rspb.royalsocietypublishing.org

An aquatic vertebrate can use amino acids from environmental water Noboru Katayama1,2,†, Kobayashi Makoto1 and Osamu Kishida1,2

Research Cite this article: Katayama N, Makoto K, Kishida O. 2016 An aquatic vertebrate can use amino acids from environmental water. Proc. R. Soc. B 283: 20160996. http://dx.doi.org/10.1098/rspb.2016.0996

Received: 6 May 2016 Accepted: 26 August 2016

Subject Areas: developmental biology, systems biology, ecology Keywords: aquatic animal, dissolved organic nitrogen uptake, nutrient flow, growth rate, Hynobius retardatus, stable isotope

Author for correspondence: Noboru Katayama e-mail: [email protected]

†

Present address: Center for Ecological Research, Kyoto University, Hirano 509-3, Otsu, Shiga 520-2113, Japan. Electronic supplementary material is available online at https://dx.doi.org/10.6084/m9.figshare.c.3469737.

1 Teshio Experimental Forest, Field Science Center for Northern Biosphere, Hokkaido University, Toikanbetsu, Horonobe, Hokkaido 098-2943, Japan 2 Tomakomai Experimental Forest, Field Science Center for Northern Biosphere, Hokkaido University, Takaoka, Tomakomai, Hokkaido 053-0035, Japan

NK, 0000-0003-1602-7497 Conventional food-web theory assumes that nutrients from dissolved organic matter are transferred to aquatic vertebrates via long nutrient pathways involving multiple eukaryotic species as intermediary nutrient transporters. Here, using larvae of the salamander Hynobius retardatus as a model system, we provide experimental evidence of a shortcut nutrient pathway by showing that H. retardatus larvae can use dissolved amino acids for their growth without eukaryotic mediation. First, to explore which amino acids can promote larval growth, we kept individual salamander larvae in one of eight different highconcentration amino acid solutions, or in control water from which all other eukaryotic organisms had been removed. We thus identified five amino acids (lysine, threonine, serine, phenylalanine, and tyrosine) as having the potential to promote larval growth. Next, using 15N-labelled amino acid solutions, we demonstrated that nitrogen from dissolved amino acids was found in larval tissues. These results suggest that salamander larvae can take up dissolved amino acids from environmental water to use as an energy source or a growthpromoting factor. Thus, aquatic vertebrates as well as aquatic invertebrates may be able to use dissolved organic matter as a nutrient source.

1. Introduction Dissolved organic nitrogen (DON), occurring as urea, peptides, amino sugars, or amino acids, is a main constituent of aquatic pools of organic matter [1]. Prokaryotes such as bacteria, cyanobacteria, and archaea absorb DON directly from environmental water to increase their populations [2–4]. The DON absorbed by prokaryotes is believed to be transported indirectly to aquatic animals at higher trophic levels via the food chain for use as a nutrient source. This pathway is known as a ‘bottom-up cascade’, in which DON increases prokaryotic abundance, the abundant prokaryotes are subsequently consumed by protozoa, which are then consumed by zooplankton and so on [5]. In other words, current food-web theory assumes that nutrients flow from DON to aquatic animals via long pathways involving multiple eukaryotic species as intermediary nutrient transporters [6]. However, some studies have suggested that shortcut nutrient flow pathways may exist from DON to marine invertebrates [7–13]. Previous studies have documented direct DON uptake by marine invertebrates, thereby shortening the nutrient flow pathway, and improving the growth of individual invertebrates in DON-enriched environmental water [14]. Most studies have used autoclaved or filtered water in their experiments to remove eukaryotic species from the environmental water, but in many cases aquatic bacteria (i.e. prokaryotes) may not have been completely removed (but see [15]). Therefore, two different shortcut pathways for nutrient flow from DON to aquatic animals may exist: (i) a direct pathway, where the focal aquatic animal absorbs DON directly and (ii) an indirect pathway, in which the aquatic animal consumes bacteria that have taken up DON. Despite this growing body of evidence for the use of shortcut nutrient flow pathways by marine invertebrates, at present, evidence for a shortcut pathway

& 2016 The Author(s) Published by the Royal Society. All rights reserved.

(a) Collection of Hynobius retardatus eggs Hynobius retardatus is a salamander species widely distributed on Hokkaido Island, Japan. In early spring, the adult salamanders, which inhabit forested lands, lay their eggs in lentic ponds. In general, the larvae hatch several weeks after the eggs are laid. During the larval period, H. retardatus typically consumes small aquatic arthropods, frog tadpoles, and conspecific larvae [21,22]. However, for a week after hatching, the larvae are not able to consume these prey items because their mouth is undeveloped [20]. In our experiments, we used hatchlings at the pre-feeding stage (referred to hereafter as ‘larvae’), thereby excluding consumption of prey items as a possible energy source or growth-promoting factor. We collected 20 H. retardatus egg clusters from a single pond in Teshio Experimental Forest of Hokkaido University, Horonobe-chou, Hokkaido, Japan, in late May 2014, and 26 egg clusters from the same pond in late May 2015. In both years, the collected egg clusters were kept in tanks (33.4  20 cm  10 cm high; 15 – 30 egg clusters per tank) filled with 2 l of aged tap water and placed at 48C in a refrigerator with a glass door under natural light conditions (an approximate 15.5 light (L) : 8.5 dark (D) cycle). Several days before the experiment, we transferred the egg clusters to an experimental room maintained at 178C to promote the development of the embryos. We used egg clusters collected in 2014 for Experiment 1 (§2b) and egg clusters collected in 2015 for Experiment 2 (§2c). For the supporting experiment (§2d), we used egg clusters collected in 2015 that had not been used in Experiment 2.

(b) Experiment 1: effects of amino acids on salamander growth The objective of this experiment was to examine whether dissolved amino acids positively affected the growth of the salamander larvae. We considered an enhanced growth rate to indicate that the larvae were using DON for a nutrient source.

