Novel mass tags for single nucleotide polymorphism detection

Article abstract of DOI:10.1021/ac071291y

A new method suitable for single nucleotide polymorphism detection and other applications based on oligonucleotide probe extension has been developed. The method is based on mass spectrometry and utilizes a single surface for affinity purification of exte

Full text of DOI:10.1021/ac071291y

Anal. Chem. 2008, 80, 2342-2350
Articles
Novel Mass Tags for Single Nucleotide
Polymorphism Detection
Klara R. Birikh,† Vladimir A. Korshun,† Pablo L Bernad,‡ Andrei D. Malakhov,† Natalie Milner,‡
Safraz Khan,‡ Edwin M. Southern,§ and Mikhail S. Shchepinov*,‡
Shemyakin-Ovchinikov Institute of Bioorganic Chemistry, RAS, Miklukho-Maklaya 16/10, 117997, Moscow, Russia, Tridend
Technologies, a division of Oxford Gene Technology Ltd., Sandy Lane, Yarnton, Oxford OX5 1PF, UK, and Department of
Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, UK
ing, as opposed to fluorescent labels, which spectra are rather
broad. This is especially important in the field of SNP genotyp-
ing.6,7
A new method suitable for single nucleotide polymor-
phism detection and other applications based on oligo-
nucleotide probe extension has been developed. The
method is based on mass spectrometry and utilizes a
single surface for affinity purification of extended probes
and matrix-independent desorption/ionization of the cleav-
able labels. A new family of sulfur-linked laser-cleavable
trityl labels with vastly improved flying abilities is imple-
mented in this study. Corresponding reagents compatible
with automated oligonucleotide synthesis are presented.
Utility of this method for SNP genotyping is demonstrated.
Trityl derivatives easily ionize, desorb, and can be efficiently
detected by MS techniques (in other words, “fly” well) due to
the stability of corresponding carbocations. Importantly, ionization
is a result of a laser irradiation and does not require the assistance
of matrix, although the presence of the latter does not adversely
affect the flying. This has been put to a practical use by employing
trityl mass tags (MT) as labels in combinatorial chemistry8 in
combination with (MA)LDI-TOF detection. That study used mass
labels comprising the standard dimetoxytrityl protecting group
with variable additional masses attached. In the present study,
we demonstrate the use of a much improved version of the trityl
mass tags for DNA labeling and SNP genotyping.
As opposed to the previously described method for SNP
genotyping with cleavable mass labels,5 the present protocol omits
minicolumn purification of the reaction mixtures and a separate
step of label cleavage by irradiation. Superb matrix-independent
flying abilities of the trityl-based labels made it possible to use
the surface of the mass spectrometry target as an affinity substrate
in the purification protocol followed by label cleavage by laser
upon desorption/ionization, altogether largely reducing sample
handling.
Mass spectrometry-based methods in molecular biology attract
substantial interest and represent a very fast developing field.
Admittedly, so far biological applications of mass spectrometry
have been mostly limited to proteomic studies,1 although there
are some examples of its penetration into the molecular biology
of DNA.2,3
Two different principles have been used for single nucleotide
polymorphism (SNP) detection by mass spectrometrysone was
based on the detection of the oligonucleotide probes themselves,4
and the other relied on detection of cleavable labels.5 The
disadvantage of flying the entire oligonucleotides is that they
have a much broader spectrum due to a large number of
isotopic variants, and a lower sensitivity, because oligonucleotides
do not desorb very well. On the other hand, this method does
not require any labels and gives a more direct indication of the
SNP nucleotide. Mass labels attract the attention of molecular
biologists due to their extremely sharp spectra, which allows one
to simultaneously detect labels that are only a few daltons apart
in mass. This opens a prospective of a large degree of multiplex-
EXPERIMENTAL SECTION
Procedures for Model Compound 11 and Modifying
Phosphoramidites 7 and 15. 5′-O-{4,4′-Dimethoxy-4′′-[2-
(N-succinimidyloxycarbonyl)ethyl]trityl}-3′-O-(diisopropy-
lamino-2-cyanethoxyphosphinyl)thymidine (7). Thymidine 5
(727 mg, 3.0 mmol) was coevaporated with dry pyridine (2 × 20
mL), dissolved in pyridine (30 mL), half-evaporated, and cooled
on an ice bath. 3-{4-[Chlorobis(4-methoxyphenyl)methyl]phenyl}
propionic acid, N-oxysuccinimide ester (6; 1.42 g, 2.8 mmol) was
added in one portion with stirring, and the stirring was continued
for 4 h. The mixture was warmed to room temperature, diluted
with CHCl3 (150 mL), washed with water (2 × 100 mL), dried
over Na2SO4, and evaporated, and the residue was purified by
column chromatography (elution with 20-35% EtOAc in toluene,
then 15-30% acetone in EtOAc-toluene (1:2), and, finally, with
* Corresponding author. E-mail: misha@tridend.com. Phone: +44-7766058079.
