XB-ART-866Dev Biol February 15, 2006; 290 (2): 246-64.
Cooperative non-cell and cell autonomous regulation of Nodal gene expression and signaling by Lefty/Antivin and Brachyury in Xenopus.
Dynamic spatiotemporal expression of the nodal gene and its orthologs is involved in the dose-dependent induction and patterning of mesendoderm during early vertebrate embryogenesis. We report loss-of-function studies that define a high degree of synergistic negative regulation on the Xenopus nodal-related genes (Xnrs) by extracellular Xenopus antivin/lefty (Xatv/Xlefty)-mediated functional antagonism and Brachyury-mediated transcriptional suppression. A strong knockdown of Xlefty/Xatv function was achieved by mixing translation- and splicing-blocking morpholino oligonucleotides that target both the A and B alloalleles of Xatv. Secreted and cell-autonomous inhibitors of Xnr signaling were used to provide evidence that Xnr-mediated induction was inherently long-range in this situation in the large amphibian embryo, essentially being capable of spreading over the entire animal hemisphere. There was a greater expansion of the Organizer and mesendoderm tissues associated with dorsal specification than noted in previous Xatv knockdown experiments in Xenopus, with consequent exogastrulation and long-term maintenance of expanded axial tissues. Xatv deficiency caused a modest animal-ward expansion of the marginal zone expression territory of the Xnr1 and Xnr2 genes. In contrast, introducing inhibitory Xbra-En(R) fusion constructs into Xatv-deficient embryos caused a much larger increase in the level and spatial extent of Xnr expression. However, in both cases (Xatv/Xlefty-deficiency alone, or combined with Xbra interference), Xnr2 expression was constrained to the superficial cell layer, suggesting a fundamental tissue-specific competence in the ability to express Xnrs, an observation with direct implications regarding the induction of endodermal vs. mesodermal fates. Our experiments reveal a two-level suppressive mechanism for restricting the level, range, and duration of Xnr signaling via extracellular inhibition by Xatv/Xlefty coupled with potent indirect transcriptional repression by Xbra.
PubMed ID: 16405884
Article link: Dev Biol
Genes referenced: admp cer1 chrd.1 frzb2 gsc lefty mixer myod1 nodal nodal1 nodal2 nrp1 otx2 shh tbxt wnt8a
Morpholinos: lefty MO1 lefty MO2 lefty MO4 lefty MO6
Article Images: [+] show captions
|Fig. 1. Specific inhibition of Xatv translation and splicing with morpholino oligonucleotides. (A) Xatv MO positions relative to the translational start sites on Xatv mRNAs (XatvMO1 and leftyB-MO) and the exon 1/intron 1 junction on Xatv genomic DNA (XatvMO2); see Materials and methods for actual MO sequences (reverse complements of those shown here). (B) XatvMO1 inhibits the translation of XatvA in vitro. While translation of XatvA is completely inhibited by XatvMO1, XatvB translation is not blocked (lane 4). Xatv* rescue RNA translation is not affected by XatvMO1 (lane 4). Neither control MO nor XatvMO2 affects translation of Xatv mRNAs (lanes 3, 5). Translation of Xnr2 RNA is used as both loading control and negative control. (C) XatvMO2 specifically inhibits Xatv pre-mRNA splicing (embryos received 50 ng of each XatvMO and were collected at the indicated stages). Primers in panel A (arrows) amplify across intron 1 using unspliced XatvA and XatvB pre-mRNA. (D) XatvMO2 preferentially inhibits XatvB pre-mRNA splicing. Embryos injected with 50 ng of XatvMO2 were collected at stage 10.5. RT-PCR was performed with A or B primer sets that amplify XatvA or XatvB pre-mRNA, respectively; see Materials and methods for primer sequences.|
|Fig. 2. Loss of Xatv function causes gastrulation defects. (A) Synergistic targeting by coinjection of XatvMOs. Coinjection of lower dose (25 ng or 16.7 ng each) of each XatvMO that targets both XatvA and XatvB RNA (XatvMO1 + leftyB-MO, XatvMO1 + XatvMO2 and XatvMO1 + XatvMO2 + leftyB-MO) is much more potent at knocking down Xatv function than single injection of each XatvMO (35 ng). Coinjection of XatvMO2 and leftyB-MO (complementary to 5′-UTR of XatvB mRNA; Branford and Yost, 2002) that mainly inhibit XatvB function is less effective than coinjection of other MO combinations. (B) Compared to (a) uninjected embryos or (b) embryos receiving 50 ng of control MO at 1-cell stage, severe morphogenetic defects at neurula stage (St. 18) are caused by (c) 25 ng each XatvMO1 and XatvMO2 (45%, n = 42), or (d) 20 pg of Xnr2 RNA (91.5%, n = 57). (C) Xatv depletion gives rise to exogastrulation. After 1-cell stage embryos received 30 ng each of XatvMO1 and XatvMO2, the vitelline membrane was removed at St. 9. (a) Uninjected embryo at St. 11.5. (b) Exogastrula caused by XatvMO1 + MO2 injection at St. 11.5 (n = 10/10). (c, d) Uninjected (c) and XatvMO1 + MO2-injected (d) embryo at St. 25. White arrowheads indicate the distal ectodermal region. (D) Xatv overexpression dose-dependently rescues the morphogenetic defects induced by XatvMO. (a) Rescue of XatvMO1 phenotypes by coinjection with the rescue RNA, Xatv* (Fig. 1A). XatvMO1 (50 ng) was injected into 1-cell stage embryos with Xatv* RNA (increasing pg dose indicated) and morphological changes scored at St. 18. Xatv* alone (200 pg), or control MO (50 ng) together with Xatv* (200 pg) did not significantly affect embryo development. In contrast, morphological defects caused by XatvMO1 injection were rescued by Xatv*, dose-dependently. (b) Dose-dependent rescue of splicing-blocker XatvMO2 phenotypes by coinjection with wild type XatvB RNA (other experimental conditions as above).|
|Fig. 3. Xatv depletion increases and prolongs Xnr expression levels and Xnr signaling. (A) XatvMO1 + MO2-injection upregulates and maintains Xnr1 and Xnr2 expression during gastrulation. (a–l) In situ hybridization (vegetal pole views) with Xnr2 in (a, e, i) uninjected or (b, f, j) control MO (60 ng)-injected embryos compared to those injected with (c, g, k) XatvMO1 alone (60 ng). [Quantitative analysis of alterations shown: c, n = 5/5; g, n = 6/6; k, n = 8/8.] (d, h, l) XatvMO1 + MO2 (30 ng each; quantitation: d, n = 14/15; h, n = 10/13; l, n = 13/13). Note the stronger Xnr2 signal in XatvMO1 + MO2-injected embryos than in XatvMO1-injected embryos. (m) RT-PCR with total RNA from whole embryos injected with 60 ng control MO, or 30 ng of each XatvMO1 and XatvMO2 (− RT, + RT controls are from uninjected embryos). (B) Xatv expression is increased and expanded in XatvMO1 + MO2-injected embryos during gastrulation. (a–c) Vegetal views. (d–l) Dorsal views. [Quantitation: c, n = 13/15; i, n = 21/23, l, n = 13/15.] (m, n) XatvMO1 + MO2-injected embryos at St. 12.5 were bisected longitudinally through the center of the Xatv expression domain after whole-mount in situ hybridization and viewed either (m) internally or (n) externally (green arrowheads, indentation of the superficial layer (see text); red arrowheads, epiboly margin).|
|Fig. 4. Xatv is essential for normal organizer and mesoderm formation during gastrulation. 1-cell stage embryos received 60 ng of control MO, XatvMO1, or 30 ng each (XatvMO1 + XatvMO2), and were assayed by in situ hybridization at the stages indicated. (A) Gsc expression. (a–l) Dorsal views. [Quantitative analysis of alterations shown: g, n = 9/9; h, n = 13/14; i, n = 9/9; j, n = 12/18; k, n = 13/14; l, n = 15/15.] Note the stronger and more expanded Gsc expression in XatvMO1 + MO2-injected embryos than in XatvMO1-injected embryos. (B) ADMP expression. (a–i) Dorsal views [Quantitation: g, n = 11/16; h, n = 15/17; i, n = 14/16.] (j, k) XatvMO1 + MO2-injected embryo at St. 12.5 was bisected through the center of ADMP expression domain along the animal–vegetal axis after in situ hybridization; lateral views are (j) internal or (k) external (green arrowheads, indentation of the superficial layer; red arrowheads, epiboly margin). (C) Cer expression. (a, d, g) Vegetal views. (b, c, e, f, h, i) Dorsal views. [Quantitation: g, n = 5/6; h, n = 7/7; i, n = 6/7.] (D) Xbra expression. (a–i) Lateral views, except panels c and f (vegetal views). (j–l) Animal views. [Quantitation: j, n = 14/15; k, n = 16/17; l, n = 15/17.] Xbra expression is massively expanded in XatvMO1 + MO2-injected embryos during gastrulation. (E) XWnt8 expression. (a, b, e, f): Vegetal views, (c, d, g, h): Same embryos viewed ventrally. [Quantitation: e, n = 13/19; f, n = 11/15.] Red arrowheads show the arc of XWnt8 non-expressing dorsal region.|
|Fig. 5. Xatv is required for normal endoderm fate specification. (A) In situ hybridization with Sox17α in uninjected and XatvMO1 + MO2-injected embryos during gastrulation. (a–c) Uninjected embryos. (d–f) XatvMO1 + MO2-injected embryos. (a–d) Vegetal view, dorsal upwards. [Quantitative analysis of alterations shown: d, n = 6/6.] (e, f) Lateral view. [Quantitation: e, n = 8/8; f, n = 9/10.] (g, h) Sox17α expression at St. 10.5 detected by in situ hybridization after bisection through the center of the dorsal lip of (g) uninjected (n = 10/10) and (h) XatvMO1 + MO2-injected (n = 9/10) embryos (dorsal to the left). (g′, h′) Magnified views, dorsal side of panels g and h, compared to (g′′, h′′) magnified ventral side views. (B) Mixer expression in (a–c) uninjected and (d–f) XatvMO1 + MO2-injected embryos during gastrulation. (a–e) Vegetal view, dorsal to the top. [Quantitation: d, n = 6/6; e, n = 8/8.] (f) Lateral view (n = 8/8). (g, h) Mixer expression at St. 10.5 detected by in situ hybridization after bisection through the center of the dorsal lip of (g) uninjected (n = 8/8) and (h) XatvMO1 + MO2-injected (n = 10/10) embryos (dorsal to the left). (g′, h′) Magnified views, dorsal side of g and h compared to (g′′, h′′) magnified ventral views. Green arrowheads: dorsal lip. Red arrowheads: most animal/anterior expression limit of Sox17α (A) and Mixer (B). White brackets: separation of blastocoel floor from the edge of Sox17α (A) and Mixer (B) expression domains in the superficial layer. Black brackets: breadth of Mixer-negative area in the dorsal mesoderm region (B: g′, h′).|
|Fig. 6. Xatv is essential for proper formation of mesodermal tissues at later embryogenesis. Classical exogastrulae were induced by incubating embryos with high salt (HS) solution. XatvMO1 + MO2-injected embryos (30ng each) were incubated in 1× SS without vitelline membranes. When sibling embryos reached stage 25, exogastrulae were collected. (A) In situ hybridization. (a–c) nrp-1 expression. [Quantitative analysis of alterations shown: b, n = 4/5; c, n = 7/8.] Yellow and blue lines in panel b indicate the protruded mesendodermal mass and ectoderm, respectively. (d–f) Xotx2 expression. White arrowheads in panel f indicate the expanded prechordal plate in the anterior end. [Quantitation: e, n = 5/5; f, n = 9/9.] (g–i) Shh expression. Red arrowheads in panel i indicate the expanded prechordal plate in head mesoderm. [Quantitation: h, n = 8/11; i, n = 14/18.] (j–l) Xatv expression. [Quantitation: k, n = 5/7; f, n = 10/15.] Yellow and blue arrowheads in panel j: Xatv expression at the hypochord and neural tube floorplate, respectively. Yellow and blue arrowheads in panel k: Xatv signal at the dorsal mesendodermal mass and mesendodermal mass/ectoderm junction, respectively. (k′, l′) Magnified views of panels k and l, respectively. (m–r) MyoD expression. [Quantitation: n, n = 6/6; o, n = 19/19.] (s–w) edd expression. Yellow arrowheads in panel w indicate the intense staining in axial mesoderm. [Quantitation: t, n = 9/9; u, n = 18/18.] (a–o, s–u) Lateral views, anterior left, dorsal upward. (p–r) Magnified dorsal view of panels m–o, respectively. (v, w) Dorsal views. (B) Histological morphology and relative volume of notochord + hypochord and mesodermal tissues in HS and XatvMO1 + MO2-induced exogastrulae. (a–d) Hematoxylin/eosin staining of exogastrulae after transverse sectioning (after whole-mount in situ hybridization for Xatv). Green dashed lines in panels a and c encircle the mesoderm area including somite and ventral mesodermal cells. (b, d) Magnified views of dorsal side of panels a and c, respectively. Yellow dashed line in panel d indicates the crescent-shaped tissue containing elongated cells and nuclei that are indicative of somitic muscle differentiation, which is less obvious within somitic mesoderm area in the HS-induced exogastrulae. (e, f) Comparison of the relative volume of hypochord + notochord area (e) and mesodermal area (f) between HS- and XatvMO1 + MO2-induced exogastrulae.|
|Fig. 7. Evidence for the expansion of Xbra expression by long-range Xnr signaling after Xatv knockdown. (A) In situ hybridization with Xbra on embryos at St. 10.5. (a) Uninjected embryo. (b) XatvMO1 + MO2-injected embryo. (c) CerS-injected embryos. (d) tALK4-injected embryo. (d′) High magnified view of panel d. (a–d) Lateral views. Red-gal staining in panels c and d detects the descendents of the cell injected with CerS and tALK4, respectively. (B) Xnr-specific inhibitors prevent ectopic Xbra expression in Xatv-deficient embryos. In situ hybridization on embryos at St. 10.5 detects Xbra expression, with red-gal staining detecting the descendants of the cell injected with RNA encoding the inhibitor (see text for detailed experimental design). Injection of β-gal (250 pg) does not affect ectopic Xbra expression (a–b′). CerS (500 pg) inhibits ectopic Xbra expression non-cell autonomously (d–e′), whereas tALK4 (500 pg) suppresses the ectopic Xbra expression cell autonomously (g–h′). (c, f, i) Simplified diagrams showing the effects on ectopic Xbra expression by β-gal, CerS, and tALK4, respectively. Red dots represent the clone of cells that express β-gal, CerS, and tALK4, respectively. “⊥” symbols in panels f and i show inhibition of Xnr signaling non-cell and cell autonomously by CerS and tALK4, respectively. (a, e) Animal views. (b, d, g, h) Lateral views skewed ∼45° animal-ward. (a′, b′, d′, e′, g′, h′) Magnified views of the yellow bracketed area of panels a, b, d, e, g, and h, respectively. Green arrowheads in panels A and B indicate animal pole. (C) In situ hybridization with Der in (a, b) uninjected and (c, d) XatvMO1 + MO2-injected embryos. Dorsal views, with (d) angled downward slightly to visualize the animal-ward extent of Der signal.|
|Fig. 8. Xatv and Xbra synergistically restrict the Xnr expression domain. Xbra-EnR RNA (1 ng) was injected ∼45° above the equator into one blastomere of 4-cell stage embryos that previously received 30 ng each of (XatvMO1 + XatvMO2) at the 1-cell stage. β-gal RNA (400 pg) was coinjected with Xbra-EnR RNA as lineage tracer. Xnr2 and Xbra expression was analyzed by in situ hybridization at St. 10.5 after red-gal staining. (A) Xnr2 expression, uninjected embryo. Red-gal staining indicates descendants of the β-gal-injected cell. (B) Xbra expression, Xbra-EnR RNA-injected embryo. (C) Xnr2 expression, XatvMO1 + MO2-injected embryo. (D) Xnr2 expression, Xbra-EnR RNA-injected embryo. (E) Xnr2 expression, XatvMO1 + MO2 and Xbra-EnR RNA-injected embryo. (C′, D′, E′) Embryos bisected through the center of the Xnr2 expression domain along animal-vegetal axis (from C, D and E). Green arrowheads, dorsal lip; blue arrowheads, anterior/animal boundary of Xnr2 expression; yellow arrowheads in panels A, C, D, E—approximate animal limit of Xnr2 expression domain. Note that Xnr2 expression remains in the superficial layer. Black print: injection or uninjection of XatvMO1 + MO2 (MO1/2); red print: reagents coinjected with the β-gal lineage tracer, except (A); blue print: probes used for in situ hybridization. (F) Model of synergistic Xatv/Xlefty and Xbra-mediated restriction of Xnr expression during gastrulation. Schematic diagrams represent dorsal midline-bissected stage 10.5 early gastrulae, showing internal expression domains of Xnr2 (red), Xatv (green), and Xbra (purple) in embryos that are (left) normal uninjected, (middle) Xatv/Xlefty MO1/MO2-coinjected, or (right) MO1/MO2 coinjected with Xbra-EnR. Increased line weight of arrows indicates elevated Xnr signaling, and increasing color intensity from light to dark red indicates increasing Xnr2 expression intensity. In normal embryos, Xnr2 expression is tightly restricted to the superficial layer of the dorsal lip (indentation), with Xnr–Xnr autoregulation maintaining this territory. Xnr2 restriction is accomplished by both Xatv and Xbra, activated by Xnr signaling diffusing from the Xnr2 expression domain. Primarily, Xatv inhibits Xnr expression non-cell autonomously. Xnr diffusing from producing cells activates expression of Xbra (purple) in both deep and superficial layers, which is stronger and earlier in the deep marginal zone compared to the superficial layer (see text). Rapid activation of Xbra provides an indirect repressive influence (Xbra is a transcriptional activator), implying that an unknown “Factor X” inhibits Xnr autoregulation, leading to the fixing of mutually exclusive expression domains for Xnr2 and Xbra. Knockdown of Xatv function using Xatv/Xlefty MO1 + MO2 allows limited expansion of Xnr2 expression, but elevated Xnr signaling range causes the massive expansion of Xatv and Xbra expression. Xnr2 expression is still restricted by the increased Xbra-mediated indirect suppression. Interfering with both Xatv and Xbra allows expansion of Xnr2 expression farther from the dorsal lip compared to either knockdown of Xatv or blocking Xbra function alone. Even in this condition, Xnr2 expression occurs only in the superficial layer.|
|Supplemental Fig. 1. Synergistic effect on Xatv knockdown by coinjecting XatvA- and XatvB-specific MOs, with dramatic effects on Xnr target gene expression beginning from early gastrula onwards. (A) Embryos were injected with 60 ng of XatvMO1 or XatvMO2, or 30 ng each of (XatvMO1 + XatvMO2) at the 1-cell stage, and assayed at St. 9.5 and St. 10.5 by whole-mount in situ hybridization. Some embryos were bisected sagittally through the center of the dorsal lip before processing. MO-based depletion of Xatv did not change the expression level/domain of Xatv or Xbra at stage 9.5 (Aa–d, i–l). Expression of both genes was markedly upregulated and expanded in Xatv MO-injected embryos at stage 10.5 (Ae–h, m–t). Note that coinjecting XatvMO1 and XatvMO2 caused a more increased and expanded expression of markers than either single MO injection, suggesting a stronger knockdown of Xatv function. (a–p) Dorsal views, except (m) (vegetal view). (q–t) Sagittally bisected embryos. Green arrowheads, dorsal lip; red and yellow arrowheads show the distal and proximal boundaries of Xatv expression from the dorsal lip, respectively; white brackets, breadth of Xbra expression domain. [Quantitative analysis of alterations in MO-injected embryos shown: b, n = 10/10; c, n = 6/9; d, n = 6/8; f, n = 9/10; g, n = 8/10; h, n = 4/4; j, n = 8/8; k, n = 5/7; l, n = 3/4; n, n = 7/9; o, n = 8/10; p, n = 6/6; r, n = 5/5; s, n = 7/7; t, n = 6/6.] (B) Comparison of knockdown effect by several combinations of XatvMOs. Embryos were injected with 60 ng of (b) leftyB-MO (Branford and Yost, 2002), 30 ng of each of the pair of (c) XatvMO1+ leftyB-MO, (d) XatvMO2+ leftyB-MO, or (e) XatvMO1 + XatvMO2, or 20 ng of each of the three (f) XatvMO1 + XatvMO2 + leftyB-MO at the 1-cell stage, and assayed at St. 10.5 by whole-mount in situ hybridization. Coinjection of MOs that target the A and B copy of Xatv induce more expanded Xatv expression than injecting leftyB-MO alone (b; 55%, n = 20). MO combinations that inhibit both XatvA and XatvB function such as XatvMO1+ leftyB-MO (c; 73.7%, n = 20), XatvMO1 + XatvMO2 (e; 85%, n = 20), and XatvMO1 + XatvMO2 + leftyB-MO (f; 72.2%, n = 18) give rise to expanded Xatv expression at higher frequency than XatvMO2+ leftyB-MO that mainly targets XatvB function (d; 55%, n = 20).|
|Supplemental Fig. 2. Expression of Xnr1 in XatvMO1 super + MO2-injected embryos. XatvMO1 and XatvMO2 (30 ng each) were injected at 1-cell stage. Xnr1 expression was analyzed by whole-mount in situ hybridization. Upper panel shows uninjected embryos at stage 10.5 (n = 5/5). Dorsal view. Lower panel shows XatvMO1+ XatvMO2-injected embryos (n = 7/7). Dorsal view. Note the limited expansion of Xnr1 expression in Xatv-deficient embryos, such as that of Xnr2 expression.|
|Supplemental Fig. 3. Expression of Chd, Xbra, and Mixer in XatvMO1 super + MO2-injected embryos. (A–D) Whole-mount in situ hybridization, Chd probe (dorsal views). During gastrulation, Chd expression is highly upregulated and expanded towards the animal side in XatvMO1 + MO2-injected embryos (C; n = 19/22, D; n = 15/17). (E–L) Comparison of the expression of organizer and mesendodermal markers. (E, F, I, J) Mixer and Xbra in situ hybridizations were performed with embryos after sagittal bisection through the center of the dorsal lip (I, n = 10/10; J, n = 9/10). (G, H, K, L) Embryos were sagittally bisected after whole-mount in situ hybridization, through the center of the Chd expression domain seen in panels A–D (green arrowheads indicate dorsal lip).|
|Supplemental Fig. 4. Xatv and Xbra synergistically regulate Xnr expression during gastrulation. The experimental method is the same as that in Fig. 8. To show more clearly the increased Xnr2 signal intensity caused by Xbra-EnR in the background of reduced Xatv/Xlefty function, the β-gal RNA injection and red-gal staining were omitted. Expression of Xnr2 and Xbra was analyzed by in situ hybridization at St. 10. 5 (A–E) and St. 11.5 (F–J). (A, C, D, E) Xnr2 expression at St. 10.5 after injecting the reagent combinations shown in black on the left side of the panels. All dorsal views, except (A) vegetal view, dorsal to the top. (B) Xbra expression at St. 10.5 after injecting Xbra-EnR. Dorsal view. (A′–E′) Magnified views of bracketed areas in panels A–E, respectively. [Quantitation: A, n = 18/20; B, n = 12/12; C, n = 20/20; D, n = 23/23; E, n = 21/22] (F, H, I, J) Xnr2 expression at St. 11.5 after injecting reagent combinations in black. Dorsal views, tilted upward. (G) Xbra expression, St. 11.5, after injecting Xbra-EnR. Vegetal view. (F′–J′) Magnified views of bracketed areas in panels F–J, respectively. [Quantitation: F, n = 19/19; G, n = 10/10; H, n = 19/19; I, n = 24/24; J, n = 27/31] Red arrowheads: animal limit of visible Xnr2 signal. The increased intensity of Xnr2 signal (punctate perinuclear staining) and expanded expression domain of Xnr2 in Xbra-EnR-injected embryos indicate that Xbra affects Xnr autoregulation at the transcriptional level (D, D′, E, E′). Xnr2 expression was expanded farther in MO1/MO2+Xbra-EnR-injected embryos with more elevated signal intensity than in the other conditions, demonstrating the multiple influences on Xnr expression by Xatv and Xbra (E, E′, J, J′). Note: The apparently minor effect on Xnr2 expansion between the MO1/MO2 + Xbra-EnR-injected condition compared to the MO1/MO2-injected condition (D and E, this figure), as compared to the greater expansion caused by Xbra-EnR in Fig. 8E may be caused by embryo batch-to-batch variation in response, which may be more noticeable during the dynamic mid-gastrula stage of patterning where the inputs on Xnr expression begin to have the most dramatic effects. In addition, the embryo represented in panel E may be slightly earlier than the one shown in Fig. 8E; analysis of later stage 11.5 embryos (F–J) clearly shows a much greater effect on intensity and domain of Xnr2 expression at this stage (compare panels H and I to panel J).|