Fluorimetric Detection of Phosphates in Water Using a Disassembly Approach: A Comparison of FeIII-, ZnII-, MnII- and MnIII-salen Complexes

Article abstract of DOI:10.1002/zaac.202000045

Details of the reaction sequence used for the fluorimetric detection of phosphates by disassembly of transition metal Schiff base complexes were investigated for [Fe^III(salen)(H₂O)]⁺, [Zn^II(salen)], [Mn^II(salen)(H₂O)₂], and [Mn^III(salen)(H₂O)]⁺. The reactivity of these compounds towards phosphorus oxoanions of differing charge, number of donor atoms and steric hindrance was detected by UV/Vis and fluorescence spectroscopy in both aprotic organic and aqueous media. Selectivity of [Fe^III(salen)(H₂O)]⁺ towards pyrophosphate over all other tested phosphorus-containing analytes was strongly supported. [Zn^II(salen)] showed a faster reactivity but was much less selective. In contrast, [Mn^III(salen)(H₂O)]⁺ proved to be more stable than the iron complex but generally showed little reactivity towards phosphorus oxoanions. The influence of the charge of the central atom was investigated using the Mn^II analogue [Mn^II(salen)(H₂O)₂]. As expected, the reduced charge resulted in a reactivity comparable to the Zn^II complex in organic solution but lead to hydrolysis of the complex in water. Finally, the reaction products of [Fe^III(salen)(H₂O)]⁺ with phosphates were characterized by IR spectroscopy and mass spectrometry, providing further insights into the reaction mechanism of the disassembly process.

Full text of DOI:10.1002/zaac.202000045

Journal of Inorganic and General Chemistry
DOI: 10.1002/zaac.202000045
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
Fluorimetric Detection of Phosphates in Water Using a
Disassembly Approach: A Comparison of FeIII-, ZnII-, MnII- and
MnIII-salen Complexes
Daniela Winkler,[a] Sophie Banke,[a] and Philipp Kurz*[a]
Dedicated to Prof. Manfred Scheer, Connoisseur of Phosphorus Chemistry, on the Occasion of his 65th Birthday
Abstract. Details of the reaction sequence used for the fluorimetric contrast, [MnIII(salen)(H2O)]+ proved to be more stable than the iron
detection of phosphates by disassembly of transition metal Schiff base
complexes were investigated for [FeIII(salen)(H2O)]+, [ZnII(salen)],
complex but generally showed little reactivity towards phosphorus
oxoanions. The influence of the charge of the central atom was investi-
[MnII(salen)(H2O)2], and [MnIII(salen)(H2O)]+. The reactivity of these gated using the MnII analogue [MnII(salen)(H2O)2]. As expected, the
compounds towards phosphorus oxoanions of differing charge, number reduced charge resulted in a reactivity comparable to the ZnII complex
of donor atoms and steric hindrance was detected by UV/Vis and in organic solution but lead to hydrolysis of the complex in water.
fluorescence spectroscopy in both aprotic organic and aqueous media. Finally, the reaction products of [FeIII(salen)(H2O)]+ with phosphates
Selectivity of [FeIII(salen)(H2O)]+ towards pyrophosphate over all
were characterized by IR spectroscopy and mass spectrometry, pro-
other tested phosphorus-containing analytes was strongly supported. viding further insights into the reaction mechanism of the disassembly
[ZnII(salen)] showed a faster reactivity but was much less selective. In process.
Introduction
sensitivity even at the typically rather low phosphate concen-
Phosphates play very important roles in a number of funda-
trations found in biological or environmental samples (e.g.
mental cellular processes. Amongst others, phosphates are im-
4–
0.3–1.3 mM of free P2O7 in E. coli cells[11]). Considerable
portant constituents of nucleic acids, phospholipids and phos-
progress was made especially in the selective detection of or-
phorylated proteins.[1–3] Moreover, they are involved in energy
thophosphate (PO43–, Pi) and pyrophosphate (P2O74–, PPi), but
metabolism, signal transduction and the regulation of protein
also examples for the recognition of nucleotides or phosphory-
synthesis.[4–6] The possibility to detect and image phosphates
lated proteins have been published.[12,13]
and organophosphates in vivo has therefore become a major
The sensing mechanisms for most sensor complexes pub-
target in analytical biochemistry.[7,8] In addition to their bio-
lished to date are either irreversible reactions with the analyte
logical importance, phosphates are also key species in environ-
(chemodosimeters), reversible phosphate coordination (chemo-
mental monitoring, as, for example, eutrophication is often as-
sensors) or displacement reactions of the indicators or metal
sociated with the use of phosphate-containing fertilizers.[9]
ions themselves. Based on the principle of such a metal dis-
placement assay, Zelder and co-workers have published an
However, a number of challenges in designing sensors for
phosphates have to be met, which often originate in their high
interesting approach for the selective detection of phosphates,
hydration energies and the presence of different protonation
which includes the hydrolysis of free H2salen ligands after de-
states in aqueous solution.[10]
metallation as central step (Scheme 1).[14,15] In this method,
Certain transition metal complexes with exchangeable
ligands exhibit a high affinity to phosphate anions and can thus
be employed as sensor molecules. The signaling unit in these
complexes usually consists of a fluorophore to ensure high
* Prof. Dr. Ph. Kurz
E-Mail: philipp.kurz@ac.uni-freiburg.de
[a] Institut für Anorganische und Analytische Chemie
Albert-Ludwigs-Universität Freiburg
Albertstr. 21
79104 Freiburg, Germany
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/zaac.202000045 or from the au-
thor.
