Proanthocyanidin Structure-Activity Relationship Analysis by Path Analysis Model

Zhaoxuan Li, Jingling Liu, Jie You, Xin Li, Zongsuo Liang, Junli DuInternational Journal of Molecular Sciences 2023 · DOI

TL;DR

This paper introduces a novel "path analysis model" to dissect the complex interplay between individual proanthocyanidin (PA) monomer units and their biological activities (antioxidant and anti-diabetes). By separating direct and indirect effects, the model offers a more nuanced understanding of how these natural compounds work, paving the way for better utilization of PA-rich resources.

What Problem Does This Paper Address?

Proanthocyanidins (PAs) are a class of natural compounds found in plants, known for their diverse health benefits like antioxidant and anti-diabetic properties. However, PAs are structurally complex, being polymers of various flavan-3-ol monomer units linked together. This complexity makes it challenging to pinpoint which specific monomer units contribute to which activity and how they do so, especially considering the potential interactions between these units. Existing methods like QSAR or response surface methodology often struggle to fully capture these intricate interactions within mixtures of compounds. This paper aims to address this gap by providing a method to quantitatively analyze the structure-activity relationship (SAR) of PAs, specifically accounting for the influence of individual monomer units and their interrelationships.

Methods

The core of this paper is the development and application of a path analysis model. This statistical approach, rooted in Wright's original work, goes beyond simple correlation by dissecting the total correlation between a monomer unit and an activity into two components:

  • Direct Effect: The direct influence of a specific monomer unit on the biological activity.
  • Indirect Effect: The influence of a monomer unit on the activity mediated through its correlations with other monomer units.

This subdivision allows researchers to understand not only if a monomer unit is important but also how it exerts its influence – whether directly or by modulating other components in the PA structure. The model uses standardized multiple linear regression and canonical equations to estimate path coefficients, which represent these direct and indirect effects. The authors then use these coefficients to rank the importance of monomer units and determine their primary mode of action (direct vs. indirect).

Datasets

The study utilized data from Rhodiola crenulata PAs. The specific monomer units analyzed were: Epicatechin (ECt), Catechin gallate (CGt), Epicatechin gallate (ECGt), Gallocatechin (GCe), Epigallocatechin (EGCe), Gallocatechin gallate (GCGe), Epigallocatechin gallate (EGCGe), Catechin (Ce), Epicatechin (ECe), and Epicatechin gallate (ECGe).

The biological activities evaluated were:

  • Antioxidant activity: Measured using Ferric reducing antioxidant activity (FRAP) and cupric ion reducing antioxidant capacity (CUPRAC).
  • Anti-diabetes activity: Assessed by the inhibition of Saccharomyces cerevisiae α-glucosidase, porcine pancreatic α-amylase, and human salivary α-amylase.

The paper mentions that raw data (Table S1) and processed data (Table S2) are available in supplementary materials, along with a preprocessing method (File S1). The degree of polymerization was not considered in the model analysis due to data size requirements.

Key Insights and Ideas

  • Beyond Simple Correlation: The path analysis model reveals that the total correlation between a monomer and an activity is only part of the story. Understanding the breakdown into direct and indirect effects is crucial for deciphering complex biological mechanisms.
  • Monomer Unit Specificity: Different monomer units have distinct roles and influence modes. For instance, some primarily act directly, while others exert their effects indirectly through interactions with other units.
  • Positional Effects Matter: The same monomer unit can have different impacts depending on its position within the PA structure (e.g., as a terminal unit vs. an extension unit). This highlights the importance of stereochemistry and linkage in determining biological activity.
  • Interactions are Key: The model explicitly accounts for interactions between monomer units, which is a significant advantage over methods that assume additivity. This is particularly relevant for complex natural product mixtures.
  • Gallate's Role: The presence of gallate groups appears to influence the activity and influence mode of monomer units, suggesting a specific role for these functional groups.

Key Results

The path analysis model was applied to Rhodiola crenulata PAs to analyze their antioxidant and anti-diabetes activities.

For Antioxidant Activity: * The model explained a remarkable 99.34% of the variation in antioxidant activity. * Monomer units ECGt and ECGe showed the strongest correlation with antioxidant activity. * ECGe had a large positive direct effect (1.1117), indicating it primarily drives antioxidant activity directly. * ECGt, despite a strong negative total correlation (-0.9822), had a very small direct effect (-0.0274), suggesting its activity is mainly driven by indirect regulation from other units.

For Anti-Diabetes Activity: * The model explained 99.49% of the variation in anti-diabetes activity. * Ce and ECt showed the strongest positive correlation with anti-diabetes activity. * Ce primarily regulates anti-diabetes activity directly (direct effect 0.6063). * ECt, conversely, mainly regulates indirectly (total indirect effect 0.5803). * CGt and ECGt showed strong negative correlations (-0.9004 and -0.9109) and primarily acted directly.

Comparison: * Monomer units like ECGt, ECGe, CGt, ECt, and Ce were identified as key players in both activities. * These key monomer units generally had a greater influence on antioxidant activity than on anti-diabetes activity. * The influence modes (direct vs. indirect) differed significantly between the two activities for the same monomer units.

Limitations and Future Work

  • Complexity of PAs: The authors acknowledge that while their model captures monomer unit interactions, it doesn't yet incorporate other crucial factors like linkage types and the degree of polymerization. Integrating these would further enhance the understanding of PA SAR.
  • Data Size: The degree of polymerization was excluded from the model analysis due to data size requirements, suggesting that future work could explore models that can handle such complexities.
  • Technological Advances: The authors suggest that technological advancements will be key to improving confidence in the results and further exploring PA SAR.
  • Broader Applications: The paper positions the path analysis model as a valuable data mining tool for PA SAR research, implying its potential for application to other PA-rich sources and biological activities.

Summary updated

Abstract

To fully explore the influence mechanism of interactions between different monomer units of proanthocyanidins (PAs) on biological activity, a path analysis model of the PA structure-activity relationship was proposed. This model subdivides the total correlation between each monomer unit and activity into direct and indirect effects by taking into account not only each monomer unit but also the correlation with its related monomer units. In addition, this method can determine the action mode of each monomer unit affecting the activity by comparing the direct and total indirect effects. Finally, the advantage of this model is demonstrated through an influence mechanism analysis of <i>Rhodiola crenulata</i> PA monomer units on antioxidant and anti-diabetes activities.
BibTeX
@article{li2023proanthocyanidin,
  author = {Zhaoxuan Li and Jingling Liu and Jie You and Xin Li and Zongsuo Liang and Junli Du},
  title = {{Proanthocyanidin Structure-Activity Relationship Analysis by Path Analysis Model}},
  journal = {International Journal of Molecular Sciences},
  year = {2023},
  doi = {10.3390/ijms24076379},
}
Figure