The adults, embryos and larvae of Pacific white shrimp (L. vannamei) were collected from the Dongfang shrimp culture center, Zhanjiang, Guangdong, China, and they were maintained in artificial seawater [30 parts per thousand (ppt) and pH 8.2] at 28°C under a 12-h dark:12-h light photoperiod. The shrimp were fed every morning, and feces and other detritus in each tank were siphoned out after feeding. Sexually immature shrimp with body weights of 7.2 ± 0.9 g and body lengths of 7.3 ± 0.6 cm were used for the molecular cloning and stress challenge experiments. Shrimp were anesthetized on ice and killed by decapitation. All animal experiments were conducted in accordance with the guidelines and approval of the Ethics Committees of the South China Sea Institute of Oceanology, Chinese Academy of Sciences.

Total RNA from the shrimp hepatopancreas was extracted using TRIzol reagent (Invitrogen), digested with DNase I (Invitrogen) and reversely transcribed into first-strand cDNA with PrimeScript^TM II Kit (TaKaRa). Based on a unigene of 1075 bp (Supplementary Data 1) that was obtained from an Illumina transcriptome of L. vannamei constructed by our lab previously (Huang et al., 2017) and shares high sequence homology with HSP40 in other species, gene-specific primers were designed to amplify a partial sequence of L. vannamei HSP40 (Lv-HSP40). The corresponding full-length sequence was obtained by 3′- and 5′-rapid amplification of cDNA ends (RACE).

Structural domains were predicted with the SMART and ScanProsite programs. Three-dimensional (3-D) model was generated by using the SWISS-MODEL server. Phylogenetic analysis for different crustacean HSPs was performed by the neighbor-joining method with 1000 bootstrap replicates using MEGA 6.0.

The tissue distribution of L. vannamei HSP40 mRNA was quantitatively detected in six individuals. The selected tissues included eyestalk, brain, thoracic nerve, abdominal nerve, gill, heart, hepatopancreas, hemolymph, muscle, leg, stomach and intestine from sexually immature shrimp, as well as ovary and testis, respectively, from female and male shrimp during gonadal development. The embryos and larvae were sampled in nine developmental stages according to their morphologies as observed using an optical microscope, including zygote, blastula, gastrula, limb bud embryo, larva in membrane, nauplius, zoea, mysis, and post-larvae. Embryos and larvae were collected when 80% of the population had reached the objective stage, and the morphology for each stage was determined as described previously (Wei et al., 2014). The tissue, embryonic and larval samples were frozen immediately in liquid nitrogen, then stored at -80°C prior to further analysis of mRNA expression.

The cellular localization of HSP40 mRNA were performed in selected tissues with high transcript levels by in situ hybridization (ISH). Briefly, samples from the muscle, gill, hepatopancreas and ovary were removed quickly and fixed overnight in 4% paraformaldehyde in PBS. Samples were dehydrated in an ethanol series, immersed in xylene, embedded in paraffin, cut into 5 μm sections and mounted onto slides. A digoxin (DIG)-labeled DNA probe of 585 bp (Supplementary Data 2) against Lv-HSP40 mRNA was synthesized with the DIG DNA Labeling Kit (Roche). The sections were deparaffinized, prehybridized and hybridized with the DIG-labeled DNA probe (0.4 ng/ml) at 42°C overnight as described previously (Chen et al., 2008). After that, the sections were incubated with horseradish peroxidase (HRP)-conjugated anti-DIG antibody (Roche), and the ISH signal were developed by a diaminobenzidine (DAB, MXB Biotechnologies) reaction and observed with a Leica DM-IRB light microscopy (Leica).

The changes in the mRNA expression of L. vannamei HSP40 and its cofactor HSP70 were measured in shrimp under temperature, acidity/alkalinity and salinity stresses. After acclimation in artificial seawater (30 ppt and pH 8.2) at 28°C for 2 weeks, fifty-five shrimp were randomly transferred into nine independent 20-L tanks (6 individuals per tank). The challenge conditions for shrimp included low temperature (13°C) and high temperature (37°C) [as validated by (Huang et al., 2017)], low pH (pH 6.8) and high pH (pH 8.9) [as validated by (Cai et al., 2017)], and low salinity (10 ppt) and high salinity (45 ppt) [as validated by (Chen et al., 2016b)]. Six shrimp from each group were killed at 6 h after challenge. The hepatopancreas, muscle and gill samples were collected for further analysis, as described above.

