Questioning the Evidence for Host-Symbiont Codiversification in Mycorrhizal Symbioses

By reanalyzing 29 mycorrhizal networks, this study reveals that while closely related plants and fungi interact (cophylogenetic signal), their evolutionary histories do not match (no phylogenetic congruence), indicating that previous evidence for host-symbiont codiversification was misinterpreted and that these symbioses are actually driven by diffuse coevolution via trait-matching.

The Gist

Imagine a massive, ancient dance floor where plants and fungi have been partnering up for over 400 million years. This is the world of mycorrhizal symbiosis, a relationship where plants and fungi help each other survive (plants get nutrients, fungi get sugar).

For a long time, scientists thought these two groups were like perfect dance twins. They believed that whenever a plant family split into two new species, its fungal partner did the exact same thing at the exact same time. This idea is called codiversification. It's like thinking that if a family of dancers splits into two groups, their dance partners must split into two matching groups simultaneously, creating two identical, mirrored dance histories.

The Big Misunderstanding

The authors of this paper, Fantine Bodin and her team, decided to check the dance floor records again. They gathered data from 29 different dance floors (ecosystems) around the world, looking at everything from tiny local gardens to the entire globe.

They found something surprising: The dancers aren't twins.

Here is the difference between what people thought was happening and what is actually happening, explained with an analogy:

1. The "Mirror Image" Myth (Codiversification)

• What people thought: If you look at the family tree of plants and the family tree of fungi, they should look like mirror images. Every branch on the plant tree should have a matching branch on the fungal tree, splitting at the exact same time.
• The Reality: The trees do not match. The plant family tree looks nothing like the fungal family tree. They didn't split apart together in a synchronized dance.

2. The "Neighborhood Effect" (Cophylogenetic Signal)

• What they actually found: While the family trees don't match, there is still a pattern. Closely related plants tend to dance with closely related fungi.
• The Analogy: Imagine a high school. Students who look similar (maybe they are cousins) tend to hang out with the same group of friends.
  • If you have a family of "Red-Headed Plants," they mostly hang out with "Red-Haired Fungi."
  • If you have a family of "Tall Plants," they mostly hang out with "Tall Fungi."
  • Why? It's not because they split apart at the same time. It's because they share traits (like height or hair color) that make them compatible. A "Red-Headed Plant" just can't dance well with a "Blonde Fungus" because their steps don't match.

The Real Story: The "Key and Lock"

The paper suggests that the reason these partners stick together isn't because of a synchronized birth, but because of trait matching.

Think of it like a key and a lock:

• Plants have evolved specific "locks" on their roots.
• Fungi have evolved specific "keys" to open them.
• Because these keys and locks are passed down through generations (they are conserved), a plant that looks like its parent will have a similar lock. Therefore, it will only accept a fungus with a similar key.

This creates a pattern where related plants hang out with related fungi, but it doesn't mean they evolved together in a synchronized way. They just evolved to fit each other's existing shapes.

Why Does This Matter?

For years, scientists looked at these patterns and said, "Aha! They codiversified! They are best friends who grew up together!"

This paper says, "Wait a minute. You're confusing hanging out with similar people (cophylogenetic signal) with growing up as twins (codiversification)."

The Takeaway:
Plants and fungi aren't synchronized twins who split apart at the same time. Instead, they are like neighbors who speak the same language. They interact because they share compatible traits (like a key fitting a lock), not because their family histories are identical.

This changes how we understand the history of life on Earth. It suggests that the evolution of these relationships is driven by functional compatibility (can we work together?) rather than synchronized history (did we split at the same time?).

In a Nutshell

• Old Idea: Plants and fungi are mirror-image twins who split apart together.
• New Discovery: They are just compatible neighbors. Related plants hang out with related fungi because they share similar "tools" (traits), but their family trees are totally different.
• The Lesson: Just because two groups look similar doesn't mean they grew up together; sometimes, they just happen to fit the same puzzle piece.

Technical Summary

1. Problem Statement

The evolutionary history of mycorrhizal symbioses (mutualistic associations between >90% of vascular plants and fungi) has long been hypothesized to involve codiversification (parallel diversification of host and symbiont lineages). Previous studies often interpreted phylogenetic congruence (mirroring tree topologies and divergence times) as evidence of this process.

However, the authors argue that a critical distinction has been overlooked in the literature:

• Cophylogenetic Signal: A pattern where closely related hosts interact with closely related symbionts. This can arise from various processes, including trait matching or biogeography, without requiring simultaneous speciation.
• Phylogenetic Congruence: A stricter pattern where host and symbiont phylogenetic trees mirror each other in topology and timing, indicative of codiversification driven by repeated cospeciation events.

The central problem is that previous studies have conflated these two patterns, leading to potentially erroneous conclusions that mycorrhizal fungi and plants codiversified. This paper aims to re-evaluate the evidence for codiversification by rigorously distinguishing between these patterns using modern cophylogenetic methods.

