A Practical Route for the Preparation of 1,4,7-Triazacyclononanyl Diacetates with a Hydroxypyridinonate Pendant Arm

The preparation of triazamacrocyclic hydroxypyridinonate (HOPO-TACN) derivatives as potential chelators for metals in biomedical applications was reported. The synthesis is based on a convergent synthetic approach, in which the key intermediate di-tert-butyl-2,2′-(1,4,7-triazonane-1,4-diyl) diacetate was coupled with a hydroxypyridinonate pendant arm. The method is suitable for rapid syntheses of metal chelator HOPO-TACNs of biomedical interest.

TACN derivatives provide versatile platforms for strongly chelating radiometals such as 90 Y, 177 Lu, 212/213 Bi, 67/68 Ga, 64/67 Cu, 99m Tc and 18 F-Al for SPECT, PET imaging and radiotherapy applications [25,26]. The strong coordination capabilities and high selectivity of the TACN derivatives toward di-or trivalent metal ions were determined by the cooperative binding mode from TACN and pendant arms. The pendent arm on TACN platform is expected to accelerate complexation with isotopes while maintaining a high level of stability of the complexes, coupled with the preorganization and macrocyclic effect of the TACN substructure [27]. So far, some researchers have been focused on the TACN derivatives with acetate pendant arms and phosphinate pendant arms. However, far less attention had been paid to other suitable pendent donor groups. In addition, selecting the correct type of chelating unit is one of the most decisive factors in achieving high selectivity toward the specific metal ion.
A class of multidentate hydroxypyridinonate (HOPO) was synthesized and studied for their outstanding abilities to scavenge lanthanide and actinide ions [28]. In particular, the octadentate ligand 3,4,3-LI(1,2-HOPO) has been identified as an effective metal chelator and is currently undergoing development as a prospective therapeutic actinide decorporation agent [29]. The pendent oxygencontaining donor hydroxypyridinonate demonstrated high affinities to hard Lewis acid.
In this context, new prospects for the synthesis of new chelating agents may be delineated while combining geometric and functional features of the TACN platform and the structural potential of common oxygen-containing donor, such as HOPO, and carboxylic acid. A triazamacrocycle with one HOPO and two carboxylic acids is ideally suited to form seven-coordinate complexes with two groups of facial donor, where the macrocyclic nitrogen atoms on the one side and the pendant donors on the other. The lanthanide ions and many other metal ions have coordination numbers larger than six, thus, these ligands have the potential to be excellent chelating agent in the MRI or radioimmunotherapy (RIT) applications.

Results and Discussion
The synthetic strategy for the preparation of the target ligands HE-NO2A and HP-NO2A involves the synthesis and coupling reaction of the fragment 3 with 4 or 5, which were obtained from readily available starting materials 1,4,7-triazacyclononane (TACN) and 3,2-hydroxypridinone (3,2-HOPO), respectively. The practical, reproducible, and readily scalable synthetic route and efficient purification of all precursor molecules for macrocyclic HOPO-TACN derivatives was developed as shown in Scheme 1.

Scheme 1.
Retrosynthetic route to HE-NO2A and HP-NO2A. Fragment 3 were synthesized from 1,4,7-triazacyclononane (6) by reacting with tert-butylbromoacetate using the reported methods (Scheme 2) [24,[30][31][32]. However, the yield was not good, because the reaction produced a mixture of mono-, di-, and trisubstituted TACN. We then optimized the reaction conditions by choosing a different solvent, optimizing the amount and ratio of reactants, and adding base to scavenge the protons formed during the reaction. In many attempts, we found that adding bases such as potassium carbonate, TEA, or sodium bicarbonate to the reaction negatively affected the yield of disubstituted compounds. The possible reason was that the protons produced in-situ would react with the unsubstituted amine to prevent oversubstitution to give trisubstituted compounds. In addition, through optimizing the ratio of reactants, the best yield was obtained when 2.0 eq. of tert-butylbromoacetate to TACN was used in the reaction. Finally, using acetonitrile as a solvent to alkylate TACN with 2 eq. tert-butyl bromoacetate gave as the major product disubstituted TACN in an acceptable yield as monitored by TLC. While the purification using column chromatography was unsuccessful, a carefully pH-controlled extraction was applied to obtain the fragment 3 in 49% isolated yield.

Scheme 2. Synthesis of fragment 3.
As shown in Scheme 3, the synthesis of fragment 4 was started from 3,2-hydroxypridinone, which was refluxed with 5 eq. ethyl bromoacetate under the protection of N2 to give 8 in 86% yield. In order to get the pure alcohol 10, the needed 9a was obtained via the hydroxyl function protection of 8 by O-benzylation. However, column chromatography must be applied to purify both 9a and 10. In another approach, 9b was directly obtained by O-benzylation and hydrolysis in one batch, and the purification by recrystallization proved very convenient. The following reduction of carboxylic acid 9b with BH3/THF was easily carried out to afforf 10 in higher yield than the approach from 9a. When we tried to convert alcohol 10 into 10-OMs by treating it with 1.2 eq. of MsCl in chloroform in the presence of 3 eq. of TEA, an unexpected chloride 4 was obtained. The same product was obtained by treating 10 with TsCl [33]. During the above reaction, two new spots were formed visualized by TLC during the first few hours, which cannot be separated from the reaction mixture. We presumed that the unidentified spot was the intermediate 10-OMs. This spot was slowly converted into the other, and the whole reaction proceeded efficiently to give the chloride 4 in one batch in 72% yield. With chloride 4 in hand, HE-NO2A (1) was prepared readily from disubstituted TACN 3 in two steps (Scheme 3). The alkylation of disubstituted TACN 3 with 4 using DIEA in refluxing MeCN gave the trisubstituted TACN 11 in excellent yield. Further reaction of 11 with the hydrolysis agent HBr/HOAc at room temperature gave the target ligand 1 in 91% yield, where the O-benzyl hydroxyl protection functional group was simultaneously removed in the reaction. Similar reaction strategy was applied in the synthesis of HP-NO2A (2), since the analogue 2 is just a one carbon extension on the HOPO side arm of 1 (Scheme 4). A Michael addition of methyl acrylate using CsF as a catalyst was employed to introduce a propanoic acid methyl ester side chain on 3,2-HOPO (7). The obtained product 12 was reacted with benzyl chloride in a NaOH/MeOH/H2O system to form O-benzyl protected 13a in 97% yield after a recrystallization from methanol. The reduction of 13a using borane-tetrahydrofuran at room temperature obtained 14 in 67% isolated yield. In order to improve the yield, the 14 could also be obtained through the reduction of 13b using borane-tetrahydrofuran at room temperature in 81% yield, after hydrolysis of 13a. This reduction method was successful in the preparation of 14 on a larger scale because of the higher yield and easier purification. Following the same conditions used for the synthesis of 4, alcohol 14 was treated with methanesulfonic chloride in chloroform in the presence of trimethylamine. However, the unusual 3,4-dihydro-2H-pyrido[2,1-b] [1,3]oxazin-5-ium product 15 was isolated. A similar reaction was also reported in another paper [34]. In another report, the electrophilic 3,2-HOPO imidate salt 15, which can be used for the facile incorporation of the HOPO moiety, reacted with a variety of nucleophiles, including amines [35]. Thus, in the current reaction, alkylation of 3 with 15 in acetonitrile in the presence of DIEA produced 16. After purification via column chromatography, pure 16 were obtained in 53% yield. The tert-butyl and benzyl groups of 16 were then removed with HBr/AcOH to obtain the target compounds HP-NO2A (2) in 90% yield. In conclusion, efficient synthetic routes for two hydroxylpyridinone TACN derivatives have been developed. The new family of monosubstituted macrocyclic ligand HE-NO2A and HP-NO2A with 3,2-HOPO moieties as pendant arms could be used as potential metal chelators in future MRI or PET research. Future study for the ligands in making new Ln(III) complexes (Gd 3+ , Eu 3+ ,Tb 3+ ,Yb 3+ ) and other metal complexes (Cu 2+ , Fe 3+ , Ga 3+ ) is in progress. All commercially available reagents were used as received.

1,4-bis(tert-Butoxycarbonylmethyl)-1,4,7-triazacyclononane (3)
To a solution of TACN (3.0 g, 23 mmol) in CH3CN (50 mL) at 0 °C t-butyl bromoacetate (9.0 g, 46 mmol) in CH3CN (100 mL) was added dropwise over 4 h. The resulting mixture was allowed to reach room temperature and stirred for 24 h. The reaction mixture was filtered, the filtrate was evaporated, and the residue was treated with DI water (15 mL). The resulting solution was adjusted to pH 3 using 1 M HCl and extracted with ether (3 × 50 mL). The aqueous layer was then adjusted to pH 8 using 1 M NaOH and extracted with CH2Cl2. The CH2Cl2 layer was dried, filtered and evaporated to dryness. The residue was treated with ether (50 mL), and the yellowish solid formed and filtered. The solid was treated with hexane (30 mL) and filtered and dried to afford pure 3 (4.0 g, 49%) as a white powder.  (8) The mixture of 3,2-hydroxypridinone (5.5 g, 50 mmol) and tert-butyl-bromoacetate (25 mL) was heated at 140 °C for 3 days. After cooling to room temperature, the solution was filtered and washed with cold MeOH. Recrystallization the solid from ethanol gave 8 as a white solid (8.4 g, 86%). 1