Final Report Summary - ORCASYN (Asymmetric Organocatalytic Cascade Reactions: <br/>Applications toward the Synthesis of Complex Natural Products)
Introduction
Many biologically active molecules, such as morphine, dextromethorphan, and strychnine, contain the morphan moiety (2-azabicylo-[3.3.1]-nonane). This 6,6-fused bicycle is also present in the Daphniphyllum alkaloid family which have been isolated from the stems and leaves of Daphniphyllum Yunnanese, a shrub endemic to the Yunnan province in China. Of primary interest are the natural products daphniyunnine B (1) and D (2) which exhibit cytotoxic activity toward a variety of cancer cell lines (e.g. IC50 3.0 μM against P-388 leukemia, IC50 0.6 μM against A-549 lung cancer, etc.). To further probe their biological profiles sufficient quantities of these natural products is required; however, the natural source contains less than 0.005% of the dry weight plant mass. Moreover, there is little available in the existing literature to define an effective strategy toward the total synthesis of these alkaloids. To address the current gap in literature we proposed to develop the requisite methods to complete the first total synthesis of these unusual 6,5,6,5,7 pentacyclic natural products (see Figure 1).
Our research methodology was split into three main objectives; the development of the 2-azabicyclo-[3.3.1]-nonenone core (3), the development toward the common tetracyclic intermediate (4), and the end-game synthesis of the natural products (4 --> 1-2; see Scheme 1).
Results and discussion:
A. Synthesis of the bicyclic core using Au-catalysis
The development of the 2-azabicyclo-[3.3.1]-nonenone core (3) is key to an expedient total synthesis of the daphniyunnine natural products. The original synthesis of this core developed previously by Dr. Fillipo Sladojevich starts with (S)-carvone, a chiral pool chemical, which is elaborated into 3 in 16-steps. While this strategy allows access to the core, the number of operations required significantly limits its synthetic appeal. Our objective was to develop a new route to the 2-azabicyclo-[3.3.1]-nonenone core (3, see Scheme 2), ideally starting with chiral pool chemicals or utilising an organocatalytic strategy. It was envisaged that the 2-azabicyclo-[3.3.1]-nonenone core (3) could be derived from a gold-catalysed hydroalkylation of allylaminocyclohexenone 5 (see Scheme 2), a reaction inspired by the work of Che and co-workers. Enantiomerically pure 5 would itself be constructed through organocatalysis or from the chiral pool.
At first, a synthetic route involving a Michael addition reaction of the readily available α-aminoacetones 6 to a γ,δ-unsaturated β–ketoester 7 was proposed (see Scheme 3, eq 1). However, a screen of reaction conditions quickly determined that this system was not reactive (Scheme 3, eq. 1). Interestingly, we found that the intramolecular Michael addition of a methyl ketone to an internal nitro olefin afforded the nitrocyclohexanone (8 --> 9) with moderate enantioselectivity (see Scheme 3, eq. 2). It was reasoned that if an optimal catalyst system for this reaction could be developed, it could be used for the synthesis of the optically pure aminocyclohexenone 5 (after a few functional group transformations). Unfortunately, only moderate levels of enantiocontrol were obtained in this reaction after a thorough screen of conditions and catalyst systems. A number of other strategies to the aminocyclohexenone were and are currently being studied involving organocatalysis.
Concurrently, the gold-catalysed hydroalkylation was examined with the allyl protected aminocyclohexenone 5. It was found that the nitrogen protecting group greatly influences the reactivity in this reaction. While the benzyl or Boc protected amine were unreactive, the tosyl, nosyl, and DPP protected amines delivered excellent reactivity profiles. Using Echavarren’s gold catalyst we were able to synthesise the the 2-azabicyclo-[3.3.1]-nonenone core (3) in excellent yield; however, it was determined that this reaction favours the undesired diastereomer. Attempts to overcome this selectivity remain elusive even after significant changes to the cyclohexanone system (see Table 1).
