Visible fluorescence chemosensor for saxitoxin

Article abstract of DOI:10.1021/jo062506r

Absorption spectra of a number of shellfish extracts have been obtained and reveal prominent absorptions in all samples at 210 and 260 nm and at 325 nm in some of them. These absorptions preclude the use of chromophores with similar absorptions in testing of shellfish samples for paralytic shellfish toxins. Two crown ether chemosensors featuring a boron azadipyrrin chromophore have been synthesized; both have absorption maxima at 650 nm, where all the shellfish extracts are transparent. The synthetic sensors feature either 18- or 27-membered crown ether rings and have been evaluated as visible sensors for the paralytic shellfish toxin saxitoxin. The binding constant for one of them is in the range of 3-9 × 10⁵ M^-1 and exhibits a fluorescence enhancement of over 100% at 680 nm in the presence of 40 μM saxitoxin.

Full text of DOI:10.1021/jo062506r

Visible Fluorescence Chemosensor for Saxitoxin
Robert E. Gawley,* Hua Mao, M. Mahbubul Haque, John B. Thorne, and Jennifer S. Pharr
Department of Chemistry and Biochemistry, UniVersity of Arkansas, FayetteVille, Arkansas 72701
bgawley@uark.edu
ReceiVed December 6, 2006
Absorption spectra of a number of shellfish extracts have been obtained and reveal prominent absorptions
in all samples at 210 and 260 nm and at 325 nm in some of them. These absorptions preclude the use
of chromophores with similar absorptions in testing of shellfish samples for paralytic shellfish toxins.
Two crown ether chemosensors featuring a boron azadipyrrin chromophore have been synthesized; both
have absorption maxima at 650 nm, where all the shellfish extracts are transparent. The synthetic sensors
feature either 18- or 27-membered crown ether rings and have been evaluated as visible sensors for the
paralytic shellfish toxin saxitoxin. The binding constant for one of them is in the range of 3-9 × 105
M-1 and exhibits a fluorescence enhancement of over 100% at 680 nm in the presence of 40 µM saxitoxin.
Introduction
Contamination of shellfish by paralytic shellfish toxins or
poisons (PSTs, PSPs) continues to be a near-worldwide health
risk.1,2 Monitoring of shellfish beds for the presence of PSTs
continues to use the mouse bioassay as the most common means
of detection,3,4 although recent advances in HPLC methods have
been made and approved for use by monitoring agencies.5 The
mouse bioassay has obvious economic and ethical drawbacks,6
and alternatives are being actively sought in a number of
laboratories.4,7-9
FIGURE 1. General design of a PET-based chemosensor for saxitoxin.
systems that are based on the concept of photoinduced electron
transfer, or PET. In PET-based chemosensors, a host for a given
analyte is designed in which a chromophore is separated from
the host by a spacer (see Figure 1).10-12 In our case, the host
component of the sensor is a crown ether that is separated from
the fluorophore by a methylene spacer. In the absence of a guest
such as saxitoxin, PET from the crown to the fluorophore
quenches the excited state after photon absorption, and fluo-
rescence is turned off (or, more accurately, is minimized). When
the host is complexed to the toxin, the relative energies of the
molecular orbitals are perturbed so that PET is no longer
Our work in this area has focused on optical methods, and
we have developed and explored a number of chemosensor
(1) Tibbets, J. J. EnViron. Health Persp. 1998, 106, A326-A331.
(2) Seafood and Freshwater Toxins. Pharmacology, Physiology, and
Detection; Botana, L. M., Ed.; Marcel Dekker: New York, 2000.
(3) Fernandez, M.; Cembella, A. D. In Manual on Harmful Marine
Microalgae, IOC Manuals and Guides No. 33; Hallegraeff, G. M., Anderson,
D. M., Cembella, A. D., Eds.; UNESCO: Paris, 1995; pp 213-228.
(4) Quilliam, M. A. J. AOAC Int. 1997, 80, 131-136.
(5) Lawrence, J. F.; Niedzwiadek, B.; Menard, C. J. AOAC Int. 2005,
88, 1714-1732.
(6) Gallacher, S.; Mackintosh, F.; Shanks, A. M.; O’Neill, S.; Riddoch,
I.; Howard, F. G. J. Shellfish Res. 1998, 17, 1647-1651.
