Dietary macronutrients influence 13 C and 15 N signatures of pinnipeds: Captive feeding studies with harbor seals (Phoca vitulina)

Transcription

1 Comparative Biochemistry and Physiology, Part A 143 (2006) Dietary macronutrients influence 13 C and 15 N signatures of pinnipeds: Captive feeding studies with harbor seals (Phoca vitulina) Liying Zhao, Donald M. Schell, Michael A. Castellini Institute of Marine Science, School of Fisheries and Ocean Sciences, University of Alaska Fairbanks, P. O. Box , Fairbanks, AK , USA Received 23 August 2005; received in revised form 22 December 2005; accepted 22 December 2005 Available online 3 February 2006 Abstract Metabolic effects of dietary macronutrients on diet tissue isotopic discrimination factors were investigated in harbor seals. Three seals were fed either high fat/low protein herring (H), or low fat/high protein pollock (P), and switched to the alternative every 4 months. This allowed each seal to be subjected to two dietary treatments in each of three metabolically defined seasons (breeding from May to September, molting from September to January, and late winter/early spring period from January to May) over a 2 year cycle, and function as its internal control regardless of physiological changes over season. One seal was fed a constant equal mix of H and P over the entire trial. Up to 1 differences in serum δ 15 N values of one seal fed alternatively on H and P were observed. Progressively more enriched serum δ 15 N values as diet switching from H to P might link to changes in seal digestive physiology and protein metabolism in response to very high protein intake on P diet. These findings demonstrate that dietary macronutrients of prey species and protein intake level of consumers also play important roles in shaping isotopic patterns of a consumer's tissues, and thus influence accurate data interpretation of stable isotope techniques in ecological applications Elsevier Inc. All rights reserved. Keywords: Controlled feeding trials; Dietary macronutrients; Diet tissue isotope discrimination factors; Harbor seals; Herring; Isotope ratios; Metabolism; Physiology; Pollock; Protein intake 1. Introduction Analysis of naturally occurring stable isotope ratios of carbon ( 13 C/ 12 C) and nitrogen ( 15 N/ 14 N) in animals' tissues and their food sources has become a powerful tool in ecological research of wildlife. One basic assumption underlying the use of stable isotope analysis in trophic dynamics and feeding ecology is that a predictable relationship in stable isotopic composition exists between consumers and their diet. However, except for commonly known natural abundance variability in stable carbon and nitrogen isotope ratios arising from ecological processes, such as changes in habitats or food sources due to seasonal migrations or prey species switching, evidence is accumulating that metabolic and physiological processes also play important roles in shaping isotopic patterns Corresponding author. Present address: Department of Animal Sciences, 132 ASL MC 630, 1207 W Gregory Dr., University of Illinois at Urbana- Champaign, Urbana, IL 61801, USA. Tel.: ; fax: address: lyzhao@uiuc.edu (L. Zhao). of an animal's tissues, and thus influence accurate estimation of diet tissue isotopic discrimination factors. For instance, previous studies have shown that diet tissue isotopic discrimination factors vary (1) with different species of consumers (DeNiro and Epstein, 1978, 1981; Macko et al., 1982; Minagawa and Wada, 1984; Hobson and Clark, 1992; Focken and Becker, 1998), (2) within same species under different physiological conditions, such as fasting or starvation (Hobson et al., 1993; Best and Schell, 1996), nursing or weaning (Polischuk et al., 2001), growth or aging (Overman and Parrish, 2001), extreme environmental stress (reviewed by Ambrose, 1991), and (3) within same individual raised on different diets with differing food quality and feeding levels (Webb et al., 1998; Fantle et al., 1999; Adams and Sterner, 2000; Pearson et al., 2003; Gaye-Siessegger et al., 2004). Variations in diet tissue isotopic discrimination occurred during biochemical reactions, associated with changes in animal metabolism and digestive physiology, have greatly complicated data interpretation of stale isotope techniques in ecological applications, and led to a general call for laboratorycontrolled studies (Gannes et al., 1997) /$ - see front matter 2006 Elsevier Inc. All rights reserved. doi: /j.cbpa

2 470 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) In recent years, stable isotope techniques have been increasingly used for studies of migration patterns and foraging ecology of pinnipeds (Hobson et al., 1996, 1997; Hobson and Sease, 1998; Burton and Koch, 1999; Hirons et al., 2001; Kurle and Worthy, 2001; Kurle, 2002). Accurate isotope data interpretation of field studies requires knowledge of both natural isotopic variability arising from ecological processes and reliable values of diet tissue isotopic discrimination factors for specific species being studied and specific tissues being analyzed (Gannes et al., 1997, 1998). As diet tissue isotopic discrimination factors are species-specific, captive studies with pinnipeds under laboratory-controlled conditions become a prerequisite, and will provide a useful model for field ecological research. Previous studies have demonstrated that physiology and metabolism of pinnipeds change significantly with season (Fadely et al., 1998). Seasonal physiological changes during molting, breeding and late winter/late spring period may cause dramatic changes in body mass and condition. This may in turn affect 13 C and 15 N patterns in pinnipeds, and ultimately influence accurate estimation of diet tissue isotopic discrimination factors. Hence, it is important to design a controlled feeding experiment that allows one to compare different dietary treatments under similar physiological conditions. Harbor seals Phoca vitulina are small pinnipeds that are widely distributed throughout the coastal regions of the Northern Hemisphere. Populations of harbor seals, Steller sea lions, and several other marine mammal species in Alaska have declined dramatically over the past 3 decades (Pitcher, 1990; Frost et al., 1994). Changes in forage fish assemblages from predominantly small fatty fish species, such as Pacific herring Clupea pallasi, to low fat/low energy density fish species, such as walleye Pollock Theragra chalcogramma, might have caused nutritional deficiency in seals and sea lions, and partially contributed to the population declines (Alaska Sea Grant, 1993). As part of studies of harbor seal health and diet (Castellini et al., 2002), the aim of this study was to investigate dietary isotopic effects on diet tissue isotopic discrimination factors by examining relationships of carbon and nitrogen isotopic variability between captive seals and their different fish diets. Herring and pollock were fed as natural prey species of harbor seals that represent nutritionally different feeding regimes due to distinct composition in their lipid contents. Blood samples were collected from seals at 2 week intervals. Carbon and nitrogen isotope ratios in blood components, i.e. serum and red blood cells (RBC), were monitored throughout the entire feeding trial. As the most commonly measured tissues, blood can be sampled repeatedly on same subjects over time, and can thus be used to track temporal isotopic variations on different time scales, depending on turnover rates of specific components. For instance, serum/plasma has a very fast turnover rate of several days, whereas RBC has relatively a slow turnover rate of more than a month (Hobson and Clark, 1993; Hilderbrand et al., 1996). Hence, isotopic variations in serum/plasma and RBC samples reflect dietary and/or seasonal physiological changes of seals from several days to several months. For accurate estimation of diet tissue isotopic discrimination factors, carbon and nitrogen isotopic compositions in different batches of both fish species used during the feeding trial were completely examined in relation to proximate composition of lipid contents, relative body length/mass, and catch location/time. Herring and pollock feed at similar trophic level. Natural abundance of carbon and nitrogen isotope ratios may differ between the two fish species or among different batches of same species, depending on geographic catch location or/and catch time (Kline, 1999). 2. Materials and methods 2.1. Animal handling and feeding trials Controlled feeding trials were conducted from September 1998 to September 2000 at the Alaska SeaLife Center (ASLC) in Seward, Alaska. All the procedures were approved by the ASLC scientific committee, and conducted as required by the Institutional Animal Care and Use Committees of the University of Alaska and ASLC. Seals were housed in large outdoor pools and fed ad libitum predominantly on herring for 3 months to acclimate to new facility before a 2 year repeated crossover feeding trial started in September 1998 (Table 1). 2 fish species, Pacific herring (Clupea pallasii pallasii) and walleye pollock (Theragra chalcogramma), were used in the trial and total 4 seals were included in this study. As shown in Table 1, Pender and Travis, both 2.5 year old males, was initially fed herring, then pollock, and switched to herring every 4 months. Poco, a 23 year old female, was first fed pollock and switched to the alternative diet at the same time period as Pender, but in reverse order. Snapper, a 15 year old male, was fed on a constant equal mix of herring and pollock for the entire trial. Seals were fed 2 3 times per day to a level of satiation to allow them to adjust food intake naturally based on their metabolic energy requirements in response to seasonal physiological changes and changes of diets, and the intake levels usually ranged from 5 10% of their body mass per day. Table 1 Feeding schedule for Phoca vitulina Starting date 15 Sept Jan June Sept Jan May 00 Ending date 15 Jan May Sept Jan May Sept 00 Pender Travis Herring (H4) Pollock (P3) Herring (H4) Pollock (P3) Herring (H4) Pollock (P5) Poco Pollock (P2) Herring (H4) Pollock (P3) Herring (H4) Pollock (P3) Herring (H6) Snapper 1:1 H4/P2 1:1 H4/P3 1:1 H6/P5 Specific batches are shown in parentheses.

3 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) Diet fish 15 H4 H6 P2 P3 P5 Individually ice-glazed fresh frozen herring and pollock were purchased from Alaskan commercial fishery companies in large batches and stored at 20 C until the day of feeding. 2 batches of herring (H4 and H6) and 3 batches of pollock (P2, P3 and P5) were used for entire feeding trials as specified in Table 1. Individual batches of herring and pollock were sampled periodically (n=10 from each batch and each species, at least once during every 4 months feeding period). Standard body length/mass of each specimen was measured prior to proximate analysis. Briefly, whole fish specimen was homogenized with a food processor or blender, sub-sampled approximately 10 g, then freeze-dried to constant mass and water content calculated on the basis of mass difference. Lipid content was determined by Soxhlet extraction with 2: 1 chloroform / methanol solvents. Energy density was determined using an adiabatic bomb calorimeter (Parr Co) (see Castellini et al., 2002 for details). Information of water and lipid contents, body length and mass, energy density, catch location and time, as well as sample size of 5 batches of diet fish are summarized in Table Stable isotope analysis Harbor seal blood samples were collected at 2 week intervals following an overnight fast. Serum (plain tubes, no additive added) and unclotted RBC were separated by clinical centrifugation immediately after blood withdrawal and frozen at 80 C. Serum and RBC samples were freeze-dried and ground to a powder for homogeneity for isotopic analysis. Fur, nails, whiskers were also collected periodically and washed in distilled water in an ultrasonic cleaner, dried, and then cut to tiny pieces for isotopic analysis. Samples of homogenized and freeze-dried whole