Abstract

   Temperature influences nearly all biochemical, physiological and life
   history activities of fish, but the molecular mechanisms underlying the
   temperature acclimation remains largely unknown. Previous studies have
   identified many temperature-regulated genes in adult tissues; however,
   the transcriptional responses of fish larvae to temperature stress are
   not well understood. In this study, we characterized the
   transcriptional responses in larval zebrafish exposed to cold or heat
   stress using microarray analysis. In comparison with genes expressed in
   the control at 28°C, a total of 2680 genes were found to be affected in
   96 hpf larvae exposed to cold (16°C) or heat (34°C) for 2 and 48h and
   most of these genes were expressed in a temperature-specific and
   temporally regulated manner. Bioinformatic analysis identified multiple
   temperature-regulated biological processes and pathways. Biological
   processes overrepresented among the earliest genes induced by
   temperature stress include regulation of transcription, nucleosome
   assembly, chromatin organization and protein folding. However,
   processes such as RNA processing, cellular metal ion homeostasis and
   protein transport and were enriched in genes up-regulated under cold
   exposure for 48 h. Pathways such as mTOR signalling, p53 signalling and
   circadian rhythm were enriched among cold-induced genes, while
   adipocytokine signalling, protein export and arginine and praline
   metabolism were enriched among heat-induced genes. Although most of
   these biological processes and pathways were specifically regulated by
   cold or heat, common responses to both cold and heat stresses were also
   found. Thus, these findings provide new interesting clues for
   elucidation of mechanisms underlying the temperature acclimation in
   fish.

Introduction

   Environmental temperature variations affect many properties and
   functions of biomolecules and structural components of the cell, such
   as folding, assembly, activity and stability of proteins [31][1],
   structure and rigidity of lipids [32][2], [33][3], and fluidity and
   permeability of cell membrane [34][4], [35][5]. Due to the ubiquitous
   temperature dependence of structures and functions of various cellular
   elements, even small temperature changes would adversely disturb
   cellular homeostasis and attenuate physiological performance. The body
   temperature of most fishes equilibrates rapidly with ambient
   temperature, so water temperature is suggested to be the abiotic master
   factor which virtually controls and limits all the biochemical,
   physiological and life history activities [36][6], [37][7]. Under
   natural conditions, fishes may experience various sources of
   temperature fluctuations, including thermocline temperature variation,
   rapid changes in solar heat, abnormal water movements, rapid
   precipitation events or changes in seasonal temperatures [38][7]. To
   combat the adverse effects elicited by temperature fluctuations and
   maintain normal cellular functions at changed temperature, fishes have
   evolved versatile mechanisms that enable them to survive harsh
   environments with temperatures ranging from −2°C in the polar oceans to
   over 45°C in hot springs [39][8]. Some eurythermal fishes may adapt
   seasonally to temperatures from near freezing to over 36°C [40][9] and
   even endure a daily temperature cycle over a 20°C range [41][10].

   The success of fishes in adaptation to an enormous range of
   environments has therefore led to considerable efforts on investigating
   how fishes respond and acclimate to temperature disturbance. Results
   from early physiological and biochemical studies have indicated that
   fishes may adapt to temperature variation through “biochemical
   restructuring”, namely changing the quantities of certain molecular
   species and the types of molecules present in the cells [42][11]. Many
   aspects of cellular biochemistry are involved in this restructuring
   process and well-defined adaptive responses include producing
   temperature specific isozymes [43][11], [44][12], changing the content
   of membrane lipid and the degree of fatty acid unsaturation [45][13],
   recruiting different muscle fiber types [46][14], synthesizing
   molecular chaperones [47][15], [48][16], changing mitochondrial
   densities and their properties [49][17], [50][18], and generating
   antifreeze protein during long-term evolutionary adaptation to freezing
   environment [51][19]. Obviously, these findings have elucidated the
   biochemical basis of adaptive responses to temperature stresses in
   fish.

