Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 16 (8), 1005-21

Venus Trap in the Mouse Embryo Reveals Distinct Molecular Dynamics Underlying Specification of First Embryonic Lineages

Affiliations

Venus Trap in the Mouse Embryo Reveals Distinct Molecular Dynamics Underlying Specification of First Embryonic Lineages

Jens-Erik Dietrich et al. EMBO Rep.

Abstract

Mammalian development begins with the segregation of embryonic and extra-embryonic lineages in the blastocyst. Recent studies revealed cell-to-cell gene expression heterogeneity and dynamic cell rearrangements during mouse blastocyst formation. Thus, mechanistic understanding of lineage specification requires quantitative description of gene expression dynamics at a single-cell resolution in living embryos. However, only a few fluorescent gene expression reporter mice are available and quantitative live image analysis is limited so far. Here, we carried out a fluorescence gene-trap screen and established reporter mice expressing Venus specifically in the first lineages. Lineage tracking, quantitative gene expression and cell position analyses allowed us to build a comprehensive lineage map of mouse pre-implantation development. Our systematic analysis revealed that, contrary to the available models, the timing and mechanism of lineage specification may be distinct between the trophectoderm and the inner cell mass. While expression of our trophectoderm-specific lineage marker is upregulated in outside cells upon asymmetric divisions at 8- and 16-cell stages, the inside-specific upregulation of the inner-cell-mass marker only becomes evident at the 64-cell stage. This study thus provides a framework toward systems-level understanding of embryogenesis marked by high dynamicity and stochastic variability.

Keywords: gene trap; lineage map; live imaging; mouse; pre‐implantation development.

Figures

Figure 1
Figure 1
Design and outcome of the fluorescence gene-trap screen

Design of the fluorescence gene-trap vector. The gene-trap cassette consists of a splice acceptor (SA) followed by stop codons in all three reading frames and an internal ribosomal entry site (IRES). Venus is tagged with a nuclear localization signal (NLS). The essential virus-specific elements include self-inactivating long terminal repeat (LTR) consisting of U5, R and truncated U3 element (ΔU3); woodchuck hepatitis post-transcriptional regulatory element (WPRE); central poly-purine tract (cPPT); Rev-responsive element (RRE); and Rous sarcoma virus promoter (RSV).

Design of the screenings. Denuded 2-cell stage embryos were transduced with the self-inactivating virus for 6–8 h. Blastocysts were first screened for fluorescence, indicating successful trapping of a gene after 3 days of in vitro culture. Positive embryos were then transferred to foster mothers to give rise to adult founders. Germline transmission of fluorescent signal-positive proviral integrants was tested in a fluorescence-based 2nd screen of expanded blastocysts of subsequent generations.

Efficiency of the Venus-trap screen.

Number of integrations identified in the founders positive in the 2nd screen.

Distribution of integration sites among mouse chromosomes. Yellow triangles mark identified integration sites positive for Venus expression at the blastocyst stage (n = 22 of 23).

None
The maps of the proviral integration sites of the VET mouse lines

A–V Maps showing the genomic structures of the trapped genes (black), the sequence identified by LM- or TAIL-PCR representing the proviral integration site (red) and the sequence identified by 5′RACE representing the trapped exons (green). The maps were established using Ensembl BLAT/BLASTN search against the mouse (Mus musculus) DNA databases of the GRCm38.p3 assembly (release 78) with standard settings. LM- and TAIL-PCR sequences were BLAT-searched against genomic sequences, and 5′RACE sequences were BLASTN- and BLAT-searched against cDNAs (transcripts/splice variants) or genomic sequences, respectively. Gene structures show a representative splice variant of the Ensembl/Havana merge transcripts or Havana transcripts.

None
Venus expression pattern of the established VET mouse lines

A–G VET lines with non-ubiquitous Venus expression during the pre-implantation stage. Brightfield and confocal live images of the developing embryos at E1.5 (2- to 4-cell stage), E2.5 (8- to 16-cell stage morula), E3.5 and E4.5 blastocyst derived from (A) PolgGt(Venus)VET25A, (B) Cd2apGt(Venus)VET36C, (C) ItpkcGt(Venus)VET50A, (D) VET53C line, (E) Ctnna1Gt(Venus)fVET3A, (F) Rbfox2Gt(Venus)fVET6C and (G) Cers6Gt(Venus)fVET12A. Arrowheads mark cells of the presumptive PrE positive for fluorescent marker expression. Scale bars are 50 μm. All fluorescent images are representative single confocal sections from n ≥ 3 embryos (n ≥ 2 experiments).

Figure 2
Figure 2
Spatio-temporal characterization of gene-trap mouse embryos

A The trapped gene and the Venus expression pattern of the established VET mouse lines, summarized as a result of n ≥ 3 independent experiments for each line. *The trapped gene for VET53A is tentatively assigned as 2610305D13Rik (see Materials and Methods for details).

B, C Brightfield and confocal live images of the developing embryos at E2.5 (8- to 16-cell stage morula), E3.5 and blastocyst cultured beyond E4.5 derived from the (B) TE-specific line Tmem50bGt and (C) ICM-specific line 2610305D13RikGt. Scale bars are 50 μm. All fluorescent images are single confocal sections, representative for n ≥ 3 independent experiments.

None
Venus-trap screen identified a novel gene essential for the pre-implantation development

Fluorescence intensity is correlated with the genotype of VET embryos. (1) Ctnna1+/+ (n = 6 embryos), Ctnna1Gt/+ (n = 10 embryos) and Ctnna1Gt/Gt embryos (n = 7 embryos) at E3.5 shown as brightfield images (upper row panels) and for Venus expression (lower panels) (n = 2 experiments). (2) Box plots depicting Venus fluorescence intensity in individual nuclei of Ctnna1+/+ (n = 5 nuclei from n = 1 embryo), Ctnna1Gt/+ (n = 10 nuclei from n = 3 embryos) and Ctnna1Gt/Gt (n = 10 nuclei from n = 2 embryos) embryos (n = 2 experiments). Kruskal–Wallis H-test, = 3 × 10−5. (3) Supt6+/+ (n = 11 embryos), Supt6Gt/+ (n = 29 embryos) and Supt6Gt/Gt embryos (n = 15 embryos) at E3.5 shown as brightfield images (upper panels) and Venus expression (lower panels) (n = 4 experiments). (4) Box plots depicting Venus fluorescence intensity in individual nuclei of Supt6+/+(n = 10 nuclei from n = 3 embryos), Supt6Gt/+ (n = 10 nuclei from n = 2 embryos) and Supt6Gt/Gt (n = 10 nuclei from n = 2 embryos) embryos. Kruskal–Wallis H-test, = 2 × 10−6.

Venus-trap insertions disrupt the formation and function of the protein of the trapped gene. (1) Immunostaining of CTNNA1 protein in Ctnna1+/+ (n = 20 embryos), Ctnna1Gt/+ (n = 17 embryos) and Ctnna1Gt/Gt (n = 8 embryos) embryos at the E4.5 blastocyst stage (n = 4 experiments). White arrowheads indicate accumulation of the CTNNA1 signal on cell–cell adhesions that was lost in Ctnna1Gt/Gt embryos. (2) Box plots indicating the fluorescence intensity at cell–cell contacts of Ctnna1+/+ (n = 10 cell–cell contacts from n = 2 embryos), Ctnna1Gt/+ (n = 15 contacts from n = 3 embryos) and Ctnna1Gt/Gt embryos (n = 11 contacts from n = 3 embryos). Kruskal–Wallis H-test, = 1 × 10−5. (3) SUPT6 distribution in Supt6+/+ (n = 4 embryos), Supt6Gt/+ (n = 10 embryos) and Supt6Gt/Gt (n = 2 embryos) embryos at E3.5 (n = 1 experiment). (4) Box plots indicating the fluorescence intensity in individual nuclei of Supt6+/+ (n = 4 nuclei from n = 1 embryo), Supt6Gt/+ (n = 8 nuclei from n = 1 embryo) and Supt6Gt/Gt (n = 6 nuclei from n = 1 embryo) embryos at E3.5. Kruskal–Wallis H-test, = 0.003.

Data information: Venus expression images in (A) are sum intensity projections and in (B) are single confocal sections. Venus expression level is shown using a fire lookup table (LUT). Scale bars are 50 μm. Fluorescence intensities in A2, A4, B2 and B4 are normalized to the highest level of fluorescence. Statistical significance is indicated as * (< 0.05) or ** (< 0.001).
Figure 3
Figure 3
ICM and TE reporter expressions are controlled under the emerging lineage-specific genes

Brightfield and confocal live images of Venus expression in Tmem50bGt in Tead4+/+ or Tead4−/− background at E3.5 (n = 109 embryos in n = 3 experiments).

Box plots depicting Tmem50bGt Venus fluorescence intensities in individual nuclei of Tead4+/+ (n = 16 nuclei from n = 4 embryos), Tead4+/− (n = 16 nuclei from n = 4 embryos) or Tead4−/− embryos (n = 19 nuclei from n = 5 embryos; Kruskal–Wallis H-test, = 10−8).

Brightfield and confocal live images of Venus expression in individual nuclei of Tmem50bGt in a Cdx2+/+ or Cdx2−/− background at E3.5 (n = 57 embryos in n = 5 experiments).

Box plots depicting Tmem50bGt Venus intensities in individual nuclei of Cdx2+/+ (n = 9 nuclei from n = 2 embryos), Cdx2+/− (n = 10 nuclei from n = 2 embryos) or Cdx2−/− embryos (n = 22 nuclei from n = 4 embryos; Kruskal–Wallis H-test, = 2 × 10−6).

Brightfield and confocal live images of Venus expression in 2610305D13RikGt in a Tead4+/+ or Tead4−/− background at E3.5 (n = 68 embryos in n = 3 experiments).

Box plots depicting 2610305D13RikGt Venus intensities in individual nuclei of TE cells in Tead4+/+ (n = 17 nuclei from n = 3 embryos), Tead4+/− (n = 15 nuclei from n = 4 embryos) or Tead4−/− embryos (n = 16 nuclei from n = 5 embryos; Kruskal–Wallis H-test, = 5 × 10−5).

Brightfield and confocal live images of Venus expression in 2610305D13RikGt in a Cdx2+/+ or Cdx2−/− background at E3.5 (n = 95 embryos in n = 4 experiments).

Box plots depicting 2610305D13RikGt Venus intensities in individual nuclei of TE cells in Cdx2+/+ (n = 15 nuclei from n = 4 embryos), Cdx2+/− (n = 16 nuclei from n = 4 embryos) or Cdx2−/− embryos (n = 14 nuclei from n = 4 embryos; Kruskal–Wallis H-test, = 0.01).

Data information: Venus intensities in (B, D, F, H) are normalized to the highest level of expression. Statistical significance is indicated as * (< 0.05) or ** (< 0.001). Scale bars are 50 μm. Venus expression images are sum intensity projections with a fire LUT. A fire LUT from 0 to 255 gray scales is shown in (A).
Figure 4
Figure 4
Four-dimensional live imaging of the pre-implantation development of ICM and TE lineage reporter mice

Representative images of a Tmem50bGt-Venus × R26-H2B-mCherry × mT embryo developing from the 8- to 64-cell stage (n = 27 embryos in n = 3 experiments). Upper row panels show Venus expression, middle panels mT and H2B-mCherry expression, and lower panels all fluorescent channels merged.

Representative images of a 2610305D13RikGt-Venus × R26-H2B-mCherry × mT embryo developing from the 16- to 128-cell stage (n = 27 embryos in n = 3 experiments). Upper panels show Venus expression, middle panels mT and H2B-mCherry expression, and lower panels all fluorescent channels merged.

Data information: Scale bars are 50 μm. All images are single confocal sections and representative for n ≥ 3 independent experiments. Venus expression images are shown using a fire LUT, and mT and mCherry are shown using a gray scale LUT.
None
The expression dynamics of lineage markers

A representative track for the expression dynamics of Tmem50bGt-Venus expression based on live imaging of a Tmem50bGt-Venus × R26-H2B-mCherry × mT embryo. Note that these data are integrated in the lineage map shown in Fig5A.

A representative track for the expression dynamics of 2610305D13RikGt Venus expression based on live imaging of a 2610305D13RikGt-Venus × R26-H2B-mCherry × mT embryo. Note that these data are integrated in the lineage map shown in Fig5B. Blue and red lines indicate outside and inside cells, respectively. Venus expression is normalized to mCherry signal.

Figure 5
Figure 5
A lineage map of mouse pre-implantation development

A, B Lineage segregation maps of all cells of two embryos indicating cell position and a lineage reporter expression. (A) Lineages from the 8- to 32-cell stage based on live imaging a Tmem50bGt-Venus × R26-H2B-mCherry × mT embryo. Venus expression intensity is normalized first to mCherry and subsequently to the maximum expression intensity of all cells in the same cell cycle stage of an embryo. Each box in the lineage tree ranges from 0 to 1 in normalized expression. Blue and red lines indicate outside and inside cells, respectively. Blue–white–red scale indicates relative distance of nuclei to the surface of the embryo. Dagger marks cells undergoing apoptosis, and cell lineages lost during imaging are labeled with an “L”. (B) Lineages from the 8- to 64-cell stage based on live imaging of a 2610305D13RikGt-Venus × R26-H2B-mCherry × mT embryo. Arrows highlight representative descendant pairs as a result of asymmetric divisions (white: essentially no difference in the expression level; yellow: higher expression in the outside cell than in the inside cell).

None
Establishment of an enhanced lineage map of mouse pre-implantation development Lineage maps are generated by tracking all nuclei of an embryo over several divisions. Binary cellular position, that is, inside (red) or outside (blue) position, as judged by the membrane contacts, and relative nuclear position, scaled by blue–white–red gradient, were combined with fluorescence intensity of the VET reporter measured in nuclei and incorporated into the lineage map.
Figure 6
Figure 6
Quantitative analyses of the expression dynamics of TE and ICM lineage markers

Venus expression dynamics, normalized to mCherry signal, in cells of Tmem50bGt-Venus × R26-H2B-mCherry × mT embryos dividing asymmetrically from 8- to 16-cell (n = 16 divisions in n = 3 embryos) and 16- to 32-cell stage (n = 7 divisions in n = 3 embryos; see text and Materials and Methods for details). Blue and red lines indicate outside and inside cells, respectively. Time point of division is set at t = 0 h.

Venus expression, normalized to mCherry signal, in cells of 2610305D13RikGt-Venus × R26-H2B-mCherry × mT embryos dividing asymmetrically from 8- to 16-cell (n = 15 divisions in n = 3 embryos) and 16- to 32-cell stage (n = 11 divisions in n = 3 embryos).

A pairwise difference of cumulative Venus expression between the descendants of asymmetric divisions (outside cell minus inside cell) at the 16- and 32-cell stages in Tmem50bGt (blue box plots; expression dynamics depicted in A; two-sided Wilcoxon rank-sum test; P (16-cell) = 0.0011; P (32-cell) = 0.0019) and 2610305D13RikGt embryos (red box plots; depicted in B). Inset depicts an example taken from (A) in which the measured value is highlighted in gray.

Mean rate of change of the normalized Venus intensity from t = 0 to 6 h after division in outside and inside cells for 16- (n = 32 and n = 16) and 32- (n = 56 and n = 40) cell stage in Tmem50bGt (left; Wilcoxon rank-sum test, P (32-cell) = 2 × 10−5) and for 16- (n = 33 and n = 15), 32- (n = 53 and n = 41) and 64- (n = 106 and n = 84) cell stages in 2610305D13RikGt (right; Wilcoxon rank-sum test, P (64-cell) = 0.025) embryos. Blue and red median bars mark outside and inside cells, respectively. Statistical significance is indicated as * (< 0.05) or ** (< 0.001).

Figure 7
Figure 7
A new model for TE and ICM lineage segregation Distinct mechanisms for the acquisition of the TE or ICM molecular identity. TE or ICM molecular identity is acquired in a non-binary fashion and at different developmental times. The specific expression of a TE fate marker (blue) is established through asymmetric divisions from 8- to 16-cell and 16- to 32-cell stage, as quantified by our reporter mice. The ICM-specific gene (red) expression is non-reciprocal to the TE marker and only becomes ICM specific at the late 64-cell stage.

Similar articles

See all similar articles

Cited by 8 PubMed Central articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback