Simultaneous quantification of multiple nucleic acid targets using chemiluminescent probes

Article abstract of DOI:10.1021/ja202221h

A novel method is described for simultaneous detection and quantification of attomoles or a few femtomoles of two (or potentially more) nucleic acid targets, without need for amplification. The technique depends on spectral-temporal resolution of chemiluminescence emitted from independent hybridization-induced chemiluminescent signal probes. The probes are internally quenched except in the presence of their specific targets, thereby allowing detection limits up to 10 000 times lower than with fluorescent probes. This is sufficient to obviate the need for amplification in many cases. The utility of the technique has been demonstrated by use of resolvable N-linked acridinium and 2,7-dimethoxyacridinium ester labeled probes in a homogeneous assay for sensitive and simultaneous independent quantification of pan-bacterial and pan-fungal target sequences in seawater.

Full text of DOI:10.1021/ja202221h

ARTICLE
pubs.acs.org/JACS
Simultaneous Quantification of Multiple Nucleic Acid Targets Using
Chemiluminescent Probes
Kenneth A. Browne,*,† Dimitri D. Deheyn,‡ Gamal A. El-Hiti,§ Keith Smith,§ and Ian Weeks||,^
†Gen-Probe Incorporated, 10210 Genetic Center Drive, San Diego, California 92121, United States
‡Marine Biology Research Division, Scripps Institution of Oceanography, University of California at San Diego, La Jolla,
California 92093, United States
§School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff, Wales CF10 3AT, U.K.
Molecular Light Technology Research Limited (now Gen-Probe Cardiff Ltd.), 5 Chiltern Close, Cardiff Industrial Park, Cardiff,
Wales CF14 5DL, U.K.
S Supporting Information
b
ABSTRACT: A novel method is described for simultaneous detec-
tion and quantification of attomoles or a few femtomoles of two (or
potentially more) nucleic acid targets, without need for amplification.
The technique depends on spectralꢀtemporal resolution of chemi-
luminescence emitted from independent hybridization-induced
chemiluminescent signal probes. The probes are internally quenched
except in the presence of their specific targets, thereby allowing
detection limits up to 10 000 times lower than with fluorescent
probes. This is sufficient to obviate the need for amplification in
many cases. The utility of the technique has been demonstrated by use of resolvable N-linked acridinium and 2,7-dimethoxy-
acridinium ester labeled probes in a homogeneous assay for sensitive and simultaneous independent quantification of pan-bacterial
and pan-fungal target sequences in seawater.
’ INTRODUCTION
complementary nucleic acids induces increased distance
between fluorophore and quenching moieties. Excess un-
bound probe exhibits lower emission than bound probe, and
so separation is unnecessary, supporting a homogeneous
assay format. Furthermore, careful selection of narrow band
excitation sources or filters coupled with appropriate emis-
sion filters advantageously supports the simultaneous discri-
mination of two or more differently labeled fluorescent
probes and, hence, can indicate the presence or absence of
multiple analyte nucleic acids in a single assay. However,
despite mitigating instrumentation and procedural designs,
organic fluorescent probes tend to have a high background
relative to the emitted signal caused by using a light source to
initiate light emission phenomena, incomplete quenching,
and the intrinsic photofluorescence of many species present
in biological samples, including nucleic acids themselves.
This means that the relationship between luminescence
intensity and target nucleic acid concentration is not linear
where the background contribution to the overall signal is
significant and, thus, limits signal-based detection typically to
nanomolar concentrations. Since organisms of interest may
be present in low abundance, fluorescent probe-based detec-
tion often requires high yield target amplification methods
The presence of important and potentially pathogenic viral,
bacterial, fungal, eukaryotic, or other species in dietary, agricul-
tural, pharmaceutical, forensic, clinical, and environmental speci-
mens requires rapid and sensitive methods for both identification
and quantification of such organisms. In this context, molecular
detection techniques such as sequencing,1 microarrays,2 and
nucleic acid probes3 are becoming increasingly important. For
rapid time-to-result, sensitivity and specificity, probes are espe-
cially useful, but in order for them to be effective tools, target-
bound probes need to be distinguished from probes that are not
bound to target. This may be accomplished by separating bound
from unbound probes (heterogeneous methods, as commonly
carried out with radioisotopically labeled probes), then querying
the bound fraction for the detectable signal. However, such
methods increase the complexity of an assay and also provide
opportunities for contamination with probes and/or target
analytes in the reaction vessel.
Methods that avoid undue handling such as separations are
favored. For example, self-reporting nucleic acid probes
labeled with interacting fluorophore and quenching moieties
(e.g., molecular beacons (MBs),3a hairpin inversion probes,4
molecular torches5) are powerful tools for analytical or
diagnostic detection of nucleic acid targets. In the case of
MBs (Scheme 1a), fluorescence induced by photoexcitation is
lower in the absence of target and increases as binding to
Received: March 10, 2011
Published: July 25, 2011
r
2011 American Chemical Society
14637
dx.doi.org/10.1021/ja202221h J. Am. Chem. Soc. 2011, 133, 14637–14648
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Comparison of Interactions and Excitations/
Emissions of Self-Quenching (a) Fluorescent or (b) Chemi-
luminescent Hairpin Probes (MBs and HICS Probes,
Respectively) with Complementary Target Nucleic Acidsa
acridinium position (C-linked) via an aminoalkyl group to a
linear oligonucleotide containing 18ꢀ30 nucleotide (nt) units.8
After the probe has hybridized to its complementary target,
bound and nonbound AE probes could be physically separated,
allowing detection of signal from only bound AE probe (in a
heterogeneous assay). Alternatively, bound and nonbound AE
probes have been differentiated by mild alkaline treatment in
which the ester moiety of the nonbound AE probe is selectively
hydrolyzed, rendering these probes nonchemiluminescent, while
the ester of the AE probe hybridized to complementary target is
protected from hydrolysis by its interaction with the nucleic
acid double strands; this is the basis of the homogeneous
hybridization protection assay.8 AE probe detection is initiated
chemically by addition of hydrogen peroxide (H2O2) followed
by a concentrated base solution, which leads to highly lumines-
cent emissions from the released, excited N-methylacridone.
The signals are typically detected with a very sensitive photo-
multiplier tube (PMT) that sums photons from across the entire
N-methylacridone emission spectrum (while typically reported
as 430 nm,8 these emissions actually range from ∼400 to
540 nm). This leads to a detection limit of ∼600 fM (about
103- to 104-fold more sensitive than typical fluorescent probes)
and a quantitative dynamic range spanning ∼4 orders of
magnitude.8 Such high sensitivity supports either direct detection
of nucleic acid analytes or detection of amplified analytes after
only limited amplification.
Recently, a new probe type, the hybridization-induced chemi-
luminescent signal (HICS) probe, has been developed.9,10 HICS
probes comprise an AE moiety conjugated through the N10
acridinium position (N-linked) to a site near one terminus of a
stem-loop oligonucleotide and a quenching moiety attached near
the opposite terminus (Scheme 1b), therefore resembling fluor-
escent MBs (Scheme 1a). Similar to MBs, upon hybridization of
the HICS probe loop region to a complementary target nucleic
acid, the stem dissociates, distancing the chemiluminophore
from the quencher and allowing light emission to be detected
upon chemical initiation. Thus, HICS probes self-report upon
binding complementary nucleic acids, and probes in excess of
target do not require physical or chemical separation from
hybridized probes, allowing use in a homogeneous assay. Unlike
with their fluorescent counterparts, however, since no methods
are available for simultaneous and independent determination of
the emissions from different probes, quantitative use of HICS
probes has been limited to measurement of only one analyte in
each assay. To overcome this limitation, some novel develop-
ments are required: (i) synthesis of two or more different emitter
moieties that can be independently detected, for example, as
a result of differences in wavelengths or kinetics of emission;
(ii) adaptation of a typical luminometer to allow it to differentiate
the emissions from the different emitters; (iii) development of
a protocol for carrying out multiple analyte detection. Here,
we report development of such a system and demonstrate its
utility by independently measuring amounts of a pan-fungal
nucleic acid sequence and a pan-bacterial nucleic acid sequence
in seawater simultaneously, in a single assay, without need for
amplification.
a Solid lines between nucleic acid strands represent base pairing. F and Q
represent fluorophore and quencher moieties, respectively. AE is an
acridinium ester chemiluminescent moiety.
such as the polymerase chain reaction (PCR)6 or transcrip-
tion-mediated amplification (TMA);7 however, amplification
necessarily increases the complexity, time, and expense of the
assay. Although time-to-detectable result can be improved by
so-called real-time-amplification procedures in which fluor-
escent probes are monitored during amplification, it would
still be advantageous to avoid amplification altogether.
