In polar aprotic solvents, the bimolecular SN2 pathway between charged species generally proceeds more rapidly than in proton solvents. For example, azide reacts ~105 times slower with trimethylsulfonic acid in methanol than in acetone. The classical explanation is that the hydrogen bonding interaction of the proton solvent stabilizes the ionic reactants better than the partially charged SN2 transition state, resulting in a higher activation barrier. Selective catalytic reactions involving the ionic SN2 mechanism must overcome the problem that ionic reactants are more stable than the transition state, which leads to a decay of the reaction rate. Of the few enzymes that promote the SN2 mechanism, most use general acid-base catalysis to activate uncharged nucleophiles or electrophiles. There is another catalytic mode for the nucleophilic halogenase 5′-fluoro-5′-deoxyadenosine synthetase (FDAS), which promotes the displacement of fluorine (or chlorine) on the cation S-adenosylmethionine (SAM). This enzyme can precisely localize the halide in a collinear relationship with the leaver group by the halide-binding active site and provide "halide holes" to counteract the energy compensation of the nucleophile dissolved from water, which increases the catalytic efficiency of the enzyme by ~106-fold. In this enzymatic mechanism, the stability of the transition state relative to the ground state has no effect on rate acceleration, instead, the enzyme active site is pre-organized to stabilize the ground state conformation, which geometrically resembles the transition state.
Eric N. Jacobsen, born in 1960, is an United States organic chemist, graduated with a Ph.D. at the age of 26, and became a professor at Harvard University at the age of 33. He has been working hard in the field of chiral organocalysis for many years, and has made a number of major breakthroughs, and is like a frequent visitor to Nature and Science.
受此启发，在该工作中，来自美国哈弗大学的Eric N. Jacobsen教授团队通过再现酶的几何预组织原理，证明了一种小分子氢键供体催化剂加速了对映选择性Michaelis-Arbuzov反应的SN2步骤。 该工作以题为“Catalysis of an SN2 pathway by geometric preorganization”发表在《Nature》上。
Figure 1. Inspiration and schematic diagram of the reaction
[Basic information of the reaction]
An empirical study of a chiral hydrogen bond donor (HBD) catalyst for the Arbuzov reaction variant showed that HCl had good enantioselectivity for the dipthylenylphenylphosphonic acid 2a dealkylation reaction. Thiourea 1c has higher enantioselectivity than its counterparts chloramide (1a) or urea (1b), facilitating model reactions. In addition, the size of aryl substituents and the α-quaternary substitution on pyrrolidine were positively correlated with enantiomeric selectivity. Ultimately, the authors found that thiourea 1 g containing a 2-phenyl substituent facilitated the dealkylation reaction with 90% EE and quantitative yields. Enantioselectivity can be significantly enhanced by changing the reaction solvent to toluene, using an equal amount of HCl, and diluting the reaction mixture to 50 mM to obtain 3a at 95% ee.
Figure 2. Reaction optimization
【Mechanism Research】
In the experiments, the authors found that catalyst 1g can speed up the reaction of 2a by about 30 times compared to the non-catalytic pathway. As a first step in elucidating the catalytic mechanism, the authors performed a simple synthetic modification of 1g to separate its anionic binding ("halide hole") domain from the putative cationic binding domain. The simple anion-bound variant 4 is constructed by substituting the tetrahydropyrrole fraction of T-isoleucine with one n-octyl group, which mitigates the activity of non-catalytic reactions when used in excess of chemical doses, and can inhibit dealkylation reactions below the background reaction rate. Thus, the simple thiourea analogue of catalyst 1g is comparable to proton solvents in its ability to inhibit ion-pair collapse. To investigate the role of the isolated t-isoleucine aryl tetrahydropyridine domain in catalysis, the authors synthesized variant 5 by S-methylation of thiourea, thereby removing the double HBD property of catalyst 1g. Compared to the blank control reaction, compound 5 did not cause a rate acceleration. At the same time, mixing 4 and 5 with 10 mol% each did not result in an increase in the rate. Taken together, these observations provide insight into the importance of the presence of these two domains and their precise relative spatial orientation in 1G for catalysis. The authors used 31P NMR spectroscopy to monitor the reaction between thiourea 1g catalyzed 2a and HCl, and proved that protonated phosphorus ions played an intermediate role in the catalytic system. The observations showed that under catalytic conditions, phosphorus chloride-6a was the resting state of the substrate.
The diffusion constant of catalyst 1g under catalytic reaction conditions was determined by NMR DOSY spectroscopy, and its molecular weight was estimated by this. After calculation, it can be determined that a single 1 g catalyst solution can provide a molecular weight consistent with the state of the monomer. In addition, the measured molecular weight of the catalyst before the reaction was consistent with the 1:1 1 g · 6a complex. After the completion of the reaction, its molecular weight was measured between the molecular weight of the monomer 1g and 1:1 1g · 3a complex, which was in line with the balance between free and product-binding catalysts, and did not conform to catalytic-product strong binding. These observations show that catalyst 1g forms a 1:1 complex with phosphoric acid 6a, and the rate-limiting dealkylation reaction is carried out by the complex.