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH &
Co. KGaA. · This is an open access article under the terms of the
Creative Commons Attribution License, which permits use, distri-
bution and reproduction in any medium, provided the original work
is properly cited.
Scheme 1. Proposed mechanism of the disassembly approach for pyro-
phosphate sensing using salen2– complexes.
Z. Anorg. Allg. Chem. 2020, 646, 1–8
1
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
salen2– transition metal complexes undergo demetallations model anions, we aimed to examine the influence of charge
after binding of the phosphate analyte to the central metal and steric effects on the process of demetallation. The use of
atoms. This eliminates the stabilization of the imine functional the aprotic solvent dmso avoids (de-)protonation of the respec-
groups by the coordinated metal ions and therefore the free tive analytes. The reactions were monitored by UV/Vis spec-
H2salen molecule hydrolyses in water yielding ethylenedi- troscopy taking advantage of the characteristic absorption
amine and salicylaldehydes. The latter show intensive green bands of H2salen, 1 and 2 (see Figure 1 and Figure S1, Sup-
fluorescence in aqueous solutions and in this way serve as porting Information). In agreement with the literature, an in-
turn-on fluorescence signaling units.[16]
tense band at 315 nm (π–π* transition) and a lower intensity
Thus far, a reactivity of ZnII- and FeIII-salen2– complexes band at 410 nm (n–π*, involving the free electron pair of the
towards various phosphates has been reported and the iron spe- imine group) could be observed for H2salen in dmso.[23] For
cies could be successfully applied for the detection of PPi in 2, the π–π* transition is shifted to 355 nm,[24] whereas 1 also
mitochondria and foodstuff.[15,17] Furthermore, the possibility exhibits an intense ligand to metal charge transfer (LMCT)
to tune selectivity and reactivity of FeIII-salen2– complexes by band at 485 nm.[25]
modifying the ligand was demonstrated. In the study presented
here, we extend this concept to MnII and MnIII complexes of
salen2–, as the coordination of phosphorus oxoanions to
MnIII-salen2– compounds is well known from the context of
molecular magnetism.[18] By examining two different
oxidation states of the same metal and comparing them to the
two established sensors, we aimed to systematically evaluate
the influence of the central ion on the reactivity of [M(salen)]
towards phosphates. Additionally, we made an attempt to
identify the metal-containing reaction products of the reaction
sequence, as we suspected that the thermodynamically favored
formation of metal-phosphate oligomers or precipitates (de-
noted as “Mx(PPi)y” in Scheme 1) is a major driving force for
the overall reaction.
Results and Discussion
The
compounds
[FeIII(salen)(H2O)]NO3
(1·NO3),
[ZnII(salen)] (2), and [MnIII(salen)(H2O)]ClO4 (3·ClO4) can be
readily obtained from established syntheses involving the free
H2salen ligand and the corresponding metal salts.[19–22] The
identity and purity of the complexes was verified by elemental
analysis, mass spectrometry, IR- and UV/Vis spectroscopy as
well as NMR spectroscopy (for 2 only).
In order to confirm the already known reactivity of the phos-
phate sensor molecules 1 and 2, solutions of pyrophosphate,
orthophosphate and the related organic compounds methyl-
and phenylphosphonic acid (H2MePO3, H2PhPO3) as well as
diphenylphosphinic acid (HPh2PO2) were added to dmso solu-
Figure 1. UV/Vis absorption spectra of 1 (50 μM, top) and 2 (20 μM,
bottom) in dmso without and with added phosphate analytes (10 eq
added, reaction time 1 min). As a reference, the UV/Vis spectrum of
H2salen (50 μM) in dmso is also shown in both figures.
In dmso solutions, an appearance of the π–π* band of free
tions of 1 and 2, respectively. The structures of the different H2salen is thus expected for cases where a demetallation of
phosphorus oxospecies are shown in Scheme 2. By using these the compounds occurs. This was indeed found for 2, where a
complete shift of the π–π* absorption maximum from 355 nm
to 315 nm was detected upon addition of almost all tested
phosphorus species. However, when Pi was added, only a de-
crease in intensity of the original bands was observed with a
simultaneous increase of absorptions at lower wavelengths,
which may be due to the coordination of phosphate to the com-
plex. A titration of 2 with PPi revealed that the demetallation
requires a 1:1 reaction stoichiometry of phosphate anion to
complex in order to reach a complete turnover (see Figure S2,
Supporting Information).
