them as sub(cid:173)
`strates for several DNA polymerases. For each of the four
`dATP. dGTP, dTTP and dCTP, the 3'-hydroxyl group
`was esterified by a distinct anthranylate or fluorescein(cid:173)
`based fluorescent residue using the corresponding anhy(cid:173)
`dride or isothiocyanate (Fig. 1 ). This resulted in different
`spectrofluorometric properties for each 3'-Reversibly
`Tagged-dNTP (3'-RTa-dNTP), as a consequence of the
`respective free acid counterpart (Table I) and allowed
`discrimination between them thanks to their different
`fluorimetric absorption or em1ss10n spectra. Nuclear
`magnetic resonance assignments are available upon
`request.
`
`absllrptinn
`
`emission
`
`H
`H
`H
`CH.1
`H
`
`H
`H
`H
`H
`CH,
`
`H
`396.5
`H
`dATP
`315
`416.5
`dGTP
`H
`CH.1
`11.li.'
`312
`dTTP
`403
`H
`CH.,
`1-1
`dCTP
`409
`317
`H
`dCTP
`403
`289
`H
`H
`dTTP
`494
`Fluorescein
`523
`- - - - - - - - - - - - - - - - - - - - ,_., _________ _
`a Methods: Absorption spectra were recorded at 25 Cina double-beam
`spectrophotometer in
`the presence of 50 mM Tris· HCI I pH 8.0 I.
`Fluorescence emission and excitation spectra were measured at 25 C
`in a LSS0B Perkin-Elmer fluorescence spectrophotometer. using a 2-ml
`cuvette or a 16-µl liquid chromatography flow cell frn111 Pcrkin-Flmer.
`All compounds were excited at their absorption maxima. The slit widths
`for excitation and emission were 2.5 11111_
`b See Fig. l for compound l with groups RI -R4.
`' n.d .. not determined.
`
`(b) Incorporation of 3'-RTa-dNTPs
`Since these nt analogs did not contain a 3'-hydroxyl
`group, their incorporation into an elongating DNA
`strand resulted in chain termination. This point was
`assessed using
`the solid-phase assay described
`111
`Methods in the legend to Fig. 3. A primed ss oligo ( Fig. 2,
`substrate 31-G,A,T,C) was incubated with the comple(cid:173)
`mentary 3' -RTa-dNTP adjacent to the primer and vari(cid:173)
`ous DNA polymerases, and the resulting product was
`
`0 T.T,G,C
`
`0
`0
`HO-P-0-P-O-P-O
`I
`1
`I
`HO HO HO
`
`Q
`
`0
`
`R4:2xNHR 1
`
`R3
`
`R2
`
`Compound 1
`
`Compound 3
`
`0
`0
`0
`N
`0
`..
`..
`..
`0-P-O-P-O-P-O\J
`Q
`I
`I
`L
`0
`0
`0
`
`0
`
`NHCH3
`
`NH,
`1~
`0 · c~ N, C..._ 6
`
`8
`
`i'i
`
`I~
`
`Compound 2
`
`Compound 4
`
`Fig. l. Chemical structure of 3'-RTa-2'-deoxy-nucleotides. Methods: An-dATP (Sarfati ct al.. 1990) was prepared from dATP and isatoic anhydride
`essentially by the same procedure as (Hiratsuka, 1982) for the synthesis of An-ATP. N-methyl-An-dGTP, 6-methyl-An-dTTP, 3-methyl-An-dCTP
`and 5-methyl-An-dCTP were synthesized, purified and characterised by the same procedure. 3-MethyL 5-methyl and 6-methyl-isatoic anhydrides
`were prepared by the procedure described by Erdmann ( 1899) for isatoic anhydride. Synthesis of compounds l. 2. 3 and 4 will be described elsewhere
`(R.S.S .. T. Berthod, C. Guerreiro and B.C., data not shown).
`
`
`
`5'-Bio-ATACTTTAAGGATATGTATCC
`TATGAAATTCCTATACATAGGNTTTTTTTTT
`
`31-G,A,T,C
`
`5•.32P-ATACTTTAAGGATATGTATCC
`TATGAAATTCCTATACATAGGCCCCC
`
`26-C
`
`Fig. 2. Diagram of the system used to measure enzymatic insertion of
`3' -modified nt. The synthetic primer is either biotinylated in 5' (substrate
`31-G,A,T,C) or 32 P-end-labelled (substrate 26-C) and annealed to a
`complementary template.