   It is well accpted that the biochemical changes induced by temperature
   stress are attributed to modulations of gene expression [52][14],
   [53][20]. In the last decade, microarray techniques have fundamentally
   revolutionized the investigations of gene regulation in organisms
   exposed to environmental challenges. Researchers have characterized the
   transcriptional responses elicited by hypo- or hyperthermia stress in
   fishes including common carp (Cyprinus carpio) [54][21], zebrafish
   (Danio rerio) [55][22], [56][23], channel catfish (Ictalurus punctatus)
   [57][9], annual killifish (Austrofundulus limnaeus) [58][10], coral
   reef fish (Pomacentrus moluccensis) [59][24], [60][25], goby
   (Gillichthys mirabilis) [61][26], [62][27], and rainbow trout
   (Oncorhynchus mykiss) [63][28]. A large number of genes were found to
   be regulated by temperature stress. In species like rainbow trout,
   channel catfish and common carp, differently expressed genes after
   exposure to temperature stress constitute about 10 percent of the
   investigated genes and most of which are upregulated. The upregulated
   genes are known to participate in a wide range of biological processes,
   such as transcriptional regulation, signal transduction, cell growth
   and differentiation, protein synthesis, stress response and metabolism
   regulation. Therefore, the transition to cold- or heat-acclimated state
   involves extensive changes in gene expression.

   Although results from previous studies significantly advanced our
   understanding in the biochemical and molecular mechanisms underlying
   the adaptive response of fishes to temperature stress, several basic
   questions remain to be addressed. First, all of the previous studies
   have focused on the transcriptional regulation in adult tissues
   including brain, heart, liver, gill, kidney, skeletal muscle and
   intestine. Fish at larval stage fish are reported to be more
   susceptible to temperature variation [64][29], so it is not clear
   whether larva-specific responses to temperature stresses exist. Second,
   most previous studies were conducted with non-model species. The
   limited genetic resources severely restricted the number of genes for
   investigation and interpretation of the expression profiles.

   Zebrafish are now widely used as a model animal for a variety of
   biological disciplines. Substantial information and resources for
   genetics, developmental biology, biochemistry, physiology and
   temporal-spatial gene expression patterns make zebrafish an ideal model
   for environmental genomics researches. Zebrafish is a small eurythermal
   fish that naturally lives in shallow fresh-water habitats and can
   withstand a wide range of daily and seasonal temperature fluctuations
   [65][30], [66][31]. It is likely that an efficient mechanism has
   evolved in zebrafish to cope with rapid changes in body temperature. In
   this study, we aimed to identify variations in gene transcriptional
   expression of zebrafish larvae during their acclimation to temperature
   changes. The Agilent Zebrafish Oligo Microarray (V2) (4×44K) was
   utilized to investigate the transcriptional responses of zebrafish
   larvae to cold (16°C) or heat (34°C) stress. This microarray platform
   contains 43603 synthesized oligonucleotide probes corresponding to at
   least 21794 unigenes. The combination of a well-established model
   organism with a microarray platform for a large number of genes would
   provide more detailed and accurate insights into molecular mechanisms
   underlying the adaptive response of fish to thermal stress. In
   addition, a parallel comparison of stress responses to temperature
   changes in different directions may identify cold- and heat-specific
   genes and signaling pathways in fish.

Results

Experimental Design and Effects of Temperature Stress on the Development of
Zebrafish Larvae

   To characterize the transcriptional responses of genes regulated by
   cold or heat stress in larval zebrafish, larvae at 96 hpf (maintained
   at 28°C from fertilization, designated as 28°C-96 hpf) were exposed to
   16 (cold), 28 (control) or 34°C (heat) for 2 (from 96 to 98 hpf) and 48
   h (from 96 to 144 hpf), respectively. The expression of genes in
   samples of these six groups (designated as 16°C-2 h, 16°C-48 h, 28°C-98
   hpf, 28°C-144 hpf, 34°C-2 h and 34°C-48 h, respectively) was analyzed
   using Agilent microarray ([67]Figure 1). This experimental design was
   aimed to reveal both immediate and later transcriptional events
   regulated by temperature stress.

Figure 1. Experimental design.

