Waldemar
Schilling
a,
Yu
Zhang
ab,
Prakash Kumar
Sahoo
b,
Samir Kumar
Sarkar
a,
Sivaraman
Gandhi
c,
Herbert W.
Roesky
a and
Shoubhik
Das
*b
aDepartment of Chemistry, Georg-August-Universität Göttingen, Tammanstraße 2-4, 37077 Göttingen, Germany
bDepartment, of Chemistry, Univeriteit Antwerpen, Groenenborgerlaan 171, 2020 Antwerpen, Belgium. E-mail: shoubhik.das@uantwerpen.be
cDepartment of Chemistry, Gandhigram Rural Institute, Gandhigra, 624032, Tamilnadu, India
First published on 23rd November 2020
We have discovered chlorophyll catalyzed 1,2-acyl migration reactions to achieve α-amino carbonyl compounds directly from the enaminones. In general, singlet oxygen is generated during photosynthesis in the photosystem II center. This singlet oxygen can readily react with the unsaturated double bonds present in biomolecules. This reactivity intrigued us to apply this concept towards unsaturated enaminones and others to achieve highly valuable compounds. Indeed, this photosensitizer is very cheap, commercially available, main group metal based and provided excellent efficiency for singlet oxygen mediated chemistry by achieving high turnover number (TON) > 300 with a high turnover frequency (TOF) of 50 h−1. Finally, a combination of DFT calculations and detailed mechanistic experiments provided the exact role of the photosensitizer and clear insights into the reaction.
α-Amino carbonyl compounds are widely represented among pharmaceutically active compounds, peptides, and complex natural products.6–9
The most straightforward way to access α-amino carbonyl compounds is the electrophilic α-amination reaction, which directly introduces a nitrogen atom to the α-position of CH-acidic carbonyl substrates (Fig. 1).10–20
This strategy relies on trapping the enolates from the corresponding ketones with an electrophilic nitrogen source. Therefore, the formation of stable enolates from the respective ketones is the prerequisite. To avoid this, excellent achievements have been made by the group of Kürti and Wirth for the selective transformation of silyl enol ether to the corresponding α-amino carbonyls (Fig. 1).21,22
Very recently, concomitant 1,2-acyl migration and the cleavage of C–C bonds to achieve valuable products have also become a strong priority in organic synthesis.23–42 However, in most of the cases, strained cyclic ketone or the presence of a directing group is necessary. To avoid these, Wei et al. recently developed a 1,2-acyl migration reaction followed by olefin isomerization to achieve α,β-unsaturated ketones.43 However, an expensive Ni-complex and strict anhydrous conditions are required to achieve a high yield of the products. In contrast, enaminones can undergo 1,2-acyl migration reactions in the presence of Ru and Pt-based expensive transition metal based photosensitizers (Ru-catalyst: 88 Eur per g, Pt-catalyst: >200 Eur per g) along with a stoichiometric amount of base to access the α-amino carbonyl compounds.44,45 Alternatively, transition metal-free compounds such as main group metal-based complexes provide an interesting avenue for further research and appeal towards the broader synthetic applications.46–59 Intrigued by all this information, we rationalized that if singlet oxygen can be generated using chlorophyll as a photosensitizer and can be applied to the unsaturated double bonds in enaminones, it should generate the corresponding hydroperoxide by reacting with the double bond (Fig. 1). This hydroperoxide can further undergo dehydration and 1,2-acyl migration to the desired α-amino carbonyl compounds. Indeed, chlorophyll is a very cheap (2.13 Eur per g), commercially available, and main group metal (Mg) based porphyrin complex. In fact, the use of natural pigments is a well-known procedure for the generation of singlet oxygen in photochemistry, which was pioneered by Günther Otto Schenck in the 1940s by utilizing chlorophyll for the synthesis of organic compounds.60–69 Based on the early advances for the utilization of singlet oxygen, the Schenck ene reaction and others were developed for a variety of different compounds.70–76
Fig. 2 Comparison of different photosensitizers with chlorophyll for the 1,2-acyl migration reaction. |
With these optimized reaction conditions in hand, the scope of the 1,2-acyl migration reactions was explored to observe the potential of the photosensitizer (Scheme 1; entries 1b–15b). At first, different alcohols were utilized to change the esterification residue of the newly generated α-amino carbonyl products via 1,2-acyl migration. It should be noted that upon increasing the chain length of the alcohols, the inclusion of branch alcohols was well tolerated and yielded the corresponding products with an excellent TON of 297 (Scheme 1; entries 1b–6b). We observed that with the increase in chain length, the reaction speed decreased slightly. This effect was assumed to be correlated with the higher viscosity of the reaction solution and thereby was slowing the overall reaction speed. In the case of benzyl alcohol, the oxidation of benzyl alcohol to benzaldehyde and benzoic acid was observed but that did not perturb the desired reaction due to the presence of large excess of the benzyl alcohol (Scheme 1; entry 6b). Additionally, all of these reactions exhibited TOF > 20 h−1 which is quite high for the photocatalytic organic synthesis.
