This study was initiated after finding an old culture of Chlamydomonas from a long-departed student (about a year earlier) on a shaker was completely bleached (and presumed dead) yet recovered when plated out on fresh media. The strategy for surviving this length of time in the absence of nutrients, and presumably an ongoing stress, brought up many interesting questions about the long-term survival under adverse conditions. Understanding how Chlamydomonas ages in batch culture and its strategy for survival is not only a relevant ecological question, but one relevant for biotechnology for the production of high-value products in cell culture (Scranton et al. 2015).
We would normally predict that Chlamydomonas doesn’t age under benign, nutrient-sufficient conditions, continuing to divide akin to an immortal cell line. Budding yeast (Saccharomyces cerevisiae), can also be propagated indefinitely, but individuals within these asymmetrically dividing unicells do eventually age and die. The budding yeast is a model system for investigating both replicative lifespan (number of cell divisions before senescence) and chronological lifespan (longevity in a non-dividing state) (Denoth Lippuner, Julou, and Barral 2014). While a chronological lifespan concept is more appropriate for Chlamydomonas, the symmetrically dividing fission yeast (Schizosaccharomyces pombe) may be a better system to compare concepts of ageing in microalgae. In fission yeast, there is no evidence of ageing in cells grown under favourable conditions; however, if these yeast cultures are stressed, evidence of cellular ageing can be detected (Coelho et al 2013). In particular, heat shock protein-associated protein aggregates are asymmetrically distributed between “symmetrically” dividing cells such that there is an old and young product of cell division (Coelho et al. 2013; 2014). The younger of these cells would be free of the protein aggregates, putting it on the path to recovery. Even the bacterium, Escherichia coli, that apparently divides symmetrically, the offspring can differentially possess new or old cell components that can affect its subsequent growth rate (Stewart et al. 2005; Lindner et al. 2008). While there are other studies looking at ageing in unicellular organisms (Florea 2017), it’s unknown whether something similar occurs with Chlamydomonas cells during cell division. However, culture heterogeneity has been described where there are different populations of cells with distinct growth-rates in the same culture (Damodaran et al. 2015), which may suggest that there is some type of ageing occurring in these cells even though these symmetrically dividing cells were grown under favourable conditions.
Our system deals more with conditional senescence, where cells senesce as nutrient levels are depleted and there is a transition from favourable to unfavourable conditions. It’s under these conditions where cell division is limited that we would predict cellular ageing to accelerate. The transition to conditional senescent conditions is first evidenced by the changing growth rate as cells run out of nutrients. This declining growth phase where cells continued to grow, though at a much-reduced rate, lasted for about 40 days. The declining growth phase is able to support limited growth by recycling internal macromolecules. In particular, protein levels per cell decline by close to 70%. This is a long-observed pattern in bacteria where protein levels decline upon entry into stationary phase when other stored nutrients are depleted (Wanner and Egli 1990). Similar protein declines are also observed in Chlamydomonas reinhardtii during nitrogen, phosphorus, or sulphur deprivation experiments (Cakmak et al. 2012; Kamalanathan et al. 2016). While part of the mechanism for this reduction in protein during senescence is dilution through cell division (Meagher et al. 2021), the activation of autophagy is likely an essential process regulating the turnover of proteins to enhance nutrient recycling. Evidence for the importance of autophagy in senescence includes the induction of ATG8 in Chlamydomonas cultures approaching stationary phase (Pérez-Pérez, Florencio, and Crespo 2010; Meagher et al. 2021). ATG8 is a ubiquitin-like protein that is conjugated to a phospholipid in a step required for autophagosome formation (Mizushima, Yoshimori, and Ohsumi 2011; Pérez-Pérez and Crespo 2010). Autophagy was found required for turnover of ribosomal proteins and regulation of lipid metabolism (Couso et al. 2018), so this process is important for metabolic restructuring in the cell. This nutrient recycling likely supported a limited cell division, but it also extended lifespan during conditional senescence— cell viability remained high during the declining growth phase.
It’s clear that the duration and characteristics of the declining growth phase is variable depending on how the cells are grown, and perhaps the strain used. For instance, Humby et al. (2103) used a cell wall-less strain under low gas-exchange conditions and observed a DGP that was very short; cells entered stationary shortly after the exponential phase. Under active bubbling for maximum gas exchange using the same strain as in this study (CC125), the declining growth phase lasted for at least 10 days following the exit from logarithmic growth (Meagher et al. 2021). The declining growth phase is often ignored or not identified in many culturing systems. In fact, these extended declining growth phases are a hallmark of batch culture growth of bacteria (Wanner and Egli 1990) and fungi (Borrow et al. 1964; Vrabl et al. 2019). It’s been proposed that entry in this phase is due to a limiting nutrient that curbs cell growth, the slope of which gets lower as other nutrients become depleted (Monod 1949; Vrabl et al. 2019). This is likely the case in the variation we observed in different studies, where there is depletion of CO2, acetate, nitrogen, phosphorous, or a micronutrient. It’s clear that by altering the culturing conditions, particularly changing the potential for gas exchange, the progression of the standard growth curve can shift dramatically.
Chlorophyll degradation/dilution in conditional senescence starts shortly after exiting exponential phase and continues through the declining growth, stationary, and death phases. The degradation of chlorophyll during senescence is a common response in photosynthetic organisms and the result of activation of several enzymes that break down chlorophyll to a variety of catabolites (Hörtensteiner 2006), and in some ways resembles the loss of chlorophyll during nutrient deficiency (Plumley and Schmidt 1989; Kamalanathan et al. 2016). Under stress conditions when the normal sink capacities are limited, chlorophyll can be a potential cellular photo-toxin when the absorbed light energy is diverted inappropriately to oxygen, leading to the production of reactive oxygen species (ROS) (Hörtensteiner and Kräutler 2011). The large and rapid drop in RuBisCO content after exiting exponential growth phase (Fig. 3B) is indicative of a lost sink capacity and the risk associated with maintaining the light harvesting machinery. Excess light in the absence of an appropriate sink could lead to an oxidative stress where ROS can damage macromolecules (Apel and Hirt 2004). It’s interesting that our earlier study using a cell wall-less strain showed no decline in chlorophyll levels as cells aged in a more closed culturing system, despite the fact that thylakoid membrane density and chloroplast size declined dramatically (Humby et al. 2013). While this was unexpected, we proposed that chlorophyll may have accumulated in the enlarging oil bodies as cells aged. However, it’s possible that this was an artifact of the cell wall-less strain that was used. These cells lyse shortly after death, perhaps leaving plastid fragments containing chlorophyll that were still harvestable in later steps given the more accelerated senescence timeline.
The drop in the Chl a/b ratio in the death phase (after day 68), and the relatively stable maintenance of LHCII levels implies there is a preferential maintenance of the LHC antenna over reaction centres during conditional senescence (Humby et al. 2013; Meagher et al. 2021). This was also observed in the leaves of senescing Arabidopsis (Nath et al. 2013). The Chl a/b ratio started to decline at about 54 days post inoculation, which corresponded with the drop in Fv/Fm and increase in the Fo/Fm, that remained high throughout stationary and death phases. As Meager et al (2021) argued, this is likely due to the detachment of the antenna from the PSII reaction centre, leading to an increase in the Fo, and subsequent drop in the Fv/Fm. In that study, the increase in the Fo/Fm was only apparent in a HL-stress culture in the DGP, not the low-light culture, indicative of a light-stress response. In this study, the Fo/Fm increase was pushed to the stationary phase under low-light conditions, indicating an indirect-light stress caused by limited sink capacity. Of course, LHCII is a major antenna complex having a dual function; light-harvesting and photo-protection (Natali and Croce 2015). LHCII is able to switch to a quenched conformation in light stress conditions thereby dissipating excess light in the form of heat (Tian et al. 2015). The abundance of LHCII well into the death phase could suggest its importance during conditional senescence, most likely in photo-protection that is triggered by LHCSR.
The ability of Chlamydomonas to deal with excess light changed throughout conditional senescence. When faced with an excess-light challenge, the quantum efficiency of PSII (φPSII) was always close to zero, indicating the reaction centers were fully reduced and all of the incoming light was being dissipated when challenged with excess light. Initially, this energy was dissipated in a manner that was not photoprotective (high φNO) when non-photochemical quenching (φNPQ) capacity was low. It is indicative of cells that are unable to properly protect themselves from excess light (Klughammer and Schreiber 2008), that could invariably lead to photoinhibition. However, φNPQ started to increase at day 11, reaching a maximum at 25 days post inoculation. φNPQ corresponds to the fraction of energy dissipated in the form of heat via regulated, nonphotochemical quenching mechanisms (Kramer et al. 2004). This increase is very likely due to the upregulation of LHCSR at day 11, which is known to be a major player in NPQ in Chlamydomonas (Peers et al. 2009).
The stress-related LHC proteins (collectively LHCSR1 and 3) were notable in their upregulation in response to senescence, peaking shortly after exponential growth. This stress-related protein is upregulated during high-light stress conditions and responsible, in part, for the induction of non-photochemical quenching in Chlamydomonas (Peers et al. 2009). LHCSR also responds to a variety of stresses, such as nutrient limitation, not just light (Toepel et al. 2013), so its induction during senescence is not surprising and previously observed (Humby et al. 2013; Meagher et al. 2021). The induction of LHCSR likely facilitates the conversion of the antenna into light-quenching centers, acting as a photo-protective mechanism for the reaction centers (Girolomoni et al. 2019). LHCSR is able to sense pH variations in the thylakoid lumen and can reversibly switch its conformation from a light-harvesting one to a dissipative one (Peers et al. 2009). This would essentially assist in neutralizing the incoming light energy to minimize the production of ROS under these stress conditions. The transient decline in LHCSR levels at days 68 and 82 was very consistent between replicates, yet a rather curious response. It’s possible that recovered nutrients from dying cells in stationary contributed to the decline, but otherwise, the signal triggering the reduction is unknown. However, the drop in LHCSR does roughly correlate to the transit decline of φNPQ during these timepoints, and the increase in φNO. This reinforces the known link of LHCSR with quenching potential (Peers et al. 2009).
This work highlights strategies used by Chlamydomonas to extend their lifespan under conditional senescence. Following the exit from exponential growth, cells maintain a low-level of cell division for an extended period through recycling macromolecules and reducing protein content and the size of the photosynthetic apparatus. During this time, photoprotective mechanisms are induced, and these remain high throughout the period, to minimize the production of reactive oxygen species. The prolonged maintenance of LHCII over a period of almost 124 days and the induction of LHCSR has an important role in this process. While we proposed a shift in the photo-protective mechanisms involving bulk fluorescence, LHCII could also potentially have a structural role in maintaining the integrity of the thylakoid membranes in ageing cultures, which might be essential for the recovery process when the conditions improve.
An important question is whether there are adaptations in the surviving cells that encourage long-term survival under nutrient depleted conditions. Certainly, sexual reproduction is activated in Chlamydomonas under nutrient stress and the resulting zygote (hypnozygotes or zygospores) is effectively a resting stage resistant to environmental stress (Daniel, Henley, and VanWinkle-Swift 2007; Ellegaard and Ribeiro 2018b). However, there are also non-sexual resting cells that are more comparable to this study. Many microalgae can form resting cells, cysts, or akinetes with a similar environmental stress resistance (Coleman 1983). In the green alga, zygonema, for instance, increasing age of the culture leads to the formation of what are called “pre-akinetes” that are the result of the accumulation of storage products and thickening of the cell wall. In zygonema, there is a metabolic restructuring that pushes metabolism toward production of storage products (Arc et al. 2020), something that also occurs in Chlamydomonas with age and nutrient deprivation (Siaut et al. 2011). We don’t have any clear evidence for formation of stress-resistant resting cells, or its equivalent. But, for instance, we did observe a gradual increase in protein per viable cell beyond 68 days. This may represent a collective partial recovery from scavenged nutrients from dying cells and perhaps these cells are more resistant to the stress condition due to some change in the cell wall or other intracellular structures. However, it is questionable whether this increase in protein is real or an artifact of the cell collection process where protein from dead cells could clearly be included, but it is nevertheless an intriguing question whether Chlamydomonas can produce something akin to a resting cell or akinete, indicating a programmed response.