Regulatory, Integrative and Comparative Physiology

Comparative physiology: a “crystal ball” for predicting consequences of global change

George N. Somero


Comparative physiology offers powerful approaches for developing causal, mechanistic explanations of shifts in biogeographic patterning occurring in concert with global change. These analyses can identify the cellular loci and intensities of stress-induced perturbation and generate predictions about ecosystem alterations in a changing world. Congeneric species adapted to different abiotic conditions offer excellent study systems for these purposes. Several findings have emerged from such comparative studies: 1) In aquatic and terrestrial habitats, the most heat-tolerant ectotherms may be most threatened by further increases in temperature, due to proximity of these species' thermal optima and tolerance limits to current maximal ambient temperatures and limited capacities for acclimatization to higher temperatures. 2) Cardiac function is a “weak link” in acute thermal tolerance. 3) Stress-induced changes in gene expression comprise a graded response involving genes linked to damage repair, lysis of irreversibly damaged molecules, and downregulation of cell proliferation. Transcriptomic and proteomic analyses provide “biomarkers” for diagnosing degrees of stress. 4) Different abiotic stresses may have synergistic or opposing effects on gene expression, a complexity needing consideration when developing integrated pictures of effects of global change. 5) Adaptation of proteins can result from one to a few amino acid substitutions, which can occur at many sites in a protein, a discovery with implications for rates of adaptive evolution. 6) Greater thermal tolerance of invasive species may favor their replacement of natives. 7) Losses of protein-coding genes and temperature-responsive gene regulatory abilities in stenothermal ectotherms of the Southern Ocean may lead to broad extinctions.

  • adaptation
  • climate change
  • gene expression
  • invasive species
  • transcriptomics

comparative physiology has a long and distinguished history, which reaches back to the foundational contributions of scholars like Claude Bernard, Christian Bohr, and August Krogh. Knut Schmidt-Nielsen (59) may have provided the most succinct statement of the overarching goals of our field, which are to elucidate the fundamental mechanisms that explain “how animals work” and to reveal how the “workings” of these physiological systems have evolved to adapt organisms to a wide range of environmental conditions (31). These two contributions of our field are coming to be increasingly appreciated as offering a relevant means for understanding how global change is likely to affect different species and, thereby, the structures and interactions of the ecosystems of which they are members. Through understanding the underlying mechanistic bases of sublethal and lethal stress and the differences that exist among species in capacities to respond to these stresses, a foundation can be constructed for developing predictions about the probable success or failure of organisms to cope with changes in physiologically significant abiotic factors like temperature, salinity, oxygen availability, and pH, all of which are aspects of global change.

Comprehensive analyses designed to provide a realistic assessment of the effects of global change require examination of a broad set of phenomena, ranging from studies of behavioral regulation of body temperature to molecular level processes, such as the gene regulatory events that comprise the cellular stress response (CSR) (39) and modulate many of the processes that underlie adaptive phenotypic plasticity. In this review, I take a reductionist approach and focus largely on cellular- and molecular-level processes, primarily those associated with the CSR, which have received less discussion in the context of global change than have processes at higher levels of biological organization (20, 24, 5254). The primary rationale for this presentational focus, however, is based on a wish to provide the most general treatment possible of how cellular- and molecular-level aspects of physiology play out in the arena of global change. Thus, a focus on the widely occurring CSR is appropriate. Here, “wide” refers to two very different aspects of the CSR. First, many different types of cellular stress, including stress from temperature, oxygen availability [and production of reactive oxygen species (ROS)], osmolality, and acidity, lead to common types of cellular damage and, therefore, trigger a common set of stress-related responses, the CSR (39). If successful, the CSR can sufficiently offset the damaging effects of stress to allow other processes to restore cellular homeostasis (the cellular homeostasis response, or CHR) (39). The second “wide” aspect of the CSR relates to its breadth of occurrence among organisms. The CSR is observed across all taxa (39); it is a fundamental property of almost all cells that has been strongly conserved during evolution. However, as discussed in the context of extreme stenotherms, long-term evolution in stable habitats may lead to loss of key elements of the CSR, which places these organisms in extreme jeopardy from global change (10).

I hope that this review will serve to complement other recent analyses of physiological aspects of global change biology (20, 24, 5254, 68) and help to further establish a pivotal role of reductionist, mechanistic analysis in guiding our understanding of one of the greatest challenges ever to confront our planet's ecosystems.

Congeners: A Kroghian “Crystal Ball” for Studying Adaptive Physiological Variation and Global Change

As is commonly the case in comparative physiology, analyses of the physiological aspects of global change biology benefit from adherence to the well-known August Krogh Principle, here as stated succinctly by Sir Hans Krebs: “For many problems there is an animal on which it can be most conveniently studied.” (38). Comparative analyses of environmental sensitivities and capacities for erecting adaptive responses, either through genetic change or by altered use of existing genetic information, have been especially effective when closely related congeneric species adapted to different abiotic environmental conditions have been examined. Congeners afford several key experimental advantages (31). They allow adaptive variation to be clearly distinguished from the effects of phylogeny per se. They provide insights into how much environmental change is sufficient to favor adaptive change in diverse physiological systems. Congeners' relatively low levels of evolutionary divergence facilitate analysis of how much genetic variation is needed to alter environmental optima and tolerance limits, including variation in gene regulatory capacities and protein amino acid sequence. The existence of suites of congeners with different environmental optima enables predictions to be made about local extinctions of species and their potential replacement with a congener with different and more appropriate environmental tolerances. In some cases, this replacement may represent the introduction of an invasive congener, with potentially large effects on the ecosystem overall. Thus, comparative studies of congeners have enabled an integrated analysis that spans all levels of biological organization, from the population to the level of fine-scale adaptive molecular variation.

Below, I illustrate the power of such comparative analyses to shed light on several key questions related to the impacts of global change on the biosphere. Studies performed with the comparative approach, including experiments that exploit new “omics” technology imported from biomedical science, can indeed provide biologists (and policy makers as well) with an effective “crystal ball” for predicting the effects of global change on individual organisms and ecosystems.

Biogeographic Patterning Correlated with Global Change: A Challenge for a Cause-Effect Physiological Analysis

Changes in biogeographic patterning, putatively caused by global change, have been documented in dozens of studies during the past two decades (for reviews, see Refs. 48, 57, and 76). These pronounced changes in organisms' latitudinal and vertical distribution patterns are widespread in both terrestrial and marine environments and involve a wide array of taxa, both ectotherms and endotherms. In terrestrial ecosystems, range expansions or contractions of 6–25 km per decade are common (48). Recent studies of lizards have revealed dramatic shifts in distribution and extremely high probabilities of extinction (35, 63). For example, it is estimated that by 2080, local extinctions of lizards could reach 39% worldwide, and species extinctions may reach 20% (63). In marine rocky intertidal habitats, which figure importantly in the discussion to follow, especially rapid shifts have been observed, between 49 and 540 km per decade (37). For one dominant rocky intertidal species in the Western Atlantic, the mussel Mytilus edulis, the southern range has contracted by ∼350 km over the past 47 years (7.5 km per year) as air- and water temperatures have risen (37). In the Eastern Pacific, the rapid changes observed in the biogeographic limits of native and invasive blue mussels, Mytilus trossulus and Mytilus galloprovincialis, respectively, are correlated with shifts in ocean temperature (30). Fortunately for investigators attempting to explain such trends at the physiological level, the genus Mytilus has been a popular study system, and it affords excellent insights into causal mechanisms behind biogeographic change, as illustrated later in this review. Many biogeographic changes also have been found in pelagic marine habitats. Ranges of North Atlantic plankton species have shifted northward by 10° of latitude since the 1960s (2). In the North Sea, 87% of demersal fish species studied extended their northern range limits and 50% of northern species either contracted their southern distribution limits or migrated to greater depths (49).

