Emx1-Cre deletes Bcl11b from cortex, and Gsx2-iCre deletes Bcl11b from striatum.

(A) Immunocytochemistry on P0 coronal forebrain section detects Bcl11b expression by cortical projection neurons in the cortical plate and by MSN in the striatum of wild-type (Bcl11bflox/flox) mice. (B) Bcl11b expression by cortical projection neurons is deleted by Emx1-Cre. (C) Bcl11b expression by MSN is deleted by Gsx2-iCre. (D) Double conditional deletion of Bcl11b by Emx1-Cre and Gsx2-iCre deletes Bcl11b from both cortical projection neurons and MSN. Scale bar, 500 µm.

Bcl11b expression by MSN in striatum, but not by projection neurons in cortex, is required for CFuPN axon fasciculation within the internal capsule.

(A–A’’’, F) CFuPN axons exit cortical plate, enter striatum, and fasciculate in the internal capsule in wild-type (Bcl11bflox/flox) mice. (B, B’, G) In Bcl11bflox/flox;Emx1IRES-Cre/+ brains, CFuPN axons lacking Bcl11b form normal fascicles through striatum, but have an abnormal border between subplate and cortical white matter (B’’’), and an ectopic external capsule (B’’). (C, C’, H) In Bcl11bflox/flox;Gsx2-iCre brains, CFuPN axon fasciculation is disturbed in striatum, while the border between subplate and cortical white matter remains normal (C’’’), and CFuPN axons enter striatum normally (C’’). (D, E, I, J) Brains lacking Bcl11b in both cortex and striatum (D–D’’’, E–E’’’) have an abnormal border between subplate and cortical white matter (D’’’, E’’’), an ectopic external capsule (D’’, E’’), and abnormally fasciculated CFuPN axons (D’, E’, I, J). A–E, Nissl staining. Scale bar, 500 µm. F–J, GAP43 immunocytochemistry. Scale bar, 100 µm. Arrowheads indicate ectopic external capsule at the border between cortex and striatum. Arrows indicate location of the border between subplate and cortical white matter.

Bcl11b deletion from SCPN disrupts outgrowth of SCPN axons, and Bcl11b deletion from MSN causes SCPN axon mistargeting into STN and dorsocaudal midbrain.

(A–M) Cortical efferent axons are visualized anterogradely by Emx1IRES-Flpo;Rosa26frt-LacZ-frt-EGFP in coronal sections of P1 brains. (A) Schematic illustrates positions of coronal sections. (B–M) EGFP-positive SCPN axons (green or white) and Bcl11b (magenta) are shown. While SCPN axons project laterally to STN in Bcl11bflox/flox and Bcl11bflox/flox;Emx1IRES-Cre/+ brains (B, C), SCPN axons mis-project into Bcl11b-labeled STN when Bcl11b is deleted from striatum (D, E; white arrowheads). (F) In Bcl11bflox/flox mice, all SCPN axons enter pons (F’) and do not project into the midbrain area dorsal to pons (F”). (G) In Bcl11bflox/flox;Emx1IRES-Cre/+ mice, SCPN axons entering pons are substantially reduced (G’), and do not project into the midbrain area dorsal to pons (G”). (H) In Bcl11bflox/flox;Gsx2-iCre mice, SCPN axons abnormally target into midbrain (H”), and there is a slight reduction of SCPN axons entering pons (H’). (I) In Bcl11bflox/flox;Emx1IRES-Cre/+;Gsx2-iCre mice, SCPN axons both have reduced entry into pons (I’) and abnormally target into midbrain (I”). Consequently, compared to control (J), SCPN axons reaching medial and caudal brainstem are substantially reduced in Bcl11bflox/flox;Emx1IRES-Cre/+ mice (K), and are slightly reduced in Bcl11bflox/flox;Gsx2-iCre mice (L). Deletion of Bcl11b from both cortex and striatum additively reduces SCPN axons reaching medial and caudal brainstem (M). See also Supplementary Figure 2A–P. Scale bars, 500 µm. (N–R) Simultaneous visualization of cortical efferent axons by Emx1IRES-Flpo;Rosa26frt-LacZ-frt-EGFP and MSN axons by Gsx2-iCre;Rosa26LSL-tdTomato in coronal sections of P1 brains. (N) Schematic illustrates positions of coronal sections. EGFP, cyan. tdTomato, magenta. Bcl11b, yellow. (O) In control brains, EGFP-positive SCPN axons and tdTomato-positive MSN axons project together lateral to STN, which expresses Bcl11b. (Q, Q’, Q”) These EGFP-positive SCPN axons subsequently project into pons at the junction of midbrain and hindbrain surrounded by tdTomato-positive axons. (P) In Gsx2 conditional KO, a subset of axons project abnormally into STN— all are positive for both EGFP and tdTomato (white arrowheads). (R, R’) EGFP-positive SCPN axons abnormally targeting into dorsocaudal midbrain colocalize with similarly mistargeting tdTomato-positive MSN axons (arrowheads). (R, R”) MSN axons reaching substantia nigra and SCPN axons entering pons are reduced compared to control. Scale bars, 200 µm.

