Control of Regioselectivity and Stereoselectivity in (4+3) Cycloadditions of Chiral Oxyallyls with Unsymmetrically Disubstituted Furans

Abstract

The regioselectivities and stereoselectivities of ZnCl₂-catalyzed (4+3) cycloadditions between chiral oxazolidinone-substituted oxyallyls and unsymmetrical disubstituted furans have been determined. The substitution pattern on the furan is found to provide a valuable tool for controlling the stereochemistry (endo-I or endo-II) of the 7-membered cycloadduct. While cycloadditions with monosubstituted furans usually favor endo-I products, from addition of the furan to the more crowded face of the oxyallyl, cycloadditions with 2,3- and 2,5-disubstituted furans instead favor the endo-II stereochemistry. Density functional theory calculations are performed to account for the selectivities. For monosubstituted furans, the crowded transition state leading to the endo-I cycloaduct is stabilized by an edge-to-face interaction between the furan and the oxazolidinone 4-Ph group, but this stabilization is overcome by steric clashing if the furan bears a 2-CO₂R group or is 2,3-disubstituted.

INTRODUCTION

Oxazolidinone-stabilized oxyallyls (3) are a fascinating class of reactive intermediates, generated by oxidation of allenamides (1).^1,2 Several valuable synthetic routes to 7-membered carbocycles rely on (4+3) cycloadditions between oxyallyl intermediates and dienes,^{3, 4} and we have found that the cycloadditions of oxyallyls 3 with monosubstituted furans (Schemes 1 and 2) display distinctive regioselectivities and stereoselectivities.^5,6 These reactions may be performed either under thermal conditions or with catalysis by ZnCl₂. The regioselectivities are summarized in Scheme 1, which shows the achiral oxyallyl 3a; cycloadditions of 3a with 2-substituted furans (G = Me or CO₂Me) favor the “syn” regiochemistry, while reactions with 3-substituted furans favor “anti” regiochemistry. The same regioselectivities are obtained with the chiral oxyallyl 3b (Scheme 2), but in this case the reactions also feature a novel mode of stereoinduction. Surprisingly, reactions of 3b with most monosubstituted furans were found to favor the endo-I stereochemistry, which entails addition of the furan to the more crowded face of 3b. Density functional theory calculations⁶ showed that the crowded transition state is favored because it contains a stabilizing edge-to-face aromatic interaction (CH–π) between furan and Ph. Of the monosubstituted furans studied, only those containing a 2-CO₂R group did not follow this behavior.

Scheme 1.

Regioselectivities of (4+3) Cycloadditions of Achiral Oxyallyl 3a with Monosubstituted Furans.^6b

Scheme 2.

Stereoselectivities of (4+3) Cycloadditions of Chiral Oxyallyl 3b with Monosubstituted Furans.^6a,c

In order to explore the generality of these stereochemical and regiochemical control elements, we have examined the (4+3) cycloadditions of 3b with more complex, unsymmetrical disubstituted furans. We report that these reactions often proceed with high levels of regio- and stereocontrol, even when the selectivities cannot be easily predicted from individual substituent effects. Complete control over stereoselectivity can be obtained through strategic choice of substituents.

RESULTS AND DISCUSSION

We investigated the (4+3) cycloadditions of oxyallyls 3 with a range of unsymmetrical disubstituted furans containing Me and CO₂Me substituents at different positions. The cycloadditions were performed under ZnCl₂-catalyzed conditions. Results are given in Table 1.

Table 1.

(4+3) Cycloadditions of Oxyallyls with Disubstituted Furans.^a

furan	syn-I	anti-I	syn-II	anti-II	yield (%)
	–	–		–	67b
	–	–		–	60b
				–	61c
	–	–		–	40b

^aIsolated yields are quoted. Isomer ratios are given in parentheses and were determined by ¹H and/or ¹³C NMR.

^bFrom reaction with allenamide 1b.

^cFrom reaction with allenamide 1c.

