A precipitator for the detection of thiophilic metals in aqua

Article abstract of DOI:10.1021/ja045798r

The detection of toxic metals including mercury and lead has become a vital analytical tool for environmental remediation and regulation of food stocks. A prevalent obstacle with the current assessment of metal ion contamination originates from the lack of adequate assay throughput. In this context, a critical concern with current analyses stems from the fact that the majority of these assays are solution-based, and thus the response is highly dependent upon the assay environment. Herein, we describe a fluorescent dye-doped crystalline assay that offers convincing metal selection and provides detection comparable to conventional solution-based ligands used for the spectrofluorometric analysis of thiophilic heavy metal ions. While comparable in analytical performance to known methodologies, the formation of crystalline analytes provides for signal amplification and, consequently, a powerful platform whose analysis is directly amenable to high-throughput video capture systems. This procedure has been tested in a variety of scenarios and shows good performance using readily available equipment, including a commercially available Universal Serial Bus (USB) CCD camera. Furthermore, when developed in a microcapillary format, this assay is capable of screening thousands of samples per day for the presence of subnanomolar concentrations of Hg²⁺ using a conventional fluorescence microscope.

Full text of DOI:10.1021/ja045798r

Published on Web 11/23/2004
A Precipitator for the Detection of Thiophilic Metals in Aqua
Tobin J. Dickerson,† Neal N. Reed,† James J. LaClair,*,‡ and Kim D. Janda*,†
Contribution from the Department of Chemistry and The Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, and
Xenobe Research Institute, P.O. Box 4073, San Diego, California 92164-4073
Received July 13, 2004; Revised Manuscript Received September 29, 2004; E-mail: kdjanda@scripps.edu; i@xenobe.org
Abstract: The detection of toxic metals including mercury and lead has become a vital analytical tool for
environmental remediation and regulation of food stocks. A prevalent obstacle with the current assessment
of metal ion contamination originates from the lack of adequate assay throughput. In this context, a critical
concern with current analyses stems from the fact that the majority of these assays are solution-based,
and thus the response is highly dependent upon the assay environment. Herein, we describe a fluorescent
dye-doped crystalline assay that offers convincing metal selection and provides detection comparable to
conventional solution-based ligands used for the spectrofluorometric analysis of thiophilic heavy metal ions.
While comparable in analytical performance to known methodologies, the formation of crystalline analytes
provides for signal amplification and, consequently, a powerful platform whose analysis is directly amenable
to high-throughput video capture systems. This procedure has been tested in a variety of scenarios and
shows good performance using readily available equipment, including a commercially available Universal
Serial Bus (USB) CCD camera. Furthermore, when developed in a microcapillary format, this assay is
capable of screening thousands of samples per day for the presence of subnanomolar concentrations of
Hg2+ using a conventional fluorescence microscope.
Introduction
correlated metal ion sensing with material synthesis (i.e.,
crystallization, precipitation, or polymerization).
The introduction of heavy metals into the environment has
occurred in tandem with the emergence of the tools of modern
society, including automobiles, industrial activities, and the
burning of coal.1 Consequently, the detection of these metals
has attained significant prominence within the scientific com-
munity and the public consciousness as a result of their varied
deleterious biological activities. These concerns have been
furthered by the discovery that aquatic organisms convert
elemental mercury to methylmercury, which subsequently
concentrates through the food chain in the tissues of fish and
marine mammals.2 Unquestionably, the development of facile
and efficient methodologies for the detection of heavy metals
in a wide range of settings is paramount in the effort to minimize
incidences of mortality attributable to heavy metal toxicity.
A plethora of indicators are known to detect metal ions
through the intramolecular modulation of a pendant chro-
mophore.3 While effective for toxic metal analysis in laboratory
environments,4 few of these probes have proven suitable for
screening within modern industrial settings.5 While the design
of colorimetric ligation has been critically evaluated by Lippard,
Haugland, and others,3,6 the predominant theme within metal
ion sensing relies on solution-based chemistry. Few efforts have
Recent advances in the study of colorimetric ligands suggest
that a viable metal indicator offers (1) the appropriate direction
of metal-mediated modulation, (2) a high degree of sensitivity,
and (3) a defined metal selectivity. We have demonstrated that
correlation of ligation-induced colorimetric response with
precipitation provides an effective vector for metal ion analysis.3b
Here, amplification of a colorimetric response through precipita-
tion provided a practical screen for mercuric ion. We now report
that the appendage of a colorimetric moiety is unnecessary for
the analysis of thiophilic metals and adapt this to provide a
digital screen for the analysis of Hg2+, Pb2+, and Cd2+
.
