On the difficult evolutionary transition from the free-living lifestyle to obligate symbiosis

Obligate symbiosis evolved from free-living individuals most likely via the intermediate stage of facultative symbiosis. However, why should facultative symbionts, who can live independently but also benefit from their partners if these are available, give up this best of both worlds? Using the adaptive dynamics approach, we analyse a simple model, focusing on one partner of the symbiosis, to gain more insight into the selective forces that make individuals forgo the ability to reproduce in the free-living state. Our results suggest that, similar to the parasitism-mutualism continuum, the free-living way of life and obligate symbiosis are two extremes of a continuum of the ability to reproduce independently of a partner. More importantly, facultative symbiosis should be the rule as for many parameter combinations completely giving up independent reproduction or adopting a pure free-living strategy is not so easy. We also show that if host encounter comes at a cost, individuals that put more effort into increasing the chances to meet with their partners are more likely to give up the ability to reproduce independently. Finally, our model does not specify the ecological interactions between hosts and symbionts but we discuss briefly how the ecological nature of an interaction can influence the transition from facultative to obligate symbiosis.


Dear Nick Davies,
Thank you for reviewing our manuscript. You will find below our response to each point that you raised. We hope that they sufficiently clarify your questions. Response: The idea that obligate symbiosis evolves from facultative symbiosis is our extension of Poulin's idea in his book: "Evolutionary Ecology of Parasites" [1]. Here, he suggested that obligate parasites may evolve from facultative parasites if the contact between the parasite and the host is frequent enough [1,Chapter 2]. We argue that the type of ecological interactions may not play a key role in this evolutionary transition, that is, obligate symbiosis may evolve from facultative symbiosis, regardless of whether the interaction is parasitism or mutualism.
We agree that for many symbiotic systems, the associations may have resulted from prey-predator interactions, and if the engulfed symbionts are then maintained purely by vertical transmission with the host then it is 1 obvious that obligate symbiosis can evolve directly from free-living organisms.
However, the prey-predator interaction at the level of microorganisms may completely be different from the interaction at the level of macroorganisms.
For instance, a deer is surely dead and cannot escape after being consumed by a tiger, but a bacteria that are engulfed inside a host may escape. In fact, some hydra strains form association with algae by engulfing the algae, but the algae symbionts have been shown to be able to escape from the hydra hosts and enter the external environment ( [2]). We argue that the mode of acquiring the symbionts is not that important, and symbiosis formed via engulfing is no different from symbiosis formed via other modes such as rhizobia entering the plant through forming nodules in the roots, or Vibrio fisherie entering the bobtail squid in the specialised light organ. Therefore, we find that conventional facultative symbiosis still holds even when it is formed via prey-predator interaction.
In the example that you raised, when protomitochondria were engulfed, it does not follow by being trapped inside their bacterial hosts forever and could only vertically transmit with the hosts and benefit from the hosts. If for any reason, the protomitochondria can escape from the host, they have to deal with the external environment, such as harsh environment, intra and interspecific competition, then they should be better off retaining the ability of independent reproduction, thus, evolving toward obligate symbiosis may not seem so obvious.
Below is our modified paragraph. It can be found in line 22-32 in the revision text file.
" Obligate symbionts have to evolve from free-living individuals, regardless of whether they are parasites or mutualists, because if there were no free-living individuals to begin with, there would be no ingredients for asso-ciations. If this process occurs due to, for instance, an event of engulfment of the symbiont by the host, followed by purely vertical transmissions of the symbionts to maintain the population then obligate symbionts may evolve directly from free-living organisms. However, many symbionts have to expe- "Theoretical studies that address the evolutionary transition toward obligate symbiosis are rare, but even there, this aspect is not the main focus. I think its more about ecto/endosymbiosis, focusing on the symbiont rather than the host, but because the introduction was quite sweepinge.g. citing in lines 49 previous work that looks very broadly at e.g. mutualisms like cleanerclient and plant-pollinatorits hard to grasp whats being modelled here. Also, the parameter sigma needs to be explained here: when an A makes another A at rate sigma, is the new A associated with the same host as its parent or a new one?
Response: First, we affirm that our model only applies to endo/ectosymbiosis, such as algae-fungi in lichen, aphid-Buchnera, Rhizobia-legume, and so on, but not to the relationships like plant-pollinator.
Furthermore, we completely agree with the reviewer that the definition of symbiosis is rather ambiguous. When symbioses were discovered in the late 19th centuries, de Bary defined symbiosis as "the living together of unlike named organisms". According to this definition, symbiosis includes both parasitism, mutualism, and associations of all level of partner dependency, from endo/ectosymbiosis, such as lichen, and insects-bacteria to short term symbiosis such as cleaner-client and plant-pollinator. However, it is not very clear whether de Bary did want to include associations such as plantpollinator or cleaning symbiosis. Zook [3] did mention this ambiguity and redefined symbiosis, that is "Symbiosis is the acquisition of an organism(s) by another unlike organism(s), and through subsequent long-term integration, new structures and metabolism(s) emerge". The symbiotic system that we modelled is closer system defined by Zook [3], that is, it includes only endo/ectosymbiosis. It is important to note that in terms of the type of interactions, we still follow de Bary's definition, that is we consider both parasitism and mutualism, even though the type of interactions cannot be taken to account in this model because we considered a fixed number of hosts for simplification.
We also add our affirmation when explaining the model schematic, which can be found in line 100-102.  [4]; in fact, many obligate parasites are only horizontally transmitted (it should be noted that we only consider symbionts with simple life cycle, that is symbionts with only one host). Secondly, early in the evolutionary transition from facultative to obligate symbiosis, vertical transmission is likely to be rare as it may require specific adaptations [5]. Therefore, considering the evolution of both types of transmission will make our model much more complicated while we are aiming for a simple model as a start to understand this transition.
# 117 -118: Does this assumption reduce the generality of your model?
Why was this assumption used to prevent the population of associations from growing without bound, instead of e.g. a saturating growth function?

Response:
We do not think that the assumption that the vertical transmission is smaller than the bound mortality rate (σ < ν) will reduce the generality of our model because as we responded in the question above, The assumption that σ < ν is to prevent the population of A from growing exponentially when the host encounter rate β evolves toward 0. This is the case of completely free-living population because even when there are free-living symbionts and hosts around, the symbionts will not encounter and establish inside the host and form association. If the mutant of zero β is created through horizontal transmission then nothing will happen but if it is created via vertical transmission, then the population of association A will grow exponentially. A saturating growth function require at least quadratic function with regard to the association population, and will prevent analytical results of the equilibrium. Response: The caption has been modified accordingly.