`
`assayed for free 3' -hydroxyl groups available for further
`extension with a chase containing a radio-labelled deoxy
`nt (Fig. 3A). Typically, this assay indicated that some
`incorporation of the 3' -modified nt might have occurred,
`as higher counts were always found in the control
`(unblocked) relative to the various enzyme/substrates
`tested. Several DNA polymerases could have extended
`the primer with a 3' -modified deoxy nt to some extent,
`the Sequenase and the M-MuLY reverse transcriptase
`being respectively the most and the least efficient under
`the experimental conditions tested here. Unmodified T7
`DNA polymerase, Taq polymerase and Kienow fragment
`of DNA polymerase I were also able to use such sub(cid:173)
`strates (data not shown).
`However, the observed blocking was not complete
`under some of these conditions (Fig. 3A), and such assays
`did not tell us what may have happened at the molecu(cid:173)
`lar level.
`Consequently, AMY-reverse transcriptase, Taq DNA
`polymerase and modified T7 DNA polymerase were
`selected for further studies using a 32P-end-labelled
`primer extension assay followed by denaturing polyacryl(cid:173)
`amide-gel electrophoresis. Fig. 3B shows the result of
`such an assay in conjunction with its corresponding solid(cid:173)
`phase assay. Modified T7 DNA polymerase was able to
`incorporate very rapidly 3'-RTa-dGTP in front of its cog(cid:173)
`nate base up to a plateau value dependent on the concen(cid:173)
`tration of the modified nt, and comparison of Fig. 3A and
`B showed that the solid-phase assay correlated well with
`the primer extension assay when relative band intensities
`were determined upon densitometric analysis (data not
`shown). The equilibrium was not exclusively dependent
`on nt concentration, since addition of more enzyme was
`able to displace this equilibrium towards further incorpo(cid:173)
`ration. The plateau value did not come from a rapid
`inactivation of the enzyme, since a chase of classical
`deoxy-nt lead to rapid extension of the remaining 21-mer
`up to 31-mer (data not shown). In Fig. 3B, concentrations
`of 3'-RTa-dGTP above 1 mM were able to displace this
`plateau value to nearly 100% incorporation in less than
`1 min. Although 3'-RTa-dNTPs did act as chain termina(cid:173)
`tors, they did not compete out significantly ddNTPs in
`classical Sanger dideoxy sequencing reactions even with
`
`3
`
`a fivefold molar excess relative to ddNTPs. For example,
`addition of 40 µM of 3'-RT-dCTP to a M13 sequencing
`ddC termination mix made of 80 µM of each dGTP,
`dATP, dCTP and dTTP, 8 µM ddCTP and 50 mM NaCl
`did not lead to significant shortening of sequencing pro(cid:173)
`ducts upon examination of polyacrylamide gel autoradio(cid:173)
`grams obtained with commercial T7 sequencing kits.
`Taken with the fact that high concentrations of 3' -RTa(cid:173)
`dNTPs were needed to reach high incorporation levels,
`this may indicate that these modified nt are not very
`efficient chain terminators, a result awaiting precise deter(cid:173)
`mination of their Km, Vmax and kcat (work in progress).
`The DNA polymerases used here had different kinetic
`behaviors ranging from very slow (e.g., AMY-RT, data
`not shown), to moderate (e.g., Taq DNA polymerase), to
`instantaneous incorporation (e.g., Sequenase) up to the
`plateau value. For example, one can compare the
`Sequenase enzyme in Fig. 3B with the Taq DNA poly(cid:173)
`merase in Fig. 5A. We do not know if this kinetic beha(cid:173)
`vior is related to the processivity of these polymerases.
`No products smaller in size than the primer were
`detected, except when DNA polymerases having a 3' to
`5' exonuclease activity were used, a finding which indi(cid:173)
`cates that such activity should not be present when one
`expects a single addition product.
`These results show that, despite a relatively bulky
`3' -group, these modified nt are still accepted by the
`enzyme. However, the nature of the 3'-substitute played
`a key role in the incorporation level, as shown in Fig. 4.
