Ethical statement
The authors declare that the research contained within this manuscript complied with the Animal Welfare Guidelines for Journal Publication. All animal housing and experimental protocols were in compliance with the Japanese Government Animal Protection and Management Law (No. 105) as well as with the Japanese Government Notification on Feeding and Safekeeping of Animals (No. 6).
Experimental animals and design
The X. tropicalis tadpoles used in this study were obtained from an inbred line prepared in the Amphibian Research Center of Hiroshima University, Japan. Twenty tadpoles of similar size (stage 51) were randomly selected from the holding tank and placed in 2 L glass aquaria (25 × 10 × 8 cm) containing 1 L of tap water treated with activated charcoal (Tsurumi Coal, Kanagawa, Japan) for 24 hr. The tadpoles were fed Sera Micron (Sera, Heinsberg, Germany), reared at 25 °C during experimental analysis, and selected for experimentation using a uniform randomized protocol previously described[18] (the complete experimental protocol is summarized in Fig. 1a). Six predator exposure conditions were used, including an unexposed control, 10 days (Ex 10day), 48 hr (Ex 48hr), 24 hr (Ex 24hr), and 6 hr (Ex 6hr) exposure as well as a group reared with the predator for five days and without the predator for another five days (5 day-Out). Eighteen aquaria were set up for predation-exposed and control tadpoles. Additionally, three, one, one, one, and two extra aquaria were prepared as backups for the Ex 10day, Ex 48hr, Ex 24hr, Ex 6hr, and Ex 5day-Out groups, respectively. Thus, each treatment group included three replicates and a variable number of backups. All aquaria (including backups) were randomly arranged, separated using cardboard as blinders, and placed in natural day–night conditions in an experimental room at 25 °C. Salamander (H. retardatus) egg masses were sampled from natural ponds in Asahikawa and Nayoro, Hokkaido, Japan (Map attached to Supplementary Figure 1), and reared to the larval stage. Predation experiments were initiated by introducing one salamander larva into each aquaria containing 20 tadpoles.
RNA extraction
RNA was extracted for RNA-seq. Three tadpoles from each replicate aquarium were pooled as a single sample. Thus, nine tadpoles were used per treatment group. Brain tissues were dissected under a microscope, soaked in RNAlater (RNA stabilization reagent, Qiagen, Hilden, Germany), stored overnight at 4 °C, and maintained at -80 °C until RNA extraction. Total RNA was purified from each sample using a RNeasy Mini Kit (Qiagen) following the manufacturer’s protocol. Total RNA quality was confirmed at an absorbance ratio A260/A280 (range, 1.7–2.1) and visual examination of 18S to 28S ribosomal RNA via gel electrophoresis.
Library preparation and sequencing
Library preparation and sequencing of extracted total RNA were performed by the BGI Group (Shenzhen, China).
Transcriptome analysis through RNA-seq and IPA
Differential expression analysis was performed using the R-based package EdgeR (v.3.8.6)[44]. EdgeR filters out low-expression genes, while those with more > 1 count per million (CPM) in at least two samples are retained. After normalizing the trimmed mean of their M-values, DEGs were extracted from pairwise comparisons of the control vs. 24 and 48 hr experimental groups using an exact test with FDR < 0.05. These extracted DEGs were visualized in a heat map (Fig. 2a). IPA (Qiagen) was used to identify functional networks of genes expressed in the brain.
Visualization of hemoglobin using a hyperspectral camera
Hemoglobin measurements in the Xenopus larvae were performed using a CosmosEye HSC1702 camera (Hokkaido Satellite Co., Ltd. Japan) mounted on a Leica (Wetzlar, Germany) stereomicroscope. Visualization of hemoglobin was achieved by analyzing the hyperspectral images generated using the normalized difference vegetation index[45]. The ratio of the oxygen–hemoglobin in tadpole brains was analyzed using the methods of Tetschke et al.[46] for the calculation of oxygen saturation with normalized absorption spectra.
