(S)-3-cyano-5-methylhexanoic acid

• CAS: 181289-37-2
• Molecular Formula: C8H13NO2
• Molecular Weight: 155.19 g/mol
• SMILES: CC(C)CC(CC(=O)O)C#N
• InChI: InChI=1S/C8H13NO2/c1-6(2)3-7(5-9)4-8(10)11/h6-7H,3-4H2,1-2H3,(H,10,11)/t7-/m0/s1
• InChIKey: MGWZYUMZVZMKTN-ZETCQYMHSA-N

(S)-3-Cyano-5-methylhexanoic acid (CAS 181289-37-2), also designated (S)-CMHA or pregabalin nitrile, is a chiral β-cyano carboxylic acid with molecular formula C₈H₁₃NO₂ and molecular weight 155.19 g/mol. It is the penultimate intermediate in the most advanced chemoenzymatic routes to (S)-pregabalin (Lyrica®), the GABA analogue with global sales exceeding $4.16 billion in 2012.

(S)-3-Cyano-5-methylhexanoic Acid (CAS 181289-37-2): A Critical Chiral Intermediate for Pregabalin API Synthesis

(S)-3-Cyano-5-methylhexanoic acid (CAS 181289-37-2), also designated (S)-CMHA or pregabalin nitrile, is a chiral β-cyano carboxylic acid with molecular formula C₈H₁₃NO₂ and molecular weight 155.19 g/mol [1]. It is the penultimate intermediate in the most advanced chemoenzymatic routes to (S)-pregabalin (Lyrica®), the GABA analogue with global sales exceeding $4.16 billion in 2012 [2]. The compound bears a single stereocenter at the 3-position (S-configuration) and is generated with high enantiomeric purity—typically >99% ee—via nitrilase-catalyzed regio- and enantioselective hydrolysis of 2-isobutylsuccinonitrile (IBSN) [3]. Its sole scale-relevant application is as the immediate precursor to (S)-pregabalin API, to which it is converted by a single heterogeneous hydrogenation step [4].

• [1] PubChem. (S)-3-Cyano-5-methylhexanoic acid, CID 9855540. National Center for Biotechnology Information. https://pubchem.ncbi.nlm.nih.gov/compound/9855540 View Source
• [2] Zheng, R.-C., et al. (2014). Enantioselective synthesis of (S)-3-cyano-5-methylhexanoic acid by a high DMSO concentration tolerable Arthrobacter sp. ZJB-09277. Biochemical Engineering Journal, 83, 97–103. https://doi.org/10.1016/j.bej.2013.12.012 View Source
• [3] Zhang, Q., et al. (2019). Highly regio- and enantioselective synthesis of chiral intermediate for pregabalin using one-pot bienzymatic cascade of nitrilase and amidase. Applied Microbiology and Biotechnology, 103(14), 5617–5626. https://doi.org/10.1007/s00253-019-09857-1 View Source
• [4] Lin, C.-P., et al. (2024). Development of a Sustainable Chemoenzymatic Process for (S)-Pregabalin Synthesis via Nitrilase-Catalyzed Hydrolysis and Continuous Flow Racemization. Organic Process Research & Development. https://doi.org/10.1021/acs.oprd.3c00441 View Source

Why (S)-3-Cyano-5-methylhexanoic Acid Cannot Be Replaced by Racemate, (R)-Enantiomer, or Alternative Intermediates in Pregabalin Manufacturing

Generic substitution of (S)-3-cyano-5-methylhexanoic acid by its (R)-enantiomer (CAS 181289-36-1), the racemate, or unsaturated precursor 3-cyano-5-methylhex-3-enoic acid is precluded by three non-negotiable constraints. First, pharmacological activity of pregabalin resides exclusively in the (S)-enantiomer; the (R)-enantiomer is inactive at the α₂δ subunit of voltage-gated calcium channels [1], making (R)-CMHA a dead-end impurity rather than a viable intermediate. Second, the nitrilase-mediated chemoenzymatic route achieves its extraordinary enantioselectivity (E > 300) only when the enzyme discriminates the prochiral or racemic substrate to yield the (S)-configured product [2]; substituting the (R)-enantiomer or racemate would forfeit this inherent stereochemical fidelity. Third, alternative intermediates such as 3-cyano-5-methylhex-3-enoic acid require a costly rhodium DuPHOS asymmetric hydrogenation step that operates at lower substrate concentrations and with air-sensitive catalysts [3], while the first-generation classical resolution process carries an E factor of 86—five-fold higher than the chemoenzymatic route [4]. These differences are not cosmetic; they determine process mass intensity, waste disposal costs, and ultimately API cost of goods.

• [1] US Patent 9,422,230. Process for the preparation of an anticonvulsant agent pregabalin hydrochloride. Justia Patents. https://patents.justia.com/patent/9422230 View Source
• [2] Chen, Z., et al. (2021). Significantly enhancing the stereoselectivity of a regioselective nitrilase for the production of (S)-3-cyano-5-methylhexanoic acid using an MM/PBSA method. Chemical Communications, 57(7), 931–934. https://doi.org/10.1039/D0CC07106D View Source
• [3] Burk, M. J., et al. (2003). An Enantioselective Synthesis of (S)-(+)-3-Aminomethyl-5-methylhexanoic Acid via Asymmetric Hydrogenation. The Journal of Organic Chemistry, 68(14), 5731–5734. https://doi.org/10.1021/jo034397b View Source
• [4] Martinez, C. A., et al. (2008). Development of a Chemoenzymatic Manufacturing Process for Pregabalin. Organic Process Research & Development, 12(3), 392–398. https://doi.org/10.1021/op7002248 View Source

Quantitative Differentiation Evidence for (S)-3-Cyano-5-methylhexanoic Acid (CAS 181289-37-2) Against Closest Analogs and Alternative Routes

Enantiomeric Fidelity: (S)-CMHA vs. (R)-CMHA in Pharmacologically Active Pregabalin Production

The (S)-configuration of 3-cyano-5-methylhexanoic acid is the sole stereochemical form that yields pharmacologically active (S)-pregabalin upon hydrogenation. Pharmacological studies establish that only the (S)-enantiomer of pregabalin is responsible for drug activity, whereas the (R)-enantiomer is inactive at the α₂δ calcium channel target [1]. Consequently, (R)-3-cyano-5-methylhexanoic acid (CAS 181289-36-1) is classified industrially as Pregabalin Impurity 62—a reference standard and impurity marker—not as a viable synthetic intermediate . No quantity of (R)-CMHA can be converted to active API.

Evidence Dimension	Pharmacological activity of derived pregabalin enantiomer
Target Compound Data	Yields exclusively (S)-pregabalin (pharmacologically active enantiomer)
Comparator Or Baseline	(R)-3-cyano-5-methylhexanoic acid (CAS 181289-36-1): yields (R)-pregabalin (pharmacologically inactive at α₂δ calcium channels)
Quantified Difference	Complete loss of therapeutic activity—(R)-enantiomer is inactive [1]
Conditions	Established pharmacological profiling of pregabalin enantiomers at voltage-gated calcium channel α₂δ subunits

Why This Matters

Procurement of the incorrect enantiomer yields material with zero synthetic utility for API manufacturing, making enantiomeric specification the single most critical procurement parameter.

• [1] US Patent 9,422,230. Process for the preparation of an anticonvulsant agent pregabalin hydrochloride. Justia Patents. https://patents.justia.com/patent/9422230 View Source

Nitrilase Enantioselectivity: (S)-CMHA from rac-IBSN vs. Classical Resolution Approaches

Engineered nitrilase variants achieve enantiomeric ratios (E) exceeding 300 for the kinetic resolution of racemic isobutylsuccinonitrile (rac-IBSN) to (S)-CMHA, with enantiomeric excess values above 99.9% and conversion ratios of 40.9–43.8% [1]. The N258D mutant of Arabis alpina nitrilase (AaNIT) independently demonstrates E > 300 at a substrate concentration of 100 g/L, yielding (S)-CMHA at 45.0% conversion with 99.3% enantiomeric excess after bienzymatic cascade with amidase to eliminate the (S)-CMHM amide byproduct [2]. The BaNITM2 variant further achieves E = 515 with a 5.4-fold activity increase and reduction of amide byproduct from 15.8% to 1.9% relative to the parent BaNITM0 enzyme . These metrics are unattainable through classical chemical resolution, which typically yields lower ee values and requires stoichiometric chiral auxiliaries.

