Genetic dissection of the glutamatergic neuron system in cerebral cortex

Diverse types of glutamatergic pyramidal neurons mediate the myriad processing streams and output channels of the cerebral cortex1,2, yet all derive from neural progenitors of the embryonic dorsal telencephalon3,4. Here we establish genetic strategies and tools for dissecting and fate-mapping subpopulations of pyramidal neurons on the basis of their developmental and molecular programs. We leverage key transcription factors and effector genes to systematically target temporal patterning programs in progenitors and differentiation programs in postmitotic neurons. We generated over a dozen temporally inducible mouse Cre and Flp knock-in driver lines to enable the combinatorial targeting of major progenitor types and projection classes. Combinatorial strategies confer viral access to subsets of pyramidal neurons defined by developmental origin, marker expression, anatomical location and projection targets. These strategies establish an experimental framework for understanding the hierarchical organization and developmental trajectory of subpopulations of pyramidal neurons that assemble cortical processing networks and output channels.

Pyramidal neurons (PyNs) constitute the large majority of nerve cells in the cerebral cortex and mediate all of the inter-areal processing streams and output channels 1,2,4 . Traditionally, PyNs have been classified into several major classes according to their laminar location and broad axon projection targets, such as intratelencephalic (IT) and extratelencephalic (ET or corticofugal), which further comprises subcerebral (including pyramidal tract; PT) and corticothalamic (CT) PyNs 1 . Within these classes, subsets of PyNs form specific local and long-range connectivity, linking discrete microcircuits to cortical subnetworks and output channels 1,5 . Single-cell transcriptome analysis suggests that there are over fifty PyN transcriptomic types 6 . However, genetic tools and strategies for experimentally accessing PyN subpopulations are limited.
All PyNs are generated from neural progenitors in the embryonic dorsal telencephalon, where regionally differentiated radial glial progenitors (RGs) undergo asymmetric divisions, giving rise to radial clones of PyNs that migrate to the cortex in an inside-out order 7 . RGs generate PyNs either directly or indirectly through intermediate progenitors (IPs), which divide symmetrically to generate pairs of PyNs 8 . A set of temporal patterning genes drive lineage progression in RGs, which unfold a conserved differentiation program in successively generated postmitotic neurons 3,4,9 . Resolving the lineage organization of diverse progenitors and their relationship to projection-defined PyN subpopulations requires fate-mapping tools with cell type and temporal resolution 2 .
Here we present strategies and a genetic toolkit in the mouse for targeting PyN subpopulations and progenitors guided by knowledge of their developmental programs. We leverage gene expression patterns of the cell-type specification and differentiation programs to target biologically significant progenitor subsets, PyN subpopulations and their developmental trajectories ( Fig. 1a-c, Extended Data Table 1). These tools and strategies provide a roadmap for accessing hierarchically organized PyN types at progressively finer resolution. They will facilitate the tracking of developmental trajectories of PyNs for elucidating the organization and assembly of neural circuits of the cerebral hemisphere, including the cortex, hippocampus and basolateral amygdala. embryonic stages to reveal these progenitors and their lineage progression, as well as their PyN progeny in the mature cortex ( Fig. 1d- At embryonic day (E) 10.5, a 24-hour tamoxifen pulse-chase in Lhx2-CreER;Ai14 embryos resulted in dense labelling of neuroepithelial cells and RGs in the dorsal pallium, with a sharp border at the cortex-hem boundary (Extended Data Fig. 1a). E12.5-E13.5 pulse-chase revealed a prominent medial high to lateral low gradient of RGs Lhx2 (Fig. 1d), suggesting differentiation of the earlier RGs. E13.5-E14.5 pulse-chase showed a similar gradient pattern at a lower cell density (Extended Data Fig. 1e). Fate-mapping from E10.5-P30, E12.5-P30 and E14.5-P30 labelled PyN progeny across cortical layers (Fig. 1e, Extended Data Fig. 1b-f, p), suggesting multipotency of RGs Lhx2 at these stages. During postnatal development, the expression of Lhx2 became postmitotic: pulse-chase in P5 labelled largely IT PyNs across layers and in the second postnatal week labelled more astrocytes (around 60%) than PyNs across layers (Extended Data Fig. 1q-r).
Article demonstrated that Tis21-Fezf2 intersection specifically labelled a set of pallial nRG Fezf2+ with enhanced green fluorescent protein (EGFP), whereas Tis21-Fezf2 subtraction labelled pallial and subpallial nRG Fezf2− with red fluorescent protein (RFP). Pallial nRGs consisted of both Fezf2 + and Fezf2 − subpopulations, suggesting heterogeneity. E12.5-P30 fate-mapping in these mice revealed three types of PyN clones (Extended Data Fig. 3i-k). RFP-only clones are likely to have derived from nRG Fezf2− in which Tis21-CreER activated RFP expression; they probably consisted of PyNs that did not express Fezf2 at any stage. EGFP-only clones are likely to have derived from nRG Fezf2+ , in which Tis21-CreER and Fezf2-Flp co-expression activated EGFP in the IS reporter allele. Mixed clones containing both EGFP and RFP cells probably derived from nRG Fezf2− in which Tis21-CreER activated RFP expression followed by postmitotic activation of EGFP through Fezf2-Flp. Together, these results indicate the presence of nRG Fezf2+ and nRG Fezf2− , both multipotent in generating PyNs across all cortical layers.

