Selective pyrophosphate detection via metal complexes

Abstract

Pyrophosphate (PPi) anions are crucial in numerous biological and ecological processes involved in energy conversion, enzymatic reactions, and metabolic regulation along with adenosine. They are also significant biological markers for various processes related to diseases. Fluorescent PPi sensors would enable visual and/or biological detection in convenient settings. However, the current availability of commercial sensors has been limited to costly enzymes that are not compatible for imaging. Sensor development has also encountered challenges such as poor selectivity and stability, and limited practical applications. In this review, we analyze the situation of PPi sensing via commercial kits and focus on sensors that use metal complexes. We address their designs, sensing mechanisms, selectivities and detection limits. Finally, we discuss limitations and perspectives for PPi detection and imaging.

Graphical Absrtact

1. We focus on PPi sensors employed by metal complexes, while addressing their design, sensing, mechanism, selectivity and detection limit.

2. In this review, we analyze the situation of commercial kits of PPi sensing.

3. We discuss the limitations and perspectives, providing the supports to PPi and its related molecular imaging.

1. Introduction

Anions are widely present in organisms and the environment, and are critical in the life sciences, environmental sciences, and medicine (1–3). Pyrophosphate (PPi) anions (Fig. 1) are a byproduct of adenosine 5’-triphosphate (ATP) hydrolysis in living cells, and participate in numerous energy transformations and enzymatic reactions to regulate metabolic processes (4–6). The PPi concentration is essential for DNA replication, and stable metabolic processes ensure normal transport in biological functions (7,8). Deviations from optimal PPi concentrations can cause various diseases. For example, a concentration higher than 1.62 mM can cause painful arthritis (9), cartilage fracture, and pyrophosphate arthritis. Whereas, concentrations below 0.85 mM can lead to diseases such as vascular calcification caused by the synthesis of vascular calcification inhibitors (10). Moreover, PPi can be a biomarker for various diseases, including calcium pyrophosphate dihydrate crystal deposition (11,12), arthritis (13), chondrocalcinosis (14), or cancer (15). Recently, non-invasive fluorescence sensors have become widespread in biomedical diagnoses and treatment research, and are expected to be clinical, radiological, and imaging tools (16–18).Therefore, it is crucial to develop and promote commercial PPi detection kits for these applications (19–21).

Fig. 1.

Structures of PPi, ATP, ADP, and cAMP

2. Commercial kits for sensing pyrophosphate (PPi)

Various methods have been used for PPi sensing, such as chromatography (22), voltammetry (23), colorimetry (24), and fluorescence assays (16,25,26). However, there are limitations and challenges that need to be addressed to ensure precise and reliable PPi detection (27). In particular, there is an urgent requirement for commercial kits for medical applications. Here, we discuss three available commercial kits for sensing PPi, all of which use indirect detection.

Commercial kit 1 (28,29) (kit 1, Fig. 2):

Fig. 2.

Detection mechanism of the PiPer pyrophosphate assay

The PiPer^™ PPi test kit contains the reagent 10-acetyl-3,7-dihydroxyphenoxazine. As shown in Fig. 2, the specific mechanism decomposes PPi maltose phosphorylase and glucose oxidase into hydrogen peroxide (H₂O₂) in turn. Then, 10-acetyl-3,7-dihydroxyphenoxazine and H₂O₂ react to produce resorcinol. Indirect detection of PPi was achieved by observing the resorcinol. By using fluorescence, the kit can detect 0.8-μM PPi, or 3 μM by absorption. Hence, the assay can be performed either fluorometrically or spectrophotometrically. However, inorganic phosphate will interfere, and glucose contamination of the reagents is a potential problem. Thus, it is not suitable for cell imaging.

Commercial kit 2 (30–32):

The main component of the EnzChek^™ pyrophosphate test kit (kit 2, Fig. 3) is 2-amino-6-methio-7-methylpurine ribonucleoside. The method converts PPi into two equal amounts of Pi, which are then consumed by the reaction of 2-amino-6–6methio-7–7methylpurine nucleoside and purine nucleoside phosphorylase (PNP in Fig. 3). The indirect detection of PPi is realized by increased absorbance at 360 nm. The sensitivity limit is approximately 1 μM (~0.2 μg/mL). However, the storage conditions and detection environment are very strict (below −20 °C). There are two enzymes involved, which is costly and not suitable for cell imaging. Moreover, if there is Pi contamination in the indirect testing, the results are unreliable.

