Energetics , thermoregulation and torpor in the Chilean mouse-opossum Thylamys elegans ( Didelphidae ) Energética , termorregulación y sopor en la yaca Thylamys elegans ( Didelphidae )

In this paper we studied the energetics and thermoregulation of the Chilean mouse-opossum Thylamys elegans (Dielphidae) a nocturnal small marsupial, endemic of southern South America. We studied their standard energetic and determined whether they exhibit shallow daily torpor or deep prolonged torpor as a function of ambient temperature and food availability. Thylamys elegans partially supports the hypothesis that Neotropical marsupials have somewhat a higher basal metabolic rate (BMR) and thermal conductance (Cm) than Australian ones. In fact, BMR was higher but Cm was lower than expected for their body mass. The higher mass-independent BMR of the Chilean mouse-opossum may be explained by its insectivorous food habits and its low Cm by its temperate habitats. Euthermic Chilean mouse-opossum showed daily fluctuations in body temperature being significantly higher during night time. In addition T. elegans entered in daily torpor and aroused spontaneously only was food was absent. That is, this species display a facultative type of daily torpor because propensity to enter in torpor was dependent of the combination of food absence and low ambient temperature. No torpor was observed when food was available. During torpor ambient temperature was slightly above ambient temperature between 0.3 to 0.5 °C. Torpor in this species as well as in marsupials in general, appears to be a flexible and an opportunistic response to unpredictable environmental conditions.


INTRODUCTION
Torpor in small endotherms is a well-studied example of energy budget regulation.Theoretically, torpor is often hypothesized to be responsible for allowing some organisms to adjust to changing biotic and abiotic environmental conditions, through increases in survival under harsh environmental conditions (Humphries et al. 2003).Thus, torpor is an energy and water-conserving mechanism used by many species of small endotherms (Schmidt-Nielsen 1987, Geiser 1994, Holloway & Geiser 1995, Cryan & Wolf 2003).Torpor is characterized by a decrease in body temperature and a decline in metabolism of c.a. 90 %, in comparison to euthermy (Geiser 1994, Ortmann & Heldmaier 1997).This relationship between torpor, ambient temperature and food availability has been well documented for many years.For example, Hainsworth et al. (1977) suggested that daily torpor is used by hummingbirds exclusively when food is scarce.Extensive use of torpor by a population should also be advantageous when food is scarce.Indeed, food supply appears to be an important factor influencing the annual period of inactivity of the rodent Perognathus longimembris, which may indicate torpor (Kenagy 1973).The mean duration of individual episodes of torpor increases for kangaroo mice when food supply is reduced (Brown & Bartholomew 1969).On the other hand, ambient temperature largely explains the reduction of metabolism of daily heterotherms and hibernators (Geiser 1986, 1988, 1994, López-Calleja et al. 1997).
Small eutherian mammals and birds are classical examples of the processes listed above (for a review see McNab 2002).However, it is not well known what happens in other endotherms such as in South American marsupials (but see Bozinovic & Rosenmann 1988a, 1988b, Caviedes-Vidal et al. 1990, Bozinovic & Marquet 1991, for examples in some South American rodents).The physiology of torpor in marsupials is similar to eutherians in several aspects (Hallam & Dawson 1993, Chappell & Dawson 1994, Gibson & Hume 2000, Holloway & Geiser 2001).Nevertheless, with respect to energetics and thermal acclimation, some differences arise.Nespolo et al (2002) demonstrated that the potential for acclimatory response of thermogenic capacity in marsupials is controversial, reporting a significant effect of thermal acclimation on energy metabolism, organ masses and digestive enzyme activity in a Chilean mouse-opposum marsupial Thylamys elegans (Didelphidae).In the same species Silva-Durán & Bozinovic (1999) tested the effect of food availability on energy expenditure and torpor (see also Roig 1971).They found that animals maintained with low food availability showed a higher frequency of torpor bouts in comparison to animals with high food availability, which never entered into torpor, hypothesizing that the reduction of energy expenditure achieved with torpor states by T. elegans is regulated through food availability.
In this study we followed the later approach to assess the effects of temperature and food availability on metabolic rate, thermoregulation and torpor bouts in this species which inhabits Mediterranean environments of central Chile.
Thylamys elegans shows conspicuous phenotypic flexibility in activity of intestinal enzymes, both, on a seasonal basis (Sabat & Bozinovic 1994) and in response to diet acclimation (Sabat et al. 1995, Bozinovic & Méndez 1997).In addition to the conceptual relevance of determining the underlying causes of observed differences in performance in a wild endothermic species, the scarcity of knowledge regarding the comparative physiology of South American marsupials will make this effort a valuable one.Consequently, the aim of this work is to determine in T. elegans the effect of changes in ambient temperature and food availability on metabolic rate, torpor state as well as on its frequency and regulation.

