Development of Ranvier nodes progressed in a similar time course along NM axons after hearing onset.

(A) Axonal projection of NM neurons in the chicken brainstem. NM axon projects to both sides of NL and forms “delay line” at contralateral NL. VIIIth, auditory nerve.

(B) Regional bias in periodic spacing of Ranvier nodes along the main trunk of NM axon. Internodal length is long at midline tract region and becomes shorter at the NL region in the axon.

(C) Development of NM neurons. NM neurons make synaptic contact with NL neurons by E10 and receive synaptic input from the auditory nerve by E12.

(D) Immunostaining of AnkG between E12 and P3. Dashed line indicates midline.

(E–F) Double immunostaining of AnkG (red) and panNav (green) for (E), and AnkG (red) and Caspr (green) for (F) at tract (upper) and NL (lower) regions from the same slices.

(G–H) Ranvier nodes matured on a similar time course across the region. Three types of Ranvier nodes immunostained with AnkG (red), panNav (green), and Caspr (blue) antibodies. (G), and their proportions between E15 and P3 (H). These types were determined according to the length of AnkG signals; the signals longer than 3 µm were defined as heminode or immature node according to the number of Nav-negative paranodal domains, and those shorter than 3 µm were defined as mature node. Note that some immature nodes had long nodal domains that exceeded 5 μm. This may correspond to a gap between the two heminodes. E15: n=183 and 177 nodes for tract and NL, N=4 chicks; E18: n=344 and 267 nodes for tract and NL, N=4 chicks; E21: n=267 and 371 nodes for tract and NL, N=4 chicks; P3: n=226 and 412 nodes for tract and NL, N=3 chicks.

Statistical analysis: Chi-square test (H) was used to compare the proportions of node types between the regions at each developmental stage. **p < 0.01.

Regional difference in internodal length was evident during development.

(A–B) Immunostaining of MAG between E9 and E21 (A). Boxes with numbers at E15 are magnified (B).

(C) Double immunostaining of MAG (red) and Olig2 (green), a marker of oligodendrocyte lineage cells, at E15.

(D) Localization of MAG (green), AnkG (red), and Nav (blue) on heminode at E15.

(E) Ranvier node is formed by restricting a gap between adjacent two heminodes during development. Note that paranodal domains, flanking the Nav-positive nodal domain, accompany fibrous AnkG clustering in heminode but not in mature node.

(F–G) Regional difference in internodal length appeared during the period of node formation. The axons were labeled with rhodamine dextran (red) and stained with Caspr antibody (green) at E18 and P9 (F). Internodal length was measured as a distance between adjacent mature/immature nodes (G). Boxes with numbers correspond to high-magnification images of each node. E18: n=41 internodes, N=4 chicks for tract, n=38 internodes, N=11 chicks for NL; P9: n=43 internodes, N=7 chicks for tract, n=62 internodes, N=10 chicks for NL.

Statistical analysis: Kruskal-Wallis test and post hoc Steel-Dwass test (G) was used to compare different developmental stages in the same region and different regions at the same developmental stage: *p < 0.05, **p < 0.01.

3D morphometry of oligodendrocytes showed distinct differences between tract and NL regions.

(A) Timeline of experiments. In ovo electroporation and A3V transfection were performed at E2 (HH stage 11–12) and E2–3 (HH stage 15–16), respectively, and brainstem was observed at E21.

(B) Two types of plasmid vectors were introduced to visualize mature oligodendrocytes using in ovo electroporation. iOn-MBP∞paltdTomato expresses tdTomato with a palmitoylation signal (paltdTomato) under the control of the MBP promoter after the genomic integration. pCAG-hyPBase integrates the above plasmid into the genome by expressing hyperactive piggyBac transposase.

(C) A3V-GFP was injected into neural tube to visualize the axon of NM neurons.

(D) GFP and paltdTomato expressions in a 200-μm-thick slice (upper) and magnified images of single oligodendrocytes at tract and NL regions (lower). NM neurons and their axons were densely labeled with GFP (green), while mature oligodendrocytes were sparsely labeled with paltdTomato (magenta). (E–G) Myelin morphometry showed concordance with internodal length and their uncorrelation with myelin diameter. Length (E) and diameter (F) of myelin sheaths were compared between regions, and their relationship (G) was evaluated with Spearman’s rank correlation coefficient (rs). These parameters did not correlate with each other at both regions. Tract: n=47 myelins, N=3 chicks; NL: n=52 myelins, N=3 chicks.

