Procedures were approved by the University of Vermont Institutional Animal Care and Use Committee Protocol Number: 14-053 and the University of Vermont Institutional Biosafety Committee Protocol Number: 14-024. Embryos were raised under standard conditions and staged as previously described (Kimmel et al., 1995; Westerfield, 2000). Strains used include: TL; Tg(Rx3:GFP) to label retinal progenitor cells (Rembold et al., 2006); Tg(pou4f3:GFP) to label sensory hair cells (Xiao et al., 2005); Tg(ngn1:GFP) to label sensory neurons (Blader et al., 2003); and Tg(mnx1:mCherry) to label motor neurons (provided by Christine Beattie, Ohio State University). Fertilized embryos were raised at 28.5 or 25°C and staged as previously described (Kimmel et al., 1995). In some cases, phenylthiourea was added to the embryo media at a concentration of 0.003% at 24 h post fertilization (hpf) in order to inhibit pigment formation.

Injections were performed at the 1-2 cell stage using an Eppendorf Femtojet 4i microinjector. A translation blocking hars morpholino (ATGGTGCTCCAGAAACACAGCCGAT), p53 morpholino (Robu et al., 2007), and GeneTools Standard Control Oligo (GeneTools, Philomath, OR, United States) were injected at the amounts indicated in results. Human HARS and zebrafish ccnd1 mRNA were injected at a dose of 200 pg.

Total RNA was isolated from manually dechorionated embryos at 48 hpf using Trizol:chloroform (Invitrogen, Carlsbad, CA, United States), and used to synthesize first strand cDNA using Reverse Transcriptase (Applied Biosystems, Foster City, CA, United States) as described in the product manual. Table 1 contains the primers used to amplify asns (NM_201163), gpt2 (NM_001098757), eif4ebp1 (NM_199645), and ccnd1 (NM_131025) using Q5 DNA Polymerase (New England Biolabs, Ipswich, MA). Q5 amplified products were cloned into pCR-Blunt-II TOPO using the Zero Blunt TOPO Cloning Kit (Invitrogen, Carlsbad, CA, United States) and sequence verified by the UVM Cancer Center Vermont Integrative Genomics Resource using M13 primers.

After sequencing, full length ccnd1 mRNA was made by linearizing the pCR-Blunt-II TOPO-ccnd1 plasmid with XbaI (New England Biolabs, Ipswich, MA, United States). Linearized DNA was used as a template for in vitro transcription using the Sp6 mMessage mMachine Kit (Invitrogen, Carlsbad, CA, United States) as described in product manual. Transcripts were polyadenylated using a Poly(A) Tailing Kit (Invitrogen, Carlsbad, CA, United States) as described in the product manual.

The ccnd1 in situ probe was made by linearizing the pCR-Blunt-II-TOPO-ccnd1 plasmid with SacI (New England Biolabs, Ipswich, MA, United States). Zebrafish asns, gpt2, and eif4ebp1 in situ probes were made from EcoRI digested pCR-Blunt-II TOPO plasmids and transcribed using a T7 RNA polymerase (Affymetrix, Santa Clara, CA, United States) and DIG-labeling mix (Roche, Indianapolis, IN). Zebrafish hars (NM_001302262) in situ probes were made by PCR amplifying hars with the primers shown in Table 1 (Dr-hars) and transcribed using a T7 RNA polymerase as above.

The human HARS coding sequence (NM_002109) was gifted to us by Dr. Anthony Antonellis, University of Michigan. The primers used to amplify the coding sequence of HARS from the pcDNA vector are shown in Table 1 (Hs-HARS). The product was then transcribed using a T7 mMessage mMachine Kit (Invitrogen, Carlsbad, CA, United States) and polyadenylated as above.

Whole uninjected and hars KD embryos were collected at 48 hpf and RNA was isolated as above. For each group, equal amounts of RNA were used to generate cDNA as above. We used the same primers for asns, gpt2, and eif4ebp1 as used to make the in situ probes (described above) and performed PCR using Q5 DNA Polymerase (New England Biolabs, Ipswich, MA, United States). As a loading control, we used the eif1α primers shown in Table 1. Equal volumes of each PCR were run on a 2% agarose gel stained with 1:10,000 SybrSafe (Invitrogen, Carlsbad, CA, United States). Gel images were captured using a Syngene GeneSys imaging system and ImageJ was used to measure mean gray values for densitometry.

In situ hybridization was carried out as in Thisse and Thisse (2014). Briefly, embryos were raised to the desired stage and fixed in 4% PFA (Kimmel et al., 1995). Fixed embryos were permeabilized with proteinase K at 10 μg/mL and incubated with RNA probes at 70°C (probe synthesis described above). Additionally, full probe sequences are provided in the Supplementary Material. Probes were labeled with anti-DIG AP primary antibody (Roche, Basel, Switzerland) at 4°C overnight and then embryos were stained with NBT/BCIP (Thermo Scientific, Rockford, IL, United States). Stained embryos were mounted in 4% methyl cellulose on glass depression slides and imaged on a Nikon SMZ800 dissecting microscope at 5× using SPOT imaging software version 5.2. Images were uniformly adjusted for brightness and contrast in Adobe Photoshop CS6.

For zebrafish histology, embryos were staged as above and fixed in 4% PFA. Embryos were embedded using the JB-4 Embedding Kit (Polysciences, Inc., Warrington, PA, United States) and sectioned at a thickness of 7 μm on a Leica RM2265 microtome. Sections were mounted on slides and stained with hematoxylin and eosin. Stained sections were imaged at 20× on an Olympus IX71 microscope. Images were uniformly adjusted for brightness and contrast in Adobe Photoshop CS6.

After morpholino injection, fish were raised to the desired stages and fixed in 4% paraformaldehyde. Embryos were washed in PBS + 0.01% Triton X-100 and permeabilized in ice-cold acetone. Primary antibodies, anti-phospho-Histone H3 (Cell Signaling Technologies, Cat. # 3377S, Danvers, MA, United States) and anti-cleaved-caspase 3 (Cell Signaling Technologies, Cat # 9579S, Danvers, MA, United States) were applied to embryos at 1:1000 and incubated at 4°C overnight. Secondary antibodies, anti-rabbit conjugated to Alexa Fluor 555 (Cell Signaling Technologies, Cat. # 4413S, Danvers, MA, United States) were also added at 1:1000 and then incubated at room temperature for 2 h. 18 hpf embryos did not require acetone permeabilization and were mounted in 4% methyl cellulose and imaged wholemount at 10x on an Olympus IX71 microscope. Labeled cells were counted using the ImageJ Cell Counter tool and eye field area was measured using SPOT imaging software version 5.2. After immunostaining, 24 h through 72 h post fertilization embryos were embedded and sectioned as for histology. Sections were counterstained with Vectashield with DAPI (Vector Labs, Burlingame, CA, United States) and imaged at 20× on a Nikon Ti confocal microscope. For the sections, labeled and total cells were counted using the ImageJ Cell Counter tool. All images were compiled and adjusted for brightness and contrast in Adobe Photoshop CS6.

After injection, Tg(Rx3:GFP) embryos were raised to the 18 somite stage and screened for GFP expression. GFP+ embryos were washed in PBS and dissociated by being pressed through a 70 μm filter. Cells were fixed in Zinc Buffer (0.05% CaAc, 0.5% ZnCl₂, 0.5% ZnAc, in 1M Tris-HCl) at 4°C overnight (Jensen et al., 2010). Fixed cells were stained with 3 μM DAPI in PBS with 0.1% Triton X-100. Flow cytometry analysis was performed on BD LSRII equipped with a 405 nm laser for excitation of DAPI, and signal detected using a 450/50 BP filter. Unstained cells were used to set up the instrument. Cytometer optimization and calibration were performed as recommended by standard guidelines (Wang and Hoffman, 2017). Fifteen thousand singlet (based on FSC-A vs. FSC-H) events were recorded. Data were acquired using BD FACSDiva software v 8.0.1 and analyzed with FlowJo software v 10.5.3, using similar univariate cell cycle model for all samples (FlowJo, LLC, Ashland, OR, United States).

All statistical analyses were performed using Graphpad Prism 7. Graphs present all data points and, in most cases, the mean and standard error of the mean (SEM). Tests used are indicated in results but include: ordinary one-way ANOVA followed by Tukey’s multiple comparisons test; two-sided, unpaired t-tests; and two-sided, paired t-tests.

We began investigating the role of hars in zebrafish development, by asking where and when the gene was most highly expressed. Using in situ hybridization, we found that hars was most strongly expressed in regions of the developing nervous system (Figure 1A–E’). At 18 h post fertilization (hpf), expression is fairly ubiquitous throughout the embryo but we noted strong expression in the optic vesicles (Figure 1A,A’). At 24, 36, and 48 hpf, there is higher expression in the developing eye, ear, and optic tectum (Figure 1B–D’). At 72 hpf, expression is weaker, however, still found in the retina, ear, and throughout the brain (Figure 1E,E’). Transverse and sagittal sections through the eye further show hars expression in areas associated with proliferation (Figure 1F–J, L–N).

In line with these expression patterns, a search through in situ hybridization data for mouse Hars show similar patterns, with striking expression in the early retina before neurons have started differentiating (Diez-Roux et al., 2011). Additionally, our zebrafish hars expression patterns are also similar to the expression of other zebrafish ARS genes, suggesting that this pattern is not exclusive to hars (Fukui et al., 2009; Cao et al., 2016; Castranova et al., 2016; Kopajtich et al., 2016; Ni and Luo, 2018; Wang et al., 2018). Overall, these patterns support the idea that ARS are uniquely important for the generation of neuronal cells.

Based on the hars expression profile, we set out to assess how neuronal development is affected by a reduction in hars expression. Zebrafish (and the other members of the teleost lineage) are unique among vertebrates in that they have a single hars gene that codes for both a mitochondrial and a cytoplasmic enzyme via alternative splicing and translation start sites (humans and other vertebrates have two separate genes) (Waldron et al., 2017). The two transcripts vary at their 5′ end, so to avoid disrupting expression of mitochondrial proteins, we used a knock-down approach designed to block translation of the cytoplasmic hars transcript but not the mitochondrial. We first looked at the effect of knocking down hars on the development of the entire zebrafish by injecting fish at the one-cell stage and letting them grow for 72 h under normal rearing conditions. Initially, hars knock-down embryos appear healthy overall, but upon closer inspection we found that compared to body length, their head and eyes are smaller than uninjected siblings in a dose-dependent manner (Figure 2A). Presumably, there is a level of synthetase at which life is no longer viable, and indeed embryos injected with the highest dose of morpholino showed much more severe, whole body defects. However, to address whether neuronal tissues are more sensitive to hars, we chose to proceed with the dose that consistently resulted in the small eye phenotype without severe effects to the rest of the embryo. To control for off-target effects of morpholinos, we used a standard control morpholino and show that these fish have no phenotype, indicating the phenotype is specific to the hars morpholino (Figure 2B). In addition, to control for non-specific activation of the p53 pathway, we co-injected a Tp53 morpholino with the hars morpholino and found that that phenotype is not Tp53 dependent (Figure 2B). We also generated mRNA for GFP that had its 5′ end replaced with the hars morpholino binding sequence or a 5 base pair mismatch version of the sequence. Injection of these mRNA with and without the morpholino showed that the morpholino could abolish GFP expression from the wildtype mRNA, but had little effect on the 5 base pair mismatch version, supporting the specificity of the morpholino (Supplementary Figure S1). Importantly, we were able to partially rescue eye size by co-injecting mRNA for human HARS, confirming the specificity of our morpholino (Figure 2C–H). Despite human and zebrafish HARS being highly conserved (77% identity), there are likely differences in tRNA recognition between the two organisms that could account for the partial rescue as opposed to full rescue (Waldron et al., 2017).

For concision, we chose to focus our analysis on the retina as it represents a well characterized, accessible subset of the central nervous system (London et al., 2013). We took transverse sections through the retina at 24, 48, 72, and 96 hpf and stained with hematoxylin and eosin to investigate whether there were morphological defects in the retina, and at what developmental stage the phenotype is observed (Figure 3A–H). By counting the number of cells per central retina section, we found that hars KD embryos had fewer retinal cells at each age, with the phenotype apparent as early as 24 hpf (Figure 3I). However, there did not appear to be any striking impact on the overall patterning of the retina as all layers are clearly present by 96 hpf.

We next asked whether the reduced cell numbers in the eyes of hars KD embryos were due to reduced proliferation or increased cell death. We performed immunohistochemistry for markers of these two processes to determine the extent to which each of them contributes to this phenotype. Because we already observed fewer cells at 24 hpf, we started with an even younger age for these experiments. At 18 hpf, brain derived eye-fields are organizing into an optic cup, which is composed of proliferative retinal progenitor cells. We used a transgenic line that expresses GFP in these cells (Tg(Rx3:GFP)) and labeled for either phospho-histone H3 (pHH3) (a marker of cells in M-phase) or cleaved-caspase 3 (cCasp3) (a marker of apoptosis) then imaged the eye-fields dorsally. At this time point, we found that there were significantly fewer proliferating cells as well as more apoptotic cells in the eye fields in hars KD embryos (Figure 4A–B”).

Quantifying labeled cells in the intact eye becomes difficult as development progresses and tissues get thicker. For 24 hpf through 72 hpf we performed whole-mount immunohistochemistry as for the 18 hpf embryos, but then took transverse sections through the eyes and quantified the percent of labeled cells in a section from the central retina. At 24 hpf we found that there is no longer a significant reduction in proliferative cells but that there are significantly more apoptotic cells (Figure 4C–D”). By 48 hpf proliferation appears to be equal in both groups, and cell death appears to be only slightly higher in the hars KD embryos (Figure 4E–F”). These differences were no longer apparent at 72 hpf (Figure 4G–H”). This data suggests that early decreases in proliferation and increases in cell death contribute to there being fewer retinal progenitor cells available to construct the retina.

Loss of ARS or their function has been shown to lead to an accumulation of uncharged tRNAs, which activate the eif2α kinase GCN2, and ultimately stalls translation which initiates a stress response termed the Amino Acid Starvation Response (AASR) (Harding et al., 2000; Wek et al., 2006). This response drives the transcription of stress-related genes, including asns, gpt2, and eif4ebp1 (Sundrud et al., 2009). To test whether this response is activated in hars KD embryos we performed in situ hybridization and semi-quantitative RT-PCR for these three genes (Figure 5). Both of these measures showed that these three genes were upregulated in hars KD embryos, though to varying extents (Figure 5A–H). Interestingly, the in situs show that they are upregulated in a tissue specific manner (Figure 5A–F’). In hars KD embryos these genes were most strongly expressed in proliferative regions of the nervous system, such as the posterior optic tectum and the region surrounding the lens in the eye (likely the ciliary marginal zone) (Figure 5A–F’). Moreover, these regions are where we saw strong hars expression, further supporting the idea that these tissues are particularly sensitive to hars KD (Figure 1).

Ultimately, one consequence of the AASR is inhibition of cyclin D1 (CCND1) accumulation (Figure 5I). A lack of CCND1 causes cells to stall in G1 and eventually undergo apoptosis. We used FACS DNA-content analysis to test whether retinal progenitor cells were indeed stalling in G1 in hars KD embryos. Analysis of GFP + cells from 18 hpf Tg(Rx3:GFP) embryos revealed that an average of 52.63% of cells were in G0/G1 in control embryos while an average of 63.8% were in G0/G1 in hars KD embryos (Figure 5J). Taken together, these findings support the idea that hars KD inhibits cell proliferation by inducing the AASR and that neuronal tissues are particularly sensitive to this stress response.

In addition to the tissue-restricted expression of the stress response genes, we also noticed that ccnd1 itself shows similar expression patterns to hars, further supporting the tissue specific phenotypes observed in hars KD embryos (Figure 6A–B’). Another study linked HARS and CCND1 when they identified a temperature-sensitive, cell cycle deficient hamster cell line with a point mutation in the gene for Hars that they found was unable to accumulate CCND1 and predictably stalled in G1 (Motomura et al., 1996). Interestingly, they also found that they could rescue the G1 stall by overexpressing Ccnd1 mRNA. To test this finding in our system, we co-injected mRNA for zebrafish ccnd1 into hars KD embryos and measured relative eye sizes at 72 hpf as in Figure 2. Surprisingly, the overexpression of CCND1 was able to fully rescue the eye size phenotype in the hars KD embryos (Figure 6C–G). These results combined with the AASR results indicate that accumulation of CCND1 is indirectly dependent on HARS function.

Finally, we chose to look at the effect that hars KD had on other neuronal cell types in the zebrafish. Because two of the diseases associated with human HARS mutations affect the peripheral nervous system and auditory system, we utilized three transgenic lines that allowed us to look at the sensory neurons, motor neurons, and sensory hair cells. hars knock-down in Tg(pou4f3:GFP) embryos caused a severe reduction in the number of sensory hair cells seen in the lateral line at 72 hpf (Figure 7A,B). We also noticed axonal branching abnormalities in sensory and motor neuron development when hars was knocked down in Tg(ngn1:GFP);Tg(mnx1:mCherry) embryos (Figure 7C,D). These results indicate that a reduction in neuronal cell numbers is not unique to the retina and that other neuronal progenitor cell populations are also sensitive to hars knock-down.