A B Late gastrula Early neurula Late neurula Figure S1 A. The neural crest gene network in vertebrates. Black arrows indicate empirically verified regulatory interactions. In vertebrates, signals from ventral ectoderm and underlying mesendoderm pattern the dorsal ectoderm, inducing expression of neural border specifiers and neural plate markers. These inductive signals then work with neural border specifiers to activate expression of neural crest specifiers. The neural crest specifiers cross-regulate and activate various effector genes, each of which mediates a different aspect of the neural crest phenotype. These include cassettes controlling neural crest delamination and migration as well as differentiation programs for neural crest derivatives like cartilage, pigment cells and peripheral neurons. B. Cross sections showing neurulation process in amphioxus embryos. In late gastrula, the embryo is ovoid with presumptive neural plate (green) becoming flattened. The boundaries of presumptive neural plate and epidermis (blue) at this stage are indicated by arrows, and the mesendoderm is colored in yellow. In early neurula, the epidermis (blue) at the edges of the neural plate detaches from the neural plate and begins to overgrow the neural plate, and the somites (pink) and notochord (red) start to form. Eventually the overgrowing dorsal epidermis fuse together at the dorsal midline, and the underneath neural plate curls up to form a dorsal, hollow neural tube in late neurula stage. Now, neurulation is complete, and the rudiment of notochord, anterior somites, as well as the endodermal tube rudiment that will form the pharyngeal tube and gut (yellow) become distinguishable.
Figure S2 Dose-dependent effect of zbmp4 treatment on amphioxus epidermal and neural gene expression. Whole-mount in situ hybridization for Dlx and SoxB1-a at late gastrula stage. All embryos are oriented in dorsal view with anterior at the top. Scale bar, 50 µm. Expression of epidermal marker Dlx was expanded by zbmp4 treatment in a dose-dependant manner; conversely, expression of neural marker SoxB1-a was reduced by low concentration of zbmp4 treatment and completely disappeared by high concentration of zbmp4.
Figure S3 Expression of Mitf, Tyrosinase, and two Tyrosinase-related proteins in amphioxus embryos. All embryos are oriented in dorsal view with anterior at the top. Scale bar, 50 µm. All these four genes are co-expressed throughout the epidermal ectoderm in gastrula stage and early neurula stage. During neurula stage, a patch of cells inside the developing neural tube (indicating by arrows) start to expressed these four genes, and the onset of Mitf transcription factor gene expression is earlier than the other melanogenic genes.
FGF8/17/18 subfamily Figure S4 Phylogenetic tree of FGF proteins constructed by MEGA program using the neighbor-joining method with default settings. Amphioxus protein is shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.2 amino acid substitutions per position. FGF8 has been implicated in playing roles in neural induction and neural crest development. Our survey identified an amphioxus candidate belonging to the FGF8/17/18 subfamily with high bootstrap value of %.
83 71 89 Q9NR61 DLL4 Hs O00548 DLL1 Hs Q9NYJ7 DLL3 Hs Delta estext fgenesh2 pg.c 1200038 Delta Bf P41 delta Dm P18168 Serrate Dm 75 e gw.243.6.1 Jagged Bf P78504 JAG1 Hs Q9Y219 JAG2 Hs Jagged/Serrate 0.1 Figure S5 Phylogenetic tree of Delta/Jagged/Serrate family proteins generated by MEGA program using the neighbor-joining method with default settings. Amphioxus proteins are shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. The unrooted tree is shown as a rooted tree for simplicity. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.1 amino acid substitutions per position. Delta and Jagged/Serrate proteins are ligands for Notch receptors. Our survey in amphioxus genome identified one candidate gene for Delta and one for Jagged/Serrate. The robust grouping of the tree supports their orthology assignment.
