London Disperse Interactions Assist Chiral Induction in the Soai Autoamplifying Reaction Provoked by 1- and 2-Aza[6]helicenes

Abstract

In this paper, DFT computations revealed the mechanisms of the asymmetric catalytic reactions of diisopropylzinc with pyrimidylaldehyde catalyzed by 1- and 2-aza[6]helicenes, which make them effective inductors of the autocatalytic chiral amplification Soai reaction. The generation of chirality takes place through the formation of adducts of aldehyde and helicenes stabilized via non-covalent disperse interactions strictly defining the orientation of the aldehyde molecule in the corresponding transition state.

1. Introduction

The autocatalytic chiral amplification Soai reaction (Scheme 1) is currently the only straightforward example of a transformation capable of effectively increasing the chirality of a catalyst [1]. In the Soai reaction, chiral alcohol 2 is a product of the alkylation of pyrimidinyl aldehyde (e.g., 1) with diisopropyl zinc, followed by acidic hydrolysis. The initial product, alcoholate 3, can effectively catalyze this transformation. Since 3 can be pre-formed by reacting 2 with diisopropylzinc, the Soai reaction is autocatalytic. Moreover, the ee of 2 obtained in this reaction is higher than that of 2 taken in catalytic amounts (Scheme 1); i.e., the autoamplification of chirality takes place. This unique feature makes the Soai reaction a convincing model for the study of possible scenarios for the evolution of homochiral life [2]. Besides the catalytic autoamplifying reaction itself, spontaneous chirality generation [3,4,5,6] and chirality generation induced by various chiral inductors [7] are well documented.

Being interested in the mechanism of chirality generation in the Soai reaction [8,9,10,11,12], we were intrigued by the results of chiral induction effectuated by aza[6]helicenes 4 and 5 (Scheme 2). The experimental results show that the handedness of the reaction products was dependent on the position of the nitrogen atom in the inductor (Scheme 2) [13]. Here, we report the results of our computations revealing the intrinsic mechanism of the inductor’s action.

2. Results and Discussion

2.1. -N[6]helicene 4 as Inductor

To be an inductor, a chemical compound must be able to promote alkylation faster than the background reaction (otherwise, the background reaction would lead to the stochastic distribution of the handedness of the product). However, the presence of an inductor should provide a reaction slower than the autocatalytic transformation; otherwise, the system will be reduced to a simple catalytic reaction known for several aminoalcohols serving as effective catalysts in the reaction of diisopropylzinc with 1 and similar aldehydes. We concentrated our efforts on identifying ways to appropriately activate substrates and generate chirality with aza[6]helicenes 4 and 5.

As can be seen in Scheme 2, N[6]helicenes with the same P-chirality induce reactions leading to opposite enantiomers of 2 and 3. This produces an important reference point that can be used to verify computed reaction pathways, thus providing insight into the mechanism of substrate activation in these reactions.

One can intuitively envisage numerous ways of arranging reagents near a molecule of 4. However, our efforts showed that it is not actually easy to locate real minima leading to chiral transition states. The experimental results clearly indicate that the nitrogen atom is directly involved in the stereoselective process. Experimentalists have suggested that this might be due to the coordination of diisopropylzinc to the nitrogen atom [13]. However, although Zn(i-Pr)₂ is capable of forming an adduct with 4, we were unable to locate further productive associates with 1.

Nevertheless, we were attracted by the possibility of chelate binding between aldehyde 1 and 1-aza[6]helicene 4 via disperse interactions of the aldehyde proton of 1 with the nitrogen atom of 4 and the oxygen atom of 1 and the N-C-H proton of 4 (Figure 1). As a result, the trimolecular associate 6 was located (Figure 1, left), which is a precursor of TS1(S) (Figure 1, right).

The starting complex 6 is evidently stabilized by two C-H⋯X disperse interactions (X = O, N). Although the strength of the weaker C-H⋯N interaction is sacrificed in TS1(S) (it becomes longer for 0.43 Å in order to accommodate the Zn(i-Pr)₂ molecule), it continues to participate to some extent in the stabilization of TS1(S), since the C-H⋯N and C-H⋯O “bonds” continue to maintain an orientation close to coplanar (dihedral angle 15.25° in TS1(S) against 2.87° in 6).

The free activation barrier for this alkylation is 23.3 kcal/mol, which is 3.0 kcal/mol lower than that for the background reaction and 6.3 kcal/mol higher than that for the autocatalyzed reaction (see

Table 1. Computed thermodynamic parameters (kcal/mol) B3LYP/6-31 G(d,p), DCM (toluene).

Parameter		ΔE(ZVPE) Activation Energy	ΔH Activation Enthalpy	ΔG Gibbs Free Activation Energy	Imaginary Frequency, cm−1
From 6	TS1(S)	20.8	19.9	23.3	i261.4
From 7	TS2(S)	23.2	22.3	25.9	i272.3
	TS2(R)	22.7	21.7	24.8	i286.4
From 8	TS3(S)	19.3	19.6	22.2	i244.5
	TS3(R)	17.3	16.4	19.9	i254.5
No catalyst		23.3	22.1	26.3	i271.8
3 as a catalyst			17.0 a

^a B3LYP/6-31 G, CPCM (toluene), Ref. [11].

