Hydrogen Transfer-Mediated Multicomponent Reaction for Direct Synthesis of Quinazolines by a Naphthyridine-Based Iridium Catalyst.

Selective linkage of renewable alcohols and ammonia into functional products would not only eliminate the prepreparation steps to generate active amino agents but also help in the conservation of our finite fossil carbon resources and contribute to the reduction of CO2 emission. Herein the development of a novel 2-(4-methoxyphenyl)-1,8-naphthyridine-based iridium (III) complex is reported, which exhibits excellent catalytic performance toward a new hydrogen transfer-mediated annulation reaction of 2-nitrobenzylic alcohols with alcohols and ammonia. The catalytic transformation proceeds with the striking features of good substrate and functional group compatibility, high step and atom efficiency, no need for additional reductants, and liberation of H2O as the sole by-product, which endows a new platform for direct access to valuable quinazolines. Mechanistic investigations suggest that the non-coordinated N-atom in the ligand serves as a side arm to significantly promote the condensation process by hydrogen bonding.

Introduction

Mass mining and consumption of fossil resources have resulted in a call for the development of new catalytic transformations, enabling production of functional chemicals from renewable resources with high step and atom efficiency (Goldemberg, 2007, Michlik. and Kempe, 2013a, Michlik and Kempe, 2013b, Kozlowski and Davis, 2013). Among various alternative feedstocks, alcohols are a category of oxidized hydrocarbons that can be extensively derived from biomass including abundantly available lignocellulose via degradation (Zakzeski et al., 2010, Sun et al., 2018, Vispute et al., 2010). N-heteroarenes represent a class of highly important compounds, and they have been extensively employed for the development of valuable products, such as bioactive molecules, pharmaceuticals, agrochemicals, dyes, ligands, sensors, and materials (Boyarskiy et al., 2016, Preshlock et al., 2016, Bandini, 2011). Consequently, the linkage of alcohols into N-heteroaromatic frameworks is of high importance, as it not only helps in the conservation of our finite fossil carbon resources but also contributes to the reduction of CO2 emission.

Over the past decade, the strategy of acceptorless dehydrogenative coupling (ADC) proceeded to renew the construction of N-heteroarenes. In this strategy, dehydrogenation is involved in the activation of alcohols via in situ formation of carbonyl intermediates, and H2 and/or H2O are generally generated as the by-products. Since 2013, significant progress has been made in this regard by the groups of Milstein (Srimani et al., 2013a, Srimani et al., 2013b, Daw et al., 2016, Daw et al., 2017), Kempe (Michlik. and Kempe, 2013a, Michlik and Kempe, 2013b, Deibl et al., 2015, Hille et al., 2014, Hille et al., 2017, Deibl and Kempe, 2017, Kallmeier et al., 2017), Beller (Zhang et al., 2013a, Zhang et al., 2013b), Kirchner (Mastalir et al., 2016), and others (Pan et al., 2016, Xu et al., 2017, Elangovan et al., 2015, Chen et al., 2014). However, it is important to note that these transformations mainly rely on the utilization of specific amines, whereas the synthesis of N-heteroarenes by combining alcohols with ammonia, an abundant and renewable nitrogen source, has been rarely explored, although the related transformations would eliminate prepreparation steps to generate active amino agents, and result in high step and atom efficiency. For instance, the Beller group has reported a Ru-catalyzed synthesis of pyrroles from ammonia, vicinal diols, and ketones (Scheme 1, Equation 1) (Zhang et al., 2013a, Zhang et al., 2013b). Milstein and the co-workers have presented a synthesis of pyrroles and pyrazines from alcohols and ammonia (Scheme 1, Equation 2) (Daw et al., 2018).

