Transcriptome and Proteome Analyses Reveal Stage-Speci c DNA Damage Response in Embryos of Endangered Sturgeon

Ievgeniia Gazo (  gazo@frov.jcu.cz ) University of South Bohemia in České Budějovice Ravindra Naraine Academy of Science of the Czech Republic Ievgen Lebeda University of South Bohemia in České Budějovice Aleš Tomčala University of South Bohemia in České Budějovice Mariola Dietrich Polish Academy of Sciences Roman Fraňek University of South Bohemia in České Budějovice Martin Pšenička University of South Bohemia in České Budějovice Radek Šindelka Academy of Science of the Czech Republic


Introduction
Genotoxicity is the property of chemical agents to induce DNA damage, modi cations, rearrangements, and mutations. The issue of genotoxicity has become increasingly important as numerous genotoxic compounds have been found in the environment [1]. Previous studies have shown that genotoxic compounds may have a long-term impact on organisms [2,3], particularly on developing embryos [4].
Embryogenesis is sensitive to DNA damage as the cell cycle is short and DNA replication is fast [5].
Furthermore, at the early stages of development most vertebrate species lack functional cell-cycle control checkpoints [6]. Subsequently, if embryos are exposed to genotoxicants before checkpoint activation, DNA synthesis and cell division continue in the presence of damaged DNA. In sh after the midblastula transition (MBT) checkpoints become activated, they can block developing embryos until DNA is repaired or apoptosis is initiated [7]. Thus, embryonic development and the cell cycle are closely associated with DNA repair.
The main DNA damage response (DDR) pathways, which are highly conserved between organisms [8], include nucleotide excision repair (NER), base excision repair (BER), DNA mismatch repair (MMR), homologous recombination (HR), and nonhomologous end joining (NHEJ) [9]. Nevertheless, the activity of these pathways during embryo development and their ability to mitigate DNA damage at different stages is still not fully described [10], particularly in non-model species.
In the current study, we have focused on the embryonic response to DNA damage in sterlet (Acipenser ruthenus). The sterlet belongs to the family of sturgeons (Acipenserids), a group of ray-nned sh, which likely evolved approximately 200 million years ago [11]. They have been classi ed as vulnerable shes by international organizations (https://www.iucnredlist.org/; [12]). Sturgeons are important aquaculture species known for their meat and caviar production. They represent an attractive model for studying DDR owing to their multiple rounds of lineage-speci c whole-genome duplication [13]. Furthermore, sturgeons have extraordinary genomic plasticity, as demonstrated through intraspeci c hybridization of individuals with different chromosome numbers [14], events of spontaneous polyploidization [15,16], and arti cial ploidy manipulation that yields the highest documented chromosome number in vertebrates [17].
According to [18], genome size is linearly related to the number of DNA repair genes and proteins. Taking all this into account, it is possible to speculate that the sturgeon's genome plasticity would require a highly effective DDR.
Several studies have already investigated the complex effects of water pollution on the transcriptome and/or proteome of model species, such as zebra sh [19,20]. Such advanced "omics" approaches can decipher the mode of action of the compound [19] and helps to understanding the organism's response to certain types of toxicity. One of the most popular "omics" techniques for the determination of adverse effects is transcriptomics (analysis of RNA transcript abundance). There are several advantages to transcriptomics, such as the low quantity of the biological material needed for analysis, high sensitivity, and robustness [21]. However, the key regulators of the cellular pathways are proteins, which can undergo post-translational modi cations and degradation; thus, the levels of functionally available proteins may be independent of transcription [22]. However, 2D gel protein electrophoresis combined with mass spectrometry (MS) has limited sensitivity and requires a large quantity of biological material.
Surprisingly, very few studies have performed complex analyses of the DDR in aquatic organisms, even though DNA damage is commonly used as a biomarker of exposure [23,24]. Therefore, in the present study we have analyzed the effect of two genotoxic compounds, camptothecin (CPT) and olaparib (Ola), on sterlet embryo development at different stages. Camptothecin is an inhibitor of topoisomerase I (Top1), an enzyme that regulates DNA topology [25]. CPT stabilizes Top1-DNA complexes leading to their conversion into DNA single-and double-strand breaks (SSB and DSB). Olaparib is an inhibitor of poly (ADP-ribose) polymerase-1 (PARP-1), a protein that plays crucial role in DNA damage sensing and repair [26]. Both chemicals are commonly used in studies on DNA damage and repair [7,27,28]. To understand the DDR in sh embryos, we have analyzed the effect of CPT and Ola on embryo morphology, DNA fragmentation, the transcriptome, and the proteome at different stages of embryo development.

Effect of Genotoxicants on Embryo DNA Integrity and Phenotype Formation
We exposed sterlet (A. ruthenus) embryos to CPT and Ola at different stages of embryonic development. The results of CPT exposure on sturgeon embryo development have been described in detail in our previous study [10]. Therefore, we will only summarize here the effects of CPT on sterlet embryo viability, hatching rate, DNA integrity, and phenotype formation. Brie y, 10 nM CPT was lethal for sterlet embryos exposed at the early stage (2-24 hpf), a mortality rate of up to 70% was induced when embryos were exposed during the gastrula stage (24-48 hpf), with a slight effect on viability following exposure during the neurula stage (48-72 hpf) (Fig. 1A). The hatching rate was unaffected when embryos were exposed to CPT during the neurula stage, whereas embryos exposed to CPT during gastrulation showed a low hatching rate (2%) (Fig. 1B). The level of DNA fragmentation reached 21.3% TailDNA in CPT 2-24 group (Fig. 2). The embryos in CPT 24-48 group showed the highest level of DNA fragmentation at 5 dpf (14.2% TailDNA), which decreased by 7 dpf (Fig. 2). The sterlet embryos in CPT 48-72 group showed no increased DNA fragmentation compared with the control. The images of the phenotypes observed following CPT exposure are presented elsewhere [10].
Nevertheless, the viability of the embryos decreased signi cantly compared with the control (Fig. 1A'). Further, the hatching rate was as low as 6.8% in embryos exposed to Ola at 2-24 hpf (Fig. 1B'). When embryos were exposed to a PARP-1 inhibitor in the later stages, the hatching rate at 8 dpf reached 32.8% for Ola 24-48 hpf and 40.4% for Ola 48-72 hpf (Fig. 1B'). Similarly, DNA fragmentation at 5 dpf was signi cantly higher for Ola 2-24 hpf and Ola 24-48 hpf compared with the control (Fig. 2). However, the group Ola 48-72 did not show a signi cant increase in DNA fragmentation. Furthermore, we did not observe the formation of phenotype in this group, unlike in embryos exposed at earlier stages (Fig. 3). Embryos exposed to Ola at the early stages showed skeletal malformations and a phenotype similar to developmental delay (Fig. 3).
For RNA sequencing and proteomic analysis, we selected the embryos from the CPT 24-48, CPT 48-72, Ola 2-24, and Ola 24-48 hpf groups. CPT exposure in the earlier stage was lethal and thus no embryos survived until hatching. Furthermore, the CPT 24-48 and CPT 48-72 showed very different patterns of DNA damage and different phenotypes, which made them of interest for further analysis. Similarly, the embryos exposed to Ola at 2-24 and 24-48 hpf showed different levels of DNA fragmentation and phenotype formation, whereas the group exposed to Ola at 48-72 hpf did not show signi cant increase in DNA damage or speci c phenotype formation and was therefore excluded from the further analysis.
We selected embryos at the hatching stage for further analysis as, at this stage, differences in phenotype formation between treatments were the most prominent. In addition, we aimed to determine the DDR that led to DNA repair observed at 7 dpf in all chosen groups (Fig. 2).
The DEGs that were detected with a lower mRNA abundance in the treatment groups compared with the control (log2 fold-change < 0) were referred to as "downregulated", and those that showed a higher abundance in the treatment compared with the control (log2 fold-change > 0) were de ned as "upregulated". The enrichment of gene ontologies associated with biological processes was determined for the downregulated and upregulated genes. There were very few downregulated GO terms associated with biological processes in the CPT 48-72 group ( Supplementary Fig. 1  There was also speci c downregulation of terms associated with development (e.g., circulatory system), transport and apoptotic cell clearance in the CPT 24-48 and Ola 2-24 groups.
The KEGG pathway enrichment analysis showed the downregulation of oxidative phosphorylation in all treatment groups (Fig. 5). Only CPT 24-48 and Ola 2-24 groups had the downregulation of the p53 signaling pathway and autophagy pathways, whereas the autophagy pathway was, in contrast, upregulated in Ola 24-48. There was downregulation of pathways associated with sugar metabolism, protein processing, and the biosynthesis of amino acids in all treatment conditions, except CPT 48-72.