Fax: +7-(495)-330-67-38.
† Shemyakin-Ovchinikov Institute of Bioorganic Chemistry.
‡ Tridend Technologies.
§ University of Oxford.
(1) Lane, C. S. Cell Mol. Life Sci. 2005, 62, 848-69.
(2) Hofstadler, S. A.; Sannes-Lowery, K. A.; Hannis, J. C. Mass Spectrom. Rev.
2005, 24, 265-85.
(3) Tost, J.; Gut, I. G. J. Mass Spectrom. 2006, 41, 981-95.
(4) Storm, N.; Darnhofer-Patel, B.; van den Boom, D.; Rodi, C. P. Methods Mol.
Biol. 2003, 212, 241-62.
(6) Marnellos, G. Curr. Opin. Drug Discovery Dev. 2003, 6, 317-21.
(7) Shi, M. M. Am J. Pharmacogenomics 2002, 2, 197-205.
(8) Shchepinov, M. S.; Chalk, R.; Southern, E. M. Tetrahedron 2000, 56, 2713-
2724.
(5) Kokoris, M.; Dix, K.; Moynihan, K.; Mathis, J.; Erwin, B.; Grass, P.; Hines,
B.; Duesterhoeft, A. Mol. Diagn. 2000, 5, 329-40.
2342 Analytical Chemistry, Vol. 80, No. 7, April 1, 2008
10.1021/ac071291y CCC: $40.75 © 2008 American Chemical Society
Published on Web 10/31/2007

acetone-toluene (1:1)) to give 5′-O-{4,4′-dimethoxy-4′′-[2-(N-
succinimidyloxycarbonyl)ethyl]trityl}thymidine. Yield 1.35 g (68%),
white amorphous solid. NMR (DMSO-d6): 11.3 (s, 1H, NH); 7.50
(s, 1H, H-6); 7.32-7.20 (m, 8H, ArH); 6.88 (m, 4H, ArH); 6.20
(apparent t, 1H, J ) 6.7 Hz, H-1′); 5.30 (d, 1H, J ) 4.4 Hz, OH);
4.31 (m, 1H, H-3′); 3.88 (m, 1H, H-4′); 3.74 (s, 6H, OCH3);
3.25-3.13 (m, 2H, H-5′); 3.03-2.89 (m, 4H, ArCH2CH2); 2.81 (s,
4H, COCH2CH2CO); 2.3-2.1 (m, 2H, H-2′); 1.45 (s, 3H, CCH3).
To a solution of 5′-O-{4,4′-dimethoxy-4′′-[2-(N-succinimidyloxycar-
bonyl)ethyl]trityl}thymidine (250 mg, 0.35 mmol) in dry DCM
(10 mL) and DIEA (85 µL, 0.49 mmol), 2-cyanoethyl diisopropy-
lchlorophosphoramidite (54 µL, 0.49 mmol) was added dropwise.
After 2 h, the reaction mixture was diluted with CHCl3 (50 mL),
washed with washed with water (2 × 50 mL), dried over Na2SO4,
and evaporated, and the residue was dissolved in DCM (2 mL)
and precipitated in hexane (200 mL). The solid was filtered off
and dried in vacuo. Yield 310 mg (97%), white amorphous solid.
NMR (DMSO-d6): 8.96 (s, 1H, NH); 7.47 (m, 1H, H-6); 7.39 (m,
2H, ArH); 7.33 (m, 4H, ArH); 7.24 (m, 2H, ArH); 6.89 (m, 4H,
ArH); 6.26 (m, 1H, J ) 6.7 Hz, H-1′); 4.64 (m, 1H, H-3′); 4.12 (m,
0.5H, H-4′); 4.07 (m, 0.5H, H-4′); 3.82-3.50 (m, 10H, OCH3, NCH,
POCH2); 3.39-3.24 (m, 2H, H-5′); 3.05-2.90 (m, 4H, ArCH2CH2);
2.78 (s, 4H, COCH2CH2CO); 2.66 (m, 1H, CH2CN); 2.54 (m, 1H,
CH2CN); 2.48-2.33 (m, 2H, H-2′); 1.54-0.88 (m, 12H, C(CH3)2).
31P NMR (MeCN-d3): 149.31, 148.28.
Model Compound 11. 4-(N-Succinimidyloxycarbonyl)-4′-
methoxytritanol (447 mg, 1.0 mmol) in DCM (20 mL) was treated
with butylamine (1.1 mmol); the solution was stirred for 1 h, then
washed with 5% citric acid (20 mL), dried over Na2SO4, and
evaporated to dryness. The residue was dissolved in 10% trifluo-
roacetic acid in DCM (5 mL), and 6-hydroxyhexanethiol (1.0
mmol) was added. The mixture was stirred at room temperature
overnight, diluted with DCM (100 mL), washed with 5% NaHCO3
(2 × 50 mL), dried over Na2SO4, and evaporated to dryness. The
residue was dissolved in dry pyridine (5 mL) and 4-(pentylami-
nocarbonyl)-4′-methoxytrityl chloride (1 mmol) (the amide was
prepared from 4-(N-succinimidyloxycarbonyl)-4′-methoxytritanol
and pentylamine and then converted to trityl chloride with AcCl)
was added, and the mixture was stirred overnight. The mixture
was poured into water (100 mL) and extracted with EtOAc (50
mL). The organic layer was washed with water (2 × 40 mL) and
5% citric acid (50 mL), dried over Na2SO4, and evaporated and
coevapotated twice with toluene (30 mL). The resulting crude 11
was not purified. An analytical sample of 11 was isolated using
preparative TLC (3% of EtOAc in toluene).
1-S-{4,4′,4′′-Trimethoxy-3-[5-(N-succinimidyloxycarbo-
nyl)pentyl]trityl}-4-O-(diisopropylamino-2-cyanethoxyphos-
phinyl)-4-hydroxypentanethiol (15). To a solution of 3-[5-(N-
succinimidyloxycarbonyl)pentyl]-4,4′,4′′-trimethoxytritanol (13) (1.50
g, 2.68 mmol) in AcOH (15 mL), 4-hydroxypenthanethiol (355 mg,
2.95 mmol) was added and the mixture was stirred for 2 h. The
mixture was poured in water (300 mL) and extracted with EtOAc
(150 mL). The organic layer was washed with water (2 × 100
mL), dried over Na2SO4, and evaporated to dryness. The residue
was purified by column chromatography (30% EtOAc in toluene)
to give the intermediate 1-S-{4,4′,4′′-trimethoxy-3-[5-(N-succinim-
idyloxycarbonyl)pentyl]trityl}-4-hydroxypentanethiol. Yield 1.52 g
(85%), colorless oil. 1H NMR (DMSO-d6): 7.19 (d, 4H, J ) 8.8 Hz,
Figure 1. Schematic representation of the SNP detection method.
Mass tags are depicted as circles; the disulfide group is indicated as
SS.
H-2′,6′,2′′,6′′); 7.07 (d, 1H, 4J ) 2.3 Hz, H-2); 7.02 (dd, 1H, J ) 8.5
4
Hz, J ) 2.6 Hz, H-6); 6.85 (m, 5H, H-5,3′,5′,3′′,5′′); 4.25 (d, 1H, J
) 4.8 Hz, OH); 3.76 (s, 3H, OCH3); 3.73 (s, 6H, OCH3); 3.43 (m,
1H, CHOH); 2.80 (s, 4H, COCH2CH2CO); 2.60 (t, 2H, J ) 7.3 Hz,
COCH2); 2.47 (t, 2H, J ) 7.5 Hz, ArCH2); 2.08 (m, 2H, SCH2);
1.60 (m, 2H), 1.47 (m, 2H), 1.38 (m, 1H), 1.30 (m, 3H), 1.23 (m,
2H) (COCH2CH2CH2CH2, SCH2CH2CH2); 0.94 (d, 3H, J ) 6.2 Hz,
CHCH3). This was phosphitylated similar to the nucleoside
derivative to give the desired amidite 15 in 86% yield, colorless
1
oil. H NMR (DMSO-d6): 7.29 (m, 4H, H-2′,6′,2′′,6′′); 7.16-7.11
(m, 2H, H-2,6); 6.87-6.81 (m, 5H, H-5,3′,5′,3′′,5′′); 3.87-3.53 (m,
14H, OCH3, POCH, POCH2, NCH); 2.78 (s, 4H, COCH2CH2CO);
2.64-2.52 (m, 6H, ArCH2, COCH2, CH2CN); 2.16 (m, 2H, SCH2);
1.69(m,2H),1.56-1.32(m,8H)(COCH2CH2CH2CH2,SCH2CH2CH2);
1.19-1.10 (m, 15H, CHCH3). 31P NMR (MeCN-d3): 147.95, 147.08.
Oligonucleotide Synthesis. Oligonucleotides were made in
an Applied Biosystems 392 DNA/RNA synthesizer using standard
phosphoramidite chemistry.
Analytical Chemistry, Vol. 80, No. 7, April 1, 2008 2343

Figure 2. Trityls as mass tags for nucleic acids. (A) Trityl cations derived from methoxy/methyl-substituted tritanols and mass spectrum
of an equimolar mixture of the tritanols; (B) synthesis of phosphoramidite of dimethoxytrityl MT and its use for oligonucleotide labeling and
detection.
A disulfide group was introduced into oligonucleotides using
dithiol phosphoramidite from Glen Research (Catalog No. 10-1937-
xx). Branching phosphoramidite reagent (“trebler”) was pur-
chased from Glen Research (Catalog No. 10-1922-xx). Incorpora-
tion of these phosphoramidites was carried out under conditions
recommended by the manufacturer.
Incorporation of the modifying phosphoramidites 7 and 15 at
the 5′-end of the oligonucleotides was carried out through a
2344 Analytical Chemistry, Vol. 80, No. 7, April 1, 2008

Figure 3. Attachment of trityl labels through a sulfide bond. (A) Model ditrityl compound (11) and its LDI mass spectrum (the corresponding
trityl cations and their masses are depicted for convenience in mass spectrum interpretation; (B) synthesis of trimethoxytrityl MT phosphoramidite.
standard synthetic cycle of automated oligonucleotide synthesis,
apart from the oxidation step, in which the concentration of Iodine
was lowered 5-fold as compared to the standard conditions.
PCR and Extension Reactions. These were carried out using
TopoTaq polymerase (Fidelity Systems) under the conditions
recommended by the manufacturer. The template PCR was carried
out in 20-µL reaction mixture with 10 ng of genomic DNA, with
primers US and DS-tail (0.1 mM each); the thermo cycle was as
follows: 45 s 94 °C, 1 min 60 °C, and 30 s 72 °C (30-40 cycles).
For the following extension reaction with probe A/C and primer
Tail-SS, the template PCR reaction mixture comprised 1/10 of the
extension reaction volume; thermo cycle for the extension reaction
was 30 s 90 °C, 1 min 55 °C, and 30 s 72 °C (6 cycles).
Thermocycling was carried out using PCR amplificator Master-
cycler (Eppendorf).