Cpd, cerebral peduncle. Ctx, cortex. Hp, hippocampus. STN, subthalamus nucleus. Mb, midbrain. Pyr, pyramidal tract. SN, substantia nigra.

Bcl11b controls expression and subcellular localization of SCPN mRNAs.

(A) Schematic for purification of SCPN somata. Green fluorescent microspheres were injected into the pyramidal tract in rostral brainstem of P0 Emx1-Cre/Rosa26-tdTomato pups, and SCPN somata were purified at P1 as cells double positive for both green microspheres and tdTomato. A representative FACS plot is shown. (B) Changes to gene expression in SCPN somata upon deletion of Bcl11b by Emx1-Cre. 515 differentially expressed genes (DEGs; q < 0.1) are shown in green. Other genes are shown in gray. Total: 10,975 genes. SCPN “identity” genes,16 Bcl11a, and Kif5c are labeled. (C) GO enrichment analysis of soma DEGs. Points are colored by q-value, and sized by number of genes per term.

Cdh13 is normally present on SCPN axons and GC surfaces, and Bcl11b deletion from SCPN reduces Cdh13 on these subcellular compartments.

(A–D) Immunocytochemistry for Cdh13 (cyan) and tdTomato fluorescence (labeling Emx-expressing cell lineage, magenta) in sagittal sections of P1 brains. (A, B, B’) Cdh13 co-localizes with tdTomato-labeled SCPN axons in wild-type brains. (C, D, D’) When Bcl11b is deleted from cortex, SCPN axon growth is perturbed, and Cdh13 abundance along SCPN axons is substantially reduced in brainstem. Cdh13 expression in olfactory bulb is unaffected by Bcl11b deletion from cortex. Scale bars, 1 mm.

(E) Schematic for purification of SCPN GCs. Crude growth cone fraction containing SCPN GCs was collected from brainstem and spinal cord of P1 Emx1-Cre/Rosa26-tdTomato pups by subcellular fractionation, followed by fluorescent small particle sorting (FSPS) for surface protein analysis or for purification and transcriptomic investigation of tdTomato-positive SCPN GCs. (F) tdTomato-positive SCPN GCs isolated from P1 Bcl11b+/+;Emx1IRES-Cre/+;Rosa26LSL-tdTomato/+ or Bcl11bflox/flox;Emx1IRES-Cre/+;Rosa26LSL-tdTomato/+ brainstem/spinal cord were incubated with control IgG antibody, anti-Cdh13 antibody, or anti-L1CAM antibody, then with Alexa488-conjugated secondary antibody, followed by FSPS. Identical gates are applied for analyzing tdTomato-positive SCPN GCs in (G). (G) Quantification was performed for relative protein abundance on tdTomato-positive SCPN GC surfaces in vivo. Dotted line indicates median. Mann-Whitney test.

BS, brainstem. Ctx, cortex. Str, striatum. Mb, midbrain. OB, olfactory bulb. Pyr, pyramidal tract. SC, spinal cord.

