3ⴕ-O-modified nucleotides as reversible terminators
`for pyrosequencing
`
`Jian Wu*†‡, Shenglong Zhang*†‡, Qinglin Meng*†‡, Huanyan Cao*†, Zengmin Li*†, Xiaoxu Li*†, Shundi Shi*,
`Dae Hyun Kim*§, Lanrong Bi*†, Nicholas J. Turro†‡¶, and Jingyue Ju*†¶
`
`*Columbia Genome Center, Columbia University College of Physicians and Surgeons, New York, NY 10032; and Departments of †Chemical Engineering,
`‡Chemistry, and §Biomedical Engineering, Columbia University, New York, NY 10027
`
`Contributed by Nicholas J. Turro, August 10, 2007 (sent for review July 3, 2007)
`
`Pyrosequencing is a method used to sequence DNA by detecting
`the pyrophosphate (PPi) group that is generated when a nucleotide
`is incorporated into the growing DNA strand in polymerase reac-
`tion. However, this method has an inherent difficulty in accurately
`deciphering the homopolymeric regions of the DNA templates. We
`report here the development of a method to solve this problem by
`using nucleotide reversible terminators. These nucleotide ana-
`logues are modified with a reversible chemical moiety capping the
`3ⴕ-OH group to temporarily terminate the polymerase reaction. In
`this way, only one nucleotide is incorporated into the growing
`DNA strand even in homopolymeric regions. After detection of the
`PPi for sequence determination, the 3ⴕ-OH of the primer extension
`products is regenerated through different deprotection methods.
`Using an allyl or a 2-nitrobenzyl group as the reversible moiety to
`cap the 3ⴕ-OH of the four nucleotides, we have synthesized two
`sets of 3ⴕ-O-modified nucleotides, 3ⴕ-O-allyl-dNTPs and 3ⴕ-O-(2-
`nitrobenzyl)-dNTPs as reversible terminators for pyrosequencing.
`The capping moiety on the 3ⴕ-OH of the DNA extension product is
`efficiently removed after PPi detection by either a chemical method
`or photolysis. To sequence DNA, templates containing homopoly-
`meric regions are immobilized on Sepharose beads, and then
`extension–signal detection– deprotection cycles are conducted by
`using the nucleotide reversible terminators on the DNA beads to
`unambiguously decipher the sequence of DNA templates. Our
`results establish that this reversible-terminator-pyrosequencing
`approach can be potentially developed into a powerful method-
`ology to accurately determine DNA sequences.
`
`nucleotide reversible terminator 兩 sequencing by synthesis
`
`DNA sequencing is a fundamental tool for biological science.
`
`The completion of the Human Genome Project has set the
`stage for screening genetic mutations to identify disease genes
`on a genome-wide scale (1). Accurate high-throughput DNA
`sequencing methods are needed to explore the complete human
`genome sequence for applications in clinical medicine and health
`care. To overcome the limitations of the current electrophoresis-
`based sequencing technology (2–5), a variety of new DNA-
`sequencing methods have been investigated with an aim to
`eventually realize the goal of the $1,000 genome. Such ap-
`proaches include sequencing by hybridization (6), mass spec-
`trometry-based sequencing (7–9), sequence-specific detection of
`DNA using engineered nanopores (10), and sequencing by
`ligation (11). More recently, DNA sequencing by synthesis
`approaches such as pyrosequencing (12), sequencing of single
`DNA molecules (13, 14), and polymerase colonies (15) have
`been widely explored.
`Pyrosequencing is a method to sequence DNA by detecting
`the pyrophosphate (PPi) that is generated when a nucleotide is
`incorporated into the growing DNA strand in polymerase reac-
`tion (12). In this approach, each of the four nucleotides is added
`sequentially with a mixture of enzymes and substrates in addition
`to the usual polymerase reaction components. If the added
`nucleotide is complementary with the first available base on the
`template, the nucleotide will be incorporated and a PPi will be
`
`released. The PPi is used by ATP sulfurylase to convert aden-
`osine 5⬘-phosphosulfate to ATP, which provides the energy to
`the luciferase-mediated conversion of luciferin to oxyluciferin,
`which generates visible light. If the added nucleotide is not
`incorporated, no light will be produced and the nucleotide will
`simply be washed away or degraded by the enzyme apyrase.
`Pyrosequencing has been widely used in single nucleotide poly-
`morphism detection and DNA methylation analysis (16, 17).