fish were collected for isotopic analysis of fish diets. Lipids were extracted in some of the above homogenized and freeze-dried whole fish samples from both species by using Bligh and Dyer's (1959) method as described by Isik et al. (1999). In addition, lipids were extracted from 5 specimens of each species (H4 and P3) in duplicate using Bligh and Dyer's (1959) and Radin's (1981) methods, respectively, to compare solvent effects of the two extraction methods on lipid-free δ 13 C and δ 15 N values. Carbon and nitrogen isotope ratios were measured using a Costech ECS4010 elemental analyzer coupled with an isotope ratio mass spectrometer (Finnigan Delta plus XL). Analytical precision of peptone standards was δ 15 N ± 0.1 for carbon and ± 0.2 for nitrogen. Results are reported using standard δnotation in parts per thousand ( ) relative to Pee Dee Belemnite (VPDB) for δ 13 C and atmospheric N 2 for δ 15 N Statistical analysis One-way ANOVA with Tukey Kramer's pairwise comparisons were conducted to examine intra-specific and interspecific differences in δ 13 C and δ 15 N values among herring and pollock. Paired t-tests were used to test if differences existed in δ 13 C and δ 15 N values between serum and RBC, between lipidfree and untreated fish samples, and between two lipid extraction methods. Effects of diet (herring, pollock and an equal mix of herring and pollock), season (breeding, molting and late winter/early spring period) and interaction between diet and season on serum δ 13 C or δ 15 N values were tested by repeated measures ANOVA. Data from the last feeding period from May 2000 to September 2000 did not included in the above analysis due to poor confidence in isotopic composition of fish batch H6 and P5. 3. Results δ 13 C Fig. 1. The δ 13 C and δ 15 N values (mean±sd) in whole fish samples. H4, n=69; H6, n=10; P2, n=15; P3, n=30; P5, n= C and 15 N isotopic composition in fish diets As Fig. 1 shown, δ 13 C values in whole fish samples spanned a much wider range in herring than those in pollock, whereas δ 15 N values were more variable in pollock than in herring. Comparisons among 5 batches of 2 fish species (H4, H6, P2, P3 Table 2 Water and lipid contents, body mass and length, and energy density (mean±sd) of diet fish Batch Water content (%, wet mass) Lipid content (%, wet mass) Body mass (g) Body length (cm) Energy density (kj g 1, wet mass) n Catch location and date H4 65.8± ± ±32 22± ± Prince William Sound, Nov 98 H6 66.0± ±0.9 87±19 19± ± Petersburg, AK, Dec 99 P2 76.6± ± ±129 36± ± The Gulf of Alaska, Mar 98 P3 77.2± ± ±89 38± ± The Gulf of Alaska, Jan 99 P5 78.2± ± ±39 29±3 4.7±0.1 6 Unknown (donated) Data were obtained from Castellini et al. (2002).

4 472 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) δ 15 N δ 13 C Bligh & Dyer Radin Untreated H4 H4 and P5) produced 3 distinct groups. The highest δ 13 C and δ 15 N values occurred in P2, with intermediate values in P3, P5 and H6 for δ 13 C and P3, P5 and H4 for δ 15 N, and the lowest δ 13 C values in H4 and δ 15 N values in H6, respectively (ANOVA, F 4, 124 =71.34, Pb for δ 13 C, F 4, 124 =21.04, Pb for δ 15 N, Tukey Kramer). No relationships were found between δ 15 N values and either body length or mass (results not shown). P3 P3 Bligh & Dyer Radin Untreated lipids Fig. 2. Effects of lipid extractions on carbon and nitrogen isotope ratios of whole fish samples. The mean±sd of δ 15 N (upper) and δ 13 C (lower) for untreated, lipid-free (Bligh and Dyer and Radin methods) and lipids of herring (H4, n=5) and pollock (P3, n=5). Lipids were extracted in duplicate for each specimen, respectively, using the above two methods. Lipids were first extracted from whole fish samples of 16 herring (H4) and 8 pollock (P3) using Bligh and Dyer's (1959) method for isotopic analysis. As expected, δ 13 C values in lipidfree samples were significantly enriched in 13 C relative to untreated samples (H4: lipid-free: 19.9 ± 1.0, untreated: 22.8±1.4, n=16; P3: lipid-free: 18.6±0.6, untreated: 20.3±1.2, n=8, paired t-test, Pb0.05 for both species). However, extraction of lipids also resulted in approximately 1 increase of δ 15 N values in lipid-free samples compared to those in untreated samples (H4: 13.3 ± 0.6 versus 12.6 ± 0.5, n=16; P3: 14.1 ±0.8 versus 13.2 ±0.9, n =8, paired t-test, P b 0.05 for both species). To compare different solvent effects on lipid-free δ 13 C and δ 15 N values, 5 specimens from each species (H4 and P3) were sub-sampled again and treated in duplicate using Bligh and Dyer's (1959) and Radin's (1981) methods, respectively (Fig. 2). No significant differences in lipid-free δ 13 C and δ 15 N values were found between the two extraction methods (paired t-test, PN0.05 for both δ 13 C and δ 15 N values and for both species, n=5). Normalized lipid-free δ 13 C values were estimated by subtracting 6 lipid content (% dry mass) from untreated δ 13 C values (Alexander et al., 1996). Comparable δ 13 C values were found between the normalized and measured lipid-free δ 13 C values in both fish species (paired t-test, P N 0.05) C and 15 N isotopic variations in blood components of captive seals Excluding the last feeding period from May 2000 to September 2000 because of insufficient isotopic data obtained from pollock batch 5 (P5), serum δ 15 N values in Travis covaried with serum δ 13 C values as diet switching between H4 and P3, though there was no statistically significant differences found in δ 15 N values between the two fish species (Fig. 3). δ 13 C values in serum and RBC of other three seals, Pender, Poco and Snapper, were shown in Fig. 4a, b, c, respectively (no complete natural abundance δ 15 N data were available for these seals due 19 15N Prior to the trial H4 P3 H4 P3 H4 P5 13C δ 15 N δ 13 C Sep-00 Jul-00 May-00 Mar-00 Jan-00 Nov-99 Sep-99 Jul-99 May-99 Mar-99 Jan-99 Nov-98 Sep-98 Jul-98 May-98 Sampling date Fig. 3. The δ 13 C and δ 15 N variations in serum of harbor seal Travis in response to the two fish diets. Specific batches of diet fish are labeled and vertical lines indicate the dates of diet switching.