Previtellogenic female shrimp, with body weights of 48.4 ± 1.6 g, body lengths of 18.3 ± 0.3 cm and a gonadosomatic index (GSI) <0.3%, were applied for artificially induced maturation with unilateral eyestalk ablation and nutrition strength (Chen et al., 2014). Ovarian development was defined into four stages (stage I-IV) based on the classification of predominant oocytes as described previously (Tinikul et al., 2011; Chen H.-Y. et al., 2018). Briefly, the anterior end of ovarian explants with the number 2–9 lobes (Bae et al., 2017) were fixed overnight in 4% paraformaldehyde. The samples were dehydrated through an ethanol series, immersed in xylene and embedded in paraffin. The sections were cut into 5 μm sections, mounted onto slides and stained with hematoxylin and eosin (H/E). In stage I, the ovaries are mainly comprised of oogonia and previtellogenic oocytes; in stage II, the ovaries majorly compose previtellogenic oocytes and endogenous vitellogenetic oocytes; in stage III, the ovaries process a large number of oocytes developing from endogenous vitellogenetic oocytes to exogenous vitellogenetic oocytes; in stage IV, exogenous vitellogenetic oocytes and mature oocytes predominate in the ovaries. In this study, the GSIs for shrimp in stage I, II, III and IV were 0.3–0.9%, 0.9–2.5%, 2.5–5.0% and >5.0%, respectively. The hepatopancreatic and ovarian samples from six individuals of each stage were collected for further analysis of HSP40 and HSP70 mRNA expression, as described above.

The effects of HSP40 on ovarian development were investigated using an RNAi approach. Small interfering RNA (siRNA) targeting sites were identified using the Ambion siRNA target finder program, and the selected sequences were subjected to global BLAST against non-redundant nucleotide sequences, as described previously (Cai et al., 2017). Three sets of siRNA (HSP40-42, HSP40-474, and HSP40-825, Supplementary Data 3) specific against Lv-HSP40 and the non-targeting nematode siRNA (NC-siRNA) were synthesized by Sangon Biotech Company. To determine the knockdown efficiencies, hepatopancreatic and ovarian Lv-HSP40 transcripts were measured by real-time PCR at 12 and 24 h after siRNA injection. Given that injection of siRNA HSP40-474 with a dosage of 300 ng/g body weight (bwt) showed the highest mRNA knockdown efficiency, it was selected for the subsequent Lv-HSP40 gene silencing experiments.

Previtellogenic female shrimp with body sizes as described above were used for RNAi experiments. The injection of siRNA was conducted as described before (Cai et al., 2017). Briefly, one-hundred fifty shrimp were transferred randomly into three independent 1-m³ tanks (fifty individuals per tank). Each shrimp was injected with 100 μL of siRNA HSP40-474 (300 ng/g bwt) in PBS, and either injection of PBS alone or NC-siRNA (300 ng/g bwt) in PBS were used as control samples. The shrimp in each group were sampled at 1, 2, 3, 4, and 5 days after siRNA injection. The GSIs were weighted and calculated after sampling, and the ovarian developmental stages were confirmed by the observation of paraffin sections with H/E staining. In parallel experiments, the mRNA levels of the cofactor HSP70 and the vitellogenesis essential genes vitellogenin (Vg) in the hepatopancreas and ovaries, and vitellogenin receptor (VgR) in the ovaries from the shrimp at 24 h after siRNA injection were measured by real-time PCR.

The transcript levels of target genes (HSP40, HSP70, Vg, and VgR) were measured in multiple tissues, embryonic, larval developmental stages, and ovarian developmental stages, as well as under different challenge/treatment conditions, by quantitative real-time PCR. Briefly, 2 μg total RNA was extracted using TRIzol reagent, digested with DNase I and reversely transcribed into first-strand cDNA with PrimeScript^TM II Kit. The cDNA samples obtained were then subjected to quantitative PCR performed on a Roche Light-Cycler 480 system (Corbett Research) by using SYBR Premix Ex Taq^TM II (Takara) according to the manufacturer’s protocol (Supplementary Data 4). In this study, β-actin was used as an internal control. Serially diluted plasmid DNAs containing the HSP40, HSP70, Vg, VgR (Supplementary Data 5) and β-actin sequences were used as the standards for real-time PCR. After the PCR reactions, the identities of PCR products were routinely confirmed by melting curve analysis. The raw data were expressed in terms of fmol target transcript detected per tube, and the data were routinely normalized as a ratio of β-actin mRNA detected in the same sample.

For mRNA expression, the raw data were simply transformed as a percentage of the mean values in the control group. The statistical analysis were performed with SPSS (IBM Software). For mRNA and GSI measurement, the data expressed as the mean ± S.E. (standard error) were analyzed by using Student’s t-test or one-way ANOVA followed by Fisher’s least significant difference (LSD) test. For staging of ovarian development, the data were examined by using Ridit analysis. Differences were considered significant at P < 0.05.

In this study, the full-length cDNA for a heat shock protein 40 (Lv-HSP40) was identified from the Pacific white shrimp. As shown in Figure 1, the Lv-HSP40 cDNA (GenBank: MH932106) is 1756 bp in size, with a 173-bp 5′-untranslated region (UTR), a 391-bp 3′-UTR region and a 1191-bp open reading frame (ORF) coding for a 396-a.a. protein precursor (∼44.4 kDa). In the 3′-UTR, a polyadenylation signal (attaaa) is located 15-bp upstream of the poly-A tail.

By prediction using the SMART and ScanProsite programs, a DnaJ domain, a glycine/phenylalanine-rich (G/F) domain, a Zn²⁺-binding CR domain and a CT domain were found in the deduced Lv-HSP40 protein sequence (Figures 1, 2A). The J domain is 56-a.a. in length and is where a highly conserved HPD tripeptide is located. Adjacent to the J-domain is a 24-a.a. G/F domain with a conserved DIF sequence. A 106-a.a. CR domain, composed of three repeats of the CXXCXGXG motif, is next to the G/F domain. In the C- terminus, a 106-a.a. CT domain and a CaaX motif are present. In addition, a 3-D model of the Lv-HSP40 dimer showed that the J domains were located in the exterior part, and the CT domains were located in the inner part (Figure 2B). Furthermore, a phylogenetic tree built with several HSPs in crustacean species showed that HSP70, HSP40, HSP10, HSP21, and HSP90 were grouped into respective branches, and our newly cloned Lv-HSP40 was clustered with the HSP40 from lobster Metanephrops japonicas (Figure 3).

The transcript expression profiles of Lv-HSP40 were detected in multiple tissues by quantitative PCR. As shown in Figure 4A, Lv-HSP40 mRNA could be ubiquitously detected in all tissues we examined. The highest expression level of Lv-HSP40 mRNA was observed in the ovaries of female shrimp, and relatively high expression levels were detected in the muscle, gill, and hepatopancreas. In addition, an ISH for Lv-HSP40 mRNA expressed cells were performed in the tissues with high transcript levels (Figure 4B). In this case, Lv-HSP40 mRNA were predominantly detected in the myocytes, epithelial cells and hepatopancreatic cells in the muscle, gill and hepatopancreas, respectively. In the ovary, the Lv-HSP40 mRNA were mainly presented in the follicle cells with only weak signals in the oocytes.

During embryonic and larval development, the expression of Lv-HSP40 mRNA showed high levels at the stages of zygote, blastula and gastrula, then reduced sharply at the stage of limb bud embryo and remained at relatively low levels at the stages of larva in membrane, nauplius, zoea, mysis and post-larvae (Figure 4C). During early embryonic development, the Lv-HSP40 transcript levels were much higher than the estimated average level (mean value of Lv-HSP40 mRNA among different tissues) in adults, while during late embryonic development and larval development, the Lv-HSP40 transcript levels were similar to the estimated average level in adults (Figure 4C).

Given that Lv-HSP40 showed high transcript levels in the hepatopancreas, muscle and gill of shrimp, the responses of Lv-HSP40 mRNA under temperature, acidity/alkalinity and salinity stresses were detected in these tissues. The expression levels of Lv-HSP40 mRNA in the hepatopancreas, muscle and gill significantly increased when shrimp were transferred to a high temperature (37°C, Figure 5A) but showed no change when they were transferred to a low temperature (13°C, Figure 5A). Under low pH challenge (pH 6.8, Figure 5B), Lv-HSP40 mRNA increased in the hepatopancreas and gill but remained stable in the muscle. In contrast, Lv-HSP40 mRNA did not respond to high pH stress (pH 8.9, Figure 5B) in any tissues we tested. Additionally, Lv-HSP40 mRNA levels in all tissues examined neither changed under high salinity (45 ppt, Figure 5C) nor low salinity stress (10 ppt, Figure 5C). In parallel experiments, mRNA levels of Lv-HSP70 were upregulated in the hepatopancreas, muscle and gill under high temperature (Figure 5D) and low pH (Figure 5E) challenges, in the gill under high pH challenge (Figure 5E), and in the muscle and gill under low temperature challenge (Figure 5F). In contrast, Lv-HSP70 mRNA expression was stable in other stress groups we tested (Figures 5D–F).

The expression levels of Lv-HSP40 mRNA were detected in the hepatopancreas and ovary from female shrimp in different ovarian developmental stages (stages I–IV). In the hepatopancreas, Lv-HSP40 mRNA was stable at stages I and II, slightly increased in stage III and reduced to the normal level at stage IV (Figure 6A). In the ovary, Lv-HSP40 mRNA increased acutely and continuously from stage I to stage IV (Figure 6A). In contrast, the Lv-HSP70 (Figure 6B) transcript levels in the hepatopancreas and ovary remained stable throughout gonadal development in female shrimp. In this case, the morphology, anatomy and histology of L. vannamei ovaries for staging were shown in Figure 6C.

Among the three siRNA we tested, siRNA HSP40-474 showed the highest knockdown efficiency on Lv-HSP40 mRNA levels in both the hepatopancreas (Figure 7A) and ovary (Figure 7B), and no significant mortality was observed with siRNA injection. After injection of siRNA HSP40-474, the ovarian developments of female shrimp were delayed when compared to that in the PBS and NC-siRNA injection groups, based on the observation of the ovarian stages (Figure 7C) with significant differences (P < 0.05) by Ridit analysis were shown at 2 and 3 day after injection. Spawning in the HSP40 siRNA injection group was also delayed but not totally abolished (Figure 7C). Similarly, the increasing of GSI induced by eyestalk ablation in the HSP40 siRNA injection group was also slower than that in the PBS and NC-siRNA injection groups (Figure 7D). In parallel experiments, the injection of HSP40 siRNA for 24 h did not alter the HSP70 (Figure 7E), Vg (Figure 7F) and VgR (Figure 7G) mRNA levels in either the hepatopancreas or ovary.