`More recently, this method was used in picoliter-sized reactors
`to produce the sequence of the known genome of Mycoplasma
`genitalium bacteria (18). However, the pyrosequencing method
`has an inherent problem in deciphering the number of bases in
`homopolymeric regions of DNA (12). The reason is that the light
`signal intensity is not exactly proportional to the amount of PPi
`released, especially when the homopolymeric region has more
`than five bases. Previously, we have reported the development of
`a general strategy to rationally design cleavable fluorescent
`nucleotide reversible terminators (NRTs) for four-color DNA
`sequencing by synthesis (19–23). In this approach, four nucleo-
`tides (A, C, G, and T) are modified as reversible terminators by
`attaching a cleavable fluorophore to the specific location of the
`base and capping the 3⬘-OH with a small chemically reversible
`moiety so that they are still recognized by DNA polymerase as
`substrates. DNA templates consisting of homopolymer regions
`were accurately sequenced by this approach (23). A recently
`developed sequencing-by-synthesis fluorescent DNA system
`based on a similar design of the cleavable fluorescent NRTs has
`already found wide applications in genome biology (24–26).
`Based on these successful results, we reasoned that we should be
`able to solve the homopolymer sequencing problem in conven-
`tional pyrosequencing by using four nucleotide analogues whose
`3⬘-OH group is capped by a reversible moiety. We report here
`the 3⬘-O-allyl and 3⬘-O-(2-
`the design and synthesis of
`nitrobenzyl)-modified nucleotides and their successful applica-
`tion as reversible terminators for pyrosequencing to accurately
`decipher the homopolymeric regions of DNA.
`
`Results and Discussion
`Design and Synthesis of Cleavable NRTs for Pyrosequencing. During
`the polymerase extension reaction, the 3⬘-OH group of the
`primer attacks the ␣-phosphate of the incoming nucleoside
`
`Author contributions: N.J.T. and J.J. designed research; J.W., S.Z., Q.M., H.C., Z.L., X.L., and
`S.S. performed research; S.Z., Q.M., H.C., Z.L., X.L., and L.B. contributed new reagents/
`analytic tools; J.W., Z.L., X.L., S.S., D.H.K., and J.J. analyzed data; and J.W., S.Z., Q.M., Z.L.,
`D.H.K., N.J.T., and J.J. wrote the paper.
`
`The authors declare no conflict of interest.
`
`Freely available online through the PNAS open access option.
`
`Abbreviations: NRT, nucleotide reversible terminator; PPi, pyrophosphate.
`¶To whom correspondence may be addressed at: Room 405A, Russ Berrie Medical Science
`Pavilion, Columbia Genome Center, Columbia University College of Physicians and Sur-
`geons, New York, NY 10032. E-mail: ju@c2b2.columbia.edu or njt3@columbia.edu.
`
`This article contains supporting information online at www.pnas.org/cgi/content/full/
`0707495104/DC1.
`
`© 2007 by The National Academy of Sciences of the USA
`
`16462–16467 兩 PNAS 兩 October 16, 2007 兩 vol. 104 兩 no. 42
`
`www.pnas.org兾cgi兾doi兾10.1073兾pnas.0707495104
`
`Illumina Ex. 1087
`IPR Petition - USP 10,435,742
`
`

`

`SCIENCES
`
`APPLIEDBIOLOGICAL
`
`CHEMISTRY
`
`site-specific introduction of the 2-nitrobenzyl group to the
`3⬘-oxygen, with 2-nitrobenzyl bromide under basic conditions
`furnished 2-nitrobenzylated compound 9-[␤-D-5⬘-O-(tert-
`butyldimethylsilyl)-3⬘-O-(2-nitrobenzyl)-2⬘-deoxyribofurano-
`syl]-6-chloropurine 2, which was desilylated and converted to
`2-deoxyadenosine derivative 3⬘-O-(2-nitrobenzyl)-2⬘-deoxyade-
`nosine 3 in a one-pot reaction. The precursor 3 was then
`transformed to the target molecule 3⬘-O-(2-nitrobenzyl)-dATP
`4 with established triphosphorylation procedures (23, 29, 30).