Gold Affinity Purification. After the extension, magnesium
concentration in the extension mixture was adjusted to 0.1 M
using 1 M MgCl2, and then 4 µL of the extension mixture was
spotted on a circular spot of a disposable gold-plated MS target
plate (ABI), left at room temperature to dry for ∼30 min. Then
the plate was rinsed with water, placed in a container with 20-30
mL of washing buffer (3 M guanidine thiocianate, 10 mM Tris-
HCl pH 9.0, 0.1 mM HSCH2CH2SO3Na, 50% THF), incubated for
10 min at room temperature with gentle shaking, rinsed thor-
oughly with 50% EtOH, and taken for MS. All solvents used for
washings were HPLC grade; salts were Molecular Biology grade.
(MA)LDI-TOF Mass Spectra. These were obtained using a
Voyager Elite Biospectrometry Research Station (PerSeptive
Biosystems, Vestec Mass Spectrometry Products) in a positive
ion mode. Laser power was set on maximum. Signal from three
to five shots was accumulated for each spectrum.
RESULTS AND DISCUSSION
Outline of the Mass Tag-Based SNP Genotyping Method.
Schematic representation of the SNP detection protocol developed
in this study is depicted on Figure 1. The first step of the
procedure, which is common in most SNP detection methods, is
PCR amplification of the fragment, containing the SNP site.
The second step involves a well-developed and widely used
principle of allel-specific primer extension in which alleles are
discriminated by the inhibited extension of a 3′-mismatched probe.
For each allele variant, a probe labeled with a certain trityl is
introduced into the reaction. Thereby, each nucleotide in the
variable position is assigned to a MassTag (the trityl, which mass
is unique in the reaction). As a result of the extension reaction,
probes that match the SNP variant are extended and the corre-
sponding trityls get incorporated into a relatively long double-
stranded DNA; whereas all mismatched probes along with an
excess of matching probes remain in the form of single-stranded
oligonucleotides.
The last and crucial step is to detect only the extended probes.
To this end, the nonextended probes have to be removed from
Analytical Chemistry, Vol. 80, No. 7, April 1, 2008 2345

Figure 4. Primers and probes for SNP 19761 of MHCIII gene. (A) Sequence information (gray bars indicate the position of the primers on the
genomic DNA, corresponding sequences are given aside, names of the oligonucleotides are in italic type, SSsa disulfide-containing group
attached at the 5′-end of the oligonucleotide, Tag1 and tag2smass tags); (B) overall structure of the trityl-labeled probes (structures that are
cleaved off and detected in the mass spectrometer are depicted in boldface type), and (C) mass spectrum of an equimolar mixture of conjugates
17C and 18C.
2346 Analytical Chemistry, Vol. 80, No. 7, April 1, 2008

Figure 5. Gel electrophoresis analysis of the extension reaction with a trityl-labeled probe and disulfide-labeled primer. Lanes: 1, PCR of the
SNP-containing fragment (∼150 nt); 2, 4, extension reaction; 3, extension reaction treated with TFA.
the mixture, because bearing cleavable labels they would produce
the same kind of signal in the mass spectrometer as the extended
probes and bury the specific signal. In order to provide means
for such a separation, a disulfide group was incorporated at the
5′-end of the reverse primer in the extension reaction. In order to
be able to use a universal SS-modified primer for any set of primers
and probes, we introduced a dangling segment (tail) at the 5′-
end of the reverse PCR primer, which provided an annealing site
for a universal primer during the extension reaction (the universal
primer is depicted on Figure 1 as a gray line with SS at one
terminus). As a result, double-stranded DNA molecules bearing
trityl labels would also bear a disulfide group on the opposite
terminus. Therefore, these molecules would be able to bind to
gold, whereas nonextended probes would not possess such
affinity. Crude reaction mixtures (with some extra MgCl2 added)
were spotted onto disposable gold-coated MS target plate (ABI)
and left to dry at room temperature for 30 min. Usually, salt-
containing solutions dry into uneven white patches; however, our
samples dried into a transparent film evenly covering the surface
of the sample circle of the plate. We attribute it to the presence
of some detergent in the reaction buffer for the TopoTaq
polymerase used in these experiments. The plate was then washed
from unbound material, rinsed with 50% ethanol, and analyzed in
a (MA)LDI mass spectrometer without any matrix added. On the
mass spectrum, each allele present in the analyzed DNA sample
should express itself as a peak of the corresponding mass tag.
