Nguyen T et al. (2022), Differential nuclear import sets the timing of ...

XB-ART-60535

Nat Commun 2022 Oct 06;131:5887. doi: 10.1038/s41467-022-33429-z.

Show Gene links Show Anatomy links

Differential nuclear import sets the timing of protein access to the embryonic genome.

Nguyen T , Costa EJ , Deibert T , Reyes J , Keber FC , Tomschik M , Stadlmeier M , Gupta M , Kumar CK , Cruz ER , Amodeo A , Gatlin JC , Wühr M .

???displayArticle.abstract???
The development of a fertilized egg to an embryo requires the proper temporal control of gene expression. During cell differentiation, timing is often controlled via cascades of transcription factors (TFs). However, in early development, transcription is often inactive, and many TF levels stay constant, suggesting that alternative mechanisms govern the observed rapid and ordered onset of gene expression. Here, we find that in early embryonic development access of maternally deposited nuclear proteins to the genome is temporally ordered via importin affinities, thereby timing the expression of downstream targets. We quantify changes in the nuclear proteome during early development and find that nuclear proteins, such as TFs and RNA polymerases, enter the nucleus sequentially. Moreover, we find that the timing of nuclear proteins' access to the genome corresponds to the timing of downstream gene activation. We show that the affinity of proteins to importin is a major determinant in the timing of protein entry into embryonic nuclei. Thus, we propose a mechanism by which embryos encode the timing of gene expression in early development via biochemical affinities. This process could be critical for embryos to organize themselves before deploying the regulatory cascades that control cell identities.

???displayArticle.pubmedLink??? 36202846
???displayArticle.link??? Nat Commun
???displayArticle.grants??? [+]

Species referenced: Xenopus laevis
Genes referenced: aurkb birc5 birc5l cdca8 creb1 cse1l ctnnb1 dbf4b gtf2b gtf2e2 gtf2h1 gtf3a h2bc21 incenp ipo11 ipo4 ipo5 ipo7 ipo9 kit kpna2 kpna3 kpna4 kpna6 kpna7 kpnb1 myc orc1 orc2 orc3 orc4 orc5 orc6 phf5a polr2a recql4 smad2 smarca4 ticrr tnpo1 tnpo2 tnpo3 trna xpo1 xpo4 xpo5 xpo6 xpo7 xpot ybx2 yy1 znf250
GO keywords: nucleocytoplasmic transport [+]

???attribute.lit??? ???displayArticles.show???