(H–M) Oligodendrocyte morphometry highlighted their regional heterogeneity. Number (H), total length (I), average length of myelin sheaths (J) per oligodendrocyte, and cross-sectional area of cell body (K) were compared between regions. The relationship between total (L) or average (M) myelin lengths and the number of myelin sheaths was evaluated with Spearman’s rank correlation coefficient (rs). Total and average myelin lengths showed increasing and decreasing tendencies, respectively, with an increase of myelin processes. H, I, J, L, M: n=31 cells, N=4 chicks for tract, n=18 cells, N=3 chicks for NL. K: n=38 cells, N=3 chicks for tract, n=65 cells, N=3 chicks for NL.

Statistical analysis: Wilcoxon rank sum test (E, H, J, K), two-tailed Welch’s t-test (F), and two-tailed Student’s t-test (I): *p < 0.05, **p < 0.01, n.s., not significant.

Axonal structure was not a major determinant of regional difference in nodal spacing.

(A) Timeline of experiments. In ovo electroporation was performed at E2 (HH Stage 11–12), and brainstem was observed at E21.

(B) Two types of plasmid vectors were introduced to visualize the axon of NM neurons using in ovo electroporation. Atoh1-Flpo expresses Flpo under the control of Atoh1 promoter, which has selectivity for NM neurons. pCAFNF-palGFP-WPRE expresses GFP with a palmitoylation signal (palGFP) in a Flpo-dependent manner.

(C) Contralateral projections of NM neurons were labeled with palGFP (green), while visualizing paranodes with Caspr antibody (magenta) in a 200-μm-thick slice.

(D) Nodal spacing across tract and NL regions along a single NM axon. Arrowheads indicate branch points of collaterals. Each number corresponds to high-magnification images of node (lower panels). Each internode (between nodes) on the axon was labeled alternately with red and green lines, and their lengths were measured.

(E–G) Internodal length was clearly shorter than the branch point interval at the NL region. Branch point interval (E) at the NL region, internodal length (F) within those intervals, and their relationship (G). Note that the internodal length was mostly below 100 µm at the NL region even when the branch point interval was above 100 µm. E: n=113 intervals, F: n=178 internodes, G: n=188 internodes (including 0 μm), N=6 chicks.

(H–I) Branch point interval correlated with the number of internodes within the interval. Number of internodes between adjacent branch points (H) and branch point interval against the number of internodes (I). “0”: n=10 intervals; “1”: n=50 intervals; “2”: n=36 intervals; “3≦”: n=17 intervals.

(J) Internodal length was not necessarily specified by branch point interval. Comparison of internodal length for branch point intervals above 220 μm (top 10% of measured values), below 220 μm at the NL region, and at the tract region. <220: n=143 internodes, >220: n=35 internodes; Tract: n=132 internodes, N=5 chicks.

(K) Diameter of main trunk of the axon at tract and NL regions was not different. Tract: n=114 axons, N=3 chicks; NL: n=81 axons, N=3 chicks.

Statistical analysis: Kruskal-Wallis test and post hoc Steel-Dwass test (J) and Wilcoxon rank sum test (K): *p < 0.05, **p < 0.01, n.s., not significant.

Region-specific facilitation of oligodendrogenesis led to higher oligodendrocyte density at NL region.

(A) Immunostainings of Olig2, a marker for oligodendrocyte lineage cells, and Nkx2.2, a marker for OPCs. Yellow and red arrowheads indicate tract and NL regions, respectively.

(B–C) Density of developing oligodendrocytes increased especially at the NL region. Developmental changes in the density of Olig2-positive (B) and Nkx2.2-positive (C) cells at each region and their ratio between the regions. E9: N=3 chicks, E12: N=4 chicks, E15: N=4 chicks, E18: N=6 chicks, E21: N=3 chicks.

(D) Immunostaining of BrdU, a marker for proliferating cells. Yellow and red arrowheads indicate tract and NL regions, respectively.

(E) Cell proliferation was facilitated at the NL region. Developmental changes in the density of BrdU-positive cells at each region and their ratio between the regions. N=3 chicks for each stage.

(F) Colocalization of BrdU (green), Nkx2.2 (red), and Olig2 (blue) signals at E12.

(G) Proliferating cells were almost exclusively OPCs. Mutual percentages between BrdU-positive cells and Nkx2.2-or Olig2-positive cells at E12 and E14. N=4 chicks.

