UHPLC Method for Sensitive Automatic Analysis of Thirty-Seven D/L-Amino Acids and for Liquor Profiling

Significance of the topic

Determination of D- and L-amino acid stereoisomers at trace levels is increasingly relevant in food science, nutrition and biomarker research. D-amino acids occur at low concentrations in fermented foods and biological samples, yet they influence taste, preservation and potential disease indicators. Reliable, sensitive and high-throughput methods for simultaneous D/L profiling of many amino acids are therefore needed for quality control, product development and exploratory biomarker discovery.

Objectives and study overview

The work presents an improved UHPLC-based method for sensitive, fully automated simultaneous analysis of thirty-seven D/L-amino acids (excluding proline enantiomers) using a single chiral derivatizing reagent. The method aims to shorten analysis time relative to prior dual-reagent approaches while preserving high sensitivity, reproducibility and practical applicability to complex food and beverage matrices (examples: beer, sake, red and white wine). Multivariate statistical profiling of samples using PCA is demonstrated to illustrate sample discrimination by stereoisomer composition.

Methodology

The assay uses pre-column fluorescence derivatization of amino acids to form diastereomers via reaction of o-phthalaldehyde (OPA) with amino acids in the presence of the chiral thiol reagent N-isobutyryl-L-cysteine (NIBC). Key operational features include:

• Automatic pre-column derivatization performed inside the autosampler needle to avoid air exposure and to keep the derivatization-to-injection time constant (critical because OPA products degrade rapidly).
• Automated mobile-phase blending to prepare the organic composition on-line, reducing manual preparation time and consumable use.
• Optimized chromatographic gradient to resolve 37 D/L pairs in a single run with good sensitivity.

Used Instrumentation

The main hardware and software elements used in the study were:

• UHPLC system: Nexera X3 (Shimadzu) to tolerate high backpressure from sub-2 µm stationary phase.
• Analytical column: CERI L-column 3 C18, 150 mm × 2.1 mm I.D., 2.0 µm (with pre-column filter).
• Autosampler with automatic pre-column derivatization and programmable pretreatment routine.
• Pump with low-pressure gradient kit and mobile-phase blending function (separate solvent ports for acetonitrile and methanol blended 15:85).
• Fluorescence detector: RF-20AXS, excitation 338 nm, emission 455 nm (semimicro cell).
• Statistical software: eMSTAT Solution for multivariate analysis (PCA).

Analytical conditions and workflow (summary)

Representative conditions: 0.22 mL/min flow, column temperature 20 °C, 1 µL injection. Mobile phase A: 10 mmol/L sodium phosphate buffer pH 6.9. Mobile phase B: blend of acetonitrile/methanol (15:85) prepared automatically. Gradient program spans about 87 minutes total, with the key region delivering separation of the 37 analytes in approximately 66 minutes. Derivatizing reagents prepared from OPA and NIBC in borate buffer (pH 9.1) and mixed immediately prior to automated reaction.

Key results and discussion

Method performance:

• Separation: Thirty-seven D/L-amino acids were baseline-resolved within ~66 minutes under the optimized gradient.
• Linearity: Calibration curves for all analytes showed excellent linearity with r2 ≥ 0.999 across the reported concentration ranges (varied by analyte; D-isomers typically quantified at lower ranges reflecting trace-level presence).
• Repeatability: Peak-area RSDs were ≤1.6% (for 0.1 µmol/L standards) and ≤0.8% (for 5 µmol/L standards) when the derivatization-to-injection interval was held constant by automation.
• Sensitivity: Optimized automated pre-column derivatization enabled detection more sensitive than prior conventional methods using the same fluorescence chemistry (quantitative LOD/LOQ not explicitly listed in the source text).

Application to real samples:

• Liquor matrix analysis: Beer (two brands), sake, red wine and white wine were analyzed after acid dilution and filtration. Between 25 and 28 amino acids were detected per sample.
• Enantiomer distribution: Overall D/(D+L) percentage was ≤6% in all tested liquor samples, with many D-isomers present at low µmol/L concentrations. Some L-amino acids in concentrated samples exceeded the calibration range and required dilution.
• PCA-based profiling: Principal component analysis (eMSTAT) discriminated sample types: beer samples clustered together, wines clustered together, and sake formed a separate group. Loading plots identified marker stereoisomers associated with each cluster (e.g., L-Trp, D-His, D-Lys influential for beers; D-Val for wines; L-Thr for sake), indicating stereoisomer patterns can serve for liquor classification and compositional insight.

Benefits and practical applications

The presented method offers several practical advantages:

• Single-reagent derivatization reduces complexity compared with dual-reagent or multidimensional LC approaches.
• Automation of mobile-phase blending and pre-column derivatization minimizes hands-on time and operator variability; example workflow comparison for 20 samples showed total active preparation time reduced from ~70 min (manual) to ~15 min (automated), saving about one hour.
• High repeatability is achieved by fixing the derivatization-to-injection timing, addressing the short lifetime of OPA-derived products.
• Fluorescence detection affords robust quantitation in complex matrices with less susceptibility to matrix effects than LC–MS, while avoiding the cost and complexity of multi-dimensional LC systems.
• Applicable to food and beverage profiling, QC, fermentation monitoring, nutritional studies and exploratory biomarker screens where D-amino acid content is informative.

Future trends and potential applications

Potential developments and extensions include:

• Integration with high-resolution MS for structural confirmation and expanded analyte panels while retaining chromatographic chiral separation for quantitation.
• Further reduction of run times via column and gradient optimization or use of shorter columns and higher pressures while maintaining resolution of enantiomeric pairs.
• Broader automation: in-line sample cleanup, dilution and filtration to increase throughput and decrease manual handling for routine QC labs.
• Application expansion to clinical or microbiome-related biomarkers where D-amino acids may reflect disease states or gut microbial activity; adaptation would require validation in biological fluids and compliant workflows.
• Development of alternative chiral derivatizing reagents or stationary phases to improve separation for difficult enantiomeric pairs or to include additional non-proteinogenic amino acids.

Conclusion

This UHPLC–fluorescence approach using OPA/NIBC pre-column derivatization and automation enables sensitive, precise and relatively fast simultaneous analysis of thirty-seven D/L-amino acids without resorting to LC–MS or multidimensional LC setups. The method demonstrates robust analytical figures (excellent linearity and low RSDs), practical time savings through automation, and effective application to complex beverage matrices with meaningful multivariate discrimination. The workflow is well suited for food and beverage profiling, routine QC and exploratory research into D-amino acid roles in health and food science.

References

1. Iwata N., Watabe Y., Horie S., Hayakawa Y., Chromatography, 42, 133–141 (2021).
2. Iwata N., Kobayashi M., Chromatography, 45, 63–72 (2024).