Multiplexed CRISPRi reveals a transcriptional switch between KLF activators and repressors in the maturing neocortex Authors: Ryan W Kirk, Liwei Sun, Ruixuan Xiao, Erin A Clark, Sacha B Nelson, Joshua Corbin, Ryan Kirk, Liwei Sun, Ruixuan Xiao, Erin A Clark, Sacha B Nelson Published: 2026 ABSTRACT A critical phase of mammalian brain development takes place after birth. Neurons of the mouse neocortex undergo dramatic changes in their morphology, physiology, and synaptic connections during the first postnatal month, while properties of immature neurons, such as the capacity for robust axon outgrowth, are lost. The genetic and epigenetic programs controlling prenatal development are well studied, but our understanding of the transcriptional mechanisms that regulate postnatal neuronal maturation is comparatively lacking. By integrating chromatin accessibility and gene expression data from two subtypes of neocortical pyramidal neurons in the neonatal and maturing brain, we predicted a role for the Krüppel-like factor (KLF) family of transcription factors in the developmental regulation of neonatally expressed genes. Using a multiplexed CRISPR interference (CRISPRi) knockdown strategy, we found that a shift in expression from KLF activators (Klf6, Klf7) to repressors (Klf9, Klf13) during early postnatal development functions as a transcriptional ‘switch’ to first activate, then repress a set of shared targets with cytoskeletal functions including Tubb2b and Dpysl3. We demonstrate that this switch is buffered by redundancy between KLF paralogs, which our multiplexed CRISPRi strategy is equipped to overcome and study. Our results indicate that competition between activators and repressors within the KLF family regulates a conserved component of the postnatal maturation program that may underlie the loss of intrinsic axon growth in maturing neurons. This could facilitate the transition from axon growth to synaptic refinement required to stabilize mature circuits. Introduction From birth onwards, neurons in the cortex continue to mature and refine their prenatally acquired identity (Molyneaux et al., 2015; Prince et al., 2024; Wallace and Pollen, 2024; Yuan et al., 2022). This is especially pronounced in the first postnatal month of rodent brain development, which is characterized by sequential overlapping periods of axon growth, synaptogenesis, and synapse refinement (Faust et al., 2021; Price et al., 2006; Südhof, 2018; Turrigiano and Nelson, 2004). Many human neuropsychiatric and neurodevelopmental disorders result from disruptions in one or more of these processes, but how these processes are regulated at a molecular level remains incompletely understood (Gonçalves et al., 2013; Krishnan et al., 2015; Marín, 2016; Sekar et al., 2016). While the precise timing and nuances of the postnatal maturation program are tailored to the circuit-specific needs of cell types in different cortical layers, many physiological outcomes are remarkably conserved across cortical cell types and species (Gonzalez-Burgos et al., 2008; Kroon et al., 2019; Marchetto et al., 2019; Shi et al., 2012). These include regulated increases in the size of the soma (Benedetti et al., 2020), the number and strength of inhibitory and excitatory presynaptic connections (Desai et al., 2002; Etherington and Williams, 2011), expansion of the dendritic arbor (Koenderink and Uylings, 1995; Wu et al., 1999), and a decline in intrinsic excitability through rebalancing of membrane ion channels (McCormick and Prince, 1987; Okaty et al., 2009; Zhang, 2004). A number of stereotyped postnatal switches in the expression of individual ion channel and neurotransmitter receptor subunits have been characterized, such as an increase in the Ca2+-permeable AMPA receptor subunit GLUA2 (Sheng et al., 1994; Williams et al., 1993), an exchange of the prenatal GABRA2 receptor for GABRA1 subtype (Fritschy et al., 1994), and a transition from GLUN2B-containing NMDA receptors to the GLUN2A-containing variety found in the adult (Brill and Huguenard, 2008; Kumar et al., 2002). Such switches appear to be conserved properties of postnatal development, as they have been identified in multiple cortical regions, cell types, and species albeit with modestly differing degrees and kinetics (Ciceri and Studer, 2024; Law et al., 2003). It has been proposed that such switches reflect functional adaptations to changing requirements for sensitivity, plasticity, and stability at maturing cortical synapses, but in many cases, the precise function of the switch is still unclear (Cho et al., 2009; Philpot et al., 2007). In the neocortex of the mouse and other altricial mammals, prenatally initiated programs of stereotyped axon guidance and growth continue throughout the first 2 weeks of postnatal life, resulting in the formation of the corpus callosum and major subcortical tracts (De León Reyes et al., 2020; Gianino et al., 1999). Following this period of axon outgrowth, cortical neurons uniformly relinquish their capacity for robust axon growth to give way to a period of axon retraction and synaptic pruning (Riccomagno and Kolodkin, 2015). This transition is terminal, as mature neurons remain incapable of re-growing their axons following injury or insult (Bradke, 2022; Bregman et al., 1989; Ramon y Cajal, 1928). The loss of regenerative capacity has been attributed to a variety of cell-extrinsic and -intrinsic factors inherent to the CNS, as peripheral neurons do not exhibit such a dramatic loss of growth potential (Huebner and Strittmatter, 2009). Extrinsically, developmental increases in extracellular matrix proteins and components of the myelin sheath prevent axon growth through both physical and molecular mechanisms (Carulli et al., 2005; Yiu and He, 2006). However, mature neurons still struggle to regrow damaged axons even when these impediments are removed and young neurons can still extend axons through the inhibitory postnatal extracellular environment, exposing the influence of a powerful intrinsic brake on axon growth (Blackmore and Letourneau, 2006; Li et al., 1995a; Steinmetz et al., 2005). The biological nature of such a brake is still under investigation, but many therapeutic strategies for restoring axon growth rely on restoring specific neonatal attributes to adult neurons (Filbin, 2006; Hilton and Bradke, 2017; Mahar and Cavalli, 2018). Thus, neuronal maturation is characterized by both the acquisition of specialized functions and loss of shared properties of immature neurons that were critical for circuit development. Viewed differently, terminal differentiation of cortical neuron subtypes can also be conceived of as a generalized process characterized by the repression of neonatal identity. Studies of genetically or anatomically defined cell types over the first 1–2 months of postnatal development have characterized the genetic and electrophysiological maturation of cortical neurons and interneurons in exquisite detail (Ciceri et al., 2024; Kroon et al., 2019; Okaty et al., 2009; Patel et al., 2022; Yuan et al., 2022). While these genetic and physiological parameters change in parallel, surprisingly little is known about the transcription factors (TFs) responsible. Unraveling transcriptional circuits in higher eukaryotes is often confounded by the presence of gene paralogs able to compensate or compete with one another by binding the same DNA sequence (Badis et al., 2009; Jolma et al., 2013; Majidi et al., 2019; Weirauch et al., 2014). One such family is the diverse Krüppel-like factor (KLF) family of TFs, which contains 17 members that share a preference for a GC-rich DNA motif frequently located in promoter regions (Kaczynski et al., 2003; Moore et al., 2011). Members of this family are expressed in unique combinations in most major cell types of the body and are known to differentially compensate and compete with one another to affect gene expression in surprising, often non-linear, ways in non-neuronal tissues (Eaton et al., 2008; Jiang et al., 2008). Therefore, ascribing a unitary role to any individual member of this diverse family is complicated by intrafamilial interactions. Despite this complexity, several KLF paralogs have been shown to play roles in promoting or repressing CNS axon growth (Moore et al., 2009). Specifically, overexpression of the pro-growth TFs Klf6 or