Environment-sensitive two-photon probe for intracellular free magnesium ions in live tissue

Article abstract of DOI:10.1002/anie.200700169

(Figure Presented) "Green" emission: A novel two-photon fluorescent probe that can detect intracellular free Mg²⁺ ions at a depth of a few hundred micrometers in live tissue was developed. The emission from the Mg²⁺-probe complex (green) shows little interference from association of the probe with other intracellular metal ions or with the cell membrane (blue).

Full text of DOI:10.1002/anie.200700169

Communications
DOI: 10.1002/anie.200700169
Fluorescent Probes
Environment-Sensitive Two-Photon Probe for Intracellular Free
Magnesium Ions in Live Tissue**
Hwan Myung Kim, Cheol Jung, Bo Ra Kim, Soon-Young Jung, Jin Hee Hong, Young-Gyu Ko,
Kyoung J. Lee, and Bong Rae Cho*
[
4,10]
Herein, we report a 2-acetyl-6-(dimethylamino)naphthalene-
derived two-photon (TP) probe—AMg1—that can detect
intracellular free Mg in living cells and tissue. Mg is one of
the most abundant divalent metal ions in cells, and it plays
crucial roles in many cellular processes such as proliferation
and cell death as well as participating in the regulation of
probes.
As the probes can be accumulated in any
membrane-enclosed structure within the cell and as the
fluorescence quantum yield should be higher in the mem-
brane than in the cytosol, it is practically difficult for the
signals from membrane-bound probes to be separated from
those of the probe–Mg complex. Therefore, there is a need
to develop efficient two-photon probes with 1) enhanced Fd
values for brighter TPM images and 2) larger spectral shifts in
different environments for better discrimination between the
cytosolic and membrane-bound probes.
2
+
2+
2
+
[
1–3]
hundreds of enzymatic reactions.
To detect intracellular
2
+
Mg , a variety of membrane-permeable fluorescent probes
have been developed with some of them being commercially
available. Most of them are used as their acetoxymethyl
AM) esters, which can readily undergo enzymatic hydrolysis
to regenerate the metal-ion probe inside the cell. How-
ever, confocal microscopy with one-photon (OP) fluores-
cent probes is limited for use near the tissue surface
[
4–6]
(
To address both of these problems, we designed the first
two-photon probe (AMg1) that has 2-acetyl-6-(dimethylami-
no)naphthalene as the two-photon chromophore and o-
[
7]
2
+
aminophenol-N,N,O-triacetic acid (APTRA) as the Mg -
selective binding site. We adopted the chromophore from C-
laurdan, a successful two-photon polarity probe for mem-
(< 100 mm).
To observe cellular events deep inside the tissue, it is
[
11]
crucial to use two-photon microscopy (TPM). TPM employ-
ing two near-infrared (NIR) photons for excitation offers a
number of advantages over one-photon microscopy, including
increased penetration depth (> 500 mm), lower tissue auto-
fluorescence and self-absorption, as well as reduced photo-
branes, and APTRA from Magnesium Green (MgG) and
[
4]
Mag-fura-2 probes, with the expectation that AMg1 would
emit strong two-photon excited fluorescence (TPEF) on
forming a complex with Mg . Moreover, if the AMg1–Mg
complex emits TPEF in a widely different wavelength range
depending on the polarity of the environment, the emission
resulting from the membrane-bound probes could be
excluded from that of the AMg1–Mg complex by using
different detection windows.
AMg1 was prepared by reaction of A and a 2-methoxy-
carbonyl-methoxy-N,N-bis(methoxycarbonylmethyl)ben-
zene-1,4-diamine derivative I (Scheme 1). To enhance the cell
permeability, the carboxylic acid moieties were converted
into AM esters (AMg1-AM).
2
+
2+
[
8]
damage and photobleaching. The extra penetration that
TPM affords is of particular interest in tissue imaging because
surface preparation artifacts such as damaged cells extends
2
+
[
9]
over 70 mm into the brain slice interior. However, most of
the OP fluorescent probes presently used for TPM have small
two-photon action cross sections (Fd) that limit their use in
TPM. Another limitation associated with tissue imaging is a
mistargeting problem, which results from membrane-bound
The fluorescence spectra of AMg1-AM showed a gradual
[
*] H. M. Kim, C. Jung, B. R. Kim, Prof. Dr. B. R. Cho
Department of Chemistry and Center for Electro- and Photo-
responsive Molecules
N [12]
bathochromic shift with the solvent polarity (E ) in the
T
order, 1,4-dioxane < DMF < EtOH < H O (Figure S1 and
2
Table S1 in the Supporting Information). The large batho-
chromic shift with increasing solvent polarity indicates the
utility of AMg1-AM as a polarity probe.