	Figure 1. a Immunofluorescence images (anti-tubulin) of early frog development show drastic changes in nuclear morphology. The large oocyte (~1.2 mm diameter) contains a proportionally large nucleus (~400 μm). After fertilization, the nucleus is only ~30 μm in diameter. Around zygotic genome activation (ZGA), the embryo contains ~4000 cells with ~18 μm diameter nuclei. The represented images were independently repeated multiple times and observed with similar results. b The nucleocytoplasmic volume ratio (NCV-ratio) in early frog development, quantified based on micrographs. The NCV-ratio drops ~10,000-fold from oocyte to fertilized embryo before increasing exponentially during early cleavage stages. After ZGA, the NCV-ratio gradually increases, approaching the value observed in the oocyte. Solid dots are the mean NCV ratios. Error bars indicate standard deviation. Volume quantification and sample sizes for each time point are provided in Source Data. Embryos developed at 16 °C. c Quantification of the fraction of the nuclear proteome replaced by new protein species from oocyte to ZGA. Despite the drastically increasing NCV-ratio from the fertilized egg to ZGA, only ~3% of the nuclear proteome is replaced by newly synthesized canonical nuclear proteins (the main contributors are high abundant histones, which increase ~2-fold).
	Immunofluorescence images (anti-tubulin) of a large Xenopus oocyte (~1.2mm diameter) which contains a proportionally large nucleus (~400m).
	Immunofluorescence images (anti-tubulin) of a large Xenopus oocyte (~1.2mm diameter) soon after fertilization, NF stage 1, where the nucleus has condensed to ~30um in diameter.
	Immunofluorescence images (anti-tubulin) of a large Xenopus oocyte (~1.2mm diameter) (c. between NF stage 8 and stage 9) at the time point of zygotic genome activation (ZGA), where the embryo contains ~4000 cells with ~18um diameter nuclei.
	Fig. 2: Quantification of nucleocytoplasmic partitioning during early development. a. Assay to quantify nucleocytoplasmic partitioning via multiplexed proteomics. Embryonic lysate, developed at 16C, was collected at various stages: nuclei were enriched by filtration through membranes of defined pore sizes. Shown are example epifluorescence images of lysate, supernatant, and flow-through stained with Hoechst (blue) and MitoTracker (red). Supernatant and flow-through fractions were digested into tryptic peptides, labeled with isobaric tags, and subjected to accurate multiplexed proteomic analysis. Xenopus illustrations Natalya Zahn (2022) [86,95]. b Nucleocytoplasmic partitioning quantification for three examples. Dots indicate measured fraction of each protein in nuclei. Solid lines indicate fit sigmoids. Blue dashed line indicates median fit from all nuclear proteins. c. Quantification of time-dependent nuclear fraction for ~2k nuclear proteins and ~600 transcription factors (TFs) [20,71,96]. Solid line indicate the median and shaded area 50% spread. d, e Measurements for multisubunit protein complexes. Solid line indicates the median and shaded area 50% spread. d. DNA repair complexes enter embryonic nuclei at various times, yet their subunits enter simultaneously. Shown are nuclear entry quantifications for homologous recombination (HR), the nonhomologous DNA end-joining repair (NHEJ), and the Fanconi anaemia (FA). An apparent time delay of nuclear entry for some DNA repair complexes explains previous observations that early embryos bypass DNA repair to accommodate fast cell divisions [30,31,32]. Oocyte nuclear fractions are shown as standard box plots. e Change in nucleocytoplasmic partitioning for nuclear pore complex (NPC) subunits. Top: Our nuclear fraction measurement for NPC proteins shows rapid incorporation onto embryos exponentially increasing nuclear surface. Middle: Quantification of NPC nuclear fraction via MS agrees well with immunofluorescence-quantified nuclear surface change. The x-axis is the relative nuclear surface measurement to that of embryos at 22.5h post-fertilization. Bars indicate standard errors. Bottom: Illustration of dynamic localization of NPC proteins in the oocyte versus the embryo. Oocyte stores NPC subunits in endoplasmic reticulum membranes embedded with pore complexes, called annulate lamellae [34,35].
