Journal of Comparative Physiology B

, Volume 183, Issue 7, pp 993–1003

Whole-body protein turnover reveals the cost of detoxification of secondary metabolites in a vertebrate browser

Authors

  • Jessie Au
    • Evolution, Ecology and Genetics, Research School of BiologyAustralian National University
  • Karen J. Marsh
    • Evolution, Ecology and Genetics, Research School of BiologyAustralian National University
    • Evolution, Ecology and Genetics, Research School of BiologyAustralian National University
  • William J. Foley
    • Evolution, Ecology and Genetics, Research School of BiologyAustralian National University
Original Paper

DOI: 10.1007/s00360-013-0754-3

Cite this article as:
Au, J., Marsh, K.J., Wallis, I.R. et al. J Comp Physiol B (2013) 183: 993. doi:10.1007/s00360-013-0754-3

Abstract

The detoxification limitation hypothesis predicts that the metabolism and biotransformation of plant secondary metabolites (PSMs) elicit a cost to herbivores. There have been many attempts to estimate these costs to mammalian herbivores in terms of energy, but this ignores what may be a more important cost—increases in protein turnover and concomitant losses of amino acids. We measured the effect of varying dietary protein concentrations on the ingestion of two PSMs (1,8 cineole—a monoterpene, and benzoic acid—an aromatic carboxylic acid) by common brushtail possums (Trichosurus vulpecula). The dietary protein concentration had a small effect on how much cineole possums ingested. In contrast, protein had a large effect on how much benzoate they ingested, especially at high dietary concentrations of benzoate. This prompted us to measure the effects of dietary protein and benzoate on whole-body protein turnover using the end-product method following an oral dose of [15N] glycine. Increasing the concentration of dietary protein in diets without PSMs improved N balance but did not influence whole-body protein turnover. In contrast, feeding benzoate in a low-protein diet pushed animals into negative N balance. The concomitant increases in the rates of whole-body protein turnover in possums eating diets with more benzoate were indicative of a protein cost of detoxification. This was about 30 % of the dietary N intake and highlights the significant effects that PSMs can have on nutrient metabolism and retention.

Keywords

DetoxificationProtein turnoverCostNitrogenHerbivorePlant secondary metabolite

Introduction

A tenet of our understanding of the interactions between mammals and their food plants is that the rate and cost of detoxification of ingested plant secondary metabolites (PSMs) influences subsequent feeding rates and how animals choose their diets. This idea is the detoxification limitation hypothesis (Marsh et al. 2006a), which makes several testable predictions. First, animals should feed so that they avoid overloading specific detoxification pathways. Experimental tests of this idea are rare, but Marsh et al. (2006b) showed that common brushtail possums ate more when they could choose from foods containing compounds that were detoxified by different pathways rather than a common pathway. The second prediction is that the ingestion and metabolism of PSMs should incur a measurable cost on animals that effectively taxes gross nutrient intake. It is this second aspect of the detoxification limitation hypothesis that we address here.

Several studies using a variety of approaches measure the costs incurred by mammals in detoxifying PSMs. In many of these studies, energy is the favoured currency. For example, Iason et al. (1996) and Thomas et al. (1988) reported that animals that ingested PSMs had elevated metabolic rates and that this effect continued long after food was withdrawn. Examining a different angle, Sorensen et al. (2005) showed that animals fed diets containing PSMs are typically less active and thus compensate for the energetic cost of detoxification. Others have taken a biochemical approach and estimated the energetic value of endogenous compounds used in the detoxification process. Glucuronidation is one of the major Phase II detoxification pathways employed by mammals and Cork et al. (1983) estimated that glucuronic acid excretion accounted for more than 25 % of the fasting glucose production in koalas fed Eucalyptus punctata foliage. Similarly, glucuronic acid represented half of the energy lost by woodrats consuming resin (Mangione et al. 2004).

Although these studies all indicate that detoxification is costly, energy may not be the only currency in which to express costs because this may ignore what may be a more important cost–increases in protein turnover and concomitant losses of amino acids. While animals may be able to adjust their metabolism, by reducing activity for instance, they may find it harder to compensate for losses of protein. The link between detoxification and protein metabolism prompted Marsh et al. (2006a) to suggest that protein might be a better currency for estimating the cost of detoxification because of its importance as a measure of food quality for herbivores (Mattson 1980; White 1978; White 1993). Much of the work, however, that examines the role of protein in diet selection tends to be simplistic and fails to account for the complex interactions with other parts of the diet, particularly PSMs. The link between dietary protein and PSMs comes from three studies, which show that animals ingest more of a PSM if supplemented with protein (Villalba et al. 2002; Villalba and Provenza 2005; Nersesian et al. 2012). However, none of these studies provide a mechanism for this response and instead link the increased tolerance to PSMs to “nutrient availability”, but fail to mention the most likely nutrient-protein.