By definition, emissions from chemiluminescent probes are
the result of one or more chemical reactions. Since their
detection is by observation without stimulation (e.g., excitation
light), such probes are necessarily not perturbed by the inter-
rogation technique. Therefore, unlike fluorescent probes, emis-
sion responses from chemiluminescent probes benefit from
much lower intrinsic background emission so that such probes
can be extremely sensitive and also offer a linear relationship
between the magnitude of luminescence and the quantity of
target nucleic acid present in specimens at low concentrations.
Acridinium ester (AE) chemiluminescent nucleic acid probes
have traditionally had the AE moiety linked through the C9
’ RESULTS AND DISCUSSION
Probe Design. In order to allow detection of two nucleic acid
analytes using pairs of HICS probes, the two probes must differ
both in their ability to recognize different oligonucleotide
14638
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
Chart 1. Comparison of acridinium ester reagent structures
sequences and in the nature of the luminescence they emit. Each
probe molecule must incorporate four features: (i) an oligonu-
cleotide strand complementary to an oligonucleotide sequence
in a particular target nucleic acid marker; (ii) extensions on either
side of the aforesaid oligonucleotide sequence by oligonucleotide
strands (stem arms) complementary to each other so that they
cause the two ends of the oligonucleotide to bind to each other
in a hairpin-like arrangement; (iii) a chemiluminophore unit
attached to one end of the oligonucleotide sequence via
an aminoalkyl group or related spacer; (iv) a quencher for the
luminescence emitted by the chemiluminophore, attached to the
other end of the oligonucleotide sequence via an appropriate
spacer (see Scheme 1). The stem arms may be partially or com-
pletely complementary to the target sequence. When the stem arms
hybridize to themselves in the absence of target, the target comple-
mentary sequence typically forms a loop structure. The starting
point was to design an appropriate pair of reagents to introduce two
differentiable chemiluminophores.
Design of Chemiluminophores for Incorporation into the Probes.
9-(4-(2-Succinimidyloxycarbonylethyl)phenoxycarbonyl)-10-methyl-
acridinium trifluoromethanesulfonate (1, Chart 1) is an unsubsti-
tuted AE labeling reagent that has been used to generate biological
probes having C-linked AE moieties.11 It comprises (i) an active
group (N-hydroxysuccinimide, NHS) capable of ready attachment to
an amino substituent in the unlabeled probe, (ii) a leaving group (the
substituted phenoxide) that is eliminated during treatment with
alkaline H2O2, and (iii) a chemiluminescent group (the AE unit,
which is converted into excited state N-methylacridone during
treatment with alkaline H2O2). Substituents on the phenoxy group
primarily influence the rate of the chemiluminescent reaction, while
those on the acridine unit have a major influence on the wavelength of
emission. We investigated one design pathway, wavelength-resolvable
AE probes, in our attempt to develop appropriate systems. This
required a substituted AE that emitted luminescence distinct from
that characteristic of compound 1.
with stability in aqueous environments and the ability to initiate
chemiluminescence under conditions that will not damage
oligonucleotide hybrids. Woodhead et al. posited that extending
the conjugation of the acridinium ring system to a 2,4-pentadie-
noic acid moiety would reduce the energy gap between the
excited and ground state N-methylacridone species, resulting in a
bathochromic shift of chemiluminescent emission wavelength
by up to 80 nm compared to an unmodified AE.12 However,
Law et al. later synthesized a closely related AE and found only a
37 nm increase in peak emission wavelength in an acetonitrile
solution.13 Law et al. also extended their studies to include
AEs fused to additional aromatic rings and reported that the
excited N-methylbenzacridone from the angular benz[a]acri-
dinium ester derivative increased peak emission by only 11 nm
while the excited N-methylbenzacridone from the linear benz-
[b]AE derivative increased peak emission wavelength by 95 nm
in acetonitrile solutions.13 In a series of N-sulfopropyl AEs,
Natrajan et al. reported that incorporation of a single methoxy
group at the acridinium ring 3 position led to a slight (8 nm)
hypsochromic shift in the maximum emission wavelength
whereas incorporation of the substituent at the 2 or 4 position
led to increases in the peak emission wavelength of 32 and 52 nm,
respectively.14 They also found that 2,7-dimethoxy substitu-
tion of their N-sulfopropyl AE led to a 58 nm increase in the
maximum emission wavelength.14
We have previously synthesized a considerable variety of
AEs with different substituents on both the phenoxide and
acridine rings.15 In the course of those studies, we sometimes
observed that, relative to an unmodified AE, the changes in
peak chemiluminescent wavelengths on introduction of a
particular substituent are not always the same as the changes
reported in the literature for similar systems. Effects of
different solvents used may account for at least part of the
differences in observed photonic behaviors. However, we
did observe that the emission wavelength of 2,7-dimethoxy-
9-(4-benzyloxycarbonylethylphenoxycarbonyl)-10-methylacri-
dinium trifluoromethanesulfonate (2, Chart 1) was 55 nm longer
than that from reagent 1 in aqueous solution, presumably
because of electron donation (+I) from the two methoxy
groups.15e
While physicochemical rationalization can assist in the design of
longer wavelength-emitting N-alkylacridone precursors, empir-
ical studies are required to discover AE derivatives that balance
the increased emission wavelength (e.g., g50 nm) needed to
distinguish them from the unsubstituted AE in aqueous solution
14639
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
Scheme 2. Synthesis of N-Linked 2,7-DimethoxyAE NHS Ester Label (4)a
a Reagents and conditions: (i) CuI, K2CO3, L-proline, DMSO, 48 h, 90 ꢀC; (ii) (ClCO)2, AlCl3, HCl; (iii) aq KOH, 72 h; (iv) HCl; (v) SOCl2; (vi)
2,6-dibromophenol; (vii) succinimidyl 4-iodobutanoate, 3 h, 150 ꢀC.
9-(2,6-Dibromophenoxycarbonyl)-10-(3-succinimidyloxycar-
bonylpropyl)acridinium iodide labeling reagent 3 was used to
produce previously reported HICS probes.10 It incorporates two
features that differ significantly from labeling reagent 1. First, it
results in N-linked attachment of the AE label to oligonucleotides
rather than through the phenolic leaving group. When chemically
initiated, light from these probes is emitted from an excited
N-alkylacridone species that remains linked to the oligonucleotide.
In the absence of target, the stem forces the emitter and the
quencher to remain proximal and, hence, allows little energy to
be emitted as detectable light upon initiation. Second, the bulky,
electron-withdrawing (ꢀI) bromo substituents help protect the
ester from facile hydrolysis and also decrease the pKa of the
leaving group phenol, thereby supporting formation of the
excited N-alkylacridone under mildly alkaline conditions that
maintain nucleic acid hybrids. We selected this reagent for
labeling our unsubstituted AE probes and the 2,7-dimethoxy
analogue (4, Chart 1) for labeling the differentiable probes.
Syntheses of Reagents 3 and 4. Compound 3 was obtained
by a published procedure.10 The preparation of compound 4
followed a similar route but with additional steps to produce
2,7-dimethoxyacridine-9-carboxylic acid (7) prior to esterification
and quaternization (Scheme 2). Bis(4-methoxyphenyl)amine
(5) was prepared in 75% yield according to a standard literature
procedure.16 Reaction of compound 5 with oxalyl chloride in
dichloromethane followed by treatment with aluminum chloride
gave N-(4-methoxyphenyl)-5-methoxyisatin (6) in 70% yield,
which on treatment with potassium hydroxide under reflux
conditions for 72 h gave 2,7-dimethoxyacridine-9-carboxylic acid
(7) in 95% yield. A mixture of compound 7 and freshly redistilled
thionyl chloride was heated under reflux for 3 h to afford the
corresponding acid chloride, which on reaction with 2,6-dibro-
mophenol in pyridine at 70 ꢀC gave 2,6-dibromophenyl
2,7-dimethoxyacridine-9-carboxylate (8) in 67% yield. Reaction
of compound 8 with succinimidyl 4-iodobutanoate at 150 ꢀC for
3 h gave 2,7-dimethoxy-9-(2,6-dibromophenoxycarbonyl)-10-(3-
succinimidyloxycarbonylpropyl)acridinium iodide (4) in 6.6%
yield. The low yield resulted from a very slow reaction, even at
150 ꢀC. Use of even more forcing conditions caused some
decomposition of the starting material, whereas under the
conditions reported most of the unreacted starting material
could be recovered without detectable side products. Thus, all
except the last step gave reasonable yields. An even lower yield
was reported in the final step of the published synthesis of
compound 3.10
Selection and Synthesis of Targets. Four target sequences
were selected (Table 1) as representatives of two different, broad
microbial groups found in environmental samples such as sea-
water (Escherichia coli and Candida albicans)17 or of two patho-
gens that may occur individually or simultaneously in human
urogenital swab or urine specimens (Chlamydia trachomatis and
Neisseria gonorrhoeae).18 In order to mimic more closely the
natural nucleic acid materials, the synthetic target oligonucleo-
tide sequences exceeded the length of the part (central in each
target in Table 1) for which complementary sequences in probes
would be constructed. These targets were synthesized on auto-
mated oligonucleotide synthesizers by established methods using
phosphoramidite monomers, then purified and quantified prior
to use. With the targets available, the next stage was to generate
appropriate oligonucleotide strands for the probes.