Fig. 3.

Detection mechanism of the EnzChek^™ pyrophosphate assay.

Commercial kit 3 (33,34):

The PPiLight^™ test kit is a widely used non-radioactive bioluminescent assay. The amount of light produced is directly proportional to the PPi concentration. As shown in Fig. 4, PPi is a substrate that promotes the conversion of AMP to ATP. Light is then produced from the newly formed ATP and luciferin via luciferase. The PPi sensitivity is typically 0.02 μM, and the kit can detect high light output within 0.1s integration time. However, because the luciferase-catalyzed reaction requires Mg²⁺as a cofactor, chelating agents such as ethylenediamine tetra-acetic acid will interfere at high concentrations. In addition, PPi can be newly generated in the presence of pyruvic acid phosphate dikinase. The reaction is complex and the reaction time may be too long to obtain rapid detection.

Fig. 4.

Detection mechanism of the PPiLight^™ Assay

As shown in Figs. 2–4 and Table 1, all three commercial kits involve indirect PPi detection based on enzymatic reactions, with high costs, that are not suitable for cell imaging and easily contaminated with low sensitivity. Hence, a direct measurement of PPi concentrations with an appropriate readout is important. In this review, we provide a comprehensive summary of the most promising PPi sensing platforms based on metal complexes, while also addressing their limitations and future prospects.

Table 1.

Commercial kits for PPi detection

Commercial kits	Assays	Detection	Optical information (nm)	Detection range
Kit 1	Fluorescence or spectrophotometry	indirect detection	λex/λem 563/587	≥0.8 μM(fluorescence) ≥3 μM(absorption)
Kit 2	spectrophotometry	indirect detection	Absmax 360	1 μM ~ 75 μM (1 mL reaction volumes)
Kit 3	non-radioactive bioluminescent	indirect detection	-	0.02 μM ~ 10 μM (100 μL sample)

3. PPi sensing via metal complexes

The design of probes for sensing ATP, ADP, and PPi anions is much more challenging than those for cation detection (35). Among various strategies (36–38), metal complexes have emerged as the most successful for selective recognition of specific phosphate anions. This has been primarily based on the strong affinity between these anions and metal cations, such as Zn²⁺. However, this affinity can be quite challenging because of structural similarities among different anions. (Fig. 1). Thus, the first crucial step will be ligands that selectively bind PPi in the presence of ATP, ADP, and other molecules. An additional challenge lies in accurately detecting changes in the signal induced by the binding process, especially fluorescent turn-on effects.

Thus, it is important to consider the concentrations of these anions in biological samples. For example, the normal plasma PPi concentration is 2.18 μM, with a range (95% confidence limits) of 0.58–3.78 μM (39). The physiological concentration of ATP in the human body ranges over 2–8 mM (7,40); but, in certain cells and tissues, it falls within the range of 1–10 μM (41). The ADP concentration has been reported to be approximately in the μM range (42,43).

Despite these challenges, significant progress has been made in inorganic PPi sensing (44–48). Here, the focus will be on PPi sensors that utilize metal complexes and their structural design, sensing mechanism, selectivity, detection limit, and potential applications. We will present a comprehensive understanding of how these sensors can detect and quantify PPi in various biological and environmental samples, along with inherent challenges and limitations, and possible improvements. The overall objective is to create highly selective and specific PPi sensors for reliable and precise detection in aqueous buffers and biological samples. (Table 2).

Table 2.

