Synthesis of 1,2,4-Triazines and the Triazinoisoquinolinedione DEF Ring System of Noelaquinone

Abstract

The intramolecular Staudinger aza-Wittig reaction is used for a general synthesis of 1,2,5,6-tetrahydro-1,2,4-triazines, a structural motif reported for the natural product noelaquinone. The DEF moiety of noelaquinone was obtained in 13 steps and 2% overall yield, and the structure of the synthetic product was confirmed by x-ray analysis.

Noelaquinone 1 was isolated in 1998 by Paul Scheuer and coworkers from the Indonesian Xestospongia sp.¹ The structure assignment of this marine metabolite was mainly based on NMR analyses, and its biological activities have not yet been explored. However, the hexacyclic noelaquinone ring system is closely related to halenaquinone, xestoquinone, and the viridin class of furanosteroids, many of which have shown potent kinase inhibitory effects.^2,3 A distinct feature of noelaquinone is its DEF 3,4-dihydro-2H-[1,2,4]triazino[2,3-b]isoquinoline-6,11-dione, a heterocycle that has hitherto never been synthesized.

The absence of any information on its biological profile and the presence of a unique heterocyclic ring system make noelaquinone an attractive target for total synthesis. Herein, we report a new method for the formation of fused 1,2,4-triazines and its application to the first preparation of the DEF ring system in 1.

As a key step for the expedient formation of 1,2,4-triazines, we planned to develop a variant of the intramolecular Staudinger-aza-Wittig reaction. This method has proven to be a useful tool in natural products synthesis,⁴ and we decided to further explore its application for the ring closure of an azido-hydrazide precursor (Scheme 1).

Scheme 1.

Retrosynthesis of 1,2,5,6-tetrahydro-1,2,4-triazines.

Due to its ready availability and high chemical reactivity, we first used N-aminophthalimide 3 as a model system (Scheme 2). Azide 5 was obtained in three steps and 47% yield from 3 by condensation with α-bromoacetaldehyde 4⁵ in the presence of 4 Å molecular sieves to give the α-bromohydrazone. Since this intermediate was found to be very labile, it was reduced without isolation with NaCNBH₃ to give the corresponding bromoethyl derivative, which was converted to alkyl azide 5 with sodium azide in DMF.

Scheme 2.

Preparation of triazino-isoindolones from N-aminophthalimide.

When azide 5 was heated in toluene at 100 °C in the presence of 2 equiv of PBu₃, a clean conversion to the desired triazinone 6 occurred. Alternatively, the hydrazide was N-protected with benzyl bromide in the presence of tetrabutylammonium hydrogen sulfate (TBAHS), and the phosphine treatment afforded the corresponding triazinone 7 in 97% yield. The N-benzylated triazinone 7 was much easier to isolate and purify than the unprotected 6.

The successful conversion of model azido-hydrazide 5 into triazino[3,2-a]isoindolones 6 and 7 encouraged us to explore the preparation of triazino[2,3-b]isoquinolines. However, we were unable to isolate the requisite intermediate α-bromohydrazone 9 from hydrazide 8 under typical condensation conditions with aldehyde 4 (Scheme 3). While the hydrazone appeared to form, decomposition during the reaction resulted in the formation of a complex mixture at workup.

Scheme 3.

Failed hydrazone formation.

After several unsuccessful attempts to add the azide moiety to the homophthalimide ring system, we were able to develop a strategy that installed the cyclic imide on a prefunctionalized hydrazide (Scheme 4). Protection of 2-hydroxyethylhydrazine with ethyl pyruvate 10 gave hydrazone 11 in 59% yield. Mesylation of the alcohol and treatment with NaN₃ led to azide 12, which was selectively benzylated and then hydrolyzed to give the azido-hydrazine 14. After acylation of 14 with o-iodobenzoylchloride 15, a C-arylation⁶ of diethylmalonate 17 in the presence of 5 mol% of Cu(I)-catalyst, 10 mol% of picolinic acid, and 3 equiv of Cs₂CO₃ in dioxane led to the α-aryl malonate intermediate, which was used crude after workup. The homophthalimide ring was readily obtained by treatment with catalytic p-toluenesulfonic acid (p-TSA) in aqueous toluene, which promoted a cascade cyclization-decarboxylation to yield 18 in 73% over 2 steps.