The pathways of detoxification involve many processes that require protein (amino acids). First, detoxification enzymes ultimately come from dietary protein (Whitlock and Denison 1995; Pass et al. 1999). Second, the excretion of PSMs often involves the loss of nitrogenous compounds either as conjugates or through the disruption of acid base balance (Edwards et al. 2010). Finally, protein is necessary to repair any tissue damaged by PSMs, which is one reason why fasting metabolic rates sometimes remain elevated after exposure to PSMs (Thomas et al. 1988). Measuring the cost of detoxification in terms of protein should better link existing models of nutritional choice and field studies explaining reproductive success in free ranging browsers (Degabriel et al. 2009; McArt et al. 2009). Although one might show that animals ingest more PSM when supplemented with protein, a protein cost of detoxification will ultimately reside in changes in protein turnover.

The rate at which the body synthesises and degrades protein is the whole-body protein turnover (Waterlow 2006). Whole-body protein turnover depends on many factors such as the age of the animal and, in some cases, the protein concentration of the diet. There are many ways to measure whole-body protein turnover but the one most suitable for non-domesticated animals, is the end-product method (Duggleby and Waterlow 2005; Waterlow 1981) because it is noninvasive. Instead, it involves dosing an animal with a known amount of labelled amino acid (usually [15N] glycine) (Fern et al. 1985), which is incorporated into the animal’s N pool and either incorporated into protein or excreted. Protein synthesis can then be estimated from the rate at which the animal excretes the label in urine. Researchers call it the end-product method because the animal excretes any isotope not incorporated into body proteins as N in the urine, typically as urea or ammonium (the end-point). Any increase in the excretion of the label at a given rate of N ingestion (and thus decrease in protein synthesis) indicates a net loss of protein from the body (negative nitrogen balance). The end-product method is widely applied in clinical studies to study the effects of various pathological and physiological states on whole-body protein turnover (Duggleby and Waterlow 2005). This suggests that it is suited to studying animals ingesting PSMs.

We investigated the interaction of the concentrations of protein and PSMs in the diet on food intake and whole-body protein turnover rate in common brushtail possums. We tested four hypotheses: first, that the concentration of dietary protein will not affect whole-body protein turnover. The second hypothesis was that the rate of protein turnover will increase with increasing concentrations of dietary PSM, leading to the third hypothesis that increasing concentrations of PSMs in the diet will decrease N balance. To separate the expected indirect effects of a toxin on protein metabolism–the synthesis of detoxification enzymes, from direct effects–the need for amino acid conjugates, we studied two PSMs—cineole and benzoic acid, with well documented modes of detoxification in common brushtail possums (Awaluddin and McLean 1985; Boyle et al. 2005; Marsh et al. 2006b). To measure indirect effects we fed the monoterpene, cineole, which the common brushtail possum detoxifies by forming oxidised metabolites of which it excretes about 50 % conjugated with glucuronic acid. To measure direct and indirect effects simultaneously we fed benzoic acid, which the possum eliminates by conjugating it with the amino acid, glycine. Thus, our fourth hypothesis was that we would expect the detoxification of benzoic acid to be more expensive, in terms of N, than the detoxification of cineole.

Methods

Animals and diets

Common brushtail possums were captured on the campus of the Australian National University and maintained in captivity as described previously (Marsh et al. 2005). We placed possums in large wire mesh cages (130 × 60 × 90 cm, L × W × H) for experiment 1 but transferred them to galvanised metabolism cages (75 × 60 × 60 cm) (Stapley et al. 2000) for experiments 2 and 3, which required accurate measurement of urine volume.

The experimental diets—a mixture of fruit and cereals, were offered to possums as a wet mash. We first mixed the dry ingredients and then added the wet ingredients. The low-protein, basal diet (0.78 ± 0.01 % N DM), formulated to provide a little more N than the daily requirement of 205 mg truly digestible N per kg of metabolic body mass (Wellard and Hume 1981), consisted of 39.9 % pureed apples, 30 % pureed bananas, 10 % pureed carrots, 8 % Solka-Floc (purified wood cellulose: International Fiber Corporation, New York), 6 % rice hulls, 2 % rolled oats, 2 % ground lucerne, 1.25 % vegetable oil, 0.25 % acid casein, 0.3 % sodium chloride, 0.25 % dicalcium phosphate (CaHPO4.2H20), 0.03 % vitamin and mineral mix (Nutriquin, Kentucky Equine Research, Australia), all on a wet matter (WM) basis. We made a “high-protein” basal diet (2.44 ± 0.04 % N DM)—one that might be eaten by a possum obtaining additional protein, by substituting some of the Solka-Floc in the low-protein diet with acid casein. Thus, the high-protein basal diet contained 3.8 % acid casein and 4.2 % Solka-Floc on a WM basis, with the rest of the ingredients remaining the same as in the low-protein basal diet. These basal diets contained roughly 30 % dry matter.

We included a PSM in the diet by substituting it for Solka-Floc. Benzoic acid (Sigma, Castle Hill, Australia) occurs as a dry powder and so was added and mixed with the dry ingredients. In contrast, 1,8-cineole is an oily liquid most conveniently added after mixing other ingredients. A sample of each diet was oven-dried and stored for analysis.

Experimental design

We used the same eight male common brushtail possums in each experiment (mean body mass = 3.1, se 0.1 kg). Because possums tend to vary in size, age and their propensity to consume the basal diet, we used crossover designs for all experiments (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00360-013-0754-3/MediaObjects/360_2013_754_Fig1_HTML.gif
Fig. 1

Time-lines for experiments one (Fig. 1a), two and three (Fig. 1b). Experiment one had three parts in which we fed diets containing benzoate and low- or high-concentration ranges of 1,8-cineole in diets containing low or high concentrations of protein. We fed the four concentrations of each of these PSMs to each possum on a single night in a Latin Square design (experimental days 1–4), having introduced the possums to the lowest concentration of each on days 5 and 6 of the first acclimation period. In the other experiments the treatments were low- and high-concentrations of protein with no PSM (experiment 2) or low- and high-concentrations of benzoate fed in the low-protein diet (experiment 3)

Animals were acclimated to either a low- or high-protein treatment for 7 days. Therefore, experiments measuring the effects of dietary protein concentration required two periods (each with their own acclimation period) to ensure animals encountered every treatment. In addition, when studying the effects of PSMs in experiment 1, we included them in the diet at the lowest concentrations to be used in the experiment (cineole 0.32 % DM; benzoate 0.61 % DM) on nights 5 and 6 of the first acclimation period. This gave possums the opportunity to familiarise themselves with the taste and smell of the compound and any consequences associated with ingesting it.

The daily routine for all experiments was the same. We fed the animals at 1700 h and collected food refusals at 0900 the next morning to determine dry matter intake. Experiments 2 and 3 were balance experiments that required the analysis of urine and faeces, which we collected at 0900 on experimental days 2–6. We stored these samples at −20 °C and at the end of this 5-day collection, we combined the daily collections for each possum and mixed and subsampled them and stored them at −20 °C pending drying and analysis of nitrogen (N).

On the sixth night of experiments 2 and 3 we administered a single dose of 40–50 mg (measured to 0.1 mg) [15N] glycine (98 atoms 15N/100 atoms N, Amersham, UK), in half a dried date (ca 4 g), immediately before feeding the possums. During the earlier acclimation, we trained all animals to eat half a dried date immediately before presenting food as training for them to consume an oral dose of the [15N] tracer in a single pulse. All animals consumed the whole dose within 1 min. At 21 h and 45 h after tracer administration (i.e., day 7 and 8), we determined the mass of urine excreted and retained 30–50 g that we stored at −80 °C. These urine samples were analysed for total N and 15N enrichment so that we could calculate the recovery of the marker after first determining the natural abundance of 15N in urinary N on a sample collected before the animal ingested the 15N.

Experiment 1

We investigated the effect of dietary protein concentration on the intake of diets containing four concentrations of either 1,8 cineole or benzoic acid in three separate trials—a low range of cineole, a high range of cineole and a range of benzoate concentrations.

Eight treatments were offered to eight common brushtail possums over eight consecutive days in a Latin square design with four concentrations of PSMs nested within each of the high- and low-protein diets. In all cases, dry matter intake (DMI) was the amount eaten by a possum in one night. The concentrations of PSMs were based on the data of Marsh et al. (2006b). Cineole was added at concentrations of 0.1, 1, 1.75 and 2.5 % WM (equivalent to 0.32, 3.2, 5.8 and 8.2 % DM). A second experiment used higher concentrations of cineole (0.32, 3.2, 9.5 and 15.8 % DM). The concentrations of benzoate were 0.61, 4.3, 8.8 and 12.6 % DM.

Experiments 2 and 3

Experiment 2 investigated the effect of dietary protein concentration on N balance (the difference between the amount of N ingested and the losses in faeces and urine) and whole-body protein turnover rate, whereas experiment 3 tested the influence of dietary benzoate on the same parameters. We were unable to keep the possums captive for longer and thus had to choose between testing cineole or benzoate and chose the latter because it had a stronger interaction with protein in experiment 1. Thus, the diets offered to the possums differed between the two experiments, but all other aspects of experimental design and analysis were the same.

In experiment 2, we used the low- and high-protein diets described for experiment 1 without any PSM to measure the effect of protein concentration alone. In experiment 3, we fed the low-protein diet supplemented with benzoate at concentrations of 0.61 and 4.3 % DM. We chose these concentrations of benzoate because neither limited intake in Experiment 1. This presented the best possibility of ensuring that possums on the different treatments ingested the same amount of protein, leaving the amount of benzoate they ingested as the only difference between the two experimental treatments.

Calculation of protein turnover

The calculation of whole-body protein turnover (whole-body N turnover) was based on a two-pool model proposed by Sprinson and Rittenberg (1949), developed by Picou and Taylor-Roberts (1969) and described by White et al. (1988).

In this model, amino acids from dietary protein (I) and body protein degradation (D) enter a metabolic pool of N. Amino acids are then either removed from this pool for protein synthesis (S), or they are deaminated and excreted (E) in the urine as nitrogenous end-products, mainly urea.

Golden and Jackson (1981) discussed the assumptions behind the ‘end-product’ method. They suggested that the most important assumption is that the rate at which the [15N] glycine isotope disappears from the metabolic pool through N synthesis and N degradation accurately reflects the fractional loss of all amino acids by these routes. This does not mean that the composition of the amino acid mixture excreted is the same as that undergoing synthesis, but the proportions of the flux (Q) that go to synthesis and excretion are the same as the proportions of the flux of the isotope. Furthermore, the method assumes that the only source of N excreted in urine is the metabolic pool of N (Zilversmit 1960). The irreversible loss (IL) of N from the metabolic pool over time is therefore:
$$ {\text{IL}} = {\text{urinary N excretion rate }}\left( {{\text{g N}}{\text{d}}^{-1}} \right)/{\text{ fraction of }}^{ 1 5} {\text{N dose recovered in urinary N}} $$
If there is no recycling of the isotope, then IL is equivalent to the total N flux (Q) through the metabolic pool. In a study of three species of small macropodid marsupials, White et al. (1988) concluded that 45 h was sufficient time for the 15N to leave the metabolic pool but short enough to avoid recycling of the isotope. Thus, the following equation enables the calculation of the other parameters of the model:
$$ {\rm Q} \, = \, {\rm S} \, + \, {\rm E} \, = \, {\rm D} \, + \, {\rm I} $$
where S is N incorporated into synthesis; E is N excreted; D is N from body protein degradation; I is dietary N intake.

Many researchers measure whole-body protein turnover by determining the abundance of 15N in one urinary compound, such as urea (Sprinson and Rittenberg 1949; Duggleby and Waterlow 2005). Since benzoate conjugates with glycine to form the excretion product, hippuric acid, and because PSMs influence acid–base balance (and thus the relative proportions of urinary urea and ammonium) (Edwards et al. 2010), it made sense to take a broad view of urinary end-products and measure the total 15N excretion in the urine.

We report values as N turnover (mmol N d−1) but have converted some values to mass of crude protein turnover per unit of metabolic body mass (g N*6.25 kg−0.75 d−1), for comparative purposes.

Analytical

We analysed the N content of all samples using the Dumas combustion procedure in a Leco TruSpec carbon/nitrogen analyzer and mass spectrometry to measure the abundance of 15N in urinary N using a CE1110 CHN-S analyser (Carlo Erba, UK) interfaced with a continuous flow isotope ratio mass spectrometer (Micromass IsoChrom, Waters, USA). We corrected for residual moisture in samples of the food, food refusals and faeces by drying 1.0 g samples in an oven at 60 °C to constant mass.

Statistical analysis

We used GenStat 14.2 (VSN International Ltd, UK) for all analyses. We used a two-way ANOVA to analyse the data from experiment 1 with treatments of protein concentration (high or low) and the concentration of PSM. The response was DMI. To incorporate the Latin square design into the statistical model, the block structure was the period and, within that, the order that the animals received each PSM concentration and the day the diet was offered.

We used the REML linear mixed model to analyse the data from experiments 2 and 3 with fixed effects being identifiable sources of variation, such as the order of treatments, period, protein or benzoate concentrations and body mass of the animals and a random effect of “possum”. We express all results in terms of nitrogen (g N d−1 or mmol N d−1) but use the terms nitrogen and protein interchangeably.

Results

Experiment 1 The effect of dietary protein concentration on the intake of diets containing graded concentrations of 1,8 cineole or benzoic acid

At moderate dietary concentrations of cineole only the highest concentration (8.2 % DM) reduced DMI (P < 0.001). There was no effect of dietary protein concentration on DMI at any concentration of cineole (P = 0.538). Thus, we conducted a second experiment using higher concentrations of cineole and in this case dietary cineole concentration strongly influenced feeding (P < 0.001; Fig. 2a). Moreover, brushtail possums were willing to eat more cineole if it was offered in a high-protein diet (P = 0.036). Cineole ingestion increased with increasing dietary concentration reaching a plateau at about 4.7 g per possum per day on the low-protein and 6.1 g per possum per day on the high-protein diet (Fig. 3a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00360-013-0754-3/MediaObjects/360_2013_754_Fig2_HTML.gif
Fig. 2

Feeding by common brushtail possums in relation to low (2.44 % DM) (black bars) and high (0.78 % DM) (grey bars) concentrations of dietary nitrogen and varying concentrations of cineole (a) or benzoate (b). The small vertical bars are the 5 % least significant difference between the cineole or benzoate treatments (black bars) or between the nitrogen treatments (grey bars)

https://static-content.springer.com/image/art%3A10.1007%2Fs00360-013-0754-3/MediaObjects/360_2013_754_Fig3_HTML.gif
Fig. 3

The effect of high (2.44 % DM) (open circles) and low (closed circles) concentrations of dietary nitrogen on the ingestion of cineole (a) or benzoate (b) by common brushtail possums. The small vertical grey bars are the 5 % least significant difference between the nitrogen treatments

Possums ate less with increasing concentrations of dietary benzoate (P < 0.001; Fig. 2b) but ate about 5.2 g more of the high protein diet across all benzoate concentrations (P < 0.001). This meant that possums ate more benzoate when eating the high protein diet (P < 0.001; Fig. 3b).

Experiment 2: The effect of dietary protein concentration on N balance and whole-body protein turnover

Nitrogen balance (Table 1)

Larger animals tended to eat more (P = 0.089) and to produce more faeces (P = 0.033). Animals ate similar amounts of the high and the low-protein diets (P = 0.398) and so those eating the high protein diet ingested more N (P < 0.001). In contrast, animals excreted similar amounts of faecal N (P = 0.396) on both diets. This, together with the slightly lower dry matter digestibility of the low-protein diet (63 versus 67 %; P = 0.053) meant that there was a large difference in N digestibility (17, 72 %; P < 0.001) between the two diets. Possums eating the high-protein diet excreted more urine (P = 0.026), more urinary N (P < 0.001) and more N in total (P < 0.001), but the higher intake of digestible N (the proportion of ingested N that did not appear in the faeces) meant that they stayed in N balance (N intake exceeded N excreted). In contrast, animals eating the low-protein diet were typically in negative N balance (P < 0.001 between diets).
Table 1

The effect of dietary nitrogen on feeding, digestion and the parameters of nitrogen balance

Dietary

Body

DM

N

Faecal

Urinary

Digestibility

N

N (% DM)

Mass (kg) (se)

Intake (g d−1)

Intake (g d−1)

N (g d−1)

N (g d−1)

DM (%)

N (%)

Balance (g d−1)

2.44

2.96 (0.15)

53.8

1.27

0.36

0.58

67.9

72.0

0.32

0.78

3.05 (0.13)

48.5

0.37

0.30

0.15

63.2

17.0

−0.08

LSD

 

9.1

0.22

0.09

0.10

4.9

12.1

0.10

Wald

 

0.84

62.2

0.84

69.2

5.87

89.9

68.2

P

 

0.40

<0.001

0.40

<0.001

0.053

<0.001

<0.001

The order of treatment also affected N balance. Animals given the low-protein diet first (i.e., in period 1) were in more positive N balance than were those fed the high-protein diet first (P = 0.042).

Protein turnover (Table 2)

Neither the period nor the treatment order influenced any measure of protein turnover. Animals eating the high-protein diet excreted more urinary N (P < 0.001) from which we recovered more of the 15N label (P = 0.004). Nitrogen turnover rate (flux) appeared to be higher in possums fed the high-protein diets, but the difference was not significant (P = 0.139). Possums synthesised similar amounts of protein on the two diets (P = 0.483) and thus degradation rates were similar. Possums eating the high-protein diet synthesised more protein than they degraded and so the net change in the protein pool was positive, whereas the opposite was true of possums eating the low-protein diet (P < 0.001), largely due to the significant difference in digestible N intake.
Table 2

The effect of nitrogen supply on the rate of whole-body nitrogen turnover (mmol N d−1)

Dietary

15N

N

N

Urinary

N

Digestible

Net change to N pool

N (% DM)

Recovery (%)

Turnover (Q)

Synthesis (S)

N (E)

Degradation (D)

N intake (I)

2.44

16.5

135.5

113.1

22.4

100.6

34.6

12.3

0.78

7.3

112.1

104.1

7.9

109.7

2.4

−5.5

LSD

4.1

32.0

32.2

4.2

30.6

4.4

2.4

Wald

21.0

2.90

0.56

53.0

0.14

222.8

222.7

P value

0.004

0.139

0.483

<0.001

0.723

<0.001

<0.001

Experiment 3: The effect of benzoate on N balance and whole-body protein turnover

Nitrogen balance (Table 3)

Dry matter digestibility did not vary across the benzoate diets (P = 0.337). Possums ate less and thus ingested less N on the high-benzoate diet (P = 0.007, P = 0.003, respectively). There were trends for possums consuming more benzoate to lose less faecal N (P = 0.092) but more urinary N (P = 0.11) and to digest less N (P = 0.12). Even so, possums eating the low-benzoate diet consumed more digestible N (P = 0.031) and consequently maintained higher N balances (but still negative) than did animals eating the high-benzoate diet, whose N balances were well below zero. A strong interaction was found between diet and N balance as animals on the low-benzoate diet were in a better N balance (P = 0.016).
Table 3