estext GenewiseH 1.C 500067 SoxB2 Bf AAY51532 SoxB2.3 Dm 79 80 Q24533 SoxB2.1 Dm AAL13542 SoxB2.2 Dm 76 61 Q9Y651 SOX21 Hs O95416 SOX14 Hs e gw.50.122.1 SoxB1-c Bf Sox B group 61 63 NP 524735 SoxNeuro Dm P41225 SOX3 Hs O00570 SOX1 Hs 56 89 P48431 SOX2 Hs estext fgenesh2 kg.c 6920001 SoxB1-a Bf 55 99 estext fgenesh2 pg.c 6920002 SoxB1-b Bf O60248 SOX15 Hs Sox G group Q05066 SRY Hs Sox A group 57 Q06945 SOX4 Hs O15370 SOX12 Hs Sox C group P35716 SOX11 Hs Q9UN79 SOX13 Hs 66 P35711 SOX5 Hs P35712 SOX6 Hs Sox D group Q9H6I2 SOX17 Hs Q9BT81 SOX7 Hs P35713 SOX18 Hs Sox F group 52 CAB63903 SoxE Dm 73 99 P57073 SOX8 Hs P56693 SOX10 Hs P48436 SOX9 Hs Sox E group e gw.5.674.1 SoxE Bf O94993 SOX30 Hs Sox H group 0.1 Figure S6 Phylogenetic tree of Sox proteins constructed by the neighbor-joining method with default settings. Amphioxus protein are shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.1 amino acid substitutions per position. Our survey of amphioxus Sox family genes was focused on the SoxB and SoxE groups. We identified an amphioxus SoxE group candidate which groups with vertebrate SoxE proteins with high bootstrap value(99%). We identified four candidates of amphioxus SoxB members. Although the bootstrap value of SoxB grouping was not very high, we designated the protein names based on the best hit analysis by blast search against a human proteome.
65 83 99 estext fgenesh2 pg.c 8060004 Olig-c Bf estext fgenesh2 pg.c 8060003 Olig-b Bf estext fgenesh2 pg.c 8060002 Olig-a Bf Olig 94 Q7RTU3 OLIG3 Hs Q13516 OLIG2 Hs fgenesh2 pg.scaffold 122000082 Beta3 Bf 98 51 71 98 Q92886 Neurogenin1 Hs Q8NFJ8 BHLH5 Hs Q9Y4Z2 Neurogenin3 Hs Q9H2A3 Neurogenin2 Hs Q8C6A8 BETA3 Mm Neurogenin Beta3 estext fgenesh2 pg.c 90049 Neurogenin Bf 66 gw.499.29.1 NeuroD Bf 99 Q9HD90 NDF4 Hs Q13562 NDF1 Hs NeuroD 97 Q15784 NDF2 Hs Q9JJR7 ASCL3 Mm 90 87 54 53 P50553 ASCL1 Hs Q99929 ASCL2 Hs P09775 Asense Dm e gw.302.11.1 Ash-a-like1 Bf e gw.302.22.1 Ash-a-like2 Bf Achaete-Scute (Ash) 78 82 83 P83 Achaete Dm P09774 Lsc Dm P84 Scute Dm 0.1 Figure S7 Phylogenetic tree of Proneural bhlh factors generated by the neighbor-joining method with default settings. Amphioxus proteins are shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudo-replications. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.1 amino acid substitutions per position. We identified three possible Olig family orthologues in the amphioxus genome and one candidate for the closely related Beta3 family. The three amphioxus Olig genes are tandemly aligned in the genome within a 62-kb region, suggesting that these genes occurred by amphioxus-specific gene duplications. We also identified amphioxus orthologues of Neurogenin and NeuroD family with strong bootstrap values. There are two possible Ash family members in amphioxus genome, however, the phylogenetic resolution was low within the family. Therefore, we designated these two factors as Ash-a-like1 and Ash-a-like2 based on the best hit analysis.