These and many other examples of shifts in biogeographic ranges in a changing world challenge comparative physiologists to provide a mechanistic explanation that can help to account for these observed changes and serve as a strong foundation for developing predictions about future shifts in distribution patterns and ecosystem structure. Such explanatory and predictive analyses need to examine several related questions. First, what are the physiological mechanisms that are responsible for acute lethality, as might occur during periods of one to a few days of extreme heating? Second, what types of physiological effects of sublethal thermal stress might so undermine an organism's physiological condition as to preclude its continued existence in a habitat? Do the effects of sublethal stress occur in a graded manner that physiological methods can identify, thereby enabling physiologists to generate a useful suite of diagnostic “biomarkers” for assessing the level of stress an organism is confronting? Third, how do species differ in potential for acclimatizing to global change? Do some species have more complete genetic “tool kits” for effecting these phenotypic responses? Fourth, what types of changes in the genome—in both protein-coding genes and gene regulatory elements—are needed to evolve enhanced physiological tolerance to abiotic stress? Do these evolutionary changes in the genome appear feasible, in light of the rapidity of global change?

Organisms Differ in Susceptibility to Global Warming: Patterns and Causal Mechanisms

The primary focus of this review is on the effects of rising temperatures (“global warming”), although the interactions of temperature with other abiotic factors likely to vary with global change, notably salinity and oxygen concentration, will be considered briefly as well. Emphasis will be placed primarily on ectothermic species, which have received the greatest amount of attention by biogeographers and physiologists interested in effects of global warming. However, it's pertinent to note that recent analyses and modeling efforts predict severe effects of global warming on endothermic homeotherms, mammals, and birds. Thus, for mammals, including many human populations, an increasing fraction of the globe is predicted to become inimical to habitation as the planet continues to warm (45, 61). It is estimated that when wet-bulb temperatures reach or exceed 35°C, many low- to mid-latitude regions that currently support a large fraction of the human population will no longer be hospitable to humans and other mammals.

Studies of ectotherms also have pointed to high probabilities of risk of extinction at low latitudes (14, 7275), and also to varying threats from warming among congeners from a single latitude that have different vertical zonation patterns (74). Latitudinal and vertical comparisons have led to an important generalization: the most warm-adapted (heat-tolerant) species may be in greatest danger from further increases in environmental temperature. This seemingly paradoxical conclusion has been reached in broad surveys of terrestrial species from different latitudes (14, 75) and in analyses of congeners of marine invertebrates found at different vertical positions in the intertidal zone and at different latitudes (7174). The surveys of terrestrial species point to a much higher susceptibility to extinction in tropical regions than in higher-latitude habitats because tropical ectotherms currently live much closer to temperatures that exceed their thermal optima than related species from higher latitudes. Warm-adapted species also may have lesser abilities to acclimate to increases in temperature, relative to more cold-adapted species. Furthermore, because warm-adapted species may have higher metabolic rates, thermal acceleration of metabolism by rising temperatures has a higher absolute effect on rates of metabolic turnover than in cold-adapted species with intrinsically lower rates of metabolism (16). Some implications of these differential Q10 effects are discussed later.

These same conclusions about relative susceptibilities of different species to global warming have been reached in studies of congeneric marine arthropods and mollusks, where the underlying mechanistic bases of these differential thermal sensitivities have been examined in some detail. The most extensive data set on thermal tolerance differences among congeners has come from studies of marine porcelain crabs (genus Petrolisthes) from subtidal and intertidal habitats at different latitudes (7274). Congeners of Petrolisthes manifest a classical pattern of adaptive variation in heat tolerance. Congeners from low-latitude (tropical) habitats and those occurring at higher sites along the subtidal-to-intertidal vertical gradient at a single latitude exhibit the greatest tolerance of high temperatures. However, the most heat-tolerant porcelain crabs are more likely than their less heat-tolerant congeners to experience maximal habitat temperatures that equal or exceed their thermal tolerance limits. In addition, the most warm-adapted congeners exhibited the least ability to increase heat tolerance through acclimation (73).

Studies of marine turban snails of the genus Chlorostoma (formerly Tegula) show additional aspects of the types of variation in thermal responses that occur among congeners with different distribution patterns (Table 1). Three congeners of Chlorostoma are common in rocky shore habitats along the midlatitude coastline of the Eastern Pacific Ocean (77, 79). The black turban snail (C. funebralis) is found in the low- to mid-intertidal zone from Vancouver Island, Canada (48°N) to central Baja California, Mexico (28°N). Chlorostoma brunnea occurs in subtidal to low-intertidal zones from Cape Arago, OR (43°N) to the Santa Barbara Channel Islands (34°N). Chlorostoma montereyi occurs almost exclusively in subtidal habitats and ranges from Sonoma County, CA (38°N) to the Santa Barbara Channel Islands. The three congeners differ significantly in upper lethal temperature (64, 71, 77), in accord with their different vertical positions and biogeographic ranges; C. funebralis withstands body temperatures ∼5–6°C higher than the two lower-occurring congeners.

View this table:
Table 1.

CTmax and CTmin (°C) of hearts of congeners of Chlorostoma: effects of acclimation and acclimatization

Studies of cardiac function at different temperatures have provided evidence for one proximate cause of these differences in thermal tolerance limits and distribution ranges. As in the case of porcelain crabs (72), congeners of Chlorostoma differ substantially in tolerance of cardiac function to both high and low extremes of temperature (71). Warm- and cold-adapted congeners of these genera also differ in the proximity of contemporary body temperatures to temperatures at which heart failure occurs and in the capacities for increasing heat tolerance of heart function through warm acclimation. These relationships are illustrated for congeners of Chlorostoma in Table 1.

Considering first the thermal tolerances of heart function in field-acclimatized specimens, there is a 6–7°C difference among species in CTmax, the temperature at which a sharp break (change in sign of the slope) occurs in an Arrhenius plot [ln heart rate in beats per minute (bpm) vs. reciprocal of temperature in Kelvins]. For the two lower occurring congeners, the CTmax occurs at temperatures well above the maximal temperatures that the species are likely to encounter in their low-intertidal and subtidal habitats, ∼22°C (79). However, for C. funebralis, which has a CTmax near 31°C in field-acclimatized specimens, field body temperatures as high as 34.5°C have been reported (78), and temperatures between 27°C and 33°C are common during low tides (79). Thus, unlike its two lower-occurring congeners, in its normal habitat, C. funebralis faces challenges to cardiac function because of a rapid fall in heart rate above CTmax.

Acclimation studies revealed additional differences between C. funebralis and the two lower-occurring congeners. Whereas all three congeners exhibited an ability to increase CTmax when acclimation temperature was increased from 14°C (a common ambient water temperature) to 22°C (the highest temperature recorded for C. brunnea), the increase in CTmax differed significantly among species. C. funebralis was able to increase its CTmax by only 1.6°C, but C. brunnea and C. montereyi increased CTmax by 6.6°C and 4°C, respectively (71). Thus, as in the case of congeners of Petrolisthes, the most warm-adapted species not only have CTmax values close to current peak body temperatures, but also have a relatively limited ability to further increase CTmax through acclimation. These findings on two phylogenetically distantly related sets of congeners suggest an important conclusion: acclimation (acclimatization) to rising temperatures is not likely to ameliorate substantially the effects of global warming in those species most threatened by further increase in body temperature.

Further aspects of the studies of congeners of Chlorostoma merit comment. First, as shown in Table 1, the field-acclimatized CTmax values for the two lower-occurring congeners are more similar to the 22°C values than the 14°C (= ambient water temperature) values. This suggests that even limited-duration exposures during low tides led to an acclimatization response that resembles the warm-acclimation response. Second, as in the case of porcelain crabs (72), the low- to mid-intertidal species has a greater tolerance of low as well as high temperatures; it is more eurythermal than its lower-occurring relatives (Table 1). This difference likely reflects the colder temperatures experienced during prolonged emersion in winter. Third, as shown by Tomanek and Sanford (78), levels of heat-shock protein 70 (HSP70) in the lower-occurring congeners are significantly lower than those in C. funebralis in field-acclimatized specimens. This finding suggests that field body temperatures for the most heat tolerant of the three congeners exceed the threshold induction temperature for genes encoding HSP70 but that this may not be the case for C. brunnea or C. montereyi. This conclusion is supported by metabolic labeling studies with the three species, which showed induction temperatures for HSP70 synthesis near 24°C for the two lower-occurring congeners, but near 27°C for C. funebralis (79). The metabolic labeling studies showed significant interspecific differences in the thermal optima and maxima for protein synthesis as well. Incorporation of 35S methionine + cysteine into newly synthesized proteins peaked near 27°C in 13°C-acclimated C. brunnea and C. montereyi and near 33°C in C. funebralis (79). Cessation of protein synthesis occurred near 33°C in the two lower-occurring congeners and near 38°C in C. funebralis. Thus, at field temperatures commonly encountered by C. funebralis, neither of the other congeners would be capable of synthesizing proteins or sustaining optimal cardiac function. In fact, when heated at rates that simulate the rise in body temperature that occurs during emersion (∼4°C/h), the two lower-occurring congeners died near 33°C. Nonetheless, because of their occurrence in habitats where temperatures rarely exceed ∼20°C, C. brunnea and C. montereyi are much less threatened by elevated temperatures than C. funebralis.

A final difference worth noting between C. funebralis and the other two congeners is a difference in heart rate (bpm) at a common temperature of measurement (13°C) (71). Hearts of C. funebralis, C. brunnea, and C. montereyi of similar body mass had heart rates of 6.8 ± 1.3 bpm, 2.0 ± 0.3 bpm, and 3.6 ± 0.4 bpm, respectively. Because rates of oxygen consumption in marine mollusks correlate with heart rate (58), the higher heart rate of C. funebralis has been conjectured to reflect higher energy demands, relative to those of its congeners, resulting from the abiotic conditions it encounters in its low- to mid-intertidal habitat (64). And, as emphasized by Dillon et al. (16), because effects of temperature on metabolism are exponential, rising temperatures will have a greater absolute effect on metabolic rates in species like C. funebralis that already have relatively high rates of metabolism.

Differences in cardiac responses to change in temperature may also play a role in governing the success of biological invasions. In the case of native (Mytilus trossulus) and invasive (Mytilus galloprovincialis) blue mussels on the coast of California, the CTmax of heart function is 23.70 ± 0.80°C for the native species and 28.30 ± 1.08°C for the invasive (specimens common-gardened acclimated to 14°C and 28 ppt salinity) (5). Acclimation to 21°C led to increases in CTmax to 26.00 ± 0.59°C and 30.70 ± 1.04°C in the native and invasive species, respectively. Difference in the heat tolerance of cardiac function is but one of the physiological differences between these two blue mussel congeners that may help to explain the marked success of the invasive in replacing the native species over the southern portion of its former biogeographic range (5). The competitive success of the invasive M. galloprovincialis at high temperatures (60) is an example of what has been proposed to be a common pattern wherein elevated temperatures enhance species invasions due to a suite of physiological and ecological factors (69, 70).

A species of blue mussel native to the East Coast of the United States, M edulis, exhibited a sensitivity of cardiac function to heat stress that was intermediate between those of the two California blue mussels (5). For 14°C-acclimated M. edulis, CTmax was 25.50 ± 0.99°C; for 21°C-acclimated specimens, CTmax rose to 28.50 ± 0.51°C. The high rates of mortality of M. edulis observed in the field following exposures to temperatures of 32°C and the associated rapid contraction of the southern distribution limit of this species' biogeographic range as air and water temperatures have increased during the past few decades (37), may find at least a partial explanation in the thermal sensitivity of this species' cardiac function, which, even in warm-acclimated specimens, is compromised at temperatures above ∼28–29°C (5).

Another significant difference in cardiac performance among the three congeners of blue mussels is in the resting heart rates of the species. For M. trossulus, M. edulis, and M. galloprovincialis, resting heart rates (bpm) at 14°C were 25.91 ± 1.17, 21.30 ± 0.87, and 17.53 ± 0.85 bpm, respectively (5). These differences in heart rate may reflect a temperature-compensatory adaptation in that rates are highest in the most cold-adapted species and lowest in the most warm-adapted congener, the invasive. Activities of enzymes associated with aerobic and anaerobic supply of ATP are higher in tissues of M. trossulus than in M. galloprovincialis (41), consistent with temperature compensation. The higher intrinsic metabolism of the native species may impose challenges at high temperatures if metabolism rises to values that cannot be supported by substrate supply or end-product removal. Also, because of the exponential rise in physiological rates with increasing temperature, thermal acceleration of metabolism will be greater in an absolute sense for M. trossulus because of its higher intrinsic metabolic rate relative to its congeners. Suffice it to say, cardiac function and metabolic capacities of differently thermally adapted congeners may play important roles in establishing environmental optima, tolerance limits, and, thereby, biogeographic patterning. The findings that recent cooling of the Eastern Pacific has been accompanied by a southward shift in the northern distribution limit of M. galloprovincialis (30) and that warming in the Western Atlantic has led to a northward shift in the southern limit of M. edulis (37) provide support to this conjecture.

Sublethal Thermal Stress and Its Consequences for Aerobic Metabolic Processes

The most prevalent threats from global change, especially from rising temperatures, are likely to be sublethal effects that, while not proving lethal to the individual in the short run, as acute cardiac failure might be, nonetheless significantly compromise the organism's performance in such critical processes as locomotory activity, growth, and reproduction. These compromised abilities to sustain a healthy population at a site may lead to local extinctions and to latitudinal range shifts.

Shortfalls in oxygen delivery have been emphasized as a critical element in sublethal thermal stress (20, 5254), as well as in the context of acute death at thermal extremes due to cardiac failure, as discussed above. There is a wealth of evidence showing that reductions in oxygen delivery capacity at thermal extremes may so impair an animal's aerobic scope (difference between minimal and maximal rates of oxygen consumption) as to reduce its chances for persistence in a habitat (5254). Reduced locomotory performance, which is key in predator-prey interactions, may be a major physiological consequence of heat stress. And, as in the case of acute cardiac failure that occurs when temperatures exceed the CTmax of heart function, species may differ in terms of the proximity of current habitat temperatures to the temperatures at which aerobic scope is compromised. In both instances, tropical species appear to be more threatened by global warming than species from higher latitudes. For example, Nilsson and colleagues (47) reported that tropical cardinal fishes lost almost 50% of their aerobic scope when water temperatures were raised only 2°C above current summer average temperatures. Similarly, tropical terrestrial ectotherms may be living closer to their thermal limits for optimal metabolic performance than species from higher latitudes (14, 75).

In the broad context of temperature-induced changes in metabolic rate and in organisms' capacities for sustaining adequate aerobic production of ATP for different physiological activities, it is important to consider the fact that allocation of available ATP production among different processes may be temperature dependent. Temperature changes thus affect metabolism both quantitatively (Q10 effects on total rates) and qualitatively (relative energy needs of different processes). For example, if cellular damage at higher temperatures increases needs for aerobic ATP turnover, the combination of reduced oxygen solubility for aquatic species (solubility falls by ∼2% per degree Celsius rise in temperature), reduced potential for oxygen delivery, and rising oxygen demands for processes linked to the CSR may confront an organism with a combination of problems that lead to suboptimal physiological performance and, at the extreme, to local extinction of a species. Furthermore, increased oxygen consumption is likely to elevate production of reactive oxygen species (ROS), which are damaging to the cell and lead to rising energy demands for repair or replacement of damaged cellular constituents like proteins and removal of irreversibly damaged cells through apoptosis (39). Elevated temperature per se may also increase ROS production, independently of effects arising from increased respiration (1). Rising temperatures thus confront ectotherms with a complex set of physiological challenges that may demand a large modification in the flow of cellular energy among competing processes linked to damage-repair, growth, cell proliferation, and reproduction.

There are many studies at the whole organism level that suggest this sort of redistribution of cellular energy. Some of these effects have been observed with only modest increases in temperature. For example, Munday et al. (46) found that growth of a damselfish was significantly reduced by a 3°C rise in temperature. The classic studies of sockeye salmon by Brett (6) show how strongly temperature, food availability, and growth are inter-related. For mussels (Mytilus californianus) in rocky intertidal habitats, abiotic stress has been shown to alter the allocation of metabolic resources, with conspecifics occurring high in the intertidal zone having relatively less energy to support reproduction (50). Below, I consider what recent studies of changes in gene expression are revealing about the cellular and subcellular processes—and the potential energy costs—associated with different intensities of sublethal stress.

Sublethal Stress at the Cellular Level: A Graded Series of Repair, Demolition, and Reconstruction Functions

Physiologists studying environmental stress at the cellular level are beginning to elucidate how the intensity of stress confronting an organism affects the types of compensatory responses made by the cell and how these responses might affect the cell's overall energy budget and its allocation of energy for different processes. Important new insights into these issues are emerging from two types of “omic” experimentation: transcriptomic studies of gene expression, which employ DNA microarrays (“gene chips”) to survey temperature-dependent transcriptional activities of hundreds to thousands of different genes (9, 25, 26, 40, 43, 44), and proteomic studies, which provide a more direct picture of changes in composition of the cellular protein pool (proteome) and the states of post-translational modification of the proteome (80). These studies are revealing new aspects of temporal patterning of transcriptional and translational changes, respectively, during exposures to increasingly stressful temperatures. Studies of differently temperature-adapted congeners are beginning to elucidate the types of adaptive changes in gene expression that distinguish transcriptomic (40) and proteomic (80) responses of species adapted to different temperatures.

Here, I review some of the relevant recent findings from microarray studies that shed light on the following questions: 1) When does rising temperature first elicit expression of genes associated with the CSR? 2) How do the recruitment patterns of different types of CSR-related genes change as the intensity of thermal stress increases? 3) When do repair processes give way to “demolition” events such as proteolysis? 4) When is cell proliferation inhibited to allow repair of DNA and protein damage to be achieved and to maximize the amount of energy that remains available for ATP-demanding processes of repair and demolition? 5) How do the transcriptional responses of congeners evolutionarily adapted to different temperature conditions vary in the face of thermal stress? 6) Is wholesale modification of gene expression needed, or are only a small number of gene regulatory events different between cold- and warm-adapted species? 7) Are transcriptional responses to temperature indicative of the responses to other forms of abiotic stress such as changes in salinity? 8) Does simultaneous exposure to two or more forms of abiotic stress lead to synergistic or antagonistic shifts in transcriptional responses?

Studies of gene expression in the eurythermal goby fish Gillichthys mirabilis (9, 43, 44) have shown that rising temperatures induce shifts in expression of hundreds of different genes that encode proteins involved in protein refolding, proteolysis, biosynthesis, cell signaling, and cell proliferation. For the purposes of this review, the most relevant of these changes in gene expression are those that 1) provide insights into the threshold temperatures at which heat stress is first evident, 2) reveal a graded response to temperature in which different categories of genes are expressed as heat stress increases, and 3) offer insights into the types of changes in energy allocation strategies that might be necessitated by a rising need for ATP-costly stress responses, which may demand suppression of processes like cell proliferation. Some of the transcriptional changes associated with the CSR in G. mirabilis are tissue specific (9). The discussion below focuses on gill tissue, which shows a robust transcriptomic change under heat stress and salinity stress as well (18).

The threshold temperatures at which genes associated with the CSR first show increased expression differ among genes associated with different elements of the stress response and with acclimation histories of specimens (8, 15, 44). For fish acclimated to 9°C, 19°C, and 28°C, threshold temperatures for induction of stress-related genes increased ∼2°C for each 9–10°C rise in acclimation temperature (44). Although many genes change expression as part of the CSR, increased transcription of genes encoding HSPs is commonly regarded as the canonical indicator of onset of heat stress (21). However, HSPs comprise several protein families that, in addition to their protein-repair activities, play multiple roles in such critical stress-related processes as regulation of programmed cell death (apoptosis) and cell proliferation (3). Thus, it is essential to examine differences in expression among classes of HSPs to elucidate how the diverse cellular processes linked to these proteins are affected as thermal stress increases. Different types of HSPs potentially can serve as “biomarkers” of different degrees of stress-induced cellular damage.

In gills of G. mirabilis, different classes of heat-shock proteins show different temperatures for onset of increased expression and different levels of upregulation (Fig. 1), consistent with the heterogeneity in function of this broad set of proteins (44). As shown in Fig. 1A, genes encoding HSP70 and HSPA9, the two genes showing the highest levels of upregulation in heat-stressed fish, differed in their temperatures of increased expression. In 9°, 19°, and 28°-acclimated fish, increased expression of HSPA9 occurred at temperatures 4°C higher than those at which increased expression of HSP70 was observed. The sequence of activation of these two genes reflects a broader pattern of sequential upregulation of genes involved in repair events and higher-level control of cellular proliferation and degradation. Thus, upregulation of HSP70 is taken as an indication that thermal stress has reached an intensity at which damage to cellular proteins can no longer be reversed by constitutively expressed molecular chaperones, thereby necessitating production of HSPs for adequate protein-refolding capacity. HSPA9, encoded by the second-most upregulated gene, is a mitochondrial paralog of the 70-kDa class of molecular chaperones. It performs a number of functions in addition to molecular chaperoning. HSPA9 is involved in the inhibition of two cellular processes that are of pivotal importance in cellular stress and restoration of cellular homeostasis: cell proliferation and apoptosis. Stress-related regulation of these two processes appears to be necessitated only at slightly higher temperatures than those at which repair of heat-damaged proteins is initiated.

Fig. 1.

Transcriptional responses (mRNA expression) of genes encoding two heat-shock proteins, HSP70 and HSPA9, in gill tissue of Gillichthys mirabilis acclimated for 4 wk to three temperatures (9°C, 19°C, and 28°C) and then subjected to a heat ramp of 4°C/h (44). mRNA expression is normalized to a reference pool of mRNA, according to methods in Logan and Somero (43, 44). Solid bars and asterisks indicate a statistically significant increase in mRNA relative to values in prestressed specimens. [Modified after Logan and Somero (44).]

Another illustration of the variation in induction temperatures among genes associated with different components of the CSR is shown in Fig. 2. Here, genes associated with protein folding (HSP70), ubiquitin-based proteasomal degradation of proteins (UBIQ), and cell cycle arrest/apoptosis (CDKN1B) are seen to show differences in onset temperatures of increased expression (44). Increased expression of genes encoding proteins involved in ubiquitin-mediated proteasomal degradation of proteins is indicative of an incomplete “rescue” of heat-damaged proteins by molecular chaperones like HSP70. Significantly increased expression of the UBIQ gene occurred at 36°C, a temperature 7°C greater than the onset temperature of increased expression of HSP70. Thus, onset of protein “rescue” may ensue well before the cell reaches a temperature at which protein degradation is activated. Increased expression of UBIQ was found in studies of steady-state acclimated G. mirabilis; levels of expression were directly correlated with acclimation temperature (43). Thus, whereas expression of HSP70 did not vary with acclimation temperature, genes linked to proteolysis were increasingly upregulated as acclimation temperature increased. This observation speaks to the issue of energy costs at different temperatures, specifically the temperature-dependence of costs of protein turnover and homeostasis (43).

Fig. 2.

Transcriptional responses of genes encoding proteins indicative of three aspects of the cellular stress response: protein folding (HSP70), proteolysis (ubiquitin-based proteasomal degradation of proteins, UBIQ), and cell cycle arrest/apoptosis (CDKN1B), in gills of heat-stressed G. mirabilis acclimated to 9°C. mRNA expression is normalized to a reference pool of mRNA, according to methods in Logan and Somero (44). Solid bars and asterisks indicate a statistically significant increase in mRNA relative to values in prestressed specimens. [Modified after Logan and Somero (44).]

Increased expression of the gene encoding cyclin-dependent kinase inhibitor 1B (CDKN1B) is associated with late-stage inhibition of the cell cycle (11). The upregulation of this gene may be an indication that cell proliferation is blocked after a certain level of cellular stress has been reached, thereby allowing increased allocation of ATP to repair processes and preventing cells with damage to DNA from proliferating.

Changed expression of genes associated with either inhibitory or stimulatory effects on apoptosis was also observed in this study. Whereas CDKN1B has been linked to induction of apoptosis, several heat-shock proteins have been shown to inhibit this process (3). Experiments that directly measure apoptotic activity, e.g., measurements of DNA damage and activation of caspase systems, are thus needed to resolve the issue of how the balance between proapoptotic and antiapoptotic factors changes with temperature stress. The lack of an unambiguous signal in these gene expression data that apoptosis is increased following heat stress may reflect the time course over which apoptosis occurs in heat-stressed cells. Induction of apoptosis may be delayed until the cell has had adequate time to try to repair stress-induced damage, especially to DNA. And, if this repair cannot be achieved, there may be a further delay in onset of apoptosis until sufficient energy is available to drive the energy-demanding events associated with programmed cell death. The rapid upregulation of genes encoding proteins involved in ATP-synthesizing pathways following heat-stress of G. mirabilis (9) suggests that increased ATP production and turnover may be an important element in the stress response.

Microarray studies of congeners of Mytilus have provided further information on temperature-dependent patterning of gene expression and initial insights into the ways in which these gene regulatory responses evolve in congeners from different abiotic environments. Comparisons of the transcriptional (40) and proteomic (80) responses of M. trossulus and M. galloprovincialis have shown that the different evolutionary histories of these congeners have led to a number of species-specific gene regulatory responses, although most stress-induced transcriptional changes were similar in the two species. Numerous genes encoding HSPs showed the canonical response to rising temperature seen in other studies of sublethal heat stress (Fig. 3A). Above a certain threshold temperature (∼24°C), production of message for commonly occurring HSPs like HSP70 rose sharply (Fig. 3A). For HSP70, there was no significant difference in response between the congeners, an observation that applies as well for almost all of the HSPs that were detected (Fig. 3B). In fact, the only large difference in expression between the native and invasive species was found for HSP24, a member of the α-crystallin family of small heat shock proteins.

Fig. 3.

Gene expression in gill tissue of blue mussels, Mytilus trossulus and Mytilus galloprovincialis. 13°C-acclimated specimens of both species were subjected to a thermal ramp, and sampling was done at 24°C, 28°C, and 32°C. A: cross-plot of expression levels of all genes encoding molecular chaperones. Except for small heat-shock proteins (HSP24), the transcriptional changes of chaperones were similar in the two species. B: transcription of the genes encoding HSP70s. mRNA for HSP70 showed a similar response to heating in both species, as indicated in Fig. 3A. C: transcription of genes encoding proteins of the proteasomal apparatus. Induction occurred at a lower temperature in M. trossulus. [Modified after Lockwood et al. (40).]

The similarities, as well as the single difference, in expression of HSPs between the congeners raise two issues. First, the similarities in responses of the two species for most HSPs suggest that evolution of gene regulatory capacities for activating the heat-shock response has not been of a wholesale nature, but rather has been restricted to a subset of HSP-encoding genes. Second, the strong upregulation of message for HSP24 seen in the more warm-adapted Mediterranean mussel provides insights into the nature of thermal stress and its downstream consequences in cellular regulation. Induction of high levels of expression of HSP24 may be an indication that the cytoskeleton is of particular sensitivity to heat stress, because this small heat-shock protein is known to be involved in refolding of cytoskeletal proteins (40). In fact, this conclusion has been reached in other studies of gene expression under heat stress and osmotic stress, which have shown that proteins associated with the cytoskeleton appear especially prone to damage from heat- (9) and osmotic shock (18). In addition, small heat-shock proteins are believed to play a role in suppressing apoptosis (3). Thus, an alternative (or additional) explanation for the rise in HSP24 message during heat stress is that apoptosis is blocked or delayed during the period of heat stress, as discussed earlier for G. mirabilis. If this is the case, M. galloprovincialis appears to be more effective than M. trossulus in inhibiting programmed cell death under heat stress.

Transcriptional studies of blue mussels also have revealed a relationship between intensity of stress and recruitment of different types of repair and demolition processes (40). As shown in Fig. 3B, HSP70 is induced in both congeners of blue mussels near 24°C. This initial phase of the CSR thus involves an emphasis on refolding of proteins that have incurred a reversible loss of conformation at high temperature. The induction of genes encoding subunits of the proteasome, which conducts nonlysosomal protein degradation, occurs only at higher temperatures and shows a species-specific pattern (Fig. 3C): M. trossulus exhibited a much higher level of expression of proteasome-encoding genes than M. galloprovincialis, and induction occurred at a lower temperature in the more cold-adapted mussel. This difference in expression of genes encoding proteasomal proteins, which mirrors the results of the proteomic analysis of these two species (80), suggests that greater damage was done to proteins in M. trossulus by heat stress. This conjecture is corroborated by the discovery of higher levels of ubiquitinated proteins in M. trossulus relative to M. galloprovincialis in common-gardened specimens (33).

A further illustration of how increasing heat stress shifts the pattern of stress-related gene expression is found in a field study of another species of Mytilus, the ribbed mussel, M. californianus (25). In field-acclimatized specimens collected over several tidal cycles from sites differing in thermal stress, induction of message for HSP70 occurred at lower temperatures than those necessary for increased transcription of genes associated with proteasomal function. Moreover, in the more heat-stressed population of M. californianus, there was temporal separation between maximal transcription of genes associated with cell proliferation and genes associated with aerobic metabolism, consistent with studies of yeast and other model organisms that have revealed temporal separation of oxidative and reductive phases in the cell cycle (81, 82). This striking periodicity in expression of genes under extreme heat stress illustrates a further, and heretofore unappreciated, aspect of thermal stress.

In summary, studies of temperature effects on gene expression in controlled laboratory studies and in a limited number of field studies have begun to reveal the graded nature of gene regulatory responses to heat stress. Activation of transcription of genes related to the CSR is not an “all or nothing” response to stress. Rather, as stress intensity increases, new classes of genes are activated that encode proteins playing a variety of functions in repair of cellular damage, removal of irreversibly damaged cellular constituents (and, most likely, of entire cells that are damaged beyond repair, notably in DNA fidelity), and regulation of cellular proliferation and life span. These transcriptomic analyses not only are revealing the mechanisms of cellular damage and repair, but in so doing are providing physiologists with molecular “biomarkers” for characterizing the degree of stress a cell has experienced and the differences among species in sensitivity to damage of the cellular apparatus. These insights into the fundamental aspects of cellular stress are starting to provide important knowledge concerning how sublethal stress might impact energy budgets and, ultimately, the abilities of ectotherms to persist in warming habitats.

Responses to Multiple Stresses: Gene Expression and Post-Translational Modifications of Proteins

Global change confronts organisms with multiple and often cooccurring forms of abiotic stress. A comprehensive analysis of the effects of a changing environment on organisms must take this complexity into account. In the case of many marine species from intertidal and estuarine habitats, changes in precipitation and storm runoff can lead to stress from osmoregulatory challenges. It is pertinent, then, to examine how changes in ambient salinity alter gene expression in these species and how these changes compare with transcriptional responses to changing temperature.

A study of blue mussels sheds light on these issues and reveals as well the degree to which congeneric species differ in their transcriptional responses to osmotic stress (42). Comparisons of transcriptional responses of native (M. trossulus) and invasive (M. galloprovincialis) blue mussels to hypoosmotic stress showed changes in expression of a variety of genes, but relatively few differences between species. The interspecific differences in response to sudden reductions in salinity that characterize the two blue mussels may include rapid behavioral changes [valve closure under salinity stress (5)] and post-translational modification of proteins (19), as well as somewhat slower changes in gene expression.

Comparisons of shifts in gene expression that occurred during either hypo-osmotic- or heat stress revealed 45 genes that changed expression in response to both stresses (42). However, almost all of the genes that were strongly upregulated by heat stress were strongly downregulated by osmotic stress. Among these oppositely responding genes were several that encode proteins involved in transmembrane movement of ions. Transport-related genes that exhibited strong upregulation under heat stress were strongly downregulated under hypoosmotic stress. Other transcriptomic studies of heat stress have shown strong upregulation of genes associated with membrane transport, a finding that may relate to disruption of membrane structure and barrier functions under acute heat stress (9).

A further aspect of the responses to reduced salinity by native and invasive blue mussels was revealed by comparisons of rapid post-translational modifications of gill proteins during hypo-osmotic stress (19). Compared with its congener, M. trossulus exhibited significantly greater amounts of protein phosphorylation by MAPKs during hypoosmotic stress. Because these signal transduction proteins are pivotal in mounting osmoregulatory responses and in suppressing apoptosis, the enhanced ability of M. trossulus to cope with sudden reductions in salinity may be more a consequence of rapid changes in protein phosphorylation than shifts in gene expression. The native and invasive congeners also displayed differences in protein phosphorylation under thermal stress (19), revealing a further level of regulatory complexity and an additional interspecific difference between the two species.

In summary, transcriptional responses and post-translational modifications of proteins show both similarities and differences under heat and osmotic stress. In some cases, these two stresses may be synergistic and create a higher level of cellular damage than found with either stress alone. Disruption of the cytoskeleton, seen in osmotic and heat stress studies of gills, is one possible example. These initial studies of transcriptional responses and rapid post-translational modifications of proteins in response to different abiotic stresses illustrate the complexities of the challenges to organisms arising from multi-faceted global change.

Orthologous Enzymes of Congeners: a Window into Protein Structure-Function Relationships and Evolutionary Adaptation of Proteins

The well-characterized high sensitivities of protein structure and function to alterations in temperature (22, 64, 65) indicate that climate change is likely to exert many of its effects through perturbation of enzymatic and structural proteins. These perturbations comprise impaired enzyme function, when temperatures reach levels at which kinetic properties like ligand binding depart from their optimal ranges; molecular chaperoning costs, which are associated with refolding of reversibly denatured proteins; and costs linked to proteolysis and protein synthesis, which arise when irreversibly damaged proteins are removed from the cell and replaced with newly synthesized proteins, through processes that have high ATP requirements. Thus, the effects of rising cell temperatures on enzymes can lead not only to metabolic impairment but also to increased costs for sustaining protein homeostasis. To the extent that protein evolution can modify protein stability and function in temperature-adaptive manners, these metabolic perturbations and energy costs could be substantially reduced.

Comparative studies of orthologous proteins from species adapted to different temperatures have addressed several questions related to global change. One of these questions—perhaps the most critical one of all—concerns the amount of change in ambient temperature that is sufficient to favor adaptive variation in proteins, to allow optimal function and stability to be maintained. A second question asks how much change in protein amino acid sequence is needed to achieve such adaptations. A follow-up question is where in the structure of the protein are adaptive changes possible and how many of these adaptable sites are available for change.

Protein orthologs of congeners adapted to different environmental temperatures have proven to be excellent study systems for elucidating these issues. For example, studies of orthologs of lactate dehydrogenase-A (LDH-A) in barracuda fishes (genus Sphyraena) with different latitudinal distributions (temperate, subtropical, and tropical) have shown that differences in habitat temperature of ∼3–5°C are sufficient to favor adaptive variation in sequence and function (34). Whether even smaller differences in average or maximal habitat temperature can favor protein evolution remains to be investigated. Suffice it to say, however, that predicted increases in global temperature are likely to be sufficient to favor adaptive evolution of proteins, if species are to sustain their current biogeographic ranges and maintain optimal metabolic capacities and acceptable costs of protein homeostasis. Conversely, the observed shifts in biogeographic ranges of many ectothermic species, including many pelagic fishes (2, 49) may be a reflection, at least in part, of suboptimal performance of proteins under newly elevated habitat temperatures.

Thermal perturbation of proteins can be separated into two categories, function and structural stability, which in fact are tightly linked (22, 64, 65). One of the most studied aspects of temperature-protein interactions in the context of function is the influence of changes in temperature on two kinetic properties of enzymes: enzyme-ligand interactions, which are often quantified by measuring Michaelis-Menten constants (Km values), and catalytic rate constants (kcat), which provide an index of the rate at which an enzyme can convert substrate to product (31). Both of these kinetic properties are at once strongly affected by temperature and highly conserved among species at their normal body temperatures (22, 31). An illustration of this relationship is given in Fig. 4, A and B, which shows how Km of pyruvate (KmPYR) and kcat differ between orthologous LDH-As of temperate (Chromis punctipinnis) and tropical (Chromis caudalis and C. xanthochira) damselfishes (36). In agreement with the results of broader phylogenetic surveys that examined orthologous LDH-As from vertebrates adapted to body temperatures between approximately −1.9°C (Antarctic notothenioid fishes) and 45°C (a thermophilic reptile) (22, 31), at a common temperature of measurement, the values of Km and kcat are higher for cold-adapted orthologs than those from warm-adapted species. These intrinsic differences in binding and catalytic power yield high degrees of conservation in these kinetic properties at species' normal body temperatures (values enclosed by ovals in Fig. 4, A and B).

Fig. 4.

Differences in kinetic and structural characteristics of lactate dehydrogenase-A (LDH-A) orthologs in tropical (Chromis caudalis and Chromis xanthochira) and temperate (Chromis punctipinnis) congeners of damselfish. A: effects of evolutionary- and measurement temperatures on the Michaelis-Menten constant for pyruvate (Kmpyr). B: effects of evolutionary and measurement temperatures on the catalytic rate constant (kcat). C: three-dimensional model of one subunit of LDH-A showing the site (219) at which an alanine residue in the tropical congeners is replaced with a threonine residue in the temperate species. [Modified after Johns and Somero (36).]

A key advantage of studying temperature-protein interactions with congeneric species is illustrated by the study of damselfish LDH-A orthologs. Because of the high similarity in nucleotide and amino acid sequences observed in these closely related species, identifying the amino acid substitution(s) responsible for adaptive change in kinetic properties is relatively straightforward. In the case of the damselfish LDH-As, four differences in amino acid sequence distinguished the temperate species from the two tropical congeners (36). Using site-directed mutagenesis to test the effects of these amino acid differences on function, we showed that a single substitution at site 219 in the sequence (alanine in the two tropical congeners; threonine in the temperate congener) was sufficient to account for the observed differences in KmPYR. This amino acid substitution resulted from a single nucleotide change (transition), the replacement of the adenine found at position 658 in the ldh-a gene of the temperate species with a guanine in the two tropical species. Thus, in a gene of ∼1,000 nucleotides, only a 0.1% change in sequence was sufficient to effect adaptive change in function.

Comparisons of a “sister” enzyme of LDH-A, cytosolic malate dehydrogenase (cMDH), another member of the Rossmann-fold family of dehydrogenases, have provided additional examples of how a single amino acid substitution can lead to temperature-adaptive modifications of Km and kcat (17, 23). Differences in kinetic properties of cMDH orthologs of native and invasive blue mussels, which are due to a single amino acid substitution, may help to explain the higher thermal tolerance and, therefore, the invasive success of the warm-adapted species, M. galloprovincialis. A similar biogeographic shift involving a northward range extension of a warm-adapted limpet, Lottia austrodigitalis, and a contraction of the southern distribution limit of a cold-adapted congener, L. digitalis, is also marked by adaptive variations in cMDH function and stability that arise from a single amino acid substitution (17).

The locations and the number of sites within an enzyme where this minimal change in sequence is adequate to foster adaptive change are also being revealed by comparisons of orthologous enzymes of congeners (17, 22, 23, 34, 36). As would be expected, sites in the catalytic vacuole of the enzyme, where ligands interact directly with amino acid side chains, are conserved among orthologs. The sites showing the highest levels of variation among orthologs typically are within regions of the protein that govern the energy changes that accompany conformational alterations essential for binding and catalysis. For LDH-A, these highly mobile and variable regions comprise the α1G- α 2G-helix, the BJ- α1G loop, and the αH helix (Fig. 4C). Movement of these regions during function affects catalytic rate (kcat) because the energy requirements needed to drive catalytically essential conformational changes establish the rate of enzyme function (22, 34). Higher rigidity of these mobile regions leads to lower kcat values. Conversely, when these mobile regions have greater intrinsic flexibility, the reaction faces lower energy barriers. However, as a consequence of these more flexible structures the binding of ligands is less probable (Km values are higher). This results from differences in the probability of the enzyme existing in a binding-competent (ligand-recognizing) conformation. Mobile regions with higher stability will more likely be in a ligand-recognizing geometry.

Because the structural flexibilities of several regions of a protein have effects on Km and kcat, there are a number of “targets” for adaptive evolution. Comparisons of several sets of orthologous LDH-As (34, 36) and cMDHs (17, 23) have shown that multiple sites of adaptive change underlie the variation in kinetic properties that have been observed. This finding, in concert with the discovery that a single nucleotide substitution is sufficient for temperature-adaptive variation, provides an initial basis for conjectures about the rate of protein evolution, the question of how likely it is that adaptive protein evolution will “keep up with” global warming. Whereas the minimal number of substitutions needed for adaptive change and the availability of numerous sites for this change to occur support a model of rapid change, a full analysis of this issue demands answers to several additional questions. These include the important—and essentially unanswerable at present—question about how many proteins “need” to adapt when temperature rises by a few degrees. The seminal work of Dennis Powers et al. (55) on allelic variation of enzymes in the teleost Fundulus heteroclitus along the East Coast of North America has shown that not all proteins show latitudinal variation in isozyme patterning (55). This suggests that some classes of proteins are more likely than others to benefit from adaptive variation in the face of small changes in temperature. This conclusion is supported by analyses such as that of Ream et al. (56), who found that the sequence of skeletal muscle actin was conserved among animals with body temperatures spanning a range of almost 40°C (Antarctic fishes to a desert reptile), whereas for LDH-A, sequence variation and concomitant adaptive change in kinetic properties was prevalent in this same suite of species. The rapid advances in sequencing technology and bioinformatic methods for sequence analysis, if coupled with appropriate examination of protein functional responses to temperature, may allow rapid increases in our appreciation of how many types of proteins are apt to benefit from “fine-tuning” of function and stability in the face of the increases in temperature expected from global change.

DNA Decay and Loss of Thermal Tolerance

The abilities of ectotherms to cope successfully with changes in temperature depend strongly on two different aspects of genome content. First, the variety of protein-coding genes in the genome may govern the ability of a species to withstand changes in temperature and other abiotic factors. Second, species' abilities to regulate the expression of this genetic information to acclimatize to changes in temperature may be critically important in determining how well they succeed in a warming world, at least in the short run before adaptive genetic variation can help to blunt the effects of stress. Elucidation of the differences that may exist in gene content and gene regulatory capacities among species may provide important insights into the fundamental genetic determinants of stenothermy and eurythermy.

A very “Kroghian” set of species for examining these questions are the cold-adapted stenothermal animals of the Southern Ocean (54). These species, for example, fish of the teleost suborder Notothenioidei, have evolved in extremely cold and thermally stable waters for ∼14 million years (54). As a consequence of this long evolutionary history under environmental conditions that did not require acclimatization to varying temperatures or salinities and that afforded the fish access to abundant dissolved oxygen, a number of types of DNA decay, the loss of genetic information, have taken place. These genetic lesions portend significant problems for these fish (and similarly genetically depauperate invertebrates) as the Southern Ocean increases in temperature.

One form of DNA decay is the loss of protein-coding elements in the genome. These losses may result from point mutations that disrupt coding regions or disappearance from the genome of all or part of a protein-coding gene. Perhaps the most striking illustration of DNA decay in protein-coding regions of the genome is the loss of hemoglobin (Hb) genes in notothenioids of the Family Channichthyidae (13), the “white-blooded icefish.” All members of this family have lost the entire gene encoding the β-chain of Hb and a fraction of the 5′ region of the α-chain as well. In addition to complete loss of circulating Hb, the expression of myoglobin (Mb) has been lost among some members of this family. All icefish so examined were found to possess the Mb-encoding gene, but expression of the gene has been lost in at least three lineages due to disruption of the reading frame of the gene or failure of the Mb message to be translated into protein (62). These genetic lesions poise the Channichthyidae in particular jeopardy from oceanic warming because this process will reduce oxygen concentration of the seawater and body fluids and elevate rates of oxygen consumption. In fact, studies of thermal tolerance of white-blooded- (Channichthyidae) and red-blooded notothenioids demonstrated that the former group has significantly lower tolerance of elevated temperatures than their Hb-containing relatives (4).

Another form of DNA decay identified in Antarctic fishes is the loss of gene regulatory capacities for initiating upregulation of heat-shock protein synthesis (10, 32). Although notothenioids possess genes for molecular chaperones, which in all organisms serve fundamental roles in the processes of protein folding and transport in the cell, there is an absence of capacity to upregulate any of these chaperone-encoding genes in response to heat stress. In a study using cDNA microarray analysis of transcriptional changes during heat stress in the notothenioid species Trematomus bernacchii, the only gene related to molecular chaperone function that exhibited upregulation was the gene encoding HSP40, which is a cochaperone that modulates activity of heat-shock proteins (10). Overall, even though a number of genes in T. bernacchii exhibited heat-induced upregulation, the responses were very muted compared with gene regulatory responses seen in temperate fish.

The loss of the ability to upregulate heat shock proteins may have several ramifications for fish experiencing rising temperatures. Not only may the processes of molecular chaperoning be impaired, but because of the role of HSPs in such important cellular processes as apoptosis, broader regulatory dysfunction may ensue. For example, HSP70 is important in suppressing apoptosis (3). Without sufficient controls on programmed cell death, stress could lead to an unregulated course of cellular destruction. Genes encoding proteins that upregulate apoptosis increased expression during heat stress in T. bernacchii, so a potential for imbalance in this aspect of cellular regulation seems possible (10).

Subsequent to the initial discovery of loss of the heat-shock response in notothenioid fishes, a number of Antarctic invertebrate species have been shown to lack this process as well (12). The loss of an ability to upregulate synthesis of molecular chaperones and the overall muted response of gene regulatory systems to increases in temperature seen in some Antarctic fish (10) may help to explain why these species are so stenothermal. Many Antarctic ectotherms show behavioral impairment at temperatures of only 1–3°C and upper lethal temperatures commonly are below 5–6°C (51, 54, 67).

In summary, a variety of physiological and genetic studies have begun to reveal some of the mechanisms underlying the extreme stenothermy of ectotherms of the Southern Ocean. Genetic lesions that disrupt or eliminate protein-coding regions and impede gene-regulatory functions would seem to put these species in extremely high risk from the effects of global warming. Reversal of these genetic lesions, notably those that involve loss of entire coding regions, to restore physiological function for life in a warming world seems highly improbable vis à vis the pace of warming of the Southern Ocean. Further studies of these cold-adapted stenotherms will provide added insights into this risk and help us to characterize the fundamental genetic differences that underlie the wide variation in environmental tolerance ranges of eury- and steno-tolerant species.

Perspectives and Significance

In preparing this review, I was again struck by the advances that have been made in comparative physiology—conceptually and technically—as well as in the massive amounts of data that are now available—since I entered this field almost 50 years ago. Had someone predicted in 1963, when I began my work in Antarctica with notothenioid fishes, that during my career there would appear technology that would allow global gene expression to be followed in these (or any!) organism; that the amino acid sequences of enzymes could be tinkered with, residue by residue, to test hypotheses about adaptation; and that the physiological studies we were doing on cold-adapted ectotherms from the Southern Ocean and elsewhere would contribute to study of a problem as broad as climate change—a concept that was first enunciated by Broecker (7) in 1975—I would have been a bit skeptical in all cases. What lies ahead for our field in terms of what we will be able to do and what the importance of our discoveries mean to our peers and to broader societal issues may be as difficult to foresee as the five-decades-off future was for me back in the early 1960s. I think, however, that with the advent of so many “omics” and biocomputational tools, in particular, comparative physiologists are on the threshold of seeing in remarkably fine detail how organisms “work,” from the level of contents of genetic tool kits to the many ways in which this information shapes the phenotype and confers upon it a species-specific level of phenotypic plasticity. The “window” on organismal function our field now possesses will allow continued growth of our understanding on these fronts and, in the process, our approach to physiology, ecology, and evolution will indeed continue to offer a powerful “crystal ball” for predicting the consequences of global change.


No conflicts of interest, financial or otherwise, are declared by the author.


This review has benefited greatly from discussions with many colleagues, especially Dr. Bruce D. Sidell, who helped to make our field one that is both intellectually vibrant and collegial, in the best sense of both of those terms. His work was an exemplar of how the comparative approach can be used most creatively and productively to address the issues that I have attempted to cover in this review.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
View Abstract