3-Methylene-2,4-chromandione in situ trapping: introducing molecular diversity on 4-hydroxycoumarin

Abstract

3-Methylene-2,4-chromandione trapped in a solid-state stable Mannich adduct, is released in solution. In the presence of different nucleophiles, this highly reactive intermediate allows the introduction of molecular diversity on C-3 position of 4-hydroxycoumarin.

---

The reactivity of ortho-quinone methides (o-QMs) has been largely exploited in natural products, total synthesis, medicinal chemistry and biochemistry. And yet, for a long time, o-QMs have been underused, mainly due to their transient nature.¹ Quinone methides have been hardly ever isolated, due to fast reactions favored by rearomatization into phenol, acting as the driving force (Scheme 1, eqn (a)). Gardner² detected their existence in 1963 by spectroscopy at −100 °C, even if more than fifty years before, in 1907, Fries got preliminary suspicion of the reality of such molecular species.³ Their high reactivity makes it attractive, in particular for multi-component reactions.⁴ Over the past years, they have regained a new interest for drug targeting or fluorescent probe.⁵ Biomimetic synthesis of Schefflone, a monoterpenoid was recently performed by oxidative trimerization of o-QM.⁶ A large spectrum for their biological activities was highlighted with the understanding of DNA alkylation, anti-tumoral or anti-bacterial properties and so on.⁷ As a notorious example, anti-diabetic Troglitazone was withdrawn in 2000 due to in vivo formation of o-QM metabolite displaying prejudicial secondary effects.⁸

Scheme 1 Structural and reactivity analogy of ortho-quinone methide and 3-methylene-2,4-chromandione.

Recently, our laboratory reported an effective procedure for C-3 reductive alkylation of 4-hydroxycoumarin by a dehydrogenative oxidation of benzylic alcohols in the presence of RuCl₂(PPh₃)₂ (5 mol%), KOH (0.2 eq.) in ^tertamyl alcohol under microwave irradiation at 140 °C in 2 hours.⁹ Supposed mechanism is described as a first step of activation of the alcoholic substrate by metallo-catalyzed dehydrogenative oxidation, followed by a Knoevenagel condensation/reduction sequence, under one pot conditions. The Knoevenagel adduct (3-methylene-2,4-chromandione) displaying similar structural analogy (Scheme 1, eqn (b)) with classical o-QM, is a versatile substrate for cycloaddition reactions.¹⁰ To a larger extent, 3-methylene-2,4-chromandione have been reported as in situ generated partner for multi-component reactions (MCRs), providing a powerful platform to access diversity as well as complexity. Following this strategy, pyranocoumarins were obtained by Knoevenagel-electrocyclic reaction starting from 4-hydroxycoumarin, an enal and an amine.¹¹ o-QMs analogs have been involved in numerous cycloadditions ([4 + 2]¹² or [4 + 1]¹³) and other multi-component or domino reactions.¹⁴ The common points of this strategy were to access the high added value molecules in a linear synthetic strategy.

The approach developed in the laboratory is to trap the 3-methylene-2,4-chromandione highly reactive intermediate in a solid-state stable Mannich adduct. This adduct in solution allows the formation of an equilibrium with the o-QM adduct which will react with various nucleophilic species (Scheme 2). Therefore, we decided to focus on an elegant approach to trap this intermediate by different methods including simple reduction methods. The methodology device should also limit the percentage of undesired DIM, which is often observable in all the synthesis reported in literature (Table 1).¹⁵

Scheme 2 Equilibrium between Mannich adduct and methylene chromandione.

Table 1 Nature of nucleophilic amine

Entry	Amine	Ratio
		Product	DIM
1	Pyrrolidine	100	0
2	Piperidine	90	10^17
3	Dimethylamine	69	31
4	Morpholine	91	9
5	Diethylamine	60	40
6	Aniline	27	73
7	Isopropylamine	93	7
8	Proline	15	85

---

Preliminary study began with the choice of the amine partner involved in the Mannich reaction (Table 1). A set of primary and secondary amines were screened and pyrrolidine was the sole amine affording the desired Mannich adduct without any trace of dimer. Other amines (piperidine, morpholine and isopropylamine) led to 10% of undesired dimer adduct (entries 2, 4, 7). Dimethylamine, diethylamine (entries 3, 5) were less efficient with almost 40% of dimer. Aniline and proline (entries 6, 8) led to partial conversion in favour of expected product with in both cases, dimer as the major product. These results were almost in accordance with the nucleophilicity scale developed by Mayr and co-workers in 2009.¹⁶

Using the conditions outlined in entry 1 (Table 1), we tested the scope of this procedure with a range of aromatic aldehydes. The three-component reaction of 4-hydroxycoumarin, pyrrolidine and a set of benzaldehydes was effective in good to excellent yields, regarding the stoichiometric ratio of each partner (Table 2). No trace of bis-coumarin product was detected for compounds 1a–k. Dichloromethane was the best solvent, allowing the precipitation of the Mannich adduct at the end of the reaction and thus facilitating its isolation, by simple filtration (Table 2).

Table 2 Mannich reaction of 4-hydroxycoumarin (and analogs) with aldehydes^a

a Conditions: 4-hydroxycoumarin (1 mmol), aldehyde (1 mmol), pyrrolidine (1 mmol), CH₂Cl₂ (2 ml), rt, 2 h.

---

The electronic nature of the substituent on aromatic ring did not affect the reaction and excellent yields were obtained for all substrates described in Table 2. Electronic effects (electron-donating group such as 1c, 1d or electron-withdrawing group for 1f, 1g) on aryl aldehyde do not affect the issue of the reaction. Halogen was also well tolerated (1e). Other substituted benzaldehydes such as 4-(pyridin-4-yl)benzaldehyde, 2,2-difluorobenzo[d][1,3]dioxole-5-carbaldehyde, 4-phenoxybenzaldehyde or 4-benzylbenzaldehyde were coupled with good to excellent yields (1h–k). Replacement of 4-hydroxycoumarin scaffold either by quinolone-2,4-diol (1l, 1m) or by 4-hydroxythiocoumarin (1n, 1o) resulted in similar yields.

To our delight, the reaction could be performed on 10 gram scale for compound 1a, our benchmark compound for the following studies. Three reactions were selected for evaluation of the reactivity of the Mannich adduct, as pointed out in Scheme 3: hydrogenation, Michael addition or tandem Michael addition–cyclization (Scheme 3). Hydrogenation allows the formation of linear C-3 alkylated compounds, which are not so obvious to obtain.¹⁸ Treating Mannich adduct 1a under H₂ atmosphere (1 bar) in anhydrous ethanol with 10% Pd/C (5 mol%) afforded compound 2a quantitatively.

Scheme 3 Overview of the reactivity of intermediate 1a.

Non-symmetrical dimer 3a was obtained in an excellent 97% yield by reaction of 1a with stoichiometric quantity of 4-hydroxythiocoumarin in acetonitrile at 50 °C during 3 hours. In this case, the mechanism reaction is supposed to be assimilated to Michael addition with nucleophilic 4-hydroxythiocoumarin reacting on o-QM liberated in solution. The last reaction explored is the MCR tandem reaction of Michael addition followed by cyclization to provide dihydropyrano[2,3-c]chromene skeleton. The reaction of 1a with malononitrile (1 eq.) in MeCN at 50 °C during 3 hours furnished derivative 4a in 98% yield. Thermal stability of compound 1a supplied important information on the equilibrium between Mannich adduct and o-QM. Heating a solution of 1a in MeCN [0.25 M] at 50 °C during 15 min only led to recovery of the starting material. After 15 min at 90 °C, a significant ratio of dimer could be quantified by ¹H NMR at a level of 10%. The 1/9 mixture of dimer/1a was heating at 15 min at 110 °C leading to a displacement of the equilibrium in favour of a more important ratio of the dimer: 33% of DIM and 67% of substrate 1a. The stability experiments were run under microwave irradiation to speed up the kinetics and evaluate the stability of 1a. In regard with these observations, to limit the formation of the dimer, classic thermal activation was selected to the detriment of microwave irradiation.

All the Mannich adducts 1 (obtained in Table 2) were submitted to hydrogenation in conditions described in Scheme 3 to allow the formation of C-3 alkylated coumarins (Fig. 1). Excellent yields were obtained for most of the cases engaged in the reduction (11 successful reactions on 15 compounds) with isolation by simple filtration (no need to additional column chromatography for increasing the compounds purity).

Fig. 1 Hydrogenation scope^a. ^a Conditions: 1a, H₂, Pd/C, EtOH, rt, 24 h.

As expected, derivatives 1e–g could not be reduced with good chemoselectivity and complete reduction of para substituent was observed. In the case of 1f, nitrile group was reduced into CH₃ within 13% yield. Substrates with quinolone diol skeleton were reduced successfully into 2i and 2j, whereas compound 2k bearing 4-hydroxythiocoumarin, was obtained in poor yield of 23%, probably due to poisoning of the catalyst. Same observation was reported for the reduction for 1n and thus was not reported.

As in solution, Mannich adduct and o-QM intermediate are in equilibrium, any nucleophile could be selected to perform Michael addition and trap irreversibly the conjugated enone, with recovery of the aromaticity. Literature has reported many conditions for the formation of dicoumarols such as DIM. To our delight, no reference was found concerning the formation of unsymmetrical dimers. Pyrrolidinyl derivatives 1 were dissolved in acetonitrile, in the presence of a stoichiometric quantity of 4-hydroxythiocoumarin, at 50 °C (Table 3). In 3 hours, the reaction ran to completion in all cases with excellent isolated yields higher than 89%. Symmetrical DIM could be also obtained successfully in these conditions in 98% yield.

Table 3 Formation of non-symmetrical dimers^a

a Conditions: 1 (1 mmol), 4-hydroxythiocoumarin (1 mmol), CH₃CN (2.5 ml), 50 °C, 3 h.

---

We decided to apply our methodology to the synthesis of 2-amino-4H-pyrans prepared by a tandem Michael addition–cyclization sequence (Table 4). The access to functionalized 2-amino-3-cyano-4H-pyrans annulated heterocycle 4 has been already published by simple multi-component reaction between aldehydes, malononitriles and 4-hydroxycoumarin in the presence of a wide range of catalysts such as (NH₄)₂HPO₄,¹⁹ proline,¹⁵ TBAB,²⁰ meglumine,²¹ fermented Baker's yeast,²² ZnFe₂O₄,²³ urea.²⁴ Simple aqueous conditions²⁵ were also efficient due to the hydrophobic effect as the driving force. Catalyst-free conditions in recyclable glycerol were recently developed.²⁶ In a direct comparison to the above cited strategies, our approach is a stepwise process, based on the isolation of the o-QM intermediate and benefits from catalyst free conditions. With the optimized conditions, the scope and generality of this protocol was next examined by employing various o-QM intermediates 1.

Table 4 Tandem Michael addition/cyclization^a

a Conditions: 1 (1 mmol), malononitrile (1 mmol), CH₃CN (2.5 ml), 50 °C, 3 h.

---

A family of eight new 2-amino-4H-pyran compounds 4 was synthesized using this protocol (Table 4). Good to excellent yields were obtained when 4-hydroxycoumarin skeleton was involved, whereas in the presence of 4-hydroxythiocoumarin, compound 4h was prepared with a poor yield of 25% due to problems in the solubility of the starting material, resulting in a low kinetics in this set of experimental conditions. In some cases such as 4b or 4e, the decrease of the yield could be also associated to the formation of the dimer present in significant quantities. As a perspective of these preliminary studies on the o-QM intermediates, we decided to study simple thiophenol nucleophiles and examine their reactivity towards o-QM intermediate 1 (Scheme 4).²⁷

Scheme 4 Introduction of simple nucleophiles.

Compounds 5a–c were obtained in moderate yield of 57–76% and conditions are under optimisation due do the problems of separation and the requirement of additional step of purification by column chromatography. Addition of acetophenone (4 eq.) was performed in a modest 50% yield under microwave activation (80 °C, 1 h).²⁸

Conclusions

An efficient synthesis of 4-hydroxycoumarin derivatives in C-3 position has been developed by taking advantage of o-QM intermediate. A wide range of reagents H₂, malononitrile, 4-hydroxythiocoumarin, thiophenols were used and broadened the scope. Due to the easily scalable step to generate derivatives 1, the challenging point is to generalize the second step of the reaction to any nucleophilic species. Aliphatic aldehydes were not tested in the first stage of the reaction and will deserve also some attention in the future.