82 P19484 TFEB Hs 66 O14948 TFECL Hs 62 64 P19532 TFE3 Hs O75030 MITF Hs e gw.492.5.1 Mitf Bf Mitf AAQ01726 Mitf Dm P22415 USF1 Hs fgenesh2 pm.scaffold 448000004 USF Bf USF 65 Q15853 USF2 Hs 0.2 Figure S8 Phylogenetic tree of Mitf family proteins generated by MEGA program using the neighbor-joining method with default settings. Amphioxus proteins are shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. The unrooted tree is shown as a rooted tree for simplicity. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.2 amino acid substitutions per position. Our survey in amphioxus genome identified one orthologue of Mitf, and it was included with other Mitf family factors as a clade with high bootstrap support (%). We also identified a candidate for the USF family member, which is a closely related family to Mitf.
82 Q9W4S7 MYC Dm P01106 MYC Hs 98 65 fgenesh2 pm.scaffold 200003 Myc Bf P04198 MYCN Hs P12524 MYCL1 Hs Myc P12525 MYCL2 Hs 87 65 fgenesh2 pg.scaffold 132000064 Max-a Bf gw.35.272.1 Max-b Bf P61244 MAX Hs Max AAF49179 Max Dm 99 fgenesh2 pg.scaffold 290000014 Mnt Bf Q99583 MNT Hs Mnt 69 96 88 AAH00745 MAD3 Hs fgenesh2 pg.scaffold 205000025 MAD Bf Q05195 MAD1 Hs Q14582 MAD4 Hs MAD 55 P50539 MXI1 Hs 0.2 Figure S9 Phylogenetic tree of Myc, Max, Mnt, and MAD family proteins generated by the neighbor-joining method with default settings. Amphioxus proteins are shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. The unrooted tree is shown as a rooted tree for simplicity. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.2 amino acid substitutions per position. Our survey in amphioxus genome identified one candidate orthologue of Myc family factor with good bootstrap support (%). We also identified orthologues of the closely related Max, Mnt, and MAD family members from amphioxus genome. There are two amphioxus Max members and the phylogenetic analysis suggests they occurred by amphioxus-specific duplication.
P40126 TYRP2 Hs 53 P29812 TYRP2 Mm 99 91 P17643 TYRP1 Hs P07147 TYRP1 Mm BAC76423 HrTYR Hr Tyrp estext fgenesh2 pg.c 2120049 Tyrp-b Bf fgenesh2 pg.scaffold 1412000002 Tyrp-a Bf BAC76424 HrTRP Hr 56 estext gwp.c 3220050 Tyrosinase Bf P14679 TYRO Hs P11344 TYRO Mm Tyrosinase 0.1 Figure S10 Phylogenetic tree of Tyrosinase and Tyrosinase-related proteins (Tyrp) generated by the neighbor-joining method with default settings. Amphioxus proteins are shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. The unrooted tree is shown as a rooted tree for simplicity. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.1 amino acid substitutions per position. Tyrosinase and Tyrosinase-related proteins are separated clearly as two groups in our phylogenetic tree. Our phylogenetic analysis suggests that the tow Tyrosinase-related proteins we identified are amphioxus specific duplications, therefore we designated them as Tyrp-a and Tyrp-b.
83 P61586 RHOA Hs 69 P08134 RHOC Hs 98 99 81 P48148 RHO1 Dm P62745 RHOB Hs fgenesh2 pg.scaffold 92000066 RhoA/B/C-b RhoA/B/C estext fgenesh2 pm.c 1920005 RhoA/B/C-a P84095 RHOG Hs 54 66 fgenesh2 pm.scaffold 283000008 RhoD/F Bf O00212 RHOD Hs Q9HBH0 RHOF Hs RhoD/F Q92730 RRHO6 Hs 87 P61587 RHOE Hs P52198 RHON Hs RhoE Q15669 RHOH Hs 0.1 Figure S11 Phylogenetic tree of Rho small G-proteins generated by MEGA program using the neighbor-joining method with default settings. Amphioxus proteins are shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. The unrooted tree is shown as a rooted tree for simplicity. Protein names are as explained in the Methods. The scale bar indicates an evolutionary distance of 0.1 amino acid substitutions per position. We identified two candidate gene models encoding Rho-like G-proteins belonging to the RhoA/B/C group (bootstrap value of 98%). There is also a possible RhoD/F orthologue in amphioxus, however, the bootstrap supporting value is not very high.
74 73 63 P21802 FGFR2 Hs P11362 FGFR1 Hs AAB33460 c-ret Gg AAG00544 c-ret Xl NP 858048 c-ret Dr P07949 RET Hs SAL fgenesh2 pg.scaffold 92000070 Ret Bf AAF53977 Ret Dm XP 396123 LOC412667 Am outgroup 0.1 Figure S12 Phylogenetic tree of Ret receptor tyrosine kinase proteins generated by the neighbor-joining method based on the alignment of a tyrosine kinase domain. Amphioxus protein is shown by the large black dot. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. Protein names are as explained in the Methods. Human FGF receptors were used as outgroup proteins. The scale bar indicates an evolutionary distance of 0.1 amino acid substitutions per position. Our survey in amphioxus genome identified one candidate belonging to Ret family.
63 62 99 P21860 ERBB3 Hs Q15303 ERBB4 Hs P00533 EGFR Hs P04626 ERBB2 Hs EGF receptor estext fgenesh2 pg.c 120089 Erbb Bf 90 P04412 EGFR Dm P08069 IGF1R Hs P09208 INSR Dm outgroup 0.2 Figure S13 Phylogenetic tree of EGF receptors generated by the neighbor-joining method based on the alignment of a tyrosine kinase domain. Amphioxus protein is shown by large black dots. The number matching each branch indicates the percentage of times that a node was supported in 0 bootstrap pseudoreplications. Protein names are as explained in the Methods. Human and fly Insulin receptors were used as outgroup proteins. The scale bar indicates an evolutionary distance of 0.2 amino acid substitutions per position. In vertebrates, there are four members of epidermal growth factor (EGF) receptor family of receptor tyrosine kinase including ERBB1 (EGFR), ERBB2, ERBB3 and ERBB4. We identified only one amphioxus Erbb candidate in the genome, and the orthology among this amphioxus protein and EGFRs from other organisms is supported by high bootstrap value of %.
Table S1. Homologues of vertebrate neural crest network genes in the amphioxus Brachiosoma floridae genome Gene name The best hit model in the version The best hit protein in the human proteome cdna clone ID 1.0 assembly Neural plate border induction signals BMP2/4 estext_fgenesh2_pg.c_3470056 P12643 BMP-2 bfga034p10 Wnt1 estext_fgenesh2_pm.c_120023 P04628 WNT-1 bfne123k05 Wnt3 estext_fgenesh2_pg.c_120044 P56704 WNT-3A bflv058m15 Wnt6 fgenesh2_pm.scaffold_12000022 Q9Y6F9 WNT-6 bflv009a17 Wnt7 fgenesh2_kg.scaffold_00002 P56706 WNT-7B bfne126m18 Wnt8 fgenesh2_kg.scaffold_413000001 Q93098 WNT-8B bfne137b06 FGF8/17/18 estext_fgenesh2_pg.c_2320058 O60258 Fibroblast growth factor 17 bfne018g21 Notch estext_gwp.c_110607 Q5SXM3 NOTCH1 bfne093p03 Delta estext_fgenesh2_pg.c_1200038 O00548 Delta-like 1 bfga003d05 Neural plate border specifiers Dlx estext_genewiseh_1.c_530082 P56177 DLX-1 bflv011k11 Msx estext_gwp.c_540071 P28360 MSX-1 bflv048l06 Zic (AmphiZic) estext_gwp.c_400285 Q2M3N1 ZIC1 bfne006l12 Pax3/7 fgenesh2_kg.scaffold_210002 Q2PJS5 PAX7 bfne062d03 Neural patterning/differentiation genes SoxB1-a * estext_fgenesh2_kg.c_6920001 P48431 SOX-2 bfne037g13 SoxB1-b estext_fgenesh2_pg.c_6920002 P48431 SOX-2 bfne072e07 SoxB1-c e_gw.50.122.1 P48431 SOX-2 bfne126o15 SoxB2 estext_genewiseh_1.c_500067 O95416 SOX-14 bfad048c06 NeuroD gw.499.29.1 Q13562 Neurogenic differentiation factor 1 N/A Neurogenin estext_fgenesh2_pg.c_90049 Q92886 Neurogenin 1 N/A Ash-a-like1 e_gw.302.11.1 P50553 Achaete-scute homolog 1 N/A Ash-a-like2 e_gw.302.22.1 P50553 Achaete-scute homolog 1 N/A Beta3 fgenesh2_pg.scaffold_122000082 Q8NFJ8 Basic helix-loop-helix protein 5 N/A Olig-a estext_fgenesh2_pg.c_8060002 Q7RTU3 Oligodendrocyte transcription factor 3 bfne042a07 Olig-b estext_fgenesh2_pg.c_8060003 Q7RTU3 Oligodendrocyte transcription factor 3 bfne044b12 Olig-c estext_fgenesh2_pg.c_8060004 Q7RTU3 Oligodendrocyte transcription factor 3 bfne004n21 Islet estext_fgenesh2_kg.c_1460001 P61371 Islet-1 bfne013h18 Hu/elav fgenesh2_pg.scaffold_219000066 Q12926 ELAV-like protein 2 bflv027g23
Neural crest specifiers Snail estext_fgenesh2_kg.c_190001 O43623 Zinc finger protein SLUG bfne095n03 FoxD estext_fgenesh2_kg.c_2440002 Q9UJU5 FOXD3 bfne002k18 Twist fgenesh2_pm.scaffold_217000008 Q8WVJ9 Twist-related protein 2 bfne115j15 Myc fgenesh2_pm.scaffold_200003 P01106 Myc proto-oncogene protein bfne011k03 Id fgenesh2_pg.scaffold_166000081 Q02363 ID-2 bfne018j05 AP-2 e_gw.327.57.1 P05549 AP-2 alpha bfne111f20 SoxE e_gw.5.674.1 P48436 SOX-9 bfne111n21 Neural crest effector genes Mitf e_gw.492.5.1 O75030 MITF bflv016p23 Tyrosinase estext_gwp.c_3220050 P14679 Tyrosinase bfga039h23 Tyrp1 fgenesh2_pg.scaffold_1412000002 P17643 TRP-1 bfga030e07 Tyrp2 estext_fgenesh2_pg.c_2120049 P40126 L-dopachrome tautomerase precursor bfga006j06 RhoA/B/C-a estext_fgenesh2_pm.c_1920005 P61586 RHOA bfga002n08 RhoA/B/C-b fgenesh2_pg.scaffold_92000066 P61586 RHOA bfne019m11 Ret SAL_fgenesh2_pg.scaffold_92000070 P07949 RET N/A Erbb estext_fgenesh2_pg.c_120089 P21860 ErbB-3 N/A Col2a fgenesh2_pg.scaffold_53000 P02458 alpha 1 type II collagen bflv014e16 ckit Myelin protein P0 Not found Not found N/A: not available * This gene was named AmphiSox1/2/3 in previous literatures This gene model was annotated manually by combining two predicted gene models fgenesh2_pg.scaffold_92000069 and fgenesh2_pg.scaffold_92000070.
Table S2. Expression of EGFP in chick embryos electroporated at stage 3+ to 4. Stage examined Somites/paraxial mesoderm Notochord Neural tube (hindbrain) Head mesenchyme Stage 8 (n=4) 4/4 4/4 0/4 4/4 Stage 10 (n=5) 5/5 3/5* 5/5 4/5** Stage 11 (n=4) 4/4 4/4 4/4 4/4 * The two embryos with no EGFP in the notochord did not show any expression of RFP in the notochord. ** The embryo with no EGFP expression in the head mesenchyme did not show any expression of RFP in the head mesenchyme.