In comparison to 2, complex 1 has been reported as a selec-
tive chemosensor for PPi over Pi as well as other phosphate
oxoanions such as ATP.[15] Fittingly, a demetallation of 1 by
Scheme 2. Structures of the phosphorus oxospecies used as analytes.
PPi was revealed by an immediate decrease of the LMCT band
Z. Anorg. Allg. Chem. 2020, 1–8
www.zaac.wiley-vch.de
2
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
at 485 nm in the UV/Vis spectrum and the appearance of the
signals of uncoordinated H2salen (Figure 1). On the other
hand, when Pi was added, this band was retained, but shifted
to ca. 410 nm, which might again indicate the coordination of
Pi (as found for 2 before) and / or also a deprotonation of the
axial aqua ligand to a hydroxide ligand. Unlike Pi or PPi, none
of the three organic P-acids used in this study induced any
changes in the UV/Vis spectrum, so that our results for dmso
solutions overall support the previously described selectivity
for PPi.
For aqueous solutions at physiological pH, it is known that
a hydrolysis of the free H2salen ligand takes place if the stabi-
lizing central metal cation is removed. This process can also
be monitored by UV/Vis spectroscopy and we therefore re-
peated the previous experiments in a buffered aqueous solution
(10 mM Tris buffer, pH 7.4). The changes in absorbance were
monitored directly after addition of the analytes and after
30 min of incubation time since the hydrolysis of free ligand
under the same experimental conditions proved to be complete
Figure 2. Spectral changes of 1 (50 μM in a 50:1-mixture of 10 mM
aqueous Tris buffer (pH 7.4) and dmso) after the addition of 10 eq of
PPi over a reaction time of ca. 1.5 h).
by this time (see Figure S3, Supporting Information). For solu-
In comparison to PPi, only minor changes of the UV/Vis
spectrum were detected when Pi, (PO3)3 or IP6 were added
3–
bility reasons, cyclotriphosphate [trimetaphosphate, (PO3)3
]
3–
and inositol hexaphosphate (IP6, Scheme 2) could be tested in
these aqueous experiments as well while reactions with the
rather hydrophobic HPh2PO2 ligand were not possible. IP6 is
an organophosphate that is relevant in numerous cellular pro-
cesses, which are to date not fully understood and is conse-
quently an analytical target of great interest.[26] We therefore
tested it in the phosphate addition studies although it is notice-
ably different from the other analytes as it is a much more
complex molecule with six phosphate moieties attached to a
six-membered cyclic alcohol.
to 1 in water, and even these required long reaction times of
more than one hour. The addition of the two phosphonic acids
lead to a slight increase of the charge-transfer band, again
possibly indicating a coordination of these compounds to the
complex.
The described reactivity of 1 and 2 with different phos-
phorus oxoanions could also be monitored by fluorescence
spectroscopy. When excited at 350 nm, the salicylaldehyde ob-
tained from the hydrolysis of H2salen shows strong green fluo-
rescence with a maximum of 490 nm in water at pH 7.4 (see
Figure S6, Supporting Information). The sensor complexes 1
For 2 in aqueous buffer, only the additions of pyrophosphate
or IP6 lead to immediate shifts of the absorption bands to and 2 themselves have quite different fluorescence properties
when excited with light of this wavelength: while the fluores-
cence of 1 is very low, the zinc compound 2 exhibits a very
pronounced emission in Tris buffer at 455 nm (see Figure 3
and Figure S6, Supporting Information). These differences can
be explained by a paramagnetic quenching effect of Fe3+,
which is absent for the diamagnetic ZnII complex.[27,28]
In agreement with the UV/Vis experiments, the highest in-
crease in salicylaldehyde fluorescence intensity was observed
400 nm corresponding to the formation of free H2salen. On the
other hand, profound UV/Vis changes could be observed for
all reaction combinations involving 2 in water at pH 7.4 after
30 min of reaction time; even in the absence of phosphorus
species (see Figures S4 and S5, Supporting Information). It can
be concluded that 2 is not suitable for a use as sensor molecule
in aqueous media under physiological conditions, as it showed
no selectivity for any of the tested P-compounds.
3–
when 1 had reacted with PPi for about 1 h (Figure 3). (PO3)3
In the case of the reaction of 1 with PPi, a marked decrease
of the LMCT band intensity was observed after the addition
of PPi, indicating that the decomplexation of Fe3+ can also
occur in water. However, the reaction appears to be slower
than in dmso since the maximum intensity of the absorption
band of the free ligand at 400 nm was only observed after a
reaction time of 5 min (Figure 2). For even longer reaction
times, this band is slowly moved to a new maximum at
385 nm, matching the signal for salicylaldehyde, one of the
expected hydrolysis products of H2salen (see Figure S3, Sup-
porting Information). A possible explanation for the longer re-
action time could be that the formation of insoluble iron pyro-
phosphates is a major driving force of the reaction and this
and IP6 also lead to a slightly higher fluorescence which con-
tinued to rise even after 90 minutes of reaction time. By con-
trast, PPi has a strong quenching effect of the fluorescence of
2 directly after addition due to the formation of non-fluo-
rescent free H2salen which was also observed for (PO3)33– and
IP6 but to a smaller extent (Figure 3).