Design and Synthesis of Probe Oligonucleotide Sequences.
HICS probe sequences (Table 1) were based on earlier fluor-
escent probes4 used in detection of pan-bacterial and pan-fungal
species or were derived from earlier probe sequences19 used to
detect Ctr and Ngo specifically but with self-complementary arm
sequences added to each pair of termini. The probes were also
synthesized on automated oligonucleotide synthesizers by estab-
lished methods using phosphoramidite monomers. Several dif-
ferent pairs of stem arms were incorporated into the
oligonucleotide sequences, all of which included GC clamps at
14640
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
Table 1. Nucleic Acid Sequences for Chosen Targets and the Probes Developed for Them
namea
nucleic acid backboneb,c
sequencec and substituentsd
EcoB1932-1947(-)HICS16
EcoB1932-1947(-)HICS17
EcoB1932-1947(-)HICS18
EcoB1921-1958(+)
DNA/OMe
DNA/OMe
DNA/OMe
RNA
50-AE-GGCTCCTG CGACAAGGAAUUUCGC CAGGAGCC-MR-30
50-AE-CCGAGGAC CGACAAGGAAUUUCGC GTCCTCGG-BHQ2-30
50-AE-CCCAGCAC CGACAAGGAAUUUCGC GTGCTGGG-BHQ2-30
30-GCCUUGAAUGG GCUGUUCCUUAAAGCG AUGGAAUCCUG-50
|
|
|
1950
1940
1930
CalA1185-1206(-)HICS86
CalA1185-1206(-)HICS87
CalA1174-1217(+)
DNA
DNA
RNA
50-AE-GGCTCCTG GTCTGGACCTGGTGAGTTTCCC CAGGAGCC-MR-30
50-dMeOAE-CCGAGGAC GTCTGGACCTGGTGAGTTTCCC GTCCTCGG-BHQ2-30
30-GGAAUAACA CAGACCUGGACCACUCAAAGGG GCACAACUCAG-50
|
|
|
|
1210
1200
1190
1180
CtrB1452-1465(+)HICS51
CtrB1447-1470(-)
OMe
OMe
50-AE-CCGAC AGAGCGAUGAGAAC GUCGG-BHQ2-30
50-C-30-30-AGGCA UCUCGCUACUCUUG CCAAU-50
|
|
|
1450
1460
1470
NgoA133-145(+)HICS62
NgoA128-150(-)
OMe
OMe
50-dMeOAE-CCGAG GUACCGGGUAGCG CUCGG-BHQ2-30
50-C-30-30-CCUUG CAUGGCCCAUCGC CCCCU-50
|
|
|
130
140
150
a Names indicating probe sequence ranges corresponding to target sequence ranges are denoted by three letter abbreviations (Genus species) of the
reference sequences, a letter indicating whether the small (A) or large (B) rRNA subunit, and the target sequence range (probes have an additional
unique designation, e.g., HICS16, from a longer list of previous probes): Eco-series sequences20 are pan-bacterial probes and targets to/from the
Escherichia coli O157:H7 23S rRNA reference sequence (NCBI accession no. E16366); Cal-series sequences20 are pan-fungal probes and targets to/from
the Candida albicans 18S rRNA reference sequence (NCBI accession no. E15168); Ctr-series sequences are probes and targets to/from TMA amplicon
sequences of Chlamydia trachomatis 23S rRNA (NCBI accession no. AM884176); Ngo-series sequences are probes and targets to/from TMA amplicon
sequences of Neisseria gonorrhoeae 16S rRNA (NCBI accession no. NC_002946). Target names are given in bold characters. b Nucleic acid backbone
sugar structures used for corresponding sequences: DNA, deoxyribose; RNA, ribose; OMe, 20-O-methylribose (OMe oligonucleotides have similar
affinities for complementary RNA sequences21 while being more resistant to nucleases than RNA oligonucleotides22). c Underlined probe sequences are
composed of DNA, and remaining probe sequences are composed of OMe. d AE and dMeOAE are N-linked chemiluminophores introduced using
labeling reagent 3 or 4, respectively. BHQ2 and MR are Black Hole Quencher-2 and methyl red quenching moieties, respectively.
their termini to maximize the affinity of the stem arms for each
other. The probe sequences were furnished with one of two
quenching moieties: (i) a 30-BHQ2 unit incorporated as part of
the oligonucleotide synthesis;(ii) aniodoacetyl-methylred labeling
reagent9,10 allowed to react with (after reduction) a 30-disulfide
group incorporated as part of the oligonucleotide synthesis. The
probe oligonucleotides were synthesized with a terminal protected
50-hexylamino phosphoramidite at the end of the oligonucleotide
synthesis sequence. After deprotection, the chemiluminophore
reagents were used to label the freed 50-NH2 group. The seven
probes produced were purified and quantified prior to use.
Testing of the Probes. Substituents on the phenoxide leaving
group of an AE have major implications for the rate of the
chemiluminescent reaction, but all of our probes had the same
leaving group, chosen to provide a balance for initiation of the
reaction at moderate pH. However, substituents on the acridin-
ium ring can influence the rate of reaction,3b so it was important
also to investigate the time courses of emissions and advanta-
geously use differences found to enhance discrimination of
emissions. Therefore, several additional characterizations were
undertaken, namely, time-resolved spectrography and time-
resolved chemiluminescent intensity measurements for individ-
ual probes in nonhybridized and fully hybridized forms and
doseꢀresponse measurements for both individual probes and
mixtures of probes in the presence of a single or two different
target oligonucleotides.
Time-Resolved Spectrography. In order to understand how
2,7-dimethoxy substitution of the AE changes photonic output
compared to the unsubstituted AE, serial emission images were
collected at regular time intervals with an SE200 echelle-type
(diffraction grating-based) low light, digital spectrograph after
chemical initiation of the different HICS probes, transformed to
spectrograms and smoothed. As evidenced from the time-
resolved spectrograms of the N-alkylacridone from probe
EcoB1932-1947(-)HICS18 hybridized to an equimolar amount
of its complementary target (EcoB1921-1958(+), Figure 1a),
chemiluminescence increased rapidly in the 400ꢀ500 nm range
(with smaller contributions from 500ꢀ540 nm) during the first
1ꢀ1.5 s (compare the initial baseline-level spectrogram at t = 0
with the two immediately subsequent spectrograms), followed
by a slower decay for at least 15 s (spectrograms 3ꢀ23). In the
absence of target, chemiluminescence from this probe was low
over the entire time course (Figure 1b), similar to the chemilu-
minescence of the buffer in the absence of probe (data not
shown). Double emission maxima at about 425 and 448 nm were
readily evident in the EcoB1932-1947(-)HICS18 plus target
14641
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
Figure 1. Examples of time-resolved spectrograms after chemical initiation: (a) 50 pmol each of EcoB1932-1947(-)HICS18 and EcoB1921-1958(+);
(b) 50 pmol of EcoB1932-1947(-)HICS18; (c) 300 pmol each of CalA1185-1206(-)HICS87 and CalA1174-1217(+); (d) 300 pmol of CalA1185-
1206(-)HICS87. Curves show the counts per second (cps) summed over the preceding 0.4 s interval (a, b) or the preceding 5 s interval (c, d). Color
patterns in graphs are to better distinguish signal magnitudes and consecutive spectrograms.
chemiluminescence spectrograms, consistent with N-alkylacri-
done fluorescence peaks at about 430 and 450 nm that we have
observed in water15e and peaks at 430 and 443 nm previously
reported in aqueous ethanol.23 The leading edges of the AE
spectrograms were steep, while the trailing edges (higher
wavelengths) were somewhat broad beyond the second peak.