Scheme 1

Alcohols and Ammonia Utilized for the Synthesis of N-Heteroarene and Amine

In recent years, the so-called hydrogen-borrowing reaction has emerged as an appealing tool in achieving the alkylation of amines (Wang et al., 2014, Xiao et al., 2019, Kaloglu et al., 2016, Elangovan et al., 2016) and activated carbon nucleophiles (Blank and Kempe, 2010, Elangovan et al., 2015, Deibl and Kempe, 2016, Peña-López et al., 2016). Interestingly, the synthesis of various alkylamines from alcohols and ammonia has also been nicely demonstrated (Scheme 1, Equation 3) (Ye et al., 2014, Pingen et al., 2010, Imm et al., 2010, Imm et al., 2011, Gunanathan and Milstein, 2008, Yamaguchi et al., 2008, Kawahara et al., 2010). In such transformations, the alcohols serve as both the hydrogen suppliers and coupling agents. So, there is no need for external reductants such as high-press H2 gas. Despite these significant advances, the construction of functional N-heteroarenes involving alcohols and ammonia feedstocks through hydrogen autotransfer as a substrate-activating strategy remains a new subject to be explored. However, such a concept would encounter the challenges of difficult proton exchanges and selectivity control, as well as catalyst deactivation by the lone pair of electrons on the nitrogen of excess ammonia (Klinkenberg and Hartwig, 2011).

Among various N-heteroarenes, quinazolines constitute a class of structurally unique compounds, which have been found to exhibit diverse biological and therapeutic activities (Parhi et al., 2013, Ugale and Bari, 2014, Juvale et al., 2013, Ple et al., 2004), and have been extensively applied for the discovery of various functional products (Zhao et al., 2013, Zhang et al., 2011). However, the existing approaches for accessing such compounds generally require preinstalled reactants (Lin et al., 2014, Malakar et al., 2012, Portela-Cubillo et al., 2008, Yan et al., 2012, Zhang et al., 2010). In this context, the search for direct synthesis of quinazolines from easily available substrates, preferably abundant and sustainable ones, would be of great significance. Enlightened by our recent work on the synthesis and functionalization of N-heterocycles (Chen et al., 2017b, Chen et al., 2018a, Chen et al., 2018b, Chen et al., 2017a, Liang et al., 2018, Liang et al., 2019, Xie et al., 2017, Xie et al., 2018, Xie et al., 2019), we wish herein to present, for the first time, a synthesis of quinazolines from 2-nitrobenzyl alcohols (Rajendran et al., 2015, Pasnoori et al., 2014), alcohols, and ammonia by a new iridium complex featuring a 2-(4-methoxyphenyl)-1,8-naphthyridyl ligand. In such a transformation, the hydrogen generated from dehydrogenation of alcohols and dehydroaromatization process is utilized for substrate activation through transfer hydrogenation (TH) of the nitro group, and there is no need for addition of external reductants.