The effect of dietary benzoate on intake, digestibility and the components of nitrogen balance

Dietary Benzoate (% WM)

Body Mass (kg) (se)

DM

N

Faecal

Urinary

Digestibility

N

Intake (g d−1)

Intake (g d−1)

N (g d−1)

N (g d−1)

DM (%)

N (%)

Balance (g d−1)

1.33

2.96 (0.12)

43.8

0.33

0.25

0.21

61

20.4

−0.134

0.18

3.03 (0.12)

56.9

0.44

0.31

0.15

59

27.6

−0.034

LSD

 

6.3

0.05

0.05

0.06

4.0

7.8

0.060

Wald

 

18.14

27.55

4.32

3.76

1.12

3.57

11.24

P

 

0.007

0.002

0.092

0.11

0.34

0.12

0.016

The timing of treatments affected several results. For example, animals excreted more urinary N in period one than in period two (P = 0.024) and N balance was lower in period one (P = 0.01). Also, possums that received the low-benzoate diet first (i.e., in period 1) had lower N balances than did those that consumed the high-benzoate diet in that period (P = 0.042).

Protein turnover (Table 4)

Possums fed the two diets excreted similar amounts of urinary N in the 45-h collection period (P = 0.306) with similar amounts of label so the recovery of the 15N label did not differ between diets (P = 0.242). There was a slower flow of N through the metabolic pools (N turnover) in possums fed the low-benzoate diets (P = 0.039). Thus, the rates of N synthesis (P = 0.041) and degradation (P = 0.034) were lower in animals consuming less benzoate. Nitrogen degradation was greater than N synthesis on both diets and thus, the net change to the metabolic pool was negative for both. However, the combination of reduced digestible N intake and increased N synthesis and degradation on the high benzoate diet resulted in a more negative net change to the N pool (P = 0.009).
Table 4

The effect of dietary benzoate on the rate of nitrogen turnover (mmol N d−1)

Dietary Benzoate (% WM)

15N

N

N

Urinary

N

Digestible

Net change to N pool

Recovery (%)

Turnover (Q)

Synthesis (S)

N (E)

Degradation (D)

N intake (I)

1.33

13.5

59.2

52.9

6.2

56.4

2.6

−3.5

0.18

19.1

31.1

25.5

5.6

26.9

4.2

−1.4

LSD

8.0

20.5

20.2

0.9

20.8

1.1

1.1

Wald

1.69

6.87

6.71

1.28

7.43

8.56

16.52

P

0.242

0.039

0.041

0.306

0.034

0.031

0.009

Again, the order of the treatments affected several measures, such as the excretion of urinary N (P = 0.041), with possums eating the low-benzoate diet in period one excreting more urinary N, which reduced the size of the N pool (P = 0.022).

Discussion

The key finding in this study was the influence of PSM ingestion on whole-body protein turnover. We found that animals ingesting more protein were in positive N balance, but that dietary protein concentration did not affect whole-body protein turnover rate. However, increasing dietary concentrations of benzoate increased the rate of whole-body protein turnover and forced animals into negative N balance. Thus, the study confirmed our original hypotheses–that the concentration of dietary protein will not affect whole-body protein turnover but that increasing concentrations of dietary PSM will amplify the rate of protein turnover with concomitant reductions in N balance. In other words, some PSMs can exert a protein cost on the animal, which suggests that measuring whole-body protein turnover is a useful way to integrate the many costs of detoxification.

The reason for measuring a protein cost of detoxification stems from the detoxification limitation hypothesis, which proposes that the rate that an animal can detoxify a PSM influences the rate at which it can ingest it (Freeland and Janzen 1974; Marsh et al. 2006a). Thus, the inference from our results–the propensity of possums to ingest more benzoate and 1,8-cineole when eating diets with more protein, is that the additional protein speeds detoxification. This result adds support to the findings of Boyle et al. (2005) and Marsh et al. (2005) that the rate at which a common brushtail possum can detoxify these PSMs influences their feeding rates and diet mixing–in this case, the tendency for an animal to switch diets when detoxification pathways becomes saturated.

Modelling of the relationship between dietary PSMs and nutrients suggest that the physiological ability of a herbivore to tolerate PSMs depends on the rate of nutrient absorption because detoxification incurs nutritional costs (Illius and Jessop 1995; Foley et al. 1995). For example, detoxification depletes protein stores in two ways: (1) in nitrogenous compounds such as glycine or glutathione that conjugate with PSMs for excretion (Awaluddin and McLean 1985; Scheline 1991); and (2) losses of N through the increased ammonium excretion used to buffer high acid loads from compounds such as glucuronic acid conjugates (Foley et al. 1995; Edwards et al. 2010). Animals must also synthesise the enzymes needed for detoxification, particularly cytochrome P450 enzymes and the conjugative enzymes, especially glucuronylytransferases (Caldwell and Jakoby 1983). Presumably, an animal eating a diet with inadequate protein will have a lower tolerance of PSMs (Foley et al. 1995; Illius and Jessop 1995) and should reduce their intake of PSMs. Therefore, the capacity of a herbivore to detoxify PSMs will depend on three main factors—the chemical nature of the compound, its concentration and the availability of protein in the diet. Tannins are the main compounds that reduce the availability of N in eucalypt leaves (Degabriel et al. 2009; Wallis et al. 2010) so common brushtail possums feeding on eucalypts may be in a precarious situation. This may explain why they occasionally seek protein-rich foods by feeding opportunistically on dead animals, especially those killed on roads (Heinsohn and Barker 2006).