Klf7 has been shown to promote regeneration of Corticospinal Tract axons in vivo after spinal cord injury (Blackmore et al., 2012; Wang et al., 2018), while removal of growth-inhibiting Klf9 has been shown to enhance Retinal Ganglion Cell axon growth after optic nerve crush (Apara et al., 2017). The convergence of opposing members of the KLF family on a singular biological function has generated speculation that they might regulate the same set of genes in opposite directions, but experimental evidence for this hypothesis is scarce (Galvao et al., 2018). Such a detailed dissection of KLF family function in neurons would require methods robust to potential compensation by paralogues. Multiplexed gene loss-of-function approaches offer an experimental workaround to the problem of TF redundancy: by eliminating a TF and its known or hypothesized compensator, previously obscured phenotypes can be unmasked (Ahmed et al., 2024; Kim et al., 2021; Majidi et al., 2019). RNA-interference has been used as an alternative to intersectional transgenic strategies to demonstrate genetic redundancy and study genetic epistasis in vitro, yet it is prone to off-target effects and shows variable potency of knockdown in vivo (Chen et al., 2021; Horn et al., 2011; Sawyer et al., 2011). CRISPR-based loss of function is an emerging alternative to RNAi, but CRISPR-mediated double-strand break repair in post-mitotic cells leads to heterogeneous outcomes complicating interpretation (González et al., 2014; Mandegar et al., 2016). CRISPR Interference (CRISPRi) has recently gained traction as a versatile tool for repressing gene expression. This approach targets a nuclease-dead Cas9 fused to a KRAB effector domain (dCas9-KRAB) to genomic loci of interest using sequence-specific single guide RNAs (sgRNAs; Gilbert et al., 2013; Qi et al., 2013). Multiplexed gene knockdown using CRISPRi has been achieved by expressing multiple sgRNAs targeting different genes (Gilbert et al., 2014; McCarty et al., 2020; Zheng et al., 2018). CRISPRi has higher efficacy and specificity than RNAi, but in vivo applications have been limited by the large size of the dCas9-KRAB effector (Lau et al., 2019; Li et al., 2020; Zheng et al., 2018). The recent creation of a mouse containing a floxed dCas9-KRAB allele enables multiplexed and cell type-specific loss of function without the AAV payload restrictions surrounding delivery of the fusion enzyme (Gemberling et al., 2021). However, this system has not yet been utilized to study the regulatory logic of neuronal TFs in vivo. In this study, we used bioinformatics to identify putative TFs regulating postnatal development in excitatory neurons of the mouse cortex. By focusing on the program shared by upper and deep layer neurons to silence genes related to pre- and neonatal functions, we identified potential roles for developmentally regulated KLF paralogs in regulating this process. Using a novel in vivo multiplexed CRISPRi approach, we demonstrate that the postnatally upregulated TFs Klf9 and Klf13 are redundant repressors of a subset of developmentally downregulated genes. Conversely, we demonstrated that the neonatally-expressed transcriptional activators Klf6 and Klf7 are required for early postnatal expression of developmentally regulated genes with cytoskeletal functions. The high degree of overlap observed among bidirectionally regulated targets of the KLF repressor and activator pairs suggests that the KLF family functions as a ‘transcriptional switch’ to turn off neonatally expressed genes with known or suggested roles in axon growth, possibly through competition among family members for the same GC-rich motifs in shared targets’ promoters. Furthermore, the timing of this switch coincides with the transition from a period of robust axon growth to one of synaptogenesis in the mouse CNS (Kiyoshi and Tedeschi, 2020; Lewis et al., 2013), suggesting a possible transcriptional mechanism contributing to this developmental process. Moreover, this study provides proof of concept for the utility of multiplexed CRISPR interference for studying redundancy and competition within large TF families in the mammalian brain. Results Pyramidal cells in Layers 4 and 6 share a common gene regulatory program in early postnatal development To identify putative transcriptional regulators of early postnatal development, we used RNA sequencing (RNA-seq) and assay for transposase-accessible chromatin with sequencing (ATAC-seq) to profile changes in gene expression and chromatin accessibility between P2 and P30 in two distinct excitatory cortical cell types: RORβ+ Layer 4 neurons and Bmp3+ Layer 6 neurons (Figure 1A; Supplementary file 1). Layer 4 neurons were purified by fluorescence-activated cell sorting (FACS) from RorbGFP/+ knock in mice, which express GFP primarily in excitatory thalamocortical input neurons of Layer 4 (Liu et al., 2013; Clark et al., 2020). Layer 6 neurons expressing mCitrine were purified by manual sorting or FACS from the 56 L mouse line, which labels a unique population of excitatory corticothalamic neurons in lower Layer 6 that project to non-primary thalamic nuclei and subcortical targets (Shima et al., 2016). Layer 4 ATAC-seq data from P30 and RNA-seq data from P2 were previously published in Clark et al., 2020 and Layer 6 ATAC-seq and RNA-seq data from P30 were published in Sugino et al., 2019. Shared and cell-type specific gene expression and chromatin accessibility changes in maturing Layer VI and Layer VI neurons. (A) Experimental workflow: Rorb-GFP+/- (L4) or 56 L (L6) mice were sacrificed at P2 or P30 and GFP +neurons were isolated by FACS for RNA-seq or ATAC-seq (left). IGV tracks of gene expression (RNA-seq) and chromatin accessibility (ATAC-seq) of neonatally expressed Doublecortin (Dcx, top right) and adult-expressed CaMKIIa (bottom right). (B) RNA-Seq of FACS-sorted Rorb+ Layer VI neurons and Bmp3+ Layer VI neurons at P2 and P30 (n=4 mice per group). B1: Heatmap of all DEGs identified in either layer between P2 and P30 identifies k=6 unique clusters (DEGs defined as |log2FoldChange|>1; p.adj<0.01 by Wald’s test with BH correction; min(TPM)>10). Data plotted as group averages of z-scored relative log gene expression. B2 (right): Top Gene Ontology terms enriched in shared upregulated genes relative to all genes eligible for differential expression (Fisher exact tests with BH adjustment). B2 (left): transcription factor motif enrichment in ATAC-seq peaks overlapping the Promoter (TSS –1000/+200 bp) of shared upregulated genes relative to all Promoter peaks using the (Castro-Mondragon et al., 2022) CORE motif set (Fisher exact tests with BH adjustment). B3: Same as B2 but for shared downregulated genes. Figure 1—figure supplement 1.Shared and cell-type-specific gene expression and chromatin accessibility changes in maturing Layer VI and Layer VI neurons.(A) Principal component analysis of RNA-Seq (A1, n=4 per condition) and ATAC-seq (A2, n=2 per condition) samples used in this analysis (B) Distributions of ATAC-seq peak annotations for all peaks, all DARs, Layer 4 DARs, Layer 6 DARs, and DARs identified in both cell types. (C) Transcription factor motif enrichment in ATAC-seq peaks overlapping the Promoter (TSS –1000/+200 bp) of shared up- (left) or downregulated (right) genes relative to all Promoter peaks using the (Castro-Mondragon et al., 2022) cluster root motifs (Fisher exact tests with BH adjustment). Cluster 18 includes AP-1 family transcription factors. Cluster 27 contains all KLF/Sp transcription factors. Consensus motif and representative members are indicated on the right for each gene set. Shared and cell-type-specific gene expression and chromatin accessibility changes in maturing Layer VI and Layer VI neurons. (A) Principal component analysis of RNA-Seq (A1, n=4 per condition) and ATAC-seq (A2, n=2 per condition) samples used in this analysis (B) Distributions of ATAC-seq peak annotations for all peaks, all DARs, Layer 4 DARs, Layer 6 DARs, and DARs identified in both cell types. (C) Transcription factor motif enrichment in ATAC-seq peaks overlapping the Promoter (TSS –1000/+200 bp) of shared up- (left) or downregulated (right) genes relative to all Promoter peaks using the (Castro-Mondragon et al., 2022) cluster root motifs (Fisher exact tests with BH adjustment). Cluster 18 includes AP-1 family transcription factors. Cluster 27 contains all KLF/Sp transcription factors. Consensus motif and representative members are indicated on the right for each gene set. Figure 1—figure supplement 2.Shared and cell-type-specific gene expression and chromatin accessibility changes in maturing Layer VI and Layer VI neurons.