By virtue of complex interactions, the behaviour of mutualistic systems is dif-12 ficult to study and nearly impossible to predict. We have developed a theoretical 13 model of a modifiable experimental yeast system that is amenable to exploring 14 self-organised cooperation while considering the production and use of specific 15 metabolites. Leveraging the simplicity of an artificial yeast system, a simple 16 model of mutualism, we develop and test the assumptions and stability of this 17 theoretical model. We examine how one-off, recurring and permanent changes 18 to an ecological niche affect a cooperative interaction and identify an ecological 19 "Goldilocks zone" in which the mutualism can survive. Moreover, we explore 20 how a factor like the cost of mutualism -the cellular burden of cooperating -21 influences the stability of mutualism and how environmental changes shape this 22 stability. Our results highlight the fragility of mutualisms and suggest the use of 23 synthetic biology to stave off an ecological collapse.
24
Introduction
25Life on earth comprises a hierarchy of units of selection. From societies to genes, we find 26 the same patterns of organisation at each of these levels [1]. However, selection is not level-27 specific. While changes may occur at low levels, such as a single nucleotide or amino acid, 28 1 selection necessarily operates at a much higher level, such as the organism. Competition 29 between entities at a specific level of organisation can spell disaster for the higher level. A 30 clear example of this breakdown of control being cancer [1].
31Mutualistic interactions, a specialised type of cooperation where replicating components 32 benefit each other, therefore, can be targeted by selection at a higher level [2]. Due to 33 conflicts of interest between entities at varying levels of selection, the origin of mutualism 34 and subsequent selection lack a clear evolutionary explanation -a complete field of research 35 in itself [3]. Even though it is hard to explain how mutualisms emerge, we see many examples 36 of them. On a global scale, prominent examples such as coral-Symbiodinium symbioses or 37 plant-rhizobia interactions are well known [4, 5, 6]. Many such mutualisms have evolved over 38 millions of years, but if mutualisms are fragile and susceptible to collapse, as hypothesised, 39 then how do they survive for aeons in constantly changing environments? 40Mutualistic interactions can emerge in numerous ways [3]. Here we focus on how they sur-41 vive, because, regardless of how they originate, mutualisms constantly face different threats.
42A common challenge for mutualistic communities is exploitation by cheater strains that 43 benefit from mutualistic interactions, but fail to contribute.
44This problem of parasitic elements, first noted by Maynard Smith [7], has been extensively 45 studied. Postulations about compensatory mechanisms that could avoid parasitic exploita-46 tion, range from conceptual [8], to mechanistic arguments [9]. Mutualism could also suffer 47 from insuff...