Chemical primer extension: Efficiently determining single nucleotides in DNA

Article abstract of DOI:10.1002/anie.200501794

Rapid replication: Non-enzymatic primer extension has previously been studied in the context of prebiotic chemistry, but not for practical applications. Reactions with primers featuring a 3′-amino-2′, 3′-dideoxy-nucleotide can be rapid and selective for all four templating nucleobases (see scheme). On a chip with immobilized capture strands, 500 fmol of template suffice for single-nucleotide determinations within 2.7 h.

Full text of DOI:10.1002/anie.200501794

Communications
enzymes with proofreading capabilities (polymerases) to
meet this challenge.Entirely chemical, non-enzymatic sys-
tems employing imidazolides as leaving groups show a certain
level of spontaneous primer extension, but successful repli-
cation has remained elusive.[3]
Known non-enzymatic replication reactions with mono-
mers are slow.For ribonucleotides and 2-methylimidazolides
as activated monomers, reaction times of days are required,
even if the monomers are employed near their solubility limit
(ꢀ 50 mm)[4] under extreme salt conditions (up to 1m Mg2+).[5]
Weakly base pairing nucleotides (A or U/Tas nucleobase) are
incorporated more slowly than those that form stable base
pairs (C or G as nucleobase).[6] Even the ligation of
oligonucleotides requires the templating of G and C resi-
dues.[7] This makes non-enzymatic primer extension seem-
ingly unattractive for practical applications, such as the
genotyping of single-nucleotide polymorphisms (SNPs) by
mass spectrometry.[8] Herein we demonstrate how these
reactions can be accelerated, so that the determination of
nucleotides at selected sites of DNA within hours becomes
feasible, starting from subpicomole quantities of analytes.
We first used 2-methylimidazolides[9] (LG = MeIm) of 2’-
deoxynucleosides 1a–t as activated monomers 2a–t
(Scheme 2), which were treated with 3’-amino-terminal
primers.Amines are known to react faster than alcohols
with activated nucleotides,[10,11] so that singly extended
primers rather than the product mixtures typical for chemical
replication reactions are formed.Assays were monitored by
DNA Replication
DOI: 10.1002/anie.200501794
Chemical Primer Extension: Efficiently
Determining Single Nucleotides in DNA**
Patrizia Hagenbuch, Eric Kervio, Annette Hochgesand,
Ulrich Plutowski, and Clemens Richert*
The replication and transcription of
genetic information requires the step-
wise extension of DNA or RNA pri-
mers in sequence-specific, template-
directed reactions (Scheme 1).Tem-
plate-directed primer extension also
underlies PCR,[1] dideoxy sequencing,[2]
and attempts to recreate life from
inanimate materials.[3] Primer extension
presents a formidable chemical chal-
lenge.Four different, richly functional-
ized monomers (the mononucleotides)
have to be converted into good electro-
philes that react in sequence-selective
oligomerization reactions in aqueous
solution without polymerizing or
hydrolyzing rapidly.Nature employs
Scheme 1. Primer extension.
pyrophosphates as leaving groups and
quantitative MALDI-TOF mass spectrometry,[12] thus allow-
ing competitive reactions with all four activated nucleotides in
the same solution and selective detection of products in mass
spectra.[13] Second-order rate constants were derived from fits
to the kinetic data.[13] We tested whether non-enzymatic
primer extension was detectable with any of the four possible
bases (A, C, G, T) at the templating position of 40-mer
templates 4a–t (Scheme 3).
[*] Dr. P. Hagenbuch, Dr. E. Kervio, A. Hochgesand, U. Plutowski,
Prof. C. Richert
Institut für Organische Chemie
Universität Karlsruhe (TH)
76131 Karlsruhe (Germany)
Fax: (+49)721-608-4825
E-mail: cr@rrg.uka.de
[**] The authors thank Dr. J. Rojas Stütz for helpful discussions. This
work was supported by the DFG (grant no. RI 1063/1-3) and the
Fonds der Chemischen Industrie (project 164431).
Sequence-selective extension was observed for each of the
four reactions (Table 1).Although this established that
chemical primer extension was sufficiently selective to
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
6588
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6588 –6592

Angewandte
Chemie
association of monomers and template.Under these condi-
tions, even a strongly base pairing nucleotide at the templat-
ing position, when offered the matched monomer alone, gave
a second-order rate constant that was almost an order of
magnitude lower than that measured at higher concentration
(compare Table 1, entry 3 with Table 2, entry 1).In the
absence of competing reactions and/or changes in binding
equilibria, second-order rate constants should, of course, be
independent of concentration.
We then focused on identifying alternative leaving groups
for the activated deoxynucleotides.Ferris and co-workers had
previously shown that oligomerization of ribonucleotides on
mineral surfaces can be accelerated through judicious choice
of leaving groups,[16] but template-directed reactions have
proven to be a challenge to accelerate.[17] With EDC[18] alone,
template 4a gives non-sequence-selective extensions.
Attempts to improve extension reactions through activation
in situ with EDC in the presence of possible leaving groups or
Scheme 2. Activated monomers 2a–t.
Table 1: Results from primer extension reactions with a mixture of the
four 2-methylimidazolides 2a–t/MeIm.[a]
through addition of organic catalysts to methylimidazolides
Primer/
template
k’
Product ratio
(primer+C/T/A/G)
[19]
failed (see Supporting Information).Proflavine
increased
[hÀ1 mÀ1 [b]
]
the conversion with 2c–MeIm minimally from 16 to 21%
after 4 h, and this effect was observed only at the highest
concentration tested (0.67 mm).
3/4a
3/4c
3/4g
3/4t
46
84
38
43
3:84:5:8
4:5:8:83
84:5:10:1
6:9:74:11
Therefore, we focused on the activation of 1c in organic
solvents and the addition of the activated monomers to 3/4g
in aqueous buffer.The tetrazolide of 2c gave a slight
acceleration over 2c–MeIm, but the only substantial accel-
eration was observed after activation with HATU[20] and
HOAt, reagents originally developed for peptide coupling.
Optimized activation conditions gave an active ester in
satisfactory yield after precipitation (see Supporting Infor-
mation).Azabenzotriazolide 2c–OAt reacts four times faster
than 2c–MeIm with 3/4g (Table 2) and does not undergo
significant side reactions with the primer.Interestingly, an
even more dramatic acceleration of primer extension is
observed in an all-RNA system when switching from meth-
ylimidazolides to azabenzotriazolides, even though the nucle-
ophile is a hydroxy group in that case.[21]
[a] [2a–t]=19.3 mm; primer/template: 360 mm; pH 7.9; buffer: HEPES
(200 mm), NaCl (400 mm), MgCl2 (80 mm). [b] k’=second-order rate
constant for the formation of desired product; oligonucleotide
complexes treated as one reactant, activated nucleotide as the other.
identify a templating nucleotide in DNA, the assay was not
attractive for genotyping or analyzing epigenomic methyla-
tion patterns,[14,15] as it requires 1.8 nmol of template for a 5-
mL assay, that is, more template than is available in typical
clinical samples after PCR.We therefore decreased the
concentration of template and primer tenfold and reduced the
concentration of the activated monomers from 19.3 mm to
3.6 mm.Dilution favors hydrolysis of the activated nucleo-
tides, an unavoidable competing reaction, and disfavors the
Scheme 3. Primer-extension system employed.
Angew. Chem. Int. Ed. 2005, 44, 6588 –6592
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6589

Communications
Table 2: Effect of activation of monomers, helper, and pH shift on the rate of non-enzymatic single-
esters employed in competitive reactions
give rapid primer extension and acceptable
sequence selectivities for all four templates
4a–t (Table 2).
nucleotide-extension reactions.[a]
Oligomers
Monomer(s)/
LG/reagent[b]
pH
k’ [hÀ1 mÀ1 [c]
]
Product ratio
t1/2 primer [h]
(primer+C/T/A/G)[d]
To make the chemical primer extension
reactions more attractive for biomedical
applications, we investigated one-pot acti-
vation/primer-extension protocols.When
solutions of 1a–t (0.4m) were treated with
EDC (5 equiv)[16] and HOAt (3 equiv) at
pH 5, the solution contained > 90% of OAt
esters after 2 h, as determined by 31P NMR.
