Tindal Rachael A, Jeffery David W, Muhlack Richard A
Australian Research Council Training Centre for Innovative Wine Production and Waite Research Institute, The University of Adelaide, PMB 1, Glen Osmond, South Australia 5064, Australia.
J Chem Inf Model. 2025 Mar 24;65(6):2834-2844. doi: 10.1021/acs.jcim.4c02300. Epub 2025 Mar 3.
Anthocyanins are pigmented polyphenolic compounds that influence the color, stability, and quality of numerous plant organs (e.g., fruit, flowers) and their derived products (e.g., natural dyes, red wine). Existing within a complex multistate system, anthocyanins can be simultaneously present as pH-dependent red, purple, or blue species that are either in a monomeric (chemically unstable) or in a self-associated (temporally stable) form. However, limitations of current analytical techniques (e.g., HPLC with a UV-vis detector) may cause experimental data to omit or misrepresent important color and stability characteristics afforded by all anthocyanin species within the natural matrix. In response, a computational spectral deconvolution method is demonstrated that increases the fidelity of spectral data collected for anthocyanins, thereby representing all existing monomeric and self-associated anthocyanin species within solution for the first time. Case studies for the developed deconvolution model are presented, based on experimental data obtained via HPLC-DAD analysis for malvidin-3--β-d-glucopyranoside (M3G) in red wines that were sampled throughout fermentation. Fourier spectral deconvolution methods were used to transform experimental spectra for pigmented anthocyanin monomers into systems that represent spectral behaviors of all pigmented monomeric and self-associated anthocyanin species in solution. The developed computational model was found to significantly increase the level of signal feature extraction for the spectral data of anthocyanins, providing key information on color expression and stability characteristics that would be otherwise unattainable with traditional approaches using HPLC with UV-vis detection. The current work increases understanding and control over key attributes of anthocyanins and has broad potential applications for the analysis and commercialization of anthocyanin-containing products.
花青素是有色素的多酚类化合物，会影响许多植物器官（如果实、花朵）及其衍生产品（如天然染料、红酒）的颜色、稳定性和品质。花青素存在于复杂的多状态系统中，可同时以pH依赖的红色、紫色或蓝色形式存在，这些形式既可以是单体形式（化学不稳定），也可以是自缔合形式（暂时稳定）。然而，当前分析技术（如配备紫外可见检测器的高效液相色谱法）的局限性可能导致实验数据遗漏或错误呈现天然基质中所有花青素种类所具有的重要颜色和稳定性特征。作为回应，本文展示了一种计算光谱去卷积方法，该方法提高了为花青素收集的光谱数据的保真度，从而首次呈现了溶液中所有现有的单体和自缔合花青素种类。基于通过高效液相色谱 - 二极管阵列检测（HPLC - DAD）分析在整个发酵过程中采样的红酒中的矢车菊素 - 3 - β - D - 葡萄糖苷（M3G）所获得的实验数据，给出了所开发去卷积模型的案例研究。傅里叶光谱去卷积方法用于将有色素的花青素单体的实验光谱转换为代表溶液中所有有色素的单体和自缔合花青素种类光谱行为的系统。结果发现，所开发的计算模型显著提高了花青素光谱数据的信号特征提取水平，提供了关于颜色表达和稳定性特征的关键信息，而这些信息使用配备紫外可见检测的传统高效液相色谱法是无法获得的。当前的工作增进了对花青素关键属性的理解和控制，并且在含花青素产品的分析和商业化方面具有广泛的潜在应用。
