HUA treatment: Experimental design

Early Xenopus embryos rely on the consumption of maternally supplied molecules, which restricts the available repertoire of cell cycle manipulations. We have chosen a systemic cell cycle arrest, achieved by a combination of HUA (from here on called HUA condition), because, together, they block S phase more rapidly and more completely than each molecule alone [33]. When applied to cells in culture, both drugs effectively block DNA replication and consequently cell division, without obvious side effects on cell viability or differentiation capacity [34,35]. This effectively leads to a cell cycle arrest at the G1/S transition. Harris and Hartenstein have used these inhibitors to demonstrate that cell division is dispensable for neural induction, neuronal differentiation, and for neural tube formation [33]. We have confirmed their finding that, whereas HUA treatment before gastrulation (i.e., at NF9) leads to aborted development, it is principally compatible with embryonic development until tadpole stages when started shortly after gastrulation onset (NF10.5) (for more information, see Embryos handling and HUA treatment in Materials and methods section and S1A Fig). In this work, we have assessed the effects of the systemic cell cycle arrest by 2 types of experiments (Fig 1). Experiment type A (“permanent arrest”) implies a continuous HUA or Mock treatment on sibling embryos from NF10.5 until NF32. Mock treatment represents incubation in 2% dimethyl sulfoxide (DMSO), which is the solvent for aphidicolin. Experiment type B (“transient arrest”) consists of continuous HUA and Mock treatments plus an additional HUA washout (HUAwo) condition. In the latter case, embryos were temporarily incubated with HUA from early (NF10.5) to late (NF13) gastrula stages, and then returned to Mock solution. Experiment type B tests the reversibility of the cell cycle arrest and its developmental consequences.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Experimental design.

Top part: a schematic representation of Xenopus laevis embryonic development. Developmental stages (NF) according to Nieuwkoop and Faber (1994). Stages used for mass spectrometry of histone modifications and embryonic analyses are characterized by the following features: NF13—late gastrula, germ layer specified; NF18—neurula, germ layer patterning and differentiation; NF25—tailbud stage, organogenesis; NF32—early tadpole, body plan established. Bottom part: We perform 2 types of experiments: Experiment type A (G1/S block)—embryos are split in 2 groups, which from NF10.5 on are continuously incubated in HUA solution or control solution (DMSO as carrier). Experiment type B (transient arrest)—embryos are split in 3 groups: continuous Mock, continuous HUA, and transient HUA. In the last group, HUA solution is replaced at NF13 with DMSO solution. DMSO, dimethyl sulfoxide; HUA, hydroxyurea and aphidicolin; HUAwo, HUA washout.

https://doi.org/10.1371/journal.pbio.3001377.g001

For all experiments, embryo siblings were collected at 4 developmental stages: NF13 (early neurula), NF18 (late neurula), NF25 (tailbud), and NF32 (tadpole) (Fig 1). By the early neurula stage, germ layers and body axes have been determined. The embryonic patterning increases the cellular diversity of the embryos during neurulation. At the tailbud stage, the first differentiated tissues such as skeletal muscle and the mucociliary epithelium of the larval skin have been formed. The end point of the analysis (NF32) is after the phylotypic stage of Xenopus (NF28 to NF31), when evolutionary conserved gene expression programs have established the vertebrate body plan [36].

HUA effects on embryo viability and cell proliferation

At the first sampling stage (NF13), there was no difference in viability between proliferating controls and HUA-arrested embryos (S1A Fig). As reported before [33], HUA treatment reduced embryonic viability after NF18, such that the survival rate was about half of that of Mock embryos at NF25 and NF32. Staining for activated Caspase 3 indicated that both Mock- and HUA-treated embryos contained apoptotic cells in a variable, but comparable extent (S2 Fig), consistent with the results of an independent study using HUA in Xenopus [37]. Depending on individual egg batches, arrested embryos survived up to stage NF41.