`
`Polymerase Extension Using 3ⴕ-O-modified Nucleotides and Charac-
`terization by MALDI-TOF MS. 3⬘-O-modified nucleotides pose a
`great challenge for incorporation by natural polymerase, espe-
`cially when the 3⬘-O-labeling group is a bulky one (31, 32). To
`verify that the NRTs can be recognized by polymerase as
`substrates in a polymerase reaction, we performed extension
`reactions with four different primers corresponding to different
`regions of a DNA template whose next complementary base was
`either A, C, G, or T. A 9°N polymerase (exo-)A485L/Y409V,
`which has been shown previously to incorporate the 3⬘-O-
`modified nucleotides (21, 23), was used in the polymerase
`extension reaction. After the reaction, the eight different primer
`extension products [four for 3⬘-O-allyl-dNTPs and four for
`3⬘-O-(2-nitrobenzyl)-dNTPs] were analyzed by MALDI-TOF
`MS, and the results are shown in Fig. 3. Single clear mass peaks
`at 6,437, 7,702, 6,500, and 8,310 (m/z) for each primer extension
`product was produced by using 3⬘-O-allyl-dNTPs with no left-
`the 3⬘-O-(2-
`over primer peak (Fig. 3 A–D). Similarly,
`nitrobenzyl)-dNTPs also produced complete primer extension
`DNA products at 8,414, 8,390, 8,430, and 5,602 (m/z) (Fig. 3
`E–H). The small peaks at 8,279, 8,255, 8,295, and 5,467 (m/z) in
`the mass spectra for the 3⬘-O-(2-nitrobenzyl)-dNTP extension
`products correspond to the photocleavage products that were
`generated by the partial photocleavage of the DNA extension
`products induced by the nitrogen laser (337 nm) used for
`ionization of the analyte in MALDI-TOF MS. These results
`indicate that the primers were quantitatively extended by the
`3⬘-O-modified-dNTPs in polymerase reaction and that the mod-
`ified nucleotides are excellent substrates for the 9°N polymerase.
`To further verify the utility of the NRTs in determining the
`homopolymeric regions of DNA sequences, we performed a
`continuous polymerase extension reaction in solution. This
`procedure allows the isolation of the DNA product at each step
`for detailed molecular characterization by MALDI-TOF MS.
`First, a polymerase extension reaction using 3⬘-O-(2-
`nitrobenzyl)-dGTP as a reversible terminator along with a
`primer and synthetic 100-mer DNA template corresponding to
`a portion of exon 7 in the human p53 gene was performed to yield
`a single-base extension product (product 2) (Fig. 4B Left ). After
`the reaction, a small portion of the extension product was
`characterized by MALDI-TOF MS. The rest of the product was
`
`Fig. 1.
`
`Structures of NRTs 3⬘-O-allyl-dNTP and 3⬘-O-(2-nitrobenzyl)-dNTP.
`
`triphosphate to produce a DNA extension product, releasing a
`PPi molecule. Thus, when the NRTs, which have a reversible
`chemical moiety capping the 3⬘-OH group, are used to perform
`the polymerase reaction, the reaction will be temporarily ter-
`minated whenever a NRT is incorporated into the growing DNA
`strand. After the removal of the capping moiety, the polymerase
`reaction will resume. Based on this rationale, we synthesized and
`evaluated two sets of nucleotide analogues as NRTs for pyro-
`sequencing: 3⬘-O-allyl-dNTPs and 3⬘-O-(2-nitrobenzyl)-dNTPs
`(Fig. 1). The allyl group can be efficiently removed by Pd-
`catalyzed deallylation, and the removal of the 2-nitrobenzyl
`moiety is readily accomplished by laser irradiation at 355 nm.
`The design and synthesis of the 3⬘-O-allyl-dNTPs has been
`described previously (23).
`It is particularly challenging to synthesize 3⬘-O-(2-nitroben-
`zyl)-dNTPs because the nucleophilic nitrogen on the base pref-
`erentially reacts with the 2-nitrobenzyl group. Using a previously
`reported method (27) for the synthesis of a 3⬘-O-(2-nitrobenzyl)-
`dATP actually led to the final nucleotide analogue with the
`2-nitrobenzyl group attached to the 6-amino group of the purine
`base (28). We have developed a selective protection strategy for
`the synthesis of four 3⬘-O-(2-nitrobenzyl)-dNTPs, and the de-
`tailed procedures are described in supporting information (SI)
`Appendix. The synthesis of 3⬘-O-(2-nitrobenzyl)-dATP is shown
`in Fig. 2 as an example. Treatment of 9-[␤-D-5⬘-O-(tert-
`butyldimethylsilyl)-2⬘-deoxyribofuranosyl]-6-chloropurine 1, in
`which both the sugar and base were modified to allow the
`
`Synthesis of 3⬘-O-(2-nitrobenzyl)-dATP. (Step a) 2-nitrobenzyl bromide, tetrabutylammonium bromide, NaOH, in CH2Cl2 at room temperature for 1 h
`Fig. 2.
