HETEROCYCLES

Review | Regular issue | Vol. 78, No. 2, 2009, pp. 281-318
Received, 20th August, 2008, Accepted, 22nd September, 2008, Published online, 2nd October, 2008.
DOI: 10.3987/REV-08-641

■ Synthesis of Nitrogen Heterocycles Utilizing Molecular Nitrogen as a Nitrogen Source and Attempt to Use Air Instead of Nitrogen Gas

Miwako Mori*

Health Sciences University of Hokkaido, Ishikari-Tobetsu 061-0293, Japan

Abstract

Nitrogen fixation using transition metals is a challenging subject. Using a titanium-nitrogen complex, discovered by Yamamoto in 1967, and titanium isocyanate complex, molecular nitrogen could be incorporated into organic compounds to afford various heterocyclic compounds. Furthermore, novel titanium-catalyzed nitrogenation procedure was developed. That is, a THF solution of TiCl₄ or Ti(OⁱPr)₄, and an excess amount of Li and TMSCl was stirred under nitrogen (1 atm) at room temperature overnight and to this solution was added phthalic anhydride. The whole solution was refluxed overnight to afford a phthalimide in over 100 % yield. Using the stoichiometric conditions of the novel nitrogenation, various indole, pyrrole, pyrrolizine, indolizine derivatives and lactams were obtained in good to moderate yields after usual workup. Although the structure of the titanium-nitrogen complex has not been determined yet, the complex is thought to be a mixture of N(TMS)₃, TiX₂N(TMS)₂ and XTi=NTMS. Nitrogen in air could be directly fixed using this method, and the natural products, such as monomoline I, lycopodine and pumiliotoxin C, could be synthesized from nitrogen in air as a nitrogen source.

CONTENTS
1. Introduction
2. Synthesis of Nitrogen Heterocycles Using Titanium Isocyanate Complex and Palladium Complex
2-1. Synthesis of Heterocycles Using Palladium-Catalyzed Carbonylation and Titanium-Promoted Nitrogenation
2-2. Synthesis of Heterocycles Using Nitrogenation-Transmetallation reaction
3. Development to Titanium-Catalyzed Nitrogenation
4. Synthesis of Heterocycles Using Novel Titanium-Promoted Nitrogenation
4-1. Synthesis of Heterocycles from Compounds Having Keto-Carbonyl Groups Using Titanium-Nitrogen Complex
4-2. Synthesis of Lactams from Compounds Having Keto-carboxyl Group
4-3. Synthesis of Heterocycles from Compounds Having Keto-Alkyne Group
5. Nitrogen Fixation Using Dry Air
5-1. Examination to Utilize Dry Air as a Nitrogen Source
5-2. Synthesis of Natural Products from Dry Air as a Nitrogen Source
6. Conclusion


1. INTRODUCTION
Since Vol’pin and Shur discovered that molecular nitrogen could be fixed by transition metals1 and reducing agents under mild conditions, various systems of nitrogen fixation have been reported.2 In 1967, Yamamoto reported the synthesis of a cobalt-nitrogen complex3 and then a titanium-nitrogen complex.4 Hidai 5 and Bercow6 later reported the synthesis of a molybdenum-nitrogen complex and a zirconium-nitrogen complex, respectively. However, there have been few reports on incorporation of molecular nitrogen into organic compounds. In 1968, Vol’pin reported the synthesis of aniline from Cp2TiCl2 and phenyl lithium under high pressure of nitrogen (eq. 1).7 Later, van Tamelen succeeded in obtaining isopropyl amine and benzonitrile from diethylketone and benzoyl chloride, respectively using Cp2TiCl2 and Mg under nitrogen (eq. 2).8 In 1977, Chatt synthesized pyrrolidine and isopropyl amine from 1,4-dibromobutane and acetone, respectively, using molybdenum or a tungsten-nitrogen complex (eqs. 3 and 4).9 Hidai recently reported the synthesis of pyrrole from a tungsten nitrogen-complex (eq. 5).10 Yamamoto reported the synthesis of an interesting titanium-nitrogen complex (1) from TiCl4 or TiCl3 using Mg as a reducing agent (eq. 6).4 In this reaction, the nitrogen-nitrogen triple bond was cleaved by a reducing agent to give a titanium-nitrogen complex (1), which reacted with benzoyl chloride to give benzoyl-titanium complex (3).11 The result is very attractive for the synthesis of nitrogen heterocycles from molecular nitrogen because one nitrogen in N2 can be used for the synthesis of heterocycles. Later, Sobota reported that the reaction of 1 with CO2 gave a titanium-isocyanate complex (2) (eq. 7).12

These results are very attractive since organometallic complexes are used for the incorporation of molecular nitrogen into organic compounds. Yamamoto’s report indicated that molecular nitrogen could be fixed using TiCl3 or TiCl4 using Mg as a reducing agent under mild conditions to produce titanium-nitrogen complex (1), and the reaction of 1 with benzoyl chloride gave benzoyl-titanium complex 3. However, in benzoyl-titanium complex (3), whether carbon-nitrogen bond is formed or not is unclear. Thus, complex (3) was prepared from 1 and benzoyl chloride according to the Yamamoto’s procedure,4 and was hydrolyzed. As a result, benzamide was obtained in 17% yield. It means that nitrogen could be incorporated into benzoyl chloride and carbon-nitrogen bond is formed in benzamide. When a THF solution of 1 and benzoyl chloride was refluxed overnight, the yield of benzamide was increased to 36% along with benzimide in 29% yield.12 Sobota reported that titanium-nitrogen complex (1) reacted with CO2 to give titanium-isocyanete complex (2) (eq. 7),13 which was thought to have the same reactivity with that of benzoyl-titanium complex (3). Thus, titanium-isocyanate complex (2) was synthesized from 1 and CO2, and a THF solution of 2 and benzoyl chloride was stirred in pyridine at room temperature overnight. As the result, benzamide was obtained in 27% yield along with benzimide in 12% yield (Scheme 1).12