Summary of recent PPi-selective and responsive probes (No.1–4 and No.6–9)

Numbers Sensors	Chemical Structures	Medium	λex/λem (nm)	LOD (Limit of detection)	Applications	Ref.
No.1		HEPES buffer (10 mM, pH 7.4)	440/591	0.8 nM	Hela cells imaging	(49)
No.2		HEPES buffer (10 mM, pH 7.4)	400/580	2×105 nM	Hela cells imaging	(50)
No.3		HEPES buffer (10 mM, pH 7.4, H2O/DMSO, 7/3, v/v)	400/515	5.37 nM	Nucleus staining Cells imaging	(51)
No.4		PB buffer (10 mM, pH 7.0)	419/525	75 nM	Synovial fluid samples	(52)
No.5		HEPES buffer (pH 7.4)	400/595	-	Hela cells imaging/Locate lysosomes	(53)
No.6		HEPES buffer (10 mM, pH 7.4)	-	-	-	(54)
No.7		HEPES buffer (10 mM, pH 7.4)	280/316	about μM	-	(55)
No.8		HEPES buffer (10 mM, pH 7.4)	316/383	95 nM	-	(56)
No.9		HEPES buffer (50 mM, pH 7.4).	561/639	43 nM	Human lung cells/Mices image	(57)

3.1. Mononuclear metal complex sensors

It is evident that the binding of mononuclear metal complexes to ATP, ADP, and PPi via two phosphate coordination sites may limit their selectivity (2,53,58). However, the fluorescent turn-on effects exhibited by terpyridine, and its significant selectivity over ADP and ATP in HEPES buffer, have stimulated further studies (49). HeLa cell imaging using this probe revealed stained nuclei (Table 2, No. 1). Terpyridine was first reported in 2011, and its zinc complex was subsequently found to exhibit fluorescence turn on and partial selectivity for PPi, AMP, and ADP under various conditions (59). Several groups reported similar terpyridine derivatives for PPi sensing (60–64). Here, we analyzed these derivatives to further understand the unique sensing mechanism (Table 2, No.1–4).

As shown in Fig. 5, sensor 1 (terpyridine) exhibited an approximately 500-fold increase in fluorescence intensity at 591 nm in the presence of PPi in a buffered solution of HEPES, with a PPi detection limit approaching 0.8 nM. No significant changes in fluorescence intensity were observed when the buffer contained various anions, such as PO₄^3-, HPO₄^2-, CO₃^2-, SO₄^2-, NO₃⁻, AcO⁻, F⁻, Br⁻, and I⁻ (Fig. 5a), which indicated that the two coordination sites of the anions were stronger than those for monovalent anions interacting with the mononuclear metal complex. Furthermore, the structurally similar ATP and ADP molecules, which also have more than two coordination sites, only caused a slight or negligible fluorescence enhancement (Fig. 5b).

Fig. 5.

(a) Emission of sensor 1 (60 nM) in the presence of various anions (10 μM); (b) Emission of probe 1 (200 nM) in the presence of PPi, ATP, ADP, CTP, and GTP (200 nM); (c) Emission of sensor 1 under a 365-nm ultraviolet lamp with different anions (15 μM)(49). Copyright American Chemical Society, 2014.

Regarding the PPi selectivity, a binding stoichiometry of 3:1 (sensor 1: PPi) was proposed, where one PPi was bound to three ligands, creating a specific binding “pocket” for PPi. This binding mode likely contributed to the significant PPi selectivity over ADP and ATP. The limited number of anions in ADP (three phosphate groups) restricted its binding to metal complexes that require multiple coordination sites. This structural difference between PPi (with four phosphate groups) and ADP accounts for the reduced binding and fluorescent response of ADP compared to PPi in the system. Although ATP has four anion coordination sites like PPi, it only exhibited a slight fluorescent turn-on (Fig. 5c) in the system. This limited response could be attributed to the larger size of ATP, which may hinder its optimal fit within the binding pocket of the sensor. The number and arrangement of coordination sites are important in determining the binding affinity and selectivity of the metal complex towards various analytes.

A crucial aspect to consider is that the system had two cationic charges (65) that affect the fluorescence response to PPi. Electrostatic interactions between the positively charged sensor and the negatively charged phosphate groups in PPi likely contributed to the preferential binding (26). However, the specific mechanism underlying the enhanced fluorescence in response to PPi binding should be investigated further. Understanding how the binding affects the sensor fluorescence will provide insights into the signaling mechanism, the factors responsible for the fluorescence enhancement, and the relationship between the binding interactions and the resulting fluorescence response.