Results and Discussion
The construct began with the synthesis of dithiopthalides 3.7
As shown in Figure 1, 3a and 3b were prepared from phthalic
(3) (a) Choi, M. J.; Kim, M. Y.; Chang, S.-K. Chem. Commun. 2001, 1664-
1665. (b) Bru¨mmer, O.; La Clair, J. J.; Janda, K. D. Org. Lett. 1999, 1,
415-418. (c) Sanceno´n, F.; Mart´ınez-Ma´n˜ez, R.; Soto, J. Chem. Commun.
2001, 2262-2263. (d) Sanceno´n, F.; Mart´ınez-Ma´n˜ez, R.; Soto, J.
Tetrahedron Lett. 2001, 42, 4321-4323. (e) Moon, S. Y.; Cha, N. R.; Kim,
Y. H.; Chang, S. K. J. Org. Chem. 2004, 69, 181-183. (f) Palomares, E.;
Vilar, R.; Durrant, J. R. Chem. Commun. 2004, 362-363. (g) Kuhn, M.
A.; Hoyland, B.; Carter, S.; Zhang, C.; Haugland, R. P. Proc. SPEI, Intl.
Soc. Opt. Eng. 1995, 2388, 238-242.
(4) (a) Baeumner, A. J. Anal. Bioanal. Chem. 2003, 377, 434-445. (b) Epstein,
J. R.; Walt, D. R. Chem. Soc. ReV. 2003, 32, 203-214. (c) Jain, K. K.
Med. DeVice Technol. 2003, 14, 10-15. (d) Ulber, R.; Frerichs, J. G.;
Beutel, S. Anal. Bioanal. Chem. 2003, 376, 342-348.
(5) Roblin, P.; Barrow, D. A. J. EnViron. Monit. 2000, 2, 385-92.
(6) (a) Descalzo, A. B.; Mart´ınez-Ma´n˜ez, R.; Radeglia, R.; Rurack, K.; Soto,
J. J. Am. Chem. Soc. 2003, 125, 3418-3419. (b) Nolan, E. M.; Lippard, S.
J. J. Am. Chem. Soc. 2003, 125, 14270-14271. (c) Xu, X.; Thundat, T.
G.; Brown, G. M.; Ji, H. F. Anal. Chem. 2002, 74, 3611-3615. (d) Zhang,
X. B.; Guo, C. C.; Li, Z. Z.; Shen, G. L.; Yu, R. Q. Anal. Chem. 2002, 74,
821-825.
† Department of Chemistry and The Skaggs Institute for Chemical
Biology, The Scripps Research Institute.
‡ Xenobe Research Institute.
(1) (a) Shotyk, W.; Weiss, D.; Appleby, P. G.; Cheburkin, A. K.; Frei, R.;
Gloor, M.; Kramers, J. D.; Reese, S.; Vanderknaap, W. O. Science 1998,
281, 1635-1640. (b) Benoit, J. M.; Fitzgerald, W. F.; Damman, A. W.
EnViron. Res. 1998, 78, 118-133.
(2) Nendza, M.; Herbst, T.; Kussatz, C.; Gies, A. Chemosphere 1997, 35,
1875-1885.
9
16582
J. AM. CHEM. SOC. 2004, 126, 16582-16586
10.1021/ja045798r CCC: $27.50 © 2004 American Chemical Society

Detection of Thiophilic Metals in Aqua
A R T I C L E S
Figure 1. Synthesis of thiophilic ligands 3 and comparison of their structures to established indicators 5-7. (a) i, ArMgBr, C6H6, 80 °C, 24 h; ii, 2 M HCl;
iii, N2H4‚H2O, EtOH, 78 °C, 24 h. (b) P4S10, xylene, 140 °C, 18 h. (c) Mn+(OAc)n or Mn+Cln, H2O, CH3CN, <1 min.
anhydride (1) in two operations.8 Initial addition to the anhydride
by phenylmagnesium bromide was followed by addition of a
second equivalent of phenylmagnesium bromide and subsequent
intramolecular ester formation to give lactone 2a. Compound
2a could then be smoothly converted to xanthane 3a by
treatment with P4S10 in refluxing xylenes. Importantly, the entire
route to 3a could be executed at the kilogram scale without
chromatographic purification.
When presented to a panel of metal ions, 3a complexed and
precipitated Hg2+, Pb2+, Cd2+, Au3+, Cu2+, Pd2+, Ni2+, Co2+
,
Mo2+, and Pt2+. Job plots indicated that precipitates 4a (M )
Hg2+, Pb2+, Cd2+) existed as a 1:1 complex. A 2:1 ligand to
metal complex could also be formed when precipitating from
solutions containing >5 mM 3a. In contrast, the precipitation
of 3b by Hg2+ and Pb2+ required over 5 min for induction, and
other metals including Cd2+ failed to provide sufficient yields
of precipitate. Consequently, we deemed that the solubility of
3b and its metal complexes do not provide a response adequate
for analytical use.