Results and Discussion

We initiated our investigations by choosing the synthesis of quinazoline 3aa from o-nitrobenzene methanol 1a, alcohol 2a, and ammonia as a model reaction. First, we tested the combinations of several metal catalysts (i.e., Ru, Mn, Co, Fe, and Ni) with various phosphine ligands such as Xantphos, DPPE, DPPB, DPPP, Binap-dp, DPEphos, and Xphos (see Table S1), the privileged catalyst systems employed for the ADC and hydrogen-borrowing reactions. However, the low yields of product (<10%) disclosed that they were not suitable systems for the current synthetic purpose. When complex [IrCp∗Cl2]2 was employed, 15% yield of 3aa was obtained. A further optimization of other reaction parameters involving solvents, bases, and temperatures (Table S2) slightly improved the yield to 18% by using t-BuONa as the base at 140°C. Enlightened by our recent synthesis of naphthyridines (Chen et al., 2017a, Chen et al., 2017b, Chen et al., 2018a, Chen et al., 2018b, Xiong et al., 2016), we believed that such compounds might serve as a class of useful N-ligands with tunable coordination modes, and the preparation of a suitable naphthyridyl Ir-complex might offer a solution to obtain the desirable catalytic efficiency. Thus, we prepared 9-cyclometalated iridium complexes, involving 8-naphthyridyl (Ir-1−Ir-8) and 1,2-phenylpyridyl (Ir-9) ones. Then, their catalytic performance toward the model reaction was evaluated (Table 1, entries 1–9). In comparison, complexes bearing a 1,8-naphthyridyl ligand (entries 1–7) exhibited appealing activity, and Ir-3 (as confirmed by single-crystal X-ray diffraction, CCDC: 1848110, for detail, see Figure S101 and Tables S5–S10) was shown to be a preferred candidate, whereas complex with a 1,5-naphthyridyl or 2-pheynlpyridyl ligand only resulted in low product yield (entries 8–9). The results imply that the N-atom at position 8 in 1,8-naphthyridyl ligands plays a crucial role in affording a satisfactory product yield. Further optimization showed that the presence of iridium is essential in affording the product (entry 10), and the gaseous ammonia is relatively superior to other nitrogen sources (entries 11–15). A decrease of base amount to 30% resulted in a diminished yield (entry 16), and 40% t-BuONa was sufficient for the reaction (entry 17). The time-conversion profile at 2, 4, 8, and 16 h showed that the satisfactory product yield is due to the catalyst robustness (entry 18). Based on the results, the optimal (standard) conditions are as indicated in entry 17 of Table 1.

Table 1

Screening of Optimal Reaction Conditions

Entry	Catalyst	NH3 Source	Yields of 3aaa,b
1	Ir-1	NH4OAc	72
2	Ir-2	NH4OAc	75
3	Ir-3	NH4OAc	82
4	Ir-4	NH4OAc	61
5	Ir-5	NH4OAc	67
6	Ir-6	NH4OAc	71
7	Ir-7	NH4OAc	68
8	Ir-8	NH4OAc	15
9	Ir-9	NH4OAc	21
10	–	NH4OAc	–
11	Ir-3	NH4Cl	5
12	Ir-3	HCOONH4	Trace
13	Ir-3	NH3⋅H2O	Trace
14	Ir-3	(NH4)2SO4	22
15	Ir-3	NH3 (g)	88c
16	Ir-3	NH3 (g)	81c,d
17	Ir-3	NH3 (g)	88c,e
18	Ir-3	NH3 (g)	(12, 40, 65, 84)f

Also see Figure S101, Tables S5−S10 and Data S3.

Unless otherwise stated, the reaction was performed with 1a (0.5 mmol), 2a (0.5 mmol), Ir (1 mol %), t-BuONa (50 mol %), NH3 sources (1.0 mmol) in toluene (1.5 mL) for 24 h under Ar protection.

Gas chromatography yields with the use of hexadecane as an internal standard.

4 bar of NH3.

t-BuONa (30 mol %).

t-BuONa (40 mol %).

Conversions for 2, 4, 8, and 16 h.

With the optimal reaction conditions established, we then examined the generality of the synthetic protocol. (2-nitrophenyl)methanol 1a was further employed to couple with various primary alcohols (2a−2t, Scheme S1) and ammonia. As illustrated in Scheme 2, all the reactions proceeded smoothly and furnished the desired quinazolines in moderate to excellent yields upon isolation (Scheme 2, 3ab−3at). Apart from the alkyl-substituted benzyl alcohols, other functional groups such as −OMe, −OH, −NH2, −Cl, −Br, −CF3, −CO2Me, −COPh, −CN, and −C=C− are well tolerated in the transformation. The retention of these functionalities offers the potential for the elaboration of complex molecules via further chemical transformations. Moreover, except for the strong electron-withdrawing group −CF3, the electronic property of these substituents has little influence on the reaction, whereas the relatively lower product yields using ortho-substituted benzyl alcohols might relate to the steric hindrance (3ac and 3ae). Furthermore, heteroaryl methanols (2o and 2p) were also amenable to the transformation and resulted in the 2-heteroaryl-substituted quinazolines (3ao and 3ap) in good yields, and the obtained products have the potential to be applied as hemilabile bidentate ligands in organometallic chemistry and catalysis. Interestingly, cinnamyl alcohol 2q underwent smooth hydrogen transfer-mediated annulation, affording the 2-alkenyl quinazoline 3aq in 46% yield. The relatively low product yield is due to partial formation of 2-alkyl quinazoline via reduction of the alkenyl group. The relatively low product yield of 3aq is due to the partial formation of 2-alkyl quinazoline via reduction of the alkenyl group. Aliphatic alcohols, such as methanol (2r), heptan-1-ol (2s), and cyclopropyl carbinol (2t), were efficiently transformed into the 2-non-substituted and 2-alkyl quinazolines (3ar, 3as, and 3at) in moderate yields.

