Transition metal cations catalyze ¹⁶O/¹⁸O exchange of catechol motifs with H₂¹⁸O

Abstract

Catechol motifs are ubiquitous in nature, as part of plant, animal and microbial metabolites, and are known to form complexes with various metal cations. Here, we report for the first time that complexation with transition metal cations, especially Fe(III), results in rapid ¹⁶O/¹⁸O exchange of the catecholic hydroxyl groups with H₂¹⁸O. We discuss the implications of this finding for mechanistic studies using H₂¹⁸O and potential relevance for production of ¹⁸O-labeled catechol derivatives.

---

Isotopic labeling studies are widely used in many fields of chemistry, typically in combination with mass spectrometric detection. For example, ¹⁸O₂ and H₂¹⁸O are often used to investigate chemical, enzymatic and microbial oxidation and degradation reactions.^1–4 In addition, when compounds of interest are available in a stable isotopically labeled form, they can be used for isotope dilution mass spectrometry,^5–7 determination of kinetic isotope effects,^8–10 mass spectrometric monitoring of reaction kinetics,^11,12 and even to follow the metabolic fate of biologically active compounds in human or animal studies.^13,14 In recent (to be published) work, we employed H₂¹⁸O to investigate degradation reactions of several catechol derivatives in the presence of metal cations, and observed unexpected rapid ¹⁶O/¹⁸O exchange between the H₂¹⁸O and the aromatic hydroxyl groups under mild conditions (37 °C, neutral or slightly acidic pH). ¹⁶O/¹⁸O exchange of catecholic hydroxyl groups with H₂¹⁸O has been previously reported, yet only under harsh catalyst-free conditions (3 M HCl, 150 °C, 20 days).¹⁵ To the best of our knowledge, rapid exchange under mild conditions has not been previously described. We investigated this metal-induced ¹⁶O/¹⁸O exchange in further detail, as it may open up a new and simple approach to produce stable ¹⁸O-labeled catechol derivatives. Additionally, this communication serves as a warning for researchers who intend to study the reactions of catechol derivatives in the presence of metal cations using H₂¹⁸O.

The starting point of our work was the observation that 3,4-dihydroxybenzoic acid (3,4-DHBA, Fig. 1) showed rapid ¹⁶O/¹⁸O exchange of two oxygen atoms in the presence of various metal cations. These ¹⁸O labels were exclusively inserted on the aromatic hydroxyl groups, as evidenced by the neutral loss of 44 Da (i.e., unlabeled CO₂) in both unlabeled and ¹⁸O labeled catechols (Fig. 2).

Fig. 1 Catechol derivatives and metal cations used for the ¹⁸O labeling study (a) and schematic representation of the observed ¹⁸O labeling (b). Unless stated otherwise, cations and catechols were used in equimolar concentrations. The R-group in figure (b) represents any of the substituents shown in figure (a).

Fig. 2 Overlayed Higher-energy Collisional Dissociation (HCD) fragmentation patterns of unlabeled (black) and doubly ¹⁸O-labeled (blue) 3,4-DHBA obtained using high resolution ESI-MS in negative ionization mode. The spectrum of doubly labeled 3,4-DHBA was obtained after incubation with FeCl₃ (equimolar, 1 h, 37 °C). Fragmentation patterns of other (un)labeled catechol derivatives can be found in Fig. S1.†

Many metal cations are known to form catecholato complexes in the presence of catechol and its derivatives.¹⁶ Seemingly, this complexation somehow catalyzes the ¹⁶O/¹⁸O exchange with H₂¹⁸O. To investigate this phenomenon in greater detail, we first screened the ability of four metal cations to catalyze this ¹⁶O/¹⁸O exchange. Hereto, catechol and 3,4-DHBA were incubated with the chloride salts of Fe(III), Al(III), Cu(II), and Zn(II) in H₂¹⁸O. We selected these metal cations because they are all able to form catecholato complexes,^17,18 but are fundamentally different in terms of redox activity. Whereas Fe(III) and Cu(II) have been described to oxidize catechol derivatives,¹⁹ Al(III) and Zn(II) are redox inactive.²⁰ Samples of the incubation mixture were taken at 2 and 24 h, and analyzed using high resolution RP-UHPLC-PDA-ESI-MS to determine the percentage of single and double ¹⁸O labeling. As complexation of catechol motifs to metal cations has been reported to be pH-dependent,²¹ these screening experiments were performed at both pH 7 and pH 3.

