Phenotypic flexibility in metabolic adjustments and digestive function in white-shouldered starlings: responses to short-term temperature acclimation Free

Phenotypic flexibility can be defined as reversible, temporary and repeatable changes in organisms in response to changes in an internal or external environment (McKechnie, 2008; Piersma and Drent, 2003; Starck and Rahmaan, 2003). Seasonal plasticity, a response to seasonal climate change developed in birds, is a good example of phenotypic plasticity (Swanson, 2010). In terms of energy expenditure, phenotypic plasticity is one of the key components of the thermoregulatory response to changes in seasonal energy requirements for birds (Swanson, 2010; Zheng et al., 2008a, 2014a). Ambient temperature is considered to be one of the most important environmental factors that influence an animal's seasonal thermoregulation and drive the evolution of a suite of morphological, physiological and behavioral adaptations (Barbatelli et al., 2010; Chi and Wang, 2011; Goymann et al., 2006; Sollid and Nilsson, 2006; Swanson et al., 2014). Low ambient temperatures represent a potential energetic problem for all homeothermic endotherms because of a substantial increase in heat loss as the temperature gradient between the environment and the interior of the body increases (Bicudo et al., 2010). For small birds that live in cold environments, increased thermogenesis is an important adaptive strategy because of the high surface-to-volume ratio compared with larger species. For an endotherm, basal metabolic rate (BMR) is the lowest metabolic rate in a rested, inactive and postabsorptive state in the absence of thermal stress (McNab, 1997; Swanson, 2010; Zheng et al., 2008b). High BMR means elevated maintenance costs required to support energetically demanding lifestyles, including living in harsh environments (Swanson et al., 2017). BMR is often, though not always, elevated in winter relative to summer for birds wintering in cold climates or in cold-acclimated relative to warm-acclimated birds (McKechnie, 2008; McKechnie and Swanson, 2010). Species responding to cold environments with an increase in BMR include hoopoe larks (Alaemon alaudipes) (Williams and Tieleman, 2000), red knots (Calidris canutus) (Vézina et al., 2006), Chinese bulbuls (Pycnonotus sinensis) (Zheng et al., 2014a), Eurasian tree sparrows (Passer montanus) (Li et al., 2020), Chinese hwameis (Garrulax canorus) (Zhou et al., 2016) and red billed leiothrix (Leiothrix lutea) (Cui et al., 2019). A high BMR arises as a correlated response to the direct selection of other physiological traits associated with high environmental costs, such as daily energy expenditure or activity or thermogenesis capacity (Swanson et al., 2017). At the cellular level, variation in mitochondrial activity within tissues and organs should contribute to the changes in metabolic rate. Therefore, state-4 respiration (S4R) and cytochrome c oxidase (CCO) activity are usually used as indicators for tracking changes in BMR (Swanson et al., 2014; Venditti et al., 2004; Zheng et al., 2014a, 2008b).

Omics studies have the potential to identify metabolic pathways that play key roles in phenotypic flexibility, but only a few studies have examined transcriptome-wide patterns of gene expression in small birds in response to cold acclimation. Stager et al. (2016) measured the pectoralis transcriptomic profiles of captive dark-eyed juncos (Junco hyemalis) and found that enhanced thermogenic performance in cold-acclimated birds is accompanied by the upregulation of genes involved in muscle hypertrophy, angiogenesis, lipid transport and oxidation, and in catabolic enzyme activity. In another study, black-capped chickadees (Poecile atricapillus) and American goldfinches (Spinus tristis) were found to have increased lipid catabolic capacity in winter compared with that in summer, as revealed by their pectoralis transcriptomic profiles (Cheviron and Swanson, 2017). Compared with other omics, metabolomics can better reflect the true phenotype of a cell or organism. Global quantification of metabolite levels can give a biochemical fingerprint of the activity of the metabolic pathway, facilitating the identification of processes that are altered during temperature acclimation (Goodacre et al., 2004; Lin et al., 2006; Morowitz et al., 2000).

The ‘energy demand’ hypothesis for short-term adjustments in BMR posits that the size of an endotherm's internal organs may undergo readjustment relative to food intake, a correlate of energy demand. Digestive phenotype flexibility, which has been well documented in several reviews, is important for the very high feeding rate in vertebrates (Karasov, 1996; Karasov and Hume, 1997; Karasov and Martinez del Rio, 2007). In cold conditions, birds respond to the variable environment, either physical or biological, by increasing food intake, accompanied by changes in the morphology of their digestive tract and improved absorption of nutrients in order to adapt to the different energy requirements (Cavieres and Sabat, 2008; Khalilieh et al., 2012; Goymann et al., 2006; Piersma and Drent, 2003; Salvante et al., 2010; Syafwan et al., 2012; Wu et al., 2014; Zhou et al., 2016). For example, red knots (C. canutus) can increase their food intake when they are either acclimated to cold or experience altered ambient temperatures, compared with birds kept in their thermal neutral zone (Vézina et al., 2006). Similar results were reported for Chinese bulbul (P. sinensis) (Wu et al., 2014), hoopoe larks (A. alaudipes) (Williams and Tieleman, 2000) and house wrens (Troglodytes aedon) (Dykstra and Karasov, 1992), of which P. sinensis and A. alaudipes were found to have a significantly larger intestine, and P. sinensis and T. aedon have longer small intestine under cold acclimation relative to warm acclimation. White-throated sparrows (Zonotrichia albicollis) (Barceló et al., 2017), red knots (C. canutus) (Vézina et al., 2017) and cedar waxwings (Bombycilla cedrorum) (McWilliams et al., 1999) also exhibit morphological changes or changes in digestive enzyme activity after temperature acclimation. However, the underlying mechanisms and the specific contributions of digestive phenotype flexibility to cold-induced changes in metabolic capacity are not fully understood.

