Synthesis of New 2-Halo-2-(1H-tetrazol-5-yl)-2H-azirines via a Non-Classical Wittig Reaction

The synthesis and reactivity of tetrazol-5-yl-phosphorus ylides towards N-halosuccinimide/TMSN3 reagent systems was explored, opening the way to new haloazidoalkenes bearing a tetrazol-5-yl substituent. These compounds were obtained as single isomers, except in one case. X-ray crystal structures were determined for three derivatives, establishing that the non-classical Wittig reaction leads to the selective synthesis of haloazidoalkenes with (Z)-configuration. The thermolysis of the haloazidoalkenes afforded new 2-halo-2-(tetrazol-5-yl)-2H-azirines in high yields. Thus, the reported synthetic methodologies gave access to important building blocks in organic synthesis, vinyl tetrazoles and 2-halo-2-(tetrazol-5-yl)-2H-azirine derivatives.


Introduction
2H-azirines are highly reactive and easily available compounds. Thus, they have been widely used as versatile building blocks for the synthesis of various nitrogen-containing compounds. They can act as nucleophiles, electrophiles, dienophiles, and dipolarophiles in a variety of organic reactions. Furthermore, selective cleavage of each of the three bonds can be achieved, and this leads to highly reactive intermediates such as vinylnitrenes, nitrile ylides, and iminocarbenes [1][2][3][4][5][6].
We have previously described a general route to tetrasubstituted alkenes via a non-classical Wittig reaction [7]. Particularly interesting was the possibility of preparing haloazidoalkenes since the study of their thermolysis led to the development of a new route to 2-halo-2H-azirines starting from α-oxophosphorus ylides [8][9][10]. This study allowed the synthesis of a range of 2-halo-2H-azirines with several substituents, including the first examples of 2-bromo and 2-iodo-2H-azirine derivatives. Since then, a few examples of halo substituted azirines prepared from haloazidoalkenes by thermal or photochemical decomposition have been reported [11][12][13][14].
Recently, we became interested in the development of synthetic routes to functionalized 5-(substituted)-1H-tetrazoles. In this context, the synthesis of novel 2-(tetrazol-5-yl)-2H-azirines using The synthesis of the target tetrazol-5-yl phosphorus ylides 6 is outlined in Scheme 2. N-Benzylchloroacetamide (2) was prepared in good yield from the reaction of benzylamine and chloroacetyl chloride by an analogous method to that described in the literature [23]. Chloroacetamide 2 was treated with phosphorus pentachloride, followed by addition of sodium azide and water to give 1-benzyl-5-chloromethyltetrazole (3) in 54% yield [24]. Reaction of chloromethyltetrazole 3 with triphenylphosphine afforded the corresponding phosphonium salt 4 in very high yield (90%), which was subsequently neutralized with aqueous sodium hydroxide solution over a short period of time with ice-cooling to give phosphorus ylide 5 bearing a tetrazolyl substituent in moderate yield (65%). As previously observed with other tetrazolic phosphorus ylides, phosphorane 5 was hydrolyzed in water to give triphenylphosphine oxide and 5-methyl-1H-tetrazole [25,26]. For this reason, in order to prevent this hydrolysis the base treatment of 4 was carried out in water for only 2 min with vigorous stirring and the resulting precipitate was filtered and immediately dried under reduced pressure. However, even with these controlled conditions mixtures of ylide and hydrolysis products were obtained making the purification procedure difficult. Reaction of phosphorus ylide 5 with ethyl oxalyl chloride and benzoyl chloride in the presence of triethylamine gave ylides 6a and 6b, respectively, in moderate yields (Scheme 2). Aiming to improve ylides of 6a and 6b yield and to overcome the difficulties observed in the synthesis of ylide 5, we tried to carry out the synthesis of ylides 6 starting directly from the phosphonium salt 4 in the Scheme 1. Synthetic strategy for the synthesis 2-halo-2-(1H-tetrazol-5-yl)-2H-azirines.

Results and Discussion
The synthesis of the target tetrazol-5-yl phosphorus ylides 6 is outlined in Scheme 2. N-Benzylchloroacetamide (2) was prepared in good yield from the reaction of benzylamine and chloroacetyl chloride by an analogous method to that described in the literature [23]. Chloroacetamide 2 was treated with phosphorus pentachloride, followed by addition of sodium azide and water to give 1-benzyl-5-chloromethyltetrazole (3) in 54% yield [24]. Reaction of chloromethyltetrazole 3 with triphenylphosphine afforded the corresponding phosphonium salt 4 in very high yield (90%), which was subsequently neutralized with aqueous sodium hydroxide solution over a short period of time with ice-cooling to give phosphorus ylide 5 bearing a tetrazolyl substituent in moderate yield (65%). As previously observed with other tetrazolic phosphorus ylides, phosphorane 5 was hydrolyzed in water to give triphenylphosphine oxide and 5-methyl-1H-tetrazole [25,26]. For this reason, in order to prevent this hydrolysis the base treatment of 4 was carried out in water for only 2 min with vigorous stirring and the resulting precipitate was filtered and immediately dried under reduced pressure. However, even with these controlled conditions mixtures of ylide and hydrolysis products were obtained making the purification procedure difficult.
Molecules 2015, 20, page-page 2 to 4-(tetrazol-5-yl)-1H-imidazoles, a class of compounds with potential biological activity [22]. Aiming to extend this approach to 5-substituted tetrazoles, we decided to prepare 2H-azirines combining halogen and tetrazole functionalities, since the presence of the extra functional group could be particularly interesting.

Results and Discussion
The synthesis of the target tetrazol-5-yl phosphorus ylides 6 is outlined in Scheme 2. N-Benzylchloroacetamide (2) was prepared in good yield from the reaction of benzylamine and chloroacetyl chloride by an analogous method to that described in the literature [23]. Chloroacetamide 2 was treated with phosphorus pentachloride, followed by addition of sodium azide and water to give 1-benzyl-5-chloromethyltetrazole (3) in 54% yield [24]. Reaction of chloromethyltetrazole 3 with triphenylphosphine afforded the corresponding phosphonium salt 4 in very high yield (90%), which was subsequently neutralized with aqueous sodium hydroxide solution over a short period of time with ice-cooling to give phosphorus ylide 5 bearing a tetrazolyl substituent in moderate yield (65%). As previously observed with other tetrazolic phosphorus ylides, phosphorane 5 was hydrolyzed in water to give triphenylphosphine oxide and 5-methyl-1H-tetrazole [25,26]. For this reason, in order to prevent this hydrolysis the base treatment of 4 was carried out in water for only 2 min with vigorous stirring and the resulting precipitate was filtered and immediately dried under reduced pressure. However, even with these controlled conditions mixtures of ylide and hydrolysis products were obtained making the purification procedure difficult. Reaction of phosphorus ylide 5 with ethyl oxalyl chloride and benzoyl chloride in the presence of triethylamine gave ylides 6a and 6b, respectively, in moderate yields (Scheme 2). Aiming to improve ylides of 6a and 6b yield and to overcome the difficulties observed in the synthesis of ylide 5, we tried to carry out the synthesis of ylides 6 starting directly from the phosphonium salt 4 in the Scheme 2. Synthesis of tetrazol-5-yl phosphorus ylides 6.
Reaction of phosphorus ylide 5 with ethyl oxalyl chloride and benzoyl chloride in the presence of triethylamine gave ylides 6a and 6b, respectively, in moderate yields (Scheme 2). Aiming to improve ylides of 6a and 6b yield and to overcome the difficulties observed in the synthesis of ylide 5, we tried to carry out the synthesis of ylides 6 starting directly from the phosphonium salt 4 in the presence of triethylamine. To our delight, carrying out the reaction of the phosphonium salt 4 with ethyl oxalyl chloride and benzoyl chloride in the presence of excess of triethylamine led to the formation of ylides 6a and 6b in 88% and 61% yield, respectively. The same methodology was applied to the synthesis of ylides 6d and 6e bearing a thiophenyl and a furanyl substituent, respectively, which were isolated in good yields (Scheme 2). On the other hand, reaction of ylide 5 with 5-nitro-furan-2-carboxylic acid in the presence of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI) and 4-dimethylaminopyridine (DMAP) afforded ylide 6c in high yield (80%).
These ylides reacted with N-halosuccinimides in the presence of azidotrimethylsilane giving the corresponding haloazidoalkenes 7a-h and 8 in yields ranging from 47% to 93% (Schemes 3 and 4). Higher yields were obtained when NCS/TMSN 3 were used as reagents in the reactions with all ylides 6. The reaction of NCS with ylide 6a in the presence of TMSN 3 led to the formation of the desired chloroazidoalkene 7a with the highest yield (93%). As for bromoazidoalkenes, the best result was obtained from the reaction of ylide 6e bearing a furanyl substituent with NBS/TMSN 3 reagent system which led to the formation of the corresponding bromoazidoalkene 7h in 57% yield. The reactions with N-chlorosuccinimide were completed after 1-1.5 h while the reactions with N-bromosuccinimide required longer periods of time (2-3 h). The azidoalkenes were obtained selectively as single isomers except in the case of 7b and 8 which was obtained as a mixture of E and Z isomers (61:39).
Molecules 2015, 20, page-page presence of triethylamine. To our delight, carrying out the reaction of the phosphonium salt 4 with ethyl oxalyl chloride and benzoyl chloride in the presence of excess of triethylamine led to the formation of ylides 6a and 6b in 88% and 61% yield, respectively. The same methodology was applied to the synthesis of ylides 6d and 6e bearing a thiophenyl and a furanyl substituent, respectively, which were isolated in good yields (Scheme 2). On the other hand, reaction of ylide 5 with 5-nitro-furan-2-carboxylic acid in the presence of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI) and 4-dimethylaminopyridine (DMAP) afforded ylide 6c in high yield (80%).
These ylides reacted with N-halosuccinimides in the presence of azidotrimethylsilane giving the corresponding haloazidoalkenes 7a-h and 8 in yields ranging from 47% to 93% (Schemes 3 and 4). Higher yields were obtained when NCS/TMSN3 were used as reagents in the reactions with all ylides 6. The reaction of NCS with ylide 6a in the presence of TMSN3 led to the formation of the desired chloroazidoalkene 7a with the highest yield (93%). As for bromoazidoalkenes, the best result was obtained from the reaction of ylide 6e bearing a furanyl substituent with NBS/TMSN3 reagent system which led to the formation of the corresponding bromoazidoalkene 7h in 57% yield. The reactions with N-chlorosuccinimide were completed after 1-1.5 h while the reactions with N-bromosuccinimide required longer periods of time (2-3 h). The azidoalkenes were obtained selectively as single isomers except in the case of 7b and 8 which was obtained as a mixture of E and Z isomers (61:39). In order to establish the stereochemistry of the synthetized alkenes compounds 7c, 7e and 7h, bearing a phenyl group, a thiophenyl group and a furanyl group at C-2′, respectively, were selected for X-ray crystallography studies. The three compounds crystallize in the same, monoclinic, space group (P21/c). The X-ray data unambiguously shows that the molecules adopt in the crystal the (Z)-configuration ( Figure 1). Although there is a significant freedom for rotation of substituents around single C-C bonds, no sign for disorder was found except for the thiophene ring in compound 7e, which features a minor disorder between two alternating positions related by a 180° rotation around the C2′′-C2′ bond with occupancies 67:33%. A selection of bond distance, bond angles and torsion angles is provided in Table 1. They are in agreement with typical average values and also to those of the XRD study of a bromo-azidoalkene reported in [17]. Cohesion of the crystal structures is provided by weak C-H···N hydrogen bonds and also C-H···Cg, Cg···Cg and Br···Cg interactions involving the aromatic rings ( Figure 2). Molecules 2015, 20, page-page presence of triethylamine. To our delight, carrying out the reaction of the phosphonium salt 4 with ethyl oxalyl chloride and benzoyl chloride in the presence of excess of triethylamine led to the formation of ylides 6a and 6b in 88% and 61% yield, respectively. The same methodology was applied to the synthesis of ylides 6d and 6e bearing a thiophenyl and a furanyl substituent, respectively, which were isolated in good yields (Scheme 2). On the other hand, reaction of ylide 5 with 5-nitro-furan-2-carboxylic acid in the presence of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI) and 4-dimethylaminopyridine (DMAP) afforded ylide 6c in high yield (80%).
These ylides reacted with N-halosuccinimides in the presence of azidotrimethylsilane giving the corresponding haloazidoalkenes 7a-h and 8 in yields ranging from 47% to 93% (Schemes 3 and 4). Higher yields were obtained when NCS/TMSN3 were used as reagents in the reactions with all ylides 6. The reaction of NCS with ylide 6a in the presence of TMSN3 led to the formation of the desired chloroazidoalkene 7a with the highest yield (93%). As for bromoazidoalkenes, the best result was obtained from the reaction of ylide 6e bearing a furanyl substituent with NBS/TMSN3 reagent system which led to the formation of the corresponding bromoazidoalkene 7h in 57% yield. The reactions with N-chlorosuccinimide were completed after 1-1.5 h while the reactions with N-bromosuccinimide required longer periods of time (2-3 h). The azidoalkenes were obtained selectively as single isomers except in the case of 7b and 8 which was obtained as a mixture of E and Z isomers (61:39). In order to establish the stereochemistry of the synthetized alkenes compounds 7c, 7e and 7h, bearing a phenyl group, a thiophenyl group and a furanyl group at C-2′, respectively, were selected for X-ray crystallography studies. The three compounds crystallize in the same, monoclinic, space group (P21/c). The X-ray data unambiguously shows that the molecules adopt in the crystal the (Z)-configuration ( Figure 1). Although there is a significant freedom for rotation of substituents around single C-C bonds, no sign for disorder was found except for the thiophene ring in compound 7e, which features a minor disorder between two alternating positions related by a 180° rotation around the C2′′-C2′ bond with occupancies 67:33%. A selection of bond distance, bond angles and torsion angles is provided in Table 1. They are in agreement with typical average values and also to those of the XRD study of a bromo-azidoalkene reported in [17]. Cohesion of the crystal structures is provided by weak C-H···N hydrogen bonds and also C-H···Cg, Cg···Cg and Br···Cg interactions involving the aromatic rings ( Figure 2). In order to establish the stereochemistry of the synthetized alkenes compounds 7c, 7e and 7h, bearing a phenyl group, a thiophenyl group and a furanyl group at C-2 1 , respectively, were selected for X-ray crystallography studies. The three compounds crystallize in the same, monoclinic, space group (P2 1 /c). The X-ray data unambiguously shows that the molecules adopt in the crystal the (Z)-configuration ( Figure 1). Although there is a significant freedom for rotation of substituents around single C-C bonds, no sign for disorder was found except for the thiophene ring in compound 7e, which features a minor disorder between two alternating positions related by a 180˝rotation around the C2 11 -C2 1 bond with occupancies 67:33%. A selection of bond distance, bond angles and torsion angles is provided in Table 1. They are in agreement with typical average values and also to those of the XRD study of a bromo-azidoalkene reported in [17]. Cohesion of the crystal structures is provided by weak C-H¨¨¨N hydrogen bonds and also C-H¨¨¨Cg, Cg¨¨¨Cg and Br¨¨¨Cg interactions involving the aromatic rings ( Figure 2).
Since 7d, 7f and 7g differ from 7c, 7e and 7h only in the nature of the halogen, the (Z)-configuration is therefore proposed for all of these compounds. In previous studies, we could confirm that our synthetic methodology allowed the synthesis of a bromoazidoalkene bearing a carboxylate group at C-1′ and a phenyl group at C-2′ with the same selectivity [17]. Thus, the stereochemistry outcome is retained when the carboxylate group is replaced by a tetrazolyl group (7c).
The synthesis of the haloazidoalkenes can be rationalized as outlined in Scheme 5. The formation of the observed products can be explained by considering isomeric halonium ions 10 and 11 as intermediates. These halonium ions can interconvert by way of acyclic cation 9. The opening of these intermediates by the TMSN3 leads to the isomeric alkenes after the elimination of triphenylphosphine oxide (Scheme 5). The observed selected formation of alkenes with (Z) configuration may result from the higher stability of halonium ion 10 in comparison with the isomeric intermediate 11.    (2)  Since 7d, 7f and 7g differ from 7c, 7e and 7h only in the nature of the halogen, the (Z)-configuration is therefore proposed for all of these compounds. In previous studies, we could confirm that our synthetic methodology allowed the synthesis of a bromoazidoalkene bearing a carboxylate group at C-1′ and a phenyl group at C-2′ with the same selectivity [17]. Thus, the stereochemistry outcome is retained when the carboxylate group is replaced by a tetrazolyl group (7c).
The synthesis of the haloazidoalkenes can be rationalized as outlined in Scheme 5. The formation of the observed products can be explained by considering isomeric halonium ions 10 and 11 as intermediates. These halonium ions can interconvert by way of acyclic cation 9. The opening of these intermediates by the TMSN3 leads to the isomeric alkenes after the elimination of triphenylphosphine oxide (Scheme 5). The observed selected formation of alkenes with (Z) configuration may result from the higher stability of halonium ion 10 in comparison with the isomeric intermediate 11.   Since 7d, 7f and 7g differ from 7c, 7e and 7h only in the nature of the halogen, the (Z)-configuration is therefore proposed for all of these compounds. In previous studies, we could confirm that our synthetic methodology allowed the synthesis of a bromoazidoalkene bearing a carboxylate group at C-1 1 and a phenyl group at C-2 1 with the same selectivity [17]. Thus, the stereochemistry outcome is retained when the carboxylate group is replaced by a tetrazolyl group (7c).
The synthesis of the haloazidoalkenes can be rationalized as outlined in Scheme 5. The formation of the observed products can be explained by considering isomeric halonium ions 10 and 11 as intermediates. These halonium ions can interconvert by way of acyclic cation 9. The opening of these intermediates by the TMSN 3 leads to the isomeric alkenes after the elimination of triphenylphosphine oxide (Scheme 5). The observed selected formation of alkenes with (Z) configuration may result from the higher stability of halonium ion 10 in comparison with the isomeric intermediate 11.
Molecules 2015, 20, page-page Scheme 5. Formation of isomeric halonium ions as intermediates of the reaction.
The formation of halophosphonium salt 9 is the expected intermediate of the halogenation of α-oxophosphorus ylides, which affords the corresponding halophosphonium salts [27]. Moreover, the synthesis of halogenated enol lactones from keto acid phosphoranes via an intramolecular non-classical Wittig reaction has also been described [28][29][30][31]. In fact, α-oxophosphorus ylides bearing a terminal carboxylic acid group react with halogenating agents leading to E-and Z-halo enol lactones. This cyclization was rationalized via a halophosphonium salt followed by loss of triphenylphosphine oxide. Indeed, bromophosphonium salt 13 could be isolated from the reaction of ylide 12 with bromine at 0 °C in the absence of NEt3. Treatment of 13 with triethylamine leads to the corresponding bromo enol lactones 14 (Scheme 6) [28]. Scheme 6. Synthesis of halogenated enol lactones from keto acid phosphoranes.
The 13 C-NMR spectra of the haloazidoalkenes 7a-h show the C-X carbon between 85.8 and 109.3 ppm and the C-N3 between 134.8 and 147.6 ppm ( Table 2). As expected, the chemical shift of C-X carbon of all bromoazidoalkenes is lower than the ones of the corresponding chloroazidoalkenes (e.g., 7b vs. 7a). The thermolysis of the haloazidoalkene derivatives 7 was then investigated (Scheme 7). Initially, attempts were made to promote these reactions in n-heptane. However, due to the low solubility of the haloazidoalkenes in this solvent, the thermolysis in n-heptane often led to complex mixtures of the desired 2H-azirines and degradation products. Nonetheless, carrying out the reaction of these haloazidoalkenes in toluene at 90 °C for 2-3 h led efficiently to the formation of new 2-halo-2-tetrazol-5-yl-2H-azirines 15. The reaction can be followed by TLC and by IR by monitoring The formation of halophosphonium salt 9 is the expected intermediate of the halogenation of α-oxophosphorus ylides, which affords the corresponding halophosphonium salts [27]. Moreover, the synthesis of halogenated enol lactones from keto acid phosphoranes via an intramolecular non-classical Wittig reaction has also been described [28][29][30][31]. In fact, α-oxophosphorus ylides bearing a terminal carboxylic acid group react with halogenating agents leading to Eand Z-halo enol lactones. This cyclization was rationalized via a halophosphonium salt followed by loss of triphenylphosphine oxide. Indeed, bromophosphonium salt 13 could be isolated from the reaction of ylide 12 with bromine at 0˝C in the absence of NEt 3 . Treatment of 13 with triethylamine leads to the corresponding bromo enol lactones 14 (Scheme 6) [28]. The formation of halophosphonium salt 9 is the expected intermediate of the halogenation of α-oxophosphorus ylides, which affords the corresponding halophosphonium salts [27]. Moreover, the synthesis of halogenated enol lactones from keto acid phosphoranes via an intramolecular non-classical Wittig reaction has also been described [28][29][30][31]. In fact, α-oxophosphorus ylides bearing a terminal carboxylic acid group react with halogenating agents leading to E-and Z-halo enol lactones. This cyclization was rationalized via a halophosphonium salt followed by loss of triphenylphosphine oxide. Indeed, bromophosphonium salt 13 could be isolated from the reaction of ylide 12 with bromine at 0 °C in the absence of NEt3. Treatment of 13 with triethylamine leads to the corresponding bromo enol lactones 14 (Scheme 6) [28]. Scheme 6. Synthesis of halogenated enol lactones from keto acid phosphoranes. The 13 C-NMR spectra of the haloazidoalkenes 7a-h show the C-X carbon between 85.8 and 109.3 ppm and the C-N3 between 134.8 and 147.6 ppm ( Table 2). As expected, the chemical shift of C-X carbon of all bromoazidoalkenes is lower than the ones of the corresponding chloroazidoalkenes (e.g., 7b vs. 7a). The thermolysis of the haloazidoalkene derivatives 7 was then investigated (Scheme 7). Initially, attempts were made to promote these reactions in n-heptane. However, due to the low solubility of the haloazidoalkenes in this solvent, the thermolysis in n-heptane often led to complex mixtures of the desired 2H-azirines and degradation products. Nonetheless, carrying out the reaction of these haloazidoalkenes in toluene at 90 °C for 2-3 h led efficiently to the formation of new 2-halo-2-tetrazol-5-yl-2H-azirines 15. The reaction can be followed by TLC and by IR by monitoring Scheme 6. Synthesis of halogenated enol lactones from keto acid phosphoranes.
The 13 C-NMR spectra of the haloazidoalkenes 7a-h show the C-X carbon between 85.8 and 109.3 ppm and the C-N 3 between 134.8 and 147.6 ppm ( Table 2). As expected, the chemical shift of C-X carbon of all bromoazidoalkenes is lower than the ones of the corresponding chloroazidoalkenes (e.g., 7b vs. 7a). The thermolysis of the haloazidoalkene derivatives 7 was then investigated (Scheme 7). Initially, attempts were made to promote these reactions in n-heptane. However, due to the low solubility of the haloazidoalkenes in this solvent, the thermolysis in n-heptane often led to complex mixtures of the desired 2H-azirines and degradation products. Nonetheless, carrying out the reaction of these haloazidoalkenes in toluene at 90˝C for 2-3 h led efficiently to the formation of new 2-halo-2-tetrazol-5-yl-2H-azirines 15. The reaction can be followed by TLC and by IR by monitoring the disappearance of the band corresponding to the azido group of the starting azidoalkenes (ν~2105-2130 cm´1). Regardless of C-3 substituents, 2-bromo-and 2-chloro-2H-azirines 15 were obtained in high yield (85%-99%).  (Table 2).
It is well established that some 2-halo-2H-azirines undergo thermal rearrangement to their azirine isomers through a [1,2]-halogen shift [32,33]. Recently, Banert et al. reported optimized reaction conditions to favor the complete and irreversible isomerization of 2-halo-2H-azirines [11]. In our case, it was possible to isolate 2H-azirines 15 as pure isomers by thermolysis of the haloazidoalkenes 7. However, after being stored at −30 °C for 3 months 2-chloro-2H-azirine 15a, bearing a carboxylate group at C-3, underwent rearrangement to a mixture of 2H-azirines 15a and 16 (Scheme 8). Carrying out NMR measurements at different temperatures (25-95 °C), the variation of the isomer ratio with increasing temperature was observed, until complete rearrangement of 2H-azirine 15a into the isomer 16a (Supplementary Materials). Similar NMR experiments with 15c and 15e did not indicate the same behavior.

General Information
NMR spectra were run in CDCl3 or DMSO-d6 on a 400 MHz Bruker Avance III spectrometer (Bruker Biospin SA, Wissembourg, France) and recorded at the following frequencies: proton ( 1 H, 400 MHz), carbon ( 13 C, 100 MHz). Chemical shifts are expressed in parts per million related to internal TMS and coupling constants (J) are in hertz. Infrared spectra (IR) were recorded on a Nicolet 6700 FTIR spectrometer (Thermo Scientific, Waltham, MA, USA). Mass spectra were recorded in electrospray ionization (ESI) mode on a Bruker FTMS APEX III spectrometer (Bruker Corporation, Bremen, Germany). Melting points were determined in open glass capillaries and are uncorrected. Thin-layer chromatography (TLC) analyses were performed using precoated silica gel plates (Merck KGaA, Darmstadt, Germany). Flash column chromatography was performed with silica gel 60 as the stationary phase.  (Table 2).

Experimental Details
It is well established that some 2-halo-2H-azirines undergo thermal rearrangement to their azirine isomers through a [1,2]-halogen shift [32,33]. Recently, Banert et al. reported optimized reaction conditions to favor the complete and irreversible isomerization of 2-halo-2H-azirines [11]. In our case, it was possible to isolate 2H-azirines 15 as pure isomers by thermolysis of the haloazidoalkenes 7. However, after being stored at´30˝C for 3 months 2-chloro-2H-azirine 15a, bearing a carboxylate group at C-3, underwent rearrangement to a mixture of 2H-azirines 15a and 16 (Scheme 8). Carrying out NMR measurements at different temperatures (25-95˝C), the variation of the isomer ratio with increasing temperature was observed, until complete rearrangement of 2H-azirine 15a into the isomer 16a (Supplementary Materials). Similar NMR experiments with 15c and 15e did not indicate the same behavior. The 13 C-NMR spectra of the 2-chloro-and 2-bromo-2-(tetrazol-5-yl)-2H-azirines 15 show the sp 2 carbon between 156.8 and 169.6 ppm and the sp 3 carbon between 33.3 and 51.9 ppm, depending on the substitution pattern (Table 2).
It is well established that some 2-halo-2H-azirines undergo thermal rearrangement to their azirine isomers through a [1,2]-halogen shift [32,33]. Recently, Banert et al. reported optimized reaction conditions to favor the complete and irreversible isomerization of 2-halo-2H-azirines [11]. In our case, it was possible to isolate 2H-azirines 15 as pure isomers by thermolysis of the haloazidoalkenes 7. However, after being stored at −30 °C for 3 months 2-chloro-2H-azirine 15a, bearing a carboxylate group at C-3, underwent rearrangement to a mixture of 2H-azirines 15a and 16 (Scheme 8). Carrying out NMR measurements at different temperatures (25-95 °C), the variation of the isomer ratio with increasing temperature was observed, until complete rearrangement of 2H-azirine 15a into the isomer 16a (Supplementary Materials). Similar NMR experiments with 15c and 15e did not indicate the same behavior.

General Information
NMR spectra were run in CDCl3 or DMSO-d6 on a 400 MHz Bruker Avance III spectrometer (Bruker Biospin SA, Wissembourg, France) and recorded at the following frequencies: proton ( 1 H, 400 MHz), carbon ( 13 C, 100 MHz). Chemical shifts are expressed in parts per million related to internal TMS and coupling constants (J) are in hertz. Infrared spectra (IR) were recorded on a Nicolet 6700 FTIR spectrometer (Thermo Scientific, Waltham, MA, USA). Mass spectra were recorded in electrospray ionization (ESI) mode on a Bruker FTMS APEX III spectrometer (Bruker Corporation, Bremen, Germany). Melting points were determined in open glass capillaries and are uncorrected. Thin-layer chromatography (TLC) analyses were performed using precoated silica gel plates (Merck KGaA, Darmstadt, Germany). Flash column chromatography was performed with silica gel 60 as the stationary phase.

General Information
NMR spectra were run in CDCl 3 or DMSO-d 6 on a 400 MHz Bruker Avance III spectrometer (Bruker Biospin SA, Wissembourg, France) and recorded at the following frequencies: proton ( 1 H, 400 MHz), carbon ( 13 C, 100 MHz). Chemical shifts are expressed in parts per million related to internal TMS and coupling constants (J) are in hertz. Infrared spectra (IR) were recorded on a Nicolet 6700 FTIR spectrometer (Thermo Scientific, Waltham, MA, USA). Mass spectra were recorded in electrospray ionization (ESI) mode on a Bruker FTMS APEX III spectrometer (Bruker Corporation, Bremen, Germany). Melting points were determined in open glass capillaries and are uncorrected. Thin-layer chromatography (TLC) analyses were performed using precoated silica gel plates (Merck KGaA, Darmstadt, Germany).
Flash column chromatography was performed with silica gel 60 as the stationary phase.

General Procedure for the Synthesis of Ylides 6
A solution of phosphonium salt 4 (10 mmol) and triethylamine (2.53 g, 25 mmol) in dry CHCl 3 (50 mL) was stirred at room temperature while a solution of the appropriate acid chloride (12 mmol) in dry CHCl 3 (10 mL) was added dropwise to it. After the addition, the mixture was stirred at room temperature for 12 h. The reaction mixture was washed with H 2 O (3ˆ50 mL), dried and evaporated to give the desired ylides 6 which were recrystallized from ethyl acetate.   2-(1-Benzyl-1H-tetrazol-5-yl)-1-(5-nitrofuran-2-yl)-2-(triphenylphosphoranylidene)ethanone (6c): Compound 6c was prepared by an analogous method to that described in the literature [34]. A solution of phosphorus ylide 5 (2.1 mmol) and 5-nitrofuran-2-carboxylic acid (2.5 mmol) in dry CHCl 3 (40 mL) was cooled in an ice bath. Then EDCI (3.2 mmol) and DMAP (catalytic) was added to it. After the addition, the mixture was stirred at room temperature for 12 h. The reaction mixture was washed with H 2 O (3ˆ50 mL), dried and evaporated. The crude product was purified by flash chromatography (ethyl acetate). Ylide 6c was obtained as a yellow solid  Ylide 6 (4.5 mmol) was dissolved in dichloromethane (50 mL) and a solution of azidotrimethylsilane (0.71 g, 6.5 mmol) and N-chloroor N-bromosuccinimide (6.5 mmol) in dichloromethane (10 mL) was added. The reaction mixture was stirred at room temperature for the appropriate time (1-3 h). After removal of the solvent, the crude product was purified by flash chromatography (ethyl acetate/hexane).