We tackled the three factors that influence detoxification rates by varying the protein content of the diet, choosing PSMs that either conjugate with an amino acid or not and by varying the concentrations of the PSMs, especially cineole. The experiments with 1,8-cineole demonstrated the importance of the concentration of PSM in mediating the responses of animals, whereby the lower concentrations of cineole (ca 0–8.2 % DM) did not saturate detoxification and thus did not influence food intake. Cineole is a common component of Eucalyptus leaves, which provide a large part of the diet of brushtail possums over much of their range, so they are adept at detoxifying this terpene. Boyle et al. (2005) found that diets containing 5.4 % DM of cineole did not affect feeding, whereas Lawler et al. (1999) showed that brushtail possums would eat diets containing 9.8 % DM of cineole after a long acclimation. The results with the low concentration of cineole prompted our second experiment with higher concentrations of cineole (ca 0–15.8 % DM) that caused a significant reduction in DMI at the high concentrations of cineole. More importantly, animals ingested more cineole when eating the high-protein diet, indicating that they increased the rate of elimination of cineole to match the increased rate of ingestion.

The finding that common brushtail possums willingly ingest more cineole when offered more protein is not new but confirms previous research in both sheep (Villalba and Provenza 2005) and possums (Nersesian et al. 2012). Although these sheep and possums did not maintain food intake at increasing concentrations of cineole, they did eat more if extra protein was available and, in the case of the possums, chose a diet with more protein. None of the authors, however, proposed a physiological mechanism for their findings even though the results suggest a protein cost in detoxification.

In contrast to cineole, detoxification of benzoate has a direct protein cost because it requires the non-essential amino acid glycine as a conjugate to form benzoylglycine (hippuric acid) (Awaluddin and McLean 1985). Possums fed diets containing benzoate ate more if their diet contained higher concentrations of protein. The effect was much stronger than with cineole and thus supports our fourth hypothesis that we would expect the ingestion of benzoic acid to cost the animal more N. In particular, the possums ate significantly more of the high- than the low-protein diet across all four concentrations of benzoate, even though their intake of benzoate did not reach a plateau. Marsh et al. (2005) found that supplementing a benzoate-containing diet with glycine increased the rate that possums could detoxify benzoate, while feeding diets containing either benzoate or glycine alone depressed feeding. The main difference between the studies is that by supplying glycine, Marsh et al. (2005) targeted only the conjugation aspect of detoxification. Glycine is a non-essential amino acid so one might expect no response to feeding it with benzoate because the animal could synthesise it. Many anabolic reactions, however, such as the synthesis of purine nucleotides, heme, glutathione, creatine and serine use glycine (Caldwell and Jakoby 1983) so one explanation is that glycine production is a rate-limiting step and, on low-protein diets, requires increased protein turnover. Our results suggest that adding whole protein to the diet has a similar effect by providing glycine and its precursor, serine and the resources needed to synthesise the detoxification enzymes.

Whole-body protein turnover: the cost of detoxification

The interaction between the concentrations of PSMs and protein in the diet points to increased protein turnover in animals ingesting toxins. The stronger response to protein in brushtail possums fed diets that also included benzoate compared with those offered cineole prompted us to measure whole-body protein turnover rates in two situations: animals eating diets with differing amounts of protein and animals offered a low-protein diet with either a low concentration of benzoate (0.61 % DM) or a higher concentration (4.3 % DM), neither of which we expected to affect food intake and thus protein intake. In contrast to our expectation, however, the higher concentration of benzoate did suppress feeding, which is something we discuss later.

The whole-body protein turnover rates of 135.5 and 112.1 mmol N d−1 measured in common brushtail possums fed the high- and the low-protein diets, respectively, did not differ. When converted, the values (10.2 and 8.5 g crude protein.kg−75 d−1) were similar to values measured by Dellow and Harris (1984) using continuous infusion of [14C]-lysine to measure whole-body protein synthesis in possums (8.6 g crude protein.kg−75 d−1).

In contrast to the rates measured in animals fed diets that differed only in protein content, the whole-body protein turnover rates measured in possums ingesting benzoate provide a much different picture. Higher concentrations of benzoate in the diet caused a change in N turnover rate. The N turnover was 7-times the digestible N intake in possums eating the low-benzoate diet but 23-times the digestible N intake in those on the high-benzoate diet. Thus, N turnover was three times greater on the high-benzoate diet. Animals can tolerate this higher turnover of N if they recover most of the N. These processes, however, are not without cost and some N leaks to the urine. Thus, as more protein turns over, the losses of urinary N increase. We found, in the balance experiment, that animals eating the high-benzoate diet lost 33 % more urinary N, or 60 mg of N per day, than they did when eating the low-benzoate diet.

Given this is the first quantification of whole-body detoxification costs in terms of protein it is pertinent to ask whether the costs are high or low. In other words, is the cost of detoxification ecologically significant or is it just a minor tax on ingested protein? The additional 60 mg of N excreted in the urine by animals eating the high-benzoate diet was about 20 % of their dietary N intake. The ecological implications of this N loss depend on whether N is limiting in the nutritional environment. The most relevant study to place this in perspective is that of Degabriel et al. (2009), who found that common brushtail possums in northern Queensland feeding predominantly on eucalypt foliage produced more young and faster growing young if inhabiting a home range containing trees whose leaves contained more available N. This was remarkable because the difference in the average available N content of trees in a relatively good home range and those in a poor one was small–about 0.20 % DM. Thus, to lose 20 % of N intake for detoxification seems detrimental to an animal with such a finely balanced N intake, especially because this N must be absorbed before the animal can use it in detoxification. In conclusion, when examined from the perspective of N balance or whole-body protein turnover, the costs of detoxification appear high.

There are two main problems to address when interpreting the results of the protein turnover measurements. The first is that the possums fed the high concentration of benzoate (4.3 % DM) in experiment 3 ate less food than they did when fed the low concentration (0.61 % DM), contrary to our finding in experiment 1 when 4.3 % DM benzoate did not suppress feeding. There is no obvious explanation for the difference although it may indicate the experience of the possums or the nature of the experiments—6 days of eating a diet with a medium concentration of benzoate versus 4 days of variable concentrations. Thus, we did not meet our aim that the only difference between the two sets of data would be the concentration of benzoate. Even so, the N intake of possums eating the low- (0.44 g N d−1) and high-benzoate (0.33 g N d−1) diets in experiment 3 encompassed the N intake of the possums eating the low-protein diet in experiment 2 (0.37 g N d−1). This suggests that if benzoate does not influence protein metabolism then we should get similar values for the various parameters in the two experiments. This was not the case. Instead, protein turnover differed markedly in the two experiments.

The second factor to consider is the assumption of the end-point method that the metabolism of the tracer (in this case [15N] glycine) reflects that of the total amino acid mixture (Sprinson and Rittenberg 1949). It is possible that the presence of benzoate influenced the metabolism of [15N] glycine so that the tracer did not reflect the activity of the amino-N pool (Waterlow 2006). If, however, detoxification did use [15N] glycine at substantial rates, one would expect to recover more tracer in animals consuming the most benzoate. This was also not the case.

We did, however, recover more tracer when measuring protein turnover in possums consuming benzoate than we did when measuring turnover in possums eating diets with different concentrations of protein. This was particularly true of the possums eating the diet with the low concentration of benzoate, from which we recovered about three times more 15N than we did from the possums eating the equivalent low-protein diet in experiment 2. This high recovery of marker explains the low flux rates of protein in the benzoate experiment. Presumably, the animal excretes some of this tracer as hippuric acid.

Hippuric acid is a common urinary metabolite formed from the metabolism of many substrates in mammals, including humans (Tremblay and Qureshi 1993; Awaluddin and McLean 1985; Bridges et al. 1970). Some applications of the end-point method measure only a single end-product (either urea or ammonium) and compute the kinetics of protein turnover based only on that product. Given that PSMs appear to influence acid–base balance (and thus the relative proportions of urinary urea and ammonium), it made sense to take a broad view of urinary end-products and measure the total 15N excretion in the urine rather than focus on a single product. Notwithstanding these considerations, we found large differences in all aspects of protein kinetics between the two benzoate diets, so it is likely that the influence of different benzoate concentrations exceeded any possible interference of benzoate on tracer metabolism. This suggests that the differences in protein kinetics in the benzoate experiment represent the costs of benzoate detoxification. Supporting this contention was the willingness of possums to consume more benzoate if they also got more protein.

Acknowledgments

The Animal Experimentation Ethics Committee of the Australian National University approved the work described in this paper, which conforms to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. We thank Ms Bori Cser for helping with the maintenance of the possums. We are grateful to Dr. Hilary Stuart-Williams (Research School of Biology, ANU) for help with mass spectrometry and to Dr. Terry Neeman of the Statistical Consulting Unit, ANU for advice on experimental design and analysis. The comments of Amanda Padovan and an anonymous reviewer greatly improved the manuscript. A DECRA Fellowship (DE120101263) to KJM from the Australian Research Council funded this work.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013