Isothiourea-Catalyzed Enantioselective Michael Addition of Malonates to α,β-Unsaturated Aryl Esters

An enantioselective Michael addition of malonates to α,β-unsaturated para-nitrophenyl esters was achieved using the Lewis basic isothiourea HyperBTM, giving excellent levels of product enantioselectivity (up to >99:1 enantiomeric ratio) in good yields and with complete regioselectivity (>20:1 regioselectivity ratio) in the presence of alternative (phenyl ketone and ethyl ester) Michael acceptors. Density functional theory calculations indicate that N-acylation is rate-limiting. This constitutes a rare example of a highly enantioselective addition of simple, readily available malonates to α,β-unsaturated esters.

Room temperature (r.t.) refers to 20−25 °C. Temperatures of 0 °C and −78 °C were obtained using ice/water and CO 2 (s)/acetone baths, respectively. Reflux conditions were obtained using a DrySyn, oil bath, or sand bath equipped with a contact thermometer.
Analytical thin layer chromatography was performed on pre-coated aluminium plates (Kieselgel 60 F 254 silica). TLC visualisation was carried out with ultraviolet light (254 nm), followed by staining with a 1% aqueous KMnO 4 solution. Manual column chromatography was performed in glass columns fitted with porosity 3 sintered discs over Kieselgel 60 silica using the solvent system stated. Automated chromatography was performed on a Biotage Isolera Four running Biotage OS578 with a UV/Vis detector using the method stated and cartridges filled with Kieselgel 60 silica.
Melting points were recorded on an Electrothermal 9100 melting point apparatus and are uncorrected.

Determination of Product Configuration by X-Ray Crystallography
X-ray diffraction data were collected at 148 K using a Rigaku XtaLAB P200 diffractometer [Cu Kα radiation (λ = 1.54184 Å)]. Data were collected using CrystalClear 5 and processed (including correction for Lorentz, polarization and absorption) using CrysAlisPro. 6 Structures were solved by dual-space (SHELXT 7 ), direct (SIR2011 8 ) or charge-flipping (Superflip 9 ) methods and refined by full-matrix leastsquares against F2 (SHELXL-2018/3 10 ). Non-hydrogen atoms were refined anisotropically, and all hydrogen atoms were refined using a riding model. All calculations were performed using the CrystalStructure 11 interface. Crystals suitable for X-ray diffraction analysis were obtained using the vapor

Methodology
All calculations were performed in Gaussian 09 suite of programs. 12 Where applicable, the initial geometries for the calculations were adapted from the optimised geometries in the study from Wang et al. 13 Optimisation were performed at the M06-2X 14 /6-31G(d,p)/IEFPCM THF level, using a polarisable continuum The enantiomeric ratio (er) was calculated from a Boltzmann equilibrium at that temperature.

