protocol
Synthesis of a highly Mg2+-selective fluorescent probe and its application to quantifying and imaging total intracellular magnesium Azzurra Sargenti1,5, Giovanna Farruggia1,2,5, Nelsi Zaccheroni3, Chiara Marraccini4, Massimo Sgarzi3, Concettina Cappadone1, Emil Malucelli1, Alessandra Procopio1, Luca Prodi3, Marco Lombardo3 & Stefano Iotti1,2 Pharmacy and Biotechnologies, University of Bologna, Bologna, Italy. 2National Institute of Biostructures and Biosystems, Rome, Italy. 3Department of Chemistry ′Giacomo Ciamician’, University of Bologna, Bologna, Italy. 4ASMN-IRCCS, Department of Transfusion Medicine, Reggio Emilia, Italy. 5These authors contributed equally to this work. Correspondence should be addressed to L.P. (
[email protected]), M.L. (
[email protected]) or S.I. (
[email protected]).
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
1Department of
published online 2 February 2017; doi:10.1038/nprot.2016.183
Magnesium plays a crucial role in many physiological functions and pathological states. Therefore, the evolution of specific and sensitive tools capable of detecting and quantifying this element in cells is a very desirable goal in biological and biomedical research. We developed a Mg2+-selective fluorescent dye that can be used to selectively detect and quantify the total magnesium pool in a number of cells that is two orders of magnitude smaller than that required by flame atomic absorption spectroscopy (F-AAS), the reference analytical method for the assessment of cellular total metal content. This protocol reports itemized steps for the synthesis of the fluorescent dye based on diaza-18-crown-6-hydroxyquinoline (DCHQ5). We also describe its application in the quantification of total intracellular magnesium in mammalian cells and the detection of this ion in vivo by confocal microscopy. The use of in vivo confocal microscopy enables the quantification of magnesium in different cellular compartments. As an example of the sensitivity of DCHQ5, we studied the involvement of Mg2+ in multidrug resistance in human colon adenocarcinoma cells sensitive (LoVo-S) and resistant (LoVo-R) to doxorubicin, and found that the concentration was higher in LoVo-R cells. The time frame for DCHQ5 synthesis is 1–2 d, whereas the use of this dye for total intracellular magnesium quantification takes 2.5 h and for ion bioimaging it takes 1–2 h.
INTRODUCTION The magnesium ion is essential to a wide range of cellular functions, and many biochemical reactions require it as a cofactor. However, its role in cellular signal transduction is not yet fully understood. Imbalance of magnesium ion regulation is observed in association with various common human diseases such as hypertension, cardiovascular diseases, neuromuscular disorders and diabetes1–3. The study of magnesium ion homeostasis gained special attention over the past several years due to the discovery and characterization of magnesium-selective ion channels and due to the use of molecular biology tools to elucidate their physiological functions at the cellular level3–6. However, tools enabling the detection and quantification of intracellular magnesium ions, which would pave the way to a better comprehension of Mg2+ regulation in live cells, have not been developed at a pace matching the growing interest in this area of research7. In addition, the Mg2+-sensitive dyes that are commercially available and most of those developed by several research groups selectively detect only the free metal ions, precluding the possibility of identifying the total pool of the element to analyze. This feature represents a serious analytical obstacle in an intracellular milieu in which free Mg2+ is only a small proportion of the total amount of magnesium ion. Competing analytical approaches and their shortcomings The method commonly used to quantify total magnesium in cells or tissue is F-AAS, conducted on acidic extracts. However, this technique necessitates samples of millions of cells or several milligrams of tissue, which is a serious drawback for many applications. Other more sensitive techniques exist, such as graphite