(A) Heatmap of all DARs in gene Promoters identified in either layer between P2 and P30, separated into k=2 clusters. (DARs defined as |log2FoldChange|>1; p.adj<0.01 by Wald’s test with BH correction; min(TMM)>10). Data plotted as group averages of z-scored log2 (TMM). (B) Transcription factor motif enrichment in gained (B1) and lost (B2) ATAC-seq Promoter peaks (TSS –1000/+200 bp) relative to all Promoter peaks using the (Castro-Mondragon et al., 2022) cluster root motif set (Fisher exact tests with BH adjustment). (C) RiP-normalized coverage of average ATAC-seq signal at Promoter peaks +/- 1.5 kb of shared downregulated (top) or shared upregulated (bottom) DEGs in all P2 or P30 samples (Shaded region = 95% CI; n=4 libraries per group, aggregate of two cell types). Shared and cell-type-specific gene expression and chromatin accessibility changes in maturing Layer VI and Layer VI neurons. (A) Heatmap of all DARs in gene Promoters identified in either layer between P2 and P30, separated into k=2 clusters. (DARs defined as |log2FoldChange|>1; p.adj<0.01 by Wald’s test with BH correction; min(TMM)>10). Data plotted as group averages of z-scored log2 (TMM). (B) Transcription factor motif enrichment in gained (B1) and lost (B2) ATAC-seq Promoter peaks (TSS –1000/+200 bp) relative to all Promoter peaks using the (Castro-Mondragon et al., 2022) cluster root motif set (Fisher exact tests with BH adjustment). (C) RiP-normalized coverage of average ATAC-seq signal at Promoter peaks +/- 1.5 kb of shared downregulated (top) or shared upregulated (bottom) DEGs in all P2 or P30 samples (Shaded region = 95% CI; n=4 libraries per group, aggregate of two cell types). Although cell-type-specific gene regulatory programs are critical for establishing characteristic physiological and morphological properties of neurons in different layers (Lodato and Arlotta, 2015), we focused on the more general program of gene expression and chromatin accessibility changes conserved across the two distinct cell types. RNA-seq and ATAC-seq data demonstrated a substantial effect of postnatal age with a lesser contribution by genetic lineage that was more pronounced in mature neurons consistent with Yuan et al., 2022; Figure 1—figure supplement 1A. RNA-seq identified 5,579 differentially expressed protein-coding genes (DEGs; criteria: absolute fold-change ≥2, adjusted p-value <0.01, TPM ≥10) changing between P2 and P30 in at least one cell type. K-means clustering separated these DEGs into six groups comprised of a larger group of shared and smaller groups of developmentally up- or downregulated cell-type-specific genes (Figure 1B1). Of 5579 DEGs, 2684 (48.1%) were similarly regulated in Layers 4 and 6 (1380 up- and 1304 downregulated). Gene Ontology (GO) overrepresentation analysis for shared upregulated genes revealed that these genes are involved in processes such as vesicle-mediated transport at the synapse (q-value=3.32 × 10–6), regulation of neurotransmitter levels (4.18×10–6), and regulation of ion transport (3.32×10–6), consistent with the synaptic and electrophysiological changes occurring during the first month of life (Figure 1B2, right). In contrast, shared downregulated genes were enriched for GO terms axon development (q-value=6.25 × 10–5), regulation of neurogenesis (3.16×10–4), and regulation of projection organization (4.09×10–4), consistent with the loss of mitotic and axon growth capacity in the maturing cortex (Figure 1B3, right). We next examined the ATAC-seq data to identify putative cis-regulatory elements and associated transcription factor binding sites (TFBSs) facilitating the observed changes in gene expression (https://github.com/rkirk822/Kirk2026_elife copy archived at Kirk and Nelson, 2026). Of the 242,127 peaks of accessible chromatin identified across conditions, we found 38,321 differentially accessible regions (DARs; criteria: absolute fold-change ≥2, adjusted p-value <0.01, reads ≥10) that gain or lose accessibility between P2 and P30 in at least one cell type (Figure 1—figure supplement 1B; Supplementary file 1). Interestingly, of the 11,434 DARs (29.9%) with significant changes in both cell types, 11,419 (99.6%) changed in the same direction, indicating the presence of a shared regulatory program amidst cell type-specific differences. We found 2378 promoter DARs which formed two clusters representing sites that either gain (1220 Gained Shared DARs) or lose (1158 Lost Shared DARs) accessibility in both cell types (Figure 1—figure supplement 2A). The lack of cell-type-specific chromatin changes at promoters is consistent with the general understanding that promoter accessibility does not distinguish between closely related cell types as well as accessibility at distal genomic elements (Gray et al., 2017; Mo et al., 2015; Yoshida et al., 2019). Notably, this property made promoters an attractive genomic feature to focus on when looking for a shared developmental regulatory program. Therefore, we looked for enrichment of TFBSs in gained or lost promoter DARs to identify putative TFs regulating chromatin accessibility changes. This revealed that motifs recognized by the AP-1 family, bHLH, and MEF family of TFs were significantly overrepresented in shared gained promoter DARs (Figure 1—figure supplement 2B1). In contrast, we found a strong enrichment for the motif recognized by Homeodomain TFs and the POU family in shared lost promoter DARs (Figure 1—figure supplement 2B2), both of which contain members that play significant roles in specifying neuronal fate during prenatal development (Briscoe et al., 2000). An identical TFBS enrichment analysis of DARs assigned to intronic and intergenic elements was performed, yielding similar results apart from strong MEF enrichment (data not shown). To relate chromatin accessibility changes to observed changes in gene expression during early postnatal development, we examined the relationship between these features. Surprisingly, fewer than 10% of shared DEGs contained a DAR in their promoter, and we observed a weakly discernable change in the aggregate accessibility of promoter elements associated with shared developmental DEGs (Figure 1—figure supplement 2C). Promoter elements were also proportionally underrepresented among DARs, suggesting that transcriptional regulation from these sites could instead arise from coordinated transitions in the TFs bound to constitutively accessible promoter sites. To identify putative TFs regulating shared DEGs through chromatin accessibility-independent mechanisms, we performed TFBS enrichment analysis on all promoter ATAC-seq peaks associated with developmentally regulated genes. In promoters of shared developmentally upregulated genes, we found significant enrichment for motifs recognized by the AP-1 TFs FOS, C-JUN, and CREB (1.47- to 1.67-fold; p adj. between 2.15×10–3 and 2.40×10–5) and for the motif recognized by the TCF family (1.26-fold; p adj.=1.21 × 10–3; Figure 1B2, left). The presence of the AP-1 motif in the promoters of genes upregulated during the early postnatal period could reflect the arrival of structured input driving activity-dependent mechanisms associated with the transcriptional maturation of neurons (Stroud et al., 2020; West and Greenberg, 2011). The most significantly overrepresented motifs associated with downregulated genes were those recognized by members of the KLF/Sp TF family in the promoters of downregulated genes (p adj. between 6.25×10–3 and 7.88×10–13; enrichment 1.29–1.40-fold; Figure 1B3, left). The KLF/Sp family contains 18 KLFs and 4 Sp factors that share a highly conserved DNA binding domain and preference for the same GC-rich motif, so the observed enrichment likely represents the activity of a subset of KLF/Sp family members (Suske et al., 2005). After simplifying the motif set by using only the root motifs from JASPAR’s motif clustering, the root motif describing the cluster that includes all KLF and Sp TFs remained the most significantly and highly enriched motif (1.35-fold, p adj.