   [68]Figure 1
   [69]Open in a new tab

   Zebrafish embryos were maintained at 28°C from fertilization to 96 hpf.
   Larvae at 96 hpf were exposed to 16, 28 or 34°C for 2 and 48 h,
   respectively. Samples exposed to different temperature were collected
   at 98 and 144 hpf and subjected to microarray analysis.

   After exposure to 16 or 34°C for 48 h (from 96 to 144 hpf), the
   morphology of treated larvae (16°C-48 h and 34°C-48 h) were compared
   with that of corresponding controls (28°C-144 hpf) to determine the
   effects of temperature stress on the development. As shown in
   [70]Figure 2A, the 16°C -48 h larvae displayed phenotypes more like
   those of 28°C-96 hpf larvae and demonstrated a larger yolk sac and
   smaller intestine lumen in comparison with those of control larvae.
   However, no obvious difference in morphology was observed between
   34°C-48 h and 28°C-144 hpf larvae ([71]Figure 2A). The standard length
   (SL) and eye diameter (ED) of larvae from different groups were
   measured and compared to represent the process of development. As shown
   in [72]Figure 2C, the SL of 16°C-48 h larvae is significantly shorter
   than those of 28°C-144 hpf and 34°C-48 h larvae. Similarly, the ED of
   16°C -48 h larvae is smaller than those of larvae in other two groups,
   but it is larger than that of 28°C -96 hpf larvae ([73]Figure 2D). Heat
   stress markedly inhibited the increase of ED, but this effect is lower
   than that of cold stress ([74]Figure 2D).

Figure 2. Effects of temperature stress on the development of zebrafish
larvae.

   [75]Figure 2
   [76]Open in a new tab

   (A) Representative images of zebrafish larvae before and after
   temperature stress. Zebrafish larvae at 96 hpf (maintained at 28°C from
   fertilization) were exposed to temperature stress for 48 h (cultured at
   16, 28 or 34°C from 96 to 144 hpf). Images were taken under a
   stereomicroscope from Zeiss with a color CCD camera. Red and yellow
   dashed line indicates intestine lumen and yolk sac, respectively. SL:
   standard length (distance from the snout to the posterior tip of the
   notochord). ED: eye diameter. Scale bar: 500 µm. (B) Alcian blue
   staining of the cartilages of jaw and branchial arches. DMC: distance
   from the inner border of Meckel’s cartilage to the anterior end of
   ceratohyal. CL: ceratohyal length. Scale bar: 200 µm. (C, D, E and F)
   Bar charts demonstrate the effect of temperature stress on the SL, ED,
   DMC and CL, respectively. Horizontal axis indicates treatments
   described in (A) and (B). Error bars indicate standard deviations
   (n = 40–60). Different letters above the error bars indicate
   statistically significant differences determined by one-way analysis of
   variance (ANOVA) followed by Duncan’s multiple range test (p<0.05).

   To address whether temperature stress influences the growth of body
   length and eye diameter but not the development of other organs, alcian
   blue staining was performed to determine the cartilage development of
   jaw and branchial arches ([77]Figure 2B) by measuring the distance from
   the inner border of Meckel’s cartilage to the anterior end of
   ceratohyal (DMC) and ceratohyal length (CL). These parameters have
   previously been used to evaluate the cartilage development of zebrafish
   larvae [78][32]. The results of statistical analysis indicate that the
   DMC and CL of 16°C-48 h larvae are significantly smaller than those of
   28°C-144 hpf larvae ([79]Figure 2E and F). Heat stress inhibited the
   increase of CL as well ([80]Figure 2F); however, the DMC and CL values
   of larvae exposed to both cold and heat are significantly larger than
   those of larvae at the beginning of exposure (28°C-96 hpf).

   Taken together, these data suggest that the development of zebrafish
   larvae was not stopped, but significantly delayed under cold stress.
   The inhibitory effect of heat stress on the development of zebrafish
   larvae is lower than that of cold stress.

