Activation of [^99(m)TcO₄]⁻ by phosphonium cations

Abstract

Phosphonium salts in solution or bound to solid phase supports interact with the robust [^99(m)TcO₄]⁻ anion in neutral water. This activated form of [^99(m)TcO₄]⁻ represents a source for complexes with the fac-{^99mTc^VIIO₃}⁺ core, which can be synthesized in high yields and purity.

---

Technetium (Tc) is the lightest artificial element, environmentally present in minute quantities resulting from spontaneous fission of uranium. Large amounts of the long-lived β⁻ emitter ⁹⁹Tc (2.13 × 10⁵ years half-life) are produced in nuclear power plants.¹⁹⁹Tc is isolated from spent fuels as highly water soluble [NH₄][⁹⁹TcO₄]. To assess the hazard potential after an accidental release of [⁹⁹TcO₄]⁻ into the environment, it is essential to elucidate parameters leading to the activation of [⁹⁹TcO₄]⁻ at the tracer level. The nuclear isomer ^99mTc is a commonly used radionuclide in nuclear medicine.^2–5 Low valent ^99mTc complexes comprising the fac-{^99mTc^I(CO)₃}⁺ core are currently investigated for molecular imaging purposes.^2,3,6 To extend the available building blocks for labeling chemistry, we introduced high-valent and water stable ^99(m)Tc complexes based on the fac-{^99(m)Tc^VIIO₃}⁺ core.^7–10 Their cycloaddition reactivities towards alkenes led to a new labeling strategy: conjugation of a targeting molecule to the ^99mTc complex at the ligand and not at the metal.^7,8 Reminiscent of [OsO₄], the fac-{^99(m)Tc^VIIO₃}⁺ core reacts with alkenes by a (3+2)-cycloaddition (vic. cis-dihydroxylation – alkene–glycol interconversion).¹¹

Pertechnetate is a robust anion. It can be “activated” by reduction or by strong Brønsted acids (e.g. H₂SO₄ or HCl) in dry, organic solvents. In situ formed (H⁹⁹TcO₄) is a reactive, moisture sensitive and volatile compound from which a series of fac-{⁹⁹Tc^VIIO₃}⁺ complexes were prepared.^12–16 Alternatively, [⁹⁹TcO₄]⁻ is activated by Lewis-acids (LA). LAs form reactive, mixed anhydrides of the {O₃⁹⁹Tc^VII-O-LA} type.^9,17 The {O-LA} group is then replaced by tridentate ligands, yielding fac-{⁹⁹Tc^VIIO₃(L³)}⁺ type complexes. Due to moisture sensitivity of Lewis acids, these reactions generally require organic solvents.

A direct synthesis from [^99(m)TcO₄]⁻ in water would be desirable and would open an alternative and more applicable pathway to new high-valent building blocks for molecular imaging. Herein, we report that phosphonium cations, albeit not “typical” Lewis acids, activate [^99(m)TcO₄]⁻ at the tracer level and enable the synthesis of [^99mTcO₃(tacn)]⁺ (tacn = 1,4,7-triazacyclononane, [1]⁺) directly in neutral water. The same strategy in organic solvents leads to [⁹⁹TcO₃(Me₃tach)]⁺ (Me₃tach = 1,3,5-trimethyl-[1,3,5]triazinane). Previously, [1]⁺ was accessible in low yields from [^99mTcO₄]⁻ and in the presence of triphenylphosphine (TPP) at low pH values (<2)^7,8 where TPP is protonated to (HTPP)⁺ (pK_a = 2.73).^18–20 We hypothesized that this phosphonium cation, generated in situ, may activate [^99mTcO₄]⁻. To support this assumption, [^99mTcO₄]⁻ and tacn were reacted in water at neutral pH in the presence of a series of phosphonium salts, such as (PPh₃Me)Br, (PPh₃Bz)Br, (PPh₃Et)Br, (PⁿBu₄)Br, (PPh₄)Br and (PPh₂Me₂)I (Table S2.1.1, ESI†). All did activate [^99mTcO₄]⁻ with ultimate formation of [1]⁺ in 19–35% yields, demonstrating that phosphonium cations do interact with [^99mTcO₄]⁻ in water and at elevated temperatures. The fate of e.g. [PPh₄]⁺ is not known. Most likely, it forms an oxy-anion [PPh₄O]⁻ from which [PPh₄]⁺ is regenerated after protonation and water cleavage. No product at all was obtained in the absence of phosphonium cations. Even sterically crowded cations, such as (PⁿBu₄)⁺ or (PPh₄)⁺ lead to product formation, albeit in low yields.