(i) Preparation of amino acid solutions In June 2014, we prepared two solutions each of eight different amino acids (tyrosine, phenylalanine, lysine, serine, threonine,

(ii) Experimental conditions Twenty egg clusters at developmental stage 40 (when Hynobius salamanders typically start hatching [23]) were first separately placed in tanks filled with 1 l of autoclaved tap water. After cutting the outer egg sac, we collected 17 larvae with similar body size from each cluster and removed their jelly coats. These larvae were then carefully rinsed with autoclaved water to remove adhesive materials and organisms. In this experiment, we raised these salamander larvae in polypropylene cases (8 cm  5 cm  4 cm high), each filled with 80 ml of one of the amino acid solutions (amino acid treatments) or autoclaved tap water (control treatment). First, the environmental water was added to each case, and then an individual larva was gently put in the case. We individually assigned the 17 full or half-sibling larvae from a single egg cluster to each of 17 experimental treatments (i.e. 8 kinds of amino acids  2 concentration levels þ 1 control). We defined the day on which the larvae were assigned to the treatments as day 1 of the experiment. Larvae from each egg cluster constituted a block, and the experiment consisted of 20 blocks. The larvae in each environmental water treatment were allowed to grow for 12 days in the 178C laboratory room under natural light conditions. On days 5 and 8 of the experiment, we exchanged the environmental water in the cases with freshly sterilized water. Using a scanner, we measured the larvae ventrally on days 1, 8, and 12. For scanning, the larvae were individually placed in small plastic Petri dishes (diameter, 3.5 cm; height, 1 cm) filled with autoclaved tap water (electronic supplementary material, figure S1). On days 8 and 12, we measured four morphological characteristics on scanned images (all units are mm): (i) length from snout to vent (snout – vent length), (ii) length from vent to tail tip (tail length), (iii) maximum gape width (gape width), and (iv) maximum head width (head width; electronic supplementary material, figure S1). Then we calculated total length as the sum of snout – vent length and tail length. Just after hatching, however, the larvae have an indeterminate body axis (i.e. they lean to one side), so we could not correctly measure their gape or head widths on day 1. Therefore, on day 1 we measured only

2

Proc. R. Soc. B 283: 20160996

2. Material and methods

glycine, alanine, and proline; Wako Pure Chemical Industries, Ltd., Japan), one with a concentration of 0.5 mM (mmol l21) and the other with a concentration of 1 mM, as environmental water for the larvae. We used amino acid solutions with high concentrations in this experiment because our primary aim was to ascertain whether this aquatic vertebrate could use dissolved amino acids as a nutrient source, and we could not exclude this possibility without first testing it by using high-concentration amino acid solutions. We autoclaved aged tap water at 1218C for 20 min to kill all organisms in the water. Chemical properties of the tap water are shown in the electronic supplementary material, table S1. After the water temperature returned to the maintenance level (178C), we dissolved 3 mmol of one of the eight amino acids in 3 l of autoclaved aged tap water to obtain a 1 mM solution of that amino acid. Next, we diluted 1 l of the 1 mM solution with 1 l of autoclaved tap water to obtain 2 l of a 0.5 mM solution (leaving 2 l of the 1 mM solution). To assess the properties of the amino acid solutions as environmental water, we also measured the pH of each solution with a pH meter (HI98128 pHepw5, Hanna Instruments, USA). In this study, to minimize the degeneration of amino acids in the environmental water, the water was freshly prepared following this procedure each time the water in a case was replaced (i.e. on days 5 and 8; see §2b(ii)). All implements used in the experiments and for preparation of the environmental water were first sterilized in 0.1% hypochlorous acid or 70% ethanol.

rspb.royalsocietypublishing.org

in aquatic vertebrates is limited, although Glover and his colleagues have shown that DON may be actively transported from environmental water to gill and skin epithelia of hagfish [16 –18]. Here, using newly hatched larvae of a salamander, we tested the hypothesis that an aquatic vertebrate also has the ability to use DON as a nutrient source. To explore whether salamander larvae can use DON as a nutrient source, we carried out laboratory experiments using environmental water free of other eukaryotic organisms in order to (i) identify which amino acids (as a source of DON) could promote growth in the larvae and (ii) determine, by using a stable nitrogen isotope tracer (15N), whether nitrogen from dissolved amino acids could be found in the larval tissue. Salamander larvae are predators [19], but small hatchlings are themselves vulnerable to predation, even from conspecifics [20]. The use of DON to enhance their growth may help them survive this vulnerable life stage. In addition, if dissolved organic matter can be shown to be a source of nutrients for other aquatic vertebrates, especially ones that are difficult to culture in the laboratory because their natural diet is unknown or unavailable, then our results may have practical application.

(iii) Statistical analysis

(c) Experiment 2: nitrogen uptake The objective of Experiment 2 was to examine whether dissolved amino acid– nitrogen was taken up into the bodies of the larvae. In Experiment 1, we confirmed that the amino acid phenylalanine could effectively promote the growth of these salamander larvae (see Results). Thus, for efficient detection of possible amino acid use by the larvae, we used 15N-labelled phenylalanine (15N-Phe; 15N concentration adjusted to 20 atom%) as a stable isotope tracer in Experiment 2. This experiment was conducted in June 2015, using 18 egg clusters.

To assure removal of any eukaryotic organisms, we used autoclaved tap water that had been filtered (filter pore size, 0.2 mm). Just before the experiment, we dissolved 10 mmol of 15 N-Phe in 10 l of the autoclaved tap water to make a 15N-Phe solution with a final concentration of 1 mM. We used 18 egg clusters at developmental stage 40 [23], and from each cluster we selected 18 larvae with similar body size. As in Experiment 1, on day 1 we scanned the larvae ventrally and measured the initial total length, and then we placed the larvae individually in cases (8 cm  5 cm  4 cm high) filled with 15 N-Phe solution or control water. The experimental conditions were the same as in Experiment 1, except for the water exchange interval; in Experiment 2, we exchanged the environmental water for fresh water every other day. On day 7, we again scanned all larvae ventrally to measure their total length.

(iii) Isotope analysis After scanning the larvae, we carefully rinsed them twice with autoclaved tap water to remove any 15N-Phe adhering to their skin. We used the second control group to confirm that the rinse was successful; we dipped the larvae in the second control group in 15N-Phe solution for 10 s before rinsing them twice with autoclaved tap water. We dipped larvae in the first control group and the 15N-Phe treatment group in autoclaved tap water, instead of 15N-Phe solution, for 10 s. Because we did not detect abnormally high 15N concentrations in the second control group samples compared with the first control group samples (see Results), we considered that our rinsing procedure successfully removed 15N-Phe adhering to the larval skin. All larvae were euthanized by immersion in a 0.1% solution of ethyl m-aminobenzoate methanesulfonate (MS-222) [26]. This euthanasia treatment served as a third rinse. Then, we excised the larval tails with scissors. The excised tail tissue was oven-dried at 608C for 48 h; then the dried samples were packed into freezer bags (Ziplocw) and put in a freezer at 2208C until the isotope analysis. After one month, 15N atom% in the samples was analysed using a mass spectrometer at SI Science Co., Ltd. (Kitakatsushika, Japan).

(iv) Statistical analysis (i) Experimental design The experiment consisted of three treatment groups: (i) 15N-Phe treatment, in which a 15N-Phe solution was used as the environmental water for the salamander larvae, (ii) control 1, in which autoclaved tap water was used for the environmental water (control water), and (iii) control 2, in which the larvae were raised in control water but dipped in 15N-Phe solution before the isotope analysis (electronic supplementary material, figure S2). The second control was used to determine whether contamination with 15N solution remaining on larval skin would affect the results. In addition, to avoid contamination from 15N remaining in the alimentary canal, we excised the tail and used this tail tissue for the 15N% analysis. The isotope analysis requires a sample mass more than 2 mg, dry weight; therefore, to obtain this sample mass, for each isotope analysis, we raised six larvae from the same egg cluster and pooled their tail tissue. For this experiment, we used 18 egg clusters (18 replicates); thus, we raised a total of 324 individual salamander larvae (3 treatments  6 individuals per treatment  18 replicates). In this experiment, the growth rate of the individual larvae was measured, and then analysed using a randomized block design (see below). The 15N% was analysed by non-block design, because the samples for this analysis were pooled within egg clusters (blocks).

The total length on day 1 and the growth rate during days 1 – 7 were compared among treatments by one-way mixed ANOVA. We used a Tukey – Kramer test for pair-wise comparison of the growth rate between treatments as a post hoc test. The 15N atom% in the tail tissue was compared among treatments by a Kruskal – Wallis test, and a U-test was used for pair-wise comparison between treatments as a post hoc test. The significance level of a ¼ 0.05 was adjusted by the Bonferroni method for the post hoc pair-wise comparisons.

(d) Supporting experiment To obtain further information about whether non-amino acid– nitrogen can promote larval growth, we maintained the larvae in environmental water containing high concentrations of one of three inorganic nitrogen compounds (NH4NO3, (NH4)2SO4, or KNO3) or another form of organic nitrogen (N-acetylglucosamine, GlcNAc, an amino sugar), and compared the growth rates of these larvae with the growth rates of the larvae maintained in a phenylalanine (Phe) solution or in control water (autoclaved water). In this supporting experiment, which was performed in June 2015, the nitrogen concentration in each solution was adjusted to 1 mM using autoclaved tap water. The pH of all solutions was neutral (average pH (mean + s.d.): NH4NO3,

3

Proc. R. Soc. B 283: 20160996

The total length of the larvae on days 1, 8, and 12, the growth rates during days 1 – 8, days 8 – 12, and days 1 – 12, and the other four morphological traits were compared among treatments by one-way mixed ANOVA in which the origin of the egg clusters was considered to be a random factor. To examine whether total length, growth rate, or any of the other four morphological traits was affected by an amino acid solution, we used a Dunnett test for pair-wise comparisons between the control treatment and each amino acid treatment (i.e. 16 separate comparisons) as a post hoc test. To describe the magnitude of the effect of each dissolved amino acid treatment on larval growth, we calculated the log response ratio of the growth rate, that is, the log-transformed ratio of the ‘growth rate in each amino acid treatment’ to the ‘growth rate in the control treatment’ in each block. The log response ratio is widely used to visualize the magnitude of a positive or negative effect after standardization [25]. We used two-way mixed ANOVA to compare the log response ratio (i.e. the effect magnitude) among the eight different amino acids and between the two concentrations of each amino acid solution (0.5 versus 1 mM).

(ii) Water preparation and experimental conditions

rspb.royalsocietypublishing.org

their length from snout to tail tip (i.e. total length). We used IMAGEJ v. 1.45 s software [24] for all measurements. The growth rate (GR) during days 1 – 8, days 8 – 12, and days 1 – 12 was calculated as follows: GR ¼ (Ly 2 Lx)/(ty 2 tx), where Lx and Ly are total larval length at times tx and ty (i.e. day 1, 8, or 12), respectively. We refer to the rate at which the total length of the salamander larvae increased (mm d – 1) as the growth rate.

(a)

4 ***

21.0

*** ***

*** ***

***

***

**

**

rspb.royalsocietypublishing.org

total length (mm)

21.5

20.5 20.0 19.5

(b)

***

**

*

24.0

**

**

*

*

*

23.5 23.0 22.5

growth rate (mm d–1)

1.0

growth rate (mm d–1)

0.1

growth rate (mm d–1)

(c)

0.90

(d)

(e)

