(A) NADPH-d histochemistry at 18 PCW revealed intensely labeled pyramidal neurons in ACC L3 and FOp L5, where stained neurons were arranged into alternating vertical columns (open arrow). Interneurons (arrowhead) and blood vessels (asterisks) were also labeled. (B) Schematic summary of the spatiotemporal dynamics of NOS1/NADPH-d staining in pyramidal neurons of L2/3 (green) and L5 (red) in the developing and adult human neocortex. (C) NOS1 immunohistochemisty of radial and tangential sections at 20 PCW. NOS1⁺ pyramidal neurons exhibited a clear columnar organization in FOp L5 but not ACC L3. (D and E) Analysis of cell density (D) and cluster spacing (E) revealed that NOS1⁺ neurons in FOp L5 were significantly distinct in cytoarchitectonic organization from those in ACC L3. *P < 0.05. Error bars represent the 5^th and 95^th percentiles of thirty measurements. (F) Serial section analysis of NADPH-d⁺ L5 columns in two brains at 18 and 20 PCW. NADPH-d⁺ columns were present bilaterally.

(A and B) Triple immunofluorescent staining for NOS1 (green), BCL11B (blue), a marker of L5 subcortical projection neurons, and SATB2 (red), a marker of upper layer corticocortical projection neurons, in 20 PCW FOp L5. NOS1⁺ and BCL11B⁺ pyramidal neurons formed alternating columns (outlined) separated by clusters of SATB2⁺ neurons. (C) Analysis of columnarity in FOp L5 neurons. NOS1⁺ neurons were significantly more columnar in organization compared to NOS1⁻ neurons. *P< 0.05. Error bars represent the 5^th and 95^th percentiles of thirty measurements. (D) NOS1 immunostaining (brown) and FEZF2 in situ hybridization (blue) of FOp L5 at 18 PCW. NOS1⁺ neurons co-expressed FEZF2 (open arrowheads). (E, F, and G) Immunofluorescent staining for NOS1 (green) and FOXP2, synaptophysin (SYP), or FOS (red in E, F, and G). NOS1⁺ neurons co-expressed FOXP2 and FOS (arrowheads in E and G), and were encircled by SYP punta (arrowheads in F). (H, I, and J) Retrograde axonal tracing at 20 PCW. Retrograde travel of Fast DiI inserted into dorsal internal capsule (red arrowhead) and Fast DiA inserted into the corpus callosum (green arrowhead) were examined after seven months in incubation. In the FOp L5 (I), DiI-labeled subcortical projection neurons formed columns similar those composed of NOS1⁺ neurons. In the ACC (J), DiA-labeled corticocortical projection neurons did not exhibit obvious columnar organization.

In the P4 mouse ACC and lateral frontal cortex (A), intense NADPH-d staining was restricted to interneurons, with neuropil staining in ACC L2/3. In the E73 macaque neocortex (B), intense NADPH-d activity was present in ACC L2/3 pyramidal neurons similar to those labeled in the human 18 PCW ACC (C). In the macaque FOp, NADPH-d⁺ L5 pyramidal neurons were arranged into vertical columns similar in organization to the human FOp columns (C). Strong interneuronal and weak neuropil NADPH-d staining was present in all cortical areas in both macaque and human neocortex.

(A) Nissl staining, FEZF2 and NOS1 in situ hybridization, NOS1 immunohistochemistry and NADPH-d histochemistry in adjacent sections from 18 PCW FOp, ACC and dorsal lateral prefrontal cortex (PFC). NOS1 mRNA was abundantly and ubiquitously present in all cortical areas and layers examined. Intense NOS1 and NADPH-d labeling in L5 pyramidal columns (asterisks) were present exclusively in the FOp. In other cortical regions, NOS1 and NADPH-d were restricted to interneurons (open arrowheads) and neuropil. (B) Nos1 in situ hybridization, NOS1 and NADPH-d staining in adjacent sections from P3 mouse frontal neocortex. Nos1 mRNA was ubiquitously present, intense NOS1 and NADPH-d staining was exclusively present in interneurons (open arrowheads). Neuropil was weakly stained. (C) Pyramidal neurons of the Emx1 lineage were isolated from the P3 mouse neocortex by fluorescent cell sorting (FACS) and analyzed by quantitative (q) RT-PCR. Nos1 mRNA was abundantly present in pyramidal neurons. Error bars represent the 5^th and 95^th percentiles of four measurements.

(A) Proteins eluted from an mRNA pull-down assay using lysates of a 21 PCW human neocortex were analyzed by silver staining and immunoblotting. NOS1, but not control (GAPDH, GFP, and NeoR), mRNA specifically associated with a ~75kDa protein that was immunopositive for FMRP. Asterisk indicates an artifact of gel transfer. (B) NOS1 (green), FMRP (red), and DAPI (blue) staining of 20 PCW FOp. FMRP was co-expressed by L5 columnar NOS1⁺ pyramidal neurons (solid arrowheads) but not most interneurons (arrow). (C and D) FMRP immunoprecipitated mRNAs from a 21 PCW human and P0 mouse CP were analyzed by quantitative RT-PCR. Relatively to control GAPDH and MAP1B mRNAs, FMRP strongly associated with NOS1 mRNA (green bar) in human but not mouse. Error bars represent the 5^th and 95^th percentiles of four measurements.

(A) Prediction of putative FMRP-binding GQ and U-rich (UR) motifs in the human NOS1 mRNA sequence and alignment of GQ1 and GQ2. GQ1 and GQ2 were highly conserved in primates but not rodents. (B) FMRP association with each putative binding motif was analyzed by an mRNA pull-down assay using 21 PCW human neocortex lysates. FMRP selectively associated with GQ1 and GQ2. (C) EMSA of GQ1 and GQ2. RNA containing both GQs exhibited a significant shift in mobility in the presence of FMRP. This shift was abolished by addition of excess unbiotinylated (“cold”) RNA or a neutralizing anti-FMRP antibody. (D) Reverse transcription termination assay. Reverse transcription paused at the expected GQ sites for both GQ1 and GQ2 in the presence of K⁺, which facilitates GQ formation, but not Li⁺, which disrupts it. (E) Colorimetric NOS assays in Neuro2a cells co-transfected with CAG-hFMR1 or CAG-hFMR1(I304N) and one of CAG-hNOS1, CAG-mNos1, or CAG-murinized-hNOS1. NOS activity from hNOS1, but not mNos1 or murinized hNOS1, increased dose-dependently with increasing wildtype FMRP. The I304N mutation in FMRP abolished its activation of hNOS1 translation. *P < 0.05. Error bars represent the 5^th and 95^th percentiles of four measurements. (F) Luciferase assays in Neuro2a cells transfected with an empty reporter construct (GL3-empty), or constructs containing the human NOS1 GQs (GL3-hNOS1(GQs)) or the orthologous sequence in mouse Nos1 (GL3-mNOS1(GQortholog)). Luciferase activity in cells transfected with GL3-hNOS1(GQs) increased dose-dependently with co-transfection of human CAG-hFMR1 (solid blue line) or mouse CAG-mFmr1 (broken blue line). Lucifase activity from the GL3-mNos1(GQortholog) decreased with increasing amounts of CAG-hFMR1 (solid red line). Error bars represent the 5^th and 95^th percentiles of six measurements.

(A, B, and C) Neocortex of wildtype or Fmr1 KO mouse electroporated in utero at E13.5 and immunostained for NOS1 (red) at P0. In wildtype neocortex, the majority of pyramidal neurons transfected with hNOS1 expressed high levels of NOS1 properly localized to the soma and apical dendrite. NOS1 protein expression from mNos1 or murinized-hNOS1 in wildtype and from hNOS1 in Fmr1 KO neocortex was dramatically reduced in comparison. The fluorescent intensity of NOS1 staining normalized to GFP (B) and the proportion of GFP⁺ cells expressing NOS1 (C) were quantified. *P < 0.05. Error bars represent the 5^th and 95^th percentiles of at least four animals. (D) Immunoblots of human fetal FXS and P0 mouse Fmr1 KO neocortex. Normalized to GAPDH levels, the neocortical expression of NOS1 protein was severely reduced in both human fetal FXS cases. In neonatal mouse neocortex, loss of Fmr1 did not alter neocortical NOS1 expression.