Scheme 4.

Preparation of homophthalimide 18 by a key copper-catalyzed C-arylation.

In preparation for the introduction of the benzylic ketone function, 18 was converted to thioacetal 20 by treatment with trimethylene dithiotosylate 19 (Scheme 5).⁷ Acetal exchange to the dimethyl acetal 21 was accomplished with iodobenzenebistrifluoroacetate (PIFA) in methanol in the presence of trifluoroacetic acid.⁸ The Staudinger aza-Wittig reaction of 21 proceeded smoothly under microwave irradiation conditions to give the desired triazine 22 in 61% yield, but all attempts to hydrolyze the acetal to unmask the carbonyl group either decomposed the product or provided a 2.5:1 mixture of the isomeric triazinoisoquinolines 23a and 23b in modest 45% combined yield. The structures of 23a and 23b were unambiguously assigned by x-ray crystallographic analysis (Figure 2).

Scheme 5.

Formation of isomeric triazinoisoquinolines under acidic acetal hydrolysis conditions.

Fig. 2.

X-ray structures of 23a (top) and 23b (bottom).

While the formation of 23b accomplished our goal to access the DEF ring system of 1, we were unable to identify acetal hydrolysis conditions avoiding the equilibration of this heterocyclic system to the isomeric 3,4-dihydro-2H-[1,2,4]triazino[4,3-b]isoquinoline-6,11-dione 23a. Furthermore, we were unable to cleave the benzyl protecting group in either 23a or 23b.

Accordingly, we modified the synthetic route shown in Schemes 4 and 5 by using the p-methoxybenzyl (PMB) protected azido hydrazine 24 (Scheme 6). Acylation of 15, C-arylation with malonate 17, and cyclization/decarboxylation proceeded uneventfully to give homophthalate 25. In two steps, the benzylic methylene group in 25 was converted to the dimethylacetal and the Staudinger-aza-Wittig reaction provided triazine 26 in high yield. Interestingly, controlled hydrolysis in cold sulfuric acid removed the PMB group in 72% yield without cleavage of the dimethyl acetal. The resulting dimethylacetal 27 was subjected to a brief (15 min) hydrolysis in conc. H₂SO₄ at room temperature to yield the desired unprotected triazino[2,3-b]isoquinoline 28 in the absence of any rearrangement products. The difference in the reactivities of 22 and 27 in the acid catalyzed hydrolysis is quite striking. We speculate that the increased solvation of the less lipophilic 27 increases the stability of the intermediate oxocarbenium ion, thus allowing for a quicker acetal cleavage at slightly lower temperature and preventing the triazine isomerization. The structural assignment of 28 was confirmed by x-ray analysis (Figure 3).

Scheme 6.

Preparation of the DEF ring system of noelaquinone from the PMB-protected azido hydrazine 24.

Fig. 3.

X-ray structure of 28.

Conclusions

We have demonstrated an efficient method to generate azido hydrazines, thus enabling the application of the Staudinger-aza-Wittig reaction of imides as a general method for the preparation of tetrahydro-1,2,4-triazines. For the synthesis of the DEF ring system of noelaquinone, the Cu(I)-catalyzed C-arylation of diethyl malonate was used to access the key intermediate homophthalimides 18 and 25, and the use of the PMB protective group and controlled acidic hydrolysis conditions led to the first preparation of the heterocycle 28, previously reported for the DEF ring moiety of the natural product noelaquinone. In the course of this synthesis, we also observed the equilibration of triazinoisoquinolines 23a and 23b under acidic conditions, most likely via an ANRORC⁹ mechanism.

Fig. 1.

Polycyclic furans isolated from the marine sponge Xestospongia.