Asymmetric localization of Cdx2 mRNA during the first cell-fate decision in early mouse development

Abstract

A longstanding question in mammalian development is whether the divisions that segregate pluripotent progenitor cells for the future embryo from cells that differentiate into extraembryonic structures are asymmetric in cell-fate instructions. The transcription factor Cdx2 plays a key role in the first cell-fate decision. Here, using live-embryo imaging, we show that localization of Cdx2 transcripts becomes asymmetric during development, preceding cell lineage segregation. Cdx2 transcripts preferentially localize apically at the late eight-cell stage and become inherited asymmetrically during divisions that set apart pluripotent and differentiating cells. Asymmetric localization depends on a cis element within the coding region of Cdx2 and requires cell polarization as well as intact microtubule and actin cytoskeletons. Failure to enrich Cdx2 transcripts apically results in a significant decrease in the number of pluripotent cells. We discuss how the asymmetric localization and segregation of Cdx2 transcripts could contribute to multiple mechanisms that establish different cell fates in the mouse embryo.

Graphical abstract

Figure 1

Exogenous Cdx2 Transcripts Recapitulate Endogenous Cdx2 Transcript Localization (A) Representative images of endogenous Cdx2 (n = 56 uncompacted and n = 144 compacted), Dab2 (n = 17 compacted), and Gata3 (n = 14 compacted) mRNAs detected by FISH in eight-cell embryos. Scale bar, 10 μm. (B) Quantification of the distribution of Cdx2 transcripts. Polar coordinates for individual cells are expressed as a single point defined by vector PI and angle θ on a polar graph. The PI vector is defined by the center of gravity of the cell and the center of gravity of the FISH signal. This vector shows an angular displacement, θ (the angular coordinate), from the apical-basal (polar) axis. The radial coordinate (PI) can be alternatively expressed along the apical-basal (polar) axis as the value BA. Apical localization of Cdx2 mRNA in compacted blastomeres (n = 27) is represented by the clustering of blue dots around 0°; red dots represent Cdx2 mRNA in individual cells before compaction (n = 20), and black dots represent a Monte Carlo simulation of random localization. (C) Density distributions of values for angles θ for indicated transcripts (measured as described in B for individual cells). (D) Representation of the distribution as the average value (±SEM) of the polar coordinate (BA) for the indicated transcripts. This has a positive value when directed apically and is negative when basal. (E and F) Localization of exogenous labeled RNAs for full-length Cdx2 mRNA (E; n = 52) and control mRNA (F; n = 12) injected into eight-cell embryo cells and imaged 1.5–2 hr after injection. Scale bar, 10 μm. (G) Stills from a representative movie of a compacted eight-cell blastomere injected with labeled Cdx2 mRNA. Time points are presented in hr:min:s format. Scale bar, 20 μm. (H) Dynamics of Cdx2 mRNA localization after cell division of the eight-cell blastomere. A cell outline is presented with a dashed line. Scale bar, 5 μm. See also Figure S1, and Movies S1 and S2.

Figure 2

Cdx2 mRNA Localization at the Eight- to 16-Cell Transition (A and B) Live imaging of exogenous Cy3-labeled Cdx2 mRNA distribution during asymmetric (A) and symmetric (B) divisions. Time points are presented in hr:min:s format. (C and D) FISH detection of endogenous Cdx2 mRNA during the eight- to 16-cell transition in mitotic cells (C) and during telophase (D). Scale bar, 10 μm. See also Figure S2, and Movies S3 and S4.

