Published online 2015 Feb 26. doi: 10.7554/eLife.04379
PMID: 25719209
Marianne E Bronner, De La’s innovative use of eclectic samples, bugged-out album skits, and multi-layered lyrical wordplay left a stamp on the music that will never die. Their “Buddy” remix became ground zero for the birth of the Native Tongues movement, helping launch the careers of A Tribe Called Quest and giving further shine to rising talents Queen.
Reviewing editorMarianne E Bronner, California Institute of Technology, UnitedStates;
This article has been cited by other articles in PMC.
Associated Data
DOI:http://dx.doi.org/10.7554/eLife.04379.029
DOI: 10.7554/eLife.04379.029
Abstract
In vertebrates, the total number of vertebrae is precisely defined. Vertebrae derivefrom embryonic somites that are continuously produced posteriorly from the presomiticmesoderm (PSM) during body formation. We show that in the chicken embryo, activationof posterior Hox genes (paralogs 9–13) in the tail-budcorrelates with the slowing down of axis elongation. Our data indicate that a subsetof progressively more posterior Hox genes, which are collinearlyactivated in vertebral precursors, repress Wnt activity with increasing strength.This leads to a graded repression of the Brachyury/T transcriptionfactor, reducing mesoderm ingression and slowing down the elongation process. Due tothe continuation of somite formation, this mechanism leads to the progressivereduction of PSM size. This ultimately brings the retinoic acid (RA)-producingsegmented region in close vicinity to the tail bud, potentially accounting for thetermination of segmentation and axis elongation.
DOI:http://dx.doi.org/10.7554/eLife.04379.001
eLife digest
In humans and other vertebrates, the number of bones (vertebrae) in the spine isdetermined early in development. The vertebrae form from blocks of tissue calledsomites that make segments along the body axis—a virtual line running from thehead to the tail-end—of the embryo. The somites form as the embryo increasesin length, with new somites forming periodically at the back near the embryo'stail-end.
A family of genes called the Hox genes are involved in controllingthe formation of the somites. However, it is not known whether they directly controlthe number of somites that form, or whether they control the length of the body ofthe embryo.
Denans et al. studied the Hox genes in chicken embryos. Theexperiments suggest that the activation of some of the Hox genes ina structure called the tail-bud, which is found at the tail-end of the embryo, slowdown the elongation of the body. The Hox genes achieve this byrepressing the activity of a signaling pathway called Wnt so that Wnt activity in thetail-bud progressively decreases as the embryo develops.
The elongation of the body stops when the levels of a molecule called retinoic acidincrease in the tail-bud, which causes the loss of the stem cells that are needed tomake the somites. Denans et al.'s findings suggest that Hoxgenes influence the timing of the halt in elongation, which in turn is important fordetermining the total number of somites that form. Understanding howHox genes control the formation of the cells that will make upthe somites and influence Wnt signaling is a major challenge for the future.
DOI:http://dx.doi.org/10.7554/eLife.04379.002
Introduction
Body skeletal muscles and vertebrae form from a transient embryonic tissue calledparaxial mesoderm (PM). The PM becomes segmented into epithelial structures calledsomites, which are sequentially produced in a rhythmic fashion from the presomiticmesoderm (PSM). The PSM is formed caudally during gastrulation by ingression of the PMprogenitors located initially in the anterior part of the primitive streak (PS) andlater, in the tail-bud (). At the end of somitogenesis, the embryonic axis issegmented into a fixed species-specific number of somites which varies tremendouslybetween species ranging from as little as ∼32 in zebrafish to more than 300 insome snakes. The somites subsequently differentiate into their final vertebral andmuscular derivatives to establish the various characteristic anatomical regions of thebody. Hox genes code for a family of transcription factors involved inspecification of regional identity along the body axis (; ). In mouse and chicken, the 39 Hox genesare organized in four clusters containing up to thirteen paralogous genes each.Hox genes exhibit both spatial and temporal collinearity, meaningthat they are activated in a sequence reflecting their position along the chromosome andbecome expressed in domains whose anterior boundaries along the body axis also reflecttheir position in the clusters.
Whether Hox genes control axis length and segment number has beencontroversial. Mouse mutants in which entire sets of Hox paralogs areinactivated show severe vertebral patterning defects but exhibit normal vertebral counts(; ). In contrast, precociousexpression of Hox13 genes in transgenic mice leads to axis truncationwith reduced vertebral numbers (). Furthermore, mouse null mutations for Hoxb13 orHoxc13 result in the production of supernumerary vertebrae (; ).
In chicken and fish embryos, the arrest of axis elongation has been linked to theinhibition of FGF and Wnt signaling in the tail-bud which leads to the down-regulationof the transcription factor T/Brachyury and of the Retinoic Acid(RA)-degrading enzyme Cyp26A1 (; ; ; ). Downregulation ofCyp26A1 in the tail-bud ultimately leads to rising RA levels and todifferentiation and death of the PM progenitors which terminates axis elongation.Premature exposure of the tail-bud to high RA levels in chicken or mouse embryosinhibits Wnt and FGF signaling and leads to axis truncation (; ; ) suggesting that the tail-bud must be protected from thedifferentiating action of RA. In the Cyp26A1 null mutant mice,RA-signaling reaches the tail-bud, prematurely inducing the downregulation of FGFsignaling and the increase of Sox2 expression, resulting in axistruncation posterior to the thoracic level (; ). Inchicken, the tail-bud starts to produce RA when explanted in culture after the 40-somitestage (). This late RAsignaling activity in the tail-bud is involved in the termination of segmentation andaxis elongation (; ). At the 40-somitestage, the mRNA for Raldh2, the RA-biosynthetic enzyme becomesexpressed in the tail-bud potentially accounting for this late RA activity. Whattriggers this late expression of Raldh2 in the chicken tail-bud ishowever unknown.
In vertebrates, the termination of axis elongation is accompanied by a progressivereduction in size of the PSM (; ). The shrinking of the PSM which brings the segmented region producing RA inthe vicinity of the tail-bud might also contribute to the raise in RA levels in thetail-bud and possibly to the late Raldh2 activation in the tail-bud.Thus in the chicken embryo, the timing of elongation arrest (and hence the total numberof somites formed) could be in part controlled by the kinetics of PSM shrinking. PSMsize depends on the velocity of somite formation which removes cells anteriorly and onthe flux of cells from the primitive streak and tail-bud generated during elongationwhich injects cells posteriorly. How this flux of progenitors ingressing in the PSM isregulated over time, and which genes are regulating this process remain poorlyunderstood. Hoxb1-9 genes have been proposed to control cell ingressionof paraxial mesoderm precursors from the epiblast during gastrulation (). However,Hoxb1-9 genes are only expressed in anterior regions of the embryoprecluding their playing a role in the control of axis termination.
Here, we first investigate the relationship between the speed of somite formation and ofaxis elongation. We show that, in the chicken embryo, activation of Abdominal B-likeposterior Hox genes in the tail-bud correlates with the slowing down ofaxis elongation, while the speed of somitogenesis remains approximately constant. Ourdata indicate that a subset of progressively more posterior Hox genes,which are collinearly activated in vertebral precursors, repress Wnt and FGF activitywith increasing strength, leading to a graded repression of theBrachyury/T transcription factor. This progressively reducesmesoderm ingression and cell motility in the PSM, thus slowing down the elongationprocess. Due to the continuation of somite formation at a steady pace, this mechanismleads to the progressive reduction of PSM size.
Results
Activation of posterior Hox genes correlates with axiselongation slowing down
We measured the variations of velocities of axis elongation and somite formation intime-lapse videos of developing chicken embryos during the formation of the first 30somites (Video 1). The velocity of somiteformation shows limited variation during this developmental window () (Figure 1A, n = 4 embryos for each condition). Incontrast, axis elongation velocity increases during the formation of the first 10somites and then it decreases until the 25-somite stage, when it drops abruptly(Figure 1A and Video 2, n = 41 embryos). The number of PSM cellsdecreases over time (Figure 1B, n = 5embryos for each condition) while no significant difference in cell proliferation orapoptosis in the PSM and tail-bud is observed (Figure1C–F, n = 4 embryos for each condition). Cell motility, whichhas been implicated in the control of axis elongation (), also decreased between 15and 27 somites (Figure 1G, n = 4embryos for each condition). Thus, a parallel decrease in cell motility and in cellflux to the PSM accompanies axis elongation slow down.
Video 1.
Time-lapse video of an embryo from Stage 5 HH to 29 somites showing thedifferent phases of axis elongation (Bright-field, ventral view, anterior isup).DOI:http://dx.doi.org/10.7554/eLife.04379.003
Video 2.
Time-lapse videos showing axis elongation slow-down around the 25-somitestage.Bright-field imaging of chicken embryos at 15–17 somites (leftpanel), 20–22 somites (middle panel), and 25–27 somites (rightpanel) (ventral view, anterior is up).
DOI:http://dx.doi.org/10.7554/eLife.04379.005
Slowing down of axis elongation correlates with decreasing cellingression in the PSM.(A) Velocity of axis elongation and of somite formation.(B) PSM cell number. (C–D)Tiling of confocal sections of 20-somite (C) and 25-somite(D) stage embryos. EdU positive cells are labeled in green,phosphorylated histone H3 (pH3) in red, and nuclei in blue (DAPI).(C′, D′) Higher magnification ofPSM regions used to quantify the proliferation. (C″,D″) Confocal sections of parasagittal cryosections oftail-bud used to quantify cell proliferation.(E–F) Quantification of cellproliferation (E) and apoptosis (F) in20–22 and 25–27 somites chicken embryos. (G) Cellmotility in the posterior PSM.
DOI:http://dx.doi.org/10.7554/eLife.04379.004
Hoxb1-9 genes were shown to regulate cell flux to the PSM bycontrolling the timing of cell ingression from the epiblast (). Activation ofHox genes in the epiblast and tail-bud is collinear and occurs intwo phases. First, the paralog groups 1 to 8 (and Hoxb9) are quicklyactivated within ten hours before the first somite formation (stage 7 HH []; Figure 2, n = 8 embryos for eachcondition). This phase is followed by a pause during formation of the first tensomites. Then between the 10 and 40-somite stage, the posterior Hoxgenes corresponding to the paralog groups 9–13 (and Hoxc8 andHoxd8) become subsequently activated in a slower phase whichtakes almost 48 hr (Figure 2, n = 8embryos for each condition). Hoxa13 is the first Hox13 activated atthe 25-somite stage, when axis elongation slows down abruptly. Thus, there is astriking correlation between the timing of posterior Hox genesactivation and the beginning of axis elongation slow down (Figures 1A and 2).
Collinear activation of Hox genes in paraxial mesodermprecursors.(A) Table showing the collinear onset of Hoxgenes expression in the epiblast/tail-bud generated from (B)Chicken embryos hybridized in whole-mount with Hoxa (blue),Hoxb (yellow), Hoxc (red), andHoxd (green) probes. Each panel shows the beginning ofactivation of each Hox gene in paraxial mesoderm precursorsin the epiblast of the anterior primitive streak or in the tail-bud.Hox probe used is indicated on the top of each panel.Anterior is up. Dorsal view.
DOI:http://dx.doi.org/10.7554/eLife.04379.006
A subset of posterior Hox genes can regulate cell ingression andaxis elongation
In order to test the role of posterior Hox genes on the control ofcell ingression and cell motility in the developing chicken embryo, we used an invivo electroporation technique, allowing to precisely target the paraxial mesodermprecursors in the epiblast of the anterior primitive streak () (Video 3). We developed a strategy allowing to overexpress twodifferent sets of constructs in largely different populations of paraxial mesodermcells by performing two consecutive electroporations of the paraxial mesoderm (PM)precursors of the epiblast of stage 4–5 HH embryos. Embryos are firstelectroporated on the left side of the primitive streak with a control Cherryconstruct, and then on the right side of the streak with a second vector expressingthe yellow fluorescent protein Venus and a Hox construct (Figure 3A). This strategy results in essentiallydifferent PM cells expressing the two sets of constructs with the Cherry expressingcells enriched on the left side whereas Hox expressing cells are mostly found on theright side. When no Hox construct is present in the Venus vector,the Cherry and Venus-expressing populations of cells were observed to extend from thetail-bud to the same antero-posterior level of the paraxial mesoderm indicating thatthey began ingressing at the same time (Figure3B, n = 8 embryos). In contrast, cells expressing Cherry were alwaysextending more anteriorly than cells expressing Venus and one of the followingposterior Hox gene: Hoxa9, Hoxc9, Hoxd10, Hoxd11, Hoxc11,Hoxa13, Hoxb13, or Hoxc13, indicating that theseHox genes can delay cell ingression of the PSM progenitors (Figure 3B–C n > 6 for eachcondition and not shown). This simply reflects the fact that cells ingressing laterbecome located more posteriorly. Strikingly, the effect on ingression wasprogressively stronger when overexpressing more 5′ genes suggesting acollinear trend (Figure 3C). Inverting theorder in which the constructs are electroporated did not affect the final phenotype.The distance between the anterior boundaries of the two domains was found toprogressively increase with more posterior Hox genes as shown bymeasuring the ratio between Venus and Cherry posterior domains (Figure 3A–C). Over-expression of Hoxa10, Hoxc10,Hoxa11, Hoxc12, Hoxd12 and Hoxd13 showed no difference with the controlCherry vector (Figure 3A–C, n >6 for each condition and data not shown). Using consecutive electroporation ofHoxd10 and Hoxc11 constructs labeled with Cherryand Venus, respectively, we observed that Hoxc11 has a strongereffect on ingression than Hoxd10 (Figure 3—figure supplement 1, n = 12 embryos). A similarresult was observed when Hoxa13 was compared toHoxc11 in the same assay (Figure 3—figure supplement 1, n = 6 embryos). Thus, a subsetof posterior Hox genes is able to delay PSM cell ingression in acollinear manner.
Video 3.
Time-lapse video showing the precise targeting of PSM progenitors andthe ingression of the epiblast cells to form the PSM.Bright-field (purple) merged with fluorescent images of PSM cell progenitorselectroporated with a control H2B-Venus (ventral view,anterior is up) from stage 6 HH onwards.