Mass Tags for DNA Labeling. A series of compounds has
been synthesized to define the physicochemical parameters that
correlate the structural features of the triarylmethyl group with
its MS performance (Korshun et al., manuscript in preparation).
It has been found that the single most important factor influencing
the stability of triarylmethyl carbocations is the number and
position of electron-donating or cation-stabilizing groups. For
instance, additional methoxy groups in ortho or para position of
the trityl’s phenyl rings stabilize the corresponding cation. Some
of the investigated trityl cations are shown in Figure 2A along
with the mass spectrum of the equimolar mixture of corresponding
tritanols. One can see that in LDI conditions tritanols display only
trityl cation signals, [M - OH]+, rather than molecular peaks [M
+ H]+. This is the case for other trityl derivatives as well, e.g.,
ethers.8
activated carboxyl with various amines.8 The central carbon atom
of the trityl is an obvious “connecting point” to an analyte (in our
case oligonucleotide).
Trityl MTs based on 3 were developed and used for oligo-
nucleotide labeling (Figure 2B). Thymidine 5 was 5′-O-tritylated
with NHS-containing dimethoxy trityl chloride 6 (Korshun et
al., manuscript in preparation) and phosphitylated further to 7
(Figure 2B). The labeling reagent 7 contains a phosphoramidite
tool for attaching it to an oligonucleotide during automated DNA
synthesis to give 8 and NHS functionality for postsynthetic
attachment of an added mass through reaction with an amine.
Thus, one phosphoramidite reagent can be used to create a variety
of mass tags due to different amines attached (conjugates 9). With
the reagent 7, the trityl block gets attached to the oligonucleotide
at the 5′-end through the central oxygen and can be cleaved off
under acidic conditions or by laser irradiation releasing trityl MT
10 (Figure 2B), easily detectable by mass spectrometry. In this
arrangement, ionization and flying properties of the mass tag
directly correlate with its lability. Therefore, a synthetic scheme
similar to 7 could not be accomplished for a better flying trityl
cation 4, because the corresponding ether proved to be too
unstable.
It has been known that carbon-sulfur bonds in trityl thioethers
show a higher stability than the corresponding trityl ethers.
However, it has not been known how efficient the trityl cation
release from trityl thioether in the course of LDI is.
In order to compare release of trityl cations from esters versus
thioesters under LDI conditions, we synthesized a model com-
pound 11 (Figure 3A). This compound comprises equimolecular
composition of two monomethoxytrityl-MTs, one of which is
connected through sulfur, while the other one is symmetrically
connected through oxygen. The mass spectrum of this compound
revealed a surprising resultsa more chemically stable thioether
turned out to be easier to brake by laser irradiation than ether.
In other words, the trityl attached through sulfur gave a higher
signal in LDI mass spectrometry (Figure 3A). This combination
of properties of trityl thioesters appears to be favorable for
designing MTs, because it might allow using labels based on
trimethoxy trityl (4) by attaching them to an analyte through
sulfur. This should increase their sensitivity in LDI and at the
same time provide enough stability in chemical and biochemical
manipulations.
Well “flying” trityl cations 3 and 4 (Figure 2A) were elucidated
as potential MT. Trityl core is responsible for the flying abilities
of the trityl MT. The latter must contain also a “mass variation
site” and must be covalently but reversibly bound to an analyte.
To obtain a mass diversity, we used the derivatization of an
These results prompted us to synthesize phosphoramidite for
oligonucleotide labeling with a trimethoxytrityl-based MT (Figure
3B). Activated ether 13 (Korshun et al. , manuscript in prepara-
Analytical Chemistry, Vol. 80, No. 7, April 1, 2008 2347

tion) reacted with 14 under acidic conditions9 and phosphitylated
to give 15. This phosphoramidite can be used for DNA labeling
similar to 7 in Figure 2B. The resulting oligonucleotides can be
decoded in LDI by the mass of the corresponding trityl cation
16.
Genetic Model System for SNP Genotyping. As a model
system, we used a polymorphic site within the MHCIII gene
(LOCUS DJ201G24; M(19761)) C or A). Primers for amplification
of the SNP-containing fragment and allele-specific probes labeled
with mass tags are presented on Figure 4A.