Cdh13 is required for SCPN axon extension in the brainstem, and Cdh13 over-expression ameliorates SCPN axon extension defects induced by cell-autonomous Bcl11b deletion.

(A–C) Fluorescence images of E18.5 wild-type brains electroporated in utero at E12.5 with a plasmid encoding control gRNA or Cdh13 gRNA #3, together with those encoding EGFP and dCas9-KRAB-MeCP2. (A, B) EGFP-labeled axons in the cerebral peduncle and brainstem. (C) Quantification at E18.5 of axon abundance in the brainstem relative to axon abundance in the cerebral peduncle reveals that Cdh13 knockdown disrupts SCPN axon extension in the brainstem. Mean ± SEM. Control gRNA vs. Cdh13 gRNA #3; p < 0.05, repeated measures two-way ANOVA. ***; p < 0.001, Bonferroni’s multiple comparison test. (D, E) Fluorescence images of P1 Bcl11bflox/flox mouse brains electroporated in utero at E12.5 with a precisely-defined cocktail of Cre-recombinase, Cre-inducible Cre-recombinase, Cre-silenced myr-Clover3 (Control; green), and Cre-activated myr-tdTomato (KO; magenta). (D) Fluorescently labeled axons in the cerebral peduncle and brainstem. (E) Quantification of axon abundance in the brainstem relative to axon abundance in the cerebral peduncle reveals that Bcl11b knockout disrupts SCPN axon extension in the brainstem. Mean ± SEM. Control vs. KO; p < 0.05, repeated measures two-way ANOVA. ***; p < 0.001, Bonferroni’s multiple comparison test.

(F, G) Fluorescence images of P1 Bcl11bflox/flox mouse brains electroporated in utero at E12.5 with a precisely-defined cocktail of Cre-recombinase, Cre-inducible Cre-recombinase, Cre-silenced myr-Clover3 (Control; green), Cre-activated myr-tdTomato, and Cre-activated Cdh13 (KO + Cdh13; magenta). (F) Fluorescently labeled axons in the cerebral peduncle and brainstem. (G) Quantification at P1 of axon abundance in the brainstem relative to axon abundance in the cerebral peduncle reveals that Cdh13 over-expression ameliorates SCPN axon extension defects in the brainstem caused by Bcl11b knockout. Mean ± SEM. Control vs. KO + Cdh13; p > 0.05, repeated measures two-way ANOVA.

Control samples from control BEAM and Cdh13 rescue BEAM experiments are indistinguishable, so they are combined to increase statistical power.

Scale bars, 100 µm.

Gsx2-iCre is expressed by neurons derived from the lateral and medial ganglionic eminences.

(A) In P0 Gsx2-iCre;Rosa26LSL-LacZ/+ forebrains, β-galactosidase is detected in striatum by LacZ labeling. Scale bar, 1 mm. (B– D) In P14 Gsx2-iCre;Rosa26LSL-LacZ/+ cortex, β-galactosidase co-localizes with parvalbumin, somatostatin, and neuropeptide Y, markers of subpallial-derived interneurons (arrowheads). Scale bar, 100 µm. (E) DARPP-32 is expressed normally in Bcl11bflox/flox;Emx1IRES-Cre/+ striatum, but is substantially reduced in Bcl11bflox/flox;Gsx2-iCre and Bcl11bflox/flox;Emx1IRES-Cre/+;Gsx2-iCre striatum. GAP43 immunolabeling reveals that fasciculation of CFuPN axons in the internal capsule is largely normal in Bcl11bflox/flox;Emx1IRES-Cre/+ striatum, but is substantially disrupted in Bcl11bflox/flox;Gsx2-iCre and Bcl11bflox/flox;Emx1IRES-Cre/+;Gsx2-iCre striatum. Scale bar, 500 µm.

Bcl11b deletion from SCPN disrupts outgrowth of SCPN axons, and Bcl11b deletion from MSN causes abnormal SCPN axon targeting into dorsocaudal midbrain.