(7) Sullivan, J. J.; Wekell, M. M.; Hall, S. In Handbook of Natural
Toxins; Tu, A. T., Ed.; Marcel Dekker: New York, 1988; Vol. 3, pp 87-
106.
(8) Beani, L.; Bianchi, C.; Guerrini, F.; Marani, L.; Pistocchi, R.;
Tomasini, M. C.; Ceredi, A.; Milandri, A.; Poletti, R.; Boni, L. Toxicon
2000, 38, 1283-1297.
(9) Jellett, J. F.; Roberts, R. L.; Laycock, M. V.; Quilliam, M. W.; Barrett,
R. E. Toxicon 2002, 40, 1407-1425.
(10) Bissell, R. A.; de Silva, A. P.; Gunaratne, H. Q. N.; Lynch, P. L.
M.; Maguire, G. E. M.; McCoy, C. P.; Sandanayake, K. R. A. S. Top. Curr.
Chem. 1993, 168, 223-264.
(11) Bissell, R. A.; de Silva, A. P.; Gunaratne, H. Q. N.; Lynch, P. L.
M.; Maguire, G. E. M.; Sandanayake, K. R. A. S. Chem. Soc. ReV. 1992,
187-195.
(12) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. ReV. 1997,
97, 1515-1566.
10.1021/jo062506r CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/14/2007
J. Org. Chem. 2007, 72, 2187-2191
2187

Gawley et al.
CHART 1
meat. Prominent absorption peaks for all extracts occur at 210
and 260 nm, with several of the extracts showing absorption
bands at 330 nm, as well. There was no correlation between
the intensity of this band and the toxicity of the sample as
indicated by mouse bioassay data obtained from the Washington
Department of Health (data not shown).
The synthesis of the boron azadipyrrin crowns 1 and 2 is
shown in Scheme 1 and is patterned after the elegant work of
O’Shea.20,21 Monoaza 18-crown-6, 3a, and monoaza 27-crown-
9, 3b,14 were N-alkylated with p-bromomethylbenzaldehyde22
to give crown-substituted aldehydes 4a,b in 96 and 94% yield,
respectively. Condensation with acetophenone afforded chal-
cones 5a,b in 92 and 90%, respectively, and Henry addition of
nitromethane gave 6a,b in equally excellent yields of 96 and
95%. Nef reaction and condensation with ammonia afforded
pyrroles 7a,b in modest yields of 49%. Condensation of 7a,b
with nitrosopyrrole 8, according to the method of O’Shea,20
afforded azadipyrrins 9a,b in good yields (69 and 65%,
respectively), as dark blue oils. Treatment with boron trifluoride
in the presence of base gave the target compounds, 1 and 2, in
yields of 77 and 78%, respectively, as dark green oils.
FIGURE 2. Absorption spectra of 12 extracts of shellfish from mouse
bioassay.
favorable, and fluorescence is “turned on”.13,14 The sensitivity
of a PET-based chemosensor is influenced by a number of
factors, including the quantum yield of the fluorophore and the
equilibrium, or binding constant, Kb, between the chemosensor
and the fluorescent complex.
In past studies, we have evaluated anthracene,13,14 cou-
marin,15,16 and acridine17 fluorophores in our chemosensors, all
of which have absorption maxima in the ultraviolet. This class
of chemosensors is selective for detection of saxitoxin over
tetrodotoxin,17 metal ions such as sodium and potassium,15 and
several other analytes.13 They can also be incorporated into self-
assembled monolayers for sensing on a surface.16,18,19 Recently,
we reported that incorporation of larger crown ether rings can
significantly increase the binding constant.14 We now report the
synthesis and evaluation of boron azadipyrrins 1 and 2, having
absorption maxima in the visible region of the spectrum, and
with both 18-crown-6 (1) and 27-crown-9 (2) ring hosts.
The absorption and emission spectra for 1 and 2 are shown
in Figure 3. As expected, both are very similar. The binding
isotherms were determined by titration with saxitoxin in
methanol, buffered to pH 7.1 with tetrabutylammonium phos-
phate. The pH of a nonaqueous buffer is a complicated
issue.23-25 In this case, pH 7.1 refers to a meter reading of 7.1
(15) Kele, P.; Orbulescu, J.; Calhoun, T. L.; Gawley, R. E.; Leblanc, R.