MATERIAL AND METHODS
Between 1982-1990 we studied a total of fifteen individuals (six males and nine females) with a mean body mass (m b ) of 40.2 ± 9.8 g (SD).Animals were captured in semiarid and Mediterranean scrublands of central Chile.This species feeds primarily on insects (Sabat & Bozinovic 1994), and is strictly nocturnal.Animals were captured with Sherman traps, transported to Santiago, and individually maintained in large plastic cages with a photoperiod of LD = 12:12 and ambient temperature of 20-25 °C.Grapes, apples, eggs, grounded beef meat, worms and water were provided ad libitum.
Metabolic rate was measured during at least 1.5 h at different T a in a computerized closed respirometric system (Morrison 1951).CO 2 and water vapor were absorbed inside the metabolic chamber through granules of Baralyme™ and Drierite™ respectively.Body mass and T b were measured before and after the metabolic runs through Cu-Constant thermocuples and BAT-12 digital thermometer.Minimal thermal conductance was calculated from the slope of the regression of VO 2 versus Ta curve below the thermoneutral zone.Extrapolated T b fell within 1 SD of the average diurnal and nocturnal euthermic measured values of T b , thus the theoretical classical assumption that T a = T b when VO 2 = 0 was satisfied.To determine BMR, animals were fasted for 3 h before beginning the measurements that were conducted in 850-mL steel chambers at 30 and 32.5 °C.
All VO 2 values were corrected for STP (standard temperature and pressure).Body temperature was measured during day and night in euthermic animals as well as in torpid individuals through Cu-Constantan thermocouples connected to a Baily digital thermometer.To test for a shallow or for a deep torpor response, continuous VO 2 records were conducted in different individuals with food as well as, deprived of food (the most usual way to induce torpor, Lyman et al. 1982).Animals were exposed to a LD = 12:12, but at different T a ranging from 5 to 25 °C.These measurements lasted for a minimum of 24 h.In addition, to determine torpor regulation by effect of temperature, we changed it during VO 2 measurements.
Comparisons between means were made through a one-way ANOVA and a post-hoc Tukey test, also a linear regression analysis was conducted to determine the thermal dependence of VO 2 .Results are reported as mean ± 1 SD.Statistical analyses were performed using STATISTICA (6.0).

