**1. Introduction**

Since the last century, scleractinian coral reef ecosystems have undergone a decrease in biodiversity and ecological functioning [1–5], formerly attributed to the direct and indirect e ffects of overfishing [6,7], pollution from agriculture, sewage runo ff, and land development [8–10]. Currently, along with the exponential increase of the human population [11] and our societal dependence on carbon fossil fuels, these local threats have been compounded by the impacts of global climate change in the oceans [12–14]. The impact of increasing greenhouse gases in the atmosphere is leading to a global increase in seawater temperatures that has caused mass bleaching events [12,14–17]. These global bleaching events are becoming more frequent (1998, 2010 and 2014–17) and severe [14,16,18–22], leaving coral reefs vulnerable and unable to recover. The 2014–2017 mass bleaching event, which lasted 36 months and spanned four calendar years, was the longest-lasting, most widespread, and probably most damaging event on record [21–29], and stands out as unique by spanning all phases of the El Niño-Southern Oscillation cycle of 2017, being the warmest non-El Niño year ever recorded [21,30].

Coral bleaching is defined as the loss of colour, due to the partial or total loss of Symbiodiniaceae dinoflagellates and/or the reduction of their photosynthetic pigments, that exposes the white calcium carbonate of the coral skeleton (Figure 1A) [31,32]. Bleaching is a generalized stress response to environmental perturbations such as aerial exposure, sedimentation, eutrophication, exposure to heavy metals, high UV radiation, and extreme changes in salinity and temperature [17,31,33,34], however, at large scales is triggered by high seawater temperatures (exceeding normal summer maxima) in combination with high solar radiation [12,15,17,31–33,35]. Scleractinian corals possess molecular protective mechanisms, such as heat shock proteins and antioxidant enzymes to resist thermal stress [17,33,36], or mycosporine amino acids (MAA) and fluorescent pigments to resist light stress (Figure 1B) [17,33,37,38]. The cellular mechanism of bleaching starts with the photoinhibition process within the photosynthetic apparatus of the endosymbionts, which results in the build-up of free electrons that react to form reactive oxygen species (ROS) [39,40]. The proliferation of harmful ROS leads to the degradation, exocytosis, or apoptosis of symbiont cells by the coral host [39], in order to avoid cellular damage [36]. If the duration of the thermal stress extends beyond their physiological ability to recover, corals cannot survive without their main symbiotic partners [15,31,35]. Even though the molecular process of bleaching is similar across coral species, variations in the mechanism to resist and survive thermal stress exist (Figure 1B,C) [17,32,35].

Resilience is the capacity of a coral colony or an entire coral reef ecosystem to absorb, resist, and recover from perturbations [41–43]. The resilience of corals to thermal stress is contingent on the mean long-term annual maximum temperature of the region they live in [17]. Much research has been done in the past decades to understand if the resilience of corals to thermal stress might be an adaptation and/or acclimatisation process (reviewed in [17,44,45]). Here, we review current research that focuses on the capabilities of coral species to adapt and/or acclimatise to thermal events, in order to understand what the future of this irreplaceable ecosystem will be. We have included only scientific studies which have clearly identified the di fferent general strategies to survive thermal stress, as presented in this review.

#### **2. Mechanisms of Resilience to Thermal Stress**

#### *2.1. Thermally Tolerant Endosymbionts*

By associating with stress-resistant symbionts, some coral species are able to acquire increased thermal tolerance. Within the Symbiodiniaceae, species like *Durusdinium* spp. (previously clade D) [46–48], *Cladocopium* C15 [49], and *C. thermophilum* C3 [50,51] are resistant to thermal stress. Dinoflagellates in the genus *Durusdinium* are extremophiles inhabiting environments of thermal stress, high temperature fluctuations, sedimentation and high-latitudinal marginal reefs [52–62]. In recent decades, *Durusdinium* spp. have generated interest because they proliferate in bleached corals [53,60,63–65], protecting against thermal stress by providing 1–1.5 ◦C of thermal tolerance [46]. *Durusdinium* spp. maintain high photochemical e fficiency when exposed to high temperatures compared to symbionts from other genera (*Breviolum* or *Cladocopium)* [48,66,67] and are able to fix more carbon and assimilate more nitrogen [68]. Furthermore, *D. trenchii* has been found to provide tolerance to cold stress too [69,70].

When exposed to thermal stress, some species of coral are capable of shifting the relative abundance of their dominant symbionts. Background symbionts, which can represent <10% of the overall Symbiodiniaceae community [60,71], become dominant, conferring thermal tolerance to the holobiont. Even though many coral species are able to associate with a heterogeneous community of Symbiodiniaceae [72,73], others do not change their dominant symbiont even when bleaching [74], showing a long-term symbiotic adaptation between coral host and dominant symbiont [75–78].

**Figure 1.** (**A**) Onset of a bleaching process in a colony of *Acropora* spp., in Kenting, Taiwan 2015 (Photo: J. Wei); (**B**) Bleaching event showing colonies with fluorescent pigments as a protective mechanism (**a**) and already bleached colonies (**b**), in Okinawa, Japan 2016 (Photo: S.-Y. Yang); (**C**) Intra-specific: between *Montipora* spp. colonies (**a**) and between *Isopora palifera* colonies (**b**), inter-specific: between *Montipora* spp. and *I. palifera* colonies (**c**) and intra-colony: within *Leptoria phrygia* colony (d) responses to thermal stress in Kenting, Taiwan 2016 (Photo: R. Carballo-Bolaños).

#### *2.2. Acclimatisation (Phenotypic Plasticity)*

Phenotypic plasticity refers to dissimilar phenotypes that can be generated from a single genotype in response to different environmental conditions [79]. These phenotypic changes are reversible and dependent on the boundaries of each organism's genotype [17]. In this context, acclimatisation refers to the phenotypic changes of corals in their natural environment, while acclimation denotes short-term phenotypic changes under manipulative experimental conditions in the laboratory. Reciprocal transplantation experiments (RTE) are a well-known method to quantify acclimatisation mechanisms by measuring di fferences in physiological parameters in specimens transplanted across environmentally distinct sites, locations or regions. For example, a RTE of *Porites lobata* between a fore reef (impacted by high wave action, oceanic swells and storms) and back reef (sheltered) in American Samoa, showed phenotypic plasticity in mean annual skeletal extension rates, bulk densities, and calcification rates after only six months, with all variables in transplanted corals approximating values of corals originally from the site [80]. In another study, Sawall et al. [81] found optimal calcification rates at 28–29 ◦C throughout all populations of *Pocillopora verrucosa* with evident di fferences in temperature fluctuations between the northern (21–27 ◦C) and southern (28–33 ◦C) parts of the Red Sea, supporting high phenotypic plasticity due to low genetic divergence between north and south coral host populations.

#### *2.3. Thermal Stress Acclimatisation*

Multiple studies have identified a direct link between thermal preconditioning and bleaching susceptibility (Table 1) [82–91]. After exposing corals to short-term thermal preconditioning experiments, only preconditioned corals did not bleach during a heat-stress experiment (Table 1) [82,83], despite maintaining their Symbiodiniaceae and the bacterial community [82]. Moreover, other studies have compared coral responses of the first major mass bleaching event in 1998 with subsequent stronger bleaching events [89,91]. Maynard et al. [91] surveyed the same sites in 1998 and after a more severe bleaching event in 2002, which featured exposure to twice as many degree heating weeks (DHW) and 15% higher solar irradiance, corals acclimatised, and exhibited less bleaching than in 1998. In a similar study, Guest et al. [89] demonstrated how coral bleaching was less severe after the 2010 large-scale bleaching event in Southeast Asia in locations that previously showed high bleaching in 1998 (Singapore and Malaysia), and had greater historical temperature variability and lower rates of warming. Meanwhile, corals in Indonesia were una ffected by bleaching in 1998, but showed high mortality in 2010. Consequently, corals acclimatised to previous thermal stress events, but also those living in sites with highly variable temperatures presented higher tolerance [89].


**Table 1.** Studies performed to test acclimatisation to thermal stress and high temperature variability at di fferent locations around the world.


**Table 1.** *Cont*.

HSE = Heat Stress Experiment, HV = Highly Variable, MV = Moderately Variable, LV = Low Variable, RT = Reciprocal Transplantation, BE = Bleaching Event, DHW = Degree Heating Weeks, d = days, h = hours, GBR = Great Barrier Reef.

Brown et al. [90] demonstrated 'long-term environmental memory' during the bleaching event in 2010. In 2000, coral colonies were rotated 180◦ in a manipulative experiment [96]. During the bleaching event of 2010, the sides of colonies exposed to high solar radiation before rotation in the 2000 experiment, retained four times as many symbionts than the sides exposed to low solar radiation, despite experiencing higher radiation for 10 years [90]. These experiments provide evidence that long-term acclimatisation to local conditions enhances thermal tolerance during bleaching events (Table 1). Coles et al. [87] showed evidence of acclimatisation to increasing seawater temperatures by replicating a bleaching experiment from 1970 at the same location in 2010. Because sea-surface temperature (SST) has steadily increased 1.13 ◦C over the last four decades, the authors experimentally increased 2.2 ◦C of ambient temperatures. Corals in 2017 showed higher calcification rates, delayed bleaching, and mortality compared to corals in 1970 (Table 1) [87]. Unfortunately, despite increased temperature tolerance in local corals, Hawaii suffered high coral mortality (34%) during the 2014–2017 global bleaching event, showing that high-temperature acclimatisation processes may not be occurring quickly enough to mitigate the projected length and intensity of future bleaching events [87].

#### *2.4. Acclimatisation to High Temperature Variability*

A series of backreef pools exhibiting tidal temperature variability on the island of Ofu, American Samoa, present a unique environment to study physiological differences between conspecific corals at small-spatial scales [97]. Using genetically identical coral fragments in a heat-stress experiment from both pools, Oliver and Palumbi [67] provided evidence of increased thermal tolerance when corals

have acclimatised to high temperature variability (Table 1). Corals from the highly variable (HV) pool showed lower mortality and higher photochemical e fficiency, while those from the moderately variable (MV) pool su ffered increased mortality and lower photochemical e fficiency related to symbiont species. Corals associated with *Durusdinium* spp. exhibited an intermediate decline in photochemical e fficiency, while those associated with *Cladocopium* spp. showed the highest decline [67]. Palumbi et al. [92] performed reciprocal transplantations of corals between HV and MV pools and subjected those corals to a heat stress experiment to test for acclimatisation responses to thermal stress (Table 1). Corals acquired heat sensitivity based on the pool they were transplanted to: MV pool corals acquired heat resistance when moved to HV pool, but not to the same extent of HV conspecifics, while HV to MV transplantees experienced reduced chlorophyll *a* retention, similar to the levels of native corals [92]. Mayfield et al. [93] performed a thermal stress experiment with corals from a site in Taiwan exhibiting high daily temperature fluctuations and found that, under HV conditions, physiological parameters behaved similarly to those in control corals, suggesting that individuals living under HV temperatures can acclimate to high temperatures that would cause bleaching and mortality in unacclimated corals from other regions (Table 1) [93].

Some studies have compiled data of past bleaching events, in an e ffort to link patterns of bleaching susceptibility within sites under high temperature variability, in a worldwide context [94,95]. Sites characterized by a high-frequency pattern of temperature variability experienced higher thermal stress during both bleaching events, with extensive bleaching reported during 1998. However, in 2005–2006, these sites experienced reduced bleaching compared to sites under low frequency patterns, due to the acclimatisation of corals to thermal stress after the 1998 bleaching event and selective adaptation of resilient corals that survived the bleaching event [94]. Safaie et al. [95] explored this concept further by collecting in situ data with remotely sensed datasets from di fferent reef locations around the globe, along with spatiotemporally coincident quantitative coral bleaching observations. Corals regularly exposed to temperature fluctuations on daily or tidal timescales became acclimatised to thermal stress and resistant to bleaching events. More importantly, these patterns of high-frequency temperature variability to bleaching occur in many reefs worldwide [95].

#### *2.5. Molecular Mechanisms for Acclimatisation*

Most studies involving transcriptomic analyses and thermal stress have shown di fferential gene expression under high temperature stress compared to controls (Table 2) [98–103]. Corals exposed to experimental thermal stress presented an upregulation of genes involved in oxidative stress responses [98,99] and carbon metabolism [98]. A comparison of di fferences in gene expression in corals preconditioned to thermal stress showed seventy di fferentially expressed genes between non-preconditioned corals and controls, 42 between preconditioned corals and controls, and nine genes between non-preconditioned and preconditioned corals (Table 2) [102].


**Table 2.** Thermal stress studies looking at differences in gene expression.

29

To understand the genomic basis of thermal resilience in corals, Barshis et al. [103] compared transcriptome-wide gene expression among thermally resilient and thermally sensitive conspecifics (Table 2). Sixty genes were up-regulated in thermally sensitive corals, while resilient corals already presented up-regulated genes under ambient conditions. These "frontloaded" genes facilitate a faster reaction to thermal stress at the protein level [103]. In a similar study, using reciprocally transplanted corals from HV and MV pools, transcriptome-wide gene expression analyses showed differential expression in 74 genes related to heat acclimation between genetically identical corals from both pools (Table 2) [93]. In a related study performed at the same sites, Ruiz-Jones and Palumbi [104] monitored the transcriptomic response of corals in the HV pool with a strong tidal cycle (high temperatures over 17 days). Their results bolstered the conclusions of Barshis et al. [103], showing that genes up-regulated during the hottest days, were enriched for "unfolded protein response", an ancient eukaryotic cellular response to endoplasmic reticulum stress, which corals use as the first line of defence against thermal stress [104].