Several manganese containing metalloenzymes are known
where organic phosphates bind to Mn cations as part of the
catalytic cycle.[29–31] Given the marked reactivity differences
found before between [FeIII(salen)(H2O)]+ and [ZnII(salen)],
we therefore synthesized and tested the suitability of the ana-
logous MnIII complex [MnIII(salen)(H2O)]+ (3) for phosphate
detection. UV/Vis and fluorescence monitoring of analyte ad-
might be more favored in less polar aprotic solvents like dmso. ditions were carried out as described above for 1 and 2. How-
Z. Anorg. Allg. Chem. 2020, 1–8
www.zaac.wiley-vch.de
3
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
Figure 4. Changes in fluorescence intensity at 490 nm of 1 and 3 and
460 nm of 2 (λex = 350 nm) after addition of analytes (10 eq reaction
time 30 min) in a 50:1 mixture of 10 mM Tris, pH 7.4 and dmso.
complex 2, analyte addition generally leads to a quenching
rather than an increase of the fluorescence intensity in this
case. However, these changes are too small to be of use as a
potential analytical tool.
Figure 3. Emission spectra of 1 (50 μM, top) and 2 (10 μM, bottom)
in a 50:1 mixture of 10 mM Tris, pH 7.4 and dmso without and with
added analytes (10 eq, reaction time 30 min, λex = 350 nm).
Comparison of the different substrates suggests that the
ever, no significant changes of the UV/Vis spectra were ob- selectivity for a particular analyte depends strongly on the sta-
served for reactions of 3 with the set of P-analytes, neither bility of the formed phosphate salt or complex. This trend is
in organic nor in aqueous medium (see Figure S7, Supporting observed in particular for the reactions with 1. The high charge
Information). In the fluorescence spectra, slight increases in density of PPi and its structure probably also contribute to the
intensity were observed, but the fluorescence intensity selectivity for this phosphate.[36] The number of available oxy-
remained fairly low and no selectivity for a particular analyte gen donor atoms plays a minor role since the observed signal
was observed (see Figure 4). The coordinative binding of intensities were significantly smaller for (PO3)3– and even for
H2salen to 3 appears to be too strong for decomplexation to IP6 which contains six phosphate groups.
take place. A likely explanation for this is the significant
Having established the surprisingly high stability of the
ligand field stabilization energy expected for Mn3+ (d4) in a MnIII-salen2– complex 3, which clearly makes it unsuitable for
distorted square-pyramidal coordination arrangement, which is fluorimetric phosphate sensing, we wondered whether the
lacking for the FeIII complex 1 (hs-d5). Hence, 1 might be analogous MnII compound might be a better choice for this
slightly less stable than 3 and can therefore be selectively de- task. Due to the lower charge of the central ion and the absence
metallated by PPi and (to a lower extend) IP6 and (PO3)3–, all of ligand field stabilization for this high-spin d5 ion, a much
of which are also known to be good chelating agents for weaker metal–ligand interaction is expected, which should fa-
iron.[32] Furthermore, the observed increasing stability from cilitate the decomplexation by phosphates. This matches the
ZnII over FeIII to the MnIII complex of H2salen matches the situation in Mn-containing metalloenzymes for phosphate-
trend found for metal-ligand bond lengths from X-ray crystal turnover, where the oxidation state of Mn is also believed to
structure experiments. For example, the average distances be- be +II in all cases. Finally, in contrast to the ZnII complex 2,
tween the metal ion and the phenolate oxygen atoms is 1.96 Å whose central ion has the same charge and similar ionic radius,
for the ZnII complex,[33] but 1.88 Å for FeIII[34] and MnIII.[35] the MnII complex is expected to be paramagnetic and this
Similarly, the metal to imine nitrogen distances decrease from could be beneficial for a use as “turn-on” fluorescence probe.
2.11 Å for ZnII via 2.10 Å for FeIII to a MnIII–N distance of Due to a quenching effect of paramagnetic metal cations the
1.98 Å.
FeIII and the MnIII complex both show significantly lower fluo-
Figure 4 summarizes our results on fluorescence intensity rescence intensities than ZnII-salen2–, which leads to a lower
changes observed after the addition of the six different P-ana- background fluorescence of the probe before phosphate ad-
lytes to solutions of complexes 1–3 in water. The d10 ion Zn2+ dition. We expect to observe a similar effect for MnII.