Solubility products were determined using conventional mass
analysis to characterize the metal ion selection of 3a (Figure
2A). Alternatively, the precipitation of 4a could be determined
by measuring the spectrophotometric loss of 3a in the super-
natant at λmax ) 304, 330, or 356 nm (Figure 2B). Quantitative
analysis with 3a was comparable in accuracy and precision to
assays developed with conventional ligands, including diphen-
ylcarbazone 5, Fluo-5N 6, and Rhod-5N 7.9 The precipitation
of 3a in 10% aqueous CH3CN was visually apparent upon
addition of 2 µM Hg(OAc)2 to 1 µM 3a, 50 µM CdCl2 to 25
µM 3a, and 10 µM Pb(OAc)2 to 5 µM 3a. The precipitation of
4a with Hg2+, Pb2+, or Cd2+ was tolerant of alkali metals (Na+,
K+), alkaline earth metals (Ca2+, Mg2+), and transition metals
(Mn2+, Mo2+, Cr2+, and Fe2+), as given by the modest change
Figure 2. Solubility products (Ksp) of metal complexes of 3a as determined
by (A) the mass of precipitate 4a, (B) spectroscopic loss of 3a, (C) the
mass of 4a obtained in the presence of 105 molar equiv of Li+, Na+, K+,
Cs+, Mg2+, Ca2+, Ba2+, V2+, Cr2+, Mn2+, Fe2+, Co2+, Rh3+, Zn2+, Al3+
,
Sn2+, As3+, Sb3+, Bi3+, La3+, Ce3+, Sm3+, Eu3+, and Yb3+. The solubility
product could also be determined using a digital displacement map analysis
as given by imaging (D) with a CCD microscope (Intel Digital Blue), (E)
on an inverted microscope (Nikon Eclipse TE300), (F) with fluorescence
crystals of 4a (Figure 3E), and (G) with fluorescent crystals of 4a grown
in capillaries (Figure 3G).
in Ksp when complexes of 4a were formed in the presence of
105 molar excesses of nonbinding metal ions (Figure 2C). This
selection permitted the analysis of Hg2+, Pb2+, and Cd2+ in
mixtures containing 500 equiv of Cu2+ in contrast to established
indicators for Hg2+ such as Rhod-5N 6, which typically offer
only a 3-5-fold selection for Hg2+ over Cu2+
.
(7) (a) Szurdoki, F.; Kido, H.; Hammock, B. D. Bioconjugate Chem. 1995, 6,
145-149. (b) Lo, J.-M.; Lee, J.-D. Anal. Chem. 1994, 66, 1242-1248. (c)
Sachsenberg, S.; Klenke, T.; Krumbein, W. E.; Zeeck, E. J. Fresenius Anal.
Chem. 1992, 342, 163-166. (d) Bond, A. M.; Scholz, F. J. Phys. Chem.
1991, 95, 7460-7465.
Metal selectivity is not the only criteria required for a practical
screen. Many indicators, including 5-7, are sensitive to pH,
ion strength, impurities, buffers, and solvents. Deviations in
these environmental factors can alter the kinetics of ligand
association or the photophysical properties of the appended
reporter. These complications are furthered by fact that the
concentration profile of many solution-based colorimetric
ligands remains nonlinear. For instance, the affinity of Fluo-
(8) (a) Nugara, P. N.; Huang, N. Z.; Lakshmikantham, M. V.; Cava, M. P.
Heterocycles 1991, 32, 1559-1561. (b) Oparin, D. A.; Kuznetsova, A. S.
Vestsi Akademii NaVuk BSSR, Seryya Khim. NaVuk 1990, 6, 109-110.
(9) Indicators 5-7 displayed high affinities (Kd ) 10-9-10-12 M-1) to Hg2+
,
Pd2+, and Co2+ and permitted detection of micromolar levels of these metal
ions in the presence of 103 equiv of nonbinding metals. See: Haugland, R.
Handbook of Fluorescent Probes and Research Products, 9th ed.; Molecular
Probes: Eugene, OR, 2001; Section 20.
9
J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004 16583

A R T I C L E S
Dickerson et al.