For applications, dissolved phosphonium salts are not desirable for toxicity reasons. Immobilization on solid phase supports would elegantly circumvent this concern. Polystyrene beads with covalently bound methyl-triphenyl-phosphonium cations (PS-TPPR⁺) or neutral triphenylphosphine (PS-TPP) groups are commercially available. On the latter resin, cations can be generated by quaternization of the phosphine groups with the respective alkyl halides (Scheme 1).

Scheme 1 Extraction and activation of [^99mTcO₄]⁻ from saline with poly-styrene bound quaternary phosphines.

These phosphonium derivatized polystyrene beads PS-[TPPR]⁺ are potent anion exchangers. Even from standard saline solution (0.9% NaCl, ∼0.15 M), [^99mTcO₄]⁻ (10⁻⁶–10⁻⁸ M) is extracted using a few mg of resin within minutes and in >98% yield. Exposure of such resins with [^99mTcO₄]⁻ as a counter ion to a neutral solution of tacn gave better [^99mTcO₃(tacn)]⁺ yields than the homogeneous reactions. As a beneficial side effect, unreacted [^99mTcO₄]⁻ remained bound to the resin. Consequently, cationic Tc^VII complexes such as [1]⁺ are obtained in very high radiochemical purity and in one single step. For instance, the reaction of PS-[PPh₃Me]⁺ with [^99mTcO₄]⁻ in the presence of tacn gave [1]⁺ in 64% yield (3% residual [^99mTcO₄]⁻ in solution), after 1 h at 95 °C and pH = 7. The “missing” 33% [^99mTcO₄]⁻ remained resin bound (Fig. S2.2.1, ESI†). This remarkable reaction is now a generally applicable approach to obtain ^99mTc^VII complexes without soluble phosphonium salts and at neutral pH.

The role of the phosphonium cation is dual: (i) trapped [^99mTcO₄]⁻ on the resin leads to a much higher, “local” concentration and, thus, to faster reactions and (ii) [^99mTcO₄]⁻ is polarized and thereby activated. Both factors lead to improved yields and purities of e.g. [1]⁺, and are crucial features for radiopharmaceuticals. The nature of the phosphonium cation has a distinct influence on product yields, i.e. the reactivity can be tuned by altering the 4th group on the quaternary phosphonium. From a functionality screen, the benzyl group PS-[PPh₃Bz]⁺ emerged as the most favourable; yields up to 88% with less than 0.5% residual [^99mTcO₄]⁻ were achieved in one step (Fig. 1).²¹ Since conditions are neutral, [1]⁺ can directly be subjected to further reactions such as cycloaddition with alkenes to yield Tc^V glycolato complexes. This has been shown by the synthesis of the model compound [^99mTcO(tacn)(O₂styrSO₃)] (Fig. S2.3.1, ESI†).^7,8

Fig. 1 HPLC trace before (dashed line) and after the reaction (solid line) of [^99mTcO₄]⁻ with PS-[PPh₃Bz]⁺ and tacn (1 h, 95 °C, pH 7). Yield: 88%.

In the preparation of ^99mTc-based molecular imaging agents, microwave assisted heating is an alternative to thermal heating.^22–24 The reaction of [^99mTcO₄]⁻ with tacn in the presence of PS-[PPh₃Me]⁺ in a microwave reactor at 120 °C for 10 min gave 78% [1]⁺ (64% thermal, 1 h, Fig. S2.2.2, ESI†) and still 7% at 50 °C after 1 h (microwave and thermal heating). The superiority of PS-[PPh₃Bz]⁺ over [^99mTcO₄]⁻ activation was evident since about 30% [1]⁺ was obtained under these conditions (50 °C).

