SBDS^R126T rescued the sbds^−/− zebrafish

We recently reported that sbds-null fish died between 15–21 dpf. We generated a zebrafish transgenic line expressing the zebrafish WT sbds under the zebrafish ubiquitin (ubi) promoter: Tg(ubi:sbds:pA), which drives constitutive transgene expression during all developmental stages and adult organs (Mosimann et al, 2011). This transgenic line in the sbds KO background was able to rescue their phenotype and was viable (Oyarbide et al, 2020). To determine if expression of SBDS^R126T could prolong survival and modify aberrant development, we created a transgenic strain Tg(ubi:SBDS^R126T:pA) and bred that against the null background (also denoted Tg(SBDS^R126T) here). We detected full-length SBDS protein (Fig 1A and B). Unexpectedly, the transgenic line Tg(ubi:SBDS^R126T:pA) in the background of the sbds KO can live for >18 mo (adulthood). We used cmlc2:EGFP as a transgenesis marker in the vector backbone, which drives cytoplasmic EGFP specifically in the heart and facilitates the screening for the presence of the transgene (Kwan et al, 2007).

To determine the number of copies of the transgene inserted in the genome, we crossed two siblings with green hearts and selected four females with green hearts from the descendants. We then outcrossed them with non-green heart fish and counted the number of green heart fish versus non-green heart fish. Two females showed a 100% of the descendants with green heart, whereas the other two had ∼50% with green heart and 50% with non-green heart. We concluded that there was only one insertion of the transgene in the new transgenic line created (Fig 1C).

Next, we determined the human SBDS^R126T mRNA levels in 2 dpf larvae with 2x and 1x copies of the transgene. We incrossed 2x Tg(ubi:SBDS^R126T:pA);sbds^+/+ and collect larvae at 2 dpf. In parallel, we outcrossed 1x Tg(ubi:SBDS^R126T:pA);sbds^+/+ with a sbds^+/+ and screened fish for green heart and non-green heart at 2 dpf. Next, we calculated the expression of SBDS in these fish and we observed a significant decrease of approximately half of the levels in the 1x comparing with 2x copies. As expected, those without a green heart showed no expression of human SBDS (Fig 1D). We also determined the protein levels in adult fish fins and observed an increase in protein levels in the 2x comparing with the 1x copy of the transgene (Fig 1E).

We then studied the phenotype of the adult fish at 1 yr. We crossed a sbds^+/− with a Tg(ubi:SBDS^R126T:pA);sbds^+/− (Fig 2A). This cross generated siblings from the same clutch without green heart and with one copy of the transgene. Because the sbds^−/− died between 10–21 dpf, we analyzed the other five phenotypes produced from this cross at 1 yr. We did not see differences in standard length (SL) or in the histologic sections including liver, pancreas, digestive tract or kidney (Fig 2B and C). Surprisingly, we observed a 7:1 male: female ratio in the Tg(ubi:SBDS^R126T:pA) with the sbds-null background, and reduced fecundity. Flow cytometric profile of hematopoietic cells prepared from 1-yr-old kidney marrow of male siblings showed no difference in any of the blood cell populations (erythrocytes, lymphocytes, myeloid, and precursors) of the five different genotypes analyzed (Fig 2D).

Because 80S ribosome formation is affected in SDS patients and animal models including zebrafish (Finch et al, 2011; Tourlakis et al, 2012; Oyarbide et al, 2020), we performed polysome profiling on the livers of 1-yr-old male fish. Compared with what we had observed in the sbds-null fish, the expression of one transgenic copy of the SBDS^R126T allele rescued the polysome profile to mostly normal (Fig 2E).

We previously reported that sbds^−/− had a decrease in Rpl5 and Rpl11 protein levels and an accumulation of Eif6 protein levels (Oyarbide et al, 2020). We evaluated the levels of these proteins by Western blotting of fin lysates from the transgenic adult fish (1 yr-old). The ribosomal protein levels were similar to those found in WT siblings. However, Eif6 protein levels were significantly increased in the Tg(ubi:SBDS^R126T:pA) sbds^−/− (Fig 2F and G), as we previously reported in sbds-null larvae at 10 dpf (Oyarbide et al, 2020). We also analyzed the mRNA levels of tp53 and cdkn1a in liver, intestine, and brain; cdkn1a was markedly up-regulated in all three organs (Fig 2H).

Because we observed accumulation of Eif6 and activation of Tp53-pathway in the 1-yr-old transgenic fish, we evaluated these pathways at earlier stages of development. We crossed a sbds^+/− with Tg(ubi:SBDS^R126T:pA) sbds^+/− (Fig 2A) and analyzed the larvae at 10 dpf. At this stage, the sbds^−/− were alive and were included in the analysis. Western blotting showed a statistically significant increase in Eif6 protein in the sbds-null fish as previously observed and a nonsignificant increase in the transgenic line with the sbds-null background. Rpl5 and Rpl11 decreased only in sbds^−/− in comparison with the WT siblings (Fig 3A and B). Sudan black staining showed a decrease in the number of neutrophils in sbds KO background fish, independently of the presence of SBDS^R126T (Fig 3C).

To understand how the SBDS^R126T allele affected the zebrafish larval development, we performed qRT-PCR to determine changes in gene expression in ribosomes (rpl11 and rps19), metabolism (fasn, pparg, srebp1, and pfkmb), and the tp53 pathway (tp53, cdkn1a, mdm2, and puma) (Fig 3D). As expected, sbds was down-regulated in sbds^−/− and Tg(ubi:SBDS^R126T:pA) sbds^−/−. Surprisingly, eif6 mRNA was up-regulated only in the Tg(ubi:SBDS^R126T:pA) sbds^−/−, whereas rpl11 and rpl5 transcripts were not changed. We did not observe changes in the metabolism markers tested. We did find Tp53 pathway activation in the sbds KO through the up-regulation of cdkn1a, mdm2, and puma. These markers were not affected in the transgenic line of neither sbds^+/+ nor sbds^−/− backgrounds (Fig 3D).

Levels of SBDS^R126T affected embryonic development in the sbds^−/− fish

Hypothesizing that levels of SBDS^R126T might modify the embryonic development and severity of SDS phenotype, we incrossed Tg(ubi:SBDS^R126T:pA);sbds^+/+ with two copies of the transgene so that all the descendants were Tg(ubi:SBDS^R126T:pA);sbds^+/+ (Fig 4A). In parallel, we incrossed Tg(ubi:SBDS^R126T:pA);sbds^−/−, to have all the descendants with the same genotype Tg(ubi:SBDS^R126T:pA);sbds^−/− (maternal zygotic mutants [MZ]) (Fig 4B). Surprisingly, after 1 dpf we observed developmental delay in some embryos and deformed embryos in the sbds^−/− background (Fig 4A and B). To characterize the genetic and phenotypic differences, we collected embryos at either 1 dpf for gene expression analysis or embryos that developed normal at 10 dpf to measure neutrophil numbers. As expected, we found a significant down-regulation of sbds in the sbds^−/− (Fig 4C). However, eif6 and all ribosomal proteins tested (rpl11, rpl5a, rps19, and rpl13a) were significantly down-regulated only in the deformed embryos. We previously showed that mRNA levels of one important enzyme in the pathway of glycolysis, phosphofructokinase (pfkmb), was down-regulated in the sbds^−/− fish (Oyarbide et al, 2020). We evaluated the pfkmb mRNA levels in the Tg(SBDS^R126T)sbds^−/−, which were decreased in the normal and deformed larvae comparing with Tg(SBDS^R126T)sbds^+/+. We did not see any change in the lipid metabolism markers. Next, we analyzed Tp53 pathway: cdkn1a and puma were up-regulated in all the sbds-null backgrounds and mdm2 in the deformed ones (Fig 4C). We also checked markers for stress response, where we found a significant up-regulation of chop and casp9 in the sbds KO background (Fig 4C). We then checked Eif6 protein levels at 2 dpf in normal and deformed fish with the sbds-null background and compared the levels with the WT sbds background. Eif6 levels were increased only in the deformed embryos (Fig 4D and E).

Almost all of the normal-appearing Tg(ubi:SBDS^R126T:pA);sbds^−/− and the WT fish survived until 10 dpf, (respectively, 80% and 76%). However, Tg(ubi:SBDS^R126T:pA);sbds^−/− were significantly smaller and had significantly lower number of neutrophils comparing with the WT group (Fig 4F and G). All developed into males.

Next, we determined the effects of one copy of the transgene SBDS^R126T (Fig 5A). Interestingly, embryonic development was defective in 25% of the MZ sbds^−/− embryos after 1 dpf, the mortality was 100% after 3 dpf (Fig 5B and C), and presence of transgene did not ameliorate the defects (Fig 5B). We collected samples at 2 dpf (i.e., at the appearance of the green heart), and determined changes in gene expression. Surprisingly, cdkn1a was only up-regulated in sbds-null embryos without the transgene. However, pfkmb, fasn, and pparg were down-regulated in sbds-null embryos with and without SBDS^R126T (Fig 5D).

Previously, we reported that sbds null fish had fewer neutrophils than WT siblings at 5 dpf (Oyarbide et al, 2020). To determine whether the SBDS^R126T rescued the neutropenia in the sbds-null background, we incrossed sbds heterozygotes with two copies of the transgene and determined the number of neutrophils at 5 dpf (Fig 6A). As expected, we did not detect differences in neutrophil counts between any of the genotypes (sbds^+/+, sbds^+/−, and sbds^−/−) in the context of the transgenic SBDS R126T (Fig 6B). With these results, we can conclude that the SBDS^R126T dose is important in the neutrophil number in our zebrafish SDS models.

Tp53-loss does not rescue neutropenia or survival of sbds mutants

The p53 tumor suppressor pathway is activated by impairment of ribosome biogenesis and aberrant protein translation (Narla & Ebert, 2010; Bursac et al, 2014). To explore the role of Tp53 in our models, we outcrossed the sbds^−/− with the tp53^M214K zebrafish mutant (Berghmans et al, 2005). We created sbds^+/−;tp53^M214K/M214K zebrafish and incrossed them (Fig 7A). The homozygous tp53^M214K/M214K background did not rescue neutropenia in sbds^−/− fish at 10 dpf (Fig 7B). Because our previous results showed a significant increase in cdkn1a levels, we next determined whether this was dependent on Tp53 activity. tp53 and cdkn1a levels in the sbds^−/− mutants with tp53^M214K/M214K background were not significantly different from their WT siblings (sbds^+/+; tp53^M214K/M214K) (Fig 7C).

To further investigate the role of Tp53, we outcrossed our transgenic line with the tp53^M214K mutant (Fig 7D). When we incrossed the 1x Tg(SBDS^R126T);sbds^−/−;tp53^M214K/M214K we saw similar results in development and survival (Fig 7E) as previously observed in the tp53 WT background (Fig 5B). We found that eif6 and the ribosomal protein mRNA levels were not affected in any of the genotypes (Fig 7F). We also observed a significant decrease of cdkn1a and mdm2 levels in all tp53^M214K/M214K backgrounds compared with the WT (sbds^+/+;tp53^+/+). Interestingly, all lipid metabolism markers analyzed were significantly decreased as previously seen in the tp53 WT background (Figs 5D and 7F).