Novel chiral tetramic acid-derived diols: organocatalytic facile synthesis and unique structural properties

Abstract

Organocatalytic tandem reactions of L-phenylalanine-derived tetramic acid with aldehydes allow a one-pot and high-yielding access to a diverse range of novel chiral diols in enantiomerically pure forms. In addition, a new entry of the diols, featuring their unique structures associated with C₂- and pseudo C₂-symmetric chiral motifs, is reported.

---

Chiral diols have been extensively used as effective ligands for chiral catalysts, which are involved in a multitude of powerful applications in asymmetric organic synthesis, and therefore belong to an important class of privileged molecules.^1–6 In general, these chiral diols are designed to have the C₂-symmetry element in conformationally rigid systems, featuring two spacially close pairs of hydroxyl groups asymmetrically disposed across the mirror planes of their entire structures to create chiral environment for expressing enantiomeric recognition. Over the past few decades, impressive progress has been made in the development of enantioselective synthetic methods promoted by asymmetric catalysis of metals and/or prochiral molecules under the influence of the chiral diols, such as diethyl tartrate (DET),² hydrobenzoin,³ α,α,α′,α′-tetraaryl-1,3-dioxolan-4,5-dimethanol (TADDOL),⁴ 1,1′-bi-2-naphthol (BINOL),⁵ and 1,1′-biaryl-2,2′-dimethanol (BAMOL)⁶ (Fig. 1). Currently available chiral diols fulfilling the structural and functional requirements for efficient use as ligands are generally limited to just a few simple compounds. Hence, it is a subject of great importance to develop a new alternative series of chiral diols for future progress in stereocontrolled syntheses of complex target molecules, which will be beneficial for chemical, biological, and medicinal applications. In the present communication, we report a new entry of chiral tetramic acid-derived diols possessing C₂- and pseudo C₂-symmetric elements, which are easily available on a gram scale by simple organocatalytic procedures and can be custom-designed with respect to their steric and conformational properties according to structural and functional attributes of the composition.

Fig. 1 Structures of representative chiral diols.

Our initial synthetic efforts have been made to exploit an interesting finding of unique reactivity of simple cyclic β-diketones toward one-pot synthesis of structurally analogous racemic diols,⁷ which can be understood to arise from the base-catalyzed tandem reactions of Knoevenagel condensation⁸ followed by Michael addition. Applying this method to biologically attractive chiral tetramic acids 1,⁹ which can be expected to react with aldehydes 2 to give a new type of chiral diols 3 via the intermediacy of α,β-unsaturated 1,3-diketones (ene-diones) 4 as illustrated in Scheme 1, one may encounter difficulties in controlling the stereochemistry of reaction products, largely due to high susceptibility of the tetramic acid cores to suffer troublesome epimerization even under mild conditions.¹⁰ In fact, when a 2 : 1 mixture of L-phenylalanine-derived tetramic acid 1a and formaldehyde 2a (37% aqueous solution) was subjected to the analogous reaction conditions with 20 mol% of piperidine (relative to 2a) in ethanol, a smooth reaction occurred at r.t. to reach completion after 3 h and the chiral diol 3a was obtained in 73% isolated yield (Table 1, entry 1). To our disappointment, the ¹H NMR analysis of this product ensured detectable loss of the stereochemical integrity, being contaminated with about 6% of its C5 epimeric isomer. Such a behavior was reproduced using the catalytic amount of pyrrolidine, which gave a 90 : 10 epimeric mixture of 3a in 78% yield after 2 h (Table 1, entry 2). For both cases, we also found that the stereochemical quality seriously suffered with prolonged reaction times, thereby resulting in considerable decrease of the diastereomeric purity of the product (see ESI†).

Scheme 1 Pathway for diols 3.

