Dietary sugar as a direct fuel for flight in the nectarivorous bat Glossophaga soricina

In evolving to be small and to hover while feeding on floral nectar,neotropical Glossophagine bats have undergone evolutionary convergence with hummingbirds. Hovering flight is energetically costly; hovering hummingbirds achieve some of the highest known mass-specific metabolic rates among vertebrates (Suarez, 1992). During hovering, >90% of whole-body metabolic rate (rate of oxygen consumption, V̇_O2)is due to exercising flight muscles(Suarez, 1992; Taylor, 1987). Hover-feeding hummingbirds are almost exclusively able to fuel their highly aerobic flight muscles using recently ingested sugar(Welch et al., 2006; Welch and Suarez, 2007). In contrast, at moderate to high exercise intensities, humans and other mammals rely heavily on endogenous fuels (glycogen and triglyceride) stored in their locomotory muscles. This is thought to be due to their limited capacities to transport and metabolize exogenous fuels (glucose and fatty acids) from the blood (Weber et al., 1996; Weber, 1988). In addition,capacities for the assimilation, transport and oxidation of dietary fuels are also limited; in humans only about 25–30% of energy expenditure can be supported by sugar ingested shortly before (or during) exercise. Given their nectarivorous diet and phylogenetic status, it is of interest to determine the extent to which nectarivorous bats rely on recently ingested sugar to fuel their energetically expensive flight. Their small size, high wing-beat frequencies, high metabolic rates and reliance upon simple sugars to supply most of their daily energy needs (Norberg et al., 1993; Voigt and Speakman, 2007; Winter and von Helversen, 1998; Winter and von Helversen, 2001), led us to hypothesize that, like hummingbirds, they are able to directly fuel their exercising muscles using recently ingested sugar. To test this hypothesis, we used a combination of respirometric and stable carbon isotope techniques to compare the use of recently ingested sugar by nectarivorous bats and hummingbirds. The evidence obtained lends further support for the idea that physiological and biochemical traits have undergone evolutionary convergence among these nectarivorous,flying vertebrates.

We evaluated the ability of Pallas' long-tongued nectar bats(Glossophaga soricina Pallas 1766) to support hovering metabolism with newly ingested sugar using a diet-switching approach. Bats were maintained for several weeks on a diet of beet sugar and powdered milk from cows fed a C3 crop. Hummingbirds were maintained for several weeks to months on a commercial hummingbird diet supplemented with beet sugar. The carbon in each of these diets displayed a stable isotope ratio(¹³C/¹²C) characteristic of C3 photosynthetic plants. During the experimental period, bats and hummingbirds were allowed to feed on a solution of cane sugar, the product of C4 photosynthesis with a significantly different ratio of ¹³C/¹²C. Sugar solutions were dispensed by a tube inside a plastic mask into which ambient air was drawn and analyzed downstream for O₂ and CO₂content. In addition to the estimation of relative rates of oxygen consumption(V̇_O2) and carbon dioxide production(V̇_CO2) during periods when the bats and hummingbirds breathed through the mask, expired air was captured to determine the ratio of ¹³C/¹²C present in the CO₂ in it. Because of the difference in carbon stable isotopic signatures of beet and cane sugars and because the expired CO₂ is derived from the carbon of oxidized fuels, we were able to determine the source (i.e. endogenous C3 or dietary C4) of oxidized fuels and the time course over which the fuel source changed as the bats and hummingbirds continued to feed (Carleton et al., 2006; Welch et al.,2006).

All capture, housing and experimental protocols were approved by the University of California, Santa Barbara Institutional Animal Care and Use Committee (Protocols 672 and 722). All data, except where noted, are presented as means ± s.e.m.

Individual Pallas' long-tongued nectar bats (Glossophaga soricinaPallas 1766; body mass at start of experiment=9.9±0.1 g; N=7,i.e. 3 males, 4 females) were captured via mist nets in banana plantations located near Tecomán, Colima, Mexico. Bats were transported to the city of Colima and were housed, indoors, in a wire-mesh cage of dimensions 0.5×0.5×0.5 m. Bats were fed ad libitum on a 20% (w/v) beet sugar solution supplemented with 5% (w/v) powdered cow's milk(Nestle Nido, Glendale, CA, USA) and 0.01% (w/v) ascorbic acid. Theδ ¹³C value of this maintenance diet was–25.70±0.12‰ (VPDB, mean ± s.d., N=10).

Data collection on bats was conducted in a mesh tent approximately 2 m long×2 m wide×1 m high. Average temperature during all experiments was 21.6±0.1°C (range= 20.7–23.8°C). Bats were weighed while inside a cloth bag on an electronic balance immediately before and after participation in the experiment. Data collection took place during February and March 2007 between 23:00 h and 06:00 h. Bats were fasted during the day and prior to data collection. This simulated their natural cycle of feeding and fasting, ensured that they were motivated to feed, and maximized their reliance on fatty acid oxidation.

Night-time experiments commenced as bats were provided with a sugar cane sucrose solution (20% w/v). The δ¹³C value of this solution was –11.22±0.12‰ (VPDB, mean ± s.d., N=10). Bats could access the sucrose solution by hovering in front of and inserting their head inside a plastic tube (derived from a 30 ml syringe)functioning as a mask. The solution was contained in a 30 ml syringe placed in a syringe pump (NE-500, New Era Pump Systems, Inc., Wantagh, NY, USA) and delivered to the mask by thin plastic tubing. An infrared emitter and detector were placed on opposite sides of the front edge of the mask such that the IR beam would be occluded whenever the bat's head was in the mask. Occlusion of IR beam triggered the release of sucrose solution by the syringe pump at a rate of 3 ml min^–1. This delivery rate was chosen because it resulted in the bats remaining in the mask for the longest period of time. The period of occlusion of the IR beam also represented the duration of the feeding event; this was the time over which measurement of reduction of[O₂] and enhancement of [CO₂] during hover-feeding occurred (see below). Air was continuously drawn into the mask at a rate of 500 ml min^–1 by a pump and passed through a column of Drierite™ (W. A. Hammond Drierite, Xenia, OH, USA) to scrub it of water vapor before entering the CO₂ analyzer (CA-2A, Sable Systems International, Las Vegas, NV, USA). After leaving the CO₂ analyzer,the air was drawn through a Drierite-Ascarite-Drierite column (Ascarite II,Arthur H. Thomas, Philadelphia, PA, USA) to scrub CO₂ and residual water from the line and into the oxygen analyzer (FOXBOX, Sable Systems International). A thermistor was placed near the mask to record ambient temperature. Output from the gas analyzers, infrared detector and thermistor were recorded by a notebook computer using Expedata (version 1.0.17, Sable Systems International).

Immediately before data collection, the oxygen analyzer was calibrated with well-mixed ambient air drawn through the mask in the absence of a bat. The carbon dioxide analyzer was calibrated with CO₂-free nitrogen gas(zero gas) and 0.5% CO₂ in nitrogen gas (Praxair, Danbury, CT,USA).

stp-corrected O₂ depletion and CO₂enrichment associated with each feeding event were determined by first subtracting baseline values (determined as the linear extrapolation of points directly before and after the feeding event in question) and then converting baseline-corrected data to ml gas by application of standard equations(Withers, 1977). Determination of absolute rates of oxygen consumption(V̇_O2) and carbon dioxide (V̇_CO2)production was not possible during this experiment because subsampling of incurrent air was attempted in each case (see below). However, as subsampling likely did not discriminate between oxygen and carbon dioxide, relative volumes (ml) of oxygen and carbon dioxide respired by the bat were determined. Relative gas exchange rates (for use in calculating RQ values) were obtained by integration of depletion or enrichment peaks over time (min).

Because subsampling of expired gas for stable isotope analysis prior to analysis precluded determination of absolute rates of O₂consumption (V̇_O2)and CO₂ production(V̇_CO2), separate measurements of V̇_O2 and V̇_CO2 during hover-feeding were performed on each individual at a flow rate of 1200 ml min^–1 without taking expired breath subsamples.