`to produce compound 2 with a 95% yield. (Step b) Tetrabutylammonium fluoride in THF at room temperature for 1 h; methanolic ammonia and dioxane at
`85–90°C for 12 h to produce compound 3 with a 56% yield. (Step c) POCl3, PO(OMe)3 at 0°C for 2 h; (Bu3NH)4P2O7, Bu3N, triethylammonium bicarbonate, and
`NH4OH at room temperature for 1.5 h to produce compound 4 with a 30% yield.
`
`Wu et al.
`
`PNAS 兩 October 16, 2007 兩 vol. 104 兩 no. 42 兩 16463
`
`

`

`Fig. 3. MALDI-TOF MS spectra of primer extension products with 3⬘-O-allyl-dNTPs (A–D) and 3⬘-O-(2-nitrobenzyl)-dNTPs (E–H). All eight 3⬘-O-modified
`nucleotides are quantitatively incorporated into the primers with high efficiency in the polymerase reaction, which indicates that the modified nucleotides are
`good substrates for the polymerase. The small peak near the 3⬘-O-(2-nitrobenzyl)-dNTP extension product corresponds to the photocleaved product generated
`during the laser desorption and ionization process used in MALDI-TOF MS.
`
`irradiated with a laser at 355 nm for 30 s to cleave the
`3⬘-O-(2-nitrobenzyl) group from the DNA to yield photocleaved
`product (product 3) (Fig. 4C Left), which was characterized by
`MALDI-TOF MS. The photocleaved DNA product (product 3)
`with a free 3⬘-OH group regenerated was then used as a primer
`for the next nucleotide extension reaction. Fig. 4 A Right–E Right
`shows the sequential mass spectrum at each step of continuous
`DNA extension reaction using 3⬘-O-(2-nitrobenzyl)-dGTP as a
`reversible terminator. The primer alone produces a peak at 6,131
`(m/z) (Fig. 4A). The mass peak at 6,594 (m/z) in Fig. 4B
`corresponds to the first extension product with a single modified
`nucleotide G incorporated in this homopolymeric region. The
`small peak at 6,459 (m/z) in Fig. 4B corresponds to the photo-
`cleavage product that was generated by the nitrogen laser (337
`nm) used for ionization of the analyte in MALDI-TOF MS. Fig.
`4C shows the photocleavage result after irradiation of the
`extension product (product 2) at 355 nm. It can be seen from the
`data that the peak at 6,594 (m/z) has completely vanished, and
`only a single peak corresponding to the DNA product (product
`3) remains at 6,459 (m/z), which indicates that the 2-nitrobenzyl
`moiety was efficiently removed to regenerate the 3⬘-OH group.
`Fig. 4D shows the MALDI-TOF MS data for the extension
`product obtained by using the photocleaved DNA product
`(compound 3) as a primer to incorporate another 3⬘-O-(2-
`nitrobenzyl)-dGTP. A dominant peak is seen at 6,922 (m/z)
`corresponding to the extension product (product 4). The small
`
`peak at 6,787 (m/z) corresponds to the photocleavage product
`that was generated by the nitrogen laser (337 nm) used for
`ionization of the analyte in MALDI-TOF MS. Upon further
`photolysis at 355 nm, the 2-nitrobenzyl moiety was removed to
`yield DNA product (product 5) at 6,787 (m/z) with a free 3⬘-OH
`group (Fig. 4E). Similar data were obtained for 3⬘-O-(2-
`nitrobenzyl)-dTTP (SI Fig. 8). The other two nucleotides,
`3⬘-O-(2-nitrobenzyl)-dATP, and 3⬘-O-(2-nitrobenzyl)-dCTP
`also were verified to be excellent reversible terminators for the
`9°N polymerase.
`
`3ⴕ-O-Modified dATP Is Not a Substrate of Luciferase. In pyrosequenc-
`ing, luciferase converts luciferin to oxyluciferin by using the energy
`provided by ATP, yielding a chemiluminescence light signal. How-
`ever, the natural nucleotide dATP also is a substrate of luciferase,
`which can produce a false positive signal to seriously interfere with
`the pyrosequencing result. To solve this problem, a sulfur-modified
`nucleotide, ␣-S-dATP, which is not a substrate for luciferase, is used
`instead of the natural dATP in conventional pyrosequencing (33).
`To our delight, the 3⬘-O-modified-dATPs [3⬘-O-allyl-dATP and
`3⬘-O-(2-nitrobenzyl)-dATP] were shown not to be substrates of
`luciferase as indicated by the data in Fig. 5. 3⬘-O-modified-dATP
`and dATP were separately added to the luciferase and luciferin
`mixtures and the corresponding light intensities were measured and
`compared. dATP (0.5 nmol) produced a light signal intensity of 80,
`whereas 0.5 nmol and 1.5 nmol of 3⬘-O-modified-dATP only led to
`
`16464 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0707495104
`
`Wu et al.