To test the efficacy of the HUA incubation, we stained embryos for the histone H3 Serine 10 phospho mark (Fig 2), which accumulates in M phase cells [38]. We determined the number of H3S10Ph-positive cells from the entire surface of the embryo with ImageJ (Fiji). Already at NF13, HUA embryos contained more than 8-fold less mitotic cells than Mock embryos, a difference that further increased to 17-fold by NF32. We conclude that the cell cycle is efficiently blocked by HUA treatment.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Continuous HUA treatment inhibits mitotic activity from gastrula to tadpole stages.

(A) ICC for the mitotic histone mark H3S10Ph. Black dots represent mitotic cells. Elongated, older embryos are recorded as anterior halves, i.e., at the same magnification as younger stages, and in whole mount views as inserts. Scale bars: 1 mm. (B) Abundance of mitotic cells in Mock- and HUA-treated embryos. Box plots based on H3S10Ph-positive cells present on the recorded surface of embryos (n = 3 biological replicates/condition; Student t test [unpaired, two-tailed]; *** p < 0.001). The individual quantitative observations can be found in S1 Data. HUA, hydroxyurea and aphidicolin; ICC, immunocytochemical staining.

https://doi.org/10.1371/journal.pbio.3001377.g002

HUA effects on morphogenesis

The cell cycle–arrested embryos completed gastrulation in synchrony with Mock controls (Fig 3A). From midneurula onward (stage NF18/19), the closure of the neural tube was delayed, consistent with previous findings [33] (S1B Fig). At later stages, HUA embryos lacked a postanal tail, melanocytes, and the eye anlage remained rudimentary. Note that the size of arrested and Mock-treated embryos was almost the same until stage NF25 (Fig 3A). Since holoblastic cleaving Xenopus embryos rely completely on maternal storage, cell proliferation subdivides the zygote into smaller and smaller cell progeny, while maintaining the maternally derived mass. Consistent with the observed mitotic arrest, we confirmed that HUA embryos consisted of bigger cells at the neurula stage (S1B Fig). This observation is consistent with the previous finding, showing that blocking cell division with HUA stops the decrease in cell size during embryonic development [39]. Confocal imaging indicates that epidermal cells of HUA embryos were clearly spread out over a larger area than proliferating cells of control embryos, and their nuclei appear enlarged (Fig 3B). H3S10Ph-positive cells were detected in Mock controls, but not in HUA embryos (S1 and S2 Videos). Overall, these observations document a remarkable compensation on the cellular level, which enables HUA embryos to undergo morphogenesis in apparent synchrony with control siblings.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Morphological development of HUA-arrested embryos.

(A) During gastrulation, Mock and HUA embryos are indistinguishable. At NF18, the latter show a delay in neural tube closure. More severe malformations are detectable at stages NF25 and NF32, most notably reduced eye formation and absence of tail bud. Scale bars: 1 mm. (B) Cell size at the tailbud stage. Flattened Z-stack images show fields from embryonic skin at constant magnification (scale bars: 20 μm). Immunofluorescence detects cell borders (beta-catenin), nuclei (DAPI), and mitotic cells (H3S10Ph). GIFs from Z-stacks are presented in the Supporting information (S1 and S2 Videos). (C) Whole mount RNA in situ hybridization for indicated marker genes. Images are representative for the majority of Mock- or HUA-treated embryos from 3 biological experiments. While mRNAs of α-tubulin and rx1 (skin, neuronal), α-globin, tnni3, actc1, and mhc-α (mesodermal) are detected in both conditions, xbra and foxd5a (tail blastema) are absent in HUA embryos. Numbers indicate embryos positive for the marker over the total number of analyzed embryos. Scale bars: 1 mm. HUA, hydroxyurea and aphidicolin.