We obtained data from Anna's hummingbirds (Calypte anna Lesson 1829; body mass at start of experiment=4.4±0.4 g; N=2, both male) and rufous hummingbirds (Selasphorus rufus Gmelin 1788; body mass at start of experiment=3.2±0.1 g, N=10, i.e. 6 males, 4 females) for comparison with bats. Capture and rearing were essentially as described previously (Welch et al.,2007; Welch and Suarez,2007). Birds were fed ad libitum on a 13% (w/v) solution of Nektar-Plus (Guenter Enderle, Tarpon Springs, FL, USA) supplemented with beet sugar (5% w/v). The δ¹³C value of the maintenance diet was –25.84±0.11‰ (VPDB, mean ± s.d., N=10). Data collection took place between December 2005 and March 2006 between 06:00 h and 11:00 h in the laboratory at 24.0±0.1°C. Prior to each experiment, the hummingbird was fasted overnight to ensure that it would oxidize primarily fat at the start of data collection(Suarez et al., 1990; Welch et al., 2006). The experimental protocol was identical to that described for bats except that the duration was 60 min and the sucrose solution was contained in a 20 ml syringe placed directly in the mask. The δ¹³C value of the sucrose solution used in this experiment was –11.69±0.11‰ (VPDB,mean ± s.d., N=10). Data for which f_a was greater than 0.5 were excluded as estimates ofδ ¹³C_sample were less robust in these cases.

Bat RQ values quickly rose as foraging continued, approaching 1.0(0.97±0.01; N=7) by about 30 min after feeding bouts commenced(Fig. 1A); this indicates that bats had switched to oxidizing predominantly carbohydrates. Because we relied on spontaneous, voluntary feeding behavior, variation among individuals in feeding frequencies (among both bats and hummingbirds) caused variation in the amount of time it took for RQ values to stabilize at or near 1.0, as well as in the amount of time required for δ¹³C values in expired breath to stabilize near the δ¹³C value of the experimental diet (below). In general, both RQ and δ¹³C had stabilized to new steady-state values at least 30 min after foraging bouts had commenced.

Similar to results reported previously for hummingbirds(Welch et al., 2006; Welch and Suarez, 2007), RQ values for Anna's and rufous hummingbirds were initially low during the first feeding following a fasting period. RQ averaged 0.72±0.00(N=2) for Anna's and 0.73±0.01 (N=5) for rufous hummingbirds during the first feeding following a fasting period. Calculated as above, using Eqn 3, f_fat values averaged 95.6±1.2% (N=2) for Anna's and 91.4±2.5% (N=5) for rufous hummingbirds during the first feeding following a fasting period, indicating the birds were oxidizing fatty acids almost exclusively to fuel their first foraging flight.

RQ values from hovering hummingbirds rose quickly with each successive foraging event, reaching average values near 1.0 (C. anna:1.01±0.00; N=2; S. rufus: 0.99±0.00; N=10) by approximately 30 min after the first feeding following the fasting period (Fig. 1A),indicating essentially exclusive reliance on the oxidation of carbohydrates.

After being maintained on a diet with a δ¹³C value of–25.70±0.12‰ (VPDB, mean ± s.d.), bats expired CO₂ that yielded correspondingly low averageδ ¹³C_breath values during the first feeding following a fast (–25.71±0.56‰, VPDB; N=7) and were not significantly different from the δ¹³C signature of their maintenance diet (t₆=–0.0102, P=0.9920). This indicates reliance on endogenous fuel stores. Like RQ values, average δ¹³C_breath values quickly rose as foraging continued (Fig. 1B),and approached (–14.11±0.64 ‰, VPDB; N=7) theδ ¹³C signature of the experimental sugar solution(–11.22‰, VPDB; t₆=–4.5307, P=0.0040). This indicates a significant shift towards reliance on exogenous sugar to fuel hovering flight.