Korea University
1-Anamdong, Seoul 136-701 (Korea)
Fax: (+82)2-3290-3544
E-mail: chobr@korea.ac.kr
Figure 1a and b show the spectral response of this probe
2
+
2+
S.-Y. Jung, Prof. Dr. Y.-G. Ko
Graduate School of Life Sciences and Biotechnology
Korea University (Korea)
toward Mg . When Mg was added to AMg1 in Tris buffer
solution (10 mm, pH 7.05), there was a slight change in the
absorption spectrum (Figure 1a). In contrast, a dramatic
increase in the fluorescence was observed with increasing
Dr. J. H. Hong, Prof. Dr. K. J. Lee
National Creative Research Initiative Center for Neurodynamics and
Department of Physics
2+
Mg concentrations probably as a result of the blocking of
the photoinduced electron-transfer (PET) process by metal-
ion complexation (Figure 1b). The fluorescence enhancement
Korea University (Korea)
[
**] This work was supported by KRF-2004-201-C00067. J.H.H. and K.J.L.
were supported by the Creative Research Initiatives of the Korean
Ministry of Science and Technology. We thank Dr. Ji Ho Kim, Institut
PasteurKo re a, forFLIM imaging.
2
+
factor was observed to be 17 in the presence of 100 mm Mg .
A nearly identical result was observed in the two-photon
process (Figure S3 in the Supporting Information). In addi-
2
+
2+
tion, Benesi–Hildebrand plots for Mg and Ca binding
showed a good linear relationship, indicating 1:1 complex-
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3
460
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 3460 –3463

Angewandte
Chemie
2
+
[4,16]
2+
Ca (10 nm–1 mm)
and as chelatable Zn is
essentially nonexistent except in specialized
areas such as the brain hippocampal CA3
[
17]
2+
region, this probe can detect Mg without
2
+
2+
interference from Ca and Zn . Furthermore,
2
+
AMg1 and AMg1–Mg are pH-insensitive in
the biologically relevant pH range (Figure S5
in the Supporting Information).
Scheme 1. Synthesis of AMg1. a) 1. I, EDCI, DMAP, DMF; 2. KOH, EtOH, H O.
2
2
+
The two-photon action spectra of the Mg
3 2 2 3 3
b) CH CO CH Br, Et N, CHCl . I=2-Methoxycarbonyl-methoxy-N,N-bis(methoxycarbonyl-
methyl)benzene-1,4-diamine; EDCI=1-ethyl-3-(3’-dimethylaminopropyl)carbodiimide;
DMAP=4-dimethylaminopyridine; DMF=N,N-dimethylformamide.
complexes of AMg1, MgG, and Mag-fura-2 in
buffer solutions are depicted in Figure 1d.
Table 1 shows that the Fd value for the
2
+
AMg1–Mg complex is 125 GM at 780 nm,
which is sevenfold larger than those for MgG–
2
+
2+
Mg and Mag-fura-2–Mg complexes. This
result indicates that TPM images would be
much brighter when stained with AMg1 than
with the commercial probes.
The TPM images of AMg1-AM-labeled
Hep3B cells are shown in Figure 2. Because the
2
+
fluorescence quantum yields of AMg1–Mg in
Tris buffer (F = 0.58) and AMg1-AM in DMF
(F = 0.32) are much higher than those of
AMg1 (F = 0.04) and AMg1-AM (F = 0.07)
in Tris buffer (Table 1), the TPEF emitted from
the cells should be mostly due to the intra-
2
+
cellular AMg1–Mg complex or membrane-
bound probes. Note that AMg1-AM in DMF is
fl
max
a good model for the latter because the l
values are similar (Figure 2d and Figure S1b in
[
11]
the Supporting Information).
Additional
evidence for this explanation was provided by
the negligible TPEF emitted from the AMg1-
AM-labeled Hep3B cells after treatment with
Figure 1. One-photon absorption (a) and emission (b) spectra of 1 mm AMg1 in the
2+
presence of free Mg ions (0–100 mm). c) One- and two-photon fluorescence titration
1
0 mm calcimycin in the presence of 2 mm
2+
curves for the complexation of AMg1 with Mg ions at various concentrations of free
ethylenediaminetetraacetic acid; the fluores-
cence increased upon treatment with 10 mm
2
+
Mg (0–100 mm). d) Two-photon action spectra of AMg1 (
2
&
), MgG (~), and Mag-fura-
(!) in the presence of 50 mm free Mg . These spectra were measured in 10 mm
2
+
calcimycin in the presence of 100 mm MgCl2
tris(hydroxymethyl)aminomethane (Tris), 100 mm KCl, 20 mm NaCl, 1 mm ethylene
glycol-bis(2-aminoethylether)-N,N,N’,N’-tetraacetic acid (EGTA), pH 7.05.
(Figure S7 in the Supporting Information).
Moreover, the images collected at 360–
OP
TP
lex =365 nm; lex =780 nm.
6
20 nm showed intense spots and bright
domains, with TPEF maxima at l = 440 (blue)
ation of metal ion to probe (Figure S2 in the Supporting
[
13]
Information).
OP
Table 1: Photophysical data formagnesium ion p ro bes.
Dissociation constants (Kd ) were calculated from the
[
a]
ð1Þ
fl
max
ð2Þ
[d]
[e]
[f]
fluorescence titration curves (Figure 1c and Figure S4 in the
Supporting Information). The Kd values for Mg and
Compound
l
l
l
F
d
max
Fd
max
max
[
14]
OP
2+
[b]
[b]
[c]
[
nm]
[nm]
[nm]
2
+
Ca were (1.4 ꢀ 0.1) mm and (9.0 ꢀ 0.3) mm, respectively,
[g]
[h]
[g]
[g]
[g]
AMg1-AM
AMg1
AMg1–Mg
360
365
365
330
506
495
498
498
ND
ND
780
780
800
0.07
0.04
0.58
0.30
0.42
ND
ND
215
56
ND
ND
125
17
[
g]
[g]
which were very similar to those measured for the two-photon
TP
2+
2+
2+
processes [Kd = (1.6 ꢀ 0.1) mm (Mg ), (11 ꢀ 1) mm (Ca )].
2
+
[i]
[i]
[i]
[i]
[i]
Mag-fura-2–Mg
491
This result indicates the operation of a similar mechanism in
2
+
[i]
[
15]
MgG–Mg
532
37
16
both processes during the binding events. The selectivity
toward other metal cations is shown in Figure S5 in the
Supporting Information. AMg1 showed a modest to strong
[a] All data were measured in 10 mm Tris buffer (100 mm KCl, 20 mm
NaCl, 1 mm EGTA, pH 7.05) in the absence and presence (50 mm) of
2
+
2+
2+
2+
MgCl ·6H
2
O. [b] lmax of one-photon absorption and emission spectra.
2
response toward Mg , Ca , Zn , and Mn , and a much
[c] lmax of two-photon excitation spectra. [d] Fluorescence quantum yield,
2
+
2+
2+
weaker response toward Fe , Cu , and Co ions. The metal-
ion selectivity of our probe is similar to those reported for
ꢁ50
4
ꢀ
10%. [e] The peak two-photon cross section in 10 cm s/photon
GM), ꢀ15%. [f] Two-photon action cross section. [g] Not determined.
(
[
4]
MgG and Mag-Fura-2. Because the intracellular concen-
tration of free Mg (0.1–6.0 mm) is much higher than that of
The two-photon excited fluorescence intensity was too weak to measure
2
+
the cross section accurately. [h] F=0.32ꢀ0.02 in DMF. [i] Ref. [18].
Angew. Chem. Int. Ed. 2007, 46, 3460 –3463
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3461

Communications
The errors arising from the membrane-bound probes
could be minimized by detecting the TPEF from the intra-
2
+
cellular AMg1–Mg complex. As shown in Figure 2d, the
shorter-wavelength band in the dissected Gaussian function
(
pale blue line) decreased to the baseline at l ꢂ 500 nm. Thus,
the TPEF emitted from the membrane-bound probe should
be negligible at l > 500 nm. On the other hand, if one uses
AMg1-AM in DMF as a model for the latter (see above), the
tail of the emission band that extends beyond 500 nm could
cause an error (Figure S1b in the Supporting Information).
However, the area of the tail at l > 500 nm accounts for about
5
% of the total emission band, indicating that it would not be
a significant problem. Consistently, the TPEF image collected
at 500–620 nm was homogeneous whereas that collected using
the shorter-wavelength window of 360–460 nm clearly showed
2
+
intense spots (Figure 2b,c). Therefore, one could detect Mg
Figure 2. TPM images collected at 360–620 nm (a), 360–460 nm (b),
ions in the 500–620 nm range with minimum contribution
from the membrane-bound probes.
and 500–620 nm (c) of AMg1-AM-labeled (2 mm) Hep3B cells. d) Two-
photon-excited fluorescence spectra (normalized) from the hydropho-
bic (blue) and hydrophilic (red) domains of AMg1-AM-labeled Hep3B
2
+
The intracellular Mg concentration was determined by
2+
using [Mg ] = K [(FꢁFmin)/(F ꢁF)] as reported (F = fluo-
d
max
cells (l =780 nm). The cell images shown are representative images
[5,20]
2+
ex
rescence intensity).
The Mg concentration in the resting
from replicate experiments. Gaussian fits are shown in pale blue and
brown. e) TPEF intensity collected at 500–620 nm (red) and 360–
Hep3B cells was (0.65 ꢀ 0.10) mm, in good agreement with
[
16]
reported values. Intracellular magnesium ions have been
qualitatively detected with TPM by using the newly devel-
4
60 nm (blue) as a function of time.
[
6]
oped probe 2,3-dicyanohydroquinone (DCHQ). However,
this is the first example of quantitative measurement of
2
+
and 498 nm (red), respectively (Figure 2a,d). Compared with
the emission spectra recorded in Tris buffer, the blue band
was significantly blue-shifted while the red one band nearly
identical (Table 1). Both spectra could be fitted to two
Gaussian functions with maxima at 439 and 488 nm (pale blue
lines) and at 426 and 498 nm (brown lines), respectively
intracellular free Mg ions with TPM.
To demonstrate the utility of this probe in deep-tissue
imaging, acute hippocampal slices from postnatal 3-day-old
mice were incubated with 5 mm AMg1-AM for 30 min at
378C. Figure 3a displays the brightfield image of a part of an
acute mouse hippocampal slice that reveals the CA1 and CA3
(
Figure 2d). It was observed that the peak positions of the
dissected spectra were similar, suggesting that the probes
might be located in two regions with different polarity.
Furthermore, the intense spot exhibited an excited-state
lifetime of 3.3 ns, which was much longer than the upper
extreme of the lifetime distribution curve centered at 2.2 ns
(
Figure S6 in the Supporting Information). From these results,
we hypothesize that the probes may be located in two
different environments, a more polar one that is likely to be
cytosol (red emission with a shorter lifetime) and a less polar
one that is likely to be membrane-associated (blue emission
with a longer lifetime).
Supporting evidence for this hypothesis was provided by
study of carbonyl cyanide m-chlorophenylhydrazone (CCCP)
activity. Because CCCP prevents the production of ATP–
2
+
Mg from ADP (ATP = adenosine triphosphate, ADP =
2
+
adenosine diphosphate), inorganic phosphate, and Mg by
uncoupling oxidative phosphorylation, CCCP-treated cells
2
+ [19]
should have higher levels of free Mg . When CCCP was
added to the AMg1-AM-labeled Hep3B cells, the TPEF
intensity in the 500–620 nm region increased immediately
after addition and then decreased to the baseline level
Figure 3. Images of an acute mouse hippocampal slice stained with
5 mm AMg1-AM. a) Brightfield image shows the CA1 and CA3 regions
as well as the dentate gyrus upon magnification (10ꢀ). b) TPM image
at the same magnification reveals the same regions at a depth of
about 270 mm. c) Magnification at 40ꢀ shows the CA1 layerat a depth
of around 150 mm. d) Magnification at 100ꢀ shows CA1 pyramidal
neurons at a depth of approximately 150 mm. Scale bars: 300 (a,b),
(
Figure 2e). Hence, the activity of CCCP is visually confirmed
by this result. In contrast, no change in the TPEF intensity was
noted in the 360–460 nm range, indicating that the TPEF
indeed arises from the AMg1-AM located in the cell
membrane.
120 (c), and 30 mm (d). The TPEF images were collected at 500–
620 nm upon excitation at 780 nm with fs pulses.
3
462 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 3460 –3463

Angewandte
Chemie
regions as well as the dentate gyrus. The TPM image revealed
[4] A Guide to Fluorescent Probes and Labeling Technologies, 10th
2
+
ed., (Ed.: R. P. Haugland), Molecular Probes, Eugene, OR, 2005.
5] H. Komatsu, N. Iwasawa, D. Citterio, Y. Suzuki, T. Kubota, K.
Tokuno, Y. Kitamura, K. Oka, K. Suzuki, J. Am. Chem. Soc.