); i.e., it exactly fits to the criteria of the chiral-induced reaction formulated above.

TS1(S) is a precursor of 2S, and it corresponds to the experimentally observed handedness induced by using 3 in the Soai reaction. Moreover, it is impossible to construct a direct analogue of 6 or TS1(S) that would provide the opposite enantiomer, since inverting the prochiral plane of 1 would result in a geometrically impossible realization of the disperse interactions stabilizing 6 and TS1(S).

In search of a transition state preceding the formation of 2R, we located trimolecular associate 7 (Figure 2). By inspecting the structure of 7, one can see that, in order to open the second prochiral place for the R-addition of the isopropyl group, the C-H⋯N contact must be lost. Moreover, although the C-H⋯O contact is kept in 7, it is not seen in either of the transition states. As a result, TS(R) and TS(S) become relatively unstable, and the more stable one, TS2(R), is about 2 kcal/mol less stable than TS1(S) (see Table 1). We concluded that this pathway is unlikely to significantly contribute to the flow of catalysis.

Thus, 5 is capable of inducing the stereoselective formation of 2S via the activation of 1 through its chelate-like coordination applying two disperse C-H⋯X interactions (Figure 1). This coordination mode leaves only one prochiral plane of 1 prone to alkylation, which determines the handedness of the product.

2.2. -N[6]helicene 5 as Inductor

Initially, we thought that the opposite handedness observed in the Soai reaction induced by 5 could be triggered via a similar coordination mode using two C-H⋯X interactions, since, in that case, the opposite prochiral plane of 1 would remain open.

However, it turned out that, during the preliminary computations, such a coordination type is impossible in the case of 2-aza[6]helicene 5 since the substituted alkynyl group would encounter the second half of the helicene molecule. Hence, we searched for another type of productive coordination.

These efforts resulted in the location of the three-molecule associate 8 in which a six-membered ring is formed by applying the coordination of Zn(i-Pr)₂ to the nitrogen atom of 2-aza[6]helicene 5 and two disperse interactions C-H⋯O and C-H⋯Zn (Figure 3).

Interestingly, both prochiral planes of the aldehyde seem to be equally inclined to the alkyl transfer. Nevertheless, the computations suggest that TS3(R) is more stable than TS3(S) for 2.3 kcal/mol, probably due to the smaller distortion of the three-molecule associate required for the alkyl group transfer.

Thus, we were able to computationally reproduce the opposite handedness of the reaction product formed in the Soai reaction induced by 1- and 2-aza[6]helicenes 4 and 5. In both computed reactions, intramolecular disperse interactions played an important role in the process of stereoselection. This feature appears to be a common aspect in various types of asymmetric catalyses [14,15].

Our computations of the alkylation of 2-aza[6]helicene 5 suggest that 5 may demonstrate catalytic activity comparable to that seen in an autocatalytic reaction (with tetrameric 3 as a catalyst) [11,16]. This observation indicates the need for further investigations of asymmetric catalytic reactions using 5.

It should be noted that the catalytic pathways located for the alkylations of 1 with Zn(i-Pr)₂ with 4 or 5 as a catalyst are not autocatalytic themselves. Our simplified approach suggests that, as soon as these reactions create a small excess of one of the enantiomers of the accumulating alcoholate 3, further amplification of chirality takes place via the normal mechanism involving oligomeric alcoholates [11]. One can imagine the involvement of, e.g., 5 in complicated equilibria seen in the reaction pool of the autocatalytic Soai reaction [11]. This might lead to the discovery of more complicated mechanisms of chiral induction.

3. Conclusions

Our results demonstrate that the induction of chirality in the Soai reaction carried out in the presence of 1-aza[6]helicene and 2-aza[6]helicene takes place via mechanisms in which P-chirality of the inductor determines the structure of the reactive precursor, whereas weak intramolecular disperse interactions play a significant role in creating asymmetry in the corresponding transition states.

We are convinced that similar computational studies of other numerous examples of chiral inductions in the Soai reaction would undercover new interesting mechanistic features and provide a theoretical background for fruitful experimental studies.

4. Computational Details

Computations were carried out using the hybrid Becke functional (B3) [17,18] for electron exchange and the correlation functional of Lee, Yang, and Parr (LYP) [19] as implemented in the GAUSSIAN 09 software package [20]. All atoms were modeled at the 6-31G(d,p) level of theory [21]. The starting geometries for the transition state search were located either by QST2 or QST3 procedures, or by a guess based on the structure of the previously found TS. The transition states were subsequently fully optimized as saddle points of first order, employing the Berny algorithm [22]. Frequency calculations were carried out to confirm the nature of the stationary points, yielding one imaginary frequency for all transition states, which represented the vector for the C−C bond formation. The solvent influence was accounted for by carrying out optimizations in the DCM force field [23].