The data from hummingbirds followed a similar pattern. During the first feeding following the fasting period, hummingbirds expired CO₂ withδ ¹³C_breath values that were not significantly different from their maintenance diet, which had a δ¹³C value of –25.84±0.11‰ (VPDB, mean ± s.d.; C. anna: –27.78±0.49‰, VPDB, N=2, t₁=–4.3345, P=0.1443; S. rufus:–26.73±0.44‰, VPDB, N=5, t₄=–2.0234, P=0.1131), indicating reliance on endogenous energy stores. Like RQ values, species-averaged hummingbirdδ ¹³C_breath values quickly rose as foraging continued (Fig. 1B), and approached (C. anna: –12.54±0.24‰, VPDB; N=2; S. rufus: –12.42±0.20‰, VPDB, N=10) the δ¹³C signature of the experimental sugar solution. After 30 min of foraging, averageδ ¹³C_breath values from Anna's hummingbirds were not significantly different from the experimental sucrose solution(–11.69‰, VPDB; t₁=–3.5087, P=0.1768). However, while averageδ ¹³C_breath values from rufous hummingbirds during this period were close to the δ¹³C signature of the experimental sugar solution, they were significantly lower(–11.69‰, VPDB; t₉=–3.5730, P=0.0060). As with the bats, this indicates a significant shift in hummingbirds towards reliance on exogenous sugar to fuel hovering flight.

The average percentage rate of isotope incorporation into the pool of expired CO₂ (k′= k×100) in bats was 7.0±1.7% per min (range: 3.4–9.5%, N=7). In Anna's hummingbirds the average percentage rate of isotope incorporation into the pool of expired CO₂ (k′) was 7.8±3.2%min^–1 (range: 3.3–12.2%, N=2). In rufous hummingbirds, the average value of k′ was 11.8±1.4%min^–1 (range: 4.0–24.1%, N=10).

RQ and δ¹³C_breath values are highly significantly correlated for each of the three species examined here (data pooled by species; Fig. 2; G. soricina: r₉₁=0.9374, P<0.0001; C. anna: r₂₂=0.9767, P<0.0001; S. rufus: r₈₁=0.9185, P<0.0001). This indicates that newly ingested sugars were the primary source of the carbohydrates oxidized during hovering flight.

Upon reaching steady state after feeding commenced (>30 min after the first feeding following a fast), M_exo averaged 3.80±0.22 mg min^–1 (N=7, Table 1). The bats'mass-specific rate of oxidation of exogenous sugars(M_exo/M_b, in mg sugar min^–1 g^–1 body mass) averaged 0.37±0.02 mg min^–1 g^–1 (N=7)during the same period (Table 1). Upon reaching steady state after feeding commenced, whole animal M_exo values were generally similar in Anna's and rufous hummingbirds, averaging 3.49±0.08 and 2.45±0.07 mg min^–1, respectively. M_exo/M_b values averaged 0.72±0.05 mg min^–1 (N=2) and 0.71±0.01 mg min^–1 (N=10) during feeding bouts upon reaching steady state after feeding commenced in Anna's and rufous hummingbirds,respectively (Table 1).

The data presented here indicate that Pallas' long-tongued nectar bats oxidize recently ingested sugar during hovering (3.80±0.22 mg min^–1) at a rate very similar to hovering hummingbirds(C. anna, 3.49±0.08 mg min^–1; S. rufus, 2.45±0.07 mg min^–1). Correcting for differences in body mass, the bats achieved mass-specific oxidation rates of recently ingested sugar (0.37±0.02 mg min^–1g^–1) that are slightly more than half of the rates achieved by the hummingbirds (C. anna, 0.72±0.05 mg min^–1 g^–1; S. rufus,0.71±0.01 mg min^–1 g^–1). In an experimental design closely approximating that used in our studies on hummingbirds and bats, Jentjens et al.(Jentjens et al., 2004b)report maximal rates of oxidation of exogenous sugars of 1220±70 mg min^–1 (N=9) in exercising humans during a period 120–150 min after the initial ingestion of a glucose + sucrose solution(Table 1). Their data yield an average mass-specific rate of oxidation of exogenous sugars of 0.02 mg min^–1 g^–1 (16.5 mg min^–1kg^–1; Table 1). The mass-specific rate of oxidation of exogenous sugar displayed by Pallas' long-tongued nectar bats is about 18 times greater than the rate estimated in humans and is the highest rate ever reported in mammals. These results add further support to the hypothesis of convergent evolution in physiological and biochemical traits among hovering nectarivorous animals. In addition, they indicate that the traditional paradigm, i.e. that exogenous,dietary fuel use is highly limited in mammals during high intensity exercise,is not without exception.

In addition to their high capacity for oxidation of exogenous sugars, bats appear to be similar to hummingbirds in being able to use exogenous sugars to fuel hovering metabolism soon after ingestion. The range of fractional rates of isotopic incorporation into the pool of expired CO₂ was large within each species. This is undoubtedly due to the variation in feeding frequency and ingestion rate observed across individuals. In addition,variation in the rapidity with which ingested sugar appears in the pool of actively metabolized substrates may be due to the transit-time of the ingested sugar solution from storage organs, such as the crop and stomach, to the small intestine, where most sugar absorption occurs. These were not measured in the present study. Nonetheless, rates of incorporation of exogenous sugar into the pool of actively metabolized substrates seen in bats and hummingbirds exceed the rates observed in humans (Jentjens et al., 2004b).

One factor limiting the use of exogenous sugars during exercise in humans is the capacity for carbohydrate absorption by the intestine(Hawley et al., 1992; Jentjens et al., 2004a). Rates of sugar absorption by hummingbird intestines are exceptionally high(Karasov et al., 1986; McWhorter et al., 2006) and a large fraction of the sugar absorption rate occurs via a paracellular pathway (McWhorter et al.,2006). Pallas' long-tongued nectar bats also possess the capacity for high rates of sugar absorption(Winter, 1998). Although data distinguishing between active transport and paracellular movement are currently not available for nectarivorous bats, studies on both the Egyptian fruit bats Rousettus aegyptiacus and the great fruit bat Artibeus literatus demonstrate that these also possess high capacities for sugar transport via a paracellular pathway(Caviedes-Vidal et al., 2004; Tracy et al., 2007). In hummingbirds, rates of active sugar transport in the intestines are insufficient and high rates of paracellular transport are required to fully meet daily energy requirements (McWhorter et al., 2006). Given the high rates of daily energy expenditure and energetically expensive hovering flight of Pallas' long-tongued nectar bats, it is reasonable to expect that they would also rely heavily on a paracellular pathway for sugar absorption. Egyptian fruit bats experience much greater rates of paracellular absorption than similar-sized laboratory rats(Tracy et al., 2007). In non-volant mammals, the ratio of paracellular absorption to active transport of glucose increased with body mass, i.e. paracellular absorption made a minimal contribution in mice, the smallest mammals examined(Pappenheimer, 1990). This pattern supports the argument that it is not small size per se but,rather, convergent evolution among these flying nectarivores that resulted in their increased capacities for paracellular absorption of sugars.

Despite their remarkable abilities, it is interesting that the bats in our study failed to support hovering metabolism with exogenous sucrose to the same extent seen in hummingbirds. The RQ values obtained from bats foraging for at least 30 min were significantly lower than 1.0 (0.97±0.01, N=7; t₆=0.01711. P=0.0017). In contrast,RQ values obtaining from hummingbirds foraging for at least 30 min were 1.01±0.00 (N=2) in the case of C. anna and 0.99±0.00 (N=10) in S. rufus. This suggests that,even when nectar is available, bats continue to rely to some extent on fat oxidation to fuel hovering flight. Using Eqn 6, we calculate that bats foraging for at least 30 min support 10.3±2.1% (N=7, t₆=4.7915, P=0.0030) of hovering metabolism with fat. In comparison, Anna's and rufous hummingbirds support essentially none of their hovering metabolism with fat (i.e. not significantly different from zero; C. anna:–4.5±0.4%, N=2, t₁=–11.711, P=0.0542; S. rufus: 1.9±1.2%, N=10, t₉=1.5142, P=0.1643).

When we constrain calculated values for the percentage of hovering metabolism supported by exogenous carbohydrate and endogenous fat to between 0–100%, the balance of bat hovering metabolism not supported by either fat or exogenous carbohydrates was 12.1±2.9% (N=7). This likely represents the fraction of metabolism fuelled by endogenous glycogen,derived from carbon in the maintenance diet. In comparison, we calculate that Anna's and rufous hummingbirds fuel lower fractions of their hovering metabolism with endogenous carbohydrates (C. anna, 3.4±1.7%, N=2; S. rufus, 4.3±0.8%, N=10). The support of nearly 1/8th of hovering metabolism with endogenous glycogen differentiates bats from hummingbirds which, during hover-feeding, appear to rely on oxidation of glycogen previously synthesized from the maintenance diet to a lesser extent (Suarez et al.,1990; Welch et al.,2006; Welch and Suarez,2007). During the first feeding following a fasting period, the bats relied on endogenous carbohydrates, likely in the form of glycogen, to fuel 27.7±3.7% of hovering metabolism. In comparison, Anna's and rufous hummingbirds supported only 8.1±1.0 and 7.1±4.0%, respectively,of hovering flight during the first feeding following a fasting period with endogenous carbohydrates. Thus, it appears that under both fasting and fed conditions, bats rely upon endogenous glycogen to fuel hovering flight to a much greater extent than hummingbirds.

A caveat to our interpretation of these data is that the floral nectars typically consumed by Phyllostomid nectarivorous bats and hummingbirds differ in the relative abundances of component sugars. On average, the sugars in `bat nectars' consist of 20% sucrose, while the sugars in `hummingbird nectars'consist of 60% sucrose (the balance consisting of the monosaccharides fructose and glucose) (Baker et al.,1998). Thus, their specialization on flowers producing low-sucrose nectars may explain why bats, despite their remarkable abilities, are unable to fuel their flight muscles with dietary sucrose to the same extent as hummingbirds. Consistent with this interpretation, Hernandez and Martínez del Rio determined that sucrase activities per unit surface area of intestine in Pallas' long-tongued nectar bats are approximately half those measured in hummingbirds (Hernandez and Martínez del Rio, 1992). Thus, it is possible that the oxidation rate of exogenous sugars by hovering bats is more constrained by limitations in the rate of sucrose hydrolysis (and subsequent absorption into the circulation) than in hovering hummingbirds. This suggests that nectarivorous bats may be capable of fueling 100% of flight metabolism when provided a mix of sucrose, glucose and fructose mimicking the nectar compositions they feed on in nature. In their recent study, Voigt and Speakman demonstrate that Pallas' long-tongued nectar bats are able to fuel a greater proportion of their resting metabolism with exogenous sugars when provided with glucose, as opposed to sucrose (Voigt and Speakman, 2007). However, extrapolation to foraging flight may not be appropriate because metabolic rates are much higher and due mainly to skeletal muscles under these conditions.

We expect that the results presented here will lead to further questions concerning the mechanistic bases for the exceptional abilities of bats to fuel exercise using recently ingested sugar, the evolution of these capacities, and their coevolution with bat-visited flowering plants. An increased understanding of the mechanisms underlying the high rates of dietary sugar metabolism among vertebrate endotherms may lead to insights into certain metabolic pathologies and their evolution in humans.

LIST OF SYMBOLS AND ABBREVIATIONS

 
• δ¹³C

  isotopic ¹³C/¹²C ratio referenced to international standard

 
• f

  fraction of expired CO₂ derived from metabolic substrate

 
• g

  volume CO₂ produced by glucose oxidation

 
• M

  amount of metabolic substrate oxidized (mg min^–1)

 
• M_b

  body mass (g)

 
• Met

  rate of caloric expenditure (cal min^–1)

 
• RQ

  respiratory quotient(=V̇_CO2/V̇_CO2)

 
• t

  time (min)

 
• V̇ _CO2

  rate of carbon dioxide production (ml CO₂g^–1 h^–1)

 
• V̇ _O2

  rate of oxygen consumption (ml O₂ g^–1h^–1)

 
• VPDB

  Vienna Pee Dee Belemnite C standard

Jazmín Osorio and Rocío García assisted in capturing bats used in this experiment. Scott Hodges provided valuable comments regarding the data. We thank Georges Paradis and the staff at the UC Santa Barbara Marine Science Institute Analytical Lab for their timely work. Funding was provided by a UC MEXUS Dissertation Research Grant to K.C.W., a UC MEXUS-CONACYT Collaboration Grant to R.K.S. and L.G.H.M., by a grant from Consejo Nacional de Ciencia y Tecnología (SEP-2004-CO2-43343) to L.G.H.M. and by NSF grant IOB 0517694 to R.K.S.