On the other hand, when the NAC was cocultured side-by-side with normal animal cap, the EB3 in the NAC cells moved toward the boundary between the two explants with strikingly biased movement, reminiscent of the boundary-contacting cells in the Keller explants. The biased EB3 movement was absent when NAC was cultured with NAC,GDC-0449or when animal cap was cultured in conjunction with animal cap. We also found that the disruption of the biased EB3 movement in NAC correlated well with aberrant cell alignment in the tissues. When NAC was conjugated with animal cap, the polarized cells were aligned perpendicular to the boundary, as described above. We also tested the conjugation of NAC and animal cap cells injected with increasing concentrations of nodal mRNA. The lowest concentrations induced neither somite nor notochord markers, and modest concentrations only induced somite markers. Higher concentrations of nodal mRNA induced both somite and notochord markers. When NAC was conjugated with S-AC, we unexpectedly found that the biased EB3 movement GDC-0879was less evident, and the ratio of the perpendicularly aligned cells representing notochord cells in vivo rather decreased. Very interestingly, however, when nodal concentration in NAC was increased even higher to ensure the notochord differentiation, cells aligned perpendicularly to the boundary were restored. These results strongly suggest that distinct nodal signaling levels, which in turn establish different populations of cells and hence inter-tissue boundaries, provide a cue for the cell polarity and cell alignment. In addition, cells in NAC combined with SN-AC tended to elongate in parallel to the tissue boundary, which is consistent with the previous finding that cells are polarized mediolaterally by sensing a shallow activin gradient generated along A-P axis. These findings further suggest that mediolateral polarity is established in notochord cells by the cells’ detection of a difference in nodal signaling, and that this polarity is essential for the notochord cells to be anchored to the boundary with nonnotochord cells and aligned vertical to the AP axis. We next confirmed that the effect of nodal is mediated by Smad2. The conjugation of Smad2-expressing animal cap and normal animal cap indicated that the influence from the tissue boundary was observed only in the chordamesoderm cells expressing Smad2. The results showing that the EB3 in animal cap did not respond in this co-culture system again suggest that chordamesoderm cells but not animal cap cells are capable to respond to the predicted polarity cue. By tracking EB3-GFP as an indicator of functional polarity, we have been able to visualize one of the intracellular events in mesoderm cells during CE.