In the homogenous reactions, we did not detect intermediates. It is likely that the cations activate [^99mTcO₄]⁻ by forming a mixed anhydride {PPh₃Me-O-^99mTcO₃} which then reacts with tacn to give [1]⁺ (Scheme 2). The equilibrium in Scheme 2 is on the ion-pair side and the concentration of the mixed anhydride 2 is low. The reaction with [1]⁺ is therefore slow and yields and rates are strongly tacn concentration dependent (Fig. S2.2.3, ESI†). Stability studies under the applied reaction conditions (saline solution, 95 °C, pH 7–8) disclosed a slow decomposition of complex [1]⁺ (⁹⁹Tc and ^99mTc) to [^99(m)TcO₄]⁻ (S2.5, ESI†). This slow decomposition accounts for reaction yields <100%. Adaption of the reaction conditions to ⁹⁹Tc is not immediately possible since phosphonium salts of [⁹⁹TcO₄]⁻ are essentially insoluble in water. Upon adding highly diluted [⁹⁹TcO₄]⁻ as a carrier to a solution of [^99mTcO₄]⁻, PS-[PPh₃Bz]⁺ and tacn, ⁹⁹Tc will compete with ^99mTc in the reaction to give [1]⁺ (⁹⁹Tc and ^99mTc). Since highly diluted, [1]⁺ (⁹⁹Tc) is not visible on HPLC, the apparent yield of [1]⁺ (^99mTc) will be reduced. This was observed, indeed, that increasing amounts of the carrier decreased yields of [1]⁺ (^99mTc, Fig. S2.2.4, ESI†). The conversion of [^99(m)TcO₄]⁻ into defined complexes in water and in the presence of phosphonium cations is remarkable and should be included in the risk assessment of [⁹⁹TcO₄]⁻.

Scheme 2 Synthesis of [^99mTcO₃(tacn)]⁺via the mixed anhydride {PPh₃Me-O-^99mTcO₃} (2).

To efficiently activate [⁹⁹TcO₄]⁻ by phosphonium cations, stronger Lewis acids than [PPh₃R]⁺ (R = Me, Bz) are required to shift the equilibrium (Scheme 2) towards the mixed anhydride {Ph₃RP-O-⁹⁹TcO₃} (2). Ph₃PBr₂ is such a strong Lewis acid with distinct affinity for oxygen. In organic solvents, Ph₃PBr₂ forms quantitatively the reactive, mixed anhydride {Ph₃BrP-O-⁹⁹TcO₃} which was observed in situ by ⁹⁹Tc and ³¹P NMR spectroscopy (Fig. S2.4.1/2.4.2, ESI†). The ⁹⁹Tc NMR spectrum showed a broad signal (ν_1/2 = 230 Hz) at ∼0 ppm which contrasts the extremely sharp (ν_1/2 = 2 Hz) and characteristic [⁹⁹TcO₄]⁻ signal at the same chemical shift. Consistently, the ³¹P NMR spectrum of the intermediate ({Ph₃BrP-O-⁹⁹TcO₃}) showed a broad signal at 27.22 ppm (ν_1/2 = 110 Hz).

The mixed anhydride {Ph₃BrP-O-⁹⁹TcO₃} can be reacted with ligands not compatible with other Lewis acids. Reactions of {Ph₃BrP-O-⁹⁹TcO₃} with bi- and tridentate ligands, such as 2,2′-bipyridine (bipy), 2,2-bis(3,5-dimethyl-1H-pyrazol-1-yl)acetic acid (bpza*), and tris(3,5-dimethyl-1H-pyrazol-1-yl)methane (tpzm*) gave the corresponding complexes [⁹⁹TcO₃(bipy)Br], [⁹⁹TcO₃(bpza*)] and [⁹⁹TcO₃(tpzm*)]Br in high yields (S1.2, ESI†).

To exemplify access to [⁹⁹TcO₃(L³)]⁺ type complexes with weakly coordinating ligands L³, we selected the small aliphatic, N-heterocycle 1,3,5-trimethyl-triazacyclohexane (Me₃tach). High-valent Me₃tach complexes are rare and only [ReO₃(Me₃tach)]⁺ and [Mo^VIO₃(R₃tach)]⁰ are reported.^25,26 Low-valent complexes [(Me₃tach)M⁰(CO)₃] (M = Mo, W) are known²⁷ and ([(R₃tach)CrCl₃]) was studied using the Phillips catalyst (CrO₃/SiO₂).^28–30

