Structure-Reactivity Relationships of Zwitterionic 1,3-diaza-Claisen Rearrangements

Abstract

Bridged bicyclic tertiary allylic amines aza-norbornene 1 and isoquinuclidene 2 add to isocyanates, isothiocyanates and in situ-generated carbodiimides to form zwitterionic intermediates that undergo 1,3-diaza-Claisen rearrangements to afford highly substituted ureas, thioureas and guanidines respectively. Aza-norbornene 1 is significantly more reactive towards 1,3-diaza-Claisen rearrangements than isoquinuclidene 2. This reactivity difference is most likely due to the inherent ring strain in the aza-bicyclo[2.2.1]heptene ring system of aza-norbornene 1. The most apparent reactivity trend of the heterocumulenes is that the most electron-deficient heterocumulenes are more reactive toward 1,3-diaza-Claisen rearrangements. The introduction of a new stereocenter α- to the nucleophilic nitrogen in aza-norbornene 1 and isoquinuclidine 2 decreases the reactivity towards 1,3-diaza-Claisen rearrangements, while the exo-diastereomers 3b and 4b are less reactive than the corresponding endo-diastereomers 3a and 4a. Isocyanates that bear an electron-withdrawing group react with allylic amines 1-3b to afford mixtures of ureas and isoureas, however, with excess isocyanate and heat, thermodynamic equilibration is possible affording ureas. Inspired by this observation a one-pot reaction of isocyanates with amines 1, 2, and 3b followed by BF₃•OEt₂-catalyzed isomerization of the urea/isourea mixture was developed that affords the corresponding ureas in excellent yields.

Introduction

Due to the importance of guanidine compounds,¹ we have been involved in the development of methods for the synthesis of the guanidine functionality. In particular, we have been interested in developing a rearrangement that affords guanidines since rearrangements often offer an efficient means for the construction of complex molecules from simple starting materials. We have previously disclosed preliminary communications on the reaction of tertiary allylic amines with isocyanates, isothiocyanates, and in situ-generated carbodiimides that afford highly substituted allylic ureas, thioureas and guanidines respectively via a zwitterionic intermediate that undergoes a 1,3-diaza-Claisen rearrangement.^2-5 Since ureas and thioureas can be easily converted to guanidines,⁶ each of these transformations provides access to guanidine products. The present work is an account of the scope and limitations of the zwitterionic 1,3-diaza-Claisen rearrangement.

A mechanistic overview of the reaction of allylic amines with heterocumulenes is shown in Figure 1. Tertiary allylic amines 1 and 2 add to isocyanates, isothiocyanates, and carbodiimides to afford the corresponding zwitterionic intermediates. The zwitterionic intermediates may then rearrange to afford the ureas, thioureas or guanidines A, or alternatively, the isoureas, isothioureas or regioisomeric guanidines B.⁷ In the present study, we investigated 1) the electronic factors that influence the reactivity of isocyanates, isothiocyanates and carbodiimides, 2) ring strain effects on reactivity of rearrangement precursors 1 and 2, 3) the reactivity effects of steric crowding proximal to the amine functionality on both the aza-norbornene and isoquinuclidine frameworks, and 4) the factors that influence regiochemistry (i.e. A vs. B).

Figure 1.

Mechanistic overview of the reaction of allylic amines 1 and 2 with heterocumulenes.

Results and Discussion

Aza-norbornenes and isoquinuclidenes 1-4b were synthesized as shown in Figure 2 through the hetero-Diels-Alder reaction of in situ-generated iminium ions with either cyclopentadiene or cyclohexadiene as described.⁸ All thioureas were synthesized by the reaction of an amine with the corresponding isothiocyanate (not shown),⁹ with the exception of N-benzyl-N′-Tf thiourea 5 which was synthesized by the reaction of triflamide with NaH and subsequent treatment of the resulting anion with BnNCS, followed by acidic workup in 88% yield (Figure 2).¹⁰ TsNCS and PmcNCS were synthesized as previously reported.^9,11

Figure 2.

Synthesis of rearrangement precursors.

I. Reaction of aza-norbornene 1 and isoquinuclidene 2 with isocyanates, isothiocyanates and in situ-generated carbodiimides

The reactions of isocyanates and isothiocyanates with aza-norbornene 1 and isoquinuclidene 2 are summarized in Table 1. TsNCO reacts with aza-norbornene 1 at room temperature to afford the urea 6 and isourea 7 in isolated yields of 53% and 41% respectively. These structural assignments are confirmed by IR. The IR spectrum of urea 6 exhibits a C=O stretch 1675 cm^-1 while the IR spectrum of isourea 7 exhibits a C=N stretch at 1567 cm^-1. BnNCO exhibited diminished reactivity with aza-norbornene 1 requiring heating at reflux in benzene, but afforded only the urea 8 in 71% yield. Interestingly, TsNCS also reacts with aza-norbornene 1 at room temperature in benzene affording thiourea 9 and isothiourea 10 in 39% and 52% yield respectively. Thus, the more electron-deficient isocyantes TsNCO and BzNCO are more reactive than the less electron deficient BnNCO. The reaction of TsNCO with the less strained isoquinuclidene 2 required heating in benzene at reflux to afford urea 11 (48%) and isourea 12 (46%). As before, assignments were made based on the C=O IR band at 1669 cm^-1 and the C=N IR band at 1557 cm^-1. Isoquinuclidene 2 failed to afford rearrangement product with the less reactive BnNCO and BzNCO (not shown) even under forcing conditions (neat, 120 °C). The lower reactivity of isoquinuclidene 2 as compared with aza-norbornene 1 may be explained by the lower ring straing in isoquinuclidene 2 and the fact that C-N bond breakage would be coupled to release of ring strain. The highly electron-deficient TsNCS smoothly reacted with isoquinuclidene 2 at room temperature, but afforded exclusively the isothiourea 13 in 81% yield. As in the case of aza-norbornene 1, isoquinuclidene 2 also exhibited the trend of a more facile rearrangement with the more electron-deficient heterocumulenes, however, the less electron deficient heterocumulenes BzNCO and BnNCO failed to react with isoquinuclidene 2.

Table 1.

Reaction of isocyanates and isothiocyanates with aza-norbornene 1 and isoquinuclidene 2.

allylic amine	R-NCXa	Conditions	product(s)b
	TsNCO	benzene, rt
1	BnNCO	benzene, reflux
1	TsNCS	benzene, rt
	TsNCO	benzene, reflux
2	BzNCO	neat 120 °C	isocyanate decomposition
2	TsNCS	benzene, rt

^a1.5-2 equiv.

^bIsolated yields.

The reaction of thioureas with a primary or secondary amine in the presence of an activating agent such as EDCI, I₂, N-methyl-2-choloropyridinium iodide, Hg(II) salts is one of the most common means by which to synthesize guanidine compounds.¹² The most widely accepted mechanism for this transformation is the conversion of the thiourea (through reaction with an activating agent) to a carbodiimide followed by addition of the amine to the carbodiimide to afford the guanidine. Accordingly, we reasoned that in situ-generated N-alkyl-N′-EWG-carbodiimides would be sufficiently electrophilic to react with tertiary amines 1 and 2 affording a zwitterionic intermediate that would then undergo a 1,3-diaza-Claisen rearrangement.

The reactions of thioureas and an activating agent with aza-norbornene 1 and isoquinuclodene 2 are summarized in Table 2. Aza-norbornene 1 was smoothly transformed to bicyclic guanidine 15 in 67% yield on treatment with thiourea 14 and EDCI at room temperature. It is worth noting that EDCI itself does not participate in a 1,3-diaza-Claisen rearrangement with aza-norbornene. This is most likely because it is not sufficiently electron-deficient. We have previously reported that N-carbamoyl thioureas in the presence of EDCI do not undergo reaction with aza-norbornenes at room temperature, but we are now pleased to report that carbamoyl thiourea 16 on activation with the Mukayiama salt underwent reaction with aza-norbornene 1 with mild heating at 60 °C affording the bicyclic guanidine 17 in 76% yield. Thus, with mild heating carbamoyl thioureas in the presence of an activating agent also react with aza-norbornene 1. In contrast, the bis-carbamoyl thiourea 18, bearing two carbamoyl electron-withdrawing groups underwent rearrangement with aza-norbornene 1 in the presence of EDCI at room temperature affording the bis-Boc-guanidine 19 in 51% yield. Single crystal X-ray analysis of guanidine 19 reveals that the imine Boc-group is syn to N1 and is thus anti to N3. At this juncture it is unclear if the 1,3-diaza-Claisen rearrangement proceeds preferentially by transition state 24 (Figure 3) in which the imine Boc-group is syn to N1 (but not coplanar due to A(1,3) strain) or through transition state 25 in which the imine Boc-group is syn to N3. Although, these transition states would yield alternate isomers 19 and 26, Kessler and Leibfritz have shown that imine syn/anti isomers of guanidines readily interconvert faster than the NMR timescale, especially when an electron-withdrawing group is on the imine nitrogen.¹³ We thus expect that the imine isomers readily interconvert and isomer 19 is selectively crystallized by crystal packing forces. Reaction of Pmc-thiourea 20 with aza-norbornene 1 in the presence of EDCI afforded guanidine 21 in 42% yield. Again, isomers 21 and 29 should readily interconvert, and 21 selectively crystallizes. The X-ray crystal structure of Pmc-guanidine 21 demonstrates the electron-withdrawing Pmc-group is on the imine nitrogen and not on N1. By analogy, we propose that the guanidines in Table 2 that are derived from unsymmetrical thioureas also possess the electron-withdrawing group (Ts, CO₂Et, etc) on the imine nitrogen and not on N1. Interestingly, under no conditions examined, did thiourea 14 react with isoquinicludene 2 in the presence of an activating agent (data not shown) to afford rearrangement product. For this reason, we were interested in investigating the reactivity of thiourea 5 bearing the highly electron-withdrawing trifluoromethanesulfonyl (Tf) group as we reasoned that the more electron-deficient N-Tf-N′-Bn-carbodiimide may react with isoquinuclidene 2 mirroring the reactivity trend of the isocyanates.

Table 2.

Reaction of thioureas with aza-norbornene 1 and isoquinuclidene 2.

allylic amine	thioureaa	conditions	productb
		EDCI, EtNi-Pr2, CHCl3, rt
1		Mukayiama salt, EtNi-Pr2, CHCl3, 60 °C
1		EDCI, EtNi-Pr2, CHCl3, rt
1		EDCI, EtNi-Pr2, CHCl3, rt
1		Mukayiama salt, EtNi-Pr2, CHCl3, rt
		Mukayiama salt, EtNi-Pr2, CHCl3, 60 °C

^a1-2 equivalents.

^bIsolated yields.

Figure 3.

X-ray crystal structures of guanidines 19 and 21 and potential transition states leading to formation of 19 and 21.

The EDCI-mediated reaction of thiourea 5 with aza-norbornene 1 did not proceed as cleanly as other reactions (data not shown), however, it was discovered that thiourea 5 on activation with the Mukayiama salt afforded N-Tf guanidine 22 in 62% yield. Furthermore, when isoquinuclidene 2 was subjected to Tf-thiourea 5 and the Mukayiama salt in CHCl₃ at 60 °C the rearrangement product 23 was obtained in 57% yield.

II. Reaction of aza-norbornenes 3a,b and isoquinuclidenes 4a,b with isocyanates

The bridged-bicyclic amines 3a-4b (Figure 4) were synthesized in order to determine what effects steric crowding adjacent to the nucleophilic nitrogen of the tertiary allylic amine would have on reactivity. Our studies began by assessing the reactivity of the diastereomeric aza-norbornenes 3a (endo-ethoxycarbonyl group) and 3b (exo-ethoxycarbonyl group) with TsNCO (highly reactive isocyanate) and BnNCO (less reactive isocyanate). The endo-aza-norbornene 3a reacted smoothly with TsNCO at room temperature to afford the urea 30 in 24% yield and the isourea 31 in 64% yield. The exo-aza-norbornene 3b also smoothly undergoes reaction with TsNCO at room temperature to afford urea 35 in 47% yield, and isourea 36 in 44% yield. Thus, with the highly reactive TsNCO, both diastereomeric aza-norbornenes 3a and 3b react at room temperature. In contrast, endo-aza-norbornene 3a did not undergo reaction with the less reactive BnNCO in benzene at reflux (conditions under which 1 underwent reaction with BnNCO). For BnNCO to react with aza-norbornene 3a, heating at 120 °C under neat conditions was required to afford the expected urea 33 in 38% yield and the hydantoin 34 in 39% yield. The exo-aza-norbornene 3b proved even less reactive toward rearrangement with BnNCO affording none of the expected urea, but affording as the sole isolable product the hydantoin 34 in 67% yield under forcing conditions (120 °C, neat). Thus, aza-norbornene 3a exhibits diminished reactivity toward rearrangement with BnNCO as compared with aza-norbornene 1, and exo-isomer 3b does not afford rearrangement product at all with BnNCO. In the case of the isomeric isoquinuclidenes 4a and 4b, the endo-isomer 4a underwent reaction with TsNCO at 80 °C under neat conditions to afford urea 37 in 48% yield, while the exo-isomer (4b) did not undergo rearrangement with TsNCO under any conditions investigated. Since isoquinuclidene 2 does not undergo rearrangement with BnNCO, we did not attempt the reaction of the less reactive isoquinuclidenes 4a and 4b with BnNCO.

Figure 4.

Reaction of aza-norbornenes 3a-4b with TsNCO and BnNCO.

In all, the presence of the ethoxycarbonyl-group on the aza-norbornene (i.e. 3a, 3b) or isoquinuclidene (i.e. 4a, 4b) frameworks regardless of the endo- or exo-stereochemistry diminishes the reactivity toward rearrangement of these substrates as compared with the unsubstituted aza-norbornene 1 and isoquinuclidene 2. This may be due to the constrained conformation of these frameworks, that enforce eclipsing interactions either between the CO₂Et and Bn groups or between the CO₂Et and −C(O^-)=N-Bn groups depending on the diastereomeric zwitterionic intermediate. These eclipsing interactions will disfavor formation of the zwitterionic intermediates as compared to zwitterionic intermediates arising from the unsubstituted amines 1 and 2. Lower concentration of zwitterionic intermediates will in turn result in a decrease of the rate of rearrangement.

III. Mechanistic observations

A key mechanistic insight was made when we initially attempted the reaction of 1 equivalent isoquinuclidene 2 with 1 equivalent TsNCO at room temperature. NMR analysis of the crude reaction mixture revealed a deshielding of the resonances α to the amine of the isoquinuclidene 2, but these changes were not consistent with a skeletal rearrangement. Instead, the shift in the resonances of the isoquinuclidene 2 suggested that the observed shift was due to the formation of the zwitterionic intermediate. The ¹H NMR spectrum of 1 equivalent isoquinuclidene 2 with 0.5 equivalents TsNCO at room temperature exhibits a shift in the resonances, but not of the same magnitude as is observed with the 1:1 2/TsNCO stoichiometry. Since the ¹H NMR spectrum of isoquinuclidene 2 with 0.5 eq. of TsNCO does not exhibit a set of resonances for 2 and another set of resonances for the zwitterionic intermediate, these data are consistent with a fast and reversible addition step on the NMR time scale. Furthermore, as the rearrangement does not take place at room temperature, the rearrangement is obviously the rate-determining step. In this scenario, the rate of formation of product = k₂[zwitterionic intermediate]. In practical terms, this indicates that conditions that increase the concentration of the zwitterionic intermediate will increase the rate of the overall reaction. These observations can in part explain the reactivity trend of the heterocumulenes (i.e. TsNCO>BzNCO>BnNCO) as it would be expected that [zwitterionic intermediate] would increase with the electron-deficiency of the heterocumulene.

IV. Proposed mechanism for the formation of hydantoin 34

A proposed scheme for the formation of hydantoin 34 is detailed in Figure 5. Addition of aza-norbornenes 3a and 3b to BnNCO will afford the corresponding diastereomeric zwitterionic intermediates 38 and 39. The retro Diels-Alder reaction of intermediates 38 and 39 affords cyclopentadiene and the isomeric 1,4-dipoles 40a and 40b. The 1,4-dipole 40b would require isomerization to 40a for attack of the anionic nitrogen of the 1,4-dipole on the ester carbonyl to give the tetrahedral intermediate 41 that undergoes ethoxide-[1,2]-migration to furnish the hydantoin 34. The retro-Diels-Alder pathway is reasonable as Grieco has shown that aza-norbornenes can readily undergo retro-Diels-Alder reactions.¹⁴ The most immediate fate of the 1,4-dipoles 40a and 40b is likely to be elimination to afford the isomeric imines and benzyl isocyante (Figure 6) but this process should be reversible such that addition of the imine nitrogen to BnNCO would result in formation of the 1,4-dipoles 40a and 40b. Indeed, when the glyoxylate imine and BnNCO were subjected to the attempted rearrangement conditions (neat, 120 °C), the hydantoin 34 was formed in 66% yield (Figure 6). Although the formation of the hydantoin 34 from the glyoxylate imine and BnNCO (Figure 6) is consistent with the proposed mechanism for hydantoin formation (Figure 5), it does not unambiguously prove that the reaction proceeds through the proposed mechanism. It is alternatively possible that the zwitterionic intermediates 38 and 39 may undergo attack of the nitrogen on the ester carbonyl prior to retro-Diels-Alder reaction. However, we have good evidence that at high temperatures, the aza-norbonene 1, can undergo a retro Diels-Alder reaction in the presence of BnNCO.³ Thus, even in the absence of the α-ester the retro-Diels-Alder is a viable pathway suggesting that zwitterionic intermediates 38 and 39 are also likely to undergo a retro-Diels-Alder reaction at high temperatures.

Figure 5.

Proposed mechanism for the formation of hydantoin 34.

Figure 6.

Proposed mechanism for elimination of 40a and 40b. Reaction of imine and BnNCO to form hydantoin 34.

In the reaction of BnNCO with the aza-norbornenes 3a and 3b, the factors that influence the ratio of the hydantoin 34 arising from the proposed retro-Diels-Alder pathway versus product arising from the rearrangement pathway are complex. A schematic that explains these factors is shown in Figure 7 for the reaction of aza-norbornene 3b with BnNCO. Aza-norbornene 3b has two possible conformers 42 and 45 arising from nitrogen lone-pair inversion. Conformer 45 can react with BnNCO to give the zwitterionic intermediate 46 whose geometry does not allow a concerted 1,3-diaza-Claisen rearrangement. Thus, zwitterionic intermediate 46 can only undergo a retro-Diels-Alder reaction with rate constant k₂ to give cyclopentadiene and the 1,4-dipole 40 that in turn leads to hydantoin 34. In contrast, conformer 42 can react with BnNCO to give the diastereomeric zwitterionic intermediate 43. Zwitterionic intermediate 43 could undergo rearrangement with rate constant k₁ to give urea 44 or undergo retro-Diels-Alder reaction with rate constant k₃. The Curtin-Hammett principle dictates that if the energies of activation for the conversion of intermediate 43 to intermediate 40b and product 44 and for the conversion of intermediate 46 to intermediate 40a are greater than the energy of activation for the interconversion of conformers 42 and 45 (as should be the case), then the energy difference between conformers 42 and 45 should not influence the ratio of ureas 34/44, instead, the ratio 34/44 = (k₂[46] + k₃[43])/k₁[43].¹⁵ As the reaction of aza-norbornene 3b with BnNCO does not afford isolable quantities of the urea 44, this indicates that k₁[43] is small relative to the term (k₂[46] + k₃[43]). Although we have not included a discussion of the product distribution in the reaction of aza-norbornene 3a with BnNCO, similar arguments may be made for the factors that influence product distribution in that reaction.

Figure 7.

Factors influencing the ratio of hydantoin 34 and urea 44.

V. Variables influencing urea/isourea distribution

In the reaction of TsNCO and BzNCO with aza-norbornenes and isoquinuclidenes, we had been observing some variation in the ratio of urea/isourea products. This prompted a thorough investigation of the reaction conditions in order to resolve what led to the variability in these ratios. Figure 8 details the examination of reaction conditions for the reaction of TsNCO with aza-norbornene 1. When aza-norbornene 1 is allowed to react with 0.75 eq. TsNCO (less than 1 equivalent) in benzene at room temperature a 1:1 ratio is obtained of urea 6 to isourea 7 in 15 minutes as determined by ¹H NMR analysis. As this reaction is complete in 15 minutes, the 0.75 eq. of TsNCO are consumed within 15 minutes. If the same reaction is carried out identically, but after 15 minutes, the reaction is brought reflux for 12 h, a 1:1 mixture of products 6/7 is also obtained. When aza-norbornene 1 is allowed to react with 1.5 eq. TsNCO in benzene at room temperature again a 1:1 ratio is obtained of urea 6 to isourea 7 within 15 minutes. However, in this case when after 15 minutes the reaction mixture is heated at reflux for 12 hours, a 2:1 ratio of products 6/7 is obtained, and if the mixture is heated at reflux for 48 hours, only the urea 6 is obtained as determined by ¹H NMR analysis. This indicates that with excess TsNCO and heat, the mixture of urea/isourea can be isomerized to afford only urea 6. Similarly, results were obtained with aza-norbornene 1 and BzNCO (not shown). These data indicate that thermodynamic control can be achieved in these rearrangements by using excess isocyanate and heat. Figure 9 details the proposed mechanism for the TsNCO-catalyzed isomerization of isourea 7 to urea 6. The imine nitrogen of isourea 7 adds to TsNCO affording the zwitterion 49. Rupture of the allylic C-O bond affords the allylic cation 50. Elimination to re-form TsNCO gives intermediate 51 that may cyclize to give either the urea 6 or the isourea 7. This proposed scheme would allow for isomerization of isourea 7 to the thermodynamically more stable urea 6. We propose that thermodynamic equilibration by this mechanism is the reason why in the reaction of amine 4a with TsNCO (Figure 4) the urea 37 is isolated without any of the corresponding isourea.

Figure 8.

Variables influencing urea/isourea ratios.

Figure 9.

Proposed mechanism for isocyanate-mediated isomerization of isourea to urea.

The isocyanate-catalyzed isomerization of isourea 7 to urea 6 appeared reminiscent of a Lewis acid-catalyzed process and thus prompted the hypothesis that a Lewis acid could coordinate the imine nitrogen of the isourea and promote an analogous allylic C-O bond rupture, and recyclization to afford the urea. Figure 10 shows the yields of urea obtained for rearrangement reactions followed by the addition of catalytic BF₃•OEt₂. When aza-norbornene 1 is allowed to react with BzNCO followed by addition of 0.1 eq. BF₃•OEt₂, the urea 47 is obtained in 88% yield. Similarly, when aza-norbornene 1 is allowed to react with TsNCO followed by the addition of catalytic BF₃•OEt₂, the urea 6 is obtained in 87% yield. In addition, when the reaction of aza-norbornene 3b with TsNCO is followed by rearrangement with catalytic BF₃•OEt₂, the urea 35 is obtained in 88% yield. The reaction of isoquinuclidene 2 with TsNCO, followed by BF₃•OEt₂-catalyzed isomerization affords the urea 11 in 94% yield. It is of note that all the examples above afford urea/isourea mixtures if the BF₃•OEt₂-catalyzed isomerization is not applied.

Figure 10.

One pot reaction of isocyanates with allylic amines 1, 2 and 3b, followed by thermodynamic equilibration with catalytic BF₃•OEt₂.

VI. Cyclic non-bridged tertiary allylic amines

We have additionally investigated the rearrangement of cyclic non-bridged tertiary allylic amines N-benzyl pyrroline 51 and N-benzyl tetrahydropyridine 52¹⁷ (Figure 11). However, under all conditions investigated, neither underwent rearrangement with the highly reactive TsNCO. We propose that the unreactivity of the amines 51 and 52 may be explained by a stereoelectronic effect. The concerted reaction would require an approximate coplanar alignment of the breaking bond with the π-bond of the alkene in the transition state. To accommodate this geometry the transition state arising from the pyrroline 51 would require severely puckering the pyrroline nitrogen out of the plane of the other four carbons as in transition state 53 (Figure 11). In addition, for proper orbital alignment, the transition state arising from the tetrahydropyridine 52 would require a boat geometry such as 54. We thus believe that these are higher energy transition states that make the rearrangement less favorable than those arising from the aza-norbornene 1 and isoquinuclidene 2 which would proceed through a better orbital alignment.

Figure 11.

Failed rearrangement of pyrroline 51 and tetrahydropyridine 52. Transition states 53 and 54 that would be required for proper alignment of the breaking bond with the alkene π-bond.

Conclusions

We have explored the scope and limitations of the reaction between isocyanates, isothiocyanates and in situ-generated carbodiimides with bridged, bicyclic tertiary allylic amines. The most apparent trend that arises from these studies regarding the reactivity of the heterocumulene component is that adding an electron-withdrawing substituent to a heterocumulene or increasing the strength of the electron-withdrawing substituent increases the rate of the overall reaction. A potential reason for this effect is that in the reaction of a tertiary allylic amine with a heterocumulene to form a zwitterionic intermediate, the more electron-deficient heterocumulene would favor a larger K_eq for formation of the zwitterionic intermediate, hence, the greater the concentration of zwitterionic intermediate, the greater the rate of the overall reaction. At this junction, it is unknown what effect if any an electron-withdrawing substituent has on the rate-determining rearrangement step. The use of more electron-deficient heterocumulenes may represent an overall strategy by which to accomplish 1,3-diaza-Claisen rearrangements with less reactive tertiary allylic amines. In comparing the reactivity of aza-norbornene 1 with isoquinuclidene 2, it is apparent that aza-norbornene 1 is more reactive towards 1,3-diaza-Claisen rearrangements and this reactivity is most likely due to the inherent ring strain in aza-norbornene 1. Interestingly, in the reactions of in situ generated carbodiimides with aza-norbornene 1, the substituent on the imino-nitrogen of the resulting guanidine crytallizes syn to N1 as determined by the X-ray crystal structures of guanidines 19 and 21, but this is attributed to crystal packing forces.

The rearrangement precursors 3a,b and 4a,b that possess an additional stereocenter α-to nitrogen exhibit diminished reactivity toward 1,3-diaza-Claisen rearrangements when compared with the rearrangement precursors 1 and 2. In comparing the reactivity of these species, the exo-isomers 3b and 4b were clearly less reactive toward 1,3-diaza-Claisen rearrangement than the corresponding endo-isomers 3a and 4a. The reasons for the reactivity profiles of these compounds are complex, but ultimately, the Curtin-Hammett principle dictates that the product distribution is determined by the concentration of each possible diastereomeric zwitterionic intermediate times the rate constant for that reaction.

We have additionally established that with excess isocyanate and heat, it is possible to achieve thermodynamic control over rearrangements and thus equilibrate urea/isourea mixtures to obtain exclusively the urea product (within our detection limits). Inspired by this observation, we developed a one-pot process that involves first the reaction of an isocyanate with a rearrangement precursor to afford a mixture of ureas/isoureas followed by BF₃•OEt₂ catalyzed isomerization of that mixture to afford the ureas in excellent yields.

Experimental

All reactions were carried out under an atmosphere of nitrogen using flame-dried glassware. All reagents and solvents were purchased and used as received. Column chromatography was carried out using standard grade silica gel P60 (40-63 μm particle size), which was purchased and used as received. ¹H and ¹³C NMR spectra were recorded on a 500 MHz spectrometer in CDCl₃ at ambient temperature unless otherwise noted. Splitting patterns were designated as s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet. Coupling constants were reported in Hz. Chemical shits were referenced to residual ¹H and ¹³C signals at 7.27 and 77.0 ppm for deuterated chloroform.

2-Benzyl-2-aza-bicyclo[2.2.1]hept-5-ene (1)

Following the procedure by Grieco,^8a,b a mixture of benzylamine hydrochloride (2.14 g, 14.9 mmol), 37% w/w aqueous formaldehyde (1.7 g, 21 mmol), freshly distilled cyclopentadiene (2.0 g, 30 mmol) and water (10 mL) was stirred vigorously for 16 h at room temperature. The resulting mixture was diluted with water (20 mL) and washed with a 1:1 Et₂O-hexanes mixture (2 × 10 mL). The aqueous layer was made basic by the addition of 4 g of KOH pellets and the mixture was extracted with Et₂O (3 × 15 mL). The Et₂O extracts were combined, dried (Na₂SO₄) and concentrated. Purification of the residue on silica gel afforded 2.78 g (100%) of 1 as a colorless oil. ¹H and ¹³C NMR spectra were consistent with those reported in the literature. ^8a,b

3-Benzyl-1-(toluene-4-sulfonyl)-1,3,4,4a,5,7a-hexahydro-cyclopentapyrimidin-2-one (6)

Toluenesulfonyl isocyanate (0.10 g, 0.5 mmol) was added at room temperature to a stirred solution of aza-norbornene 1 (0.10 g, 0.5 mmol) in CHCl₃ (3 mL) under N₂. After stirring for 1 h, the reaction mixture was concentrated in vacuo and dissolved in EtOAc (10 mL). The organic layer was washed with water (2 × 10 mL) and brine (10 mL). The organic layer was dried over Na₂SO₄ and concentrated in vacuo. Purification of the crude product over silica gel, eluting with 3:2 (hexanes-EtOAc) gave 0.10 g (53%) of urea 6 as a colorless foam and 0.08 g (41%) of the isourea 7 as yellow oil. (6): mp 171-173 °C; ¹H NMR (500 MHz, CDCl₃) δ 8.02 (d, J = 8.3 Hz, 2H), 7.33 (d, J = 7.8 Hz, 2H), 7.27-7.26 (m, 3H), 7.15 (d, J = 5.8 Hz, 2H), 5.77-5.76 (m, 1H), 5.63-5.62 (m, 1H), 5.42 (d, J = 8.7 Hz, 1H), 4.50 (d, J = 14.6 Hz, 1H), 4.41 (d, J = 14.6 Hz, 1H), 3.30-3.27 (m, 1H), 2.93-2.86 (m, 2H), 2.43 (s, 3H), 2.41-2.40 (m, 1H), 1.83 (d, J = 14.6 Hz, 1H); ¹³C NMR (125 MHz, CDCl₃) 153.9, 143.8, 137.3, 136.3, 134.3, 129.9, 129.2, 128.4, 128.2, 127.5, 126.4, 65.1, 50.8, 47.5, 36.6, 36.1, 21.9 ppm; IR (ATR) υ_max 2920, 1675, 1342, 1163 cm^-1; MS (ESI-TOF) calcd for C₂₁H₂₂N₂O₃S [M + Na]⁺ 405.1249, found 405.1252.

N-(3-Benzyl-4,4a,5,7a-tetrahydro-3H-cyclopenta[e][1,3]oxazin-2-ylidene)-4-methyl-benzenesulfonamide (7)

Yield 41%; ¹H NMR (500 MHz, CDCl₃) δ 7.83 (d, J = 8.3 Hz, 2H), 7.31-7.26 (m, 5H), 7.21 (d, J = 8.3 Hz, 2H), 5.89-5.85 (m, 1H), 5.60-5.57 (m, 1H), 4.74 (d, J = 15.1 Hz, 2H), 4.59 (d, J = 15.1 Hz, 1H), 3.32 (dd, J = 5.8 Hz, 1H), 2.99 (dd, J = 4.8 Hz, 1H), 2.38 (s, 3H), 1.98-1.84 (m, 1H), 1.43 1.36 (m, 2H); ¹³C NMR (125 MHz, CDCl₃) 158.4, 141.7, 141.0, 137.3, 135.1, 128.8, 128.7, 128.5, 128.3, 128.1, 127.0, 87.8, 53.5, 46.2, 36.6, 35.8, 21.4 ppm; IR (ATR) υ_max 2919, 1567, 1471, 1086, 886 cm^-1; MS (ESI-TOF) calcd for C₂₁H₂₂N₂O₃S [M + Na]⁺ 405.1249, found 405.1245.

3-Benzyl-1-(toluene-4-sulfonyl)-1,3,4,4a,5,7a-hexahydro-cyclopentapyrimidine-2-thione (9)

p-Toluenesulfonyl isothiocyanate (0.11 g, 0.5 mmol) was added to a solution of aza-norbornene 1 (0.09 g, 0.5 mmol) in CHCl₃ (3 mL) under N₂ with stirring at room temperature. After stirring for 2 h, the reaction mixture was concentrated in vacuo and dissolved in EtOAc (10 mL). The organic layer was washed with water (2 × 10 mL) and brine (10 mL), dried over Na₂SO₄ and concentrated in vacuo. Purification of the residue over silica gel, eluting with 60:40 (hexanes-EtOAc) afforded the thiourea 9 (0.08 g, 39%) and isothiourea 10 (0.11 g, 52%) as a colorless solid. (9): mp: 56-58 °C; ¹H NMR (500 MHz, CDCl₃) δ 8.08 (d, J = 8.3 Hz, 2H), 7.33 (d, J = 8.3 Hz, 2H), 7.31-7.26 (m, 5H), 5.75-5.73 (m, 1H), 5.70-5.69 (m, 1H), 5.32-5.29 (m, 1H), 5.23 (d, J = 14.6 Hz, 1H), 4.98 (d, J = 14.6 Hz, 1H), 3.60-3.56 (m, 1H), 3.26 (m, 1H), 3.13 (d, J = 12.6 Hz, 1H), 2.44 (s, 3H), 2.42-2.40 (m, 1H), 1.63-1.60 (m, 1H); ¹³C NMR (125 MHz, CDCl₃) 182.4, 143.9, 136.8, 135.8, 135.3, 129.1, 129.0, 128.6, 128.5, 128.1, 127.5, 66.3, 56.6, 51.7, 39.7, 37.1, 21.6 ppm; IR (ATR) υ_max 2920, 1599, 1484, 1159, 1096 cm^-1; MS (ESI-TOF) calcd for C₂₁H₂₂N₂O₂S₂ [M + Na]⁺ 421.1021, found 421.1015.

N-(3-Benzyl-4,4a,5,7a-tetrahydro-3H-cyclopenta[e][1,3]thiazin-2-ylidene)-4-methyl-benzenesulfonamide (10)

Yield 52%; mp: 157-160 °C; ¹H NMR (500 MHz, CDCl₃) δ 7.84 (d, J = 8.3 Hz, 2H), 7.32-7.23 (m, 7H), 5.65-5.63 (m, 1H), 5.22-5.19 (m, 1H), 4.90 (d, J = 14.1 Hz, 1H), 4.74 (d, J = 14.1 Hz, 1H), 4.29 (d, J = 9.27 Hz, 1H), 3.44-3.41 (m, 1H), 3.22-3.18 (m, 1H), 2.91-2.85 (m, 1H), 2.57-2.48 (m, 1H), 2.40 (s, 3H), 1.91 (d, J = 17.1 Hz, 1H); ¹³C NMR (125 MHz, CDCl₃) 166.4, 141.9, 139.9, 135.5, 132.6, 129.8, 128.7, 128.6, 128.4, 128.0, 127.1, 55.0, 53.4, 52.6, 39.0, 36.9, 21.4 ppm; IR (ATR) υ_max 2917, 1522, 1441, 1144, 907 cm^-1; MS (ESI-TOF) calcd for C₂₁H₂₂N₂O₂S₂ [M + Na]⁺ 421.1021, found 421.1011.

2-Benzyl-2-aza-bicyclo[2.2.2]oct-5-ene (2)

The same procedure was used as in the synthesis of 1 except that 1,3-cyclohexadiene (2.38 g, 29.8 mmol) was used instead of cyclopentadiene. Purification of the residue over silica gel, eluting with 70:30 (hexanes-EtOAc) gave 1.69 g (57%) of 2 as a colorless oil. ¹H and ¹³C NMR spectra were consistent with those reported in the literature.^8a,b

3-Benzyl-1-(toluene-4-sulfonyl)-3,4,4a,5,6,8a-hexahydro-1H-quinazolin-2-one (11)

Toluenesulfonyl isocyanate (0.09 g, 0.5 mmol) was added to a solution of azabicyclo[2.2.2]octene 2 (0.1 g, 0.5 mmol) in C₆H₆ (1.5 mL) at room temperature. The mixture was then heated at reflux for 12 h. The resulting reaction mixture was concentrated in vacuo and dissolved in EtOAc (15 mL). The organic layer was washed with water (2 × 15 mL) and brine (15 mL), dried over Na₂SO₄ and concentrated in vacuo. Purification of the residue over silica gel, eluting with 7:3 (hexanes-EtOAc) gave (0.091 g, 48%) of urea 11 as a pale yellow solid and (0.095 g, 46%) of the isourea 12 as a yellow oil. 11: yield 48%; mp: 141-144 °C; ¹H NMR (500 MHz, CDCl₃) δ 7.95 (d, J = 8.1 Hz, 2H), 7.32 (d, J = 7.9 Hz, 2H), 7.27-7.23 (m, 3H), 7.13 (d, J = 6.2 Hz, 2H), 5.81(d, J = 10.1 Hz, 1H), 5.73 (d, J = 10.1 Hz, 1H), 5.20-5.15 (m, 1H), 4.65 (d, J = 15.0 Hz, 1H), 4.27 (d, J = 15.2 Hz, 1H), 3.25 (t, J = 11.8 Hz,1H), 2.94-2.87 (m, 1H), 2.42 (s, 3H), 2.40-2.37 (m, 1H), 2.07-1.98 (m, 1H), 1.97-1.88 (m, 2H), 1.76-1.65 (m, 1H); ¹³C NMR (125 MHz, CDCl₃) 151.2, 143.7, 138.0, 136.4, 129.1, 128.5, 128.2, 127.7, 127.4, 127.3, 54.8, 50.8, 44.5, 29.5, 23.4, 21.5, 20.3 ppm; IR (ATR) υ_max 2921, 1669, 1491, 1163 cm^-1; MS (ESI-TOF) calcd for C₂₂H₂₄N₂O₃S [M + Na]⁺ 419.1405, found 419.1404.

N-(3-Benzyl-3,4,4a,5,6,8a-hexahydro-benzo[e][1,3]oxazin-2-ylidene)-4-methyl-benzenesulfonamide (12)

yield: 46%; ¹H NMR (500 MHz, CDCl₃) δ 7.83 (d, J = 8.32 Hz, 2H), 7.33-7.23 (m, 5H), 7.21 (d, J = 8.3 Hz, 2H), 5.89-5.84 (m, 1H), 5.60-5.56 (m, 1H), 4.74-4.70 (m, 2H), 4.59 (d, J = 14.6 Hz, 1H), 3.34-3.27 (m, 1H), 3.01-2.94 (m, 1H), 2.38 (s, 3H), 2.16-2.07 (m, 1H), 1.98-1.95 (m, 2H), 1.44-1.38 (m, 2H); ¹³C NMR (125 MHz, CDCl₃) 154.6, 141.3, 141.0, 135.2, 133.3, 128.7, 128.5, 128.2, 128.0, 127.1, 122.9, 73.9, 53.5, 46.6, 29.3, 23.3, 21.4, 21.1; IR (ATR) υ_max 2923, 1570, 1473, 1288, 1135 cm^-1; MS (ESI-TOF) calcd for C₂₂H₂₄N₂O₃S [M + Na]⁺ 419.1405, found 419.1404.

N-(3-Benzyl-3,4,4a,5,6,8a-hexahydro-benzo[e][1,3]thiazin-2-ylidene)-4-methyl-benzenesulfonamide (13)

p-Toluenesulfonyl isothiocyanate (0.053 g, 0.25 mmol) was added to a solution of azabicyclooctene 2 (0.049 g, 0.25 mmol) in benzene (2 mL) at room temperature. The reaction was maintained at room temperature. After stirring for 2 h, the reaction mixture was concentrated in vacuo and dissolved in EtOAc (5 mL). The organic layer was washed with water (2 × 5 mL) and brine (5 mL), dried over Na₂SO₄ and concentrated in vacuo. The residue was purified over silica gel, eluting with 7:3 (hexanes-EtOAc) and then 10:90 (CH₂Cl₂ –acetone) to afford isothiourea 13 (0.083 g, 81%) as a yellow oil. 13: ¹H NMR (500 MHz, CDCl₃) δ 7.8 (d, J = 8.1 Hz, 2H), 7.31-7.25 (m, 3H), 7.23-7.19 (m, 4H), 5.90-5.85 (m, 1H), 5.62-5.56 (m, 1H), 4.78 (s, 2H), 3.91-3.85 (m, 1H), 3.46 (dd, J = 4.7 Hz, 1H), 3.21 (dd, J = 6.8 Hz, 1H), 2.39 (s, 3H), 2.26-2.18 (m, 1H), 2.08-1.96 (m, 2H), 1.66-1.59 (m, 1H), 1.44-1.36 (m, 1H); ¹³C NMR (125 MHz, CDCl₃) 164.8, 141.9, 140.0, 135.6, 131.3, 128.8, 128.6, 128.2, 127.9, 126.8, 123.8, 55.0, 52.5, 42.4, 33.0, 23.4, 22.0, 21.3 ppm; IR (ATR) υ_max 2922, 1523, 1403, 1143, 912 cm^-1; MS (ESI-TOF) calcd for C₂₂H₂₄N₂O₂S₂ [M + Na]⁺ 435.1177, found 435.1173.

N-(1,3-Dibenzyl-1,3,4,4a,5,7a-hexahydro-cyclopentapyrimidin-2-ylidene)-4-methylbenzenesulfonamide (15)

Aza-norbornene 1 (0.10 g, 0.54 mmol) and EtNi-Pr₂ (0.09 mL, 0.54 mmol) were added to a solution of EDCI (0.10 g, 0.54 mmol), thiourea 14 (0.17 g, 0.54 mmol) in CHCl₃ (5 mL) under N₂ at room temperature. After stirring overnight at room temperature, the reaction mixture was poured into a mixture of EtOAc (10 mL) and 0.25 M aqueous citric acid (10 mL). The layers were separated; the organic layer was washed with 0.25 M aqueous citric acid (5 mL), water (2 × 5 mL), brine (5 mL), dried (Na₂SO₄) and concentrated. Purification of the crude product over silica gel, eluting with 2:3 (hexanes-EtOAc) gave 0.170 g (67%) of 15 as a colorless oil. 15: ¹H NMR (500 MHz, CDCl₃) δ 7.74 (d, J = 8.1 Hz, 2H), 7.66 (d, J = 8.4 Hz, 2H), 7.31- 7.05 (m, 10H), 5.76 (s, 2H), 5.22 (d, J = 15.2 Hz, 1H), 4.79 (d, J = 14.6 Hz, 1H), 4.62 (d, J = 14.5 Hz, 1H), 4.39 (d, J = 15.2 Hz, 2H), 4.26 (d, J = 9.0 Hz, 1H), 3.20 (m, 1H), 2.98 (m, 1H), 2.54 (m, 1H), 2.30 (s, 3H), 1.72 (d, J = 17.1 Hz, 1H); ¹³C NMR (125 MHz, CDCl₃) 157.9, 143.5, 140.5, 137.2, 136.4, 133.5, 128.9, 128.8, 128.7, 128.6, 128.4, 127.9, 127.8, 127.5, 125.3, 64.0, 54.7, 53.6, 47.4, 37.2, 36.3, 20.9 ppm; FTIR (film) 1527 cm^-1; MS (FAB-magnetic sector) m/z 478.2128 (MLi, 478.2141 calcd for C₂₈H₂₉N₃O₂SLi).

(3-Benzyl-1-hexyl-1,3,4,4a,5,7a-hexahydro-cyclopentapyrimidin-2-ylidene)-carbamic acid ethyl ester (17)

Aza-norbornene 1 (0.26 g, 1.4 mmol) and EtNi-Pr₂ (0.37 g, 2.8 mmol) were added to a solution of 2-chloro-1-methyl pyridinium iodide (0.37 g, 1.4 mmol) and thiourea 16¹⁷ (0.34 g, 1.4 mmol), in CHCl₃ (10 mL) under N₂ at room temperature. The mixture was then allowed to heat at 80 °C for 5h with constant stirring. The reaction mixture was then cooled to room temperature and concentrated in vacuo to remove chloroform, and the residue was poured into EtOAc (10 mL) and washed with water (3 × 10 mL) and brine (10 mL). The layers were separated; the organic layer was dried over Na₂SO₄ and concentrated in vacuo. Purification over silica gel, eluting with 3:7 (hexanes-EtOAc) gave 0.43 g (76%) of 17 as a brown oil. 17: ¹H NMR (500 MHz, CDCl₃) δ 7.31-7.29 (m, 5H), 5.82 (s, 2H), 4.57 (dd, J = 14.6 Hz, J = 37.1 Hz 2H), 4.45 (d, J = 9.7 Hz, 1H), 4.07 (q, J = 6.8 Hz, 2H), 3.56-3.50 (m, 1H), 3.39-3.34 (m, 1H), 3.11-3.07 (m, 1H), 2.94-2.90 (m, 1H), 2.66-2.60 (m, 1H), 2.45-2.40 (m, 1H), 1.83 (d, J = 15.6 Hz, 1H), 1.70-1.62 (m, 2H), 1.31 (s, 6H), 1.24 (t, J = 6.8 Hz, 3H), 0.89 (t, J = 6.3 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 161.5, 158.7, 136.5, 132.6, 129.4, 128.5, 128.4, 127.5, 65.0, 60.3, 53.3, 49.4, 46.8, 36.4, 36.3, 31.3, 27.9, 26.4, 22.4, 14.8, 13.8 ppm; IR (ATR) υ_max 2928, 2233, 1632, 1515, 1064 cm^-1; MS (ESI-TOF) calcd for C₂₃H₃₃N₃O₂ [M + H]⁺ 384.2651, found 384.2645.

3-Benzyl-2-tert-butoxycarbonylimino-2,3,4,4a,5,7a-hexahydro-cyclopentapyrimidine-1-carboxylic acid tert-butyl ester (19)

The same procedure was used as in the synthesis of guanidine 15 except that thiourea 18¹⁸ (0.24 g, 0.86 mmol) was used instead of 14. Purification of the residue over silica gel eluting with 1:1 (hexanes-EtOAc) afforded the guanidine 19 (0.19 g, 51%) as a colorless solid: mp 113-115 °C; ¹H NMR (500 MHz, CDCl₃) δ 7.30-7.26 (m, 5H), 5.79 (d, J = 23.9 Hz, 2H), 5.50 (d, J = 9.2 Hz, 1H), 4.85 (d, J = 14.6, 1H), 4.58 (d, J = 14.64 Hz, 1H), 3.35 (dd, J = 4.3 Hz, 1H), 2.95-2.90 (m, 1H), 2.88 (d, J = 13.1 Hz, 1H), 2.42 -2.36 (m, 1H), 1.66 (d, J = 16.1 Hz, 1H), 1.51 (d, J = 5.3 Hz, 18 H); ¹³C NMR (125 MHz, CDCl₃) 159.7, 154.0, 151.5, 136.6, 134.4, 129.8, 128.8, 128.4 127.7, 82.1, 78.3, 63.2, 52.7, 48.9, 38.0, 37.1, 28.5, 28.2 ppm; IR (ATR) υ_max 2975, 2852, 1714, 1674, 1579, 1366, 1131 cm-1; MS (ESI-TOF) calcd for C₂₄H₃₃N₃O₄ [M + H]⁺ 428.2549, found 428.2545.

2,2,5,7,8-Pentamethyl-chroman-6-sulfonic acid (1,3-dibenzyl- 1, 3, 4, 4a, 5, 7a-hexahydro-cyclopentapyrimidin-2-ylidene)-amide (21)

The same procedure was used as in the synthesis of guanidine 15 except that thiourea 20⁹ (0.519 g, 1.20 mmol) was used instead of 14. Purification of the residue over silica gel eluting with 2:3 (hexanes-EtOAc) gave 0.596 g (85%) of 21 as a colorless solid. 21: ¹H NMR (500 MHz, CDCl₃) δ 7.32-7.20 (m, 10H), 5.83-5.80 (m, 1H), 5.78-5.75 (m, 1H), 5.28 (d, J = 15.4 Hz, 1H), 4.82 (d, J = 14.7 Hz, 1H), 4.53 (d, J = 14.7 Hz, 1H), 4.43 (d, J = 15.4 Hz, 1H), 4.28-4.25 (m, 1H), 3.25-3.21 (m, 1H), 3.00-2.96 (m, 1H), 2.57 (s, 4H), 2.54 (s, 3H), 2.38-2.31 (m, 1H), 2.07 (s, 3H), 1.76 (t, J = 6.7 Hz, 3H), 1.26-1.25 (m, 8H); ¹³C NMR (125 MHz, CDCl₃) 157.7, 152.2, 137.5, 136.5, 136.4, 134.3, 133.5, 133.1, 129.1, 128.6, 128.3, 128.1, 127.8, 127.6, 127.3, 123.0, 117.2, 73.3, 64.3, 55.1, 54.2, 48.0, 37.9, 37.0, 33.3, 26.7, 21.4, 18.8, 17.5, 12.0 ppm.

N-(1,3-Dibenzyl-1,3,4,4a,5,7a-hexahydro-cyclopentapyrimidin-2-ylidene)-C,C,C-trifluoro-methanesulfonamide (22)

Aza-norbornene 1 (0.098g, 0.53 mmol) and EtNi-Pr₂ (0.14 g, 1.07 mmol) were added to a solution of 2-chloro-1-methyl pyridinium iodide (0.13 g, 0.53 mmol) and thiourea 5 (0.15 g, 0.53 mmol) in CHCl₃ (10 mL) under N₂ at room temperature. After stirring overnight at room temperature, the reaction mixture was concentrated in vacuo to remove chloroform, and the residue was poured into EtOAc (10 mL) and washed with water (2 × 15 mL) and brine (10 mL). The layers were separated; the organic layer was dried over Na₂SO₄ and concentrated in vacuo. Purification of the crude product over silica gel, eluting with 60:40 (hexanes-EtOAc) gave 0.14 g (62%) of 22 as a colorless oil 22: ¹H NMR (500 MHz, CDCl₃) δ 7.39-7.29 (m, 10H), 5.83 (d, J = 12.2 Hz, 2H), 5.41 (d, J = 15.1 Hz, 1H), 4.98 (d, J = 15.1 Hz, 1H), 4.66 (d, J = 14.6 Hz, 1H), 4.44 (d, J = 15.1 Hz, 1H), 4.29 (d, J = 8.7 Hz, 1H), 3.20 (dd, J = 4.8 Hz, J = 13.1 Hz, 1H), 3.04 (dd, J = 5.8 Hz, J = 12.7 Hz, 1H), 2.57-2.53 (m, 1H), 2.40-2.35 (m, 1H), 1.83 (d, J = 15.1 Hz, 1H); ¹³C NMR (125 MHz, CDCl₃) 158.3, 136.0, 135.2, 134.0, 128.9, 128.8, 128.3, 128.2, 121.3 (q), 64.1, 55.3, 53.8, 47.6, 36.9, 36.4 ppm; IR (ATR) υ_max 2929, 1541, 1507, 1320, 737 cm^-1 MS (ESI-TOF) calcd for C₂₂H₂₂F₃N₃O₂S [M + Na]⁺ 472.1283, found 472.1282.

N-(1,3-Dibenzyl-3,4,4a,5,6,8a-hexahydro-1H-quinazolin-2-ylidene)-C,C,C-trifluoromethanesulfonamide (23)

Aza-bicyclo[2.2.2]octene 2 (0.19 g, 1 mmol) and EtNi-Pr₂ (0.51 g, 4 mmol), were added to a solution of 2-chloro-1-methyl pyridinium iodide (0.51 g, 2 mmol) and thiourea 5 (0.59 g, 2 mmol) in CHCl₃ (8 mL) at room temperature under N₂ with stirring. The resulting mixture was heated at 80 °C for 10h after which time it was concentrated. The resulting crude was dissolved in EtOAc (15 mL) and washed with water (2 × 20 mL) and brine (15 mL). The combined organic layers were dried over Na₂SO₄ and concentrated. Purification of the crude product over silica gel eluting with 60:40 (hexanes-EtOAc) afforded 0.26 g (57%) of the guanidine 23 as a yellow oil. ¹H NMR (500 MHz, CDCl₃) δ 7.37-7.29 (m, 10 H), 5.79-5.75 (m, 1H), 5.64-5.62 (m, 1H), 5.53 (d, J = 15.1 Hz, 1H), 5.09 (d, J = 14.6 Hz, 1H), 4.69 (d, J = 14.6 Hz, 1H), 4.44 (d, J = 15.1 Hz, 1H), 3.81-3.79 (m, 1H), 3.14-31.0 (m, 2H), 2.12-2.02 (m, 1H), 1.96-1.88 (m, 1H), 1.61-1.56 (m, 2H), 1.25-1.23 (m, 1H); ¹³C NMR (125 MHz, CDCl₃) 155.0, 135.4, 135.1, 130.3, 128.9, 128.7, 128.3, 128.1, 128.2, 123.7, 121.4 (q), 55.9, 53.7, 52.8, 46.0, 28.8, 21.9, 21.4, 21.0 ppm; IR (ATR) υ_max 2929, 1553, 1497, 1314, 1206, 911 cm^-1; MS (ESI-TOF) calcd for C₂₃H₂₄F₃N₃O₂S [M + H]⁺ 464.1620, found 464.1617.

2-Benzyl-2-aza-bicyclo[2.2.1]hept-5-ene-3-carboxylic acid ethyl ester (3a, 3b)

Benzylamine hydrochloride (29 g, 0.21 mol) was dissolved in water (25 mL), to which a 50% solution of ethyl glyoxylate in toluene (25 g, 0.24 mol) and cyclopentadiene (26 g, 0.40 mol) were added. The mixture was stirred at room temperature until TLC indicated no further accumulation of adduct (35 h). The resulting aqueous mixture was washed with ether (2 × 30 mL), the pH adjusted to 9-10 with NaHCO₃ and the layer extracted with EtOAc (30 mL). The organic layer was separated, washed with brine (30 mL), dried over MgSO₄ and concentrated in vacuo. Purification of the crude on silica gel, eluting with 70:30 (hexanes-EtOAc) gave (30.3 g, 53%) of the exo isomer and (14.0 g, 27%) of the endo isomer. The spectral data was consistent with published data.^8c

3-Benzyl-2-oxo-1-(toluene-4-sulfonyl)-2,3,4,4a,5,7a-hexahydro-1H-cyclopentapyrimidine-4-carboxylic acid ethyl ester (30)

p-Toluenesulfonyl isocyanate (0.11 mL, 0.77 mmol) was added to a solution of the endo-ethoxycarbonyl aza-norbornene derivative 3a (0.20 g, 0.77 mmol) in CHCl₃ (5 mL) under N₂ with stirring at room temperature. After stirring for 2 h, the reaction mixture was concentrated in vacuo and dissolved in EtOAc (10 mL). The organic layer was washed with water (2 × 10 mL) and brine (10 mL), dried over Na₂SO₄ and concentrated in vacuo. Purification of the residue over silica gel, eluting with 70:30 (hexanes-EtOAc) gave (0.08 g, 24%) of the urea 30 and (0.22 g, 64%) of isourea 31 as colorless oil. (30): ¹H NMR (500 MHz, CDCl₃) δ 8.00 (d, J = 8.3 Hz, 2H), 7.33 (d, J = 8.3 Hz, 2H), 7.29-7.27 (m, 3H), 7.19-7.16 (m, 2H), 5.74-5.71 (m, 1H), 5.53-5.49 (m, 1H), 5.42 (d, J = 9.7 Hz, 1H), 5.17 (d, J = 14.6 Hz, 1H), 4.26-4.16 (m, 2H), 3.90 (d, J = 14.6 Hz, 1H), 3.58 (s, 1H), 3.54-3.48 (m, 1H), 2.43 (s, 3H), 2.39-2.32 (m, 1H), 1.63-1.56 (m, 1H), 1.29 (t, J = 6.8 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 167.4, 157.3, 141.7, 140.6, 135.1, 134.3, 128.8, 128.7, 128.67, 128.6, 128.2, 126.9, 85.5, 61.7, 56.7, 52.6, 38.7, 33.4, 21.3, 13.4 ppm; IR (ATR) υ_max 2924, 1735, 1696, 1422, 1164 cm^-1; MS (ESI-TOF) calcd for C₂₄H₂₆N₂O₅S [M + Na]⁺ 477.1460, found 477.1445.

3-Benzyl-2-(toluene-4-sulfonylimino)-2,3,4,4a,5,7a-hexahydro-cyclopenta[e][1,3]oxazine-4-carboxylic acid ethyl ester (31)

Yield 64%; ¹H NMR (500 MHz, CDCl₃) δ 7.84 (d, J = 8.3 Hz, 2H), 7.35-7.30 (m, 3H), 7.27-7.23 (m, 4H), 5.86-5.82 (m, 1H), 5.36-5.31 (m, 3H), 4.20-4.07 (m, 2H), 4.05 (d, J = 14.6 Hz, 1H), 3.69 (s, 1H), 3.34-3.29 (m, 1H), 2.39 (s, 3H), 2.36-2.35 (m, 1H), 1.60 (d, J = 17.5 Hz, 1H), 1.15 (t, J = 6.8 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 168.9, 157.5, 141.7, 140.7, 138.3, 134.2, 129.3, 128.6, 128.4, 127.9, 126.8, 86.3, 62.2, 58.5, 52.8, 37.9, 36.5, 21.1, 13.7 ppm; IR (ATR) υ_max 2249, 1738, 1563, 1465, 1151 cm^-1; MS (ESI-TOF) calcd for C₂₄H₂₆N₂O₅S [M + Na]⁺ 477.1460, found 477.1452.

1,3-Dibenzyl-2-oxo-2,3,4,4a,5,7a-hexahydro-1H-cyclopentapyrimidine-4-carboxylic acid ethyl ester (33)

Benzyl isocyanate (0.51 g, 3.8 mmol) was added to the endo-ethoxycarbonyl aza-norbornene derivative 3a (0.49 g, 1.9 mmol) under N₂ and the mixture was allowed to heat at 120 °C for 12 h under solvent free conditions. Purification of the resulting deep brown crude over silica gel, eluting with 65:35 (hexanes-EtOAc) gave (0.23 g, 38%) of the urea 33 as a colorless oil. 33: ¹H NMR (500 MHz, CDCl₃) δ 7.37-7.20 (m, 10H), 5.84-5.80 (m, 1H), 5.77-5.73 (m, 1H), 5.38 (d, J = 14.6 Hz, 1H), 5.24 (d, J = 15.1 Hz, 1H) 4.20-4.07 (m, 3H), 3.92-3.84 (m, 2H), 3.61 (s, 1H), 3.24-3.15 (m, 1H), 2.38-2.29 (m, 1H), 2.10-1.95 (m, 1H), 1.21 (t, J = 7.3 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 171.6, 158.5, 138.2, 137.6, 134.4, 129.1, 128.8, 128.4, 128.3, 128.1, 127.5, 127.1, 61.5, 61.3, 59.2, 50.8, 48.8, 38.7, 36.7, 14.2 ppm; IR (ATR) υ_max 1730, 1645, 1227 cm^-1; MS (ESI-TOF) calcd for C₂₄H₂₆N₂O₃ [M + Na]⁺ 413.1841, found 413.1837.

1,3-dibenzyl-5-ethoxyimidazolidine-2,4-dione (34)

Yield 39%. ¹H NMR (500 MHz, CDCl₃) δ 7.35-7.28 (m, 10H), 4.92-4.88 (d, J = 14.65 Hz, 1H), 4.86 (s, 1H), 4.75- 4.63 (q, J = 14.40 Hz, 2H), 4.24 - 4.20 (d, J = 14.91 Hz, 1H), 3.61- 3.54 (m, 1H), 3.42- 3.35 (m, 1H), 1.19- 1.15 (t, J = 7.07 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 169.1, 155.2, 135.2, 135.7, 135.3, 128.8, 128.7, 128.6, 128.1, 127.9, 82.6, 62.2, 44.0, 42.3, 14.8 ppm; IR (ATR) υ_max 3078, 2922, 177, 1683, 1557, 1202, 695 cm^-1; MS (ESI-TOF) calcd for C₁₉H₂₀N₂O₃ [M + Na]⁺ 347.1372, found. 347.1374.

3-Benzyl-2-oxo-1-(toluene-4-sulfonyl)-2,3,4,4a,5,7a-hexahydro-1H-cyclopentapyrimidine-4-carboxylic acid ethyl ester (35)

p-Toluenesulfonyl isocyanate (0.207 g, 1.05 mmol) was added to a solution of the exo-ethoxycarbonyl aza-norbornene derivative 3b (0.27 g, 1.05 mmol) at room temperature. After stirring for 1 h, the reaction mixture was concentrated in vacuo and dissolved in EtOAc (15 mL). The organic layer was washed with water (2 × 15 mL) and brine (15 mL), dried over Na₂SO₄ and concentrated in vacuo. Purification of the crude product over silica gel, eluting with 3:2 (hexanes-EtOAc) gave (0.22 g, 47%) of the urea 35 and (0.20 g, 44%) of isourea 36 as a white solid. 35: mp: 125- 127 °C; ¹H NMR (500 MHz, CDCl₃) δ 7.96 (d, J = 8.7 Hz, 2H), 7.34 (d, J = 8.3 Hz, 2H), 7.29-7.23 (m, 3H), 7.14 (d, J = 7.8 Hz, 2H), 6.12-6.06 (m, 1H), 5.77-5.73 (m, 1H), 5.27-5.23 (m, 1H), 4.97 (d, J = 14.6 Hz, 1H), 4.01-3.90 (m, 2H), 3.88 (d, J = 15.1 Hz, 1H), 3.65 (d, J = 5.8 Hz, 1H), 2.94-2.86 (m, 1H), 2.67-2.59 (m, 1H), 2.55-2.44 (m, 1H), 2.44 (s, 3H), 1.14 (t, J = 7.3 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 169.2, 152.7, 143.9, 137.4, 135.6, 133.4, 130.9, 129.2, 128.6, 128.5, 128.2, 127.7, 63.6, 61.7, 57.8, 50.1, 39.1, 32.8, 21.5, 13.8 ppm; IR (ATR) υ_max 1738, 1557, 1298, 733 cm^-1; MS (ESI-TOF) calcd for C₂₄H₂₆N₂O₅S [M + Na]⁺ 477.1460, found 477.1462.

3-Benzyl-2-(toluene-4-sulfonylimino)-2,3,4,4a,5,7a-hexahydro- cyclopent[e] [1,3]oxazine-4-carboxylic acid ethyl ester (36)

Yield 44%; mp: 54-56 °C; ¹H NMR (500 MHz, CDCl₃) δ 7.89 (d, J = 8.3 Hz, 2H), 7.30-7.26 (m, 5H), 7.25 (d, J = 9.2 Hz, 2H), 5.89-5.84 (m, 1H), 5.63-5.58 (m, 1H), 5.31-5.26 (m, 1H), 5.01 (d, J = 14.8 Hz, 1H), 4.33 (d, J = 14.8 Hz, 1H), 3.90-3.84 (m, 2H), 3.81-3.74 (m, 1H), 3.04-2.98 (m, 1H), 2.61 (d, J = 15.4 Hz, 1H), 2.46-2.40 (m, 1H), 2.38 (s, 3H), 0.93 (t, J = 7.0 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 167.4, 157.3, 141.7, 140.7, 135.1, 134.3, 128.8, 128.7, 128.6, 128.5, 128.2, 126.9, 85.5, 61.7, 56.7, 52.6, 38.7, 33.4, 21.3, 13.4 ppm; IR (ATR) υ_max 1735, 1676, 1168, 663 cm^-1; MS (ESI-TOF) calcd for C₂₄H₂₆N₂O₅S [M + Na]⁺ 477.1460, found 477.1452.

2-Benzyl-2-aza-bicyclo[2.2.2]oct-5-ene-3-carboxylic acid ethyl ester (4a, 4b)

The same procedure was followed as in the synthesis of 3a, 3b except 1,3-cyclohexadiene (21 g, 0.26 mmol) was used instead of cyclopentadiene. Purification of the crude product over silica gel, eluting with 70:30 (hexanes-EtOAc) gave (22.7 g, 42%) of the exo isomer and (8.1 g, 23%) of the endo isomer. The spectral data was consistent with published data.^8c

3-Benzyl-1-(toluene-4-sulfonyl)-3,4,4a,5,6,8a-hexahydro- 1H-quinazolin-2-one (37)

The same procedure was followed as in the synthesis of 30 except p-TsNCO (0.19 g, 0.96 mmol) was added to the aza-bicyclooctene 4a (0.1 g 0.5 mmol) at room temperature. The resulting mixture was heated at 80 °C for 12 h under solvent-free conditions. Purification of the residue over silica gel, eluting with 65:35 (hexanes-EtOAc) gave (0.09 g, 48%) of the urea 37 as a colorless oil. ¹H NMR (500 MHz, CDCl₃) 7.97 (d, J = 8.5 Hz, 2H), 7.33 (d, J = 8.5 Hz, 2H), 7.26-7.24 (m, 3H), 7.14-7.12 (m, 2H), 5.72 (m, 2H), 5.11 (d, J = 15 Hz, 1H), 5.02 (s, 1H), 4.22-4.11 (m, 2H), 3.82 (d, J = 15 Hz, 1H), 2.68-2.64 (m, 1H), 2.44 (s, 3H), 1.95-1.68 (m, 4H), 1.26 (t, J = 7 Hz, 3H); ¹³C NMR (125 MHz, CDCl₃) 170.6, 151.7, 144.0, 137.6, 135.3, 129.2, 128.9, 125.9, 61.9, 58.5, 52.9, 49.9, 36.3, 24.1, 21.5, 20.8, 14.1 ppm; IR (ATR) υ_max 1619, 903 cm^-1; MS (ESI-TOF) calcd for C₂₅H₂₈N₂O₅S [M + Na]⁺ 491.1617, found 491.1621.

1-Benzoyl-3-benzyl-1,3,4,4a,5,7a-hexahydro-cyclopentapyrimidin-2-one (47)

The same procedure was used as in the synthesis of 6 except that benzoyl isocyanate (0.17 g, 1.1 mmol) was used instead of p-TsNCO. Purification of the residue over silica gel, eluting with 2:3 (hexanes-EtOAc) gave 0.35 g (90%) of 47 as a colorless oil. ¹H NMR (500 MHz, CDCl₃) δ 7.59-7.24 (m, 10 H), 5.97-5.82 (m, 2H), 5.50-5.36 (m, 1H), 4.61 (d, J = 14.6 Hz, 1H), 4.36 (d, J = 14.6 Hz, 1H), 3.62-3.46 (m, 1H), 3.20-3.09 (m, 1H), 3.03-2.93 (m, 1H), 2.68-2.50 (m, 1H), 2.03-1.88 (m, 1H); ¹³C NMR (125 MHz, CDCl₃) 173.0, 156.7, 137.1, 136.6, 134.0, 131.1, 130.1, 128.6, 128.5, 128.0, 127.9, 127.7, 63.2, 52.0, 48.8, 37.4, 36.5 ppm; FTIR (film) 1712, 1656 cm^-1; MS (ESI-ion trap) calcd for C₂₁H₂₀N₂O₂ [M+H]⁺ 333.1603, found 333.1598.

N- (benzylcarbamothioyl)trifluoromethanesulfonamide (5)

A solution of TfNH₂ (500 mg, 3.35 mmol) in THF (3 mL) was added drop wise to a suspension of NaH (148 mg, 3.68 g, 60% dispersion in mineral oil) in THF (2 mL) at 0 °C. The cooling bath was removed and the mixture was maintained at rt for 1 h. BnNCS (0.44 mL, 3.4 mmol) was added and the mixture was heated at reflux overnight. The reaction mixture was allowed to cool to rt and H₂O (2 mL) was added. The THF was removed on a rotary evaporator. The residue was taken up in H₂O (8 mL) and washed with CH₂Cl₂ (2 × 15 mL). Aqueous 2 M HCl (3 mL) was added to the aqueous layer and a precipitate formed. This mixture was extracted with CH₂Cl₂ (3 × 15 mL). The last three organic fractions were combined, dried over Na₂SO₄ and concentrated to afford thiourea 5 (840 mg, 88%) which was used without further purification. Note: Thiourea 5 is unstable and undergoes rapid decomposition and it exhibited decomposition on flash chromatography over silica gel. ¹H NMR (500 MHz, CDCl₃) 7.41-7.29 (m, 5H), 4.80 (d, J = 5.5 Hz, 2H); ¹³C NMR (125 MHz, CDCl₃) 175.9, 134.6, 128.9, 128.3, 127.7, 119.2 (q, J = 320 Hz), 50.2 ppm; IR (ATR) 3007, 1543, 1497 cm^-1.

General Procedure for BF₃·OEt₂ Catalyzed Isomerization

In a typical procedure, to a solution of aza-norbornene 1 (54.1 mg, 0.291 mmol) in dry benzene (3 mL) under nitrogen was added drop wise p-TsNCO (57.4 mg, 0.291 mmol) at ambient temperature. After stirring for 20 min, analysis by TLC indicated complete consumption of starting material and the presence of the urea/isourea isomers 6 and 7 respectively. The solution was then cooled to 0 °C followed by the drop wise addition of BF₃·OEt₂ (4.1 mg, 0.029 mmol). After the addition was complete, the cooling bath was removed and the reaction mixture was allowed to stir at room temperature. At the end of 15 min, TLC indicated complete disappearance of the isourea 7 and indicated the presence of only one isomer. The reaction mixture was then diluted with water (5 mL). The benzene was removed in vacuo, and the residue was diluted with EtOAc (5 mL) washed with water (2 × 5 mL). The combined organic layers were washed with brine, dried over MgSO₄ and concentrated. The residue was purified by coloumn chromatography on silica gel (20 % ethyl acetate in hexanes) to afford 6 96.5 mg, (87%) as a white foam identical to that prepared using the method for compounds in Table 1.