The magnitudes of chemiluminescent spectrograms from the
N-alkyldimethoxyacridone of probe CalA1185-1206(-)HICS87
hybridized to an equimolar amount of its complementary target
(CalA1174-1217(+), Figure 1c) increased rapidly in the
440ꢀ550 nm range {compare the initial spectrogram at t = 0
with the immediately subsequent spectrogram; chemilumines-
cence from this probe in the absence of target was low over the
entire time course (Figure 1d), again comparable to buffer alone
(data not shown)}. However, chemiluminescence from the N-
alkyldimethoxyacridone decayed much more slowly than the N-
alkylacridone (spectrograms 3ꢀ23), nearing background levels
after 120 s. Single peak emissions were at ∼483 nm, very nearly
the same as the 485 and 484 nm peaks found for related
compounds in water15 and dimethylformamide,14 respectively.
However, the present results also showed a pronounced shoulder
at ∼525 nm. This shoulder from the N-alkyldimethoxyacridone
was separated substantially more than the double peaks from the
unsubstituted N-alkylacridone (about 40 nm versus 23 nm,
respectively). The leading and trailing edges of the N-alkyldi-
methoxyacridone were more symmetric than those of N-alkyl-
acridone. The emission rise and fall was uniform across
spectrograms of both HICS probes, without any isosbestic
points, supporting emitted chemiluminescence arising from a
single species. Although partially overlapping, the emissions from
these two acridinium esters were substantially separated below
∼450 nm and above ∼550 nm. In addition to defining the
wavelength ranges expected from probes with these two labels,
the spectrography results suggested initial ranges for quantifying
time-based constants for the different probes.
Time-Resolved Chemiluminescence. Constants for total chem-
iluminescent emissions from HICS probes were determined
after hybridization to excess complementary target oligonucleo-
tides. Preliminary hybridization time courses indicated that
incubation for 40 min at 60 ꢀC and 10 min at room temperature
was sufficient to maximize signal output from each probe under
the conditions employed (data not shown). Plots of counts
per second (cps) versus time indicated a rapid increase in
emissions to a maximum followed by a slow decrease trending
toward the baseline (Figure 2). Background (probe but no
target) emissions followed similar time course patterns but with
substantially attenuated magnitudes (Figure 2, insets; note the
different scales compared to those in the presence of targets).
Multiple series of chemical events occur upon addition of
alkaline peroxide to AEs (Scheme S1 in Supporting In-
formation). In order to be able to compare different probes
quantitatively, it was necessary to simplify the kinetic scheme for
the reactions leading to the increase and decrease in chemilumi-
nescence and then to test this scheme against the observed data.
AEs are chemically similar to lucigenin in that both classes of
compounds emit light through the same type of chemilumines-
cent intermediate, an excited N-alkylacridone. In the reaction of
lucigenin with basic oxygen, Maskiewicz et al. demonstrated that
14642
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
Figure 2. Representative time courses of chemiluminescent emissions for (a) 0.5 pmol of EcoB1932-1947(-)HICS18 plus 1 pmol of EcoB1921-1958(+)
and (b) 0.5 pmol of CalA1185-1206(-)HICS87 plus 1 pmol of CalA1174-1217(+). Each point shows the cps summed over a 0.25 s interval. Insets
are output from 0.5 pmol probes in the absence of target. Lines through data points are nonlinear best fits to a linear combination of exponential
functions (eq 1).
despite photon emission from the excited N-methylacridone
being the result of a minor reaction path, a pair of apparent first-
order reactions for the formation and decay of chemiluminescent
emissions accounted for their data based on a pair of competing,
parallel reactions.24 We adapted their kinetic scheme of processes
discernible by chemiluminescence for our needs (Scheme 3).
This scheme only considers rate limiting formation of a pre-
luminescent intermediate A from the AE label prior to initiation
of chemiluminescence, rapid formation of a postluminescent
product C with emission of light, and a competing reaction to
give a nonchemiluminescent product B. If the dark process
competes effectively with the process leading to intermediate A,
the concentration of A decreases rapidly from its peak and the
overall emission process is of short duration. If the dark process is
much slower than the process leading to intermediate A, then the
emission profile reflects more closely the rate of formation of A
and the overall emission process is of longer duration. Data in
plots of light emission versus time (e.g., Figure 2) were fitted to a
linear combination of exponential functions based on this
scheme (eq 1, where a and b are pre-exponential factors for ka
and kb, respectively; ka = kx ꢀ kꢀx is the apparent first-order rate
constant for the rate determining step for increase in light
intensity; kb = kz ꢀ kꢀz is the apparent first-order rate constant
for the decrease in light intensity; ky is the constant for the rapid
emission of light as the excited intermediate decays to its ground
state; ky > kb; and t is time in seconds). The constants ka and kb,
specific activities, and signal-to-background ratios are compared
along with other parameters in Table 2.
Scheme 3. Discernible Reaction Events Supporting
Observed Chemiluminescent Emissions from AEs
1ꢀ8, indicating that the dark reaction was much less significant
for 2,7-dimethoxyAEs than for unmodified AEs. The time to
peak, a balance of formation and decay of light emissions, varied
∼3.3-fold among the probes. Not surprisingly, the probes with
the longest times to peak and smallest kb values were those
comprising 2,7-dimethoxyAE units, which clearly exhibited emis-
sion over longer periods of time (Figure 2). Probes bearing
unsubstituted AE units and quenchers with a wavelength absorp-
tion profile that was well matched with the emitter wavelength
profile (e.g., EcoB1932-1947(-)HICS16, CalA1185-1206(-)
HICS86) had longer times-to-peak, smaller values of kb, and
lower S/B than the probes for which the quencher absorption
profile was less well matched to the emitter profile (e.g.,
EcoB1932-1947(-)HICS17, EcoB1932-1947(-)HICS18,CtrB1452-
1465(+)HICS51). With the exception of the pan-bacterial probe
with a methyl red quencher, which showed an SA of only ∼4 ꢁ
107 ∑cps/pmol and a S/B of less than 5, SA and S/B of all probes
were high, exceeding 108 ∑cps/pmol and 40, respectively.
Chemiluminescent emission constants were used to select
probe constructs with preferred attributes. For the current studies,
the Ctr/Ngo pair of distinguishable probes and their target was
not chosen as a model system because of the very long data
collection that would be required for NgoA133-145(+)HICS62
(5τ1/2 ≈ 7.8 min). The only pan-fungal probe that could be clearly
distinguished from the pan-bacterial AE-labeled probes was
2,7-dimethoxyAE-labeled CalA1185-1206(-)HICS87. From the pan-
bacterial probes, the SA and S/B from EcoB1932-1947(-)HICS16
were substantially lower than from EcoB1932-1947(-)HICS17 or
EcoB1932-1947(-)HICS18. Preliminary experiments with
EcoB1932-1947(-)HICS17 and CalA1185-1206(-)HICS87 re-
sulted in increased quenching of each other as a function of
increased specific probeꢀtarget hybrid, suggesting interactions
between the arms of these two different probes in the open state
(data not shown). Therefore, EcoB1932-1947(-)HICS18 and
CalA1185-1206(-)HICS87 were selected for demonstrating util-
ity for simultaneous detection, discrimination, and quantification
of two targets in individual samples.
cps ¼ að ꢀ expꢀk tÞ þ bðexpꢀk t
Þ
ð1Þ
a
b
The emission increase constant (ka in Scheme 3 and eq 1)
varied little more than ∼2-fold despite the different probes
varying in terms of target-specific sequence, ribose backbone,
labels and quenching moieties (Table 2). This is consistent with
all of the probes containing the same 2,6-dibromophenoxy
leaving group. The constant for decreasing emission (kb), on
the other hand, varied ∼125-fold among probes, leading to τ1/2
for decay in light emission ranging from 0.75 to 94 s. However,
for a given AE label-quencher combination, kb varied only up to
∼3-fold (e.g., 1.1-fold for probes containing AE and methyl red;
3.2-fold for probes containing AE and BHQ2; and 2.5-fold for
probes containing 2,7-dimethoxyAE and BHQ2). The average
value of kb for the two probes bearing a 2,7-dimethoxyAE unit
was almost 30 times lower than the average value for all probes
bearing an AE unit without modifications at acridinium positions
14643
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
Table 2. Chemiluminescent Emission Constants for HICS Probesa
name
time-to-peak (s)
ka (sꢀ1
)
kb (sꢀ1
)
SA b (∑cps/pmol)
S/B b
EcoB1932-1947(-)HICS16
EcoB1932-1947(-)HICS17
EcoB1932-1947(-)HICS18
CalA1185-1206(-)HICS86
CalA1185-1206(-)HICS87
CtrB1452-1465(+)HICS51
NgoA133-145(+)HICS62
1.3 [16]
0.63 [16]
0.70 [6]
1.0 [18]
1.8 [16]
0.75 [14]
2.1 [14]
2.57 ( 0.14 [18]
4.53 ( 0.15 [18]
3.99 ( 0.07 [6]
3.49 ( 0.17 [20]
2.92 ( 0.27 [18]
2.11 ( 0.09 [14]
2.86 ( 0.21 [17]
0.157 ( 0.004 [16]
0.339 ( 0.019 [16]
0.291 ( 0.008 [14]
0.141 ( 0.003 [18]
0.0188 ( 0.0004 [16]
0.924 ( 0.019 [14]
0.00735 ( 0.00032 [14]
(3.93 ( 0.22) ꢁ 107 [16]
(2.08 ( 0.06) ꢁ 108 [16]
(3.34 ( 0.09) ꢁ 108 [14]
(3.74 ( 0.08) ꢁ 108 [18]
(2.27 ( 0.04) ꢁ 108 [2]
(1.72 ( 0.04) ꢁ 108 [14]
(1.67 ( 0.07) ꢁ 108 [3]
4.60 ( 0.33 [16]
150 ( 6 [16]
285 ( 10 [14]
73.3 ( 2.8 [18]
93.6 ( 2.5 [2]
276 ( 11 [14]
40.4 ( 1.8 [3]
a Values are the mean ( standard deviations. Numbers in brackets are numbers of replicates. ka and kb are from nonlinear best fits to eq 1 from at least 5 ꢁ
first-order half-lives (τ1/2) of each constant. b Specific activities (SA) are from the summed cps from 5τ1/2 of the decay process divided by the picomoles
of probe, and signal-to-background ratios (S/B) are from the summed cps of 5τ1/2 of the decay process for probeꢀtarget hybrids divided by the same
sum for probes alone.
