Dose and life stage-dependent effects of dietary beta-carotene supplementation on the growth and development of the Booroolong frog

This study is the first to show that dietary beta-carotene supplementation has a dose and life stage-dependent effect on vertebrate growth and development. Knowledge in this domain could assist with the captive breeding of threatened species.


Introduction
Dietary carotenoids are expected to directly impact vertebrate growth and development by acting as powerful antioxidants. Growth is associated with high metabolic activity and oxygen consumption, accompanied by free radical production and potential for oxidative stress . By scavenging and quenching free radicals produced during metabolism, carotenoids can reduce the oxidative damage caused to cells, tissues and DNA, enabling more efficient cell 1 division and differentiation (Edge et al., 1997), and consequently increased growth . In turn, an increase in growth (and more rapid changes in body size) should facilitate more rapid progression through developmental stages (Gillooly et al., 2002;Marri and Richner, 2014). Additionally, dietary carotenoids might influence growth and development indirectly by improving immune function and general health , or foraging performance (Toomey and McGraw, 2011).
Despite sound theoretical arguments suggesting that carotenoid supplementation will improve vertebrate growth and development, this remains strongly debated because evidence for beneficial effects remains equivocal, both within and between taxa. In some species, clear benefits of carotenoid supplementation have been reported (e.g. Torrissen, 1984;Christiansen and Torrissen, 1996;Cucco et al., 2006;Arulvasu et al., 2013). However, in other species there have been no detectable benefits (e.g. Ř ehulka, 2000; Amar et al., 2004;Wang et al., 2006;Byrne and Silla, 2017), or negative effects (Cothran et al., 2015). This interspecific variation might exist because species differ in their capacity to utilize carotenoids, reflecting different evolutionary histories in carotenoid consumption. For instance, because herbivores consume higher amounts of carotenoids than carnivores, they might have a greater physiological capacity to absorb and process these micronutrients (Tella et al., 2004). Alternatively, much of the interspecific variation reported in the literature might simply reflect the fact that very few studies have tested for dose-response relationships.
Within species, the effect of dietary carotenoids on growth and development may shift depending on the quantities of carotenoids consumed. Based on optimization theory, we should expect that benefits will be restricted to a specific range of metabolically-relevant concentrations, with growth and development optimized at intermediate concentrations (Christiansen et al., 1995;Christiansen and Torrissen, 1996). If carotenoids are provided at extremely low concentrations, effects on growth and development might be negligible, and any benefits might remain undetected. In contrast, carotenoids at extremely high concentrations may stunt growth and development because high concentrations could be toxic, and large amounts of energy might be required to detoxify and remove excess nutrients, limiting growth (Palozza et al., 1997;Cothran et al., 2015). High concentrations could also have pro-oxidant effects, whereby excess carotenoids elicit an increase in oxidative stress, either directly by generating additional ROS (Stoss and Refstie, 1983), or indirectly by suppressing the endogenous antioxidant system (Bouayed and Bohn, 2010). Adding to this complexity, is the possibility that optimal carotenoid concentrations differ between life stages. For most vertebrate species growth will be most rapid during juvenile life stages, so we should expect that benefits associated with ROS scavenging will be most pronounced during early development. Surprisingly, however, past studies testing for effects of carotenoids on growth have largely ignored the potential for life-stage effects, instead focussing on very narrow periods of either juvenile or adult development.
Anuran amphibians (frogs and toads) provide a model group for investigating carotenoid dose-response relationships across different life stages because most species have a biphasic life cycle, with juvenile and adult life stages separated by a distinct period of metamorphosis (Wilbur and Collins, 1973). At each life stage, anurans consume different types of food (in general larvae eat algae and adults eat invertebrates), creating life stage specific differences in carotenoid intake. Growth is usually several orders of magnitude faster during the larval life stage, and assuming that this is when ROS production is highest, this is the life stage when the antioxidant properties of carotenoids will be most beneficial. Antioxidant properties may also be valuable during and immediately after metamorphosis. At this developmental phase, rapid cell division and morphological change can dramatically increase ROS production (Blount et al., 2000). Moreover, a pronounced spike in cellular oxygen concentration during the transition from aquatic to terrestrial life would be expected to cause oxidative stress . To date, the effect of dietary carotenoids on growth and development has been examined in six anuran species, with a focus on the combined effects of five common carotenoids; astaxanthin, cryptoxanthin beta-carotene, lutein and zeaxanthin Dugas et al., 2013;Cothran et al., 2015;Byrne and Silla, 2017). As for studies in other vertebrates, studies in amphibians have mainly tested for effects at one dose or life stage, and findings have been inconsistent. Despite these inconsistencies, benefits to growth and development have been reported for a number of anuran species from several families Dugas et al., 2013), suggesting that carotenoid-mediated growth benefits could be widespread in amphibians.
From an applied perspective, knowledge of carotenoid dose effects on amphibian growth and development may have considerable conservation value. Globally, amphibians are declining faster than any other vertebrate class and the recommended conservation action for threatened species is captive breeding and reintroduction (Gascon et al., 2007). To date, however, the success of captive breeding programs has been limited, in part due to a lack of knowledge regarding amphibian nutritional requirements. Identifying whether specific doses of carotenoids improve anuran growth and development, and whether the importance of carotenoids varies between life stages, may help recovery teams more effectively generate mature individuals needed to maintain viable captive insurance colonies. Furthermore, rapid generation of healthy animals could increase the number of individuals available for release, which is known to be an important predictor of reintroduction success (Tarszisz et al., 2014). The present study aimed to investigate the influence of beta-carotene on the growth and development of the critically endangered Booroolong frog Litoria booroolongensis. The specific aim was to determine whether effects of dietary beta-carotene supplementation on growth and development are either dose (0, 0.1, 1 and 10 mg g −1 ) or life-stage (larval and post-metamorphic) dependent.

