Remembrance of phytoplankton’s past

A snapshot of a population of cells—with cilia, vacuoles, and flagella along with a kaleidoscope of shapes and colors—does not do justice to the complex dynamics and subcellular structures teeming within. The internal workings of cells are capable of complex processing, computation, and coordination. These processes can be witnessed within phytoplankton that have subsumed chloroplasts and mitochondria for combined production of energy and to help determine allocation toward or away from growth and new biomass (1–3). A compelling paper in PNAS by Anderson et al. (4) shows that phytoplankton are even capable of a type of remembering—known as phenotypic memory—and they devise theory and experiments to discover its mechanistic underpinnings and climatic consequences.

The balance of energy, nutrient, resource, and biomass fluxes through the cell largely determines its fate, as it does for all of life. Teasing apart the tradeoffs among these different fluxes is therefore needed to understand the future of a single cell or population of cells or their global distribution. Delicate balances of fluxes allow phytoplankton to uptake resources for energy and growth while avoiding becoming overly bulky and sinking. However, these balances are neither perfect nor immutable. The fact that the balance is imperfect—here perfect means input exactly matches output—allows the cell to store resources that are useful for surviving future stresses and gaining an evolutionarily advantage. The balance is mutable because, for instance, changes in resources and nutrients—such as dumping large amounts of phosphorus into a lake (5)—are well known to cause fundamental shifts in biomass growth. This is tantamount to changing the input to the system of the cells.

However, another way to alter a cell and a system, even when the input resources are unchanged, is to tune the cell’s internal fluxes. An omnipresent and eternal dial for doing this tuning is temperature (6). In the simplest scenario that all chemical reactions (including ones for nutrient production) behave the same, temperature would affect all organelles, processes, and fluxes identically. Consequently, increasing temperature would be like fast forwarding through a movie, and decreasing temperature would be like slow motion. That is, none of the actions or sequences would change. We would just observe them being stepped through more slowly or more quickly.

In contrast, if some fluxes, certain organelles, and even processes within the cell and organelles are differentially tuned by temperature, then the action and sequences—corresponding to the net output of the cell—can change, sometimes dramatically. Adding even more complexity is the resource storage alluded to above. Properly understood, the amount of resource storage is a running tally of the difference between the previous inputs to and outputs from the cell. From that perspective resource storage partly serves as a record of the cell’s history, and said differently, as a form of memory.

Treating the accumulation of resources within a single cell as a form of memory may at first feel discomfiting to a neuroscientist or according to the usual definitions. This form of memory is not in the neuronal or digital sense of storing and retrieving specific information for conscious processing or learning or calculation. Single cells clearly have no brain or hard drive. But resource storage allows the cell’s past experiences to impact its present and future states. And from a mathematical and practical viewpoint the cell’s ability to integrate the past to impact the present is indeed a type of memory. In physics terms, this can be thought of as hysteresis in which even inanimate magnets exhibit a type of memory (7). More biologically, the muscle mass and fat content of our bodies are partly due to our past diet and exercise, both representing a record of the past and retaining information for the future via things like muscle memory (8). Similarly, tree rings are a record of the age and growth of a tree that can also be deciphered as a key to past environments (9).

Anderson et al. (4) connect these ideas to the concept of phenotypic memory (10)—a phenotype persisting due to nongenetic causes and potentially across generations even after the cell or organism is removed from the environment that induced that phenotype. In this sense, it can also be considered as a type of plasticity (11) that lingers and thus encodes a memory. In many cases, phenotypic memory applied to the level of bacteria and single cells has been tied to epigenetics, meaning changes in gene expression. In contrast, Anderson et al. (4) tie phenotypic memory of phytoplankton directly to resource storage and how it depends on differential temperature responses of cellular fluxes such as nutrient uptake and assimilation of biomass. I propose that this could be termed a nonneuronal type of “physiological” memory (12) that would be a subset of phenotypic memory.

By analogy to the work of Anderson et al. (4), resource uptake rate can be thought of as work income, assimilation to biomass as funds spent on infrastructure, and resource storage as savings and possibly investments. Consequently, increasing temperature increases the spending rate on infrastructure (assimilation of biomass) due to faster decomposition and the need for replacement, while it also increases the rate of work (resource uptake). The delicate balance of these rates and accumulated resource storage is governed by how these relative rates change with temperature, and through theory and experiment, Anderson et al. assess which fluxes change most quickly. They find that assimilation increases with temperature faster than resource uptake rate, meaning the output increases faster than the input and thus a depletion of resource storage at higher temperatures will occur. Analogously and ignoring other factors (e.g., endothermy versus ectothermy), we might work more when it is warm outside but we eventually become exhausted and have to rest, whereas there is no real rest for the relentless decomposition and degradation that is happening. Consequently, higher temperatures lead to lower levels of resource storage that serve as the savings and investments for the cell, hence leaving the phytoplankton more vulnerable to future stresses. These future stresses could be lower levels of resources and even higher temperatures. The result of Anderson et al. means that the stresses are not just due to the current instantaneous temperature, but instead a cumulative result of more frequent higher temperatures that the phytoplankton has not yet adapted to handle. Similarly, a sudden job loss or a recession might be impossible to compensate for with our current budget if we do not have savings in reserve that would allow us to survive until there is a reprieve. For the phytoplankton experiments conducted in Anderson et al., this inability to compensate with resource storage reserves is exactly evident in the changes in the thermal breadth via shifts in the minimum and maximum temperatures (T_min and T_max) at which phytoplankton can grow and how those depend on the phytoplankton’s thermal history.