0.9

**

*** *** *** *** ***

*** ***

***

***

***

*

0.8 0.7 0.6

0.9 0.8 0.7 0.6

0.85

**

*** *

***

*** ***

***

*** **

0.80 0.75 0.70 0 0.5 1 water Tyr

0.5 1 Phe

0.5 1 Lys

0.5 1 Ser

0.5 1 Thr

0.5 1 Gly

0.5 1 Ala

0.5 1 Pro

solution type

Figure 1. Effects of dissolved amino acids on the growth of Hynobius retardatus larvae (n ¼ 20 for each treatment). Larval total length on days (a) 8 and (b) 12. Growth rate during days (c) 1–8, (d) 8–12, and (e) 1–12. Amino acid concentrations (mM) are shown on the horizontal axis: Tyr, tyrosine; Phe, phenylalanine; Lys, lysine; Ser, serine; Thr, threonine; Gly, glycine; Ala, alanine; Pro, proline. Error bars denote standard error (s.e.). Dashed lines show the average for the control treatment (water), and asterisks above the error bars indicate the level of statistical significance compared with the control treatment: *p , 0.05, **p , 0.01, ***p , 0.001 (Dunnett test). 6.67 + 0.05; [NH4]2SO4, 6.60 + 0.15; KNO3, 6.62 + 0.07; GlcNAc, 6.74 + 0.04; Phe, 6.70 + 0.06; control water, 6.63 + 0.12). We used a randomized block design for the experiment. We selected 18 larvae from each of the eight egg clusters, and assigned three individuals to each of the six treatment groups (3 inorganic N compound treatments þ GlcNAc þ Phe þ control) (6 treatments  3 individuals per treatment  8 replicates). After measuring their initial total length, we placed the larvae individually in cases (8 cm  5 cm  4 cm high) filled with one of the prepared environmental waters. We exchanged the environmental water with fresh water every other day. On day 7 of the experiment, we measured the total length of the larvae. The total lengths and the growth rates of the larvae were compared among treatments by one-way mixed ANOVA. The Tukey– Kramer test was used as a post hoc test.

3. Results We found significant block effects on all dependent variables in the block-design tests ( p , 0.001). In all experiments, there was no difference in the initial total length of the larvae among treatments at the start of the experiment (Exp. 1, F16,304 ¼ 1.42, p ¼ 0.13; Exp. 2, F2,323 ¼ 0.07, p ¼ 0.94; supporting Exp., F5,143 ¼ 0.83, p ¼ 0.53; electronic supplementary material, figure S3).

(a) Experiment 1: effects of amino acids on salamander growth Total length on days 8 and 12 differed among treatments (day 8: F16,304 ¼ 5.28, p , 0.001, figure 1a; day 12: F16,304 ¼ 3.07, p ,

Proc. R. Soc. B 283: 20160996

total length (mm)

24.5

The growth rate of larvae was significantly greater in the 15NPhe treatment than in the control treatments (F2,323 ¼ 78.37, p , 0.001; figure 2a). Except for one sample, 15N atom% in the tail tissue was constant (0.369%) in both control groups (figure 2b).15N atom% was 0.370% in one control 2 sample, so it is possible that not all of the amino acid adhering to the skin was completely rinsed off. However, because 15N atom% in this sample was almost the same as in the other control samples, we considered 15N contamination in this experiment to be negligible. 15N atom% in the tail tissue of larvae maintained in 15N-Phe solution was significantly greater than that of larvae maintained in control water (x 2 ¼ 48.7, p , 0.001; figure 2b); therefore, we judged that dissolved amino acid–nitrogen was taken up into larval tissue.

average growth rate (mm d–1)

5 1.15 a 1.10 1.05 b

1.00

b

0.95 0.90

15N-Phe

control 1

control 2

15N-Phe

control 1

control 2

(b) 0.390

Atom%

0.380

15N

0.385

0.375 0.370 0.365

Figure 2. Nitrogen uptake from dissolved amino acids by Hynobius retardatus larvae: (a) Larval growth rate (n ¼ 108 for each treatment) and (b) 15N concentration in the larval tail tissue (n ¼ 18 for each treatment). 15N-Phe, 15N-labelled phenylalanine treatment; control 1, water control; control 2, water control in which the larvae were dipped in 15N phenylalanine solution before the isotope analysis. Different letters in (a) indicate a significant difference between treatments (Tukey–Kramer test, p , 0.05). Error bars denote s.e. 1.0

growth rate (mm d–1)

c 0.9 b

b

b b

0.8 a 0.7

0.6

water

NH4NO3 (NH4)2SO4 KNO3 inorganic N

GlcNAc

Phe

organic N

Figure 3. Growth rates during days 1 – 7 of Hynobius retardatus salamander larvae reared in water, inorganic nitrogen solutions, an amino sugar solution, or a phenylalanine solution. NH4NO3, ammonium nitrate; (NH4)2SO4, ammonium sulfate; KNO3, potassium nitrate; GlcNAc, N-acetylglucosamine; Phe, phenylalanine (n ¼ 48 for each treatment). The nitrogen concentration in each solution (except for the control water) was 1 mM. Different letters indicate a significant difference between treatments (Tukey– Kramer test, p , 0.05). Error bars denote s.e.

(c) Supporting experiment In the supporting experiment, there were differences in the growth rate among treatments (F5,143 ¼ 8.32, p , 0.001; figure 3). The growth rate in the Phe treatment was higher than the growth rates in all other dissolved nitrogen treatments as well as that in the control treatment. However, growth rates did not differ between the (NH4)2SO4, KNO3, and GlcNAc treatments and the control, and was lower in the NH4NO3 treatment compared with the control.

Proc. R. Soc. B 283: 20160996

(b) Experiment 2: nitrogen uptake

(a)