Simultaneous Chemiluminescent Responses from Probes
Hybridizing to Target Sequences. Since the probes were
designed, and shown (vide supra), to be distinguishable within
a mixture by differences in wavelengths of emission, it was
necessary to adapt a luminometer to distinguish multiple
wavelengths simultaneously. This was achieved for two probe
systems by incorporation of a second PMT into a lumi-
nometer and separately attenuating the pair of PMTs with two
different filters, one of which allowed light of relatively short
wavelength (e450 nm) to pass while the other allowed only
light of longer wavelength (g550 nm) to pass.
In an attempt to emulate assay conditions for simultaneous,
direct detection of bacteria and/or fungi in an environmental
sample, constant amounts of both EcoB1932-1947(-)HICS18
and CalA1185-1206(-)HICS87 were allowed to hybridize to
increasing amounts of one, the other, or both target nucleic acids
in a mixture of seawater and hybridization reagent. As is evident
from Figure 1, some longer wavelength emissions from
EcoB1932-1947(-)HICS18 overlapped with some shorter wave-
length emissions from CalA1185-1206(-)HICS87. Such overlaps
could be problematic for distinguishing emissions at higher
concentrations of each target. However, emission decays from
EcoB1932-1947(-)HICS18 were rapid after a high peak emission,
while those from CalA1185-1206(-)HICS87 were considerably
slower and had a substantially lower maximum peak intensity
(Figure 2 and Table 2). Taking advantage of wavelength separa-
tions as well as these timing differences facilitated considerable
reduction in overlapping emissions from these probes at high
concentrations without detrimental reduction in specific target
signal intensities.
Figure 3 demonstrates spectralꢀtemporal resolved emissions
from one example of several target dilutions. The deconvoluted
response from EcoB1932-1947(-)HICS18 (corresponding to the
first 12 s of emission measured at e450 nm, Figure 3a) was
linearly quantitative from 500 amol to 200 fmol of bacterial
nucleic acid EcoB1921-1958(+) in the absence or presence of a
high concentration (50 fmol) of fungal CalA1174-1217(+);
EcoB1921-1958(+) curves in the absence or presence of
CalA1174-1217(+) are nearly indistinguishable except at the
lowest target concentrations. As evidenced by the near zero slope
(m < 0.02) of the line linking emissions from solutions contain-
ing 50 fmol of bacterial target and various amounts of fungal
target, the presence of fungal target (200 amol to 200 fmol) did
not appreciably influence the emission responses of EcoB1932-
1947(-)HICS18 from bacterial target. The deconvoluted
response from CalA1185-1206(-)HICS87 (corresponding to the
emission measured at g550 nm between 12 and 184 s,
Figure 3b) was also linearly quantitative from ∼500 amol to
200 fmol of fungal target in the absence or presence of a high
concentration (50 fmol) of bacterial target; CalA1174-1217(+)
curves in the absence or presence of EcoB1921-1958(+)
are nearly indistinguishable except at the lowest target
concentrations. Very high concentrations (e.g., 200 fmol) of
bacterial target in the absence of fungal target yielded substantial
long wavelength and long-lived emissions, effectively limiting the
quantitative detection range of CalA1185-1206(-)HICS87 in the
presence of possible bacterial target to just over 5 fmol to
200 fmol. Similar to EcoB1932-1947(-)HICS18, the emission
responses from CalA1185-1206(-)HICS87 plus 50 fmol fungal
target were minimally influenced by the presence of 200 amol to
200 fmol bacterial target (m < 0.02). The slopes and inter-
cepts of the specific target dilutions were similar to each other
in the absence or presence of the other target, and repeated
experiments demonstrated similar results (Table S1 in Sup-
porting Information).
From a single analyte point of view, the current results are
close to or exceed the sensitivity reported in many literature
examples. For example, in a heterogeneous, single-analyte,
sandwich-type antibody assay with eight complex 2,7-di-
alkoxyAE labels (similar relative quantum yields as
2,7-dimethoxyAE) per secondary antibody, as little as 28
pmol of theophylline could be detected.14 In a homogeneous,
single-analyte molecular assay with a single C-linked AE label
per linear DNA probe, 30 amol of Ctr target could be
detected.25 Finally, in another molecular assay in which single
analytes were captured on magnetic particles in a ternary
complex conjugated to horseradish peroxidase, chemilumi-
nescent detection of DNA target was down to 10 amol.26
However, compared with such tests, the present method has
the major advantage of being suitable for simultaneous
measurement of two analytes.
Other label and detection strategies have been used to
quantify multiple analytes in individual, nonamplified sam-
ples. For example, Adamczyk et al. captured bovine serum
albumin and myoglobin from a common solution, removed
unbound materials by washing, and detected 100 fmol to 100
pmol of the analytes in the same solution by sequentially
triggering aequorin label with Ca2+, recording luminescence,
then triggering an AE label with alkaline peroxide and recording
luminescence.27 Using acridinium- and benz[b]acridinium-
labeled antibodies to follicle-stimulating hormone (FSH) and
luteinizing hormone (LH), respectively, in a heterogeneous
14644
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
Figure 3. Spectralꢀtemporal deconvolution of chemiluminescent emissions due to EcoB1932-1947(-)HICS18 and CalA1185-1206(-)HICS87 from
mixtures containing 200 fmol of both probes plus various amounts of target (calibration curves) in a 1:1 mixture of seawater and 2ꢁ hybridization
reagent. Open symbols are for EcoB1921-1958(+) (O) or CalA1174-1217(+) (0) alone, while closed symbols are for EcoB1921-1958(+) (b) or
CalA1174-1217(+) (9) plus 50 fmol of the other target. (a) Each point shows the wavelength-resolved cps (<450 nm) summed over 5t1/2 of EcoB1932-
1947(-)HICS18 (0ꢀ12.0 s). (b) Each point shows the wavelength-resolved cps (>550 nm) summed from the end of the first 5t1/2 of EcoB1932-1947(-)
HICS18 to the end of the first 5t1/2 of CalA1185-1206(-)HICS87 (12.0ꢀ184 s). All conditions were performed in at least duplicate. Lines through
background-subtracted data points are linear best fits (Table S1). The intersections of the solid target-response and dashed lines indicate the minimum of
one target that can be distinguished in the presence of a high quantity (200 fmol) of the other target.
assay, Law et al. disclosed dual, wavelength-resolved detection of
FSH and LH down to about 10ꢀ20 mIU/mL over about 10- and
5-fold concentration ranges.13 Nelson et al. used pairs of linear
nucleic acid probes labeled with N-methyl AE that differed in the
pKa of their phenolic leaving groups and, hence, in their emission
time constants to homogeneously detect and quantify by time
resolution as little as 500 amol of Ngo or 150 amol of Ctr over a
20-fold concentration range in the absence or presence of the
other analyte.3b While this last example demonstrated higher
sensitivity for one of a pair of dual analytes, the dynamic range
for the other analyte was narrower than in the current study
(e.g., Figure 3).