=1.26 × 10–6; Figure 1—figure supplement 1C). While members of the KLF/Sp family have been implicated in a variety of neuronal functions, a role for this family in pyramidal cell maturation has not been described with the exception of Klf9, which has been shown to influence dentate granule cell maturation (Scobie et al., 2009). Of note, multiple members of the KLF family have established roles in regulating axon growth, which was one of the most enriched GO terms among shared downregulated DEGs (Moore et al., 2009). Altogether, this made the KLFs an attractive TF family for further investigation. Significant enrichment in these promoters was also observed for motifs recognized by members of the SOX family of TFs (1.33- to 1.79-fold; p adj. between 9.24×10–3 and 2.09×10–7). Sox1-3 are required for the maintenance of neural stem cell pluripotency, while Sox4 and Sox11 have established roles in establishing early neuronal identity, so the presence of their recognized motif in downregulated genes likely reflects their regulation of prenatally expressed genes (Bergsland et al., 2006; Bylund et al., 2003). An identical motif enrichment analysis focused on intronic or intragenic peaks with proximity to DEGs was uninformative, likely due to the difficulty of accurately assigning peaks distal to an annotated TSS to their target gene(s) (Fulco et al., 2019; data not shown). Together, these results highlight the existence of a developmental program shared by upper and deep layer neurons amidst their cell type-specific gene regulatory networks. The shared program may involve the upregulation of genes required for the acquisition of mature electrophysiological properties concomitant with downregulation of genes required for early developmental processes of cell migration and axon growth (Figure 1B1). The expression of KLF activators and repressors is oppositely regulated during early postnatal development The striking enrichment of the KLF/Sp motif in shared downregulated DEG promoters, along with the KLF family’s functional agreement with enriched GO terms and their understudied role in the developing cortex, made them appealing candidates for loss-of-function study. While motif enrichment analyses can predict TFBSs that are abundant in regions of interest undergoing some notable biological change, they rarely offer insight into the specific TFs underpinning that change since individual motifs are typically recognized by several TFs. This is particularly true for the KLF family, which consists of 18 members with conserved DNA binding domains resulting in a preference for the same motif (Suske et al., 2005). To parse this family, we used two common-sense criteria for identifying candidate TFs that impact developmental cortical gene expression: sufficient expression of the TFs in pyramidal cells and developmental regulation of the TFs themselves. We first looked at the expression levels of all KLF paralogs in six major excitatory cell types of the adult mouse cortex using additional RNA-seq data published in Sugino et al., 2019. This narrowed our pool of candidate TFs to six moderately- to highly-expressed KLFs: Klf6, Klf7, Klf9, Klf10, Klf12, and Klf13 (Figure 2—figure supplement 1A). Appropriate to our interest in studying transcriptional networks shared across pyramidal cells, the expression of all KLF paralogs did not differ substantially across excitatory neuron subtypes in different layers in our data or in the Allen Institute’s single-cell. RNA-seq data from mouse brain (Figure 2—figure supplement 1B, Yao et al., 2021). Next, we investigated how the expression of expressed KLFs changes during postnatal development. Previous studies have found evidence for downregulation of Klf6 and Klf7 and upregulation of KLF9 during a similar postnatal period in the brain (Laub et al., 2001; Moore et al., 2009; Scobie et al., 2009; Wang et al., 2018). Consistent with these observations, we detected a significant increase in expression of Klf9 and Klf13 in both Layer 4 (Klf9: 5.67-fold, p adj.=1.7 × 10–100; Klf13: 3.96-fold, 8.5×10–53) and Layer 6 (Klf9: 4.11-fold, 3.32×10–66; Klf13: 2.55-fold, 3.46×10–26) populations and a modest but significant decrease in Klf7 in both cell types (Layer 6: 3.13-fold, 1.29×10–40; Layer 6: 1.49-fold, 3.50×10–6) and Klf6 in Layer 4 (3.31-fold, 4.87×10–87) in our RNA-seq data (Figure 2—figure supplement 1C). To examine the expression of the highest expressed activating (Klf7) and repressive (Klf9) KLFs at finer temporal resolution in other genetically defined excitatory cortical cell types, we leveraged recent RNA-seq data from sorted Layer II/III and Layer 6 neurons across postnatal development published in Yuan et al., 2022. This clearly demonstrated a sharp and synchronous decline in Klf7 and steep increase in Klf9 expression between P7 and P21 among cells in Layers II/III and upper Layer 6, suggesting that the switch between KLF activators and repressors is a conserved feature across cortical cell types (Figure 2A). We also examined human RNA-seq data from the BrainSpan atlas (Kang et al., 2011) to see if similar trends in KLF expression occur in the human cortex and observed a similar decline in KLF6 and KLF7 concurrent with a rise in KLF9 and KLF13 expression between birth and the 4th month of postnatal life in four primary sensory cortical regions (Figure 2B). This suggests that the developmental regulation of this TF family is a conserved feature of mammalian cortical development. Activating (Klf6 and Klf7) and Repressive (Klf9 and Klf13) members of the KLF family display opposing patterns of gene expression throughout the developing cortex. (A) Developmental dynamics of Klf7 and Klf9 in Layer II/III and Layer VI neurons between E18 and P48. Plotted as mean FPKM from n=2 mice (data replotted from Yuan et al., 2022). (B) Expression of Klf6, Klf7, Klf9, and Klf13 mRNA across the human lifespan in four sensory cortical regions taken from the human BrainSpan Atlas (Kang et al., 2011). Data are plotted as smoothed loess fits +/- SE. (C) RNAScope of Klf7 and Klf9 mRNA P2, P7, and P30 C57 mouse brains demonstrates Klf7 mRNA is more abundant than Klf9 mRNA at P2 (C1) while Klf9 mRNA is more abundant than Klf7 mRNA at P30, where laminar distribution is more uniform (C3). (w.m.=white matter; Scale bar: 100 µm; Insert: 25 µm). Data quantified in C2 (n=3–4 mice per age, plotted as puncta normalized by nuclei density across the dorsal-ventral axis with 95% confidence intervals; **:p<0.001, Mann-Whitney U Test). Figure 2—source data 1.Developmental expression of KLF activators and repressors by fluorescent in situ hybridization.Klf9 and Klf7 puncta and nuclei counts in 2000 bins across the dorsal-ventral axis of wild-type P2, P7, and P30 cortices. Developmental expression of KLF activators and repressors by fluorescent in situ hybridization. Klf9 and Klf7 puncta and nuclei counts in 2000 bins across the dorsal-ventral axis of wild-type P2, P7, and P30 cortices. Figure 2—figure supplement 1.Activating (Klf6 and Klf7) and repressive (Klf9 and Klf13) members of the KLF family display opposing patterns of gene expression throughout the developing cortex.