B) Organocatalytic strategy toward the synthesis of the bicyclic core
An alternative strategy to the bicyclic core was planned after limited success was obtained using the Au-catalysed methodology (see Scheme 4). We envisioned that we could overcome the inherent bias of the methyl group facial selectivity in the aforementioned reaction by introducing the carbon atom from the opening of a cyclopropane (11 --> 10). This cyclopropane 11 would arise from the Simmons-Smith reaction from the less hindered face of the bicyclic olefin 12. The key reaction, an organocatalytic intramolecular aldol reaction would generate the desired stereochemistry required for preparation of the desired core (14 --> 13). This aldol reaction also desymmetrises an achiral amine 14 that is easily available from commercial chemicals (ie, 17).
This strategy was realised starting from the commercially available ketone 17 starting with a reductive amination with amine 18 using NaBH4 as the reductant (see Scheme 5). The CBz protection of the free amine gave rise to 15 in excellent conversion. Next an acid mediated ketal/acetal deprotection using optimised conditions (ie, 5% HCl in MeCN) allowed access for the keto-aldehyde 14.
A thorough screen of asymmetric catalysts for the key intramolecular aldol desymmetrisation reaction was performed (see Figure 2). After looking at a variety of primary and secondary amines, chiral lewis acids, chiral sulphuric, and chiral carboxylic acids, we found that the highest levels of asymmetry were offered by chiral phosphoric acid catalysts. After a solvent, temperature and concentration screen we were delighted to see very high enantioselectivity for this reaction. Most importantly, this is the first example of a chiral phosphoric acid catalysed aldol reaction between a ketone and an aldehyde. This reaction requires a thorough investigation and future studies in the Dixon research group may lead to impactful results in this field.
For the synthesis of the natural product, we found that the binol based phosphoric acid catalyst in toluene afforded the desired aldol adduct 13 in 70% yield and >99% ee after crystallisation (see Scheme 6). Next, the elimination was performed with Tf2O/pyridine but when the reaction was scaled up the yields dropped significantly due to an excessive exotherm. Thus, a new protocol involving POCl3/pyridine was used which after significant optimization can be used for large scale preparation of the elimination product 12.
With the elimination product in hand, a Simmons-Smith cyclopropanation was required. After looking at a variety of reaction conditions and reagents for this transformation it was found that an activated non-coordinating reagent was required for a single diastereomer of the cyclopropane product 11 in high yields (see Scheme 7). Next, the reductive CBz deprotection/cyclopropane opening proved to be more difficult than originally thought. In this reaction, the formation of π-allyl species leading to β-hydride elimination follow by reduction of the olefin gave rise to a mixture of products. Later, it was found that using Rh/C under 5 bar of hydrogen gas pressure gives rise to the desired product in >10:1 d.r. At this stage the final steps of installing an α,β-unsaturation and addition of the methyl group at the β-position remain in this concise synthesis to the key bicyclic intermediate.
C) The end-game strategy
The end-game strategy to daphniyunnines B and D is currently being applied (see Scheme 8). First access to 25 was achieved through a key tandem intramolecular Michael addition/O-allylation reaction (21 --> 23). After a Claisen rearrangement (23 --> 24) and a ring closing metathesis access was quickly secured to 25.
Specifically, with the tetracyclic core (25) in hand we were able to perform an O-allylation followed by a Claisen rearrangement to afford 26 (see Scheme 9). After deprotection and hydrolysis we obtained 27 as a crystalline solid. Most of our efforts have been focused around the cross-metathesis reaction of 27 with β,γ-unsaturated esters to afford 28 and its subsequent transformations. We proposed that after a successful metathesis reaction, we could force the olefin into conjugation with the carbonyl system, followed by an intramolecular Michael addition reaction to afford the desired pentacyclic compound 30. However, we found that the isomerisation/Michael reaction with the β,γ-unsaturated ethyl ester was not possible. While the conjugative isomerisation/Michael addition reaction did proceed, a rapid undesired intramolecular Claisen condensation reaction subsequently occurs. To stop this undesired reaction a tert-butyl ester was used, which gave access to the β,γ-unsaturated ester 28 and after treatment with a unique mixture of KOtBu/DBU led to the isomerised product 29. To affect the cyclisation, it was found that treatment with Quadrasil AP, a polymer supported primary amine, gave access to the pentacycle 30. This reaction is novel and was further investigated by a Masters student under my supervision. The oxidation of 30 with IBX led to mixture of products including the overoxidised product 31. At this time three reductions remain before access to the natural products.
Once completed this synthesis would represent a set of novel synthetic methodologies that could have a wide ranging impact to scientists around the world. In addition, biological evaluation will give us a better understanding of the pharmaceutical potential of this class of natural products.