Later, our interest to observe the different substitution patterns at the aromatic ring revealed that a wide variety of aromatic rings could be tolerated and could generate high TON for this 1,2-acyl migration reaction (Scheme 1; entries 7–15b). For example, the reaction exhibited moderate to excellent TON (up to 273) when the aromatic ring contained alkyl (7–9b), methoxy groups (10b) and halides (12b, 14–15b). Additionally, ortho substitution in the aromatic ring with respect to the amino group did not hamper the migration (9b).
Inspired by these above results, we became further interested to explore the 1,2 migration reactions of the longer alkyl chain in the carbonyl part and also on the α-carbon part of the enaminones. To our delight, a pentanone moiety was migrated and generated the corresponding α-amino carbonyl compound with a TON of 237 (Scheme 2, entry 16b). Further extension of the alkyl chain next to the amino group at α-carbon in the starting material led to the TON of 270 (Scheme 2, entry 17b). Interestingly, benzylic amine-substituted amines also exhibited an excellent TON of 237 (Scheme 2, entry 18b). However, changing the aromatic amines to aliphatic amines led to the decomposition of the starting materials under the photocatalytic conditions.
Later, an investigation of the impact of different wavelengths on this reaction revealed that the photosensitizer can also be highly reactive under the irradiation of red LED (low in energy); however, it exhibited less reactivity compared to the blue LED (Fig. 3). The observed yields are in accordance with the measured absorbance spectrum of the utilized photosensitizer. These results intrigued us to apply this reaction under the irradiation of direct sunlight. Indeed, sunlight is renewable and available throughout the entire world. The average intensity of the total solar irradiance is about 1366.1 W m−2 which provides roughly 4.3 × 1020 J energy only in 1 h.77–84 Therefore, if it is harvested properly and is utilized in organic synthesis, it can solve many of the sustainable issues. To our delight, the utilization of direct sunlight exhibited up to 82% yields of the products within 5 h (Scheme 3). Additionally, a g-scale reaction under the irradiation of sunlight was also performed efficiently.
Fig. 3 Absorbance spectrum of chlorophyll and the yields of the model substrate at different wavelengths after a 6 h reaction time. |
After having excellent solar energy driven 1,2-acyl migration reactions, we sought to diversify chlorophyll as a photosensitizer for various singlet oxygen mediated organic transformations (Scheme 4).85–90 For this purpose, we applied this photosensitizer for the oxidative dearomatisation of phenol substrate to achieve para-quinol. To our delight, chlorophyll exhibited an excellent yield of 81% for the formation of the desired product (19b) and this photosensitizer system avoided previously reported transition metals as well as stoichiometric oxidants and reagents.91–95 Furthermore, chlorophyll was applied onto anthracene to obtain endo peroxide (20b) with a yield of 78%.96,97 Furthermore, chlorophyll can selectively cleave both the C–C and C–N bonds in the imidazole moiety to access amide (21b) with a moderate yield of 62%.98–100 Later, our efforts to transform α-terpinene in a single step to ascaridole, an anthelmintic drug (22b), was highly successful with the yield of 83%.101–103
Scheme 4 Chlorophyll as a robust photosensitizer for singlet oxygen mediated organic transformations. |
Inspired by the generality of this photosensitizer, we sought to gather mechanistic information about the role of the light, oxygen, and the photosensitizer in our reactions (Table 1). According to the control experiments, no product was observed in the absence of light, photosensitizer and oxygen/air. Furthermore, the effect of different quenchers was investigated to figure out the reactive oxygen species and possible intermediates in our reactions (Table 1).104–107 When 2,6-di-tert-butyl-4-methylphenol (BHT) or 2,2,6,6-tetramethyl-1-piperidinyloxyl (TEMPO) was added, the yield of the reaction was not changed which proved that the reaction was not following a radical pathway. Further addition of sodium azide in the reaction displayed the presence of singlet oxygen.