Figure 3. Microscopic analysis. (A) The precipitate generated by mixing 2 µM Hg(OAc)2 with 1 µM 3a in 10% aqueous CH3CN, (B) an image (inset) and
the Delaunay triangulation map of 100 µm microspheres, (C) an image (inset) and Delaunay map of precipitates 4a (M ) Hg), (D) fluorescent microcrystals
4a generated by the mixing of 1 mM Hg(OAc)2, 1 mM 3a, and 0.5 µM Rhod-5N in 10% aqueous CH3CN, (E) globular crystals formed from 20 mM Hg2+
,
20 mM 3a, and 0.5 µM Rhod-5N, (F) a single fluorescent crystal of 4a (M ) Hg) generated in capillaries by placing a 1 ( 0.05 mm long section of 10 µm
i.d. fused silica capillary containing 200 µM 3a and 2 µM rhodamine B in CH3CN into a 100 µL aliquot of 0.2 nM Hg(OAc)2. Images were collected on
a Nikon Eclipse TE300 using Y-2E/C (560BP40 excitation and 595 LP 630/60 BP emission) filter. Comparable images were also generated when precipitating
with equimolar amounts of Pb(OAc)2, Cd(OAc)2, HgCl2, PbCl2, and CdCl2.
5N 6 to Cd2+ is 10-fold larger than La3+ at 1 µM, while the
affinity for the complexes changes to favor La3+ by 2-fold over
Cd2+ at 100 µM.3g,9 The fact that Ksp values (Figure 2A-C)
were reproducible within 1% deviation over a wide range of
concentrations (1 mM to 0.1 µM), pH (3-10), and temperature
(0-50 °C) indicated that the precipitation of 3a offered increased
resistance to environmental factors relative to comparable
solution-based methodologies. This does not imply, however,
that a precipitation-based assay can supplant existing solution
technologies for heavy metal detection. Current fluorescent
probes that operate in solution are directly amenable to in situ
analysis and thus can assay samples that cannot be removed
from their natural environment. In contrast, our precipitation-
based methodology is very resistant to signal degradation by
environmental factors (vide supra), yet requires extraction of
the sample from its environment for testing.
Rhod5N or rhodamine B during precipitation to yield needles
(Figure 3D) or globular crystals (Figure 3E).
A digital displacement map method was developed to
determine the mass of precipitate within each image. A
Delaunay triangulation10 was used to transpose each image
(Figure 3) into a 3D vector map.11 This process provided a net
volume of precipitate generated per image using vector analysis.
This volume deviated 3.9 ( 1.5 and 6.1 ( 2.2% when imaging
standard 100 µm microspheres (Figure 3B), on an inverted
microscope (Nikon Eclipse TE300) or inexpensive CCD mi-
croscope (Intel Digital Blue), respectively. Once calibrated with
microspheres, the volumes of precipitates were quickly deter-
mined from images of precipitates 4a (Figure 3C inset) and their
corresponding 3D maps (Figure 3C). Using the conditions
described in Figure 3A, the CCD microscope (Figure 2D) and
inverted microscope (Figure 2E) provided Ksp values that were
comparable to that obtained by conventional assays (Figure
2A,B).12
The formation of metal precipitates was rapidly screened
using high-throughput video-capture techniques. When con-
ducted on a glass slide or in a fused-silica capillary, crystals 4a
(M ) Hg2+, Pb2+, or Cd2+) associated on the surface of a glass
slide or wall of a capillary within a few seconds after formation
(Figure 3). Elemental analysis of the resulting supernatant
indicated that the mercuric ion was quantitatively removed from
the solution during this process. Indeed, vapor atomic absorption
measurements demonstrated that the amount of Hg2+ (0.58 (
0.03 µM) after reacting 6.7 ( 0.01 mM Hg(OAc)2 with an 6.7
( 0.01 mM 3a was well below the error threshold of the
pipettors used to prepare the reaction ((0.1 mL or (4 µM).
To improve visualization, the crystals were doped with a
fluorescent dye, as given by the addition of 10-3 equiv of
The method was capable of detecting part per billion levels
of thiophilic metals when examined in small volume elements.
As shown in Figure 3F, single crystals of 4a were reproducibly
generated upon the addition of microcapillaries filled with 3a
into aqueous solutions of metal ion. Displacement map analysis
(10) (a) Wohlberg, B.; de Jager, G. IEEE T Image Process. 1999, 8, 1716-
1729. (b) Lohner, R. Finite Elem. Anal. Des. 1997, 25, 111-134.
(11) Zhu, W.; Wang, Y.; Yao, Y.; Chang, J.; Graber, H. L.; Barbour, R. L. J.
Opt. Soc. Am. A 1999, 14, 799-802.