DOI:http://dx.doi.org/10.7554/eLife.04379.007
Posterior Hox genes can regulate cell ingression ina collinear fashion.(A) Consecutive electroporation protocol. The ratio of thegreen domain (green bar, Hox expressing) over the reddomain (red bar, control vector) measures the ingression delay.(B) Embryos consecutively electroporated first withCherry and then with Venus together with control, Hoxa9,Hoxc11, or Hoxb13 vectors.Arrowheads: anterior boundary of Cherry (red) and Venus (green) domains.(C) Ratio of Venus over Cherry domains for posteriorHox genes. Dots: electroporated embryos. Barindicates the mean. Stars: p-value of two-tailed Student'st-test applied between the different conditions.*p < 0.05; **p < 0.01;***p < 0.005. Error bars: standard error tothe mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.008
Figure 3—figure supplement 1.
The posterior Hox genes regulate cell ingressionwith increasing strength.(A) Embryos consecutively electroporated withHoxd10-Cherry and Hoxc11-Venus(left) and with Hoxc11-Cherry andHoxa13-Venus (right). Arrowheads: anterior boundaryof Cherry (red) and Venus (green) domains. (B) Ratio ofVenus over Cherry domains corresponding to A. This showsthat Hoxa13 retains the cell longer in the epiblast thanHoxc11 which retains the cell longer thanHoxd10.
DOI:http://dx.doi.org/10.7554/eLife.04379.009
To analyze the effect of posterior Hox genes on ingression, PMprecursors were electroporated with Venus and a Hoxa13 or a controlconstruct and harvested after 5 hr when the electroporated cells start to ingress. Noectopic expression of laminin (Figure4A,B–C″), acetylated tubulin (Figure 4A,D–E″), or E-cadherin (data not shown) was observedafter Hoxa13 over-expression. We compared the number ofVenus-positive cells in epiblast vs primitive streak and mesoderm in embryo sections.The majority of Hoxa13-expressing cells were still found in theepiblast while control cells have ingressed into the primitive streak and mesodermindicating that Hoxa13 delays ingression by retaining cells in theepiblast (Figure 4F–H, n = 4embryos for each condition). No up-regulation of the neural markerSox2 was observed in the tail-bud of embryos electroporated withHoxa13 (Figure4I–J, n = 8 embryos for each condition) and very few cells wereobserved in the neural tube of embryos electroporated with Hox constructs (see Figure 3B, Figure4I–J, Figure 6A–B, Figure 7D–J and Figure 9A andVideos 4–8). This indicates that the effect on ingression is not caused by therecruitment of PM precursor cells to a neural fate. Ingression of cells from theepiblast to the primitive streak occurs via an epithelium to mesenchyme transitionwhich involves first destabilization and then complete loss of basal microtubules inthese cells. This process has been shown to be regulated by a basally localizedactivity of the small GTPase RhoA (). In order to test if the effect of the posterior Hoxgenes on delaying PSM progenitors ingression could involve regulation of microtubulestability, we used a dominant negative form of RhoA (DN-RhoA) as atool to destabilize basal microtubules in the epiblast (). We performed consecutive electroporationsat stage 5 HH to overexpress a control Cherry vector in one population of PSMprogenitors and Hoxa13 with DN-RhoA vectors inanother population and allowed the embryos to develop for 20 hr. We observed that thetwo populations of cells reach the same anterior level (Figure 4K, n = 5/5 embryos) indicating that these cellsingressed at the same time. Altogether, theseresults suggest that Hox genes control PSM progenitors ingressionthrough the regulation of basal microtubule stability in the epiblast.
Epiblast cells overexpressing Hox genes do not convertto a neural fate.(A) Transverse section of a stage 7 HH chicken embryo labeledwith phalloidin (white) to highlight the actin network and with laminin(red) to identify the epiblast basal membrane. Colored boxes indicate thedifferent phases of differentiation of the mesoderm: epiblast (purple),ingressing cells (yellow), and mesoderm (blue).(B–E″) Transverse sections at thePSM progenitors level 5 hr after electroporation of a control Venus or ofHoxa13. (B-C”) Lamininimmunolabeling (red) after Venus(B–B″) or Hoxa13over-expression (C–C″).(D–E″) Acetylatedα-tubulin immunolabeling (red) after Venus(D–D″) or Hoxa13(E–E″) over-expression.(F–G) Transverse cryosections of theanterior primitive streak of an embryo electroporated with Venus(F) or with Venus and Hoxa13(G). White arrow: cells ingressed in the primitive streak(F) and non-ingressed epiblast cells (G).Green: Venus; red: laminin; blue: nuclei. (H) Quantification ofingression in embryos electroporated with control orHoxa13-expressing constructs.(I–J) In situ hybridization of 2-day oldchicken embryos electroporated with Venus (I) orHoxa13-Venus (J) expressing vectors. Leftpanel shows Sox2 expression in the neural tube andtail-bud, and right panels show GFP immunohistochemistry. (K)Chicken embryo consecutively electroporated with a control (Cherry, red) anda mix of Hoxa13+DN-RhoA (Venus, green). Arrowheads:anterior boundary of Cherry (red) and Venus (green) domains. Stars: p-valueof two-tailed Student's t-test applied between thedifferent conditions. ***p < 0.005. Error bars:standard error to the mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.010
We next tested the effect of over-expressing posterior Hox genes onaxis elongation (Figure 5A–C, Video 4, n = 47 embryos).Over-expression of either Hoxa9, Hoxc9, Hoxd10, Hoxd11, Hoxc11, Hoxa13,Hoxb13 or Hoxc13 but not of Hoxa10, Hoxc10,Hoxa11, Hoxc12, Hoxd12 and Hoxd13 in PM precursorscaused a significant decrease of elongation velocity (Figure 5A–C and not shown). The effect of Hoxgenes becomes progressively stronger for more posterior genes (Figure 5C and not shown, Video 4). Therefore, the same posterior Hox genes canalter cell ingression and axis elongation with a similar collinear trend (Figures 3C and 5C). The cell-autonomouscontrol of ingression by posterior Hox genes (Figure 4F–H) is expected to reduce the supply of motilecells in the posterior PSM. This should slow down elongation movements and couldexplain why such a non-cell autonomous effect on axis elongation is observed whileonly 30–50% PM cells express the Hox constructs. These datasuggest that a subset of posterior Hox genes controls the slowingdown of axis elongation by regulating ingression of PM precursors.
Video 4.
Effect of Hoxa9, Hoxc11, and Hoxa13 electroporation on axis elongationand ingression.Bright-field (purple) merged with fluorescent images of PSM cell progenitorselectroporated with either a control H2B-Venus (first panelfrom the left), Hoxa9-ires2-H2B-Venus (second panel fromthe left), Hoxc11-ires2-H2B-Venus (third panel) or aHoxa13-ires2-H2B-Venus (right panel) constructs (green)(ventral view, anterior is up) from Stage 6 HH onwards. Over-expression ofHoxa9, c11, and a13 affects ingression and axiselongation in a collinear fashion.
DOI:http://dx.doi.org/10.7554/eLife.04379.015
Posterior Hox genes control the axis elongationvelocity in a collinear fashion.(A–B) Time-lapse series of chicken embryoselectroporated either with control (A) orHoxa13 (B). Red line: position ofHensen's node. ss = somite-stage. (C) Velocity ofaxis elongation of embryos electroporated with either a control,Hoxa9, Hoxc9, Hoxd10,Hoxd11, Hoxc11, Hoxa13, Hoxb13, orHoxc13 expressing constructs. Stars: p-value oftwo-tailed Student's t-test applied between thedifferent conditions. *p < 0.05. Error bars: standard error tothe mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.020
In Drosophila and vertebrates, Hox genes expressedposteriorly can suppress the function of more anterior ones, a property termedphenotypic suppression or posterior prevalence (). We previously showed that posterior prevalenceapplies for the control of cell ingression by Hoxb1-9 genes (). To testwhether this property also applies to the posterior Hox genes withan effect on axis elongation, we performed consecutive electroporations first with amix of Hoxd10 and Hoxc11 constructs (leading toexpression in the same cells, in green Figure6A) and then with a mix of Hoxc11 and a control construct(a mutated Hoxc11 unable to bind DNA (Hoxc11mutH),in red Figure 6A). We observed that cellsover-expressing the two functional Hox genes reach the same anteriorposition as cells over-expressing Hoxc11 and control (Figure 6A,C, n = 10 embryos). ThusHoxc11 function is dominant over Hoxd10.Similarly, we observed dominance of Hoxa13 overHoxc11 in the same assay (Figure6B,C, n = 8 embryos). Therefore, posterior prevalence appears togenerally apply for Hox control of cell ingression in the mesoderm(). As aresult, the effect of Hox genes on cell retention in the epiblastshould become progressively stronger as more posterior genes become activated.
Posterior prevalence of posterior Hox genes.(A) Embryos consecutively electroporated first withHoxc11-Cherry +Hoxc11mutH-Cherry and with Hoxd10-Venus+ Hoxc11-Venus shown 24 hr after reincubation.(B) Embryos consecutively electroporated first withHoxa13-Cherry +Hoxa13mutH-Cherry and then withHoxa13-Venus + Hoxc11-Venusshown 24 hr after reincubation. Red arrowheads: anterior boundary ofCherry-expressing cells. Green arrowheads: anterior boundary ofVenus-expressing cells. (C) Quantification of the ratio ofVenus over Cherry expressing domains for the experiments shown inA and B. Each dot corresponds to oneelectroporated embryo and bar indicates the mean.(D–E) Luciferase assay measuringWnt/βcatenin pathway activity after over-expression of the BATLucconstruct together with a Renilla-expressing vector and either(D) control, Hoxa9, Hoxa13or the combination of Hoxa9 and Hoxa13expressing vectors. (E) Blow-up of the samples shown in(D). (F) BATLuc assay with serial dilutions ofthe Hoxa13 plasmid (in μg/μl on the x axis).(G) Western blot labeled with an anti-HA antibody showingembryos electroporated with Hoxa13 under the control of adoxycycline-responsive promoter activated with different doses ofdoxycycline (in μg/ml). (H) BATLuc assay afterHoxa13 over-expression under the control of adoxycycline-responsive promoter activated with different doses ofdoxycycline (in μg/ml on the x axis). Stars represent the p value ofthe two-tailed Student's t-test applied between thedifferent conditions. **p < 0.01;***p < 0.005. Error bars represent the standarderror to the mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.011
Pbx1 acts as a cofactor regulating cell ingression controlled by anteriorHox genes
Expression of anterior Hox genes in the primitive streak ismaintained during the fast axis elongation phase occurring during the formation ofthe first ten somites, suggesting that there must be a mechanism blocking theireffect on ingression during this time window (Figure1A). TALE (Three Amino-acid Loop Extension) family members have been shownto differentially interact with anterior and posterior Hox genes(; ). In chicken, the onlyTALE gene expressed in PM precursors is Pbx1 which is detected inthe primitive streak from stage 4 to 7 HH (Figure7A, n = 8 embryos for each condition []). Electroporation of a siRNA targetingPbx1 in the epiblast resulted in strong down-regulation ofPbx1 (Figure 7B–C,n = 4 embryos for each condition). In consecutive electroporations performedfirst with Cherry and a control siRNA and then with Venus and a siRNA targetingPbx1, cells electroporated with the Pbx1 siRNAwere found extending more anteriorly than control cells (Figure 7D,K, n = 19 embryos). The effect ofPbx1 siRNA on ingression could be rescued by co-expressingPbx1 (Figure 7E,K, n= 16 embryos). We compared in consecutive electroporations the effect ofexpressing first a control siRNA with either Hoxb7, Hoxb9,Hoxa9, Hoxc9, Hoxd10, Hoxd11, Hoxc11, Hoxa13, Hoxb13 orHoxc13, and then the Pbx1 siRNA with the sameHox gene. Cells co-expressing Hoxb7 orHoxb9 and the Pbx1 siRNA reached more anteriorlevels than cells co-expressing these Hox genes and the controlsiRNA (Figure 7F–G, K, n = 10and 15 embryos respectively). In contrast, cells co-expressing either a control orthe Pbx1 siRNA together with a posterior Hox genewere found to extend up to the same anterior level (Figure 7H–K and not shown, n > 8 embryos for eachcondition). Over-expression of Pbx1 in PM precursors after the3-somite stage slowed down axis elongation (Figure7L and Video 5, n = 12embryos), suggesting that Pbx1 can restore the effect of anteriorHox genes on ingression during this time window. Thus,Hox-dependent control of ingression in the paraxial mesodermrequires Pbx1 for anterior but not for posteriorHox genes.
Video 5.
Effect of Pbx1 over-expression between the 5- and 9-somitestage.Bright-field (purple) merged with fluorescent images of PSM cell progenitorselectroporated with either a control pBIC (left panel) or aPbx1pBIC (right panel) construct (green) (ventral view,anterior is up). Over-expression of Pbx1 slows down axiselongation.
DOI:http://dx.doi.org/10.7554/eLife.04379.016
Control of ingression of PM precursors by anterior Hoxgenes is dependent on Pbx1.(A) Pbx1 expression during somitogenesis. Redsquares: region of PM progenitors. White dashed line: level of transversesection shown in bottom left. (B–C)Pbx1 expression in stage 6–7 HH chicken embryoselectroporated with Venus and control siRNA (B) orPbx1 siRNA (C). Left panels: Venusexpression. (D–J) 2-day-old chicken embryosconsecutively electroporated first with Cherry and a control siRNA and thenwith a Pbx1 siRNA and a Venus construct either alone(D) or together with Pbx1 (E),Hoxb7 (F), Hoxb9(G), Hoxc9 (H),Hoxc11 (I), Hoxa13(J). Arrowheads: anterior boundary of Cherry (red) and ofVenus (green) domains. (K) Ratio of Venus over Cherryexpressing domains. Dots: electroporated embryos. Bar indicates mean.(L) Effect of Pbx1 over-expression on axiselongation rate. Stars represent the p value of the two-tailedStudent's t-test applied between the differentconditions. ***p < 0.005. Error bars: standarderror to the mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.012
Hox genes regulate axis elongation through collinear repressionof Wnt/βcatenin
To identify effector targets regulated by posterior Hox genes, weelectroporated epiblast PM progenitors in stage 5 HH embryos, either with a controlH2B-Venus or with aHoxa13-IRES-H2B-Venus vector, and we harvestedembryos at 9 somites (Figure 8A).Venus-positive cells were sorted by Fluorescence Activated Cell Sorter (FACS)following tail dissociation and their transcriptome was analyzed using Affymetrixmicroarrays (Figure 8A, n = 2 ×2 arrays for each condition). The Wnt/βcatenin pathway targets, Axin2,Fgf8, and Sp8 were down-regulated inHoxa13 over-expressing cells (Table 1, Supplementary file 1, Figure 8B)suggesting that posterior Hox genes might control axis elongationrate by progressively down-regulating the Wnt/βcatenin pathway. To test thishypothesis, we first performed in situ hybridizations (ISH) for Axin2,Fgf8, and T/Brachyury that show thattheir expression in the tail-bud is down-regulated when Hoxa13becomes activated (Figure 8—figuresupplement 1A–D, n = 8 embryos for each condition). Since theISH technique is not quantitative enough to resolve slight differences, we performedquantitative Reverse Transcription PCR (qRT-PCR) on micro-dissected tail-buds from10, 15, 20, and 25-somite stages for Axin2, Fgf8, andT/Brachyury. These experiments show a slightprogressive down-regulation of these genes from the 10 to 20-somite stage followed bya significant decrease in gene expression at the 25-somite stage correlating with theslowing down of axis elongation as well as with the timing of posterior genesexpression (Figure 8C–E, n = 5embryos for each stage). Co-electroporation of Hoxd10, Hoxc11, orHoxa13 with βcatLEF (which activates theWnt/βcatenin pathway []) rescues axis elongation (Figure8F–H, Videos6–8, n = 41). We co-electroporated a Wnt/βcateninfirefly luciferase reporter (BATLuc) and a CMV-Renilla luciferase construct in PMprogenitors together with either Venus or Hoxa9,Hoxc9, Hoxd10, Hoxd11,Hoxc11, Hoxa13, Hoxb13, or Hoxc13. TheseHox genes induced a down-regulation of luciferase activity whichincreased in a collinear fashion (except for Hoxd10 andHoxd11 which showed a weaker effect) (Figure 8I and Figure8—figure supplement 2A, n = 83 embryos). All together, theseresults strongly suggest that the posterior Hox genes control axiselongation by modulating Wnt/βcatenin signaling activity. When co-expressingHoxa9 and Hoxa13, the Wnt-repressive effect wasequivalent to that of Hoxa13, indicating that posterior prevalencealso applies to Wnt repression (Figure6D–E, n = 30 embryos). By expressing various amounts ofHoxa13, we observed that Wnt repression is independent of thequantity of protein expressed (Figure6G–H, n = 62 embryos), suggesting that Hox proteins levelsare saturating in our experiments. Therefore, the same posterior Hoxgenes can regulate ingression, axis elongation, and Wnt signaling with strikinglysimilar collinear trends.
Video 6.
Activation of Wnt/βcatenin signaling and T over-expression rescueHoxa13 axis elongation phenotype.Bright-field (purple) merged with fluorescent images of PSM cell progenitorselectroporated with Hoxa13mutH-ires2-H2B-Venus (leftpanel), Hoxa13-ires2-H2B-Venus (second panel),T and Hoxa13-ires2-H2B-Venus construct(third panel) or βcatLEF andHoxa13-ires2-H2B-Venus construct (right panel) (green)(ventral view, anterior is up) from Stage 6 HH onwards.
DOI:http://dx.doi.org/10.7554/eLife.04379.017
Video 7.
Activation of the Wnt/βcatenin pathway rescues the axiselongation phenotype due to Hoxd10 over-expression.Bright-field (purple) merged with fluorescent images of PSM cell progenitorselectroporated with Hoxd10mutH-ires2-H2B-Venus (leftpanel), Hoxd10-ires2-H2B-Venus (middle panel) orβcatLEF andHoxd10-ires2-H2B-Venus construct (right panel) (green)(ventral view, anterior is up) from Stage 6 HH onwards.
DOI:http://dx.doi.org/10.7554/eLife.04379.018
Video 8.
Activation of the Wnt/βcatenin pathway rescues the axiselongation phenotype due to Hoxc11 over-expression.Brightfield (purple) merged with fluorescent images of PSM cell progenitorselectroporated with Hoxc11mutH-ires2-H2B-Venus (leftpanel), Hoxc11-ires2-H2B-Venus (middle panel), orβcatLEF andHoxc11-ires2-H2B-Venus construct (right panel) (green)(ventral view, anterior is up) from Stage 6 HH onwards.
DOI:http://dx.doi.org/10.7554/eLife.04379.019
Collinear repression of Wnt/βcatenin signaling by posteriorHox genes.(A) Design of the microarray experiment. (B)Validation by Q-RT PCR of selected Hoxa13 targetsidentified in the microarray experiment. (C-E) Q-RT PCR for(C) T/Brachyruy, (D)Axin2, and (E) Fgf8 at10, 15, 20, and 25-somite stage from microdissected tail-buds.(F–H) elongation velocity of embryosover-expressing (F) Hoxd10mutH,Hoxd10 orHoxd10+βcatLEF, (G)Hoxc11mutH, Hoxc11 orHoxc11+βcatLEF, (H)Hoxa13mutH, Hoxa13 orHoxa13+βcatLEF. (I)Luciferase assay measuring Wnt/βcatenin activity afterover-expression of BATLuc together with CMV-Renilla and either control,Hoxa9, Hoxc9,Hoxd10, Hoxd11, Hoxc11,Hoxa13, Hoxb13, or Hoxc13.(J–M) Luciferase assay measuringWnt/βcatenin activity after over-expression of BATLuc andCMV-Renilla and control, Hoxa13,Hoxa13+dBC, orHoxa13+Lrp6ΔN (J), orcontrol, Hoxa13, Hoxa13+Wnt3a orHoxa13+Wnt5a (K), or control,Fzd2, Hoxa13, orHoxa13+Fzd2 (L) or control andDact2 (M). Firefly luciferase intensityvalues have been normalized to their respective Renilla values (RLU).Controls have been set to 1. Stars: p value of the two-tailedStudent's t-test applied between the differentconditions. *p < 0.05; **p < 0.01;***p < 0.005. Error bars represent standarderror to the mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.021
Figure 8—figure supplement 1.
The Wnt signaling is repressed when posterior Hoxgenes are activated.(A–F) In situ hybridization of 15-somite(left panels) and 25-somite stage (right panels) embryos hybridized withHoxa13 (A),Axin2(B), Fgf8 intronic(C), T intronic (D),Fzd2 (E), and Dact2(F) (red arrowhead: tail-bud) showing a repression of theWnt targets and components as well as an upregulation of the Wntinhibitor Dact2 when Hoxa13 start to beexpressed in the tail-bud.
DOI:http://dx.doi.org/10.7554/eLife.04379.022
Figure 8—figure supplement 2.
Collinear repression of Wnt signaling and cell motility by posteriorHox genes.(A) Graph showing Figure8I samples after removal of control and Hoxd10and Hoxd11 (which have a weaker effect) to highlight thecollinear trend of this set of Hox genes on Wntrepression. (B) Cell motility measured in the posterior PSMof embryos electroporated with H2B-Venus and either a control,Hoxb1, Hoxa5,Hoxc11 or Hoxa13. Stars representthe p-value of the two-tailed Student's t-test applied between thedifferent conditions. *p < 0.05; ***p< 0.005. Error bars represent the standard error to the mean(SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.023
Table 1.
List of selected genes of the Wnt and FGF pathways down-regulated orup-regulated following over-expression of Hoxa13 intail-bud cells
DOI:http://dx.doi.org/10.7554/eLife.04379.024
Gene | Average (control) | Standard Dev (control) | Average (Hoxa13) | Standard dev (Hoxa13) | Fold change |
---|---|---|---|---|---|
Sp8 | 949.9 | 279.2 | 483.8 | 21.4 | 0.51 |
Fzd2 | 139.7 | 10.7 | 78.5 | 8.6 | 0.56 |
Axin2 | 857.8 | 42.5 | 677.0 | 99.1 | 0.79 |
Dact2 | 415.8 | 134.4 | 989.8 | 270.1 | 2.38 |
Cyp26a1 | 625.1 | 258 | 102.9 | 13 | 0.16 |
Fgf8 | 1523.9 | 159.3 | 591.2 | 65 | 0.39 |
Etv1 | 296.6 | 113.2 | 155.1 | 23.8 | 0.52 |
Fgfr1 | 145.9 | 5.8 | 80 | 0.6 | 0.55 |
Rasgrp3 | 1441.3 | 671.8 | 362.8 | 218.8 | 0.25 |
We next analyzed how Hox genes interfere with Wnt function.Hoxa13 ingression phenotype is rescued by an activated form ofLrp6 or a stabilized form of Ctbbn1 (Figure 8J, n = 42 embryos) but not byWnt3a or Wnt5a (Figure 8K, n = 30 embryos). This suggests that, genetically,Hox genes act on Wnt signaling at the membrane level.Over-expression of the Wnt receptor Fzd2 (down-regulated inHoxa13 over-expressing cells (Figure 8B, Table 1 and Supplementary file 1))with Hoxa13 rescued Wnt repression (Figure 8L, n = 29 embryos). Fzd2 is expressed inthe tail-bud at 15 somites and down-regulated after 25 somites (Figure 8—figure supplement 1E, n = 8 embryos).Over-expression of the Wnt pathway component Dact2 (which isexpressed in the tail-bud from 25 somites onward and up-regulated inHoxa13 over-expressing cells [Figure 8—figure supplement 1F, n = 8 embryos, Table 1, Supplementary file 1]),repressed Wnt activity (Figure 8M, n =9 embryos). In Hoxa13 over-expressing cells, the FGF receptorFGFR1, its ligand Fgf8, and its targetsEtv1 and Cyp26A1 as well as the FGF pathwaycomponent Rasgrp3, were down-regulated while the FGF/MAPK inhibitor,Spred2, was up-regulated (Figure8B, Table 1, Supplementary file 1),indicating that Hoxa13 can also inhibit FGF signaling. Thisinhibition is consistent with the down-regulation of PSM cell motility observed afterHoxc11 or Hoxa13 over-expression (Figure 8—figure supplement 2B, n= 20 embryos) (). FGF down-regulation is expected since FGF and Wnt signalingreciprocally regulate each other in PM precursors (; ). Down-regulation of Cyp26A1, which degrades RA, canup-regulate RA signaling leading to repression of the Wnt pathway noncell-autonomously (;; ). Together, these data suggest thatposterior Hox genes act on a gene network converging towardautonomous and non-autonomous negative Wnt regulation.
Gradual repression of T/Brachyury by posteriorHox genes regulates cell ingression and axis elongation
The T-box transcription factor T (aka Brachyury) isa well-characterized Wnt target which has been shown to control cell ingression tothe mesoderm (; ). Q-PCR analysis ofmicro-dissected tail buds shows that T expression levels decrease between 10 and20-somite stage and then significantly drop at the 25-somite stage (Figure 8C). Over-expressing T byelectroporation often resulted in PM-expressing cells extending more anteriorly thancontrol cells suggesting that they ingress earlier (Figure 9A–B, n = 6 embryos). Over-expression ofT together with either Hoxa9,Hoxd10, Hoxc11 or Hoxa13rescued the ingression delay (Figure9A–B, n = 6, 11, 10 and 7 embryos respectively).T also rescued the elongation slow down observed afterHoxa13 over-expression (Video6, Figure 9C, n = 4 embryos).A lower dose of T (0.5 μg/μl) only led to partialrescue of the Hoxa13 phenotype (Figure 9C, n = 4 embryos). Endogenous T expressionis down-regulated in Hoxa13 over-expressing cells FACS-sorted fromelectroporated embryos (Figure 9D, n =2 FACS sorted cell samples for each condition). Over-expression of a reportergenerated by fusing one kilobase of the chicken T promoter to thefirefly luciferase (cTprLuc) together with Hoxc11 orHoxa13 and the CMV-Renilla luciferase show Trepression which is stronger for Hoxa13 (Figure 9E, n = 19 embryos). Over-expression ofβcatLEF leads to T up-regulation (Figure 9F, n = 20 embryos) andco-expression of Hoxa13 with βcatLEF totallyrescues T repression (Figure9F, n = 20 embryos) suggesting that Hox genesdown-regulate T expression by repressing the Wnt/βcateninpathway. Over-expressing T had no effect on BATluc activation (Figure 9—figure supplement 1, n = 8 embryos).This argues that the effect of Hox genes on epiblast ingressioninvolves quantitative regulation of T expression levels.
Hox genes effect on axis elongation involvesBrachyury regulation downstream of theWnt/βcatenin pathway.(A) Consecutive electroporation of PM precursors with Cherryand then with Venus together with T (left panel),Hoxa13 (middle), or a combination of the two vectors(right). Arrowheads: anterior boundary of Cherry (red) and Venus (green)domains. (B) Ratio of Venus over Cherry domains. Dots:electroporated embryos. Bar indicates the mean. (C) Axiselongation velocity of embryos electroporated with control,Hoxa13, or co-electroporated withHoxa13 and either high or low dose ofT. (D) Q-RT PCR quantification ofT expression in control orHoxa13-expressing PM progenitor cells.(E–F) Luciferase activity (RLU) afterover-expression of cTprLuc and CMV-Renilla together with either(E) control, Hoxc11 orHoxa13 or (F) control,βcatLEF, Hoxa13 orHoxa13+βcatLEF.(G–I) Luciferase assay measuringWnt/βcatenin activity after over-expression of BATLuc andCMV-Renilla and (G) Hoxa13mutH, Hoxa13dn,Hoxa13+Hoxa13mutH orHoxa13+Hoxa13dn. (H)Hoxd10mutH, Hoxd10dn,Hoxd10+Hoxd10mutH orHoxd10+Hoxd10dn, (I)Hoxc11mutH, Hoxc11dn, Hoxc11+Hoxc11mutH orHoxc11+Hoxc11dn. (J) Luciferaseassay measuring Wnt/βcatenin activity from 28-somite stagedissected tail-buds after over-expression of BATLuc and CMV-Renillaconstructs and either Hoxa13mutH orHoxa13dn, or Hoxa13mutH withHoxc11mutH or Hoxa13dn withHoxc11dn, or Hoxa13mutH withHoxc11mutH and Hoxd10mutH orHox13dn with Hoxc11dn andHoxd10dn. (K, L) Q-RT PCRquantification of T, Axin2(K), and Fzd2 (L)expression in PM progenitors co-expressing eitherHoxa13mutH with Hoxc11mutH andHoxd10mutH or Hoxa13dn withHoxc11dn and Hoxd10dn. Stars:p-value of the two-tailed Student's t-testapplied between the different conditions. *p < 0.05;**p < 0.01; ***p <0.005. Error bars: standard error to the mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.013
Figure 9—figure supplement 1.