	Fig. 3: In early development, proteins enter nuclei sequentially correlating with the onset of their nuclear functions. a–c Embryos developed at 16 °C. a Left: Protein abundance dynamics of previously reported ZGA key regulatory factors change little from fertilization to the ZGA [15,16]. Right: These regulators are among the earliest proteins to titrate into embryonic nuclei. Shown are fits of individual ZGA regulators and the median fit of the core histones (H2a, H2b, H3, and H4) and their isoforms with 50% spread. b RNA polymerase III (Pol III) and II (Pol II) enter nuclei at different times of development, which corresponds to the respective appearances of their first downstream transcripts. Left: Proteomics data show that Pol III subunits titrate into nuclei before Pol II subunits. Shown are median and 50% spread of unique subunits. Right: The first tRNA transcripts, transcribed by Pol III, and snRNA, transcribed by Pol II, show corresponding timing to the nuclear entry of their corresponding polymerases. The RNA measurements are quantified from Newport and Kirschner’s RNA gel of newly synthesized transcripts [13]. Error bars indicate standard deviation of five snRNA species (snRNA U1, U2, U4, U5, U6). c Timing of the nuclear import of TFs in the mesendoderm gene regulatory network corresponds to the timing of their activation. Left: Transcriptional activation order in early Xenopus embryos (Adapted drawing from Charney et al. [24] and [43,44]). Middle: MS-quantification of nuclear import in early development for these TFs. Right: Scatter plot of rank between the measured nuclear entry Tembryo1/2 by MS and the reported temporal gene regulatory network shows strong agreement (Spearman correlation of 0.87, two-tailed p-value = 0.005).
	Fig. 4: The affinity of proteins to importin contributes significantly to their ordering of nuclear entry in early development. a Estimation of proteome-wide affinity to importin a/B. We quantified changes in protein abundance associated with importin beads among conditions with varying amounts of RanQ69L. Abundance of known importin a/B substrates [53], including histones, decreases with increasing RanQ69L concentration. Large dots represent the median protein fraction of a protein subgroup at each RanQ69L concentration, while small dots represent measurements for individual proteins. We applied a linear fit for each protein with a fixed y-intercept and used the slope to proxy for a protein’s affinity to importin. b Scatter plot of triplicate affinity proxy measurements from experiments outlined in (a). We integrated these measurements to one dimension using cross-validated canonical correlation analysis [49]. c Importin affinity can explain a significant fraction of the timing of nuclear entry in early development. The scatter plot shows T embryo 1/2 versus importin a/B affinity proxy. The observed Pearson correlation suggests that importin affinities can explain >46% of the variance of the timing of nuclear entry or NLS containing proteins in early embryonic development. d Schematic of our proposed model in which the differential affinity of proteins to importin controls the timing of genomic access in embryonic development. A high-affinity protein titrates into the nucleus faster than a low-affinity protein, resulting in the corresponding DNA access of proteins. This ordering could determine the timing for the onset of downstream transcription. e Simulation of the model proposed in (d). We model competitive binding of substrates with varying affinity to a limiting number of importin. The proposed model provides a simple explanation for the timing of protein access to the embryonic genome in early development.
	Supplementary Figure 1: Changes in nucleocytoplasmic volume (NCV) ratio for cells in the frog ovary and early development. a, Left: Immunofluorescence (-tubulin and Yo-Pro-1) of a frog ovary shows cells varying by 3-orders of magnitude in diameter (from ~10m to ~1200um). A zoomed-in section of the white framed box is shown to the right. The represented images were independently repeated multiple times for reproducibility and observed with similar results. Right: Quantification of the NCV-ratio versus cell volume for maturing oocytes and early developing embryos. Compared to the rapidly changing embryos, the oocyte NCV-ratios remain approximately constant (decreasing by 2.3-fold as the cell volume increases ~500-fold (from ~2nL to ~1L)) throughout oogenesis. The slight decrease might come from some cytoplasmic volume being excluded by the forming yolk platelets [1,2]. In contrast, the NCV-ratios increase by ~34,000-fold in developing embryos from a ~1uL fertilized egg to ~30pL cells at 46 hours post-fertilization (16C). (Supplementary Data 5). b, k-means clustering (k = 5) of changes in relative protein abundance of the entire proteome (left) and nuclear proteins (right) from the time series of the oocyte to ZGA embryo measured by multiplexed proteomics. The line thickness scales with the number of proteins within a cluster. While the NCV-ratio drastically increases from the fertilized egg to the ZGA, protein levels change little. c, The detected transcription factors reported being involved in the gene regulatory network of the mesendoderm formation do not change their expression levels between fertilization to the ZGA [3]