forms the least stable complexes with H2salen due to the lower
With the aim to first establish a feeling for the oxidation
charge and no ligand field stabilization. 2 therefore reacts sensitivity of such a [MnII(salen)]-species, cyclic voltammo-
faster, but less selectively, with phosphorus oxoanions than 1. grams of the MnIII complex 3 were recorded in dmso, where
Owing to the high intrinsic fluorescence intensity of the ZnII a reversible reduction event was observed at –650 mV vs. Fc/
Z. Anorg. Allg. Chem. 2020, 1–8
www.zaac.wiley-vch.de
4
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
Fc+ (Figure 5, inset). With a peak split of ca. 90 mV and a
pc / ipa ratio very close to unity, this reduction process can be
i
assigned with great certainty to the reduction of 3 to a
[MnII(salen)] species, which in water we propose to be an octa-
hedral complex of the composition trans-[MnII(salen)(H2O)2]
(4). Furthermore, the reduction of 3 was monitored by spectro-
electrochemistry using a modified 1 mm UV/Vis cuvette with
a platinum net as working electrode (Figure 5). When applying
a constant potential of –850 mV vs. Fc/Fc+, we observed a
decrease of the LMCT band at 400 nm together with a shift of
the π–π* band at 310 nm to 360 nm. The final spectrum thus
shows the π–π* transition of the salen2– ligand as only promi-
nent feature, while LMCT and / or d-d absorptions are absent
(as expected for a hs-d5 MnII complex). Isosbestic points at
340 nm and 395 nm confirm the reduction of the central ion
with no further side reactions.
Figure 6. UV/Vis absorption spectra of 4 (50 μM) in dmso without
and with added phosphate analytes (10 eq added, reaction time 1 min).
free H2salen for all analytes except Pi (Figure 6). This con-
firmed that as intended the lower Mn oxidation state indeed
led to a decrease of the stability of the complex, which strongly
facilitates decomplexation in comparison to the MnIII complex
3.
However, for aqueous solutions the complex is now too
labile as we detect hydrolysis even in the absence of added
phosphates. The UV/Vis spectrum of 4 dissolved in aqueous
buffer just shows absorption bands at similar wavelengths as
free H2salen, which again are shifted towards the characteristic
absorption maxima of salicylaldehyde at 325 nm and 380 nm
within minutes (see Figure S8, Supporting Information). The
initial fluorescence spectrum in water shows a broad and weak
emission between 440 nm and 530 nm. After about 5 min, this
band is not visible any more, as the much more intense emis-
sion band at 490 nm (indicating the formation of salicylalde-
hyde) appears and again takes ca. 30 min to reach its maximum
intensity. Overall, we thus attribute the detected spectral
changes to the fast demetallation of complex 4 in aqueous
buffer followed by the hydrolysis of H2salen. Given this result,
it is also understandable that even though numerous entries in
the literature exist about MnII complexes of H2salen and
similar ligands including their properties in the solid state or
in organic solvents, there are to the best of our knowledge no
reports describing MnII-salen type complexes in water.
For potential applications of the sensor complex in vivo, it
is of interest to know the reaction product of the metal with
phosphates. We therefore repeated the previous experiments in
aqueous solution using higher concentrations in order to isolate
and characterize products formed after addition of phosphates
to 1 which were analyzed by IR spectroscopy, mass spectrome-
try and XRD. For the reaction with Pi, a red-orange precipitate
was isolated. The reaction with PPi yielded a pale yellow pre-
cipitate. None of the XRD diffractograms obtained for these
precipitates did feature any sharp reflections, indicating that
the solids are amorphous materials. The IR spectra of salen2–
Figure 5. Spectral changes of 3 (50 μM) during reduction at –850 mV
and cyclic voltammogram (scan rate: 100 mV·s–1, 1 mM complex) in
dmso with [Bu4N][PF6] (0.2 m) as electrolyte.
As the low reduction potential of 3 indicated that the corres-
ponding [MnII(salen)] complex should be quite oxidation sen-
sitive, the synthesis and handling of 4 had to be carried out
under the exclusion of air. As previously reported, MnII com-
plexes of salen type ligands can be isolated from the reaction
of MnII acetate and H2salen under anaerobic conditions in the
presence of a base[37,22] and this synthetic route could be suc-
cessfully reproduced. When a UV/Vis spectrum of the thus
obtained complex 4 was measured in dmso, it is virtually iden-
tical to the one observed at the end of the electrochemical re-
duction of 3 shown in Figure 5, thus confirming our previous
assignment of the Mn redox state.
The MnII complex 4 was now tested in the same phosphate
addition experiments monitored by UV/Vis spectroscopy as
described above for 1–3, with the only difference that argon-
purged solvents and a gas tight cuvette were used to prevent
oxidation. In dmso, 4 showed the same reactivity as 2, i.e. a complexes show characteristic bands for the ν(C–O) stretching
shift of the absorption maximum to 315 nm corresponding to vibration of the ligand. In the case of complex 1 this band is
Z. Anorg. Allg. Chem. 2020, 1–8
www.zaac.wiley-vch.de
5
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
located at around 1290 cm–1.[38] The spectrum of the precipi-
Conclusions
tate formed after addition of Pi to 1 exhibits a blueshift of this
band to 1305 cm–1 which may indicate a deprotonation of the
aqua ligand (see Figure S9, Supporting Information). This
analysis is supported by ESI-MS data showing signals at m/z
362.03 and 683.07 (with characteristic isotope patterns for one
or two iron atoms, respectively), which can be assigned to so-
dium adducts of [FeIII(salen)(OH)] and [FeIII2(salen)2(O)]
(Figure 7). As for the initial solution of 1 before addition of
Pi, the complex [FeIII(salen)]+ without any additional ligand is
observed at m/z 322.04.