Notes and references

1. R. W. van de Water and T. R. R. Pettus, Tetrahedron, 2002, 58, 5367.
2. P. D. Gardner, H. S. Rafsanjani and L. Rand, J. Am. Chem. Soc., 1959, 81, 3364.
3. K. Fries and K. Kann, Justus Liebigs Ann. Chem., 1907, 353, 335.
4. (a) N. J. Willis and C. D. Bray, Chem.–Eur. J., 2012, 18, 9160; (b) M. S. Singh, A. Nagaraju, N. Anand and S. Chowdhury, RSC Adv., 2014, 4, 55924.
5. (a) S.-T. Huang, K.-N. Ting and K.-L. Wang, Anal. Chim. Acta, 2008, 620, 120; (b) S. Gnaim and D. Shabat, Acc. Chem. Res., 2014, 47, 2970.
6. D. Liao, H. Li and X. Lei, Org. Lett., 2012, 14, 18.
7. (a) V.-S. Li, D. Choi, M.-S. Tang and H. Kohn, J. Am. Chem. Soc., 1996, 118, 3765; (b) G. B. Evans, R. H. Furneaux, M. B. Gravestock, G. P. Lynch and G. K. Scott, Bioorg. Med. Chem., 1999, 7, 1953; (c) S. B. Herzon and C. M. Woo, Nat. Prod. Rep., 2012, 29, 87.
8. (a) K. Kassahun, P. G. Pearson, W. Tang, I. McIntosh, K. Leung, C. Elmore, D. Dean, R. Wang, G. Doss and T. A. Baillie, Chem. Res. Toxicol., 2001, 14, 62; (b) V. A. Dixit and P. V. Bharatam, Chem. Res. Toxicol., 2011, 24, 1113.
9. A. Montagut-Romans, M. Boulven, M. Lemaire and F. Popowycz, New J. Chem., 2014, 1794.
10. G. Appendino, G. Cravotto, L. Toma, R. Annunziata and G. Palmisano, J. Org. Chem., 1994, 59, 5556.
11. G. Cravotto, G. M. Nano and S. Tagliapietra, Synthesis, 2001, 49.
12. V. Nair, C. N. Jayan, K. V. Radhakrishnan, G. Anilkumar and N. P. Rath, Tetrahedron, 2001, 57, 5807.
13. (a) V. Nair, R. S. Menon, A. U. Vinod and S. Viji, Tetrahedron Lett., 2002, 43, 2293; (b) A. Kumar, S. Srivastava and G. Gupta, Green Chem., 2012, 14, 3269.
14. (a) G. Appendino, G. Cravotto, A. Minassi and G. Palmisano, Eur. J. Org. Chem., 2001, 3711; (b) I. Yavari, M. Sabbaghan and Z. Hossaini, Synlett, 2008, 8, 1153; (c) S. Tu, C. Li, F. Shi, D. Zhou, Q. Shao, L. Cao and B. Jiang, Synthesis, 2008, 369; (d) Y. Gu, J. Barrault and F. Jérôme, Adv. Synth. Catal., 2009, 351, 3269; (e) J.-N. Tan, M. Li and Y. Gu, Green Chem., 2010, 12, 908; (f) X.-S. Wang, J. Zhou, K. Yang and M.-M. Zhang, Synth. Commun., 2010, 40, 3332; (g) W. Ma, X. Wang, F. Yan, L. Wu and Y. Wang, Monatsh. Chem., 2011, 142, 163; (h) M. Khoobi, A. Foroumadi, S. Emami, M. Safavi, G. Dehghan, B. H. Alizadeh, A. Ramazani, S. K. Ardestani and A. Shafiee, Chem. Biol. Drug Des., 2011, 78, 580; (i) A. Kumar, M. Kumar, M. Kumar Gupta and L. P. Gupta, RSC Adv., 2012, 2, 8277; (j) S. M. Rajesh, S. Perumal, J. Carlos Menénedez, S. Pandian and R. Murugesan, Tetrahedron, 2012, 68, 5631; (k) M. S. Singh, G. C. Nandi and S. Samai, Green Chem., 2012, 14, 447; (l) P. P. Ghosh, G. Pal, S. Paul and A. R. Das, Green Chem., 2012, 14, 2691; (m) C.-J. Lee, Y.-J. Jang, Z.-Z. Wu and W. Lin, Org. Lett., 2012, 14, 1906; (n) M. G. Dekamin, M. Eslami and A. Maleki, Tetrahedron, 2013, 69, 1074; (o) A. T. Khan, M. Lal and R. S. Basha, Synthesis, 2013, 406; (p) S. Ahadi, M. Zolghadr, H. R. Khavasi and A. Bazgir, Org. Biomol. Chem., 2013, 11, 279.