White-shouldered starlings (Sturnus sinensis) are a small migratory passerine bird that mainly breeds in Southern China and Northern Vietnam, but will migrate to Southeast Asia, the Philippines and Borneo in winter (MacKinnon et al., 2000). They mainly live in the open areas of low mountains, plains and hills, especially in dry land within a forest environment, but they also appear in farmland, towns and villages. These birds are omnivorous and feed on figs as well as other flowers and fruits. Numerous studies have focused on the regulation of heat production as the adaptation of small birds to temperature acclimation (Vézina et al., 2006; Noakes and McKechnie, 2020; Cui et al., 2019; Jetz et al., 2008; Stager et al., 2016; White et al., 2007). However, these studies either did not include the metabolomic changes of thermogenic organs or ignored the effects of digestive tract structure and function. In order to fill this gap, we systematically measured several morphological, physiological and biochemical markers in two groups of white-shouldered starlings that were acclimated to either 30 or 10°C. We also used metabolomic techniques to compare the metabolite profiles of the pectoral muscle tissue under different temperature conditions. This approach enabled us to perform an integrated analysis to shed light on the changes in the metabolome of the muscle tissue that may play important roles in the metabolic response to a change in energy requirement. We hypothesized that adjustment of energy metabolism and thermogenesis could be important strategies for small birds to cope with temperature acclimation based on the ‘energy demand’ hypothesis. We predicted the birds acclimated to 10°C would show higher body and organ masses, a higher level of respiratory enzymes, and enhanced catabolism in the pectoral muscle compared with the birds acclimated to 30°C. Moreover, birds acclimated to 10°C would have an increased energy budget and improved digestion of food and absorption of nutrients compared with those acclimated to 30°C, accounting for their increased BMR.

This study was carried out in Wenzhou City, Zhejiang Province (27°29′N, 120°51′E, 14 m in elevation), China. Wenzhou has a warm temperate climate with an average annual rainfall of 1500 mm spread across all months, with slightly more precipitation during spring and summer. Mean daily temperature ranges from 28 to 39°C in July and from 3 to 8°C in January (Zheng et al., 2008a; Wu et al., 2015; Zhou et al., 2016).

A total of 16 male Sturnia sinensis (J. F. Gmelin 1788) were captured in July 2019. In the laboratory, these birds were housed individually in cages (50×30×20 cm³) at 25±1°C under a 12 h:12 h light:dark photoperiod with the lights on at 06:00 h (Zhou et al., 2016). Following capture, the birds were habituated to the laboratory conditions for 2 weeks. During this period, they had ad libitum access to food and water. After habituation, the birds were randomly divided into two groups: one group was housed at 30±1°C (warm-acclimated group, n=8) and the other group was kept at 10±1°C (cold-acclimated group, n=8). These two temperatures are comparable to the mean of the daily maximum temperatures experienced by birds in the field in summer and winter. The BMR of each bird was measured at 30°C before and after 4 weeks of acclimation. BMR was measured at 30°C because this temperature is within the thermoneutral zone for this species (unpublished data). All experimental procedures were carried out in compliance with the Animal Care and Use Committee of Wenzhou University.

Metabolic rate was measured using an open-circuit respirometry system (TSE Systems, Berlin, Germany). Each bird was allowed to rest in a 1.5 l plastic metabolic chamber kept inside a temperature-controlled cabinet capable of regulating the temperature to ±0.5°C (Artificial Climatic Engine BIC-300, Shanghai, China). Dry air was used as a control, and the air was scrubbed of water before being pumped into the chamber at a rate of 1000 ml min⁻¹ (flow rate was measured with a mass flow meter in the machine) through a sealed cylindrical Perspex chamber. The gas leaving the chamber was also dried using a special drier (TSE Systems) and directed through an oxygen analyzer at a flow rate of 600 ml min⁻¹ (Wu et al., 2015; Zhou et al., 2016). The data were recorded and averaged every 10 s by a computer connected via an analog-to-digital converter (Lab Master, TSE Systems), and then analyzed using standard software (TSE systems). The software connected to the metabolic instrument compared the composition of the sampled gas with the components in the atmosphere to automatically calculate oxygen consumption. This signal was then translated into a digital signal by the Lab Master software (Wen et al., 2019). All measurements of gas exchange were obtained during the rest phase of the birds' circadian cycles (between 20:00 h and 04:00 h) in dark chambers. Food was removed 4 h before each test to create postabsorptive conditions. Measurement of oxygen consumption commenced when the birds were observed perching calmly in the chamber, and each animal was placed in the metabolic chamber for at least 2 h (Wu et al., 2015; Zhou et al., 2016; Hu et al., 2017; Li et al., 2022). BMR was calculated from the lowest rate of oxygen consumption over 10 min. Body temperature was measured before metabolic measurements using a lubricated thermocouple inserted in the cloaca, and the output was digitized using an Oakton thermocouple meter (Eutech Instruments, Singapore). The mean body mass of each experimental group was calculated by measuring the body mass of each bird before and after the experiment.

