A general fluorescence resonance energy transfer (FRET) method for observation and quantification of organometallic complexes under reaction conditions

Article abstract of DOI:10.1021/om900629s

A two-fluorophore FRET system provided a more general approach than previously described fluorescence techniques to observing and quantifying organometallic complexes under reaction conditions. Over the concentration range of 3 × 10^-7 to 5 × 10^-6 M, this method provided quantification with faster time resolution and greater sensitivity than is possible with N M R spectroscopy.

Full text of DOI:10.1021/om900629s

Organometallics 2009, 28, 4643–4645 4643
DOI: 10.1021/om900629s
A General Fluorescence Resonance Energy Transfer (FRET) Method for
Observation and Quantification of Organometallic Complexes under
Reaction Conditions
Sung-Gon Lim and Suzanne A. Blum*
Department of Chemistry, University of California, Irvine, California 92697-2025
Received July 17, 2009
Scheme 1. Tagging of One Ligand with a Donor Fluorophore and
a Second Ligand with an Acceptor Fluorophore To Permit
Quantification of a Metal Complex with Both Ligands Bound,
through FRET
Summary: A two-fluorophore FRET system provided a more
general approach than previously described fluorescence tech-
niques to observing and quantifying organometallic complexes
under reaction conditions. Over the concentration range of
3 ꢀ 10-7 to 5 ꢀ 10-6 M, this method provided quantification
with faster time resolution and greater sensitivity than is
possible with NMR spectroscopy.
Small quantities of transition-metal complexes play key
roles in catalytic reactions,1 and therefore new methods for
the sensitive detection of transition-metal complexes can
lend insight into catalytic and other metal-mediated pro-
cesses. For example, Sohn and Ihee recently reported that
the metal-based quenching of a ligand’s fluorescence could
be used to monitor formation of an intermediate in the
Grubbs enyne metathesis reaction at low concentration.2
The increased sensitivity of a fluorescence measurement
over an NMR spectroscopy measurement3 was critical for
their detection. We envisioned a fluorescence resonance
energy transfer (FRET)4,5 method more general than pre-
viously reported fluorescence detection methods. Specifi-
cally, a two-fluorophore system would be employed: a
donor fluorophore attached to a ligand through an alkyl
spacer and an acceptor fluorophore similarly attached to a
second ligand. When both ligands were bound to the metal
center, their proximity would result in a strong FRET signal
(Scheme 1), permitting quantification of that specific com-
plex at low concentrations. In contrast to the extensive
literature pertaining to the detection of specific metal ions
that relies on fluorescence quenching and enhancement by
the metal atom,2,6-12 this new method would use spectator
fluorophores which would not rely on changing photophy-
sical properties from interaction with the metal, thereby
creating a more general method for studying reactions of
diverse substrates and metal centers. The current dearth of
fundamental studies and applications of two-fluorophore
FRET in transition-metal systems stands in sharp contrast
to the abundant studies and applications in biochemical
systems.13-15 Two-fluorophore FRET’s employment in
transition-metal complexes will open up new areas of study
and application; however, the fundamental properties and
feasibility of the method had not been demonstrated pre-
viously. Herein, we report fundamental studies of this
method, with the goal of establishing its useful concentra-
tion range, determining its sensitivity in comparison to
NMR spectroscopy, quantifying its FRET efficiency, and
discovering the role of the metal (if any) in moderating the
FRET process. These fundamental studies provide the
groundwork for the general employment of two-fluoro-
phore FRET to study transition-metal complexes at low
concentrations.
In order to explore the useful concentration range of this
two-ligand FRET method, palladium complex 4 was
synthesized with a donor and an acceptor fluorophore.
Complexes analogous to 4 are intermediates in palladium-
catalyzed nucleophilic allylic substitution reactions,16 and
therefore 4 was chosen as a representative complex for
this fundamental study. We examined the tetramethyl
BODIPY (dipyrromethene boron difluoride) core as a
green FRET donor (λex =498 nm, λem =504 nm) and the
tetrahydroindole BODIPY core as an orange FRET accep-
tor (λex=541 nm, λem=545 nm) (eq 1).17 Displacement of
*To whom correspondence should be addressed. E-mail: blums@
uci.edu.