`A spacer arm was esterified in the 3' position of dTTP
`and dCTP, allowing facile' coupling of N-methyl(cid:173)
`anthranylic and fluorescein derivatives (compounds 2 and
`3), respectively (R.S.S., T. Berthod, C. Guerreiro and B.C.,
`unpublished results). This led to very slow and incom(cid:173)
`plete incorporation for the 3' -fluorescein derivative of
`dTTP, no matter which enzyme was used, but had a dra(cid:173)
`matic effect for
`the N-methyl-anthranylic derivative
`( Fig. 4A and not shown). Indeed, 500 µM of this
`3'-substituted nt were sufficient to drive the incorporation
`reaction close to completion with modified T7 DNA
`polymerase. To our surprise, two band products in equi(cid:173)
`librium with one another were obtained. The fact that
`this was again
`independent of
`the enzyme used
`(Sequenase, AMY-RT or Taq DNA polymerase) sug(cid:173)
`gested that these two products were probably conformers
`or differed only by a net electric charge under the electro(cid:173)
`phoretic conditions used here, but this awaits further
`characterization of the addition products.
`
`(c) The tags can be chemically or enzymatically removed
`Chemical or enzymatic removal of such a tag is shown
`in Fig. 5. Panel A shows incorporation of 3'-RTa-dGMP
`in front of its cognate dC base (template 26-C of Fig. 2),
`
`
`
`4
`
`A
`
`'O .,
`.,.
`u
`0 :a
`'$.
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`B
`
`0.4mM
`
`0.7mM
`
`0 1 2 3 4 5 6 1 2 3 4 5 6
`
`+1
`p
`
`•111a 1n1n• • J• wuP
`. ._1 •·••••W.
`
`lmM
`
`2mM
`
`0 1 2 3 4 6 0 1 2 3 4 6
`
`+1
`p .,,
`
`time (min)
`
`Fig. 3. Incorporation assay into DNA. IA) Incorporation of 3'-RTa-dNTPs with a solid-phase assay. ~Icthods: Approx. 2 pmol uf :'>'-biotinylatcd
`21-mer (5'-Bio-ATACTTTAAGGATATGTATCC) were bound to M-280 Dynabeads as described by the manufacturer and hybridized tu an excess
`( 50 pmol I of a complementary oligo presenting a 5' tail I Fig. 2). Annealing was for 1 h at room temperature in the presence of 1 \1 NaCl 5 111 M
`Tris·HCl pH 7.5/0.5 mM EDTA. After removal of the unbound oligo. washed beads werc suspended in 50 m\1 Tris·HCI pH 8.0>10 mM MgCl 2 ·10 mM
`dithiothreitol/100 µg BSA per ml and incubated in the presence of one 3'-RTa-dNTP at various concentrations and a DNA polymerase al 37 C. The
`reaction was terminated with 20 mM EDTA.10.01 ''.-;, Triton X-100, the beads washed and their concentration determined under the microscope with
`a hematocyter cell before being assayed for free 3'-hydroxyl group with radiolabel incorporation as follows: heads carrying the hybridized oligos were
`incubated in the same buffer as for 3'-RTa-dNTPs supplemented with a mix of dNTPs containing [-Y:-"SldATP and AMY-RT at -,,7 C. The beads
`were washed until radioactive counts reached background level in the supernatants and the concentration uf the beads was determined in an aliquot
`as above. The beads were then dispersed in scintillation counting cocktail (Aquasafe 300. Zinsser Analytic). The amount of radiolahcl ,1 as estimated
`relative to the unblocked control in pre-set channels for the corresponding isotopes in a scintillation counter. Concentration uf 3"-RTa-dGTP: ((cid:127))
`0.4 mM; ( •) 0.7 mM; (,6.) 1 mM; I' ) 2 mM. ( B) End-labelled primer extension and gel assay. Simple standing start reactions were performed exact Iv
`as described (Boosalis ct al., 1987) using the primer:template 26-C of Fig. 2 and 5 units of modified T7 DNA polymerase. Incubation \las for O min
`(lane O ). 2 min ( lane 1 ). 4 min ( lane 2 ), 6 min ( lane J ), 8 min (lane 4 ), 10 min ( lane 5 ). After IO min. 5 units were added and the incubation extended
`to 20 min (lane 6 ). P: primer ( 21-mer). The reaction products were subjected to electrophoresis through a 15" o denaturing polyacrylamidc gcL which
`was subsequently autoradiographed.
`
`+1 p
`
`0
`
`1
`
`2 3 4 5
`
`0 1 5 6
`
`+1
`p
`
`T (cid:127)
`
`Ill L --
`
`lf
`
`Fig. 4. Incorporation of compound 2 (2 mM, left panel I and compound
`3 (0.5 mM. right panel) into primer:template 31-A and Jl-G of Fig. 2.
`respectively. using 5 units of modified T7 DNA polymerase. Incubation
`times are 0, I. 2. 3. 4. 5 and 10 min for lanes 0. 1. 2. J. 4, 5 and 6.