The proposed mechanism explains the limited response to ATP and ADP and the specific binding mode in achieving such selectivity. The combination of both charge effects and the unique binding mode was likely responsible for the selectivity (66), and suggests new ways for designing and optimizing sensors for specific analytes. Therefore, we provide a summary below of published terpyridine derivatives for PPi sensing (Table 2, No. 2–4).

In addition to the Zn (II) probe, Kar et al. developed a triplet pyridine-Cd (II) complex sensor 2 (Table 2, No.2) for detecting PPi (50). As shown in Fig. 7(a,b), its fluorescence intensity at 580 nm was enhanced 40-fold in a HEPES buffer solution after the addition of PPi. The reported binding stoichiometry of 1:3 (PPi:sensor 2) had a limit of detection (LOD) of 2×10⁵ nM. This was not as sensitive as that exhibited by the Zn (II)-probe (LOD=0.8 nM)(49). Although the binding stoichiometry between PPi and sensor 2 was the same as that for sensor 1, there was no mention of the response to ATP, ADP, or other phosphate derivatives in the presence of PPi. Sensor 2 was used for PPi detection in HeLa cells, where most of the orange-yellow emission occurred in the nuclei (Fig. 8). Because of the high concentrations of ATP, GTP, and CTP, we highly recommend testing the probe responses to these anions. In any case, we conclude that that the Zinc (II) complex was superior to the Cd (II) complex.

Fig. 7.

(a) Emission spectra of sensor 2 (50 μM) with various anions. (b) Emission spectra of sensor 2 (50 μM) upon addition of PPi (200–460 mmol) (c) Emission of sensor 2 (50 μM) under a 365-nm ultraviolet lamp with various anions(50). Copyright Elsevier B.V, 2017.

Fig. 8.

Fluorescence microscopy images of (a) HeLa cells incubated with sensor 2, (b)a single HeLa cell(50). Copyright Elsevier B.V, 2017.

Aggregation-induced emission of terpyridine-Zn sensor 3 (Table 2, No. 3) was reported for selective detection of nano-molar levels of PPi (51). Sensor 3 initially exhibited weak 570-nm orange emission in HEPES buffer (pH 7.4). Upon addition of PPi, there was a blue shift with a significant enhancement of green emission at 515 nm (Fig. 9b), while ATP and ADP both induced slight fluorescence. Thus, the fluorescence activation effect of PPi on sensor 3 far exceeded those of ATP and ADP. It was speculated that there was a donor-acceptor (D-A) structure in which the carbazole group was the donor and the terpyridine-Zn (II) portion was the acceptor. When sensor 3 recognized PPi, ICT (intramolecular charge transfer) occurred. Because of the conjugated structure and its low solubility, the supramolecular complex (PPi and probe) easily aggregated, leading to intense luminescence emission (Fig. 10). The LOD was as low as 5.37 nM, and sensor 3 was brightest in the nuclei of the HeLa cells, similar to sensor 1 (49).

Fig. 9.

(a) Fluorescence spectra of sensor 3 (10 μM) upon addition of PPi (0–5 μM). (b) Fluorescence spectra of sensor 3 (10 μM) in the presence of anions (4 μM for PPi and 30 μM for other anions). (c) Corresponding photographs under a 365-nm ultraviolet lamp(51).(Ref 51 is licensed under a Creative Commons Attribution 4.0 International License)

Fig. 10.

Strategy for the detection of PPi in aqueous media (51). (Ref 51 is licensed under a Creative Commons Attribution 4.0 International License)