Evidence Dimension	Enantioselectivity (E value) of chiral intermediate production
Target Compound Data	E > 300 (N258D AaNIT); E = 515 (BaNITM2); ee > 99.9% (W57F/V134M and W57Y/V134M mutants); 45.0% conversion with 99.3% ee (bienzymatic cascade) [1][2]
Comparator Or Baseline	Classical chemical resolution using chiral auxiliaries or diastereomeric salt formation: yields typically <95% ee, requires stoichiometric resolving agents, and generates equimolar waste
Quantified Difference	E value improvement from baseline non-enzymatic approaches (<50 typical) to >300–515; ee improvement of ≥4–5 percentage points above classical resolution ceiling
Conditions	Nitrilase-catalyzed hydrolysis of rac-IBSN at 100 g/L substrate loading, pH 7.5–8.5, 25–30 °C [2]

Why This Matters

Ultra-high enantioselectivity (E > 300) translates directly to reduced purification burden, higher isolated yield of enantiopure product, and elimination of costly chiral separation steps in downstream processing.

• [1] Chen, Z., et al. (2021). Significantly enhancing the stereoselectivity of a regioselective nitrilase for the production of (S)-3-cyano-5-methylhexanoic acid using an MM/PBSA method. Chemical Communications, 57(7), 931–934. https://doi.org/10.1039/D0CC07106D View Source
• [2] Zhang, Q., et al. (2019). Highly regio- and enantioselective synthesis of chiral intermediate for pregabalin using one-pot bienzymatic cascade of nitrilase and amidase. Applied Microbiology and Biotechnology, 103(14), 5617–5626. https://doi.org/10.1007/s00253-019-09857-1 View Source

Process Mass Efficiency: Chemoenzymatic (S)-CMHA Route vs. First-Generation Pregabalin Manufacturing

The chemoenzymatic process employing (S)-CMHA as the key intermediate achieves a 5-fold reduction in E factor (kg waste per kg product) compared to the first-generation pregabalin manufacturing process. Specifically, the E factor decreased from 86 (first-generation, classical resolution route) to 17 (chemoenzymatic route) [1]. The overall yield after one recycle of the undesired (R)-enantiomer reaches 40–45% in the second-generation chemoenzymatic process, with substantial reductions in organic solvent usage enabling a predominantly aqueous process [1]. In the nitrilase-mediated route from IBSN, the E factor is further demonstrated to be significantly lower than both first- and second-generation routes, with a two-step process flow replacing the four-step second-generation sequence [2].

Evidence Dimension	Environmental process mass intensity (E factor)
Target Compound Data	E factor = 17 (chemoenzymatic route via (S)-CMHA) [1]; process flow: 2 steps (IBSN → (S)-CMHA → pregabalin) [2]
Comparator Or Baseline	First-generation process: E factor = 86, classical resolution of rac-3-aminomethyl-5-methylhexanoic acid; Second-generation process: 4 synthetic steps [1][2]
Quantified Difference	E factor reduction of 69 units (86 → 17); 5-fold improvement; process step reduction from 4 to 2 (50% reduction) [1][2]
Conditions	Full manufacturing-scale process comparison as reported by Pfizer process development team [1]

Why This Matters

A 5-fold E factor reduction directly impacts manufacturing cost structure through lower solvent procurement, reduced waste treatment expenditure, and improved regulatory sustainability profile—critical factors in large-volume generic API procurement.

• [1] Martinez, C. A., et al. (2008). Development of a Chemoenzymatic Manufacturing Process for Pregabalin. Organic Process Research & Development, 12(3), 392–398. https://doi.org/10.1021/op7002248 View Source
• [2] Zhang, Q., et al. (2019). Efficient Chemoenzymatic Synthesis of Optically Active Pregabalin from Racemic Isobutylsuccinonitrile. Organic Process Research & Development, 23(9), 2042–2049. https://doi.org/10.1021/acs.oprd.9b00273 View Source

Industrial Substrate Loading and Racemization-Enabled Cycle Efficiency: (S)-CMHA vs. Static Resolution Processes

The nitrilase-catalyzed production of (S)-CMHA operates at high substrate loadings of 100–150 g/L (700 mM IBSN) with sustained enantioselectivity. In the state-of-the-art continuous process, IBSN (700 mM) is converted to (S)-CMHA with 45.3% conversion and 99.5% ee, while unreacted (R)-IBSN is continuously racemized using γ-alumina-based solid base catalysts (ASBCs) in a packed bed reactor, achieving a space-time yield of 6.76 mol L⁻¹ day⁻¹ [1]. The ASBCs retained 85.9% catalytic activity after 20 days of continuous operation [1]. In batch mode, immobilized E. coli cells harboring nitrilase BrNIT achieved E > 150 and >41.1% conversion at 100 g/L substrate loading over 12 reaction batches, with (R)-IBSN recycled by racemization at 94.5% yield [2]. This continuous racemization-recycle architecture is unavailable in classical stoichiometric resolution processes.

Evidence Dimension	Substrate loading, conversion, and racemization-enabled recycle efficiency
Target Compound Data	700 mM (≈108 g/L) IBSN → 45.3% conversion, 99.5% ee (S)-CMHA; space-time yield 6.76 mol L⁻¹ day⁻¹; racemization catalyst retains 85.9% activity after 20 days [1]; 100 g/L → >41.1% conversion, (R)-IBSN racemization yield 94.5%, 12 batch reuse [2]
Comparator Or Baseline	Classical resolution: typically ≤50 g/L effective substrate loading; stoichiometric resolving agent required; no continuous racemization of undesired enantiomer possible without additional chemical steps
Quantified Difference	Space-time yield of 6.76 mol L⁻¹ day⁻¹ in continuous mode; racemization recycling efficiency of 94.5% enables theoretical yield approaching 90% with full recycle, vs. 50% maximum for static resolution [1][2]
Conditions	Continuous flow packed bed reactor with immobilized nitrilase on epoxy resin; γ-alumina solid base catalyst for racemization; membrane separation for (S)-CMHA purification [1]

Why This Matters

High substrate loading combined with continuous racemization-recycle transforms a theoretical 50% yield ceiling (static kinetic resolution) into a >90% theoretical yield process, dramatically reducing raw material cost per kg of (S)-CMHA.

• [1] Lin, C.-P., et al. (2024). Development of a Sustainable Chemoenzymatic Process for (S)-Pregabalin Synthesis via Nitrilase-Catalyzed Hydrolysis and Continuous Flow Racemization. Organic Process Research & Development. https://doi.org/10.1021/acs.oprd.3c00441 View Source
• [2] Zhang, Q., et al. (2019). Efficient Chemoenzymatic Synthesis of Optically Active Pregabalin from Racemic Isobutylsuccinonitrile. Organic Process Research & Development, 23(9), 2042–2049. https://doi.org/10.1021/acs.oprd.9b00273 View Source

Single-Step Hydrogenation to Pregabalin API: (S)-CMHA vs. Multi-Step Intermediate Routes

(S)-CMHA is converted to (S)-pregabalin API in a single heterogeneous hydrogenation step, preserving stereochemical integrity. In the 2024 sustainable process, (S)-CMHA purified by membrane separation (95% yield) was hydrogenated to (S)-pregabalin in 88% yield [1]. In the BrNIT-based process, the resultant (S)-CMHA was hydrogenated directly to pregabalin with 99.6% chemical purity and 99.4% optical purity [2]. By contrast, the 3-cyano-5-methylhex-3-enoic acid route requires two sequential hydrogenation steps—first asymmetric hydrogenation over expensive Rh-MeDuPHOS catalyst (60 psi, 45 °C) to set the stereocenter, followed by heterogeneous nickel-catalyzed nitrile reduction—with reported pregabalin purity of only 97.82% (S-isomer) [3].

Evidence Dimension	Downstream hydrogenation efficiency and final API purity
Target Compound Data	(S)-CMHA → single-step hydrogenation → 88% yield, 99.6% purity, 99.4% optical purity [1][2]
Comparator Or Baseline	3-cyano-5-methylhex-3-enoic acid → two-step hydrogenation (Rh-MeDuPHOS then Ni) → 97.82% S-isomer [3]
Quantified Difference	Optical purity advantage of 1.58 percentage points (99.4% vs. 97.82%); one fewer hydrogenation step; elimination of expensive chiral rhodium catalyst
Conditions	Heterogeneous nickel or sponge nickel catalyst, H₂ atmosphere; (S)-CMHA free acid or salt form [1][2][3]

Why This Matters

Single-step nitrile reduction to the primary amine avoids a second hydrogenation reactor, eliminates the cost and air-sensitivity burden of chiral rhodium catalysts, and delivers API-grade optical purity directly—reducing both capital expenditure and per-batch processing time.

• [1] Lin, C.-P., et al. (2024). Development of a Sustainable Chemoenzymatic Process for (S)-Pregabalin Synthesis via Nitrilase-Catalyzed Hydrolysis and Continuous Flow Racemization. Organic Process Research & Development. https://doi.org/10.1021/acs.oprd.3c00441 View Source
• [2] Zhang, Q., et al. (2019). Efficient Chemoenzymatic Synthesis of Optically Active Pregabalin from Racemic Isobutylsuccinonitrile. Organic Process Research & Development, 23(9), 2042–2049. https://doi.org/10.1021/acs.oprd.9b00273 View Source
• [3] Burk, M. J., et al. (2003). An Enantioselective Synthesis of (S)-(+)-3-Aminomethyl-5-methylhexanoic Acid via Asymmetric Hydrogenation. The Journal of Organic Chemistry, 68(14), 5731–5734. https://doi.org/10.1021/jo034397b View Source

Biocatalytic Stability Under Industrial Conditions: (S)-CMHA-Producing Nitrilases vs. Solvent-Sensitive Esterases

Multiple biocatalytic routes to (S)-CMHA exhibit tolerance to industrially relevant co-solvent concentrations that would inactivate competing enzyme systems. Arthrobacter sp. ZJB-09277 whole-cell esterase retains full activity in 50% (v/v) DMSO and tolerates DMSO concentrations up to 80% (v/v), enabling a substrate loading of 100 mM to yield (S)-CMHA at 44.6 mM with 95.1% ee—the enantiomeric ratio (E) increasing from 33 to 80 upon DMSO addition [1]. In the nitrilase route, cross-linked enzyme aggregates (CLEAs) of Thermomyces lanuginosus lipase remain stable in diverse organic solvents and at temperatures up to 50 °C, achieving ethyl (S)-3-cyano-5-methylhexanoate in 90–92% theoretical yield and >99% ee, with enzyme reuse demonstrated for at least 11 cycles [2]. The immobilized nitrilase on epoxy resin in the continuous flow process further enables 12-batch reuse without loss of enantioselectivity at 100 g/L substrate loading [3].

Evidence Dimension	Biocatalyst tolerance to organic co-solvents and operational stability under high substrate loading
Target Compound Data	Arthrobacter sp. esterase: 80% (v/v) DMSO tolerance, E = 80 in 50% DMSO [1]; CLEA-lipase: stable to 50 °C, 11 recycles, >99% ee [2]; immobilized BrNIT: 12 batches at 100 g/L, E > 150 [3]
Comparator Or Baseline	Conventional lipases/esterases: typically inactivated at >10–20% (v/v) organic co-solvent; most commercial immobilized enzymes: 5–8 recycles before significant activity loss
Quantified Difference	DMSO tolerance 4–8× higher than typical esterases; enzyme reuse extended to 11–12 cycles without ee degradation; DMSO addition increases E value 2.4-fold (33 → 80) [1]
Conditions	Whole-cell Arthrobacter sp. ZJB-09277 in phosphate buffer with 50% (v/v) DMSO, 30 °C [1]; CLEAs in calcium acetate buffer pH 7.5, 30 °C, 50% substrate loading [2]

Why This Matters

Extraordinary co-solvent tolerance and operational stability enable process intensification at industrial scale, reduce enzyme replacement frequency, and allow higher substrate concentrations that improve overall volumetric productivity.

• [1] Zheng, R.-C., et al. (2014). Enantioselective synthesis of (S)-3-cyano-5-methylhexanoic acid by a high DMSO concentration tolerable Arthrobacter sp. ZJB-09277. Biochemical Engineering Journal, 83, 97–103. https://doi.org/10.1016/j.bej.2013.12.012 View Source
• [2] Basetty, S., & Kumaraguru, T. (2022). Preparation of enantiopure pregabalin intermediate using cross linked enzyme aggregates (CLEAs) in basket reactor. Taylor & Francis Figshare. https://doi.org/10.6084/m9.figshare.19096506 View Source
• [3] Zhang, Q., et al. (2019). Efficient Chemoenzymatic Synthesis of Optically Active Pregabalin from Racemic Isobutylsuccinonitrile. Organic Process Research & Development, 23(9), 2042–2049. https://doi.org/10.1021/acs.oprd.9b00273 View Source

Optimal Deployment Scenarios for (S)-3-Cyano-5-methylhexanoic Acid in Pregabalin API Manufacturing and Analytical Science

Large-Scale Generic Pregabalin API Manufacturing via Nitrilase Chemoenzymatic Route

(S)-3-Cyano-5-methylhexanoic acid is the indispensable penultimate intermediate for generic pregabalin manufacturers seeking to compete on process mass efficiency. The chemoenzymatic route from rac-IBSN via immobilized nitrilase achieves an E factor of 17—a 5-fold improvement over the first-generation classical resolution process (E factor = 86)—with a two-step process flow (IBSN hydrolysis → hydrogenation) replacing the legacy four-step sequence [1][2]. Continuous racemization of the unreacted (R)-IBSN using γ-alumina solid base catalysts enables space-time yields of 6.76 mol L⁻¹ day⁻¹ and an effective theoretical yield approaching 90% [3]. This scenario is optimal for high-volume generic API producers for whom waste disposal costs and raw material utilization directly determine competitive pricing.

Enzymatic Process Development and Nitrilase Engineering for Chiral Intermediate Production

The (S)-CMHA production system serves as an exceptional model platform for nitrilase engineering and biocatalytic process development. The availability of multiple characterized nitrilase variants—AaNIT N258D (E > 300, 100 g/L substrate) [1], BaNITM2 (E = 515, 5.4-fold activity enhancement, amide byproduct reduced from 15.8% to 1.9%) [2], and W57F/V134M/W57Y/V134M double mutants (E > 300, ee > 99.9%, 40.9–43.8% conversion) [3]—provides a rich dataset for structure-activity relationship studies. Researchers developing nitrilases for other prochiral dinitrile substrates can benchmark their engineered variants against these established performance metrics.

Enantiopure Reference Standard for Pregabalin Impurity Profiling and Regulatory Submission

High-enantiopurity (S)-CMHA (>99% ee) is essential as both a process intermediate and as the reference standard for detecting and quantifying the (R)-CMHA impurity (CAS 181289-36-1, Pregabalin Impurity 62) in API batches [1]. The single-step hydrogenation of (S)-CMHA to pregabalin with 99.6% chemical purity and 99.4% optical purity [2] means that the enantiomeric purity of the (S)-CMHA intermediate directly determines the optical purity of the final API. Analytical laboratories supporting ANDA filings require (S)-CMHA of verified enantiomeric excess as the chiral reference marker for HPLC and SFC method validation.

Continuous Flow Manufacturing Process Intensification for (S)-Pregabalin

The 2024 integrated continuous flow process combining immobilized nitrilase-catalyzed IBSN hydrolysis with continuous racemization in a packed bed reactor represents the most advanced manufacturing paradigm for (S)-pregabalin [1]. In this configuration, (S)-CMHA is generated at 45.3% conversion with 99.5% ee from 700 mM IBSN, purified via membrane separation in 95% yield, and hydrogenated to pregabalin in 88% yield [1]. The solid base racemization catalyst retains 85.9% activity after 20 days of continuous operation [1]. This scenario is ideally suited for manufacturers implementing ICH Q13 continuous manufacturing guidelines for pregabalin API, offering reduced equipment footprint, lower solvent inventory, and real-time quality assurance capability.

• [1] Martinez, C. A., et al. (2008). Development of a Chemoenzymatic Manufacturing Process for Pregabalin. Organic Process Research & Development, 12(3), 392–398. https://doi.org/10.1021/op7002248 View Source
• [2] Zhang, Q., et al. (2019). Efficient Chemoenzymatic Synthesis of Optically Active Pregabalin from Racemic Isobutylsuccinonitrile. Organic Process Research & Development, 23(9), 2042–2049. https://doi.org/10.1021/acs.oprd.9b00273 View Source
• [3] Lin, C.-P., et al. (2024). Development of a Sustainable Chemoenzymatic Process for (S)-Pregabalin Synthesis via Nitrilase-Catalyzed Hydrolysis and Continuous Flow Racemization. Organic Process Research & Development. https://doi.org/10.1021/acs.oprd.3c00441 View Source