`
`

`

`SCIENCES
`
`APPLIEDBIOLOGICAL
`
`CHEMISTRY
`
`Comparison of reversible terminator-pyrosequencing using 3⬘-O-allyl-
`Fig. 6.
`dNTPs with conventional pyrosequencing using natural nucleotides. (A) The
`self-priming DNA template with stretches of homopolymeric regions (five A, two
`T, two G, and two C bases) was sequenced by using 3⬘-O-allyl-dNTPs. The ho-
`mopolymeric regions are clearly identified, with each peak corresponding to the
`identity of each base in the DNA template. (B) Pyrosequencing data using natural
`nucleotides. The homopolymeric regions produced one large peak correspond-
`ing to the stretch of T bases and three smaller peaks for stretches of A, C, and G
`bases. However, it is very difficult to decipher the exact sequence from the data.
`
`luciferase-catalyzed reaction. This hydroxyl group may interact with
`the active catalytic site of the luciferase. This interaction is inter-
`rupted when the 3⬘-OH group is modified with an allyl group or a
`2-nitrobenzyl group (or without a 3⬘-hydroxyl group as in ddATP),
`thereby preventing luciferase from using 3⬘-O-modified-dATP and
`ddATP as a substrate. Thus, 3⬘-O-modified-dATP can be directly
`used in pyrosequencing without any further modification.
`
`Pyrosequencing with 3ⴕ-O-Modified NRTs. To verify that the NRTs
`can be successfully used in pyrosequencing, we carried out a
`sequencing reaction on a self-priming DNA template, which con-
`tained multiple homopolymeric regions, immobilized on Sepharose
`beads (SI Fig. 9). The pyrosequencing reaction was initiated by
`extending the DNA template using a polymerase extension reaction
`mixture containing the NRTs. The extension of the primer by only
`the complementary NRT was confirmed by subsequent enzymatic
`cascade reactions to convert the released PPi into a light signal. For
`the 3⬘-O-allyl-dNTPs, after detection of the light signal, the DNA
`beads were immersed in a Pd deallylation solution and incubated
`for 2 min to cleave the 3⬘-O-allyl group to regenerate a free 3⬘-OH
`for further extension. In the case of 3⬘-O-(2-nitrobenzyl)-dNTPs,
`after NRT incorporation, the DNA beads were irradiated with a
`laser at 355 nm to remove the 2-nitrobenzyl group for further
`extension. After washing the beads, the next extension cycle was
`initiated. Extension–signal detection–deprotection cycles were per-
`formed multiple times to decipher unambiguously the homopoly-
`meric sequences in the DNA template.
`The pyrosequencing data generated by 3⬘-O-allyl-dNTPs are
`shown in Fig. 6A. The 11 bases in the homopolymeric regions
`(five T, two A, two C, and two G bases) are clearly identified,
`whereas the pyrosequencing data obtained by using natural
`
`The polymerase extension scheme using 3⬘-O-(2-nitrobenzyl)-dGTP (A
`Fig. 4.
`Left–E Left) and MALDI-TOF MS spectra of the two consecutive extension
`products and their photocleavage products (A Right–E Right). (A) Primer for
`the polymerase extension reaction. (B) Primer extended with 3⬘-O-(2-
`nitrobenzyl)-dGTP to yield DNA extension product 2. (C) Product 2 photo-
`cleaved to yield photocleavage product 3. (D) Product 3 extended with an-
`other 3⬘-O-(2-nitrobenzyl)-dGTP to yield product 4. (E) Product 4 photocleaved
`to yield photocleavage product 5. After 30 s of irradiation with a laser at 355
`nm, photocleavage is complete with all of the 3⬘-O-(2-nitrobenzyl)-group
`cleaved from the DNA extension products.
`
`light intensities near background level. These results confirmed that
`3⬘-O-modified-dATP is not a substrate of luciferase. Fig. 5 also
`shows that ddATP is not a substrate to luciferase. These results
`indicate that the 3⬘-OH group may play a significant role in
`
`Signal intensity of luciferase catalyzed reactions using 0.5 nmol of
`Fig. 5.
`dATP, 0.5 nmol of 3⬘-O-(2-nitrobenzyl)-dATP, 1.0 nmol of 3⬘-O-(2-nitrobenzyl)-
`dATP, 0.5 nmol of 3⬘-O-allyl-dATP, 1.5 nmol of 3⬘-O-allyl-dATP, and 1.5 nmol
`of ddATP. The results show that 3⬘-O-(2-nitrobenzyl)-dATP and 3⬘-O-allyl-
`dATP are not substrates of luciferase (NB, 2-nitrobenzyl).