15. C. Wolfrum and F. Bohlmann, Liebigs Ann. Chem., 1989, 295.
16. T. Kanzian, T. A. Nigst, A. Maier, S. Pichl and H. Mayr, Eur. J. Org. Chem., 2009, 6379.
17. C. Mukhopadhyay, S. Rana and R. J. Butcher, Synth. Commun., 2012, 42, 3077 . Isolated yield of 78–81% when piperidine was used instead of pyrrolidine. No information if DIM was detected, but the yields are in accordance with the ratio measured in our experiments.
18. General methods to access linear C-3 alkylated 4-hydroxycoumarins: (a) P. Bravo, C. Ticozzi and G. Cavicchio, Synthesis, 1985, 894; (b) A. S. Kende, K. Koch and C. A. Smith, J. Am. Chem. Soc., 1988, 110, 2210; (c) G. Cravotto, G. M. Nano, G. Palmisano and S. Tagliapietra, Synthesis, 2003, 1286; (d) M. Gebauer, Bioorg. Med. Chem., 2007, 15, 2414; (e) R.-Y. Yang, D. Kizer, H. Wu, E. Volckova, X.-S. Miao, S. M. Ali, M. Tandon, R. E. Savage, T. C. K. Chan and M. A. Ashwell, Bioorg. Med. Chem., 2008, 16, 5635; (f) C. R. Reddy, N. Kiranmai, K. Johny, M. Pendke and P. Naresh, Synthesis, 2009, 399; (g) D. B. Ramachary and M. Kishor, Org. Biomol. Chem., 2010, 8, 2859; (h) G. Palmisano, F. Tibiletti, A. Penoni, F. Colombo, S. Tollari, D. Garella, S. Tagliapietra and G. Cravotto, Ultrason. Sonochem., 2011, 18, 652; (i) Y.-J. Shue and S.-C. Yang, Tetrahedron Lett., 2012, 53, 1380; (j) M. Baron, E. Métay, M. Lemaire and F. Popowycz, J. Org. Chem., 2012, 77, 3598.
19. S. Abdolmohammadi and S. Balalaie, Tetrahedron Lett., 2007, 48, 3299.
20. J. M. Khurana and S. Kumar, Tetrahedron Lett., 2009, 50, 4125.
21. R.-Y. Guo, Z.-M. An, L.-P. Mo, R.-Z. Wang, H.-X. Liu, S.-W. Wang and Z.-H. Zhang, ACS Comb. Sci., 2013, 15, 557.
22. M. Saha and A. K. Pal, Synth. Commun., 2013, 43, 1708.
23. P. Das, A. Dutta, A. Bhaumik and C. Mukhopadhyay, Green Chem., 2014, 16, 1426.
24. G. Brahmachari and B. Banerjee, ACS Sustainable Chem. Eng., 2014, 2, 411.
25. A. Shaabani, S. Samadi and A. Rahmati, Synth. Commun., 2007, 37, 491.
26. H. R. Safaei, M. Shekouhy, S. Rahmanpur and A. Shirinfeshan, Green Chem., 2012, 14, 1696.
27. (a) K. Nawaz, M. A. Munawar and M. Siddiq, J. Chem. Soc. Pak., 1991, 13, 272; (b) K. Tasneem and K. Z. Khan, Asian J. Chem., 2008, 20, 929.
28. Analytical data were in accordance with those reported in literature: I. Manolov and N. Danchev, Arch. Pharm. Pharm. Med. Chem., 2003, 2, 83.

---

Footnote

† Electronic supplementary information (ESI) available: Synthetic procedures, characterisation of compounds, NMR spectra. See DOI: 10.1039/c5ra22440c

---