CE-TOF-MS analysis of X. tropicalis tadpole brains
Tadpoles were anesthetized in an ice water bath and immediately dissected. The brains were excised under a microscope, placed into a tube, and instantly frozen with liquid nitrogen. Two brains were combined into a single sample, and three samples were treated as one experimental group. A detailed description of the CE-MS procedure can be found in Soga et al.[47].
Measurement of free radicals and sugar content in the X. tropicalis tadpoles
Xenopus tropicalis tadpoleswere sampled, dried, weighed, immediately frozen in liquid nitrogen, and maintained at -80 °C until analysis. Tadpoles were homogenized using a Polytron homogenizer (Kinematica, Switzerland) at maximum speed on ice and then centrifuged at 12,000 × g for 15 min at 4 °C. Thereafter, 20 μL of supernatant was used for free radical analysis according to the d-ROMs test (Wismerll, Tokyo, Japan), and 50 μL was used for sugar content estimation using the Dri-Chem system (FDC3030, Fujifilm, Tokyo, Japan).
Real-time PCR
Real-time PCR (Rotor-Gene thermal cycler, Qiagen) was performed to determine the gene expression profiles of sod3, phosphoserine phosphatase, g6pc3, hba3, alas2, fumarylacetoacetate hydrolase, phosphoserine aminotransferase, and pygm. The primers for these genes are summarized in Supplementary Table 4. PCR was run for 35 cycles as follows: incubation at 95 °C for 5 min, annealing at 95 °C for 5 s, and extension at 60 °C for 10 s. DNA content was calculated as per the ΔΔCT method.
Statistical analysis for PCA
Fifteen body parts of X. tropicalis were measured in the five experimental groups, including controls. Relationships between the measured variables and the five groups were subjected to PCA using SPSS (version 20). The mean first and second principal components of each experimental group were plotted against those of the total samples, which included 30 control, 65 Ex10 day, 27 Ex 5 day-Out, 28 Ex 48hr, and 44 Ex 24hr samples.
Immunohistochemistry
Tadpoles were fixed with 4 % paraformaldehyde solution overnight at 4 ℃, prepared for whole-body immunostaining using sequential dehydration with ethanol, and stored in 100 % ethanol at -80 ℃ until use. For whole-body immunostaining, a sample was depigmented with hydrogen peroxide solution (methanol:H2O2, 2:1) and rehydrated with ethanol and PBSTW (1X PBS + 0.1 % Tween 20). The samples were then washed with PBSTW(x) (PBSTW+5 % dimethyl sulfoxide), blocked with PBSTW(x) + 10 % normal goat serum (NGS), and incubated overnight with the primary antibody HBE1 anti-rabbit (PA5-77997; Thermo Fisher) diluted (1:500) in PBSTW(x) and 1 % NGS at 4 ℃. After washing the primary antibody with PBSTW(x) and 1 % NGS three times for 5 min, the Alexa Fluor 546 anti-rabbit secondary antibody diluted (1:1,000) in PBSTW(x) and 1 % NGS was incubated overnight at 4 ℃. After further washing the secondary antibody with PBSTW(x) and 0.1 % NGS, the samples were embedded in DPX mountant for histology (Sigma-Aldrich) and observed via confocal microscopy (Leica TCS SP8).
Statistics and reproducibility
RNA-seq read data were analyzed using AfterQC (v.2.7)[48]. Mapping to the X. tropicalis reference genome was performed using HISAT2 (v.2.1.0)[49]. To analyze gene alignment, the SAM file, which is in text format, must first be converted to its binary counterpart, BAM format, a very tractable format for SAMtools (v.1.9)[50]. Reads mapped to each gene were counted using StringTie (v.1.3.4)[51], and the differential expression analysis was performed in R. DEGs were extracted from pairwise comparisons of the control vs. 24 hr and 48 hr experimental groups using an exact test with FDR < 0.05 in EdgeR.
IPA (Qiagen) was used to identify functional networks of genes, where only those with fold changes > 1.5 were analyzed (Content version: 48207413). Statistical analyses of PCR, CE-MS, hemoglobin, glucose concentrations, and ROS levels were performed via one-way analysis of variance, followed by post hoc Bonferroni or Dunnett’s T3 tests.