rspb.royalsocietypublishing.org

0.001, figure 1b). Growth rates during days 1–8 also differed among treatments (F16,304 ¼ 10.73, p , 0.001; figure 1c). In particular, the growth rates of larvae in five of the amino acid solutions (i.e. tyrosine, phenylalanine, lysine, serine, and threonine) during the first 8 days were higher than the growth rates of those in control water. Although the growth rates during days 8–12 differed among treatments (F16,304 ¼ 1.90, p ¼ 0.020), the growth rates of larvae in the amino acid solutions were not higher than those of larvae in control water (figure 1d). The pattern of the overall growth rate (i.e. during days 1–12) in response to each amino acid was almost the same as the pattern of the growth rate during days 1–8 (F16,304 ¼ 6.05, p , 0.001; figure 1e). The log response ratios of larval growth (electronic supplementary material, figure S4) showed that the magnitude of the effect of the dissolved amino acids on larval growth differed among the different amino acids (F7,285 ¼ 14.95, p , 0.001), and between the two amino acid concentrations tested (F1,285 ¼ 4.64, p ¼ 0.032). A significant interaction effect between amino acid and concentration was also detected (F7,285 ¼ 2.49, p ¼ 0.017). Serine and alanine promoted larval growth more at the higher concentration (0.5 versus 1 mM) (serine: F1,19 ¼ 8.83, p ¼ 0.008; alanine: F1,19 ¼ 5.94, p ¼ 0.025), but the log response ratios of the other amino acids did not differ between the two concentrations ( p . 0.1). At the higher concentration, no amino acid inhibited larval growth. The average pH of the control water was 6.62, and the average pH of six of the seven amino acid solutions was within the range from 6.55 to 6.62 (electronic supplementary material, table S2). The lysine solution, however, was alkaline, with an average pH . 9. Although, based on the average of the log response ratios at the two concentrations, lysine was the most effective in promoting larval growth, there was no consistent trend between the pH of the amino acid solutions and larval growth (electronic supplementary material, table S2). Similar to the total length, the magnitudes of all four morphological traits (snout –vent length, tail length, gape width, and head width) differed among treatments on day 8 (all traits: F16,304 . 3.30, p , 0.001; electronic supplementary material, figure S5). In the tyrosine, phenylalanine, lysine, serine, and threonine treatments, the magnitudes were larger than in the control treatment, except for snout –vent length in the tyrosine treatments and the lower concentration (0.5 mM) serine treatment.

4. Discussion

We did not find any positive effects of the inorganic nitrogen compounds (NH4NO3, (NH4)2SO4, or KNO) or the amino sugar (N-acetylglucosamine) on larval growth (figure 3). Although aquatic bacteria and cyanobacteria can use these forms of dissolved nitrogen [1,31], it is unlikely that the salamander larvae can use them as a nutrient resource. Although only amino acids promoted larval growth in our experiments, the magnitude of the effect differed among amino acids (figure 1). This result raises the question of what properties of amino acids determine the extent of the effect. Among the amino acids tested, lysine, an alkaline amino acid, promoted larval growth most effectively (electronic supplementary material, table S2). In general, water pH is an influential determinant of the growth of aquatic animals. A pH higher or lower than the optimal is harmful to aquatic animals [32], and probably decreases the rate of DON transport to invertebrates [15]. However, in our experiments, pH was probably not a key determinant of the overall effects of the amino acids because the neutral amino acids threonine and serine exhibited effects of the same magnitude as lysine (electronic supplementary material, table S2). The properties of a dissolved amino acid that enable it to meet the energy or nutrient demands of the larvae probably determine the magnitude of the effect (electronic supplementary material, table S3). Among the five amino acids that exhibited large effects, lysine, phenylalanine, and tyrosine are ketogenic amino acids, which can be converted to fatty acids in starvation-stressed organisms [33], and lysine, threonine, and phenylalanine are essential amino acids, ones that animals must obtain from their diet [34]. By contrast, alanine, proline, and glycine, which exhibited a small effect on growth, are neither ketogenic nor essential amino acids. Alternatively, the role of amino acids as a source of dissolved organic carbon may be relevant, because the amino acids that exhibited large effects tended to have higher numbers of carbons per molecule, a potential energy index for a metabolic resource, than those that exhibited small effects. The vital roles played by different amino acids may also influence the magnitude of the effect. For example, lysine can serve as a substrate for growth hormones [35], and tyrosine and phenylalanine are inducers of catecholamine, which forms the basic structure of neurotransmitters [36]. The ability of each amino acid to penetrate the surfaces of the gut, skin, and gills might also determine the effect magnitude; amino acids with a better penetrative ability could have better usability as a source of nutrients or growth-promoting factors.

Proc. R. Soc. B 283: 20160996

(a) Properties of amino acids that promote larval growth

6

rspb.royalsocietypublishing.org

It has generally been assumed that aquatic vertebrates do not absorb DON directly or consume bacteria, but that nitrogen from DON reaches aquatic vertebrates via a long pathway involving multiple eukaryotic species. Here, using larvae of the salamander H. retardatus as a model species, we demonstrated that an aquatic vertebrate could use dissolved amino acids without eukaryotic mediation. Of the eight amino acids tested in Experiment 1, five (lysine, threonine, serine, phenylalanine, and tyrosine) had a remarkably positive effect on the growth of salamander larvae (figure 1). In Experiment 2, using a stable nitrogen isotope tracer (15N), we obtained evidence that nitrogen derived from an amino acid ( phenylalanine) was actually taken up into the tissues of the salamander larvae (figure 2). We carefully designed these experiments to prevent biological factors other than aquatic bacteria from contributing to the effects of the amino acids. Although we could not exclude the possibility that aquatic bacteria might have increased again in the environmental water, our procedure ensured that no living eukaryotes, excepting the salamander larvae, were present in the environmental water. In fact, in another experiment with H. retardatus larvae, using similar procedures, fluorescence in situ hybridization revealed no eukaryotes in the environmental water (N Katayama 2015, personal observation). Hence, the results of this study suggest that a shortcut pathway exists via which DON can be taken up by an aquatic vertebrate without eukaryotic species acting as intermediaries in the nutrient flow. We suggest that there are two possible pathways of nitrogen uptake of the dissolved amino acids by the larval tissue. The first possible pathway is the direct absorption of dissolved amino acids by the salamander larvae via their digestive organs, gills, or skin. Although these early larvae cannot eat, they may drink the amino acid solution and thereby absorb amino acids through the wall of the stomach or gut; their direct absorption through the gills or skin is also possible. Indeed, in hagfish, an ancient vertebrate, the uptake of dissolved amino acids across both gill and skin epithelia has been experimentally shown [16,18]. Although little is known about absorption of amino acids by amphibians, DON in the form of urea is capable of being transported across amphibian skin [27]. Alternatively, an indirect absorption pathway involving bacteria is also possible. The salamander larvae may take up amino acid-derived materials by consuming bacteria that had absorbed the dissolved amino acids. Changes in gut microbiomes, which generally affect the survival and growth of the host animal [28], are also a possible mechanism of the growth effects of the amino acids. At present, we have no evidence supporting any particular uptake mechanism or absorption pathway. Future studies are necessary to identify the specific process or mechanism of uptake of the dissolved amino acids. Regardless of the precise pathway, our experimental results suggest that the nutrients (here, amino acids) transferred into the salamander larvae improved larval growth, which implies that the larvae might use the nutrients as an energy source. Alternatively, the nutrients might have some other growth-promoting effect: amino acids may, for example, serve as substrates for growth hormones [29] or contribute to osmoregulation [18]. The way in which the dissolved amino acids influence the growth of this aquatic animal should be examined in the future.