Chen, Shen, et al. fabricated an aggregated nanoparticles sensor 4 (Table 2, No. 4) with terpyridine-Cu²⁺ for PPi sensing (52). Upon addition of PPi, the fluorescence intensity increased 4.8-fold at 525 nm (Fig. 11a). When strong interfering anions of ATP, ADP, AMP, H₃PO₄⁻, HPO₄^2-, PO₄^3-and CO₃^2- were present, the fluorescence intensity did not change significantly, indicating good specificity (Fig. 11b). The authors investigated the detection mechanism (Fig. 11c). Upon addition of Cu²⁺ to the ligand, the strong green fluorescence at 421 nm was quenched because of an intramolecular PET (photoinduced electron transfer) effect, and the metal sensor 4 was transformed into nanoparticles. During PPi recognition, the PPi chelated with Cu²⁺. The 421-nm fluorescence did not recover as expected, but, instead, the nanoparticle aggregation induced a significant fluorescence enhancement at 525 nm, with a LOD as low as 75 nM. The aggregated nanoparticles formed by the ligand after probe recognition could improve both the sensitivity and selectivity. Although the Schiff base system was unstable, the Schiff base terpyridine ligand exhibited strong chelation with Cu²⁺, resulting in the formation of nanoparticle aggregates to detect PPi. Therefore, micro- and nano-probes based on terpyridine could be used for PPi sensing(67). In addition, sensor 4 was also used to detect PPi in synovial fluid, and thus displayed great potential for biological fluids. Therefore, non-invasive fluorescence sensors are expected to serve as auxiliary diagnostic tools for early detection of lesions.

Fig. 11.

(a) Fluorescence spectra of sensor 4 in the presence of PPi (0–100 μM); (b) Interference experiments: fluorescence intensities for different samples (50 μM) a-h: PPi, ATP, ADP, AMP, H₂PO₄⁻, HPO₄^2-, PO₄^3-, CO₃^2-. (c-d) Interaction mechanism of sensor 4 with PPi (52). Copyright Elsevier B.V, 2021.

Ahn et al. reported sensor 5 (Table 2, No.5) for ATP sensing for nuclei staining (53), which provided insights in the design of PPi sensors. Sensor 5 was weakly fluorescent in HEPES buffer (pH 7.4), but then displayed a strong “turn-on” fluorescence upon the addition of ATP (Fig. 12a). As shown in Fig. 12b, probe 5 had negligible responses to various nucleotides, oxygen compounds, anions, metal ions, reactive oxygen compounds, and bio-thiols. When treated with the homologous series PPi, UTP, GTP, TTP, and CTP, the fluorescence intensity increased 25- and 16-fold with the addition of ATP and ADP, respectively, while others exhibited no change. Apyrase catalyzes the hydrolysis of ATP that is not bound to probe 5 to AMP and PPi; thus, the addition of apyrase decreased the fluorescence (Fig. 12c). This indicated that PPi produced by hydrolysis did not interfere. Relative to probe 4, the results for probe 5 indicated that the binding mode between the probe and the target was critical for selective sensing. With the Zn site, probe 5 coordinated with pyridinamine and phosphate, and the base region of ATP overlapped the coumarin region of the probe (Fig. 12d). This may be attributed to the optimal coordination between the larger ATP and sensor 5, and the number and arrangement of coordination sites are key to determining the binding affinity and selectivity of the metal complex towards various analytes. The nucleus-targeting mechanism is depicted in Fig. 12e, where the probe senses nuclear ATP. Although sensor 5 does not detect PPi, it provides insights regarding the design of PPi sensors. For example, the structure of sensor 5 can be controlled by the linker between the luminescent groups and the recognition sites, which enhances the selective recognition of ATP (adenosine region and the phosphate ion). This actually showed that the “pocket” sensors were more sensitive to the smaller PPi(49).

Fig. 12.

(a) Fluorescence titration of sensor 5 (10 μM) with ATP (0–100 μM) in pH 7.4 HEPES buffer; (b) fluorescence response of sensor 5 toward various biologically relevant compounds; (c) fluorescence intensity changes vs. time when apyrase (0–1.0 U) was added to a solution of the probe (10 μM) saturated with ATP (100 μM); (d) detection scheme of ATP with sensor 5; (e) fluorescence images of PC3 cells co-incubated with sensor 5 (53). Copyright Wiley-VCH GmbH, 2023.

To conclude this section, mononuclear metal complexes may not be ideal candidates for PPi sensing, because they may lack sufficient selectivity (68–71). However, terpyridine derivatives are promising with significant fluorescent turn-on and moderate selectivity (72–74). Continued research may lead to enhanced selectivity and sensitivity for PPi, thus enabling a better understanding of its biological roles and potential diagnostic applications.

Two or three terpyridine units connected by a suitable linker could be a better alternative to single terpyridines for PPi sensing (Fig. 13). This may lead to the formation of a more stable PPi-metal complex, enhancing the selectivity and sensitivity. A linker between the terpyridine units could provide additional structural flexibility, enabling optimal binding interactions with PPi and potentially improving the overall performance.

Fig. 13.

Possible polymetallic ligands as PPi sensors

3.2. Binuclear metal-complex sensors

In previous reports (35–38), we showed that utilization of optimally positioned binuclear zinc complexes is a successful strategy for PPi sensing. The strong binding affinity of metal complexes towards aqueous PPi results in a stable complex for reliable sensing and accurate measurements of low-level PPi concentrations. This binding mode also enables tunable selectivity over ATP and ADP. Below, we introduce some bimetallic metal complex sensors (Table 2, No. 6–9).

Hong et al. reported a colorimetric sensor 6 (Table 2, No. 6) with two Zn²⁺-bis(2-pyridylmethyl) amine (DPA) units that recognize PPi in HEPES buffer (54). As shown in Fig. 14(a–c), the addition of PPi resulted in bathochromic absorbance shifts from 417 nm to 465 nm that changed the solution of sensor 6 from yellow to red. The binding mode was confirmed by the X-ray structure, which revealed that two sets of oxygen anions on each P atom of PPi were bound to the di-nuclear zinc complex. They acted as bridges between the two metal ions and resulted in the formation of two hexacoordinated Zn²⁺ ions. According to X-ray structural analysis, the six-coordinated Zn ion structure generated by the combination of the positively charged probes and the negatively charged PPi phosphate groups was more stable. This binding mode prompted further study of this system with regard to binding affinity, selectivity, and fluorescent responses.

Fig. 14.

(a) Absorbance changes of sensor 6 upon addition of PPi. (b) Ultraviolet-visible absorption spectra of sensor 6 in aqueous 10-mM HEPES buffer (pH 7.4) in the presence of various anions. (c) Color changes of sensor 6 in 10 mM aqueous HEPES buffer solution (pH 7.4). (d) Binding mode and crystal structure of the complex between sensor 6 and PPi (54). Copyright American Chemical Society, 2003.

A naphthalene-dpa unit sensor 7 fluorescent response to PPi was reported by Hong et al. (Table 2, No. 7) (55). When PPi was added to sensor 7, the fluorescent emission exhibited a red shift from 436 nm to 456 nm, and increased 9.5-fold (Fig. 15a). However, the fluorescence did not change in the presence of other anions such as CH₃CO₂⁻, HCO₃⁻, F⁻, and Cl⁻. The binding between sensor 7 and PPi and ATP had a 1:1 stoichiometry via a Job plot, and the association constant was (2.9±0.7)×10⁸ M⁻¹, and K_a for the ATP- sensor 7 was (7.2±1.0)×10⁶ M⁻¹. In contrast, the association constant between ATP and sensor 7 was 40-fold lower than that for PPi, which indicated that sensor 7 could detect nano-molar concentrations of aqueous PPi. Homologous ATP and ADP exhibited minimal interference in PPi detection (Fig. 15b). Therefore, the underlying factor was attributed to the charge density at the recognition region of the sensor 7 bimetallic complex. Despite the structural similarity between ATP and PPi, the total anionic charge density of the four O-P oxygen atoms involved in ATP complexation with sensor 7 was lower than that of the four O-P oxygen atoms in PPi. This significantly reduced the binding affinity of ATP, enabling specific PPi recognition. In the presence of large amounts of ATP (50–250-fold excess), sensor 7 was able to detect low nM concentrations of PPi in HEPES buffer. Consequently, the bimetallic sensor 7 can selectively detect PPi in water without interference from micromolar ATP, which enables biochemical applications and analytical enzyme assays involving ATP and PPi.

Fig. 15.

(a) Changes in fluorescence emission from sensor 7 (6 μM) upon addition of PPi (0–7.2 μM); (b) fluorescence emission spectra of sensor 7 (6 μM) in the presence of various anions (8 μM). (c) Proposed mechanism for complexation of sensor 7 with PPi and ATP(55). Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2004.

The introduction of several hydrogen-bonding sites could significantly increase the binding affinity by more than two orders of magnitude. Hong et al. (75) synthesized compound Zn₂L (Fig. 16a) that had four amides enabling additional hydrogen bonding with PPi, yielding pM binding affinity. This system exhibited the strongest PPi binding. Figure 16b depicts the X-ray crystal structure of the PPi complex, with additional hydrogen bonds formed by amide groups. The improved binding affinity (K_d was approximately 20 pM) was obtained with the hydrogen bonds with PPi coordinated by two Zn (II) ions.

Fig. 16.

(a) Structure of Zn₂L. (b) X-ray crystal structure of Zn₂L-PPi (75). Copyright American Chemical Society, 2007.

Fiedler et al.(76) demonstrated that the double-zinc complexes a-c (Fig. 17i) exhibited strong PPi binding. Compared to compounds a-c (Fig. 17ii), which incorporate smaller hydrogen-bond donors, complexes with high selectivity and affinity toward ADP were obtained. The selectivity for the diphosphate ester moiety was significant for selective fluorescence detection of PPi.

Fig. 17.

i) Structures of compounds a-c, and ii) association constants for zinc complexes with various anions (76). Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2015.

Another strong-binding (Table 2, No. 8) for PPi with DPA-Zn (II) was sensor 8, as reported by Yu et al. (56). It exhibited weak fluorescence emission at 383 nm. However, the fluorescence increased almost 4-fold for the free probe in the presence of PPi, and reached saturation at 3 μM (0.3 equiv.) (Fig. 18a). The fluorescence intensity did not change significantly when anions such as PO₄^3-, H₂PO₄⁻, SO₄^2-, HCO₃⁻, NO₃⁻, AcO⁻, F⁻, Br⁻, I⁻, SCN⁻, and N₃⁻ were present. When the sensor 8 concentration was increased from 1.0×10⁻⁵ to 2.5×10⁻³M, PPi could be detected visually via precipitation, with a detection limit of 95 nM. The PPi recognition was tracked using scanning electron microscopy. Upon the addition of PPi, rough and scaly solids with large surface areas confirmed the interaction between sensor 8 and PPi. The PPi selectivity was attributed to the rigidity and distance between the two DPA–Zn (II) units. However, the selectivity for PPi over ATP (4.1/2.8) was not as significant as that for sensor 7 (55). ATP also produced an approximately three-fold fluorescent turn-on effect. Nuclei staining was demonstrated in HeLa cell imaging (Fig. 18d). Thus, in addition to four amide hydrogen bonds to improve binding affinity, the distance between the binuclear recognition units in sensor 8 could be controlled to obtain specific PPi recognition.

Fig. 18.

(a) Fluorescence emission spectra of in situ prepared sensor 8 (10 μM) upon addition of PPi (0–4μM) Fluorescence responses of sensor 8 towards various anions (100 μM, 10 equiv.), (b) and the presence of analogs anions (0.275 mM.) (c) Structures of chemo sensors and their proposed PPi binding modes; (d) Fluorescence images of Hela cells treated with sensor 8 followed by 20 μM of PPi(56). Copyright The Royal Society of Chemistry, 2015.

Feng et al. developed a benzopyran- and phenol-bridged dipicolylamine (DPA) near-infrared fluorescent sensor 9 for PPi (Table 2, No. 9) (57). The PPi binding group was quite strong and sensor 9 exhibited peak fluorescence emission at 639 nm. When the PPi concentration was increased from 0 to 0.5 equiv., the fluorescence intensity decreased; then enhanced (>six-fold) emission at 654 nm was observed after four equivalents of excess PPi were added to the solution (Fig. 19a). These characteristics may introduce complexities for in vitro imaging. The selectivities over ATP and ADP were similar to those of sensor 7. The fluorescence quenching and enhancement were attributed to PET between the DPA nitrogen atoms and the phenyl ring, and ICT between the phenolate oxygen and the fluorophore with the addition of PPi (Fig. 19d). When a small amount of PPi was bound to sensor 9, the electron density on the nitrogen atoms of the two DPA regions increased, inducing PET from the nitrogen atoms to the fluorophore, which partially quenched the fluorophore. However, with increasing PPi anion concentrations, a tight sensor 9-PPi complex was formed that weakened the bond between the phenolate oxygen atom and the two Zn²⁺ ions, leaving a negatively charged phenolate oxygen atom. That donated more electron density to the fluorophore phenyl ring and significantly enhanced ICT and fluorescence. The 42-nM LOD and the sensor selectivity enabled in vitro and in vivo PPi detection levels in healthy mice and human lung (A549) cancer cells, respectively (Fig. 20). The near-infrared chromogenic benzopyran group and the stable metal recognition unit (DCMB: dicyanomethylene-benzopyran), which enabled “turn-on” PPi sensing, provided better design ideas for the further commercialization of PPi sensors. This complex system should be further tuned for more sensitivity and a stronger fluorescent group, and the mechanism should be studied in more detail. Finally, the nuclei were not stained as much those by other sensors; basically, the cytoplasm was stained by this sensor.