Overexpression of T has no effect on Wnt activity.Luciferase assay measuring Wnt/βcatenin pathway activity 20 hrafter over-expression of BATLuc and Renilla constructs together withcontrol, T, Hoxa13, orHoxa13+T. Stars represent the p-value of thetwo-tailed Student's t-test applied between thedifferent conditions. *p < 0.05; ***p< 0.005. Error bars represent the standard error to the mean(SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.014
A Hoxa13 truncated form is responsible for the dominantHand-Foot-Genital syndrome in man (). A similar truncation in the chicken homolog(Hoxa13dn) acts as a dominant-negative inhibiting the function ofall Hox13 genes (). When over-expressed before activation ofHox13 paralogs, Hoxa13dn had no effect on BATlucactivity (Figure 9G, n = 18 embryos).However, co-expression with Hoxa13 in similar conditions abolishedWnt repression (Figure 9G, n = 18embryos). Similar truncations in chicken Hoxd10(Hoxd10dn) and Hoxc11 (Hoxc11dn) also exert a dominant-negative effect on theirwild-type counterparts (Figure 9H–I, n= 38 embryos). We over-expressed Hoxa13dn alone or combinedwith either Hoxc11dn or with Hoxc11dn andHoxd10dn along with BATLuc and CMV-Renilla constructs in PMprecursors of the streak at stage 8 HH. Embryos were harvested at the 28-somite stagewhen most Hox10-13 paralogs are expressed. Increasing the number ofdominant-negative constructs results in a corresponding increase in luciferaseactivity (Figure 9J, n = 35 embryos).We next co-expressed the three dominant-negative vectors Hoxd10dn,Hoxc11dn, and Hoxa13dn together, along with Venus andFACS-sorted dissociated Venus-positive cells from tail-buds of 28-somite embryos.qRT-PCR analysis of T, Axin2, andFzd2 in the Venus-positive cells shows up-regulation of the threegenes (Figure 9K-L, n = 4 embryos foreach condition). All together these results argue that a subset of posteriorHox genes gradually represses Wnt/βcatenin signaling andconsequently T/Brachyury in paraxial mesoderm precursors of theepiblast. This progressive repression leads to reduced cell ingression and cellmotility in the PSM, resulting in a slowing down of axis elongation.
The N-terminal region of posterior Hox genes but not thehomeodomain is responsible for the repression of T/Brachyury
In order to identify the domain of posterior Hox proteins involved in repressingT/Brachyury expression, we generated chimera proteins where thedifferent regions (N-terminal, homeodomain and C-terminal) of different posterior Hoxproteins are swapped with the equivalent region of Hoxa5 which has no effect on axiselongation, Wnt activity and T/Brachyury expression (Figure 10A–B). Over-expression of cTprLucalong with a chimera where the homeodomain of Hoxa5 has been swapped with the onefrom Hoxa13 (Hoxa5Ha13) does not show any repression of luciferase activity whileover-expression of a chimera where the homeodomain of Hoxa13 has been swapped withthe one from Hoxa5 (Hoxa13Ha5) shows a strong repression of luciferase activity(Figure 10A–B, n = 35embryos) suggesting that the homeodomain does not contain the major domainresponsible for T/Brachyury repression. We nexttested if either the N-terminal domain (N-ter) or the C-terminal domain (C-ter) isresponsible for T/Brachyury repression. Overexpression of a chimerawhere the N-ter of Hoxa5 is swapped with the N-ter of Hoxa13 (NHoxa13HCa5) shows astrong repression of luciferase activity while a chimera where the C-ter of Hoxa5 isswapped with the C-ter of Hoxa13 (Hoxa5Ca13) does not show any repression (Figure 10C–D, n = 16 embryos)suggesting that the N-ter region of Hoxa13 contains the domain responsible for therepression of T/Brachyury. Sequence alignment of the N-terminalregions of Hoxa9, d10, c11, and a13 shows little conservation at the amino acid levelsuggesting that it is not a conserved amino acid domain but rather a structuraldomain that is responsible for the repression activity of these proteins (Figure 10—figure supplement 1). We nexttested if the nature of the homeodomain could have a role in refining the level ofrepression of T by designing chimeras where the homeodomain of Hoxc11 and Hoxa13 wasreplaced by the homeodomain of Hoxa5 (Hoxc11Ha5 and Hoxa13Ha5, respectively) (Figure 10E). With the wild-type proteins, weobserve a stronger downregulation of T/Brachyury with Hoxa13 thanwith Hoxc11 (Figure 9E). Surprisingly, when weoverexpress Hoxc11Ha5 or Hoxa13Ha5 along with cTprLuc, we observe a strongerdown-regulation of T/Brachyury with the chimera containing theHoxc11 N-ter than with the one containing the Hoxa13 N-ter (Figure 10E–F, n = 26 embryos) suggesting that thehomeodomain could be responsible for fine tuning T/Brachyuryrepression. Altogether our data suggest that the progressivedeployment of posterior Hox genes in PM precursors during axiselongation leads to a collinear repression of the Wnt/βcatenin pathway and itstarget T/Brachyury.
The N-terminal region of posterior Hox genes contains the repressivedomain.(A, C, E) Design of the Hoxchimeras. N-ter is in blue, the homeodomain in white, and the C-ter inred. (B, D, F) Luciferase assaymeasuring T/brachyury expression 20 hr after over-expression of cTprLucand Renilla constructs together with (B) control,Hoxa5, Hoxa13, Hoxa5Ha13, orHoxa13Ha5, (D) control, Hoxa13,NHox13HCa5, or Hoxa5Ca13, (E)control, Hoxc11Ha5, or Hoxa13Ha5. Starsrepresent the p-value of the two-tailed Student's t-test appliedbetween the different conditions. ***p <0.005. Error bars represent the standard error to the mean (SEM).
DOI:http://dx.doi.org/10.7554/eLife.04379.025
Figure 10—figure supplement 1.
The N-ter region of posterior Hox genes is poorly conserved at theamino-acid level.ClustalW alignment of the N-ter region of Hoxa9, Hoxd10, Hoxc11, andHoxa13 shows poor conservation at the amino acid level.
DOI:http://dx.doi.org/10.7554/eLife.04379.026
Discussion
Here, we show that a subset of posterior Hox genes represses Wnt/Tsignaling with increasing strength showing a collinear trend. We observe a similarcollinear effect of these posterior genes overexpressed in PM precursors on delayingtheir ingression in the PSM and on the slowing down of axis elongation. This inhibitionof Wnt signaling is accompanied by a down-regulation of FGF signaling which was shown tocontrol elongation velocity by regulating cell motility in the PSM (). Thissuggests that posterior Hox genes are involved in the slowing down ofaxis elongation by acting both on the flux of cells in the posterior PSM and on themotility of PSM cells.
In the chicken embryo, PM precursors originate initially from the lateral epiblast whichmigrate toward the midline during formation of the primitive streak ( ; ). Around stage 4 HH, somite precursorsbegin to ingress from the superficial epiblast of the anterior primitive streak andposterior Node region (; ). Two typesof PM precursors have been identified in chicken and mouse embryos (). A first set derives from theNode/primitive streak border and exhibits long-term self-renewal properties (; , ). These cells express Sox2 andBrachyury and they can contribute both to the PM (mostly to themedial part of the somites) and to the neural tube (Ordahl, 1993; ;; ; ). A second set derives from theanterior portion of the primitive streak and contributes to shorter clones restricted tothe PM (; ; ). After stage 4 HH, in the chicken embryo,the primitive streak begins to regress and after stage 13-14 HH, it becomes part of thetail-bud (). At the 25-somitestage (stage 15 HH), the posterior neuropore closes and the tail-bud becomes enclosedinto the tail fold. During these stages, PM precursors are continuously produced firstby the primitive streak and then by the tail-bud. Fate mapping of the 25-somite stagetail-bud with quail-chick chimeras and diI labeling showed that the formation of the PMfollows morphogenetic movements very similar to that seen earlier at the primitivestreak level during gastrulation (; ). Afterthis stage, the remnant of the primitive streak becomes localized ventrally to form astructure known as the Ventral Ectodermal Ridge (VER) (; ).
Whether cell ingression continues after posterior neuropore closure to generate the PMis not well established. Knezevic et al. reported that cell ingression from the VERstops at stage 16 HH (26–28 somites) () but Ohta et al. demonstrated that ingression into the mesodermcontinues in the VER up to the 40-somite stage (stage 20 HH) (). There is also some lineage continuity at thelevel of the PM precursors of the Node/primitive streak border which were shown tobecome internalized to become part of the chordo-neural hinge in the tail-bud. DiIlabeling of the late chordo-neural hinge in stage 20–22 HH (40–45 somites)embryos showed that mesoderm cells are produced by this structure at late stages (). The cellularorganization of the chordo-neural hinge has not been characterized and whether mesodermproduction by this structure occurs through ingression movements involving an epitheliumto mesenchyme transition as is seen for the production of paraxial mesoderm from theprimitive streak is not established. Overall, very little is known about the movementsof cells in the tail-bud after the 25-somite stage in chicken and mouse embryos. Therespective contribution of the VER and the CNH to the PM at these late stages has notbeen characterized.
If Knezevic et al. are correct, it could be that the action of posterior Hox genes oningression ends at the 25-somite stage when Hoxa13 is first expressed.The strong effect of Hoxa13 on ingression might trigger the arrest ofcell ingression leading to the slowing down of axis elongation observed at this stage.The resulting imbalance between the velocity of somitogenesis and of axis elongationcould account for the progressive shortening of the PSM observed during the productionof the next 25–28 somites. Alternatively, it could be that, as suggested by Ohtaet al. and by Olivera-Martinez et al., ingression continues up to the 40–45somite stage. In this case, Hoxa13 expression at the 25-somite stagewould only significantly reduce the rate of cell ingression into the PM. At the40–43 somite stage, Hoxb13 and Hoxc13 becomeexpressed in the tail-bud potentially terminating further ingression. Whether Hox13genes might regulate the late ingression of PM at the level of the VER, as shown by Ohtaet al., or at the level of the CNH as proposed by Olivera-Martinez remains to beestablished. In both cases, however, the PSM is expected to shrink in response to Hox13genes.
Even though our experiments do not directly address the process whereby axis elongationstops, they suggest that Wnt and FGF repression in the tail-bud, which signalstermination of axis formation (), could be mediated by posterior Hox genes. Byreducing the flux of cells to the PSM and their motility, posterior Hoxgenes can indirectly control its progressive shortening. Furthermore, the inhibition ofFGF and Wnt signaling which are required for the segmentation clock oscillationsprovides an explanation for the arrest of somite formation before the completeexhaustion of the PSM described in avians (Bellairs,1986; ). In vivo,the downregulation of the FGF target Cyp26A1 downstream of Hox13 genes(this report, ) would leavethe tail-bud more vulnerable to the increase of RA. Whether, the raise in RA levelscaused by bringing the segmented region closer is also responsible forraldh2 activation in the late tail-bud remains to be explored. Insuch scenarios, posterior Hox genes indirectly control the terminationof axis elongation and hence the segment number in the chicken embryo.
In mouse embryos, over-expression of Hox13 genes results in axis truncation posterior tothe thoracic level ().Remarkably, overexpression of Hoxa13, b13, andc13 from the same promoter in transgenic mice results in truncationsat different antero-posterior levels (), arguing for different truncation efficiency of the mouse Hox13proteins. This is highly reminiscent of our observations showing different quantitativeeffects of the overexpression of the same three Hox13 genes in chicken embryos.Duplications and deletions of regions of the mouse Hoxd cluster lead toheterochronic expression of posterior Hoxd genes in the tail-bud yetthey do not seem to alter segment numbers (; ; ; ). This is also consistent with ourobservations that Hoxd genes have limited effect on axis elongation inour experiments.
In transgenic mice overexpressing Hoxc13, Wnt targets and the FGFtarget Cyp26A1 were also found to be down-regulated () as observed in chicken embryosoverexpressing Hoxa13. This argues for a conserved role of posteriorHox proteins in the repression of the Wnt and FGF pathway between chicken and mouseembryos. In mouse embryos however, no strong raldh2 expression or lateRA production is detected in the tail bud () and axis elongation continues for a longer time resulting in tailformation. Moreover, raldh2 −/− mouse embryos which lackRA production during posterior body formation can form normal tails, suggesting that RAis not involved in axis termination in mouse (). In mouse embryos, initially reported an arrest of ingression whenthe posterior neuropore closes at the 30-somite stage (), but subsequently provided evidence for continued ingression ofcells in the PM after this stage (). Thus, in mouse embryos, termination of axis elongation could simplyresult from exhaustion of PM progenitors caused by the slowing of axis elongationtriggered by posterior Hox genes acting on cell ingression andmotility. Remarkably, among amniotes, many species such as lizards, rodents, or monkeysbear a long tail whereas others such as birds or humans do not. Closely related speciessuch as monkeys and apes can differ by the presence of a tail suggesting that thegenetic switch involved in the control of tail formation is quite simple. Whether thisswitch involves an RA-dependent elongation arrest mechanism as seen in chicken andwhether this control depends on posterior Hox genes is an attractivepossibility which remains to be investigated.
Our work provides evidence for functional collinearity in the control of axis elongationby posterior Hox genes. Our data also suggest that our overexpressionconditions are saturating (Figure 6), abolishingany effect of gene dosage of the overexpressed Hox genes. This confirmsprevious results published in showing that Hoxb1-9 gene expressiondriven by promoters of different strength (CMV, TK, and CAGGS) leads to similaringression phenotypes. Together, this suggests that the information driving thequantitative effects of Hox proteins on Wnt repression, cell ingression, and elongationis built in the structure of the proteins themselves rather than reflecting the actualamounts of Hox proteins present. This functional collinearity might be related to therecently described structural collinearity of binding specificities reported for fly Hoxproteins ().
Our work suggests that low amounts of posterior Hox protein levels could be saturatingin vivo. This is consistent with the analysis of paralog knock-out experiments showingthat leaving only one single wild-type allele leads to a much milder phenotype than thedeletion of an entire paralog group (; ). Also, increasing Hox doses by adding an extra mouse or human HoxD clusterdoes not alter the vertebral formula (; ; ; ). The fact that low levels ofHox proteins are saturating could confer great robustness to thesystem consistent with the extreme stability of intraspecific vertebral formula. That 8of the 16 posterior Hox genes from all posterior paralog groups exceptHox12 show an effect in the ingression, elongation, and Wnt signaling assays argue foran extreme redundancy of the system that could further explain the intraspecificrobustness of the vertebral formula.
We observe a trend showing an increasing strength of the effects on cell ingression, Wntrepression, and axis elongation when overexpressing progressively more 5′ Hoxgenes. This increasing trend can be partly accounted for by the posterior prevalence ofposterior Hox genes observed in the control of ingression and in therepression of Wnt signaling for genes of different paralog groups. However, differentquantitative effects were observed for genes from the same paralog groups arguingagainst a simple posterior prevalence model. This is in line with the result ofinactivation of the entire paralogs groups such as Hox10 orHox11 which demonstrates specific properties of each of theseparalog groups arguing against a simple posterior prevalence model functioning invertebral patterning (; ).
We identify a role for the TALE protein Pbx1 in the control of cell ingression from theepiblast into the PSM by anterior but not posterior Hox genes. TALEhomeoproteins have been shown to act as co-factors able to enhance DNA bindingspecificity of Hox genes (). Pbx proteins bind anterior Hox proteins via a specific hexapeptidesequence (). The nullmutation of Pbx1 in mouse leads to patterning defects of the axialskeleton but axis length appears essentially normal (). In contrast, double mutants for Pbx1and Pbx2 often show a smaller number of somites suggesting that thesetwo genes could act redundantly in patterning the axial skeleton (). In Pbx1−/−;Pbx2+/− mutants, anterior shifts of Hoxexpression boundaries in the paraxial mesoderm have been reported (). Such shifts are consistent with aprecocious ingression of cells normally fated to a more posterior identity.
Genetic studies on mouse T mutants have shown that graded T activity is required forbody axis formation (; ). Embryos withprogressively lower quantities of T exhibit more severe axis truncations (). Similar graded truncationsare also observed for Wnt3a allelic series (), indicating that precise quantitativeregulation of this pathway is required for completion of body axis elongation.Repression of the Wnt pathway and of T together with axis truncationswas also observed in Hox13 over-expressing transgenic mice (). Our data suggest that thegradient of T activity is established by the graded regulation of Wnt signaling byposterior Hox genes, (Figure11) thus providing a possible explanation for these complex phenotypes. At thecellular level, it argues that the Hox-dependent regulation of T levelsin the epiblast is critical to control the balance between cell ingression andmaintenance of a self-renewing paraxial mesoderm progenitor pool in theepiblast/tail-bud. Cell ingression requires an EMT that involves destabilization of thebasal microtubules of epiblast cells followed by basal membrane breakdown (). Inhibiting Rhoa activity canrescue the ingression delay caused by Hoxa13 overexpression, suggestingthat posterior Hox genes can control cell flux to the PM by acting onbasal microtubule stabilization in epiblast cells. As T is also able to rescueHoxa13 phenotype on elongation, it could act upstream of thisprocess and the details of such a molecular pathway remain to be investigated.
Model representing the 3 phases (I, II, and III) of Hoxaction in PM precursors in the epiblast/tail-bud during axiselongation.Model representing the 3 phases (I, II, and III) of Hox actionin PSM precursors in the epiblast/tail-bud during body axis elongation.Anterior Hox genes (paralogs 1–9) are expressed duringphase I. They control cell ingression in a Pbx1-dependentmanner leading to the collinear positioning of Hox genesexpression domains in the anterior region of the embryo. NoHox genes are activated during phase II, allowing fastelongation of the embryonic axis. During phase III, posteriorHox genes (paralogs 9–13) are collinearly activatedin PSM precursors. Our data suggest that collinear activation of posteriorHox genes leads to repression of Wnt signaling and itstarget T/Brachyury, which progressively increases in strength.This results in a progressive arrest of cell ingression in the PSM, leading toa decrease in axis elongation rate. Since the velocity of somite formation isroughly constant, PSM size starts to decrease when elongation velocity becomesslower than that of somite formation. During this latter phase the control ofcell ingression by posterior Hox genes appears to beindependent of Pbx1.
DOI:http://dx.doi.org/10.7554/eLife.04379.027
Wnt signaling was proposed to promote the paraxial mesoderm fate at the expense of theneural fate in a population of bipotential neuro-mesodermal stem cells in the tail bud(, ; ; ).Thus, the Wnt repression experienced by epiblast cells in response to posteriorHox genes overexpression could induce these cells toward a neuralfate hence preventing them to ingress. However, Hoxa13 overexpressiondoes not lead to up-regulation of the neural marker Sox2 inelectroporated cells as detected by in situ hybridization and in our microarrayanalysis. Furthermore, electroporated cells are seen to enter the PM and do not enterthe neural tube (Figure 4 and supplementaryvideos). This therefore suggests that posterior Hox genes are unlikelyto control cell ingression by promoting acquisition of a neural fate in epiblast cells.While we cannot completely rule out that a subpopulation of these cells remains in thetail-bud as Sox2-positive cells, it is unlikely that this contributesto the dramatic axis elongation slow down observed after posterior Hoxgenes overexpression.
Thus our data suggest that disruption of the balance between epiblast and ingressingcells can be achieved by interfering with T levels either directly orindirectly by altering Wnt levels or Hox expression. The graded repressive activity ofposterior Hox genes on the Wnt/T pathway might provide an evolutionaryconstraint that led to the selection of collinearity of posterior Hoxgenes ().
Materials and methods
Chicken embryo culture ex ovo
Fertilized chicken eggs were obtained from commercial sources. Eggs were incubated at38°C in a humidified incubator for approximately 24 hr. Embryos were preparedfor Early Chick (EC) cultures () and then electroporated. Embryos were staged following the Hamburgerand Hamilton (HH) table ()and by counting somites (somite stage: ss).
Electroporation
Electroporation of the paraxial mesoderm (PM) precursors of the epiblast was carriedout as described in . For the axis elongation assay, the electroporation success is firstmonitored 3 hr after by examining the embryos under a fluorescent stereomicroscope.Only embryos successfully electroporated in the PSM progenitors (90–100%) areprocessed for videos (see the axis elongation measurement section). For theluciferase assay embryos are electroporated in the PM progenitors and examined 20 hrafter electroporation (see the Luciferase assay section). In both assays, we finallyobtain 90–100% of the electroporated embryos showing reporter expressionrestricted to the paraxial mesoderm. In rare cases (less than 10%), we observed fewcells in the lateral plate mesoderm. These embryos were discarded. To betterillustrate the accuracy of the electroporations performed, we now present a videoshowing the expression of a control construct electroporated in the anterior epiblastand showing the ingressing PSM cells (Video3).
Consecutive electroporation
To over-express two sets of constructs in different somitic precursors of theepiblast, two consecutive electroporations of the anterior primitive streak (PS) werecarried out. Embryos were first prepared for EC culture at room temperature whichpause their development. The first construct mixed with the control vectorpCX-MyrCherry (gift from X Morin) was microinjected on one side of the PS groove andthe first electroporation was carried out as described above right after injection.Only 10 s after the first electroporation, the second construct, containing the geneto over-express (Hox, T…) cloned in thepCAGGS-I2-MyrVenus, was then microinjected at the same level on the other side of thePS groove and immediately electroporated. This procedure targets the entire paraxialmesoderm territory of the epiblast of the anterior primitive streak on theelectroporated side. Thus, in these experiments, we only track the timing ofingression of the most anterior epiblast cells which give rise to the most anteriorlylocalized progeny in the paraxial mesoderm. By biasing the position of the electrodeon the right or on the left side of the primitive streak, we can ensure that theelectroporation is biased on one side allowing easier observation of the anteriorboundaries of the over-expressing cells. In most cases, however, some electroporatedcells are seen on both sides despite this bias. The control Cherry vector is usuallyfound to be more bilateral than the Hox expressing Venus-positivecells as can be seen from the pictures, which reflects an effect specific toHox genes. Inverting the order of the electroporated plasmids inconsecutive electroporation has no effect on the outcome of the experiment, andsimilar results are observed when the entire somitic territory of the epiblast of theanterior streak is electroporated. Following consecutive electroporation, embryoswere cultured in a humidified incubator at 38°C to resume their development. 3hr after electroporation, the embryos were screened based on fluorescence to ensurethat both constructs were expressed and that the correct region was targeted for eachconstruct. At this stage, up to 70% of the embryos are successfully electroporated.The embryos were then reincubated at 38°C until they reach the 15-somite stage(∼20 hr). As can be seen on the embryos shown in Figure 3, Figure 6, andFigure 9 or in the videos, virtually nocells are seen outside of the paraxial mesoderm meaning that the electroporationaccurately targeted the anterior streak epiblast. Since expression of the constructsis driven by the ubiquitous CAGGS promoter, even a slight inaccuracy in positioningthe electrode would result in expression in the neural tube or the lateral plate.
A fluorescent stereomicroscope (Leica M205 FA) equipped with a color camera was usedto track anterior boundaries of both control Cherry-expressing cells and mutantVenus-expressing cells. At this stage, up to 60% of the electroporated embryos areexclusively electroporated in the PSM and retained for further analyses (∼10%are discarded because of developmental defects due to the consecutive electroporationprocedure). We used the ‘Measure’ plugin of ImageJ to measure thedistance between the tail-bud and anterior boundary of both Venus andCherry-expressing domains in the same embryos. We used these measures to calculatethe ratio of Venus over Cherry domains. The Dot-Plot resulting from these ratios wasgenerated using Graphpad 5 (Prism).
Quantitative analysis of cell ingression
Paraxial Mesoderm (PM) progenitors in the anterior primitive streak wereelectroporated in Stage 5 HH embryos with either a control vectorpCAGGS-Venus or pCAGGS-Hoxa13-IRES2-Venus andcultured in a humidified chamber at 38°C for 5 hr. Embryos were then fixed in4% paraformaldehyde at room temperature for 40 min and immunolabeled for GFP,laminin, and DAPI as described below. Embryos were then mounted and imaged with aZeiss 510 NLO equipped with a 20× NA 0.8 objective. 80 µm z-stack wereacquired (one section every 0.42 µm), and cells were scored with respect totheir position in the primitive streak or in the epiblast after 3D reconstruction andoptical transverse sections using Imaris software. Epiblast cells were counted as‘non-ingressed’ and primitive streak/mesodermal cells as‘ingressed’.
RNA in situ hybridization and probes
Whole mount RNA in situ hybridizations were carried out as described (). Pictures of wholeembryos were made using a macroscope (Z16APOA, Leica) with a 1× planapoobjective (Leica) and a high resolution color camera (DFC 420C, Leica). ChickenPbx1 RNA probe is described in . The Fgf8 intronic probe isdecribed in . A 750-bp fragment of the coding sequence (from nucleotide 301 to1061) of chicken Fzd2, the last 800 bp of the coding sequence ofchicken Dact2, intron 6 of chicken T, a 948-bpfragment of cSox2 coding sequence (gift from B Pain) were used asprobes. Chicken Hox RNA probes were Hoxa2 (), Hoxa3, b3,d4 (gift from R Krumlauf), Hoxa10, a11, a13, c4, c5, c6, c8, c9,d8, d9, d10, d11, d12, andd13 (gift from C Tabin), Hoxb1, b4, b7, b9(described in ), Hoxb8 (gift from A Kuroiwa), Hoxb5, c10,c11, c12, c13, and b13 (cloned by PCR using ENSEMBLsequence informations), Hoxa4 (chEST 427p4, Geneservice),Hoxa5 (chEST 382m24, Geneservice), Hoxa6 (chEST338h20, Geneservice), Hoxa7 (chEST 259o11, Geneservice),Hoxa9 (chEST 333f16, Geneservice), Hoxb2 (chEST194e4, Geneservice), Hoxb6 (ChEST147L22, Geneservice),Hoxd3 (chEST 195d1, Geneservice).
Plasmid construction
Full-length coding sequences for chicken Hox genes,Pbx1, T, Wnt3a, Wnt5a,Dact2, and mouse Fzd2 were PCR-amplified from chicken ormouse cDNAs using the proofreading Accuprime pfx DNA polymerase(Invitrogen, Grand Island, NY). PCR fragments were then cloned in either, GrandIsland, NY P221 (Invitrogen) or pENTR-D/TOPO (Invitrogen) to generate gateway(Invitrogen) entry clones. The constitutively active version of lef1(βcatLEF) (gift from R Grosschedl) (), a dominant activated form of Lrp6(Lrp6ΔN) (gift from S Aaronson) (), a stabilized form of Ctbbn1 (dBC) (), and a dominant negativeform or Rhoa (DN-Rhoa) (Gift from P Kulesa) was PCR-amplified andsub-cloned in pENTR-D/TOPO (Invitrogen). Hoxa13, Hoxc11, andHoxd10 mutated versions unable to bind DNA (HoxmutH) weregenerated by mutating amino acids 50, 51, and 53 of the homeodomain to alanine (). When over-expressed inparaxial mesoderm precursors, these HoxmutH constructs show no effect on cellingression, elongation velocity, and Wnt signaling (data not shown). Thedominant-negative forms of Hoxd10, c11, anda13 (respectively Hoxd10dn, Hoxc11dn andHoxa13dn) were generated by inserting a stop codon instead of theamino acid 50 of the homeodomain. Chimeras for Hox genes were generated by fusionPCR. The homeodomain sequence of Hoxa13 was fused to the N-ter and C-ter of Hoxa5 togenerate Hoxa5Ha13. The homeodomain of Hoxa5 was fused to the N-ter and C-ter ofHoxa13 to generate Hoxa13Ha5. The N-ter of Hoxa13 was fused to the homeodomain andC-ter of Hoxa5 to generate NHoxa13HCa5. The Cter of Hoxa13 was fused to the Nter andhomeodomain of Hoxa5 to generate Hoxa5Ca13. The homeodomain of Hoxa5 was fused witheither the Nter and Cter of Hoxc11 or the Nter and Cter of Hoxa13 to generateHoxc11Ha5 and Hoxa13Ha5, respectively. The chimeras were cloned in pENTR-D/TOPO togenerate entry clones. Entry clones were then cloned in destination vectors(depending on the experiments) using Gateway technology (Invitrogen).