Scheme 2

Variation of Alcohols

Also see Scheme S1, Figures S1−S60 and Data S3.

Subsequently, we turned our attention to the transformation of different 2-nitrobenzyl alcohols 1. First, various related substrates (1b−1i) in combination with different primary alcohols 2 and NH3 were tested. As shown in Scheme 3, all the reactions smoothly delivered the multi-substituted quinazolines in moderate to excellent isolated yields. The electronic property of the substituents on the aryl ring of substrates 1 significantly influenced the product yields. In general, 2-nitrobenzyl alcohols 1 with electron-donating groups afforded the products in higher yields (3ba–3ca and 3ea–3fi) than with electron-deficient ones (3ga–3ia). This phenomenon is rationalized as the catalyst has better stability toward the electron-rich aniline intermediates, arising from the TH of nitro group of substrates 1. Gratifyingly, secondary alcohols, such as 1j and 1k, also underwent smooth annulation to give the 2,4-disubstituted quinazolines in good yields (3ja, 3jl, and 3ka). Similar to the results described in Scheme 2, a wide array of functionalities such as −Me, −OMe, −F, −Cl, −Br, −CN, −Ph, and ester are well tolerated in the transformation (Schemes 2 and 3). Noteworthy, the halogen groups did not undergo hydrodehalogenation, showing that the developed catalytic system exhibits unique chemoselectivity.

Scheme 3

Variation of o-Nitroaryl Alcohols

Also see Scheme S1, Figures S61−S91 and Data S3.

To demonstrate the significance and practicality of the developed synthetic methodology, a gram-scale synthesis of compound 3aa could be achieved by performing the reaction with 8 mmol of 1a and 9 mmol of benzyl alcohol 2a, which still afforded a good isolated product yield (78%) even with lower catalyst loading (Scheme 4, Equation a, 0.2 mol%). Furthermore, compound 3la, a key ingredient used as a herbicide with the activity on Toll-like receptors, 20 could be prepared through the reduction of commercially available acifluorfen acid to 2-nitrobenzyl alcohol 1l (Scheme S3) followed by the annulation reaction of 1l with alcohol 2a and ammonia (Equation b), and such a synthesis is far superior to the reported multi-step synthetic protocol (Mc Gowan et al., 2012, Munro and Bit, 1987, Sumida et al., 1995). Moreover, the extended π-conjugated system like compound 5ja was successfully prepared by the halocyclization (Tan et al., 2016) of compound 3ja and further Sonogashira coupling (Equation c), which offers a valuable basis for further development of optoelectronic materials.

Scheme 4

The Synthetic Utility of the Developed Chemistry

Also see Scheme S1 and S3, Figures S92−S97 and Data S3.

To gain mechanistic insights into the catalytic transformation, a time-concentration profile of the model reaction is depicted in Figure 1 (also see Data S1). Substrates 1a and 2a with ammonia were converted into 3aa in maximum yield within 24 h. 2-Aminobenzaldehyde 1a-4 and 1,2-dihydroquinazoline 3aa-1 were observed during the reaction, but they were consumed up after completion of the reaction (Figure 1). The subjection of compound 1a-4 with benzaldehyde 2a-1 and NH3 or direct treatment of 3aa-1 under the standard conditions afforded product 3aa in almost quantitative yields (see Equations 1 and 2 of Scheme S2, also see Data S1). These results support the fact that compounds 1a-4, 2a-1, and 3aa-1 are the reaction intermediates. Furthermore, both the iridium catalyst and base play crucial roles in the dehydrogenation of 3aa-1 to product 3aa (Equation 2). An iridium hydride complex (Ir-H) was obtained from the reaction of equimolar Ir-3 and benzyl alcohol, which can efficiently catalyze the reaction to afford 3aa, showing that Ir-H as a catalytic species is involved in the reaction (Equations 3 and 4, Scheme S2, also see Data S2 and Figure S98).

Figure 1

Time-Concentration Profile of the Model Reaction

With the above-mentioned preliminary experimental evidence in hand, the mechanism was further scrutinized by density functional theory calculations (geometry optimizations using B3LYP and single-point energy calculations using M06). For details, see Figures S99 and S100, Tables S3 and S4, Schemes S4–S8, and Data S4. The calculated free energy profile for the first TH (first TH) of 1a to 2-nitrosobenzaldehyde 1a-2 is shown in Figure 2. Initially, the anion exchange between Ir-3 and t-BuONa generates the alkoxy complex Ir-O1. One of the arms in 1,8-naphthyridyl ligand of Ir-O1 dissociates, allowing the Ir center to coordinate with the hydroxyl group of 1a. O–H bond cleavage occurs via transition state TS1 with an energy barrier of 21.4 kcal/mol to give Ir-alkoxide complex IN2, which then undergoes β-hydride elimination by overcoming an energy barrier of 28.0 kcal/mol (TS2) relative to IN2, and generates complex Ir-H and o-nitrobenzaldehyde 1a-1. The nitro group of 1a-1 further acts as a sacrificial hydrogen acceptor of Ir-H through two transition states TS3 and TS4. Finally, 2-nitrosobenzaldehyde 1a-2 is generated with the regeneration of Ir-O1. In addition, the base-promoted intramolecular Meerwein-Ponndorf-Verley-Oppenauer-type transfer hydrogenation (MPV-O TH) is calculated to have an energy barrier of 33.1 kcal/mol (see Scheme S4), which is 3.5 kcal/mol higher than the overall barrier of the pathway shown in Figure 2. Thus, the MPV-O TH pathway is kinetically unfavorable.

Figure 2

Calculated Energy Profiles for First TH

o-Nitrobenzene methanol 1a → 2-nitrosobenzaldehyde 1a-2. Values shown are relative free energies in kcal/mol. Also see Tables S3 and S4 and Data S4.

The calculated free energy profiles for the second TH of 2-nitrosobenzaldehyde 1a-2 to 2-(hydroxyamino)benzaldehyde 1a-3 and the third TH of 1a-3 to 2-aminobenzaldehyde 1a-4 are shown in Figures S99 and S100 (also see Tables S3 and S4, Schemes S5–S8 and Data S4). In consideration that both 2-aminobenzaldehyde 1a-4 and benzaldehyde 2a-1 can condense with ammonia, two plausible pathways toward the formation of imines were investigated. For the reaction of 2a-1 and ammonia (black line in Figure 3), ammonia approaches benzaldehyde through the C–N bond linkage (TS14) giving IN18. The TH of the ammonia using other ammonia as the proton-transferring shuttle then takes place via TS15 and leads to IN20. The calculated free energy barrier of transition state TS15 is 22.8 kcal/mol relative to IN16. After rearranging to more stable IN21 featuring two hydrogen bonds, the dehydration occurs via TS16, giving the imine complex IN22. Meanwhile, we performed calculations for the dehydration without the hydrogen-bonding between the OH group and the non-coordinated N-atom in the ligand (green line in Figure 3). The calculated free energy of transition state TS16″ is −58.4 kcal/mol, which is higher than that of TS16. Therefore, the non-coordinated N-atom in the 1,8-naphthyridyl ligand plays a crucial role in the reaction, as it serves as a side-arm to significantly promote the dehydration by hydrogen bonding. An alkoxyl anion ligand rebounds to Ir center to give imine 2a-2 with regeneration of the Ir-O2 catalyst. The reaction of 1a-4 and ammonia (purple line in Figure 3) follows similar mechanisms to those for 2a-1. The relevant mechanistic details are therefore not discussed again, for simplicity. The highest energy point for the reaction of 1a-4 and ammonia is TS16′, which is energetically less favorable by 1.8 kcal/mol compared with that of TS16 for the reaction of 2a-1 and ammonia. Therefore, from a kinetic point of view, the reaction of 1a-4 and ammonia is less kinetically favorable.

Figure 3

Calculated Energy Profiles for Coupling of Alcohol with Ammonia

Black line for benzaldehyde 2a-1 and purple line for 2-aminobenzaldehyde 1a-4. The dehydration without the assistance of the non-coordinated N-atom in the ligand is shown in green line. Values shown are relative free energies in kcal/mol. Also see Tables S3 and S4 and Data S4.

Based on all the above-mentioned findings, a plausible reaction pathway for the formation of product 3aa is illustrated in Scheme 5. In the first TH process, the Ir-catalyzed dehydrogenation of 1a via alkoxy anion exchange of Ir-O1 with 1a gives IN2, which is followed by β-H elimination to form the 2-nitrobenzaldehye 1a-1 and the Ir–H species. The successive TH to the nitro group of 1a-1 and t-BuOH-assisted dehydration forms 2-nitrosobenzaldehyde 1a-2 and regenerates the Ir-O1 species. In the second TH process, the anion exchange of Ir-O1 with 2a gives a benzyloxy complex Ir-O2. The subsequent β-H elimination of Ir-O2 followed by TH to the nitroso group and alcoholysis with 2a delivers 2-(hydroxyamino)benzaldehyde 1a-3 and regenerates complex Ir-O2, respectively. In the third TH process, the Ir-H and benzaldehyde 2a-1 are generated via β-H elimination of Ir-O2. Subsequently, the Ir-promoted dehydration of 1a-3 forms a nitrene complex IN13, and the TH using 2a as the proton-transferring shuttle generates 2-aminobenzaldehyde 1a-4 (Qu et al., 2014, Hou et al., 2017). Next, the successive formation of imine 2a-2 via the condensation of benzaldehyde 2a-1 with NH3 and the cyclization between 2a-2 and 1a-4 affords the dihydroquinazoline 3aa-1. Finally, the iridium alkoxy complex-catalyzed dehydroaromatization of 3aa-1 gives rise to product 3aa, and the in situ-generated Ir-H and alcohol further take part in the TH of the nitro group.

Scheme 5

Plausible Reaction Pathway

Also see Data S1 and S2 and Figure S98.

Conclusion

In summary, we have prepared a series of cyclometalated iridium complexes. Among them, Ir-3 featuring a 2-(4-methoxyphenyl)-1,8-naphthyridyl ligand exhibits the best catalytic performance toward the hydrogen transfer-mediated annulation of 2-nitrobenzyl alcohols with readily available alcohols and ammonia, which allows direct synthesis of a wide array of valuable quinazolines. Mechanistic investigation suggests that the non-coordinated N-atom in the ligand serves as a side arm to significantly promote the condensation step by hydrogen bonding. The catalytic transformation proceeds with the striking features of good substrate and functional group compatibility, liberation of H2O as the sole by-product, high atom and step efficiency, and no need for additional reductants. The developed chemistry paves the avenues for further development of hydrogen transfer-mediated coupling reactions by design of catalysts bearing N-side arm ligands.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.