https://doi.org/10.1371/journal.pbio.3001377.g003

Differentiation of cell cycle–arrested embryos

Cell proliferation and differentiation are, in most cases, mutually exclusive processes [40]. To investigate the developmental relationship of Mock- and HUA-treated embryos, we investigated the tissue-specific expression of a panel of marker genes (Fig 3C and S1 Table). Whole mount RNA in situ hybridization revealed that most of these markers (17 out of 19) were expressed at the correct site, although in smaller domains. Only 2 genes were not transcribed in the majority of HUA-treated embryos. Xbra and foxd5 are normally expressed in the tail, which HUA embryos fail to form. To rule out that the absence of these mRNAs in cell cycle–arrested embryos was due to a developmental delay, we compared by quantitative reverse transcription polymerase chain reaction (qRT/PCR) the timing and relative levels of gene transcription between Mock- and HUA-treated embryos (S1C Fig). We examined 6 genes, which become induced at different time points—pax6 and actc1 at gastrulation, twist and myt1 during neurulation, and tnni3 and fabp2 in tadpoles—and normalized the relative mRNA levels to the housekeeping gene odc. A significant difference in gene expression between Mock and HUA embryos was found only for pax6, which had disappeared by stage NF32. The mRNA levels of the other 5 genes, including the late-induced fabp2 gene, were proportionate between the 2 conditions. This suggests that genes were activated around the proper time, arguing against a general delay in the development of HUA embryos.

Although HUA-treated tadpoles were morphologically impaired, they responded to touch stimulation by muscle twitching, and at least some of them were capable of mounting a flight response (S3 and S4 Videos). This burst of swimming activity involves a physiological connection between sensory neurons, the CNS, and the body wall musculature [41].

The morphological and molecular analyses demonstrated that embryonic development proceeds in the absence of cell proliferation, although the anlagen of some organs like tail, fin, and eyes were compromised. These findings are in agreement with and extend previous observations on HUA-treated embryos [33]. We conclude that blocking the cell cycle from gastrula stages onward is compatible with largely normal and apparently synchronous differentiation of the embryo at least until tadpole stage (NF32).

Stage-specific histone modification profiles in HUA embryos

To investigate the impact of the cell cycle arrest on embryonic chromatin, we compared the posttranslational HMPs in proliferating and HUA-arrested embryos by quantitative mass spectrometry (for more information, see Mass spectrometry analysis with scheduled PRM method and Histone PTM quantification in Materials and methods section). In total, this study quantified 64 histone modification states across 4 developmental stages in 2 experimental conditions, each in 3 biological replicates. The combination of all modification states at a certain developmental time point constitutes the so-called stage-specific HMP [20]. To visualize differences in the HMPs between cell cycle–arrested and control embryonic chromatin, we performed hierarchical clustering analysis with all 24 samples to build a single, stage-specific heatmap (Fig 4A). This clustering was based on the absolute intensity values of endogenous histone peptides.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Mitotic activity shapes stage-specific HMPs.

(A) Heatmap of the absolute histone modification states abundance, normalized to the corresponding R10 spiketides. Data in columns represent an average from 3 biological replicates/condition. Color key: row Z-score. The individual quantitative observations can be found in S2 Data. (B) Ratio heatmap of relative histone modifications abundance for selected histone marks. In the first step, relative distributions of the indicated modification states were calculated in percentages within each tryptic peptide (see S4 Fig). Then, HUA values were divided by Mock values to produce the relative change for each histone modification state between the 2 conditions. Color key is based on the Log10 scale. Numerical values in the cells indicate the fold-change. Values greater than 1.0 indicate an increase in HUA condition; values smaller than 1.0 indicate a higher abundance in Mock condition. NA, not applicable, when values in Mock were 0. Asterisks (*), adj.p < 0.1 (Benjamini–Hochberg procedure). “p,” propionylated (naturally unmodified); “un,” tryptic peptide, which has no modification state. The direct comparison of relative histone PTMs abundance is shown in S4 Fig, with p-values indicated in S5 Table. HMP, histone modification profile; HUA, hydroxyurea and aphidicolin; PTM, posttranslational modification of histone.

https://doi.org/10.1371/journal.pbio.3001377.g004

The first 3 levels of the dendrogram on the left axis of the heatmap revealed 5 clusters with the following features. Cluster 1 grouped histone PTMs, which undulate in control embryos, i.e., they were more abundant at stages NF13 and NF25 than at stages NF18 and 32. In cell cycle–arrested embryos, this pattern collapsed into a single high point at NF25. Cluster 2 included modification states, which gradually increased their abundance under both experimental conditions. Cluster 3 represented modifications, which were more abundant in HUA-arrested chromatin compared to control chromatin, whereas in clusters 4 and 5, highly abundant modification states in Mock embryos were down-regulated in HUA samples.