Conditions | Quencher | Quenching | Yield |
---|---|---|---|
a STD = standard reaction conditions: 0.3 mol% of photosensitizer, 0.3 mmol of substrate, 3.5 mL MeOH in air, 20 W blue LED, 6 h, rt. | |||
No oxygen | — | — | 0 |
No catalyst | — | — | 0 |
No light | — | — | 2 |
STD | NaN3 | Singlet oxygen | 0 |
STD | BHT | Radical | 89 |
STD | TEMPO | Radical | 87 |
Furthermore, a Stern–Volmer plot revealed that the quenching of the excited state of the chlorophyll was dependent on both the oxygen saturation as well as the concentration of the substrate (Fig. S2–4†).108 Detailed investigation of the absorption and emission spectra with and without the presence of substrate exhibited that both of the spectra underwent significant changes and indicated that the substrate was forming a substrate–catalyst intermediate, which provided access to the fast energy transfer between the catalyst and substrate (Fig. 4). After the addition of the substrate, an additional band was observed with high intensity, which was not present for the pure compounds alone (Fig. 4, left). The quenching of the photosensitizer by oxygen without the presence of the substrate further supported the involvement of singlet oxygen species in the reaction. To further prove the presence of the singlet oxygen species in the reactions, EPR-spin trap (TMP = tetramethyl piperidine) investigations were performed (Fig. 5). Performing the reaction without the oxygen atmosphere revealed no EPR signal which was in strong accordance with the control experiments from Table 1. However, as soon as the reaction was performed under an oxygen atmosphere (with and without the presence of the substrate), typical singlet oxygen signals arrived in the EPR spectra which showed clear proof of the involvement of singlet oxygen in this reaction.109–113
Fig. 4 3D-absorption and emission spectra for the migration reaction. Left: Substrate–chlorophyll mixture in MeOH (dilution 1:80). Right: Chlorophyll in MeOH (dilution 1:80). |
Further labeling experiments under an 18O2 atmosphere generated a single labeled product and indicated that only one oxygen atom was incorporated from the oxygen atmosphere (see the ESI†). This result combined with the different esterified products from the respective alcohols (Scheme 1) provided evidence that the second oxygen was originated from the alcohol. The oxygen in the acyl group seemed therefore intact and indicated an acyl shift during the reaction to obtain the desired products. Primary kinetic experiments were also performed and a linear time dependence of 16% yield/hour was observed (Fig. S5†). Additionally, ‘on and off’ experiments revealed that under darkness, an additional product (up to a maximum of 8%) was generated (Fig. S6†).
Detailed investigation of the conversion and yield over time in the dark revealed that the conversion was increased during the first 10 min followed by the generation of the product and then further decreased the conversion (Fig. 6). This observation could be explained by assuming two different pathways. The main pathway could activate the substrate and generate singlet oxygen at the same time resulting in a fast reaction. Based on the fact that oxygen was present in high excess and obviously, more singlet oxygen was produced. The singlet oxygen reacted with the substrate and formed a reversible intermediate at a lower rate.
Fig. 6 Time-dependent conversion and yield of the reaction after 2 h of the irradiation and samples were taken in the dark phase. |
To gain further mechanistic insights into this reaction, computational studies were undertaken to employ the density functional theory (DFT) calculations at the ub3lyp/6-311++g(d,p) level of theory in conjunction with the CPM solvent model for MeOH (Fig. 7 and Scheme 5). All the ground state geometries were optimized followed by a subsequent frequency test to ascertain stationary points. The absorption of a photon upon light irradiation (blue LED = 460 nm) excited the chlorophyll molecule from its ground state to the excited state (chlorophyll*) which interacted with the molecular oxygen. The excitation energy migration from chlorophyll* converted the ground-state oxygen (O2) to singlet oxygen (1O2) and chlorophyll* returned to its ground state by the energy transfer pathway. The frontier molecular orbitals of the O2 bound chlorophyll in the excited state, the highest occupied molecular orbital (HOMO) was spread over the π moiety of the porphyrin ring whereas the lowest unoccupied molecular orbital LUMO resided on O2 alone. It clearly indicated the charge/energy transfer from the porphyrin ring to the oxygen molecule in the excited state (Fig. S7†). The HOMO and LUMO for the chlorophyll catalyst were mainly observed in the π moiety of the porphyrin ring in both the ground and in the excited state (Fig. S8†).