2. Methodology

The authors conducted a meta-analysis of 29 diverse mycorrhizal networks derived from 13 different publications. These networks covered various spatial scales (local to global), evolutionary scales (14 Myr to >450 Myr), and the four main types of mycorrhizae (arbuscular, ectomycorrhizal, orchid, and ericoid).

Data Processing:

• Phylogenetic Reconstruction: For networks lacking robust trees, the authors retrieved DNA sequences (plants: trnL, ITS, matK, matR, 18S, ATPa; fungi: SSU rRNA, ITS, 28S, etc.).
• Alignment & Tree Building: Sequences were aligned using MAFFT, trimmed with trimAl, and phylogenies reconstructed using IQ-TREE.
• Handling Uncertainty: For fungal trees with limited resolution (e.g., ITS only), the authors used a "backbone approach" (constraining deep nodes with robust multi-gene trees from literature) to ensure topological reliability.

Analytical Framework:
The study employed two distinct analytical approaches to test for different evolutionary patterns:

1. Testing for Cophylogenetic Signal (Global-fit):

   • Method: PACo (Procrustean Approach to Cophylogeny) and ParaFit.
   • Goal: To determine if closely related plants interact with closely related fungi more often than expected by chance.
   • Metric: $R^2$ and p-values derived from 10,000 permutations.

2. Testing for Phylogenetic Congruence (Event-based):

   • Method: eMPRess (event-based reconciliation).
   • Goal: To reconstruct the most parsimonious evolutionary scenario involving cospeciation, host transfer, duplication, and loss.
   • Criteria for Congruence: A scenario was considered significant only if:
     • The total cost was significantly lower than permuted data ($p < 0.05$).
     • The number of cospeciation events exceeded the number of transfer events (ratio > 1).
   • Note: The authors acknowledge that while eMPRess tests topology, true codiversification also requires matching divergence times, which could not be fully assessed due to a lack of time-calibrated fungal trees. However, the absence of topological congruence is sufficient to reject codiversification.

3. Key Contributions

• Conceptual Clarification: The paper explicitly disentangles "cophylogenetic signal" from "phylogenetic congruence," providing a clear framework (Table 1) to prevent future misinterpretation of cophylogenetic data.
• Methodological Rigor: By applying event-based models (eMPRess) alongside global-fit models (PACo) to a large, diverse dataset, the study demonstrates that global-fit methods alone are insufficient to prove codiversification.
• Correction of Literature: The authors systematically re-evaluate previous claims of codiversification in mycorrhizal systems, suggesting that these findings were likely artifacts of detecting cophylogenetic signal rather than true cospeciation.

4. Results

• Significant Cophylogenetic Signal: 13 out of 29 networks (45%) showed significant cophylogenetic signal using PACo. This confirms that closely related plants do tend to interact with closely related fungi across various scales and mycorrhizal types.
• Absence of Phylogenetic Congruence: Zero of the 29 networks showed significant phylogenetic congruence according to eMPRess.
  • In all cases, the number of inferred host transfer events exceeded the number of cospeciation events.
  • The reconciliation scenarios were not significantly better than random permutations when requiring a dominance of cospeciation.
• Robustness: Results remained consistent regardless of the phylogenetic reconstruction method (e.g., using ITS-only vs. multi-gene backbone trees) or the specific cost parameters used in eMPRess.
• Interpretation: The observed cophylogenetic signal is likely driven by diffuse coevolution and trait matching (e.g., conserved root architecture or molecular signaling) rather than strict one-to-one cospeciation. Biogeography (vicariance) may also contribute at larger scales, but trait matching explains the signal even at local community scales.

5. Significance

• Paradigm Shift: The study challenges the prevailing view that mycorrhizal symbioses are driven by codiversification. Instead, it suggests that these interactions are shaped by diffuse coevolution, where evolutionary conservatism of functional traits allows related species to interact, without requiring simultaneous speciation.
• Implications for Evolutionary Biology: It highlights that "phylogenetic tracking" (where symbionts track host diversification) is not the dominant mode in mycorrhizal networks. The high degree of partner sharing and generalization in these networks likely prevents the strict isolation required for frequent cospeciation.
• Future Directions: The authors emphasize that understanding the evolution of mycorrhizal symbioses requires focusing on the functional traits that mediate compatibility (trait matching) rather than solely relying on phylogenetic tree topology. This framework should be applied to other plant-associated interactions (pollination, seed dispersal, parasitism) to reassess the prevalence of codiversification in nature.

In conclusion, the paper provides robust evidence that while mycorrhizal plants and fungi exhibit phylogenetic signal (relatedness correlates with interaction), they have not undergone codiversification. The evolutionary history of these symbioses is characterized by complex networks and trait-based selection rather than strict parallel speciation.