In comparison with the control embryos, all treatment groups also shared several upregulated GO main terms ( Supplementary Fig. 2, Supplementary Table. 1). These were related to biological processes, such as the increase in the regulation of cellular and metabolic processes, RNA biosynthesis and macromolecules. There were also upregulated terms associated with the organization of cellular components, chromatin, and the cytoskeleton in all treatment groups. Upregulation was also observed for GO terms associated with the cell cycle and cell cycle phase transition, but these were only seen in the Ola 2-24 and Ola 24-48, a few in CPT 24-48, and completely lacking in CPT 48-72.
The following KEGG pathways were upregulated in all treatment groups: adrenergic signaling in cardiomyocytes, cell cycle, Wnt signaling pathway, cellular senescence pathway, MAPK signaling pathway, melanogenesis, Notch signaling pathway, and mRNA surveillance pathway (Fig. 5). RNA degradation and mismatch repair was only observed in the Ola-treatment groups. The FoxO and ErbB signaling pathways were only upregulated in the CPT 48-72 and Ola 24-48.
We have analyzed enrichment of GO terms for the downregulated DEGs in the "developmental processes" category ( Supplementary Fig. 3). The results of this analysis revealed very limited downregulation of GO terms. The upregulated GO terms included anatomical structure development, animal organ development, brain development, eye development, cell development, head development, and sensory system development in all treatment groups ( Supplementary Fig. 3). The upregulation of GO terms related to embryonic development ending in birth or egg hatching, chordate embryonic development, and embryo development was detected only in the CPT 24-48, Ola 2-24, and Ola 24-48 treatment groups.

Differentially Expressed Genes Associated with Cell Cycle and DNA Repair Pathways
The upregulation of the KEGG pathways involved in NER, BER, and HR was observed for CPT 24-48, Ola 2-24, and Ola 24-48 groups. The upregulation of mismatch repair was observed only in Ola 2-24 and Ola 24-48 (Fig. 5). The upregulation of GO terms related to DNA repair and response to stress was detected only in the CPT 24-48 and Ola-treatment groups ( Supplementary Fig. 2).
We have compared the changes in the expression of genes associated with cell cycle, checkpoints, and DNA repair (Fig. 6). The results of this analysis showed that numerous cell cycle genes were signi cantly upregulated in the Ola 2-24, Ola 24-48, and CPT 24-48 groups (Fig. 6, Supplementary Fig. 4). However, the largest fold-change in upregulation was observed in Ola 2-24. All treatment groups showed signi cant upregulation of tfdp2 and downregulation of ccnd1. In general, CPT 24-48 and Ola resulted in similar trends in cell cycle gene expression.
Among the genes associated with DNA repair, ddb1, parp8, apex1, rad52, and ogg1 were upregulated in all treatments, whereas xrcc5 and parp9 were downregulated (Fig. 6). Sterlet embryos from Ola 2-24, Ola 24-48, and CPT 24-48 groups showed increased expression of xpc, rad50, xpa, xrcc1, msh2, rpa1, ercc5, pold3, ercc2, fen1, blm, rad51ap1, nbn, and eme1. In contrast, the expression of genes associated with DNA repair was only slightly increased in CPT 48-72 group. Mapping the KEGG pathways of BER We compared the 2D-PAGE protein patterns between the control embryos and embryos exposed to genotoxicants (N = 3 for each group) ( Supplementary Figures 9, 10). The results of the mass spectrometric analysis of differently abundant proteins (DAPs) are presented in Supplementary Tables 2  and 3. We have identi ed 25 DAPs following exposure to CPT. Interestingly, from these proteins only two followed the same trend as mRNA levels, namely co lin-2-like (upregulated relative to control) and translationally controlled tumor protein homolog (downregulated relative to control). The most upregulated group of proteins in both CPT-treated groups (CPT 24-48 and CPT 48-72) were zona pellucida sperm-binding proteins (Supplementary Table 2). Several proteins associated with metabolic pathways were downregulated in CPT-treated embryos: putative aminopeptidase W07G4.4, alpha-enolase isoform X2, and creatine kinase M-type. This result agreed with the results of RNA sequencing from the CPT 24-48, where the DEGs associated with metabolic pathways were downregulated (Fig. 5, Table 1). Table 1 Summary of the results obtained in the current study for sterlet (A. ruthenus) embryos exposed to 10 nM camptothecin (CPT) and 20 Table 3). One protein was upregulated in the same way as its corresponding mRNA, namely protein disul de-isomerase A3. In contrast to CPT, embryos treated with Ola had a higher number of upregulated proteins involved in oxidoreductase and antioxidant activity.
We have summarized the results obtained in the current study in the Table 1. Our data indicate that treatment with Ola upregulated proteins associated with KEGG metabolic pathways and peroxisome. In contrast, treatment with CPT downregulated cytoskeletal proteins (Table 1).
We have also analyzed correlation between viability, hatching rate, DNA fragmentation, number of DEGs, and the number of DAPs (Table 2). According to Spearman's test, there was a strong positive correlation between viability and hatching rate. Furthermore, a signi cant negative correlation was observed between viability and the number of DEGs, viability and DAPs, as well as hatching and DEGs.

Discussion
The current study aimed to describe the DDR in sturgeon embryos induced by exposure to genotoxicants.
Considering the endangered status of many sturgeon species and their importance to aquaculture, the studies on sturgeon embryo development are highly demanded. Furthermore, the results of our research indicated that sturgeon embryos were most sensitive to stress in the early stages of development. Following neurulation, the embryos acquire mechanisms to cope with chemical exposure and/or DNA damage. This information could be helpful to aquaculture, where sturgeon embryos are produced and reared in large amounts in vitro.
We have previously shown that CPT exposure led to different levels of DNA damage and different malformation rates, depending on the stage of exposure [10]. Similarly, Ola exposure in the early stages of development led to decreased survival and hatching rate, whereas exposure after neurulation did not have a signi cant effect on sturgeon embryos. In contrast to CPT, Ola exposure at early stages (2-24 hpf) did not lead to total mortality and embryos developed up to 7 dpf. Nevertheless, these embryos failed to hatch, with a high rate of malformations, indicating that the proper functioning of PARP-1 is required for embryo development. This was agreed with previous studies on zebra sh embryos that showed how the inhibition of PARP-1 could lead to increased DNA damage, cell death, and problems with development [26,28,29].
In general, gastrulation is a critical phase for sh embryos, as, at this point, cells acquire the ability to undergo apoptosis [7]. When sh embryos were treated with genotoxicants after gastrulation they showed a higher rate of survival and hatching than the group treated before gastrulation ( [7,30]; current study). There are several possible hypotheses to explain this gradual increase in the resistance of post-MBT embryos. First, it is possible to suggest that the activation of cell cycle checkpoints and apoptosis after the MBT transition allows embryos to effectively repair DNA or remove damaged cells [7]. On the other hand, it is known that, upon gastrulation, sh embryos develop an epidermis that acts as the major barrier to harmful molecules [31]. The increased expression of ATP-binding cassette transporters in the epidermis could result in a higher resistance of embryos after gastrulation [32]. This is supported by the report of [26], who observed the decreased uptake of the genotoxicant doxorubicin in post-MBT zebra sh embryos compared with the early stages of development. Further studies are needed to understand kinetics of genotoxic compounds in sh embryos at different stages of development.
We compared the changes in the transcriptome between control sturgeon embryos and embryos treated with CPT at 24- [21], who suggested that the adverse effects of chemicals should be de ned based on impairment of functional capacity and pathology formation, rather than only alterations at transcriptomic level. Thus, it is possible to draw a correlation between embryo phenotypes and omics data in the case of genotoxic treatments applied in the early stages of embryo development ( Table 2).
In contrast, the CPT 48-72 group developed no phenotype, but several pathways, such as developmental pathways, oxidative phosphorylation, metabolism, and the regulation of the cytoskeleton, were signi cantly altered at the transcriptome and proteome levels. In this case, it was di cult to distinguish between the adverse effect of genotoxicity and adaptive response. However, we observed signi cant upregulation of developmental processes and nervous system development in this treatment. Previous studies indicated that early life stress may have large effects on the developing brain and nervous system in vertebrates [33,34]. Further studies are needed to understand if exposure to genotoxicants after neurulation may affect sturgeons later in life and subsequently induce adverse effects.
The analysis of gene expression associated with cell cycle and DNA repair showed that several genes were upregulated in all treatment groups, such as tfdp2, tp53bp1, ddb1, parp8, apex1, rad52, and ogg1, whereas ccnd1, tp53, parp9, and xrcc5 were downregulated. Thus, cell cycle, p53 signaling, and some DDR pathways were affected by all test conditions. The observed decrease in mRNA levels of p53 (encoded by tp53) and cyclin D1 (encoded by ccnd1) may indicate that embryonic cells overcome cell cycle arrest and undergo proliferation [35,36]. However, the increased expression of check1, atm, and atr in CPT the 24-48, Ola 2-24, and Ola 24-48 groups indicated that at least some portions of embryonic cells still undergo checkpoint arrest. This assumption is supported by the low hatching rate in those groups.
We observed the upregulation of several DNA repair pathways in response to CPT and Ola exposure. Thus HR, BER, and NER were upregulated in embryos exposed to CPT at 24-48 hpf and to Ola at 2-24 and 24-48 hpf. A previous study showed that DSBs induced by Ola are repaired through the HR pathway in zebra sh embryos [28]. Similarly, CPT has been shown to induce DSBs repaired through HR in embryonic stem cells [37]. The activation of several DDR pathways in response to genotoxicity could be attributed to the cooperation between these pathways. Thus, the previous study showed the interaction between the HR and NER pathways in removing the inter-strand crosslinks (ICL) [38]. Furthermore, interplay between BER and NER pathways is well-documented [39]. Overall, in the developing sterlet embryos, it is possible that exposure to genotoxicants induced not only DSB, but also SSB and ICL formation, as well as oxidative stress, which in turn required activation of several DDR pathways.
The exposure of sturgeon embryos to genotoxicants has affected the expression of genes associated with metabolism and the cell cycle. Transcriptomic changes were similar to the proteomic changes, in which the abundance of several metabolic proteins was modi ed. The association between DDR and metabolism is well known [40,41]. It has been previously shown that growth factors govern metabolism, while metabolism is a factor that drives cell cycle and proliferation [42]. Furthermore, it is known that DNA damage causes cells to rewire their metabolism [41]. The obtained results suggest that DDR induces the adaptive response of metabolic pathways in developing sh embryos, which also affects cell cycle and growth. In agreement with this conclusion, we also observed changes in vitellogenin (Vtg) abundance in embryos exposed to CPT and Ola (Supplementary Tables 2 and 3). In the oocyte, Vtg serves as one of the main sources of nutrients as well as free amino acids for protein synthesis during early embryogenesis [43]. A previous study reported changes in Vtg cleavage products in zebra sh embryos in response to toxic stress [44]. The stress response in sturgeon embryos altered the expression of genes associated with metabolism, and glycolysis, as well as the protein pro les of Vtg and glycolytic enzymes. Thus, DNA damage in sturgeon embryos is repaired at the cost of metabolic adjustments, changes in cell cycle, and proliferation.
Another highly abundant protein in sturgeon embryos treated with genotoxicants compared with control embryos was zona pellucida (Zp) sperm-binding protein. In general, Zp proteins are expressed in egg chorion and provide thickness and hardness [45]. The study of [13] on sterlet genome showed that there were 116 Zp genes in sterlet. The authors also noted that the possible reason for Zp gene family expansion in sturgeons is a protection mechanism against physical forces for the developing embryos. Based on the results of our study, it is possible to speculate that Zp proteins may also be involved in stress-response or protect embryos against xenobiotic exposure. This may indicate the neofunctionalization of Zp proteins in sturgeon.
Although we have observed many similarities in the embryo response to CPT and Ola, there were also compound-speci c changes in the transcriptome and proteome of sturgeon embryos. For example, the analysis of protein pro les showed that treatment with Ola upregulated the expression of several proteins associated with oxidoreductase and antioxidant activities. These results agreed with previous studies showing that PARP-1 inhibition induced the formation of reactive oxygen species and oxidative DNA damage [46]. In contrast, exposure to CPT at 24-48 and 48-72 hpf affected proteins associated with metabolism, actin laments, and cytoskeletal organization. A previous study has shown that CPT can interfere with microtubule polymerization and actin laments during mitosis [47]. These side effects of CPT and Ola could explain the low Spearman's correlation coe cients between DNA fragmentation and omics response ( Table 2).

Conclusions
This study is the rst report on transcriptomic and proteomic changes induced by DNA damage in sturgeon embryos. We have revealed that several DNA repair pathways were activated (HR, BER, NER, MMR) and DNA repair was closely associated with metabolic changes (oxidative phosphorylation, sugar metabolism, etc.). Thus, apart from DNA damage, embryo genotoxicant exposure can be associated with oxidative stress, changes in metabolism, and affected organ development. The effect of genotoxicants and the level of DNA damage were stage-speci c with prior-neurulation stages being more sensitive.
Our results highlighted correlation between phenotype formation and the omics response. Treatments that induced DNA damage and phenotype formation were associated with changes in expression of the DDR genes: check1, atm, atr, xpc, rad50, xpa, xrcc1, msh2, rpa1, ercc5, pold3, ercc2, fen1, blm, rad51ap1, nbn, and eme1. The presented pieces of evidence suggest that these genes could be used as markers of DNA damage and DDR. In contrast, several genes were modi ed irrelevant of phenotype formation: tfdp2, tp53bp1, ddb1, parp8, apex1, rad52, ogg1, ccnd1, tp53, parp9, and xrcc5. It can be concluded that the second group of genes is related to adaptive stress-response. We found that the sturgeon DDR mechanisms have some degree of similarity with previously reported zebra sh DDR, that agrees with the high level of DDR evolutionary conservation. Thus, it could be speculated that the obtained markers could nd its use in toxicological studies on broad variety of sh species.
Although DDR mechanisms were mainly conserved in sturgeons, we have observed a species-speci c response to genotoxicant exposure, namely

Ethics
Gametes were collected from live sterlet during the natural spawning period. Manipulations with broodstock were performed according to the law on the protection of animals against cruelty (Act no. 246/1992 Coll.) in the certi ed workplace (ref. number 6OZ3066/2020-18134). Fish embryos were exposed to genotoxicants before reaching an independently feeding larval form. Therefore, no special approval from the local ethical committee was needed according to the laws to protect animals against maltreatment (Act no. 246/1992 Coll.).
The embryo phenotypes were observed under a microscope and imaged at 2.5× magni cation. A minimum of 150 embryos was analyzed per treatment.

Comet Assay
DNA integrity was assessed using alkaline comet assays, or single cell gel electrophoresis assays, as described in our previous study [10]. Brie y, embryos at different stages were transferred into 1. mM EDTA, 10 mM Tris, 1% Triton X-100, 10% DMSO, pH 10) and incubated overnight at 4°C. After the cells were lysed, the buffer was removed, and the slides were placed in a horizontal gel tank lled with freshly prepared electrophoresis buffer (90 mM Tris base, 90 mM boric acid, 2.5 mM EDTA Normalization of the raw counts and detection of the differentially expressed genes (DEGs) between the treatment and the control samples were determined using R (v. 4.1.0) software, DESeq2 (v. 1.32.0), using the median-of-ratios method and a 0.1 p-adjusted (padj) cutoff [52]. Differentially upregulated and downregulated genes were de ned as those with padj < 0.1 and log2foldchange greater than one or less than one respectively. Gene ontology (targeting Biological Processes) and pathway (KEGG) enrichment analysis was performed using gpro ler2 (v. 0.2.0) with the default parameters [53]. All the annotatable genes for A. ruthenus were used as a custom statistical background, with multiple testing correction performed using Set Counts and Sizes (g:SCS) with a signi cance cutoff of 0.05. Related Gene Ontology (GO) terms were then clustered and summarized using the simplifyEnrichment package (v. followed by a second equilibration with a solution as above, but containing 2.5% iodacetamide instead of dithiothreitol, for a further 15 min. The strip was loaded on a 10% acrylamide gel for SDS-PAGE, and the proteins were separated by molecular weight using the Bio-Rad Mini-PROTEAN vertical electrophoresis system. After electrophoresis, the gels were stained with Coomassie Brilliant Blue R-250 to visualize protein spots and scanned using the documentation system Fusion Solo 7S Edge (Vilber Lourmat, France). The intensity of spots was analyzed using ImageJ (NIH, Bethesda, MD, USA). Only spots that were signi cantly changed between the control and treatment groups were selected for in-gel digestion and MS.

In-Gel Digestion and Protein Mass Spectrometry
After the gels were washed in water, the selected protein spots were cut from the gels. The gel pieces were  DNA damage analyzed by the comet assay as %TailDNA. Embryos were exposed to CPT and Ola at 2-24, 24-48, and 48-72 hpf. Percentage DNA in the comet tail was analyzed at 2, 5, and 7 dpf. Capital letters indicate signi cant difference from the control at the same dpf (ANOVA, p < 0.05). Lowercase letters indicate signi cant differences among time points (dpf) for each treatment (ANOVA, p < 0.05).  Overlap in signi cantly (padj < 0.1) differentially expressed genes between treatments. (a) All differentially expressed genes. (b) Only unique gene symbols.  Genes associated with DNA damage, cell cycle, and cell cycle checkpoint control that were signi cantly (padj < 0.1) differentially expressed in at least one of the treated samples relative to the control. Color scale represents the log2 fold-change between the treatment and the control and is capped at maximum of 3 and minimum of −3. A black dot represents the genes that were differentially expressed.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.