(1) Hegedus, L. S. Transition Metals in the Synthesis of Complex
Organic Molecules, 2nd ed.; University Science Books: Sausalito, CA, 1999.
(2) Sohn, J.-H.; Kim, K. H.; Lee, H.-Y.; No, Z. S.; Ihee, H. J. Am.
Chem. Soc. 2008, 130, 16506.
(3) Jarrell, H. C. J. Magn. Reson. 2009, 198, 204.
(4) Vogel, S. S.; Thaler, C.; Koushik, S. V. Sci. STKE 2006, 331, re2.
(5) Jares-Erijman, E. A.; Jovin, T. M. Nat. Biotechnol. 2003, 21, 1387.
(6) Lee, S. Y.; Kim, H. J.; Wu, J.-S.; No, K.; Kim, J. S. Tetrahedron
Lett. 2008, 49, 6141.
(7) Lee, M. H.; Kim, H. J.; Yoon, S.; Park, N.; Kim, J. S. Org. Lett.
2008, 10, 213.
(8) de Silva, A. P.; de Silva, S. A. J. Chem. Soc., Chem. Commun. 1986,
1709.
ꢀ
(9) Sanz-Medel, A.; Alonso, J. I. G.; Gonzalez, E. B. Anal. Chem.
(13) Roy, R.; Hohng, S.; Ha, T. Nat. Methods 2008, 5, 507.
(14) Sekar, R. B.; Periasamy, A. J. Cell Biol. 2003, 160, 629.
(15) Tuschi, T.; Gohlke, C.; Jovin, T. M.; Westhof, E.; Eckstein, F.
Science 1994, 266, 785.
(16) Trost, B. M. Acc. Chem. Res. 1980, 13, 385.
(17) Canham, S. M.; Bass, J. Y.; Navarro, O.; Lim, S.-G.; Das, N.;
Blum, S. A. Organometallics 2008, 27, 2172.
1985, 57, 1681.
(10) Baxter, P. N. W. J. Org. Chem. 2004, 69, 1813.
(11) Winkler, J. D.; Bowen, C. M.; Michelet, V. J. Am. Chem. Soc.
1998, 120, 3237.
(12) Burdette, S. C.; Frederickson, C. J.; Bu, W.; Lippard, S. J. J. Am.
Chem. Soc. 2003, 125, 1778.
r
2009 American Chemical Society
Published on Web 07/29/2009
pubs.acs.org/Organometallics

4644 Organometallics, Vol. 28, No. 16, 2009
Lim and Blum
chloride from complex 3 with the tagged phosphine 1
produced complex 4 in quantitative yield (eq 1).
Figure 1. (A) Comparison of fluorescence intensities of sepa-
rate solutions of 1, 1 þ 7, 2, and 4, at 2.5 ꢀ 10-6 M, showing
FRET in 4. (B) Red and black lines denoting the linear regions
for detection of 4 at 545 and 504 nm, respectively.
We next investigated FRET as a tool to detect 4, which we
anticipated would occur only when both fluorophores were
attached through ligands to the palladium center. No sig-
nificant changes in the photophysical properties of the
BODIPY cores of 1 or 218 were observed upon separate
coordination to the metal through a ligand and a spacer in
the monotagged complexes 6 and 5, respectively, confirming
their behavior as photophysical spectators.19 Excitation of a
2.5 ꢀ 10-6 M solution of 4 at 460 nm resulted in a decrease of
the emission signal at 504 nm (green) and a concomitant
increase in the emission signal at 545 nm (orange), in
comparison to the emission intensities of 1 and 2. This
decrease at 504 nm corresponded to a FRET efficiency of
92% (Figure 1A).
Figure 2. (A) Addition of 1 to 3 in one portion. (B) Calculated
intensity corresponding to full conversion to 4. (C) Addition of
1 to 3 in four portions. (D) Control, with 1 added to 7 in four
portions.
from FRET,21 the emission intensity of 8 at 545 nm was not
significantly changed relative to 4, establishing different
criteria for FRET in palladium complexes than in labeled
DNA (119 au for 8; 119 au for 4).