The reactivity of FeIII-, ZnII-, MnII- and MnIII-salen2–
towards demetallation by different phosphorus oxoanions was
investigated using UV/Vis and fluorescence spectroscopy.
Based on the results, the suitability of the complexes as phos-
phate sensors using the presented disassembly strategy was
evaluated. [ZnII(salen)] and [MnII(salen)(H2O)2] showed a
high, but unselective reactivity towards all tested analytes in
UV/Vis experiments in dmso. However, the MnII complex was
not stable in aqueous buffer, making it unsuitable for an appli-
cation under physiological conditions. The ZnII complex was
more stable but still showed signs of hydrolysis after reaction
times of approximately 30 min. Furthermore, [ZnII(salen)]
possesses a much higher intrinsic fluorescence than the manga-
nese and iron complexes due to the diamagnetic nature of the
Zn2+ ion. As previously reported, [FeIII(salen)(H2O)]+ showed
a high selectivity towards PPi over Pi and other substrates. In
accordance with this, the phosphonic- and phosphinic acids
used in this study did not induce a detection signal and only
slow reactions were observed with good chelate ligands such
3–
as (PO3)3 and IP6. [MnIII(salen)(H2O)]+ did not react with
any of the phosphorus oxoanions tested.
Overall, the studied series of salen2– complexes with four
different metal centers points to the fact, that decomplexation
reactivity can often be correlated with stability trends expected
from charge and ligand field stabilization effects. This observa-
tion could serve as an indication for the design of further salen
based phosphate sensors. Besides the use of other metal ions,
substitutions at the ligand are conceivable, which could be de-
signed with the aim to (de-)stabilize the metal–salen interac-
tion. This was previously shown for FeIII-salen2– [17] and could
likewise be possible for the MnIII- and ZnII complexes.
Figure 7. ESI-MS (positive ion mode) of the product obtained from 1
(660 μM in in a 15:1 mixture of 10 mM Tris, pH 7.4 and dmso) after
addition of 5 equiv. Pi using H2O as the spray solvent.
Experimental Section
Although no significant changes were observed in the UV/
Vis- and fluorescence spectra, mass spectrometry of the solu-
tion of 1 after addition of phenylphosphonic acid showed a
signal at m/z 801.09 indicating the formation of a complex
consisting of two [FeIII(salen)]+ moieties bridged by one
HPhPO3 ligand (see Figure S10, Supporting Information).
This indicates that phosphonic acids can coordinate to (but not
demetallate) complex 1.
General: All reagents were used as purchased without further
purification. The syntheses of H2salen[41] and the complexes
[19,38]
[FeIII(salen)(H2O)]NO3
(1·NO3),
[ZnII(salen)][20]
(2),
[MnIII(salen)(H2O)]ClO4
(3·ClO4) and [MnII(salen)(H2O)2][37]
[21,22]
(4) were adapted from published procedures. All products were charac-
terized by elemental analysis, mass spectrometry, IR- and UV/Vis
spectroscopy as well as NMR spectroscopy in the case of the ligand
and 2 (see Figures S11 and S12, Supporting Information). Since the
–
The IR spectrum of the reaction product with PPi did not MnII complex is not stable in presence of O2, it was handled under
inert atmosphere in all experiments. All other complexes are stable
under aerobic conditions.
show any bands of C–H or N–H vibrations neither in the fin-
gerprint area nor at higher wavenumbers. The spectrum is
similar to iron pyrophosphates with a broad vibration band in
the area of PO3 stretching frequencies between 1000 cm–1 and
1200 cm–1 (see Figure S9, Supporting Information). Two
bands at 890 cm–1 and 750 cm–1 can be tentatively assigned
to the asymmetric and symmetric P–O–P stretching vibrations
respectively, which are typically observed for pyrophos-
phates.[39,40] This result confirms that the product is an inor-
ganic complex or salt of iron and PPi and is in accordance with
the reaction mechanism proposed by Zelder and co-workers.
However, overall these results clearly show the complex nature
of the analysis of reaction products of the metal with phos-
phates which will therefore demand detailed future studies.
