Straightforward Access to Enantioenriched cis-3-Fluoro-dihydroquinolin-4-ols Derivatives via Ru(II)-Catalyzed-Asymmetric Transfer Hydrogenation/Dynamic Kinetic Resolution

Herein we report a practical method for the asymmetric transfer hydrogenation/dynamic kinetic resolution of N-Boc 3-fluoro-dihydrotetrahydroquinolin-4-ones into the corresponding cis-fluoro alcohols in 70–96% yields, up to 99:1 diastereomeric ratio (dr) and up to >99% ee (enantiomeric excess) by using the ruthenium complex Ts-DENEB and a formic acid/triethylamine (1:1) mixture as the hydrogen donor under mild conditions.


Introduction
The number of bioactive molecules containing fluorine approved by the Food and Drug Administration (FDA) has greatly increased over time. The first decade of this century saw the introduction of 40 new compounds having a fluorine atom. This number represented at the time an important increase of 20% of the commercial drugs bearing at least one fluorine atom in their structure [1]. Compared to this, in 2018 alone, 17 fluorinecontaining pharmaceuticals were approved by the FDA [2]. In a single year, the number of new fluorinated drugs covered almost half of the increase of a whole decade.
This fluorine rush is easily explained by fluorine's ability to affect important parameters for a drug candidate, such as permeability or pK a , and to modify its pharmacokinetics and pharmacodynamics. Fluorine is thus an element that gives control to medicinal chemists over tailoring the properties of a molecule. Organic chemists must then work in consideration of this increasing demand and develop new methodologies to introduce fluorine into complex molecules and into versatile building blocks [3].

Results and Discussion
The study began with the synthesis of fluorinated dihydroquinolin-4-ones 2a-j via an electrophilic fluorination of the N-Boc-protected heterocyclic ketones 1a-j by using lithium bis(trimethylsilyl)amide as a base and NFSI as the fluorine source (Scheme 2).

Scheme 2. Synthesis of fluorinated dihydroquinolin-4-ones 2a-j
3-Fluoro-dihydroquinolin-4-one carboxylate 2a was chosen as the standard substrate for the optimization of asymmetric transfer hydrogenation. Based on previous studies, the ATH was set at 40 °C in acetonitrile with a (1:1) molar mixture of formic acid and triethylamine as the hydrogen source. Under these conditions, a set of commercially available

Results and Discussion
The study began with the synthesis of fluorinated dihydroquinolin-4-ones 2a-j via an electrophilic fluorination of the N-Boc-protected heterocyclic ketones 1a-j by using lithium bis(trimethylsilyl)amide as a base and NFSI as the fluorine source (Scheme 2).

Scheme 2. Synthesis of fluorinated dihydroquinolin-4-ones 2a-j
3-Fluoro-dihydroquinolin-4-one carboxylate 2a was chosen as the standard substrate for the optimization of asymmetric transfer hydrogenation. Based on previous studies, the ATH was set at 40 °C in acetonitrile with a (1:1) molar mixture of formic acid and triethylamine as the hydrogen source. Under these conditions, a set of commercially available 3-Fluoro-dihydroquinolin-4-one carboxylate 2a was chosen as the standard substrate for the optimization of asymmetric transfer hydrogenation. Based on previous studies, the ATH was set at 40 • C in acetonitrile with a (1:1) molar mixture of formic acid and triethylamine as the hydrogen source. Under these conditions, a set of commercially available Ru(II) catalysts and a Rh(III) complex were screened (Table 1). After 3 h with a catalyst loading of 0.5 mol%, the Ru and Rh catalysts allowed a full conversion of the α-fluoro ketone into the corresponding fluoro alcohol 3a. All the complexes delivered excellent isolated yields and enantiomeric excesses for the desired fluorohydrin 3a (entries 1 to 4). The main difference appeared in the diastereoselectivity outcome. Whereas (R,R)-Rh-teth-TsDPEN ((R,R)-A; entry 1) led to a moderate 79:21 dr, (R,R)-Ru(p-cymene)TsDPEN ((R,R)-B; entry 2), (R,R)-Ru(mesitylene)TsDPEN ((R,R)-C; entry 3) and (R,R)-TsDENEB ((R,R)-D; entry 4), all gave an excellent diastereoisomeric ratio dr of 97:3 or 98:2, indicating that the reaction was neither affected by the nature of the arene ligand nor the tethered or untethered characteristic of the complex. From this series of results, the (R,R)-TsDENEB ((R,R)-D) catalyst was chosen for the optimization conditions. Ru(II) catalysts and a Rh(III) complex were screened (Table 1). After 3 h with a catalyst loading of 0.5 mol%, the Ru and Rh catalysts allowed a full conversion of the α-fluoro ketone into the corresponding fluoro alcohol 3a. All the complexes delivered excellent isolated yields and enantiomeric excesses for the desired fluorohydrin 3a (entries 1 to 4). The main difference appeared in the diastereoselectivity outcome. Whereas (R,R)-Rh-teth-TsDPEN ((R,R)-A; entry 1) led to a moderate 79:21 dr, (R,R)-Ru(p-cymene)TsDPEN ((R,R)-B; entry 2), (R,R)-Ru(mesitylene)TsDPEN ((R,R)-C; entry 3) and (R,R)-TsDENEB ((R,R)-D; entry 4), all gave an excellent diastereoisomeric ratio dr of 97:3 or 98:2, indicating that the reaction was neither affected by the nature of the arene ligand nor the tethered or untethered characteristic of the complex. From this series of results, the (R,R)-TsDENEB ((R,R)-D) catalyst was chosen for the optimization conditions. The solvent of the asymmetric transfer hydrogenation was then investigated. By using a catalyst loading of 0.5 mol% of (R,R)-D, formic acid/triethylamine (1:1) mixture and a reaction time of 3 h, we tested a series of different solvents ( Table 2). Several polar aprotic (CH3CN, CH2Cl2, THF, Me-THF and EtOAc) and polar protic (MeOH, iPrOH and HFIP) as well as aromatic solvents were screened (toluene, chlorobenzene and trifluoromethylbenzene). The diastereo-and enantioselectivities remained mostly unchanged in all the tests. The conversion rate and the yield showed to be highly dependent on the solvent used. Isopropanol and acetonitrile led to 95% and 100% conversions, respectively, with excellent stereoselectivities in both solvents, acetonitrile slightly surpassing isopropanol in all of them. This prompted us to set acetonitrile as the solvent for the asymmetric transfer hydrogenation. The solvent of the asymmetric transfer hydrogenation was then investigated. By using a catalyst loading of 0.5 mol% of (R,R)-D, formic acid/triethylamine (1:1) mixture and a reaction time of 3 h, we tested a series of different solvents ( Table 2). Several polar aprotic (CH 3 CN, CH 2 Cl 2 , THF, Me-THF and EtOAc) and polar protic (MeOH, iPrOH and HFIP) as well as aromatic solvents were screened (toluene, chlorobenzene and trifluoromethylbenzene). The diastereo-and enantioselectivities remained mostly unchanged in all the tests. The conversion rate and the yield showed to be highly dependent on the solvent used. Isopropanol and acetonitrile led to 95% and 100% conversions, respectively, with excellent stereoselectivities in both solvents, acetonitrile slightly surpassing isopropanol in all of them. This prompted us to set acetonitrile as the solvent for the asymmetric transfer hydrogenation.
We continued the optimization by varying the hydrogen source of the reaction ( Table 3). The acid-to-base ratio was examined (entries 1 to 4). Changing the formic acid to triethylamine ratio from 1:1 to 2:5 (entry 2) showed no impact on the reaction. However, switching to a HCO 2 H/Et 3 N ratio of 5:2 (entry 3) was detrimental to the diastereoselectivity which dropped to a 54:46 dr. Moreover, when this ratio was further increased to 12:1 (entry 4), no conversion was observed. This strong influence of the acid/base ratio of the hydrogen donor reflects the one that was observed in the ATH/DKR of 3-fluorochromanones [6]. Other organic bases associated to formic acid (DBU, entry 5, and DABCO, entry 6) as well as formate salts (ammonium formate and calcium formate, entries 7 and 8, respectively) were screened, all giving lower conversions and similar stereoselectivities than Et 3 N. Finally, the catalyst loading (0.2 mol%; entry 9) and the reaction temperature (room temperature; entry 10) were lowered. Both tests gave lower conversions in spite of longer reaction times, so we set the optimized conditions for the ATH/DKR as 0.5 mol% of (R,R)-D as the catalyst, HCO 2 H/Et 3 N (1:1) mixture as the hydrogen donor, acetonitrile as the solvent and 40 • C as the reaction temperature.  We continued the optimization by varying the hydrogen source of the reaction ( Table  3). The acid-to-base ratio was examined (entries 1 to 4). Changing the formic acid to triethylamine ratio from 1:1 to 2:5 (entry 2) showed no impact on the reaction. However, switching to a HCO2H/Et3N ratio of 5:2 (entry 3) was detrimental to the diastereoselectivity which dropped to a 54:46 dr. Moreover, when this ratio was further increased to 12:1 (entry 4), no conversion was observed. This strong influence of the acid/base ratio of the hydrogen donor reflects the one that was observed in the ATH/DKR of 3-fluorochromanones [6]. Other organic bases associated to formic acid (DBU, entry 5, and DABCO, entry 6) as well as formate salts (ammonium formate and calcium formate, entries 7 and 8, respectively) were screened, all giving lower conversions and similar stereoselectivities than Et3N. Finally, the catalyst loading (0.2 mol%; entry 9) and the reaction temperature (room temperature; entry 10) were lowered. Both tests gave lower conversions in spite of longer reaction times, so we set the optimized conditions for the ATH/DKR as 0.5 mol% of (R,R)-D as the catalyst, HCO2H/Et3N (1:1) mixture as the hydrogen donor, acetonitrile as the solvent and 40 °C as the reaction temperature.  We continued the optimization by varying the hydrogen source of the reaction ( Table  3). The acid-to-base ratio was examined (entries 1 to 4). Changing the formic acid to triethylamine ratio from 1:1 to 2:5 (entry 2) showed no impact on the reaction. However, switching to a HCO2H/Et3N ratio of 5:2 (entry 3) was detrimental to the diastereoselectivity which dropped to a 54:46 dr. Moreover, when this ratio was further increased to 12:1 (entry 4), no conversion was observed. This strong influence of the acid/base ratio of the hydrogen donor reflects the one that was observed in the ATH/DKR of 3-fluorochromanones [6]. Other organic bases associated to formic acid (DBU, entry 5, and DABCO, entry 6) as well as formate salts (ammonium formate and calcium formate, entries 7 and 8, respectively) were screened, all giving lower conversions and similar stereoselectivities than Et3N. Finally, the catalyst loading (0.2 mol%; entry 9) and the reaction temperature (room temperature; entry 10) were lowered. Both tests gave lower conversions in spite of longer reaction times, so we set the optimized conditions for the ATH/DKR as 0.5 mol% of (R,R)-D as the catalyst, HCO2H/Et3N (1:1) mixture as the hydrogen donor, acetonitrile as the solvent and 40 °C as the reaction temperature. Once the aforementioned conditions were identified, we subjected the family of substituted tert-butyl 3-fluoro tetrahydroquinoline-4-ones carboxylates (2a-j) previously synthesized to the asymmetric transfer hydrogenation. The scope (Scheme 3) showed that the fluoro dihydrotetrahydroquinolin-4-ones bearing electron-donating or electron-withdrawing groups all led to good yields (90-96%), diastereo-(94:6 to 99:1 dr) and enantioselectivities (99-> 99% ee) to the corresponding fluorohydrins (3a-j). Notably, the asymmetric transfer hydrogenation tolerated substituents such as chlorine, bromine or iodine along with other groups such as methyl, methoxy and trifluoromethyl. Concerning the outcome of the reaction, the two substrates bearing a methoxy-group in position 7 of the aromatic ring (2d and 2e) gave a slightly lower diastereoselectivity compared to the rest of the dihydroquinolin-4-ones. The ATH of compound 2e, possessing two methoxy groups in positions 6 and 7, also required a higher catalyst loading (1.0 mol%) and a longer reaction time (24 h) to achieve a full conversion.
The absolute configuration of compound 3a was unambiguously assigned as (3R, 4S) by the X-ray crystallographic analysis. We conjectured that the other products of the developed ATH/DKR reaction led to 3-fluoro-dihydrotetrahydroquinolin-4-ols derivatives with the same absolute configuration (Scheme 3). The usefulness of the studied asymmetric transfer hydrogenation was evaluated by performing the reaction on a gram-scale with substrate 2a (Scheme 4). 1.0 g of tert-butyl 3-fluoro-dihydrotetrahydroquinolin-4-one carboxylate was subjected to the optimized reaction conditions which led to the cis-fluorohydrin 3a in 96% yield, 97:3 dr and > 99% ee. In addition, the post-functionalization of compound 3j using a Sonogashira cross-coupling reaction with phenylacetylene was carried out in the presence of PdCl 2 (PPh 3 ) 2 and CuI to afford the corresponding alkyne 4 in 68% yield. Finally, N-Boc deprotection of compound 3a was readily achieved by heating in dioxane/H 2 O and the desired amine 5 was isolated in 90% yield.

General Information
All air-and/or water-sensitive reactions were carried out under an argon atmosphere. THF, CH 2 Cl 2 and toluene were dried over alumina columns in a solvent purification apparatus (Innovative Technology, Oldham, UK). Methanol, isopropanol, chlorobenzene, trifluoromethylbenzene, ethyl acetate and acetonitrile from Sigma-Aldrich (Darmstadt, Germany) were used without further purification. Hexafluoroisopropanol (HFIP) was purchased from TCI and was used without further purification. Formic acid/triethylamine (1:1) mixture was purchased from Fluka or Alfa Aesar and was used without further purification. Reactions were monitored by thin layer chromatography carried out on precoated silica gel plates (Merck 60F254, Darmstadt, Germany) and revealed with either an ultraviolet lamp (λ = 254 nm) or a potassium permanganate solution. Proton nuclear magnetic resonance ( 1 H-NMR) spectra were recorded using a Bruker AC 400 (400 MHz). The chemical shifts are expressed in parts per million (ppm) referenced to residual chloroform (7.26 ppm). Data are reported as follows: chemical shifts (δ), multiplicity (recorded as s, singlet; d, doublet; t, triplet; q, quadruplet; quint, quintuplet; sext, sextuplet; hept, heptuplet; m, multiplet and br, broad), coupling constants and integration. Carbon-13 nuclear magnetic resonance ( 13 C-NMR) spectra were recorded using a Bruker AC 400 (101 MHz). The chemical shifts are expressed in parts per million (ppm) relative to the centre line of the triplet at 77.16 ppm for CDCl 3 . Melting points (m. p.) were determined on a Köfler melting point apparatus. Optical rotations were measured on a Jasco P-1010 polarimeter. High resolution mass spectrometric (HRMS) analyses were measured on LTQ-Orbitrap (Thermo Fisher Scientific, Waltham, Massachusetts, US) at Sorbonne Université.

Method for the Synthesis of Compounds 2
To a 0 • C solution of LiHMDS (1.0 M in THF, 1.05 equiv) in a round-bottom tube set under argon, was slowly added a solution of the ketone (1.2 mmol, 1.0 equiv) in THF (2 mL) over 5 min. The mixture was stirred at 0 • C for 1.5 h. The resulting solution was then added dropwise over 5 min via a cannula to a −78 • C solution of NFSI (1.4 mmol, 1.2 equiv) in THF (4 mL). The reaction mixture was stirred at −78 • C for 30 min and then allowed to come to room temperature overnight. The reaction was diluted with 3 mL of CH 2 Cl 2 , quenched with 5 mL of saturated NH 4 Cl aqueous solution and extracted with CH 2 Cl 2 (3 × 10 mL). The combined organic layers were washed with brine, dried over MgSO 4 , filtered, and concentrated under reduced pressure [25]. The pure products were isolated by flash column chromatography on silica gel (see Supplementary Materials for NMR spectra). In a round-bottom tube charged with the corresponding 3-fluoro-4-oxo-3,4dihydroquinoline-1(2H)-carboxylate (1.0 equiv) set under argon, the necessary volume of a solution of (R,R)-A in acetonitrile (1.0 mL of a 1.24 mg/mL solution; 0.005 equiv) was added. The mixture was stirred for one minute before adding, by syringe, the necessary volume of formic acid/triethylamine (1:1) mixture (6.0 equiv). The reaction mixture was stirred at 40 • C for the time needed and then quenched with 3 mL of saturated NaHCO 3 aqueous solution. The media was extracted with CH 2 Cl 2 (2 × 4 mL) and the organic layers were dried over MgSO 4 , filtered, and concentrated under vacuum. The diastereoisomeric ratio was determined by 1 H-NMR analysis of the crude product. The product was purified with a flash column chromatography on silica gel (petroleum ether/EtOAc) and the enantiomeric excess was determined by SFC analysis (CHIRALPAK IE or IF column) (see Supplementary Materials for NMR and SFC spectra).