In our experiments, the effect of amino acids on growth promotion in the salamander larvae disappeared eight days after hatching (figure 1). The disappearance of this effect may be due to the balance between energy supply and demand. Because the energy demands of organisms increase with the size of the organism [30], once the larvae have grown to a large body size, they might not be able to obtain adequate energy from dissolved amino acids. Alternatively, if the larvae take up the amino acids through their skin or gills, a developmental change in the function of these organs might explain the disappearance of growth promotion. In any case, there is a strong growth-promoting effect of dissolved amino acids in the early post-hatch stage (figure 1).

(c) Perspectives Up until now, DON has been generally assumed to follow a long nutrient flow pathway to vertebrates, via multiple eukaryotic species linked by successive prey–predator interactions. Thus, food-web theory has been developed on the basis of this assumption. This study offers a novel perspective, namely, that an aquatic vertebrate can use DON as a nutrient source without the mediation of other eukaryotic species. The present results are clear evidence of the existence of a shortcut nutrient pathway from DON to aquatic vertebrates. This novel perspective may require the revision of current theories about energy flow as well as community organization in aquatic systems. For example, whereas it is usually supposed that the growth and survival of aquatic vertebrates are supported by nutrients derived from vegetable primary production in grazing food-chains, this novel view leads to the idea that nutrients to support vertebrate growth and survival can be, at least partly, provided by the initial links of a detrital food-chain, in which DON is released from organic debris and absorbed by aquatic bacteria. Many aquatic vertebrates, such as salamander larvae, are predators, and larger predators require a larger amount of food obtained from a variety of prey items [22]; therefore, predators that can jump-start their growth by using DON as a nutrient source can have a great impact on aquatic prey in future prey– predator interactions and thus drive organizational changes in the aquatic community. The utilization of DON by aquatic vertebrates also has potential application in aquaculture. For fish cultivation, understanding the nutritional demands of fishes is an important area of research [44]. For example, hatchery fishes often suffer from nutritional deficiencies because of insufficient quantities of essential amino acids [45]. To correct for this nutritional deficiency, fish farmers are encouraged to supplement the feed of hatchery fishes with an artificial amino acid-enriched diet [46]. If hatchery fishes can absorb amino acids from environmental water as well as from their diet, the addition of amino acids to their environmental water would be a convenient alternative way of improving the growth and survival of hatchery fishes. In addition, efforts to cultivate some economically valuable fishes have failed owing to our lack of knowledge about their natural diets. For such fishes, dissolved organic matter may have potential as an alternative supplemental diet. Of course, the extent to which aquatic vertebrates other than H. retardatus can use DON as an energy source is unknown. Future investigations targeting other aquatic vertebrates should breathe new life into studies of food webs and aquaculture. Ethics. Our work conforms to Japanese guidelines for the proper conduct of animal research.

Data accessibility. Data of Experiments 1 and 2 and the electronic supplementary material are deposited in the Dryad Digital Repository doi:10.5061/dryad.3g7n2 [47].

7

Proc. R. Soc. B 283: 20160996

Since Pu¨tter [37] hypothesized that DON is available as an energy source for marine invertebrates over 100 years ago, many researchers have searched for evidence of DON (especially amino acid) utilization by invertebrates (for reviews, see [38,39]). For example, Stephens & Schinske [40] reported that when marine invertebrates, such as Porifera, Ectoprocta, Annelida, Mollusca, and Echinodermata, were reared in seawater to which DON had been added, the amino acid concentration in the seawater decreased. They attributed this result to the consumption of dissolved amino acids by the invertebrates. Thereafter, tracer experiments using 14C-labelled amino acids demonstrated that dissolved amino acids are transported into the bodies of invertebrates [41]. Although in many early studies, researchers were unable to remove bacteria from marine invertebrates, Manahan et al. [15] overcame this worrisome problem by raising bacteria-free sea urchin larvae from axenic eggs; as a result, they were able to demonstrate that in a sterile environment the bacteria-free larvae could directly absorb dissolved amino acids. Such an approach should be effective in our system to determine the relative importance of direct and indirect (via bacteria) effects of dissolved amino acids on the growth of vertebrate species. Furthermore, genes encoding functional amino acid transporters have recently been found in a sea urchin [42] and in sea stars [43]. Thus, it is now widely accepted not only that dissolved amino acids improve invertebrate performance, but also that aquatic invertebrates can take up dissolved amino acids from environmental water to satisfy their energy demands. In contrast to the many studies arguing that aquatic invertebrates can use DON, only a few reports have suggested that hagfish, an aquatic vertebrate, can actively take up DON from environmental water into its own tissue [16,18]. However, whether aquatic vertebrates can use DON as an energy source remains an open question. In this study, although we did not identify the precise pathway, we showed that H. retardatus larvae incorporated nitrogen from dissolved amino acids into their body tissue (figure 2) and that the growth of larvae kept in an amino acid solution was improved (figures 1 and 2). These findings support the possibility that these larvae can use DON as a nutrient resource. Larvae of H. retardatus salamander usually inhabit cannibalistic environments, in which large individuals consume smaller conspecifics [20,26]. If intraspecific variation in the ability to use amino acids exists, then the advantage of using amino acids to salamander larvae in conspecific interactions can be demonstrated; larvae with a strong utilization ability would grow more rapidly than those with only a weak ability and thus would, by eating their smaller conspecifics, achieve greater success as cannibals. For their survival in such an environment, it is critical for newly hatched larvae to obtain as much energy as possible in order to grow rapidly during the early post-hatching stage. However, newly hatched salamander larvae cannot feed on solid prey items because their mouths remain incompletely developed; thus, they are forced into an oligotrophic situation. Therefore, the ability to use dissolved amino acids might be an adaptation that helps them survive during this pre-feeding life stage. Although this ecological argument is plausible, the concentrations of the amino acids solutions used in our experiments were considerably

higher than typical dissolved amino acid concentrations in natural ponds (usually less than 1 mM). Therefore, to evaluate the early uptake of amino acids in an ecological context, future experiments should be designed to test whether dissolved amino acids at the levels found in natural pond water have growth-promoting effects in salamander larvae.