M. Tetrahedron Lett. 2002, 43, 4413-4416.
(16) Kele, P.; Orbulescu, J.; Gawley, R. E.; Leblanc, R. M. Chem.
Commun. 2006, 1494-1496.
(17) Gawley, R. E.; Shanmugasundaram, M.; Thorne, J. B.; Tarkka, R.
M. Toxicon 2005, 45, 783-787; Corrigendum 46, 477.
(18) Wang, S.; Zhang, Q.; Datta, P. K.; Gawley, R. E.; Leblanc, R. M.
Langmuir 2000, 16, 4607-4612.
Results
The absorption spectra of extracts of 12 shellfish, collected
in Washington in July and August of 2004, are shown in Figure
2. The data are from extracts of blue mussels (Mytilus trossulus),
geoduck clams (Panope generosa), butter clams (Saxidomus
giganteus), and littleneck clams (Protothaca staminea). The PSP
number for these samples ranged from <38 to 2484 µg/100 g
(19) Kele, P.; Orbulescu, J.; Calhoun, T. L.; Gawley, R. E.; Leblanc, R.
M. Langmuir 2002, 18, 8523-8526.
(20) Hall, M. J.; McDonnell, S. O.; Killoran, J.; O’Shea, D. F. J. Org.
Chem. 2005, 70, 5571-5578.
(21) Hall, M. J.; Allen, L. T.; O’Shea, D. F. Org. Biomol. Chem. 2006,
4, 776-780.
(22) Wen, L.; Li, M.; Schlenoff, J. B. J. Am. Chem. Soc. 1997, 119,
7726-7733.
(23) Rondinini, S.; Mussini, P. R.; Mussini, T.; Vertova, A. Pure Appl.
Chem. 1998, 70, 1419-1422.
(13) Gawley, R. E.; Pinet, S.; Cardona, C. M.; Datta, P. K.; Ren, T.;
Guida, W. C.; Nydick, J.; Leblanc, R. M. J. Am. Chem. Soc. 2002, 124,
13448-13453.
(14) Mao, H.; Thorne, J. B.; Pharr, J. S.; Gawley, R. E. Can. J. Chem.
2006, 1273-1279.
(24) Tindall, G. W. LCGC North Am. 2002, 20, 1028-1032 (www.chro-
matographyonline.com).
(25) Tindall, G. W. LCGC North Am. 2002, 20, 1114-1118 (www.chro-
matographyonline.com).
2188 J. Org. Chem., Vol. 72, No. 6, 2007

Visible Fluorescence Chemosensor for Saxitoxin
FIGURE 3. Excitation and emission spectra of boron azadipyrrins 1 (left) and 2, in methanol; λex ) 652 and 681 nm; λem ) 650 nm.
FIGURE 4. Binding isotherms for titration of saxitoxin against fluorescence response for 1 (left) and 2 (right), in methanol. The curve fit assumes
a 1:1 binding stoichiometry. For 1, Kb ) 8.9 × 105 M-1; for 2, Kb ) 1.4 × 104 M-1
.
SCHEME 1. Synthesis of Boron Azadipyrrin Chemosensors
in methanol, after calibration of the meter in aqueous buffer. In
methanol, the pKa for 1 and 2 is approximately 6.5. The
concentration of the crown chemosensors was held constant at
40 µM; typical binding isotherms are shown in Figure 4.
Nonlinear least-squares analysis,26 assuming a 1:1 host/guest
stoichiometry, reveals binding constants of 6.2 × 105 M-1 for
J. Org. Chem, Vol. 72, No. 6, 2007 2189

Gawley et al.
1 (average of three runs), and Kb ) 1.4 × 104 M-1 for 2 (best
the contrary, the binding constant fell to a disappointing 1.4 ×
104 M-1. Moreover, 2 exhibited properties that we have not
seen before. For example, fluorescence emission of 40 µM 2
in buffered methanol or in pure methanol (in the absence of
toxin) decreased over a period of 3-10 min, suggesting an
(unknown) equilibration. The fluorescence enhancement was
minimal at best, and sometimes was negligible. The reasons
for this difference in behavior are unknown.
run; see Discussion).
Discussion
The chromophores we have incorporated into the sensor
design in the past all absorb in the ultraviolet: coumarin at 328
nm, acridine at 350, and anthracene with three bands from 360
to 390 nm. The 330 nm UV absorption band of the shellfish
extracts excludes coumarin as a viable chromophore for evaluat-
ing PSPs in shellfish. The tailing of this band to 350 nm in
several samples also suggests that acridine is not ideal either.
Anthracene has three prominent absorption bands, one of which
is at 390 nm, and is in a region of little absorption by the
shellfish matrix. Nevertheless, we decided to investigate a
chromophore with absorption in the visible, far from any
absorption bands due to the matrix. With this in mind, we were
intrigued by the recent reports from the O’Shea group, in which
boron azadipyrrins were incorporated into PET sensors (for pH
sensing) having absorption bands in the region of 650 nm.21,27
Particularly appealing was the recent report of a general
synthesis that permits preparation of a variety of unsymmetrical
azadipyrrins.20 Chemosensors 1 and 2 incorporate the boron
azadipyrrin as the chromophore, with crown ethers as the host
moiety. Previously, we have found that larger crown rings
produce larger binding constants; with anthracene as the
fluorophore, the binding constant for the 18-crown-6 host was
5.3 × 104 M-1; for the 27-crown-9 host, it was 1.69 × 105
We have previously suggested, based on molecular modeling
studies, that the PET mechanism in these systems involves
π-stacking of one of saxitoxin’s guanidiniums to the fluorophore,
and that this π-stacking perturbs the relative energies of the
pertinent molecular orbitals to “turn off” photoinduced electron
transfer. π-Stacking of guanidiniums to arenes is well-known
in DNA-binding proteins,28-30 in proteins engineered to bind
guanidinium ion,31 and even in binding of arginine to C60, with
the arginine sandwiched between the buckyball and a tryp-
tophan.32 Although the precise orientation of saxitoxin’s guani-
dinium to the arene is not known, it should be similar among
coumarin, acridine, and anthracene chromophores we have
studied in the past. In the present case, the fluorophore is much
larger, and features a zwitterionic boron-nitrogen dative bond.
In such systems, although the negative charge is formally on
the boron and the positive charge on the nitrogen, the negative
charge density is mostly on the fluorines, while the positive
charge is delocalized by resonance. One can easily imagine that
having such a functionality proximal to the binding site could
alter the geometry of the π-stacking of the toxin to the
fluorophore, thereby affecting the ability of the toxin to inhibit
the PET. We speculate that the larger crown, being able to offer
more hydrogen-bonding acceptor heteroatoms, holds the toxin
more firmly in an orientation that precludes the guanidinium
π-stack that we believe is necessary to turn off PET.
M-1 14
. We decided to prepare boron azadipyrrin chemosensors
with both 18-crown-6 and 27-crown-9 rings.
The syntheses went smoothly in six steps, as illustrated in
Scheme 1, with an overall yield of 22% for 1 and 20% for 2.
The lowest-yielding steps were the conversion of nitroketones
6a,b to pyrroles 7a,b and subsequently to dipyrrins 9a,b. Our
observed yields for these steps are somewhat lower than those
O’Shea observed in similar reactions,20 but since adequate
amounts were obtained for evaluation, we did not attempt further
optimization.
The optical properties of 1 and 2 were as expected (Figure
3). The fluorescence response of these chromophores is not as
large as we have seen with coumarin and anthracene chro-
mophores, possibly due to the instrumentation used. In order
to see a good fluorescence response, it was necessary to employ
40 µM concentrations of sensors 1 and 2 in the titrations. At
this concentration, 1 showed excellent sensitivity to very low
concentrations of saxitoxin. At ∼1:1 toxin/crown stoichiometry,
the fluorescence enhancement, F/F0, is well over 100%. This
level of fluorescence enhancement is reproducible in methanol,
buffered at pH 7.1. The legal limit of saxitoxin equivalents in
shellfish extracts is 80 µg/100 g meat, but the detection limit
in the mouse bioassay is about half that. This detection limit in
the mouse bioassay corresponds to an approximately 1 µM
solution of toxin.
In summary, we have shown that boron azadipyrrins, func-
tionalized with an 18-crown-6 ring, make excellent visible
chemosensors for the marine toxin saxitoxin, the most toxic
component of paralytic shellfish poisons. Analysis of shellfish
extracts for individual paralytic shellfish toxins (saxitoxin and
its 20+ analogues) and the fluorescence response to each by
our chemosensors is underway and will be reported in due
course.
Experimental Section
Extracts of four species of shellfish used in routine mouse
bioassays were obtained courtesy of Bob Lona, Department of
Health, Seattle, Washington. Saxitoxin was obtained from Sherwood
Hall, U.S. Food and Drug Administration. Details of the chemical
syntheses of the boron azadipyrrin crowns, including characteriza-
tion data and copies of the NMR spectra for all intermediates, are
found in the Supporting Information. Titrations and binding constant
calculations were done as described previously,13 with the excitation
and emission slits set at 10 and 15 nm, respectively. Methanol is
buffered with tetrabutylammonium phosphate to pH 7.1 (phosphate
concentration ) 300 mM) by adding 600 µL of 5 M phosphoric
The average binding constant for 1 to saxitoxin, 6.2 × 105
M-1, is among the highest we have observed for any of our
chemosensors. Since changing from an 18-crown-6 ring to a
27-crown-9 ring improved the binding with an anthracene
fluorophore,14 we anticipated a similar increase with 2. As
shown in Figure 4, a similar enhancement was not observed; to
(28) Gromiha, M. M.; Santhosh, C.; Ahmad, S. Int. J. Biol. Macromol.
2004, 34, 203-211.
(29) Sathyapriya, R.; Vishveshwara, S. Nucleic Acids Res. 2004, 32,
4109-4118.
(30) Biot, C.; Wintjens, R.; Rooman, M. J. Am. Chem. Soc. 2004, 126,
6220-6221.
(26) Connors, K. A. In Binding Constants. The Measurement of Molec-
ular Complex Stability; Wiley-Interscience: New York, 1987; pp 339-
343.
(31) Baldwin, E.; Baase, W. A.; Zhang, X.; Feher, V.; Matthews, B. W.
J. Mol. Biol. 1998, 277, 467-485.
(32) Noon, W. H.; Kong, Y.; Ma, J. Proc. Natl. Acad. Sci. U.S.A. 2002,
99 (Suppl. 2), 6466-6470.
(27) Killoran, J.; O’Shea, D. Chem. Commun. 2006, 1503-1505.
2190 J. Org. Chem., Vol. 72, No. 6, 2007

Visible Fluorescence Chemosensor for Saxitoxin
acid to ∼7 mL of methanol, and raising the pH to 7.1 (meter reading
on a pH meter calibrated with aqueous buffer) with Bu4NOH (55-
60% in water). The crown sensor 1 or 2 is added such that the
final concentration will be 40 µM (e.g., 140 µL of a 2.9 mM solution
in methanol) and the solution diluted to 10 mL with methanol.
UV-Visible absorption spectra of shellfish extracts were ob-
tained as follows: 100 µL of shellfish extract was diluted with 10
mL of pH 7.1 phosphate buffer (1:100); buffer solutions were then
filtered using a 0.45 µM PTFE syringe filter and scanned on a UV-
vis spectrophotometer.
Center for Research Resources (P20 R15569), and the Arkansas
Biosciences Institute. We thank Bob Lona of the Washington
Department of Health for sharing shellfish extracts and for help
with taxonomical assignments, and Sherwood Hall of the U.S.
Food and Drug Administration for gifts of saxitoxin. Finally,
we are grateful to Professor Donal O’Shea for a helpful
discussion and for sharing unpublished work.
Supporting Information Available: Details of the synthesis
procedures and characterization data; proton, carbon, and DEPT-
135 NMR spectra of all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. This work was supported by the Envi-
ronmental Protection Agency (GR-83238201). Core facilities
were provided by the National Institutes of Health, National
JO062506R
J. Org. Chem, Vol. 72, No. 6, 2007 2191