’ EXPERIMENTAL SECTION
(1) Organic Synthesis. (1.a) General. Chemicals and reagents
were obtained from Aldrich Chemical Co. and were used without further
purification unless otherwise stated. Tetrahydrofuran (THF) was dis-
tilled from sodium benzophenone ketyl. Other solvents were purified by
standard procedures.28 Melting point determinations were performed
by the open capillary method using a Gallenkamp melting point
apparatus and are reported uncorrected. 1H and 13C NMR spectra were
recorded on an AV500 spectrometer operating at 500 MHz for 1H or
125 MHz for 13C measurements. Chemical shifts δ are reported in parts
per million (ppm) relative to tetramethylsilane, and coupling constants
J are in Hz. Low- and high-resolution mass spectra were recorded on a
GCT-Premier spectrometer, electron impact (EI) at 70 eV. Atmospheric
pressure chemical ionization (APCI) mass spectra were acquired on a
Waters LCT Premier XE instrument. Electrospray (ES) mass spectro-
metric analyses were performed on a ZQ4000 instrument in positive
ionization mode. Column chromatography was carried out using Fisher
Scientific silica 60A (35ꢀ70 μm).
’ CONCLUSIONS
A new chemiluminescent label incorporating a 2,7-dimethoxy-
AE unit has been synthesized and used to generate HICS
probes. Emission constants of these probes have been character-
ized and compared to an existing HICS probe label design.
Unimodal rise-and-decay of spectrograms over time from both
probe types supports that emissions are from the generation of
excited acridone species from AE- and from 2,7-dimethoxyAE-
labeled probes. Peak emissions from 2,7-dimethoxyAE-labeled
probes shift bathochromatically by ∼58 nm from those of AE-
labeled probes. Time-resolved chemiluminescent emissions in-
dicate that 2,7-dimethoxyAE-labeled probes have emitter forma-
tion rates that are similar to those from AE-labeled probes but
that they have substantially slower emission decay rates because
of slower dark reactions of the AE. These differences in emission
wavelengths and time courses advantageously support simulta-
neous spectralꢀtemporal separation of signals from two probes
in a dual wavelength luminometer. This allowed demonstration
of a rapid (∼60 min/sample), sensitive (about 500 amol to 5
fmol), and homogeneous detection and quantification (dynamic
range spanning up to 3 log concentration units) process without
enzymatic amplification of two microbial target nucleic acids
representing those that may be present in single environmental
samples. In principle, the method lends itself to providing further
assay value by addition of extra detection channels to the
luminometer and by including one or more additional spectrally
and temporally resolved probes to other analytes or internal
controls.
(1.b) Bis(4-methoxyphenyl)amine (5). Compound 5 was prepared in
75% yield according to a standard literature procedure:16 mp
102ꢀ103 ꢀC (lit. 102ꢀ103 ꢀC,29 99.5ꢀ101.5 ꢀC 30).
(1.c) N-(4-Methoxyphenyl)-5-methoxyisatin (6). A solution of 5
(3.21 g, 14.0 mmol) in dichloromethane (CH2Cl2, 40 mL) was added
dropwise to a stirred, refluxing solution of oxalyl chloride (3.53 g, 28.0
mmol) in CH2Cl2 (70 mL). The mixture was heated under reflux for 1 h,
and the excess oxalyl chloride and CH2Cl2 were then removed under
reduced pressure. To the residue, CH2Cl2 (100 mL) was added,
followed by anhydrous AlCl3 (4.32 g, 32.4 mmol) portionwise over 10
min. The mixture was heated under reflux for 1 h. The solvent was
removed under reduced pressure, and to the residue was added dilute
HCl (40 mL of ∼1 M). The mixture was stirred for 30 min and was then
extracted with CHCl3 (3 ꢁ 30 mL). The extracts were combined and
dried (MgSO4), and the solvent was removed under reduced pressure.
The solid obtained was washed with CH2Cl2 to give pure 6 (2.78 g, 9.81
mmol; 70%): mp 203ꢀ205 ꢀC; 1H NMR (CDCl3) δ 7.34 (d, J = 9.0 Hz,
2H), 7.25 (d, J = 2.8 Hz, 1H), 7.13 (dd, J = 2.8 and 8.6 Hz, 1H), 7.07 (d,
J = 9.0 Hz, 2H), 6.79 (d, J = 8.6 Hz, 1H), 3.88 (s, 3H), 3.84 (s, 3H); 13C
NMR (CDCl3) δ 183.5, 159.5, 157.7, 156.7, 146.1, 127.3, 125.5, 125.0,
117.8, 115.2, 112.3, 109.1, 56.0, 55.6; HRMS (APCI) calcd for
C16H14NO4 (MH+) 284.0923, found 284.0935.
(1.d) 2,7-Dimethoxyacridine-9-carboxylic Acid (7). A mixture of 6
(1.42 g, 5.00 mmol) and KOH (7.00 g, 125 mmol) in water (70 mL) was
14645
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
refluxed for 72 h. The resulting mixture, after cooling, was poured into a
mixture of concentrated HCl (15 mL) and ice (30 g). The yellow solid
obtained was collected by filtration, washed with water and CH2Cl2, and
then dried in a vacuum oven at 60 ꢀC overnight to give 7 (1.35 g,
(N,N-diisopropyl) phosphoramidite (part no. 10-1906, Glen Re-
search, Sterling, VA) was used to functionalize probe oligonucleo-
tides with 50-terminal amine linker arms. The dimethoxytrityl
group was removed (as trityl groups were for subsequently incor-
porated phosphoramidites) by machine-automated detritylation
with trichloroacetic acid in dichloromethane according to vendor
recommendations. After completion of their syntheses, oligonu-
cleotides were cleaved from the CPG column under standard
conditions with ammonium hydroxide, followed by purification
by standard polyacrylamide gel electrophoretic separation as pre-
viously described.4
(2.b) Labeling of the Oligonucleotides. The oligonucleotides were
deprotected to liberate the terminal amino group according to vendor
recommendations. The amino groups were then labeled through the
succinimidyl groups of 9-(2,6-dibromophenoxycarbonyl)-10-(3-succin-
imidyloxycarbonylpropyl)acridinium iodide (3)9,10 or 9-(2,6-dibromo-
phenoxycarbonyl)-2,7-dimethoxy-10-(3-succinimidyloxycarbonylpropyl)
acridinium iodide (4).10 The 30-disulfides were reduced according to
vendor recommendation (http://www.glenresearch.com//Technical/
TB_Thiol_Modifier_S-S.pdf, accessed Nov 3, 2010) with dithiothreitol
and coupled with iodoacetyl-methyl red9 according to the literature
procedure.9,10 The fully labeled oligonucleotides were purified by
reversed phase HPLC as previously described.4 The structures of the
oligonucleotides (labeled or not) were confirmed by MALDI-TOF MS
(Supporting Information), and their concentrations were determined by
conventional absorption spectroscopy using a DU 640B spectropho-
tometer (Beckman Coulter, Inc., Brea, CA).
(2.c) Chemiluminescent Spectrographic Emissions. Time-resolved,
spectrographic images of HICS probe chemiluminescent emissions
were acquired on a low light, echelle-type SE200 spectrograph
(Optomechanics Research Inc., Vail, AZ) using KestrelSpec software
(Catalina Scientific Corp., Tucson, AZ). The images were converted to
spectrograms at 1 nm resolution and then smoothed (4%) with a locally
weighted scatter plot smoothing algorithm31 (Excel Add-In, Peltier
Technical Services, http://peltiertech.com/, accessed Jul 31, 2009).
An amount of 50 pmol/100 μL of AE-bearing probes (Table 1) or 300
pmol/100 μL of 2,7-dimethoxyAE-linked probes (Table 1), either alone
or together with 50 or 300 pmol/100 μL of the corresponding specific
synthetic target (Table 1), was allowed to incubate for 15 min at 60 ꢀC in
a low pH, surfactant-containing 1ꢁ hybridization reagent (indicating
hybridization or predetection concentrations; 95 mM succinic acid,
1.5 mM ethylenediamine-N,N,N0,N0-tetraacetic acid, 1.5 mM ethylene
glycol-bis(2-aminoethylether)-N,N,N0,N0-tetraacetic acid, 312 mM
lithium dodecyl sulfate, 125 mM LiOH, pH 5.2)9 and cooled for at
least 15 min at room temperature. These conditions allowed probe loop
hybridization to target sequences to occur in the solutions containing
target and probe stem hybridization to occur in solutions without target.
Chemiluminescence from 100 μL of each of the solutions (probes
incubated with or without complementary target) was initiated by a
single 100 μL injection of 240 mM H2O2/2 M Tris-HCl, pH 9.0,
solution, and data collection was started immediately (for 25 ꢁ 0.4 s
or 5 s intervals separated by 0.37 s pauses for the AE-linked or
2,7-dimethoxyAE-linked probes, respectively).