furnace AAS and inductively coupled plasma mass spectroscopy (MS), which require smaller sample sizes. However, these techniques require instrumentation that is not easily available in many laboratories, specific technical skills and complex analytical procedures. An additional shortcoming of the analytical techniques just mentioned, and one that prevents a deeper comprehension of the mechanisms governing cellular homeostasis of magnesium, is the lack of information they provide with respect to the intracellular repartition between free and bound magnesium ions8. In fact, the intracellular compartmentalization of Mg2+ has not yet been thoroughly elucidated, mainly as a consequence of the inadequacy of analytical techniques aimed at mapping total intracellular magnesium distribution7,9. The absence of fluorescent dyes specific to the assessment of total intracellular magnesium ions has until recently represented a methodological hurdle in determining the intracellular content and distribution of this cation. Convincing evidence indeed demonstrates the peculiarity of the cellular magnesium ion homeostasis: as a consequence of different hormonal and nonhormonal stimuli, substantial amounts of free Mg2+ have been shown to flow across the cell membrane in both directions, resulting in substantial changes in the amount of cation present in plasma8. However, this ion trafficking results in small changes in free Mg2+ intracellular concentration, although large variations in total magnesium content have been found in subcellular organelles and tissues10. These results suggest that alterations in magnesium availability, influencing several physiological functions, could be attained by modifications of the extent of magnesium binding by cellular ligands and of the ion’s distribution among cellular nature protocols | VOL.12 NO.3 | 2017 | 461
protocol Cl
Acetone 5 h, RT
N OH
Ph
N OTs
1
K2CO3, PhB(OH)2 DMF, 4 h, 150 °C
2 (90–95%)
1. NaOH (aqueous) EtOH/acetone 2.5 h, 80 °C
2.5
Pd(PPh3)4 (5 mol%)
2.5
OTs
3 (85–90%)
2
N OH
4 (85–90%)
O O
N
N
O
4 +
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
H
O
5
H
+ (CH2O)n 6
Option A Toluene 120 °C, 18 h
OH N
Option B Toluene MW (600 W), 4 h
Ph
Ph
O O
N
N
O O
2
500
1.5
320
1
160
0.5 0
1.5
×105
80 0
200 400 Mg2+ (µM)
600
5 2 1 Dye
OH
subcompartments, rather than by major modifications of the overall free ion fraction11. Mitochondria have been reported to represent the main intra cellular store for Mg2+ (ref. 12), and disrupting the mitochondrial membrane potential has been found to cause a release of Mg 2+ into the cytosol9,12. These results lead to the hypothesis that mitochondria have a central role in the intracellular homeostasis of this element10. Over the past few years, some novel fluorescent dyes with peculiar features, such as high selectivity for mitochondrial Mg2+ (ref. 13) and enhanced red-shifted ratiometric Mg2+ detection14, have been developed. However, these very advanced fluorescent dyes were not designed to provide quantitative data, so they cannot be used in analytical assays. DCHQ5-based fluorescent dye for total intracellular magnesium quantification We recently designed and developed a DCHQ5-based fluorescent dye with the unique analytical capability to accurately quantify total intracellular magnesium content15. This probe is composed of an N,N′-bis-((8-hydroxy-7-quinolinyl)methyl)-1,10-diaza-18crown-6 ether bearing a phenyl group as a substituent in position 5 of each hydroxyquinoline (HQ) arm16, and is named DCHQ5. It showed enhanced characteristics as compared with those of the original reference DCHQ1 previously synthesized9, as it displays greater fluorescence intensity upon cation binding, more intense membrane staining and higher intracellular retention than DCHQ1; furthermore, in contrast to DCHQ1, it affords the option of being excited by light in both the UV and visible spectrum ranges17. DCHQ5 is prepared in a 60–65% overall yield by a four-step linear synthetic sequence, starting from commercial 5-chloro-8hydroxyquinoline (1; Fig. 1). The phenyl substituent in position 5 is introduced by a palladium-catalyzed Suzuki cross-coupling on protected HQ (2), followed by basic hydrolysis to afford the desired 5-phenyl-8-hydroxyquinoline (4) in high yield18. DCHQ5 is prepared through a one-pot Mannich condensation reaction of 4 with 1,4,10,13-tetraoxa-7,16-diazacyclooctadecane (5), and paraformaldehyde (6), either thermally (see Step 37(A) of the
Mg2+ (µM)
10
1
N
DCHQ5 (90–95%)
40 20
0.5
Figure 1 | Synthetic route to DCHQ5 fluorescent dye. MW, microwave heating; RT, room temperature.
462 | VOL.12 NO.3 | 2017 | nature protocols
DCHQ5/MeOH:MOPS
N
Ph
2. HCl (aqueous)
×105
Fluorescence intensity (a.u.)
NaOH (aqueous) TsCl
Fluorescence intensity (a.u.)
Cl
0 400
Buffer 450
500
� (nm)
550
600
650
Figure 2 | Emission spectra (λex = 360 nm) of DCHQ5 (10 µM) in titration with Mg2 + in MeOH:MOPS upon addition of increasing MgSO4 concentrations (1, 2, 5, 10, 20, 40, 80, 160, 320 and 500 µM) reported in arbitrary units (a.u.). Inset: fluorescence intensity of DCHQ5 at 510 nm in MeOH:MOPS, reported as a function of Mg2 + concentration. a.u., arbitrary units.
PROCEDURE) or by microwave-assisted reaction (see Step 37(B) of the PROCEDURE). The photophysical properties of this class of compounds are determined by the presence of the derivatives of the well-known 8-HQ dye19,20. The absorption spectrum of DCHQ5 presents a very intense band attributed to a π–π* transition, with a maximum at 250 nm (ε = 81,000 per M/cm) and a less intense, broader and nonstructured band at 330 nm (ε = 9,000 per M/cm), which is associated with a charge transfer from the oxygen atom to the quinoline moiety17. We had already proven that the substitution of the hydrogen in the 5 position of the quinoline structure of DCHQ1 generally causes a red-shift of both these peaks, and, in the case of DCHQ5, the introduction of two aromatic groups, which results in an overall charge delocalization, causes an increase in the molar absorption coefficients17. Addition to the dye solution of increasing amounts of Mg 2+ induces a decrease in the intensity of both bands and the appearance of a new band at 390 nm, which indicates the deprotonation of the hydroxyl group in the 8 position of HQ via complexation, even if, in the case of Mg2+ binding, this deprotonation is incomplete17. The emission spectrum of DCHQ5 presents a nonstructured fluorescence band centered at 512 nm, whose intensity and energy are strongly affected by the nature of the solvent. As with all HQ-based compounds, the emission band’s intensity increases with the lipophilicity of the solvent21; therefore, results obtained in MeOH–H2O (1:1) solution buffered at pH 7.4 with 4-morpholinepropanesulfonic acid (reported in Table 1) are not fully representative of the highly complex cellular environment, and this is especially true for the highly aromatic DCHQ5. The emission quantum yield of the free ligand is quite low (3.0 × 10−4), as expected, as, in HQ derivatives, intramolecular photoinduced proton transfer (PPT) between the photoacid hydroxyl and the nearby photobase quinoline nitrogen quenches the emission. Moreover, in protic media, intermolecular PPT processes
protocol Table 1 | Photophysical properties of DCHQ5 and its complexes in MeOH–H2O (1:1) solution buffered at pH 7.4 with MOPS. Compound
Log Ka (stoichiometry)
Absorbance
max (nm) DCHQ5
DCHQ5·Mg2 +
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
DCHQ5·Zn2 +
DCHQ5·Cd2 +
DCHQ5·Hg2 +
DCHQ5·Cu2 +
DCHQ5·Ni2 +
—
5.08 ± 0.06 (ML)
6.6 ± 0.2 (ML)
9.4 ± 0.2 (ML2)
10.0 ± 0.1 (ML2)
9 ± 1 (ML)
10.8 ± 0.1 (ML2)
DCHQ5·Na +
No complexation
DCHQ5·K +
No complexation
DCHQ5·Ca2 +
No complexation
DCHQ5·Fe2 +
No complexation
DCHQ5·Fe3 +
No complexation
Fluorescence
ε (per
M/cm − 1)
249
81,100
330
9,000
265
39,600
390
3,300
263
66,500
385
7,300
263
114,000
390
12,900
266
110,000
390
12,900
265
66,000
395
7,400
266
124,800
395
15,500
involving solvent molecules can also occur, further decreasing the emission quantum yield. Upon binding of Mg2+, the DCHQ5 emission band shifts to 517 nm, and it undergoes a marked enhancement with respect to the free ligand (Fig. 2). In fact, the fluorescence quantum yield of the Mg2+-bound ligand becomes 3.5 × 10−2, an expected result as both PPT and photoinduced electron transfer are inhibited by ion complexation22. Remarkably, DCHQ5 fluorescence is not substantially affected by the presence of many divalent cations (Table 1), including Ca 2+, the most important of these, or by pH changes within the physiological range. It is worth mentioning that complexation of Zn2+ and Cd2+ by DCHQ5 induces comparable relative fluorescence intensity changes and presents binding constants similar to those observed with the complexation of Mg2+. However, Zn2+ and Cd2+ are not expected to cause substantial interference, as the intracellular Zn2+ concentration is ~100 times lower than that of Mg2+, and the cadmium ion is not present in nonpoisoned cells.
max (nm)
512
3.0 × 10 − 4
517
3.5 × 10 − 2
551
1.4 × 10 − 2
544
2.1 × 10 − 2
—