Since isocyanate complex (2) is more stable compared with that of titanium-nitrogen complex (1), synthesis of nitrogen containing compounds was examined using this complex 2. When phthaloyl chloride was reacted with 2 in pyridine, phthalimide (4) was obtained in 14% yield. Use of phthalic anhydride instead of phthaloyl chloride improved the yield of 4 to 55%. As a solvent, NMP (N-methyl pyrrolidone) gave a good result and desired phthalimide (4) was obtained in 78% yield. It was interesting that benzoxazinone (5), which is thought to be a kind of a mixed anhydride, was treated in a similar manner to give quinazolinone (6) in 55% yield.12

On the basis of these results, synthesis of heterocycles using titanium isocyanate complex (2) was investigated and novel nitrogenation method was developed. Furthermore, we found that dry air can be used as a nitrogen source for the synthesis of heterocycles.14

1. SYNTHESIS OF NITROGEN HETEROCYCLES USING TITANIUM ISOCYANATE COMPLEX AND PALLADIUM COMPLEX
2-1. Synthesis of Heterocycles Using Palladium-Catalyzed Carbonylation and Titanium-Promoted Nitrogenation
It was well known that an oxidative addition of an aryl halide to Pd(0) afforded aryl palladium complex, which was converted into acyl palladium complex (7) under carbon monoxide.15 If acyl palladium complex (7) is in a state of equilibrium with acid chloride and Pd(0), titanium-isocyanate complex (2) may react with acyl palladium complex (7) to give amide or imide (Scheme 3). The problem on this idea is whether titanium-isocyanate complex (2) and Pd(PPh3)4, which are added in the same reaction vessel, can act for each role. When a NMP solution of bromobenzene, Pd(PPh3)4, titanium-isocyanate complex (2), and K2CO3 was stirred upon heating at 100 °C for 24 h under carbon monoxide (1 atm), benzimide was obtained in 22% yield. It means that carbon monoxide and nitrogen could be incorporated into aryl halide to give benzimide in one-pot reaction, although the yield was moderate. When 2-bromobenzoic acid was treated in a similar manner, phthalimide (4) was obtained in 82% yield.16

On the basis of these results, the synthesis of lactam (9) from aryl halide (8) having a keto-carbonyl group at the 2-position was planned (Scheme 4). If aryl halide (8) is reacted with Pd(0) under carbon monoxide, acylpalladium complex (11) would be formed via arylpalladium complex (10). Since the keto-carbonyl group is in a state of equilibrium with an enol form, the enol part would react with acylpalladium complex to produce enol lactone (12). If the enol lactone (12) can react with titanium-isocyanate complex (2), isoindolinone (9) would be produced.

The idea was realized when o-bromobutylophenone (8a) was reacted with Pd(PPh3)4, titanium-isocyanate complex (2) and K2CO3 in NMP under carbon monoxide at 120 °C for 24 h. As the result, isoindolinone (9a) was obtained in 70% yield. To clarify this reaction course, the reaction was carried out in stepwise. When compound (8a) was reacted with Pd(PPh3)4 in NMP at 100 °C for 12 h under carbon monoxide, enol lactone (12a) was obtained in 87% yield.17 Then, compound (12a) was reacted with 2 in NMP at 120 °C for 24 h to give isoindolinone (9a) in 80% yield.

Various isoindolinones could be synthesized using this procedure and the results are shown in Table 1. In some cases, hydroxylated isoindolines (13) were produced along with isoindolinones (9). Presumably, compound (13) would be formed from 9 during the workup. The total yields of isoindolinones (9) and (13) were high in each case.
Using this procedure, a natural product, glycosminine, could be synthesized from 14 in a one-pot reaction. A NMP solution of N-acetyl-2-bromoaniline derivative (14), Pd(PPh3)4, 2 and K2CO3 was heated under carbon monoxide at 100 °C overnight to afford glycosminine in 40% yield. The reaction course would be same as that of isoindolinone, that is, acylpalladium complex (16) would react with an enol part of 16 to afford benzoxazinone (17). Then it reacts with 2 to afford glycosminine.18

2-2. Synthesis of Heterocycles Using Nitrogenation-Transmetallation Reaction
van Tamelen succeeded in obtaining isopropylamine and diisopropylamine from diethylketone using Cp2TiCl2 and Mg under nitrogen (eq. 2).8

If an imino part on titanium-imine complex (20) generated from keto-carbonyl compound and 2 is transmetallated to arylpalladium complex (21), which was derived from aryl halide (18) and Pd(0), and then reductive elimination occurs from resultant palladium complex (22), aniline derivative (19) would be formed, that is, novel C-N-C bond formation would be developed. When a NMP solution of diketone (23), 4-bromobenzene (18a), titanium-isocyanate complex (2), Pd(PPh3)4, and K2CO3 was heated at 100 °C for 12 h, aniline derivative (19a) was obtained in 39% yield along with enaminone (24) in 30% yield (Scheme 8). The latter compound (24) would be obtained from 20a by hydrolysis. Similarly, 4-bromotoluene (18b) afforded aniline derivative (19b) in 37% yield. These results indicated that novel C-N-C bond is formed from ketone (23), aryl halide (18) and titanium-isocyanate complex 2 using Pd(0), and aniline derivative (19) could be obtained.

To demonstrate the applicability of this nitrogenation transmetallation process, synthesis of heterocycles by intramolecular reaction was planned as shown in Scheme 9.

When a NMP solution of 1,3-diketone (25a) having vinyl halide in a tether and 2 was stirred upon heating at 100 °C in the presence of Pd(pph3)4 for 12 h, indole derivative (26a) was obtained in 87% yield (Table 2, entry 1). The intermediate should be imino-titanium complex (27a), and transmetallation of the imino group on titanium to palladium metal would occur to afford 28a and then reductive elimination affords desired compound (26a).

Thus, various indole derivatives (26) could be synthesized from 1,3-diketone (25) in high yields, respectively, and the novel C-N-C bond formation could be developed (Table 2).19

3. DEVELOPMENT TO TITANIUM-CATALYZED NITROGENATION
Various heterocycles could be synthesized from molecular nitrogen as a nitrogen source using titanium isocyanate complex (2). Although these reactions are unique and interesting, an equimolar amount of titanium-isocyanate complex (2) was required. Thus, the next subject is whether this reaction proceed by a catalytic amount of titanium complex. To realize a catalytic reaction, the working hypothesis was shown in Scheme 10. Titanium-nitrogen complex [THF·Mg2Cl2·TiN] (1) can be represented as 1’. If TMSCl reacts with 1’, titanium-trimethylsilylamine complex (29) would be formed. σ-Bond metathesis of 29 with TMSCl and then oxidative addition of the TMSCl to the resultant complex would afford 31 and tris(trimethylsilyl)amine (30) [N(TMS)3]. Complex (31) would be reduced to complex (1’) in the presence of excess amount of Mg under nitrogen because complex (31) should be the same oxidation state as that of TiCl3.