Results and Discussion
Our target system bears considerable resemblance to that from Wang et al. 13 for the synthesis of pyridones and pyranones. Both systems are catalysed by isothiourea HyperBTM (5) and include Michael addition reactions to fluorinated α,β-unsaturated esters. The key differences are that we use a different aryl alcohol ester (para-nitrophenolate, OPNP, instead of 2,4,6-trichlorophenolate) and a different nucleophile for Michael addition (dimethyl malonate ester 7 instead of 2-acylbenzazole). In the system of Wang et al.,13 catalytic turnover is achieved from an intramolecular cyclisation step, whereas in our system, catalytic turnover is driven by the free aryl oxide. Because of this close similarity of both systems, it is reasonable to assume that they would have similar transition state and intermediate geometries in the initial steps. We therefore used the study from Wang el al. as inspiration for our own calculations, which were performed at the same level of theory. The resulting profile including the key steps is summarised in Figure S1. Figure S1: Reaction profile for Michael addition of dimethyl malonate (7') to α,β-unsaturated ester (6) catalysed by HyperBTM (5)  Following the procedure from Wang et al. 13 the free energies are reported relative to an encounter complex between catalyst 5 and our model reactant 6 (R = CF 3 in Scheme 4 in the main paper), denoted 5•6. This is to minimise artifacts from entropies of associative steps that are evaluated from standard thermodynamic expressions based on ideal-gas approximation. At 0°C, formation of 5•6 is computed endergonic by ∆G = 23.2 kJ mol -1 . 17 The first key intermediate, M1, is obtained as contact ion pair between a cationic isothiouronium complex and the OPNPleaving group. Depending on which face of the prochiral enone moiety the phenolate is located, two diastereomeric forms are possible, of which the si form is less stable despite a potentially favourable − interaction (see Figure S2), presumably because the steric clash between the two aromatic moieties leads to an unfavourably large charge separation in the zwitterion. The lowest barrier leading to one of the diastereomeric M1 intermediates (the more stable of the two, actually), is found via re-TS1 at ∆G ‡ = 52.8 kJ mol -1 .
Again following the protocol of Wang et al. 13 we assume facile exchange between the phenolate and the Michael nucleophile, the deprotonated malonate ester (7'), and have located the transition states for attack of the latter at the -carbon of the , -unsaturated ketone moiety in M1, as well as the resulting zwitterionic intermediates, M2. This is the point where the stereochemistry of the final product, (R or S) is determined.
In addition to the stereochemistry in the product, there is some conformational flexibility about the newly formed C-C single bond. We have trialled several such conformations (for both R and S intermediates M2 and transition states, TS2) and report only the results for the most stable of each in Figure S1.
It turns out that S-M2 is more stable than R-M2, but only by ∆∆G = 3.6 kJ mol -1 . The steric clash that, arguably, favours S-M2 over R-M2 is illustrated in Figure S3. There is also some conformational flexibility in the malonate moiety itself. In the M2 products, like in the free neutral malonate ester 7, the two carbonyl groups adopt a gauche conformation (as opposed to an anti orientation with C 2 -or pseudo C 2 -symmetry). These conformations would connect to the deprotonated malonate 7' in its cis configuration (C 2v -symmetry), however for the free anion 7', the trans configuration (C s -symmetry) turned out to be slightly more stable (by ∆G = -4.9 kJ mol -1 ). 11 We located a total of 10 transition states for TS2 involving both cis-7' and trans-7'; the relative free energies of these TSs are collected in Table S1. Those leading to S products tend to be more stable than those leading to R. The free energy difference between the lowest of each, S-TS2 and R-TS2 (both involving trans-7'), is ∆∆G ‡ = 17.5 kJ mol -1 .
steric clash S22 The structures of S-TS2 and R-TS2 are shown in Figure S4 (which is a stereo version of Figure 1 in the main paper). A similar steric clash as in the product R-M2 ( Figure S3) is seen in R-TS2 (Figures 1 and S4).
It is arguably this clash that causes an elongated C ... C distance in that TS (2.57 Å at the M06-2X level), compared to the same distance in S-TS2 (2.40 Å). In fact, there appears to be a loose correlation between that distance and the barrier height, the latter tending to increase with the former (see r C ... C values in Table   S1). Note that S-and R-TS2 do not connect directly to S-and R-M2, but to slightly higher-lying rotamers (S-M2' and R-M2', not shown), which can convert to the more stable intermediates S-M2 and R-M2 via simple rotation about C-C single bonds. This step, addition of the malonate, is computed so exergonic (e.g. ∆G = -78.9 kJ mol -1 from re-M1 to S-M2) that the barriers for the reverse reaction are essentially unsurmountable under the mild reaction conditions (e.g. ∆G ‡ = 91.9. kJ mol -1 from S-M2 to S-TS2). This addition is therefore irreversible and under kinetic control. The stereochemistry of the final product should thus be determined by the free-energy difference between the (selectivity-determining) transition states S-TS2 and R-TS2. This aforementioned difference of ∆∆G ‡ = 17.5 kJ mol -1 corresponds to a computed er of 99.95:0.05 at 273.15 K, in excellent agreement with the experimental values exceeding 99:1 (see Table 1 in the main paper). For a full prediction of the stereocontrol one could calculate the amount of S-and R-product from appropriate Boltzmann averages over all relative barriers leading to these products (∆∆G ‡ rel values in Table S1, weighted by the Boltzmann equilibrium of nucleophiles cis -7' and trans-7'), which would afford a very similar outcome, namely essentially exclusive formation of S-product. 19

S24
The reaction is completed by the protonation of the enolate moiety in M2 and substitution of the HyperBTM moiety with the aryl enolate, affording the final product (8) and regenerating the organocatalyst (5). Starting from the lowest intermediate S-M2, the intermediate of the first step of this sequence was modelled via reaction with free aryl phenol, HOPNP affording a contact ion pair S-M3 (again following the procedure of Wang et al.,13 in order to avoid artifacts from charge separation with the simple solvation model). The energetics of this step include the driving force for formation of the phenol via deprotonation of the neutral malonate ester (7), according to , ∆G = 54.4 kJ mol -1 (eq 2), which produces the Michael nucleophile 7' needed earlier. Formation of S-M3 and its decay into the final products, S-8 and 5, are computed to be so favourable (with driving forces for each elementary step between ∆G  -26 to -33 kJ mol -1 , see the last two steps on the profile in Figure S1) that no kinetic hindrance and, thus, no bearing on the stereocontrol is to be expected. Therefore, no other stereoisomers were considered for M3 and no transition states connecting to it were located at this stage.