`respectively. P. primer (21-mer). The reaction products were subjected
`to electrophoresis through a 15% denaturing polyacrylamide gel. which
`was subsequently autoradiographed.
`
`saponification with 0.1 M NaOH, neutralization,
`re-annealing and re-incorporation of ddAM P using Taq
`DNA polymerase. In both incorporations, a single-nt
`adduct is detected, at the expexted level, indicating the
`
`easy deprotection of the 3'-end of the growing DNA chain
`( lane 6 ), to give a functional 3' -hydroxyl end. Omission
`of the alkaline treatment did not allow a second primer
`extention ( lane 5 ).
`Fig. 5B shows a thin-layer chromatogram ot the time(cid:173)
`course of compound 4 reacted with proteinase K. After
`2 h, complete removal of the tag was obtained, and
`absence of a ninhydrin positive spot indicated that the
`ester linkage was indeed the cleaved bond.
`
`(d) The chemical nature of the 3' bond is important for a
`convenient re-incorporation
`Of particular importance was the nature of the chemi(cid:173)
`cal bond between the ribosyl moiety and the anthranyloyl
`substituents. Ethers or esters are both expected to restore
`a hydroxyl group upon deprotection. We reasonned that
`ether bonds would be hard to cleave under mild condi(cid:173)
`tions compatible with DNA chemical stability. whereas
`chemical deprotection using alkali has the present disad-
`
`
`
`A
`1 2 3 4 5 6
`
`B
`
`- +2
`- +l
`
`-P
`
`Fig. 5. Removal of the 3'-tag. (A) Chemical cleavage of the 3'-ester after
`incorporation of 3'-RT-dGMP using primer:template 26-C and 5 units
`of Taq DNA polymerase. Incubation time was O min (lane 1 and 4), 5
`min (lane 2 ), 10 min (lane 3 ). After I 5 min, 200 µM of ddATP was
`added. Half of the mixture was incubated for another 10 min (lane 5)
`and the other half was treated with 0.1 M NaOH for 10 min at 3TC,
`neutralized with HCl, ethanol precipitated and dried under vacuum.
`The oligos were reannealed for 2 min at 60°C and left at room temper(cid:173)
`ature for at least 30 min. Re-incorporation was with 200 µM ddATP
`and 5 units of Taq polymerase. (B) Enzymatic cleavage of the 3'-ester.
`3'-RT-TMP (5 mM) was dissolved in Tris·HCl buffer (50 mM, pH 8)
`and incubated at 50°C in the presence of 100 µg of proteinase K per ml.
`Reaction products were separated by thin-layer chromatography on
`silica gel plates using isopropanol/ammonium hydroxide/water (7:1:2)
`or dichloromethane/ammonium hydroxide/methanol ( 65:35: IO) as
`eluent and detected upon UV examination. Incubation time was 120
`min ( lane I) and O min ( lane 2 ). Arrows indicate the origin of migration.
`
`vantage of melting the primer-template duplex. However,
`we foresee several ways of remediating to this problem,
`such as 'locking' the primer covalently by means of a
`cross-linking agent, or by ligating a 5'-phosphorylated
`'hairpin-like' primer to the 3' -end of the ss template. In
`both cases, alkali denaturation followed by neutralization
`would lead to immediate intra-molecular reannealing
`compatible with cycling of the process. Hence, the ester
`bond is a very attractive candidate to fullfill appropriate
`conditions for attachment of the tag, and it is amenable
`to reaction with hydroxyl groups through several acti(cid:173)
`vated forms. Moreover, esterases of broad specificity are
`ubiquitous in nature, and we anticipated that it would
`be possible to find such an enzymatic activity in the wide
`group of serine enzymes (Fersht et al., 1985). This point
`illustrated with our results using proteinase K.
`is
`Although deprotection on a free 3' -esterified-dNMP was
`complete after 2 h of incubation, we propose that it is
`not out of reach to optimize this time of removal in the
`growing DNA chain, or to find a better adapted enzyme,
`or to esterify the 3' position with a tag carrying a spacer
`arm specifically designed to be recognized quickly and
`efficiently removed by an enzyme.