IPs
IPs and indirect neurogenesis have evolved largely in the mammalian lineage and have further expanded in primates 14,15 . Along the neural tube, IP-mediated indirect neurogenesis is restricted to the telencephalon and is thought to contribute to the expansion of cell numbers and diversity in the neocortex. The majority of PyNs in mouse cortex are produced through IPs 16,17 , but the link between indirect neurogenesis and PyN types remains unclear. The T-box transcription factor Tbr2 (also known as Eomes) is expressed in pallial IPs throughout indirect neurogenesis 18 . E16.5 pulse-chase in the Tbr2-CreER driver line specifically labelled IPs (Extended Data Fig. 3a, c). E16.5 and E17.5 fate-mapping labelled PyNs in L2/3 and upper L2, respectively (Fig. 2c, d, Extended Data Fig. 3d). Therefore, the Tbr2-CreER driver enables highly restricted laminar targeting of PyN subpopulations in supragranular layers. Furthermore, Tis21-CreER and Tbr2-FlpER intersection enabled specific targeting of neurogenic but not the transit-amplifying IPs (Extended Data Fig. 3a, e). Altogether, these progenitor driver lines facilitate dissecting progenitor diversity and tracking the developmental trajectories of PyNs from their lineage origin to circuit organization.

Targeting PyN subpopulations
We generated driver lines targeting PyN subpopulations and characterized these in comparison to existing lines where feasible (Fig. 2 Tables 1, 2), Cux1-CreER is unique in targeting predominantly cortex-but not striatum-projecting IT subpopulations.
The supragranular layers comprise diverse IT types 20 , but only a few L2/3 drivers have been reported so far 24 and none distinguish L2 versus L3 PyNs. We used a lineage and birth dating approach to dissect L2/3 PyNs. In our Tbr2-CreER driver targeting IPs, tamoxifen induction at E16.5 and E17.5 specifically labelled PyNs in L2/3 and L2, respectively (Fig. 2c, d). Combined with the CreER to Flp conversion strategy that converts lineage and birth timing signals to permanent Flp expression 13 , this approach enables adeno-associated virus (AAV) manipulation of L2 and L3 IT neurons.

PT drivers
After early expression in a subset of dorsal pallial progenitors, Fezf2 becomes restricted to postmitotic L5/6 corticofugal PyNs, with higher levels in L5B PT neurons and lower levels in a subset of CT neurons 10,27 . At postnatal stages, Fezf2 drivers labelled projection neurons in the cerebral cortex, hippocampus, amygdala, and olfactory bulb (Extended Data Fig. 4 vibrissal secondary motor area (MOs)) and subcortical regions including the striatum, thalamic ventral posteromedial nucleus of the thalamus (VPM) and posterior complex of the thalamus (PO), anterior pretectal nucleus, ipsilateral superior colliculus (iSC), pontine nucleus, corticospinal tract (CST) and contralateral spinal trigeminal nucleus (cSp5) (Fig. 3b, Fig. 6). It remains to be determined whether PyNs Tbr1 with contralateral projections (Fig. 3f) represent a distinct type.
Tle4 is a transcription corepressor that is expressed in a subset of CT PyNs 31,32 . Our Tle4-CreER driver specifically labelled L6 CT PyNs across the cortex (Fig. 2i Foxp2 is expressed in many CT neurons from the postmitotic stage to the mature cortex 33-35 . In adult Foxp2-IRES-Cre mice 36 , systemic injection of Cre-dependent AAV9-DIO-GFP specifically labelled L6 PyNs; Foxp2 + cells were also found in the striatum, thalamus, hypothalamus, midbrain, cerebellum and inferior olive (  To further characterize several PyN driver lines, we performed a set of histochemical analyses (Extended Data Fig. 6). PyNs targeted in Fezf2, Tcerg1l and Adcyap1 drivers extensively co-labelled with PT markers. PyNs targeted in Tle4 and Tbr1 drivers co-labelled with CT markers. The laminar patterns and class-specific marker expression in these driver lines precisely recapitulated endogenous patterns (in situ hybridization data in the Allen Brain Map: Mouse Brain Atlas; https://mouse.brain-map.org/search/index), providing further evidence of the reliability and specificity of these driver lines.