After determining the molecular composition of the resting-state complexes, the authors sought to determine the kinetic dependence on [1g]T and [6a] to elucidate the stoichiometry of the rate-determining transition complexes. In-situ infrared indicates that the reaction rate is first-order dependent on the catalyst [1g]T. In addition, in the first 80% of the reaction, both the consumption of phosphoric acid 6a and the formation of product 3a follow a zero-order kinetic rate behavior.
Figure 3. Possible reaction mechanisms
[Computational Research]
The authors simulated the catalytic and non-catalytic dealkylation reactions of phosphonium chloride 6a using a low dielectric (PCMtoluene, e=2.38) continuous solvent method, and analyzed the interaction between the catalyst and phosphonium chloride. In the absence of a catalyst, 6a is a tight ion-pairing with a cage-like structure in which the chloride anion undergoes a multi-stage stable interaction with the cation. In order to perform the SN2 reaction and meet its linear geometry requirements, it is necessary to sacrifice the stable hydrogen bond and Coulomb attraction interaction. Considering this geometric recombination, the synergistic pathway of the dealkylation group can be divided into an ion-pair recombination phase followed by an ion-pair collapse phase. By analysis, more than 75% of the total electron activation barrier is due to the recombination of chloride ions in the first phase, while the ion-pair collapse corresponding to covalent bond cleavage and formation contributes < 25% (about 4 kcal/mol) to the overall electron activation barrier.
In the computational model of the catalytic pathway, it was found that nucleophilic substitution was carried out in two discrete steps. Three critical points can be distinguished along the reaction coordinates: the binding of 1g to ion-to-6a, the binding of recombinant ion-to-6a', and the transition state of the dealkylation step. In all three structures, the cationic and anionic components of the catalyst with chloride-phosphorus ion pairs are involved in a network of stable non-covalent interactions. Chlorine is bound to thiourea and phosphorus is bound to arylpyrrolidine. These interactions stabilize the two charged components of the ion pair. The geometry of the phosphonium chloride resting-state complex 1g·6a' is very similar to that of 6a', i.e., the chlorine is positioned in an almost optimal pre-transition geometry by hydrogen bonding. Therefore, catalyst 1g can be seen as participating in a non-covalent interaction network to obtain a relatively stable ground-state complex 1g·6a', which provides the preparation for the dealkylation reaction. Through computational analysis, the effect of catalyst 1g on ion-pair recombination and ion-pair collapse can be quantitatively evaluated. Compared to the non-catalytic pathway, catalyst binding improves the barrier to ion-pair collapse. However, the inhibition of ion-pair collapse is offset by the catalyst because it reduces the necessary geometric pre-organization energy of the phosphonium chloride ion pair, resulting in an acceleration of the reaction relative to the blank control.
Figure 4. Computational research
【Scope of application of substrate】
Finally, the authors tested the substrate applicability for this reaction. In this work, dibenzylphosphonate was shown to be a compatible substrate that could provide a chiral H-phosphonate product that is stable to air and water. Various paragonoyl arylphosphonate dealkylation reactions have good yields and enantioselectivity (3a-3g), and low enantioselectivity for substrates with high electron-withdrawing substituents (3H-3J). Meta-substituted phenylphosphonates undergo a dealkylation reaction with enantioselectivity comparable to that of their para-substituted isomers (3k-3m). Conversely, ortho-substitutions with a large steric hindrance have been shown to be disadvantageous for enantioselectivity. Ortho-fluorine substitution reduces enantiomeric selectivity from 90% (3F) to 73% EE (3N), while o-phenyl substitution reduces enantiomer selectivity (30). However, substrates containing ortho-fused cyclic aryl substituents exhibit high enantioselectivity for dealkylation reactions (3p and 3q). The method is also compatible with a variety of heteroaromatic and multi-aromatic substituents (3R-3T and 3V), including acid-labile functional groups (N-Boc-protected indole, 3U) with non-aryl substituents for the dealkylation reaction of phosphonates such as isopropylene (3W) and adamantyloxy (3X) with moderate levels of enantioselectivity, while the dealkylation reactions of alkyl substituents such as cyclopropyl and methyl groups have lower enantioselectivity of 54% and 30% EE, respectively. The enantioselective dealkylation reaction of 2A was successfully carried out in grams at 3 mol % 1 g, with high enantioselectivity (93% EE), yield (98%) and efficient catalyst recovery (95%).
Figure 5. Substrate application
In summary, this work describes a novel method for a small molecule hydrogen bond donor catalyst to accelerate the SN2 dealkylation step in a chiral Michaelis-Arbuzov reaction. Mechanistic and computational studies have shown that the catalyst accelerates the rate-determining step by recombining the geometric configurations of phosphium cation and chloride anion. This is the first time to achieve catalytic enantioselective control of the dealkylation step of phosphorus chiral compounds, which provides a new synthesis platform for the synthesis of P-chiral compounds. This method is suitable for a variety of substituted dibenzylphosphonate substrates and can be performed efficiently at gram scales.
--Testing Services--
Source: Frontiers of Polymer Science