Despite the fact that the cell cycle arrest alters the abundance of nearly every analyzed modification to varying extent, HMPs are clearly discernible for both conditions (Fig 4A). To obtain further insight into their difference, we performed principal component analysis (PCA) for the whole data set (S3A Fig). The 2 conditions are partly separating, particularly clear for late gastrula control and tadpole HUA samples. In general, chromatin features of younger HUA samples seem to cosegregate with those of older controls.

HUA effects on individual histone marks

The global, absolute heatmap indicated that the HUA block resulted in a different developmental histone modification profile. To obtain a better understanding of how the histone modification states differed between the 2 conditions, we converted the absolute abundance of individual histone modifications into relative proportions for those modification states, which are located on a single tryptic histone peptide (S4 Fig and S5 Table). We then investigated the distribution of the relative histone PTM abundance over their fold-change between Mock and HUA (S3B Fig). Lower abundant modifications tend to have larger variability. Significantly different modifications (highlighted in red) are spread through the entire range of abundance but are more prevalent at the high abundance end. One-quarter of the analyzed histone PTMs is significantly different between Mock and HUA conditions (S3C Fig and S5 Table). We then generated a second heatmap to display the fold-change of some selected histone PTMs between Mock and HUA conditions at a given stage (Fig 4B). These PTMs were selected based on their biological relevance for embryonic development and established impact on gene activity. PTMs, whose fold-change between Mock and HUA conditions is statistically significant (adj.p < 0.1, Benjamini–Hochberg procedure), are marked in Fig 4B (and S5 Table).

The relative ratio heatmap supports the permanent nature of the G1/S cell cycle arrest by revealing on average a more than 10-fold reduction of the mitotic chromatin mark H3S10Ph. These data are in line with the assumption that HUA treatment affected not only superficial cell layers, but all embryonic tissues. Phosphorylation of Threonine 3 on histone H3 represents a second PTM that is mitotically enriched [42]. For unknown reasons, this modification shows a large variability between biological replicates in our data set, resulting in a very similar averaged abundance for Mock and HUA conditions in all 4 stages (S5 Table). Therefore, it cannot confirm our conclusions from the H3S10Ph mark.

Furthermore, the overall levels of H3K4me2/me3 were maintained in HUA embryos at all 4 stages, suggesting that the number of active and poised promoters can be maintained in nonproliferating embryos, consistent with the largely correct expression of cell-type specific markers (Figs 3C and 4B). Other modifications deviated between the 2 conditions. The embryonic epigenome of HUA-treated embryos contained reduced levels for some modifications, notably the acetylated states of K9 and K27 on histone H3. These modifications are thought to prevent the writing of repressive methylation marks on chromatin harboring cis-regulatory DNA elements [43,44], a prerequisite for de novo enhancer activation [45]. Beyond the H3S10Ph mark, the next largest differences in HUA-arrested chromatin were found for the trimethylated states of lysines 9 and 27 on H3, and lysine 20 on H4. These sites became globally elevated by more than 2.5-fold compared to the chromatin of control embryos. While higher levels of H3K27me3 and H4K20me2/3 were anticipated, since their levels are linked to cell cycle progression [46], the increase in H3K9me3 suggests that the HUA cell cycle arrest occurs partly in S phase [47].

Altogether, we demonstrated that incubating Xenopus embryos in HUA solution was an efficient way to systemically block cell proliferation. The HUA-induced cell cycle block had consequences on defined morphological aspects of development but also altered stage-specific HMPs. Most notably, this had an impact on specific histone modifications known to be involved in epigenetic memory and enhancer activity [45,48]. The chromatin differences between HUA and control embryos suggest that the cell cycle state—possibly cell proliferation as such—coordinates the developmental decoration of chromatin with covalent histone modifications.