Combinatorial targeting of projection types
To further dissect driver-line-defined subpopulations according to projection targets, we first used retrograde tracing. Within the PT population, retroAAV and fluorogold injections in the spinal cord of Fezf2-CreER mice specifically labelled L5B corticospinal PyNs in the sensorimotor cortex (Extend Data Fig. 10a-c, Supplementary Table 6). To explore PyNs Fezf2 subpopulations jointly defined by projection targets and sublaminar position, we used the IS reporter 13 . Consistent with previous findings 37 , PyNs Fezf2 that project to the thalamus and medulla resided in the upper and lower sublamina of L5B in the primary motor area (MOp), respectively (Extended Data Fig. 9d-f). In SSp-bfd, PyNs Fezf2 with collaterals to the striatum resided in upper L5, those with collaterals to the superior colliculus or cSp5 resided in the middle and lower portion of L5B, and those projecting to thalamic POm resided both in middle to lower L5B and in L6 (Extended Data Fig. 9g-h, l-o). We then distinguished subsets of L5B PyNs Fezf2 according to their expression of the calcium-binding protein parvalbumin using Fezf2-CreER;Pv-Flp;IS mice that differentially labelled PyNs Fezf2+/PV− and PyNs Fezf2+/PV+ (in which PV represents parvalbumin; this gene is also known as Pvalb) (Extended Data Fig. 9i, j). Compared to PyNs Fezf2+/PV− , PyNs Fezf2+/PV+ exhibited more depolarized resting membrane potentials. In addition, we designed a strategy (triple trigger) to target PyNs Fezf2 jointly defined by a driver line, a projection target and a cortical location (Extended Data Fig. 10).
We also used retroAAV to dissect the CT and IT populations. In Tle4-CreER;IS mice, retrograde tracing from the thalamic VPM revealed two subpopulations of L6 PyNs Tle4 , one extending apical dendrites to the L4/5 border, the other to L1 (Extended Data Fig. 9q), suggesting differential inputs. In Plxnd1-CreER mice (Extended Data Fig. 9p, r-w), whereas L5A PyNs Plxnd1 projected to both the ipsi-and the contralateral striatum, L2/3 PyNs Plxnd1 projected mostly to the ipsilateral striatum.
In addition, consistent with the finding that some PyNs Fezf2 extend contralateral cortical and striatal projections (Fig. 3e), retrograde cholera toxin subunit B (CTB) tracing from the striatum labelled a set of contralateral PyNs Fezf2+ at the L5A-L5B border (Extended Data Fig. 11a-e), a characteristic IT feature. Indeed, a small set of PyNs at the L5A-L5B border co-expressed Fezf2 and Plxnd1 mRNAs; these PyNs Fezf2/Plxnd1 occupied the very top sublayer of the PyN Fezf2 population (Extended Data Fig. 11f-h), and thus probably contributed to their contralateral cortical and striatal projections (Fig. 3a, b, d, e). Single-cell reconstruction may reveal whether PyNs Fezf2/Plxnd1 are typical IT cells or also project subcortically and represent an 'intermediate PT-IT' type.

Discussion
Together with previous resources 24,38,39 , the PyN driver lines we present here provide much improved specificity, coverage and robustness for a systematic dissection of PyN organization from broad subclasses to finer types. By focusing on driver lines that recapitulate the expression of key transcription factors and effector genes that are implicated in specification and differentiation, these tools will enable the dissection and fate-mapping of biologically significant subpopulations of PyNs through their inherent developmental, anatomical and physiological properties; that is, 'carving nature at its joints'. The precision and reliability of these drivers also allows the combinatorial targeting of finer projection types through the intersection of molecular, developmental and anatomical properties. The inducibility of driver lines enhances the specificity and flexibility of cell targeting, manipulation and fate-mapping. Inducibility also allows control over the density of labelling and manipulation, from dense coverage to single-cell analysis-the ultimate resolution for clarifying the stereotypical and variable features of neurons within marker-defined subpopulations 40-42 . Temporal control allows gene manipulations at different developmental stages to discover the cellular and molecular mechanisms of circuit development and function. Together, these tools and strategies establish a roadmap for dissecting the hierarchical organization of PyN types on the basis of their inherent biology. The incorporation of recently developed enhancer AAVs 43 with these driver lines may further increase the specificity, ease and throughput of cell-type access.
Several transcription factors used in this study (for example, Cux1, Fezf2, Tbr1, Tbr2 and Foxp2) continue to evolve and diverge in primates 44 and are implicated in developmental disorders such as autism 23,35 . Our transcription factor driver lines provide handles to track the developmental trajectories of PyN subpopulations in cortical circuit assembly, with implications in the cross-species evolution of PyNs and for deciphering the genetic architecture of neurodevelopmental disorders.

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-021-03955-9.

Data reporting
No statistical methods were used to predetermine sample size. The experiments were not randomized and the investigators were not blinded to allocation during experiments and outcome assessment.

Generation of knock-in mouse lines
Driver and reporter mouse lines listed in Supplementary Table 1 were generated using a PCR-based cloning, as described before and below 13,45 . All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the Cold Spring Harbor Laboratory (CSHL) in accordance with NIH guidelines. Mouse knock-in driver lines are deposited at The Jackson Laboratory for wide distribution. Knock-in mouse lines were generated by inserting a 2A-CreER or 2A-Flp cassette in-frame before the STOP codon of the targeted gene. Targeting vectors were generated using a PCR-based cloning approach as described before. In brief, for each gene of interest, two partially overlapping BAC clones from the RPCI-23&24 library (made from C57BL/b mice) were chosen from the Mouse Genome Browser. The 5′ and 3′ homology arms were PCR-amplified (2-5 kb upstream and downstream, respectively) using the BAC DNA as template and cloned into a building vector to flank the 2A-CreERT2 or 2A-Flp expressing cassette as described 47 . These targeting vectors were purified and tested for integrity by enzyme restriction and PCR sequencing. Linearized targeting vectors were electroporated into a 129SVj/B6 hybrid embryonic stem (ES) cell line (v6.5). ES clones were first screened by PCR and then confirmed by Southern blotting using appropriate probes. DIG-labelled Southern probes were generated by PCR, subcloned and tested on wild-type genomic DNA to verify that they give clear and expected results. Positive v6.5 ES cell clones were used for tetraploid complementation to obtain male heterozygous mice following standard procedures. The F 0 males were bred with reporter lines (Supplementary Tables 1, 3, 4) and induced with tamoxifen at the appropriate ages to characterize the resulting genetically targeted recombination patterns.

Tamoxifen induction
Tamoxifen (T5648, Sigma) was prepared by dissolving in corn oil (20 mg ml −1 ) and applying a sonication pulse for 60 s, followed by constant rotation overnight at 37 °C. Embryonic inductions for most knock-in lines were done in the Swiss-Webster background; inductions for Tis21-CreER, Fezf2-Flp intersection experiments were done in the C57BL6 background. E0.5 was established as noon on the day of vaginal plug and tamoxifen was administered to pregnant mothers by gavage at a dose varying from 2-100 mg kg −1 at the appropriate age. For embryonic collection (12-24 h pulse-chase experiments), a dose of 2mg kg −1 was administered to pregnant dams via oral gavage. For postnatal induction, a 100-200 mg kg −1 dose was administered by intraperitoneal injection at the appropriate age.