RESULTS
The effect of T a on VO 2 of T. elegans is shown in Fig. 1.The resting VO 2 below thermoneutrality increased in a linear form as T a decreased being represented by the following linear regression, VO 2 = 4.94 -0.138 (± 0.004) T a , r = 0.99, P < 0.0001.The line intersects the abscissa at 35.8 °C a value within the range of diurnal and nocturnal recorded T b (see below).The slope of the curve (0.138 mLO 2 g -1 h -1 °C-1 ) represents the minimal thermal conductance (C m ) which was: 87.5 5 % of the expected value for a similar size (40.2 g) standard mammal (i.e., C m = 1.0 m b -0.5 , Herreid & Kessel 1967).
The basal metabolic rate was 1.07 ± 0.01 mLO 2 g -1 h -1 °C-1 (n = 12) which was a 112.3.9 % of the predicted value for a standard metatheria (BMR = 2.49 m b -0.26 ) and 85.3 % of the predicted value for a standard eutheria (BMR = 3.53 m b -0.28 ) see (McNab 2002).The lower critical temperature of thermoneutrality was 27.5 °C.A slight increase in VO 2 was recorded at 35 °C (Fig. 1).In addition T. elegans exhibit torpor between 5 to 25 °C and only when food was absent (Fig. 1).During torpor T b was slightly above T a , between 0.3 to 0.5 °C.
In an experimental T a of 15 °C and only when deprived of food, T. elegans'entered in daily torpor and aroused spontaneously every night (Fig. 2).When in torpor VO 2 falls nearly 1/100 th of that during euthermia.Body Relación entre la tasa metabólica y la temperatura ambiente en Thylamys elegans durante eutermia (círculos cerrados) y sopor (círculos abiertos).Cada punto representa una media de doce animales (± 1 DE).Véase el texto para detalles.
Metabolic rate (mLO 2 g -1 h -1 ) Ambient temperarure ( o C) temperature during deep torpor was 0.5-1.0'°C(n = 12) above T a .Interestingly when we changed T a from 15 to 0 °C, torpid animals aroused spontaneously (Fig. 3).Also, torpor bouts length lasted from 7 to 19 h.When T a was changed to the initial conditions (15 °C) animals become torpid again (Fig. 3).The same patterns of daily torpor in the absence of food were observed during all seasons, revealing a facultative type of torpor.hypothesized that this is because Dromiciops is a well-differentiated lineage from South American opossums within the Australidelphian radiation of metatherians (Palma 2003).That is, hypothetically its physiological energetics seems to be phylogenetically closer to the Australian marsupials.
Nevertheless, T. elegans partially supports McNab's (1978) predictions, because BMR was higher but C m was lower than expected for their body mass.The higher mass-independent BMR of the Chilean mouse-opossum may be explained by its insectivorous food habits and its low C m by its temperate habitats (McNab (1978(McNab ( , 2002)).In addition, Nespolo et al. ( 2002) measured thermoregulatory maximum metabolic rate and BMR in T. elegans, before and after acclimation to cold or warm conditions.These authors observed no significant effect of thermal acclimation on maximum metabolic rate however; BMR of cold acclimated animals was 30 % higher than that for warm acclimated individuals, suggesting that BMR shows plasticity.Cambios continuos en la tasa metabólica (eutermia y sopor) en un individuos representativo de Thylamys elegans en dos temperaturas diferentes y sin alimento.Las barras negras indican las horas de oscuridad mientras que las barras grises el tiempo entre cambios de temperatura.Note que los animales sales del sopor a la temperatura de congelamiento.Véase el texto para detalles.

