Distinct Structuring Mechanisms of Extrinsic and Intrinsic Factors on the Wild Marine Fish Gut Microbiome

Technical Summary: Distinct Structuring Mechanisms of Extrinsic and Intrinsic Factors on the Wild Marine Fish Gut Microbiome

Problem Statement
The fish gut microbiome (GM) is critical for nutrient absorption, immune homeostasis, and host physiology. While it is established that both extrinsic factors (e.g., season, diet, water temperature) and intrinsic factors (e.g., host species, genetics) shape the GM, the distinct mechanisms by which these factors drive GM structure remain unclear. Previous studies have largely relied on single metrics (such as taxonomic composition or alpha/beta diversity) to demonstrate that a factor influences the GM. However, these approaches fail to distinguish whether factors act similarly or differently—specifically, whether they primarily reorganize microbial abundance without phylogenetic similarity, or shape phylogenetic similarity without distinct shifts in composition. This study addresses the gap in understanding how extrinsic and intrinsic factors differentially structure the fish GM through distinct processes.

Methodology
The researchers characterized the GM of 21 wild marine fish species collected from the South Sea of Korea across two seasons (Summer 2021 and Winter 2022).

• Sample Collection: 118 gut samples were obtained from 21 species. Fish were anesthetized, and hindgut contents were excised and stored at -80°C.
• Sequencing: 16S rRNA V3-V4 amplicon sequencing was performed on an Illumina MiSeq platform.
• Data Processing: Sequences were processed using QIIME2 (v2024.05) with DADA2 for denoising and chimera removal. Taxonomic classification was performed against the SILVA database (v138.1).
• Statistical Analysis:
  • Diversity Metrics: Alpha diversity (Shannon, Simpson, Faith's PD) and beta diversity were calculated.
  • Dissimilarity Metrics: Three metrics were employed to dissect different aspects of community structure: Bray-Curtis (BC) for taxon abundance, Unweighted UniFrac for phylogenetic presence/absence, and Weighted UniFrac for phylogenetic abundance.
  • Drivers: The study analyzed the effects of Season (extrinsic), Trophic Level (extrinsic proxy for diet), and Fish Species (intrinsic).
  • Functional Prediction: PICRUSt2 was used to predict functional pathways (KEGG orthologs). A Random Forest classifier with 5-fold cross-validation (repeated 10 times) evaluated the predictive power of these factors on functional profiles, using the Area Under the Precision-Recall Curve (AUPRC) to handle class imbalance.
  • Core ASVs: Fish species-core amplicon sequence variants (ASVs) were identified to assess host-specific filtering.

Key Results

1. Extrinsic Factors Drive Abundance Shifts (Bray-Curtis):

   • Season: Season significantly influenced GM composition based on taxon abundance (BC dissimilarity). Catenococcus and Enterovibrio were indicator taxa for summer, while Aliivibrio and Shewanella characterized winter. Notably, while genera like Vibrio and Photobacterium were present in both seasons, specific ASVs within these genera showed distinct seasonal enrichment.
   • Trophic Level: Trophic level also significantly influenced GM abundance. Photobacterium was an indicator for high trophic levels (>3.4), while Vibrio was associated with lower trophic levels (<3.4). A Mantel test confirmed a significant correlation between trophic level (as a continuous variable) and BC dissimilarity.
   • Diversity: Neither season nor trophic level significantly altered alpha diversity (Shannon/Simpson), indicating that these factors reorganize existing communities rather than changing overall diversity richness.

2. Intrinsic Factors Drive Phylogenetic Similarity (Unweighted UniFrac):

   • Fish Species: While distinct species-specific clustering was not apparent in abundance-based ordinations (PCoA), fish species identity showed significant specificity when analyzed via Unweighted UniFrac (phylogenetic distance).
   • Core ASVs: Analysis of "fish species-core ASVs" (ASVs consistently detected within a host species) revealed phylogenetic clustering within host species. For example, core ASVs of certain species clustered within Vibrio/Catenococcus clades, while others clustered in Enterovibrio.
   • Mechanism: This suggests that host-specific filtering (intrinsic factors) selects for phylogenetically similar microbes regardless of their relative abundance, a pattern that remained conserved even when seasonal changes altered which specific ASVs were detected as "core."

3. Functional Stratification by Season:

   • Predictive Power: The Random Forest classifier demonstrated that Season was the strongest predictor of functional pathway profiles (AUPRC = 0.8571), whereas trophic category was a weaker predictor (AUPRC = 0.5791). Fish species could not be evaluated due to sample size constraints.
   • Functional Shifts:
     • Summer: Enriched in pathways related to carbohydrate digestion/absorption, styrene degradation, and polycyclic aromatic hydrocarbon (PAH) degradation. Antibacterial pathways (phosphonate metabolism, siderophore biosynthesis) were also enriched, coinciding with higher pathogen proliferation in summer.
     • Winter: Enriched in protein digestion/absorption, atrazine degradation (potentially linked to agricultural runoff), and steroid biosynthesis.

Key Contributions

• Differentiation of Mechanisms: The study provides evidence that extrinsic factors (season, trophic level) and intrinsic factors (host species) structure the GM through fundamentally different mechanisms. Extrinsic factors primarily drive shifts in microbial abundance (captured by Bray-Curtis), while intrinsic factors drive phylogenetic similarity and host-specific filtering (captured by Unweighted UniFrac and core ASV analysis).
• Resolution of Taxonomic Signals: The study highlights that seasonal and trophic signals for certain genera (e.g., Vibrio, Photobacterium) are only resolvable at the ASV level, not the genus level, emphasizing the need for high-resolution sequencing.
• Functional Dynamics: It demonstrates that seasonal environmental changes are the primary driver of predicted functional shifts in the fish GM, particularly regarding digestion (carbohydrate vs. protein) and defense mechanisms (antibacterial pathways).

Significance and Claims
The authors claim that their findings reveal that extrinsic and intrinsic factors differentially structure fish GM through distinct processes. Specifically, they conclude that:

1. Host-specific filtering drives the phylogenetic structure of the GM within fish species, maintaining a conserved evolutionary relationship even when seasonal fluctuations alter the specific microbial composition.
2. Seasonal changes are the dominant force shaping the predicted functional capacity of the GM, likely driven by environmental changes in prey availability (seaweed vs. animal prey) and pathogen pressure.
3. Understanding these distinct mechanisms is essential for predicting fish GM responses to environmental changes and for designing effective probiotics that account for the primary drivers of GM variation.

The paper maintains a modest tone regarding its limitations, noting that host traits were not directly measured and functional pathways were inferred rather than experimentally validated. Consequently, the authors state that further studies incorporating host physiological measurements and direct functional assays are required to confirm these mechanisms.