The in situ generated reactive 1O2 was then reacted with the electron-rich double bond of enaminone (A) and formed an unstable four-membered dioxetane intermediate (B) which was +6.92 kcal mol−1 uphill in comparison with the starting enaminone substrate (A). The computed energy barrier for this step was +6.92 kcal mol−1 and it is possible to overcome it under the irradiation of blue LED (λ = 460 nm, energy = 2.69 eV). Our model suggested that the four-membered dioxetane intermediate possessed ring strain energy. Thus, the dioxetane intermediate immediately converted to the hydroperoxide intermediate (C) via a ring-opening pathway after abstracting the hydrogen atom from the allylic hydrogen adjacent to the nitrogen atom. The formation of the hydroperoxide intermediate step was favorably exergonic, at −3.06 kcal mol−1, with respect to the four-membered dioxetane intermediate. The hydroperoxide intermediate subsequently underwent a stepwise elimination of H2O to form the dicarbonyl imine intermediate (D). This dehydration step was thermodynamically more favorable and highly exergonic, giving dicarbonyl imine intermediate at −52.76 kcal mol−1 after the removal of H2O molecule which is considered to be the rate determining step. The nucleophilic addition of methanol or MeO− to dicarbonyl imine produced the intermediate (E). Two nucleophilic addition pathways were located and investigated using methanol and MeO− as nucleophiles. Computationally, both of the pathways were feasible, although the barrier height of methanol was 2 kcal mol−1 higher than MeO− for the nucleophilic addition reaction. Proceeding via the more favorable MeO− assisted pathway, the nucleophilic addition reaction was more feasible by −11.77 kcal mol−1 to yield the intermediate (E). The subsequent intramolecular 1,2-acyl migration insertion of (E) to form the final product (F) was computed to occur by protonation with a barrier of −19.4 kcal mol−1, suggesting that the product formation (F) was facile and exergonic by intramolecular 1,2-acyl migration from the β- to α-position of the nitrogen atom.
Free energy calculations were also carried out on various transition states such as molecular oxygen bound chlorophyll (TS-I), dioxygen diradical (triplet O2 bound chlorophyll (TS-II) and mono oxygen bound catalyst (TS-III) (Fig. S9†) in this reaction. The free energy of activation for the formation of different transition states followed the order of mono oxygen bound catalyst (TS-III) > dioxygen diradical bound catalyst TS-II > dioxygen TS-I with the B3LYP functional sets. The overall activation barrier of the TS-III was 24.3 kcal mol−1, which was 12.2 kcal mol−1 higher than TS-II (Fig. S10†). Intrinsic reaction coordinate (IRC) calculations clearly predicted that the conversion of molecular oxygen to the singlet oxygen took place via the dioxygen diradical pathway by energy transfer (ET).114
Combining all this information, the mechanism of this reaction was finally proposed (Scheme 5). At first, chlorophyll was excited by light and the excited state of the chlorophyll underwent (a) an energy transfer (ET) to oxygen to form singlet oxygen which was clear from the EPR spectroscopy and DFT calculations and (b) an energy transfer (ET) to the substrate which was clear from the Stern–Volmer plot. Literature reports also suggested that the oxidation potential of the substrate (1a) (E1a.+/1a) is +1.24 V vs. SCE and is quite higher than the potential of the excited chlorophyll (+0.94 V vs. SCE).44,45,115,116 Therefore, chlorophyll in the excited state could not undergo single electron transfer (SET) with the substrate and ET was more preferable. The substrate was reacted with singlet oxygen to form the peroxo intermediate via the dioxetane intermediate and formed a reversible diketone intermediate after the dehydration. The labeling experiment confirmed the single labeled oxygen provided by the oxygen atmosphere. Therefore, the dicarbonyl compound was irreversibly attacked by the methoxide anion and finally, the 1,2-acyl shift generated the desired product.
Footnote |
† Electronic supplementary information (ESI) available: Detailed synthetic procedures, analytical data for compounds, optimization studies and additional mechanistic experiments. See DOI: 10.1039/d0gc03555f |
This journal is © The Royal Society of Chemistry 2021 |