To examine whether this reaction proceeds with a catalytic amount of TiCl4, the reaction was carried out using an equimolar amount of TiCl4 under nitrogen in the presence of an excess amount of reducing agent, TMSCl, and benzoyl chloride. Since benzoyl chloride would react with titanium-nitrogen complex fixed under these reaction conditions to afford benzamide and benzimide, the amount of incorporated nitrogen would be estimated by the total yields of benzamide and benzimide. Initially, a mixture of TiCl4 (1 equiv.), 4 equiv. of Mg, 5 equiv. of TMSCl in THF was stirred at room temperature overnight, and then PhCOCl (10 equiv.) was added. The whole solution was refluxed overnight. After the usual workup, benzamide and benzimide were obtained in 29% yield (Table 3, entry 2). In the absence of TMSCl, the yield is almost same (entry 1). When an excess amount of Mg and TMSCl were used for this reaction, the total yields of amide and imide raised to 56% (entry 4). When Li was used instead of Mg as a reducing agent, the yield is almost same, but the TLC of the reaction mixture was clear compared with that of the use of Mg (entry 5). However, the total yields were almost same as those using Mg and it was not over 100%. The lower and higher reaction temperature did not give the good results (entries 6 and 7). Since it was thought that all of the titanium-nitrogen complexes and N(TMS)3 generated in this reaction could not react with benzoyl chloride under these reaction conditions, the reaction procedure was changed, that is, the reaction mixture was hydrolyzed with 10% HCl to convert titanium-nitrogen complex into NH4Cl and to the aqueous solution was added an excess amount of benzoyl chloride and K2CO3, and the whole solution was stirred overnight. After the usual workup, the yield of benzamide obtained in this reaction was determined based on TiX4 (entries 8-16). As a result, when Mg was used as a reducing agent, benzamide was obtained in 75% yield (entry 8). On the other hand, when Li was used instead of Mg, benzamide was obtained in 250% yield based on TiCl4 (entry 9). The result indicated that the catalytic cycle in regard to TiCl4 was established.20

Shiina reported the reductive silylation of molecular nitrogen.21,22 He obtained tris(trimethylsilyl)amine [N(TMS)3] from various transition metals, TMSCl and Li under nitrogen. When CrCl3 was used as the metal, 5.4 equiv. of N(TMS)3 was produced, but TiCl4 afforded only 0.8 equiv. of N(TMS)3. Shiina’s and our result mean that in the case of TiCl4, other titanium-nitrogen complex, such as ClTi=NSiMe3 or Cl2TiN(TMS)2, would be produced under these reaction conditions.

Since the reaction mechanism was not clear, various amounts of Li and TMSCl were examined to determine the amount of them required for this reaction (entries 10-15). In the absence of TMSCl, benzamide was not obtained (entry 10) and the use of 4 equiv. of Li and 5 equiv. of TMSCl did not afford the product (entry 11). When 10 equiv. of Li and 10 equiv. of TMSCl were used for this reaction, PhCONH2 was obtained in 79% yield (entry 12), and increasing the amount of Li raised the yield of the desired product (entries 12-15).
In order to promote the reaction of benzoyl chloride and the titanium-nitrogen complex formed during the nitrogen fixation, various additives were examined under the stoichiometric reaction conditions modified the conditions of entry 12 [TiCl4 (1.0 equiv.), Li (10 equiv.), TMSCl (10 equiv.) and benzoyl chloride (10 equiv.)]. As the result, addition of CsF improved the total yields of benzamide and benzimide (Table 4).
Various titanium complexes and the reducing agents were examined under the same reaction conditions (Table 5). It was interesting that Ti(OiPr)4 could be used for this reaction (entry 4).23

Whether nitrogen of titanium-nitrogen complex (32) can be incorporated into phthalic anhydride was examined.20 To a solution of nitrogenation fixation complex, prepared from TiCl4 (1 equiv.), TMSCl (10 equiv.), and Li (10 equiv.) in THF, was added phthalic anhydride (3 equiv.), and the solution was stirred at room temperature overnight. After hydrolysis, phthalimide (4) was obtained in 23% yield (Table 6, entry 1). When the reaction was carried out using 50 equiv. of Li and TMSCl at room temperature, the yield of 4 improved to 51% (entry 2). Interestingly, the yield of phthalimide (4) rose up to 204% when the solution was refluxed overnight in the presence of CsF. These results suggest that phthalimide (4) could be synthesized from phthalic anhydride using novel titanium-catalyzed nitrogenation.20

Although the reaction species are not clear, the possible reaction course is considered as shown in Scheme 11.

Titanium complex TiX4 would be reduced with Li to afford TiX2 33, which would react with N2 to give 34. The nitrogen-nitrogen double bond of 34 is cleaved with Li and TMSCl to give titanium-imide complex (35), which is reacted with TMSCl to give titanium-amide complex (36). In the presence of Li, 36 reacts with TMSCl to give N(TMS)3 (30), and TiX2 (33) would be regenerated.

4. SYNTHESIS OF HETEROCYCLES USING NOVEL TITANIUM-PROMOTED NITROGENATION
4-1. Synthesis of Heterocycles from Compounds Having Keto-Carbonyl Group Using Titanium-Nitrogen Complex
4-1-1. Synthesis of Indole, Quinoline and Pyrrole Derivativves from Compounds Having Diketone.
Although various heterocyclic compounds could be synthesized from molecular nitrogen as a nitrogen source using titanium-isocyanate complex (2), synthesis of titanium-isocyanate complex (2) was not easy, that is, titanium-nitrogen complex 1 was synthesized from TiCl4 and Mg under nitrogen and was isolated. Then isolated 1 was reacted with CO2 to produce 2, which was isolated. However, complex (1) is not stable and it was occasionally difficult to obtain the reproducibility of the preparation of 2. Since a solution of titanium-nitrogen complex (32)24 could be easily prepared from TiCl4 or Ti(OiPr)4, TMSCl, and Li in THF under nitrogen (1 atm) at room temperature, whether nitrogen fixed by new method can be used to synthesize various heterocycles in situ was examine.