Birds were euthanized at the end of the experiment, and their pectoral muscles, liver, heart, kidneys and small intestine were weighed to the nearest 0.1 mg (Zheng et al., 2008b; Zhou et al., 2016). The small intestine was washed in normal saline to remove all contents before being weighed. The liver, pectoral muscle, heart and kidney samples were roughly chopped with scissors and then nine volumes of ice-cold medium (250 mmol l⁻¹ sucrose, 5 mmol l⁻¹ Tris HCl, 1 mmol l⁻¹ MgCl₂, 0.5 mmol l⁻¹ EDTA and 0.5 mg PBS, pH 7.4) were mixed with each sample, followed by homogenization using a high-throughput tissue homogenizer (SCIENTZ-48L, Ningbo Scientz Biotechnology Co., Ltd, Ningbo, China). S4R was measured with a Clark electrode (DW-1, Hansatech Instruments Ltd, King's Lynn, UK) at 30°C in 1.96 ml of respiration medium (225 mmol l⁻¹ sucrose, 50 mmol l⁻¹ Tris HCl, 5 mmol l⁻¹ MgCl₂, 1 mmol l⁻¹ EDTA and 5 mmol l⁻¹ KH₂PO₄, pH 7.2) essentially as described by Estabrook (1967). S4R was measured for 1 h with succinate as a substrate (Zheng et al., 2008a). CCO activity in the organs/tissues was measured at 30°C in respiration medium (100 mmol l⁻¹ KCl, 20 mmol l⁻¹ TES, 1 mmol l⁻¹ EGTA, 2 mmol l⁻¹ MgCl₂, 4 mmol l⁻¹ KH₂PO₄, 60 mmol l⁻¹ BSA, pH 7.2). For each sample, 10 μl of suspension and 30 μl of cytochrome c (37.9 mg ml⁻¹) were added to the electrode, after which CCO activity was measured in a final volume of 2 ml (Zheng et al., 2014b; Zhou et al., 2016). S4R and CCO activity were expressed as mean mass-specific values (μmol O₂ min⁻¹ g⁻¹ tissue) and as whole-organ activity (μmol O₂ min⁻¹) (Zhou et al., 2016; Cui et al., 2019). Whole-organ activity (μmol O₂ min⁻¹) was defined as being equal to the mass-specific value (μmol O₂ min⁻¹ g⁻¹ tissue) multiplied by the organ mass.

The pectoral tissue (100 mg) sampled from each bird was ground into powder in liquid nitrogen. The powder was then resuspended in 500 µl of prechilled 80% methanol containing 0.1% formic acid. The sample was kept at −20°C for 30 min and then centrifuged at 15,000 g at 4°C for 15 min. After centrifugation, 400 µl supernatant was diluted with liquid chromatograph mass spectrometer (LC-MS)-grade water to reduce the methanol concentration in the sample to 60%. The sample was filtered through a 0.22 μm filter and then centrifuged at 15,000 g at 4°C for 10 min. Finally, the sample was subjected to metabolite analysis via LC-MS/MS.

LC-MS/MS analysis was performed using a Vanquish UPLC system (Thermo Fisher), and an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher) operated in data-dependent acquisition (DDA) mode. The sample was injected into an Accucore HILIC column (100×2.1 mm, 2.6 μm) at a flow rate of 0.2 ml min⁻¹. The column was eluted by a gradient consisting of buffer A and buffer B under positive and negative polarity modes. For positive polarity mode, buffer A consisted of 0.1% formic acid in water, while buffer B consisted of just methanol. For negative polarity mode, buffer A consisted of 5 mmol l⁻¹ ammonium acetate in methanol (pH 9.0) whereas buffer B was just methanol. The gradient was set as follows: 2% B, 1.5 min; 2–80% B, 15 min; 100% B, 14 min, 100–2% B, 14 min; 2% B, 15 min. The Q-Exactive HF-X mass spectrometer was operated in positive/negative polarity mode with the spray voltage, capillary temperature, sheath gas flow and aux gas flow set to 3.2 kV, 320°C, 35 arb, and 10 arb, respectively. In order to control the quality of the experiment, a quality control (QC) sample was prepared during sample processing. The QC sample consisted of an equal portion of each experimental sample from both groups, and it was used to balance the chromatographic–mass spectrometry system and monitor the state of LC-MS system performance, as well as for evaluating the stability of the system during the whole experimental process. At the same time, blank samples were set up to allow subtraction of background ions.

The raw data generated from UPLC-MS/MS were processed using Compound Discoverer 3.0 (CD 3.0, Thermo Fisher) to perform peak alignment, peak selection and quantification for each metabolite. The main parameters were set as follows: retention time tolerance, 0.1 min; actual mass tolerance, 5 ppm; signal intensity tolerance, 30%; signal:noise ratio, 3; and minimum intensity, 100,000. After that, peak intensities were normalized to the total spectral intensity. The normalized data were used to predict the molecular formula of each species based on the additive ions, molecular ion peaks and fragment ions, and the peaks were then matched to compounds in the mzCloud (https://www.mzcloud.org/) and ChemSpider (http://www.chemspider.com/) database to obtain the accurate qualitative and relative quantitative results. Differential metabolites were identified as follows: (1) variable importance in the projection (VIP) in the partial least-squares discrimination analysis (PLS-DA model) >1.0; (2) the adjusted P-value was used for the non-parametric Mann–Whitney U-test (PASW Statistics 19, SPSS Inc., Chicago, IL, USA) in order to reduce the possibility of false positives with an adjusted P-value <0.05; and (3) PASW statistic 19 was used to calculate the value of AUC-ROC curve (the area under the receiver operating characteristic curve). Metabolites with AUC-ROC≤0.75 were discarded. Metabolites with fold change (FC) >1.5 or ≤0.667 were considered to be the differential metabolites. Changes in metabolites in each group were evaluated by the volcano map. The Kyoto Encyclopedia of Genes and Genome (KEGG) database was used to perform the enrichment analysis and pathway analysis of differential metabolites.

The whole small intestine was removed immediately after the bird was killed. The preparation of tissue samples used for histological observations was performed as described by Lv et al. (2014). Briefly, about 1 cm of intestine cut from the intermediate section of the duodenum was fixed in 10% buffered neutral formalin and processed for paraffin wax sectioning. Slices of approximately 5 µm thickness were stained with hematoxylin and eosin. Each slide was scanned with a scanner (Pannoramic MIDI, 3DHISTECH, Budapest, Hungary), and the digital image generated was then processed with K-ViEWER software. The wall area (section area minus lumen area), wall thickness (the difference between the outer and inner radius of the cross-section), mucosa area, mucosa thickness, number of columnar epithelial cells per 100 µm length along the longitudinal axis of a villus, and the proportion of villi with branches and villus length (the curvilinear distance from the top to the bottom along the longitudinal axis of a villus, including branches if present) were measured with ImageJ (version 1.8.8) software.

To measure the activity of digestive enzymes, the duodenum was first sliced open in a longitudinal direction to expose the inside of the tract, which was subsequently washed in normal saline to remove all contents. After that, the specimen was weighed to the nearest 0.1 mg and then homogenized in 9 volumes of normal saline. The homogenate was centrifuged at 500 g for 5 min and then the supernatant was collected. These steps were conducted at 4°C. Finally, the activity of sucrase (EC 3.2.1.48), maltase (EC 3.2.1.20), lipase (EC 3.1.1.3) and aminopeptidase-N (EC 3.4.11.2) in the extract was measured using commercial kits (Jiancheng Biotech Co. Ltd, Nanjing, China, art. no. A082-2, A082-3, A054-1-1, E009-1-1) according to the manufacturer's protocol.

The data were analyzed using SPSS (version 20.0 for Windows). The distribution of all variables was tested for normality using the Kolmogorov–Smirnov test. Non-normally distributed data were transformed into natural logarithms. Within each group, the body mass of each bird taken before acclimation (initial) was compared with the body mass taken after acclimation (final) using the paired sample t-test. Between the 10°C and 30°C groups, the body mass of the birds taken before and after acclimation was compared using an independent sample t-test. In order to test the differences in BMR, we tested the effects of temperature acclimation, captivity and their interaction in a general linear model, where the temperature acclimation was either cold (10°C acclimation) or warm (30°C acclimation) and captivity was ‘yes’ for after acclimation (final) or ‘no’ for before acclimation (initial). BMR was analyzed with body mass as a covariate. Tissue/organ (pectoral muscle, liver, kidney and heart) mass was also analyzed with body mass minus organ mass as a covariate to avoid statistical problems with part-whole correlations (Christians, 1999). Differences in other variables between the 10°C and 30°C groups were all evaluated using independent sample t-tests. All results are expressed as means±s.e.m., and P<0.05 was considered statistically significant.

Prior to acclimation, body mass did not differ significantly between the 30°C and 10°C groups (t₁₄=0.216, P=0.832). Temperature treatment did not have a significant influence on body mass between the two groups either (t₁₄=0.228, P=0.823). However, the results of paired sample t-tests showed that the body mass of birds was significantly reduced both in the 10°C group (t₇=4.146, P=0.004) and the 30°C group (t₇=3.599, P=0.009) following acclimation (Fig. 1A).

With body mass as a covariate, BMR was significantly affected by both captivity (F_1,27=7.150, P_C=0.013; Fig. 1B) and the interaction between temperature acclimation and captivity (F_1,27 =37.502, P_T×C<0.001; Fig. 1B). Although BMR was not significantly affected by temperature acclimation, the P-value approached statistical significance (F_1,27=4.000, P_T=0.056; Fig. 1B). Birds acclimated for 4 weeks underwent a significant decrease in BMR compared with the initial stage. After acclimation, the birds in the 10°C group showed a significantly higher BMR compared with those in the 30°C group when body mass was used as a covariate. These results suggest that warm acclimation decreased BMR to a greater extent than cold acclimation in these captive birds.

As shown in Table 1, significant differences in the mass of pectoral muscle and small intestine were detected between the 30°C and 10°C groups. The birds in the 10°C group had significantly heavier intestines than those in the 30°C group. However, their pectoral muscles were significantly lighter compared with those in the 30°C group. No significant differences in liver, heart and kidney mass were found between the 30°C and 10°C groups.

Mass-specific S4R (µmol O₂ min⁻¹ g⁻¹ tissue) in the liver (t₁₄=2.303, P=0.037; Fig. 2B), heart (t₁₄=2.527, P=0.024; Fig. 2C) and kidney (t₁₄=3.724, P=0.002; Fig. 2D) varied significantly between the two groups, with higher levels in the 10°C group; in the pectoral muscle, the increase in S4R was not significant compared with the 30°C group (t₁₄=0.98, P=0.344; Fig. 2A). Similarly, the birds in the 10°C group showed significantly higher whole-organ S4R (µmol O₂ min⁻¹) for the heart (t₁₄=2.444, P=0.028; Fig. 2C) and kidney (t₁₄=3.138, P=0.007; Fig. 2D); however, whole-organ S4R for the pectoral muscle (t₁₄=−0.165, P=0.869; Fig. 2A) and liver (t₁₄=0.894, P=0.386; Fig. 2B) did not change markedly between the two groups.

Mass-specific CCO activity (µmol O₂ min⁻¹ g⁻¹ tissue) in the liver (t₁₄=4.007, P=0.001; Fig. 2F) and kidney (t₁₄=3.314, P=0.005; Fig. 2H) showed significant variation, with a higher level in the 10°C group compared with the 30°C group; however, CCO activity in the pectoral muscle (t₁₄=1.705, P=0.110; Fig. 2E) and heart (t₁₄=0.478, P=0.640; Fig. 2G) did not differ between the two groups. Whole-organ CCO activity (µmol O₂ min⁻¹) in the liver (t₁₄=2.174, P=0.047; Fig. 2F) and kidney (t₁₄=3.472, P=0.004; Fig. 2H) showed significant variation between groups, with a higher level in the 10°C group compared with the 30°C group, whereas CCO activity in the pectoral muscle (t₁₄=0.931, P=0.368; Fig. 2E) and heart (t₁₄=0.253, P=0.804; Fig. 2G) did not differ between the two groups.

As shown in Table 2, a total of 258 (negative polarity mode) or 232 (positive polarity mode) metabolites from the pectoral tissue were annotated. These metabolites are mainly involved in five types of biological functions: cellular processes, environmental information processing, genetic information processing, metabolism and organismal systems. In metabolism, global and overview maps, lipid metabolism, amino acid metabolism, metabolism of cofactors and vitamins, and metabolism of other amino acids were the top five pathways acquired from both the negative polarity and the positive mode.

In order to investigate the temperature-induced changes in metabolic signatures in the pectoral muscle tissue of white-shouldered starlings, the differences in the levels of small molecular metabolites were analyzed. Under the positive polarity mode, a total of 80 differentially enriched metabolites were identified. Compared with the 30°C group, 42 metabolites were significantly down-regulated, while 38 metabolites were up-regulated in the 10°C group (Fig. 3A). Under the negative polarity mode, a total of 78 differentially enriched metabolites were identified, and 47 metabolites were significantly down-regulated while 31 metabolites were up-regulated in the 10°C group (Fig. 3B).

Some of the differential metabolites could be annotated by the KEGG database, and the results are shown in Table 3. Compared with the 30°C group, a total of 16 metabolites were up-regulated, whereas 17 metabolites were down-regulated in the 10°C group. Among these, 12 were related to lipid metabolism, eight were related to amino acid metabolism, three to purine metabolism, and one to biotin metabolism.

For lipid metabolism, the 10°C-acclimated birds displayed changes in the fatty acid metabolism pathway that contrasted with those for the 30°C-acclimated birds. These changes included down-regulated levels of 3-oxopalmitoyl-coenzyme A (CoA) and (S)-3-hydroxydecanoyl-CoA and up-regulated levels of hexanoyl-CoA and butyryl-CoA in the 10°C-acclimated birds (Fig. 4Aa and Table 3). Changes that occurred in the 10°C-acclimated birds that were specific to the synthesis pathway of unsaturated fatty acids included significantly lower levels of erucic acid, 14(Z)-eicosenoic acid, (13Z,16Z)-docosadienoic acid and 11(Z),14(Z)-eicosadienoic acid (Table 3 and Fig. 4Ab). Other changes displayed by the 10°C-acclimated birds were a higher level of jasmonic acid in the pathway of α-linolenic acid metabolism (Table 3 and Fig. 4Ac), and significantly lower levels of (10E,12Z)-9-hydroperoxy-10,12-octadecadienoic acid and 13S-hydroxyoctadecadienoic acid in the pathway of linoleic acid metabolism (Table 3 and Fig. 4Ad).

For amino acid metabolism, the 10°C group also showed higher levels of uric acid, l-glutamic acid, l-saccharopine, α-aminoadipic acid, isovaleryl-CoA, 3-methylbut-2-enoyl-CoA and 4-guanidinobutyric acid in multiple pathways, including purine metabolism, glutathione metabolism, lysine degradation, valine, leucine and isoleucine degradation, and arginine and proline metabolism (Table 3 and Fig. 4B).

For purine metabolism, the 10°C-acclimated group of birds exhibited a significant increase in the level of adenine and a significant decrease in the level of guanosine monophosphate. These birds also exhibited a marked decrease in the level of biocytin in the case of biotin metabolism (Table 3 and Fig. 4C).

Prior to acclimation, no significant difference in energy budget was found between the two groups. However, after 4 weeks of acclimation, significant increases in GEI (t₁₄=15.021, P<0.001; Fig. 5A), FE (t₁₄=11.172, P<0.001; Fig. 5B) and DEI (t₁₄=15.516, P<0.001; Fig. 5C) were observed for the group acclimated to 10°C compared with the group acclimated to 30°C. In contrast, the digestive efficiency decreased significantly for the group acclimated to 10°C compared with the group acclimated to 30°C (t₁₄=−4.451, P=0.001; Fig. 5D).

Seven microstructural indexes were measured to evaluate the morphological differences in the duodenum between the two groups of birds. The wall area and the mucosa area of the duodenum in the 10°C group were, respectively, 33% and 42% larger after 4 weeks of acclimation, although these increases were not statistically significant relative to the 30°C group (see Table 1 for statistics). The duodenum of the 10°C group (Fig. 6E,F) appeared to have the same wall thickness, mucosa thickness and the number of columnar epithelial cells per 100 µm of villus length as in the 30°C group (Fig. 6A,B, Table 1). Although the proportion of villi with branches in the 10°C group was 1.35 times higher than that in the 30°C group after acclimation, no significant difference was found between the two groups in this regard (Table 1). However, the villus length of the duodenum in the 10°C group was significantly greater compared with that of the 30°C group (Table 1). Furthermore, we did observe noticeable differences in the structural characteristics of the duodenum between the two groups of birds in the following ways. (1) A single villus of the small intestine was repeatedly curved at the base and extended in a circular direction, and then it formed two or more branches after being turned inward toward the small intestinal cavity (Fig. 6G versus C). (2) The adjacent small intestinal villi combined with each other in the vertical direction, forming an H-shaped structure that made up a network (Fig. 6H versus D).

In order to indicate the ability of the small intestine to digest and assimilate carbohydrates, proteins and lipids, we measured sucrase (hydrolyzing sucrose), maltase (hydrolyzing maltose), aminopeptidase-N (digesting small peptides) and lipase activity. At the mass-specific level (µmol min⁻¹ g⁻¹ tissue), maltase (t₁₄=2.463, P=0.027; Fig. 7B) showed significant variation in activity, with a higher activity level in the 10°C group compared with the 30°C group. In contrast, mass-specific aminopeptidase-N (t₁₄=2.251, P=0.041; Fig. 7C) in the 30°C group was significantly higher than that in the 10°C group. No significant difference in mass-specific sucrase (t₁₄=0.339, P=0.739; Fig. 7A) or lipase (t₁₄=1.338, P=0.202; Fig. 7D) was found between the two groups. At the whole-organ level, birds acclimated to 10°C showed a significantly higher level of maltase activity (t₁₄=3.602, P=0.003; Fig. 7B) compared with those acclimated to 30°C. However, no significant difference in the activity of sucrase (t₁₄=0.797, P=0.439; Fig. 7A), aminopeptidase-N (t₁₄=−0.578, P=0.573; Fig. 7C) or lipase (t₁₄=1.986, P=0.067; Fig. 7D) was found between the two groups, although the 10°C group seemed to have higher total sucrase and lipase activity compared with the 30°C group. These results suggest that cold acclimation could change the activity of digestive enzymes in the duodenum in white-shouldered starlings.

Ambient temperature is considered to be a key cue for thermoregulation and a driving force for the evolution of a range of adaptations for birds, from biochemical to morphological (Deville et al., 2014; Klaassen et al., 2004; Liknes and Swanson, 2011; Petit et al., 2014). To test the ‘energy demand’ hypothesis, which states that when a small bird is acclimated to different temperatures, multiple levels of organization are involved in the adjustment of its energy demand, various indicators ranging from the organismal level to the biochemical level were evaluated for white-shouldered starlings acclimated to 10°C and 30°C. Temperature acclimation was found to change body, pectoral muscle and small intestine mass, S4R level in the liver, heart and kidney, and CCO activity in the liver and kidney. Untargeted metabolomics analysis found that under low-temperature conditions, 16 metabolites were up-regulated, while 17 metabolites were down-regulated, and these metabolites were related to amino acid metabolism, fat metabolism, purine metabolism and biotin metabolism. In addition, GEI and FE in the energy budget, villus length and maltase activity in the duodenum were also found to increase significantly in the cold-acclimated group. Changes in these physiological parameters may be the reason for the change in BMR during temperature acclimation.

A cold ambient temperature is one of the significant environmental factors threatening the survival of small birds (Zhou et al., 2016). Numerous studies have demonstrated that small birds have developed a suite of morphological, physiological, biochemical and behavioral strategies to help them to survive at low temperatures (Deville et al., 2014; Hammond et al., 1999; Klaassen et al., 2004; Liknes and Swanson, 2011). Enhancing BMR is one of the physiological responses of some birds under cold acclimation or acclimatization (McKechnie et al., 2007; Swanson, 2001; Vézina et al., 2006; Zhou et al., 2016). For example, Chinese hwamei (G. canorus) acclimated to 15°C showed a higher BMR level than those kept at 35°C after 1–4 weeks (Zhou et al., 2016) and, similarly, the BMR of red knots (C. canutus) acclimated to 4°C was 27% higher compared with acclimation at 25°C (Vézina et al., 2006). In this study, BMR was significantly affected by both captivity and the interaction between temperature acclimation and captivity, but not by temperature acclimation alone. Compared with the initial values, the BMR measured after treatment was significantly reduced for both the 10°C and 30°C groups (Fig. 1B). It cannot be ignored that many previous studies have indeed reported that keeping birds in cages can cause various changes, and these include changes in stress levels, changes in body composition and organ quality caused by captive diets, and reduced flight activity (Piersma et al., 1996; Al-Mansour, 2004; McKechnie et al., 2007). Consistent with the above, we found that the body mass of white-shouldered starlings was also affected by captivity as both groups of birds exhibited significantly reduced body mass after treatment (Fig. 1A). Therefore, the significant reduction in BMR after treatment was partly due to the loss of body mass caused by captivity. Moreover, the birds in the 10°C-acclimated group showed a significantly higher BMR by day 28 compared with the birds in the 30°C-acclimated group, indicating that positive modulation of basal metabolism occurred in the birds during temperature acclimation. A change in BMR during temperature acclimation is usually the result of a change in thermogenic capacity (McKechnie and Swanson, 2010; Rezende et al., 2004; Swanson, 2010). Thus, the effect of temperature on the BMR of white-shouldered starlings appeared to be mediated by the variable energy demands used for thermoregulation, consistent with what has been proposed by previous investigators (Swanson et al., 2017).

S4R is an indicator of mitochondrial proton leak capacity, whereas CCO is a key regulatory enzyme in oxidative phosphorylation, and its level is a direct indication of cellular metabolism (Madelaire et al., 2020; Pena-Villalobos et al., 2014; Swanson et al., 2017; Wikstrom et al., 2018). S4R and CCO activity are commonly used as indicative markers of BMR at the cellular level because they contribute significantly to cellular aerobic metabolism and overall metabolic rate. This thermogenesis mainly occurs in the visceral organs such as the liver, heart, brain and kidney. Although these internal organs represent less than 10% of the body mass, they account for 50–70% of the energy expenditure in the resting state (Clapham, 2012; Zeng et al., 2022). In the case of white-shouldered starlings, the birds that were acclimated to 10°C displayed higher levels of S4R in the heart and kidney and enhanced levels of CCO activity in the liver and kidney compared with those acclimated to 30°C (Fig. 2). The results are consistent with those of other small birds, including but not limited to Chinese bulbuls (P. sinensis) (Zheng et al., 2014a), red-billed leiothrix (L. lutea) (Cui et al., 2019), Eurasian tree sparrow (P. montanus) (Li et al., 2020) and Chinese hwamei (G. canorus) (Zhou et al., 2016), indicating that alterations in the level of cellular metabolism in the liver, kidney and heart could contribute to significant changes in BMR.

In volant birds, the skeletal muscle typically comprises nearly 40% of the body mass, and it is an important contributor to thermogenesis (Weber and Piersma, 1996; Swanson et al., 2014; Zheng et al., 2008b, 2014a). Zhang et al. (2015) found that house sparrows (Passer domesticus) had heavier pectoral muscles after cold acclimation. Previous studies conducted by our group found that some small birds do not show significant differences in the pectoral muscle mass but display marked changes in S4R and/or CCO activity in the pectoral muscle during temperature acclimation or seasonal acclimatization. The birds that were captured during winter or those acclimated to low temperature (cold acclimation) were found to exhibit higher levels of S4R and/or CCO activity compared with those captured during summer or acclimated to warm temperature (warm acclimation), including the Eurasian tree sparrow (P. montanus) (Li et al., 2020; Hu et al., 2017), Chinese bulbuls (P. sinensis) (Zheng et al., 2014a), Chinese hwamei (G. canorus) (Zhou et al., 2016) and red-billed leiothrix (L. lutea) (Cui et al., 2019). Unlike these results, cold-acclimated white-shouldered starlings had significantly lighter pectoral muscles (Table 1). For these birds, neither the S4R level nor CCO activity changed significantly in the pectoral muscle after acclimation to 10°C compared with acclimation to 30°C. However, the birds acclimated to 10°C showed a slight increase in CCO activity in the pectoral muscle compared with those acclimated to 30°C. To find out whether the pectoral muscle might contribute to the phenotypic flexibility in metabolic adjustments during temperature acclimation, we further examined the effect of temperature on metabolite levels in the pectoral muscle.

Energy-rich lipids are the main source of fuel for avian flight muscles during various strenuous movements, and the lipid pools provide 80% of the energy for shivering in cold conditions in birds (Vaillancourt et al., 2005). Our data appeared to show that lipid catabolism in the pectoral tissue was increased while lipid anabolism was reduced in the 10°C-acclimated birds compared with the 30°C-acclimated ones, and various lines of evidence supporting this claim were obtained. (1) Down-regulated levels of 3-oxopalmitoyl-CoA and (S)-3-hydroxydecanoyl-CoA, and up-regulated levels of hexanoyl-CoA and butyryl-CoA (Table 3 and Fig. 4Aa), which are involved in fatty acid metabolism, were found. As 3-oxopalmitoyl-CoA and (S)-3-hydroxydecanoyl-CoA are long-chain fatty acyl CoAs, their down-regulated levels suggested decreased synthesis of fatty acids. In contrast, hexanoyl-CoA and butyryl-CoA are short-chain fatty acyl CoAs, and their upregulated levels suggested increased breakdown of fatty acids. This could suggest a lower level of fatty acid synthesis coupled with a higher level of degradation occurred in the 10°C-acclimated birds compared with the 30°C-acclimated ones. (2) Erucic acid, 14(Z)-eicosenoic acid, (13Z,16Z)-docosadienoic acid and 11(Z), 14(Z)-eicosadienoic acid all take part in the synthesis of polyunsaturated fatty acids. Their levels decreased significantly in the 10°C-acclimated birds (Table 3 and Fig. 4Ab), indicating that the synthesis of unsaturated fatty acids in the pectoralis was reduced in the 10°C-acclimated birds compared with the 30°C-acclimated ones. (3) Jasmonic acid, the end-product of the degradation of α-linolenic acid, showed a higher level in the 10°C-acclimated group (Table 3 and Fig. 4Ac), indicating that the degradation of α-linolenic acid in the cold-acclimated group was significantly enhanced compared with that in the warm-acclimated group. Similar to our results, the abundance of cytosolic fatty acid binding protein (FABP) in the pectoralis of Poecile atricapillus and Sitta carolinensis has also been found to increase in the winter compared with the summer (Liknes et al., 2014). In addition, the genes involved in fatty acid oxidation in the pectoralis of the dark-eyed junco (Junco hyemalis) have been documented to be upregulated during the winter (Stager et al., 2016). Taken together, these findings suggest that the relative differences in the levels of metabolites involved in lipid oxidation could indicate a change in the lipid oxidation capacity in the pectoralis of small birds subjected to temperature acclimation.

In this study, increases in the concentrations of various products of amino acid catabolism were detected from the 10°C-acclimated group compared with the 30°C-acclimated group. These included uric acid, l-glutamic acid and specific metabolites of leucine, lysine and arginine (Table 3 and Fig. 4B). Uric acid is a marker of amino acid catabolism in birds, whereas l-glutamic acid is an important carrier for the combined deamination in the degradation of amino acids (Hird and Marginson, 1966). l-Saccharopine and α-aminoadipic acid, isovaleryl-CoA and 3-methylbut-2-enoyl-CoA, and 4-guanidinobutyric acid are the degradation products of lysine, leucine and arginine, respectively. Taken together, these results indicate that temperature acclimation could lead to a change in amino acid catabolism. Although amino acids can be oxidated to provide energy by entering the tricarboxylic acid cycle (Powar and Chatwal, 2008; Wu, 2013), lipids are the main substrate for energy metabolism in birds during cold exposure (Klaassen et al., 1989). However, Vaillancourt et al. (2005) reported that the total ATP produced in the ruff (Philomachus pugnax) is unequally shared among lipids (82%), carbohydrates (12%) and proteins (6%) when the bird is exposed to cold. Additionally, it has been reported that humans use mixed fuels under cold exposure, in which proteins normally play a minor but significant role (∼10% of the heat produced) (Weber, 2011). Based on these results, we speculate that changing the degradation of amino acids might be one of the adaptive strategies for white-shouldered starlings to adapt to temperature acclimation, a feature that could help the birds successfully cope with a wide range of fuel supplies. Contrary to the above results, the level of 3-methylhistidine was significantly lower in the 10°C group (Table 3). 3-Methylhistidine is thought to be a biomarker of skeletal muscle protein breakdown (Abasht et al., 2016; Sansbury et al., 2014; Zampiga et al., 2018). The reason for this observation needs further study.

Some metabolites involved in purine and biotin metabolism were also differentially enriched in white-shouldered starlings under cold acclimation, suggesting that the energy supply level of the pectoral muscle was higher during cold acclimation than during warm acclimation. Adenine is one of the end products of adenosine triphosphate (ATP) breakdown (Dzeja et al., 1998; Shimamoto et al., 2020). Therefore, the increased concentration of adenine in the 10°C group might be the result of enhanced ATP hydrolysis (Table 3 and Fig. 4Ce). This is consistent with the hypothesis that in birds, temperature acclimation induces a change in thermogenesis. Equally important was the decreased level of guanosine monophosphate (the product of GTP decomposition) under cold acclimation (Table 3 and Fig. 4Ce), which might suggest an increase in the synthesis of GTP. GTP is an energy carrier during the conversion of succinic acid CoA to succinic acid in the Krebs cycle (Davey, 1961). Additionally, the content of biocytin (the degradation product of biotin) was reduced significantly under cold acclimation (Table 3 and Fig. 4Cf), suggesting an enhanced utilization of biotin. The importance of biotin in animal metabolism is, therefore, due to the fact that it is a coenzyme for many vital processes, such as gluconeogenesis, biosynthesis of fatty acids and amino acid metabolism (Friedrich, 2013; Zempleni et al., 2009).

Taken together, these results suggest that white-shouldered starlings might regulate multiple pathways such as lipid metabolism, amino acid metabolism, and purine and biotin metabolism in the skeletal muscle to ensure a flexible level of metabolism under varying temperature conditions to help them maintain a balance in energy homeostasis.

Energy balance plays a crucial role in animal survival, reproduction, evolution, determination of food intake, digestion and heat production, among others (Daan et al., 1990; Kersten and Piersma, 2015). A number of studies have shown that birds can adapt to an increased energy requirement by increasing both energy intake and the efficiency of food usage at low temperatures (Williams and Tieleman, 2000). In this study, GEI and DEI increased by 124% and 81%, respectively, for the birds in the 10°C group compared with those in the 30°C group (Fig. 5A,C), indicating that the energy budget of white-shouldered starlings was significantly affected by ambient temperature. These results are consistent with the adaptive changes in many other small birds in response to cold temperatures (Goymann et al., 2006; Salvante et al., 2010; Syafwan et al., 2012; Zhou et al., 2016). For example, Cui et al. (2019) found that red-billed leiothrix (L. lutea) acclimated to 15°C increased their energy intake and DEI by 81% and 40%, respectively, compared with those acclimated to 35°C for 2 weeks. Zhou et al. (2016) also showed that Chinese hwamei (G. canorus) kept at 15°C consumed more food than those kept at 35°C (a temperature within the thermoneutral zone of this species). These findings, together with our data, suggest that increased energy intake and utilization may be an important adaptive strategy to cold conditions for small birds (Goymann et al., 2006; Piersma, 2002; Vézina et al., 2006; Wu et al., 2014).

The digestive tract of birds and small mammals can also be adjusted by changing its morphological features, nutrient transport mechanisms and other physiological indicators (Karasov and McWilliams, 2005; McWhorter et al., 2009; Sailer and Fietz, 2009) to maximize the energy release from ingested food and to supply more energy for thermogenesis, reproduction, migration or other physiological processes (Cavieres and Sabat, 2008; Khalilieh et al., 2012; Piersma and Drent, 2003). In northeastern China, tree sparrows (P. montanus) tend to show greater luminal and wall cross-sectional area in spring and autumn than in summer. The villus length reaches a maximum in winter and shows similar seasonal dynamic changes to those of mucosal thickness (Lv et al., 2014). We also observed this phenomenon in white-shouldered starlings under cold acclimation, such that the villus length was significantly higher for birds acclimated to 10°C (Table 1). Furthermore, a single villus in the mucosa layer seemed to have a significantly greater total length by extending in a circular direction and forming more branches (Fig. 6G). Moreover, adjacent small intestinal villi (Fig. 6H) combined to form a network in the form of an H-shaped structure. These changes could be beneficial through improving the total surface area for digestion and absorption in the duodenum under a cold-temperate climate in order to maximize the energy released from ingested food and satisfy the higher level of thermogenesis. However, the lack of significant changes for other indicators in Table 1 may be related to the fact that all the birds in the experiments ate the same food.

The adaptation of digestive enzyme activity to changes in energy demands (e.g. lactation, exercise and thermogenesis) has been extensively studied in a variety of vertebrates (e.g. Karasov and Hume, 1997; Karasov, 1992). An increase in energy requirement often induces hyperphagia and an improvement in nutrient absorption (Karasov and Hume, 1997; Weiss et al., 1998). Therefore, according to the adaptive adjustment hypothesis, the activity of digestive enzymes should be positively correlated with the amount of diet the birds consume (Buddington et al., 1991; Ferraris and Diamond, 1989; Karasov and Diamond, 1988). Consistent with the adaptive modulation hypothesis, we found that the mass-specific activity and total activity of maltase in white-shouldered starlings were significantly upregulated in the 10°C-acclimated group compared with the 30°C-acclimated group. The total sucrase and lipase activities of starlings were higher in the 10°C group, although no significant difference was found between the two groups. This trend in enzyme activity was consistent with the levels of BMR, GEI and DEI. This suggests that starlings resorted to physiological adjustment via activity changes in digestive enzymes under temperature acclimation to cope with the significant differences in energy intake, and similar results have also been observed for other birds (McKechnie, 2008; McWilliams and Karasov, 2014). For example, the rufous-collared sparrow (Zonotrichia capensis) has been shown to have higher levels of total maltase and sucrase activity in the intestine in winter than in summer (Ramírez-Otarola et al., 2018), and the location where this study was carried out has a mean temperature of 8.3°C in winter and 18.9°C in summer. Similar changes in digestive enzymes associated with seasonal acclimatization and temperature acclimation have been observed in rodents such as Mongolian gerbils (Meriones unguiculatus) (Zhao et al., 2014) and striped hamsters (Cricetulus barabensis) (Liu et al., 2013).

Taken together, our results show that when white-shouldered starlings were exposed to temperature acclimation, a number of physiological and physical changes occurred, which included changes in pectoral muscle mass and small intestine mass, S4R and CCO activity in the liver, heart and kidney, as well as lipid oxidation, amino acid decomposition, and purine and biotin metabolism in the pectoral tissue. Furthermore, variations in the energy budget, villus length and total surface area of the duodenal mucosa, and maltase activity were also prominent. Significant changes in all of the above indicators may have led to changes in the BMR during temperature acclimation. Thus, white-shouldered starlings could deploy a wide array of adaptive strategies at multiple levels (including at the whole-organism, organ, cellular and molecular level) to acclimatize to variable ambient temperatures. These results supported the hypothesis that temperature is an important environmental cue for the adaptive adjustment of energy metabolism and thermogenesis in birds.

The authors acknowledge Dr Alan K. Chang (Wenzhou University) for his helpful discussion and for editing the language of the manuscript. Two anonymous reviewers supplied their valuable comments on the manuscript. We also thank Tianjin Novogene Bioinformatic Technology Co., Ltd (China) for their technical support with the LC-MS/MS.