Statistical analysis: Two-tailed paired t-test (B, C, E), Wilcoxon rank sum test (G): *p < 0.05, **p < 0.01.

Inhibition of vesicular release caused unmyelinated segments via suppression of oligodendrogenesis at NL region.

(A) Timeline of experiments. In ovo electroporation and A3V transfection were performed at E2 (HH Stage 11–12) and at E2–3 (HH Stage 15–16), respectively, and brainstem was observed at E15 in (L– M) and at E21 in (C–K).

(B) A3V-eTeNT expressing GFP-tagged eTeNT was used to inhibit vesicular release from NM axons. A3V-GFP expressing only GFP was used as a control

(C–I) eTeNT caused unmyelinated segments at the NL region without affecting internodal length. Most of the axons were labeled with A3V-GFP (C) or A3V-eTeNT (D), while nodal spacing was analyzed along one of the axons. Arrowheads indicate branch points of collaterals. Each number corresponds to high-magnification images of the node. Each internode was labeled alternately with red and green lines, while unmyelinated segments were labeled with yellow broken lines for A3V-eTeNT. “Unmyelinated segment” was identified as a non-overlapping axonal segment formed by a pair of heminodes, while internodes were classified into “internode w/o heminode” and “internode w/ heminode” according to the types of nodes at the ends (E) (see Methods). Percentages of heminode and mature/immature node (F), and internodal length (H) at tract and NL regions. Percentage of unmyelinated segments in each axon traced over 200 μm (G), and length of each segment including internode w/o and w/ heminode, and unmyelinated segment for A3V-eTeNT (I) at the NL region. F: GFP (n=124 nodes) and eTeNT (n=160 nodes) at tract, GFP (n=186 nodes) and eTeNT (n=270 nodes) at NL. G: n=29 axons for GFP, n=38 axons for eTeNT. H: GFP (n=62 internodes) and eTeNT (n=80 internodes) at tract, GFP (n=178 internodes) and eTeNT (n=180 internodes) at NL. I: Internode w/o (n=115) and w/ (n=63) heminode and unmyelinated segment (n=63). Tract: N=3 chicks for GFP, N=4 chicks for eTeNT, NL: N=5 chicks for GFP, N=5 chicks for eTeNT.

(J–K) eTeNT did not affect oligodendrocyte morphology at the NL region. Magnified images of single oligodendrocytes after A3V-eTeNT transfection (J). Number, total length, average length of myelins per oligodendrocyte were compared between A3V-eTeNT and A3V-GFP (Figure 4 H–J) and shown as a ratio (K; n=31 cells, N=4 chicks).

(L–M) eTeNT suppressed oligodendrogenesis at the NL region. Immunostainings of Nkx2.2 (OPC marker) and BrdU (proliferating cell marker) at E15 for A3V-GFP (upper) and A3V-eTeNT (lower)

(L). Yellow and red arrowheads indicate tract and NL regions, respectively. Relative density of Nkx2.2- and BrdU-positive cells between the tract and NL regions (M). A3V-eTeNT reduced the density of these cells specifically at the NL region and abolished the difference between the regions. M: N=4 chicks for GFP, N=6 chicks for eTeNT.

Statistical analysis: Chi-square test (F), Wilcoxon rank sum test (G, H, K), Kruskal-Wallis test and post hoc Steel-Dwass test (I) and two-tailed student’s t-test (M): *p < 0.05, **p < 0.01, n.s., not significant.

Regional heterogeneity of oligodendrocytes and adaptive oligodendrogenesis underlie the biased nodal spacing pattern along NM axons.

(A) Morphology of oligodendrocytes, such as the number and length of myelins, is determined intrinsically at each region; those at the NL region have larger numbers of short myelins compared to those at the tract region. In addition, adaptive oligodendrogenesis increases the density of oligodendrocytes specifically at the NL region.

(B–C) Nodal spacing primarily reflects the length of myelins at each region (B). Inhibition of vesicular release from NM axons by eTeNT suppressed adaptive oligodendrogenesis and caused unmyelinated segments at the NL region without altering oligodendrocyte morphology and internodal length (C), suggesting the importance of adaptive oligodendrogenesis in myelinating the entire axons with the short myelins at the NL region. Thus, intrinsic and adaptive properties of oligodendrocytes play a pivotal role in shaping the region-specific nodal spacing along NM axons.