Figure 3

A Cdx2 mRNA Localization Element Resides in the Last 97 nt of the Coding Sequence (A) Schematic representation of the Cdx2 and fusion constructs along with their controls used for injections. (B–I) Representative images of localization patterns observed for the truncated forms of Cdx2 mRNA. Scale bar, 10 μm. (J) Quantification of the injection experiments and numbers of injected blastomeres. Statistically significant differences compared with the control are marked with asterisks (χ² test, ^∗p < 0.05, ^∗∗∗p < 0.001). (K) Quantification of average (±SEM) apical persistence time of ORF and Δ97 Cdx2 transcripts (t test, p < 0.001). (L and M) Stills from representative movies of labeled Cdx2 ORF RNA (L; n = 15) and Δ97 form (M; n = 14) injected into compacted blastomeres of an eight-cell embryo. Time points (hr:min:s) starting 40 min postinjection are indicated above the panels. Scale bar, 10 μm. See also Figure S3, and Movies S5 and S6.

Figure 4

Localization of Cdx2 mRNA Requires an Intact Cytoskeleton (A and B) Embryos treated with nocodazole (A) and cytochalasin D (B) and with the vehicle control (DMSO). FISH detection of endogenous Cdx2 mRNA in eight- to 16-cell-stage embryos (lower panels in A [18 embryos] and B [12 embryos]). Immunostaining for α-tubulin (upper panels in A; n = 14). F-actin staining with phalloidin (upper panels in B; n = 15). Scale bar, 15 μm. (C) Quantification of the polarization, represented as average BA value (±SEM), in individual cells of embryos treated with nocodazole and cytochalasin D. (D and E) Localization of labeled Cdx2 mRNA following injection of anti-pan-kinesin (D; n = 0/20) or dynein intermediate chain (E; n = 4/7) antibodies. Scale bar, 5 μm. See also Figure S4.

Figure 5

Disruption of Cell Polarity Abolishes Apical Localization of Cdx2 mRNA (A–C) dn-aPkc mRNA along with lineage tracer (GFP mRNA) was injected into one blastomere of two-cell embryos that were then cultured until the late eight-cell stage. (A) FISH detection of endogenous Cdx2 mRNA in control (left panel; n = 51/80, ten embryos) and dn-aPKC injected (right panel; n = 18/56, seven embryos) embryos. Cdx2 mRNA localized apically is indicated by white arrows, and its absence in some blastomeres is shown by red arrows. Scale bar, 5 μm. (B) Quantification of the blastomeres that showed apical Cdx2 mRNA localization in the experiment shown in (A) and (B) (mean value ± SEM; Student’s t test; ^∗p < 0.05, ^∗∗∗p < 0.001). (C) Effect of dn-aPKC (marked with GFP fluorescence, first panel) on localization of aPKC (second panel) and actin (third panel). Scale bar, 15 μm. (D and E) Effect of aPKC inhibition with Gö6983 in compacted eight-cell blastomeres on Cdx2 mRNA localization (D; n = 8 embryos) and on actin (E, upper panel; n = 6 embryos) and aPKC (E, lower panel; n = 6 embryos) localization. Scale bar, 15 μm. (F and G) Effect of premature compaction of four-cell embryos resulting from activation of PKCs with DiC8 on mRNA localization (F; n = 6 embryos), and on actin (G, upper panel; n = 6 embryos) and aPKC (G, lower panel; n = 6 embryos) localization. Scale bar, 15 μm. See also Figure S5.

Figure 6

Mislocalization of Cdx2 mRNA in the Eight-Cell Embryo Results in CDX2 Protein in Inside Cells and a Decrease in Epiblast Cell Number (A) Coinjection of excess of RNA containing competitor Cdx2 97 nt localization signal (97nt comp) with labeled Cdx2 mRNA abolishes apical localization of Cdx2 mRNA, which now localizes to basolateral regions. (B) Quantification of cells showing apical localization of labeled Cdx2 mRNA upon coinjection of 97nt comp (competitor; n = 56) or control unrelated RNA of similar size (n = 41); the difference is statistically significant (χ² test, p < 0.001). (C–F) All cells of the four-cell embryo were injected with control or 97nt comp RNA and allowed to develop until the early (C) or mature blastocyst (D–F) stage. (C) Representative images of embryos stained for Cdx2 (green) and actin (red). Yellow arrow marks an inside cell positive for Cdx2 in the embryo injected with 97nt comp RNA (n = 10/13); control injected embryos do not express Cdx2 in inside cells (n = 0/14). (D and E) Examples of embryos injected with control (D) and 97nt comp (E) RNA stained for CDX2 (green), SOX17 (red), or NANOG (white). (F) Quantification of the percentage (±SEM) of cells positive for Cdx2, Sox17, or Nanog in blastocyst (∼96-cell stage). (G–L) Single inside (G, J, and L; n = 17) or outside (H; n = 15) cells of a 16-32 cell embryos injected with Cdx2 mRNA and DxRed (G, H, and J) or with DxRed alone (I; n = 19). Progeny of inside cells injected with Cdx2 mRNA shows nuclear CDX2 at blastocyst stage (H, arrow; green plus red = yellow). Scale bar, 10 μm. (K) Summary of injection data (Fisher’s exact test, ^∗∗∗p < 0.001). (L) Time-lapse series of an embryo with one inside cell injected with Cdx2 mRNA and DxRed. See also Figure S6.

Figure 7

Working Model for How the Localization and Partitioning of Cdx2 Transcripts together with Differential Transcription Lead to Lineage Segregation (Top) Symmetric cell division generates two outside cells (green) that inherit similar amounts of Cdx2 transcripts. Asymmetric division generates an outside daughter that inherits more Cdx2 transcripts (green) than the inside daughter (yellow). (Bottom) Cell polarization leads to the asymmetric localization of Cdx2 transcripts and their differential inheritance, contributing to elevated levels of CDX2 in outside cells (as we show here). This CDX2 asymmetry is then reinforced by a transcriptional response to TEAD4 and the nuclear localization of YAP in outside cells (Yagi et al., 2007; Nishioka et al., 2008, 2009). In turn, the elevated CDX2 level reinforces cell polarization (Jedrusik et al., 2008) and represses the expression of pluripotency factors (Niwa et al., 2005) to promote differentiation into the TE lineage. In contrast, the reduced CDX2 level in inside cells is a consequence of (1) reduced inheritance of Cdx2 transcripts, (2) prevention of Cdx2 transcription as a result of the cytoplasmic retention of YAP, and (3) the repressive effects of OCT4 and NANOG that promote pluripotent fate.

Figure S1

Localization of Endogenous and Exogenous Cdx2 mRNA in Eight-Cell-Stage Embryos, Related to Figure 1 (A) Quantification of blastomeres showing apical localization of endogenous Cdx2 mRNA by FISH in eight-cell-stage embryos, using visual observation and using the image analysis software. In the latter method, blastomeres were categorized based on the angle (θ) between the basal-apical axis of the blastomere and the polarization vector of the FISH staining as: showing apical localization for θ ≤ 45 deg, ambiguous for 45 < θ < 90 deg, and not showing apical localization for θ > = 90 deg. Note, the two methods being equally good in detecting apically localized Cdx2 transcripts. Examples of the three categories are shown on the right. (B and C) Fluorescence intensity plots along the apical-basal axis derived for top, intermediate, and middle focal planes from confocal images of full-length Cdx2 mRNA (B) and control RNA (C), injected into compacted eight-cell-stage blastomeres. Notice the prevalent peak at the position corresponding the apical side of blastomere. For the most cortical plane the cell curvature leads to the appearance of two peaks one at the projected basal side and one at the apical side. (D) Coinjection of full-length wt Cdx2 mRNA and RNA transcribed from the pBluescript vector, labeled with two different fluorescent dyes into compacted eight-cell blastomeres results in partitioning of the two RNAs similarly to single injections (see Figure 1). Scale bar, 5 μm. (E) Injection of full-length wt Cdx2 mRNA to precompacted eight-cell blastomere does not result in the apical localization of the transcripts.

Figure S2

Inheritance of Cdx2 mRNA during Eight- to 16-Cell Division, Related to Figure 2 (A and B) Time-lapse imaging of asymmetric (A) and symmetric divisions (B). Time-points (in hh:mm:ss format) were as indicated above the panels. Upper row shows the images taken in the red channel that results from the fluorescence of the Cy3-labeled Cdx2 mRNA. Lower panels represent the merge of the DIC and red channels. Note that outside cells inherits majority of the labeled mRNA during differentiative cell division, while both cells arising from conservative division inherit comparable amounts of the RNA. Cell boundaries highlighted in dashed lines. Scale bar, 10 μm. (C and D) Localization of Cdx2 mRNA detected by FISH (white) in the eight-to-16 transiting cell. Four different focal planes are shown. DNA (blue) is stained with DAPI. (E) Relative amounts of Cdx2 and Nanog mRNAs measured by qRT-PCR in outside and inside cells isolated after the eight-to-16 cell division. Bars represent the average fold enrichment of particular mRNA in outside cells (14 samples of three to four cells) as compared with inside cells (nine samples of three to four cells).

Figure S3

Localization Element of Cdx2 mRNA, Related to Figure 3 (A) Fluorescently labeled RNA comprising 362 bases from 3′ of Cdx2 ORF and the whole 3′UTR localizes apically upon injection to compacted eight-cell blastomere (n = 12/18). (B and C) RNA secondary structure prediction of the (B) WT and (C) mutated form of the 97 bases element from the 3′ end of the Cdx2 coding sequence (plus 4 additional bases after stop codon); arrow head and arrow mark the 5′ and 3′ end of RNA, respectively; base-pair probabilities are color coded. Runs of purins in WT version are boxed. (D) The 97 bases long 3′ terminal part of Cdx2 coding sequence (including codon stop) is highly conserved in live-bearing, especially placental, mammals but not in monotremes or lower vertebrates and invertebrates.

Figure S4

Effect of Cytoskeleton Disruption on Cdx2 mRNA Expression and aPKC Localization, Related to Figure 4 (A) Quantitative RT- PCR was performed to embryos treated with 30 mM Nocodazole for 30 min and to the DMSO vehicle control group. Normalization was done against the gene Chuk. The Fold change between the Nocodazole treated group and the untreated group is shown. There is no significant difference in Cdx2 transcripts level between the treated and the untreated group (Pair Wise Fixed Reallocation Randomization Test, P(H1) = 0.12). Data represent the average of two biological replicates. (B and C) Embryos treated with Nocodazole (B) and Cytochalasin D (C) and with the vehicle control (DMSO). F-actin staining with phalloidin (upper panels in (B) and immunostaining for α-tubulin (upper panels in (C). Immunostaining for aPKC (lower panels in (B) and (C); Note - the differences in aPKC staining of control embryos in (B) and (C) result from different fixation protocols used in the two experiments. In panel (C), methanol fixation was used to preserve microtubules and this resulted in non-specific staining of the zona. Scale bar, 15 μm.

Figure S5

Overexpression of dn-aPkc and Floped, Related to Figure 5 (A) Effect of overexpression of dominant-negative (dn) form of aPKC on microtubule cytoskeleton. Both, injected and un-injected eight-cell stage blastomeres are shown (upper panels). Blow up of the dn-aPKC expressing blastomeres. Notice, disruption of the aPKC staining but not difference in alpha-tubulin staining (lower panels). (B) Effect of overexpression of Floped protein on localization of aPKC and Cdx2 mRNA in compacted eight-cell stage embryos.

Figure S6

Overexpression of Cdx2 in the Inside Cells of a Morula-Stage Embryo, Related to Figure 6 (A and B) Time-lapse series of a typical embryo with one inside cell injected with either Cdx2 mRNA and dextran red – DxRed – (A) or DxRed alone (B). Embryos are expressing membrane form of GFP (green) to visualize cell boundaries to track cell behavior.