Normal development
The diameter of SC and DV eggs was similar: 0.58 ± 0.13 cm and 0.50 ± 0.15 cm, respectively (t-test df = 98 p = 0.889). In SC group, the hatching occurred at 335370 (50% achievement ~ 350) day*degree post fertilization, whereas in DV group it occurred later, at 360400 (~ 380) day*degree post fertilization. The size and weight characteristics of the newly hatched SC and DV were commensurable, FL = 1.61.9 cm and W = 6580 mg. The transition to larva stage proceeded in SC at 105165 (50% ~150) day*degree post hatching (= ddph), in DV at 130190 (~ 170) ddph. The onset of the fry stage occurred in SC at 450530 (~ 510) ddph, and 490570 (~ 540) ddph in DV (Fig. 1, upper inset). For the description of the postnatal developmental stages see Methods section (also Supplementary Fig. S1).
The significant differences in FL and W between the experimental SC and DV were found at 500 ddph (Fig. 1a and b). These discrepancies increased 215 dd later at the end of the experiment, and FL / W of SC fry were larger about 15% / 40%, respectively, than those characteristics of DV fry. In nature the differences in FL and W were found as well. The wild SC parrs (~ 850 ddph) were significantly larger than the same-aged DV parrs (Table S2).
The total whole-body triiodothyronine (T3) content was similar in the newly hatched SC and DV. Then, prior to the onset of external feeding, both groups displayed a synchronous drop of the hormonal level. In SC, the decrease was more pronounced (H-W test H5;131 = 21.27 p = 0.0160 for SC and H5;186 = 18.32 p = 0.0242 for DV). As the result, at 160 ddph SC and DV prelarvae significantly differed in T3 content (Fig. 2). In further development, until the transition to the fry stage, SC and DV demonstrated similar positive dynamics of T3. At the fry stage, the hormonal level within the groups did not display any serious fluctuations, but differences between the groups increased. At the end of the experiment, SC fry had significantly higher level of T3. In nature SC parrs demonstrated the significantly higher values of T3 content relative to DV parrs. The T3 content values in both SC and DV wild parrs were 1.31.4 times higher than those detected for the experimental fry (U test df = 39 p = 0.0052 for SC and df = 59 p = 0.0094 for DV) (Fig. 2).
The ontogenetic dynamics of the whole-body thyroxin (T4) content was similar in the experimental SC and DV. Both morphs displayed a gradual increase of T4 level from the hatching to the fry stage (H3;17 = 6.58 p = 0.0310 for SC and H4;20 = 8.11 p = 0.0397 for DV). Nevertheless, we found a significantly lower hormone level in SC larvae and fry compared with the same-aged DV (Fig. S2).
We did not reveal any significant fluctuations of the biochemical parameters within the groups of experimental fry (H tests p > 0.05 for all parameters). However, the stable differences between the artificially reared SC and DV fry, as well as between the wild SC and DV parrs were noted (Fig. 3). In the wild and experimental DV, the blood glucose level was significantly higher than in the wild and experimental SC. The phospholipid content was significantly higher in the wild DV than in the wild SC. The same but insignificant differences in the phospholipid content were detected for the experimental DV and SC. In contrast, the tissue antioxidant activity in the experimental DV was significantly lower than in SC.
We failed to find distinguishable discrepancies in the head and body shape between the experimental SC and DV at the early life stages: in prelarvae Procrustes ANOVA F1;49 = 1.26 p = 0.1756, SSind/res = 0.023/0.068; in larvae F1;49 = 1.13 p = 0.2005, SSind/res = 0.030/0.077, and in late larvae F1;44 = 3.66 p = 0.06122, SSind/res = 0.019/0.054. The first significant differences occurred at the fry stage, after 500 ddph (F1;49 = 5.24 p = 0.0322, SSind/res = 0.024/0.080). At the end of the experiment these discrepancies became more pronounced (F1;49 = 12.08 p = 0.0010, SSind/res = 0.029/0.071).
The plot of principal component scores depicted the ontogenetic channels of SC and DV, and the timing once morphological differences appeared (Fig. 4a). At the early stages (prelarva, larva and late larva) DV and SC had partially overlapping channels. During the transition to the fry stage, the channels of both morphs changed the direction and, at the fry stage (> 650 ddph), became separate. According to the landmark loadings on PC2 (Table S1), the most expressed differences between SC and DV fry occurred in the eye diameter, maxilla length, head height, and ventral fins position. These morphological discrepancies were reflected in the ‘consensus’ body shapes stretching along the canonical root (Fig. 4b and Table S1), which clearly demonstrated the head enlargement and the caudal shift of the fins in SC fry relative to DV fry. Similar morphometric differences were revealed for the wild parrs of SC and DV (Fig. 4c). The additional data concerning the main differences in the head proportions and fin position in experimental fry and wild parrs are present in Table S2.
SC and DV differed in the ossification rate of skull bones. The MANCOVA showed that, despite the strong effect of age (p ≤ 0.0321), a ‘group’ predictor was significant in determining the ossification rate of supraethmoid, frontal bone, maxilla, vomer and preopercle (Table). Two ‘modules’ of cranial ossifications - neurocranium and teeth-armed bones - displayed significant differences in the developmental rates between SC and DV. Thus, the initial advancement in ossification of the neurocranium in SC prelarvae was replaced by the retardation during the subsequent developmental stages (Fig. 5a). As the result, DV demonstrated a significantly more ossified neurocranium starting from the larva stage. In contrast, the teeth-armed bones displayed an accelerated ossification in SC from the early stages (Fig. 5b). The ossification rate of preopercle resembled the ossification pattern of the neurocranium bones being advanced in DV. The intraoral tongue-bite apparatus did not demonstrate a significant difference in ossification rate. We discovered a minor advancement in the ossification of these bones in the SC late fry (Fig. 5c). The comparative analysis of postcranial skeleton revealed differences in counts of ossified vertebrae: 61.3 ± 0.4 (from 59 to 63) in SC and 62.4 ± 0.3 (6064) in DV. Other structures did not display any discrepancies.
Table. Results of MANCOVA for 13 skeletal elements developmental rate in experimental SC and DV
Skeletal element | Group (SC vs DV) | Age (ddph) | Group * age |
F12;170 | p | F96;1155 | p | F12;170 | p |
Supraethmoid | 4.51 | 0.0442 | 9.82 | 0.0001 | 14.322 | < 0.0001 |
Frontal bone | 8.92 | 0.0007 | 8.88 | 0.0001 | 20.48 | < 0.0001 |
Orbital series | 1.98 | 0.4740 | 17.13 | < 0.0001 | 11.14 | < 0.0001 |
Parasphenoid | 1.41 | 0.7191 | 13.84 | < 0.0001 | 9.52 | 0.0001 |
Premaxilla | 4.03 | 0.0622 | 16.6 | < 0.0001 | 21.64 | < 0.0001 |
Maxilla | 4.53 | 0.0452 | 7.63 | 0.0002 | 6.63 | 0.0003 |
Dentary | 2.71 | 0.1551 | 14.62 | < 0.0001 | 10.62 | 0.0001 |
Vomer | 7.94 | 0.0014 | 11.38 | < 0.0001 | 4.62 | < 0.0001 |
Palatine bone | 3.81 | 0.0712 | 12.72 | < 0.0001 | 9.63 | 0.0001 |
Lingual bone | 1.2 | 0.9710 | 7.79 | 0.0002 | 6.04 | 0.0121 |
Gill arch elements | 1.21 | 0.9520 | 4.69 | 0.0321 | 5.12 | 0.0222 |
Gill rakers | 1.19 | 0.9754 | 7.73 | 0.0002 | 7.72 | 0.0002 |
Preopercle | 5.12 | 0.0345 | 12.17 | < 0.0001 | 13.51 | < 0.0001 |
Development Under Altered Thyroid Status
The goitrogen treatment and administration of T3 during the larva-fry transition (from 160 to 515 ddph) changed the thyroid status of experimental fish (Fig. 2 and S2, in boxes). In SC and DV reared in the 0.5 g l− 1 solution of thiourea, three-times decline of T4 content was detected. The hypothyroid SC also demonstrated a significant decrease in T3 content (H2;33 = 19.39 p = 0.0002), but the hypothyroid DV did not (H2;38 = 4.36 p = 0.1127). T3 administration did not affect T4 content (H2;14 = 1.45 p = 0.4829), but significantly increased T3 level (H2;28 = 7.78 p = 0.0211) in DV.
In both SC and DV reared under the 0.2 g l− 1 and 0.5 g l− 1 solutions of thiourea, the mortality rate was 10%, and 20%, respectively. The mortality rate in DV reared under T3 treatment was higher: 25% in the 1.0 µg l− 1 solution, and 55% in the 5.0 µg l− 1 solution. In the control groups of DV and SC, the mortality rate was drastically lower, 1% of individuals only.
The goitrogen-caused decrease in the biochemical parameters was statistically insignificant in both groups (K-W test p > 0.50), but, in fact, more pronounced in SC than in DV: H2;58 = 2.85 / H2;81 = 0.37 for the blood glucose, = 2.62 / 0.33 for the tissue antioxidant activity, = 0.19 / 0.05 for the phospholipid content, respectively. T3 administration provoked two times stronger relative to the thiourea treating but still insignificant physiological response in DV (Fig. 3, in boxes).
The goitrogen negatively affected the fish growth rate (Fig. 1c and d). The significant differences in size were observed between normal and thiourea-treated SC (Table S3; ANOVA for length F2;110 = 19.04 p ≤ 0.0250). In DV, the reaction to the thiourea treatment was less pronounced; the significant size differences were revealed between the control fish and the fish reared under the 0.5 g l− 1 solution only (F2;143 = 4.86 p ≤ 0.0011). In contrast to thyroxin, T3 administration did not affect growth rate (Fig. 1c and d; Table S3).
The alterations of thyroid status resulted in the developmental abnormalities. In the majority of hypothyroid SC late larvae and early fry (90%), the body walls did not close on the belly around the remains of the yolk sac. However, we failed to find similar defects in DV late larvae and fry. The goitrogen affected development of the neural arches associated with the preural-1 vertebra. The majority of the normally developing SC (77%) possessed the neural arches, whereas in the hypothyroid groups the frequency of their occurrence was 30% and 0% in 0.2 and 0.5 g l− 1 of thiourea, respectively. In the control DV, the frequency of fish with the neural arches was 27%. In the hypothyroid DV reared under 0.2 and 0.5 g l− 1 of thiourea, it was 25% and 13%, respectively. The hyperthyroid DV group differed in excessive amount of the scoliotic individuals; the frequency of this abnormality reached 40% under 1.0 µg l− 1 concentration and then 90% under 5.0 µg l− 1 concentration.
The changes in the skull ossification advancement in fish with altered TH status had a doze-dependent and group-specific character (Fig. 5 ac, in boxes). In the thiourea-treated SC, the postponed ossification of teeth-armed bones was observed; the relative sum ossification ranks differed significantly between three SC groups, i.e. control − 0.2 g l− 1 − 0.5 g l− 1 of thiourea (H2;49 = 23.69 p = 0.0058). Within this ‘module’, premaxilla demonstrated the most evident developmental delay. The intraoral tongue-bite apparatus and the skull dermal bones were less affected by the thiourea (H2;49 = 7.12 and 6.79 respectively), only the SC reared under the 0.5 g l− 1 concentration differed from the control group (p = 0.0231 and 0.0322). In the hypothyroid DV, the dermal bones ossification displayed a significant retardation (H2;66 = 12.28 p = 0.0185) in comparison with the control group. The teeth-armed and the intraoral bones displayed similar but less pronounced reaction (H2;66 = 11.28 and 9.89 respectively). A significant difference (p ≤ 0.0230) was detected only for the teeth-armed bones ossification ranks between the control and the 0.5 g l− 1 thiourea group. In the hyperthyroid DV, all bone ‘modules’ displayed accelerated ossification, and only the teeth-armed bones developmental advancement was significant (H2;58 = 9.55 p = 0.0301). The skull dermal bones and the intraoral bones displayed relatively weak reaction (H2;58 = 2.65 and 1.18 respectively). The supraethmoid ossification in the hyperthyroid DV did not exceed that in the control group.
The geometric morphometrics revealed that the hypo- and hyperthyroid DV displayed different vectors of morphological changes and were aligned along the first canonical root in CV morphospace in the following order: 0.5 g l− 1 of thiourea – 0.2 g l− 1 of thiourea – control – 1.0 µg l− 1 of T3 – 5.0 µg l− 1 of T3 (scoliotic individuals were excluded from the analysis). The vector of morphological changes in the thiourea-treated SC coincided with that in the thiourea-treated DV (Fig. 6a). The appropriate loading on landmarks (Table S1) and stretching of the ‘consensus’ body shape along the first root (Fig. 6b) manifested the head shortening and the anterior shift of fins in the hypothyroid fish, and precisely the opposite morphometric changes in the hyperthyroid fish. As the result of CV scaling, the hyperthyroid DV (5.0 µg l− 1 of T3) significantly overlapped with the control SC, whereas hypothyroid SC (0.2 g l− 1 of thiourea) overlapped with the control DV. The Procrustes distances between the groups and significance level of differences from the Procrustes ANOVA comparison are presented in Table S4.
SC demonstrated additional morphometric shift along the second canonical root (Fig. 6a). The appropriate loading on the landmarks (Table S1) and the stretching of the ‘consensus’ body shape along the second root (Fig. 6c) displayed the reduction of the eye diameter and the maxilla length, elongation of the mandible, and the caudal shift of the ventral fins in the hypothyroid SC relative to the control one.
Coniferous Infusion Impact
The exposure of fry to the coniferous infusion resulted in the acute toxicosis and the increase of mortality. Approximately 11% (from 7 to 14% in different tanks) of DV fry died in the 1.0% solution, and 18 (323)% of individuals died in the 1.5% solution. The SC mortality rate was 3 (04)% in the 1.0% solution as well as in the 1.5% solution. The 100% mortality of SC and DV fry was observed in the 3.0% solution in four days. The dead fry were characterized by typical pale anemic gills. In the control groups the mortality rate was 0%.
The infusion administration induced group-specific physiological responses (Fig. 2, in boxes). It provoked a significant surge of the blood glucose level in DV (H3;158 = 42.33 p = 0.0003), but not in SC (H3;115 = 16.96 p = 0.0530). DV also demonstrated a significant increase in the tissue antioxidant activity (H2;67 = 8.38 p = 0.0263), while SC did not (H2;61 = 3.18 p = 0.0771). The decrease in muscle phospholipid content reflecting the level of biomembrane oxi(peroxi)dation was more pronounced in DV (H3;97 = 7.97 p = 0.0510) than in SC (H3;87 = 7.21 p = 0.0653). In both groups, we failed to find differences (U test p > 0.05) in the biochemical parameters between the control groups at the onset and at the end of the experiment.
The goitrogen and T3 treatments altered DV response to the impact of 1.5% coniferous infusion. In the euthyroid fry, the mean mortality rate was 23% in the infusion and 0% in the pure water. In the survived euthyroid fish, an increase of T3 content (+ 1.91 ng g− 1) was detected, but intra-sample variance remained stable. In the hyperthyroid fry, the mean mortality rate was 10% only. The survived individuals demonstrated a two-times increase of T3 content accompanied by a splash in the intra-sample variance. In the hypothyroid fry, the mean mortality rate was 54%. They also manifested elevated T3 content (+ 0.34 ng g− 1) and 1.5-fold drop in the variance. The alterations of T4 content were also detected (Fig. S2).