`
`5
`
`(e) Conclusions
`(J) Taken together, these results clearly indicate that
`3' -anthranyloyl-deoxy nt can be used as reversible chain(cid:173)
`terminators in a stable and reproducible fashion. Indeed,
`they act as substrates for several DNA polymerases to
`extend a DNA primer at its 3' end by only one nt. Many
`chain-terminating nt analogs are substrates for different
`DNA polymerases. Proper base pairing of the nt sub(cid:173)
`strate with its template DNA strand and formation of
`the phosphodiester bond has been shown to occur even
`with P-L-ribosides enantiomers (Van Draanen et al.,
`1992 ), indicating that binding of the sugar portion by the
`enzyme was probably not specific.
`(2) Upon chemical or enzymatic hydrolysis, the 3' sub(cid:173)
`stituent can be removed, and the 3'-hydroxyl regenerated
`with a good yield. This 3' -hydroxyl end can now be used
`as a new DNA primer. It is important to note that the
`tag released by hydrolysis is specific of the incorporated
`nt. Its identification would mean the identification of the
`nucleobase corresponding to the DNA template using
`standard rules of base pairing, and thus provide a very
`easy and rapid way for determining a nt sequence in a
`single tube or column, provided the process could be
`efficiently cycled. In this respect, each step of the process
`should be performed with a good yield and be compatible
`with the next step. A full cycle would include incorpora(cid:173)
`tion, tag removal and tag identification. Our results show
`that it is possible to reach high incorporation levels
`required to perform several cycles in a row. Modified T7
`DNA polymerase shows a versatile ability in using
`3' -modified substrates at a very high rate (Tabor and
`Richardson, 1989; this work), and it is of great interest
`that the presence of an ester at the 3' -position of a given
`dNTP does not slow down dramatically the incorpora(cid:173)
`tion reaction, making total chain termination within
`reach in less than 1 min. However, this is clearly not the
`case for all 3' -esters, as exemplified by the bulky fluores(cid:173)
`cein moiety, and this finding might be of interest in the
`fine mapping of functional domains of DNA polymerases.
`(3) Solid-phase sequencing reactions on magnetic
`beads utilize 2 pmol of immobilized template (Hultman
`et al.,1991). The quantitative release of 2 pmol of fluoro(cid:173)
`phore can be characterized with a classical, commercially
`available spectrofluorimeter and liquid chromatography
`detection cell (see Methods in the legend to Table I).
`Thus, signal-to-noise ratio should not be a problem, con(cid:173)
`sidering that detection of a single fluorescent molecule is
`under way (Davis et al., 1991). Moreover, as fluorescence
`chemistry is presently a rapidly growing field, we predict
`that the spacer arm of our 3' -derivatized nt will allow the
`coupling (through its primary amine) of powerful fluoro(cid:173)
`phores that are compatible with DNA polymerases. This
`should lead to convenient identification of tags present
`in microtiter wells.
`
`
`
`6
`
`Sequencing technology is under intense study, mainly
`due to the development of genome projects. Most
`improvements are directed to the dideoxy method at sev(cid:173)
`eral of its steps ( Prober et aL 1987: Venter et aL 1992:
`Mathies and Huang, 1992). Emerging new approaches,
`such as sequencing by hybridization (Strezoska et al..
`1991) or scanning tunneling microscopy ( Driscoll et al..
`1990) may become widespread if elimination of the gel
`electrophoresis is replaced by a cost-effective and simple
`alternative.
`We have synthesized 3'-modified-dNTPs substrates for
`DNA polymerases. They can go through three distinct
`reactions,
`(namely
`incorporation, deprotection and
`re-incorporation), that can be extended as a new non-gel(cid:173)
`based, non-radioactive method to determine ant sequence
`from a point mutation to a whole DNA fragment.
`
`ACKNOWLEDGEMENTS
`
`We would like to thank the numerous colleagues from
`the Institut Pasteur and elsewhere who spent time in con(cid:173)
`structive discussions, support, and help in some of the
`experiments: Octavian Barzu, Veronique Bourdon,
`Georges Carle, Alain Chaffotte, Stewart Cole, Bruno
`Frey, Tranh Huynh-Dinh, Abdelkader Namanc, Gregor
`Sagner, Joel Ughetto-Montrin, Agnes Ullmann and
`Isabelle Varlet. Special thanks to Frederick Hecht for
`improving the English of this manuscript. This work was
`supported by
`the Comite Consultatif pour
`les
`Applications de la Recherche de l'Institut Pasteur de
`Paris and the Groupe de Recherches et d'etudes des
`Genomes ( France).
`
`REFERENCES
`
`Boosalis. M.S.. Petruska. J. and Goodman. M.F.: DNA polymerase
`insertion fidelity: gel assay for site specific kinetics. J. Biol. Chem.
`262 ( 1987) 14689-14696.