The emission intensity of a solution of 4 at 504 and 545 nm
was examined over 4 orders of magnitude: 1 ꢀ 10-4 to 1 ꢀ
10-8 M. The solution’s emission intensity was proportional
to the concentration of 4 over the range 3 ꢀ 10-7 to 5 ꢀ
10-6 M (Figure 1B). For example, at 3.1 ꢀ 10-7 M, the
fluorescence intensity at 545 nm was 16.2 ( 1.2 au and at
1.3 ꢀ 10-5 M the intensity was 62.8 ( 5.5 au. Standard
deviations from triplicate runs were consistently below 10%
of the value of the measurement. At concentrations greater
than 1 ꢀ 10-5 M, intermolecular FRET and self-quenching
became competitive with the desired intramolecular FRET,
as determined by detection of FRET in a mixed solution of
free ligands 1 and 7 at 1.0ꢀ10-5 M and self-quenching22 in a
separate solution of 1 at 1.0 ꢀ 10-5 M.19 At concentrations
less than 5ꢀ10-7 M, quantification of the small differences
in emission signal approached the limits of the spectrometer.
With the fundamental concept established, the FRET
method was applied to the quantification of complex
4 during a chemical reaction. Addition of equal volumes of
A control experiment examined the excitation of a mixed
solution of 1 and bromide 7, both at 2.5 ꢀ 10-6 M. Bromide 7
was examined as a nonreactive surrogate to allyl chloride 2
and had indistinguishable photophysical properties. No
decrease in the signal at 504 nm occurred upon mixing of
1 and 7. Thus, the decrease in the signal at 504 nm from a
solution of 4 arose from intramolecular FRET within com-
plex 4, since intermolecular FRET was not significant at this
concentration.
Complexes analogous to the (η1-allyl)chlorobis(phos-
phine)palladium species 8, which was formed upon exposure
of complex 3 to 4 equiv of phosphine 1, have been proposed
as the reactive electrophiles in palladium-catalyzed nucleo-
philic allylic substitution reactions.20 In order to build a
predictive model for the photophysical properties of metal
complexes tagged with two donors and one acceptor,4 we
investigated the fundamental FRET properties of complex 8.
In contrast to Watanabe’s report that labeling DNA with
two BODIPY donors and one acceptor enhanced the signal
a 5.0 ꢀ 10-6 M solution of phosphine 1 to a 2.5 ꢀ 10-6
M
solution of complex 3 was monitored using 460 nm excitation
and detection of fluorescence intensity at 545 nm (Figure 2,
line A). Data were acquired every 0.5 s. In the first experi-
ment, solutions of 1 and 3 were mixed in one portion. The
first data point, acquired 10 s after the mixing of 1 and 3,
showed a high fluorescence intensity that did not change over
(18) Mixture of regioisomers (see the Supporting Information for
details).
(19) See the Supporting Information for data, comparisons, and
calulations.
(20) Amatore, C.; Jutand, A.; M’Barki, M. A.; Meyer, G.; Mottier, L.
Eur. J. Inorg. Chem. 2001, 873.
(21) Okamura, Y.; Kondo, S.; Sase, I.; Suga, T.; Mise, K.; Furusawa,
(22) Lakowicz, J. R. Principles of Fluorescence Spectroscopy;
I.; Kawakami, S.; Watanabe, Y. Nucleic Acids Res. 2000, 28, e107.
Springer: New York, 2006; p 69.

Communication
Organometallics, Vol. 28, No. 16, 2009 4645
time, implying an extremely rapid reaction between 1 and
3 to form 4, even at these low concentrations. The mea-
sured fluorescence intensity of the solution, 125 au, corre-
sponded to 2.5ꢀ10-6 M 4 and was within 2% of the expected
fluorescence value for full conversion (127 au, Figure 2, line B).
Half-second resolution of the initial stages of the reaction was
obtained by adding four aliquots of a 5.0ꢀ10-6 M solution of
1 (2.0 equiv) to a 2.5ꢀ10-6 M solution of 3 (1.0 equiv) inside
the fluorimeter. Addition of each sequential aliquot resulted in
a rapid increase in FRET, consistent with a reaction rate to
fully form 4 of less than 2 s (Figure 2, line C). Since mixing was
slow inside the fluorimeter, the signal required about 10 s
to fully stabilize after each addition, during which time the
solution homogenized. Intensity measurements at less than
10 s were within 15% of the homogenized value. A control
experiment examined the fluorescence intensity of a 5.0 ꢀ
Figure 3. (A) Concentrated 31P NMR spectrum at 1.3 ꢀ 10-3 M,
312 scans, 8 min acquisition (eq 3). (B) After dilution, 31P NMR
spectrum at 1.3 ꢀ 10-6 M, 1086 scans, 27 min acquisition; no
signalobserved (eq3). (C) Fluorescence spectrumat 1.3ꢀ 10-6
M
(eq 2).
acquisition times (1086 scans, 27 min, 31P; Figure 3B).