Dilution and adjustment of the pH value to
8.9 followed by addition of DNA strands led
to reactions similar to those with HATU-
activated OAt esters (Table 2), at least in
the presence of 5.Because the excess of
activation reagents allows reactivation of
nucleotides, the kinetics of these reactions
are more complex, but the one-pot proce-
dure makes assays easier to perform by
personnel not trained as synthetic chemists.
The lower reactivity and selectivity of
the reaction templated by adenine was then
addressed.We increased the concentration
of 2t–OAt fivefold to compensate for the
noncompetitive reactions
3/4g
3/4g
2c/MeIm
2c/OAt
7.9
7.9
7.9
7.9
7.9
8.9
9.5
5
19
20
112
205
381
393
–
–
–
–
–
–
–
38.5
10.0
9.5
1.7
0.9
3/4g/5
3/4g/5
3/4g/5
3/4g/5
3/4g/5
2c/MeIm
2c/OAt
2c/OAt[e]
2c/OAt[e]
2c/OAt[e]
0.5
0.5
competitive reactions
3/4a/5
3/4c/5
3/4g/5
3/4t/5
2a–t/OAt
8.9
8.9
8.9
8.9
94
259
212
217
4:64:14:18
1:1:5:93
86:1:8:5
1:1:89:9
2.0
0.7
0.9
0.9
2a–t/OAt
2a–t/OAt
2a–t/OAt
competitive reactions, one-pot activation/extension
3/4a/5
3/4c/5
3/4g/5
3/4t/5
2a–t/OAt/EDC
2a–t/OAt/EDC
2a–t/OAt/EDC
2a–t/OAt/EDC
8.9
8.9
8.9
8.9
81
177
230
134
4:75:18:3
1:3:1:95
97:1:1:1
1:1:97:1
2.4
1.1
0.8
1.4
competitive one-pot assays with dT-boosted monomer mixture
3/4a/5
3/4c/5
3/4g/5
3/4t/5
2a–t/OAt/EDC
2a–t/OAt/EDC
2a–t/OAt/EDC
2a–t/OAt/EDC
8.9
8.9
8.9
8.9
47
146
217
117
2:90:5:3
1:1:1:97
97:1:1:1
1:1:97:1
0.8
1.3
0.9
1.6
[a] Template and primer: 36 mm; activated 2: 3.6 mm; aqueous solution; buffer:
HEPES (200 mm), NaCl (400 mm), MgCl2 (80 mm). [b] LG=leaving group;
MeIm=2-methylimidazolide; OAt=oxyazabenzotriazolide, EDC=excess carbo-
diimide from one-pot activation/extension assay. [c] k’=second-order rate con-
stant; oligonucleotide complexes treated as one reactant, activated nucleotide as
the other. [d] Perfectly matched primer-extension products are highlighted in
boldface. [e] Activated monomer purified by HPLC.
Next, we studied whether an additional or helper oligo-
nucleotide 5 that binds immediately downstream from the
templating base would create a tighter binding site for the
incoming monomer (Scheme 3).The rate was accelerated
several fold for both 2c–MeIm and 2c–OAt (Table 2).For
azabenzotriazole-activated 2c–OAt, the rate enhancement
over the helper-free reaction with 2c–MeIm is 22-fold.A
further twofold acceleration resulted from purification of the
azabenzotriazolide 2c–OAt by HPLC prior to use in our
assay, an effort invested only for the kinetic study.The
azabenzotriazolide-driven primer-extension reaction benefits
further from elevated pH values.At pH 95. , the acceleration
of the helper-assisted reaction over that of the unassisted
reaction with 2c–MeIm is 79-fold (Table 2).A pH value of 8.9
gives a similar effect, but lowers the risk of losing sequence
selectivity through deprotonation of dT and dG.
We then established HATU/HOAt activation for mono-
mers 1a, 1g, and 1t (CH3CN was required as solvent for 1t as
cleaner activations were observed when starting from slur-
ries).Hydrolysis of OAt esters, despite their reactivity
towards amino and ribo primers,[21] is sufficiently slow.An
exploratory study of the hydrolysis of 2t–OAt in the assay
buffer in D2O by NMR spectroscopy shows a half-life of
16.2 h (see Supporting Information). Mixtures of all four OAt
Figure 1. MALDI-TOF mass spectra from competitive primer extension
reactions with 3/4a–t/5 and a dT-boosted mixture of 2a–t/OAt at
pH 8.9, 8 h. a) Template 4a, b) template 4c, c) template 4g, d) tem-
plate 4t. The asterisk denotes the peak for unconverted primer.
6590 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6588 –6592

Angewandte
Chemie
lower target affinity.The dT-boosted mixture gives sequence-
selective reactions for all four templates 4a–t in which the
correctly extended primer constitutes ꢀ 90% of the elonga-
tion products.Furthermore, the half-life of the reactions is
ꢁ 1.6 h and within a factor of 2 for all four reactions (Table 2).
More-refined reactivity-adjusted mixtures may make the rate
differences even smaller.Representative MALDI spectra and
kinetics are shown in Figures 1 and 2.
Figure 3. System for determining single-nucleotide identity on DNA
microarrays for on-chip MALDI-TOF mass spectrometry.[22] DNA
strands used: 7 (capture strand): 5’-taaaagataccatcaa-3’; 8 (dA tem-
plate): 5’-cagcgtgaaattagggtAagaacagaatgattgatggtatcttttaggaacctt-
taggtc-3’; 9 (dG template): 5’-cagcgtgaaattagggtGagaacagaatgattgatgg-
tatcttttaggaacctttaggtc-3’; 10 (3’-amino-primer): 5’-tcattctgttct-3’; 11
(helper): 5’-accctaatttcacgctg-3’. Lower-case letters denote
Figure 2. Kinetics of competitive primer-extension reactions with 3/4a–t/5 and
dT-boosted mixture of 2a–t/OAt at pH 8.9. a) Template 4a, b) template 4c,
deoxynucleotides, SNP sites are highlighted as upper-case letters.
~
~
*
c) template 4g, d) template 4t. 6c, 6t, + 6a, 6g.
Finally, we performed exploratory primer extension
reactions with the microarray system shown schematically in
Figure 3.It consists of capture oligonucleotide 7, immobilized
on spots of flame-smoothed gold surfaces background-
passivated by oligo(ethylene glycol ether) self-assembled
monolayers.[22] The spots were exposed to solutions contain-
ing one of two different DNA 60-mers 8 or 9 (which contain
either nucleotide A (8) or G (9) at the site of primer
extension), primer 10, and helper 11.After hybridization,
template-directed reactions were induced that employed
EDC-activated mononucleotides (3.6 mm each) in the exten-
sion buffer at pH 8.9 and 88C.After 160 min, the surface was
washed, treated with MALDI matrix solution, and subjected
to mass-spectrometric analysis.Figure 4 shows spectra
obtained in assays starting from solutions containing tem-
plates 8 or 9 (500 fmol).Even at 50 fmoles, an SNP call could
be made for the A-template (see Supporting Information).
Our optimized activation of nucleotides is a one-step
process employing commercial reagents.No significant non-
nucleoside extensions of primers have been detected.Given
the rate accelerations observed, practical applications of
chemical-primer extension, such as SNP genotyping[23] can be
envisioned.Chemical-primer extension on a chip does not
require purification after the elongation step and avoids
sample transfer, as all steps are performed on the same
Figure 4. MALDI-TOF mass spectra from primer extensions performed with the
system shown in Figure 3 and 60 mer templates 8/9 featuring the nucleotides
A (left) or G (right) at the SNP site after 160 min. Spectra were obtained
directly from the surface on which the primer-extension reaction had been
performed. Surfaces had been incubated with 500 fmol of template 8 or 9. See
Supporting Information for further details.
surface.This should lower the cost of high-throughput assays.
Heavily modified nucleotides, including those that form non-
natural base pairs, may be incorporated into primers without
the constraints imposed by active sites of polymerases.An
extension of this work to assays with fluorophore-labeled
monomers that allow optical detection on DNA microarrays
Angew. Chem. Int. Ed. 2005, 44, 6588 –6592
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6591

Communications
[9] T.Inoue, L.E.Orgel, J. Am. Chem. Soc. 1981, 103, 7666 – 7667.
[10] L.E.Orgel, R.Lohrmann, Acc. Chem. Res. 1974, 7, 368 – 377.
is being actively pursued in these laboratories.A fluorophore-
labeled form of activated 1t suitable for non-enzymatic
primer extension reactions was recently described.[24] Explor-
atory experiments with 3’-Cy3-labeled 2c–OAt, and 3/4g/5
gave 42% primer conversion after 1 h and 72% conversion
after 3.5 h,[25] demonstrating the feasibility of fluorophore-
based genotyping.
[11] A.Kanavarioti, M.W.Stronach, R.J.Ketner, T.B.Hurley,
J.
Org. Chem. 1995, 60, 632 – 637.
[12] D.Sarracino, C.Richert,
2543 – 2548.
Bioorg. Med. Chem. Lett. 1996, 6,
[13] J.A. Rojas Stütz, C. Richert, J. Am. Chem. Soc. 2001, 123,
12718 – 12719.
[14] C.Grunau, W.Hindermann, A.Rosenthal,
2000, 9, 2651 – 2663.
[15] G.Egger, G.N.Liang, A.Aparicio, P.A.Jones,
457 – 463.
Hum. Mol. Genet.
Received: May 24, 2005
Revised: July 1, 2005
Published online: September 21, 2005
Nature 2004, 429,
[16] a) K.J.Prabahar, T.D.Cole, J.P.Ferris,
J. Am. Chem. Soc. 1994,
J. Am. Chem.
Chem.
116, 10914 – 10920; b) K.J.Prabahar, J.P.Ferris,
Soc. 1997, 119, 4330 – 4337; c) W.Huang, JP. .Ferris,
Commun. 2003, 1458 – 1459.
Keywords: DNA recognition · non-enzymatic replication ·
.
oligonucleotides · primer extension · single-nucleotide
polymorphisms
[17] L.Rodriguez, L.E.Orgel, J. Mol. Evol. 1991, 32, 278 – 281.
[18] J.Sulston, R.Lohrmann, L.E.Orgel, H.T.Miles,
Proc. Natl.
Acad. Sci. USA 1968, 59, 726 – 733.
[19] S.S.Jain, F.A.Anet, C.J.Stahle, N.V.Hud,
116, 2038 – 2042; Angew. Chem. Int. Ed. 2004, 43, 2004 – 2008.
[20] L.A.Carpino, J. Am. Chem. Soc. 1993, 115, 4397 – 4398.
[1] K.B. Mullis, Angew. Chem. 1994, 106, 1271 – 1276; Angew.
Chem. Int. Ed. Engl. 1994, 33, 1209 – 1213.
[2] F.Sanger, S.Nicklen, A.R.Coulson, Proc. Natl. Acad. Sci. USA
1977, 74, 5463 – 5467.
[3] L.E. Orgel, Crit. Rev. Biochem. Mol. Biol. 2004, 39, 99 – 123.
[4] M.Hey, C.Hartel, M.W.Göbel, Helv. Chim. Acta 2003, 86, 844 –
854.
Angew. Chem. 2004,
[21] S.R.Vogel, C.Deck, C.Richert,
10.1039/B510775.
Chem. Commun. 2005, DOI:
[22] U.Plutowski, C.Richert, Angew. Chem. 2005, 117, 627 – 631;
Angew. Chem. Int. Ed. 2005, 44, 621-625.
[5] T.Wu, L.E.Orgel, J. Am. Chem. Soc. 1992, 114, 317 – 322.
[23] K.Tang, D-.J.Fu, D.Julien, A.Braun, C.R.Cantor, H.Köster,
Proc. Natl. Acad. Sci. USA 1999, 96, 10016 – 10020.
[24] N.Griesang, E.Kervio, C.Richert, Synthesis 2005, 2327 – 2334.
[25] N.Griesang, C.Richert, unpublished results.
[6] T.Wu, L.E.Orgel,
[7] A.Luther, R.Brandsch, G.von Kiedrowski, Nature 1998, 396,
245 – 248.
J. Am. Chem. Soc. 1992, 114, 7963 – 7969.
[8] J.Tost, I.G.Gut, Mass Spectrom. Rev. 2002, 21, 388 – 418.
6592 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6588 –6592