UV/Vis and Fluorescence Sensing Experiments: UV/Vis absorption
spectra were recorded with a Varian Cary 50 spectrophotometer. Fluo-
rescence spectra were obtained with a Perkin–Elmer LS55 lumines-
cence spectrometer. Stock solutions of 1, 3, 4 (2.5 mM) and 2 (1 mM)
in dmso were added to dmso or buffer (10 mM Tris, pH 7.4). The final
concentration of complexes 1, 3 and 4 used for UV/Vis and fluores-
cence measurements was 50 μM in water/dmso (50:1, 10 mM Tris, pH
7.4) and dmso. The final concentration of 2 was 20 μM in UV/Vis and
10 μM in fluorescence experiments. Stock solutions of the analytes
dissolved in buffer (10 mM Tris, pH 7.4) or dmso were added after-
wards. Sodium salts of orthophosphate (Pi), pyrophosphate (PPi),
cyclotriphosphate [(PO3)33–] and inositol hexaphosphate (IP6) were
used for the experiments performed in aqueous medium. For studies
Z. Anorg. Allg. Chem. 2020, 1–8
www.zaac.wiley-vch.de
6
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
[7] J. Wongkongkatep, A. Ojida, I. Hamachi, Top. Curr. Chem. 2017,
in dmso, tetrabutylammonium phosphate, tributylammonium pyro-
375, 30.
phosphate as well as methylphosphonic acid, phenylphosphonic acid
and diphenylphosphinic acid were used. The excitation wavelength in
all fluorescence experiments was 350 nm.
[8] S. Lee, K. K. Y. Yuen, K. A. Jolliffe, J. Yoon, Chem. Soc. Rev.
2015, 44, 1749.
[9] M. Chen, T. E. Graedel, Glob. Environ. Chang. 2016, 36, 139.
[10] Y. Marcus, Biophys. Chem. 1994, 51, 111.
[11] J. K. Heinonen, H. L. Drake, FEMS Microbiol. Lett. 1988, 52,
205.
[12] F. J. Williams, D. Fiedler, ACS Chem. Biol. 2015, 10, 1958.
[13] H. Singh, S. Sreedharan, R. Tiwari, C. Walther, C. Smythe, S. K.
Pramanik, J. A. Thomas, A. Das, Cryst. Growth Des. 2018, 18,
7199.
Electrochemistry: Cyclic voltammograms of complexes 1, 3 (both
1 mM) and 2 (0.5 mM) in dmso with [Bu4N][PF6] (0.2 m) as electrolyte
were measured using a Biologic VSP potentiostat. The working electrode
was a platinum disc (d = 2 mm) and the counter electrode a platinum
wire. A Ag/Ag+ electrode (0.01 m in MeCN) was used as reference.
Ferrocene was added to calibrate the potential of the reference electrode.
[14] N. Kumari, F. Zelder, Chem. Commun. 2015, 51, 17170.
[15] N. Kumari, H. Huang, H. Chao, G. Gasser, F. Zelder, Chem-
BioChem 2016, 17, 1211.
[16] S. Hisaindee, O. Zahid, M. A. Meetani, J. Graham, J. Fluoresc.
2012, 22, 677.
[17] F. H. Zelder, G. Gasser, P. Yadav, M. Jakubaszek, B. Spingler, B.
Goud, Chemistry 2020.
[18] T.-T. Wang, M. Ren, S.-S. Bao, B. Liu, L. Pi, Z.-S. Cai, Z.-H.
Zheng, Z.-L. Xu, L.-M. Zheng, Inorg. Chem. 2014, 53, 3117.
[19] F. Lloret, M. Julve, M. Mollar, I. Castro, J. Latorre, J. Faus, X.
Solans, I. Morgenstern-Badarau, J. Chem. Soc., Dalton Trans.
1989, 729.
Spectroelectrochemistry: The spectroelectrochemistry experiment
was carried out using a Princeton Applied Research VersaSTAT 4 po-
tentiostat and a Varian Cary 50 spectrophotometer. A modified 0.1 cm
UV/Vis cuvette with a platinum net as working electrode, a platinum
wire as counter electrode and a Ag/Ag+ electrode (0.01 m in MeCN)
as reference electrode was used. A 0.5 mM solution of 3 in dmso (also
containing [Bu4N][PF6] (0.2 m) as electrolyte) was prepared in an ar-
gon atmosphere. The potential was set at –850 mV and UV/Vis spectra
were recorded at 30 s intervals.
X-ray Powder Diffraction: XRD analysis was performed using a
STOE STADI-P diffractometer with Cu-Kα radiation in transmission
mode. The samples (ca. 20 mg) were held by a polyacetate film in
round sample holders (inner diameter 15 mm). Data analysis was car-
ried out using the Match! software package (Version 3.3).
[20] A. Dalla Cort, L. Mandolini, C. Pasquini, K. Rissanen, L. Russo,
L. Schiaffino, New J. Chem. 2007, 31, 1633.
[21] P. Przychodzen´, K. Lewin´ski, M. Bałanda, R. Pełka, M. Rams,
T. Wasiutyn´ski, C. Guyard-Duhayon, B. Sieklucka, Inorg. Chem.
2004, 43, 2967.
[22] K. Srinivasan, P. Michaud, J. K. Kochi, J. Am. Chem. Soc. 1986,
108, 2309.
[23] L. Zhou, P. Cai, Y. Feng, J. Cheng, H. Xiang, J. Liu, D. Wu, X.
Zhou, Anal. Chim. Acta 2012, 735, 96.
