Direct conversion of amino acids to oxetanol bioisosteres via photoredox catalysis

Carboxylic acids are an important structural feature in many drugs, but are associated with a number of unfavorable pharmacological properties. To address this problem, carboxylic acids can be replaced with bioisosteric mimics that interact similarly with biological targets but avoid these liabilities. Recently, 3-oxetanols have been identified as useful carboxylic acid bioisosteres that maintain similar hydrogen-bonding capacity while decreasing acidity and increasing lipophilicity. However, the installation of 3-oxetanols generally requires multistep de novo synthesis, presenting an obstacle to investigation of these promising bioisosteres. Herein, we report a new synthetic approach involving direct conversion of carboxylic acids to 3-oxetanols using a photoredox-catalyzed decarboxylative addition to 3-oxetanone. Two versions of the transformation have been developed, in the presence or absence of CrCl3 and TMSCl cocatalysts. The reactions are effective for a variety of N-aryl α-amino acids and have excellent functional group tolerance. The Cr-free conditions generally provide higher yields and avoid the use of chromium reagents. Further, the Cr-free conditions were extended to a series of N,N-dialkyl α-amino acid substrates. Mechanistic studies suggest that the Cr-mediated reaction proceeds predominantly via in situ formation of an alkyl-Cr intermediate while the Cr-free reaction proceeds largely via radical addition to a Brønsted acid-activated ketone. Chain propagation processes provide quantum yields of 5 and 10, respectively.


Supporting Information
A. Supplementary Tables S1 and S2 and Supplementary Figures S1 and S2 S2 B. Materials and Methods S8 C. Synthesis of N-Aryl-a-Amino Acids (1c-g,l-t,x) S10 D. Conversion of N-Aryl-a-Amino Acids (1) to 3-Oxetanols (3) S15 E. Conversion of N,N-Dialkyl-a-Amino Acids (4) to 3-Oxetanols ( a Yields based on 1 H-NMR analysis of crude reaction product and are relative to N-phenyl glycine. b 1 equiv N-phenyl glycine to 1 equiv 3-oxetanone. c 1 equiv N-phenyl glycine to 2 equiv 3-oxetanone. d 1 equiv Nphenyl glycine to 3 equiv 3-oxetanone. e Reaction mixture was not degassed with Ar. f Reaction mixture was not irradiated with blue LEDs.  Figure S1 above for full-scale reaction scheme. Reaction arrows are in blue; propagation cycles are in teal; relative flux is represented by arrow size. (a) The Crmediated reaction proceeds primarily through the alkyl-Cr addition pathway (8 + 2 ® 10), including the propagation cycle via Cr-carboxylate 12. (b) In the presence of methanol but absence of oxetanone, only the protodecarboxylation products 13a/b can be formed. (c) In the presence of both methanol and oxetanone, flux increases through the radical pathway, thus paradoxically decreasing protodecarboxylation to 13a/b. (d) In the absence of TMSCl, the Cr-dependent propagation cycle is blocked or decreased, thus decreasing the quantum yield (F = 1.6 vs 5.1 with TMSCl); propagation may still occur to some extent via TMSCl-independent formation of Cr-carboxylate 12 and/or by the SET propagation cycle in the radical addition pathway. (e) The Cr-free reaction proceeds primarily through the radical addition pathway (6 + 2 ® 7), including the SET propagation cycle that regenerates radical 6. (f) In the presence of methanol but absence of oxetanone, the radical intermediate 6 has no productive pathway forward other than to shunt to photocatalyzed reduction to carbanion intermediate 9 and then to protodecarboxylation products 13a/b. (g) In the presence of both methanol and oxetanone, flux increases through the radical addition pathway, resulting in increased yield of the 3-oxetanol product 3a.

Supporting Information
Page S8

Reagents
Reagents were obtained from Aldrich Chemical (www.sigma-aldrich.com) or Acros Organics/Tokyo Chemical Industry (www.fishersci.com) and used without further purification.

Reactions
All reactions were performed in flame-dried glassware under positive Ar pressure with magnetic stirring unless otherwise noted. Liquid reagents and solutions were transferred thru rubber septa via syringes flushed with Ar prior to use.
Photochemical reactions were performed using a PR160 Rig with fan kit and irradiated with four PR160L LEDs (40 W, lmax = 456 nm) from Kessil (www.kessil.com). This apparatus was enclosed within an aluminum covered box. The reaction vials were placed 5 cm from the LEDs and the temperature was measured to be between 25 °C to 30 °C using this configuration.

C. SYNTHESIS OF N-ARYL-a-AMINO ACIDS (1c-g,l-t,x)
N-Aryl a-amino acids 1c-g, 1l-t, and 1x, which were not commercially available, were synthesized using a literature protocol 2 with the minor modification of running the reaction at 60 °C instead of rt.

General Procedure for Ullmann Coupling
In a 50-mL roundbottom flask, CuI (23 mg, 0.25 mmol, 10 mol%), Cs2CO3 (1.63 g, 5 mmol, 2 equiv), and the appropriate a-amino acid (3 mmol, 1.2 equiv) were combined. The mixture was evacuated and refilled with Ar three times, then dissolved in DMF (3.7 mL). The appropriate aryl iodide (2.5 mmol, 1 equiv) and 2-isobutyrylcyclohexanone (0.82 mL, 0.5 mmol, 20 mol%) were added. The reaction mixture was heated to 60 °C and stirred until complete conversion as determined by TLC and 1 H-NMR analysis. The mixture was allowed to cool to rt, then diluted with 1N HCl to pH = 4, and extracted with 25 mL EtOAc (3x). The combined organic extracts were washed with 30 mL of water (2x) and brine, dried (Na2SO4), filtered, and concentrated by rotary evaporation. Purification by silica flash chromatography (10-30% EtOAc in hexanes, with 1% AcOH) afforded the N-aryl a-amino acid product 1.    N-Phenyl-L-aspartic acid (1t): Reaction solvent was changed to methanol. Prepared from iodobenzene and L-aspartic acid in 15% yield as an off-white solid. The NMR spectra of 1t matched those reported in literature. 4

Supporting Information
Page S19
The following control experiment was also carried out to exclude formation of 4b by H/Dexchange between 4a and CH3OD in the presence of CsOPiv.

DIFFERENTIAL PULSE VOLTAMMETRY (DPV) MEASUREMENTS OF 3-OXETANONE
DPV measurements were performed using a 660E potentiostat/galvanostat model from CH Instruments. A standard three-electrode configuration was used for these experiments, which were a glassy carbon working electrode, a platinum wire counter electrode, and a Ag/AgCl reference electrode. The supporting electrolyte used was tetrabutylammonium tetrafluoroborate (TBAPF6) dissolved in acetonitrile (0.1 M) and analyte concentration was 20 mM. Standard reduction potentials were obtained against Ag/AgCl and converted against standard calomel electrode (SCE). The 3-oxetanone wave was confirmed using Fc/Fc + as an internal standard.
Supplementary Figure S3. Differential pulse voltammetry of 3-oxetanone and 3-oxetanone with 1 equiv. pivalic acid. This data indicates that N-phenyl glycine (cesium salt), cesium pivalate, 3-oxetanone, and Cr(III)Cl3, and TMSCl quench the excited state photocatalyst, however, the Stern-Volmer constant of N-phenyl glycine (cesium salt) is much larger than that of the other reagents.  This data indicates that N-phenyl glycine (cesium salt), cesium pivalate, and 3-oxetanone quench the excited state photocatalyst, however, the Stern-Volmer constant of N-phenyl glycine (cesium salt) is much larger than that of the other reagents.

QUANTUM YIELD DETERMINATION
The quantum yield of the reaction was determined using a published protocol. 6

Determination of photon flux
The photon flux of the LED (Kessil PR160L, 40 W, lmax = 456 nm) was measured via standard ferrioxalate actinometry. 7,8 A solution of ferrioxalate (0.15 M) was prepared by dissolving 2.21 g of potassium ferrioxalate hydrate in 30 mL of 0.05 M H2SO4. A buffered solution of 1,10phenanthroline was prepared by dissolving 25 mg of phenanthroline and 5.63 g of sodium acetate in 25 mL of 0.5 M H2SO4. Solutions were stored in the dark. 3 mL of the ferrioxalate solution was added to 4 mL vials and irradiated for 60 seconds. After irradiation, 0.525 mL of the phenanthroline solution was added and the sample was allowed to rest for 1 hour for coordination. Next, the mixture was transferred to a quartz cuvette and the absorbance was measured at 510 nm. Non-irradiated samples as controls were also prepared. Photoconversion of Fe 3+ to Fe 2+ was calculated using eq 1.

1.
!" = V * DA * V is the total volume (0.003525 L), DA is the difference in absorbance at 510 nm between the irradiated and non-irradiated samples, is the path length (1 cm), and is the molar absorptivity at 510 nm (11,100 L mol -1 cm -1 ). After the !" was calculated from the equation 1, the photon flux was determined using eq 2. The average photon flux from 3 experiments was determined to be 1.79 x 10 -8 einsteins per second.
The quantum yield (Φ) was determined to be 5.2.
The quantum yield (Φ) was determined to be 1.

Conditions 2 (Cr-free)
In a 4-mL glass vial, 4CzIPN (1.58 mg, 2 µmol, 1 mol%), cesium pivalate (56 mg, 0.24 mmol, 1.2 equiv), and the appropriate N-aryl a-amino acid 1 (0.2 mmol, 1 equiv) were combined. Methylene chloride (0.4 mL) was added followed by 3-oxetanone (2) (38.4 µL, 0.6 mmol, 3 equiv). The mixture was degassed with Ar for 1 min. The reaction was exposed to blue LED light (456 nm) at 25 °C for 3 minutes. After irradiation, the yield was determined using 1 H-NMR analysis of crude reaction product in the presence of mesitylene as an internal standard, relative to limiting reagent 3-oxetanone. The yield obtained was 16% (3.2 x 10 -5 mol of product). The quantum yield of the reaction was calculated using eq 4, where photon flux is 1.79 x 10 -8 einsteins per second, is time (180 seconds), and is the fraction of light absorbed by the reaction mixture in the conditions described (0.96).
The quantum yield (Φ) was determined to be 10.