`
`Wu et al.
`
`PNAS 兩 October 16, 2007 兩 vol. 104 兩 no. 42 兩 16465
`
`

`

`NRTs in extension–signal detection–deprotection cycles is that
`higher efficiency can be achieved with multiple extensions or
`deprotections without any dephasing in the sequence determina-
`tion or a reduction in the sequencing accuracy. Therefore, in
`principle, one can achieve ⬎99% efficiency in each cycle to reach
`read lengths of at least several hundred. The signal reduction in our
`preliminary pyrosequencing data generated with the NRTs is
`mainly due to the loss of DNA beads during each washing step
`because the reaction was performed manually. Therefore, longer
`read lengths can be achieved when using single DNA-bead exten-
`sion and automated washing systems, such as the 454 genome
`sequencer (18). It is well established that PCR templates can be
`generated on millions of beads through emulsion PCR (18, 34).
`Thus, future implementation of the reversible-terminator pyrose-
`quencing on a high-density bead array platform will provide a
`high-throughput and accurate DNA sequencing system with wide
`applications in genome biology and biomedical research.
`
`Materials and Methods
`Synthesis of 3ⴕ-O-Allyl-dNTPs and 3ⴕ-O-(2-Nitrobenzyl)-dNTPs. 3⬘-O-
`allyl-dNTPs were synthesized according to the literature (23), and
`the synthesis of 3⬘-O-(2-nitrobenzyl)-dNTP is described in SI
`Appendix. An enzymatic method was used to yield ultrapure
`3⬘-O-modified nucleotide analogues based on the literature (35)
`(also see SI Appendix).
`
`Incorporation of 3ⴕ-O-Modified NRTs in Solution and Characterization
`by MALDI-TOF MS. Each polymerase extension reaction solution
`consists of 40 pmol of templates, 40 pmol of primers (the template
`and primer sequences are described in SI Table 1), 100 pmol of
`NRTs, 2 ␮l of 10⫻ Thermopol II reaction buffer (New England
`Biolabs, Ipswich, MA), 2 ␮l of 20 mM MnCl2, and 2 ␮l (4 units) of
`9°N polymerase (exo-)A485L/Y409V in a total volume of 20 ␮l.
`After an initial incubation at 95°C for 5 min and 4°C for 5 min, the
`reaction was performed at 95°C for 15 seconds, 55°C for 15 seconds,
`and 65°C for 1 min for 20 cycles. The resulting DNA products were
`purified for MALDI-TOF MS analysis by using a previously
`reported procedure (23). We also characterized 3⬘-O-(2-
`nitrobenzyl)-dGTP by performing a continuous DNA extension
`reaction using a primer (5⬘-GTTGATGTACACATTGTCAA-3⬘)
`and a synthetic DNA template (SI Table 1). The detailed procedure
`is described in SI Appendix. The other 3⬘-O-(2-nitrobenzyl)-dNTPs
`were similarly characterized.
`
`Pyrosequencing Using the NRTs. Each extension reaction consisted of
`Sepharose bead-immobilized DNA (the procedure to prepare the
`DNA beads is described in SI Appendix), 200 pmol of NRTs, 1.2 ␮l
`of 50 mM MnCl2, 1 ␮l (2 units) of 9°N polymerase (exo-)A485L/
`Y409V, and 20 ␮l of annealing buffer (20 mM Tris-acetate/5 mM
`magnesium acetate, pH 7.6). Extension was conducted in a thermal
`cycler and incubated at 65°C for 20 min with occasional stirring to
`prevent the beads from settling. After the polymerase reaction, the
`beads were pelleted by centrifugation for 20 s, and the supernatant
`was carefully removed. The beads were washed with 30 ␮l of
`annealing buffer, and the PPi of the combined supernatant was
`detected on a 96PSQ Pyrosequencer (Biotage, Uppsala, Sweden)
`for sequence determination (33). After detection of the signal, the
`beads were washed three times with 180 ␮l of deionized water. For
`3⬘-O-allyl-dNTP extensions, deallylation was conducted under
`aqueous-Pd-catalyzed conditions (23). After deallylation, the beads
`were washed three times with 180 ␮l of 1 M Tris-acetate buffer (pH
`7.7) and three times with 180 ␮l of annealing buffer and the next
`extension–signal detection–deprotection cycle was initiated. For
`3⬘-O-(2-nitrobenzyl)-dNTP extensions, extended DNA beads were
`suspended in 1 ml of annealing buffer in a cuvette with stirring and
`irradiated with a laser at 355 nm (3 W/cm2) for 1 min. After
`photocleavage, the beads were washed two times with annealing
`
`Comparison of reversible terminator-pyrosequencing using 3⬘-O-(2-
`Fig. 7.