Similarly, 1H NMR spectroscopy was not able to identify or
quantify the complexes, even with long acquisition times (512
10-6 M solution of 7 when four aliquots of a 5.0 ꢀ 10-6
M
1
scans H, see the Supporting Infornation). In contrast, our
solution of 1 were added, in the absence of palladium, and
established that no increase in fluorescence intensity was
observed (Figure 2, line D).
FRET method can detect and quantify 4 in 1 s (Figure 3C; full
spectrum clearly observable at 1.2 ꢀ 10-6 M). This result
confirmed that the FRET technique permitted rapid quantifi-
cation of complex 4 at concentrations where the lower sensi-
tivity of NMR spectroscopy was not adequate.
Formation of 4 could be observed and quantified in a
reaction that produced a mixture of products (eq 2). Addi-
tion of a mixture of phosphines to 3 was anticipated to
produce a mixture of complexes 4, 5, 9, and 10. During the
reaction, dimer 3 produces 2 equiv of (η3-allyl)palladium;
thus, a total of 2 equiv of phosphines was added. After
mixing, excitation of the solution at 460 nm resulted in a
fluorescence intensity of 13 au at 504 nm and 65 au at 545 nm,
corresponding to a measured concentration of 4 of 1.2 ꢀ
10-6 M (1.3ꢀ10-6 M expected, on the basis of complete and
rapid consumption of phosphine 1 as identified in Figure 2).19
This result established that low concentrations of side pro-
ducts did not interfere with this method’s quantitative appli-
cation: for example, by excited-state quenching.23-25
In conclusion, FRET with spectator fluorophores pro-
vided a more general method for studying transition-metal
complexes than the metal-ligand quenching method pre-
viously reported. The high signal from FRET permitted
quantification of complexes with the rapid time resolution
of 1 s at 1.3 ꢀ 10-6 M, a concentration at which the lower
sensitivity of NMR spectroscopy was not sufficient. No
evidence for photophysical involvement of the metal was
observed. More broadly, these fundamental studies provide
the groundwork and feasibility demonstration for future
applications, such as studying transition-metal intermediates
using ensemble FRET methods, or single-molecule FRET
studies of transition-metal reactivity,17 in analogy to single-
molecule biophysical FRET studies.13,26
To directly compare the sensitivity of the FRET technique
1
with that of 31P and H NMR spectroscopy, a mixture of
phosphines was added to dimer 11 at 2.1ꢀ10-3 M to produce
analogous nonfluorophore-tagged complexes 12-14 at con-
centrations high enough for NMR spectroscopic detection
(eq 3). After confirmation of the reaction by 31P NMR
spectroscopy (Figure 3A), the mixture of products was then
diluted to the identical concentrations of the previous FRET
experiment. At these low concentrations, the less sensitive
technique of 31P NMR spectroscopy could not detect any
of the resonances from the complexes, even with long
Acknowledgment. We thank the U.S. Department of
Energy, Office of Basic Energy Sciences, Grant No. DE-
FG02-08ER1534, for funding and Mr. Yong Wang and
Mr. Yili Shi for helpful discussions.
Supporting Information Available: Text and figures giving
experimental procedures, compound characterization data, and
sample fluorescence spectra. This material is available free of
charge via the Internet at http://pubs.acs.org.
(23) Okamoto, M. J. Phys. Chem. A 2000, 104, 7518.
(24) Sashuk, V.; Schoeps, D.; Plenio, H. Chem. Commun. 2009, 770.
(25) Varnes, A. W.; Dodson, R. B.; Wehry, E. L. J. Am. Chem. Soc.
1972, 94, 946.
(26) Weiss, S. Science 1999, 283, 1676.