Mg distributions in the same regions at 100–300 mm depth
Figure 3b and Figure S8 in the Supporting Information),
[
(
2
+
however, we cannot rule out the possibility that Zn may
have contributed in the CA3 region (see above). Moreover,
the images taken at higher magnifications resolved Mg
distributions in the pyramidal neuron layer of the CA1
region (Figure 3c and d), where intracellular Zn is essen-
tially nonexistent. Furthermore, a closer examination of the
image shown in Figure 3d revealed that this probe could also
detect Mg ions in the nucleus in the deep tissue. These
results demonstrate that AMg1 is capable of detecting
endogenous stores of labile Mg at 100–300 mm depth in
live tissues using TPM.
2004, 126, 16353 – 16360.
[
17]
[
6] G. Farruggia, S. Iotti, L. Prodi, M. Montalti, N. Zaccheroni, P. B.
Savage, V. Trapani, P. Sale, F. I. Wolf, J. Am. Chem. Soc. 2006,
128, 344 – 350.
2
+
2
+
[7] R. Y. Tsien, Nature 1981, 290, 527 – 528.
[8] a) W. R. Zipfel, R. M. Williams, W. W. Webb, Nat. Biotechnol.
[
17]
2
2
003, 21, 1369 – 1377; b) F. Helmchen, W. Denk, Nat. Methods
005, 2, 932 – 940.
2
+
[
9] R. M. Williams, W. R. Zipfel, W. W. Webb, Curr. Opin. Chem.
Biol. 2001, 5, 603 – 608.
2
+
[10] a) C. L. Slayman, V. V. Moussatos, W. W. Webb, J. Exp. Biol.
1994, 196, 419 – 438; b) M. J. Petr, R. D. Wurster, Cell Calcium
1997, 21, 233 – 240.
In conclusion, we have developed a two-photon probe for
2
+
[11] H. M. Kim, H.-J. Choo, S.-Y. Jung, Y.-G. Ko, W.-H. Park, S.-J.
the detection of intracellular free Mg by TPM. This probe
2
+
Jeon, C. H. Kim, T. Joo, B. R. Cho, ChemBioChem 2007, 8, 553 –
showed 17-fold TPEF enhancement in response to Mg , a
559.
TP
dissociation constant (Kd ) of 1.6 mm, and sevenfold stronger
[
[
12] C. Reichardt, Chem. Rev. 1994, 94, 2319 – 2358.
13] H. A. Benesi, J. H. Hildebrand, J. Am. Chem. Soc. 1949, 71,
2703 – 2707.
[14] a) J. R. Long, R. S. Drago, J. Chem. Educ. 1982, 59, 1037 – 1039;
b) K. J. Hirose, J. Inclusion Phenom. Macrocyclic Chem. 2001,
TPEF than Mg-fura-2 and MgG commercial probes, and can
2
+
detect intracellular free Mg in live cells and tissue without
interference from other metal ions and membrane-bound
probes.
39, 193 – 209.
[
15] a) H. M. Kim, M.-Y. Jeong, H. C. Ahn, S.-J. Jeon, B. R. Cho, J.
Org. Chem. 2004, 69, 5749 – 5751; b) H. C. Ahn, S. K. Yang,
H. M. Kim, S. Li, S.-J. Jeon, B. R. Cho, Chem. Phys. Lett. 2005,
Received: January 13, 2007
Published online: April 2, 2007
4
10, 312 – 315.
16] a) J. G. Fitz, A. H. Sostman, J. P. Middleton, Am. J. Physiol.
994, 266, G677-G684; b) M. R. Cho, H. S. Thatte, M. T. Silvia,
Keywords: fluorescent probes · imaging agents · magnesium ·
two-photon processes
[
[
.
1
D. E. Golan, FASEB J. 1999, 13, 677 – 683.
17] C. C. Woodroofe, R. Masalha, K. R. Barnes, C. J. Frederickson,
S. J. Lippard, Chem. Biol. 2004, 11, 1659 – 1666.
[18] H. Szmacinski, J. R. Lakowicz, J. Fluoresc. 1996, 6, 83 – 95.
[19] D. L. Nelson, M. Cox, M. Lehninger, Principles of Biochemistry,
4th ed., W. H. Freeman & Company, New York, 2005, p. 707.
[20] R. E. London, Annu. Rev. Physiol. 1991, 53, 241 – 258.
[
[
1] H. Rubin, BioEssays 2005, 27, 311 – 320.
2] R. Eskes, B. Antonsson, A. Osen-Sand, R. Montessuit, C.
Richter, R. Sadoul, G. Mazzei, A. Nichols, J. C. Martinou, J. Cell
Biol. 1998, 143, 217 – 224.
[
3] Magnesium and the Cell (Ed.: N. J. Birch), Academic Press, San
Diego, CA, 1993.
Angew. Chem. Int. Ed. 2007, 46, 3460 –3463
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3463