	Supplementary Figure 2: Rapid nuclear filtration outperforms nuclear isolation method based on differential sedimentation in quantifying the nucleocytoplasmic (NC) partitioning in early frog embryos. a, The newly developed rapid filtration method can enrich nuclear protein signals from the rest of the embryonic cell. Shown is the histogram of nuclear fraction, quantified with multiplexed proteomics, after the fractionation of ZGA embryos using the nuclear filtration method. b, Nuclear extraction based on differential sedimentation poorly separates nuclear proteins from the proteins of other organelles. Shown is the quantification as in (a) using a commercial nuclear extraction kit (Abcam). c, Receiver operating characteristics (ROC) for the measured nuclear fractions, comparing filtration and sedimentation-based nuclear enrichment methods. We use proteins that were measured nuclear in the oocyte as true positives and proteins that were measured cytoplasmic in the oocytes as false positives4. Measuring NC partitioning via rapid filtration (pink) appears superior to sedimentation (green) in early frog embryos.
	Supplementary Figure 3: The level of reproducibility of the embryonic nuclear proteome experiments on the proteome-wide scale and on individual groups of protein discussed throughout the manuscript. a, Scatter plot of two biological replicates to determine Tembryo1/2 of nuclear entry into embryonic nuclei. We observe a Pearson correlation of 0.81 (two-tailed p-value < 1e-325) with lower reproducibility at late Tembryo1/2. b-i, Correlation plots for highlighted proteins discussed in the manuscript are separated into different panels, corresponding to the associated figures in the main text. Below are the pairs of the presented panel and the corresponding figure in the manuscript: panel (b) – Fig. 2b, panel (c) – Fig. 2e & Supplementary Fig. 3b, panel (d) – Fig. 3a, panel (e) – Fig. 3b & Supplementary Fig. 4, panel (f) - Fig. 3c, panel (g) – Fig. 5c, panel (h) – Supplementary Fig. 4a, panel (i) – Fig. 2d.
	Supplementary Figure 4: Quantifying protein NC partitioning reveals sequential nuclear entry in early embryos. a-d, embryos developed at 16°C. Solid lines indicate median nuclear titration patterns and shaded areas 50% spread. a, Core subunits of the origin of replication complex (ORC) enter the nucleus at different times. Left: Relative protein abundances of ORC subunits stay approximately constant from fertilization to ZGA. Right: Orc6 enters embryonic nuclei later than other components, consistent with previous observations of cytoplasmic function in Drosophila and mammals [5]. While Orc1-6 localize in somatic nuclei, Orc1-5 are cytoplasmic and Orc6 is nuclear in the oocyte [4,6]. In early embryos, Orc1-5 rapidly enter the nuclei, presumably to accommodate for function in Xenopus embryos’ DNA replication process [7]. Bottom: Cartoon model summarizing NC partitioning of ORC core subunits in the oocyte, embryo, and somatic cells. b, Nuclear basket proteins show different NC partitioning in early development than other nuclear pore complex (NPC) proteins. Top: Yet, nuclear basket proteins demonstrate constant expression levels during early development. Bottom: The nuclear basket is mostly nucleus-localized, contrasting NPC proteins being generally equidistributed between nucleus and cytoplasm in the oocyte. Nuclear basket proteins titrate into the nucleus slightly faster than other subunits. Previous observations indicate nuclear baskets are not part of the annulate lamellae but interact with actin filamentous networks connecting NPC to chromatin in oocyte nuclei [8]. c, The CPC complex was equidistributed between the nucleus and cytoplasm in the oocyte. However, CPC components is sequestrated into the embryonic nuclei. The transcription factor beta-catenin (Ctnnb1) is cytoplasmic in the oocyte but is among early nuclear-imported proteins [9,10]. d, Protein complex subunits co-import into nuclei. The standard deviation of nuclear entry time of each complex measured in the embryonic assay (blue) versus those of complexes with shuffled subunits (orange)11 (Wilcoxon-rank two-sided test with N=311 biological independent complexes p- value ~8e-26). Box plots of the range from Q1 to Q3 and 1.5* IQR whiskers are shown.