(A–P) Cortical efferent axons are visualized anterogradely by Emx1IRES-Flpo;Rosa26frt-LacZ-frt-EGFP in coronal sections of P1 brains. Schematic illustrates positions of coronal sections. (A–D) Cortical efferent axons are visualized anterogradely by Emx1IRES-Flpo;Rosa26frt-LacZ-frt-EGFP in coronal sections of P1 brains. EGFP expression in Emx1IRES-Flpo/+;Rosa26frt-LacZ-frt-EGFP/+ cortex is indistinguishable between wild-type and any of the Bcl11b mutant cortices. (A’–D’) Bcl11b expression in cortex and striatum is ablated by Emx1IRES-Cre and Gsx2-iCre, respectively. (E–H) Innervation of thalamus by CThPN axons is largely unchanged in any of the mutant mice compared to wild-type mice. Compared to wild type (I, M), EGFP-labeled SCPN axons reaching the medulla and decussating are substantially reduced in Bcl11bflox/flox;Emx1IRES-Cre/+ mice (J, N), and are slightly reduced in Bcl11bflox/flox;Gsx2-iCre mice (K, O). (L, P) Deletion of Bcl11b from both cortex and striatum additively reduces SCPN axons reaching the medulla and decussating. Scale bars, 500 µm.

(Q–T) DiI crystals placed in sensorimotor cortex of fixed P0 brains anterogradely label efferent axons, including SCPN axons, as schematized in a sagittal view. (Q–Q’’) DiI labeled axons reaching caudal to thalamus are unequivocally SCPN axons. (Q’) In wild-type (Bcl11bflox/flox) brains, DiI labeled axons fasciculate tightly in the cerebral peduncle (bracket). (Q, Q’’) While some SCPN axons (i.e., corticotectal axons) branch and turn dorsally to enter the lenticular fascicle (yellow arrowheads), others extend further caudally into pons, and leading SCPN axons reach caudal brainstem. (R–R’’) In Bcl11bflox/flox;Emx1IRES-Cre/+ brains, SCPN axons lacking Bcl11b fail to fasciculate tightly in the midbrain, corticotectal axons do not turn precisely to enter the lenticular fascicle (R’, bracket, yellow arrowheads), fewer SCPN axons reach rostral brainstem, and only a few leading axons reach caudal brainstem. (S–S’’) In Bcl11bflox/flox;Gsx2-iCre brains, SCPN axons branch normally and turn toward tectum, and fasciculate largely normally in the cerebral peduncle (S’, bracket, yellow arrowheads). However, a small subset of labeled SCPN axons abnormally target into dorsocaudal midbrain (S, red arrowheads), and there are slightly fewer SCPN axons reaching caudal brainstem (S’’). (T–T’’) SCPN axons in brains lacking Bcl11b in both cortex and striatum are more severely affected than either single mutant. SCPN axons are defasciculated in the cerebral peduncle (T’, bracket), do not branch toward tectum (T’, yellow arrowheads), abnormally target dorsocaudal midbrain (T, red arrowheads), and do not appropriately reach caudal brainstem (T, T’’). Scale bar, 1 mm.

BS, brainstem. Cpd, cerebral peduncle. Ctx, cortex. Str, striatum. Hp, hippocampus. Pyr, pyramidal tract. Th, thalamus.

Bcl11b deletion in cortex substantially reduces corticospinal projections.

(A, B) Fewer SCPN are retrogradely labeled in P25 Bcl11bflox/flox;Emx1IRES-Cre/+ cortex by Alexa-conjugated cholera toxin B (CTB) injection into the cervical spinal cord at P21. (C) Quantification of retrogradely labeled neurons reveals that loss of Bcl11b reduces the number of SCPN axons reaching the spinal cord. Mean ± SEM, n = 4. Mann-Whitney test. *; p < 0.05. Scale bar, 1 mm.

Purification of SCPN somata from wild-type and cortex-specific Bcl11b null mice.

(A) A representative P1 brain image showing green-fluorescent microsphere (cyan) injection into the rostral brainstem of a P0 Emx1-Cre;Rosa26tdTomato mouse, and retrogradely labeled SCPN in cortex. The micropipet was inserted rostrally from caudal brainstem. tdTomato, magenta. DAPI, yellow. (B) Distribution of retrogradely labeled SCPN is largely normal in Bcl11bflox/flox;Emx1IRES-Cre/+ cortex, with minor reduction of retrogradely labeled SCPN in lateral cortex. Green-fluorescent microspheres (cyan) were injected into P1 Emx1-Cre;Rosa26tdTomato mouse, and brains were fixed at P2. tdTomato, magenta. Bcl11b (N-terminus), yellow. (C) FACS-isolated SCPN (tdTomato+ and green microsphere+) per hemisphere is not significantly reduced by Bcl11b deletion from cortex. Green fluorescent microspheres were injected at P0, and SCPN were collected at P1. Mean ± SEM, n = 5. Student’s t-test. (D) Clustering of soma samples by genotype, using the 500 most variable genes. Lower sample-to-sample distances have darker colors, and higher sample-to-sample distances have lighter colors. WT, SCPN somata from Bcl11b+/+;Emx1IRES-Cre/+;Rosa26-tdTomatoflox/+ mice. KO, SCPN somata from Bcl11bflox/flox;Emx1IRES-Cre/+;Rosa26-tdTomatoflox/+ mice. (E) Exon 2 of Bcl11b is deleted in SCPN of Bcl11bflox/flox;Emx1IRES-Cre/+;Rosa26-tdTomatoflox/+ mice. (F) Efficient deletion of Bcl11b in cortex of P1 Bcl11bflox/flox;Emx1IRES-Cre/+ mice, demonstrated by an antibody detecting the C-terminus of Bcl11b (encoded by exon 4). Scale bars, 500 µm.

Cdh13 abundance along SCPN axons decreases after Bcl11b deletion in cortex.

Schematic illustrates positions of coronal sections. Immunocytochemistry of Cdh13 (cyan) and L1CAM (yellow), and tdTomato fluorescence (labeling Emx-expressing cell lineage, magenta) of P1 coronal brain sections. After deletion of Bcl11b from cortex, fewer tdTomato-positive SCPN axons reach brainstem, and Cdh13 abundance along SCPN axons is substantially reduced in brainstem. Note that reduction in L1CAM abundance along SCPN axons is proportional to the reduction in SCPN axons after Bcl11b deletion from cortex. Scale bar, 500 µm.

Cpd, cerebral peduncle. Ctx, cortex. Hp, hippocampus. IC, internal capsule. Pyr, pyramidal tract. Str, striatum.Th, thalamus.

Growth cone surface labeling enables quantitative investigation of antigens on GC surfaces (slFSPS).

(A) Schematic illustrating slFSPS to investigate antigens on GC surfaces. Here, we labeled CPN unilaterally by in utero electroporation at E14.5 with a plasmid encoding EGFP-GPI or cytosolic EGFP, and isolated bulk GCs from the contralateral hemisphere at P3 by subcellular fractionation. After incubation with an antibody conjugated to Alexa647, we diluted and centrifuged GC fractions to remove unbound antibodies, and performed FSPS. (B) EGFP-GPI CPN GCs are labeled by anti-GFP and anti-NCAM1 antibodies. (C) Cytosolic EGFP CPN GCs are labeled by anti-NCAM1 antibody, but not by anti-GFP antibody.

SCPN GC purification and subcellular transcriptomic analyses.

(A) Representative images showing dissection of brainstem and spinal cord. Upper panel shows a sagittal section of P1 Emx1-Cre;Rosa26tdTomato CNS tissue without brainstem/spinal cord dissection. Lower panel shows a sagittal section of P1 Emx1-Cre;Rosa26tdTomato CNS tissue after brainstem/spinal cord dissection. tdTomato, magenta. DAPI, yellow. Scale bar, 500 µm.

(B) A representative FSPS plot is shown.

(C–E) (C) Differences in transcript abundance in Bcl11b+/+;Emx1IRES-Cre/+ (WT) SCPN GCs vs. somata. 2,049 GC-enriched genes (q < 0.1) are shown in orange, and 1,938 soma-enriched (q < 0.1) are shown in purple. Other genes are shown in gray. Total: 6,396 genes. These compartments show highly distinct transcriptomes, highlighting tightly regulated transport of these transcripts along SCPN axons. (D, E) GO enrichment analysis of GC-enriched (D) and soma-enriched (E) genes in Bcl11b+/+;Emx1IRES-Cre/+ SCPN. All (D) and the 10 most significantly enriched (E) terms are shown. Points are colored by q-value, and sized by number of genes per term. GC-enriched genes (q < 0.1) are over-represented for GO terms associated with translation and mitochondrial functions (see also Supplementary Table 9), consistent with previous findings from distinct neuronal populations,31,62 suggesting that GCs across distinct neuronal subtypes have partially shared transcriptomic characteristics. Soma-enriched genes (q < 0.1) are over-represented for GO terms associated with neuronal differentiation and functions (see also Supplementary Table 9), reflecting SCPN soma processes at P1. Together, these findings demonstrate successful SCPN GC purification.

(F, G) (F) Clustering of GC samples, using the 500 most variable genes. Lower sample-to-sample distances have darker colors, and higher sample-to-sample distances have lighter colors. WT, SCPN GC from Bcl11b+/+;Emx1IRES-Cre/+;Rosa26-tdTomatoflox/+ mice. KO, SCPN GC from Bcl11bflox/flox;Emx1IRES-Cre/+;Rosa26-tdTomatoflox/+ mice. (G) Changes to transcript abundance in SCPN GCs upon deletion of Bcl11b by Emx1-Cre. 88 differentially abundant genes (DAGs; q < 0.1) are shown in red. Other genes are shown in gray. Total: 7,794 genes. DAGs include genes relevant to cytoskeleton (Septin4, 49314228F04Rik (recently named Matcap1)), trafficking (Rab3d, Grip2), signaling (Tradd), protein modification (Josd1, Uba3), glycosylphosphatidylinositol (GPI)-anchor biosynthesis (Pigq), and adhesion (Clstn3).

Manipulation of Cdh13 expression in vivo.

(A) Fluorescence images of E17.5 wild-type brains electroporated in utero at E14.5 with a plasmid encoding the indicated gRNA, together with those encoding EGFP and dCas9-VPR. A littermate brain without electroporation is shown as a negative control. All 4 Cdh13 gRNAs induce ectopic Cdh13 expression (arrowheads), while control gRNA does not. None of the gRNAs alter L1CAM expression. EGFP, cyan. Cdh13, magenta. L1CAM, yellow. Scale bar, 200 µm.

(B) Fluorescence images of E18.5 wild-type brains electroporated in utero at E12.5 with a plasmid encoding control gRNA or Cdh13 gRNA#3, together with those encoding EGFP and dCas9-KRAB-MeCP2. Areas in the inset are enlarged in B’. EGFP-labeled axons in striatum derived from neurons expressing control gRNA colocalize with Cdh13 and L1CAM (B’, arrowheads). EGFP-labeled axons in striatum derived from neurons expressing Cdh13 gRNA#3 exhibit reduced Cdh13, and unchanged L1CAM immunolabeling (B’, arrowheads). EGFP, cyan. Cdh13, magenta. L1CAM, yellow. Scale bar, 500 µm.

(C) Immunoblot analysis of Cdh13 overexpression in HEK293T cells. Cells were transfected with plasmids encoding Cre recombinase and/or Cre-activated Cdh13 (FLEX-Cdh13), then cultured for 1 day.

(D) Fluorescence images of P1 Bcl11bflox/flox brains electroporated in utero at E12.5 in control BEAM or Cdh13 rescue BEAM experiments. Fluorescently labeled cells in the cortex are shown. mClover3, green. tdTomato, magenta. Scale bar, 100 µm.

(E) Fluorescence images of P1 Bcl11bflox/flox brains electroporated in utero at E12.5 in Cdh13 rescue BEAM experiments. Fluorescently labeled cells and axons in the forebrain are shown. Cdh13 abundance is increased in Cre+ tdTomato-labeled somata and axons (insets 1 and 2, yellow arrowheads). mClover3, green. tdTomato, magenta. Cdh13, white. Scale bar, 200 µm. Ctx, cortex. Str, striatum. Hp, hippocampus.