Methods
The procedures outlined below were performed following approval by the University of Wollongong's Animal Ethics Committee (approval number AE 14/21).

Study species
Litoria booroolongensis is listed by the IUCN as critically endangered (Hero et al., 2004) and in 2008, under the recommendation of the NSW Department of Environment and Heritage (OEH), a captive breeding program for the species was established at Taronga Zoo (Sydney, Australia).

Experimental animals
A total of 360 L. booroolongensis eggs (from four discrete clutches) were collected from a captive colony maintained at Taronga Zoo. Egg clutches were laid between 02 January 2015 and 17 January 2015, and all eggs hatched within 3 days of oviposition. Within 1-2 days of being laid, egg clutches were transported to the Ecological Research Centre at the University of Wollongong. Individual clutches were held in~700-1000 ml of carbon-filtered water in plastic aquarium bags enriched with atmospheric oxygen. Hatchling tadpoles from each clutch were held communally in 20-l aquaria (one clutch per aquaria) for a 2-week period prior to being transferred to individual experimental containers. Immediately after hatching, tadpoles were fed a basal diet of ground fish flake mixture (75:25 Sera Flora/Sera Sans; SERA, Heinsberg, Germany), ad libitum, every 2 days before the beginning of the experiment. A 2-week acclimation period was imposed in order to ensure all tadpoles were feeding properly and were of similar size at the beginning of the experiment. All tadpoles were 9-15 days old post-hatching at the commencement of the experiment.

Experimental design
Individuals were reared on one of four diets supplemented with different concentrations of beta-carotene. Beta-carotene was selected because it is one of the major carotenoids found in algae and herbivorous insects consumed by anurans (Takaichi, 2011), has been shown to significantly impact anuran reproduction , and has been identified in the skin tissue of wild Booroolong frogs (B. Tinning, 2016, pers. comm.). Larval and post-metamorphic basal diets (containing 0.015 and 0.005 mg g −1 total carotenoids, respectively) were supplemented with beta-carotene at the following concentrations: 0 mg g −1 (control), 0.1 mg g −1 (low), 1 mg g −1 (medium) and 10 mg g −1 (high). These supplement doses were chosen because they encompassed, and expanded on those previously used to investigate the effects of dietary carotenoids on amphibian growth, and other fitnessdetermining traits Dugas et al., 2013;Silla et al., 2016). Each treatment included 72 replicate individuals (288 individuals in total) and individuals remained on the same diet treatment throughout both larval and post-metamorphic life stages. The experiment commenced on 19 January 2015 and the larval stage of the experiment was completed by 13 July 2015, when the last tadpole metamorphosed. The post-metamorphic stage of the experiment commenced when the first individual metamorphosed on 3 March 2015 and was completed by 26 January 2016, when the last frog reached 7-months of age postmetamorphosis, and sex could be determined.
Animals (n = 288) were individually housed in plastic containers (10 cm diameter and 10.5 cm high) throughout both life stages. Each experimental container was enclosed in a black plastic ring so that there was no visual contact between individuals. The experimental containers were held in sets of nine in plastic trays, which were positioned across three shelves in a constant temperature room. The experimental containers were aligned in rows of three, with each diet treatment alternating between rows in the following order; 0, 0.1, 1 and 10 mg g −1 (see Supplementary Fig. S1). The room was artificially illuminated on a 15:9-h, day: night cycle (including twilight lighting for half an hour each cycle to simulate dawn and dusk). In addition to overhead room lighting, UV-B lights (Reptisun 10.0 UV-B 3600 bulb; Pet Pacific, Emu Plains, Australia) were suspended~20 cm above each container, providing 6.5 h per day of UV-B light (between 10 am to 4:30 pm). Ambient temperature in the room was maintained at 22°C (range was 21.9-23°C).

Larval husbandry and nutrition
During the larval life stage, containers were filled with 550 ml of carbon-filtered water. Water changes and water quality testing were conducted according to methods described previously (Keogh et al., 2018). Over the duration of the experiment, ammonia concentrations remained low (<0.50 mg l −1 ), nitrate remained at 0 mg l −1 and pH remained at 7.4. Tadpoles were fed one of the four diets supplemented with different concentrations of beta-carotene (0, 0.1, 1 and 10 mg g −1 ; Supplementary  Table S1). Each diet consisted of 990 mg of ground fish flake mixture and a combination of beta-carotene (0.1, 1 and 10 mg g −1 ) and cellulose microcrystalline powder (435 236; Sigma-Aldrich, Castle Hill, NSW; Supplementary Table S1). Total carotenoid concentrations in the basal diet were <0.015 mg g −1 (see Suppementary Table S2 for a carotenoid profile). Cellulose was used to ensure that each experimental diet consisted of the exact same quantity of feed. Cellulose is commonly used to create balanced designs in dietary studies because it has no nutritional value, is odourless, tasteless and cannot be digested (Dias et al., 1998;Deng et al., 2006). Importantly, in a dietary study run in parallel to our main experiment, cellulose was found to have no effect on the survivorship, growth or development of L. booroolongensis larvae (see Supplementary Table S3). Experimental diets were pre-prepared by suspending 1000 mg of feed in 10 ml of reverse osmosis water (RO water; Sartorius Stedim Biotech, Goettingen, Germany). Feed suspensions were stored in 20 ml syringes and frozen in opaque containers at −20°C until required. On feeding days, syringes were defrosted at room temperature, homogenised and administered in a dropwise manner to ensure an even proportion of feed was administered to each individual. Tadpoles were fed two drops of feed suspension (dry mass; 0.0184-0.0291 g) for the first 4weeks of the experiment, and then three-drops (dry mass; 0.0415-0.0706 g) for the remainder of the experiment (to support increased energetic demands). Tadpoles were fed treatment diets bi-weekly (Monday and Friday) for the duration of the experiment. This feeding regime ensured animals were fed ad libitum.

Quantifying larval survival, growth and development
To quantify tadpole survivorship, each individual was observed every second day and scored as alive or dead. To quantify tadpole growth, each individual was photographed once per fortnight using a Panasonic Lumix DMC-FT20 camera. To ensure accurate photos, water volume in each experimental container was dropped to 100 ml before taking the photo, and then immediately refilled to 550 ml. The development of individual tadpoles was assessed every 2 days using Gosner staging tables. At the emergence of forelimbs (Gosner stages 43-46), the water volume in each experimental container was dropped to 100 ml. To facilitate metamorphosis, a piece of sponge (diameter 9 cm) and a thin layer (<1 cm) of aquarium stones was added to the bottom of each container, and placed on an angle to allow metamorphosing tadpoles to climb from the water. The tadpoles were not fed during this developmental period as tail reabsorption met their nutritional needs. Tadpoles were checked daily, and once an individual had completely reabsorbed its tail, it was blotted with absorbent tissue (kimwipes), weighed and photographed next to a scale. The day of complete tail reabsorption was used to quantify time to metamorphosis. Images of tadpoles and metamorphs were used to determine body length using image analysis software (ImageJ version 1.45s). For each tadpole, head and tail lengths were measured separately, and total tadpole length was calculated by summing these two measures. For each metamorph, snout-vent length was measured. For each individual, two measurements were made and values averaged.

Post-metamorphic husbandry and nutrition
At metamorphic climax (Gosner stage 46), individuals were moved from their larval housing into new containers, which remained in original shelf positions. The post-metamorphic containers and cleaning regime used have been described previously (Keogh et al., 2018). Post-metamorphosis, frogs were fed twice a week with 2-4 day old crickets (Acheta domestica) until they reached sexual maturity at 7 months of age. Total carotenoid concentrations in the basal diet were <0.006 mg g −1 (see Supplementary Table S2 for a carotenoid  profile). Beta-carotene was supplemented by dusting the crickets with 0, 0.1, 1 or 10 mg g −1 of beta-carotene powder. Diets were again balanced using cellulose to control the overall quantity of feed that frogs received. In addition, to ensure healthy bone growth all individuals received calcium supplementation by adding 200 mg of calcium powder containing vitamin D3 (Repti-Cal, Aristopet, Melbourne, Australia) to each gram of treatment powder.
Quantifying post-metamorphic survival, growth and development To quantify adult survivorship, individuals were observed every second day and scored as alive or dead. To quantify adult growth rate, each individual was photographed and weighed every four weeks post-metamorphosis. Frogs were removed from their experimental containers, blotted dry with absorbent tissue paper and weighed to the nearest 0.01 g. Frogs were then photographed using a Panasonic Lumix DMC-FT20 digital camera. The snout-vent length [measured from the tip of the nose (snout) to the bottom of the anus (vent)] was measured for each individual from digital images using image analysis software (ImageJ, version 1.45s). For each individual, two measurements were made and values averaged. An index of body condition was calculated by taking the residuals of an ordinary least squares linear regression of body mass (g) against SVL (mm). Once the frogs reached sexual maturity (at 7 months post-metamorphosis), they were euthanised and sex determined following dissection.

Statistical analyses
The effect of supplement dose on: (i) larval survival until metamorphosis (0-8 weeks post-hatching), and (ii) metamorph survival until sexual maturity (0-7 months post-metamorphosis) was tested using likelihood ratio chi-squared tests. Two tadpoles and six metamorphs were excluded from each analysis because they died as a result of unexpected adverse events. Linear mixed effects models fitted with restricted maximum likelihood (REML) were used to test the effect of supplement dose (0, 0.1, 1 and 10 mg g −1 ) on: (i) tadpole body length (measured from snout tip to tail tip) at experimental weeks 0, 2, 4, 6 and 8, (ii) the body mass (g) of individuals at metamorphosis, and (iii) the time taken (days) until metamorphosis. In each model the treatment dose was the fixed categorical effect and clutch number (1-4) was included as a random effect. Linear mixed effects models fitted with restricted maximum likelihood (REML) were also used to test the effect of treatment dose on metamorph mass, and metamorph body condition each month for seven months post-metamorphosis. In these models, supplement dose and individual sex were the fixed categorical effects, and clutch number (1-4)  data and stabilize variances, we log transformed all response variables before running analyses. For LME models that detected significant effects, differences between treatment levels were compared using post hoc Tukey HSD tests. All statistical analyses were performed using JMP 11.0 software package (SAS Institute Inc. North Carolina, USE).

Time to metamorphosis
The time tadpoles took to metamorphose was highly variable, ranging from 49 to 175 days (mean ± SEM = 81.06 ± 2.042 days), and there was a significant difference in time to metamorphosis among diet treatments (LME: F 3,253 = 3.85, P = 0.010). Tadpoles supplemented with a dose of 1 mg g −1 metamorphosed 1-8 days faster than all other supplement doses (0, 0.1 and 10 mg g −1 ), and the difference in time to metamorphosis between tadpoles supplemented with a dose of 0 mg g −1 and those supplemented with a dose of 1 mg g −1 was significant (Tukey's HSD test: P = 0.006; Fig. 2).
Our finding that larval growth and development was not improved at the low or high-supplement doses indicates that the effects of dietary beta-carotene were dose dependent. A lack of any effect at the low supplement dose may have resulted because beta-carotene was preferentially invested in other fitness-determining traits, such as immune function or exercise performance (as per the carotenoid trade-off hypothesis; Lin et al., 2010). Alternatively, the concentration of beta-carotene used might simply have been too low to significantly alter cellular function and cause any significant change Byrne and Silla, 2017). The lack of any beneficial effect at the high-supplement dose might have several explanations. First, above a certain threshold concentration beta-carotene might begin to have toxic effects, resulting in energy being diverted away from growth and into detoxifying and eliminating excess beta-carotene. Such toxic effects of carotenoids have been reported in other species (Palozza et al., 1997;Cothran et al., 2015). Alternatively, high concentrations might have elicited a mild pro-oxidant effect. Research with cell models has suggested that relatively high doses of beta-carotene can induce an overproduction of ROS due to either: (i) autoxidation of carotenoid metabolites, (ii) changes in the activity of cytochrome P450 enzymes and/or (iii) alterations to endogenous antioxidant defences (Palozza et al., 2003). If high concentrations of beta-carotene have a similar pro-oxidant effect in Booroolong frogs, oxidative damage to cell structures could have slowed somatic growth, resulting in an overall reduction in developmental rate.
Dose-dependent effects were also evident post-metamorphosis. At the low supplement dose (0.1 mg g −1 ) there was no effect on post-metamorphic growth and development, but at both the intermediate (1 mg g −1 ) and high (10 mg g −1 ) doses there was evidence for negative growth effects. Significantly reduced body mass and body condition indicates that postmetamorphic frogs supplemented with intermediate to high doses of beta-carotene had reduced energy and protein reserves and were relatively unhealthy (Ferrie et al., 2014). As for larvae, negative effects of high beta-carotene doses may have resulted from beta-carotene either having a toxic effect or causing pro-oxidant activity, with these effects being more pronounced in post-metamorphic animals. Irrespective of the exact cause, our finding that the effect of beta-carotene dose differed in larval versus postmetamorphic animals supports our second hypothesis that the effects of dietary beta-carotene supplementation are life stage-dependent. The most likely reason for this difference is that individuals produce much higher levels of ROS as larvae than as adults due to significantly faster rates of somatic growth prior to metamorphosis, making the protective antioxidant capacity of beta-carotene more important during early life. In support of this argument, the average percent weight gain of tadpoles per week was more than double that of post-metamorphic individuals. In principle, ROS production in tadpoles is also likely to have increased exponentially as larvae approached metamorphic climax because this is a developmental period that requires extremely high metabolic activity (Pough and Kamel, 1984).
Our findings have important conservation implications. Booroolong frogs are listed as critically endangered, and captive breeding and reintroduction are important recovery actions for this species. Supplying captive Booroolong frogs with dietary beta-carotene has the potential to support these actions in a number of ways. First, the capacity to more quickly generate large numbers of metamorphs might help managers to maintain colonies at optimal sizes and avoid issues associated with natural attrition and inbreeding. Second, more rapid generation of frogs could increase the number of individuals available for release, which could improve reintroduction success by overcoming issues linked to high dispersal, demographic stochasticity or low individual fitness at low population densities (i.e. the Allee effect; Armstrong and Seddon, 2008). Beyond benefits to growth, supplying frogs with dietary beta-carotene also has the potential to significantly improve reproductive success through positive effects on female fecundity and egg quality (see Dugas et al., 2013). Such improvements to captive breeding protocols could reduce the costs of recovery programs and make them more financially viable, and sustainable (Tarszisz et al., 2014). Given that anuran amphibians are declining faster than any other vertebrate group, and that captive breeding is a standard recovery action for threatened species globally, dietary supplementation of beta-carotene (and other carotenoids) could potentially benefit conservation programs for various threatened species. At present, however, it remains uncertain whether benefits to frogs are likely to be widespread. To date, the effect of carotenoids on amphibian growth and development has only been investigated in seven anuran species (including L. booroolongensis), and findings have been equivocal Dugas et al., 2013;Cothran et al., 2015;Byrne and Silla, 2017).
More broadly, our results add to the very small but growing body of evidence that dietary carotenoid supplements improve vertebrate growth and development. Carotenoids have been shown to influence growth and development in fish and birds, however, most studies have examined no more than two supplement doses, or have used a mixture of carotenoid compounds, or carotenoid derived substances, such as spirulina, shrimp head meal, and marigold petal meal (Boonyaratpalin and Unprasert, 1989;James et al., 2006;Sinha and Asimi, 2007;Ezhil et al., 2008;Güroy et al., 2012;Arulvasu et al., 2013). Without testing the effects of carotenoid compounds supplemented at three or more doses, it is impossible to attribute beneficial effects to specific carotenoids and identify optimal supplement doses. To the best of our knowledge, our study is the first to dem- dietary beta-carotene supplementation on vertebrate growth and development. Considering the effect of carotenoids more broadly, only two studies (both in Salmo salar) have tested the effects of individual carotenoid compounds on growth and development at more than two doses. These studies both reported dose effects; fish fed higher doses of astaxanthin exhibited improved growth and survivorship (Christiansen et al., 1995, Christiansen andTorrissen, 1996). Furthermore, these studies provided evidence for life-stage effects; juvenile fish required a higher dose of astaxanthin to exhibit enhanced growth and survivorship. In most vertebrates, growth rates will differ significantly between juvenile and adult life stages. As such, we predict that consideration of the effects of different doses of carotenoids across different life stages will reveal that dietary carotenoid supplementation has more widespread and profound effects on vertebrate growth and development than currently demonstrated.

Supplementary material
Supplementary material is available at Conservation Physiology online.