Immunohistochemistry
Postnatal and adult mice were anaesthetized (using Avertin) and intracardially perfused with saline followed by 4% paraformaldehyde (PFA) in 0.1 M PB. After overnight post-fixation at 4 °C, brains were rinsed three times and sectioned at a 50-75-µm thickness with a Leica 1000s vibratome. Embryonic brains were collected in PBS and fixed in 4% PFA for 4 h at room temperature, rinsed three times with PBS, dehydrated in 30% sucrose-PBS, frozen in OCT compound and cut by cryostat (Leica, CM3050S) in 20-50-µm coronal sections. Early postnatal pups were anaesthetized using cold shock on ice and intracardially perfused with 4% PFA in PBS. Post-fixation was performed similarly to older mice. Postnatal mice aged 1-2 months were anaesthetized using Avertin and intracardially perfused with saline followed by 4% PFA in PBS; brains were post-fixed in 4% PFA overnight at 4 °C and subsequently rinsed three times, embedded in 3% agarose-PBS and cut to a 50-100-µm thickness using a vibrating microtome (Leica, VT100S). Sections were placed in blocking solution containing 10% normal goat serum (NGS) and 0.1% Triton-X100 in PBS1X for 1 h, then incubated overnight at 4 °C with primary antibodies diluted in blocking solution. Sections were rinsed three times in PBS and incubated for 1 h at room temperature with corresponding secondary antibodies (1:500, Life Technologies). Sections were washed three times with PBS and incubated with DAPI for 5 min (1:5,000 in PBS, Life Technologies, 33342) to stain nuclei. Sections were dry-mounted on slides using Vectashield (Vector Labs, H1000) or Fluoromount (Sigma, F4680) mounting medium.
To perform molecular characterization of GeneX-CreER mouse lines, we stained 40-µm vibratome sections for CUX1 and CTIP2, that were imaged in a Nikon Eclipse 90i fluorescence microscope. Focusing on the somatosensory cortex, we counted tdTomato + cells in a column of around 300-µm width and determined their relative position along the dorso-ventral axis that goes from the ventricular surface (0) to the pia (100%). As a reference, CTIP2 + and CUX1 + regions were plotted as For colocalization determination, we obtained confocal z-stacks centred in layer 5 or 6 of the somatosensory cortex, of 320 × 320 × 40 µm 3 volumes. For all tdTomato + cells in the volume, we manually determined whether they were also positive for the desired markers by looking in individual z-planes. The percentage of positive cells was calculated for each area. Average number of tdTomato + cells quantified per staining:

Validation of PyN driver lines
ViewRNA tissue Assay (Thermo Fisher Scientific) fluorescent in situ hybridization (FISH) was carried out as per the manufacturer's instructions on genetically identified PyNs expressing H2bGFP nuclear reporter (GeneX-CreER;LSL-H2bGFP) to validate the expression of PyN mRNA within Cre-recombinase dependent H2bGFP expressing cells in adult tissue (p24). Antibody validation with Cre-recombinase dependent reporter (GeneX-CreER;Ai14) was also used as it was available for use in adult tissue. For both FISH and antibody validation experiments, the percentage of total recombinase-dependent reporter-positive cells co-expressing PyN driver transcript or antibody was quantified.

Viral injection and analysis
Stereotaxic viral injection. Adult mice were anaesthetized by inhalation of 2% isofluorane delivered with a constant air flow (0.4 l min −1 ). Ketoprofen (5 mg kg −1 ) and dexamethasone (0.5 mg kg −1 ) were administered subcutaneously as preemptive analgesia and to prevent brain oedema, respectively, before surgery, and lidocaine (2-4 mg kg −1 ) was applied intra-incisionally. Mice were mounted in a stereotaxic headframe (Kopf Instruments, 940 series or Leica Biosystems, Angle Two). Stereotactic coordinates were identified (Supplementary Table 5). An incision was made over the scalp, a small burr hole drilled in the skull and brain surface exposed. Injections were performed according to the strategies delineated in Supplementary Table 5. A pulled glass pipette tip of 20-30 µm containing the viral suspension was lowered into the brain; a 300-400 nl volume was delivered at a rate of 30 nl min −1 using a Picospritzer (General Valve Corp); the pipette remained in place for 10 min preventing backflow, prior to retraction, after which the incision was closed with 5/0 nylon suture thread (Ethilon Nylon Suture, Ethicon) or Tissueglue (3M Vetbond), and mice were kept warm on a heating pad until complete recovery.

Systemic AAV injection.
Foxp2-IRES-Cre mice were injected through the lateral tail vein at 4 weeks of age with a 100 µl total volume of AAV9-CAG-DIO-EGFP (UNC Viral Core) diluted in PBS (5 × 10 11 vg per mouse). Three weeks after injection, mice were transcardially perfused with 0.9% saline, followed by ice-cold 4% PFA in PBS, and processed for serial two-photon (STP) tomography.

Viruses
AAVs serotype 8, 9, DJ PHP.eB or rAAV2-retro (retroAAV) packaged by commercial vector core facilities (UNC Vector Core, ETH Zurich, Biohippo, Penn, Addgene) were used as listed in Supplementary Table 5. In brief, for cell-type-specific anterograde tracing, we used either Cre-or Flp-dependent or tTA-activated AAVs combined with the appropriate reporter mouse lines 28 (Supplementary Table 7), or dual-tTA ( Fig. 4 and Extended Data Fig. 10) to express EGFP, EYFP or mRuby2 in labelled axons. retroAAV-Flp was used to infect axons at their terminals 46 in target brain structures to label PyNs retrogradely according to the experiments detailed in Supplementary Table 5.

Microscopy and image analysis
Imaging was performed using Zeiss LSM 780 or 710 confocal microscopes, Nikon Eclipse 90i or Zeiss Axioimager M2 fluorescence microscopes, or whole-brain STP tomography (detailed below). Imaging from serially mounted sections was performed on a Zeiss LSM 780 or 710 confocal microscope (CSHL St. Giles Advanced Microscopy Center) and Nikon Eclipse 90i fluorescence microscope, using objectives ×63 and ×5 for embryonic tissue, and ×20 for adult tissue, as well as ×5 on a Zeiss Axioimager M2 System equipped with MBF Neurolucida Software (MBF). Quantification and image analysis was performed using Image J/FIJI software. Statistics and plotting of graphs were done using GraphPad Prism 7 and Microsoft Excel 2010.

Twenty-four-hour pulse-chase embryonic experiments.
For 24-hour pulse-chase embryonic experiments (Fig. 1, Extended Data  Fig. 1), high-magnification insets are not maximum intensity projections. To observe the morphology of RGs, only a few sections from the z-plane in low-magnification images have been projected in the high-magnification images.
Target-specific cell depth measurement. Cell depth analysis for retrogradely labelled projection-specific genetically identified PyNs (GeneX-CreER) were obtained using 5× MBF fluorescent widefield images of 65-µm thick coronal sections in MO and SSp-bfd. MO cell depths are presented in micrometres owing to the absence of a defined white matter border in frontal cortical areas and SSp-bfd depth ratio measurements were normalized to the distance from pia to white matter. For each condition we quantified at least four sections taken from two mice.

Whole-brain STP tomography and image analysis
Perfused and post-fixed brains from adult mice were embedded in oxidized agarose and imaged with TissueCyte 1000 (Tissuevision) as described 48,49 . We used the whole-brain STP tomography pipeline previously described 48,49 . Perfused and post-fixed brains from adult mice, prepared as described above, were embedded in 4% oxidized-agarose in 0.05 M PB, cross-linked in 0.2% sodium borohydrate solution (in 0.05 M sodium borate buffer, pH 9.0-9.5).The entire brain was imaged in coronal sections with a 20× Olympus XLUMPLFLN20XW lens (NA 1.0) on a TissueCyte 1000 (Tissuevision) with a Chameleon Ultrafast-2 Ti:Sapphire laser (Coherent). EGFP/EYFP or tdTomato signals were excited at 910 nm or 920 nm, respectively. Whole-brain image sets were acquired as series of 12 (x) × 16 (y) tiles with 1 µm × 1 µm sampling for 230-270 z sections with a 50-µm z-step size. Images were collected by two PMTs (PMT, Hamamatsu, R3896), for signal and autofluorescent background, using a 560-nm dichroic mirror (Chroma, T560LPXR) and band-pass filters (Semrock FF01-680/SP-25). The image tiles were corrected to remove illumination artifacts along the edges and stitched as a grid sequence 47,49 . Image processing was completed using ImageJ/FIJI and Adobe/Photoshop software with linear level and nonlinear curve adjustments applied only to entire images.
Cell body detection from whole-brain STP data. PyN somata were automatically detected from cell-type specific reporter lines (R26-LSL-GFP or Ai14) by a convolutional network trained as described previously 48 . Detected PyN soma coordinates were overlaid on a mask for cortical depth, as described 48 .
Axon detection from whole-brain STP data. For axon projection mapping, PyN axon signal based on cell-type-specific viral expression of EGFP or EYFP was filtered by applying a square root transformation, histogram matching to the original image, and median and Gaussian filtering using Fiji/ImageJ software 50 so as to maximize signal detection while minimizing background auto-fluorescence, as described before 51 . A normalized subtraction of the autofluorescent background channel was applied and the resulting thresholded images were converted to binary maps. Three-dimensional rendering was performed on the basis of binarized axon projections and surfaces were determined based on the binary images using Imaris software (Bitplane). Projections were quantified as the fraction of pixels in each brain structure relative to each whole projection.
Axon projection cartoon diagrams from whole-brain STP data. To generate cartoons of axon projections for a given driver line, axon detection outputs from all individual experiments were compared (sorting the values from high to low), and analysed side-by-side with low-resolution image stacks (and the CCFv3 registered to the low-resolution dataset for brain area definition) to get a general picture of the injection, as well as high-resolution images for specific brain areas.
Registration of whole-brain STP image datasets. Registration of brain-wide datasets to the Allen reference Common Coordinate Framework (CCF) version 3 was performed by 3D affine registration followed by a 3D B-spline registration using Elastix software 52 , according to established parameters 52 . For cortical depth and axon projection analysis, we registered the CCFv3 to each dataset so as to report cells detected and pixels from axon segmentation in each brain structure without warping the imaging channel.
Patch-clamp recording in brain slices. Patch pipettes were pulled from borosilicate glass capillaries with filament (1.2 mm outer diameter and 0.69 mm inner diameter; Warner Instruments) with a resistance of 3-6 MΩ. The pipette recording solution consisted of 130 mM potassium gluconate, 15 mM KCl, 10 mM sodium phosphocreatine, 10 mM HEPES, 4 mM ATP·Mg, 0.3 mM GTP and 0.3 mM EGTA (pH 7.3 adjusted with KOH, 300 mOsm). Dual or triple whole-cell recordings from tdTomato + and EGFP + PyNs were made with Axopatch 700B amplifiers (Molecular Devices) using an upright microscope (Olympus, BX51) equipped with infrared-differential interference contrast optics (IR-DIC) and a fluorescence excitation source. Both IR-DIC and fluorescence images were captured with a digital camera (Microfire, Optronics). All recordings were performed at 33-34 °C with the chamber perfused with oxygenated working ACSF.
Recordings were made with two MultiClamp 700B amplifiers (Molecular Devices). The membrane potential was maintained at −75 mV in the voltage clamping mode and zero holding current in the current clamping mode, without the correction of junction potential. Signals were recorded and filtered at 2 kHz, digitalized at 20 kHz (DIGIDATA 1322A, Molecular Devices) and further analysed using the pClamp 10.3 software (Molecular Devices) for intrinsic properties.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this paper.

Extended Data Fig. 3 | Fate-mapping neurogenic and intermediate progenitors. a, Fate-mapping strategy for intermediate progenitors (IPs) and
indirect neurogenesis, used in c-e. Tbr2-CreER labels an IP and all of its progeny with a fluorescent marker when combined with Ai14 (Left). The intersection of Tis21-CreER and Tbr2-FlpER specifically targets neurogenic IPs when combined with the Ai65 intersectional reporter (Right). b, Simultaneous fate-mapping of different molecularly defined neurogenic RGs using an intersection/ subtraction reporter (IS) combined with Tis21-CreER and Fezf2-Flp drivers. This scheme is used in h-k. In RGs Tis21+Fezf2− , Cre activates RFP expression only. In RGs Tis21+Fezf2+ , Cre and Flp recombinations remove the RFP cassette and activate EGFP expression. At a later stage when Fezf2 is only expressed in postmitotic deep layer PyNs, Tis21-CreER in RGs activates RFP expression in all of its progeny, but RFP is then switched to EGFP only in Fezf2 + PyNs expressing Flp. c, E16.5 IPs densely labelled by 12-hour pulse-chase in Tbr2-CreER;Ai14 mice; magnified view shows IP somata (arrowhead) away from the lateral ventricle (dashed line) lacking radial fibers and endfeet. d, Fate-mapping E16.5 IPs in Tbr2-CreER;Ai14mice labels PyNs in L2-3 cortex at P28. e, Intersectional fate-mapping of neurogenic IPs at E16.5 in Tis21-CreER;Tbr2-FlpER;Ai65 mice, as depicted in a, labelled L2-3 PyN progeny in P28 cortex. f, 48-hr pulse-chase in E10.5 Tis21-CreER;Ai14 embryo labels Tis21 + neurogenic progenitors (nRGs) and their postmitotic progeny throughout the neural tube, including dorsal pallium (high magnification). Self-renewing RGs are identified by their endfeet at the ventricular surface (arrowheads) and radial fibers (arrow). g, Fate-mapping of E10.5 nRGs to mature cortex reveals PyNs are distributed throughout cortical layers. Note that multipolar GABAergic interneurons (some in layer 1) derived from subpallium nRGs are also labelled (arrowheads). h, The presence of nRGs Fezf2− and nRGs Fezf2+ at E11.5 is revealed by intersection/ subtraction fate-mapping with 24-hour pulse-chase in Tis21-CreER;Fezf2-Flp;IS mice, schematized in b; magnified view shows RFP-labelled nRGs Fezf2− and EGFP-labelled nRGs Fezf2+ . i, Fate-mapping E11.5 nRGs using Tis21-CreER; Fezf2-Flp;IS mice. The mixed RFP and EGFP clone is likely to have derived from a nRG Fezf2− , which activated RFP expression in all progeny and EGFP expression was then switched on only in Fezf2 + postmitotic deep layer PyNs expressing Flp. j, k, More examples of differential fate-mapping of nRGs Fezf2− and nRGs Fezf2+ from E12.5 to the mature cortex using the scheme in b. The majority of clones consist of mixed RFP and EGFP PyNs (j'), and rarely RFP-only (j") or EGFP-only PyNs (k). RFP-only clones (1 of 29) probably derive from nRG Fezf2− whose progeny were all Fezf2− (j"). EGFP-only clones (2 of 29) are derived from nRGs Fezf2+ , suggesting multipotency of RGs Fezf2+ (k). Mixed RFP/EGFP clones are most prominent and are likely to result from Cre activation of RFP in nRGs Fezf2− and subsequent Flp activation of EGFP in Fezf2 + L5/6 postmitotic PyNs (i,j'). Scale bars: 500µm in d, j, k; 100µm in d (high mag), e, f, g, h, i, j',j", k (high mag); 20µm in c, e, h.

Article
Extended Data Fig. 4 | Comparison of new with existing driver lines in terms of areal and laminar patterns. a, Side-by-side comparison of Cre recombination patterns from 8 mouse driver lines characterized in this study (blue font) and 4 existing driver lines (black font) visualized through reporter expression (green; background autofluorescence in red), grouped according to IT, PT and CT projection classes. First row: coronal hemisections at Bregma -1.7 mm. Second row: Image panel showing cortical depth detailing cell body distribution pattern of PyN subpopulations within SSp-bfd taken from the hemisection above at level Bregma -1.7 mm. Third row: coronal hemisections at Bregma 0 mm. Image panel showing cortical depth detailing cell body distribution pattern of PyN subpopulations within MOp taken from the hemisection above at level Bregma 0 mm. For comparison to the PT driver Sim1-Cre transgenic line, see reference 37. b, Cortex-wide distribution patterns of PyN subpopulations viewed as cortical flatmaps in a side-by-side comparison of 8 newly generated (blue font) and 4 existing driver lines (black font): first row, normalized for each dataset's total number of cells detected; second row, absolute scale per flatmap grid area, with maximum number of cells for any PyN subpopulation. Arrowheads indicate gaps in expression and labelling. b' shows cortical flat-mapping coordinate space and two exemplary coronal hemisections describing the demarcations used to generate the cortical grid for flatmapping 48 . c, Overview of brain-wide cell body distribution patterns for each driver line. This table provides an overall impression of the recombination patterns in major adult brain regions in selected lines. d, Histograms showing normalized laminar distribution for six genetically targeted PyN subpopulations by cortical area. Brain-wide cortical depth quantification was performed based on cell detection by convolutional networks from GeneX-CreER driver lines crossed to Ai14 (R26-LSL-tdTomato), R26-LSL-h2b-GFP or Snap25-LSL-EGFP reporters and induced at the ages specified in Fig. 2, and P7 for Sema3E. The normalized cortical depth (0-1) was divided into 24 bins for the left histogram and 124 bins for the right plot in each panel. Abbreviations explained in the Supplementary Information. Scale bars: Last panel of first and third rows applies to all hemisections, 1mm; last panel of second and fourth rows applies to all cortical depth image panels, 200µm.

Extended Data Fig. 5 | Comparison of new with existing driver lines in terms of axon projections from SSp-bfd somatosensory cortex. a-c, STP images at
the SSp injection site (first row, arrow head) and at selected subcortical projection targets for eight driver lines characterized in this study (coloured gene names code for IT-red, PT-blue and CT-purple) compared to seven existing driver lines (black gene names), with EGFP or EYFP expression from Cre-activated viral vector (green) and background autofluorescence (red). Arrows point to axons. a, IT drivers project to cortical and striatal targets. PyNs Plxnd1 project bilaterally to cortex and striatum; PyNs Cux1 project bilaterally to cortex but not to striatum. b, PT drivers project to many corticofugal targets including brainstem and spinal cord. PyNs Fezf2 , PyNs Adcyap1 and PyNs Tcerg1l project to multiple ipsilateral targets and to the contralateral brainstem (arrows). c, CT drivers project predominantly to the thalamus. PyNs Tbr1 project bilaterally to cortex and to ipsilateral thalamus, PyNs Foxp2 and PyNs Tle4 project to the ipsilateral cortex and thalamus. Scale bars: first row in c (applies to first row), 1 mm; second to eighth rows in c (applies to each respective row), 200 µm; CST panel (bottom row) in c applies to entire row, 100 µm. Asterisks in b & c indicate presence of passing fibers. A side-by-side list of axon projection matrix for all these lines is presented in Fig. 3d. Fig. 6 | Molecular validation and developmental characterization of PyN driver lines. a, d, Low magnification images of sections of somatosensory cortex (SSp) at P7 stained with antibodies against CTIP2, CUX1, and tdTomato from PyN-CreER;Ai14 mice induced with tamoxifen. Inset to the right shows markers and Tomato + cell distributions across layers. Fezf2-, Tcerg1l-, Adcyap1-and Tle4-CreER;Ai14 were induced at E16.5 and collected at P7. Tbr1-, Cux1-and Plxnd1-CreER;Ai14 were induced at P4 and collected at P7. Tbr2-CreER;Ai14 was induced at E16.5 (not shown) or E17.5 and collected at P7. Lhx2-CreER;Ai14 was induced at P3 and collected at P7. a', d', Histograms showing radial distribution of Tomato + cells in the cortical plate, in the region corresponding to SSp. In brief, in CUX1-and CTIP2-stained sections, Tomato + cell depths relative to the thickness of the cortex were measured, as well as the limits of the areas occupied by CUX1 + or CTIP2 HIGH cells, shown in green (layers 2-4) and blue (layer 5b) bars, respectively (average relative values for the same sections, gray shading corresponds to 1 SD). For Tbr2-CreER;Ai14, mice induced at E16.5 or E17.5 were quantified separately, showing the later induction (darker red) results in more superficial labelling. Quantifications were made from 4-10 sections from 2-3 different mice for each line. b, e, Magnification of Tomato + cells in sections co-stained against CUX1 and CTIP2, LDB2 (enriched in PT), FOG2 (expressed in CT), BRN2, or SATB2 (expressed in IT). Arrowheads show double-positive cells; asterisks show Tomato + cells not expressing the marker. c, f, Percentage of Tomato + cells stained with each antibody. Quantifications were done in equivalent areas (320µm by 320µm) within the SSp centered in the specified layers. Each dot is an area from a different section, for which the percentage of double positive cells was calculated. Bars are mean+SD. Quantifications were made from 4-8 sections from 2-3 different mice for each line. Tbr2-CreER;Ai14 labelled CUX1 + , SATB2 + , BRN2 + IT in the most superficial layers 2-3, irrespective of their induction time. Lhx2-CreER;Ai14 and Cux1-CreER;Ai14 labelled CUX1 + , SATB2 + , BRN2 + IT deeper in layers 2-3. Plxnd1-CreER;Ai14 labelled SATB2 + , BRN2 + cells in layer 5A, as well as CUX1 + , SATB2 + , BRN2 + cells in layer 4. No cells were found in layer 5B. Tcerg1l-CreER;Ai14 and Adcyap1-CreER;Ai14 labelled sparse LDB2 + , CTIP2 + PT in layer 5. Fezf2-CreER;Ai14 extensively labelled PT in layer 5 that were LDB2 + , CTIP2 + , as well as some CT in layer 6 expressing CTIP2 and FOG2. Tle4-CreER;Ai14 and Tbr1-CreER;Ai14 labelled CT expressing FOG2 and CTIP2 (and lower levels of LDB2) in layer 6. Tle4-CreER;Ai14 also labelled some LDB2 + , CTIP2 + cells in layer 5 (PT), whereas Tbr1-CreER;Ai14 also labelled some CUX1 + , BRN2 + IT in layer 2/3. g, Fate-mapping of PyNs Tbr1 using Tbr1-CreER; Ai14 mice. Tamoxifen induction at E14.5 densely labelled L6 CT cells with minor labelling of cells in layers 3-5. h, Tamoxifen induction at E15.5 labelled L6 CT cells. i, Tamoxifen induction at P4 labelled L6 CT cells and also a subset of L2/3 cells. A subset of adult PyNs Tbr1 labelled from E14.5, E15.5 and P4 induction project to contralateral cortex via the corpus callosum (arrowheads, g', h', i'). Scale bars: g-i, low magnification, 500µm; high magnification, 100µm. Fig. 7 | Anterograde tracing, registration to CCFv3 and analysis of PyN projections from SSp-bfd. a, Summary table of driver lines and viral vectors used for anterograde tracing from PyNs in primary somatosensory cortex, related to Fig. 3. b, Virus injection centroid coordinates across single driver experiments in CCFv3 space on a dorsal whole-brain view. c-e, Whole-brain 3D renderings of axon projections registered to CCFv3 and main projection targets for each PyN subpopulation in the SSp-bfd. f, Axon projection matrix from SSp-bfd to 321 ipsilateral and 321contralateral targets (in columns), each grouped under 12 major categories (top row) for each of the driver lines generated in this study highlighted in Fig. 3, and presented alongside several previously published driver lines (IT lines: Cux2, Sepw1, Rasgrf2, Tlx3; PT lines: Rbp4, Sim1; CT lines: Ntsr1) for comparison (see Extended Data Fig. 5 for images). Colour shades in each row represent fraction of total axon signal measured from a single experiment per brain area; signal in the inj. site (white) was subtracted from total axon signal to show the fraction of projections outside the inj. site. PyNs Plxnd1 project bilaterally to CTX and Str; PyNs Cux1 project bilaterally to CTX but minimally to Str; PyNs Fezf2 , PyNs Adcyap1 and PyNs Tcerg1l project to multiple ipsilateral targets, and contralateral brainstem (arrows); PyNs Tbr1 project bilaterally to CTX and ipsilaterally to thalamus, PyNs Foxp2 and PyNs Tle4 project to the ipsilateral CTX and thalamus. Fezf2-Flp (green) and Tle4-CreER (red) with co-injection of Flp-and Cre-dependent AAVs expressing EGFP and mCherry, respectively (g-l). g, PyNs Fezf2 and PyNs Tle4 at the injection site occupying mainly L5B and L6, respectively. h, PyN Fezf2 and PyN Tle4 projection patterns converge in primary thalamus, VPM, whereas PyNs Fezf2 collaterals (asterisk) extend medially to higher order thalamic nuclei. i, PyNs Fezf2 (green) extend axon collaterals in Str, whereas PyNs Tle4 (red) pass through en route to thalamus. j-l, PyNs Fezf2 but not PyNs Tle4 project to multiple other corticofugal targets, including SC, Pn and cSp5. m, Schematic showing simultaneous anterograde tracing from PyNs targeted by Plxnd1-Flp (green) and Fezf2-CreER (red) with co-injection of Flpand Cre-dependent AAVs expressing EGFP and mCherry, respectively (n-r). n, PyNs Plxnd1 and PyNs Fezf2 at the injection site in motor cortex occupying mainly L5A and L5B/L6, respectively. o-p, PyNs Fezf2 but not PyNs Plxnd1 project to Thal (o) and medulla (p). q-r, PyNs Plxnd1 and PyNs Fezf2 project to ipsilateral Str with overlapping terminals (r), whereas PyNs Plxnd1 but not PyNs Fezf2 project to contralateral Str (q). Asterisks indicate PyN Fezf2 collaterals and arrows indicate PyN Plxnd1 collaterals. Scale bars: a, 2 mm; b, 1 mm; c-e, 200 µm; e', 100µm; l (applies to g-l), 200µm; n-r, 200µm; q' & r', 100µm.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Data Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A list of figures that have associated raw data -A description of any restrictions on data availability Raw and stitched whole-brain STP imaging data is available from the BICCN Brain Image Library (BIL) (http://www.brainimagelibrary.org/download.html) at the Pittsburgh Supercomputing Center, based on the accession codes detailed in Supplementary Tables 1 & 2. Anterograde projection datasets can be visualized on the Mouse Brain architecture website (http://brainarchitecture.org/cell-type/projection) as detailed in Supplementary Tables 1 & 2. All accession codes will be included  within Supplementary Tables 1 & 2 prior to publication. All other datasets generated during this study are available from the corresponding author upon request.

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative. Data exclusions All analyzed data was included in the study. Prior to STP analysis, datasets were screened for standard quality control by two independent reviewers according to pre-established criteria. Blinding Not relevant to this study because there was no group allocation.

Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.