`nitrobenzyl)-dNTPs with conventional pyrosequencing using natural nucleo-
`tides (NB, 2-nitrobenzyl). (A) The self-priming DNA template with stretches of
`homopolymeric regions was sequenced by using 3⬘-O-(2-nitrobenzyl)-dNTPs.
`The homopolymeric regions are clearly identified, with each peak correspond-
`ing to the identity of each base in the DNA template. (B) Pyrosequencing data
`using natural nucleotides. The homopolymeric regions produced two large
`peaks corresponding to the stretches of G and A bases and five smaller peaks
`corresponding to stretches of T, G, C, A, and G bases. However, it is very
`difficult to decipher the exact sequence from the data.
`
`nucleotides show a single large peak corresponding to a stretch
`of Ts and three smaller peaks corresponding to stretches of A,
`C, and G bases (Fig. 6B). However, it is very difficult to identify
`the exact sequence from this conventional pyrosequencing data.
`The pyrosequencing results using 3⬘-O-(2-nitrobenzyl)-dNTPs
`are shown in Fig. 7A. Twenty-one bases in the homopolymeric
`regions (five G, five A, three T, two G, two C, two A, and two G
`bases) are clearly identified, whereas the pyrosequencing data
`obtained by using natural nucleotides shows two large peaks
`corresponding to stretches of G and A bases and five smaller peaks
`corresponding to stretches of T, G, C, A and G bases (Fig. 7B),
`leading to ambiguity to identify the sequence. To further verify the
`utility of the reversible terminator-pyrosequencing method, we
`used 3⬘-O-(2-nitrobenzyl)-dNTPs to sequence a PCR DNA tem-
`plate produced by amplification on Sepharose beads to unambig-
`uously decipher 11 bases in the DNA templates containing ho-
`mopolymeric sequences (SI Appendix and SI Fig. 10).
`
`Conclusion
`We have developed two sets of NRTs, 3⬘-O-allyl-dNTP and 3⬘-O-
`(2-nitrobenzyl)-dNTP, for pyrosequencing, which are able to ac-
`curately decipher the homopolymeric sequences in DNA templates.
`The reversible terminators were efficiently incorporated, and they
`terminated the polymerase reactions, and the released PPi for each
`extension was detected with a standard luciferase assay. We have
`generated preliminary feasibility sequencing data of 11 bases with
`3⬘-O-allyl-dNTPs and 21 bases with 3⬘-O-(2-nitrobenzyl)-dNTPs on
`DNA templates consisting of multiple homopolymer regions.
`Longer read length should be possible with further optimization in
`nucleotide incorporation efficiency and deprotection efficiency
`coupled with automation. Also, other alternative reversible chem-
`ical groups can be explored for further optimization of the NRTs
`for pyrosequencing. In addition to solving the homopolymer issues
`in conventional pyrosequencing, the other advantage of using the
`
`16466 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0707495104
`
`Wu et al.
`
`

`

`SCIENCES
`
`APPLIEDBIOLOGICAL
`
`CHEMISTRY
`
`buffer for the continuation of the subsequent extension reactions.
`Conventional pyrosequencing data shown in Figs. 6B and 7B were
`generated by using the same instrument in parallel to compare the
`data with those of pyrosequencing by using the NRTs.
`
`We thank Dr. Steffen Jockusch, Dr. James J. Russo, and Mr. Liyong Deng
`for generous discussions and technical support. This work was supported by
`National Institutes of Health Grants P50 HG002806, R01 HG003582, and
`R21HG004404 and by the Packard Fellowship for Science and Engineering.
`
`1. Collins FS, Green ED, Guttmacher AE, Guyer MS (2003) Nature 422:835–847.
`2. Sanger F, Nicklen S, Coulson AR (1977) Proc Natl Acad Sci USA 74:5463–5467.
`3. Maxam AM, Gilbert W (1977) Proc Natl Acad Sci USA 74:560–564.
`4. Smith LM, Sanders JZ, Kaiser RJ, Hughes P, Dodd C, Connell CR, Heiner C,
`Kent SB, Hood LE (1986) Nature 321:674–679.
`5. Ju J, Ruan C, Fuller CW, Glazer AN, Mathies RA (1995) Proc Natl Acad Sci
`USA 92:4347–4351.
`6. Drmanac S, Kita D, Labat I, Hauser B, Schmidt C, Burczak JD, Drmanac R
`(1998) Nat Biotechnol 16:54–58.