(12) Ksp values were calculated on the basis of comparison of the measured
amount of precipitate generated with the known aliquot of metal ion. The
amount of precipitate generated was calculated from the volume of
precipitate using a density of 7.3, 3.8, and 4.3 g/mL, for the precipitate
generated by the addition of 3a to Hg(OAc)2, Pb(OAc)2, and CdCl2,
respectively. An average of 20 repetitions was provided.
9
16584 J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004

Detection of Thiophilic Metals in Aqua
A R T I C L E S
C20H14S2: 319.0610. Found: 319.0609. UV/vis (CH3CN): λmax (ꢀ) )
509 (1850), 340 (5600), 220 nm (13000).
indicated that the crystal in Figure 3F contained 16.5 ( 2.9
fmol of 4a. Assuming a 1:1 complex, this finding represented
the detection of 0.17 ( 0.3 nM Hg2+ (0.3 ppb) and indicated
an 82% yield of 4a upon exposure to a 100 µL aliquot of 0.2
nM Hg2+. Using the procedure in Figure 3F, comparable
precipitates were obtained when exposed to solutions that
contained greater than 0.2 nM (0.4 ppb) Hg2+, 1.5 nM (0.31
3,3-Bis(4-decylphenyl)isobenzofuran-1(3H)-one (2b). A solution
of 4-decylphenylmagnesium bromide (∼1 M in THF) was prepared
from magnesium (90 mg, 3.7 mmol) and 4-(decylphenyl)bromide (1
g, 3.4 mmol). This Grignard reagent was added to a solution of phthalic
anhydride (201 mg, 1.36 mmol) in toluene (10 mL) and the solution
heated to reflux for 24 h. The dark red solution was then cooled to
room temperature and quenched with 20% aqueous HCl (10 mL). The
organic phase was separated, washed with water (2 × 10 mL), saturated
NaCl (10 mL), and concentrated. The resulting residue was then
dissolved in EtOH (10 mL), and hydrazine hydrate (1 mL) was added
and heated to reflux for 24 h. The reaction was then cooled to room
temperature, concentrated in vacuo, and purified by radial chromatog-
raphy (5:95 EtOAc:hexane) to give 103 mg (14%) of the desired lactone
ppb) Pb2+, or 2.5 nM (0.28 ppb) Cd2+
.
Conclusion
In summary, we have developed an assay for thiophilic heavy
metals that uses precipitation to decrease interference and
increase detection. Analysis in droplets or capillaries provides
an effective tool for determining the solubility product of metal
complexes using femtomoles of ligand 3a. This assay was
conducted with common imaging systems. This finding suggests
that the combination of ligand synthesis, crystal engineering,
and fluorescent imaging can provide an information-rich
platform for toxic metal analyses.
1
2b. H NMR (CDCl3, 500 MHz): δ 7.93 (dd, J ) 0.75 Hz, 7.7 Hz,
1H), 7.67 (dt, J ) 1.1 Hz, 7.7 Hz, 1H), 7.55 (m, 2H), 7.23 (d, J ) 8.2
Hz, 4H), 7.12 (d, J ) 8.2 Hz, 4H), 2.58 (t, J ) 7.5 Hz, 4H), 1.59 (m,
5H), 1.29 (m, 32H), 0.88 (t, J ) 7 Hz, 6H). 13C NMR (CDCl3, 125
MHz): δ 170.4, 152.9, 143.8, 138.6, 134.5, 129.6, 128.8, 127.5, 126.4,
126.1, 124.6, 92.3, 36.0, 32.4, 31.8, 30.1, 30.0, 29.9, 29.8, 23.2, 14.6.
MALDI-FTMS (M + H+). Calcd for C40H54O2: 567.4196. Found:
567.4184.
Experimental Section
General Methods. Unless otherwise stated, all reactions were
performed under an inert atmosphere with dry reagents and solvents
and flame-dried glassware. Analytical thin-layer chromatography (TLC)
was performed using 0.25 mm precoated silica gel Kieselgel 60 F254
plates. Visualization of the chromatogram was by UV absorbance,
iodine, dinitrophenylhydrazine, ceric ammonium molybdate, ninhydrin,
or potassium permanganate as appropriate. Preparative and semi-
preparative TLC was performed using Merck 1 mm or 0.5 mm coated
silica gel Kieselgel 60 F254 plates, respectively. Methylene chloride and
chloroform were distilled from calcium hydride. Tetrahydrofuran (THF)
was distilled from sodium/benzophenone. Methanol was distilled from
magenesium. 1H and 13C NMR spectra were recorded on a Varian
INOVA-399 spectrometer at 400 and 100 MHz, respectively, and are
reported in parts per million, unless otherwise noted. All spectra were
processed with 0.5 Hz line broadening. Matrix-assisted laser desorption/
ionization (MALDI) FTMS experiments are performed on an IonSpec
FTMS mass spectrometer. Electrospray ionization (ESI) mass spec-
trometry experiments were performed on an API 100 Perkin-Elmer
SCIEX single quadrupole mass spectrometer.