(2.d) Chemiluminescent Constants. Chemiluminescent emission
time courses were acquired on a Gen-Probe Incorporated LEADER
HC+ luminometer. Solutions containing 0.5 pmol/100 μL probe
with or without 1 pmol/100 μL specific synthetic target in 1ꢁ
hybridization reagent plus 100 μL of silicone oil were mixed by
vortexing in 12 mm ꢁ 75 mm polypropylene tubes and allowed to
incubate in a 60 ꢀC water bath for 40 min to facilitate hybridization of
probes and targets (when present). After the mixtures had cooled to
room temperature for at least 10 min, chemiluminescent emissions
were initiated by a 200 μL injection of detect 1 solution (240 mM
H2O2/1 mM HNO3) followed by a 2 s pause and then a 200 μL
1
4.76 mmol; 95%): mp 192ꢀ193 ꢀC; H NMR (DMSO-d6) δ 8.32
(d, J = 9.4 Hz, 2H), 7.75 (dd, J = 2.6 and 9.4 Hz, 2H), 7.29 (d, J = 2.6 Hz,
2H), 3.97 (s, 6H). The material was highly insoluble in solvents, and the
13C NMR spectrum was not recorded. HRMS (EI) calcd for
C16H13NO4 (M+) 283.0845, found 283.0845.
(1.e) 2,6-Dibromophenyl 2,7-Dimethoxyacridine-9-carboxylate (8).
A mixture of 7 (1.21 g, 4.27 mmol) and freshly distilled thionyl chloride
(20 mL, 69 mmol) was refluxed under anhydrous conditions for 3 h. The
excess thionyl chloride was removed under reduced pressure to leave an
orange solid. 2,6-Dibromophenol (1.33 g, 5.30 mmol) was dissolved in
pyridine (20 mL) by heating at 70 ꢀC, cooled, and transferred into the
flask containing the acid chloride. The mixture was heated at 70 ꢀC
overnight. The solvent was removed under reduced pressure, and the
residue obtained was extracted with CH2Cl2. The extract was concen-
trated, and the residue obtained was purified by column chromatography
(silica gel; Et2Oꢀhexane, 1:3) to give 8 (1.48 g, 2.86 mmol; 67% yield):
mp 245ꢀ247 ꢀC; 1H NMR (CDCl3) δ 8.17 (d, J = 9.4 Hz, 2H,), 7.89 (d,
J = 2.6 Hz, 2H), 7.74 (d, J = 8.1 Hz, 2H), 7.48 (dd, J = 2.6 and 9.4 Hz,
2H), 7.18 (t, J = 8.1 Hz, 1H), 4.02 (s, 6H); 13C NMR (CDCl3) δ 164.0,
158.9, 146.7, 144.25, 132.9, 131.8, 128.9, 128.2, 125.0, 124.3, 117.9,
101.3, 56.1; HRMS (EI) calcd for C22H15NO479Br2 (M+) 514.9368,
found 514.9367.
(1.f) Succinimidyl 4-Iodobutanoate. Succinimidyl 4-iodobutano-
ate was prepared in 84% yield according to a standard literature
procedure:10 mp 86ꢀ87 ꢀC (lit. 86ꢀ87 ꢀC 10).
(1.g) 9-(2,6-Dibromophenoxycarbonyl)-2,7-dimethoxy-10-(3-succin-
imidyloxycarbonylpropyl)acridinium Iodide (4). A well-mixed mix-
ture of 8 (50 mg, 0.091 mmol) and succinimidyl 4-iodobutanoate (50
mg, 0.16 mmol) was heated in an oil bath (150 ꢀC) for 3 h under
nitrogen. The mixture turned to dark brown, and acetonitrile (3 ꢁ 1 mL)
was used to extract the product. The crude product was precipitated with
diethyl ether (10 mL), then subjected to extensive washing with diethyl
ether and diethyl ether/CH2Cl2 mixture (5/1 by volume) until TLC
showed that all the starting materials were removed, leaving 4 as a
brownish solid (3 mg, 0.0036 mmol; 4%). The washings were evaporated
under reduced pressure. The residue was subjected to the same reaction
procedure again to give further 4 (1 mg), and the cycle was repeated yet
again to give further 4 (1 mg). The total yield of 4 was 6.6%: mp
159ꢀ162 ꢀC;1H NMR(CD3CN) δ 8.64 (d, J = 9.5 Hz, 2H), 8.07 (dd, J =
2.7 and 9.5 Hz, 2H), 7.98 (d, J = 2.7 Hz, 2H), 7.90 (d, J = 8.2 Hz, 2H), 7.36
(t, J = 8.2 Hz, 1H), 5.44 (t, J = 7.0 Hz, 2H), 4.08 (s, 6H), 3.16 (t, J = 7.0Hz,
2H), 2.85 (s, 4H), 2.57 (m, 2H); HRMS (ES+) calcd for
C30H25N2O879Br2 ([M ꢀ I]+) 698.9978, found 698.9982.
(2) Oligonucleotides. (2.a) Synthesis. Oligonucleotides were
synthesized in-house using an Expedite model 8909 nucleic acid synthesis
system {PerSeptive Biosystems (now part of Life Technologies Corpora-
tion, Carlsbad, CA)}. The syntheses of the probe oligonucleotides utilized
substituted 500 Å controlled pore glass (CPG) substrates packed in
automated synthesizer columns. Deoxy CPG were used for RNA
target syntheses, resulting in 30-deoxyribonucleotides on these oli-
gonucleotides. Probe sequences incorporating a 30-terminal BHQ2
unit started with Black Hole Quencher-2 attached to CPG via a
glycolate linker and with a dimethoxytrityl (DMT) protecting the
terminal hydroxyl (part no. CG5-5042G, Biosearch Technologies,
Inc., Novato, CA), and sequences destined for eventual conjugation
to 2-(4-dimethylaminophenylazo)-N-[2-(2-iodoacetylamino)ethyl]-
benzamide (iodoacetyl-methyl red) to incorporate the quenching
agent started with 1-O-dimethoxytrityl-propyl-disulfide,10-succin-
yl-[long chain alkylamino]-CPG (part no. 20-2933, Glen Research,
Sterling, VA). 6-(4-Monomethoxytritylamino)hexyl (2-cyanoethyl)
14646
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
injection of detect 2 solution (2 M Tris-HCl, pH 9.0), and data
collection was started after the first 40 ms mixing time and continued
for at least 5τ1/2 of the apparent first order rate constant kb (10ꢀ500 s
total) without interinterval delays. Final pH after detection was
always 9.0. Constants were calculated as described in the Results and
Discussion. Interpretation of the kinetic constant ka may be limited
by the hardware and firmware configurations of the lumimometer.
(2.e) Competitive Wavelength-Resolved Chemiluminescence. Chem-
iluminescent emission time courses from 0.2 pmol/100 μL of both
EcoB1932-1947(-)HICS18 and CalA1185-1206(-)HICS87 probes in
hybridization reagent were acquired simultaneously on a luminometer
modified for dual wavelength-range detection. This luminometer was
equipped with two high-count PMT modules (28 mm diameter, head-
on, bialkali cathode, peak cathode radiant sensitivity at 420 nm; Hama-
matsu Photonics, Hamamatsu City, Japan) on opposite sides of and
directed toward a light-tight detection chamber fitted with injector
tubing from two reagent pumps. Filters {25.4 mm diameter; 417/60
BrightLine bandpass filter (transmits ∼95% of light from 385 to
450 nm) from Semrock, Inc. (Rochester, NY) for AE, and OG550
cut-on filter (transmits ∼90% of light from 550 to 700 nm) from
Newport Corporation (Irvine, CA) for 2,7-dimethoxyAE} were fitted
between PMT 1 and PMT 2, respectively, and the detection chamber.
Custom VisualBasic software controlled reagent injections and chemi-
luminescent data acquisition from the two PMT channels. Hybridiza-
tions were performed substantially as for Chemiluminescent Constants
(above) except an amount of 50 μL of 0 or 200 amol to 200 fmol/50 μL
targets in seawater (0.2 μm filtered; Scripps Institution of Oceanography
Pier, La Jolla, CA) was mixed with 50 μL of 2ꢁ hybridization reagent
containing 0.5 pmol/50 μL of both probes. Chemiluminescence was
initiated by a 200 μL injection of detect 1 solution followed by a 2 s
pause, then a 200 μL injection of detect 2 solution, and data collection
was started immediately from the dual wavelength luminometer for
230 ꢁ 0.8 s intervals (184 s total) without interinterval delays. Final pH
after detection was always 9.0.
Scientific Research (AFOSR) under Award No. FA9550-07-
1-0027 to D.D.D.