`7. Fu DJ, Tang K, Braun A, Reuter D, Darnhofer-Demar B, Little DP, O’Donnell
`MJ, Cantor CR, Koster H (1998) Nat Biotechnol 16:381–384.
`8. Roskey MT, Juhasz P, Smirnov IP, Takach EJ, Martin SA, Haff LA (1996) Proc
`Natl Acad Sci USA 93:4724–4729.
`9. Edwards JR, Itagaki Y, Ju J (2001) Nucleic Acids Res 29:E104.
`10. Kasianowicz JJ, Brandin E, Branton D, Deamer DW (1996) Proc Natl Acad Sci
`USA 93:13770–13773.
`11. Shendure J, Porreca GJ, Reppas NB, Lin X, McCutcheon JP, Rosenbaum AM,
`Wang MD, Zhang K, Mitra RD, Church GM (2005) Science 309:1728–1732.
`12. Ronaghi M, Uhlen M, Nyre´n P (1998) Science 281:363–365.
`13. Braslavsky I, Hebert B, Kartalov E, Quake SR (2003) Proc Natl Acad Sci USA
`100:3960–3964.
`14. Greenleaf WJ, Block SM (2006) Science 313:801.
`15. Mitra RD, Shendure J, Olejnik J, Olejnik EK, Church GM (2003) Anal
`Biochem 320:55–65.
`16. Nordstrom T, Ronaghi M, Forsberg L, de Faire U, Morgenstern R, Nyren P
`(2000) Biotechnol Appl Biochem 31:107–112.
`17. Tost J, Dunker J, Glynne GY (2003) BioTechniques 35:152–156.
`18. Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, Berka
`J, Braverman MS, Chen Y-J, Chen Z, et al. (2005) Nature 437:376–380.
`
`19. Ju J, Li Z, Edwards J, Itagaki Y (2003) US Patent 6,664,079.
`20. Li Z, Bai X, Ruparel H, Kim S, Turro NJ, Ju J (2003) Proc Natl Acad Sci USA
`100:414–419.
`21. Ruparel H, Bi L, Li Z, Bai X, Kim DH, Turro NJ, Ju J (2005) Proc Natl Acad
`Sci USA 102:5932–5937.
`22. Seo TS, Bai X, Kim DH, Meng Q, Shi S, Ruparel H, Li Z, Turro NJ, Ju J (2005)
`Proc Natl Acad Sci USA 102:5926–5931.
`23. Ju J, Kim DH, Bi L, Meng Q, Bai X, Li Z, Li X, Marma MS, Shi S, Wu J, et
`al. (2006) Proc Natl Acad Sci USA 103:19635–19640.
`24. Mikkelsen TS, Ku M, Jaffe DB, Issac B, Lieberman E, Giannoukos G, Alvarez
`P, Brockman W, Kim TK, Koche RP, et al. (2007) Nature 448:553–560.
`25. Johnson DS, Mortazavi A, Myers RM, Wold B (2007) Science 316:1497–1502.
`26. Barski A, Cuddapah S, Cui K, Roh TY, Schones DE, Wang Z, Wei G, Chepelev
`I, Zhao K (2007) Cell 129:823–837.
`27. Metzker ML, Raghavachari R, Richards S, Jacutin SE, Civitello A, Burgess K,
`Gibbs RA (1994) Nucleic Acids Res 22:4259–4267.
`28. Welch MB, Martinez CI, Zhang AZ, Jin S, Gibbs R, Burgess K (2005) Chem
`Eur J 11:7145.
`29. Kawate T, Allerson CR, Wolfe JL (2005) Org Lett 7:3865–3868.
`30. Ludwig J, Eckstein F (1989) J Org Chem 54:631–635.
`31. Welch MB, Martinez CI, Zhang AJ, Jin S, Gibbs R, Burgess K (1999) Chem
`Eur J 5:951–960.
`32. Lu G, Burgess K (2006) Bioorg Med Chem Lett 16:3902–3905.
`33. Ronaghi M, Karamohamed S, Pettersson B, Uhlen M, Nyren P (1996) Anal
`Biochem 242:84–89.
`34. Kim JB, Porreca GJ, Song L, Greenway SC, Gorham JM, Church GM, Seidman
`CE, Seidman JG (2007) Science 316:1481–1484.
`35. Metzker ML, Raghavachari R, Burgess K, Gibbs RA (1998) BioTechniques
`25:814–817.
`
`Wu et al.
`
`PNAS 兩 October 16, 2007 兩 vol. 104 兩 no. 42 兩 16467
`
`