5 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) (a) -16 Pender serum Pender RBC Prior to the trial H4 P3 H4 P3 H4 P5-17 δ 13 C δ 13 C (b) Jul-00 May-00 Mar-00 Jan-00 Nov-99 Sep-99 Jul-99 May-99 Mar-99 Jan-99 Nov-98 Sep-98 Jul-98 May-98 Sampling date Poco serum Poco RBC Prior to trial P2 H4 P3 H4 P3 H4 Sep Sep-00 Jul-00 May-00 Mar-00 Jan-00 Nov-99 Sep-99 Jul-99 May-99 Mar-99 Jan-99 Nov-98 Sep-98 Jul-98 May-98 (c) -16 Sampling date Snapper serum Snapper RBC Prior to trial 1:1 H4/P3 1:1 H6/P5-17 δ 13 C Sep-00 Jul-00 May-00 Mar-00 Jan-00 Nov-99 Sep-99 Jul-99 May-99 Mar-99 Jan-99 Nov-98 Sep-98 Jul-98 May-98 Sampling date Fig. 4. The δ 13 C values in serum and RBC of (a) Pender, (b) Poco and (c) Snapper. Specific batches of diet fish are labeled and vertical lines indicate the dates of diet switching.

6 474 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) Table 3 The δ 13 C and δ 15 N values (mean±sd a ) in different tissue types of captive seals Seals Diet Isotopes Serum RBC Fur Nails Whiskers b Pender H4 δ 13 C 18.9± ± ± ±1.3 δ 15 N 17.5± ± ± ±0.5 (n=8) (n=8) (n=2) (n=44) Snapper 1:1 mix δ 13 C 18.2± ± ± ± ±0.2 H4/P3 δ 15 N 16.8± ± ± ± ±0.3 (n=8) (n=8) (n=2) (n=2) (n=45) Poco P3 δ 13 C 18.1± ± ± ± ±0.2 δ 15 N 17.3± ± ± ± ±0.4 (n=8) (n=8) (n=4) (n=2) (n=29) a Calculated based on repeated measures of same seal that sampled periodically on a specific fish diet. b The δ 13 C and δ 15 N values represent the mean±sd of multiple measurements from one whisker for each seal, as each whisker was sectioned into 2 mm pieces and isotope ratios measured in each of them. The means of δ 13 C and δ 15 N values did not include those from the first 1 cm near the base due to fairly large differences compared to the rest along whisker length. to concurrent 15 N-labeled tracer experiments). Overall, serum δ 13 C values in seals varied with diet as δ 13 C values in herring were generally more depleted than those in pollock. For instance, depleted serum δ 13 C values were observed in Travis and Pender (Figs. 3, 4a) when they were fed H4. After diet was switched to P3, serum δ 13 C values increased progressively, and then gradually decreased upon switching back to H4. Reverse patterns were observed in Poco (Fig. 4b), who initially started with pollock and switched to herring. As expected, no dietary and seasonal patterns in serum δ 13 C values were found in Snapper fed on a constant equal mix of herring and pollock (Fig. 4c). Statistical analysis with repeated measures ANOVA confirmed that serum δ 13 C values and δ 15 N values were affected by diet but not by season for all seals (repeated measures ANOVA, P b0.0001, P =0.306). There was no statistically significant interaction between diet and season. In contrast, δ 13 C values in RBC showed little variability in seal Pender and Poco in response to changes of diets and were generally more enriched than those in serum (paired t-test, Pb0.001). δ 13 C and δ 15 N values in different tissue samples that were taken prior to 15 N-labeled tracer experiments were compared (Table 3). The most enriched δ 15 N values were found in serum, followed by RBC or whiskers, and the most depleted δ 15 N values were in fur or nails, both keratinous proteins. δ 13 C values were more enriched in whiskers, followed by fur or nails, then RBC, and the most depleted δ 13 C value was found in serum Diet tissue isotopic discrimination factors Diet tissue isotopic discrimination factors were estimated by comparing individual isotope ratios in seals' serum with mean isotope values of their fish diet. To reduce interspecies and intraspecies isotopic variability in δ 13 C arising from varying lipid contents in different fish species and different batches of same fish species, the normalized lipid-free δ 13 C values were used to calculate δ 13 C diet tissue discrimination factors (see below in Discussion for details). As expected, δ 13 C diet serum isotope discrimination factors showed small variations, ranging from 0.6 to 1.7. Considering possible δ 15 N alteration followed by lipid extraction procedure, δ 15 N values measured in untreated fish samples were used for calculation of δ 15 N diet 10 food protein Prior to trial H4 P3 H4 P3 H4 P5 25 Daily food intake (kg) Daily protein intake (g/kg) 0 0 Sep-00 Jul-00 May-00 Mar-00 Jan-00 Nov-99 Sep-99 Jul-99 May-99 Mar-99 Jan-99 Nov-98 Sep-98 Jul-98 May-98 Sampling date Fig. 5. Daily protein ( ) and food ( ) intakes of harbor seal Travis in response to different diets. Each point represents weekly average. Specific batches of diet fish are labeled and vertical lines indicate the dates of diet switching.

7 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) tissue isotope discrimination factors for Travis. δ 15 N diet serum isotope discrimination factors were more pronounced in Travis fed on pollock diet (mean ± SD = 4.6 ± 0.3, n = 22) than those fed on herring diet (mean±sd=3.9 ±0.4, n=23). More enriched serum δ 15 N values, and thus larger δ 15 N diet serum discrimination factors on pollock diet might link to high protein intake, as food intakes were dramatically increased as diet switching from herring to pollock, which corresponded to approximately 3 times higher in protein intake level in the seal fed on pollock diet ( 17 g kg 1 day 1 ) than on herring diet ( 6 gkg 1 day 1 )(Fig. 5). 4. Discussion As mentioned earlier, physiology and metabolism of harbor seals may change significantly with season due to varying energetic demands and activity levels (Fadely et al., 1998). It is critical that captive feeding studies are designed in such a way that seasonal physiological effects can be compensated for a better understanding of metabolic effects of different fish diets on 13 C and 15 N patterns in seals' tissues. The repeated crossover feeding trials of this study allowed each seal to subject to two different dietary treatments in each of three metabolically defined seasons: breeding from May to September, molting from September to January, and late winter/early spring period from January to May over 2 annual cycles. Thus, each seal functioned as its own internal control relative to seasonal physiological changes. For instance, except for Snapper who had a constant diet of an equal mix of herring and pollock over the entire feeding trial, the other three seals were fed alternatively herring and pollock in year 1 and 2 during the same physiological defined seasons, so that any isotopic variations in seals were caused mainly by changes of diet. Ideally, same batch of each fish species should be used so that food quality and isotopic composition remained constant throughout the entire experimental period. Our data showed that energy density remained the same among different batches of same species; whereas isotopic composition differed. Since P2, P5 and H6 were fed only for a relative short period (see Table 1), H4 and P3 were used as representatives in most cases. As shown in Fig. 1, more depleted δ 13 C values were observed in H4 whole fish samples than those in P3, owing partially to greater lipid contents in herring (49% of dry matter) than in pollock (22% of dry matter). Because δ 13 C values in lipids are generally depleted by 6 compared with those in proteins (DeNiro and Epstein, 1978; Tieszen et al., 1983), lipids are generally extracted from samples being analyzed for 13 C to reduce isotopic variability arising from variable lipid contents in different tissues or species. This is particularly true for this study because herring and pollock contained distinctly different lipid contents. Indeed, our data showed that more than 2 differences in δ 13 C values between H4 and P3 were observed when using untreated fish samples, compared with approximately 1 differences when using lipid-free fish samples or normalized δ 13 C values. Thus, if δ 13 C values of untreated fish samples were used for estimation of diet tissue isotopic discrimination factors, it would yield greater errors for herring than that for pollock. These results were consistent with previous findings that lipid removal from different prey items, particularly from those with distinct lipid contents, was a necessary step for accurate estimations of diet tissue isotope discrimination factors when using stable isotope analysis in marine mammal diet reconstruction (Hobson et al., 1996). However, our data also showed that lipid extraction treatments might alter δ 15 N values in both fish samples. Radin (1981) noted that chloroform/methanol extraction could remove significant amount of protein from some fish species. Ideally, two separate sets of samples are better analyzed: untreated samples for 15 N analysis and lipid-extracted samples for 13 C analysis. Alternatively, if data of lipid contents are available for untreated samples, normalized lipid-free δ 13 C values, estimated by subtracting 6 lipid content (% dry mass) from δ 13 C values of untreated samples (Alexander et al., 1996), may be used to reduce cost of isotopic analysis. Our data showed that comparable δ 13 C values were found between normalized and measured lipid-free δ 13 C values in both fish species. As shown in Fig. 3, upto1 differences in serum δ 15 N values were found in one seal fed alternatively between H4 and P3, although no statistically significant differences in δ 15 N values were found between the two fish diets. Serum δ 15 N values in Travis increased progressively and tended to become more enriched as diet switching from H4 to P3, decreased as switching back to H4. These findings indicated that different composition of macronutrients in prey fish species and the resulting differing protein intake levels in seals might affect serum δ 15 N patterns, due to changes in digestive physiology and protein metabolism in response to the two distinct feeding regimes, i.e. high fat/low protein herring and low fat/high protein pollock. It is well known that marine mammals generally consume very high fat diets and lipids are the major energy source for these species, although glucose and protein also account for a small portion of energy supply (Kirby and Ortiz, 1994). As a critical part of this controlled feeding trial study, seals were allowed to adjust their food intake naturally on the basis of their metabolic energy requirements in response to seasonal physiological conditions and different diets. As shown in Fig. 5, the seal increased its food intake dramatically as diet switching from herring to pollock, as energy density was on average 1.8 times higher in herring than in pollock (refer to Table 2). Consequently, the seal consumed approximately 3 times higher dietary protein on pollock diet than on herring diet. We hypothesize that changes in seal protein metabolism, as a result of changes in protein intake level due to differing composition of macronutrients in fish diets, affect isotope ratios in seals' tissues, which in turn impact accurate estimation of diet tissue discrimination factors. Variations in urea excretion rate in humans, in response to different nitrogen intake, are so straightforward that they have been recognized for decades (Epstein et al., 1957). The ability to reduce urinary water loss by increasing urine concentrations in response to water stress (Livingston et al., 1962) and the ability to restrict urea nitrogen excretion while on a low nitrogen diet (Ambrose, 1986; Ambrose and DeNiro, 1987) are well documented physiological adaptations. Recent studies also reveal that urea synthesis rate in rats increased at high protein

8 476 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) intake level and decreased at low protein intake level (Sick et al., 1997). As isotopic discrimination mainly occurs in processes of transamination and oxidative deamination (Macko et al., 1986), amino acid catabolism leading to urea is assumed to be a major branch that results in greater isotopic fractionation (Gaebler et al., 1966). In general, δ 15 N values in urea are up to 10 more depleted than that of serum (Sick et al., 1997). Consequently, a high rate of excretion of 15 N-depleted urea yields more enriched δ 15 N values in other tissues based on nitrogen isotope mass balance model proposed by Ambrose (1991). In this case, elevated rates of protein catabolism, urea synthesis and nitrogen disposal, due to very high dietary protein intake on pollock diet, caused elevated 15 N-depleted urea excretion, resulting in more enriched δ 15 N values in the seal's tissues. These findings are in good agreement with other results that isotope ratios in consumers' tissues may vary within one species raised on different diets with differing food quality and feeding levels due to changes in biochemical reactions associated with changes in animal metabolism and digestive physiology (Webb et al., 1998; Fantle et al., 1999; Adams and Sterner, 2000; Pearson et al., 2003; Gaye-Siessegger et al., 2004). Moreover, studies with the same subjects confirmed that seal protein digestibility did change in response to feeding frequency and different diets between herring and pollock (Trumble et al., 2003). It is interesting to notice that nutritional stress and high dietary protein intake can both result in more enriched δ 15 N values in consumers' tissues. For instance, significantly enriched δ 15 N values were found in avian blood components when birds were raised in a laboratory with restricted food intake or when wild birds underwent natural fasting (Hobson and Clark, 1992; Hobson et al., 1993). Similar findings were also evident in southern right whales during winter breeding season with little or no feeding (Best and Schell, 1996) and in a slow growing juvenile blue crab that fed on a protein-poor detritus (Fantle et al., 1999). Evidence showed that fasting or starvation could trigger protein sparing mechanisms in marine mammals by regulation of protein metabolic pathways (Castellini and Rea, 1992). The enriched δ 15 N values in fasting or starvation animals were assumedly caused by the increased loss of isotopically lighter 14 N by breakdown of their own muscle proteins under negative nitrogen balance due to absence of exogenous dietary protein intakes (Gannes et al., 1997). Animals thus become progressively more enriched in 15 N over the course of starvation. Unlike a previous feeding study with juvenile Steller sea lions, in which it was found that sea lions were unable to maintain body mass on exclusively pollock diet because they failed to adjust food intake to compensate for the decreased energy intake on relatively low energy density fish diet of pollock (Rosen and Trites, 2000, 2002). Harbor seals instead were able to adjust food intake in response to differing dietary macronutrients of fish diets. There was no evidence that demonstrated any nutritional stress in body mass and condition that could be linked to pollock diet (Castellini et al., 2002). The more enriched serum δ 15 N values when the seal was fed a very high protein diet of pollock, compared with those when it was fed a protein intake level not exceeding requirements or under nitrogen balance, i.e. herring, are rather explained as elevated rates of urea synthesis, i.e. elevated 15 N-depleted urea excretion must have resulted in more enriched δ 15 N values in serum. Therefore, the magnitude of δ 15 N diet tissue isotopic discrimination factors in a consumer reflects both physiological conditions of the consumer and food quality and/or availability, as protein metabolism regulates on the basis of endogenous requirements and exogenous intakes. Coupling of physiological conditions and protein metabolism, linked to positive or negative nitrogen balance as determined by endogenous requirements and exogenous dietary intake, affects significantly the extent to which dietary isotopic signatures can be conserved in a consumer. As a result, ultimate isotope patterns in wild animals will be affected by both natural abundance of 13 C and 15 N isotopic variability arising from ecological factors, such as changes in habitats or food sources due to seasonal migrations or prey species switching, and internal biochemical controls, associated with metabolic regulations on the basis of physiological conditions of a consumer and food nutritional quality and/or availability (Gannes et al., 1997). Whereas possible modifications of isotopic signatures through these multiple processes have made field isotope data interpretation more complicated, subtle isotopic variability that links to metabolic and physiological changes might make stable isotope techniques as a useful tool in studies of animal physiology (Gannes et al., 1998). Our results demonstrated that δ 13 C and δ 15 N values in serum reflected dietary changes immediately due to its rapid metabolic turnover rates, and thus served as very sensitive tissues for tracking short-term dietary changes. Overall, δ 13 C values in RBC were more enriched than those in serum. This was consistent with previous studies on other species (Hobson et al., 1997; Kurle, 2002). The repeated measurements also revealed that δ 13 C values in RBC varied little in response to dietary changes (see Fig. 4). This is expected as turnover rates of RBC were much slower than those of serum (Hobson and Clark, 1993). The average life span of a red cell in harbor seals was approximately 120 days (Zhao, 2002). Hence, RBC could integrate changes in dietary isotopic composition over a much longer period up to several months. Differences in δ 13 C and δ 15 N values were found in serum, RBC, fur, nails and whiskers of the captive seals, independent of diets. The tissuespecific carbon and nitrogen isotope ratios are due to different chemical composition that constitutes the specific tissues, as isotope ratios in different chemical components (lipid, protein, carbohydrate, etc.) or individual organic compounds (amino acids, fatty acids, etc.) vary greatly (Hare et al., 1991; Fogel et al., 1997). For instance, our data showed that RBC was always enriched in δ 13 C, but depleted inδ 15 N compared to serum. This is because (1) protein hydrolysable amino acid mole percent composition of RBC differs to that of serum, particularly in glutamate and histidine (Zhao, 2002), and (2) carbon and nitrogen isotope ratios differ greatly among individual amino acids (Zhao, 2002). The differences in amino acid mole percent composition between serum and RBC result in different δ 13 C and δ 15 N values based on isotopic mass balance model.

9 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) In summary, long-term controlled feeding studies with captive harbor seals provided an experimental model for a better understanding of metabolic and physiological effects on carbon and nitrogen isotopic relationships between consumers and their different diets, by excluding ecological factors. Our results demonstrated that changes in protein metabolism and digestive physiology, as a result of changes in food and protein intake level due to differing composition of macronutrients in diets, also played important roles in shaping isotopic patterns of a consumer's tissues, and thus influenced accurate estimation of diet tissue discrimination factors. Reliable values of species- and tissue-specific diet tissue discrimination factors are critical for field applications of stable isotope techniques in trophic dynamics and feeding ecology. Further study using compound specific isotopic analytical techniques may provide fine-scale information on biochemical reactions that influence diet tissue isotopic discrimination across trophic levels. Acknowledgments We thank Drs. S Henrichs and P Barboza for their valuable suggestions. Special thanks are given to JM Castellini for providing fish samples and proximate composition data and extended to SJ Trumble, TL Mau, HL Harmon and staff of ASLC for assistance with sample collections. Captive harbor seal blood samples were authorized through EVOS project and collected under MMPA permit # with UAF and ASLC-IACUC approval. Fish samples were obtained through EVOS project with the cooperation of the ASLC. The fish were food items for the marine mammals at the ASLC and stored under USDA inspection. The research described in this paper (Restoration Project 01371) was supported by the Exxon Valdez Oil Spill Trustee Council. However, the findings and conclusions presented by the authors are their own and do not necessarily reflect the views or position of the Trustee Council. References Adams, T.S., Sterner, R.W., The effect of dietary nitrogen content on trophic level 15 N enrichment. Limnol. Oceanogr. 45, Alaska Sea Grant, Is it food? Addressing marine mammal and seabird declines. Workshop Summary Alaska Sea Grant Report Alaska Sea Grant College Program. University of Alaska Fairbanks, Fairbanks, AK. Alexander, S., Hobson, K.A., GrattoTrevor, C.L., Diamond, A.W., Conventional and isotopic determinations of shorebird diets at an inland stopover: the importance of invertebrates and Potamogeton pectinatus tubers. Can. J. Zool. 74, Ambrose, S.H., Stable carbon and nitrogen isotope analysis of human diet in Africa. J. Hum. Evol. 15, Ambrose, S.H., Effects of diets, climate and physiology on nitrogen isotope abundance in terrestrial foodwebs. J. Archaeol. Sci. 18, Ambrose, S.H., DeNiro, M.J., Bone nitrogen isotope composition and climate. Nature 325, 201. Best, P.B., Schell, D.M., Stable isotopes in southern right whale (Eubalaena australis) baleen as indicators of seasonal movements, feeding and growth. Mar. Biol. 124, Bligh, E.G., Dyer, W.J., A rapid method of total lipid extraction and purification. Can. J. Biochem. Phys. 37, Burton, R.K., Koch, P.L., Isotopic tracking of foraging and long-distance migration in northeastern Pacific pinnipeds. Oecologia 119, Castellini, M.A., Rea, L., The biochemistry of natural fasting at its limits. Experientia 48, Castellini, M.A., Castellini, J.M., Trumble, S.J., Recovery of harbor seals Phase II: controlled studies of health and diet. Exxon Valdez Oil Spill Restoration Project Final Report (01341). Alaska Department of Fish and Game, Anchorage, AK. DeNiro, M.J., Epstein, S., Influence of diet on the distribution of carbon isotopes in animals. Geochim. Cosmochim. Acta 42, DeNiro, M.J., Epstein, S., Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. Cosmochim. Acta 45, Epstein, F., Kleeman, C.R., Pursel, S., Hendrikx, A., Effects of feeding protein and urea on the renal concentrating process. J. Clin. Invest. 36, Fadely, B.S., Castellini, J.M., Castellini, M.A., Recovery of harbor seals from EVOS: condition and health status. Exxon Valdez Oil Spill Restoration Project Final Report (97001). Alaska Department of Fish and Game, Anchorage, AK. Fantle, M.S., Dittel, A.I., Schwalm, S.M., Epifanio, C.E., Fogel, M.L., A food web analysis of the juvenile blue crab, Callinectes sapidus, using stable isotopes in whole animals and individual amino acids. Oecologia 120, Focken, U., Becker, K., Metabolic fractionation of stable carbon isotopes: implications of different proximate compositions for studies of the aquatic food webs using δ 13 C data. Oecologia 115, Fogel, M.L., Tuross, N., Johnson, B.J., Miller, G.H., Biogeochemical record of ancient humans. Org. Geochem. 27, Frost, K.J., Lowrey, L.F., Sinclair, E.H., Ver Hoef, J., McAllister, D.C., Impacts on distribution, abundance and productivity of harbor seals. In: Loughlin, T.R. (Ed.), Marine Mammals and Exxon Valdes. Academic Press, San Diego, pp Gaebler, O.H., Trieste, G.V., Vukmirowich, R., Isotope effects in metabolism of 14 N and 15 N from unlabeled dietary proteins. Can. J. Biochem. 44, Gannes, L.Z., O'Brien, D.M., Martínez del Rio, C., Stable isotopes in animal ecology: assumptions, caveats, and a call for more laboratory experiments. Ecology 78, Gannes, L.Z., Martínez del Rio, C., Koch, P.L., Natural abundance variations in stable isotopes and their potential in animal physiological ecology. Comp. Biochem. Physiol. A 119, Gaye-Siessegger, J., Focken, U., Abel, S.M.H., Becker, K., Feeding level and individual metabolic rate affect δ 13 C and δ 15 N values in carp: implications for food web studies. Oecologia 138, Hare, P.E., Fogel, M.L., Stafford Jr., T.W., Mitchell, A.D., Hoering, T.C., The isotopic composition of carbon and nitrogen in individual amino acids isolated from modern and fossil proteins. J. Archaeol. Sci. 18, Hilderbrand, G.A., Farley, S.D., Robbins, C.T., Hanley, T.A., Titus, K., Servheen, C., Use of stable isotopes to determine diets of living and extinct bears. Can. J. Zool. 74, Hirons, A., Schell, D.M., Finney, B.P., Temporal records of δ 13 C and δ 15 N in North Pacific pinnipeds: inferences regarding environmental change and diet. Oecologia 129, Hobson, K.A., Clark, R.G., Assessing avian diets using stable isotopes II: factors influencing diet tissues fractionation. Condor 94, Hobson, K.A., Clark, R.G., Turnover of 13 C in cellular and plasma fractions of blood: implications for nondestructive sampling in avian dietary studies. Auk 110, Hobson, K.A., Sease, J.L., Stable isotope analyses of tooth annuli reveal temporal dietary records: an example using Steller sea lions. Mar. Mamm. Sci. 14, Hobson, K.A., Alisauskas, R.T., Clark, R.G., Stable nitrogen isotope enrichment in avian tissues due to fasting and nutritional stress: implications for isotopic analyses of diet. Condor 95, Hobson, K.A., Schell, D.M., Renouf, D., Noseworthy, E., Stable carbon and nitrogen isotopic fractionation between diet and tissues of captive seals: implications for dietary reconstructions involving marine mammals. Can. J. Fish. Aquat. Sci. 53,

10 478 L. Zhao et al. / Comparative Biochemistry and Physiology, Part A 143 (2006) Hobson, K.A., Sease, J.L., Merrick, R.L., Piatt, J.F., Investigating trophic relationships of pinnipeds in Alaska using stable isotope ratios of nitrogen and carbon. Mar. Mamm. Sci. 13, Isik, O., Sarihan, E., Kusvuran, E., Gül, Ö., Erbatur, O., Comparison of the fatty acid composition of the freshwater fish larvae Tilapia zillii, the rotifer Brachionus calyciflorus, and the microalgae Scenedesmus abundans, Monoraphidium minitum and Chlorella vulgaris in the algae rotifer fish larvae food chains. Aquaculture 174, Kirby, V.K., Ortiz, C.L., Hormones and fuel regulation in fasting elephant seals. In: LeBoeuf, B.J., Laws, R.M. (Eds.), Elephant Seals: Population Ecology, Behavior and Physiology. University of California Press, Berkeley, pp Kline Jr., T.C., Temporal and spatial variability of 13 C/ 12 C and 15 N/ 14 Nin pelagic biota of Prince William Sound, Alaska. Can. J. Fish. Aquat. Sci. 56, Kurle, C.M., Stable isotope ratios of blood components from captive northern fur seals (Callorhinus ursinus) and their diet: applications for studying the foraging ecology of wild otariids. Can. J. Zool. 80, Kurle, C.M., Worthy, G.A.J., Stable isotope assessment of temporal and geographical differences in feeding ecology of northern fur seals (Callorhinus ursinus) and their prey. Oecologia 126, Livingston, H., Payne, W., Friend, M., Urea excretion in ruminants. Nature 194, Macko, S.A., Lee, W.Y., Parker, P.L., Nitrogen and carbon isotope fractionation by two species of marine amphipods: laboratory and field studies. J. Exp. Mar. Biol. Ecol. 63, Macko, S.A., Estep, M.L.F., Engel, M.H., Hare, P.E., Kinetic fractionation of stable nitrogen isotopes during amino acid transamination. Geochim. Cosmochim. Acta 50, Minagawa, M., Wada, E., Stepwise enrichment of 15 N along food chains: further evidence and the relation between δ 15 N and animal age. Geochim. Cosmochim. Acta 48, Overman, N.C., Parrish, D.L., Stable isotope composition of walleye: 15 N accumulation with age and area-specific differences in δ 13 C. Can. J. Fish. Aquat. Sci. 58, Pearson, S.F., Levey, D.J., Greenberg, C.H., Martinez del Rio, C., Effects of elemental composition on the incorporation of dietary nitrogen and carbon isotopic signatures in an omnivorous songbird. Oecologia 135, Pitcher, K.W., Major decline in number of harbor seals, Phoca vitulina richardsi, on Tugidak Island, Gulf of Alaska. Mar. Mamm. Sci. 6, Polischuk, S.C., Hobson, K.A., Ramsay, M.A., Use of stable carbon and nitrogen isotopes to assess weaning and fasting in female polar bears and their cubs. Can. J. Zool. 79, Radin, N.S., Extraction of tissue lipids with a solvent of low toxicity. Methods Enzymol. 72, 5 7. Rosen, D.A.S., Trites, A.W., Pollock and the decline of Steller sea lions: testing the junk-food hypothesis. Can. J. Zool. 87, Rosen, D.A., Trites, A.W., Changes in metabolism in response to fasting and food restriction in the Steller sea lion (Eumetopias jubatus). Comp. Biochem. Physiol. B 132, Sick, H., Ross, N., Saggau, E., Hass, K., Meyn, V., Walch, B., Trugo, N., Amino acid utilization and isotope discrimination of amino nitrogen in nitrogen metabolism of rat liver in vivo. J. Nutr. Sci. 36, Tieszen, L.L., Boutton, T.W., Tesdahl, K.G., Slade, N.A., Fractionation and turnover of stable carbon isotopes in animal tissues: implications for 13 C analysis of diet. Oecologia 57, Trumble, S.J., Barboza, P.S., Castellini, M.A., Digestive constraints on an aquatic carnivore: effects of feeding frequency and prey composition on harbor seals. J. Comp. Physiol. B 173, Webb, S.C., Hdges, R.E.M., Simpson, S.J., Diet quality influences the δ 13 C and δ 15 N of locusts and their biochemical components. J. Exp. Biol. 201, Zhao, L., Tracing amino acid metabolism of harbor seals (Phoca vitulina) using stable isotope techniques. PhD thesis, University of Alaska Fairbanks, Fairbanks, AK.