Formation Mechanism of Triketo Cage Compounds from Reaction of Phencyclone with Benzoquinones: Cascade Reaction of Intermolecular [4+2]π and Intramolecular [2+2]π Cycloadditions

Abstract

The [4+2]π cycloadduct (3aa) of phencyclone (1a) and benzoquinone (2a) readily transformed into the corresponding [2+2]π cycloadduct (4aa) during purification. The structure of the triketo cage compound was determined by single crystal X-ray analysis. The structure–reactivity relationships are discussed based on the PM6-calculated energy profiles for the whole photocycloaddition of the [4+2]π cycloadducts of some cyclopentadienones and benzoquinones.

Cyclopentadienones are reactive and versatile dienes, whose cascade reactions with various dienophiles such as non-conjugated dienes,¹⁾ conjugated polyenes,^2,3) allyl alcohols,⁴⁾ 2-alkynyl alcohols,⁵⁾ 2-alkynyl amines,⁶⁾ and but-3-yn-2-ones,⁷⁾ have been extensively studied. A previous report showed that the endo[4+2]π cycloadduct of 2,5-bis(methoxycarbonyl)-3,4-diphenylcyclopentadienone (1b) and benzoquinone (2a) in benzene underwent photocyclization to yield an oxetane derivative (5).^8,9) On the other hand, irradiation of UV light on the powdered crystals of the [4+2]π cycloadduct gave the intramolecular [2+2]π cycloadduct (triketo cage compound, 4ba) (Chart 1).

Chart 1

Our recent studies on the [4+2]π cycloadducts of phencyclone (1a) as clathrate host^10–13) showed that the [4+2]π cycloadduct of 1a and 2a readily transformed into the corresponding [2+2]π cycloadduct during purification, despite being thermally forbidden.¹⁴⁾ Wooi and White reported that the same reaction occurred in dull light.¹⁵⁾ However, the mechanistic details of the reaction is still unclear. Therefore, we explored the reason for the difference in the reactivity of several cyclopentadienone-[4+2]π cycloadducts to form the [2+2]π cycloadduct.

In this study, we determine the structure of the cage compound by X-ray crystallographic analysis and discuss a reaction mechanism based on molecular orbital (MO) calculations.

Results and Discussion

The [4+2]π cycloadducts of 1a and quinones (2) were prepared according to a previously reported method¹⁶⁾ (Chart 2, Table 1).

Chart 2

Table 1. [4+2]π Cycloaddition of Phencyclone (1a) with Quinones (2) in Toluene

Diene	Quinones	Temp. (°C)	Time (h)	Yield of 3
1a	2a	80	24	99 (3aa)
	2b	60	48	73 (3ab)
	2c	60	48	84 (3ac)
	2d	60	68	—a)
	2e	105	72	95 (3ae)

a) 9,10-Dibenzoylphenanthrene was obtained (yield 67%).

A mixture of 1a and 2a in toluene was heated at 80°C for 24 h to give a 1 : 1 endo[4+2]π cycloadduct (endo-3aa, 99%). Similar reactions between 1a and compounds 2b, 2c and 2e afforded the corresponding endo[4+2]π cycloadducts (3ab, 3ac, 3ae) in moderate yield. In contrast, the reaction between 1a and 2d yielded 9,10-dibenzoylphenanthrene as the predominant product, possibly derived from the [4+2]π cycloadduct of 1a and molecular oxygen.¹⁷⁾

Experiments to form an inclusion complex of the cycloadduct 3aa were carried out with the following solvents: o-xylene, m-xylene, p-xylene, and acetone. Interestingly, the above treatments led to the formation of a compound 4aa, instead of the expected inclusion of a guest solvent molecule. The IR and ¹³C-NMR spectra of 4aa indicated the absence of enedione moiety. The formation of the 2 : 1 cycloadduct of 3aa via retro Diels‒Alder reaction was initially thought to occur on heating. The structure of the compound was determined by X-ray diffraction analysis. The single crystal suitable for the analysis was obtained from CHCl₃ by slow-evaporation. As shown in Fig. 1, the newly formed compound (4aa) was the intramolecular [2+2]π cycloadduct: a triketo cage compound containing a cyclobutane ring.¹⁸⁾ This reaction is thermally inhibited and results in the loss of resonance energy of phenanthrene moiety. The cage compound had highly strained σ-bonds between the phenanthrene ring and the α carbon attached to the bridged carbonyl [C(4)–C(5) and C(7)–C(11) bonds]. The bond lengths were significantly elongated to 1.608(5) Å and 1.609(6) Å. Density function theory (DFT) calculations¹⁹⁾ at B3LYP/6-31G(d) level also showed bond elongation (1.620 Å) for the bond in question, indicating that there existed steric repulsions between the face-to-face oriented aromatic rings.²⁰⁾ The biphenyl moiety of dihydrophenanthrene is oriented at a dihedral angle [C(16)–C(17)–C(18)–C(19)] of −9.1(7)°, indicating that the ring is considerably twisted. The structure was found to be non-planar, and the interplanar angle between the two phenyl rings was 18.5°.

Fig. 1. Crystal Structure of 4aa·CHCl₃ Determined by X-Ray Diffraction

Protons are omitted for clarity.

Close inspection of the ¹H-NMR spectrum of the crude 3aa revealed the presence of small amounts of 4aa, indicating that the [2+2]π cycloaddition took place readily even under very mild reaction conditions and in the absence of artificial light. ¹H-NMR was used to monitor the reaction, and the concentration of 4aa was found to increase over time. In control experiments where the test tube was covered with aluminum foil, the cage compound was not formed even on heating. This observation suggested that the reaction was induced by natural light. In fact, the irradiation of 3aa with a fluorescent lamp for indoor use (60 W) in CDCl₃ through a Pyrex filter afforded 4aa. To maintain the irradiation conditions of light, the cycloaddition reaction of 3 was examined using a transparency overhead projector (160 W halogen lamp). The reactions of 3aa–ac were completed within 4 h and resulted in the formation of the corresponding triketo cage compounds (4ab–ac) (Table 2). The cycloadduct 3ae derived from the reaction of 1a and naphthoquinone (2e) was inert to irradiation by visible light, presumably owing to the aromaticity of the benzene ring.

Table 2. Intramolecular [2+2]π Photocycloaddition of 3 in CDCl₃

Cycloadduct, 3	Time (h)	Yielda) of 4
3aa	0.5	32 (4aa)
	1.5	76 (4aa)
	3	100 (4aa)
3ab	3	100 (4ab)
3ac	4	100 (4ac)
3ae	4	—
3ba	4	—

a) Determined based on integrated intensity of the methine signals in ¹H-NMR.

In stark contrast to 3aa, the [4+2]π cycloadduct derived from compounds 1b and 2a (3ba) was photochemically unreactive under the reaction conditions used.^8,9) To understand the difference in the reactivity of 3aa and 3ba, we performed PM6²¹⁾ semi-empirical MO calculations on the energy profiles for photocycloaddition of the [4+2]π adducts. It is known that the photocycloaddition of enones to alkenes proceed via the triplet excited state of enones.²²⁾ The photoexcitation of an enone followed by intersystem crossing, results in the lowest energy triplet state, which participates in the [2+2]π cycloaddition reaction. Based on theoretical studies which have looked at the [2+2]π photocycloaddition of acrolein to ethylene,²³⁾ and intramolecular [2+2]π cycloaddition of α,β-unsaturated furanones,²⁴⁾ we postulate the initial step in the reaction to be the interaction between a triplet state of enedione moiety and a ground-state of alkene moiety. The possible pathway of the intramolecular photocycloaddition of enedione moiety to the alkene moiety is shown in Fig. 2.

Fig. 2. Potential Energy Profile for the Intramolecular Photocycloaddition of 3

We postulate that the photocycloaddition proceeds via a singlet biradical intermediate.²²⁾ The reaction starts with the attack of the excited triplet state of an enedione (³3) on a ground-state alkene. The first step corresponds to bond formation between the triplet state of enedione and the double bond of 3, and gives rise to a transition state (TS1). The biradical intermediate formed from TS1 either closes to form the triketo cage compound (4) via a second transition state (TS2) or goes back to the starting material (3) via a different transition statre (TS3). The PM6-calculated potential energies and TS1 geometries are summarized in Table 3 and Fig. 3, respectively. The energy barrier associated with the formation of TS2 was much smaller than the barrier for the biradical reverting back to starting material (3) via TS3. Therefore, the rate-determining step is formation of biradical intermediate through TS1. The TS barrier of 3aa was considerably smaller (ca. 3.8 kcal/mol) than that of 3ba.²⁵⁾ The difference in the reaction barriers of 3aa and 3ba can be explained by an attractive CH/O type interaction^26–31) between the substituents in the transition state. The distances between the hydrogen of the phenyl group and the oxygen atom on the enone group of TS1_3aa are 2.242 Å and 2.342 Å, respectively, indicating the presence of TS stabilizing by CH/O interactions.

Table 3. PM6-Calculated Potential Energies (ΔH_f, kcal/mol)

Reaction	3	33	TS1	Biradical	TS2	TS3	4
3aa→4aa	50.2	95.7	111.4	86.8	86.9	93.1	40.6
		(0.0)a)	(15.7)a)
				[0.0]b)	[0.1]b)	[6.3]b)
3ba→4ba	−154.4	−111.9	−92.4	−122.4	—c)	−110.5	−167.9
		(0.0)a)	(19.5)a)
				[0.0]b)		[11.9]b)

a) Energy barrier relative to triplet-state of 3 (³3). b) Energy barrier relative to biradical. c) Stationary point was not obtained.

Fig. 3. PM6-Calculated Structures for First Transition State (TS1)

Applying frontier molecular orbitals (FMO)^32,33) theory to the [4+2]π cycloadducts provided a clue for understanding the difference in the reactivity in the first step of intramolecular cyclization. FMO calculations were performed using the analogues of each reaction site: stilbene (A), phenanthrene (B), and 1,4-enedione (C). As shown in Fig. 4, the energy difference between lowest singly occupied molecular orbital (LSOMO)_enedione and highest occupied molecular orbital (HOMO)_alkene was found to be smaller than that of the corresponding highest singly occupied molecular orbital (HSOMO)_enedione and lowest unoccupied molecular orbital (LUMO)_alkene, thus making the former an effective frontier orbital interaction. Compared to the stilbene analogue (A), the HOMO of the phenanthrene analogue (B) was raised, bringing it closer in energy to the LSOMO of enedione, resulting in favourable the orbital interaction for the cycloaddition.

Fig. 4. PM6-Calculated FMO Energy Levels for Enedione Triplet and Alkenes

In addition to this, the degree of orbital overlapping between the reaction sites contributes to the reactivity. The PM6-optimized triplet-state structure of 3aa showed that the cyclohexene-1,4-dione moiety was bent considerably inwards (below the phenanthrene ring), in comparison to the structure of ³3ba³⁴⁾ (Fig. 5), where CH/O type interactions between the carbonyl oxygen and 2′-hydrogen of the phenyl group might be present. On the other hand, the two phenyl groups (stilbene moiety) in ³3ba were no longer planar and interfere an approach of an enedione moiety. As a result, one of the carbon atoms of stilbene is closer to the oxygen atom of the enone (C···O, 3.136 Å) than the carbon atom of the alkene (C···C, 4.085 Å). Thus, the Paterno–Büchi reaction of 3ba, which yields an oxetane, is favourable. The distance between interacting carbon atoms of ³3aa is 0.95 Å shorter than that of ³3ba, which gave the [2+2]π cycloadduct only upon UV irradiation. In such the geometry, the orbital overlap of ³3aa is large enough to undergo stepwise [2+2]π cycloaddition.

Fig. 5. PM6-Calculated Triplet-State Structures of 3aa and 3ba

The UV-Vis spectrum of 3aa suggests the through-space interaction between the two π-systems. The absorption spectra in CHCl₃ showed a weak and broad absorption band (ca. 380–450 nm) without a clearly defined absorption maximum (Fig. 6). On the other hand, 4aa, 2a, and 1a-norbornene cycloadduct¹⁴⁾ of having no enedione moiety did not show any absorption band at the region. These results suggest that the observed broad absorption in 3aa is due to intramolecular charge-transfer band arising from the interaction between phenanthrene ring and enedione moiety.³⁵⁾

Fig. 6. Absorption Spectra of 3aa, 4aa, 2a, and 1a-Norbornene Cycloadduct in CHCl₃ (2.0 mmol/L)

Table 4. X-Ray and Calculated Bond Lengths (Å) of 4aa

Bond	X-Ray	B3LYP/6-31G(d)
C1–C2 (C8–C9)	1.517 (1.512)	1.528
C1–C5 (C8–C7)	1.547 (1.559)	1.556
C1–C8	1.592	1.599
C2–C3 (C9–C10)	1.508 (1.502)	1.518
C3–C4 (C10–C11)	1.579 (1.564)	1.570
C3–C10	1.565	1.585
C4–C5 (C11–C7)	1.608 (1.609)	1.620
C4–C11	1.563	1.576
C5–C6 (C7–C6)	1.516 (1.539)	1.538

In conclusion, the [4+2]π cycloadduct of phencyclone (1a) and benzoquinone (2a) readily transformed into the corresponding [2+2]π cycloadduct. The intramolecular cycloaddition reactivity depends on the energy of the first transition state (TS1), and sterically favourable orbital interactions between the alkenes in the triplet-state.

Experimental

Melting points were uncorrected. The IR spectra were taken with a Hitachi 270-30 spectrophotometer. ¹H- and ¹³C-NMR spectra were taken with JEOL JNM-A 500 (500 MHz) spectrometers for ca.10% solution with tetramethylsilane (TMS) as an internal standard; chemical shifts are expressed as δ values and the coupling constants (J) are expressed in Hz. Mass spectra were obtained with JEOL JMS-T100LC Accu TOF and JMS-DX303HF instruments.

Materials

Phencyclone (1) was prepared by the previously reported method.³⁶⁾

Reaction of 1a with 2a (General Procedure)

A solution of 0.38 g (1.0 mmol) of 1a in 4 mL of toluene and 0.16 g (1.5 mmol) of 2a was heated at 80°C for 24 h under light-shielded condition. The mixture was cooled and the solvent was removed under vacuum. The products separated were recrystallized from acetone in darkness.

3aa: Yield 99%, pale yellow prisms; mp 263–266°C. IR (KBr): 1788 (bridged >C=O), 1676 (enone >C=O) cm⁻¹. ¹H-NMR (CDCl₃) δ (ppm): 4.46 (s, 2H, CH), 5.75 (s, 2H,=CH–), 7.08–7.66 (m, 14H, aromatic H), 8.29 (d, 2H, J=7.9 Hz, aromatic H), 8.68 (d, 2H, J=8.5 Hz, aromatic H); ¹³C-NMR (CDCl₃) δ (ppm): 48.1 (CH), 65.6 (>C<), 123.2, 126.0, 126.3, 126.7, 127.1, 128.2, 128.3, 128.9, 131.2, 133.3, 134.0 (aromatic C), 141.5 (=CH–), 194.7 (enone >C=O), 198.1 (bridge >C=O). m/z (electrospray ionization (ESI⁺)) high resolution (HR)-MS Calcd for C₃₅H₂₂O₃Na (M⁺+Na): 513.14666. Found: 513.14793.

3ab: Yield 73%, pale yellow prisms; mp and decomp. 260–265°C. IR (KBr): 1791 (bridged >C=O), 1669 (enone >C=O) cm⁻¹. ¹H-NMR (CDCl₃) δ (ppm): 0.75 (s, 3H, CH₃), 4.53 (d, 1H, J=8.6 Hz, CH), 4.57 (d, 1H, J=8.6 Hz, CH), 5.50 (s, 1H,=CH–), 7.12–7.66 (m, 14H, aromatic H), 8.34 (d, 2H, J=6.8 Hz, aromatic H), 8.68 (d, 2H, J=8.5 Hz, aromatic H); ¹³C-NMR (CDCl₃) δ (ppm): 15.5 (CH₃), 47.8, 48.8 (CH), 65.7, 65.9 (>C<), 123.4, 123.5, 126.0, 126.1, 126.5, 126.8, 127.0, 127.3, 128.2, 128.3, 128.5, 128.6, 129.1, 131.2, 131.3, 133.6, 133.8, 134.3 (aromatic C), 139.0 (=CH–), 152.3 (=C–CH₃), 194.3, 195.9 (enone >C=O), 198.5 (bridged >C=O). m/z (ESI⁺) HR-MS Calcd for C₃₆H₂₄O₃Na (M⁺+Na): 527.16231. Found: 527.16666.

3ac: Yield 84%, pale yellow prisms; mp and decomp. 208–212°C. IR (KBr): 1785 (bridged >C=O), 1647 (enone >C=O) cm⁻¹. ¹H-NMR (CDCl₃) δ (ppm): 0.81 (s, 3H, CH₃), 1.91 (s, 3H, CH₃), 4.11 (s, 1H, CH), 5.45 (s, 1H, =CH–), 7.02–7.71 (m, 14H, aromatic H), 8.35 (d, 1H, J=7.5 Hz, aromatic H), 8.68 (d, 3H, J=8.0 Hz, aromatic H); ¹³C-NMR (CDCl₃) δ (ppm): 15.8, 22.8 (CH₃), 54.5 (–C–CH₃), 59.7 (CH), 65.4, 67.5 (>C<), 123.2, 123.5, 124.9, 126.0, 126.6, 127.0, 127.3, 128.0, 128.3, 128.5, 128.7, 129.2, 129.5, 131.3, 132.6, 133.9 (aromatic C), 138.8 (=CH–), 152.4 (=C–CH₃), 194.7, 198.7 (enone >C=O), 199.0 (bridged >C=O). m/z (ESI⁺) HR-MS Calcd for C₃₇H₂₆O₃Na (M⁺+Na): 541.17796. Found: 541.18245.

3ae: Yield 95%, yellow prisms; mp 270–271°C. IR (KBr): 1794 (bridged >C=O), 1689 (enone >C=O) cm⁻¹. ¹H-NMR (CDCl₃) δ (ppm): 4.80 (s, 2H, CH), 6.66–6.71 (m, 2H, aromatic H), 6.73–6.78 (m, 2H, aromatic H), 7.13–7.74 (m, 14H, aromatic H), 8.35 (d, 2H, J=8.3 Hz, aromatic H), 8.48 (d, 2H, J=7.9 Hz, aromatic H); ¹³C-NMR (CDCl₃) δ (ppm): 49.1 (CH), 65.6 (>C<), 122.8, 124.4, 125.6, 126.4, 126.7, 126.9, 128.1, 128.2, 128.3, 129.0, 130.7, 131.1, 132.3, 133.5, 134.0, 136.3 (aromatic C), 194.6 (enone >C=O), 198.5 (bridged >C=O). Anal. Calcd for C₃₉H₂₄O₃: C, 86.65; 4.47. Found: C, 86.66; H, 4.49.

The ¹H- and ¹³C-NMR spectral data for 3aa and ¹³C-NMR spectral data for 3ae were identical to those reported by Wooi and White,¹⁵⁾ whereas the ¹H-NMR data of the aromatic signals for 3ae appear to be partially different from our data confirmed by the X-ray analysis.

Photocyclization of 3a (General Procedure)

A CDCl₃ (5 mL) solution of 3aa (2 mg) in NMR tube was irradiated with a 160 W halogen lamp (using a transparency overhead projector) at room temperature. The solvent was removed under reduce pressure and purified by recrystallization. The yield was determined based on the integrated intensity of the methine signals of 3aa and 4aa in ¹H-NMR. UV-Vis spectra were taken with a Shimadzu UV-2500 spectrometer.

4aa¹⁵⁾: Colorless prisms; mp >280°C (from CHCl₃). IR (KBr): 1758, 1740 (>C=O) cm⁻¹. ¹H-NMR (CDCl₃) δ (ppm): 3.30 (d, 2H, J=2.4 Hz, CH (cyclobutane)), 3.81 (d, 2H, J=2.4 Hz, CH), 5.69 (d, 2H, J=7.3 Hz, aromatic H), 6.67 (t, 2H, J=7.3 Hz, aromatic H), 6.92 (s, 4H, aromatic H), 7.17–7.29 (m, 8H, aromatic H), 7.88 (d, 2H, J=7.9 Hz, aromatic H); ¹³C-NMR (CDCl₃) δ (ppm): 51.5 (>C< (cyclobutane)), 52.1 (CH (cyclobutane)), 54.2 (CH), 61.2 (>C<), 122.9, 127.4, 128.1, 128.6, 129.0, 129.5, 129.7, 130.2, 131.5, 131.8 (aromatic C), 204.7, 207.9 (bridge >C=O). m/z (ESI⁺) HR-MS Calcd for C₃₆H₂₆O₄Na (M⁺+CH₃OH+Na): 545.17288. Found: 545.17657.

4ab: Colorless prisms; mp >280°C (from CHCl₃). IR (KBr): 1778, 1751 (>C=O) cm⁻¹. ¹H-NMR (CDCl₃) δ (ppm): 1.05 (s, 3H, CH₃), 2.80 (d, 1H, J=3.4 Hz, CH (cyclobutane)), 3.79 (dd, 1H, J=3.4, 10.3 Hz, CH), 3.84 (d, 1H, J=10.3 Hz, CH), 5.38 (d, 1H, J=8.0 Hz, aromatic H), 5.68 (d, 1H, J=7.5 Hz, aromatic H), 6.66 (t, 1H, J=8.1 Hz, aromatic H), 6.69 (d, 1H, J=8.0 Hz, aromatic H), 7.17–7.30 (m, 12H, aromatic H), 7.88 (d, 1H, J=8.6 Hz, aromatic H), 7.94 (d, 1H, J=8.6 Hz, aromatic H); ¹³C-NMR (CDCl₃) δ (ppm): 15.1 (CH₃), 49.4, 55.3, 56.1 (>C< (cyclobutane)), 50.9, 52.0 (CH), 60.4, 61.1 (>C<), 61.2 (CH (cyclobutane)), 122.9, 123.1, 126.6, 127.4, 128.0, 128.1, 128.5, 128.7, 129.5, 130.1, 130.3, 131.5, 131.6, 131.7, 131.8, 132.6 (aromatic C), 205.3, 208.5, 210.0 (>C=O). m/z (ESI⁺) HR-MS Calcd for C₃₇H₂₈O₄Na (M⁺+CH₃OH+Na): 559.18853. Found: 559.18839.

4ac: Colorless prisms; mp >280°C (from acetone). IR (KBr): 1775, 1735 (>C=O) cm⁻¹. ¹H-NMR (CDCl₃) δ (ppm): 1.16 (s, 3H, CH₃), 1.50 (s, 3H, CH₃), 2.80 (d, 1H, J=3.4 Hz, CH (cyclobutane)), 3.18 (d, 1H, J=3.4 Hz, CH), 5.54 (d, 1H, J=8.0 Hz, aromatic H), 5.69 (d, 1H, J=8.1 Hz, aromatic H), 6.52 (t, 1H, J=7.8 Hz, aromatic H), 6.66 (t, 1H, J=8.1 Hz, aromatic H), 7.07–7.26 (m, 12H, aromatic H), 7.86 (d, 2H, J=8.6 Hz, aromatic H); ¹³C-NMR (CDCl₃) δ (ppm): 15.3, 17.8 (CH₃), 49.5, 54.6, 55.4 (>C< (cyclobutane)), 54.9 (C–CH₃), 61.1 (CH), 61.1 (CH (cyclobutane)), 61.6, 64.5 (>C<), 122.8, 123.0, 126.4, 127.3, 127.8, 128.1, 128.2, 128.4, 129.3, 130.5 (aromatic C), 205.2, 208.9, 212.3 (>C=O). m/z (ESI⁺) HR-MS Calcd for C₃₈H₃₀O₄Na (M⁺+CH₃OH+Na): 573.20418. Found: 573.20794.

Crystal Structure Analysis

All measurements were performed on a RAXIS RAPID imaging plate area detector with graphite-monochromated MoKα radiation (λ=0.7107 Å). The structure was solved by direct method (SIR92), and all hydrogen atoms were located at calculated positions. The structure was refined by a full-matrix least-squares technique using anisotropic thermal parameters for non-hydrogen atoms and a riding model for hydrogen atoms. All calculations were performed using the crystallographic software package Crystal Structure.³⁷⁾ These X-ray crystallographic data have been deposited at the Cambridge Crystallographic Data Centre (CCDC).

Crystal data of 4aa: C₃₅H₂₂O₃·CHCl₃, M=609.93, monoclinic, Space group P2₁/n, a=15.494 (1), b=10.0396 (7), c=18.422 (1) Å, β=99.787 (3)°, V=2823.9 (4) ³, D_c=1.435 g cm⁻³, Z=4, R=0.049, R_w=0.093, CCDC ref. No. 929648.

Molecular Orbital (MO) Calculations

Semi-empirical MO (PM6) and Complete Neglect of Differential Overlap/Spectroscopic parameterization (CNDO/S) calculations were run through Winmostar³⁸⁾ interface using MOPAC2009.²¹⁾ All transition-states were confirmed by the presence of one imaginary frequency. DFT calculations at B3LYP/6-31G(d) level were carried out using Gaussian09 program package.¹⁹⁾ The calculation data (atomic coordinates) are available upon request (e-mail: [e-mail]).

Acknowledgment

We thank Mr. T. Miyagoe and Ms. K. Aizawa for experimental assistance.

References

• 1)  Harano K., Uchida K., Izuma M., Aoki T., Eto M., Hisano T., Chem. Pharm. Bull., 36, 2312–2322 (1988).
• 2)  Jikyo T., Eto M., Harano K., J. Chem. Soc., Perkin Trans. 1, 3463–3470 (1998) and cited therein.
• 3)  Jikyo T., Eto M., Harano K., Chem. Pharm. Bull., 45, 1961–1969 (1997).
• 4)  Eto M., Harano K., Oya T., Hisano T., Chem. Pharm. Bull., 41, 97–107 (1993).
• 5)  Yamaguchi K., Utsumi K., Yoshitake Y., Harano K., Tetrahedron Lett., 47, 4235–4239 (2006).
• 6)  Yoshitake Y., Yamaguchi K., Kai C., Akiyama T., Handa C., Jikyo T., Harano K., J. Org. Chem., 66, 8902–8911 (2001).
• 7)  Yamaguchi K., Eto M., Higashi K., Yoshitake Y., Harano K., Tetrahedron Lett., 52, 6082–6085 (2011).
• 8)  Harano K., Okamoto Y., Yasuda M., Ueyama K., Kanematsu K., J. Org. Chem., 48, 2728–2733 (1983).
• 9)  Okamoto Y., Harano K., Yasuda M., Osawa E., Kanematsu K., Chem. Pharm. Bull., 31, 2526–2529 (1983), and references cited therein.
• 10)  Eto M., Setoguchi K., Harada A., Sugiyama E., Harano K., Tetrahedron Lett., 39, 9751–9754 (1998).
• 11)  Yoshitake Y., Misaka J., Setoguchi K., Abe M., Kawaji T., Eto M., Harano K., J. Chem. Soc., Perkin Trans. 2, 1611–1619 (2002).
• 12)  Yoshitake Y., Misaka J., Abe M., Yamasaki M., Eto M., Harano K., Org. Biomol. Chem., 1, 1240–1249 (2003).
• 13)  Abe M., Eto M., Yamaguchi K., Yamasaki M., Misaka J., Yoshitake Y., Otsuka M., Harano K., Tetrahedron, 68, 3566–3576 (2012).
• 14)  Presented in part: Yamasaki M., Miyagoe T., Yoshitake Y., Harano K., 20th Annual Meeting of the Pharmaceutical Society of Japan (Kyushu branch), Fukuoka, December 2003. Unpublished master’s thesis (Yamasaki M., Mar. 2004, Graduate School of Pharmaceutical Sciences, Kumamoto University).
• 15)  Wooi G. Y., White J. M., Org. Biomol. Chem., 3, 972–974 (2005).
• 16)  Compound 3aa was first prepared by Sasaki et al. probably without knowing that the intramolecular [2+2] cycloaddition took place during purification by recrystallization. Sasaki T., Kanematsu K., Iizuka K., J. Org. Chem., 41, 1105–1112 (1976).
• 17)  Marchand A. P., Srinivas G., Watson W. H., ARKIVOC, 8–15 (2003).
• 18)  Compound 4aa enclathrated guest CHCl₃ molecule with a 1 : 1 host/guest stoichiometric ratio.
• 19)  Gaussian 09, Revision C.01, Frisch M. J., Trucks G. W., Schlegel H. B., Scuseria G. E., Robb M. A., Cheeseman J. R., Scalmani G., Barone V., Mennucci B., Petersson G. A., Nakatsuji H., Caricato M., Li X., Hratchian H. P., Izmaylov A. F., Bloino J., Zheng G., Sonnenberg J. L., Hada M., Ehara M., Toyota K., Fukuda R., Hasegawa J., Ishida M., Nakajima T., Honda Y., Kitao O., Nakai H., Vreven T., Montgomery Jr., J. A., Peralta J. E., Ogliaro F., Bearpark M., Heyd J. J., Brothers E., Kudin K. N., Staroverov V. N., Kobayashi R., Normand J., Raghavachari K., Rendell A., Burant J. C., Iyengar S. S., Tomasi J., Cossi M., Rega N., Millam J. M., Klene M., Knox J. E., Cross J. B., Bakken V., Adamo C., Jaramillo J., Gomperts R., Stratmann R. E., Yazyev O., Austin A. J., Cammi R., Pomelli C., Ochterski J. W., Martin R. L., Morokuma K., Zakrzewski V. G., Voth G. A., Salvador P., Dannenberg J. J., Dapprich S., Daniels A. D., Farkas Ö., Foresman J. B., Ortiz J. V., Cioslowski J., Fox D. J. Gaussian, Inc., Wallingford CT, 2009.
• 20)  The B3LYP/6-31G(d) calculations tend to overestimate strained bond lengths (ca. 0.01–0.02 Å) of the cage skeleton in comparison to what is observed experimentally. The X-ray and calculated bond lengths of 4aa are listed below.
• 21)  MOPAC2009 Version 9.062M: Stewart J. J. P., Stewart Computational Chemistry. ‹http://OpenMOPAC.net›.
• 22)  Schuster D. I., Lem G., Kaprinidis N. A., Chem. Rev., 93, 3–22 (1993).
• 23)  Bertrand C., Bouquant J., Pete J. P., Humbel S., J. Mol. Struct. THEOCHEM, 538, 165–177 (2001).
• 24)  Omar H. I., Shimo T., Somekawa K., J. Mol. Struct. THEOCHEM, 763, 115–121 (2006).
• 25)  The single-point-energy calculation at UB3LYP levels based on the PM6 geometry [UB3LYP/6-31G(d)//PM6] gave poor results for explaining the difference in reactivity. The height of the energy barrier for TS1 of 3aa and 3ba were 7.7 and 7.4 kcal/mol, respectively.
• 26)  Taylor R., Kennard O., J. Am. Chem. Soc., 104, 5063–5070 (1982).
• 27)  Desiraju G. R., Acc. Chem. Res., 24, 290–296 (1991).
• 28)  Desiraju G. R., Acc. Chem. Res., 29, 441–449 (1996).
• 29)  Steiner T., Kanters J. A., Kroon J., J. Chem. Soc., Chem. Commun., 1277–1278 (1996).
• 30)  Steiner T., Crystallogr. Rev., 6, 1–51 (1996).
• 31)  Steiner T., Angew. Chem. Int. Ed., 41, 48–76 (2002).
• 32)  Fukui K., “Kagaku Hanno to Densi no Kido (Chemical Reactions and Electron Orbitals),” Maruzen, Tokyo, 1976.
• 33)  Fleming I., “Frontier Orbitals and Organic Chemical Reactions,” Wiley, London, 1976.
• 34)  The inversely puckered conformation of the benzoquinones moiety (the carbonyl groups toward the endo site) was not optimized because it transformed to another conformation (the carbonyl groups toward the exo site) during the optimization.
• 35)  The CNDO/S calculations²¹⁾ of 3aa suggested the presence of a charge–transfer interaction. The calculation using the PM6-optimized structure predicted absorption wavelength arisen from HOMO→LUMO transition at 351 nm (oscillator strength: 0.03). Similar calculation of exo-3aa isomer did not show the corresponding absorption at the region.
• 36)  Forsyth W. R., Weisenburger G. A., Field K. W., Trans. Ill. State Acad. Sci., 89, 37–40 (1996).
• 37)  CrystalStructure 3.6.0: Crystal Structure Analysis Package, Rigaku and Rigaku/MSC (2000–2004). 9009 New Trails Dr. The Woodlands TX 77381 U.S.A.
• 38)  Senda N., Idemitsu Technical Report, 49, 106 (2006).