For both catechol and 3,4-DHBA, incubation with FeCl₃ resulted in extensive labeling at pH 3 and 7, already after 2 hours (Fig. 3). ZnCl₂ at pH 7 and CuCl₂ at both pH 3 and 7 were found to catalyze ¹⁶O/¹⁸O exchange in the case of catechol, but no effect was observed in the case of 3,4-DBHA. The higher labeling yield at pH 7 as compared to pH 3 is in line with the expected increased complexation of catechol at elevated pH (Fig. S2†).²² Despite its ability to form catecholato complexes,²³ Al(III) did not catalyze ¹⁸O labeling (Fig. 3 and Fig. S2†). The observed ¹⁶O/¹⁸O exchange in the presence of Fe(III) and Cu(II) could be the result of valence tautomerism within the catecholato complexes. Such tautomerism has, indeed, been described for several complexes of transition metal cations with partly filled 3d orbitals.^24–26 A H₂¹⁸O molecule coordinated to the metal or present in the bulk could attack the activated keto-tautomer of the catechol and thereby replace the original ¹⁶O hydroxyl group (Fig. 4). Such a mechanism seems, furthermore, fully in line with the lack of ¹⁶O/¹⁸O exchange in the presence of the redox-inactive Al(III). The proposed mechanism, however, cannot explain the observation of ¹⁶O/¹⁸O exchange in the presence of Zn(II). Since Zn(II) has a completely filled 3d orbital, valence tautomerism, forming Zn(I), is expected to be highly unfavorable. In addition, when catechol incubations were performed with Mn(II), rapid ¹⁶O/¹⁸O exchange was observed at pH 7 (Fig. S3†). As we expect valence tautomerism between Mn(II) and Mn(I) to be highly unlikely, this finding also challenges valence tautomerism as the only mechanism. To explore other potential mechanisms, we investigated whether molecular oxygen could, somehow, promote ¹⁶O/¹⁸O exchange, by repeating the incubations of catechol with FeCl₃, CuCl₂, and ZnCl₂ after N₂ purging of the H₂¹⁸O solvent and headspace. In none of the incubations, a decrease in ¹⁸O labeling was observed after N₂ purging (Fig. S4†), excluding a potential role of O₂ in the ¹⁶O/¹⁸O exchange. We therefore propose that the exchange in the presence of Fe(III) and Cu(II) occurs via the mechanism depicted in Fig. 4, whereas the exact mechanism of the Zn(II)-catalyzed exchange remains to be elucidated.

Fig. 3 Percentages of unlabeled, singly labeled, and doubly labeled catechol (a–d) and 3,4-DHBA (e–h) after equimolar incubations (1 mM) with FeCl₃, CuCl₂, AlCl₃, or ZnCl₂ at 37 °C for 2 or 24 h, as determined by RP-UHPLC-PDA-ESI-MS. Data are presented as average and standard deviation of two separate incubations.

Fig. 4 Proposed mechanism for the Fe(III) or Cu(II) catalyzed ¹⁶O/¹⁸O exchange of catechol motifs, here depicted with Fe(III) as an example. The H₂¹⁸O attacking the catecholato ligand may also originate from the bulk. Examples of valence tautomerism as a mechanism underlying catalysis in Fe and Cu–catecholato complexes have been reported in earlier research.^25,26

As Fe(III) showed the highest extent of ¹⁸O labeling in all cases, we continued our research using only FeCl₃. No notable differences were observed between the labeling percentages at pH 3 and pH 7 in FeCl₃ incubations. Because of the relatively higher solubility of Fe(III) at acidic conditions,²⁷ further experiments were performed at pH 3.

Since the initial screening showed limited differences between the timepoints tested, and the ¹⁶O/¹⁸O exchange of catechol seemed already complete within 2 h, we proceeded to follow the exchange kinetics by using time-resolved ESI-MS. Hereto, catechol, 3,4-DHBA, 4-chlorocatechol, 4-methylcatechol, and dopamine solutions in H₂¹⁸O (pH 3) were mixed with equimolar amounts of FeCl₃ in H₂¹⁸O (pH 3), and infused for ∼60 min into an ion trap MS. The kinetics were found to be strongly dependent on the substitution of the aromatic ring. Catechol, 4-chlorocatechol, and 3,4-DHBA were found to undergo near-to-complete ¹⁸O labeling within 30 min (Fig. 5a, b and d), whereas 4-methylcatechol and dopamine were labeled considerably slower and reached around 60% labelling within the duration of the experiment (Fig. 5c and e).