Our work commenced with the 2-CO₂Me,3-Me-substituted furan, 4. Based on the selectivities previously observed with monosubstituted furans (Schemes 1 and 2), it is not possible to predict the major product for 4; a 2-CO₂Me substituent favors syn-II but a 3-Me substituent favors anti-I. The cycloaddition of 3b with 4 was found to give exclusively the syn-II cycloadduct, 8, in 67% yield. The structure of 8 was assigned unambiguously by X-ray (Figure 1). The cycloaddition of 3b with the 2-Me,3-CO₂Me-substituted furan 5 also gave exclusively a syn-II adduct (9). Confirmation of the structure of 9 was obtained following reduction/esterification, which afforded a rearranged derivative (12) whose X-ray structure was determined (Figure 1).⁷ The regioselectivities observed for both 4 and 5 match those previously obtained^6b with the achiral oxyallyl 3a, which displayed syn selectivities of ≥95:5.

Figure 1.

X-Ray structures of cycloadduct 8, ester 12 (derived from 9), and cycloadducts 10a and 10b.

We next examined the 2-Me,3-Me-substituted furan, 6. In order to deal with competing oxidation of this more electron-rich furan, we had to switch to the more reactive allenamide, 1c,⁸ which contains Close's chiral imidazolidinone auxiliary.⁹ Previous experiments⁵ indicated that Close's imidazolidinone controls stereoselectivity in the same way as 4-Ph-oxazolidinone. Cycloaddition of 3c with 6 afforded mainly the syn-II product, 10c, but also gave two other cycloadducts: syn-I (10a) and anti-I (10b). X-ray structures were determined for the two minor products (Figure 1), while the major product was assigned unambiguously by NOESY and COSY (see Supporting Information).

In 2,3-disubstituted furans, the substituents have opposite regio-directing influences (Scheme 1) and the selectivities obtained with 4–6 indicate that it is the 2-substituent that dictates the outcome. In order to discern the relative directing power of Me and CO₂Me groups at the 2-position, we examined the reaction of 3b with 7. The only cycloadduct detected was 11, where CO₂Me is syn to nitrogen. Its structure was confirmed by NOESY and COSY.

The regioselectivities and stereoselectivities observed here with disubstituted furans are quite distinct from those reported previously for monosubstituted furans.⁶ We performed density functional theory calculations to explore the selectivities. Transition states for ZnCl₂-catalyzed cycloadditions of the oxyallyls 3b and 3c with furans 4–7 were computed at the B3LYP/6-31G(d) level (with LANL2DZ on Zn), as we have previously done in our studies of related oxyallyl (4+3) cycloadditions. The transition states for the reaction of 3b with 4 are representative and are shown in Figure 2. Regardless of the regiochemistry, the transition states each show the same sense of asynchronicity, with more advanced bonding at the unsubstituted terminus of the oxyallyl. The HOMO coefficient of the oxyallyl is larger at this terminus.

Figure 2.

Transition structures for ZnCl₂-catalyzed cycloadditions of oxyallyl 3b with furan 4, optimized with B3LYP. Below each structure are given the activation energies obtained from M06-2X single-point calculations. Distances in Å, ΔH^‡ and ΔG^‡ in kcal/mol.

We previously found that B3LYP activation energies provide good predictions of regioselectivity and stereoselectivity for oxyallyl (4+3) cycloadditions. In order to provide a better treatment of dispersion energies in the cycloadditions of 4–7, which are likely to vary significantly between stereoisomeric transition states I and II, we calculated the activation barriers from single-point energies at the M06-2X/6-311+G(d,p) level. Solvent effects were modeled by means of CPCM calculations. The M06-2X activation energies in the gas phase and in dichloromethane are given in Table 2.

Table 2.

Calculated Activation Energies for ZnCl₂-Catalyzed Cycloadditions of 3b with Disubstituted Furans.^a

Reactants	ΔH‡ (ΔG‡) [ΔG‡soln]
	syn-I	anti-I	syn-II	anti-II
4 + 3b·ZnCl2	–13.2 (5.2) [11.5]	–12.7 (4.9) [10.1]	–16.1 (2.2) [7.2]	–9.9 (6.8) [11.0]
5 + 3b·ZnCl2	–13.2 (4.3) [12.5]	–15.5 (3.6) [11.9]	–14.6 (2.6) [7.3]	–13.0 (5.6) [14.1]
6 + 3b·ZnCl2	–12.8 (3.9) [8.4]	–15.0 (2.3) [7.4]	–12.1 (3.6) [8.0]	–12.4 (3.9) [8.0]
6 + 3c·ZnCl2	–7.0 (9.7) [14.3]	–8.3 (9.0) [14.4]	–6.1 (9.6) [14.2]	–6.0 (10.4) [14.9]
7 + 3b·ZnCl2	–9.6 (8.4) [16.1]	–11.5 (4.8) [9.9]	–14.4 (3.8) [11.0]	–8.3 (7.0) [11.4]

^aM06-2X/6-311+G(d,p)//B3LYP/6-31G(d)–LANL2DZ. Solution-phase data incorporate CPCM solvation energies in CH₂Cl₂, computed at the M06-2X/6-31G(d)–LANL2DZ level. ΔH^† and ΔG^‡ in kcal/mol at 298.15 K.

The calculated values of ΔG^‡ in the gas phase and in solution predict a strong preference for the observed product (syn-II) for both furans 4 and 5. The syn-II TS from 4 is favored by ≥2.9 kcal/mol in dichloromethane, and the syn-II TS from 5 is favored by ≥4.6 kcal/mol. Lower selectivity is predicted for furan 6. Indeed, the cycloaddition of 6 with 3b is calculated to favor the anti-I TS by 0.6 kcal/mol. However, the cycloaddition of the same furan with 3c, which was studied experimentally, is correctly predicted to favor syn-II. The syn-II TS is preferred by 0.1–0.2 kcal/mol over syn-I and anti-I. Experimentally, a mixture of products was obtained containing the latter two isomers as minor products. The agreement between experiment and theory is not as good for 7, where the calculations predict a 1.1 kcal/mol preference for anti-I in dichloromethane. The selectivity in the gas phase does correctly predict syn-II, however. Experimentally, the syn-II cycloadduct was obtained in 40% yield.

Activation energies were also computed with a second dispersion-inclusive functional, B3LYP-D3 (Supporting Information). The predicted selectivities mirror those from M06-2X. B3LYP alone also correctly predicts the major products for furans 4–6 (and from 7 in the gas phase), but fails to predict the presence of minor products in the cycloaddition involving 6.

Our previous calculations showed that a 2-substituent on the furan favors syn regiochemistry because the shorter forming bond in the transition state is located at the less-hindered carbon (C-5) of the furan.⁶ When the 2-substituent is electron-withdrawing, this steric effect is reinforced by a strong oxyallyl–furan interaction energy, related to the large furan LUMO coefficient at C-5. For 3-substituted furans, by contrast, the anti regiochemistry is favored because its transition state avoids steric clashing between the 3-substituent and the oxazolidinone.

The syn selectivities obtained experimentally from furans 4–7 (Table 1) indicate that the regio-directing influence of a 2-substituent outweighs that of a 3-substituent, regardless of whether the substituent is Me or CO₂Me. A 2-CO₂Me group exerts a stronger regio-directing effect than a 2-Me group (cf. 7), because the oxyallyl–furan interaction in the TS is stronger when CO₂Me is syn to nitrogen.

The stereoselectivities observed with furans 4–7 cannot be predicted from those of the corresponding monosubstituted furans. Most monosubstituted furans favor diastereomer I, due to the stabilizing edge-to-face aryl–aryl interaction in the crowded (I) transition state (Scheme 2).^6a,c Only for 2-CO₂Me-furan was II favored; in that case, the (syn-)I TS is destabilized by electrostatic repulsion between the CO₂Me group and the Ph π-cloud.

A similar type of repulsive electrostatic effect influences the stereoselectivities for 4 and 7, favoring II. More generally, however, the stereoselectivities for all three 2,3-disubstituted furans (4–6) are dictated by the furan 3-substituent. This group eliminates the possibility of an edge-to-face CH–π interaction in the crowded syn-I TS, and replaces it with a repulsive electrostatic interaction (3-CO₂Me) or a weaker CH–π interaction (3-Me). The result is a preference for II. Only for 6, where no CO₂Me group is present, are isomers of I found as minor products.

CONCLUSION

The investigations detailed herein provide a valuable new approach to controlling the stereoselectivities of (4+3) cycloadditions. The different classes of substituent effects operating for monosubstituted and disubstituted furans are outlined in Scheme 3, and can be summarized as follows: cycloaddition of oxyallyl 3b or 3c with a 2-substituted furan (including a disubstituted furan) favors the syn-I cycloadduct, unless (a) the 2-substituent is CO₂R or (b) there is also a substituent present at furan C-3. In the latter two cases, syn-II is preferred instead. These principles enable cycloadditions of oxyallyls 3 with complex furans to be achieved with high levels of regioselectivity and stereoselectivity.

Scheme 3.

Summary of Substituent-Controlled Regioselectivity and Stereoselectivity in (4+3) Cycloadditions of Oxyallyl 3b with Furans

EXPERIMENTAL SECTION

General Procedure for Allenamide (4 + 3) Cycloadditions

To a solution of a respective allenamide in CH₂Cl₂ (0.1 M) was added 3-9 equiv of the appropriate furan and 4Å powdered molecular sieves [0.5 g]. The reaction solution was cooled to –78 °C, and 2.0 equiv of ZnCl₂ (1.0 M in Et₂O) was added. 4.0 - 6.0 equiv of dimethyldioxirane [DMDO] in acetone was then added as a chilled solution [at –78 °C] via a syringe pump over 3-4 h. The syringe pump was cooled using dry ice during the entire addition time. After the addition, the reaction mixture was stirred for an additional 14 h before it was quenched with sat aq NaHCO₃, filtered through Celite™, concentrated in vacuo, and extracted with CH₂Cl₂ [4 × 20 mL]. The combined organic layers were dried over Na₂SO₄ and concentrated in vacuo. The crude residue was purified via silica gel column chromatography [gradient eluent: 10% to 75% ethyl acetate in hexane].

8: Reaction Scale: 0.50 mmol of allenamide 1b; Yield: 67%; mass: 119.6 mg. R_f = 0.13 (50% EtOAc in hexane); [α]_D²³ –169.6 ° [c 4.6, CH₂Cl₂]; white solid; mp = 119 – 120 °C; ¹H NMR (400 MHz, CDCl₃) δ 2.04 (s, 3H), 2.54 (d, 1H, J = 16.0 Hz), 2.83 (dd, 1H, J = 5.4, 16.0 Hz), 3.29 (s, 3H), 3.91 (s, 1H), 4.20 (t, 1H, J = 9.0 Hz), 4.62 (t, 1H, J = 9.0 Hz), 4.79 (t, 1H, J = 9.0 Hz), 4.90 (dt, 1H, J = 5.2, 1.6 Hz), 7.00 (d, 1H, J = 1.6 Hz), 7.31 – 7.45 (m, 5H); ¹³C NMR (100 MHz, CDCl₃) δ 15.3, 44.3, 52.6, 65.9, 68.5, 70.5, 76.8, 88.3, 128.7, 129.3, 129.7, 130.2, 136.6, 142.9, 158.5, 166.9, 200.9; IR (thin film) cm^-1 3280w, 2911w, 1766s, 1437w; mass spectrum (APCI): m/e (% relative intensity) 358.1 (100) (M + H)⁺; HRMS (MALDI-TOF): m/e calcd for C₁₉H₁₉NO₆Na⁺ (M+Na⁺) 380.1105, found 380.1097.

9: Reaction Scale: 1.00 mmol of allenamide 1b; Yield: 60%; mass: 214.2 mg. R_f = 0.25 (50% EtOAc in hexane); [α]_D²³ –74.5 ° [c 4.0, CH₂Cl₂]; ¹H NMR (500 MHz, CDCl₃) δ 1.30 (s, 3H), 2.60 (d, 1H, J = 17.5 Hz), 2.88 (dd, 1H, J = 6.0, 17.5 Hz), 3.45 (s, 1H), 3.80 (s, 3H), 4.36 (t, 1H, J = 9.0 Hz), 4.71 (t, 1H, J = 9.0 Hz), 4.83 (t, 1H, J = 9.0 Hz), 4.99 (dd, 1H, J = 6.0, 2.0 Hz), 7.08 (d, 1H, J = 0.25 Hz), 7.48 – 7.52 (m, 5H); ¹³C NMR (125 MHz, CDCl₃) δ 15.4, 44.3, 52.6, 66.0, 68.6, 70.6, 76.9, 88.4, 126.4, 128.8, 129.4, 129.8, 130.3, 143.0, 158.6, 167.0, 201.0; IR (thin film) cm^-1 2954s, 2930s, 1764s, 1726s; mass spectrum (APCI): m/e (% relative intensity) 358.2 (100) (M + H)⁺; HRMS (MALDI-TOF): m/e calcd for C₁₉H₁₉NO₆Na⁺ (M+Na⁺) 380.1105, found 380.1095.

10a: Reaction Scale: 0.44 mmol of allenamide 1c; Yield: 9%; mass: 13.5 mg. R_f = 0.22 (50% EtOAc in hexane); [α]_D²³ +14.9 ° [c 0.75, CH₂Cl₂]; mp = 162 – 163 °C; ¹H NMR (400 MHz, CDCl₃) δ 0.79 (d, 1H, J = 6.8 Hz), 1.46 (s, 3H), 1. 81 (s, 3H), 2.49 (d, 1H, J = 17.8 Hz), 2.60 (dd, 1H, J = 5.8, 17.8 Hz), 2.69 (s, 3H), 3.29 (s, 1H), 3.82 (dq, 1H, J = 6.8, 9.2 Hz), 4.70 (d, 1H, J = 9.2 Hz), 4.75 (d, 1H, J = 5.8 Hz), 5.91 (s, 1H), 7.13 (br, 1H), 7.32 (br, 4H); ¹³C NMR (100 MHz, CDCl₃) δ 14.0, 15.5, 21.4, 29.3, 43.8, 56.6, 65.2, 69.3, 75.9, 86.8, 128.5, 128.8, 128.9, 129.5, 135.3, 144.8, 159.3, 200.9; IR (thin film) cm^-1 2929m, 1725s, 1687s, 1432m, 1263m, 1161m, 736m, 694s; mass spectrum (APCI): m/e (% relative intensity) 341.2 (60) (M + H)⁺; HRMS (MALDI-TOF): m/e calcd for C₂₀H₂₄N₂O₃Na⁺ (M+Na⁺) 363.1679, found 363.1680.

10b: Reaction Scale: 0.44 mmol of allenamide 1c; Yield: 21%; mass: 31.4 mg. R_f = 0.17 (50% EtOAc in hexane); [α]_D²³ –184.3 ° [c 0.79, CH₂Cl₂]; white solid; mp = 194 – 195 °C; ¹H NMR (400 MHz, CDCl₃) δ 0.71 (d, 1H, J = 6.8 Hz), 1.34 (s, 3H), 1.59 (s, 3H), 2.44 (d, 1H, J = 15.6 Hz), 2.57 (d, 1H, J = 15.6 Hz), 2.73 (s, 3H), 3.96 (dq, 1H, J = 6.8, 8.8 Hz), 4.46 (d, 1H, J = 8.8 Hz), 4.49 (d, 1H, 4.2 Hz), 4.66 (s, 1H), 5.10 (d, 1H, J = 4.2 Hz), 7.03 (br, 1H), 7.31-7.33 (m, 1H); ¹³C NMR (100 MHz, CDCl₃) δ 12.1, 15.3, 21.6, 29.3, 50.8, 57.2, 59.7, 64.5, 79.2, 85.9, 125.6, 127.4, 128.4, 129.2, 138.8, 146.3, 162.6, 203.8; IR (thin film) cm^-1 2941m, 1728s, 1686s, 1431m, 1255m, 1196s, 1025m, 855s, 735s, 700s; mass spectrum (APCI): m/e (% relative intensity) 339.2 (45) (M - H)^-; HRMS (MALDI-TOF): m/e calcd for C₂₀H₂₄N₂O₃Na⁺ (M+Na⁺) 363.1679, found 363.1678.

10c: Reaction Scale: 0.44 mmol of allenamide 1c; Yield: 31%; mass: 46.4 mg. R_f = 0.23 (50% EtOAc in hexane); [α]_D²³ -105.7 ° [c 0.83, CH₂Cl₂]; colorless oil; ¹H NMR (400 MHz, CDCl₃) δ 0.85 (d, 1H, J = 6.8 Hz), 1.13 (s, 3H), 1. 89 (s, 3H), 2.47 (d, 1H, J = 17.2 Hz), 2.71 (dd, 1H, J = 5.8, 17.2 Hz), 2.79 (s, 3H), 3.43 (s, 1H), 3.92 (dq, 1H, J = 6.8, 9.2 Hz), 4.61 (d, 1H, J = 9.2 Hz), 4.74 (d, 1H, J = 5.8 Hz), 5.86 (s, 1H), 7.27-7.34 (m, 5H); ¹³C NMR (100 MHz, CDCl₃) δ 14.4, 15.1, 21.4, 29.1, 44.7, 56.3, 68.1, 74.9, 75.9, 87.7, 121.1, 128.5, 128.8, 129.5, 137.0, 146.9, 161.7, 204.2; IR (thin film) cm^-1 2933m, 1703s, 1431, 1400s, 1262m, 1159m, 762m; mass spectrum (APCI): m/e (% relative intensity) 341.2 (60) (M + H)⁺; HRMS (MALDI-TOF): m/e calcd for C₂₀H₂₄N₂O₃Na⁺ (M+Na⁺) 363.1679, found 363.1674.

11: Reaction Scale: 0.35 mmol of allenamide 1b; Yield: 40%; mass: 50.0 mg. R_f = 0.22 (50% EtOAc in hexane); [α]_D²³ –31.4 ° [c 6.4, CH₂Cl₂]; ¹H NMR (500 MHz, CDCl₃) δ 1.43 (s, 3H), 2.59 (d, 1H, J = 16.0 Hz), 2.68 (dd, 1H, J = 5.2, 16.0 Hz), 3.65 (s, 3H), 3.93 (s, 1H), 4.18 (t, 1H, J = 9.0 Hz), 4.65 (t, 1H, J = 9.0 Hz), 4.82 (t, 1H, J = 9.0 Hz), 6.07 (d, 1H, J = 6.0 Hz), 6.66 (d, 1H, J = 6.0 Hz), 7.28 – 7.41 (m, 5H); ¹³C NMR (100 MHz, CDCl₃) δ 23.2, 51.3, 64.4, 67.2, 70.5, 72.8, 86.1, 89.4, 126.4, 128.4, 129.9, 133.4, 135.6, 136.5, 158.1, 167.5, 199.8; IR (thin film) cm^-1 3520w, 2957w, 2927w, 1759s, 1726s; mass spectrum (APCI): m/e (% relative intensity) 358.2 (35) (M + H)⁺; HRMS (MALDI-TOF): m/e calcd for C₁₉H₁₉NO₆Na⁺ (M+Na⁺) 380.1105, found 380.1090.

Synthesis of Ester 12 from Cycloadduct 9

To a solution of cycloadduct 9 (53.6 mg, 0.150 mmol) in anhyd MeOH (5 mL) was added NaBH₄ (34.2 mg, 0.904 mol) portion-wise within 1 h at 0 °C. The reaction mixture was further stirred at room temperature for 6 h until the consumption of the starting material. Sat aq NaCl (10 mL) was added to quench the reaction. After removal of the volatile solvent under reduced pressure, the aqueous residue was extracted with CH₂Cl₂ (3 × 10 mL). The combined organic layer was dried over Na₂SO₄ and concentrated in vacuo, and the crude residue was directly used for the next step.

The above crude residue was taken up in CH₂Cl₂ (5 mL), and to this solution were added DMAP (1.1 mg, 0.009 mmol) and Et₃N (0.06 mL, 0.431 mmol) with vigorous stirring under ice-water cooling bath. After which, 2,4,6-trichlorobenzoyl chloride (0.07 mL, 0.450 mmol) was then added in one portion. The reaction mixture was stirred at rt for 30 h, and was then quenched with cold H₂O (10 mL), and extracted with CH₂Cl₂ [4 × 20 mL]. The combined organic layers were dried over Na₂SO₄ and concentrated in vacuo. The crude residue was purified via silica gel column chromatography [gradient eluent: 10% to 30% EtOAc in hexane] to give the pure ester 12 (35.3 mg, 0.062 mmol, 42% over two steps) as a white solid.

12: R_f = 0.21 (30% EtOAc in hexane); [α]_D²³ + 29.9 ° [c 1.05, CH₂Cl₂]; white solid; mp = 176 – 177 °C; ¹H NMR (400 MHz, CDCl₃) δ 1.66 (s, 3H), 2.01 – 2.14 (m, 2H), 2.29 (dt, 1H, J = 8.4, 12.0 Hz), 2.56 (ddd, 1H, J = 2.4, 8.8, 12.0 Hz), 2.79 (dd, 1H, J = 8.8, 11.6 Hz), 3.35 (s, 3H), 3.37 (d, 1H, J = 7.2 Hz), 4.35 (br, 1H); 4.39 (t, 1H, J = 6.4 Hz); 4.84 – 4.89 (m, 2H); 4.95 (q, 1H, J = 7.2 Hz), 7.54 (d, 1H, J = 1.4 Hz), 7.52 (d, 1H, J = 1.4 Hz), 7.35 – 7.41 (m, 3H), 7.31 (s, 2H); ¹³C NMR (100 MHz, CDCl₃) δ 26.8, 31.2, 33.7, 51.8, 52.3, 60.9, 63.7, 64.4, 72.0, 74.1, 82.2, 128.3, 129.1, 129.2, 131.7, 132.8, 135.2, 136.7, 159.3, 163.6, 169.8 (one carbon peak was missing due to overlapping); IR (thin film) cm^-1 2955m, 1755s, 1579m, 1373m, 1265s, 1209m, 1117m, 734m; mass spectrum (APCI): m/e (% relative intensity) 574.0 (7) (M + 6 + H)⁺; 572.1 (33) (M + 4 + H)⁺; 569.2 (30) (M + 2)⁺, 568.1 (96) (M + H)⁺; 567.1 (90) (M)⁺, 419.4 (9), 344.2 (22), 329.0 (85), 328.1 (21), 327.0 (100); HRMS (MALDI-TOF): m/e calcd for C₂₆H₂₄Cl₃NO₇Na⁺ (M+Na⁺) 590.0516, found 590.0511.

THEORETICAL CALCULATIONS

Density functional theory calculations were performed in Gaussian 09.¹⁰ Geometries were optimized in the gas phase using the B3LYP¹¹ functional, with a mixed basis set consisting of LANL2DZ on Zn and 6-31G(d) on all other atoms. The lowest-energy conformer of each species was identified through conformational searching. The B3LYP vibrational frequencies were used to characterize species as minima or transition states, and to obtain scaled¹² zero-point energies and thermal contributions to enthalpy and entropy. Single-point energies were then computed at the M06-2X/6-311+G(d,p) level,¹³ and used in conjunction with the B3LYP thermochemical corrections to obtain gas-phase activation enthalpies and free energies. Free energies of activation in dichloromethane were calculated by incorporating CPCM¹⁴ solvation energies computed at the M06-2X/6-31G(d)–LANL2DZ level (UAKS radii). A standard state of 1 mol/L was used. Activation energies were also computed from single-point energy calculations at the B3LYP-D3¹⁵ level, with zero-damping. The activation energies predicted by B3LYP and B3LYP-D3 for cycloadditions involving 3b are provided in the Supporting Information, along with M06-2X activation energies for cycloaditions of 3c with 4–7.