Probes for SNP Detection. Probe C labeled with two different
trityls has been synthesized using phosphoramidite reagents 7
and 15. To further increase sensitivity, three trityl labels were
attached to each DNA molecule during automated synthesis using
branching reagent Trebler (Glen Research).10 PEG-hexamer
linkers (Glen Research) were introduced between the oligonucle-
otide and the Trebler as well as between the Trebler and trityl
labels (Figure 4B). These linkers were employed to prevent
interference of bulky hydrophobic trityl residues with oligonucle-
otide hybridization and extension. After the synthesis, added
masses were introduced by reacting amines with the NHS
functionality of the trityl blocks. Thus, masses of the flying ions
of the labels were 472 Da for 17C-based MT and 558 Da for 18C.
After ammonolysis, fractions containing three labels were purified
by HPLC. An equimolar mixture of oligonucleotides 17C and 18C
was prepared and analyzed by MS without matrix.
As expected, probe 18C labeled with trimethoxytrityl-based
reagent 15 produced substantially higher signal then the probe
labeled with dimethoxytrityl based reagent 7 (Figure 4C).
Extension of Tritylated Probes. Both labels described in the
previous section proved to be stable under PCR conditions in a
buffer with pH 9.0. Considering higher sensitivity of the label
based on reagent 15, this label was chosen for further experi-
ments.
An extension reaction with labeled probe C (18C in Figure
4B) is demonstrated in Figure 5. The PCR reaction with primers
US and DS-tail (Figure 5, lane 1) was diluted 10-fold and used as
a source of template for the extension reaction with 18C and
primer Tail-SS. Six PCR cycles were carried out to produce double-
stranded extension product (calculated length 115 bp). The
reaction was analyzed by gel electrophoresis in 2% agarose gel.
The major product of the extension reaction demonstrated a much
lower mobility than the corresponding unmodified double-stranded
DNA (Figure 5 lanes 2 and 4, compare to PCR template lane 1,
and length markers). Control reactions that lacked either probe
C or Tail-SS did not give any prominent products (data not shown).
We attributed the changed mobility of the extension product to
the presence of branched linkers and labels. In order to confirm
the presence of trityl labels in the extension product, the reaction
mixture was treated with TFA (1:10 v/v of 10% TFA was added to
the reaction, incubated for 30 min at room temperature, and DNA
precipitated by EtOH). Indeed, removal of the trityls from the
extension product caused a reversal change in its mobility (lane
3), making it more similar to the corresponding double-stranded
DNA. Somewhat increased mobility of this detritylated product
Figure 6. Detection of labeled and extended probes by MS. (A)
Detection of equimolar mixture of labeled probes by MS. An equimolar
mixture (0.1 µM) of probes 18A and 18C in 50% MeOH in water was
spotted on the MS target and dried. Mass spectra were taken from
four different locations on two different spots; a representative
spectrum is shown. Intensity of 18C was always taken for 100%, and
relative intensity (see the diagram) and standard deviation (gray tip
of the diagram pyramid) of the second probe 18A was calculated.
(B) Nonspecific binding of the tritylated probe to gold. A nonextended
probe (0.5 µM) was added to the extension mixture after the reaction.
The mixture was loaded on gold, washed, and analyzed by MS.
Signals from the extended probe in each experiment were taken for
100%. The data were averaged from 12 spectra obtained in three
different experiments. Gray tip of the pyramid represents standard
deviation.
(9) Filippi, J. J.; Fernandez, X.; Lizzani-Cuvelier, L.; Loieseau, A. M. Tetrahedron
Lett. 2002, 43, 6267.
(10) Shchepinov, M. S.; Udalova, I. A.; Bridgman, A. J.; Southern, E. M. Nucleic
Acids Res. 1997, 25, 4447-54.
2348 Analytical Chemistry, Vol. 80, No. 7, April 1, 2008

Figure 7. SNP genotyping. Representative spectra of the samples obtained from homozygous DNA (C at position 19761), top spectrum;
homozygous DNA (A at position 19761), middle spectrum; and heterozygous DNA (both alleles present) , bottom spectrum.
can be attributed to the linkers, which remain attached to the DNA
after the trityl removal.