Fig. 5 Kinetics of FeCl₃ catalyzed ¹⁸O labeling of (a) catechol, (b) 4-chlorocatechol, (c) 4-methylcatechol, (d) 3,4-DHBA, and (e) dopamine. Relative intensities of unlabeled, singly ¹⁸O-labeled, and doubly ¹⁸O-labeled catechols are shown in black, red, and blue, respectively. In absence of FeCl₃, no labeling was observed as indicated in (f) for 4-chlorocatechol. Note: the relative intensities shown are not corrected for the fact that 97% ¹⁸O-labeled water was used as solvent.

No labeling was observed in H₂¹⁸O in absence of FeCl₃, confirming that ¹⁶O/¹⁸O exchange with H₂¹⁸O only occurs in the presence of a suitable catalyst (Fig. 5f). The latter also indicates that the isotopic labeling will be stable after removal of FeCl₃ from the product. When phenol was used instead of the catechols, no labeling was observed after 24 h, which underscores the importance of the catechol motif for the isotope exchange (data not shown).

Although the above described ¹⁶O/¹⁸O exchange may provide an interesting new approach for production of stable ¹⁸O-labeled catechols, Fe(III) has been reported to induce oxidative degradation of various catechols.²⁸ Therefore, we simultaneously determined the labeling yield and recovery of all eight catechol derivatives, shown in Fig. 1, at various time points, using high resolution RP-UHPLC-PDA-ESI-MS. As can be observed from Table 1, complete (i.e., 100% doubly labeled compound) or near-to-complete labeling (i.e., >90% doubly labeled compound) was obtained for catechol, 4-chlorocatechol, 4-methylcatechol, 3,4-DHBA and dopamine with high (i.e., >69%) to very high (i.e., >95%) recoveries. In the case of catechol and 4-chlorocatechol, complete labeling and very high recoveries were obtained for incubations at 0.5 and 1 h. Prolonged incubations (24 h) decreased the recoveries, presumably due to oxidative degradation. For 4-methylcatechol, 3,4-DHBA and dopamine, incubations of 4 h were required to achieve >90% doubly ¹⁸O labeled products. The double labeling yields of 3,4-DBHA and dopamine stagnated around 95%, and addition of extra iron or incubating at 1 : 3 catechol : Fe ratio did not substantially increase the labeling yield, but mainly resulted in a decreased recovery. Significantly lower labeling yields were obtained for 4-nitrocatechol and 3,4-dihydroxybenzaldehyde (3,4-DHBZ), even though both compounds possess an electron withdrawing substituent. No general trend could be observed on the effect of electron withdrawing and electron donating substituents on the labeling yield. The reason for the poor yields of the ¹⁶O/¹⁸O exchange of 4-nitrocatechol and 3,4-DHBZ remains to be investigated. For 4-methoxycatechol, 66% ¹⁸O labeling was observed after 24 h, along with a remarkably low recovery. Various new peaks corresponding to compounds of decreased molecular weight were observed in the RP-UHPLC-PDA-ESI-MS chromatograms, suggesting extensive oxidative degradation of 4-methoxycatechol in the presence of FeCl₃ (Fig. S5i†).

Table 1 Labeling yields and recoveries of the five catechols after equimolar incubations (0.1 mM) with FeCl₃ at 37 °C, as determined by RP-UHPLC-PDA-ESI-MS. Results are shown as the average and standard deviation of two independent incubations. For the 1 h incubation of dopamine and incubations of 4-methoxycatechol no duplicates were included. The labeling percentage has been corrected for the fact that the H₂¹⁸O contained 97% ¹⁸O. Recovery represents relative concentration of the compound (sum of labeled and unlabeled) compared to the initial concentration. Chromatograms and mass spectra are shown in Fig. S5†