Metabolic rate (mLO
The combinations of high BMR and low C m by implying a continuous T b regulation, is likely to be advantageous for survival in temperate habitats (McNab 1992).The minimal temperature differential between body and ambient (∆T m ) is equal to BMR/C m (McNab, 1979).Indeed low mass-independent C m values (high thermal insulation) coupled to high massindependent BMR values may contribute to a higher thermoregulatory capacity by allowing heat conservation.The predicted ∆T m for marsupials follow the equation: 2.49 m b 0.28 (McNab 2002).In T. elegans the ratio BMR/C was 7.8 °C representing a 111.4 % of the expected value for a 40.2 g standard marsupial.
The high thermoregulatory capability of T. elegans may explain why this species never became torpid when food is available (Silva-Durán & Bozinovic 1999).Following Geiser's (1994) classificatory scheme of torpor in marsupials, T. elegans seems to display a facultative type of daily torpor.The propensity to enter in torpor was dependent of the combination of food absence and T a .Torpor in this species as well as in marsupials in general appears to be a flexible and an opportunistic response to unpredictable environmental conditions (Geiser 1987).It is interesting to note that the tendency to become torpid at freezing temperatures was reduced (Fig. 3).We hypothesized that under these environmental conditions, this species should exhibit a controlled daily torpor avoiding physiologically risky conditions.Overall, the thermoregulatory energetic response of T. elegans during daily torpor was similar to the one reported for Australian species (Geiser (2003) as well as to that shown by some small Metabolic rate (mLO 2 g -1 h -1 ) Time (h) eutherian hibernators (Kayser 1961), representing a fall to 1/100th of the metabolic rate of the euthermic animals.Silva-Durán & Bozinovic (1999) reported that the frequency of torpor in T. elegans varied with food availability.That is, these authors observed that with amounts of food that exceed daily energy requirements, animals were never torpid in contrast to animals confronted to lower availability of food.Moreover, the relationship between the length of torpor bouts and digestible energy intake indicates that food availability influences not only the frequency of torpor but also its length.In fact, Holloway & Geiser (1995) indicate that the amount of energy saved during torpor was dependend not only on the level of metabolic rate, but also on the duration of torpor as well as on the amount of energy required for rewarming.We postulated that the reduction of energy expenditure achieved by a torpid animal is induced by food availability, i.e., regulation by energy supply (Ricklefs 1996).Besides, it has been shown that nutrient absorption in the digestive tract of torpid endotherms decreased when T b drops (Holloway & Geiser 1995).Then, in terms of energy savings, it would be risky to maintain torpidity for long periods, given the low absorption of nutrients that it implicates.Nevertheless, in the absence of food it may allow an economy of energy.In addition, Silva-Durán & Bozinovic (1999) also observed that, in spite of an increase in food availability, animals did not increase energy expenditure, probably because such augmentation results in extra energy costs.On the other hand, when food availability was low, animals decreased metabolic rate to a minimum that should be maintained even under the lowest food availability.Apparently, there is a physiological constraint set by the energy needed for arousal, which represents the greatest energy cost within the torpid cycle (Opazo et al. 1999).Interestingly, winteracclimated Djungarian hamsters (Phodopus sungorus) enter torpor spontaneously with food ad lib when their thermogenic capacity is high (Heldmaier & Steinlechner 1981).
The habitat of T. elegans may be described as an evergreen scrub, locally know as matorral.The climate is Mediterranean, with an annual mean rainfall of 376.4 mm, concentrated (65 %) during the winter months.On average, it rains every month of the year, but precipitation is scant during late spring and summer (3 % of the yearly total).Mean temperature is highest from late spring and summer and lowest during the austral winter.The combination of high temperature and low precipitation occurs during summer, thus determining a summer drought.Since this habitat is seasonal in both water and food availability, torpor appears to be important in water and energy economy.Thus, depending of the local environmental conditions, we postulate that there is a trade off between the need to satisfy energy/water requirements through foraging and the necessity to increase energy/water savings through torpor.Furthermore, as postulated by Geiser (1986) torpor appears to have importance in the survival of individuals mainly in a highly seasonal and unproductive habitat (see Humphries et al. 2003).
In summary, the observed flexibility in torpor states dealing with energy and water economy may favour survival and reproduction.News studies dealing with the variability of thermoregulatory strategies during torpor are needed.For example, Cryan & Wolf (2003) demonstrated that during reproduction females of heterothermic bats defend euthermia, presumably to expedite embryonic growth, while males use torpor, presumably to minimize energy and water deficits.

Fig. 1 :
Fig. 1: Relationship between metabolic rate and ambient temperature in Thylamys elegans in euthermia (closed circles) and in torpor (hollow circles).Values are a mean of 12 animals (± 1 SD).See text for details.

DISCUSSION
Fig. 2: Daily torpor in a representative individual of Thylamys elegans measured over a 3 days period at 15 °C without food.Bars indicated dark hours.Sopor diario en un individuo representativo de Thylamys elegans medido durante 3 días a 15 °C sin alimento.Las barras indican las horas de oscuridad.

Fig. 3 :
Fig. 3: Continuous changes in metabolic rate (euthermia and torpor) in a representative individual of Thylamys elegans at two different temperatures without food.Black bars indicated dark hours while the vertical grey bars indicate the time lag between changes in ambient temperature.Note that animals arouse from torpor at freezing temperature.See text for details.