HUA effects on embryogenesis are reversible

To check whether the observed differences in morphology, gene expression, and histone modification landscape are reversible, we performed a transient cell cycle arrest (HUAwo) (Fig 1). Only 4 hours after returning to Mock condition (i.e., at midneurula stage—NF18), H3S10Ph-positive cell numbers were significantly increased compared to siblings maintained in HUA (Fig 5A and S5 Fig). At the endpoint of the experiment, mitotic cells were equally abundant in HUAwo and Mock conditions, indicating full recovery of the mitotic activity within 13 hours after inhibitor removal. In addition, the survival rate of HUAwo embryos increased, compared to continuously arrested embryos, and was similar to that of Mock embryos (S1A Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. The HUA effects on embryogenesis are reversible.

Embryos, which were transiently incubated in HUA solution and thus mitotically arrested, were returned to Mock solution at NF13. (A) Abundance of mitotic cells in Mock, HUA-treated, and HUAwo embryos. Box plots based on H3S10Ph-positive cells present on the recorded surface of embryos, displayed in S5 Fig (n ≥ 3 biological replicates/condition; Student t test [unpaired, two-tailed]; *** p < 0.001; * p < 0.05; n.s., not significant). The individual quantitative observations can be found in S3 Data. (B) Morphological and molecular features of embryos at the early tadpole stage. In contrast to continuous HUA-treated embryos, HUAwo embryos regain eye cups, fin, and tailbud structures. In addition, they express foxd5a and xbra mRNAs in the growth zone of the tailbud, comparable to Mock embryos. Numbers indicate embryos positive for the marker over the total number of analyzed embryos (n = 3 biological replicates/condition). Scale bars: 1 mm. HUA, hydroxyurea and aphidicolin; HUAwo, HUA washout.

https://doi.org/10.1371/journal.pbio.3001377.g005

Morphological hallmarks also recovered toward normality in HUAwo embryos (Fig 5B). Unlike continuously arrested embryos, the HUAwo group frequently developed postanal tails of nearly normal length, as well as a clearly distinguishable fin surrounding trunk and tail. The recovery of the tail structure was accompanied by the recurrence of xbra and foxd5a mRNAs at the tail tip of all HUAwo embryos (Fig 5B). These findings indicate that the major morphological defects of the HUA arrest are reversible, at least when the inhibitors are removed during neurulation.

Encouraged by the results, we next performed mass spectrometry analysis on sibling cohorts of Mock, permanent HUA, and HUAwo embryos at the tadpole stage (NF32). Experimental series type A and B include 2 similar conditions: Mock and permanent HUA. To compare results of the same conditions between Exp.A and Exp.B, we looked at the distribution of histone PTM fold-change at NF32 (S3D Fig). We did not observe a significant difference, supporting the robustness of the mass spectrometry measurements.

Next, we visualized the results of the Exp.B (HUAwo) in a heatmap, based on relative values (Fig 6A and S6 Table). As for the experiment type A, the fold-change of functionally well-annotated histone PTMs was calculated from their relative proportions (Fig 6B and S6 Fig). In total, almost three-quarters of the modifications were closer to the levels found in control embryos than to HUA chromatin, although in most cases, they do not reach wild-type abundance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. HUA effects on stage-specific HMPs are reversible.

(A) Heatmap of the relative histone modification states abundance under the indicated conditions. As before, the relative distributions of the indicated modification states are calculated in percentages within each tryptic peptide. Data in columns are collected from early tadpole stage embryos NF32 (n = 1 biological replicate), independently from the data of experiment type A. Note: Relative differences between Mock and HUA samples of experiment type B are highly similar to the results of experiment type A (S6 Fig). Color key, row Z-score. The individual quantitative observations can be found in S4 Data. (B) Ratio heatmap of relative histone modification abundance for selected histone marks. First, the relative distributions of the indicated modification states are calculated as percentages within each tryptic peptide (S6 Fig). Then, the relative values are presented as ratios between sample pairs indicated on top, showing the relative change for each histone modification state between the 2 conditions. Color key is based on the Log10 scale. Numerical values in the cells indicate the fold-change. “p,” propionylated (naturally unmodified). Values greater than 1.0 indicate an increase in HUA condition; values smaller than 1.0 indicate a higher abundance in Mock condition. NA, not applicable, when values in Mock were 0. The first 2 columns illustrate the similarity between experiment types A and B; the information in the third column shows the similarity between HUAwo and Mock samples. Asterisks (*), adj.p < 0.1 (Benjamini–Hochberg procedure) between indicated conditions. The direct comparison of relative histone PTMs abundance is shown in S6 Fig. HMP, histone modification profile; HUA, hydroxyurea and aphidicolin; HUAwo, HUA washout.

https://doi.org/10.1371/journal.pbio.3001377.g006

In detail, we found that H3S10Ph levels were 8-fold higher in transiently arrested (66% of Mock) versus constantly arrested (8% of Mock) embryos. The return to proliferation is also reflected in increased H4K20me1 levels. In proliferating cells, this modification reaches the maximum in G2 and M phases, when PR-Set7 monomethylates new histones incorporated during the previous S phase [49,50]. The observed doubling in H4K20me1 levels (Fig 6B and S6 Fig) indicates that upon inhibitor washout, cells can go through S phase and mitosis. Also, the H3K27me1 mark, which promotes the transcription of H3K36me3 decorated genes in ES cells [10], is recovering, as well as the abundance of acetyl marks on H3K9 and H3K27. The latter marks protect histone H3 from the repressive influence of trimethylation at these lysines, consistent with a transcriptionally more permissive state of chromatin. Indeed, levels of the repressive chromatin modifications H3K9me3, H3K27me2/3, and H4K20me3 were lower in the HUAwo condition, compared to permanently arrested embryos. Whether these changes are causally connected, i.e., acetylation marks replace methyl marks, or occur at independent chromatin regions, is not known. Nevertheless, the readjustments in the chromatin landscapes of embryos released from the G1/S block are in agreement with the observed morphological improvement.

Overall, these results document that HUA-arrested embryos can reinitiate cell proliferation, when the inhibitors are washed out at late gastrula. In contrast to embryos, treated continuously with the cell cycle inhibitors, the HUAwo embryos show a higher vitality with improved tissue formation and gene expression. These changes are accompanied by a partial restoration of their normal histone modification landscape. Taken together, the results of the inhibitor washout experiment support the hypothesis that cell proliferation has an impact on the development of stage-specific HMPs during early embryogenesis of Xenopus.

Ethics statement

Protocols of Xenopus care and experimental use are approved by the Government of Oberbayern (approval number: ROB-55.2-2532.Vet_03-17-102).

Embryos handling and HUA treatment

X. laevis eggs were collected, in vitro fertilized, and handled as described in “Early Development of Xenopus Laevis: A Laboratory Manual” [72]. Embryos were staged accordingly to Nieukoop and Faber [73]. When they reached the desired stage, the embryos were immersed into a solution with 2 DNA synthesis inhibitors (HUA): 20 mM hydroxyurea and 150 μM aphidicolin (made from a frozen stock at 10 mg/ml in DMSO) in 0.1x MBS solution. Hydroxyurea blocks ribonucleotide diphosphate reductase, an enzyme that catalyzes the reductive conversion of ribonucleotides to deoxyribonucleotides, a crucial step in the biosynthesis of DNA. Aphidicolin blocks eukaryotic DNA Pol-alpha. To establish a full and irreversible cell cycle block, HUA treatment takes 4 to 6 hours, if kept in the solution continuously [33]. Within this time period, embryos develop from NF10.5 to NF13, if cultivated at 16C. Controls in all cases were embryos from the same batch that were treated identically, except they were pipetted into 2% DMSO in 0.1x MBS solution (Mock). HUA- and Mock-treated embryos were kept continuously in the HUA and Mock solutions, respectively, while in the HUAwo condition, the embryos were transiently incubated in the HUA from NF10.5 until NF13, and then moved in the Mock until the end point of the analysis, NF32.

RNA in situ hybridization and immunocytochemistry

Whole-mount RNA in situ hybridization was performed as described in “Early Development of Xenopus Laevis: A Laboratory Manual” [72]. For immunocytochemistry, anti-H3S10Ph antibody (1:500, Active Motif, Carlsbad, CA, USA) and anti-mouse alkaline phosphatase–conjugated (1:2,000, Chemicon/Thermo Fisher Scientific, Waltham, MA, USA) secondary antibody were used. Embryos were photographed with a Leica M205FA stereomicroscope. Signal from H3S10Ph+ cells was counted using ImageJ software. For confocal imaging, anti-H3S10Ph antibody (1:500, Active Motif), anti-x-β-catenin antibody (1:100, Elizabeth Kremmer), and DAPI were used. Embryos were photographed with a Leica TCS SP5II confocal microscope, using Z-stack montage function. Z-stack depth, 30 μm.

Nuclear histone extraction

Around 50 to 200 embryos developed to desired stages (NF13, NF18, NF25 and NF32) were harvested and washed with 110 mM KCl, 50 mM Tris/HCl (pH 7.4 at 23 uC), 5 mM MgCl2, 0.1 mM spermine, 0.1 mM EDTA, 2 mM DTT, 0.4 mM PMSF, and 10 mM Na-butyrate. Nuclei of the embryos were prepared by centrifugation with 2,600xg, 10 minutes (3 to 18, Sigma) after homogenization by a 5-ml glass–glass douncer (Braun, Melsungen). The nuclear pellets of Xenopus embryos were resuspended in 1 ml 0.4 M HCl, incubated on a rotating wheel overnight and dialysed against 3 l of 0.1 M acetic acid/1 mM DTT. The dialysed histone solution was vacuum dried in a Concentrator Plus (Eppendorf) and stored at −20°C. Each developmental stage in experiment type A is represented by 3 biological replicates; one biological replicate for experiment type B. Each biological replicate derives from a different mating pair.

Histone acid extraction and sample preparation

The pellet from the nuclear histone extraction was dissolved in an appropriate amount of Laemmli Buffer to reach 1.37e6 nuclei/μl in each sample. A volume of 15 μL were loaded on an 8% to 16% gradient SDS-PAGE gel (SERVA Lot V140115-1) and stained with Coomassie Blue to visualize the histone bands. Histone bands were excised and propionylated as described before [74]. Propionylation blocks all endogenously unmodified and monomethylated lysine residues from being cleaved in the subsequent trypsin digest, thereby creating an optimized peptide pool for mass spectrometry analysis. Due to this step, naturally unmodified lysine residues are labeled as “p,” tryptic peptides which have no modification states indicated as “un.” Arginine10-labeled forms of the PTMs of interest (500 fmols/peptide, JPT Peptide Technologies, Berlin, Germany) were added to the gel samples together with trypsin. Digested peptides were sequentially desalted using C18 Stagetips (3M Empore, St Paul, MN, USA) and porous carbon material (TopTip Carbon TT2CAR (Glygen Corporation, Columbia, MD, USA)) as described elsewhere [75] and resuspended in 15 μl of 0.1% FA.

Mass spectrometry analysis with scheduled PRM method

To study histone PTMs during Xenopus development, we analyzed the absolute and relative abundance of histones from HUA treated and control cohorts of embryos from the 4 developmental stages using LC–MS (S7 Fig).

As an internal and intersample control, a library consisting of isotopically labeled peptides (R10s) was used (S2 Table) to normalize for ionization differences between peptides. The R10 peptides were mixed in the library at equimolar concentration, and the mix was added to each analyzed sample. In total, the library consists of 64 peptides, 58 of them represent different histone H3 modification states, 6—histone H4 modification states. Considering only confirmed methylation and acetylation modifications on H3 and H4 histone tails, the library covers 76% of the histone H3 and 50% of the histone H4 modification states.

In order to identify and measure the abundance of the histone PTMs, we used a parallel reaction monitoring (PRM) method [76]. This method allows to isolate a target peptide ion based on the mass and retention time window, fragment it, and analyze the masses of all fragment ions simultaneously. Therefore, we could distinguish and identify peaks of isobaric peptides.

Peptide mixtures (5 μL) were subjected to nanoRP-LC-MS/MS analysis on an Ultimate 3000 nano chromatography system (Thermo Fisher Scientific, Dreieich Germany) coupled to a QExactive HF mass spectrometer (Thermo Fisher Scientific). The samples were directly injected in 0.1% formic acid into the separating column (150 × 0.075 mm, in house packed with ReprosilAQ-C18, Dr. Maisch GmbH, 2.4 μm) at a flow rate of 300 nL/min. The peptides were separated by a linear gradient from 3% ACN to 40% ACN in 90 minutes. The outlet of the column served as electrospray ionization emitter to transfer the peptide ions directly into the mass spectrometer. The QExactive HF was operated in a scheduled PRM mode to identify and quantify specific fragment ions of N-terminal peptides histone proteins. In this mode, the mass spectrometer automatically switches between one survey scan and 9 MS/MS acquisitions of the m/z values described in the inclusion list containing the precursor ions, modifications, and fragmentation conditions (S3 Table). Survey full scan MS spectra (from m/z 270 to 730) were acquired with resolution 60,000 at m/z 400 (AGC target of 3 × 10⁶). PRM spectra were acquired with resolution 30,000 to a target value of 2 × 10⁵, maximum IT 60 ms, and isolation window 0.7 m/z and fragmented at 27% or 30% normalized collision energy. Typical mass spectrometric conditions were spray voltage, 1.5 kV; no sheath and auxiliary gas flow; and heated capillary temperature, 250°C.

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository [77] with the data set identifier PXD021803.

Histone PTM quantification

Data analysis was performed with the Skyline (version 3.7) [78] by using doubly and triply charged peptide masses for extracted ion chromatograms (XICs). Selection of respective peaks was identified based on the retention time and fragmentation spectra of the spiked in heavy-labeled peptides. Integrated peak values (total area MS1) were exported as.csv file for further calculations. Total area MS1 from endogenous peptides was normalized to the respective area of heavy-labeled peptides. The sum of all normalized total area MS1 values of the same isotopically modified peptide in one sample resembled the amount of total peptide. The relative abundance of an observed modified peptide was calculated as percentage of the overall peptide.

RNA extraction and qPCR sample preparation

Total RNA of 10 embryos was extracted using Trizol (Ambion/Thermo Fisher Scientific, Dreieich, Germany) and phenol/chloroform. The RNA was precipitated with 70% isopropanol and cleaned using the RNeasy Cleanup Kit (Qiagen GmbH, Quiagen Strasse 1, D-40724 Hilden Germany) including DNAse I-on-column digestion. For qPCR analysis, 1 μg of total RNA was transcribed with the DyNAmo cDNA Synthesis Kit (Thermo Fisher Scientific). For qPCR, 5 to 20 ng cDNA was mixed with the Fast SYBR Green Master mix (Applied Biosystems/Thermo Fisher Scientific, Dreieich, Germany) and amplified with a LightCycler (Roche Diagnostics, Mannheim, Germany). Primer sequences are given in S4 Table.

Statistical analysis and data visualization

For embryonic quantitative analysis (morphological phenotype, qRT/PCR, quantitative MS analysis), SEMs are displayed. Statistical analysis was performed using two-tailed, paired Student t test. In case of multiple testing, the Benjamini–Hochberg FDR procedure was performed—indicated as “adj.p.” For boxplots in Figs 2B and 5A, “ggplot2” R package was used. PCA in S3A Fig was performed using R, without scaling. MA-plot, volcano plot, and boxplot in S3B–S3D Fig were generated using GraphPad Prism v9.0.1.

Heatmap generation

Mass spectrometry intensity values of the spiketides were log2 transformed, and missing values imputed using 2 nearest neighbors (library “knnImputation” in R). Endogenous peptides intensities were log2 transformed after adding 1 to all values. For heatmap display in Fig 4A, the normalized log ratios were quantile normalized across all samples and subsequently averaged per condition (n = 3 biological replicates per stage per condition). The values were scaled row-wise per peptide and hierarchically clustered using the “complete” method on euclidean distances. For heatmap generation and hierarchical clustering in Fig 6A, values of relative histone PTM abundance were used (n = 1 biological replicate at developmental stage NF32).