Compound	Time (h)	Doubly labeled (%)	Recovery (%)
a These incubations were performed at a catechol : FeCl3 ratio of 1 : 3.
Catechol	0.5	100 ± 0.1	101 ± 1
Catechol	1	100 ± 0.1	102 ± 5
Catechol	24	100 ± 1	88 ± 5
4-Chlorocatechol	0.5	100 ± 1	95 ± 4
4-Chlorocatechol	1	100 ± 0.5	91 ± 1
4-Chlorocatechol	24	100 ± 0.1	70 ± 3
4-Methylcatechol	1	84 ± 1	69 ± 1
4-Methylcatechol	4	100 ± 0.2	79 ± 2
4-Methylcatechol	24	100 ± 0.1	70 ± 1
3,4-DHBA	1	91 ± 10	86 ± 13
3,4-DHBA	4	94 ± 8	88 ± 15
3,4-DHBA^a	4	98 ± 2	49 ± 3
3,4-DHBA	24	95 ± 6	80 ± 0.3
Dopamine	1	92	98
Dopamine	4	96 ± 1	92 ± 11
Dopamine^a	4	97 ± 2	59 ± 3
Dopamine	24	95 ± 2	72 ± 13
4-Nitrocatechol	24	0.2 ± 0.04	65 ± 13
4-Nitrocatechol^a	24	0.7 ± 0.03	60 ± 4
3,4-DHBZ	24	9 ± 1	87 ± 9
3,4-DHBZ^a	24	24 ± 9	59 ± 6
4-Methoxycatechol	24	66	2
4-Methoxycatechol^a	24	N.D.	0

---

Results in Table 1 show that FeCl₃ is an excellent catalyst for the ¹⁸O labeling of various catechols. Although isotopically labeled compounds are mainly used in mass spectrometric studies, requiring low concentrations, the above described experiments (at 0.1 mM) would only allow synthesis of ∼10–15 μg of labeled compound per 1 mL of solvent. Therefore, we investigated whether the ¹⁸O labeling of catechol could be upscaled to concentrations that are more meaningful for lab-scale synthesis. We note that at increased concentrations the catecholato-complexes partly precipitated, but could be completely resolubilized in ethanol after incubation. RP-UHPLC-PDA-ESI-MS analysis of the resolubilized samples showed that using 10 and 50-fold higher concentrations (1 and 5 mM) still resulted in complete and near-to complete labeling, respectively, within 0.5 h (Table 2). Further upscaling to 50 mM, however, showed a drop in the labeling yield. Presumably this is caused by the poor water solubility of the catechol–iron mixtures at increased concentrations. We are currently investigating approaches to circumvent this. Nonetheless, even partially ¹⁸O-labeled compounds can be highly valuable as mechanistic probes and quantification standards.

Table 2 Labeling yields and recoveries of catechol at various concentrations after equimolar incubations with FeCl₃ at 37 °C, as determined by RP-UHPLC-PDA-ESI-MS. Results at 0.1, 1 and 5 mM are shown as the average and standard deviation of two independent incubations. The labeling percentage has been corrected for the fact that the H₂¹⁸O contained 97% ¹⁸O. Recovery presents relative concentration of the compound (sum of labeled and unlabeled) compared to the initial concentration. For incubations at 50 mM no duplicates were included

Concentration (mM)	Time (h)	Doubly labeled (%)	Recovery (%)
0.1	0.5	100 ± 1	101 ± 1
1	0.5	100 ± 0.8	101 ± 8
5	0.5	95 ± 2	90 ± 15
50	4	8	100
50	24	16	85

---

In summary, we showed that Fe(III), Cu(II), Zn(II) and Mn(II) cations can catalyze ¹⁶O/¹⁸O exchange of catechol motifs with H₂¹⁸O. This may provide a new route for facile production of ¹⁸O labeled catechol derivatives. At the same time, our findings indicate that extra caution is required when drawing conclusions from studies on the reactions of catechol derivatives that employ H₂¹⁸O labeling in the presence of transition metals.

Conflicts of interest

There are no conflicts to declare.