In situ prepared {Ph₃BrP-O-⁹⁹TcO₃} reacts with Me₃tach immediately to give Ph₃PO (Fig. S2.4.3, ESI†) and [⁹⁹TcO₃(Me₃tach)][⁹⁹TcO₄] ([3][⁹⁹TcO₄]) which is precipitated as a yellow powder (Scheme 3). As solid, [3][⁹⁹TcO₄] can be stored for weeks. It is soluble in dry acetonitrile and CH₂Cl₂. [3][TcO₄] is reactive towards reducing agents such as bromide. Thus, fast separation from the reaction solution is crucial. An X-ray diffraction analysis of single crystals, obtained from slow evaporation of a CH₂Cl₂ solution of [3][⁹⁹TcO₄], gave a structure with two [3]⁺ cations, one [⁹⁹TcO₄]⁻, and one [⁹⁹TcOBr₄]⁻ anion. Both cations show identical structural features. Details of data collection and structure calculations are given in the ESI† (Table S3.1). Fig. 2 shows the structure of the [3]⁺ cation. The Tc–N bond lengths in [3]⁺ are elongated (2.265(2)–2.281(2) Å) as compared to the water stable fac-[⁹⁹TcO₃]⁺ complex [⁹⁹TcO₃(tacn)]⁺ (2.239(4) Å).¹⁰ These long bonds are in agreement with the reactivity of [3]⁺ and imply the possibility of substituting Me₃tach with other ligands. The highly distorted coordination geometry around the ⁹⁹Tc centre (d⁰ system) is induced by the strain of the small Me₃tach ligand (N–Tc1–N angles: 59.85(9)–60.40(9)°). The steric demand of the Me₃tach ligand is significantly smaller than that for tacn. The O–Tc–O angles in [3]⁺ are increased (108.84(12)–109.32(14)°) in comparison to [⁹⁹TcO₃(tacn)]⁺ (106(2)°) to adopt the favoured tetrahedral conformation.

Scheme 3 Synthesis of [⁹⁹TcO₃(Me₃tach)][⁹⁹TcO₄] ([3][⁹⁹TcO₄]).

Fig. 2 ORTEP representation³¹ of one [⁹⁹TcO₃(Me₃tach)]⁺ ([3]⁺) cation of the [3]₂[TcOBr₄][TcO₄]·CH₂Cl₂ structure (two molecules per asymmetric unit). Thermal ellipsoids represent 50% probability. Hydrogen atoms are omitted for clarity. Selected bond lengths [Å] and angles [°]: Tc1–O1 1.697(2), Tc1–O2 1.684(3), Tc1–O3 1.687(2), Tc1–N1 2.265(2), Tc1–N2 2.274(2), Tc1–N3 2.281(2), O1–Tc1–O2 109.32(14), O1–Tc1–O3 108.84(12), O2–Tc1–O3 109.06(13), N1–Tc1–N2 60.12(8), N1–Tc1–N3 60.40(9), N2–Tc1–N3 59.85(9).

The ⁹⁹Tc NMR spectrum of [3][⁹⁹TcO₄] exhibits one broad signal at 362 ppm (ν_1/2 = 913 Hz), in a similar region as that for [⁹⁹TcO₃(tacn)]⁺ (358 ppm, ν_1/2 = 2500 Hz).¹⁰ An additional sharp signal at 0 ppm represents the [⁹⁹TcO₄]⁻ counter ion (1 : 1 ratio, Fig. S2.4.4, ESI†). Since the small bite angles of Me₃tach render its complexes prone to hydrolysis or ligand substitution, trace amounts of water hydrolyse the compound to [⁹⁹TcO₄]⁻. The sensitivity of [3]⁺ towards hydrolysis was confirmed by ⁹⁹Tc NMR studies. In 0.2 ml acetonitrile, [3]⁺ is fully converted to [⁹⁹TcO₄]⁻ within minutes upon the addition of 10 μl of H₂O as evident from the single signal at 0 ppm ([⁹⁹TcO₄]⁻).

The robust [^99mTcO₄]⁻ anion can be activated in aqueous solution by in situ mixed anhydride formation in the presence of quaternary phosphonium cations. Tridentate ligands such as 1,4,7-triazacyclononane (tacn) can replace the {R₄P-O}⁻-group, thereby forming the water stable [^99mTc^VIIO₃(tacn)]⁺ complex. Substituted tacn ligands or other ligands are expected to show a similar reactivity pattern. If the phosphonium is bound to a solid phase support, the heterogeneous reaction leads to high purity compounds since residual [^99mTcO₄]⁻ remains bound to the resin. The overall procedure is carried out at neutral pH, thus, the product can be further directly reacted with e.g. alkenes. [⁹⁹TcO₄]⁻ at the tracer level is transformed into defined complexes in water in the presence of phosphonium cations. In the future, this important finding has to be included in the risk assessment of [⁹⁹TcO₄]⁻ release into the environment. We are currently studying the highly reactive [⁹⁹TcO₃(Me₃tach)]⁺ complex as a source for complexes with the fac-{⁹⁹Tc^VIIO₃}⁺ core.

Dr H. Braband acknowledges financial support from the Swiss National Science Foundation (PZ00P2_143102).

