An efficient oxa-Michael addition to diethyl vinylphosphonate under mild reaction conditions

Abstract

An oxa-Michael addition of various secondary and branched primary alcohols to diethyl vinylphosphonate was systematically studied and optimized. This efficient method precedes using of harsh reaction conditions (e.g. strong bases, high temperatures) and gains access to an important class of biologically active compounds in one step.

---

Acyclic nucleoside phosphonates (ANPs)¹ represent catabolically stable nucleotide analogues with a wide variety of biological properties.² Although ANPs bearing an unbranched phosphono-ethoxyethyl (PEE) moiety³ usually lack any antiviral activity, 6-oxopurine PEE derivatives have been found to display very promising antimalarial activity.⁴ To further study structure–activity relationship of ANPs, an efficient synthetic method for the introduction of the phosphonoethyl motif to various secondary alcohols is urgently needed.

Only few successful methods on how to introduce the phosphonoethyl moiety to alcohols have been reported, but with limited applicability and/or poor yields. Reactions of bromoethyl-phosphonate (and the corresponding tosyl or mesyl analogues) with secondary alcohol of 1,2-di-O-tert-butyldimethylsilyl-L-threose did not afford the expected alkylation product, possibly due to β-elimination of the ethylphosphonate derivatives under the basic conditions.⁵ An introduction of the hydroxyethyl group, followed by the subsequent reaction with trialkylphosphite under Arbuzov reaction conditions afforded dialkyl ethylphosphonate derivatives in low overall yields under harsh reaction conditions.^5,6 Finally, 2-hydroxyethylphosphonates were directly introduced to a phenolic hydroxyl under the Mitsunobu reaction conditions, although in 26–54% yields only.⁷

Thus, the most promising method for the attachment of the ethylphosphonate group is addition reaction to dialkyl vinylphosphonate (VP). Recently, aza-Michael⁸ and sulfa-Michael⁹ addition to VP derivatives has become a popular synthetic way for preparation of various biologically active molecules, since amines and thiols are usually very good nucleophiles. On the other hand, VP has been neglected as an acceptor in oxa-Michael addition reactions due to the reversibility of the reaction and low reactivity of alcohols, requiring harsh conditions and thus limiting the synthetic scope of the reaction.

An addition of primary alcohols to dialkyl VP has been reported using strong bases¹⁰ or carbonates at elevated temperature.¹¹ Similarly, the oxa-Michael additions of secondary alcohols to dialkyl VP (or vinyl phosphinates) were performed with strong bases at elevated temperatures and with moderate yields only.¹²

Herein we report on the efficient introduction of the ethylphosphonate moiety via optimized oxa-Michael addition of structurally diverse secondary alcohols to diethyl VP.¹³

First of all, an extensive study of the reaction of fluorohydrine 1 with diisopropyl bromo- and chloroethylphosphonates (2 and 3, Scheme 1) with various bases (LiH, NaH, t-BuOK, K₂CO₃, Cs₂CO₃, DBU, NaOH, BuLi, TBAF, Et₃N, MeONa, 1.0–1.5 eq.), in various solvents (DMF, ether, dioxane, toluene), and at various reaction temperatures verified that haloethylphosphonates¹⁴ are not suitable reagents for O-alkylation reactions.⁵ In all cases, an evident formation of the side products 5 and 6 was observed, affording the desired product 4 in 10% yield at best.

Scheme 1 Alkylation of 1 with haloethylphosphonates 2 and 3.

Next, the reaction of fluorohydrine 1 with commercially available diethyl vinylphosphonate (7, DEVP, Scheme 2) under various reaction conditions was studied and thoroughly optimized. The best result (52% of the desired product 8) was observed under the catalysis with KOH in dioxane (entry 5, Table 1).

Scheme 2 General addition of fluorohydrine 1 to DEVP 7.

Table 1 Optimization of the reaction of 1 with DEVP (7) (Scheme 2)

Entry	Base (eq.)	t (°C)	Solvent	Yield of 8 [%]^a
a Yield (GC-MS). Reaction conditions: DEVP (1.5 eq.), 24 h.
1	Cs2CO3 (1.0)	rt to 50	dioxane	22
2	BuLi (0.2)	−70 to rt	THF	13
3	NaH (0.2)	−70 to rt	THF	33
4	NaOH (0.2)	rt	dioxane	20
5	KOH (0.2)	rt	dioxane	52
6	RbOH (0.2)	rt	dioxane	31
7	CsOH (0.2)	rt	dioxane	34
8	t-BuOK (0.2)	rt	dioxane	39
9	DBU (0.1)	rt to 50	dioxane	0
10	Ba(OH)2 (0.2)	rt to 50	dioxane	0

---

The increasing size of the cation had a positive effect on the yields of product 8 (Table 2, Scheme 2), as the reactions with t-BuOK and CsOH reached ~75% conversion in 24 h (with t-BuOLi or t-BuONa less than 5%). A prolonged treatment only led to the increased rate of the retro-Michael reaction (GC-MS).

Table 2 Optimization of the reaction of 1 with DEVP (7) (Scheme 2)

Time [d]	Yield of 8 [%]^a
	t-BuOLi	t-BuONa	t-BuOK	CsOH·H2O
a Yield determined by GC-MS. Reaction conditions: DEVP (1.5 eq.), base (0.2 eq.), dioxane, rt.
1	0.0	3.7	74.6	74.2
3	3.8	3.8	76.0	67.8
5	0.0	3.3	76.8	69.0

---

Sekine et al.¹⁵ have described a reaction of the ribonucleoside derivatives with acrylonitrile in t-BuOH in the presence of Cs₂CO₃ to give 2′-O-cyanoethylated ribonucleosides. Later, their methodology was extended to acrylate esters.¹⁶ Although non-activated DEVP (7) is much poorer electrophile compared to acrylates, we decided to apply the Sekine's reaction conditions (1 eq. of Cs₂CO₃ and 20 eq. of DEVP in t-BuOH) for the addition of fluorohydrine 1 to DEVP (7) and, surprisingly, a quantitative conversion to phosphonate 8 was observed (Scheme 2). The finding, that DEVP (7) can serve as a good Michael acceptor for secondary alcohols under mild reaction conditions, encouraged us to further optimize this reaction.

To reduce amount of the reagents, a set of experiments with gradually decreasing amount of DEVP (7) was carried out (Table 3, Scheme 2). The reaction was completed in 10 h when 5 eq. of starting DEVP (7) were used, in 24 h with only 2 eq. of DEVP, and in prolonged reaction time (48 h) only 1 eq. of DEVP (7) suffices to obtain a 96% conversion to the product 8 (Table 3).

Table 3 Effect of DEVP (7) amount on the yield of 8 (Scheme 2)

Time [h]	Yield of 8 [%]^a
	1 eq.	2 eq.	5 eq.	10 eq.	20 eq.
a Yield determined by GC-MS. Reaction conditions: DEVP (1.0 eq.), Cs2CO3 (1.0 eq.), t-BuOH, rt.
1	0.2	0.0	1.5	3.0	2.2
2.5	1.3	3.5	13.3	24.2	16.6
5	7.1	27.8	69.4	89.5	71.0
10	27.0	76.2	98.8	99.4	96.5
24	81.4	98.3	99.5	99.9	99.9
48	96.0	98.6	99.5	99.9	99.9

---

The concentration profile of DEVP (7) in the studied oxa-Michael reaction follows the sigmoid curve typical for autocatalytic reactions (Fig. 1).¹⁷ Since the studied reaction (Scheme 2) is heterogeneous due to low solubility of Cs₂CO₃ in t-BuOH, we speculate that the reaction profile is a result of the solvation effects as phosphonates are known to strongly bind metal ions. Thus, binding of the cesium ion by phosphonate 8 formed during the reaction would increase the solubility of Cs₂CO₃ and subsequently the reaction rate.

Fig. 1 Reaction profile of the oxa-Michal reaction of fluorohydrine 1 as a function of the DEVP (7) concentration.

On the other hand, the catalysis of the studied oxa-Michael addition (Scheme 2) with K₂CO₃ turned out to be completely ineffective due to even worse solubility of K₂CO₃ in t-BuOH compared to Cs₂CO₃. Not even the addition of 1 eq. of [18]crown-6 to the reaction mixture really improved the yield of product 8 (8%, entry 1, Table 4) and the reaction mixture contained mainly the starting compound 1 and epoxide 5 (Scheme 2). Actually, in all experiments using crown ethers (entries 1–4, Table 4), the presence of the “naked” carbonate anion resulted in the preferential formation of epoxide 5. Catalysis of the reaction with KF on alumina (entry 5, Table 4) afforded compound 8 in a satisfactory 68% yield but the results are inferior to the above optimized methodology using Cs₂CO₃ in t-BuOH. All other experiments with various more or less soluble cesium salts, as well as attempts to reduce the amount of Cs₂CO₃, resulted in lack of reactivity or extension of the reaction times.

Table 4 Optimization of the reaction of 1 with DEVP (7) (Scheme 2)

Entry	Base	Additive	Yields (%)^a
			1	8	5
a Yield determined by GC-MS. Reaction conditions: DEVP (1.5 eq.), base (1.0 eq.), crown ether (1.0 eq.), t-BuOH, rt, 24 h.
1	K2CO3	[18]crown-6	60	8	32
2	Cs2CO3	[18]crown-6	9	32	59
3	Cs2CO3	DB[21]crown-7	2	33	65
4	Cs2CO3	DB[24]crown-8	3	52	45
5	KF/Al2O3	none	13	68	19
6	KHDMS	none	28	14	58

---

Finally, our optimized reaction conditions of the oxa-Michael addition (alcohol, 1.5 eq. of DEVP, 1 eq. of Cs₂CO₃, t-BuOH, rt) were subsequently applied to various compounds containing secondary or primary hydroxyl groups (Table 5). Thus, reaction of symmetrically protected glycerol 9 gave high yield of phosphonate 19 (86%, entry 1), while its more sterically hindered derivative 10 afforded phosphonate 20 in lower, but still satisfactory yield (55%, entry 2, Table 5). Structurally diverse secondary alcohols 11–13 were converted to corresponding phosphonates 21–23 in moderate to high yields (51–88%, entries 3–5, Table 5). Selective alkylation of secondary hydroxyl next to tertiary hydroxyl group in compound 14 led to the formation of compound 24 (entry 6), while both primary and secondary hydroxyls of the vicinal diol 15 react to give bisphosphonate 25 (entry 7, Table 5).

Table 5 Application of the oxa-Michael addition to DEVP (7)^a

Entry	Substrate	Product	Yield [%]^b
a Reaction conditions: DEVP (1.5 eq.), Cs2CO3 (1.0 eq.), t-BuOH, rt. b Isolated yields. c DEVP (2.0 eq.).
1			86
2			55
3			88
4			51
5			63
6			45
7			56^c
8			12
9			13
10			14

---

An analogous treatment of acyclic nucleosides 16 and 17 and ANP 18 afforded the desired phosphonoethylated products 26–28 in 12–14% yields only (entries 8–10, Table 5), due to the low solubility of the starting compounds in t-BuOH and presence of the reactive amino group in the C–6 position of the purine ring. To improve synthesis of such potentially biologically active compounds, protection of the amino group with a suitable lipophilic protecting group is strongly advisable.

In summary, the first efficient oxa-Michael reaction of secondary and sterically hindered primary alcohols with diethyl vinylphosphonate (7) under very mild reaction conditions is reported. To demonstrate the importance of our optimized one-step protocol in the field of medicinal chemistry, it is currently being employed for the preparation of wide range of compounds of biological interest.

This study was performed as a part of research project OZ40550506 of the IOCB, v.v.i. and was supported by Ministry of the Interior of the Czech Republic (VG20102015046) and by Gilead Sciences.

