Annuloselectivity and stereochemistry in the sulfa-Staudinger cycloadditions of cyclic imines

Abstract

The annuloselectivity and the stereochemistry in the sulfa-Staudinger cycloadditions of cyclic imines are controlled by the ring size of the cyclic imines. Intrinsically, it is the steric hindrance of cyclic imines that controls the annuloselectivity, as well as the stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annulations. A stepwise [4 + 2] annulation mechanism, which incorporates an intermolecular addition, C=S bond isomerization, and subsequently intramolecular addition, is proposed to explain the different stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annulations. The intermolecular addition is regarded as the key stereo-determining step. Firstly, the C3 and C5 stereochemistry is kinetically controlled by the endo or exo addition of imines to the key zwitterionic 2,3-thiaza-1,4-butadiene-type intermediates, and then the C5 and C6 stereochemistry is thermodynamically controlled by the isomerization of the C=S bond in the zwitterionic endo- or exo-adducts generated from the previous step. The intramolecular addition does not affect the stereochemical outcomes of the [2^s + 2ⁱ + 2ⁱ] annulations.

---

Introduction

The sulfa-Staudinger cycloadditions, namely the [2^s + 2ⁱ] annulations of sulfenes (or their equivalents) with imines,^1,2 represent a classic method to construct the β-sultam backbone.³ β-sultams are synthetically important⁴ and biologically active compounds.⁵ Recently, the sulfa-Staudinger cycloadditions have received much attention in both the synthetic and medicinal community. However, the reactions between sufenes and imines do not always give the [2^s + 2ⁱ] annuladducts β-sultams, in some cases the [2^s + 2ⁱ + 2ⁱ] annuladducts are forged.^3b,6 Both of the two types of annuladducts show a wide spectrum of bioactivities.⁷ Thus, the annuloselectivity, namely [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annulation selectivity, comes to be one of the most concerned issues in this field.

Our recent studies revealed that the annuloselectivity was controlled by the α-substituents of sulfonyl chlorides and the nucleophilicity of imines.^2b In the reactions of sulfonyl chlorides with strongly electron-withdrawing α-substituents such as ethoxycarbonylmethanesulfonyl chloride, depending on the nucleophilicity of imines, there exist three kinds of annuloselective results: (1) the imines with larger N-substituents than methyl afford exclusively [2^s + 2ⁱ] annuladducts; (2) N-methyl imines give both [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annuladducts; (3) cyclic imines provide exclusively [2^s + 2ⁱ + 2ⁱ] annuladducts.

In these studies, our attention was mainly focused on the linear imines (Fig. 1a).^2b We only reported one example on six-membered cyclic imine 3,4-dihydroisoquinoline, of which the [2^s + 2ⁱ + 2ⁱ] annuladducts were a pair of diastereomers in (3,5)-cis-(5,6)-cis- and (3,5)-cis-(5,6)-trans-configurations. However, the annuloselectivities and the stereochemistry involved in the sulfa-Staudinger cycloadditions of other ring-size-different cyclic imines still remain mysterious. As our continuing interests in sulfa-Staudinger cycloadditions, we studied this problem, and found that the annuloselectivity and the stereochemistry were also controlled by the ring size of cyclic imines (Fig. 1b). Herein, we report our results, hoping they will not only complete our recently proposed annuloselective empirical rule, but also provide practical guidelines to predict the diverse products and the stereochemistry of the [2^s + 2ⁱ + 2ⁱ] annuladducts in the sulfa-Staudinger cycloadditions of cyclic imines.

Fig. 1 Factors that control the annuloselectivity and stereochemistry in the sulfa-Staudinger cycloaddition.

Results and discussion

Selection of the cyclic imine probes

To begin the investigation, we need to select some representative cyclic imines to probe the annuloselectivity and stereochemistry in their sulfa-Staudinger cycloadditions. Since the five-, six-, and seven-membered cyclic imines are most accessible, and possess synthetically significant applications, we decided to select the probes from these three kinds of cyclic imines. Based on our previous work, the cyclic imines selected should satisfy the following two requirements: (1) the N-terminal should be substituted by an alkyl group; (2) the C-terminal should be mono-substituted.² Thus, the selected cyclic imines will exhibit small steric hindrance and strong basicity and nucleophilicity,⁸ which are required for the occurrence of the reactions between cyclic imines and sulfonyl chlorides.

Actually, during the past six years, we have been studying the reactions between sulfonyl chlorides and many kinds of cyclic imines.^6,9,10 Unfortunately, no five-membered cyclic imines can undergo either [2^s + 2ⁱ] or [2^s + 2ⁱ + 2ⁱ] annulation.⁹ Out of our experience, the following cyclic imines 1 and 2 in Fig. 2 are chosen as the probes in the current studies.¹⁰ In fact, some reactions between cyclic imines 1 and 2 with certain sulfonyl chlorides were dispersed in our previous publications,^2b,6 concerning other issues in the sulfa-Staudinger cycloaddition. Herein, with the freshly conducted reactions and the collected previously dispersed ones,^2b,6 the annuloselectivity and the stereochemistry in the sulfa-Staudinger cycloadditions of cyclic imines could be disclosed.

Fig. 2 Selected cyclic imines as probes.

Experimental studies

Since the annuloselectivity are closely associated with the α-substituent effect of sulfonyl chlorides,^2b the above two cyclic imines 1 and 2 were reacted with four representative sulfonyl chlorides, namely ethanesulfonyl chloride (3), phenylmethanesulfonyl chloride (4), ethoxycarbonylmethanesulfonyl chloride (5), and (methanesulfonyl)methanesulfonyl chloride (6), substituted by weakly electron-donating methyl group, weakly electron-withdrawing phenyl group, strongly electron-withdrawing ethoxycarbonyl and methanesulfonyl groups, respectively. The designed reactions are between cyclic imines (1 and 2) and sulfonyl chlorides (3–6), and the results are listed in Schemes 1 and 2.

Scheme 1 Reactions between the six-membered cyclic imine 1 and sulfonyl chlorides 3–6.

Scheme 2 Reactions between the seven-membered cyclic imine 2 and sulfonyl chlorides 3–6.

In Scheme 1, the reaction of 1 with 3 gave a chloride-addition product 7 in 16% yield from the key intermediate N-sulfonyl iminium chloride (Scheme 1, eqn (1)).⁶ To our surprise, in our more careful studies, more products have been detected in the reaction of imine 1 with sulfonyl chloride 4, that is, three [2^s + 2ⁱ + 2ⁱ] annuladducts 8, 9, 10, one [2^s + 2ⁱ] annuladduct 11, and one water-addition product 12 (generated during workup) in 5%, 5%, 6.5%, 16%, and 11% yields, respectively (Scheme 1, eqn (2)). Among those products, the structures of 8, 10, and 12 were well identified in our previous publication.⁶ In current reinvestigation on the reaction in eqn (2), [2^s + 2ⁱ] annuladduct 11 can not be successfully isolated from the mixture with two [2^s + 2ⁱ + 2ⁱ] annuladducts 8 and 9. However, the ¹H NMR spectrum of the mixture shows a couple of double-split peaks at δ 5.31 and 4.82 (J = 4.5 Hz, see ESI†). Subsequent HRMS determination also presents a peak at 286.0893 (calculated at 286.0896 for product 11). These convincing data successfully demonstrate the generation of [2^s + 2ⁱ] annuladduct 11. In addition, the relative structures of [2^s + 2ⁱ + 2ⁱ] annuladducts 8 and 9 were also assigned by means of the NOE analyses. In eqn (2), the [2^s + 2ⁱ + 2ⁱ] annulation dominates. When cyclic imine 1 was reacted with sulfonyl chloride 5, two diastereoisomeric [2^s + 2ⁱ + 2ⁱ] annuladducts 13 and 14 were formed in 20% and 26% yields, respectively, and their stereostructures were clearly established in our previous report.^2b The similar [2^s + 2ⁱ + 2ⁱ] annuladducts 15 and 16 were also accessible in the reactions of imine 1 with sulfonyl chloride 6, but the ratio and total yield were 34 : 66 and 30%, respectively. The lower total yield than that in eqn (3) was probably caused by the large steric hindrance of the α-sulfonyl group in 6.

The seven-membered cyclic imine 2 reacted with ethanesulfonyl chloride (3) only to give a hydrolyzed product 17a in 19% yield (Scheme 2, eqn (5)).^2b However, phenylmethanesulfonyl chloride (4) and imine 2 reacted smoothly, giving both [2^s + 2ⁱ] annuladduct tricyclic β-sultam 19 and hydrolyzed product 17b in 16% and 20% yields, respectively (Scheme 2, eqn (6)).^2b,11 Our previous work showed that a variety of arylmethanesulfonyl chlorides can undergo the above two types of reactions with 2, as key evidence, disclosing the reasonable mechanism for [2^s + 2ⁱ] annulation in the sulfa-Staudinger cycloaddition.^2b In the subsequent studies by reacting imine 2 with sulfonyl chloride 5, two [2^s + 2ⁱ + 2ⁱ] annuladducts 20 and 21 were isolated as a pair of diastereoisomers in 5% and 20% yields, respectively. The NOE analysis of the C3 and C5 protons, together with the coupling constants between C5 and C6 protons of the newly-forged six-membered rings, indicated that the stereochemical configurations of 20 and 21 were (3,5)-cis-(5,6)-trans and (3,5)-trans-(5,6)-cis, respectively (Scheme 2, eqn (7)). In contrast with the results in eqn (4), seven-membered cyclic imine 2 did not match well with the (methanesulfonyl)methanesulfonyl chloride (6), with neither [2^s + 2ⁱ] nor [2^s + 2ⁱ + 2ⁱ] annuladduct formed, possibly because of the large steric hindrance of the α-methanesulfonyl group (Scheme 2, eqn (8)).

The results in the reactions of sulfonyl chlorides and representative cyclic imines are summarized in Table 1. Comparing the above results leads to an insight into the annuloselectivity and stereochemistry in the reactions of sulfonyl chlorides with cyclic imines: ethanesulfonyl chloride (3) cannot produce any annulated product with five- to seven-membered cyclic imines. For cyclic imines, (1) the five-membered cyclic imines do not favour either [2^s + 2ⁱ] or [2^s + 2ⁱ + 2ⁱ] annulations (see ref. 9); (2) the six-membered cyclic imines favour the [2^s + 2ⁱ + 2ⁱ] annulations (Scheme 2, eqn (5) and (6)); (3) the seven-membered cyclic imines favour both the [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annulations, depending on the α-substituent effect of the sulfonyl chlorides employed (Scheme 3, eqn (8) and (9)). Another important observation is that the stereochemistry of the [2^s + 2ⁱ + 2ⁱ] annuladducts from seven-membered cyclic imines is quite different from that of the [2^s + 2ⁱ + 2ⁱ] annuladducts from the six-membered cyclic imines (Scheme 2, eqn (5) and (6) vs. Scheme 3, eqn (9)). The above annuloselective and stereoselective issues are quite mechanistically interesting but not yet systematically explored.

Table 1 Reactions of representative sulfonyl chlorides and cyclic imines

Entry	Sulfonyl chloride	Cyclic imine	“Hydrolyzed” product yield (%)	[2^s + 2^i] product yield (%)	[2^s + 2^i + 2^i] products
					(3,5)-cis-(5,6)-trans Yield (%)	(3,5)-cis-(5,6)-cis Yield (%)	(3,5)-trans-(5,6)-trans Yield (%)	(3,5)-trans-(5,6)-cis Yield (%)
a The yield of the chloride-addition product 7.b The corresponding products were not obtained.
1	3		7 (16)^a	—^b	—	—	—	—
2	4		12 (11)	11 (6.5)	8 (5)	9 (5)	10 (16)	—
3	5		—	—	13 (20)	14 (26)	—	—
4	6		—	—	15 (10)	16 (20)	—	—
5	3		18a (19)	—	—	—	—	—
6	4		18b (20)	19 (16)	—	—	—	—
7	5		—	—	20 (5)	—	—	21 (20)
8	6		—	—	—	—	—	—

---

Scheme 3 Key intermediates in the [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annulations.

Intermediates in the [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annulations

According to our recently established mechanisms for the sulfa-Staudinger cycloaddition,^2a the general routes toward the [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annuladducts were proposed. As outlined in Scheme 3, there exist two paths to the α-anionic N-sulfonyl iminium intermediates D (path A and path B), depending on the acidity of the α-protons in sulfonyl chlorides A, as described in our recent publication.^2c These two paths do not affect the subsequent annuloselectivity or stereochemistry. The key 2,3-thiaza-buta-1,3-diene-type intermediates E,¹² resonating from D, constitutes the only active intermediates that undergo either intramolecular conrotation to form [2^s + 2ⁱ] annuladducts trans-β-sultams F, or intermolecular [4 + 2] annulation to form [2^s + 2ⁱ + 2ⁱ] annuladducts I. It is noteworthy that intermediates G do not exist in the reactions. If they exist, cis-bicyclic β-sultams H would be generated. However, in our experiments no compound of such a type was observed, eliminating the existence of G. Probably it is the conjugated 4π system that prevents the isomerization of E to G over the C=S bond. The unambiguty of intermediates E has been proved by the formation of the trans-tricyclic sultams 11 and 19 in the reactions listed in Scheme 1, eqn (2), and Scheme 2, eqn (6), respectively.

Mechanistic insights into the stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annulations of different cyclic imines

In our previous work, we uncovered that the nature of the [2^s + 2ⁱ + 2ⁱ] annulations is a stepwise [4 + 2] annulation between 2,3-thiazabuta-1,3-diene-type intermediates E and a second molecule of strongly nucleophilic imines.^2b Since all the products exhibit (3,5)-cis-configurations, we simply put it that the stereochemistry of the C3- and C5-positions of the [2^s + 2ⁱ + 2ⁱ] annuladducts was generated as a result of the chair transition-state conformation, with the C3- and C5-protons on the axial positions to decrease the steric congestion of the intermediates.^2b However, in current studies, the (3,5)-trans-products 10, 20, and 21 were obtained, revealing that the stereochemistry of the [2^s + 2ⁱ + 2ⁱ] annuladducts is not as simple as we previously observed. Therefore, a new model is in demand to explain and further predict the complex stereochemical outcomes in the [2^s + 2ⁱ + 2ⁱ] annulations.

As delineated in Scheme 4, the [4 + 2] annulation constitutes three steps, that is, (1) an intermolecular Mannich-like addition between imines and 2,3-thiazabuta-1,3-diene-type intermediates E from endo or exo side to afford zwitterionic adducts E-1 or E-2, respectively, (2) the C=S bond isomerization of the zwitterionic intermediates E-1 and E-2 generated from the above step to give intermediates E-3 and E-4, respectively, and (3) an intramolecular nucleophilic cyclization inside the four zwitterionic intermediates to afford the corresponding [2^s + 2ⁱ + 2ⁱ] annuladducts I-1, I-2, I-3, or I-4, respectively. The first intermolecular addition, is regarded as a rate-determining step, because it competes with the [2^s + 2ⁱ] annulation. It is crucial to the occurrence of the [2^s + 2ⁱ + 2ⁱ] annulations. Herein, as a stereo-determining step, it also plays an extremely important role in deciding the C3 and C5 stereochemistry of the [2^s + 2ⁱ + 2ⁱ] annulations. The imines can initiate the intermolecular addition to the 2,3-thiazabuta-1,3-diene-type intermediates E from either exo or endo direction, consequently leading to (3,5)-trans- or (3,5)-cis-products, respectively. The C=S bond isomerization and intramolecular cyclization are in competition, which may be affected by the ring size and/or steric hindrance of imines. The occurrence of the intermolecular addition and isomerization steps is controlled by the ring geometry of the cyclic imines, consequently deciding the final stereochemistry of the [2^s + 2ⁱ + 2ⁱ] annuladducts.

Scheme 4 Proposed stepwise mechanism for the [2^s + 2ⁱ + 2ⁱ] annulations.

As depicted in Scheme 5, the addition of imine 1 to the corresponding 2,3-thiazabuta-1,3-diene intermediates J from the endo side (endo addition) gives rise to intermediates K, which directly undergo intramolecular cyclization to form (3,5)-cis-(5,6)-trans-8, 13, and 15, or isomerize over their C=S bond followed by intramolecular addition to afford (3,5)-cis-(5,6)-cis-9, 14, and 16 via intermediates L. The exo addition also occurs, delivering intermediates M, which go through a sequence of complete isomerization and intramolecular addition inside intermediates N to evolve into (3,5)-trans-(5,6)-trans-10.

Scheme 5 Rationalization of the stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annuladducts from 3,4-dihydroisoquinoline (1).

The stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annulations of the seven-membered imine 2 is also rationalized in Scheme 6. The intermolecular addition of 2 occurs from the endo side of intermediates O, giving intermediates P, of which the direct intramolecular addition gives (3,5)-trans-(5,6)-cis-20. Similarly, the exo addition and subsequent intramolecular cyclization afford 3,5-trans-5,6-trans-21, through intermediates Q. However, the isomerization of the C=S bond of intermediates P and Q did not occur, mainly because of the steric effect of the congested cyclic iminium moieties. In addition, the large steric hindrance of 2 and O makes the intermolecular addition between 2 and O very sensitive to the steric hindrance of the α-substituent of sulfonyl chlorides. For example, in eqn (8), the sterically bulky α-methanesulfonyl group imposed disastrous effect, with no [2^s + 2ⁱ] or [2^s + 2ⁱ + 2ⁱ] annuladduct formed.

Scheme 6 Rationalization of the stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annuladducts from seven-membered cyclic imine 2.

The above stereochemical elucidation discloses that the stereochemistry between the C3 and C5 stereocenters is kinetically controlled by the endo or exo addition of imines, while that between the C5 and C6 stereocenters is thermodynamically controlled by the isomerization of the C=S bond in the zwitterionic endo- or exo-adducts K, M, P and Q generated from the previous step. The intramolecular cyclization does not affect the stereochemical outcomes of the [2^s + 2ⁱ + 2ⁱ] annulations. It is also interesting that the C=S bond isomerization step only occurs in the reactions of the six-membered imine 1, possibly controlled by the ring size and steric hindrance of the cyclic imines.

Mechanistic insights into the ring-size-controlled annuloselectivity

When reacted with phenylmethanesulfonyl chloride (2), the six-membered cyclic imine 1 is prone to the [2^s + 2ⁱ + 2ⁱ] annulations, while the seven-membered analogue 2 [2^s + 2ⁱ] annulation. The annuloselectivity is mainly governed by the steric hindrance of the imines (1 and 2) and intermediates (J and O). As shown in Fig. 3, the sterically smaller 1 and J favour the intermolecular addition, thus [2^s + 2ⁱ + 2ⁱ] annulations are preferred to the [2^s + 2ⁱ] annulation. In contrast, the sterically larger 2 and O prevent the intermolecular addition, and consequently, [2^s + 2ⁱ + 2ⁱ] annulations do not occur. On one hand, the intermediate O undergoes conrotatory ring closure to give [2^s + 2ⁱ] annuladduct 19 in low yield;² on the other hand, it hydrolyzes to give aldehyde 17b.

Fig. 3 Steric hindrance of imines 1 and 2 and intermediates.

In the reactions with sulfonyl chloride 5, both of the imines exclusively undergo [2^s + 2ⁱ + 2ⁱ] annulations. This annulospecificity is controlled by the α-substituent effect of sulfonyl chlorides, as pointed out in our previous work.^2b Detailedly, when R¹ is a strongly electron-withdrawing group, the direct conrotation of intermediates J and O is drastically decelerated. As a result, the strongly nucleophilic imines 1 and 2 would have enough probability to initiate the intermolecular addition, if sterically permitted, leading to the exclusive [2^s + 2ⁱ + 2ⁱ] annuladducts 13, 14, 15, 16, 20, and 21 in eqn (3), (4), and (7).

Kinetic proposals for the stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annulations of linear imines

In our previous studies, the stereochemistry of the [2^s + 2ⁱ + 2ⁱ] annuladducts of N-methyl linear imines 24 was attributed to the thermodynamical properties of the intermediates S.^2b Herein, from a kinetic perspective, the stereochemistry is more easily understood. As delineated in Scheme 7, the endo addition of 24 to R directly leads to S, and subsequently the intramolecular cyclization of S successfully explains the stereochemistry of (3,5)-cis-(5,6)-trans-25.

Scheme 7 Rationalization of the stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annuladducts from N-methyl linear imines.

Conclusions

By using representative six- and seven-membered cyclic imines 1 and 2 as probes, the annuloselectivity and stereochemistry in the sulfa-Staudinger cycloadditions of cyclic imines were studied. The results reveal that the annuloselectivity and stereochemistry are closely associated with the ring size of the cyclic imines. When reacted with sulfonyl chlorides with weakly electron-withdrawing substituents, the six-membered cyclic imines afford both [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annuladducts, with the latter dominating, while the seven-membered cyclic imines give only [2^s + 2ⁱ] annuladducts. When reacted with sulfonyl chlorides with strongly electron-withdrawing substituents, regardless of the ring size, all the cyclic imines undergo [2^s + 2ⁱ + 2ⁱ] annulations. It is the steric hindrance of the cyclic imines that controls not only the [2^s + 2ⁱ] and [2^s + 2ⁱ + 2ⁱ] annuloselectivity in the reactions of weakly electron-withdrawing-substituted sulfonyl chlorides, but also the stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annulations of strongly electron-withdrawing-substituted sulfonyl chlorides.

A stepwise [4 + 2] annulation mechanism, which incorporates an intermolecular addition, C=S bond isomerization, and an intramolecular cyclization, is proposed to explain the different stereochemistry in the [2^s + 2ⁱ + 2ⁱ] annulations of the six- and seven-membered imines. The intermolecular addition is regarded as the key stereo-determining step. Firstly, the C3 and C5 stereochemistry is kinetically controlled by the endo or exo intermolecular addition of imines to the key zwitterionic 2,3-thiaza-1,4-butadiene-type intermediates, and then the C5 and C6 stereochemistry is thermodynamically controlled by the isomerization of the C=S bond of the zwitterionic endo- or exo-adducts (for example, K, M, P, Q) generated from the previous step. The intramolecular cyclization does not affect the stereochemical outcomes of the [2^s + 2ⁱ + 2ⁱ] annulations. The six-membered imines undergo the [4 + 2] annulations predominantly or exclusively in an endo way with C=S bond isomerization, giving mainly (3,5)-cis-(5,6)-cis- and (3,5)-trans-(5,6)-cis-products, while the seven-membered imines in both endo and exo way without C=S bond isomerization, giving (3,5)-cis-(5,6)-trans- and (3,5)-trans-(5,6)-cis-products, respectively. In addition, the current stereochemical model successfully explains the stereochemistry of the [2^s + 2ⁱ + 2ⁱ] annulations of the linear imines.

Experimental section

General information

Tetrahydrofuran was refluxed over sodium with diphenyl ketone as indicator and freshly distilled prior to use. Melting points were obtained on a melting point apparatus and are uncorrected. ¹H and ¹³C NMR spectra were recorded on a Bruker 400 MHz spectrometer in CDCl₃ with TMS as an internal standard and the chemical shifts (δ) are reported in parts per million (ppm). The one-dimention selected NOE experiments were conducted on a Bruker 600 MHz spectrometer. The IR spectra (KBr pellets, v [cm⁻¹]) were taken on a FTIR spectrometer. HRMS measurements were carried out on an LC/MSD TOF mass spectrometer. TLC separations were performed on silica gel GF₂₅₄ plates, and the plates were visualized with UV light.

Sulfonyl chlorides 3 and 4 were prepared according to the methods in our previous reports,¹³ while 5 was prepared according to the method reported by Du Bois et al.,¹⁴ and 6 is commercially available. The cyclic imines 1 and 2 were prepared according to Cava's¹⁵ and our procedures,^2b respectively.

Typical procedure

The experiments were conducted by the following procedures.² To a solution of cyclic imine 1 (197 mg, 1.5 mmol) in 2 mL of dry tetrahydrofuran was dropwise added a solution of (methanesulfonyl)methanesulfonyl chloride 6 (96 mg, 0.5 mmol) in 0.5 mL of dry tetrahydrofuran. Upon addition, the mixture was allowed to stand at room temperature for another 24 h. Ether (10 mL) was added, and large amount of white solids precipitated. After washing with brine (10 mL), the organic phase was dried over Na₂SO₄, and concentrated at vacuum. The residue was purified by column chromatography on silica gel with petroleum ether and ethyl acetate as eluent.

In Scheme 1, 3,4-dihydroisoquinoline 1 (197 mg, 1.5 mmol) was reacted with 0.5 mmol of sulfonyl chloride 3, 5, or 6 following the typical procedure, but 3 mmol of 1 was used in the reaction of 1 mmol of phenylmethanesulfonyl chloride (4) in 5 mL of dry THF. In the reactions in Scheme 2, 3 mmol of cyclic imine 2, 1 mmol of sulfonyl chloride 3, 4, 5, or 6, and 5 mL of dry THF were used.

1-Chloro-2-(ethylsulfonyl)-1,2,3,4-tetrahydroisoquinoline (7)⁶

Colorless crystals, mp 170–172 °C. Yield 8 mg (16%). R_f = 0.5 (PE : EA = 1 : 1). ¹H NMR (400 MHz, CDCl₃) δ: 7.15–7.69 (m, 4H, ArH), 6.74 (s, 1H, CH), 3.89 (ddt, J = 5.6, 13.2, 1.2 Hz, 1H in CH₂), 3.56 (dt, J = 3.6, 12.8 Hz, 1H in CH₂), 3.04–3.10 (q, J = 7.2 Hz, 2H, CH₂), 2.98–3.16 (m, 1H in CH₂), 2.77 (ddd, J = 1.6, 3.6, 16.4 Hz, 1H in CH₂), 1.26 (t, J = 7.2 Hz, 3H, CH₃). ¹³C NMR (100 MHz, CDCl₃) δ: 135.6, 130.1, 129.6, 129.0, 128.3, 126.7, 85.9, 47.4, 38.8, 28.9, 7.7.

Compounds 8, 9, and 11 were obtained as an inseparable mixture (60 mg, about 16.5% total yield), and only trace amount of 8 was separated. But the characteristic ¹H NMR data of the three products, the NOE analyses, and the HRMS data of 10 clearly demonstrate the structures (for details, see ESI†). The ¹H NMR indicated that the ratio of 8, 9, and 11 was 1 : 1 : 1.3. The yields of 8, 9 and 10 were calculated to be 5%, 5%, and 6.5%, respectively.

rel(4bR,5R,13bR)-5-Phenyl-4b,5,9,13b,15,16-hexahydro-8H-[1,2,4]thiadiazino[3,2-a:5,4-a′]diisoquinoline 6,6-dioxide (8)⁶

Colorless crystals, mp 126–128 °C, 20 mg (5%). R_f = 0.8 (PE : EA = 1 : 2). ¹H NMR (400 MHz, CDCl₃) δ 7.27–6.17 (m, 13H, ArH), 6.07 (s, 1H), 5.12 (d, J = 10.4 Hz, 1H), 4.57 (d, J = 10.4 Hz, 1H), 3.78–3.72 (m, 1H), 3.65–3.60 (m, 1H), 3.26–3.19 (m, 1H), 3.06–2.96 (m, 2H), 2.86–2.76 (m, 2H), 2.70–2.64 (m, 1H). ¹³C NMR (100 MHz, CDCl₃) δ 135.7, 134.6, 132.5, 132.1, 129.8, 129.1, 129.0, 128.9, 128.7, 128.6, 128.0, 127.9, 127.3, 127.1, 124.5, 75.2, 64.9, 63.9, 39.6, 37.8, 29.3, 29.2.

rel(4bR,5S,13bR)-5-Phenyl-4b,5,9,13b,15,16-hexahydro-8H-[1,2,4]thiadiazino[3,2-a:5,4-a′]diisoquinoline 6,6-dioxide (9)

White solid (5% yield). ¹H NMR (600 MHz, CDCl₃) δ 5.69 (s, 1H in NCHN), 5.18 (d, J = 3.6 Hz, 1H in CHN), 4.84 (d, J = 3.6 Hz, 1H in CHS).

(4bR,5R,13bS)-5-Phenyl-4b,5,9,13b,15,16-hexahydro-8H-[1,2,4]thiadiazino[3,2-a:5,4-a′]diisoquinoline 6,6-dioxide (10)⁶

Colorless crystals, mp = 145–147 °C. Yield 64 mg (16%). R_f = 0.7 (PE : EA = 1 : 2). ¹H NMR (400 MHz, CDCl₃) δ: 7.54–5.92 (m, 13H, ArH), 5.48 (s, 1H, CH), 4.74 (d, 1H, J = 9.6 Hz), 4.74 (q, 1H, J = 9.6 Hz), 4.36–4.31 (m, 1H), 4.27–4.19 (m, 1H), 3.46–3.28 (m, 4H), 3.03–2.97 (m, 1H), 2.91–2.86 (m, 1H). ¹³C NMR (100 MHz, CDCl₃) δ 136.3, 136.045, 135.3, 134.4, 132.7, 130.0, 129.5, 128.5, 128.2, 127.4, 126.5, 125.6, 80.6, 72.1, 60.1, 48.1, 41.8, 30.8, 29.3.

rel(1S,9bS)-1-Phenyl-1,4,5,9b-tetrahydro[1,2]thiazeto[3,2-a]isoquinoline 2,2-dioxide (11)

White solid (6.5% yield). ¹H NMR (600 MHz, CDCl₃) δ 5.31 (d, J = 4.5 Hz, 1H in CHN), 4.82 (d, J = 4.5 Hz, 1H in CHS). HRMS (ESI) calcd for C₁₆H₁₆NO₂S [M + H]⁺ m/z 286.0896, found 286.0893.

2-(Benzylsulfonyl)-1,2,3,4-tetrahydroisoquinolin-1-ol (12)⁶

Colorless crystals, mp 204–205 °C. Yield 33 mg (11%). R_f = 0.5 (PE : EA = 1 : 2). ¹H NMR (400 MHz, CDCl₃) δ 7.09–7.31 (m, 9H), 6.02 (d, J = 5.4 Hz, 1H), 4.34 (s, 2H), 3.55 (dddd, J = 0.4, 2.4, 5.6, 12.8 Hz, 1H), 3.30 (dt, J = 4.0, 12.4 Hz, 1H), 2.93 (d, J = 5.4 Hz, 1H), 2.77 (ddd, J = 5.6, 11.6, 16.4 Hz, 1H), 2.65 (dt, J = 16.0, 3.2 Hz, 1H). ¹³C NMR (100 MHz, CDCl₃) δ 134.0, 133.8, 130.9, 130.7, 128.7, 128.6, 128.4, 126.7, 77.1, 59.7, 38.8, 29.0.

Ethyl rel(4bR,5R,13bS)-4b,5,9,13b,15,16-hexahydro-8H-[1,2,4]thiadiazino[3,2-a:5,4-a′]diisoquinoline-5-carboxylate 6,6-dioxide (13)^2b

Colorless crystals, mp 134–135 °C. Yield 42 mg (20%). ¹H NMR (400 MHz, CDCl₃) δ 7.48–7.44 (m, 1H), 7.31–7.25 (m, 2H), 7.25–7.18 (m, 3H), 7.11 (dd, J = 12.4, 7.5 Hz, 2H), 5.97 (s, 1H), 5.18 (d, J = 10.4 Hz, 1H), 4.32 (d, J = 10.4 Hz, 1H), 4.28–4.23 (m, 1H), 4.23–4.18 (m, 1H), 3.84 (dt, J = 11.4, 5.0 Hz, 1H), 3.52–3.42 (m, 1H), 3.09–3.01 (m, 1H), 2.99–2.90 (m, 2H), 2.85–2.73 (m, 2H), 2.70–2.61 (m, 1H), 1.20 (t, J = 7.1 Hz, 3H). ¹³C NMR (400 MHz, CDCl₃) δ 164.7, 135.3, 134.4, 132.6, 131.1, 129.3, 128.7, 128.2, 128.1, 127.9, 127.8, 127.2, 125.9, 74.2, 63.1, 62.3, 61.6, 40.4, 38.4, 29.3, 29.1, 13.9.

Ethyl rel(4bR,5R,13bR)-4b,5,9,13b,15,16-hexahydro-8H-[1,2,4]thiadiazino[3,2-a:5,4-a′]diisoquinoline-5-carboxylate 6,6-dioxide (14)^2b

Colorless crystals, mp 146–148 °C. Yield 54 mg (26%). ¹H NMR (400 MHz, CDCl₃) δ 7.38–7.25 (m, 4H), 7.24–7.17 (m, 3H), 7.09 (d, J = 7.2 Hz, 1H), 5.94 (s, 1H), 5.15 (d, J = 4.6 Hz, 1H), 4.35 (d, J = 4.6 Hz, 1H), 4.08–4.02 (m, 1H), 3.95–3.87 (m, 1H), 3.84–3.72 (m, 2H), 3.56–3.47 (m, 1H), 3.16–3.08 (m, 1H), 2.96–2.90 (m, 1H), 2.75–2.65 (m, 2H), 2.55–2.47 (m, 1H), 0.78 (t, J = 7.1 Hz, 3H). ¹³C NMR (400 MHz, CDCl₃) δ 166.8, 135.8, 135.4, 131.9, 131.0, 129.2, 129.1, 128.2, 128.1, 127.5, 126.87, 126.86, 126.0, 73.8, 64.8, 61.5, 60.4, 42.0, 38.6, 29.9, 29.7, 13.3.

Compounds 15 and 16 were obtained as an inseparable white solid mixture (mp 165–170 °C; 60 mg, 30% total yield). The ratio of 15 : 16 was 34 : 66, identical to the ratio obtained from the crude reaction mixture.

rel(4bR,5S,13bR)-5-(Methylsulfonyl)-4b,5,9,13b,15,16-hexahydro-8H-[1,2,4]thiadiazino[3,2-a:5,4-a′]diisoquinoline 6,6-dioxide (15)

Yield 10%. ¹H NMR (400 MHz, CDCl₃) δ 7.81 (d, J = 7.2 Hz, 1H), 7.46–7.41 (m, 1H), 7.32–7.27 (m, 3H), 7.23–7.19 (m, 2H), 7.09 (d, J = 7.2 Hz, 1H), 5.99 (s, 1H), 5.39 (d, J = 9.6 Hz, 1H), 4.76 (d, J = 9.6 Hz, 1H), 3.97 (dt, J = 11.7, 4.9 Hz, 1H), 3.55–3.50 (m, 1H), 3.17 (s, 3H), 2.94–2.78 (m, 4H), 2.71–2.65 (m, 2H). ¹³C NMR (400 MHz, CDCl₃) δ 135.0, 134.8, 133.9, 131.5, 130.4, 128.84, 128.77, 128.69, 128.5, 128.1, 127.4, 125.4, 76.51, 73.6, 61.6, 46.5, 41.2, 38.8, 29.4, 28.2. IR (film) v cm⁻¹ 3029, 2929, 1496, 1454, 1428, 1405, 1340, 1324, 1277, 1139, 1165, 1116, 1089, 1044, 1006, 970, 786, 754, 732, 690, 647, 607. ESI-HRMS [M + H]⁺ calcd for C₂₀H₂₃N₂O₄S₂ 419.1094, found 419.1094.

rel(4bR,5R,13bR)-5-(Methylsulfonyl)-4b,5,9,13b,15,16-hexahydro-8H-[1,2,4]thiadiazino[3,2-a:5,4-a′]diisoquinoline 6,6-dioxide (16)

Yield 20%. ¹H NMR (400 MHz, CDCl₃) δ 7.46–7.41 (m, 1H), 7.32–7.27 (m, 2H), 7.23–7.19 (m, 4H), 7.16–7.13 (m, 1H), 6.02 (s, 1H), 5.44 (d, J = 4.0 Hz, 1H), 4.66 (d, J = 4.0 Hz, 1H), 4.04 (ddd, J = 2.5, 5.3, 12.7 Hz, 1H), 3.59 (dt, J = 12.4, 3.3, 1H), 3.43–3.36 (m, 1H), 3.21 (s, 3H), 3.11–3.01 (m, 2H) 2.94–2.78 (m, 3H). ¹³C NMR (400 MHz, CDCl₃) δ 135.8, 132.1, 131.1, 130.2, 129.4, 129.2, 128.4, 127.9, 127.6, 127.1, 125.9, 125.5, 76.48, 73.8, 63.2, 45.8, 42.9, 38.7, 30.0, 29.0. IR (film) v cm⁻¹ 3029, 2929, 1496, 1454, 1428, 1405, 1340, 1324, 1277, 1139, 1165, 1116, 1089, 1044, 1006, 970, 786, 754, 732, 690, 647, 607. ESI-HRMS [M + H]⁺ calcd for C₂₀H₂₃N₂O₄S₂ 419.1094, found 419.1094.

5-Ethoxy-4-(ethylsulfonyl)-1-methyl-4,5-dihydro-1H-benzo[f][1,4]diazepin-2(3H)-one (18a)^2a

Yield 59 mg (19%). R_f = 0.4 (PE : EA = 2 : 1, v/v). Yellowish oil. ¹H NMR (400 MHz, CDCl₃) δ = 7.51–7.29 (m, 4H), 5.90 (s, 1H), 4.11 (d, J = 14.0 Hz, 1H), 3.57 (d, J = 14.0 Hz, 1H), 3.45 (dq, J = 15.2, 7.6 Hz, 1H), 3.38 (d, J = 15.2, 7.6 Hz, 1H), 3.35 (s, 3H), 3.30 (dq, J = 14.4, 7.2 Hz, 1H), 3.12 (dq, J = 14.4, 7.2 Hz, 1H), 1.46 (t, J = 7.6 Hz, 3H), 1.09 (t, J = 7.2 Hz, 3H); ¹³C NMR (100 MHz, CDCl₃) δ = 166.7, 141.8, 130.6, 130.0, 129.1, 126.5, 123.8, 88.7, 62.7, 49.1, 47.9, 35.3, 14.9, 7.5.

4-(Benzylsulfonyl)-5-ethoxy-1-methyl-4,5-dihydro-1H-benzo[f][1,4]diazepin-2(3H)-one (18b)^2a

Yellowish oil. Yield 75 mg (20%). R_f = 0.4 (PE : EA = 2 : 1, v/v). ¹H NMR (400 MHz, CDCl₃) δ 7.50–7.29 (m, 9H), 5.92 (s, 1H), 4.61 (d, J = 13.6 Hz, 1H), 4.53 (d, J = 13.6 Hz, 1H), 4.20 (d, J = 14.0 Hz, 1H), 3.57 (d, J = 14.0 Hz, 1H), 3.50 (dq, J = 14.6, 7.3 Hz, 1H), 3.36 (s, 3H), 3.19 (dq, J = 14.6, 7.3 Hz, 1H), 1.14 (t, J = 7.3 Hz, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 166.6, 141.8, 131.0, 130.6, 130.0, 129.1, 128.7, 128.5, 128.3, 126.5, 123.9, 89.0, 62.8, 60.5, 48.1, 35.4, 14.9.

trans-6-Methyl-1-phenyl-1,10b-dihydro-6H-benzo[f][1,2]thiazeto[2,3-d][1,4]diazepin-5(4H)-one 2,2-dioxide (19)^2a

Colorless crystals, mp 237–239 °C. Yield 53 mg (16%). R_f = 0.5 (PE : EA = 2 : 1, v/v). ¹H NMR (400 MHz, CDCl₃) δ 7.64–7.26 (m, 9H, ArH), 5.82 (d, J = 3.2 Hz, 1H), 4.93 (d, J = 3.2 Hz, 1H), 3.83 (d, J = 12.0 Hz, 1H), 3.79 (d, J = 12.0 Hz, 1H), 3.40 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 165.1, 142.7, 130.6, 130.3, 130.0, 129.4, 128.9, 127.1, 126.8, 126.2, 123.6, 78.7, 55.3, 46.1, 36.1.

Ethyl rel(2S,10R,11S)-1,5,8,14-tetraza-9-thiadibenzo[c,l]tricyclo[9.5.0.0 ^2,8]hexadecane-6,15-dione-10-carboxylate 9,9-dioxide (20)

Yield 25 mg (5%). TLC R_f = 0.7 (CH₂Cl₂ : MeOH = 40 : 1). Colorless crystals, mp > 300 °C (dec.). ¹H NMR (400 MHz, CDCl₃) δ 7.56–7.19 (m, 8H), 5.85 (s, 1H), 4.91 (d, J = 11.2 Hz, 1H), 4.51 (d, J = 11.2 Hz, 1H), 4.37 (d, J = 14.4 Hz, 1H), 4.17 (m, 2H), 3.66 (d, J = 14.4 Hz, 1H), 3.42 (s, 3H), 3.14 (s, 3H), 2.93 (d, J = 14.0 Hz, 1H), 2.70 (d, J = 14.0 Hz, 1H), 1.15 (dd, J = 7.2, 7.2 Hz, 3H); ¹³C NMR (400 MHz, CDCl₃) δ 166.9, 166.5, 143.6, 141.8, 132.5, 127.7, 126.7, 125.5, 124.0, 121.2, 80.4, 65.6, 62.5, 59.8, 51.6, 49.1, 35.3, 33.5, 13.8; IR (film) v cm⁻¹ 2976, 2929, 1739, 1668, 1602, 1495, 1462, 1368, 1297, 1174, 1150, 1097, 1027, 998, 766, 736, 666; ESI-HRMS [M + H]⁺ calc. for C₂₄H₂₇N₄O₆S 499.1651, found 499.1651.

Ethyl rel(2R,10S,11S)-1,5,8,14-tetraza-9-thiadibenzo[c,l]tricyclo[9.5.0.0 ^2,8]hexadecane-6,15-dione-10-carboxylate 9,9-dioxide (21)

Yield 99 mg (20%). TLC R_f = 0.4 (CH₂Cl₂ : MeOH = 40 : 1). Colorless crystals, mp > 300 °C (dec.). ¹H NMR (400 MHz, CDCl₃) δ 8.45 (d, J = 7.6 Hz, 1H), 7.52–7.15 (m, 7H), 5.34 (s, 1H), 5.10 (d, J = 7.2 Hz, 1H), 5.03 (d, J = 7.2 Hz, 1H), 4.43 (d, J = 14.4 Hz, 1H), 4.14 (dq, J = 14.4, 7.2 Hz, 1H), 4.06 (dq, J = 14.4, 7.2 Hz, 1H), 3.65 (d, J = 14.4 Hz, 1H), 3.52 (s, 3H), 3.33 (s, 3H), 2.64 (d, J = 14.2 Hz, 1H), 2.40 (d, J = 14.2 Hz, 1H), 0.97 (dd, J = 7.2, 7.2 Hz, 3H); ¹³C NMR (400 MHz, CDCl₃) δ 166.9, 166.6, 163.3, 143.5, 142.6, 130.7, 129.7, 129.3, 126.8, 126.5, 125.4, 124.3, 121.3, 63.6, 62.3, 58.7, 54.0, 49.0, 34.8, 13.6; IR (film) v cm⁻¹ 2978, 2930, 1743, 1669, 1601, 1493, 1461, 1424, 1368, 1285, 1172, 1153, 1001, 766, 734, 648; ESI-HRMS [M + H]⁺ calc. for C₂₄H₂₇N₄O₆S 499.1651, found 499.1650.

Acknowledgements

This work was supported in part by the National Basic Research Program of China (No. 2013CB328905), the National Natural Science Foundation of China (Nos 21372025 and 21172017).

Notes and references

1. H. Staudinger and E. Pfenninger, Chem. Ber., 1916, 49, 1941.
2. (a) Z. H. Yang and J. X. Xu, J. Org. Chem., 2014, 79, 10703; (b) Z. H. Yang, N. Chen and J. X. Xu, J. Org. Chem., 2015, 80, 3611; (c) Z. H. Yang and J. X. Xu, Tetrahedron, 2015, 71, 2844.
3. For selected reactions of sulfonyl chlorides with imines, see: (a) O. Tsuge and S. Iwanami, Bull. Chem. Soc. Jpn., 1970, 43, 3543; (b) T. Hiraoka and T. Kobayashi, Bull. Chem. Soc. Jpn., 1975, 48, 480; (c) P. Loiseau, M. Bonnafous and Y. Adam, Eur. J. Med. Chem., 1984, 19, 569; (d) M. J. Szymonifka and J. V. Heck, Tetrahedron Lett., 1989, 30, 2869; (e) E. Grunder and G. Leclerc, Synthesis, 1989, 135; (f) E. Grunder-Klotz and J.-D. Ehrhardt, Tetrahedron Lett., 1991, 32, 751; (g) M. F. Gordeev, E. M. Gordon and D. V. Patel, J. Org. Chem., 1997, 62, 8177; (h) T. Iwama, T. Kataoka, O. Muraoka and G. Tanabe, J. Org. Chem., 1998, 63, 8355; (i) T. Iwama, A. Takag and T. Kataoka, Chem. Pharm. Bull., 1998, 46, 757; (j) M. Zajac and R. Peters, Org. Lett., 2007, 9, 2007.
4. For reviews, see: (a) P. Hudhomme. Comprehensive Heterocyclic Chemistry III, Vol. 2, Four-membered Rings with one Sulfur and One Nitrogen Atom. Elsevier Ltd, 2008, pp. 729–759; (b) J. Chanet-Ray and R. Vessiere, Org. Prep. Proced. Int., 1986, 18, 157 For selected examples, see:; (c) H. H. Otto and P. Schwenkkraus, Tetrahedron Lett., 1982, 23, 5389; (d) F. Cavagna, W. Koller, A. Linkies, H. Rehling and D. Reuschling, Angew. Chem., 1982, 94, 549–550; (e) M. Müller and H. H. Otto, Liebigs Ann. Chem., 1991, 171; (f) T. Iwama, A. Karaoka, O. Muraoka and G. Tanabe, J. Org. Chem., 1988, 63, 8355; (g) T. Iwama, A. Karaoka, O. Muraoka and G. Tanabe, Tetrahedron, 1998, 54, 5507; (h) N. J. Baxter, A. P. Laws, L. J. H. Rigoreau and M. I. Page, Chem. Commun., 1999, 2401.
5. For reviews, see: (a) M. I. Page, Acc. Chem. Res., 2004, 37, 297; (b) M. I. Page, W. Y. Tsang and N. Ahmed, J. Phys. Org. Chem., 2006, 19, 446 For selected examples, see:; (c) W. Y. Tsang, N. Ahmed, P. S. Hinchliffe, J. M. Wood, L. P. Harding, A. P. Laws and M. I. Page, J. Am. Chem. Soc., 2005, 127, 17556; (d) W. Y. Tsang, N. Ahmed, L. P. Harding, K. Hemming, A. P. Laws and M. I. Page, J. Am. Chem. Soc., 2005, 127, 8946; (e) W. Y. Tsang, N. Ahmed, K. Hemming and M. I. Page, Org. Biomol. Chem., 2007, 5, 3993; (f) M. Gersch, R. Kolb, F. Alte, M. Groll and S. A. Sieber, J. Am. Chem. Soc., 2014, 136, 1360.
6. J. Liu, S. L. Hou and J. X. Xu, Phosphorus, Sulfur Silicon Relat. Elem., 2011, 186, 2377 and unpublished results.
7. For recent examples, see: (a) A.-B. Nørholm, P. Francotte, E. Goffin, I. Botez, L. Danober, P. Lestage, B. Pirotte, J. S. Kastrup, L. Olsen and C. Oostenbrink, J. Chem. Inf. Model., 2014, 54, 3404; (b) P. Francotte, A. B. Nørholm, Y. T. Deva, L. Olsen, K. Frydenvang, E. Goffin, P. Fraikin, P. de Tullio, S. Challal, J.-Y. Thomas, F. Iop, C. Louis, I. Botez-Pop, P. Lestage, L. Danober, J. S. Kastrup and B. Pirotte, J. Med. Chem., 2014, 57, 9539 , and the references cited therein.
8. B. Maji and H. Z. Mayr, Naturforsch, 2013, 68b, 693.
9. The monocyclic and bicyclic imines were designed and synthesized, and their electron density and ring extension or steric hindrance were adjusted by introducing hetero atoms or new substituents into the ring. The representative five-membered cyclic imines that failed to undergo [2^s + 2ⁱ] or [2^s + 2ⁱ + 2ⁱ] annulation are listed as following.
   .
10. The reactions of cyclic imine 2a with various benzyl sulfonyl chlorides failed to give [2^s + 2ⁱ] or [2^s + 2ⁱ + 2ⁱ] annuladducts, instead, the corresponding hydrolyzed products N-(3-(2-formylphenyl)propyl)-substituted phenylmethanesulfonamides generated.⁶ We surmised that it was the strong ring extension and the resulting bulky steric hindrance in the intermediates that caused the failure. Therefore, we decided to modify the ring extension of the cyclic imine 2a by introducing a functional group into the structure. Illuminated by linear imines employed in Loiseau's report,^3c we devised a new bicyclic imine 2b with an ester group embedded. However, numerous attempts toward this structure failed, partly because of the instability of the ester group under basic conditions. So, further modification of the structure by replacing the ester with an amide directed us to the imine 2.
    .
11. Since 17a and 17b were difficult to purify, converting them into 18a and 18b in refluxed ethanol easily tackled the problem.
12. For structural properties of the analogue from ketenes, see: (a) F. P. Cossio, A. Arrieta and M. A. Sierra, Acc. Chem. Res., 2008, 41, 925; (b) L. Jiao, Y. Liang and J. X. Xu, J. Am. Chem. Soc., 2006, 128, 6060; (c) Y. Liang, L. Jiao, S. W. Zhang, Z. X. Yu and J. X. Xu, J. Am. Chem. Soc., 2009, 131, 1542.
13. (a) Z. H. Yang and J. X. Xu, Synthesis, 2013, 45, 1675; (b) Z. H. Yang, Y. P. Zheng and J. X. Xu, Synlett, 2013, 24, 2165; (c) Z. H. Yang, B. N. Zhou and J. X. Xu, Synthesis, 2014, 46, 225; (d) Z. H. Yang and J. X. Xu, Org. Synth., 2014, 91, 116.
14. (a) S. A. Wolckenhauer, A. S. Devlin and J. Du Bois, Org. Lett., 2007, 9, 4363; (b) Z. H. Yang and J. X. Xu, Chem. Commun., 2014, 50, 3616.
15. J. C. Pelletier and M. P. Cava, J. Org. Chem., 1987, 52, 616.

---

Footnote

† Electronic supplementary information (ESI) available: Copies of ¹H, ¹³C, and NOE NMR spectra. See DOI: 10.1039/c5ra15717j

---