If cyclohexenone (37) bearing a keto-carbonyl group in a tether at the 2-position is treated with titanium nitrogen complex (32a), nitrogen in 32 would attack the 3-position of cyclohexenone to produce imine (38), whose imino group would further react with the carbonyl group in a tether intramolecularly to give cyclized compound (39) (Scheme 12).

When to a THF solution of 32a, prepared from TiCl4, TMSCl, Li, and CsF under nitrogen using the stoichiometric reaction conditions, was added 37aa and the whole solution was refluxed overnight, indole derivative (39a) was obtained in 51% yield (Table 7, entry 1). Various indole derivatives (39) could be synthesized from diketone derivative (37) and 32a prepared in situ in high yields by one operation. The hydroxyl group was suitable as a leaving group (entries 2 and 4). Elongation of the methylene group in a tether gave quinoline derivative (40) in moderate yield (entry 6). In this reaction, the spot on the TLC of the reaction mixture was different from that of the purified product (40). Presumably, tetrahydroquinoline derivative (39d) would be formed, and it was converted into dihydroquinoline derivative (40) by an air oxidation during the workup. Treatment of triketone (37db) in a similar manner afforded 40 in the same yield (entry 7).

The reaction procedure for the synthesis of indole 39 and quinoline derivatives 40 is very simple and the reproducibility was easily obtained, that is, to a THF solution containing Li (12 equiv.), TMSCl (12 equiv,) and CsF (6 equiv.) was added TiCl4 (1.25 equiv.) under nitrogen (1 atm) and the solution was stirred under nitrogen overnight. The colorless solution was changed to black. To this solution was added substrate (37) (1 equiv.) and the whole solution was refluxed overnight under argon. After the usual workup, desired compound (39) or (40) was obtained.25

Since it is thought that titanium-nitrogen complex (32a) can react with the carbonyl group to form imine, the imino group of 42, generated from 1,4-diketone (41) and 32a, would further react with the keto-carbonyl group intramolecularly to form pyrrole derivative (43) (Scheme 13).

Various 1,4-diketones (41) were treated with 32a prepared from TiCl4, Li, TMSCl and CsF under nitrogen in THF to give pyrrole derivatives (43) in good yields (Table 8).25 These results indicated that titanium-nitrogen complex 32a could react with two keto-carbonyl groups intramolecularly to give pyrrole derivatives.

4-1-2. Synthesis of Pyrrolizine and Indolizine Derivatives from Compounds Having Triketone.
Next try is the reaction of triketone (44) with titanium-nitrogen complex (32a). If this reaction proceeds, pyrrolizine or indolizine derivative (46) would be synthesized from compound (44) having tri-ketocarbonyl groups and 32a in one-pot reaction (Scheme 14).

Reaction of triketone (44) with 32a prepared in a usual manner gave pyrrolizine derivative (46a) in 30% yield (Table 9, entry 1). Although the yield is moderate, it is interesting that pyrrolizine derivative (46a) could be synthesized from straight chain compound (44a) having triketone and 32a by a one-pot reaction. Triketone (44b) was treated with 32a in a similar manner to give tricyclic compound (46b) as a mixture of two inseparable isomers in 31% yield (entry 2). Triketone (44c), whose carbon chain length was elongated, was reacted with 32a to give indolizine derivative (46c) in 29% yield. When 2 equiv. of 32a was used for this reaction, the yield of 46c was increased to 41% (run 3). In a similar treatment of 44d, 44e, and 44f with 32a (2 equiv.), indolizine derivatives 46d, 46e and 46f were obtained in 56%, 30% and 30% yields, respectively (entries 4-6).25 Various pyrrolizine and indolizine derivatives could be easily synthesized from straight chain compounds 44 having triketone and titanium-nitrogen complex 32 by a one-pot reaction.

4-2. Synthesis of Lactams from Compounds Having Keto-Carboxyl Group
It is thought that if compound 47 having the keto-carbonyl group and the carboxyl group reacts with 32 to produce imine 48 and an imine part of 48 would react with the carboxyl group intramolecularly, lactam (49) would be obtained (Scheme 15).

When a THF solution of acid chloride (47aa) and 32a, which was prepared using a TiCl4-Li-TMSCl system under nitrogen, was refluxed in the presence of CsF overnight, lactam (49a) was obtained in 28% yield (Table 10, entry 1). Use of titanium-nitrogen complex (32b) prepared using a Ti(OiPr)4-Li-TMSCl system slightly increased the yield of 49a (entry 2). In this reaction, various carboxylic acid derivatives, such as chloride (47aa), mixed anhydrides (47ab) (entries 3-5), ester (47ae) (entry 7), and even the carboxylic acid (47ad) (entry 6), could be used and lactam (49a) was produced in good yields (entries 1-7).

Various bicyclic lactams (49b-d) were obtained in good yields from corresponding keto-carboxylic acid derivative (47b-d) using this method (Table 11, entries 1-5). In a similar manner, piperidone derivatives (49e-g) were also obtained in moderate yields from corresponding straight chain compounds (47e-g) (entries 6-8).26 It is interesting that the carboxylic acids 47cd and 47dd afforded lactams 49c and 49d in moderate yields.

4-3. Synthesis of Heterocycles from Compounds Having Keto-Alkyne Group
If compound 50 having the keto-carbonyl group and an alkyne part can react with 32, and the imine part of 51 generated from the keto-carbonyl group can react with an α,β-unsaturated ester moiety intramolecularly by Michael-type addition, cyclic compound (52) would be formed after hydrolysis (Scheme 16).

A solution of keto-alkyne (50a) (1 equiv.) and titanium-nitrogen complex (32a), which was prepared from TiCl4 (1.2 equiv.), Li (10 equiv.), and TMSCl (10 equiv.) in THF under nitrogen, was refluxed in the presence of CsF (6 equiv.) for 17 h to give indole derivative (52a) in 90% yield (Table 12, entry 1). When the reaction was carried out at room temperature for 24 h, desired indole derivative (52a) was obtained in 59% yield (entry 2). On the other hand, when titanium-nitrogen complex (32b) prepared from Ti(OiPr)4, Li and TMSCl was used for this reaction, the reaction proceeded at room temperature for only 50 min and 52a was obtained in 82% yield (entry 3). In the absence of CsF, the yield was slightly decreased (entry 4). The reaction was carried out using N(TMS)3 instead of 32, no product was formed. It means that the real species for formation of indole derivative (52a) is titanium-nitrogen complex, not N(TMS)3.

Indole derivative (50b) having an acetoxy group on a cyclohexane ring was also synthesized from keto-alkyne (52b) (Table 13, entry 1). The effect of a substituent on alkyne was examined. In the case of keto-alkyne (50c) having a nitrile group on the alkyne, the yield of 52c was also high (entry 2), and 50d or 50e bearing the amide or keto-carbonyl group on the alkyne gave desired indole derivative (52d) or (52e) in moderate yield (entries 3 and 4). The keto-alkyne (50f) having the methyl group on the alkyne gave desired indole derivative (52f) in only 3% yield, but the phenyl group on the alkyne gave desired compound (52g) in 35% yield (entry 6). The yield of 52h having trifluoromethyl group on an aromatic ring is higher than that of 52i having the methyl group (entries 7 and 8). These results indicate that the electron-withdrawing group on the alkyne gave a good result. Presumably, theses groups would accelerate the addition of an imine part of 51 to an alkyne part.

Various keto-alkynes (53) having an ester group on the alkyne were reacted with titanium-nitrogen complex (32b), prepared using Ti(OiPr)4-Li-TMS system under nitrogen, in THF at room temperature (Table 14). As the results, the desired heterocyclic compounds such as pyrrole derivatives (54a) and (54b) (entries 1 and 2), quinoline derivative (54c) (entry 3), and piperidine derivatives (54d) and (54e) (entries 4 and 5) were obtained from corresponding keto-alkynes (53) in good to moderate yields.27

The possible reaction course for the formation of heterocyclic compound (52) or (54) from keto-alkynes (50) or (53) is shown in Scheme 17. Two possible pathways should be considered. If the reaction of keto-alkyne (50) or (53) with N(TMS)3 proceeds in the presence of CsF, imine (51) would be formed. Then Michael addition of nitrogen of imine would afford 55, which would isomerize to give 52 or 54 (path A). However, when 50 was reacted with N(TMS)3 in the presence of CsF at room temperature for 16 h, no cyclized product was formed, indicating that N(TMS)3 in titanium-nitrogen complex (32b) is not an active species. Thus, the active species in this reaction would be titanium-imide complex [XTi=N(TMS)] (35) or titanium-amide complex [X2Ti-N(TMS)2] (36), and the complex reacts with the carbonyl carbon to give imine (51) via 56 by σ-bond metathesis.28,29 Then the Michael addition of nitrogen gives 55. The mechanism of this reaction is similar to that of the synthesis of a pyrrole derivative by Arcadi, who obtained a pyrrole derivative from keto-alkyne and primary amine.30

Subsequently, the reaction of 1,4-diketo-alkyne (57) with 32b was examined and indolizidines (58) and (59) were obtained in 20% and 10% yields, respectively, along with pyrrole derivative (60) in 11% yield. The structures of them were determined by NOE experiments. The ester group of indolizine (58) would be converted into the nitrile group by titanium nitrogen complex (32) to give indolizine (59).31

5. NITROGEN FIXATION USING DRY AIR
5-1. Examination to Utilize Dry Air as a Nitrogen Source
It is well known that nitrogenase catalyzes the reduction of atmospheric nitrogen to NH3. Nitrogen accounts for 80% of the gases in air, the other gases being oxygen and carbon dioxide. If nitrogen in air can be fixed directly using our nitrogenation method, the results is interesting. Since early transition metals are not so sensitive to oxygen but are very sensitive to water, whether dry air could be used for nitrogenation was examined.

Initially, it was examined whether titanium-nitrogen complex (32) is generated from dry air, TiX4, Li and TMSCl. To confirm the fixation of nitrogen in air, preparation of benzamide from benzoyl chloride and 32 generated from dry air was tried. A THF solution of TiCl4 (1.0 equiv) and TMSCl (10.0 equiv) was stirred in the presence of Li (10 mol equiv, porous) under dry air passed through a calcium chloride tube and a molecular sieve tube at room temperature for 24 h. The solution turned black with a green tinge. After hydrolysis of the reaction mixture, benzoyl chloride and K2CO3 were added to this aqueous solution. It was very pleased to find that benzamide could be isolated in 89% yield, although the use of nitrogen gas in a similar manner gave benzamide in 96% yield (Table 15, entry 1). The result obtained using a Ti(OiPr)4-Li-TMSCl system under dry air was the same as that obtained using a TiCl4-Li-TMSCl system (entry 2). It had already been shown that this nitrogen fixation reaction proceeded catalytically based on TiCl4.20

Thus, an experiment was carried out to determine whether the nitrogen fixation reaction using dry air proceeds catalytically. Titanium-nitrogen complex (32b) was synthesized from Ti(OiPr)4 (1 equiv), excess amounts of Li (50 mol equiv) and TMSCl (50 equiv) under dry air for 2 days. After the solution had been hydrolyzed and then benzoyl chloride (50 equiv.) and K2CO3 were added. The solution was stirred at room temperature overnight to give benzamide in 356% yield based on an amount of Ti(OiPr)4 (entry 3). The results indicated that nitrogen in air could be directly fixed from dry air.32

Subsequently, whether the heterocycle could be synthesized from dry air as a nitrogen source was examined. To a THF solution of 1,4-diketone (41e) and CsF was added a THF solution of 32a prepared from TiCl4, Li, and TMSCl under dry air, and the whole solution was refluxed overnight. After the usual workup, pyrrole derivative (43e) was obtained in 37% yield based on 1,4-diketone (41e) (Table 16, entry 1). When Ti(OiPr)4 was used instead of TiCl4, 43e was obtained in 38% yield (entry 2). Although these yields were lower compared with those used N2 gas, dry air could be used directly as a nitrogen source.

Various heterocycles were synthesized using titanium-nitrogen complex (32) prepared from dry air, TiX4, Li, and TMSCl, and in each case, desired heterocycles could be synthesized in good yields (Table 17). The yield obtained from dry air was slightly lower than that obtained using nitrogen gas.26 These results indicate that dry air could be used for the synthesis of various nitrogen heterocycles as the nitrogen source instead of nitrogen gas.

5-2. Synthesis of Natural Products from Dry Air as a Nitrogen Source
5-2-1. Synthesis of Monomoline I
Since various heterocycles could be prepared from dry air as a nitrogen source, synthesis of the natural products using this complex (32) was examined. At first, the synthesis of the natural product, monomorine I33 from indolizine derivative (46) was planned. The retro synthetic analysis of monomoline I using this method is shown in Scheme 20. Monomorine I would be obtained by hydrogenation of indolizine (46g) because hydrogen on the catalyst would approach from the backside of the substituents. Indolizine (46g) would be obtained from triketone (44g) using the present nitrogenation. Ozonolysis of 6134 followed by treatment with Me2S gave triketone (44g) in 85% yield. A THF solution of triketone (44g) and titanium nitrogen complex (32a) (2 equiv), prepared from TiCl4, Li and TMSCl under dry air, was refluxed for 24 h. After the usual workup, desired indolizine derivative (46g) was obtained in 22% yield. In this reaction, same compound (46g) was obtained in 30% yield from nitrogen gas as the nitrogen source. Hydrogenation of 46g with Rh on alumina (20 atm) afforded monomorine I as a main product in 32% yield along with indolizidine 195B in 4% yield.35 Thus, a short-step synthesis of monomorine I and indolizidine 195B was achieved using titanium-nitrogen complex (32a) prepared from dry air as a nitrogen source.36

5-2-2. Formal Total Synthesis of Lycopodine
Since quinolone derivative (49a) could be synthesized from 47ac and dry air as a nitrogen source, the synthesis of lycopodine37 was planed. The retrosynthetic analysis was shown in Scheme 21. The total synthesis of (±)-lycopodine was achieved by Stork and co-workers38 from tetracyclic compound (62), which was obtained from lactam (63). The key compound (63) should be synthesized from carboxylic acid (64) using dry air as the nitrogen source. Carboxylic acid (64) would be synthesized from cyclohexenone (65).

Ketalization of 66, which was easily prepared from 5-methyl-2-cyclohexenone (65),39 followed by hydroboration gave alcohol (67) in high yield. Conversion of the hydroxyl group of 67 into the carboxyl group with the usual method afforded 64, which was treated with diethyl phosphorochloridate to give 68. When a THF solution of (68) and titanium-nitrogen complex (32a), prepared from Ti(OiPr)4, Li, and TMSCl under dry air, was refluxed for 36 h, lactam (63) was isolated in 40% yield along with the stereoisomers (63’) in 7% yield. When the reaction was carried out using nitrogen gas, 63 was obtained in 39% yield. It means that dry air could be used as the nitrogen source instead of nitrogen gas. The melting point of 63 was in complete agreement with that reported in the literature.38 Compound (63) was treated with H3PO4 in HCOOH according to the Stork’s procedure38 to afford tetracyclic compound (62) . Thus, a formal total synthesis of (±)-lycopodine could be achieved using titanium-nitrogen complex (32) prepared from dry air.26

5-2-3. Synthesis of Pumiliotoxine C
Perhydroquinoline derivatives 54c could be synthesized from the compound (53c) having the keto-alkyne part and 32 prepared from air as the nitrogen source. Thus, the total synthesis of (±)-pumiliotoxine C was planned. Pumiliotoxin C was isolated40 from skin extracts of the Panamanian poisonous frog Dendrobates pumilio41 as the first member of one major class of dendrobatid alkaloids and has a cis-decahydroquinoline skeleton. The retrosynthetic analysis of pumiliotoxin C using air as the nitrogen gas is shown in Scheme 23. Pumiliotoxin C would be synthesized from quinoline derivatives (54f), which should be able to be synthesized from keto-alkyne (53f) and air using the present method. From commercially available 3-methylcyclohexenone (69), keto-alkyne (53f) would be synthesized.

Allylation of 3-methylcyclohexenone (69)42 followed by ketalization afforded compound (70). Hydroboration followed by oxidation with PCC and then treatment with CBr4 and PPh3 gave 71.

Treatment of 71 with BuLi gave lithium acetylide, which was reacted with methyl chloroformate followed by deketalization to give keto-alkyne (53f). Synthesis of quinoline (54f) was successfully achieved by the reaction of keto-alkyne (53f) with titanium-nitrogen complex (32b) prepared from air in high yield. Hydrogenation of compound (54f) was carried out using 5 mol % of palladium on charcoal (STD-type) and then protection of the secondary amine with carbobenzyloxy chloride afforded 72a and 72b in 57% yields in a ratio of 7 to 1. The major isomer (72a) was treated with DIBAL-H followed by the reaction with a Wittig reagent to give compound (73). Hydrogenation of (73) with 5% palladium on charcoal followed by treatment with Et2O·HCl afforded pumiliotoxine C hydrochloride, whose spectral data agreed with those reported in the literature. Thus the total synthesis of pumiliotoxin C was achieved using air as a nitrogen source.43

6. Conclusion
Nitrogen fixation and utilization of the fixed nitrogen are a challenging theme in synthetic organic chemistry, and transition metal complexes should play an important role in nitrogen fixation. Titanium-nitrogen complex (1) reported by Yamamoto is very interesting because one atom pressure of nitrogen could be fixed by TiCl4 or TiCl3 and Mg at room temperature, and nitrogen-nitrogen triple bond was cleaved under these reaction conditions. Sobota reported that that titanium-nitrogen complex (1) easily reacted with CO2 to produce titanium-isocyanate complex 2. Since complex (2) is more stable and handling of 2 is easier compared with that of 1, various heterocyclic compounds could be synthesized using titanium-isocyanate complex (2) and palladium catalyst. To develop to the catalytic reaction in regard to the titanium complex, an excess amount of Li and TMSCl were added to the THF solution of TiCl4 or Ti(OiPr)4 solution under nitrogen. As the result, the novel titanium-catalyzed nitrogenation method could be developed. Since this procedure was very simple and it was not required to isolate titanium-nitrogen complex (32), the stoichiometric reaction conditions were confirmed by modification of the catalytic reaction conditions, that is, a THF solution of TiCl4 or Ti(OiPr)4 (1 equiv.) was stirred in the presence of Li (10 equiv.) and TMSCl (10 equiv.) under nitrogen at room temperature overnight to afford a solution of titanium nitrogen complex (32). To this THF solution of 32 were added a THF solution of the substrate and CsF, and the solution was stirred an appropriate temperature under argon. After the usual work up, various heterocyclic compounds could be synthesized. Furthermore, nitrogen in air instead of nitrogen gas could be fixed directly using this method. As the results, various heterocyclic compounds could be synthesized from air as the nitrogen source, and the several natural products, such as monomoline I, lycopodine and pumiliotoxine C, could be synthesized from dry air as the nitrogen source. These results indicated that titanium-nitrogen complex (32) prepared from air could be used as a nitrogen source in synthetic organic chemistry.

Furthermore, we have already reported that non-substituted aniline derivatives or benzolactams could be synthesized by use of palladium-catalyzed coupling reaction and palladium catalyzed-carbonylation in the presence of an aryl halide and 32.44 Presumably, nitrogen-containing compounds such as amino acid and nucleic acid will be synthesized from air as a nitrogen source. It is also known that many transition metals can react with nitrogen to form transition metal-nitrogen complexes, which should be used as a nitrogenation agent in the future. As the results, nitrogen gas or air will be directly used as a nitrogen source in synthetic organic chemistry in the future.

ACKNOWLEDGEMENT
These works were performed by Dr. Y. Uozumi, Dr. M. Akashi, Dr. K. Ueda, Ms. M. Kawaguchi , Mr. M. Hori, and Mr. K. Hori. I express my heartfelt thanks to them. I also thank to Prof. M. Shibasaki (The University of Tokyo), Prof. Y. Sato, Dr. M. Nishida, and Dr. M. Takimoto for their helpful discussion.

References

1. M. E. Vol’pin and V. B. Shur, Dokl. Akad. Nauk SSSR., 1964, 156, 1102. [Chem. Abstr., 1964, 61, 8933a]; M. E. Vol’pin and V. B. Shur, Nature, 1966, 209, 1236. CrossRef
2. M. Hidai and Y. Mizobe, Chem. Rev., 1995, 95, 1115. CrossRef
3. A. Yamamoto, S. Kitazume, and S. Ikeda, J. Chem. Soc., Chem. Commun., 1976, 79.
4. A. Yamamoto, M. Ookawa, and S. Ikeda, J. Chem. Soc., Chem. Commun., 1969, 841; CrossRef A. Yamamoto, S. Go, M. Ookawa, M. Takahashi, S. Ikeda, and T. Keii, Bull. Chem. Soc. Jpn., 1972, 45, 3110. CrossRef
5. M. Hidai, U. Tominari, Y. Uchida, and A. Misono, J. Chem. Soc., Chem. Commun., 1969, 1392; CrossRef M. Hidai, K. Tominari, and Y. Uchida, J. Am. Chem. Soc., 1972, 94, 110. CrossRef
6. J. M. Manriquez and J. E.Bercaw, J. Am. Chem. Soc., 1974, 96, 6229; CrossRef J. M. Manriquez, R. D. Sanner, R. E. Marsh, and J. E. Bercaw, J. Am. Chem. Soc., 1976, 98, 8351. CrossRef
7. M. E. Vol’pin, V. B. Shur, R. V. Kudryavtsev, and L. A. Prodayko, J. Chem. Soc., Chem. Commun., 1968, 1038. CrossRef
8. E. E. van Tamelen and H. Rudler, J. Am. Chem. Soc., 1970, 92, 5253. CrossRef
9. P. C. Bevan, J. Chatt, G. J. Leigh, and E. G. Leelamani, J. Organomet. Chem., 1977, 139, C 59; CrossRef H. Seino, Y. Ishii, T. Sasagawa, and M. Hidai, J. Am. Chem. Soc., 1995, 117, 12181; CrossRef M. Hidai and Y. Ishii, Bull. Chem. Soc. Jpn., 1996, 69, 819. References are therein. CrossRef
10. H. Seino, Y. Ishii, and M. Hidai. J. Am. Chem. Soc., 1994, 116, 7433; CrossRef H. Seino, Y. Ishii, T. Sasagawa, and M. Hidai, J. Am. Chem. Soc., 1995, 117, 12181. CrossRef
11. Recently, hydrogenation and cleavage of dinitrogen to ammoinia with zirconium complex was reported. In this reaction, nitrogen-nitrogen triple bond was cleaved to ammonia. J. A. Pool, E. Lovkovsky and P. J. Chirik, Nature, 2004, 427, 527. CrossRef
12. M. Mori, Y. Uozumi, and M. Shibasaki, Tetrahedron Lett., 1987, 28, 6187. CrossRef
13. P. Sobota, J. Trzebiantowska, and Z. Janas, J. Organomet. Chem., 1976, 118, 253. CrossRef
14. M. Mori, J. Heterocycl. Chem., 2000, 37, 623; CrossRef M. Mori, A. Akashi, M. Hori, K. Hori, M. Nishida, and M. Sato, Bull. Chem. Soc. Jpn., 2004, 77, 1655; CrossRef M. Mori, J. Organomet. Chem., 2004, 689. 4210. CrossRef
15. For review; R. F. Heck, Palladium Reagent in Organic Synthesis; Academic Press: New York, 1985; Chapter 8.
16. M. Mori, Y. Uozumi, and M. Shibasaki, J. Organomet. Chem., 1990, 395, 255. CrossRef
17. Y. Uozumi, E. Mori, M. Mori, and M. Shibasaki, J. Organomet. Chem., 1990, 399, 93. CrossRef
18. Y. Uozumi, N. Kawasaki, E. Mori, M. Mori, and M. Shibasaki, J. Am. Chem. Soc., 1989, 111, 3275. CrossRef
19. Y. Uozumi, M. Mori, and M. Shibasaki, J. Chem. Soc., Chem. Commun., 1991, 81; CrossRef M. Mori, Y. Uozumi, and M. Shibasaki, Heterocycles, 1992, 33, 819. CrossRef
20. M. Kawaguchi, S. Hamaoka, and M. Mori, Tetrahedron Lett., 1993, 34, 6907; CrossRef M. Mori, M. Kawaguchi, M. Hori, and S. Hamaoka, Heterocycles, 1994, 39, 729. CrossRef
21. K. Shiina, J. Am. Chem. Soc., 1972, 94, 9266. CrossRef
22. Catalytic conversion of molecular nitrogen into silylamide using molybdenum- and tungsten-dinitrogen complexes was achieved by K. Hidai, H. Komori, Y. Oshita, Y. Mizobe, and M. Hidai, J. Am. Chem Soc., 1989, 111, 1939. CrossRef
23. Unpublished data.
24. Although titanium nitrogen complex 32 would be a mixture of titanium-imide complex, titanium-amide complex and others, a mixture of them is called as titanium-nitrogen complex 32.
25. M. Hori and M. Mori, J. Org. Chem., 1995, 60, 1480. CrossRef
26. M. Mori, K. Hori, A. Akashi, M. Hori, Y. Sato, and M. Nishida, M. Angew. Chem. Int. Ed., 1998, 37, 636. CrossRef
27. A. Akashi, M. Nishida, and M. Mori, Chem. Lett., 1999, 465; CrossRef M. Akashi and Mori, Heterocycles, 2003, 59, 661. CrossRef
28. [2+2] Cycloaddition of a carbonyl group and metal-imide has been reported. See: S. M. Rocklage and R. R. Schrock, J. Am. Chem. Soc., 1980, 102, 7808. CrossRef
29. [2+2] Cycloaddition of an alkyne and imide complex has been reported. See: P. L. McGrane, M. Jensen, and T. Livinghouse, J. Am. Chem. Soc., 1992, 114, 5459; CrossRef D. Duncan and T. Livinghouse, Organometallics, 1999, 18, 4421. CrossRef
30. A. Arcadi and E. Rossi, Synlett, 1997, 667; CrossRef A. Arcadi and E. Rossi, Tetrahedron, 1998, 54, 15253. CrossRef
31. Unpublished data.
32. Unpublished data.
33. F. J. Ritter, I. E. Rotgans, E. Talman, P. E. J. Verwiel, and F. Stein, Experientia, 1973, 29, 530. CrossRef
34. T. T. Shawe, C. J. Sheils, S. M. Gray, and J. L. Canard, J. Org. Chem., 1994, 59, 5481. CrossRef
35. P. E. Sonnet, D. A. Netzel, and R. Mendoza, J. Heterocycl. Chem., 1979, 16, 1041; CrossRef H. Tanaka, H. Bandoh, and T. Momose, Tetrahedron, 1993, 49, 11205. CrossRef
36. M. Mori, M. Hori, and Y. Sato, J. Org. Chem., 1998, 63, 4832. CrossRef
37. K. Wiesner, Fortschr. Chem. Org. Nartust., 1962, 20, 271; D. B. MacLean, Alkaloids (N.Y.), 1968, 10, 305.
38. Total synthesis of (±)-lycopodine: G. Stork, R. A. Kretchmer, and R. H. Schlessinger, J. Am. Chem. Soc., 1968, 90, 1647; CrossRef G. Stork, Pure Appl. Chem., 1968, 36, 383; CrossRef C. H. Heathcock, E. F. Kleinman, and E. S. Binkley, J. Am. Chem. Soc., 1982, 104, 1054; CrossRef E. Wenkert and C. A. Broka, J. Chem. Soc., Chem. Commun., 1984, 714; CrossRef W. A. Ayer, W. R. Bowman, T. C. Joseph, and P. Smith, J. Am. Chem. Soc., 1968, 90, 1648; CrossRef D. Schuman, H. J. Muller, and A. Nauman, Liebigs Ann. Chem., 1982, 1700; CrossRef G. A. Karus and Y. S. Hon, J. Am. Chem. Soc., 1985, 105, 4341. CrossRef
39. W. Oppolzer and X. Petrzilka, Helv. Chim. Acta, 1978, 61, 2755. CrossRef
40. Synthesis of racemic pumiliotoxine C, see T. Ibuka, N. Masaki, I. Saji, K. Tanaka, and Y. Inubushi, Chem. Pharm. Bull., 1975, 23, 2779; T. Ibuka, Y. Mori, and Y. Inubushi, Chem. Pharm. Bull., 1978, 26, 2442; A. I. Meyers and G. Milot, J. Am. Chem. Soc., 1993, 115, 6652, references are therein; CrossRef Synthesis of (-)-pumiliotoxin C, W. Oppolzer and E. Flaskamp, Helv. Chim. Acta, 1977, 60, 204; CrossRef S. Murahashi, S. Sasano, E. Saito, and T. Naota, J. Org. Chem., 1992, 57, 2521; CrossRef D. L. Commins and A. J. Dehghani, J. Chem. Soc., Chem. Commun., 1993, 1838; CrossRef T. G. Back and K. Nakajima, J. Org. Chem., 1998, 63, 6566; CrossRef Synthesis of (+)-pumiliotoxin C, see M. Toyota, T. Asoh, and K. Fukumoto, J. Org. Chem., 1996, 61, 8687. CrossRef
41. J. W. Daly, T. Tokuyama, G. Habermehl, I. L. Karle, and B. Witkop, Liebigs Ann. Chem., 1969, 729, 198; CrossRef T. Tokuyama, T. Tsujita, A. Shimada, H. M. Garaffo, T. F. Spande, and J. W. Daly, Tetrahedron, 1991, 47, 5401. CrossRef
42. D. Caine, T. Chao, and A. Smith, Org. Syn., 1988, Coll. Vol., 6, 51.
43. M. Akashi, Y. Sato, and M. Mori, J. Org. Chem., 2001, 66, 7873. CrossRef
44. K. Hori and M. Mori, J. Am. Chem. Soc., 1998, 120, 7651; CrossRef K. Ueda, Y. Sato, and M. Mori, J. Am. Chem. Soc., 2000, 122, 10722; CrossRef K. Ueda and M. Mori, Tetrahedron Lett., 2004, 45, 2907. CrossRef