`Davis. L Fairfield. F.R., Harger. C.A .. Jett. J.H .. Keller. R.A.. Hahn.
`
`,u11,·1. ll I
`rvlar1111. .I (
`.I.IL k.ra1n11,ki. L.i\ .. \1arrnpnc. B.I
`Ratcliff. R.L.. Shera. B.L.. Simpson. D..l. and Snpcr. S \. Rapid
`DNA sequencing based upon ,inglc molecule detection. ( ,encl. -\nal
`Tech. Appl. 8 ( 1991 i 1 :
`Driscoll. R.J .. Youngquist. M.G. and Baldeschwielcr. JD. -\tt,mic-sc·alc·
`imaging of DNA using scanning tunelling miernsw11:1. Nature .14h
`( 1990) 294 2%.
`Erdmann. E.: Constitution und Verhalten dcr ""lsatosiiurc'". Chem. Ber.
`.132 I 1899) 2159 2172.
`Fersht. r\ .. En1yme Structure and Mechanism. 2nd ed. Freeman. New
`York. NY. 1985. pp.405 4L1.
`Hiratsuka. T. · New fluorescent anologs of cAM P and cGM P available
`as substrates for cyclic nucleotide phosphodiesterase . .I. Biol. Chem.
`257 I 1982 I 13354 13358.
`Hultman. T .. Bergh. S .. :vlob. T. and Uhlcn. M.: Bidirectionnal solid(cid:173)
`phase sequencing of in vitro-amplified plasmid DNA. Biotcchniques
`10 i 1991) 84 93.
`Jencks. \V.P. I Fd.): Catalysis in Chemistry and Enzymology. Dover.
`'-Jc1\ York. NY, 1969. pp. 508 517.
`'vlathics. R . .-\. and Huang. X.C.: Capillary array clectrophore,i,: an
`approach to high-speed. high-throughput DNA sequencing. Nature
`35911992) 167 169.
`Maxam. A.M. and Gilhert. W.: A new method l'nr ,cquencing DNA.
`Proc. Natl. Acad. Sci. USA 74 11977 I 560 564.
`Prohcr. J.M .. Trainor. G.l. .. Dam. R.L Hobbs. F.W .. Rohertson. C.W ..
`Zagursky. R.J .. Cocuua. A.J .. Jensen. M.A. and Baumeister. K.: ;\
`rapid DNA
`system
`l'or
`sequencing with
`fluorescent chain(cid:173)
`terminating dideoxynucleotidcs. Science 238 ( 19871 3.16 341.
`Sanger. F .. Nicklen. S. and Coulson. A.R : DNA sequencing with chain(cid:173)
`terminating inhihitors. Proc. Natl. Acad. Sci. LS;\ 74 119771
`546.1 5467.
`Sarfati. R.S.. Kansai. V.K .. Munier. H .. Glaser. P .. (iillcs. A.'\1..
`l.ahruycre. L Mock. :\1 .. Danchin. A. and Barzu. 0.: Binding of
`3'-anthraniloyl-2'deoxy-ATP
`to calmodulin-activated adcnylatc
`cyclase from Rordetd/a pertussis and Racil/11s anthruci-' . .I. Biol.
`Chem. 265 11990 I 18902 18905.
`Stre;oska. Z.. Pauncscu. T.. Radnsalvjcvic. D .. Lahat. I.. Drmanac, R.
`and Crkvenjakov. R.: DNA sequencing hy hybridin1tion: JOO hasc,
`read h) a non-gel-based method. Proc. Natl. Acad. Sci l IS;\ 88
`11991) 10089 10093
`Tahor. S. and Richardson. C.: Effect ol' manganese ions ,rn the incorpo(cid:173)
`ration of dideoxynuclcotides by bacteriophage T7 DNA polymerase
`and t;sclwrichia coli DNA polymerase I. Proc. Natl. Acad. Sci. lJSA
`86 I 1989) 4076 4080.
`Van Draanen. N .. Tucker. C .. Boyd. L.. Trotter. W. and Reardon. J..
`il-1 -thymidine 5'-triphosphatc analogs as polymerase substrates.
`J. Biol. Chem. 267 ( 1992) 25019 25024.
`Venter. C.J .. Adams. M.D .. Martin-Gallardo. A .. McCombic. R.W. and
`Fields. C.: Genome sequence analysis: scientific ohjcctivcs and prac(cid:173)
`I I.
`tical strategies. Trends Biochem. Sci. 10 ( 1992) 8
`
`