For consecutive electroporations and luciferase assays, a pCAGGS-IRES2-Venus-RFAdestination vector was generated as follows: a yellow fluorescent protein (YFP),Venus, with two sites of myristoylation that target the fluorescent protein to themembrane (Venus, gift from K Hadjantonakis) (), was fused to an Internal Ribosomal Entry Site (IRES2)(Clontech) by PCR. The primers used contained an EcoRI site in 5′ and a NotIsite in 3′. The EcoRI/IRES2-Venus/NotI fragment was then cloned into theEcoRI-NotI restriction sites of pCAGGS. A Gateway cassette (RFA, Invitrogen) was theninserted into the EcoRV site of the pCAGGS-I2-Venus, upstream of the IRES2.
For axis elongation measurements and cell tracking experiments, a pCI2HV-RFAdestination vector was generated as follows: a YFP protein, Venus, was first fused tothe full-length coding sequence of histone H2B to target the fluorescent protein tothe nucleus (H2B-Venus). The H2B-Venus PCR fragment was then fused by PCR to an IRES2(Clontech). The primers used contain an EcoRI site in 5′ and a NotI site in3′. The EcoRI/IRES2-H2B-Venus/NotI fragment was then cloned into theEcoRI-NotI restriction sites of pCAGGS. A Gateway cassette (RFA, Invitrogen) was thencloned in the EcoRV site of the pCI2HV, upstream of the IRES2.
For luciferase assay experiments, the chicken T promoter (1 kbupstream of the ATG) was PCR-amplified and cloned upstream of thefirefly luciferase in the pGL4.10 (luc2) vector (Promega) togenerate the cTprLuc reporter. Expression driven by this promoter fragment in chickenembryo recapitulates the PM expression of T (not shown). TheWnt/βcatenin pathway activity reporter (seven TCF/LEF binding sites +siamois minimal promoter) was PCR amplified from the BAT-GALplasmid (Addgene plasmid 20889) () and cloned upstream of a firefly luciferase in thepGL4.10(luc2) vector (Promega) to generate the BATLuc reporter.
Pbx1 siRNA
RNA interference experiments were performed using 21-nucleotide dsRNAs (Dharmacon,Option A4). To identify electroporated cells, siRNAs (suspended in TE to a finalconcentration of 5 mg/ml) were mixed with a pCAGGS-Venus or Cherry expression plasmid(1.0 mg/ml). The target sequence against chick Pbx1 was as follows:5′- GTGTGAAATCAAAGAGAAA-3′. As a control siRNA, we used a siRNAtargeting chick Pbx1 containing two point mutations (underlined inthe sequence):5′-ACACAAAGCTGAAGAAGTA-3′that show no effect on Pbx1 expression.
To monitor the Pbx1 siRNA efficiency, the anterior primitive streakof stage 4 HH embryos was electroporated with either control siRNA orPbx1 siRNA mixed with a pCAGGS-Venus expression plasmid (1.0mg/ml). Embryos were reincubated at 38°C until they reach stage 7 HH when theywere harvested and processed for ISH for Pbx1 and immunofluorescenceagainst GFP.
Pbx1 over-expression
A pBIC control vector (derived from the pBI-tet [clontech] in which Cherry has beencloned) (gift from J Chal) that allows simultaneous expression of two proteins at thesame level once activated by doxycycline (Tet-on, Clontech) or the pBIC vectorcontaining the full length Pbx1 along with a vector expressing thertTA (Clontech) were electroporated in PSM progenitors at Stage 5 HH. Embryos werereincubated until they reach the 3-somite stage. They were then placed on imagingplates containing 0.5 μg/ml doxycycline for 1 hr at 38°C beforestarting acquisition. Axis elongation measurements were performed as described belowbetween 5 and 9-somite stages.
Time-lapse microscopy
Electroporated embryos were cultured ventral side up on a microscope stage. We used acomputer controlled, wide-field (10× objective) epifluorescent microscope(Leica DMR) workstation, equipped with a motorized stage and cooled digital camera(QImaging Retiga 1300i), to acquire 12-bit grayscale intensity images (492 ×652 pixels). For one embryo, several images at different focal planes and differentfields were captured at a single time-point (frame). The acquisition rate used was 10frames per hour (6 min between frames). Image processing, including focal plane‘collapsing’ field merging and registering, was performed to createhigh-resolution, 2D time-lapse sequences for cell tracking and axis elongationmeasures (see ,for details). To correct for the gradual drift of the embryo position or suddenchanges due to repositioning of the microscope stage, images were registered to theembryo center.
Axis elongation measurements
Variation of the distance between a formed somite and the node was used to determinethe velocity of body axis elongation. The coordinates of the different points weredetermined on bright-field images of the time-lapse experiments using the cellulartracking option of ImageJ. ImageJ is a public domain, Java-based image processingprogram developed at the National Institutes of Health.
For wild-type embryo measurements, axis elongation velocity was measured between 1and 3 somites (n = 8), between 5 and 7 somites (n = 8), between 9 and11 somites (n = 6), between 15 and 17 somites (n = 5), between 20 and22 somites (n = 6), and between 25 and 27 somites (n = 8). For15–17 somites measurements, embryos were cultured starting at 13 somites andimaged until 18 somites. For 20–22 somite measurements, embryos were culturedstarting at 18 somites and imaged until 23 somites. For 25–27 somitesmeasurements, embryos were cultured starting at 23 somites and imaged until 28somites. For measurements of axis elongation velocity after Hox orT over-expression, electroporated embryos at stage 5 HH werecultured in a humidified incubator at 38°C for 3 hr and then placed on themicroscope stage, as described above, for 18 hr. Axis elongation velocity wasmeasured for 10 hr, starting from the 5-somite stage. Student's t-tests wereapplied to evaluate the differences between conditions.
Cell tracking
Cells electroporated with either a control or a Hox gene and anuclear fluorescent protein (H2B-Venus or H2B-GFP) were automatically tracked usingthe Imaris software's cell tracking module (version 7.3.1). Cells weresegmented based on nucleus size (set at 5 μm) and fluorescence intensity. Thetracking algorithm was based on Brownian motion. Only cells in the posterior PSM weretracked for 10 hr. To substract the tissue motion to the single cell motion, theaverage speed of all tracked cells, that represent the tissue motion, has beensubstracted from the average speed of each individual cell (as described in ). Studentt-tests were applied to evaluate the differences recorded between the differentconditions.
Luciferase assay
Embryos were harvested at stage 5 HH and electroporated with a DNA mix containingeither cTprLuc or BATLuc (1 μg/μl final), CMV-Renilla (Promega,Madison, WI) (used as a control to normalize the differences of electroporationintensity between embryos [0.2 μg/μl final]), a control pCAGGS-Venusvector (gift from K Hadjantonakis) or a gene of interest cloned in pCAGGS-IRES2-Venus(5 μg/μ; final). Electroporated embryos were cultured in a humidifiedincubator at 38°C for 20 hr. Embryos were analyzed using a fluorescentmicroscope and only embryos showing restricted expression of Venus in the paraxialmesoderm were selected (90–100% of the electroporated embryos) for luciferaseassay (between 3 and 5 embryos for each condition). The posterior region (from somite1 to tail-bud) of the selected embryos was dissected and lysed in passive lysisbuffer (Promega) for 15 min at room temperature. Lysates were then distributed in a96-well plate and luciferase assays were performed using a Centro LB 960 luminometer(Berthold Technology, France) and the dual luciferase kit (Promega) followingmanufacturer's instructions. Raw intensity values for Firefly luciferasesignal were normalized with corresponding Renilla luciferase values (RLU) and thecontrol experiment was set to 1. Student t-tests were applied to evaluate thedifferences between conditions.
For Hox dominant-negative experiment, embryos were electroporated atst8 HH with a mix containing BATLuc, CMV-Renilla and either aHoxa13mutH or a mix of Hoxc11mutH andHoxa13mutH or a mix Hoxd10mutH,Hoxc11mutH and Hoxa13mutH (in pCAGGS-I2-Venus[control condition]) or Hoxa13dn, or a mix ofHoxc11dn and Hoxa13dn or a mix ofHoxd10dn, Hoxc11dn and Hoxa13dn (inpCAGGS-I2-Venus [mutant condition]). Embryos were reincubated until they reach the28-somite stage. The tail-bud of each embryo was dissected and used for theluciferase assay as described above.
Histology, immunohistochemistry, and imaging
Stage 5 HH embryos electroporated with either a control pCI2HV or apCI2HVHoxa13 vector were cultured in a humidified incubator at38°C for 6 hr. Embryos were then selected using a fluorescencestereomicroscope based on electroporation efficiency. Selected embryos were fixed for30 min at room temperature and then cryo-preserved in 30% sucrose in PBS at4°C. Embryos were then transferred in a solution containing 7.5% gelatin and15% sucrose in PBS and placed at 42°C. Embryos were then included in acryosection mold and flash frozen in a dry ice-ethanol bath. 12-μm transversecryosections of the electroporated region were prepared using a Leica CM3050 Scryostat. Sections were collected on superfrost slides and stored at−20°C. For immunocytochemistry, sections were placed in warm PBS(42°C) for 5 min to remove gelatin. Sections were incubated with the primaryantibody in PBS/BSA (2%)/Triton (0.1%) for 2 hr in a humidified chamber at roomtemperature. Slides were then washed four times for 15 min in PBS and incubated withthe secondary antibody in PBS/BSA (2%)/Triton (0.1%) for 45 min in a humidifiedchamber.
For the cell ingression assay and the labeling of the extracellular matrix (ECM), weused, respectively, a rabbit anti-GFP (abcam, #ab290, UK) at 1/2000 and themouse anti-laminin (DSHB, #3H11, Iowa City, IA) at 1/200. The secondaryantibodies were anti-rabbit Alexafluor488 (Invitrogen) and anti-mouseIgG1Alexafluor555 (Invitrogen), respectively, used at 1/1000. DAPI (Invitrogen,1/1000 dilution) and an Alexafluor 633 phalloidin (Invitrogen) were applied at thesame time as the secondary antibodies to label the nuclei and the F-actin,respectively. For tubulin labeling, we used the mouse anti-acetylated alpha-tubulin(sigma T6793) at 1/1000. The secondary antibody was an anti-mouse IgG2b Alexafluor546(Invitrogen), used at 1/1000. Slides were mounted in Fluoromount-G (SouthernBiotech)and analyzed with a LSM 510 NLO inverted confocal microscope (Carl ZEISS, Germany)using a plan apochromat 63× (NA 1.4) immersion (oil) objective (CarlZeiss).
Hoxa13 protein quantification
A pBIC control vector (described above) or the pBIC vector containing the full-lengthHoxa13 with a C-terminal HA tag along with a vector containingEGFP under the CAGGS promoter and a vector expressing the rtTA (Clontech, France)were electroporated in the PM progenitors at stage 5 HH. A drop of 50 μl ofdifferent doses of doxycyclin (from 50 μg/ml to 0.5 μg/ml) was appliedon top of the embryos immediately after electroporation and the embryos werereincubated for 20 hr. Three embryos for each condition were individually lysedfollowing standard procedure and each lysate was loaded on a different well of anSDS-page gel. Western blot analysis was done following standard procedure. Ananti-HA-HRP antibody was used to detect Hoxa13 (Roche #12013819001, dilution1/1000, Germany). An anti-GFP antibody (abcam ab6556, dilution 1/2000) was used todetect GFP from the pCAGGS-EGFP used as an electroporation control. An anti-βactin antibody (Sigma A5441, dilution 1/5000, Germany) was used to verify that thesame amount of tissue was loaded in each well. This experiment has been repeatedtwice independently.
Cell proliferation analysis
A 20 μl drop of 100 µM EdU (Click-iT EdU kit, Cat. #C10083Invitrogen) was applied on the posterior region of 20–22- and25–27-somite stage embryos cultured in vitro for 45 min. Embryos were thenimmediately fixed in 4% paraformaldehyde (PFA) for 45 min at room temperature (RT)and were then processed as described in . Phospho-histone H3 (pH3) (Millipore, #06-570, 1/1000dilution, France) immunolabelling was performed after the EdU reaction. Single planesections were generated, and the PSM region was manually segmented. For the tail-budproliferation assay, parasagittal cryosections (20 µm) were made. Nucleilabeled with DAPI and EdU and/or pH3 were manually counted. Sections were imagedusing a Zeiss 510 NLO and a 20× dry NA0.8 objective.
Apoptosis quantification
Embryos were harvested at 20–22- and 25–27-somite stage and processedas described ()using the ApopTag Red In Situ kit (#S7165; Millipore). Single plane sectionswere generated and the PSM region was manually segmented. For tail-bud apoptosisassay, parasagittal cryosections (20 µm) were made. Nuclei labeled with DAPIand/or apoptotic labeling were manually counted. Labeled embryos were imaged using aZeiss 510 NLO and a 20× dry NA0.8 objective.
Microarray analysis
PM precursors of the anterior primitive streak of Stage 5 HH embryos wereelectroporated as previously described either with a control vector coding for aH2B-venus fusion (pCI2HV) or a vector coding forHoxa13 and a H2B-venus fusion(Hoxa13pCI2HV). Embryos were reincubated for 14 hr in ahumidified incubator at 38°C until they reach the 9-somite stage. The regioncontaining the PM progenitors was dissected from several embryos and pooled in a dropof PBS/FCS1% (seven embryos per condition) on ice. Dissected tissues were thentransferred in a drop of diluted trypsin and incubated at 38°C for 10 min toallow efficient enzymatic dissociation of cells. Cell dissociation was completedmechanically by pipetting up and down. Cells were then transferred into 500 μlof PBS/FCS 1% on ice and sorted based on YFP fluorescence using a FACS DIVA (BDtechnologies, France). For each condition, one thousand YFP+ cells werecollected directly in Trizol (Invitrogen) and immediately frozen at−80°C. This experiment was repeated twice independently.
Extraction of total RNA was performed according to manufacturer's instructions(Trizol, Invitrogen). Biotinylated cRNA targets were prepared from total RNA using adouble amplification protocol according to the GeneChip Expression Analysis TechnicalManual: two-Cycle Target Labeling Assay (P/N 701021 Rev.5, Affymetrix, Santa Clara,USA). Following fragmentation, cRNAs were hybridized for 16 hr at 45°C onGeneChip Chicken Genome arrays. Each microarray (one microarray per condition= two control microarrays and 2 Hoxa13 microarrays) was thenwashed and stained on a GeneChip fluidics station 450 and scanned with a GeneChipScanner 3000 7G. Finally, raw data (.CEL Intensity files) were extracted from thescanned images using the Affymetrix GeneChip Command Console (AGCC) version 3.1. CELfiles were further processed with MAS5 and RMA algorithms using the Bioconductorpackage (version 2.8) available through R (version 2.12.1). Probe sets were filteredbased on their expression intensity value (MAS5 value). Probe sets with an intensityvalue under 100 were discarded. Probe sets were ranked based on fold change betweenthe intensity value of the control condition and the Hoxa13over-expression condition. The microarrays raw data are available on the GEO website(http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE38107).
Q-RT PCR analysis of FACS-sorted cells over-expressingHoxa13
RNAs from the microarray experiments were used as templates for cDNA synthesis usingthe Qantitect kit (Qiagen). 3 μl of cDNA was mixed with 6 μl of2× Lightcycler 480 SYBR green I master (Roche) and 1 μM of primers(listed in Table 2) in a total volume of 12μl. The Q-PCR reactions were run on a Lightcycler 480 (Roche) with theLightcycler 480. Each sample was run in duplicate and gapdh was usedas a control gene. The CT values obtained for each gene were normalized against theCT value obtained for gapdh.
Table 2.
List of primers used for Q-RT PCR
DOI:http://dx.doi.org/10.7554/eLife.04379.028
Gene name | Gene reference | Primers sequence 5′→3′ | Size of the amplicon |
---|---|---|---|
Gapdh | NM_204305.1 | F: GCTGAGAACGGGAAACTTGTG | 62 bp |
R: GGGTCACGCTCCTGGAAGA | |||
T | NM_204940.1 | F: CGAGGAGATCACAGCTTTAAAAATT | 75 bp |
R: TCATTTCTTTCCTTTGCGTCAA | |||
Axin2 | NM_204491.1 | F: GCGCAAACGATAGTGAGATATCC | 76 bp |
R: CCATCTACACTGCTGTCTGTCATTG | |||
Sp8 | NM_001198666.1 | F: CATGGCGCACCCCTACGAGTC | 131 bp |
R: CGTTGGGGGCACGTCGATCCA | |||
Fzd2 | NM_204222.1 | F: CCCTGCCCGCTGCACTTCAC | 190 bp |
R: CCGCTCACACCGTGGTCTCG | |||
Cyp26a1 | NM_001001129.1ik | F: AGGAGCCCGAGGGTGGCTACA | 138 bp |
R: TGGCAGTGGTTTCATGACCTCCAA | |||
Fgf8 | NM_001012767.1 | F: CGCTCTTCAGCTACGTGTTCATGC | 108 bp |
R: TGGTAGGTGCGCACGAGCC | |||
Etv1 | NM_204917.1 | F: ATGGACCACAGATTTCGCCGCC | 145 bp |
R: TTGGACGTCCTTCCCTCGGCA | |||
Fgfr1 | NM_205510.1 | F: CACGCTGCCCGACCAAGCTC | 168 bp |
R: GTGATGCGCGTGCGGTTGTT | |||
Rasgrp3 | NM_001006401.1 | F: AACGGCATCTCCAAGTGGGTCCA | 111 bp |
R: GAGATGAAGGAGCTTCTGTGCAACA |
Q-RT PCR analysis of FACS-sorted cells over-expressing the dominant-negativeconstructs
PM precursors of the anterior primitive streak of Stage 8 HH embryos wereelectroporated as previously described either with a mix of control vectors codingfor Hoxd10mutH, Hoxc11mutH and Hoxa13mutH in pCAGGS-IRES2-Venus or amix of vector coding for Hoxd10dn, Hoxc11dn andHoxa13dn in pCAGGS-IRES2-Venus. Embryos were reincubated in ahumidified incubator at 38°C until they reached the 28-somite stage. Tail-budregions containing the PM progenitors were then dissected and pooled in a drop ofPBS/FCS1% (three embryos per condition) on ice. Dissected tissues were thentransferred into a drop of diluted trypsin and incubated at 38°C for 10 min toallow efficient dissociation of the cells. The dissociation of cells was completedmechanically using a glass micropipette by pipetting up and down. Cells were thentransferred into 500 μl of PBS/FCS1% on ice and sorted based on YFPfluorescence using a FACS DIVA (BD technologies). For each condition, one thousandYFP+ cells were collected directly in Trizol (Invitrogen) and immediatelyfrozen at −80°C. This experiment was repeated four timesindependently.
Extraction of total RNA was performed according to manufacturer's instructions(Trizol, Invitrogen). RNAs were used as templates for cDNA synthesis using theQuantiTect kit (Qiagen). 3 μl of cDNA was mixed with 6 μl of 2×Lightcycler 480 SYBR green I master (Roche) and 1 μM of primers (listed inTable 2) in a final volume of 12μl. The Q-RT PCR reactions were run on a Lightcycler 480 (Roche). Each samplewas run in duplicates, and gapdh was used as a control. The CTvalues obtained for each gene were normalized against the CT value obtained forgapdh.
Q-RT PCR analysis of microdissected tailbuds
Embryos were harvested in PBS at different stages (10, 15, 20, or 25-somite stages)and pined using 0.10 mm minutiens on a silicon-coated petri dish. The tailbud regionwas then microdissected using a sharpened tungsten needle and care was taken toremove the endoderm and the ectoderm. Each individual tailbud was immediatelytransferred in 500 μl of Trizol (Invitrogen) in a 1.5 ml RNAse free tube(Ambion) on ice until five individual tailbuds were collected per stage (resulting infive tubes per stage). Then the tubes were immediately frozen at −80°C.Extraction of total RNA was performed according to manufacturer's instructions(Trizol, Invitrogen). RNAs were used as templates for cDNA synthesis using theiScript reverse transcriptase Supermix (Biorad). 3 μl of cDNA was mixed with 5μl of 2× SSoAdvanced universal SYBR green supermix (Biorad) and 1μM of primers (listed in Table 2) ina final volume of 10 μl. The Q-RT PCR reactions were run on a CFX384 (Biorad).Each sample was run in triplicates, and gapdh was used as a control.The CT values obtained for each gene were normalized against the CT value obtainedfor gapdh.
Acknowledgements
The authors thank V Wilson, D Wellik, L Selleri, and members of the Pourquiélaboratory for comments. The authors acknowledge P François, P Moncuquet, C Ebel,and O Tassy for help. This research was supported by the Howard Hughes MedicalInstitute, the Stowers Institute for Medical Research, a NIH grant R02 HD043158 and aChaire d'excellence of the Agence Nationale pour la Recherche (ANR) as well asthe European Research Council.
Funding Statement
The funders had no role in study design, data collection andinterpretation, or the decision to submit the work forpublication.
Funding Information
This paper was supported by the following grants:
- Howard Hughes Medical Institute (HHMI) to Olivier Pourquié.
- Stowers Institute for Medical Research to Nicolas Denans, Tadahiro Iimura, Olivier Pourquié.
- National Institutes of Health (NIH)R02 HD043158 to Olivier Pourquié.
- Agence Nationale de la Recherche to Olivier Pourquié.
Additional information
Competing interests
The authors declare that no competing interests exist.
Additional files
Supplementary file 1.
List of genes regulated by Hoxa13 in our microarray screen.(A) List of genes upregulated in the PSM precursors after Hoxa13overexpression. (B) List of genes downregulated in the PSM precursorsafter Hoxa13 overexpression.
DOI:http://dx.doi.org/10.7554/eLife.04379.029
Major dataset
The following previously published dataset was used:
Denans N, MoncuquetP, 2012, HoxA13 gain-of-function inchicken embryo PSM progenitors, ; Publiclyavailable at NCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/).
References
- Abu-Abed S, Dollé P, Metzger D, Beckett B, Chambon P, Petkovich M. The retinoic acid-metabolizing enzyme, CYP26A1, is essential fornormal hindbrain patterning, vertebral identity, and development of posteriorstructures. Genes & Development. 2001;15:226–240. doi: 10.1101/gad.855001.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Aulehla A, Wehrle C, Brand-Saberi B, Kemler R, Gossler A, Kanzler B, Herrmann BG. Wnt3a plays a major role in the segmentation clock controllingsomitogenesis. Developmental Cell. 2003;4:395–406. doi: 10.1016/S1534-5807(03)00055-8. [PubMed] [CrossRef] [Google Scholar]
- Bellairs R. The tail bud and cessation of segmentation in the chickembryo. In: Bellairs E, Lash JW, editors. Somitogenesis in developing embryos. New York and London: Plenum Press; 1986. pp. 161–178. [Google Scholar]
- Bénazéraf B, Francois P, Baker RE, Denans N, Little CD, Pourquié O. A random cell motility gradient downstream of FGF controls elongationof an amniote embryo. Nature. 2010;466:248–252. doi: 10.1038/nature09151.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Bénazéraf B, Pourquié O. Formation and segmentation of the vertebrate body axis. Annual Review of Cell and Developmental Biology. 2013;29:1–26. doi: 10.1146/annurev-cellbio-101011-155703. [PubMed] [CrossRef] [Google Scholar]
- Cambray N, Wilson V. Axial progenitors with extensive potency are localised to the mousechordoneural hinge. Development. 2002;129:4855–4866. [PubMed] [Google Scholar]
- Cambray N, Wilson V. Two distinct sources for a population of maturing axialprogenitors. Development. 2007;134:2829–2840. doi: 10.1242/dev.02877. [PubMed] [CrossRef] [Google Scholar]
- Capellini TD, Zewdu R, Di Giacomo G, Asciutti S, Kugler JE, Di Gregorio A, Selleri L. Pbx1/Pbx2 govern axial skeletal development by controlling Polycomband Hox in mesoderm and Pax1/Pax9 in sclerotome. Developmental Biology. 2008;321:500–514. doi: 10.1016/j.ydbio.2008.04.005.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Catala M, Teillet MA, Le Douarin NM. Organization and development of the tail bud analyzed with the quail-chick chimaera system. Mechanisms of Development. 1995;51:51–65. doi: 10.1016/0925-4773(95)00350-A. [PubMed] [CrossRef] [Google Scholar]
- Chang CP, Shen WF, Rozenfeld S, Lawrence HJ, Largman C, Cleary ML. Pbx proteins display hexapeptide-dependent cooperative DNA bindingwith a subset of Hox proteins. Genes & Development. 1995;9:663–674. doi: 10.1101/gad.9.6.663. [PubMed] [CrossRef] [Google Scholar]
- Chapman SC, Collignon J, Schoenwolf GC, Lumsden A. Improved method for chick whole-embryo culture using a filter papercarrier. Developmental Dynamics. 2001;220:284–289. doi: 10.1002/1097-0177(20010301)220:33.0.CO;2-5. [PubMed] [CrossRef] [Google Scholar]
- Coy SE, Borycki AG. Expression analysis of TALE family transcription factors during aviandevelopment. Developmental Dynamics. 2010;239:1234–1245. doi: 10.1002/dvdy.22264. [PubMed] [CrossRef] [Google Scholar]
- Cunningham TJ, Zhao X, Duester G. Uncoupling of retinoic acid signaling from tailbud development beforetermination of body axis extension. Genesis. 2011;49:776–783. doi: 10.1002/dvg.20763.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Czirók A, Rupp PA, Rongish BJ, Little CD. Multi-field 3D scanning light microscopy of earlyembryogenesis. Journal of Microscopy. 2002;206:209–217. doi: 10.1046/j.1365-2818.2002.01032.x. [PubMed] [CrossRef] [Google Scholar]
- de Santa Barbara P, Roberts DJ. Tail gut endoderm and gut/genitourinary/tail development: a newtissue-specific role for Hoxa13. Development. 2002;129:551–561.[PMC free article] [PubMed] [Google Scholar]
- Duboule D. The rise and fall of Hox gene clusters. Development. 2007;134:2549–2560. doi: 10.1242/dev.001065. [PubMed] [CrossRef] [Google Scholar]
- Duboule D, Morata G. Colinearity and functional hierarchy among genes of the homeoticcomplexes. Trends in Genetics. 1994;10:358–364. doi: 10.1016/0168-9525(94)90132-5. [PubMed] [CrossRef] [Google Scholar]
- Dubrulle J, Pourquié O. fgf8 mRNA decay establishes a gradient that couples axial elongationto patterning in the vertebrate embryo. Nature. 2004;427:419–422. doi: 10.1038/nature02216. [PubMed] [CrossRef] [Google Scholar]
- Economides KD, Zeltser L, Capecchi MR. Hoxb13 mutations cause overgrowth of caudal spinal cord and tailvertebrae. Developmental Biology. 2003;256:317–330. doi: 10.1016/S0012-1606(02)00137-9. [PubMed] [CrossRef] [Google Scholar]
- Galceran J, Fariñas I, Depew MJ, Clevers H, Grosschedl R. Wnt3a-/–like phenotype and limb deficiency inLef1(-/-)Tcf1(-/-) mice. Genes & Development. 1999;13:709–717. doi: 10.1101/gad.13.6.709.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Galceran J, Hsu SC, Grosschedl R. Rescue of a Wnt mutation by an activated form of LEF-1: regulation ofmaintenance but not initiation of Brachyury expression. Proceedings of the National Academy of Sciences of USA. 2001;98:8668–8673. doi: 10.1073/pnas.151258098.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Gehring WJ, Qian YQ, Billeter M, Furukubo-Tokunaga K, Schier AF, Resendez-Perez D, Affolter M, Otting G, Wuthrich K. Homeodomain-DNA recognition. Cell. 1994;78:211–223. doi: 10.1016/0092-8674(94)90292-5. [PubMed] [CrossRef] [Google Scholar]
- Godwin AR, Capecchi MR. Hoxc13 mutant mice lack external hair. Genes & Development. 1998;12:11–20. doi: 10.1101/gad.12.1.11.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Goldman DC, Martin GR, Tam PP. Fate and function of the ventral ectodermal ridge during mouse taildevelopment. Development. 2000;127:2113–2123. [PubMed] [Google Scholar]
- Gomez C, Ozbudak EM, Wunderlich J, Baumann D, Lewis J, Pourquié O. Control of segment number in vertebrate embryos. Nature. 2008;454:335–339. doi: 10.1038/nature07020. [PubMed] [CrossRef] [Google Scholar]
- Gomez C, Pourquié O. Developmental control of segment numbers invertebrates. Journal of Experimental Zoology Part B, Molecular and DevelopmentalEvolution. 2009;312:533–544. doi: 10.1002/jez.b.21305.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Gouti M, Tsakiridis A, Wymeersch FJ, Huang Y, Kleinjung J, Wilson V, Briscoe J. In vitro generation of neuromesodermal progenitors reveals distinctroles for wnt signaling in the specification of spinal cord and paraxial mesodermidentity. PLOS Biology. 2014;12:e1001937. doi: 10.1371/journal.pbio.1001937.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Hamburger V. The stage series of the chick embryo. Developmental Dynamics. 1992;195:273–275. doi: 10.1002/aja.1001950405. [PubMed] [CrossRef] [Google Scholar]
- Hamburger V, Hamilton HL. A series of normal stages in the development of the chick embryo(1951) Developmental Dynamics. 1992;195:231–272. doi: 10.1002/aja.1001950404. [PubMed] [CrossRef] [Google Scholar]
- Harada N, Tamai Y, Ishikawa T, Sauer B, Takaku K, Oshima M, Taketo MM. Intestinal polyposis in mice with a dominant stable mutation of thebeta-catenin gene. The EMBO Journal. 1999;18:5931–5942. doi: 10.1093/emboj/18.21.5931.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Hatada Y, Stern CD. A fate map of the epiblast of the early chick embryo. Development. 1994;120:2879–2889. [PubMed] [Google Scholar]
- Henrique D, Adam J, Myat A, Chitnis A, Lewis J, Ish-Horowicz D. Expression of a Delta homologue in prospective neurons in thechick. Nature. 1995;375:787–790. doi: 10.1038/375787a0. [PubMed] [CrossRef] [Google Scholar]
- Iimura T, Pourquié O. Collinear activation of Hoxb genes during gastrulation is linked tomesoderm cell ingression. Nature. 2006;442:568–571. doi: 10.1038/nature04838. [PubMed] [CrossRef] [Google Scholar]
- Iimura T, Yang X, Weijer CJ, Pourquie O. Dual mode of paraxial mesoderm formation during chickgastrulation. Proceedings of the National Academy of Sciences of USA. 2007;104:2744–2749. doi: 10.1073/pnas.0610997104.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Iulianella A, Beckett B, Petkovich M, Lohnes D. A molecular basis for retinoic acid-induced axialtruncation. Developmental Biology. 1999;205:33–48. doi: 10.1006/dbio.1998.9110. [PubMed] [CrossRef] [Google Scholar]
- Kmita M, Fraudeau N, Hérault Y, Duboule D. Serial deletions and duplications suggest a mechanism for thecollinearity of Hoxd genes in limbs. Nature. 2002;420:145–150. doi: 10.1038/nature01189. [PubMed] [CrossRef] [Google Scholar]
- Knezevic V, De Santo R, Mackem S. Continuing organizer function during chick taildevelopment. Development. 1998;125:1791–1801. [PubMed] [Google Scholar]
- Kondoh H, Takemoto T. Axial stem cells deriving both posterior neural and mesodermal tissuesduring gastrulation. Current Opinion in Genetics & Development. 2012;22:374–380. doi: 10.1016/j.gde.2012.03.006. [PubMed] [CrossRef] [Google Scholar]
- Liu G, Bafico A, Harris VK, Aaronson SA. A novel mechanism for Wnt activation of canonical signaling throughthe LRP6 receptor. Molecular and Cellular Biology. 2003;23:5825–5835. doi: 10.1128/MCB.23.16.5825-5835.2003.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Mallo M, Wellik DM, Deschamps J. Hox genes and regional patterning of the vertebrate bodyplan. Developmental Biology. 2012;344:7–15. doi: 10.1016/j.ydbio.2010.04.024.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Maretto S, Cordenonsi M, Dupont S, Braghetta P, Broccoli V, Hassan AB, Volpin D, Bressan GM, Piccolo S. Mapping Wnt/beta-catenin signaling during mouse development and incolorectal tumors. Proceedings of the National Academy of Sciences of USA. 2003;100:3299–3304. doi: 10.1073/pnas.0434590100.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Martin BL, Kimelman D. Brachyury establishes the embryonic mesodermal progenitorniche. Genes & Development. 2010;24:2778–2783. doi: 10.1101/gad.1962910.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Martin BL, Kimelman D. Canonical Wnt signaling dynamically controls multiple stem cell fatedecisions during vertebrate body formation. Developmental Cell. 2012;22:223–232. doi: 10.1016/j.devcel.2011.11.001.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- McGrew MJ, Sherman A, Lillico SG, Ellard FM, Radcliffe PA, Gilhooley HJ, Mitrophanous KA, Cambray N, Wilson V, Sang H. Localised axial progenitor cell populations in the avian tail bud arenot committed to a posterior Hox identity. Development. 2008;135:2289–2299. doi: 10.1242/dev.022020. [PubMed] [CrossRef] [Google Scholar]
- McIntyre DC, Rakshit S, Yallowitz AR, Loken L, Jeannotte L, Capecchi MR, Wellik DM. Hox patterning of the vertebrate rib cage. Development. 2007;134:2981–2989. doi: 10.1242/dev.007567. [PubMed] [CrossRef] [Google Scholar]
- Moens CB, Selleri L. Hox cofactors in vertebrate development. Developmental Biology. 2006;291:193–206. doi: 10.1016/j.ydbio.2005.10.032. [PubMed] [CrossRef] [Google Scholar]
- Mortlock DP, Innis JW. Mutation of HOXA13 in hand-foot-genital syndrome. Nature Genetics. 1997;15:179–180. doi: 10.1038/ng0297-179. [PubMed] [CrossRef] [Google Scholar]
- Naiche LA, Holder N, Lewandoski M. FGF4 and FGF8 comprise the wavefront activity that controlssomitogenesis. Proceedings of the National Academy of Sciences of USA. 2011;108:4018–4023. doi: 10.1073/pnas.1007417108.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Nakaya Y, Sukowati EW, Wu Y, Sheng G. RhoA and microtubule dynamics control cell-basement membraneinteraction in EMT during gastrulation. Nature Cell Biology. 2008;10:765–775. doi: 10.1038/ncb1739. [PubMed] [CrossRef] [Google Scholar]
- Noordermeer D, Duboule D. Chromatin architectures and Hox gene collinearity. Current Topics in Developmental Biology. 2013;104:113–148. doi: 10.1016/B978-0-12-416027-9.00004-8. [PubMed] [CrossRef] [Google Scholar]
- Ohta S, Suzuki K, Tachibana K, Tanaka H, Yamada G. Cessation of gastrulation is mediated by suppression ofepithelial-mesenchymal transition at the ventral ectodermal ridge. Development. 2007;134:4315–4324. doi: 10.1242/dev.008151. [PubMed] [CrossRef] [Google Scholar]
- Olivera-Martinez I, Harada H, Halley PA, Storey KG. Loss of FGF-dependent mesoderm identity and rise of endogenousretinoid signaling determine cessation of body axis elongation. PLOS Biology. 2012;10:e1001415. doi: 10.1371/journal.pbio.1001415.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Ordahl CP. Myogenic lineages within the developing somite. In: Bernfield M, editor. Molecular basis of morphogenesis. New-York: John Wiley and Sons; 1993. pp. 165–176. [Google Scholar]
- Prince V, Lumsden A. Hoxa-2 expression in normal and transposed rhombomeres: independentregulation in the neural tube and neural crest. Development. 1994;120:911–923. [PubMed] [Google Scholar]
- Psychoyos D, Stern CD. Fates and migratory routes of primitive streak cells in the chickembryo. Development. 1996;122:1523–1534. [PubMed] [Google Scholar]
- Rhee JM, Pirity MK, Lackan CS, Long JZ, Kondoh G, Takeda J, Hadjantonakis AK. In vivo imaging and differential localization of lipid-modifiedGFP-variant fusions in embryonic stem cells and mice. Genesis. 2006;44:202–218. doi: 10.1002/dvg.20203.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Sakai Y, Meno C, Fujii H, Nishino J, Shiratori H, Saijoh Y, Rossant J, Hamada H. The retinoic acid-inactivating enzyme CYP26 is essential forestablishing an uneven distribution of retinoic acid along the anterio-posterioraxis within the mouse embryo. Genes & Development. 2001;15:213–225. doi: 10.1101/gad.851501.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Schoenwolf GC. Histological and ultrastructural observations of tail bud formation inthe chick embryo. The Anatomical Record. 1979;193:131–147. doi: 10.1002/ar.1091930108. [PubMed] [CrossRef] [Google Scholar]
- Selleck MA, Stern CD. Fate mapping and cell lineage analysis of Hensen's node in thechick embryo. Development. 1991;112:615–626. [PubMed] [Google Scholar]
- Selleri L, Depew MJ, Jacobs Y, Chanda SK, Tsang KY, Cheah KS, Rubenstein JL, O'Gorman S, Cleary ML. Requirement for Pbx1 in skeletal patterning and programmingchondrocyte proliferation and differentiation. Development. 2001;128:3543–3557. [PubMed] [Google Scholar]
- Slattery M, Riley T, Liu P, Abe N, Gomez-Alcala P, Dror I, Zhou T, Rohs R, Honig B, Bussemaker HJ, Mann RS. Cofactor binding evokes latent differences in DNA binding specificitybetween Hox proteins. Cell. 2011;147:1270–1282. doi: 10.1016/j.cell.2011.10.053.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Smith SM, Cartwright MM. Spatial visualization of apoptosis using a whole-mount in situ DNAend-labeling technique. Biotechniques. 1997;22:832–834. [PubMed] [Google Scholar]
- Spitz F, Gonzalez F, Peichel C, Vogt TF, Duboule D, Zákány J. Large scale transgenic and cluster deletion analysis of the HoxDcomplex separate an ancestral regulatory module from evolutionaryinnovations. Genes & Development. 2001;15:2209–2214. doi: 10.1101/gad.205701.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Stott D, Kispert A, Herrmann BG. Rescue of the tail defect of Brachyury mice. Genes & Development. 1993;7:197–203. doi: 10.1101/gad.7.2.197. [PubMed] [CrossRef] [Google Scholar]
- Tarchini B, Duboule D. Control of Hoxd genes' collinearity during early limbdevelopment. Developmental Cell. 2006;10:93–103. doi: 10.1016/j.devcel.2005.11.014. [PubMed] [CrossRef] [Google Scholar]
- Tenin G, Wright D, Ferjentsik Z, Bone R, McGrew MJ, Maroto M. The chick somitogenesis oscillator is arrested before all paraxialmesoderm is segmented into somites. BMC Developmental Biology. 2010;10:24. doi: 10.1186/1471-213X-10-24.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Tsakiridis A, Huang Y, Blin G, Skylaki S, Wymeersch F, Osorno R, Economou C, Karagianni E, Zhao S, Lowell S, Wilson V. Distinct Wnt-driven primitive streak-like populations reflect in vivolineage precursors. Development. 2014;141:1209–1221. doi: 10.1242/dev.101014.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Tschopp P, Tarchini B, Spitz F, Zakany J, Duboule D. Uncoupling time and space in the collinear regulation of Hoxgenes. PLOS Genetics. 2009;5:e1000398. doi: 10.1371/journal.pgen.1000398.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Tzouanacou E, Wegener A, Wymeersch FJ, Wilson V, Nicolas JF. Redefining the progression of lineage segregations during mammalianembryogenesis by clonal analysis. Developmental Cell. 2009;17:365–376. doi: 10.1016/j.devcel.2009.08.002. [PubMed] [CrossRef] [Google Scholar]
- Warren M, Puskarczyk K, Chapman SC. Chick embryo proliferation studies using EdU labeling. Developmental Dynamics. 2009;238:944–949. doi: 10.1002/dvdy.21895.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Wellik DM, Capecchi MR. Hox10 and Hox11 genes are required to globally pattern the mammalianskeleton. Science. 2003;301:363–367. doi: 10.1126/science.1085672. [PubMed] [CrossRef] [Google Scholar]
- Wilson V, Beddington R. Expression of T protein in the primitive streak is necessary andsufficient for posterior mesoderm movement and somitedifferentiation. Developmental Biology. 1997;192:45–58. doi: 10.1006/dbio.1997.8701. [PubMed] [CrossRef] [Google Scholar]
- Wilson V, Beddington RS. Cell fate and morphogenetic movement in the late mouse primitivestreak. Mechanisms of Development. 1996;55:79–89. doi: 10.1016/0925-4773(95)00493-9. [PubMed] [CrossRef] [Google Scholar]
- Wilson V, Manson L, Skarnes WC, Beddington RS. The T gene is necessary for normal mesodermal morphogenetic cellmovements during gastrulation. Development. 1995;121:877–886. [PubMed] [Google Scholar]
- Wilson V, Olivera-Martinez I, Storey KG. Stem cells, signals and vertebrate body axis extension. Development. 2009;136:1591–1604. doi: 10.1242/dev.021246. [PubMed] [CrossRef] [Google Scholar]
- Yamaguchi TP, Takada S, Yoshikawa Y, Wu N, McMahon AP. T (Brachyury) is a direct target of Wnt3a during paraxial mesodermspecification. Genes & Development. 1999;13:3185–3190. doi: 10.1101/gad.13.24.3185.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
- Young T, Rowland JE, van de Ven C, Bialecka M, Novoa A, Carapuco M, van Nes J, de Graaff W, Duluc I, Freund JN, Beck F, Mallo M, Deschamps J. Cdx and Hox genes differentially regulate posterior axial growth inmammalian embryos. Developmental Cell. 2009;17:516–526. doi: 10.1016/j.devcel.2009.08.010. [PubMed] [CrossRef] [Google Scholar]
Marianne E Bronner, California Institute of Technology, UnitedStates;