References

1. S. E. Davis, B. N. Zope and R. J. Davis, Green Chem., 2012, 14, 143–147.
2. K. Norrman, S. A. Gevorgyan and F. C. Krebs, ACS Appl. Mater. Interfaces, 2009, 1, 102–112.
3. G. Vaaje-Kolstad, B. Westereng, S. J. Horn, Z. Liu, H. Zhai, M. Sørlie and V. G. Eijsink, Science, 2010, 330, 219–222.
4. M. D. Hilton and W. J. Cain, Appl. Environ. Microbiol., 1990, 56, 623–627.
5. C. G. Arsene, R. d. Ohlendorf, W. Burkitt, C. Pritchard, A. Henrion, D. M. Bunk and B. Güttler, Anal. Chem., 2008, 80, 4154–4160.
6. N. Tretyakova, M. Goggin, D. Sangaraju and G. Janis, Chem. Res. Toxicol., 2012, 25, 2007–2035.
7. J. Villanueva, M. Carrascal and J. Abian, J. Proteomics, 2014, 96, 184–199.
8. P. Liuni, E. Olkhov-Mitsel, A. Orellana and D. J. Wilson, Anal. Chem., 2013, 85, 3758–3764.
9. D. A. Singleton and A. A. Thomas, J. Am. Chem. Soc., 1995, 117, 9357–9358.
10. R. Hilgers, A. Van Dam, H. Zuilhof, J.-P. Vincken and M. A. Kabel, ACS Catal., 2020, 10, 8650–8659.
11. L. Jasikova, M. Anania, S. Hybelbauerova and J. Roithova, J. Am. Chem. Soc., 2015, 137, 13647–13657.
12. R. Hilgers, S. Y. Teng, A. Bris, A. Pereverzev, P. White, J. Jansen and J. Roithová, Angew. Chem., Int. Ed., 2022, 61, e202205720.
13. M. D. Neinast, C. Jang, S. Hui, D. S. Murashige, Q. Chu, R. J. Morscher, X. Li, L. Zhan, E. White and T. G. Anthony, Cell Metab., 2019, 29, 417–429.
14. B. Kluger, C. Bueschl, N. Neumann, R. Stückler, M. Doppler, A. W. Chassy, A. L. Waterhouse, J. Rechthaler, N. Kampleitner and G. G. Thallinger, Anal. Chem., 2014, 86, 11533–11537.
15. K. L. Clay and R. C. Murphy, Biomed. Mass Spectrom., 1980, 7, 345–348.
16. Z. Xu, Sci. Rep., 2013, 3, 1–7.
17. L. Malacaria, C. La Torre, E. Furia, A. Fazio, M. C. Caroleo, E. Cione, L. Gallelli, T. Marino and P. Plastina, J. Mol. Liq., 2022, 345, 117895.
18. A. Primikyri, G. Mazzone, C. Lekka, A. G. Tzakos, N. Russo and I. P. Gerothanassis, J. Phys. Chem. B, 2015, 119, 83–95.
19. E. Nkhili, M. Loonis, S. Mihai, H. El Hajji and O. Dangles, Food Funct., 2014, 5, 1186–1202.
20. M. E. Bodini, G. Copia, R. Robinson and D. T. Sawyer, Inorg. Chem., 1983, 22, 126–129.
21. J. Bijlsma, W. J. de Bruijn, J. A. Hageman, P. Goos, K. P. Velikov and J.-P. Vincken, Sci. Rep., 2020, 10, 1–11.
22. M. Elhabiri, C. Carrër, F. Marmolle and H. Traboulsi, Inorg. Chim. Acta, 2007, 360, 353–359.
23. L. Malacaria, G. A. Corrente, A. Beneduci, E. Furia, T. Marino and G. Mazzone, Molecules, 2021, 26, 2603.
24. T. Tezgerevska, K. G. Alley and C. Boskovic, Coord. Chem. Rev., 2014, 268, 23–40.
25. N. Shaikh, S. Goswami, A. Panja, X.-Y. Wang, S. Gao, R. J. Butcher and P. Banerjee, Inorg. Chem., 2004, 43, 5908–5918.
26. G. Speier, Z. Tyeklar, P. Toth, E. Speier, S. Tisza, A. Rockenbauer, A. Whalen, N. Alkire and C. Pierpont, Inorg. Chem., 2001, 40, 5653–5659.
27. P. N. Johnson and A. Amirtharajah, J. – Am. Water Works Assoc., 1983, 75, 232–239.
28. J. Bijlsma, W. J. de Bruijn, K. P. Velikov and J.-P. Vincken, Food Chem., 2022, 370, 131292.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2ob01884e

---