rspb.royalsocietypublishing.org

(b) Dissolved organic nitrogen utilization by aquatic animals

Acknowledgements. We thank K. Takasu, A. Yamaguchi, M. Nosaka,

N.K. and O.K. developed the methodology and wrote the manuscript. N.K. carried out all experiments and analysed the data and K.M. provided editorial advice.

T. Sato, H. Ohiwa, N. Sakai, and the staff at Teshio Experimental Forest of Hokkaido University for their support during this study. We are grateful to the staff of Horonobe Town Office for providing data of chemical properties of the tap water used in our experiments. We also thank S. Duhon for correcting the English of the text.

Competing interests. We have no competing interests. Funding. This work was supported from KAKENHI grant no. 26650152 from the Japan Society for the Promotion of Science.

References Berman T, Bronk DA. 2003 Dissolved organic nitrogen: a dynamic participant in aquatic ecosystems. Aquat. Microb. Ecol. 31, 279–305. (doi:10.3354/ame031279) 2. Antia NJ, Harrison PJ, Oliveira L. 1991 The role of dissolved organic nitrogen in phytoplankton nutrition, cell biology, and ecology. Phycologia 30, 1– 89. (doi:10.2216/i0031-8884-30-1-1.1) 3. Jørgensen NOG, Kroer N, Coffin RB, Yang XH, Lee C. 1993 Dissolved free amino acids, combined amino acids, and DNA as sources of carbon and nitrogen to marine bacteria. Mar. Ecol. Prog. Ser. 98, 135–148. (doi:10.3354/meps098135) 4. Bronk DA. 2002 Dynamics of DON. In Biogeochemistry of marine dissolved organic matter (eds DA Hansell, CA Carlson), pp. 153–247. San Diego, CA: Academic Press. 5. Chauhan A, Cherrier J, Williams HN. 2009 Impact of sideways and bottom-up control factors on bacterial community succession over a tidal cycle. Proc. Natl Acad. Sci. USA 106, 4301 –4306. (doi:10.1073/pnas. 0809671106) 6. Li Y, Gal G, Makler-Pick V, Waite AM, Bruce LC, Hipsey MR. 2014 Examination of the role of the microbial loop in regulating lake nutrient stoichiometry and phytoplankton dynamics. Biogeosciences 11, 2939– 2960. (doi:10.5194/bg11-2939-2014) 7. Ferguson JC. 1980 The non-dependency of a starfish on epidermal uptake of dissolved organic matter. Comp. Biochem. Physiol. 66, 461 –465. (doi:10. 1016/0300-9629(80)90192-9) 8. Ferrier MD. 1991 Net uptake of dissolved free amino acids by four scleractinian corals. Coral Reefs 10, 183–187. (doi:10.1007/BF00336772) 9. Jaeckle WB, Manahan DT. 1992 Experimental manipulations of the organic composition of seawater: implications for studies of energy budgets in marine invertebrate larvae. J. Exp. Mar. Biol. Ecol. 156, 273–284. (doi:10.1016/00220981(92)90251-5) 10. Grover R, Maguer JF, Allemand D, Ferrier-Page`s C. 2008 Uptake of dissolved free amino acids by the scleractinian coral Stylophora pistillata. J. Exp. Biol. 211, 860–865. (doi:10.1242/jeb.012807) 11. Skikne SA, Sherlock RE, Robison BH. 2009 Uptake of dissolved organic matter by ephyrae of two species of Scyphomedusae. J. Plank. Res. 31, 1563 –1570. (doi:10.1093/plankt/fbp088) 12. Gori A, Grover R, Orejas C, Sikorski S, Ferrier-Page`s C. 2014 Uptake of dissolved free amino acids by four cold-water coral species from the

13.

14.

15.

16.

17.

18.

19. 20.

21.

22.

23.

24.

Mediterranean Sea. Deep Sea Res. Part II 99, 42 –50. (doi:10.1016/j.dsr2.2013.06.007) Brothers CJ, Lee RW, Nestler JR. 2015 The uptake of dissolved organic material by the sea cucumber Parastichopus californicus (Stimpson) and its potential role in visceral regeneration. J. Exp. Mar. Biol. Ecol. 469, 69 –75. (doi:10.1016/j.jembe.2015. 04.016) Johnson CH, Wendt DE. 2007 Availability of dissolved organic matter offsets metabolic costs of a protracted larval period for Bugula neritina (Bryozoa). Mar. Biol. 151, 301–311. (doi:10.1007/ s00227-006-0492-0) Manahan DT, Davis JP, Stephens GC. 1983 Bacteriafree sea urchin larvae: selective uptake of neutral amino acids from seawater. Science 220, 204– 206. (doi:10.1126/science.220.4593.204) Glover CN, Bucking C, Wood CM. 2011 Adaptations to in situ feeding: novel nutrient acquisition pathways in an ancient vertebrate. Proc. R. Soc. B 278, 3096– 3101. (doi:10.1098/rspb.2010. 2784) Bucking C, Glover CN, Wood CM. 2011 Digestion under duress: nutrient acquisition and metabolism during hypoxia in the Pacific hagfish. Physiol. Biochem. Zool. 84, 607–617. (doi:10.1086/662630) Glover CN, Bucking C, Wood CM. 2013 The skin of fish as a transport epithelium: a review. J. Comp. Physiol. B 183, 877 –891. (doi:10.1007/s00360-0130761-4) Wells KD. 2007 The ecology and behavior of amphibians. Chicago, IL: Chicago University Press. Kishida O, Tezuka A, Ikeda A, Takatsu K, Michimae H. 2015 Adaptive acceleration in growth and development of salamander hatchlings in cannibalistic situations. Funct. Ecol. 29, 469 –478. (doi:10.1111/1365-2435.12361) Nosaka M, Katayama N, Kishida O. 2015 Feedback between size balance and consumption strongly affects the consequences of hatching phenology in size-dependent predator–prey interactions. Oikos 124, 225 –234. (doi:10.1111/oik.01662) Takatsu K, Kishida O. 2015 Predator cannibalism can intensify negative impacts on heterospecific prey. Ecology 96, 1887–1898. (doi:10.1890/141616.1) Iwasawa H, Yamashita K. 1991 Normal stages of development of a hynobiid salamander, Hynobius nigrescens Stejneger. Jap. J. Herpetol. 14, 39 –62. Schneider CA, Rasband WS, Eliceiri KW. 2012 NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671 –675. (doi:10.1038/nmeth.2089)