Thus, successful incorporation of the trityl-labeled probe along
with disulfide labeled primer into the extension product has been
demonstrated.
A second probe, probe A, was also synthesized and labeled
with the same type of trityl, but with a different added mass (18A
in Figure 4B, flying ion mass 572 Da).
When an equimolar mixture of the probes 18A and 18C was
spotted on a gold MS target plate, it gave approximately the same
relative MS signal intensities for both probes (Figure 6A), whereas
absolute values of the signal intensity varied in orders of
magnitude. To further address the issue of reproducibility of
relative signal intensities in a mixture of multiple probes, we
prepared a mixture of probes composed of the same oligonucle-
otide labeled with five different MTs and monitored signal
intensities in different locations on the target plate. Relative
intensities were plotted as percent of the lightest probe intensity.
Standard deviations of these values for all probes were within 10%
(see Supporting Information). This confirmed an important notion
that relative intensity of MT signals is area independent.
Gold Surface Affinity Purification and Detection of the
Extended Probes. According to the principle of SNP genotyping
protocol depicted in Figure 1, signals from the probes correspond-
ing to genotypes, which are absent from the DNA sample, should
not appear on the spectrum. False positive signals, however, might
appear either as a result of extension through a 3′-mismatch or
due to incomplete washout of the nonextended probes from the
surface. In order to distinguish these two backgrounds and
develop optimal washing conditions, we carried out a number of
experiments where only one probe was taken to the extension
reaction, while the second probe (at the same concentration) was
added after the reaction, hence remained totally single-stranded.
The washing conditions developed in this study provided fairly
low nonspecific signal in this setup (within 15% of the main peak
intensity, Figure 6B).
Genotyping. The background related to the extension through
a 3′-mismatch is known to largely depend on the probe structure/
SNP environment. Development of efficient allele-specific probes
has been extensively studied yet remains mainly empirical. The
structure of the probes used in this study (probe A and probe C)
has been designed on the basis of experiments with unmodified
DNA probes analyzed by conventional methods (data not shown).
To demonstrate the utility of the method depicted in Figure
1, DNA samples with all three known variants at the investigated
SNP position were tested using this protocol sa heterozygous
Analytical Chemistry, Vol. 80, No. 7, April 1, 2008 2349

DNA (A and C) and two homozygoussone with A and one with
C nucleotide at the position of interest. Human genomic DNAs
representing these genotypes were kindly provided by Dr. I.A.
Udalova of Wellcome Trust Center for Human Genetics (Univer-
sity of Oxford, Oxford, UK).
As expected, homozygous DNAs produced only one major
peak corresponding to the A or C probe, whereas the heterozy-
gous DNA showed both peaks (Figure 7). Slight difference in
intensity of the signals from probe C and probe A was consistent.
It could not be attributed to the mismatch extension, because in
that case we would have also seen such nonspecific extension of
the probe C in the homozygous sample (top spectrum). Appar-
ently, this difference in intensities stems from somewhat higher
extension efficiency of probe C (probably due to G-C pair at the
3′-end rather then A-T).
laser desorption/ionization has been successfully used to imple-
ment this method. A more efficient release of trityl cation from
trityl thioether as compared to ether in the course of laser-assisted
desorption/ionization has been demonstrated.
ACKNOWLEDGMENT
We thank Dr. Susan Wheeler and Mrs. Kaajal Patel from
Oxford Gene Technology Ltd. for their helpful advice in mass-
spectrometry, and Dr. Irina Udalova from Wellcome Trust Centre
for Human Genetics, Oxford Universuty, for her help in selecting
the genetic model.
SUPPORTING INFORMATION AVAILABLE
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
Received for review June 19, 2007. Accepted September
14, 2007.
CONCLUSIONS
These experiments demonstrated the utility of the suggested
SNP genotyping method, which allows use of multiple mass labels.
A new generation of trityl-based mass tags for matrix-independent
AC071291Y
2350 Analytical Chemistry, Vol. 80, No. 7, April 1, 2008