Notes and references

1. K. Yoshihara, Top. Curr. Chem., 1996, 176, 17–35.
2. S. Liu and S. Chakraborty, Dalton Trans., 2011, 40, 6077–6086.
3. R. Alberto and U. Abram, Handbook of Nuclear Chemistry, Kluwer Academic Publishers, Dordrecht, 2003.
4. J. R. Dilworth and S. J. Parrott, Chem. Soc. Rev., 1998, 27, 43–55.
5. I. Amato, Chem. Eng. News, 2009, 87, 58–64.
6. R. Alberto, H. Braband and H. W. P. N'Dongo, Curr. Radiopharm., 2009, 2, 254–267.
7. H. Braband, Y. Tooyama, T. Fox, R. Simms, J. Forbes, J. F. Valliant and R. Alberto, Chem.–Eur. J., 2011, 17, 12967–12974.
8. H. Braband, Y. Tooyama, T. Fox and R. Alberto, Chem.–Eur. J., 2009, 15, 633–638.
9. Y. Tooyama, H. Braband, B. Spingler, U. Abram and R. Alberto, Inorg. Chem., 2008, 47, 257–264.
10. H. Braband and U. Abram, Inorg. Chem., 2006, 45, 6589–6591.
11. R. M. Pearlstein and A. Davison, Polyhedron, 1988, 7, 1981–1989.
12. J. A. Thomas and A. Davison, Inorg. Chim. Acta, 1991, 190, 231–235.
13. A. Davison, A. G. Jones and M. J. Abrams, Inorg. Chem., 1981, 20, 4300–4302.
14. H. J. Banbery, W. Hussain, I. G. Evans, T. A. Hamor, C. J. Jones, J. A. McCleverty, H.-J. Schulte, B. Engles and W. Kläui, Polyhedron, 1990, 9, 2549–2551.
15. J. A. Thomas and A. Davison, Inorg. Chem., 1992, 31, 1976–1978.
16. J. G. H. Du Preez, T. I. A. Gerber and M. L. Gibson, J. Coord. Chem., 1990, 22, 33–42.
17. E. Oehlke, R. Alberto and U. Abram, Inorg. Chem., 2010, 49, 3525–3530.
18. C. A. Streuli, Anal. Chem., 1960, 32, 985–987.
19. W. A. Henderson and C. A. Streuli, J. Am. Chem. Soc., 1960, 82, 5791–5794.
20. T. Allman and R. G. Goel, Can. J. Chem., 1982, 60, 716–722.
21. A vial containing 10 mg benzyltriphenylphosphonium bromide, polymer bound, was tightly closed and flushed with N₂ for 10 min. 1,4,7-triazacyclononane trihydrochloride (15 mg, 0.06 mmol) was dissolved in 1 ml [^99mTcO₄]⁻ eluate and neutralized with 1 M NaOH. The solution was added to the polymer and the mixture was heated for 1 h at 95 °C. The reaction solution was cooled and the resin filtered. Yield: 88%, radiochemical purity >99%.
22. P. W. Causey, T. R. Besanger, P. Schaffer and J. F. Valliant, Inorg. Chem., 2008, 47, 8213–8221.
23. P. W. Causey, T. R. Besanger and J. F. Valliant, J. Med. Chem., 2008, 51, 2833–2844.
24. A. E. C. Green, P. W. Causey, A. S. Louie, A. F. Armstrong, L. E. Harrington and J. F. Valliant, Inorg. Chem., 2006, 45, 5727–5729.
25. D. V. Partyka, R. J. Staples and R. H. Holm, Inorg. Chem., 2003, 42, 7877–7886.
26. H. Braband, S. Imstepf, M. Felber, B. Spingler and R. Alberto, Inorg. Chem., 2010, 49, 1283–1285.
27. M. V. Baker and M. R. North, J. Organomet. Chem., 1998, 565, 225–230.
28. R. D. Köhn, D. Smith, M. F. Mahon, M. Prinz, S. Mihan and G. Kociok-Köhn, J. Organomet. Chem., 2003, 683, 200–208.
29. R. D. Köhn, M. Haufe, G. Kociok-Köhn, S. Grimm, P. Wasserscheid and W. Keim, Angew. Chem., Int. Ed., 2000, 39, 4337–4339.
30. R. D. Köhn, M. Haufe, S. Mihan and D. Lilge, Chem. Commun., 2000, 1927–1928.
31. L. J. Farrugia, ORTEP-3 for Windows – a version of ORTEP-III with a Graphical User Interface (GUI), 1997.

---

Footnote

† Electronic supplementary information (ESI) available: Ligand and complex synthesis including IR data, NMR data as well as crystallographic details. CCDC 976753. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4cc00718b

---