’ REFERENCES
(1) (a) Meller, A.; Nivon, L.; Brandin, E.; Golovchenko, J.; Branton,
D. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1079–1084. (b) Shendure, J.;
Porreca, G. J.; Reppas, N. B.; Lin, X.; McCutcheon, J. P.; Rosenbaum,
A. M.; Wang, M. D.; Zhang, K.; Mitra, R. D.; Church, G. M. Science 2005,
309, 1728–1732. (c) Ozsolak, F.; Platt, A. R.; Jones, D. R.; Reifenberger,
J. G.; Sass, L. E.; McInerney, P.; Thompson, J. F.; Bowers, J.; Jarosz, M.;
Milos, P. M. Nature 2009, 461, 814–818.
(2) (a) Southern, E. M. J. Mol. Biol. 1975, 98, 503–517. (b) Fulton,
R. J.; McDade, R. L.; Smith, P. L.; Kienker, L. J.; Kettman, J. R., Jr. Clin.
Chem. 1997, 43, 1749–1756. (c) Gunderson, K. L.; Steemers, F. J.; Lee,
G.; Mendoza, L. G.; Chee, M. S. Nat. Genet. 2005, 37, 549–554.
(3) (a) Tyagi, S.; Kramer, F. R. Nat. Biotechnol. 1996, 14, 303–308.
(b) Nelson, N. C.; Cheikh, A. B.; Matsuda, E.; Becker, M. M. Biochem-
istry 1996, 35, 8429–8438. (c) Gaylord, B. S.; Heeger, A. J.; Bazan, G. C.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 10954–10957. (d) Storhoff, J. J.;
Lucas, A. D.; Garimella, V.; Bao, Y. P.; M€uller, U. R. Nat. Biotechnol.
2004, 22, 883–887.
(4) Browne, K. A. J. Am. Chem. Soc. 2005, 127, 1989–1994.
(5) Becker, M. M.; Schroth, G. P. U.S. Patent 6,534,274, 2003.
(6) Saiki, R. K.; Gelfand, D. H.; Stoffel, S.; Scharf, S. J.; Higuchi, R.;
Horn, G. T.; Mullis, K. B.; Erlich, H. A. Science 1988, 239, 487–491.
(7) Brentano, S. T.; McDonough, S. H. Nonradioactive Analysis of
Biomolecules; Kessler, C., Ed.; Springer-Verlag: New York, 2000; pp
374ꢀ380.
(8) Nelson, N. C.; Reynolds, M. A.; Arnold, L. J., Jr. Nonisotopic
Probing, Blotting, and Sequencing; Kricka, L. J., Ed.; Academic Press: San
Diego, CA, 1995; pp 391ꢀ428.
(9) (a) Rutter, A. J.; Weeks, I.; Li, Z.; Smith, K. U.S. Patent
7,169,554, 2007. (b) Weeks, I.; Rutter, A. J.; Li, Z.; Smith, K. U.S.
Patent Application 10/570,175, 2007.
(10) Brown, R. C.; Li, Z.; Rutter, A. J.; Mu, X.; Weeks, O. H.; Smith,
K.; Weeks, I. Org. Biomol. Chem. 2009, 7, 386–394.
(11) Campbell, A. K.; Woodhead, J. S.; Weeks, I. U.S. Patent
’ ASSOCIATED CONTENT
4,946,958, 1990.
S
Supporting Information. AE reaction scheme and table
b
(12) Woodhead, J. S.; Weeks, I.; Batmanghelich, S. U.S. Patent
5,656,207, 1997.
(13) Law, S.-J.; Jiang, Q.; Fischer, W.; Unger, J. T.; Krodel, E. K.; Xi,
of linear fit constants and additional experimental details. This
material is available free of charge via the Internet at http://
pubs.acs.org.
J. U.S. Patent 5,879,894, 1999.
(14) Natrajan, A.; Sharpe, D.; Costello, J.; Jiang, Q. Anal. Biochem.
2010, 406, 204–213.
’ AUTHOR INFORMATION
(15) (a) Batmanghelich, S.; Woodhead, J. S.; Smith, K.; Weeks, I.
J. Photochem. Photobiol., A 1991, 56, 249–254. (b) Batmanghelich, S.; Brown,
R. C.; Woodhead, J. S.; Smith, K.; Weeks, I. J. Photochem. Photobiol., B 1992,
12, 193–201. (c) Smith, K.; Yang, J.-J.; Weeks, I.; Woodhead, J. S. J.
Photochem. Photobiol., A 2000, 132, 181–191. (d) Smith, K.; Yang, J.-J.; Li,
Z.; Weeks, I.; Woodhead, J. S. J. Photochem. Photobiol., A 2009, 203, 72–79.(e)
Li, Z. Ph.D. Thesis, University of Wales, Swansea, U.K., 1998.
(16) Zhang, H.; Cai, Q.; Ma, D. J. Org. Chem. 2005, 70, 5164–5173.
(17) (a) Hartz, A.; Cuvelier, M.; Nowosielski, K.; Bonilla, T. D.;
Green, M.; Esiobu, N.; McCorquodale, D. S.; Rogerson, A. J. Environ.
Qual. 2008, 37, 898–905. (b) Papadakis, J. A.; Mavridou, A.; Richardson,
S. C.; Lampiri, M.; Marcelou, U. Water Res. 1997, 31, 799–804.
(18) Johnson, S. J.; Green, D. H.; Reed, D. A.; Wood, L. S. Clin.
Chem. 2001, 47, 760–763.
(19) Hogan, J. J.; Smith, R. D.; Kop, J. A.; McDonough, S. H. U.S.
Patent 5,693,468, 1997.
(20) Hogan, J. J. U.S. Patent 6,821,770, 2004.
(21) Lesnik, E. A.; Freier, S. M. Biochemistry 1998, 37, 6991–6997.
(22) Sproat, B. S.; Lamond, A. I.; Beijer, B.; Neuner, P.; Ryder, U.
Nucleic Acids Res. 1989, 17, 3373–3386.
(23) McCapra, F.; Richardson, D. G.; Chang, Y. C. Photochem.
Photobiol. 1965, 4, 1111–1121.
Corresponding Author
Ken.Browne@gen-probe.com
Present Address
^School of Medicine, Cardiff University, Tenovus Building,
Heath Park, Cardiff, Wales CF14 4XN, U.K.
’ ACKNOWLEDGMENT
M. Majlessi, the Oligonucleotide Synthesis Group, A. Maheson,
and T. Luo at Gen-Probe Incorporated have our gratitude for their
syntheses and MS analyses of the nucleic acid constructs presented
in this article. We are also indebted to R. Pacheco for assistance
designing the dual wavelength luminometer and writing custom
data acquisition software. In addition, we thank Molecular Light
Technology Research Limited for supplying 9-(2,6-dibromo-
phenoxycarbonyl)-10-(3-succinimidyloxycarbonylpropyl)acridinium
iodide and 2-(4-dimethylaminophenylazo)-N-[2-(2-iodoacetyl-
amino)ethyl]benzamide. The results of this project are par-
tially based upon work supported by the Air Force Office of
14647
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648

Journal of the American Chemical Society
ARTICLE
(24) Maskiewicz, R.; Sogah, D.; Bruice, T. C. J. Am. Chem. Soc. 1979,
101, 5355–5364.
(25) Arnold, L. J., Jr.; Hammond, P. W.; Wiese, W. A.; Nelson, N. C.
Clin. Chem. 1989, 35, 1588–1594.
(26) Cai, S.; Lau, C.; Lu, J. Anal. Chem. 2010, 82, 7178–7184.
(27) Adamczyk, M.; Moore, J. A.; Shreder, K. Bioorg. Med. Chem.
Lett. 2002, 12, 395–398.
(28) (a) Vogel, A. I. Vogel’s Textbook of Practical Organic Chemistry,
5th ed.; Longman: Harlow, U.K., 1989. (b) Perrin, D. D.; Armarego,
W. L. F. Purification of Laboratory Chemicals, 3rd ed.; Pergamon Press:
Oxford, U.K., 1988.
(29) McNulty, J.; Cheekoori, S.; Bender, T. P.; Coggan, J. A. Eur. J.
Org. Chem. 2007, 1423–1428.
(30) Wolfe, J. P.; Timori, J.; Sadighi, J. P.; Yin, J.; Buchwald, S. L.
J. Org. Chem. 2000, 65, 1158–1174.
(31) (a) Cleveland, W. S. J. Am. Stat. Assoc. 1979, 74, 829–836.
(b) Cleveland, W. S.; Devlin, S. J. J. Am. Stat. Assoc. 1988, 83, 596–610.
14648
dx.doi.org/10.1021/ja202221h |J. Am. Chem. Soc. 2011, 133, 14637–14648