3,3-Diphenylisobenzofuran-1(3H)-one (2a). To a solution of ph-
thalic anhydride (5 g, 33.8 mmol) in benzene (100 mL), phenylmag-
nesium bromide (84.5 mmol) was added slowly and the solution heated
to reflux for 24 h. The reaction was then cooled and 2 M HCl (100
mL) added slowly. The organic phase was separated, washed with water
(3 × 20 mL), dried on MgSO4, and concentrated. The resulting residue
was then dissolved in EtOH (75 mL), and hydrazine hydrate (3 mL)
was added and heated to reflux for 24 h. The solution was then cooled
to 4 °C, and the resulting yellow crystals were collected and dried in
vacuo to give 3.7 g (38%) of the desired lactone 2a. 1H NMR (CDCl3,
300 MHz): δ 7.94 (d, J ) 7.5 Hz, 1H), 7.69 (t, J ) 7.2 Hz, 1H), 7.58
(d, J ) 7.5 Hz, 1H), 7.53 (d, J ) 7.2 Hz, 1H), 7.33 (m, 10H). 13C
NMR (CDCl3, 75 MHz): δ 169.8, 152.0, 141.0, 134.3, 129.5, 128.7,
128.6, 127.3, 126.2, 125.7, 124.4, 92.0. MALDI-FTMS (M + H+).
Calcd for C20H14O2: 287.1067. Found: 287.1058.
3,3-Bis(4-decylphenyl)benzo[c]thiophene-1(3H)-thione (3b). Lac-
tone 2b (47 mg, 83.0 µmol) was dissolved in xylene (2 mL). To this
solution, P4S10 (30 mg, 67.6 µmol) was added and the reaction heated
to reflux for 18 h. The solution was then cooled, filtered, and
concentrated to give the desired product 3b as a red oil. This was further
purified by chromatography on silica using hexane as the eluent to
give 22 mg (44%). 1H NMR (CDCl3, 500 MHz): δ 8.08 (dd, J ) 0.75
Hz, 8 Hz, 1H), 7.59 (dt, J ) 0.5 Hz, 8 Hz, 1H), 7.48 (dt, J ) 0.5 Hz,
8 Hz, 1H), 7.27 (d, J ) 8.8 Hz, 1H), 7.20 (d, J ) 8.5 Hz, 4H), 7.11 (d,
J ) 8.5 H z, 4H), 2.58 (t, J ) 7.7 Hz, 4H), 1.59 (m, 4H), 1.28 (m,
29H), 0.88 (t, J ) 7 Hz, 6H). 13C NMR (CDCl3, 125 MHz): δ 226.0,
153.4, 142.8, 142.3, 139.0, 132.7, 128.5, 128.2, 127.4, 125.0, 35.5,
31.9, 31.3, 29.6, 29.5, 29.4, 29.3, 22.7, 14.1. MALDI-FTMS (M +
H+). Calcd for C40H54S2: 599.3739. Found: 599.3742.
General Procedure for Thiophilic Metal Precipitation. A 10 µL
aliquot of a 200 mM stock of metal ion in water was incubated at
room temperature with 200 µL of a 10 mM stock of ligand 3a in CH3-
CN. Red precipitates from 3a appeared upon increasing the amount of
water in the final reaction mixture to over 20% (v/v). The formation
of these red precipitates provided an ideal tool for verification of sample
quality. Red precipitates from 3a were readily removed by triturating
the precipitate with acetonitrile.
The lack in solubility of the complexes of 4a (M ) Hg2+, Pb2+, and
Cd2+) not only offered metal selection but also allowed one to increase
the selectivity of the method as trituration could be used to remove
more soluble complexes. For instance, Mo2+ complexes of 3a were
readily extracted from Hg2+ complexes 4a by trituration with hot 10%
acetic acid. Spectroscopic analyses indicated that the rate of precipitation
did not correlate with the Ksp and was as given by Hg2+ > Pb2+
>
Cd2+ > (Au3+ ∼ Cu2+)> (Pd2+ ∼ Ni2+) > (Co2+ ∼ Mo2+ ∼ Pt2+).
Determination of the Extent of Metal Precipitation. A 167 µL
aliquot of a 20 mM stock of Hg(OAc)2 in water was incubated at room
temperature with 333 µL of a 10 mM stock of ligand 3 in CH3CN.