Effects of Temperature Stress on the Body Composition

   The main body composition parameters of zebrafish larvae were analyzed
   to elucidate the effects of temperature stress on the nutrient
   consumption and energy conversion. As shown in [81]Table 1, the wet
   mass of zebrafish larvae was not significantly affected by temperature
   stress exposure for 48 h; however, the dry mass, protein, lipid and
   glycogen content of 16°C-48 h larvae were all higher than those of
   28°C-144 hpf larvae. It is interesting that the glycogen content of
   16°C-48 h samples was found to be higher than that of 28°C-96 hpf
   larvae. In comparison with 28°C-144 hpf samples, exposure to heat
   stress for 48 h significantly reduced the dry mass, protein and
   glycogen contents of zebrafish larvae. These results suggest that cold
   and heat stress exert opposite effects on the nutrient consumption in
   developing larvae, and this is consistent with the delayed development
   under cold stress.

Table 1. Body composition of zebrafish larvae before and after temperature
stress exposure.

   Components 28°C-96 hpf 16°C-48 h 28°C-144 hpf 34°C-48 h
   Wet mass (n = 5) 367.67±13.87^a 413.67±22.68^b 421.33±10.83^b
   398.33±22.58^b
   Dry mass (n = 5) 62.67±0.91^a 59.67±2.17^b 56.33±0.75^c 50.33±3.21^d
   Protein (n = 10) 55.37±4.74^a 55.72±3.13^a 49.98±3.96^b 45.07±3.82^c
   Lipid (n = 5) 10.40±0.80^a 10.12±0.40^a 7.82±0.62^b 7.87±0.37^b
   Glycogen (n = 5) 0.48±0.03^a 0.56±0.05^b 0.33±0.07^c 0.21±0.04^d
   [82]Open in a new tab

   Notes: 1) Zebrafish larvae at 96 hpf were exposed to 16, 28 or 34°C for
   48 h. Samples exposed to different treatments were designated as 28°C
   -96 hpf, 16°C -48 h, 28°C -144 hpf and 34°C -48 h, respectively. 2)
   Data (µg/individual) are given as means ± standard deviations. The
   number of samples is displayed in parenthesis. 3) The numbers of
   individuals used for the measurement of wet mass, dry mass, protein,
   lipid and glycogen are 80, 80, 1, 20 and 20 respectively. 4) Different
   letters in the same row indicate statistically significant differences
   determined by ANOVA followed by Duncan’s multiple range test or
   Dunnett’s T3 test (p<0.05).

Global Effects of Temperature Stress on Gene Expression

   To examine overall effects of cold or heat stress on gene
   transcription, signal intensity values of all probes with significant
   expression were subjected to principle component analysis (PCA). The
   results of PCA indicate that 72.5% of the transcriptional variations
   can be explained by the first three components. As shown in [83]Figure
   3, the projections of all samples in the principle component space
   clearly revealed an obvious consistency within the same group and a
   clear discrepancy among groups. The first principle component (PC1)
   mainly captured the difference between 98 and 144 hpf larvae maintained
   at 28°C ([84]Figure 3), indicating the expression variations of genes
   associated with the normal development process. A similar trend of
   variation in gene expression between larvae exposed to heat stress for
   2 and 48 h was also explained by PC1 ([85]Figure 3). PC2 mainly
   captured the difference in gene expression between cold-treated and
   control samples ([86]Figure 3). PC3 mainly explained the variation in
   gene expression between larvae maintained at 28°C and 34°C ([87]Figure
   3).

Figure 3. Principle component analysis (PCA) of gene expression profiles.

   [88]Figure 3
   [89]Open in a new tab

   Normalized signal intensity values of all the probes called “present”
   in at least 2/3 (12) arrays were subjected to PCA using ArrayTrack. The
   x-, y- and z-axes represent PC1, PC2 and PC3, respectively. The colors
   and shapes of data points indicate temperature treatment and time of
   exposure: blue for cold (16°C), red for heat (34°C) and green for
   control (28°C), sphere for 2 h and square for 48 h after exposure. One
   of two overlapped heat-treated samples at 48 h is shown as yellow and
   sphere. The sample names were displayed within the figure.

Validation of Microarray Data by Quantitative Real-time PCR (qPCR)

   To validate the expression profiles from microarray analysis, relative
   mRNA levels for 15 genes were measured by qPCR. To choose the most
   stable genes as internal references for qPCR data normalization, five