Mutualism & Host-Parasite Coevolution
The eco-evolutionary logic of mutualisms, antagonistic coevolution, and the Red Queen
A Brief History
Species Interactions Revisited
In Interacting Species, we explored competition and predation as the primary modes of ecological interaction. But biological interactions span a continuum, from pure antagonism to pure cooperation. The framework of interaction coefficients captures this diversity: the sign and magnitude of effects one species exerts on another's fitness.
Consider two species with population dynamics modified by their interaction:
$$\frac{dN_1}{dt} = r_1 N_1 \left(1 - \frac{N_1}{K_1} - \alpha_{12} \frac{N_2}{K_1}\right)$$ $$\frac{dN_2}{dt} = r_2 N_2 \left(1 - \frac{N_2}{K_2} - \alpha_{21} \frac{N_1}{K_2}\right)$$The coefficients $\alpha_{12}$ and $\alpha_{21}$ determine the outcome:
- Competition ($\alpha_{12} > 0, \alpha_{21} > 0$): each species inhibits the other (−/−)
- Predation ($\alpha_{12} > 0, \alpha_{21} < 0$): prey harmed, predator benefits (+/−)
- Commensalism ($\alpha_{12} = 0, \alpha_{21} < 0$): one benefits, one unaffected (+/0)
- Parasitism ($\alpha_{12} > 0, \alpha_{21} < 0$): host harmed, parasite benefits (−/+)
- Mutualism ($\alpha_{12} < 0, \alpha_{21} < 0$): both benefit (+/+)
Mutualism and parasitism are not separate phenomena but points on a continuum of coevolving interactions. A relationship that appears parasitic at one moment may shift toward mutualism if partners evolve to exploit the interaction more efficiently. Conversely, a stable mutualism can collapse if one partner evolves to cheat. The key question is not whether an interaction is mutualistic or parasitic, but what mechanisms maintain cooperation versus exploitation, and how evolutionary dynamics allow transitions between states.
This chapter explores how such interactions evolve, persist, and sometimes break down. The Red Queen hypothesis—that species must constantly evolve just to keep pace with their partners—provides a unifying framework for understanding coevolutionary arms races and the maintenance of genetic diversity in natural populations.
The Mathematics of Mutualism
The Lotka-Volterra framework extends naturally to mutualism. If species 1 and 2 both benefit from interaction, the coefficients $\alpha_{12}$ and $\alpha_{21}$ are negative, representing a reduction in mortality or an increase in fecundity:
$$\dot{N}_1 = r_1 N_1 \left(1 - \frac{N_1}{K_1} + \alpha_{12} \frac{N_2}{K_1}\right)$$ $$\dot{N}_2 = r_2 N_2 \left(1 - \frac{N_2}{K_2} + \alpha_{21} \frac{N_1}{K_2}\right)$$Here $\alpha_{12} < 0$ and $\alpha_{21} < 0$. The positive coefficients in the interaction terms mean that each species increases the other's growth rate. But this creates a fundamental instability: positive feedback can lead to unbounded growth, especially if mutualism is obligate (each species requires the other for survival).
For a stable interior equilibrium with both species present, the system must satisfy:
$$1 - \alpha_{12} \alpha_{21} > 0$$
If $|\alpha_{12} \alpha_{21}| > 1$ (strong reciprocal benefits), the system becomes unstable and populations either grow without bound or collapse. This is the paradox of mutualism: strong mutual benefits are destabilizing!
How then do mutualisms persist? The answer lies in saturating benefit functions. Rather than linear benefits, real mutualisms often show diminishing returns as density increases:
$$\dot{N}_1 = r_1 N_1 \left(1 - \frac{N_1}{K_1} + \alpha_{12} \frac{N_2}{K_1(1 + \beta N_2)}\right)$$The term $(1 + \beta N_2)$ in the denominator causes benefits to plateau. This non-linear saturation stabilizes the system while preserving mutualistic effects at low density. It reflects biological reality: a plant pollinated by 100 insect visits gains less extra seed set than a plant pollinated by 10, because fruit set becomes limited by other factors (water, nutrients, etc.).