Precipitation occurred immediately. After 5 min the sample was
centrifuged for 5 min at 14000g. The supernatant was removed and an
aliquot submitted to vapor atomic absorption analysis (Galbraith
Laboratories, Knoxville, TN).
3,3-Diphenylbenzo[c]thiophene-1(3H)-thione (3a). Lactone 2a (489
mg, 1.71 mmol) was dissolved in xylene (25 mL). To this solution,
P4S10 (380 mg, 0.86 mmol) was added and the reaction heated to reflux
for 18 h. The solution was then cooled, filtered, and concentrated to
give the desired product 3a in excellent yield (540 mg, 99%). 1H NMR
(CDCl3, 300 MHz): δ 8.08 (d, J ) 8 Hz, 1H), 7.59 (dt, J ) 0.5 Hz,
8 Hz, 1H), 7.48 (dt, J ) 0.5 Hz, 8 Hz, 1H), 7.30 (m, 11H). 13C NMR
(CDCl3, 75 MHz): δ 225.5, 153.0, 142.4, 142.0, 132.9, 128.9, 128.7,
128.5, 128.1, 127.6, 125.3. MALDI-FTMS (M + H+). Calcd for
Determination of Solubility Products (Ksp). (1) Weight Analysis
(Figure 2A). Ksp values were determined by weighing the amount of
precipitate 4a. Scaleup was required to provide sufficient material for
analysis on conventional microbalances. The following procedure was
used for this analysis: a 100 µL aliquot of a 200 mM stock of metal
9
J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004 16585

A R T I C L E S
Dickerson et al.
ion in water was incubated at room temperature with 2 mL of a 10
mM stock of ligand 3a in acetonitrile. After 10 min the precipitate
was isolated by centrifugation for 5 min at 2000g, washed with
acetonitrile (5 mL) and methanol (5 mL), and dried in vacuo.
(2) Spectrophotometric Analysis (Figure 2B). A 10 µL aliquot of
a 200 mM stock of metal ion in water was added to 200 µL of a 10
mM stock of ligand 3a in acetonitrile in a spin filter (Millipore). After
10 min at room temperature, the precipitate was removed by centrifuga-
tion at 2000g. Spectroscopic analysis of the supernatant was performed
on a conventional microarray reader (PerSeptive Biosystems CytoFluor
or Perkin-Elmer HST 7000 plate reader).
Heavy Metal Competitions. A 167 µL aliquot of a 20 mM stock
of metal ion in water was incubated at room temperature with 333 µL
of a 10 mM stock of ligand 3a in acetonitrile. After 10 min the sample
was centrifuged for 5 min at 2000g or filtered through a 0.8 µm filter
plate. The supernatant was analyzed on a PerSeptive Biosystems
CytoFluor or Perkin-Elmer HST 7000 plate reader using excitation at
510 nm. The metals presented were prepared using LiCl (EM Chemicals
OmniPure), NaCl (Baker), KCl (EM), CsCl (Aldrich), MgCl2‚6H2O
(EM), CaCl2‚2H2O (EM), Ba(OAc)2 (Alfa AESAR), VCl3 (Alfa), CrCl3‚
6H2O (EMD Chemicals), MoCl3 (Alfa), Mn(OAc)2 (Alfa), FeCl3‚6H2O
(EMD), CoCl2‚6H2O (Aldrich), RhCl3 (Alfa), NiCl2‚6H2O (Aldrich),
PdCl2 (Alfa), PtCl2 (Alfa), CuCl2 (EMD), AgCl (Aldrich), AuCl3 (ICN),
ZnCl2 (EMD), CdCl2 (EMD), HgCl2 (Aldrich), Al(OAc)3 (Alfa), Sn-
(OAc)2 (Alfa), Pb(OAc)2‚3H2O (EMD), AsCl3 (Aldrich), SbCl3 (Alfa),
BiCl2 (Alfa), La(OAc)3 (Aldrich), CeCl3 (Aldrich), Sm(OAc)3 (Aldrich),
Eu(OAc)3 (Aldrich), and Yb(OAc)3 (Aldrich).
Fluorescent Doping of Precipitation Reactions. Fluorescent com-
plexes 4a were prepared by the addition of 200 mM metal in water to
1 mM 3a in the presence of 5 µM Rhod5N (Molecular Probes R-14207)
or 5 µM rhodamine B. Advantageously, the addition of Rhod5N or
rhodamine B led to the formation of needles or globular crystals. Doping
the reactions in this manner reduced the amount of dye required while
providing sufficient fluorescence for analysis on a fluorescence
microscope (Nikon Eclipse TE300). After aggregating on the glass
surface, the precipitate was then washed with H2O (3 × 300 µL). A
Nikon Eclipse TE300 was used for this study. White light images were
collected using Hoffman Modulation Contrast at 100× or 1000×.