“A compelling paper in PNAS by Anderson et al. (4) shows that phytoplankton are even capable of a type of remembering-known as phenotypic memory-and they devise theory and experiments to discover its mechanistic underpinnings and climatic consequences.”

Part of what sets the Anderson et al. study apart is not just the rich concepts that are verbally described here, but their rigorous quantitative analysis via both theory and experiment. In particular, they construct a mathematical model to decompose these fluxes and temperature effects into terms that can be manipulated and used to calculate predictions. These predictions are then compared with empirical data from incisive thermal experiments using phytoplankton. The results and conclusions align in ways that convincingly show i) assimilation rate to biomass changes faster with temperature than does resource uptake rate, ii) resource storage within the cell is depleted more at higher temperatures, iii) ability to compensate for future stress is thus diminished following a history of high temperatures, and iv) therefore, there is a narrower range of temperatures at which the phytoplankton can survive. This narrowing of temperatures at which phytoplankton can grow is the impact of the phenotypic memory due to previous thermal stress that is materially mediated by the reduced resource storage.

On the modeling side, Anderson et al. employ a coupled set of differential equations to track the dynamics of biomass growth (nutrient assimilation rate) and nutrient cell density (also called the quota and representing the resource storage). The authors are able to find solutions by deriving an equilibrium that corresponds to the resource storage that occurs at an experimental temperature to which the phytoplankton are acclimated. They then substitute this equilibrium back into the dynamical equations to determine how the cells will respond and survive when subjected to acute exposure to a range of temperatures. The experimental setup maps to a mathematical method that is akin to a separation of time scales between the acclimated and acute period, making the problem both mathematically tractable and experimentally practical. This clever matching of experimental feasibility with mathematical approximation permits concrete quantitative questions to be asked and answered that the authors use to elucidate how phenotypic memory is manifest for phytoplankton and what its consequences will be.

Nevertheless, several questions remain. For example, how do resource storage and memory change under more complicated temperature time courses—more stochastic changes, faster changes, multiple layers of time scales for changes—such as those in nature? Despite the successes of the current model, it would be challenging to answer these questions without further development of theory and more experiments. In parallel, it may be essential to assess what proportion of the cells’ fluxes and resource storage are due to memoryless stochastic processes versus processes with memory.

There are a variety of modeling techniques to tackle these questions, most notably the Langevin equation (13) and Fokker–Planck equation (14) as well as fractional diffusion and fractional derivatives (15). Further complicating the paper’s approach, however, is that memory or hysteresis typically alters the differential equations via time-delay terms or some sort of integration over time, transforming the ordinary differential equations into delay-differential equations (16) or integrodifferential equations (17) that each have their own technical difficulties. This inclusion of time lags and previous time scales into the equations could also alter our current thinking about system stability (18, 19) and the potential for chaotic dynamics (19). More technically elaborate models—that would be capable of dealing with more realistic histories of temperature—could be developed in the future that draw on these other techniques.

Returning to the tree ring analogy, the phenotypic memory of the phytoplankton does not provide a physical visual snapshot that can be read as a timeline in the same way as for tree rings or as sedimentation of rock strata and fossils. Yet, it is a memory that even these single cells can hold and that affects global ecological health. Just as trees in the Amazon rainforest are vital for ecosystem function, phytoplankton in lakes and oceans are ubiquitous primary producers that underlie global functioning and health (20). Understanding the memories of phytoplankton and how climate is influencing them and can stress them beyond their limits will be essential for discerning how climate change will impact biological systems and life at a global scale.

One lesson that can possibly be generalized from this study is that it is crucial to understand that the impacts of climate change are cumulative. It is not only the immediate damage we see from a single heat wave or hurricane. It is not only higher average temperatures but also, and perhaps more importantly, the increased frequency of heat waves that could deplete internal cellular resources, effectively shorten phytoplankton memories, and potentially lead to die-offs and perilous outcomes. There is a sentiment that if we can rest now, it is because of work we have done in the past. But it is not just our work; it is the resultant effects of evolution and the Earth. It may be both the limited phenotypic short-term memory of phytoplankton as well as its few billion years of evolution and linkages to global health and life that matter.