Loss and resiliency of social amoeba symbiosis under simulated warming

Abstract Anthropogenic global change is increasingly raising concerns about collapses of symbiotic interactions worldwide. Therefore, understanding how climate change affects symbioses remains a challenge and demands more study. Here, we look at how simulated warming affects the social ameba Dictyostelium discoideum and its relationship with its facultative bacterial symbionts, Paraburkholderia hayleyella and Paraburkholderia agricolaris. We cured and cross‐infected ameba hosts with different symbionts. We found that warming significantly decreased D. discoideum's fitness, and we found no sign of local adaptation in two wild populations. Experimental warming had complex effects on these symbioses with responses determined by both symbiont and host. Neither of these facultative symbionts increases its hosts’ thermal tolerance. The nearly obligate symbiont with a reduced genome, P. hayleyella, actually decreases D. discoideum's thermal tolerance and even causes symbiosis breakdown. Our study shows how facultative symbioses may have complex responses to global change.

The symbiosis between social amebae and certain Paraburkholderia bacterial species is a promising system for gaining insight into how facultative mutualisms respond to global warming. The soil-dwelling ameba Dictyostelium discoideum is a good model to address eukaryote-microbe interactions because of its dynamic relationship with bacteria. In a nutrient-rich environment, D. discoideum lives as independent haploid amebae that reproduce by binary fission. When food is scarce, cAMP-mediated aggregation occurs, leading to the formation of multicellular slugs that move to a favorable location to develop into fruiting bodies. In these fruiting bodies, approximately 20% of the cells die to form a long thin stalk that supports a spherical structure called the sorus, while the remaining 80% ascend into the sorus and turn into spores (Kessin, 2001). D. discoideum is a predator of bacteria and a popular model for studying biological phenomena, including multicellularity, chemical signaling, and social phenomena (Chen et al., 2016;DiSalvo et al., 2015;Ho et al., 2013;Kessin, 2001;Zhang et al., 2016).
In addition to eating bacteria, D. discoideum can also form symbiotic associations with some bacterial species (Brock et al., 2011;DiSalvo et al., 2015;Strassmann & Shu, 2017). About one-third of wild-collected clones of D. discoideum, which are referred to as "primitive farmers," have stable associations with their symbiotic bacteria throughout their life cycle (Brock et al., 2011). These farmer clones can carry bacteria during spore dispersal and seed them as new food sources (Figure 1). Later studies found that farming status is induced by symbiotic bacteria belonging to the genus Paraburkholderia (DiSalvo et al., 2015;Haselkorn et al., 2019;) (named P. agricolaris, P. hayleyella, and P. bonniea ). These Paraburkholderia are not edible themselves, but they facilitate further carriage of food bacteria that on their own would be digested. The inedible symbionts actively find their ameba hosts through chemotaxis, reside within food vacuoles, and form very stable associations ( Figure 1) (Shu et al., 2015;Haselkorn et al., 2019;. Therefore, we also define their association as "bacterial carriage" by social ameba. Both D. discoideum and their Paraburkholderia symbionts can live independently, making them facultative symbioses. However, P. hayleyella shows three indications of being more obligate than P. agricolaris. First, it is a sister species comprising a very long branch in the phylogeny, suggesting that it has been associated with amebas for a long time Haselkorn et al., 2019).
Second, consistent with greater dependence on the host, it grows slowly on its own under laboratory conditions compared to P. agricolaris. P. hayleyella also has greatly reduced carbon usage compared to P. agricolaris . Finally, it shows the genome size reduced by over one half compared to close relatives . This system gives us an opportunity to investigate how increased temperatures associated with global warming could potentially affect facultative symbioses.
Facultative symbioses could be more vulnerable to global warming compared to obligate symbioses because their relationships are less stable. Alternatively, facultative symbioses may be more resilient to global warming because both partners can live on their own and therefore may be more resilient to environmental changes. We will test whether these facultative symbionts help or harm their hosts under warming, and also whether the symbiosis is less or more resilient with the more facultative species P. agricolaris versus the more obligate species P. hayleyella. We first tested the thermal tolerance of social amebas using common garden experiments.

| Choosing experimental temperature for simulating warming
We wanted to choose an experimental temperature that is stressful to social amebae but does not cause complete death. We tested growth conditions of D. discoideum (three clones: QS11, QS70, and QS9) under different temperatures ranging from 21 to 30°C. We found that almost no clone can survive above 28°C, while there were drastic changes between 27 and 28°C ( Figure 2b). Therefore, we chose 27.5°C as the thermal stress temperature for this experiment.
We want to test how extreme warming event (from D. discoideum ameba's perspective) affects the social ameba symbiosis and whether its bacterial symbionts could help.

| Effects of thermal stress on two wild D. discoideum populations
We used two D. discoideum populations from geographic and climate divergent locations Texas (N29°46′, W95°27′; elevation, 11 m; annual temperatures: 5.7-34.7°C; average temperatures: 20.6°C) and Virginia (N37°21′, W80°31′; elevation, 1,160 m; annual temperatures: −15-25°C; average temperatures: 5.2°C) to investigate how D. discoideum responds to simulated thermal stress and whether they could locally adapt to it. We randomly chose 10 Texas clones and 10 Virginia clones of wild D. discoideum and plated those (2 × 10 5 spores) in association with K. pneumoniae (200 µl, OD1.5) on SM/5 plates. F I G U R E 2 (a) Diagram of symbiosis experimental design. The experiment explores how thermal stress affects D. discoideum-Paraburkholderia symbiosis by mixing and matching D. discoideum with two facultative symbionts P. agricolaris and P. hayleyella. (b) Spore count (mean ± 95% CI) of three D. discoideum clones under different temperatures ranging from 21 to 30°C. QS9, naïve host; QS11, native host carrying P. hayleyella B2qs11, and QS70, native host carrying P. agricolaris B1qs11; (c) Spore count (mean ± 95% CI) of two D. discoideum populations (Texas and Virginia) under two temperature treatments (27.5 and 21°C). All tested Texas and Virginia clones are naïve host which do not carry any Paraburkholderia symbionts We incubated these clones at room temperature 21°C (control) and 27.5°C (thermal stress treatment), respectively. We harvested spores from each plate after one week. We flooded the plate with 10 ml KK2 + 0.1%NP-40 and collected spores into 15 ml falcon tubes. We counted spores on a hemocytometer using a light microscope. This design resulted in a total of 2 (populations) × 10 (clones) × 2 (temperatures) × 3 (replicates) = 120 experimental units. The mean of three replicates was used for further statistical analyses.

| Effects of thermal stress on D. discoideum-Paraburkholderia symbiosis
We generated symbiont-free native host clones (QS70C, QS159C, NC21C, QS11C, QS21C, and NC28C) by curing them of their bacteria with tetracycline, or by ampicillin-streptomycin treatment as previously described (Brock et  them out on bacteria-free plates and confirming that the social amebae could not proliferate, a test we call a spot test (Brock et al., 2011).
We mixed and matched (Figure 2)

| P. agricolaris had no effect on D. discoideum's thermal tolerance
When P. agricolaris clones were mixed with their native hosts, thermal stress decreased D. discoideum's fitness, as indicated by the Overall, these results suggest that the more facultative P. agricolaris neither helps nor harms D. discoideum under thermal stress.
In addition, there is no difference between native and naïve hosts.
There was no significant temperature*symbiont interaction (GLM, stress. In addition, 2 out of 3 tested naïve hosts showed zero growth under thermal stress when mixed with P. hayleyella, indicating symbiosis breakdown, while this did not happen in any of the native hosts.
Taken together, these results suggest that adding P. hayleyella, like thermal stress, can decrease D. discoideum's fitness. In addition, it further decreases host fitness under thermal stress. We also found evidence of symbiosis breakdown when P. hayleyella was mixed with naïve hosts, while this does not happen in the native hosts. This indicates potential partner adaptation between P. hayleyella and their native hosts.

| D ISCUSS I ON
Overall, we show that increased temperature affects symbiotic interactions. Increased temperature can significantly decrease D. discoideum's fitness. We found no adaptive divergence to thermal stress in two wild populations. Neither symbiont increased its hosts' F I G U R E 3 Spore counts (mean ± 95% CI) of D. discoideum hosts (with and without P. agricolaris and P. hayleyella) under two temperature treatments ( thermal tolerance. Our study shows that facultative symbioses can also have complex responses to warming. Previous studies found that facultative symbionts provide greater flexibility in response to temperature change compared to obligate symbioses (Burke et al., 2010;Renoz et al., 2019). For example, facultative bacterial symbionts benefit aphids under heat stress (Montllor et al., 2002) and may protect both host and obligate symbiont from thermal stress (Burke et al., 2010). However, in the social ameba symbiosis system, we find no evidence that facultative Paraburkholderia symbionts increase D. discoideum hosts' thermal tolerance.
We find that different symbionts behave differently within the same host under simulated warming, and we also find evidence of host adaptation. Of the two symbionts, the more facultative P. agricolaris has no effects on the thermal tolerance of either native or naïve D. discoideum hosts. On the other hand, the more obligate P. hayleyella induces a significant difference to the host's thermal tolerance, imposing a higher cost to D. discoideum. Our study shows that the addition of P. hayleyella to its native host decreases host fitness at both temperatures indicating that native hosts suffer a fitness cost when they carry P. hayleyella. In addition, P. hayleyella harms and even kills naïve hosts exposed to thermal stress, disrupting the symbiosis. The more severe fitness costs exerted by P. hayleyella colonization in naïve hosts compared to native hosts suggest potential host adaptation between P. hayleyella and their native host clones.
One potential drawback of this study is that we did not monitor the population dynamics of K. pneumoniae and Paraburkholderia symbionts under different temperatures. Simulated warming can directly affect the interactions between food bacteria and symbionts, which in turn affects the growth of amebae. Indeed, a recent study reported that the optimal growth temperature for both Paraburkholderia symbionts is 30°C, and P. agricolaris grows faster than P. hayleyella . Therefore, in this study, both

ACK N OWLED G M ENTS
Many thanks to Mountain Lake Biological Station of the University of Virginia where we collected the samples.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.