Many biologically active molecules, such as morphine, dextromethorphan, and strychnine, contain the morphan moiety (2-azabicylo-[3.3.1]-nonane). This 6,6-fused bicycle is also present in the Daphniphyllum alkaloid family which have been isolated from the stems and leaves of Daphniphyllum Yunnanese, a shrub endemic to the Yunnan province in China. Of primary interest are the natural products daphniyunnine B (1) and D (2) which exhibit cytotoxic activity toward a variety of cancer cell lines (e.g. IC50 3.0 μM against P-388 leukemia, IC50 0.6 μM against A-549 lung cancer, etc.). To further probe their biological profiles sufficient quantities of these natural products is required; however, the natural source contains less than 0.005% of the dry weight plant mass. Moreover, there is little available in the existing literature to define an effective strategy toward the total synthesis of these alkaloids. To address the current gap in literature we proposed to develop the requisite methods to complete the first total synthesis of these unusual 6,5,6,5,7 pentacyclic natural products (see Figure 1).
Our research methodology was split into three main objectives; the development of the 2-azabicyclo-[3.3.1]-nonenone core (3), the development toward the common tetracyclic intermediate (4), and the end-game synthesis of the natural products (4 --> 1-2; see Scheme 1).
Results and discussion:
A. Synthesis of the bicyclic core using Au-catalysis
The development of the 2-azabicyclo-[3.3.1]-nonenone core (3) is key to an expedient total synthesis of the daphniyunnine natural products. The original synthesis of this core developed previously by Dr. Fillipo Sladojevich starts with (S)-carvone, a chiral pool chemical, which is elaborated into 3 in 16-steps. While this strategy allows access to the core, the number of operations required significantly limits its synthetic appeal. Our objective was to develop a new route to the 2-azabicyclo-[3.3.1]-nonenone core (3, see Scheme 2), ideally starting with chiral pool chemicals or utilising an organocatalytic strategy. It was envisaged that the 2-azabicyclo-[3.3.1]-nonenone core (3) could be derived from a gold-catalysed hydroalkylation of allylaminocyclohexenone 5 (see Scheme 2), a reaction inspired by the work of Che and co-workers. Enantiomerically pure 5 would itself be constructed through organocatalysis or from the chiral pool.
At first, a synthetic route involving a Michael addition reaction of the readily available α-aminoacetones 6 to a γ,δ-unsaturated β–ketoester 7 was proposed (see Scheme 3, eq 1). However, a screen of reaction conditions quickly determined that this system was not reactive (Scheme 3, eq. 1). Interestingly, we found that the intramolecular Michael addition of a methyl ketone to an internal nitro olefin afforded the nitrocyclohexanone (8 --> 9) with moderate enantioselectivity (see Scheme 3, eq. 2). It was reasoned that if an optimal catalyst system for this reaction could be developed, it could be used for the synthesis of the optically pure aminocyclohexenone 5 (after a few functional group transformations). Unfortunately, only moderate levels of enantiocontrol were obtained in this reaction after a thorough screen of conditions and catalyst systems. A number of other strategies to the aminocyclohexenone were and are currently being studied involving organocatalysis.
Concurrently, the gold-catalysed hydroalkylation was examined with the allyl protected aminocyclohexenone 5. It was found that the nitrogen protecting group greatly influences the reactivity in this reaction. While the benzyl or Boc protected amine were unreactive, the tosyl, nosyl, and DPP protected amines delivered excellent reactivity profiles. Using Echavarren’s gold catalyst we were able to synthesise the the 2-azabicyclo-[3.3.1]-nonenone core (3) in excellent yield; however, it was determined that this reaction favours the undesired diastereomer. Attempts to overcome this selectivity remain elusive even after significant changes to the cyclohexanone system (see Table 1).
B) Organocatalytic strategy toward the synthesis of the bicyclic core
An alternative strategy to the bicyclic core was planned after limited success was obtained using the Au-catalysed methodology (see Scheme 4). We envisioned that we could overcome the inherent bias of the methyl group facial selectivity in the aforementioned reaction by introducing the carbon atom from the opening of a cyclopropane (11 --> 10). This cyclopropane 11 would arise from the Simmons-Smith reaction from the less hindered face of the bicyclic olefin 12. The key reaction, an organocatalytic intramolecular aldol reaction would generate the desired stereochemistry required for preparation of the desired core (14 --> 13). This aldol reaction also desymmetrises an achiral amine 14 that is easily available from commercial chemicals (ie, 17).
This strategy was realised starting from the commercially available ketone 17 starting with a reductive amination with amine 18 using NaBH4 as the reductant (see Scheme 5). The CBz protection of the free amine gave rise to 15 in excellent conversion. Next an acid mediated ketal/acetal deprotection using optimised conditions (ie, 5% HCl in MeCN) allowed access for the keto-aldehyde 14.
A thorough screen of asymmetric catalysts for the key intramolecular aldol desymmetrisation reaction was performed (see Figure 2). After looking at a variety of primary and secondary amines, chiral lewis acids, chiral sulphuric, and chiral carboxylic acids, we found that the highest levels of asymmetry were offered by chiral phosphoric acid catalysts. After a solvent, temperature and concentration screen we were delighted to see very high enantioselectivity for this reaction. Most importantly, this is the first example of a chiral phosphoric acid catalysed aldol reaction between a ketone and an aldehyde. This reaction requires a thorough investigation and future studies in the Dixon research group may lead to impactful results in this field.
For the synthesis of the natural product, we found that the binol based phosphoric acid catalyst in toluene afforded the desired aldol adduct 13 in 70% yield and >99% ee after crystallisation (see Scheme 6). Next, the elimination was performed with Tf2O/pyridine but when the reaction was scaled up the yields dropped significantly due to an excessive exotherm. Thus, a new protocol involving POCl3/pyridine was used which after significant optimization can be used for large scale preparation of the elimination product 12.
With the elimination product in hand, a Simmons-Smith cyclopropanation was required. After looking at a variety of reaction conditions and reagents for this transformation it was found that an activated non-coordinating reagent was required for a single diastereomer of the cyclopropane product 11 in high yields (see Scheme 7). Next, the reductive CBz deprotection/cyclopropane opening proved to be more difficult than originally thought. In this reaction, the formation of π-allyl species leading to β-hydride elimination follow by reduction of the olefin gave rise to a mixture of products. Later, it was found that using Rh/C under 5 bar of hydrogen gas pressure gives rise to the desired product in >10:1 d.r. At this stage the final steps of installing an α,β-unsaturation and addition of the methyl group at the β-position remain in this concise synthesis to the key bicyclic intermediate.
C) The end-game strategy
The end-game strategy to daphniyunnines B and D is currently being applied (see Scheme 8). First access to 25 was achieved through a key tandem intramolecular Michael addition/O-allylation reaction (21 --> 23). After a Claisen rearrangement (23 --> 24) and a ring closing metathesis access was quickly secured to 25.
Specifically, with the tetracyclic core (25) in hand we were able to perform an O-allylation followed by a Claisen rearrangement to afford 26 (see Scheme 9). After deprotection and hydrolysis we obtained 27 as a crystalline solid. Most of our efforts have been focused around the cross-metathesis reaction of 27 with β,γ-unsaturated esters to afford 28 and its subsequent transformations. We proposed that after a successful metathesis reaction, we could force the olefin into conjugation with the carbonyl system, followed by an intramolecular Michael addition reaction to afford the desired pentacyclic compound 30. However, we found that the isomerisation/Michael reaction with the β,γ-unsaturated ethyl ester was not possible. While the conjugative isomerisation/Michael addition reaction did proceed, a rapid undesired intramolecular Claisen condensation reaction subsequently occurs. To stop this undesired reaction a tert-butyl ester was used, which gave access to the β,γ-unsaturated ester 28 and after treatment with a unique mixture of KOtBu/DBU led to the isomerised product 29. To affect the cyclisation, it was found that treatment with Quadrasil AP, a polymer supported primary amine, gave access to the pentacycle 30. This reaction is novel and was further investigated by a Masters student under my supervision. The oxidation of 30 with IBX led to mixture of products including the overoxidised product 31. At this time three reductions remain before access to the natural products.
Once completed this synthesis would represent a set of novel synthetic methodologies that could have a wide ranging impact to scientists around the world. In addition, biological evaluation will give us a better understanding of the pharmaceutical potential of this class of natural products.