Fluorescent images were collected using Y-2E/C (560BP40 excitation
and 595 LP 630/60 BP emission) filter.
Displacement Map Analysis. Positioning of the 1000 regions of
each well was regulated by the assistance of an XY stage (XY stage
85-16, Linos). Precise movements ((5 µm) about the surface of the
well were regulated by mounting the sample on the XY stage and
attaching this stage to the microscope using a small (10 cm2)
microbench. A digital micrometer could be added for automation. The
predicted volume was calculated by multiplying the number of
microspheres by bead volume. Microspheres such as 30-100 µm of
poly(methyl methacrylate) (Sigma) or polystyrene microspheres (Sigma)
were routinely for this analysis.
The volume of the resulting 3D meshes was determined using
conventional space-filling algorithms. As shown in Figures 3A,B,
D-meshes provide an accurate representation of their 2D microscopic
image.
The accuracy of this method was established by comparing a simple
CCD camera (Digital Blue) to an inverted microscope (Nikon Eclipse
TE300). Assays were conducted in wells (surface area of 0.9 cm2) of
a chamber slide system (Lab-Tek*). To reduce optical errors, the bottom
of each reaction well was divided into 1000 individual 300 µm × 300
µm regions and each region was filmed at 20 frames for 5 s. Each film
was compiled into a single image using a scatter correction algorithm
and processed using displacement map analysis. The volume of
precipitate found in each region was tabulated, and the amount of
precipitate found per assay was determined by summation. Calibrations
using 100 µm microspheres indicated that the Digital Blue camera
provided only modest accuracy, delivering a volume of precipitate that
deviated within 6% of the predicted volume, as compared to 4% for
the inverted microscope. This method was then used to determine the
Ksp of metal complexes 4a. The error in the Ksp the CCD camera was
only modestly greater than that obtained with the inverted microscope.
Although this error was considerable, the method was far more sensitive
than visual analysis, and thereby provided an effective system for
analyses in nonlaboratory settings.
Capillary Analysis of Heavy Metal Precipitation. A 50 µL aliquot
of a stock solution of Hg(OAc)2 (10 mM) and Rhod-5N (50 µM;
Molecular Probes R-14207) in H2O was added to 100 µL of a 10 mM
ligand 3a stock in CH3CN. Immediately after mixing, a sample of this
solution was loaded into a 1 mm long segment of a 10 µm i.d. capillary
(Polymicro Technologies Ltd.). The loading was conducted under a
microscope to ensure that crystal formation occurred after reaction
within the capillary. After incubation at room temperature for 15 min,
the capillary was imaged using both an inverted and conventional
fluorescent microscopes.
The material requirements of this assay were reduced using
micrometer-sized capillaries. Fluorescent crystals of 4a (M ) Hg) were
generated in capillaries by placing a 1 ( 0.05 mm long section of 10
µm i.d. fused silica capillary (Polymicro Technologies Inc.) containing
200 µM 3a and 2 µM rhodamine B in acetonitrile into a 100 µL aliquot
of aqueous metal ion. Single or multiple crystals of 4a were generated
by placing capillaries loaded with 3a and dye into aqueous solutions
containing 0.2 nM Hg2+. Using displacement map analysis, the capillary
used for the experiment contained 16.5 ( 2.9 fmol of 4a. Assuming a
1:1 metal to ligand complex in 4a, this represented the detection of
0.17 ( 0.3 nM Hg2+ (0.3 ppb). This finding indicated an 82% yield of
4a upon exposure to a 100 µL aliquot of 0.2 nM Hg2+. Using the same
capillaries, precipitates were obtained when exposed to solutions that
contained greater than 0.2 nM (0.4 ppb) Hg2+, 1.5 nM (0.31 ppb) Pb2+
or 2.5 nM (0.28 ppb) Cd2+
,
.
Acknowledgment. The authors are grateful to Dr. Olivier
Pertz and Prof. Klaus Hahn for microscopy assistance and for
financial support from The Skaggs Institute for Chemical
Biology, an Eli Lilly Graduate Fellowship, and a Norton B.
Gilula Graduate Student Fellowship.
While analytical determinations of precipitation can be preformed
in laboratory settings using microbalances or spectroscopic analysis,
such machines are far too expensive for remote applications. To solve
this problem, we developed a digital displacement routine to determine
the volume, and hence the weight, of precipitate. This technique was
based on conventional 3D volume analysis tools that use Delaunay
triangulation projections to generate at 3D mesh from a 2D image.
JA045798R
9
16586 J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004