The Evolution of Mutualism
How does parasitism become mutualism? The transition is mediated by partner choice and partner sanctions—the ability of one species to preferentially interact with, reward, or punish partners based on their cooperativeness.
Consider a host in a parasitic interaction with a pathogen. If the host evolves resistance (a private benefit), and simultaneously evolves to preferentially associate with less virulent pathogen strains (partner choice), then selection favours virulence reduction in the parasite. Over time, the parasite's transmission rate may become dependent on host cooperation: a parasite that exploits too heavily is abandoned by the host, reducing its transmission. This selection pressure converts an antagonist into a mutualist.
A striking experimental and theoretical result: mutualistic populations are not stable endpoints. In laboratory yeast-algae cocultures, obligate mutualisms spontaneously break down when cheater variants arise—cells that gain mutualistic benefits without reciprocating. These cheaters initially increase in frequency, but as they proliferate, the cost to the other partner becomes unsustainable. The mutualism collapses, followed by selection for renewed cooperation as populations face mutual extinction. The system oscillates between mutualistic and parasitic states on ecological timescales, refuting the notion that mutualism represents a stable attractor.
The tragedy of the mutualism commons is the core problem: in a mutualistic interaction, natural selection favours individuals that gain benefits without providing them (cheaters). Yet if cheating becomes too common, the mutualism collapses, punishing all. This tension—between individual selection for cheating and group-level selection for stability—defines the evolutionary dynamics of cooperation.
Mechanisms that prevent the tragedy include:
- Host sanctions: Hosts withhold or withdraw resources from non-cooperating partners (e.g., legumes nodulating rhizobia allocate more carbon to nodules of cooperative strains).
- Partner fidelity feedback: Each partner's fitness directly depends on its partner's fitness, aligning individual and group interests.
- Biological markets: Multiple potential partners allow "shopping" for the best cooperators, driving competition for resources among partners.
The Red Queen Hypothesis
Leigh Van Valen's (1973) Red Queen hypothesis—named after Lewis Carroll's character who runs constantly just to stay in place—provides a unifying vision of coevolution. In a coevolving system, organisms are locked in an evolutionary arms race: as one species evolves to exploit or resist another, the other must evolve in response. The result is perpetual evolutionary change with no net long-term increase in fitness— evolution for the sake of staying competitively equivalent.
Hamilton extended this insight (1980), proposing that sexual reproduction persists because genetic recombination generates novel allele combinations faster than asexual clones, allowing faster coevolution with parasites. A host population of clones is rapidly outpaced by an evolving parasite; a sexually reproducing host population, by contrast, maintains genetic diversity that keeps parasites off-balance.
Consider a simple two-allele system in both host and parasite, with a matching-allele model where each host allele has highest fitness against a different parasite allele:
Host allele $i$ has fitness $W_i = 1 - s$ when matching parasite allele $i$ (matched), and fitness $W_i = 1$ when mismatched. A parasite that matches the common host allele can spread; but as it increases, the common host allele becomes a liability, and selection favours the alternative host allele. This drives parasite evolution in turn. The result is an oscillation in allele frequencies with no stable equilibrium—the Red Queen cycle.
Mathematically: $\dot{x}_i = x_i(\bar{W}_i - \bar{W})$ and $\dot{y}_j = y_j(\bar{V}_j - \bar{V})$, where $\bar{W}_i$ is average host fitness for allele $i$, and $\bar{V}_j$ is average parasite fitness for allele $j$. If fitness is determined by genotype matching, these equations generate deterministic chaos or limit cycles.
An alternative model is the gene-for-gene system, where a host resistance allele $R_i$ confers immunity against parasite virulence allele $V_i$. Here, specific coevolutionary pairs—$(R_1, V_1)$, $(R_2, V_2)$, etc.— determine the dynamics. This model often produces multiple stable equilibria and more complex coevolutionary trajectories than the matching-allele model.
Host-Parasite Coevolution
Real host-parasite systems combine epidemiological and evolutionary dynamics. May and Anderson (1983) developed a framework coupling population growth with genotypic evolution, showing how virulence and transmission rate evolve together.