25. Hedges LV, Gurevitch J, Curtis PS. 1999 The metaanalysis of response ratios in experimental ecology. Ecology 80, 1150– 1156. (doi:10.2307/177062) 26. Takatsu K, Kishida O. 2015 Allometric equations for estimation of energy contents from body length for common amphibians (Hynobius retardatus and Rana pirica) in Hokkaido, Japan. Herpetol. Notes 8, 187–191. 27. Dytko G, Smith PL, Kinter LB. 1993 Urea transport in toad skin (Bufo marinus). J. Pharmacol. Exp. Ther. 267, 364–370. 28. McFall-Ngai M et al. 2013 Animals in a bacterial world, a new imperative for the life sciences. Proc. Natl Acad. Sci. USA 110, 3229–3236. (doi:10.1073/ pnas.1218525110) 29. Harvey S, Scanes CG, Daughaday WH. 1995 Growth hormone. Boca Raton, FL: CRC Press. 30. Nagy KA. 2005 Field metabolic rate and body size. J. Exp. Biol. 208, 1621–1625. (doi:10.1242/jeb. 01553) 31. Wheeler PA, Kirchman DL. 1986 Utilization of inorganic and organic nitrogen by bacteria in marine systems. Limnol. Oceanogr. 31, 998 –1009. (doi:10.4319/lo.1986.31.5.0998) 32. Morris R, Taylor EW, Brown DJA, Brown JA. 2008 Acid toxicity and aquatic animals. Cambridge, UK: Cambridge University Press. 33. Manninen AH. 2004 Metabolic effects of the verylow-carbohydrate diets: Misunderstood ‘Villains’ of human metabolism. J. Int. Soc. Sports Nutr. 1, 7–11. (doi:10.1186/1550-2783-1-2-7) 34. Schmidt-Nielsen K. 1997 Animal physiology: adaptation and environment. Cambridge, UK: Cambridge University Press. 35. Liao SF, Wang T, Regmi N. 2015 Lysine nutrition in swine and the related monogastric animals: muscle protein biosynthesis and beyond. SpringerPlus 4, 147. (doi:10.1186/s40064-015-0927-5) 36. Flatmark T. 2000 Catecholamine biosynthesis and physiological regulation in neuroendocrine cells. Acta. Physiol. Scand. 168, 1–17. (doi:10.1046/j. 1365-201x.2000.00596.x) 37. Pu¨tter A. 1908 Die Ernahrung der Wassertiere. Z. Allg. Physiol. 7, 283. 38. Stephens GC. 1988 Epidermal amino acid transport in marine invertebrates. Biochim. Biophys. Acta 947, 113–138. (doi:10.1016/0304-4157 (88)90022-6) 39. Wendt DE, Johnson CH. 2006 Using latent effects to determine the ecological importance of dissolved organic matter to marine invertebrates. Integr. Comp. Biol. 46, 634–642. (doi:10.1093/icb/icl025)

Proc. R. Soc. B 283: 20160996

1.

8

rspb.royalsocietypublishing.org

Authors’ contributions. K.M. and O.K. originally formulated the idea, and

45. Takagi S, Murata H, Goto T, Endo M, Yamashita H, Ukawa M. 2008 Taurine is an essential nutrient for yellowtail Seriola quinqueradiata fed non-fish meal diets based on soy protein concentrate. Aquaculture 280, 198–205. (doi:10.1016/j.aquaculture.2008.05.012) 46. Jennings S, Kaiser M, Reynolds JD. 2001 Marine fisheries ecology. Oxford, UK: Blackwell Science. 47. Katayama N, Makoto K, Kishida O. 2016 Data from: An aquatic vertebrate can utilize amino acids from environmental water. Dryad Digital Repository. http://dx.doi.org/10.5061/dryad.3g7n2

9

Proc. R. Soc. B 283: 20160996

developing sea urchins (Strongylocentrotus purpuratus). Biol. Bull. 217, 6 –24. 43. Applebaum SL, Ginsburg DW, Capron CS, Manahan DT. 2013 Expression of amino acid transporter genes in developmental stages and adult tissues of Antarctic echinoderms. Polar. Biol. 36, 1257–1267. (doi:10.1007/s00300-013-1345-1) 44. Li P, Mai K, Trushenski J, Wu G. 2009 New developments in fish amino acid nutrition: towards functional and environmentally oriented aquafeeds. Amino Acids 37, 43– 53. (doi:10.1007/s00726-0080171-1)

rspb.royalsocietypublishing.org

40. Stephens GC, Schinske RA. 1961 Uptake of amino acids by marine invertebrates. Limnol. Oceanogr. 6, 175–181. (doi:10.4319/lo.1961.6.2. 0175) 41. Jaeckle WB, Manahan DT. 1989 Feeding by a ‘nonfeeding larva’: uptake of dissolved amino acids from seawater by lecithotrophic larvae of the gastropod Haliotis rufescens. Mar. Biol. 103, 87 –94. (doi:10.1007/BF00391067) 42. Meyer E, Manahan DT. 2009 Nutrient uptake by marine invertebrates: cloning and functional analysis of amino acid transporter genes in