[24] M. Cano, L. Rodríguez, J. C. Lima, F. Pina, A. Dalla Cort, C.
Pasquini, L. Schiaffino, Inorg. Chem. 2009, 48, 6229.
[25] H. Fujii, Y. Funahashi, Angew. Chem. 2002, 114, 3790.
[26] G. J. Miller, Inositol Phosphates. Methods and Protocols, Hu-
mana Press, New York, 2020.
[27] J. Cheng, X. Ma, Y. Zhang, J. Liu, X. Zhou, H. Xiang, Inorg.
Chem. 2014, 53, 3210.
[28] K. Kim, Y. Ha, L. Kaufman, D. G. Churchill, Inorg. Chem. 2012,
51, 928.
[29] G. Schenk, L. R. Gahan, L. E. Carrington, N. Mitic, M. Valizadeh,
S. E. Hamilton, J. de Jersey, L. W. Guddat, Proc. Natl. Acad. Sci.
USA 2005, 102, 273.
Supporting Information (see footnote on the first page of this article):
Additional UV/Vis and fluorescence spectra of complexes 1–4 and
phosphate addition experiments, IR and mass spectra of reaction prod-
ucts of 1 with analytes and NMR spectra of H2salen and 2.
Acknowledgements
Many thanks to Dr. Felix Zelder (Universität Zürich), Prof. Dr.
Henning Jessen (Albert-Ludwigs-Universität) and their respective
research groups for introducing us to the interesting topic of phosphate
sensing. Their valuable insights and experimental advice is highly
appreciated. Furthermore, this project profited from equipment and
laboratory procedures made available in our institute through the
BMBF cluster project MANGAN (FKZ 03SF0511A). We also thank
the Deutsche Bundesstiftung Umwelt and Elke Hanisch for providing
the illustration used for the table of contents figure.
[30] S. J. Charnock, G. J. Davies, Biochemistry 1999, 38, 6380.
[31] S. Jonas, B. van Loo, M. Hyvönen, F. Hollfelder, J. Mol. Biol.
2008, 384, 120.
[32] K. Subba Rao, B. S. Narasinga Rao, Food Chem. 1985, 17, 13.
[33] J. Reglinski, S. Morris, D. E. Stevenson, Polyhedron 2002, 21,
2175.
[34] G.-M. Yang, P. Cheng, D.-Z. Liao, Z.-H. Jiang, S.-P. Yan, G.-L.
Wang, X.-K. Yao, H.-G. Wang, J. Cluster Sci. 1998, 9, 465.
[35] H.-L. Shyu, H.-H. Wei, Y. Wang, Inorg. Chim. Acta 1999, 290, 8.
[36] D. H. Lee, S. Y. Kim, J.-I. Hong, Angew. Chem. Int. Ed. 2004,
43, 4777.
Keywords: Phosphate; Fluorimetric sensing; Disassembly
approach; Iron; Zinc; Manganese
References
[37] B. Mabad, P. Cassoux, J. P. Tuchagues, D. N. Hendrickson, Inorg.
Chem. 1986, 25, 1420.
[38] Y. Yang, J. Guan, P. Qiu, Q. Kan, Trans. Met. Chem. 2010, 35,
[1] M. W. Bowler, M. J. Cliff, J. P. Waltho, G. M. Blackburn, New J.
Chem. 2010, 34, 784.
[2] W. Dowhan, Annu. Rev. Biochem. 1997, 66, 199.
[3] C. T. Walsh, Posttranslational Modification of Proteins. Ex-
panding Nature’s Inventory, Roberts and Company Publishers,
Englewood, Colorado, 2006.
[4] R. Moreno-Sánchez, A. Marín-Hernández, E. Saavedra, J. P.
Pardo, S. J. Ralph, S. Rodríguez-Enríquez, Int. J. Biochem. Cell
Biol. 2014, 50, 10.
263.
[39] B. S. Parajón-Costa, R. C. Mercader, E. J. Baran, J. Phys. Chem.
Solids 2013, 74, 354.
[40] E. Steger, B. Käßner, Spectrochim. Acta Part A 1968, 24, 447.
[41] P. S. Hariharan, S. P. Anthony, Spectrochim. Acta Part A 2015,
136, 1658.
[5] J. Stebbing, L. C. Lit, H. Zhang, R. S. Darrington, O. Melaiu, B.
Rudraraju, G. Giamas, Oncogene 2014, 33, 939.
[6] P. W. Majerus, Annu. Rev. Biochem. 1992, 61, 225.
Received: January 31, 2020
Published Online: ᭿
Z. Anorg. Allg. Chem. 2020, 1–8
www.zaac.wiley-vch.de
7
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
D. Winkler, S. Banke, P. Kurz* ......................................... 1–8
Fluorimetric Detection of Phosphates in Water Using a Disas-
sembly Approach: A Comparison of FeIII-, ZnII-, MnII- and
MnIII-salen Complexes
Z. Anorg. Allg. Chem. 2020, 1–8
www.zaac.wiley-vch.de
8
© 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim