Skip to main content
NCBI home page
Nature Communications logoLink to Nature Communications
. 2026 Aug 3;17:9324. doi: 10.1038/s41467-026-76222-y

Divergent concerted and stepwise cycloadditions via enantioselective cross-conjugated iminium-ion catalysis

Erlaitz Basabe Obregón 1,#, Ida Rygaard Kocemba 1,#, Chi Zhang 2,#, Mikk Kaasik 1, Metin Cakiroglu 1, Philipp Waser 1, Pan-Pan Chen 2, K N Houk 2,✉, Karl Anker Jørgensen 1,✉
PMCID: PMC13530200  PMID: 42675065

Abstract

Cycloadditions are described as either concerted or stepwise processes, where a catalyst can influence both the mechanism and the peri-, regio- and stereoselectivity. We report experimental and DFT-computational investigations that reveal the versatile reactivities of cross-conjugated unsaturated iminium-ion intermediates. These undergo cycloadditions with cyclic and linear dienes via both stepwise and concerted pathways, respectively, providing two distinct classes of enantioenriched cycloadducts. Activation of 2-formyl-1,3-cyclohexadiene with a chiral 1,2-diaminocatalyst generates an acyclic cross-conjugated iminium-ion intermediate, which by utilizing 4π-electrons, undergoes enantioselective inverse-electron-demand [4 + 2] cycloadditions with cyclic dienes via a stepwise mechanism, as well as indene and a styrene, affording cycloadducts in high yields and up to 96% ee. In contrast, only 2π-electrons of the cross-conjugated iminium-ion intermediate are involved in the [4 + 2] cycloaddition with linear dienes, proceeding via an ambimodal asynchronous concerted pathway, yielding decalin derivatives in up to 96% ee. The generality of this catalytic enantioselective concept is expanded to also include acyclic cross-conjugated iminium-ion intermediates, which react with both cyclic and linear dienes following the same two distinct reaction pathways. In contrast, the analogous iminium-ion, generated from 1-formylcyclohexene, does not react under the same conditions.

Subject terms: Asymmetric catalysis, Organocatalysis

The authors report experimental and DFT-computational investigations that reveal the versatile reactivities of cross-conjugated unsaturated iminium-ion intermediates. These undergo cycloadditions with cyclic and linear dienes via both stepwise and concerted pathways, respectively, providing two distinct classes of enantioenriched cycloadducts.

Introduction

The Diels-Alder reaction is very general and synthetically valuable for applications in both total synthesis and industry1–4. The [4 + 2] cycloaddition is allowed by the Woodward-Hoffmann rules5, which leads to a fundamental mechanistic understanding and prediction of the stereochemical outcomes of cycloadditions6.

The Diels-Alder reaction usually entails the reaction of an electron-rich diene with an electron-deficient olefin, resulting in a six-membered carbocyclic adduct in one step with 100% atom efficiency. The inverse-electron-demand Diels-Alder reaction is the cycloaddition of an electron-poor diene with an electron-rich olefin and has many applications, especially in biorthogonal chemistry7–9.

Organocatalysis has added a new dimension to cycloadditions10–13 as a versatile activation mode for aldehydes, enals, and polyenals. These synthetic pathways and enhanced reactivity based on nucleophilic enamine-14–16, dienamine-17–19, cross-dienamine-20–23, and polyenamine intermediates24–26 have been demonstrated, where the control of peri-, chemo-, diastereoselective, and enantioselectivity in cycloadditions has led to the construction of complex molecular scaffolds.

Enals and polyenals are also attractive substrates for undergoing organocatalytic electrophilic activation, generating reactive intermediates, such as iminium-27–31 and poly-iminium-ions (Fig. 1a)32,33. Therefore, organocatalysis can generate intermediates, which act as both electron-rich and electron-poor reaction partners in cycloadditions, enamines, and iminium ions, respectively.

Fig. 1. Activation modes.

Fig. 1

Open in a new tab

a Main activation modes in iminium-ion catalysis. b Two different pericyclic reactions of 2-formyl-1,3-cyclohexadiene with cyclic dienes and styrene (left) versus linear dienes (right) catalyzed by a primary aminocatalyst.

Activated polyenals can react by competing pathways, and the stereoselectivity and periselectivity may depend on the reaction partner34–38. Recent studies have shown that formyl cycloheptatrienes react by different pericyclic reactions with 6,6-dimethylfulvene, and the cycloadduct formed is dependent on the position of the formyl group relative to the cycloheptatriene moiety39.

In this work, we study the unique reactivity of enantioselective cross-conjugated iminium-ion catalysis by activating 2-formyl-1,3-cyclohexadiene with a primary aminocatalyst, demonstrating its potential in asymmetric synthesis, and unraveling the underlying mechanistic details for different classes of [4 + 2] cycloadditions with dienes. Surprisingly, the periselectivity of the cycloaddition is dependent on the diene reaction partner (Fig. 1b). For cyclic dienes and styrenes, the cross-conjugated iminium-ion intermediate acts as a 4π-component, and the cycloaddition proceeds with only one of the olefinic bonds in the cyclic diene and styrene (Fig. 1b, left). While the cycloaddition with linear dienes (Fig. 1b, right) proceeds through the same intermediate, only the 2π olefin directly in conjugation with the formyl group reacts with all 4π-electrons of the linear diene.

We also demonstrate that a common intermediate, the cross-conjugated iminium ion, can react by different mechanisms–stepwise or ambimodal concerted reaction pathways–depending on the nature of the reacting diene partner37,40–42.

Results

Reaction concept

We first examined the condensation of 2-formyl-1,3-cyclohexadiene 1 with a primary aminocatalyst, which was envisioned to generate the cross-conjugated iminium-ion intermediate I in Fig. 2. The LUMO of I is polarized towards C6 (vide infra), which allows for a [4 + 2] cycloaddition at C6 and C1 in I with a conjugated system. However, an inverse-electron demand [4 + 2] cycloaddition at C6 and C3 is also feasible.

Fig. 2. Screening of aminocatalysts 3a-h and reaction conditions for the enantioselective [4 + 2] cycloaddition of 2-formyl-1,3-cyclohexadiene 1 with 6,6-dimethylfulvene 2a.

Fig. 2

Open in a new tab

a Reaction conditions unless otherwise noted:1 (1.0 equiv), 2a (2.0 equiv), 3a-h (0.2 equiv), TFA (40 mol%) in CH2Cl2 (0.5 M) for 3 h. Conversion of 1 and NMR yield of 4a measured by 1H NMR analysis of the crude reaction mixture relative to the internal standard mesitylene. ee was determined by chiral stationary phase UPC2. b Isolated yield. c TCA used instead of TFA. d No TFA added. eCalculated LUMO of the cross-conjugated iminium-ion intermediate I formed by reaction of 2-formyl-1,3-cyclohexadiene 4 with (1 R,2 R)-1,2-diphenylethane-1,2-diamine 3a. The percentage densities in the LUMO are given next to each orbital, and the green and blue correspond to positive and negative regions of the wavefunction. The orbital is plotted as isodensity surfaces.

Reaction development

To determine whether one of the envisioned cycloadditions occurs with dienes, we screened reaction conditions for the reaction of 2-formyl-1,3-cyclohexadiene 1 with 6,6-dimethylfulvene 2a in the presence of several aminocatalysts 3a-h.

As shown in Fig. 2, the reaction of 2-formyl-1,3-cyclohexadiene 1 with 6,6-dimethylfulvene 2a proceeds in the presence of (1 R,2 R)-1,2-diphenylethane-1,2-diamine 3a as an inverse-electron-demand [4 + 2] cycloaddition, affording cycloadduct 4a in 87% yield and excellent enantioselectivity of 96% ee (Fig. 2, Entry 1). Cycloadduct 4a is formed by reaction of C6 and C3 in the cross-conjugated iminium-ion intermediate I with one of the olefinic double bonds in the cyclic moiety in 2a participating in the cycloaddition.

The more hindered catalyst 3b, as well as the pyrrolidine-containing catalyst 3c, also catalyzed the desired reaction affording 4a in good yield, albeit with reduced enantioselectivity (Entries 2,3). The chiral 1,2-diaminocyclohexane catalyst 3d also proved to be suitable for the formation of 4a; however, in significantly lower yield and enantioselectivity (Entry 4). Diarylprolinol catalysts 3e,f could also catalyze the desired cycloaddition, but with low to no enantioselectivity (Entries 5,6). Cinchona alkaloid catalysts 3 g,h also efficiently catalyzed the reaction; however, catalyst 3a remained the most efficient one (Entries 7,8). Therefore, the optimization process was continued using 3a. Using TCA as the acid additive led to slightly lower yield and enantioselectivity (Entry 9), while lowering the reaction temperature resulted in a slight improvement in the enantioselectivity (97% ee); however, cycloadduct 4a was obtained in lower yield (Entries 10,11). Furthermore, both acid and aminocatalyst were found to be essential for the reaction to proceed, as demonstrated in entries 12, 13, also showing negligible background reactivity. Finally, the effect of TFA mol% on the reaction yield and enantioselectivity was investigated, showing that 40% was the optimal acid amount. Higher amounts of TFA led to lower yield due to faster degradation of 1, while the lower enantioselectivity observed can be attributed to increased background reactivity. (Fig. 2, right, for more optimization results see Supporting Information).

Within 6,6-dimethylfulvene 2a, there are several conjugated systems that can react with the conjugated systems of intermediate I. During the optimization process in Fig. 2, cycloadduct 4a was the only observed product. The observed reactivity of 2a prompted us to expand this [4 + 2] cycloaddition by reacting 2-formyl-1,3-cyclohexadiene 1 with a series of cyclic dienes and styrene 2.

Scope for reaction of cyclic dienes and styrene 2, and linear dienes 5

Figure 3 shows examples of the [4 + 2] cycloaddition of 2-formyl-1,3-cyclohexadiene 1 with cyclic dienes and styrene 2 (Fig. 3a) and linear dienes 5 (Fig. 3b) catalyzed by (1 R,2 R)- 1,2-diphenylethane-1,2-diamine 3a.

Fig. 3. Scope of the [4 + 2] cycloaddition of 2-formyl-1,3-cyclohexadiene 1 catalyzed by (1 R,2 R)-1,2-diphenylethane-1,2-diamine 3a.

Fig. 3

Open in a new tab

a Cyclic dienes and styrene 2. b Linear dienes 5.a Reaction conditions with cyclic dienes and styrene 2: 1 (1.0 equiv), 2 (2.0-4.0 equiv), 3a (20 mol%), TFA (40 mol%) in CH2Cl2 (varying, 0.5-1 M) or neat at T (varying, –40-rt °C). ee was determined by chiral stationary phase UPC2. (E):(Z)-ratio determined by 1H NMR.bReaction conditions with linear dienes 5: 1 (1.0 equiv), 5 (4.0 equiv), 3a (20 mol%), TCA (40 mol%) in CH2Cl2 (varying, 0.5-1 M) at T (varying, –25-40 °C) or neat. ee was determined by chiral stationary phase UPC2. d.r. determined by 1H NMR.c Catalyst 3 h in TFA (40 mol%) at –40 °C (Result in gray corresponds to standard conditions with catalyst 3a).dCatalyst 3 f in TFA (40 mol%) at –10 °C (Result in gray correspond to standard conditions with catalyst 3a).eIsolated yield of the aldehyde product in parentheses.

The aminocatalytic [4 + 2] cycloaddition of 2-formyl-1,3-cyclohexadiene 1 with 6,6-dimethylfulvene 2a afforded the inverse-electron-demand cycloadduct 4a in high yield (87%) with excellent enantioselectivity (96% ee, Fig. 3a) in the presence of catalyst 3a. The cyclohexyl analog cycloadduct 4b was obtained in 83% yield and with a minor decrease in enantiomeric excess–89% ee. The corresponding 6,6-diarylfulvenes 2c,d also reacted smoothly with 1 and the cycloadducts 4c,d were obtained in good yields and 94-95% ee for both cycloadducts. The absolute configuration was determined by X-ray crystallography of 4d (Fig. 3a) after recrystallization ( > 99% ee) and assigned in analogy.

Next, we investigated the cycloadditions of 6-arylfulvenes having electron-donating or withdrawing substituents in the para-position; cycloadducts 4e-j were obtained in high yields (78-90%) and 86-93% ee for the major (E)-isomer, while the (Z)-isomer was obtained with a significantly lower enantioselectivity. It should be noted that both isomers of 4e-j can be isolated as isomeric pure products (see Supporting Information).

We have found that the (E):(Z) ratio of the cycloadducts 4e-j is dependent on the electronic nature of the substituents in the 6-arylfulvenes. As shown to the right in Fig. 3a, electron-withdrawing substituents slightly favor the formation of (E)-isomer in contrast to electron-donating substituents. As shown in the Supporting Information, the computational investigations account for the trend.

To test the [4 + 2] cycloaddition with simpler cyclic dienes, we also studied the reaction of 2-formyl-1,3-cyclohexadiene 1 with cyclopentadiene 2h. We were pleased to find the 2h also underwent the [4 + 2] cycloaddition smoothly, affording cycloadduct 4k in 68% yield and with a minor decrease in the enantiomeric excess to 59% ee (see Supporting Information for the specific optimization of this substrate). The desired [4 + 2] cycloaddition also proceeds for 1H-indene 2 l, affording cycloadduct 4l in 71% yield and with a high enantiomeric excess (87% ee).

To further challenge the reaction concept, 4-methoxystyrene 2m was tested and also found to participate in the cycloaddition. Catalyst 3a provided the cycloadduct 4m in low yield and enantiomeric excess (32%, 12% ee). Pleasingly, by introducing catalyst 3f, a significant increase in enantiomeric excess to 82% ee was found along with an increase in yield to 43% (see Supporting Information for the specific optimization of this substrate).

After the successful reactivity shown for 4-methoxystyrene 2m, we next investigated the reactivity of trans-1-phenyl-1,3-butadiene 5a with 2-formyl-1,3-cyclohexadiene 1, applying the reaction conditions used for the cyclic dienes 2. Surprisingly, a remarkable change in periselectivity took place when changing the substrate from a cyclic (or 4-methoxystyrene) to a linear diene, as now only 2π-electrons of 2-formyl-1,3-cyclohexadiene 1 reacted with the 4π-electrons of 5a, leading to the formation of chiral decalin 6a. Such decalin scaffolds are of importance in natural products43. Thus, in this reaction, the cross-conjugated iminium-ion intermediate I reacted at C6 and C1 (see Fig. 2).

The reaction conditions for the scope in Fig. 3a were successfully extended to linear dienes 5 (Fig. 3b), with only a minor modification of the acid additive from TFA to TCA (see Supporting Information for more optimization results).

Figure 3b shows the results for the [4 + 2] cycloaddition of 2-formyl-1,3-cyclohexadiene 1 with linear dienes 5 catalyzed by 3a. All the aldehyde products were found to be stable and could be isolated, as exemplified for 6a or 6g-j (see Supporting Information). The yields for 6a-f in Fig. 3b are shown after reduction to the corresponding alcohol (for aryl-containing dienes). For non-aryl-containing dienes, subsequent hydrazone formation (for an analytical quantity of 6g-j) was performed for UPC2-separation purposes (see Supporting Information).

The linear dienes 5 react smoothly with 2-formyl-1,3-cyclohexadiene 1 in the presence of aminocatalyst 3a. trans-1-Phenyl-1,3-butadiene 5a provided cycloadduct 6a in high yield with 96% ee, and for different para-substituted trans-1-phenyl-1,3-dienes 5b-d, the [4 + 2] cycloaddition also proceeded smoothly, affording 6b-d in 43-83% yield with 90–94% ee. Increased steric hindrance in the ortho-position did not affect the outcome of the cycloaddition, as 6e was obtained in 68% yield with 91% ee. The reaction with 3-methyl-substituted butadiene 5f also occurred, and cycloadduct 6f was formed in good yield and enantioselectivity.

The reaction also proceeded successfully for aliphatic dienes 5g-j as 4π-components. For 1,3-pentadiene and 1,3-hexadiene, the reaction proceeded, affording the cycloadducts 6g,h in good yield and high diastereoselectivity and enantioselectivity. We also studied simple aliphatic dienes; butadiene 5i and 2,3-dimethylbutadiene 5j provided the two simple cis-decalins 6i,j in moderate to good yields and good enantiomeric excesses.

The absolute configuration of cycloadducts 6 were assigned based on ECD (see Supporting Information).

Finally, we tested the cycloaddition applying cis-1-p-MeO-phenyl-1,3-butadiene cis-5b in the reaction of 2-formyl-1,3-cyclohexadiene 1 under the same reaction conditions (Fig. 3b). However, no cycloadducts were formed, indicating that the reaction is only feasible for the trans-1-aryl dienes 5 as presented in Fig. 3b.

The results in Fig. 3 demonstrate that the [4 + 2] cycloaddition of cyclic dienes and 4-methoxystyrene with 2-formyl-1,3-cyclohexadiene 1 catalyzed by (1R,2 R)-1,2-diphenylethane-1,2-diamine 3a proceeds as an inverse-electron-demand cycloaddition, while for the reaction with linear dienes, a normal electron-demand [4 + 2] cycloaddition takes place. To gain a deeper understanding of the activation mode and the variability of reaction pathways with different dienes, computational investigations were conducted. The role of the proposed intermediate I (Fig. 2), the influence of organocatalyst 3a on the two very different reaction pathways, and the factors determining the regio-, peri-, diastereo- and enantioselectivity were all investigated.

Computational investigations of the aminocatalytic cycloaddition of the cyclic dienes

Computational investigations were performed with DFT methods at the M06-2X/def2-TZVP-CPCM(CH2Cl2)//M06-2X/def2-SVP-CPCM(CH2Cl2) level of theory (Computational details are included in the Supporting Information).

The calculated LUMO of the optimized structure of intermediate I has large coefficients at C6 and C7, consistent with the high, nearly equal electrophilicity of these carbons (see Fig. 2 and Supporting Information). Interestingly, the calculated structure for I reveals that the aminocatalyst shows no face-shielding at C6, and therefore, the origin of the high enantioselectivity obtained for the cyclic dienes is not obvious.

Carbon C6 is more electrophilic than C7, because attack on C7 experiences an antibonding overlap with C8. C6 is the carbon attacked by the electron-rich carbon–Ca of 6,6-dimethylfulvene 2a. If 2a reacted as a 4π-component and I as a 2π-component, a [4 + 2] cycloaddition could take place with second-bond formation to C1 in I to Cd in 2a, thereby affording cycloadduct 4a’ (Fig. 4)44. This is clearly the only product predicted by the application of simple FMO theory, since there is no contribution of the additional double bond orbitals to the LUMO. On the other hand, if I did act as a 4π-component, the reaction could proceed as a [4 + 2] cycloaddition by reaction with Ca and Cb in 2a, leading to 4a (Fig. 4). Both of these reaction paths could arise from the strong interaction of nucleophilic Ca of 2a with C6 of the cross-conjugated iminium-ion intermediate.

Fig. 4. The two potential [4 + 2] cycloadditions between 2-formyl-1,3-cyclohexadiene 1 and 6,6-dimethylfulvene 2a catalyzed by 1,2-diphenylethane-1,2-diamine 3a.

Fig. 4

Open in a new tab

Shown are the two possible reaction products 4a’ (top) and 4a (bottom).

In the event, computations predict that both products can be formed from a single transition state and a carbocationic intermediate! Fig. 5 shows DFT-computed free-energy changes of two cycloaddition pathways of 2-formyl-1,3-cyclohexadiene 1 with 6,6-dimethylfulvene 2a catalyzed by (1R,2R)-1,2-diphenylethane-1,2-diamine 3a (see Supporting Information). The condensation of the aminocatalyst 3a with 1 leads to the formation of the cross-conjugated iminium-ion intermediate I is thermodynamically favored by 2.4 kcal/mol. The first C–C bond formation, the primary interaction affording zwitterion II proceeds through TS1 in which the C6–Ca bond-forming distance is 2.07 Å. The distances between C1–Cd and C3–Cb are 3.09 Å and 3.22 Å, respectively, indicating TS1 has a similar structure to other ambimodal transition states40–43, exactly as predicted from FMO theory (see Fig. 2 and Supporting Information). Despite this ambimodal TS resemblance, only the acyclic intermediate is formed, and the reaction is a non-pericyclic stepwise process, due to the high stabilization of the carbocationic intermediate. (If the dienophile had been the aldehyde, this would be a much less stable zwitterion. This was not observed.) From intermediate II, two reaction paths were found: one that proceeds by TS2 (red line) forming the C3–Cb bond affording the inverse-electron-demand Diels-Alder product 4a, or another via TS3 (blue line), leading to C1–Cd bond formation providing intermediate V. Subsequent hydrolysis affords the Diels-Alder cycloadduct 4a’ (Fig. 5). During the formation of 4a, TS1 is the rate-determining step, but only 0.4 kcal/mol higher in energy than TS2.

Fig. 5. DFT-computed free energy changes of [4 + 2] cycloadditions of 2-formyl-1,3-cyclohexadiene 1 and 6,6-dimethylfulvene 2a catalyzed by (1R,2R)−1,2-diphenylethane-1,2-diamine 3a.

Fig. 5

Open in a new tab

Energies are given in kcal/mol. Key distance is given in Å.

The reaction pathway where I reacts at C6 and C1 with Ca and Cd in 2a leading to intermediate V proceeds through TS3, having higher energy than TS2 by 3.5 kcal/mol. Intermediate III can generate the more stable conformer IV by bond rotation. Final hydrolysis releases the [4 + 2] cycloadduct 4a and regenerates the active catalyst 3a-H+.Fig. 6

Fig. 6. Distortion/interaction-activation strain (DIAS) analysis and independent gradient model based on Hirshfeld partition (IGMH) analysis of TS1 and its corresponding diastereomer TS1’.

Fig. 6

Open in a new tab

Energy reference points are defined as the separated reactants. Fragment distortion and interaction energies are electronic energies calculated at the M06-2X level of theory with the def2-TZVP basis set without the inclusion of solvation energy corrections (black, activation energies; green, distortion energies of cyclohexadiene iminium ion; blue, distortion energies of 6,6-dimethylfulvene; red, interaction energies). For reference, the free energy of activation for TS1’ is 0.8 kcal/mol higher than that of TS1. Energies are given in kcal/mol. Key distance is given in Å. Relative energies are not to scale.

From both thermodynamic and kinetic perspectives, the formation of [4 + 2] cycloadduct 4a is preferred over 4a’, which aligns with the experimental observations and crystal structure. The 3.5 kcal/mol lower barrier for formation of III rather than V is complemented by the 19.6 kcal/mol lower energy of III compared to V. However, V can be converted to III by a Cope rearrangement; in this case involving the cationic intermediate II. Other less favored pathways are included in Fig. S1 in Supporting Information.

The enantioselectivity observed in this reaction originates from the energy differences between the diastereomeric pathways during the formation of the first transition state, TS1 (Fig. 6) (also see Supporting Information). Distortion/interaction-activation strain (DIAS) analysis45 of TS1 and the corresponding diastereomer TS1’ reveals that the energy differences are primarily attributed to the distortion of the 6,6-dimethylfulvene 2a. Independent Gradient Model based on Hirshfeld partition (IGMH) analysis illustrates that the CH-π interaction between 2a and aminocatalyst 3a in TS1 is very similar to that of in TS1’.

However, this non-covalent interaction in TS1 induces a further rotation of the methyl group of the fulvene, hence enlarging the dihedral angle between the two C–H bonds highlighted in green in Fig. 6 and avoiding the disfavored eclipsed conformation to reduce the 1,3-allylic strain. As a result, the fulvene fragment is more stable compared to that of TS1’. We propose that these arguments might also account for why cyclopentadiene only gives 20% ee under identical conditions, compared to the significantly higher enantioselectivity obtained for the other cyclic dienes in Fig. 3a.

Computational investigations of the aminocatalytic cycloadditions of linear dienes

We also performed a computational analysis of the [4 + 2] cycloadditions of 2-formyl-1,3-cyclohexadiene 1 with linear dienes 5 catalyzed by (1R,2R)-1,2-diphenylethane-1,2-diamine 3a.

Linear dienes 5a also have the possibility for two types of [4 + 2] cycloadditions with 2-formyl-1,3-cyclohexadiene 1 in the presence of (1R,2R)-1,2-diphenylethane-1,2-diamine aminocatalyst 3a. These two reaction pathways are presented for trans-1-phenyl-1,3-butadiene 5a in Fig. 7.

Fig. 7. The two potential [4 + 2] cycloadditions of 2-formyl cyclohexadiene 1 with trans-1-phenyl-1,3-butadiene 5a catalyzed by (1R,2R)-1,2-diphenylethane−1,2-diamine 3a.

Fig. 7

Open in a new tab

Shown are the two possible reaction products 6a (top) and 6a’ (bottom).

DFT-computed free-energy changes of the two [4 + 2] cycloadditions of 2-formyl-1,3-cyclohexadiene 1 with trans-1-phenyl-1,3-butadiene 5a in the presence of the aminocatalyst (1R,2R)-1,2-diphenylethane-1,2-diamine 3a are shown in Fig. 8 (see Supporting Information).

Fig. 8. DFT-computed free energy changes of [4 + 2] cycloadditions of 2-formyl-1,3-cyclohexadiene 1 and trans-1-phenyl-1,3-butadiene 5a in the presence of (1R,2R)−1,2-diphenylethane-1,2-diamine 3a.

Fig. 8

Open in a new tab

Energies are given in kcal/mol. Key distance is given in Å.

A change in periselectivity takes place when changing the substrate from a cyclic to a linear diene. The free energy profile in Fig. 8 shows that the most favored pathway for the reaction of trans-1-phenyl-1,3-butadiene 5a involves the principal primary orbital interaction between the most nucleophilic terminus of the diene and the electrophilic site of the iminium ion, exactly like 6,6-dimethylfulvene 2a described earlier. As before, and for ambimodal transition states in general, the primary orbital interaction distance is 2.10 Å, while the conditional primary interactions are 3.02 Å, in the α-position to the iminium-ion, and 3.15 Å to its γ-position. For the reaction with 2a, these three distances are 2.07, 3.09, and 3.22 Å, respectively. (Fig. 5). While the reaction of 2a proceeds from this type of transition state to a carbocationic intermediate, the acyclic diene 5a reacts via this ambimodal transition state to connect directly to the Cope rearrangement transition state TS5. This connects the two [4 + 2] intermediates, VI and VII, where any generated VII will rapidly undergo this Cope rearrangement to form the more stable VI. Once intermediate VI is formed, final hydrolysis regenerates the active catalyst 3a-H+ and delivers cycloadduct 6a.

The difference between the reaction with 6,6-dimethylfulvene 2a and that of the acyclic diene 5a is the high energy of the norbornene-containing adduct V, which prevents concerted formation of this product. Instead, the stabilized carbocation is generated from the intended ambimodal TS-1. (Fig. 5). Other less favored pathways are included in Fig. S2 in Supporting Information.

DIAS and second-order perturbation theory analysis were applied to the diastereomer of TS4 to investigate the factors contributing to the enantioselectivity of the [4 + 2] cycloaddition. Transition state TS4 is 1.4 kcal/mol lower in energy than TS4’ in terms of Gibbs free energy, corresponding to an enantiomeric excess of ca. 83% ee. The DIAS results suggest that the interaction energy is the primary factor driving selectivity. When the diene approaches from the Si-face of the imine group in TS4’, the bulky substituents of the aminocatalyst block the approaching substrate, resulting in a larger Cd–C7 distance (3.16 Å), which weakens the secondary-orbital interactions (Fig. 9b). Conversely, in the favorable TS4, the substrate approaches from the Re-face of the imine group, where the substituents of the catalyst are positioned in an open space, minimizing steric effects. As a result, the Cd–C7 distance becomes shorter (2.89 Å), enhancing the secondary-orbital interactions, ultimately determining the enantioselectivity.

Fig. 9. Optimized structures of VI and VII and distortion/interaction-activation strain.

Fig. 9

Open in a new tab

a DFT-Optimized structures of VI and VII. b Distortion/interaction-activation strain (DIAS) analysis and HOMO orbitals of TS4 and corresponding diastereomer TS4’. The legend of DIAS analysis is the same as that of Fig. 6. Fragment distortion and interaction energies are electronic energies, with energy reference points defined as the separated reactants. For reference, the free energy of activation for TS4’ is 1.4 kcal/mol higher than that of TS4. Energies are given in kcal/mol. Key distance is given in Å. Relative energies are not to scale.

The driving factor that determines the enantioselectivity changes when moving from a cyclic to a linear diene. In the first case, fulvene distortion determines the facial selectivity, but in the latter case, it is determined by the interaction component between the two species. Interestingly, although the catalyst does not offer any direct shielding effect to the cyclohexadiene core, the drive to maximize secondary-orbital interactions forces the diene to approach in a manner that utilizes the bulky backbone of the catalyst as a remote shield.

Experimental verification of computational studies for the linear dienes

The calculations of the reaction pathways for cyclic and linear dienes with 2-formyl-1,3-cyclohexadiene reveal some intriguing results: i) The cyclic and linear dienes react through two distinct reaction pathways, the cyclic involving a carbocationic intermediate, the acyclic an ambimodal transition state. Both reactions have very similar rate-determining transition states; the differences in products are determined by thermodynamics that control the second bond formation, which is not very advanced in the rate-determining transition state. ii) The cross-conjugated iminium-ion intermediate also undergoes two different reactions, depending on the type of diene—for the cyclic dienes, the intermediate reacts utilizing its 4π-electrons, while for the linear dienes, only 2π-electrons from the intermediate are involved.

Control experiments and proof-of-reaction concept

The calculations of the reaction of 2-formyl-1,3-cyclohexadiene 1, through the cross-conjugated iminium-ion intermediate I with trans-1-phenyl-1,3-butadiene 5a, showed that I acts as an ‘isolated’ iminium-ion (Fig. 8). Intrigued by this, we investigated the reaction of 1-formyl cyclohexene 7 with trans-1-p-MeO-phenyl-1,3-butadiene 5b and with 6,6-dimethylfulvene 2a in the presence of catalyst 3a to investigate the role of a simple iminium-ion. As shown in Fig. 10a, no reaction was observed for both dienes. The unsuccessful reaction with 2a is likely due to being thermodynamically unfavorable (see Fig. 5). The lack of reactivity with 5b for this reaction indicates that the secondary conditional primary interaction with intermediate I in the TS (the 3.15 Å partial bond in TS4, Fig. 8) facilitates the reaction, even when it does not lead to a bond in the cycloadduct.

Fig. 10. Control experiments.

Fig. 10

Open in a new tab

a Control experiments with 1-formyl cyclohexene 7. b Modification of the amine scaffolds in catalyst 3. c Proof-of-reaction concept with 2-formyl-1-phenyl-1,3-butadiene 10. d Reactivity of 5-phenylpenta-2,4-dienal 13 under the same reaction conditions. Reaction conditions: 10 or 13 (1.0 equiv), 2a and 5b (4.0 equiv), 3a (20 mol%) with 40 mol% TCA or TFA in CH2Cl2 (1.0 M) at room temperature. ee was determined by chiral stationary phase UPC2.

Extensive computational attempts to generate a potential [4 + 2] transition state for the reaction of 1-formyl cyclohexene 7 with linear dienes 5, failed to give a reasonable structure for this transition state, perhaps because the putative zwitterionic structures are not stable at all and collapse back to reactants.

We also wanted to further investigate the origin of stereoselectivity by modifying catalyst 3a by first protecting one of the amines with a tosyl-functionality (3i) and then replacing the primary amine functionality with an alcohol (3j) (Fig. 10b). It appears from the results that the presence of both basic nitrogen functionalities is crucial for obtaining high stereoselectivity (see also Fig. 2, Entry 3) for both reaction pathways.

To expand this reaction concept, we explored other systems than 2-formyl-1,3-cyclohexadiene 1. We were pleased that 2-formyl-1-phenyl-1,3-butadiene 10 reacts in the presence of (1 R,2 R)- 1,2-diphenylethane-1,2-diamine 3a with 6,6-dimethylfulvene 2a and 1-p-MeO-phenyl-1,3-butadiene 5b through the same two reaction pathways as observed for 1 (Fig. 10c).

The reaction with 6,6-dimethylfulvene 2a afforded the inverse-electron demand [4 + 2] cycloadduct 11 in 46% yield, 1.8:1 d.r., and 61% ee (for the major diastereoisomer)–the same class of cycloadducts as obtained for the cyclic dienes in Fig. 3a. Changing from a cyclic diene to a linear diene, 10 reacted with 5b in a similar type of [4 + 2] cycloaddition as for the linear dienes in Fig. 3b, affording cycloadduct 12 in 55% yield, >20:1 d.r. and 93% ee.

Thus, the dual reactivity observed for 2-formyl-1,3-cyclohexadiene 1 with cyclic and linear dienes in the presence of (1 R,2 R)- 1,2-diphenylethane-1,2-diamine 3a as the catalyst is not restricted to cyclic systems, as the acyclic system 2-formyl-1-phenyl-1,3-butadiene 10 in the presence of catalyst 3a reacts under the reaction conditions in the same dual manner.

Finally, we wanted to compare the reactivity of cross-conjugated iminium-ion intermediates with a related linear-conjugated iminium-ion intermediate, by applying identical conditions than the previous acyclic cross-conjugated system 10. The reaction of linear dienal 13 with 6,6-dimethylfulvene 2a, which now reacted as a 4π component, proceeded very sluggishly, while no cycloadducts were formed in the reaction with the linear diene 5b, showing the superiority of the cross-conjugated iminium-ion intermediates (Fig. 10d).

Discussion

We have presented a development in organocatalytic enantioselective cycloaddition chemistry–a common cross-conjugated iminium-ion intermediate that reacts either by a stepwise mechanism or an ambimodal asynchronous concerted pathway, providing attractive chiral cycloadducts. The divergence between mechanisms is observed for cyclic and linear substrates. 2-Formyl-1,3-cyclohexadiene reacts with (1R,2R)-1,2-diphenylethane-1,2-diamine to generate a cross-conjugated iminium-ion intermediate that, acting as a 4π-electron-deficient component, undergoes enantioselective inverse-electron-demand [4 + 2] cycloadditions with cyclic dienes and styrene. The cycloadditions proceed in good to high yields and good to excellent enantioselectivities. The cross-conjugated iminium ion reacts with 6,6-dimethylfulvene in a stepwise manner, generating a zwitterionic intermediate. The periselectivity of the [4 + 2] cycloaddition is determined by the second-bond formation, affording the cycloadduct by utilizing the 4π-electrons in the cross-conjugated iminium ion. The reactions of linear dienes resulted in a change in the mechanism, also leading to a different cycloadduct. For this system, all 4π-electrons in the linear diene are involved in the cycloaddition reacting with the 2π-electrons in direct conjugation with the iminium-ion functionality in the common intermediate. This reaction proceeds through a concerted asynchronous ambimodal [4 + 2]/[2 + 4] transition state. Linear dienes reacted smoothly, affording the cycloadducts in moderate to good yields, and generally excellent diastereoselectivity and enantioselectivity. The generality of the reaction concept was demonstrated for acyclic 2-formyl-1-phenyl-1,3-butadiene, which also undergoes the same two reaction pathways with cyclic and linear dienes. These reactions of cross-conjugated iminium-ion intermediates expand the scope of cycloaddition reactions and open new dimensions for cycloadditions.

Methods

General procedures for the enantioselective [4 + 2] cycloaddition with cyclic dienes

A 2 mL flame dried vial was charged with a solution of 2-formyl-1,3-cyclohexadiene 1 in benzene (20% v/v, 54 mg, 0.10 mmol, 1.0 equiv), 100 μL dry degassed CH2Cl2 was added and stirred for 1 min. Catalyst 3a (4.3 mg, 0.020 mmol, 0.20 equiv) and fulvene 2a-i (0.20 mmol, 2.0 equiv) were added. Thereupon, a solution of TFA in dry CH2Cl2 (22.8 mg/mL, 100 μL, 0.040 mmol, 0.40 equiv) was added, and the vial was purged with argon and stirred at rt for 3 h. The crude product was directly column chromatographed (silica gel) to afford enantioenriched cycloadducts 4a-i.

Note: Different procedures were used for cyclopentadiene 2k, indene 2 l and 4-methoxystyrene 2 m.

General procedure for the enantioselective [4 + 2] cycloaddition with linear aromatic dienes

A 2 mL flame dried vial was charged with a solution of 2-formyl-1,3-cyclohexadiene 1 in benzene (20% v/v, 54 mg, 0.10 mmol, 1.0 equiv), 100 μL dry degassed CH2Cl2 was added and stirred for 1 min. Catalyst 3a (4.3 mg, 0.020 mmol, 0.20 equiv) and linear diene 5a-f (0.20 to 0.40 mmol, 2.0 to 4.0 equiv) were added. Thereupon, a solution of TCA in dry CH2Cl2 (32.7 mg/mL, 100 μL, 0.040 mmol, 0.40 equiv) was prepared, the vial was purged with argon, and stirred at rt for 16 h. Then, sodium borohydride (2 equiv) was added alongside 200 μL of MeOH and stirred for 1 h at rt. The crude product was directly column chromatographed (silica gel) to afford enantioenriched cycloadducts 6a-f.

Note: For reactions performed at 1 M CH2Cl2 (6b,f), pure TCA was added instead of the solution in dry CH2Cl2.

General procedure for the enantioselective [4 + 2] cycloaddition with linear aliphatic dienes

A 2 mL flame dried vial was charged with a solution of 2-formyl-1,3-cyclohexadiene 1 in benzene (20% v/v, 54 mg, 0.10 mmol, 1.0 equiv), catalyst 3a (4.3 mg, 0.020 mmol, 0.20 equiv), and linear diene 5g-j (0.30 to 0.40 mmol, 3.0 to 4.0 equiv). Thereupon, TCA (4.10 μL, 0.040 mmol, 0.40 equiv) was added; the vial was purged with argon and stirred at the corresponding temperature. Upon full conversion of 1, the crude product was directly column chromatographed (silica gel) to afford enantioenriched cycloadducts 6g-j.

Supplementary information

Source data

Source Data (101.3KB, xlsx)

Acknowledgements

We thank René S. Bitsch for helping with X-ray analysis. Thanks are expressed to Jesper Bendix, Copenhagen University for assistance with X-ray and ECD analysis.

Author contributions

K.A.J. and K.N.H. supervised the project. M.K. conceived the idea. E.B.O., I.R.K., M.K., M.C., and P.W. designed and performed the experiments. E.B.O., I.R.K., M.K., M.C., and P.W. analyzed the data. C.Z. and P.P.C. performed DFT calculations. K.A.J., K.N.H., E.B.O., C.Z., and I.R.K. wrote the manuscript. All authors discussed the results and revised the manuscript.

Peer review

Peer review information

Nature Communications thanks Shiqi Jia, Tingshun Zhu the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Funding

KAJ thanks Villum Investigator grant (no. 25867), Novo Nordisk Foundation (NNF 21OC0072047), FNU (2032-00165B) and Aarhus University. KNH thanks the U.S. National Science Foundation (CHE 2153972) for financial support of this research. We thank René S. Bitsch for helping with X-ray analysis.

Data availability

Crystallographic data for the structures reported in this article have been deposited at the Cambridge Crystallographic Data Center (CCDC), under deposition numbers CCDC 2440544 (4 d). These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Experimental procedures, characterization of new compounds, and all other data supporting the findings are available in the Supplementary Information. All data are available from the corresponding author upon request. Source data of DFT studies are provided with this paper. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Erlaitz Basabe Obregón, Ida Rygaard Kocemba, Chi Zhang.

Contributor Information

K. N. Houk, Email: houk@g.ucla.edu

Karl Anker Jørgensen, Email: kaj@chem.au.dk.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76222-y.

References

  • 1.Diels, O. & Alder, K. Synthesen in der hydroaromatischen Reihe. Justus Liebigs Ann. Chem.460, 98–122 (1928). [Google Scholar]
  • 2.Nicolaou, K. C., Snyder, S. A., Montagnon, T. & Vassilikogiannakis, G. The Diels–Alder reaction in total synthesis. Angew. Chem. Int. Ed.41, 1668–1698 (2002). [DOI] [PubMed] [Google Scholar]
  • 3.Takao, K. I., Munakata, R. & Tadano, K. I. Recent advances in natural product synthesis by using intramolecular Diels−Alder reactions. Chem. Rev.105, 4779–4807 (2005). [DOI] [PubMed] [Google Scholar]
  • 4.Funel, J. A. & Abele, S. Industrial applications of the Diels–Alder reaction. Angew. Chem. Int. Ed.52, 3822–3863 (2013). [DOI] [PubMed] [Google Scholar]
  • 5.Hoffmann, R. & Woodward, R. B. Selection rules for concerted cycloaddition reactions. J. Am. Chem. Soc.87, 2046–2048 (1965). [Google Scholar]
  • 6.Houk, K. N., Liu, F., Yang, Z. & Seeman, J. I. Evolution of the Diels–Alder reaction mechanism since the 1930s: Woodward, Houk with Woodward, and the influence of computational chemistry on understanding cycloadditions. Angew. Chem. Int. Ed.60, 12660–12681 (2021). [DOI] [PubMed] [Google Scholar]
  • 7.Bachmann, W. E. & Deno, N. C. The Diels-Alder reaction of 1-vinylnaphthalene with α,ß- and α,β,γ,δ-unsaturated acids and derivatives. J. Am. Chem. Soc.71, 3062–3072 (1949). [Google Scholar]
  • 8.Jiang, X. & Wang, R. Recent developments in catalytic asymmetric inverse-electron-demand Diels–Alder reaction. Chem. Rev.113, 5515–5546 (2013). [DOI] [PubMed] [Google Scholar]
  • 9.Png, Z. M., Zeng, H., Ye, Q. & Xu, J. Inverse-electron-demand Diels-Alder reactions: principles and applications. Chem. Asian J.12, 2142–2159 (2017). [DOI] [PubMed] [Google Scholar]
  • 10.Moyano, A. & Rios, R. Asymmetric organocatalytic cyclization and cycloaddition reactions. Chem. Rev.111, 4703–4832 (2011). [DOI] [PubMed] [Google Scholar]
  • 11.Pellissier, H. Asymmetric organocatalytic cycloadditions. Tetrahedron68, 2197–2232 (2012). [Google Scholar]
  • 12.Klier, L., Tur, F., Poulsen, P. H. & Jørgensen, K. A. Asymmetric cycloaddition reactions catalysed by diarylprolinol silyl ethers. Chem. Soc. Rev.46, 1080–1102 (2017). [DOI] [PubMed] [Google Scholar]
  • 13.Otevrel, J., Eugui, M., Ričko, S. & Jørgensen, K. A. Enantioselective organocatalytic cycloadditions for the synthesis of medium-sized rings. Nat. Synth.2, 1142–1158 (2023). [Google Scholar]
  • 14.Juhl, K. & Jørgensen, K. A. The first organocatalytic enantioselective inverse-electron-demand hetero-Diels–Alder reaction. Angew. Chem. Int. Ed.42, 1498–1501 (2003). [DOI] [PubMed] [Google Scholar]
  • 15.Han, B., Li, J.-L., Ma, C., Zhang, S.-J. & Chen, Y.-C. Organocatalytic asymmetric inverse-electron-demand aza-Diels–Alder reaction of N-sulfonyl-1-aza-1,3-butadienes and aldehydes. Angew. Chem. Int. Ed.47, 9971–9974 (2008). [DOI] [PubMed] [Google Scholar]
  • 16.Jiang, X., Fu, D., Shi, X., Wanga, S. & Wang, R. PPh3-Catalyzed synthesis of dicyano-2-methylenebut-3-enoates as efficient dienes in catalytic asymmetric inverse-electron-demand Diels–Alder reaction. Chem. Commun.47, 8289–8291 (2011). [DOI] [PubMed] [Google Scholar]
  • 17.Bertelsen, S., Marigo, M., Brandes, S., Dinér, P. & Jørgensen, K. A. Dienamine catalysis: organocatalytic asymmetric γ-amination of α,β−unsaturated aldehydes. J. Am. Chem. Soc.128, 12973–12980 (2006). [DOI] [PubMed] [Google Scholar]
  • 18.Hong, B.-C. et al. Organocatalytic asymmetric Robinson annulation of α,β-unsaturated aldehydes: applications to the total synthesis of (+)-palitantin. J. Org. Chem.72, 8459–8471 (2007). [DOI] [PubMed] [Google Scholar]
  • 19.Li, J.-L., Kang, T.-R., Zhou, S.-L., Li, R. & Chen, Y.-C. Organocatalytic asymmetric inverse-electron-demand Diels–Alder reaction of electron-deficient dienes and crotonaldehyde. Angew. Chem. Int. Ed.49, 6418–6420 (2010). [DOI] [PubMed] [Google Scholar]
  • 20.Ramachary, D. B., Chowdari, N. S. & Barbas, C. F. Organocatalytic asymmetric domino Knoevenagel/Diels–Alder reactions: a bioorganic approach to the diastereospecific and enantioselective construction of highly substituted spiro[5,5]undecane−1,5,9-triones. Angew. Chem. Int. Ed.42, 4233–4237 (2003). [DOI] [PubMed] [Google Scholar]
  • 21.Thayumanavan, R., Dhevalapally, B., Sakthivel, K., Tanaka, F. & Barbas, C. F. Amine-catalyzed direct Diels–Alder reactions of α,β-unsaturated ketones with nitro olefins. Tetrahedron Lett.43, 3817–3820 (2002). [Google Scholar]
  • 22.Ramachary, D. B., Chowdari, N. S. & Barbas, C. F. Amine-catalyzed direct self Diels–Alder reactions of α,β-unsaturated ketones in water: Synthesis of pro-chiral cyclohexanones. Tetrahedron Lett.43, 6743–6746 (2002). [Google Scholar]
  • 23.Mose, R. et al. Organocatalytic stereoselective [8+2] and [6+4] cycloadditions. Nat. Chem.9, 487–492 (2017). [DOI] [PubMed] [Google Scholar]
  • 24.Jia, Z.-J. et al. Trienamines in asymmetric organocatalysis: Diels-Alder and tandem reactions. J. Am. Chem. Soc.133, 5053–5061 (2011). [DOI] [PubMed] [Google Scholar]
  • 25.Zhou, Z. et al. Switchable regioselectivity in amine-catalysed asymmetric cycloadditions. Nat. Chem.9, 590–594 (2017). [DOI] [PubMed] [Google Scholar]
  • 26.Donslund, B. S. et al. Organocatalytic enantioselective higher-order cycloadditions of in situ generated amino isobenzofulvenes. Angew. Chem. Int. Ed.57, 1246–1250 (2018). [DOI] [PubMed] [Google Scholar]
  • 27.Ahrendt, K. A., Borths, C. J. & MacMillan, D. W. C. New strategies for organic catalysis: the first highly enantioselective organocatalytic Diels−Alder reaction. J. Am. Chem. Soc.122, 4243–4244 (2000). [Google Scholar]
  • 28.Jen, W. S., Wiener, J. J. M. & MacMillan, D. W. C. New strategies for organic catalysis: the first enantioselective organocatalytic 1,3-dipolar cycloaddition. J. Am. Chem. Soc.122, 9874–9875 (2000). [DOI] [PubMed] [Google Scholar]
  • 29.Ishihara, K. & Nakano, K. Design of an organocatalyst for the enantioselective diels−alder reaction with α-acyloxyacroleins. J. Am. Chem. Soc.127, 10504–10505 (2005). [DOI] [PubMed] [Google Scholar]
  • 30.Sakakura, A., Suzuki, K., Nakano, K. & Ishihara, K. Chiral 1,1’-binaphthyl-2,2’-diammonium salt catalysts for the enantioselective Diels−Alder reaction with α-acyloxyacroleins. Org. Lett.8, 2229–2232 (2006). [DOI] [PubMed] [Google Scholar]
  • 31.Frankowski, S. & Skrzyńska, A. & Albrecht, Ł. Inverting the reactivity of troponoid systems in enantioselective higher-order cycloaddition. Chem. Commun.55, 11675–11678 (2019). [DOI] [PubMed] [Google Scholar]
  • 32.Poulsen, P. H., Vergura, S., Monleón, A., Jørgensen, D. K. B. & Jørgensen, K. A. Controlling asymmetric remote and cascade 1,3-dipolar cycloaddition reactions by organocatalysis. J. Am. Chem. Soc.138, 6412–6415 (2016). [DOI] [PubMed] [Google Scholar]
  • 33.Gao, Y., Song, X., Yan, R.-J., Du, W. & Chen, Y.-C. Asymmetric β,γ′-regioselective [4+3] and [4+2] annulations of α-vinylenals via cascade iminium ion-dienamine catalysis. Org. Biomol. Chem.19, 151–155 (2021). [DOI] [PubMed] [Google Scholar]
  • 34.Palazzo, T. A., Mose, R. & Jørgensen, K. A. Cycloaddition reactions: why is it so challenging to move from six to ten electrons? Angew. Chem. Int. Ed.56, 10033–10038 (2017). [DOI] [PubMed] [Google Scholar]
  • 35.Martin-Somer, A., Xue, X.-S., Jamieson, C. S., Zou, Y. & Houk, K. N. Computational design of a tetrapericyclic cycloaddition and the nature of potential energy surfaces with multiple bifurcations. J. Am. Chem. Soc.145, 4221–4230 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Yu, P. et al. Mechanisms and origins of periselectivity of the ambimodal [6+4] cycloadditions of tropone to dimethylfulvene. J. Am. Chem. Soc.139, 8251–8258 (2017). [DOI] [PubMed] [Google Scholar]
  • 37.McLeod, D. et al. Expanding the frontiers of higher-order cycloadditions. Acc. Chem. Res.52, 3488–3501 (2019). [DOI] [PubMed] [Google Scholar]
  • 38.Hammer, N. et al. An experimental stereoselective photochemical [1s,3s]-sigmatropic silyl shift and the existence of Silyl/Allyl conical intersections. J. Am. Chem. Soc.142, 6030–6035 (2020). [DOI] [PubMed] [Google Scholar]
  • 39.Kaasik, M., Chen, P. P., Ričko, S., Jørgensen, K. A. & Houk, K. N. Asymmetric [4+2], [6+2], and [6+4] cycloadditions of isomeric formyl cycloheptatrienes catalyzed by a chiral diamine catalyst. J. Am. Chem. Soc. 145, 23874–23890 (2023). [DOI] [PubMed]
  • 40.Caramella, P., Quadrelli, P. & Toma, L. An unexpected bispericyclic transition structure leading to [4+2] and [2+4] cycloadducts in the endo dimerization of cyclopentadiene. J. Am. Chem. Soc.124, 1130–1131 (2002). [DOI] [PubMed] [Google Scholar]
  • 41.Xue, X.-S. et al. Ambimodal trispericyclic transition state and dynamic control of periselectivity. J. Am. Chem. Soc.141, 1217–1221 (2019). [DOI] [PubMed] [Google Scholar]
  • 42.Zhou, Q., Thøgersen, M. K., Rezayee, N. M., Jørgensen, K. A. & Houk, K. N. Ambimodal bispericyclic [6+4]/[4+6] transition state competes with diradical pathways in the cycloheptatriene dimerization: dynamics and experimental characterization of thermal dimers. J. Am. Chem. Soc.144, 22251–22261 (2022). [DOI] [PubMed] [Google Scholar]
  • 43.Li, G., Kusari, S. & Spiteller, M. Natural products containing the “decalin” motif in microorganisms. Nat. Prod. Rep.31, 1175–1201 (2014). [DOI] [PubMed] [Google Scholar]
  • 44.Ebner, C. & Carreira, E. M. Pentafulvene for the synthesis of complex natural products: total syntheses of (±)-pallambins A and B. Angew. Chem. Int. Ed.54, 11227–11230 (2015). [DOI] [PubMed] [Google Scholar]
  • 45.Bickelhaupt, F. M. & Houk, K. N. Analyzing reaction rates with the distortion/interaction-activation strain model. Angew. Chem. Int. Ed.56, 10070–10086 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Crystallographic data for the structures reported in this article have been deposited at the Cambridge Crystallographic Data Center (CCDC), under deposition numbers CCDC 2440544 (4 d). These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Experimental procedures, characterization of new compounds, and all other data supporting the findings are available in the Supplementary Information. All data are available from the corresponding author upon request. Source data of DFT studies are provided with this paper. Source data are provided with this paper.

Articles from Nature Communications are provided here courtesy of Nature Publishing Group

RESOURCES