References

1. E. De Clercq, A. Holý, I. Rosenberg, T. Sakuma, J. Balzarini and P. C. Maudgal, Nature, 1986, 323, 464–467.
2. (a) A. Holý, Curr. Pharm. Des., 2003, 9, 2567–2592; (b) A. Holý, Antiviral Res., 2006, 71, 248–253; (c) E. De Clercq, Antiviral Res., 2007, 75, 1–13.
3. A. Holý, I. Rosenberg and H. Dvořáková, Collect. Czech. Chem. Commun., 1990, 55, 809–818.
4. (a) D. T. Keough, D. Hocková, A. Holý, L. Naesens, T. S. Skinner-Adams, J. de Jersey and L. W. Guddat, J. Med. Chem., 2009, 52, 4391–4399; (b) D. Hocková, A. Holý, M. Masojídková, D. T. Keough, J. de Jersey and L. W. Guddat, Bioorg. Med. Chem., 2009, 17, 6218–6232; (c) D. T. Keough, D. Hocková, M. Krečmerová, M. Česnek, A. Holý, L. Naesens, I. M. Brereton, D. J Winzor, J. de Jersey and L. W. Guddat, Mol. Biochem. Parasitol., 2010, 173, 165–169; (d) J. de Jersey, A. Holý, D. Hocková, L. Naesens, D. T. Keough and L. W. Guddat, Curr. Top. Med. Chem., 2011, 11, 2085–2102.
5. Q. Huang and P. Herdewijn, Nucleosides, Nucleotides Nucleic Acids, 2009, 28, 337–351.
6. (a) G. D. Diana, E. S. Zalay, U. J. Salvador, F. Pancic and B. Steinberg, J. Med. Chem., 1984, 27, 691–694; (b) H. Liu, W. Li and C. U. Kim, Bioorg. Med. Chem. Lett., 1997, 7, 1419–1420; (c) D. V. Yashunsky, Y. E. Tsvetkov and A. V. Nikolaev, Tetrahedron Lett., 2000, 41, 3665–3668; (d) P. Jurečka, P. Vojtíšek, K. Novotný, J. Rohovec and I. Lukeš, J. Chem. Soc., Perkin Trans. 2, 2002, 1370–1377; (e) S. Iyer, J. M. Younker, P. G. Czyryca and A. C. Hengge, Bioorg. Med. Chem. Lett., 2004, 14, 5931–5935.
7. Int. Pat. WO2005/117904 A2, 2005..
8. (a) G. David, B. Boutevin and Y. Hervaud, Phosphorus, Sulfur Silicon Relat. Elem., 2005, 180, 2201–2209; (b) E. V. Matveeva, P. V. Petrovskii and I. L. Odinets, Tetrahedron Lett., 2008, 49, 6129–6133; (c) E. V. Matveeva, A. E. Shipov and I. L. Odinets, Phosphorus, Sulfur Silicon Relat. Elem., 2011, 186, 698–706.
9. F. Gao, X. Yan and K. Auclair, Chem.–Eur. J., 2009, 15, 2064–2070.
10. (a) H.-J. Kleiner and C. Schumann, Phosphorus and Sulfur, 1982, 13, 363–370; (b) H. J. Cristau and D. Virieux, Tetrahedron Lett., 1999, 40, 703–706; (c) L. Delain-Bioton, D. Villemin and P.-A. Jaffrès, Eur. J. Org. Chem., 2007, 1274–1286; (d) J. Köhler, H. Keul and M. Möller, Chem. Commun., 2011, 47, 8148–8150.
11. (a) E. Brunet, M. J. Mata, O. Juanes, H. M. H. Alhendawi, C. Cerro and J. C. Rodríguez-Ubis, Tetrahedron: Asymmetry, 2006, 17, 347–354; (b) J. L. Kelley, J. A. Linn, E. W. McLean and J. V. Tuttle, J. Med. Chem., 1993, 36, 3455–3463.
12. (a) A. N. Pudovik and R. G. Kuzovleva, J. Gen. Chem. USSR, 1963, 33, 2755–2760; (b) Von G. Märkl and B. Merkl, Tetrahedron Lett., 1981, 22, 4463–4466.
13. O. Baszczyňski, P. Jansa, D. Hocková, Z. Janeba, M. Dračínský, A. Holý, D. T. Keough, J. de Jersey and L. W. Guddat, Collect. Symp. Ser., 2011, 12, 302–303.
14. P. Jansa, A. Holý, M. Dračínský, O. Baszczyňski, M. Česnek and Z. Janeba, Green Chem., 2011, 13, 882–888.
15. H. Saneyoshi, K. Seio and M. Sekine, J. Org. Chem., 2005, 70, 10453–10460.
16. T. Yamada, N. Okaniwa, H. Saneyoshi, A. Ohkubo, K. Seio, T. Nagata, Y. Aoki, S. Takeda and M. Sekine, J. Org. Chem., 2011, 76, 3042–3053.
17. R. Plasson, A. Brandenburg, L. Jullien and H. Bersini, J. Phys. Chem. A, 2011, 115, 8073–8085.

---

Footnotes

† Electronic Supplementary Information (ESI) available. See DOI: 10.1039/c2ra00938b/

‡ This work is dedicated to Prof. Antonín Holý on the occasion of his 75th birthday.

---