The key insight is the virulence-transmission trade-off: a parasite that reproduces rapidly within a host damages the host severely, reducing host survival time and transmission opportunities. Conversely, a benign parasite may transmit poorly but persists longer in the host. The optimal virulence $\alpha^*$ balances these costs and benefits, maximizing the basic reproduction number $R_0$.
Coevolutionary cycling arises when host resistance increases (reducing parasite fitness), driving selection for higher parasite virulence to overcome resistance. But increasing virulence makes the parasite more detectable by the host immune system, selecting for further resistance. The result is an arms race with no stable endpoint.
A landmark study by Schenk, Traulsen & Gokhale (2017) revealed chaotic provinces in host-parasite coevolution: regions of parameter space where genotype frequencies follow deterministic chaos rather than simple cycles. These chaotic dynamics explain why some disease dynamics are unpredictable on long timescales despite being fully deterministic.
Czuppon & Gokhale (2018) further disentangled ecological and evolutionary effects, showing that in finite populations, the fixation of traits depends not just on selection but on demographic stochasticity. A mutation that increases virulence may be beneficial in the long run (by resisting host immunity) but causes immediate population collapse, stochastically driving it to extinction before selection can act.
Explore: Red Queen Coevolution
Eco-Evolutionary Feedbacks
When ecology and evolution happen on the same timescale, their feedbacks become critical. Traditional models treat ecology (population dynamics) and evolution (genotype frequencies) separately: first populations reach equilibrium, then selection acts slowly. But in many real systems, allele frequencies and population sizes change together, creating bidirectional feedbacks.
Gokhale & Traulsen (2014) developed a framework for "evolutionary multiplayer games in ecological settings," showing how a genotype's fitness depends not just on its frequency relative to other genotypes, but on absolute population density and resource availability. A cooperative genotype might be stable at low density (where the costs of cooperation are easily met) but unstable at high density (where resources become limiting). This creates a density-dependent switch in evolutionary stability.
Denton & Gokhale (2019) extended this to include environmental heterogeneity, showing how spatial variation in resources couples ecological dynamics to evolutionary stability. In patchy environments, genotypes that excel locally may spread, only to crash when populations expand into unfavourable patches, creating dynamic spatial-temporal fluctuations in both ecology and evolution.
A striking parallel exists with cancer as an eco-evolutionary process (Cancer & Cell Dynamics): tumours are ecological communities of cells with different genotypes, evolving under strong selection in a nutrient-limited environment. The tumour's ecology (microenvironment, vasculature, immune infiltration) feeds back on evolutionary dynamics, favouring genotypes that exploit the environment most efficiently. Understanding this feedback is key to cancer treatment: a therapy that kills the majority of cells creates an ecological collapse that selects for highly virulent resistant mutants, regenerating the tumour more aggressively than before.
The unifying insight: whenever populations reproduce, experience selection, and modify their environment, ecological and evolutionary dynamics are inseparable. The mathematics of coevolution, mutualism, and host-parasite arms races all reflect this deep coupling.
Exercises
Conceptual Questions
- Explain the Red Queen Hypothesis and why it applies to both host–parasite and mutualistic relationships. Why must species "run just to stay in place"?
- What is the fundamental difference between an obligate mutualism (where both partners cannot survive alone) and a facultative one? Give biological examples of each and discuss their stability.
- In host–parasite coevolution, what does it mean for virulence to evolve? Why might a parasite evolve reduced virulence over time, and under what conditions would it instead become more virulent?
- How do gene-for-gene matching systems (like in plant-pathogen interactions) maintain genetic variation in both host and pathogen populations? Why do such systems prevent selective sweeps?
- In fig-wasp mutualisms, partner choice and host sanctions maintain cooperation. Explain how these mechanisms prevent the cheater problem where a partner exploits the relationship.
Computer Problems
- Lotka–Volterra Mutualism Dynamics. Implement a mutualistic system where both partners benefit: $\frac{dN_1}{dt} = r_1 N_1(1 + \beta_1 N_2 - N_1/K_1)$ and $\frac{dN_2}{dt} = r_2 N_2(1 + \beta_2 N_1 - N_2/K_2)$ with $r_1 = r_2 = 0.8$, $K_1 = K_2 = 100$, and $\beta_1 = \beta_2 = 0.01$. Plot trajectories for weak and strong mutualism, and show when unbounded growth occurs versus stable coexistence.
- Host–Parasite Coevolution with Gene-for-Gene. Implement a discrete-time host–parasite model where host genotype $H$ resists parasite genotype $P$ with specificity: fitness $w_H = 1 - s \cdot \mathbb{1}(H \text{ susceptible to } P)$ and $w_P = 1 + b - c \cdot \mathbb{1}(P \text{ incompatible with } H)$. Simulate allele frequency dynamics over 1000 generations with mutation rate $\mu = 10^{-4}$. Plot the outcome: coevolutionary cycling or balanced polymorphism?
- Parasite Virulence Evolution. Model a tradeoff where parasite transmissibility increases with virulence: $\beta(v) = \beta_0(1 + av)$ and virulence cost $\mu(v) = \mu_0 + cv$. Implement $\frac{dI}{dt} = \beta(v) S I - \mu(v) I$ and compute the optimal virulence $v^*$ that maximizes the basic reproduction number $R_0 = \beta(v^*) / \mu(v^*)$. Show how $v^*$ changes with host life expectancy.
- Red Queen Cycling in Coupled Populations. Implement a continuous-time system coupling host allele frequency $p$ and parasite allele frequency $q$ with selection: $\dot{p} = p(1-p)[w_{11}(q) - w_{12}(q)]$ where fitnesses depend on matching (e.g., $w_{11}(q) = 1 - s q$). Observe limit cycles or chaotic dynamics as parameters vary. Compute Lyapunov exponents to characterise cycling.
- Mutualism Breakdown and Partner Sanctions. Extend the basic mutualism model to include cheater strains that receive mutualistic benefits without reciprocating. Implement $\frac{dN_c}{dt} = r N_c(1 + \beta N_m - N_c/K)$ and $\frac{dN_m}{dt} = r N_m(1 + \beta(N_c/(\theta + N_c)) - N_m/K)$ where $\theta$ controls host-sanction severity. Show at what sanction strength cheaters are eliminated.
References
- Van Valen, L. (1973). A new evolutionary law. Evol. Theory, 1, 1–30.
- Hamilton, W. D. (1980). Sex versus non-sex versus parasite. Oikos, 35, 282–290.
- May, R. M. & Anderson, R. M. (1983). Epidemiology and genetics in the coevolution of parasites and hosts. Proc. R. Soc. Lond. B, 219, 281–313.
- Herre, E. A., Knowlton, N., Mueller, U. G. & Rehner, S. A. (1999). The evolution of mutualisms: exploring the paths between conflict and cooperation. Trends Ecol. Evol., 14, 49–53.
- Gokhale, C. S. & Traulsen, A. (2014). Evolutionary multiplayer games. Dynamic Games and Applications, 4, 468–488.
- Czuppon, P. & Gokhale, C. S. (2018). Disentangling eco-evolutionary effects on trait fixation. Theoretical Population Biology, 124, 93–107.
- Gokhale, C. S., Frean, M. & Rainey, P. B. (2023). Eco-evolutionary logic of mutualisms. Dynamic Games and Applications, 13, 1066–1087.
- Denton, J. A. & Gokhale, C. S. (2019). Synthetic mutualism and the intervention dilemma. Life, 9, 15.
- Schenk, H., Traulsen, A. & Gokhale, C. S. (2017). Chaotic provinces in the kingdom of the Red Queen. J. Theor. Biol., 431, 1–10.
- Song, Y., Gokhale, C. S., Papkou, A., Schulenburg, H. & Traulsen, A. (2015). Host-parasite coevolution in populations of constant and variable size. BMC Evol. Biol., 15, 212.
- Sachs, J. L., Mueller, U. G., Wilcox, T. P. & Bull, J. J. (2004). The evolution of cooperation. Q. Rev. Biol., 79, 135–160.
- Bronstein, J. L. (2001). The costs of mutualism. Am. Nat., 157, 445–454.