Metabolic substrate use and the turnover of endogenous energy reserves in broad-tailed hummingbirds (Selasphorus platycercus)

The nutrients that animals assimilate can follow several pathways. They can be stored for future use, immediately oxidized to fuel metabolism, or used to synthesize materials for reproduction, growth and tissue maintenance. The sugars assimilated by a non-reproductive nectar-feeding bird, for example, can be stored as glycogen, used to synthesize lipid, or oxidized immediately(Alexander, 1999). Historically, studies investigating metabolic substrate use have relied on the respiratory quotient(RQ=V̇_CO2^.V̇_O2^-1)to assess whether carbohydrates, lipids or proteins support respiration(Surarez et al., 1990; Powers,1991). Advances in elemental stable isotope analysis now allow an alternative/complementary method to RQ to determine metabolic substrate use. Recently, stable isotopes have been used to clarify nutritional, physiological and ecological questions that respirometry could not, for instance, to quantify the relative contribution of ingested (`income') to stored(`capital') nutrients for reproduction in adults of both moths and butterflies(O'Brien et al., 2000; O'Brien et al., 2004). Stable isotopes have also been used in migratory birds to discriminate between nutrients ingested on the wintering grounds and during migration versus those ingested on the breeding grounds(Hobson et al., 2000). These studies relied on animals that either naturally or artificially switched between diets of distinct isotopic composition.

We used a similar diet-shifting approach to clarify metabolic substrate use in hummingbirds. Carleton et al. found that the carbon stable isotope composition (δ¹³C) of respired CO₂ from feeding rufous hummingbirds (Selasphorus rufus) closely resembled that of dietary nutrients (Carleton et al.,2004). When they switched birds to a diet with a contrasting isotopic composition, δ¹³C of respired CO₂ was intermediate between diets, which indicated that hummingbirds were metabolizing both exogenous nutrients and endogenous reserves. Here, by measuring δ¹³C of exhaled CO₂ in animals that were shifted between diets with contrasting carbon isotope compositions, we were able to quantify the fraction of metabolism fueled by income (assimilated sugars) and capital (endogenous reserves) in broad-tailed hummingbirds(Selasphorus platycercus Swainson). Additionally, because stable isotopes allow determining isotopic incorporation of assimilated nutrients into an organism's tissues (Carleton et al., 2004), we examined both the isotopic incorporation of carbon and the mean residence time of a carbon atom in the endogenous reserves of broad-tailed hummingbirds.

We captured male broad-tailed hummingbirds Selasphorus platycercusSwainson (N=8) with mist-nets in Albany County, Wyoming, USA(41°20′N, 106°15′W). Birds were housed under a 15 h:9 h photoperiod (photophase: 05:30 h-20:30 h MST) in a room at 20±2°C. Hummingbirds fed ad libitum on a diet derived from C₃plants (Nektar-Plus®, Guenter Enderle, Tarpon Springs, FL, USA;δ ¹³C=-24.2±0.09, N=10) for ≈90 days prior to experiments. This was to ensure that the δ¹³C of their endogenous reserves reflected that of the C₃-derived diet (see List of abbreviations and symbols). Our experiment had three phases. During phase 1(day -11 to -1), we verified on three dates (day -11, -8 and -4) that birds were in isotopic equilibrium with their C₃ diet(Fig. 1). During phase 2 (day 0 to 19), birds fed ad libitum on a 20% (mass percent) sucrose solution derived from C₄ plants (δ¹³ C=-11.4±0.07, N=10) and fruit flies (δ¹³C=-23.0±0.4, N=11). In phase 3 (day 20 to 43), birds were shifted back to the C₃-derived diet. We weighed birds (at 08:30 h-09:00 h) periodically and measured their food intake throughout the experiment.

For phases 1, 2 and 3, we measured the δ¹³C in the breath of birds after an overnight fast (`fasted') at 05:30 h and on birds that had ad libitum access to food (`fed') at 12:00 h. Briefly, birds were taken from their cage, lightly restrained within a sleeve of laboratory tissue and introduced into 50 ml polypropylene centrifuge tube. This tube had an internal diameter of 28 mm and was fitted with a one-way stopcock valve(Fig. 1). After introducing the bird in the tube, we gently flushed the tube with ≈500 ml of CO₂-free air over a 30 s period. Tubes were then sealed and exhaled CO₂ was allowed to accumulate for 1 min. A 30 ml air sample was then extracted using a gastight syringe. Withdrawing the gas causes a sudden change in the pressure inside the tube. To avoid injuring the birds we immediately re-pressurized the chamber following gas extraction by opening the stopcock valve. Birds were unaffected by this procedure. Samples were gathered within 3 min after birds were taken from their cages. We injected air samples into pre-evacuated gastight vials (Exetainer®; Labco Ltd, Buckinghamshire,UK) until a positive pressure was achieved.

We measured the isotopic composition of expired CO₂ on a Micromass VG Optima continuous flow mass spectrometer coupled to a micro gas injector (GV Instruments, Manchester, UK) at the Mass Spectrometry Isotope Facility, Colorado State University (Fort Collins, CO, USA). The precision of these analyses was ±0.2‰ and our standard was gaseous CO₂ (δ¹³C=-37.8‰, VPDB). Our method is similar to that developed by Hatch et al.(Hatch et al., 2002) and applied by Podlesak et al. (Podlesak et al., 2005), except that we did not use party balloons.

Carbon isotope ratios of food were measured on a continuous flow isotope ratio mass spectrometer with samples combusted in a Carlo Erba NA 1500 elemental analyzer (Milan, IT). The precision of these analyses was±0.2‰. Our standards were vacuum oil(δ¹³C=-27.5‰, VPDB) and Australian National University sucrose (δ¹³C=-10.5‰, VPDB, NIST 8542). We included standards in every run to correct raw values obtained from the mass spectrometer. Isotope ratios in this paper are reported as δ values on a per mil (‰) basis relative to the International Atomic Energy Agency carbon isotope standard, Vienna Pee Dee Belemnite (VPDB).

We used standard least squares linear regression to estimate rates of change in body mass. Unless noted to the contrary, N=8 for our analyses. We report data as means ± s.d.

Birds ingested significantly different quantities of energy during the three phases of our experiment (RM-ANOVA: F_2,6=6.6, P=0.03). A Tukey's post hoc test, however, revealed that energy consumption was only greater in phase 1 (C₃; 25.9±2.5 kJ day^-1); birds had similar energy consumption rates in phases 2(C₄; 23.8±1.8 kJ day^-1) and 3 (C₃;23.5±2.3 kJ day^-1). Birds maintained body mass in phase 1 and lost mass at a slow rate in phase 2 (0.02±0.01 g day^-1,mean r_mass,time=-0.85, P<0.001; Fig. 2). However, they regained mass at a constant rate during the first 10 days of phase 3 (0.04±0.02 g day^-1, mean r_mass,time=0.84, P<0.001), after which they maintained relatively constant body mass (Fig. 2).

During phase 1, the birds breath had a distinctly C₃ signature. The exhaled CO₂ of birds when fasted overnight was significantly more negative (δ¹³C=-27.1±0.2) than after 6 h of feeding (δ¹³C=-25.5±0.2; paired t-test: t₇=19.2, P<0.001; Fig. 3). On the days of a diet shift (day 0 and 20), the δ¹³C in exhaled CO₂ of fed and fasted birds was dramatically different(Fig. 3). Fasted birds exhaled CO₂ with δ¹³C that closely resembled that of their previous diet, whereas fed birds exhaled CO₂ withδ ¹³C that closely resembled that of the new diet(Fig. 3). On day 0, fasted and fed birds exhaled CO₂ with δ¹³C=-26.6±0.6 and -13.2±1.1‰, respectively (paired t-test: t₇=37.5, P<0.001); on day 20, fasted and fed birds exhaled CO₂ with δ¹³C=-13.1±0.6 and-23.2±1.2‰, respectively (paired t-test: t₇=25.9, P<0.001).

On a diet-shift day, the δ¹³C of breath in fed birds was very similar to that of their diet; however, these two values were not identical. On day 0, δ¹³C of breath in fed birds was slightly, but significantly, more negative than that of their diet (one sample t-test: t₇=4.5, P=0.003; Fig. 3); on day 20,δ ¹³C of breath in fed birds was slightly, but significantly,more positive than that of their diet (one sample t-test: t₇=2.5, P=0.04; Fig. 3). During phases 2 and 3,the δ¹³C of fed birds' breath changed through time and eventually came to resemble that of their current diet(Fig. 3).

The change in δ¹³C of CO₂ exhaled by fasted birds was well described by Eqn 1 (the coefficients of determination ranged from 0.89 to 0.97; Fig. 3). The fractional rate of isotopic incorporation, k, was significantly higher in phase 3 than in phase 2 (k=0.86±0.16 and 0.47±0.19,respectively; Table 1) and the asymptotic carbon isotopic composition of the breath[δ¹³C(∞)] of fasted birds was significantly more depleted in ¹³C than that of their food (one sample t-tests: phase 1, t₇=14.4, P<0.0001;phase 2, t₇=24.2, P<0.0001; Fig. 1, Table 1).

In our discussion, we compare the findings of our stable isotope approach to those of respirometric studies on metabolic substrate usage in hummingbirds. We then use δ¹³C of expired breath to quantify the fraction of metabolism fueled by endogenous and exogenous nutrients. We also estimate isotopic incorporation rates and carbon atom residence times in hummingbirds, and consider how energy balance affects them. We conclude our discussion by addressing the possibilities and limitations of our stable isotope approach to ecological studies.

On the days that birds shifted diets (day 0 and 20), fasted birds exhaled CO₂ with δ¹³C that resembled that of their previous diet, whereas fed birds exhaled CO₂ withδ ¹³C that closely resembled that of the new diet(Fig. 3). These results support the notion that fed hummingbirds fuel their metabolism primarily with recently ingested sugars, whereas fasted hummingbirds use endogenous reserves(Suarez et al., 1990). Our results are consistent with measurements on Anna's (Calypte anna) and Costa's (Calypte costae) hummingbirds(Powers, 1991). During the day, when birds were feeding, Powers found that their RQ was approximately 1.0, which indicates that birds were oxidizing sugars; after an overnight fast, their RQ was close to 0.7, which indicates that birds were oxidizing lipids (Powers, 1991). Although our stable isotope approach does not allow identifying the endogenous substrates used by fasted hummingbirds, the significant difference between theδ ¹³C of food and breath in birds in isotopic equilibrium suggests that a large fraction of the substrates oxidized by fasted hummingbirds were lipids (Table 1). In general, lipids are depleted in ¹³C relative to the carbohydrates from which they are synthesized(DeNiro and Epstein, 1977). Our hypothesis, that endogenous lipids fuel the fasting metabolism of hummingbirds, can be tested by measuring δ¹³C in expired breath and RQ concurrently.

Our results are also consistent with the predictions of Suarez et al.(Suarez et al., 1990), who proposed that active, fed hummingbirds should oxidize carbohydrates preferentially to fuel respiration and rapidly shift to lipids after even very short fasts (Suarez and Gass,2002). Using dietary sugars as fuel when feeding is advantageous because using synthesized fat to fuel respiration entails an approximately 16%cost of synthesis. However, hummingbirds are small and have high metabolic rates. Thus, in order to spare their small glycogen reserves, they must shift to the oxidation of lipids even after short fasts(Suarez and Gass, 2002). Changes in the δ¹³C in breath of fasting hummingbirds can reveal the details of shifts in substrate oxidation during the transition from the absorptive to the postabsorptive state.

We only estimated P for day 0 and 20 because here the end-points of the mixing model were sufficiently distinct to allow using Eqn 2 with confidence. Endogenous reserves contributed 11.6±7.3 and 8.5±11.0% (paired t-test: t₇=0.7, P>0.5) to total respiration on day 0 and 20, respectively. Although fed hummingbirds fueled respiration primarily (≈90%) with dietary sugars, they oxidized a small fraction of endogenous reserves as well(Carleton et al., 2004). Surprisingly, during phase 2, there was no evidence of a significant contribution of the isotope composition of fruit flies in theδ ¹³C of breath of hummingbirds. We hypothesize that hummingbirds routed the protein contained in this component of their diets directly into the manufacture of body protein and thus spared the protein in fruit flies from oxidation(Martínez del Rio and Wolf,2005).

Hummingbirds incorporated the isotope signal of their diet into endogenous reserves very rapidly (Table 1). The average residence time of a carbon atom in a hummingbird's endogenous reserves can be estimated as k^-1. On average, between assimilation, storage and oxidation, a dietary carbon atom remained in a hummingbird's energy reserves only between 1 and 2 days. The remarkably high mass-specific metabolic rates of hummingbirds(Suarez and Gass, 2002)explain their high rates of isotope incorporation, and hence the high rates of carbon flux, into energy reserves. Carpenter et al. estimated that non-migrating hummingbirds store between 0.2 and 0.5 g of lipids(Carpenter et al., 1993). Assuming endogenous reserves comprise primarily lipids, hummingbirds turned over 0.10 to 0.24 g lipid day^-1 in phase 2, when they were losing body mass. These numbers are within the range of lipid masses lost overnight by congeneric rufous hummingbirds(Carpenter et al., 1993).

A serendipitous result of our experiment allowed us to address how isotopic incorporation is affected by energy balance. The fractional rate of isotope incorporation (k) was almost twice as high in phase 3 compared to phase 2(Fig. 3). This disparity has a relatively straightforward explanation. Birds were losing body mass,presumably including endogenous energy reserves, during phase 2, but gaining it during phase 3. Fry and Arnold (Fry and Arnold, 1982) and Hesslein et al.(Hesslein et al., 1993)recognized that the value of k is determined by both the addition of new material (`net growth') and by the replacement of material exported from the tissue as a result of catabolism (`catabolic turnover'). If the animal is losing endogenous reserves, k equals the fraction of new materials from the diet that partially replace the materials lost by catabolism. However, if the animal is increasing the size of its endogenous reserves, then k has two components: the fractional rate of storage, which represents a net addition to endogenous reserves, and the fractional rate of oxidation, which represents replacement. Therefore, an increase in the size of endogenous reserves equates to higher fractional rates of isotope incorporation.

Stable isotopes have been used to investigate what animals eat(Dalerum and Angebörn,2005) and to assess the temporal variation in their diets(Hatch et al., 2002). A particularly informative approach is to use tissues that incorporate dietary isotope signatures at different rates within a single individual(Podlesak et al., 2005; Dalerum and Angerbörn,2005). Our experiment established that fed hummingbirds oxidized primarily, albeit not exclusively, dietary sugars. Thus, the carbon isotope composition of breath in a fed hummingbird provided a snapshot of the isotopic composition of what the animal was eating. Our experiments also allowed us to measure the turnover of endogenous reserves and revealed it was brisk. Therefore, we were able to use the δ¹³C values in breath to distinguish between what animals were eating currently and what they ate previously, but only on the days of a diet shift (day 0 and 20). Two days after a diet shift, the CO₂ of the breath of fasted birds contained a mixture of carbon from the old and the new diet. Hummingbirds incorporate dietary carbon into their energy reserves so rapidly that theδ ¹³C in the breath of fed and fasted animals cannot be used to assess temporal variation in the isotope composition of their diet -except at very short time scales. Carleton and Martínez del Rio demonstrated that the rate of isotopic incorporation into blood is an allometric function of body mass (Carleton and Martínez del Rio, 2005). Therefore, we expect that larger animals will have slower carbon fluxes into their energy reserves. Measuringδ ¹³C in absorptive and postabsorptive individuals of larger species will likely resolve temporal variation in diet over longer time scales(Hatch et al., 2002). However,measuring δ¹³C in absorptive and postabsorptive animals to resolve temporal variation in the isotopic composition of diet will require determining the allometric relationship between carbon atom residence time in energy reserves and body mass.

List of abbreviations and symbols
 
• δ¹³C

  [(¹³C/¹²C_sample-¹³C/¹²C_standard^-1]×10³(‰, VPDB)

 
• δ¹³C_diet

  isotope composition of the diet (‰, VPDB)

 
• δ¹³C_tissue

  isotope composition of the tissue (‰, VPDB)

 
• δ¹³C_fasted

  isotope composition of breath from a fasted bird (‰,VPDB)

 
• δ¹³C_fed

  isotope composition of breath from a fed bird (‰, VPDB)

 
• δ¹³C(∞)

  asymptotic equilibrium isotope composition (‰, VPDB)

 
• δ¹³C(0)

  initial isotope composition (‰, VPDB)

 
• δ¹³C(t)

  isotope composition at time t (‰, VPDB)

 
• ¹²C

  carbon isotope (6 protons, 6 neutrons) (moles)

 
• ¹³C

  carbon isotope (6 protons, 7 neutrons) (moles)

 
• C₃

  C₃ photosynthetic pathway

 
• C₄

  C₄ photosynthetic pathway

 
• k

  fractional rate of isotopic incorporation (day^-1)

 
• k^-1

  residence time (days)

 
• MST

  mountain standard time (24 h)

 
• NIST

  National Institute of Standards and Technology

 
• P

  proportion

 
• RM-ANOVA

  repeated-measures analysis of variance

 
• RQ

  respiratory quotient (dimensionless)

 
• t

  time (days)

 
• V̇_CO2

  carbon dioxide production rate (ml min^-1)

 
• V̇_O2

  oxygen consumption rate (ml min^-1)

 
• VPDB

  Vienna Pee Dee Belemnite

Annie Hartman Bakken sketched the hummingbird breathalyzer. We thank Robert Carroll for assisting us with experiments and Pancho Bozinovic and Todd McWhorter for their thoughtful comments on this work. Hummingbirds were collected under United States Fish & Wildlife and Wyoming Game & Fish permits issued to C.M.R. Capture, care and experimental protocols were approved by the University of Wyoming's Institutional Animal Use and Care Committee. This work was supported by a National Science Foundation grant(IBN-0110416) to C.M.R.