	Supplementary Figure 5: Nuclear entry of Pol III and its transcription factors during early development. Embryos developed at 16°C. For complexes with multiple subunits, we represent the median nuclear titration pattern by a solid line and a 50% spread by shaded area with the corresponding color. Source data are provided as Supplementary Data 1,2. Left: Relative protein abundance of Pol III and its associated transcription factors, Gtf3a and Gtf3c1-5, stay approximately constant from fertilization to the ZGA. Middle: Although the core Pol III titrates in the nucleus early, the nuclear entry times of its associated transcription factors vary. In our proteomics analysis, Gtf3c1-5 titrate in the nucleus before Gtf3a. Under Pol III, the transcription factors Gtf3c1-5 (and Pol III) are upstream of tRNAs, while Gtf3a (and Pol III) are upstream of 5S rRNA and 7S rRNA. Right: Quantification of the newly transcribed RNA indicates that tRNA is transcribed much earlier than 5S rRNA and 7S rRNA, which corresponds to the nuclear entry order of their upstream transcription factors. Quantification is from Newport and Kirschner’s RNA gel of newly synthesized transcripts [12].
	Supplementary Figure 6: Quantification of proteome-wide affinities to plasmid DNA and their correlation with nuclear entry times in early embryos. a, Quantification of proteome-wide protein affinity to DNA. We exposed Xenopus egg lysate to magnetic beads that were covered with plasmid DNA. The pull-down was collected and subjected to MS quantification of the relative protein signal. The experiment was repeated with different DNA to lysate ratios. The DNA affinities at different ratios were projected on one dimension of the canonical coordinate space with cross-validation to produce a DNA affinity proxy for each identified protein [13]. b, Scatter plot of the projected DNA affinity proxy versus Tembryo1/2. The correlation suggests that plasmid DNA affinity explains 8% of the nuclear import variance in the embryos. For technical reasons, we could not perform these assays with frog DNA. Therefore, this assay does not capture protein affinity to frog- specific DNA sequences. c, Scatter plot of the importin affinity proxy versus Tembryo1/2 for all quantified proteins suggests that importin affinities can explain 29% of the of the nuclear import variance in the embryos. d, The combined DNA affinities and importin affinities projected onto a single dimension result in an improved correlation with nuclear entry time, suggesting that together DNA and importin affinities can explain >35% of the variance of observed timing of nuclear entry in early embryos. Shown is the scatter plot between Tembryo1/2 and the projected affinity proxy.
	Supplementary Figure 7. The expression levels and subcellular localization of nuclear transport receptors observed in the frog embryos and their dynamic changes in early development. Source data are provided as Supplementary Data 1,2,4. a, The absolute abundance of transport receptors in the frog egg. Left: importin Beta family, right importin alpha's (Supplementary Data 4 [4]). b, Relative protein abundance (left) and the subcellular localization (right) of the importin Beta family importins as a function of developmental progression. We observed that the expression levels of importins stay constant throughout development and that the proteins locate preferentially in the cytoplasm throughout early development. The preferential cytoplasmic partition of importins was similarly observed in the frog oocyte [4]. c, Relative protein abundances of importin alpha’s in early frog embryos show that their levels remain approximately constant. d, Relative protein abundance changes of exportins and biportins as a function of developmental progression. Most exportins and biportins remain approximately constant. The exception is exportin 6 (Xpo6), which has actin as the primary substrate. Xpo6 is absent in Xenopus oocytes, which results in nuclear actin localization that supports the physical integrity of the large oocyte nucleus [14,15]. Upon fertilization, Xpo6 expression level increases, and actin is excluded from embryonic nuclei.
	Supplementary Figure 8: Validation of importin affinity assay (Figure 4) using NLS peptides with orthogonally measured KDs [16]. We defined the importin affinity proxy as the free peptide concentration difference between a condition with RanQ69L and a condition without RanQ69L (importin affinity proxy = [free protein+RanQ69L] – [free protein–RanQ69L]). We observed a correlation of 0.90 (R2 = 0.80, two-tailed p-value = 0.001) between our measured importin affinity proxy with the log KDs from Hodel et. al.[16]. The line is a linear fit. Solid dots are the means of importin affinity proxy and error bars indicate standard error (calculated from a set of biological independent replicates, n = 3 controls and 5 RanQ69L added conditions, resulting in 15 biological independent measurements of importin affinity proxy).
	Supplementary Figure 9: Changes of total nuclear volume in embryos and droplet assay; predictions of nuclear concentration for proteins with varying importin affinity in embryos; Scatter plot of nuclear import timing cell-free assay and embryos. a, Total nuclear volume, and nuclear protein flux as functions of time in developing X. laevis embryos. Each dot indicates measurements of total nuclear volume based on immunofluorescence from one embryo. A spline fit (blue curve) provides the functional form of the changes in nuclear volume over time. The time derivative of nuclear volume expansion is the nuclear flux due to nuclear import (orange curve). The maximum volume during this time normalizes the nuclear volume (y-axis left), and the nuclear flux is normalized by the maximum flux (y-axis right). The decline in the total nuclear import rate might be due to the previously reported sequestration of importin α to the cellular membrane17,18. b, Our model predicts the changes in nuclear concentration over increasing embryonic nuclear volume for three proteins with different affinities to importin. The three proteins sequentially reach their maximal nuclear concentration from the highest affinity protein to the lowest affinity protein. After the maximum is reached, a further nuclear volume increase leads to a decline in nuclear concentration. c, Nuclear volume and nuclear flux as functions of time in oil encapsulated cytoplasmic droplets. Similar to (a), the nuclear volume per cell-free droplet is shown to increase with time in the raw imaging data (blue dots) and is fit by a spline function (blue curve). The function’s time derivative is the nuclear flux over time (orange curve). T = 0 is when the cytoplasm is taken off ice, and nuclear formation is initiated. d, Comparing the time measurement of nuclear import in (Tdroplet1/2) from the cell-free assay and the log import time (Tembryo1/2) for the nine TFs show strong agreement (Pearson correlation of 0.77, two- tailed p-value = 0.01).
	Supplementary Figure 10: Import kinetics of NLS-GFP fusions correlate with import kinetics of the NLS-origin proteins. a, We transferred the bioinformatically predicted NLS sequence from proteins assayed in Figure 5, Yy1 (a fast-imported protein ΔTdroplet1/2 = -15.0 min) and the slow-imported proteins Gtf2h1 (ΔTdroplet1/2 = 41.1 min), and Gtf3a (ΔTdroplet1/2 = 33.6 min), to bacterially expressed GFP [19]. b, Experimental procedure. We expressed GFP with the predicted NLS signals from Yy1, Gtf2h1, and Gtf3a to perform nuclear import assays. We added these expressed constructs at the same concentration as the reference SV40-NLS-mCherry into Xenopus egg extracts doped with sperm DNA. We collected samples every 5 minutes and imaged nuclear import with confocal microscopy. c, We monitored the nuclear import kinetic of NLS-GFP and SV40-NLS-mCherry, and fit the data with a sigmoid to extract the time (T1/2). We calculate the import time difference (ΔT) between SV40-NLS- mCherry and the NLS of interest to overcome extract variability. Markers represent the raw measurements (n ≥ 5 replicates), and the box plot shows the spread of measurement data for all the nuclei at each time point. Lines are sigmoid fits. Source and analyzed data are provided as Supplementary Data 7. d, A scatter plot summarizes the result from the import kinetics of GFP with transferred NLS and import kinetics of the corresponding protein. We observe a good correlation (Pearson correlation of 0.94, two-tailed p-value = 0.006) between T1/2 of NLS-GFP constructs measured in the bulk extract and the ΔTdroplet1/2 of proteins measured in the droplet assay. Box plots show the range from the first quartile to the third quartile with the center mark representing the median value. The whiskers represent 1.5*IQR of the distribution.

References [+] :

Abdallah, Ketamine Treatment and Global Brain Connectivity in Major Depression. 2017, Pubmed