Table 1 Catalyst screening for the reaction of 1a

Entry	2	Catalyst [mol%]	Solvent	T	t [h]	3	Yield^a [%]
a Isolated yield.b Epimeric products were obtained in 3a:epi-3a ratios of 94 : 6 (for entry 1) and 90 : 10 (for entry 2). See ESI.†c The reaction was conducted on a gram scale (1.23 g of 3b).
1	2a	Piperidine (20)	EtOH	r.t.	3	3a^b	73
2	2a	Pyrrolidine (20)	EtOH	r.t.	2	3a^b	78
3	2a	L-Proline (20)	EtOH	r.t.	5	3a	96
4	2b	L-Proline (20)	EtOH	0 °C	2	3b	100
5	2b	Pyrrolidine/AcOH = 1/1 (20)	EtOH	0 °C	3	3b	100
6	2b	L-Serine (20)	EtOH	0 °C	49	3b	99
7	2b	L-Lysine (20)	EtOH	0 °C	20	3b	83
8	2b	L-Aspartic acid (20)	EtOH	0 °C	291	3b	91
9	2b	L-Proline (20)	H2O	r.t.	5	3b	93
10^c	2b	L-Proline (1)	EtOH	r.t.	1	3b	100

---

However, a dramatic change was observed when the catalytic amount of L-proline was used instead of the above two bases. In this case, the reaction proceeded smoothly at r.t. to allow efficient production of 3a as an enantiomerically pure form in 96% yield without any other reaction products after a reaction time of 5 h (Table 1, entry 3). This catalytic system was also effective for the reaction with propanal 1b. Indeed, the reaction proceeded very rapidly at r.t. to reach completion within 20 min with a quantitative product yield of the enantiopure diol 3b, and still took place very efficiently over 2 h even at 0 °C to give the essentially same result (Table 1, entry 4). In order to search for a replacement of L-proline, we next explored the possibility of using neutral species as a catalyst. When an equimolar mixture of pyrrolidine and acetic acid, which represents structural components of the proline architecture, was employed for the reaction at 0 °C, closely comparable results were obtained with quantitative yield of enantiopure 3b (Table 1, entry 5). The other amino acids such as L-serine, L-lysine, and L-aspartic acid were also found effective but underwent rather slow reactions to reach 99%, 83%, and 91% product yields for prolonged reaction periods of 49, 20, and 291 h, respectively (Table 1, entries 6–8). Obviously, the cyclic amino acid, proline, has substantial promise for the efficient catalysis of the reactions to afford the chiral diols,¹¹ where chemical reactivity of carbonyl electrophiles present in the reaction mixture would be enhanced through formation of proline-derived iminium intermediates.^12,13 Also worthy of note is that with the use of L-proline, the approach provided another option to use aqueous media for the reaction as well as allowed easy access in ethanolic solution at r.t. to a gram scale quantity (1.23 g, 100% isolated yield) of 3b with a catalyst loading as low as 1 mol% (Table 1, entries 9 and 10).

Having identified proline as the optimal catalyst, investigation into the substrate scope of this methodology was next explored with a number of aldehydes. The results presented in Table 2 demonstrate that the proline catalyzed reaction has a high degree of functional-group tolerance with respect to the aldehyde substituents (Table 2, entries 1–7). It is remarkable to note that sterically hindered aldehydes such as 2c, 2h, and 2i also reacted with 1a, albeit sluggishly, leading to good to high yields of the corresponding diols 3c, 3h, and 3i, respectively (Table 2, entries 1, 6, and 7). The versatility of this catalytic system was further enhanced by applicability of the reaction to a Boc-protected tetramic acid 1b, a substitute for 1a, which resulted in rapid and high-yield production of the desired Boc derivative of 3b (denoted as 3b′) (Table 2, entry 8). In contrast to these results, we must also note that the most sterically hindered 9-anthraldehyde 2j exhibited exceptional reactivity, ultimately leading to exclusive formation of the ene-dione 4j. Additionally, the reaction with acetone as a representative example of ketones in the presence of catalytic even stoichiometric amounts of L-proline also gave the relevant ene-dione 5 quantitatively in each case (Scheme 2).

Table 2 Scope of proline-catalyzed reaction of 1 with aldehydes 2^a

Entry	1	2	R	Time [h]	3	Yield^b [%]
a Conditions: aldehyde (0.12–0.30 mmol), tetramic acid (0.24–0.60 mmol), L-proline (20 mol%), EtOH (0.5–0.9 mL), r.t.b Isolated yield.c N-Boc derivative of 1a.
1	1a	2c	2-Methylbutyl	26	3c	82
2	1a	2d	Ph	3	3d	100
3	1a	2e	4-tBuPh	2	3e	90
4	1a	2f	3-ClPh	3	3f	99
5	1a	2g	4-NO2Ph	2	3g	99
6	1a	2h	1-Naphtyl	26	3h	96
7	1a	2i	2-Naphtyl	20	3i	96
8	1b^c	2b	Et	1	3b′	96

---

Scheme 2 Reaction of 1 with 2j and acetone.

Since the above result had given an indication for a reaction mechanism involving intermediacy of 4, which would generate 3, we next attempted to confirm this mechanistic possibility by reconsidering the reaction trajectory. In the case of entry 7, the reaction progress could be followed visually through monitoring the reaction mixtures by thin layer chromatography (TLC), which showed a spot assignable to 4h, together with those of 1a and 3h. At the reaction time of 5 h, the resulting mixture allowed efficient column separation where the fraction was flushed through the silica-gel column, isolating 4h in 62% yield as an inseparable mixture of cis- and trans-isomers.¹⁴ With 4h in hand, we then examined the reaction of 4h with 1.0 equiv. of 1a at r.t. in the presence and absence of the catalytic amount (20 mol%) of L-proline. As a result, the catalytic conditions proved potentially reactive to give rise to 3h with 83% isolated yield after a period of 50 h, whereas no reaction was found to occur without L-proline. Thus, the mechanistic pathway can be understood as the tandem Knoevenagel condensation/Michael addition sequence, in which either the cis- or trans-isomer of the resultant ene-diones equally reacts with the remaining tetramic acids to converge to form the uniform structures with the given molecular chirality.

Fig. 2 illustrates ¹H NMR spectroscopic characteristics of 3a and 3b, obtained with DMSO-d₆ as a solvent. There is a marked difference in the spectral features between these compounds. As represented in Fig. 2a, the methylene-bridged diol 3a displays simple resonance patterns, indicating a genuine spectroscopic equivalence of the structural units. This observation is well in accordance with our assumption that 3a should adopt the C₂-symmetric structure. In contrast to this, the ¹H NMR spectrum for 3b shows more complex patterns with two sets of proton resonances for the relevant tetramic acid segments, indicating unsymmetrical nature of the tetramic acid cores (Fig. 2b). This suggests that the tetramic acid units are less free to rotate at r.t. around the single bonds connecting the methine junction. To our surprise, this spectral appearance essentially remains unchanged upon changing the solvent to acetone-d₆ and methanol-d₄ as well as upon heating up to 55 °C in methanol-d₄ (see ESI†), indicating that energy barrier for the internal rotations is too high to allow coalescence of the separated resonances and the molecule adopts a severely restricted conformation. This behavior can be considered as a typical characteristic of the overall geometry, which should be categorized as a pseudo C₂-symmetric system.

Fig. 2 ¹H NMR spectra (300 MHz, DMSO-d₆) for (a) 3a and (b) 3b.

In an attempt to have a better understanding of conformational preference of the pseudo C₂-symmetric chiral diols, a single crystal X-ray diffraction study was undertaken on p-nitrophenyl-substituted diol 3g, whose X-ray quality crystals were grown by slow evaporation from a chloroform–methanol solution. Indeed, the X-ray results show that the compound adopts the chiral space group P2₁2₁2₁ where the asymmetric unit cell contains one constituent molecule with an encapsulated methanol solvent as depicted in Fig. 3.^15,16 In this structure, two tetramic acid groups are oriented preferentially in the opposite direction so as to form intramolecular hydrogen bond between one of the amide oxygen atoms and enolic proton of another tetramic acid group with a short O(3)⋯O(6) distance of 2.47 Å.¹⁷ It is of interest to note that the other set of the hydrogen-bond donor and acceptor is separated by a long O(4)⋯O(5) distance of 4.26 Å and is substantially in a geometry unfavorable to interact with each other. Thus, as suggested by its ¹H NMR spectrum, the two tetramic acid groups are shown to behave differently due to the chemical non-equivalence of the enolic functions. This symmetry breaking must be a result of deformation imposed by the p-nitrophenyl ring, which favors CH–π interactions with the two side phenyl groups to give a W-shaped conformation in the crystalline state,¹⁸ and brings the pseudo C₂-symmetric nature along the vertical axis of the model into its structural property.

Fig. 3 ORTEP diagram for the absolute configuration of 3g with 50% thermal ellipsoid probability.

In conclusion, the results of these extensive studies have shown that the proline-catalyzed methodology allows a one-pot and high-yielding access to a diverse range of the chiral diol analogues in enantiomerically pure forms via the tandem processes, and provide an entry of novel tetramic acid-derived diols that present unique structural features associated with the C₂- and pseudo C₂-symmetric motifs. In addition, we put particular emphasis on the synthetic utility of the current reaction, offering ready availability for the preparation of custom-designed chiral diols on a gram scale, which will provide new opportunities for the future development of effective and potentially versatile chiral ligands.

Acknowledgements

We thank Dr Toshiyasu Inuzuka (Gifu University) for performing mass spectral analyses. We also thank Ms Saki Matsuura for partial support of this research. This work was supported financially in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan and the JGC-S Scholarship Foundation.

Notes and references

1. K. C. Bhowmick and N. N. Joshi, Tetrahedron: Asymmetry, 2006, 17, 1901–1929.
2. For selected examples of enantioselective reactions using diethyl tartrate, see: (a) T. Katsuki and K. B. Sharpless, J. Am. Chem. Soc., 1980, 102, 5974–5976; (b) R. M. Hanson and K. B. Sharpless, J. Org. Chem., 1986, 51, 1922–1925; (c) Y. Gao, J. M. Klunder, R. M. Hanson, H. Masamune, S. Y. Ko and K. B. Sharpless, J. Am. Chem. Soc., 1987, 109, 5765–5780; (d) C. L. Elston, R. F. W. Jackson, S. J. F. MacDonald and P. J. Murray, Angew. Chem., Int. Ed. Engl., 1997, 36, 410–412 (Angew. Chem., 1997, 109, 379–381).
3. (a) C. Rosini, L. Franzini, D. Pini and P. Salvadori, Tetrahedron: Asymmetry, 1990, 1, 587–588; (b) P. N. Devine and T. Oh, J. Org. Chem., 1992, 57, 396–399; (c) K. R. K. Prasad and N. N. Joshi, Tetrahedron: Asymmetry, 1996, 7, 1957–1960; (d) K. Ishihara, D. Nakashima, Y. Hiraiwa and H. Yamamoto, J. Am. Chem. Soc., 2003, 125, 24–25.
4. (a) D. Seebach, A. K. Beck and A. Heckel, Angew. Chem., Int. Ed., 2001, 40, 92–138 (Angew. Chem., 2001, 113, 96–142); (b) H. Pellissier, Tetrahedron, 2008, 64, 10279–10317.
5. For selected reviews of BINOL, see: (a) J. M. Brunel, Chem. Rev., 2005, 105, 857–897; (b) M. Shibasaki and S. Matsunaga, in Privileged Chiral Ligands and Catalysts, ed. Q.-L. Zhou, WILEY-VCH, Weinheim, 2011, ch. 8, pp. 295–332.
6. A. K. Unni, N. Takenaka, H. Yamamoto and V. H. Rawal, J. Am. Chem. Soc., 2005, 127, 1336–1337.
7. (a) K. Matsuo and K. Tanaka, Yakugaku Zasshi, 1984, 104, 1004–1008; (b) M. J. Begley, N. G. Clemo and G. Pattenden, J. Chem. Soc., Perkin Trans. 1, 1985, 2393–2397; (c) K. Gorlitzer, J. Trittmacher and U. Bartke, Pharmazie, 2002, 57, 606–609.
8. For examples of acid-catalyzed condensation giving 1,3-diketones (ene-diones) 4, see: (a) H. G. Henning, A. Koppe and S. Briehn, Pharmazie, 1988, 43, 473–474; (b) M. N. S. Saudi, M. M. A. El Semary and G. El Sawaf, Pharmazie, 2002, 57, 519–522; (c) C. A. Holloway, C. J. Matthews, Y.-C. Jeong, M. G. Moloney, C. F. Roberts and M. Yaqoob, Chem. Biol. Drug Des., 2011, 78, 229–235.
9. (a) B. J. L. Royles, Chem. Rev., 1995, 95, 1981–2001; (b) R. Schobert and A. Schlenk, Bioorg. Med. Chem., 2008, 16, 4203–4221.
10. (a) S. V. Ley, S. C. Smith and P. R. Woodward, Tetrahedron, 1992, 48, 1145–1174; (b) T. Sengoku, Y. Nagae, Y. Ujihara, M. Takahashi and H. Yoda, J. Org. Chem., 2012, 77, 4391–4401.
11. As expected, racemic proline gave results very comparable to those with L-proline.
12. (a) Y. Oikawa, H. Hirasawa and O. Yonemitsu, Chem. Pharm. Bull., 1982, 30, 3092–3096; (b) B. List and C. Castello, Synlett, 2001, 1687–1689; (c) G. Cardillo, S. Fabbroni, L. Gentilucci, M. Gianotti and A. Tolomelli, Synth. Commun., 2003, 33, 1587–1594; (d) A. Lee, A. Michrowska, S. Sulzer-Mosse and B. List, Angew. Chem., Int. Ed., 2011, 50, 1707–1710 (Angew. Chem., 2011, 123, 1745–1748).
13. (a) N. Halland, T. Hansen and K. A. Jørgensen, Angew. Chem., Int. Ed., 2003, 42, 4955–4957 (Angew. Chem., 2003, 115, 5105–5107); (b) A.-B. Xia, D.-Q. Xu, S.-P. Luo, J.-R. Jiang, J. Tang, Y.-F. Wang and Z.-Y. Xu, Chem.–Eur. J., 2010, 16, 801–804; (c) N. C. Ganguly, S. Roy, P. Mondal and R. Saha, Tetrahedron Lett., 2012, 53, 7067–7071.
14. The product ratio, cis-/trans-isomers, was not determined because the observed ¹H NMR data for the isomeric mixture of the ene-dione are essentially indistinguishable from each other.
15. Crystal data for 3g: orthorhombic, space group P2₁2₁2₁, a = 10.4426(6) Å, b = 12.0149(7) Å, c = 21.1390(13) Å, V = 2652.2(3) ų, Z = 4, ρ = 1.359 M g m⁻³, μ(CuKα) = 0.809 mm⁻¹, T = 173 K; in the final least-squares refinement cycles on F², the model converged at R₁ = 0.1071 (I > 2ρ(I)), wR₂ = 0.2861, and GOF = 1.107 for 4427 reflections and 363 parameters (CCDC deposition number 963494).†.
16. The elemental analysis data of 3g (C, 65.90; H, 5.13; N, 8.12) are in good agreement with the calculated values for C₂₉H₂₅N₃O₆·CH₃OH (C, 66.29; H, 5.38; N, 7.73).
17. The O(3)⋯HO(6) bond length is estimated to be 1.93 Å.
18. S. Paliwai, S. Geib and C. S. Wilcox, J. Am. Chem. Soc., 1996, 116, 4497–4498.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, characterization data, ¹H and ¹³C NMR spectra for the compounds 3a–i, 3b′, 4h, 4j and 5. CCDC 963494. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4ra05405a

---