Fig. 19.

Changes in fluorescence spectra of sensor 9 (10 μM) upon addition of PPi (0–100 μM) (a) and various anions (500 μM) (b). (c) 654-nm fluorescence intensity changes of sensor 9 at vs. concentrations of PPi, ATP, ADP, and AMP. (d) Sensor 9 for PPi sensing(57). Copyright Elsevier B.V, 2018.

Fig. 20.

Confocal fluorescence images of PPi (10 equiv) using sensor 9 in A549 cells(57). Copyright Elsevier B.V, 2018.

In summary, various efforts have been devoted to sensing phosphate anions, and optimally positioned binuclear zinc complexes were one of the most successful strategies. In addition to metal regulation, hydrogen-bonding sites are expected to further enhance PPi binding. Various groups can be used for near-infrared fluorescence PPi biosensors (77–80).

4. Conclusion and Prospective

This review presented an overview of recent advancements in fluorescent PPi sensors based on metal complexes (sensors 1–4 and 6–9), with a specific focus on sensing via mono/binuclear designs.

The detection of inorganic pyrophosphate (PPi) using metal complex sensors has been studied extensively. These sensors exhibited fluorescence turn-on or colorimetric changes upon binding PPi, enabling selective and sensitive detection in water. Terpyridine derivatives have shown promising results, with significant fluorescence enhancement upon PPi binding (Sensor 1–4). However, the selectivity of these sensors for PPi over other phosphate derivatives, such as ATP and ADP, must be improved.

Binuclear metal complexes, particularly those containing zinc (II) ions, have exhibited strong PPi binding and increased selectivity. They form stable PPi-metal complexes that serve as reliable sensing platforms (Sensors 6–9). The introduction of additional hydrogen-bonding sites in these complexes has enhanced PPi binding. However, the fluorescent responses should be improved. A combination of the terpyridine response and the strong binding affinity of binuclear metal complexes might be optimal for PPi sensing.

Metal-complex sensors have been successfully used for PPi detection in cellular imaging studies. They can selectively stain the nuclei (Sensor 1-3) or cytoplasm (Sensor 8–9), providing valuable -insights into PPi distributions and cellular processes. However, both ATP-sensor- and PPi sensor-stained nuclei should examined for further validation. In conclusion, metal complex sensors offer a promising approach for PPi detection. Further efforts are needed to enhance their sensitivity and selectivity for PPi, particularly over other phosphate derivatives in biological samples.

In summary, we first analyzed commercial kits for PPi sensing. Then, we focused on PPi sensors based on metal complexes and discussed the limitations to facilitate the development of the next generation of sensors. We hope that this will promote the development of better PPi commercial kits for biomedicine and PPi-related disease sensing.

Fig. 6.

Confocal fluorescence microscopy images of (a) HeLa cells incubated with sensor 1, (b) a single HeLa cell (49). Copyright American Chemical Society, 2014.

Abbreviations:

PPi

pyrophosphate

ATP

adenosine 5’-triphosphate

ADP

adenosine diphosphate

AMP

adenosine monophosphate

cAMP

cyclic adenosine monophosphate

Data Availability Statement

Data sharing is not applicable to this article as no data sets were generated or analyzed during the current study.