(A) Expression of all KLF family members in major excitatory neuronal cell types of the cortex obtained by bulk RNA-seq (left, Sugino et al., 2019) and in the principal cell clusters of the mouse cortex and hippocampus obtained by scRNA-seq (right, Yao et al., 2021) (B) Expression patterns of developmentally regulated KLFs between P2 and P30 in Layer VI and Layer VI neurons (RNA-seq, n=4 mice per group). Activating (Klf6 and Klf7) and repressive (Klf9 and Klf13) members of the KLF family display opposing patterns of gene expression throughout the developing cortex. (A) Expression of all KLF family members in major excitatory neuronal cell types of the cortex obtained by bulk RNA-seq (left, Sugino et al., 2019) and in the principal cell clusters of the mouse cortex and hippocampus obtained by scRNA-seq (right, Yao et al., 2021) (B) Expression patterns of developmentally regulated KLFs between P2 and P30 in Layer VI and Layer VI neurons (RNA-seq, n=4 mice per group). We then examined Klf7 and Klf9 expression across early postnatal development using fluorescent in situ hybridization (FISH) to simultaneously interrogate the temporal and spatial dynamics of the transcriptional switch. As expected, we detected a much higher expression of Klf7 than Klf9 mRNA at P2, particularly in Layer 4 (Figure 2—figure supplement 1C1). By P7, both TFs had similar expression across cortical layers, with neurons in Layer 4 maintaining a small but significantly higher level of Klf7, potentially due to their unique maturation process (Clark et al., 2020; Figure 2—figure supplement 1C2). At P30 Klf9 expression had clearly overtaken Klf7 expression by 1.65- to 4.39-fold in all cortical layers, and the expression of both KLFs examined appeared uniform across lamina (Figure 2—figure supplement 1C3). Together, these results confirm that members of the KLF family, particularly the activator Klf7 and the repressor Klf9, undergo a coordinated switch in their expression during early postnatal cortical development which coincides with a downregulation of genes enriched for the KLF/Sp motif in their promoters. Notably, the downregulated KLFs (Klf6 and Klf7) are closely structurally related and known to activate transcription of their targets while the upregulated KLFs (Klf9 and Klf13) share similar functional domains and have been primarily described as repressors of gene expression (McConnell and Yang, 2010). We theorized that these opposing expression patterns could enable the KLFs to function as a ‘transcriptional switch’ at a shared set of developmentally downregulated targets through the combined loss of activating drive and increase in repressive drive. Furthermore, this switch could operate without significant changes to the chromatin structure at promoters containing KLF/Sp motifs through a simple competition mechanism between repressors and activators at the same DNA-binding sites. Multiplexed knockdown of Klf9 and Klf13 in vivo using CRISPR interference To achieve cell-type-specific loss of function of KLF TFs, we crossed homozygous 129S6-Gt(ROSA)26Sortm2(CAG-dcas9/KRAB)Gers/J (also known as Rosa26-LSL-dCas9-KRAB) to Emx1cre/cre mice, which generated progeny expressing dCas9-KRAB in excitatory neurons of the forebrain. We refer throughout to mice carrying the knock-in dCas9-KRAB fusion allele as "CRISPRi" mice and further specify whether they are hetero- or homozygous for this allele to clarify breeding. Using this strategy, cell-type-specific knockdown of target transcripts can be achieved through delivery of sgRNA alone. We therefore cloned sgRNA(s) targeting our promoter(s) of interest individually or in tandem into custom AAV transfer vectors also containing a FLOXed fluorescent reporter (mCherry or EGFP) driven by the human SYN promoter to enable reporter expression exclusively in cre-positive, infected neurons. Transfer vectors were used to generate high-titer rAAV9. To ensure widespread infection, we bilaterally injected neonatal Emx1Cre/+; heterozygous CRISPRi pups intracerebroventricularly within 12 hr of birth (P0.5), then allowed 18–20 days for viral infection and maximal knockdown (Zheng et al., 2018) before their brains were harvested and processed for purification of reporter-positive neurons by FACS (Figure 3A). Cortical infection following neonatal injection was widespread in juvenile mice as previously described (see Methods), establishing the technique’s utility for manipulating large populations of neurons with medium throughput (Kim et al., 2013). A CRISPR interference strategy for studying cell-autonomous transcription factor knockdown in excitatory cortical neurons. (A) Schematic of experimental approach: Mice harboring the dCas9-KRAB transgene are crossed with Emx1-Cre mice and double heterozygous offspring are intracerebroventricularly injected at postnatal day 0.5 with AAV9 containing guide RNA's targeting gene(s) of interest and a floxxed fluorophore. After a 10–20 waiting period, infected neurons are sorted based on fluorescence and processed for RT-qPCR or RNA-seq. (Scale bars = 500 µm). (B) CRISPRi produces consistent and effective knockdown of Klf9 (99.0±1.7%) and Klf13 (98.6±1.7%) individually or in combination (Klf9: 99.4±0.6%; Klf13: 97.9±2.2%) as measured by RT-qPCR (n=6–8 mice/condition; data plotted as average fold change relative to scramble sgRNA control ± SD, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test, * p<0.05 **p<0.01, ***p<0.001) (C) Example RNA-scope image of P20 mouse cortex infected at P0.5 with AAV9-Klf9g1,g2-Klf13g1,g2-hSyn-DIO-EGFP and probed for EGFP (green) and Klf9 (magenta). (Scale bar = 50 µm). We began by testing this approach against Klf9, which is the most abundant KLF in the adult brain and sharply upregulated in excitatory neurons in early postnatal development. We reasoned that this TF was well positioned to effect gene expression in maturing cortex if knocked down prior to its’ rise in expression. We targeted Klf9 by injecting P0.5 pups with our rAAV9-Klf9sg1,sg2-SYN-DIO-mCherry virus (Klf9-mCherry) containing two sgRNAs targeting the TSS of Klf9 to ensure thorough knockdown (Qi et al., 2013). As a control, we injected littermates with rAAV9-Scrsg1-SYN-DIO-mCherry (Scr-mCherry) containing a non-targeting scrambled sgRNA. After 18–20 days, we observed a 99.0 ± 1.7% (p=0.002) knockdown of Klf9 by qPCR in neurons from mice injected with the Klf9-mCherry virus relative to mice that received the Scr-mCherry virus (Figure 3B). While Klf9 is the most abundant KLF in the adult mouse brain by far, studies in other tissues suggest that its loss can be partially compensated for by Klf13 (Ávila-Mendoza et al., 2020a; Heard et al., 2012; Bernhardt et al., 2022). Because Klf13 exhibits the same developmental upregulation as Klf9 and they are among the most structurally similar KLFs (McConnell and Yang, 2010), we reasoned that a knockdown of both TFs could be more effective at uncovering a transcriptional phenotype. Therefore, we used identical methods to knock down Klf13 or Klf9 and Klf13 simultaneously using AAV9-Klf13sg1,sg2-SYN-DIO-mCherry (Klf13-mCherry) or AAV9-Klf9sg1,sg2-Klf13sg1,sg2-SYN-DIO-mCherry (Klf9/13-mCherry) respectively. We detected a 98.6 ± 1.7% (p=0.0009) knockdown of Klf13 in mice injected with Klf13-mCherry and a 99.4 ± 0.6% (p=0.002) knockdown of Klf9 and a 97.9 ± 2.2% (p=0.001) knockdown of Klf13 in mice injected with Klf9/13-mCherry relative to scramble controls, demonstrating the utility of CRISPRi for in vivo multiplexed knockdown experiments (Figure 3B). Furthermore, we observed a 1.5±0.4 fold (p=0.025) increase in Klf13 expression in mice receiving the Klf9-mCherry virus and a similar 2.1±1.2 fold (p=0.01) increase in Klf9 expression in Klf13-mCherry samples, suggesting that compensation following loss of a single factor could involve a release of mutual repression. Finally, to confirm that our knockdown was restricted to infected cells, we used RNA-Scope to visualize Klf9 expression in the brains of P20 mice neonatally injected with an EGFP-expressing rAAV9-Klf9sg1,sg2-Klf13sg1,sg2-SYN-DIO-EGFP virus containing identical sgRNAs as our Klf9/13-mCherry construct. We detected a near-total absence of Klf9 puncta in EGFP-positive cells interspersed with EGFP-negative nuclei surrounded by dense Klf9 signal (Figure 3C, further quantified in Figure 8). This indicates that our manipulation is highly penetrant and restricted to our cells of interest. Klf9 and Klf13 are partially compensatory transcriptional repressors To identify putative targets of Klf9 and Klf13 and examine the transcriptional consequences of knocking down the major KLF repressors in the juvenile mouse brain, we prepared RNA-seq libraries from sorted neurons receiving sgRNAs targeting Klf9, Klf13, Klf9, and Klf13 or scrambled controls. Principal component analysis of the resultant data showed a clean separation of all treatment groups along PC1 that mirrored the fraction of endogenous Klf9 +Klf13 mRNA remaining in these groups relative to the scramble control (Klf9/13>Klf9>>Klf13; Figure 4A; Supplementary file 2). Differential gene expression analysis revealed 5 DEGs following Klf13 KD, 28 DEGs following Klf9 KD, and 212 DEGs in the Klf9/13 KD samples relative to scramble controls (DEGs: protein-coding genes with absolute fold-change ≥2, adjusted p value <0.05; Figure 4—figure supplement 1A–B). The gradient observed along PC1 and the discrepancy in DEGs identified among single- versus double-knockdown samples indicates that compensation for loss of Klf9 by Klf13 likely occurs in vivo. Volcano plots comparing Klf9/13 KD samples to scramble controls further underscored the efficacy and specificity of CRISPRi knockdown, as Klf9 and Klf13 stood out as the most significantly downregulated DEGs (Figure 4B). Furthermore, the greater fraction of upregulated DEGs relative to all DEGs following Klf9/13 KD (85.4%, 181 genes) is consistent with Klf9 and Klf13’s known roles as repressors of gene expression, as upregulated DEGs could plausibly result from a loss of repression. Klf9 and Klf13 show compensatory activity and additive regulation of target genes. (A) Principal component for top 500 most variable genes in all knockdown samples and scramble controls shows graded effect of KLF knockdown along PC1 (n=4 mice per condition). (B) De-repression of target genes following Klf9/13 KD. Volcano plot of gene expression changes in Klf9/13 KD relative to scramble control mice (DEGs defined as |log2FoldChange|>1; p.adj<0.05 by Wald’s test with BH correction). (C1) Heatmap of all DEGs identified in Klf9/13 KD (212 genes) across all libraries. Data plotted as z-scored, regularized log-transformed counts. Right bar represents KLF/Sp root cluster motif counts identified in ATAC-seq peaks overlapping respective Promoters (TSS –1000/+200 bp) by FIMO (qval <0.05, quantified in C3, Fisher’s exact test). C2: Line plots of all upregulated Klf9/13 targets across all libraries. Grey lines are z-scored, regularized log-transformed RNA-seq counts of individual genes, blue lines represent average of all genes +/- SEM. (***:p<0.0001, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test) (D) Additive regulation of selected targets with roles in cytoskeletal or synaptic function (n=4 mice per group, average TPM ± SD). (E) Example IGV traces of RNA-seq coverage at the Dpysl3 locus. Arrows indicate KLF/Sp motif matches (FIMO, q<0.05). Figure 4—figure supplement 1.Klf9 and Klf13 show compensatory activity and additive regulation of target genes.(A) Volcano plots of gene expression changes in Klf9 KD (A1) or Klf13 KD (A2) relative to scramble controls (DEGs defined as |log2FoldChange|>1; p.adj<0.05 by Wald’s test with BH correction) (B) Venn diagrams comparing All DEGs (B1), Upregulated DEGs (B2), and Downregulated DEGs (B3) identified in Klf9, Klf13, and Klf9/13 KD samples relative to scramble controls (C) Line plot of all downregulated Klf9/13 targets across all libraries. Grey lines are z-scored, regularized log-transformed RNA-seq counts of individual genes, red lines represent average of all genes ± SEM. (***:p<0.0001, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test) (D) transcription factor motif enrichment in ATAC-seq peaks overlapping the Promoter (TSS –1000/+200 bp) of Klf9/13 targets (upregulated DEGs) relative to Promoter peaks in non-DEGs using the (Castro-Mondragon et al., 2022) CORE motif set (Fisher exact tests with BH adjustment). Klf9 and Klf13 show compensatory activity and additive regulation of target genes. (A) Volcano plots of gene expression changes in Klf9 KD (A1) or Klf13 KD (A2) relative to scramble controls (DEGs defined as |log2FoldChange|>1; p.adj<0.05 by Wald’s test with BH correction) (B) Venn diagrams comparing All DEGs (B1), Upregulated DEGs (B2), and Downregulated DEGs (B3) identified in Klf9, Klf13, and Klf9/13 KD samples relative to scramble controls (C) Line plot of all downregulated Klf9/13 targets across all libraries. Grey lines are z-scored, regularized log-transformed RNA-seq counts of individual genes, red lines represent average of all genes ± SEM. (***:p<0.0001, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test) (D) transcription factor motif enrichment in ATAC-seq peaks overlapping the Promoter (TSS –1000/+200 bp) of Klf9/13 targets (upregulated DEGs) relative to Promoter peaks in non-DEGs using the (Castro-Mondragon et al., 2022) CORE motif set (Fisher exact tests with BH adjustment). Figure 4—figure supplement 2.Normal intrinsic and synaptic properties of Klf9/13 KD neurons.(A) Example biocytin fill of Layer 5 pyramidal neuron (green) infected with Klf9/13-mCherry KD virus (red) at P0.5 (Scale bar = 200 µm) (B) Intrinsic excitability is unchanged between Klf9/13 KD neurons (N=15) and scrambled controls (N=10). (p=0.787; Virus::Current effect from two-way ANOVA for current and virus) (C) Miniature excitatory postsynaptic current (mEPSC) properties of Klf9/13 KD neurons (red, N=12) and scrambled controls (blue, N=15) are not different (top) (amplitude: Klf9/13-mCherry: 12.9±0.614 pA, Scr-mCherry: 12.7±0.47 pA, p=0.90, Mann-Whitney U test; frequency: Klf9/13-mCherry: 5.85±0.76 Hz; Scr-mCherry: 5.4±42 Hz p=0.83, Mann-Whitney U test). Figure 4—figure supplement 2—source data 1.Electrophysiology of Klf9/13 knockdown neurons.Intrinsic excitability and miniature EPSC recordings of Klf9/13 KD and scrambled control neurons at P18-20. Normal intrinsic and synaptic properties of Klf9/13 KD neurons. (A) Example biocytin fill of Layer 5 pyramidal neuron (green) infected with Klf9/13-mCherry KD virus (red) at P0.5 (Scale bar = 200 µm) (B) Intrinsic excitability is unchanged between Klf9/13 KD neurons (N=15) and scrambled controls (N=10). (p=0.787; Virus::Current effect from two-way ANOVA for current and virus) (C) Miniature excitatory postsynaptic current (mEPSC) properties of Klf9/13 KD neurons (red, N=12) and scrambled controls (blue, N=15) are not different (top) (amplitude: Klf9/13-mCherry: 12.9±0.614 pA, Scr-mCherry: 12.7±0.47 pA, p=0.90, Mann-Whitney U test; frequency: Klf9/13-mCherry: 5.85±0.76 Hz; Scr-mCherry: 5.4±42 Hz p=0.83, Mann-Whitney U test). Figure 4—figure supplement 2—source data 1.Electrophysiology of Klf9/13 knockdown neurons.Intrinsic excitability and miniature EPSC recordings of Klf9/13 KD and scrambled control neurons at P18-20. Electrophysiology of Klf9/13 knockdown neurons. Intrinsic excitability and miniature EPSC recordings of Klf9/13 KD and scrambled control neurons at P18-20. K-means clustering of Klf9/13 DEGs across all treatment groups identified two clusters representing down- and upregulated gene sets. When visualized as a heatmap, the upregulated gene cluster revealed a gradient in gene expression from scramble control samples to Klf9/13 KD samples akin to the gradient observed in PC1 (Figure 4A), suggesting that there were subthreshold effects of Klf9 and Klf13 KD on Klf9/13 DEGs (Figure 4C1). The increased effect size produced by depleting increasing amounts of KLF repressive drive was particularly evident when we plotted the z-scored expression of the 181 upregulated DEGs across all samples (Figure 4C2). Viewed this way, even Klf13 KD produces a small but highly significant (p=5.44 × 10–26) effect on Klf9/13 target genes. A similar trend was observed among the 31 downregulated DEGs (Figure 4—figure supplement 1C). Together, these results support a model in which Klf9 and Klf13 additively cooperate to repress shared targets and can partially compensate for one another in a dose-dependent manner. Although GO analysis did not return any enriched terms among up- or downregulated gene sets (data not shown), manual inspection of the genes with the greatest ‘KLF Effect’ (absolute loadings along PC1) led to the identification of putative targets with synaptic and cytoskeletal functions as well as ion channels (Figure 4C1, D and E). The abundance of cytoskeletal regulators among targets of Klf9/13 coincides with the established roles of these factors as inhibitors of neurite outgrowth. We did not detect a significant change in Klf6 or Klf7 expression in the double knockdown, making it unlikely that the rise in Klf9 and Klf13 expression drives the decline in Klf6 and Klf7 expression. The dysregulation of genes related to cytoskeletal regulation and synaptic function prompted us to look for a synaptic phenotype in Klf9/13 KD cells. However, whole cell patch clamp recordings from infected neurons in Klf9/13 KD mice and scramble controls showed no significant difference in miniature excitatory postsynaptic current (mEPSC) amplitude (Klf9/13-mCherry: 12.9±0.614 pA; Scr-mCherry: 12.7±0.47 pA) or frequency (Klf9/13-mCherry: 5.85±0.76 Hz; Scr-mCherry: 5.44±0.20 Hz) between groups (Figure 4—figure supplement 2A and C). Intrinsic excitability of Klf9/13 KD neurons was also unchanged (Figure 4—figure supplement 2B). Because the KLF family was selected based on enrichment of the KLF/Sp motif in developmental DEG promoters, we reasoned that we should be able to recover this motif enrichment using ATAC-seq peaks overlapping the promoters of DEGs identified from Klf9/13 KD experiments. Indeed, we found a significant (1.98-fold, p=2.18 × 10–10) enrichment of the root motif of the cluster containing all KLFs in the JASPAR 2022 CORE collection in promoters of upregulated DEGs but not among downregulated DEGs (p=0.46, Figure 4C3). Because this approach was biased towards the expected motif, we also performed an unbiased motif enrichment analysis on upregulated DEG promoters using the entire Castro-Mondragon et al., 2022 CORE collection and found a nearly exclusive enrichment of KLF/Sp family motifs ranging from to 1.47- to 2.76-fold (p adj. <<0.01) (Figure 4—figure supplement 1D). These data support the idea that the upregulated genes are likely direct targets of Klf9 and Klf13 and validate CRISPRi as a novel tool to identify TF targets in vivo. Targets of Klf9 and Klf13 are developmentally downregulated genes Developmental upregulation of Klf9 and Klf13 coincides with downregulation of genes containing an enrichment of the KLF/Sp motif in their promoters. We therefore reasoned that putative targets of Klf9 and Klf13 should themselves be developmentally downregulated. To investigate this, we examined RNA-seq data from sorted Rorb +Layer 4 neurons published in Clark et al., 2020, which afforded us an additional timepoint between P2 and P30. We observed a clear and significant decrease in the normalized expression of our 181 putative targets of Klf9/13 between P2 and P30, with intermediate expression at P7 (Figure 5A). Furthermore, the rise of Klf9 and Klf13 during this period is perfectly antagonistic to the decrease in their targets. Plotting genome-wide effects of Klf9/13 KD against expression changes between P2 and P30 from our original dataset revealed a significant negative correlation (R=–0.56, p<2.2 × 10–16; Figure 5B). Putative Klf9/13 targets are developmentally downregulated in the maturing cortex. (A) Expression of upregulated DEGs in Klf9/13 KD at early, mid, and late postnatal developmental time points. Grey lines are z-scored, regularized log-transformed RNA-seq counts of individual genes from sorted Rorb+ (Layer IV) excitatory neurons, blue lines represent the average of all genes ± SEM (n=4 mice per group, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test, ***p<0.001). (B) Gene expression changes in Klf9/13 KD relative to scramble controls at P18-20 (n=4 mice per group) relative to gene expression changes between P2 and P30 (n=8 mice per group, aggregate of two cell types). Points represent individual genes, and red points are DEGs identified in Klf9/13 KD samples. The line is a linear fit ± 95% CI to Klf9/13 DEGs (red). (C) RiP-normalized coverage of average ATAC-seq signal at Promoter peaks around DEGs identified from Klf9/13 KD RNA-seq (shaded region = 95% CI; n=4 libraries per group, aggregate of two cell types). (D) Example IGV traces from P2 and P30 ATAC-seq and RNA-seq for Klf9/13 targets Atat1 (top, Developmental DEG without Promoter DAR) and Tubb2b (bottom, Developmental DEG with Promoter DAR). Arrows indicate putative KLF/Sp binding site identified by FIMO (q<0.05). Figure 5—figure supplement 1.Putative Klf9/13 targets are developmentally downregulated in the maturing cortex.(A) Gene expression changes in Klf9/13 KD relative to scramble controls at P18-20 (n=4 mice per group) relative to chromatin accessibility changes between P2 and P30 (n=4 mice per group, aggregate of two cell types). Only DEGs or genes with at least one differentially accessible promoter peak are plotted. Points represent individual genes, and colors indicate the direction of change in Klf9/13 KD samples. Genes with more than one Promoter peak may appear multiple times. (B) Example IGV traces from P2 and P30 ATAC-seq and RNA-seq for Klf9/13 targets Plppr1 (B1, Developmental DEG without Promoter DAR) and Dpysl3 (B2, Developmental DEG with Promoter DAR). Arrows indicate putative KLF/Sp binding site identified by FIMO (q<0.05). Putative Klf9/13 targets are developmentally downregulated in the maturing cortex. (A) Gene expression changes in Klf9/13 KD relative to scramble controls at P18-20 (n=4 mice per group) relative to chromatin accessibility changes between P2 and P30 (n=4 mice per group, aggregate of two cell types). Only DEGs or genes with at least one differentially accessible promoter peak are plotted. Points represent individual genes, and colors indicate the direction of change in Klf9/13 KD samples. Genes with more than one Promoter peak may appear multiple times. (B) Example IGV traces from P2 and P30 ATAC-seq and RNA-seq for Klf9/13 targets Plppr1 (B1, Developmental DEG without Promoter DAR) and Dpysl3 (B2, Developmental DEG with Promoter DAR). Arrows indicate putative KLF/Sp binding site identified by FIMO (q<0.05). Transcriptional repression by Klf9 and Klf13 could occur via direct inhibition of RNA Polymerase II’s binding or activity, or through dynamic changes to the chromatin accessibility at target gene promoters, or through both mechanisms. In support of the latter model, in vitro studies on Klf13’s mechanism of action have suggested that it may promote histone modifications leading to heterochromatin formation through its’ Sin3a domain, which it shares with Klf9 (Kaczynski et al., 2001). To differentiate between these possibilities, we looked for evidence of changes in chromatin accessibility between P2 and P30 at promoter elements of Klf9/13 targets. Although only a minority (21.6%) of peaks in target promoters met the threshold for DARs, coverage plots of ATAC-seq signal centered on promoter peaks revealed a noticeable loss in accessibility between P2 and P30 in genes upregulated by Klf9/13 KD (Figure 5C2). Furthermore, examination of the relationship between changes in gene expression following Klf9/13 KD and chromatin accessibility between P2 and P30 revealed a weak negative correlation (R=–0.2) between these variables bolstered by large changes in chromatin accessibility at a minority of highly upregulated Klf9/13 targets (Figure 5—figure supplement 1A). Inspection of ATAC-seq signal at promoters of selected Klf9/13 targets revealed a notable loss of accessibility at promoter peaks containing the KLF/Sp motif upstream of Tubb2b and Dpysl3, but no change in accessibility in motif-positive peaks upstream of Plppr1 and Atat1 (Figure 5—figure supplement 1B). These data suggest that a Klf9/13-mediated loss of promoter accessibility could lead to the repression of a subset of targets during early postnatal development, but the mechanism of repression may differ in a target- or site-specific manner. Klf6 and Klf7 promote expression of developmentally regulated genes in the perinatal neocortex The transcriptional activators Klf6 and Klf7 display an inverse pattern in their mRNA and protein expression relative to Klf9 and Klf13, with high prenatal expression that declines gradually after the first week of life. This suggests that Klf6 and Klf7 could function in opposition to Klf9 and Klf13 by promoting the transcription of at least some of the same neonatally expressed targets, possibly by binding to the same KLF/Sp motifs in their promoters. To test this possibility, we utilized a similar CRISPRi loss of function approach by creating an AAV construct designed for multiplex knockdown of Klf6 and Klf7 (Klf6/7-GFP; AAV9-Klf6sg1,sg2-Klf7sg1,sg2-SYN-DIO-EGFP). We elected to pursue a multiplex KD strategy from the outset following our observation that Klf9/13 knockdowns have reduced penetrance due to redundancy. Klf6 and Klf7 have highly homologous protein sequences, can compensate for one another in zebrafish, and have non-additive effects on neurite branching in vitro (Moore et al., 2009; Veldman et al., 2007). As before, we bilaterally injected P0.5 offspring of homozygous CRISPRi and Emx1Cre/Cre mice with targeting rAAV9 or a scramble control (Scr-GFP) before sorting and collecting mRNA from infected neurons. Because expression of KLF activators is highest in early development and their targets might already be downregulated by KLF repressors by P18-20 (P20), we also sorted neurons from P10-12 mice (P10). This was the minimum latency of infection we deemed appropriate given the time required for mRNA and protein turnover following repression by dCas9-KRAB. We found a near-complete knockdown of both Klf6 and Klf7 at P10 (Klf6: 99.6 ± 0.5%, p=0.0005; Klf7: 97.6 ± 3.1%, p=0.0006) and at P20 (Klf6: 99.8 ± 0.1%, p=0.0005; Klf7: 94.7 ± 12.6%, p=0.004) relative to scramble controls, further demonstrating the general applicability of CRISPRi for manipulation of a range of targets (Figure 6A). In addition, we were able to detect a 2.76-fold (p=0.0007) reduction in Klf6 and 2.89-fold (p=0.0009) reduction in Klf7 expression between P10 and P20, indicating that the KLFs are still undergoing their switch during this period. Klf6 and Klf7 promote expression of developmentally regulated genes in the perinatal cortex. (A) RT-qPCR quantification of Klf6 and Klf7 in mice injected with Klf6 and Klf7-targeting sgRNAs at P0, collected at early (P10-12) or late (P18-20) postnatal timepoints (n=7–9; data plotted as average fold-change relative to P10 scramble control ± SD, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test with BH adjustment, **p<0.005, ***p<0.001). (B) Volcano plot of DEGs identified in Klf6/7 KD samples relative to scramble controls at P10 (B1) and P20 (B2, DEGs defined as |log2FoldChange|>1; p.adj<0.05 by Wald’s test with BH correction). (C1) Heatmap of all DEGs identified in Klf6/7 KD at P10 (374 genes, k=3). Data plotted as z-scored, regularized log-transformed counts. Right bar represents KLF motif counts identified in ATAC-seq peaks overlapping respective Promoters (TSS –1000/+200 bp) by FIMO (qval <0.05, quantified in C3, Fisher’s exact test). Gene Ontology (GO) overrepresentation analysis of downregulated targets shown in C2. Gene names in C1 colored by GO term membership. (D) Developmental regulation of Klf6/7 targets D1: Expression of downregulated DEGs in Klf6/7 KD at early, mid, and late postnatal developmental time points. Grey lines are z-scored, regularized log-transformed RNA-seq counts of individual genes from sorted Rorb+ (Layer IV) excitatory neurons, blue lines represent average of all genes ± SEM (n=4 mice per group, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test, ***p<0.001). D2: Gene expression changes in Klf6/7 KD relative to scramble controls at P10-12 (n=4 mice per group) relative to gene expression changes between P2 and P30 (n=8 mice per group, aggregate of two cell types). Points represent individual genes, and green points are DEGs identified in Klf6/7 KD samples. Lines are linear fits ± 95% CI to all points (black) or Klf6/7 DEGs (green) with correlation indicated in the respective colors. Figure 6—figure supplement 1.Klf6 and Klf7 promote expression of developmentally regulated genes in the perinatal cortex.(A) Transcription factor motif enrichment in ATAC-seq peaks overlapping the Promoter (TSS –1000/+200 bp) of Klf6/7 P10 targets (downregulated DEGs in clusters 1 and 2) relative to all Promoter peaks in non-DEGs using the Castro-Mondragon et al., 2022 CORE motif set (Fisher exact tests with BH adjustment). (B) RiP-normalized coverage of average ATAC-seq signal at Promoter peaks around DEGs identified from Klf6/7 P10 KD RNA-seq (shaded region = 95% CI; n=4 libraries per group, aggregate of two cell types). (C) Gene expression changes in Klf6/7 KD relative to scramble controls at P10-12 (n=4 mice per group) relative to chromatin accessibility changes between P2 and P30 (n=4 mice per group, aggregate of two cell types). Only DEGs or genes with at least one differentially accessible promoter peak are plotted. Points represent individual genes, and colors indicate the direction of change in Klf6/7 KD samples. Genes with more than one Promoter peak may appear multiple times. Klf6 and Klf7 promote expression of developmentally regulated genes in the perinatal cortex. (A) Transcription factor motif enrichment in ATAC-seq peaks overlapping the Promoter (TSS –1000/+200 bp) of Klf6/7 P10 targets (downregulated DEGs in clusters 1 and 2) relative to all Promoter peaks in non-DEGs using the Castro-Mondragon et al., 2022 CORE motif set (Fisher exact tests with BH adjustment). (B) RiP-normalized coverage of average ATAC-seq signal at Promoter peaks around DEGs identified from Klf6/7 P10 KD RNA-seq (shaded region = 95% CI; n=4 libraries per group, aggregate of two cell types). (C) Gene expression changes in Klf6/7 KD relative to scramble controls at P10-12 (n=4 mice per group) relative to chromatin accessibility changes between P2 and P30 (n=4 mice per group, aggregate of two cell types). Only DEGs or genes with at least one differentially accessible promoter peak are plotted. Points represent individual genes, and colors indicate the direction of change in Klf6/7 KD samples. Genes with more than one Promoter peak may appear multiple times. We next conducted RNA-sequencing of P10 and P20 excitatory neurons infected with Klf6/7-GFP and Scr-GFP AAV with the goals of identifying likely targets of the KLF activators and examining how their regulatory role changes across early postnatal development. Using identical criteria as above, we identified 374 DEGs in P10 Klf6/7 KD neurons and 115 DEGs at P20 relative to age-matched scramble controls, with both sgRNA targets once again being the most significantly affected DEGs (Figure 6B; Supplementary file 2). In accordance with Klf6 and Klf7’s reported functions as transcriptional activators, the majority of DEGs were downregulated at both ages (80.5% at P10, 85.2% at P20). Furthermore, the enhanced effect of Klf6/7 KD at P10 is consistent with an elevated KLF activating drive at this age resulting from higher expression of Klf6 and Klf7. The elevated Klf9/13 repressive drive at P20 could have repressed Klf6/7 targets that would otherwise have exhibited loss of activation in the KDs as at P10. Therefore, downstream analysis focused on the 374 DEGs identified at P10 when the role of Klf6 and Klf7 is most pronounced. The inclusion of an additional developmental time point allowed us to ask how Klf6 and Klf7’s regulatory activity and the expression of their putative targets may differ between ages. To identify DEG sets with different trajectories across groups, we performed k-means clustering of our 374 DEGs and visualized the results as a heatmap. We detected three clusters: downregulated DEGs that are also developmentally downregulated between P10 and P20, downregulated DEGs that do not change between P10 and P20, and genes upregulated by Klf6/7 KD (Figure 6C1). Notably, the largest cluster of 190 DEGs downregulated by age and Klf6/7 KD displayed a gradient from high expression at P10 in scramble control neurons to low expression at P20 in Klf6/7 KD neurons, approximating the ratio of KLF activator to repressor abundance across groups (Figure 6C1). The upregulated gene cluster exhibited the opposite trend, and P20 targets were largely consistent with those identified at P10. GO analysis of all downregulated DEGs identified enrichment for functions related to neurite growth and cell motility such as actin filament organization (q-value=0.040), membrane rafts (0.004), and focal adhesion (0.035), consistent with previously described roles for Klf6 and Klf7 in vivo (Figure 6C1 and C2; Blackmore et al., 2012; Hong et al., 2023; Kajimura et al., 2007; Wang et al., 2018). Genes contributing to these terms include the actin nucleator Cobl, the actin-binding proteins Ablim3, Cd44, Dpysl3, and Iqgap1, the actin capping protein Eps8, the growth-cone enriched microtubule destabilizing factor Stmn2 (STG10) tubulin isoforms Tubb3, Tubb5, Tuba1a, and Tubb2b, and the myosin motor Myo1c. Most notably, putative Klf6/7 targets were enriched for genes associated with the distal axon (q-value=0.020), which includes previously identified targets of Klf7 such as Rac3, Stmn2, Gap43, Cdkn1a (P21), and L1cam. This indicates that CRISPRi can reproduce molecular phenotypes of null mutants in vivo (Hong et al., 2023; Kajimura et al., 2007). If DEGs identified by RNA-seq are enriched for direct targets of Klf6/7, we would expect to see an enrichment of the KLF/Sp motif in ATAC-seq peaks overlapping the promoters of DEGs. Indeed, we found the KLF cluster root motif was 1.69-fold enriched in downregulated gene promoters (p=1.02 × 10–9) but not in upregulated gene promoters (p=0.77), consistent with logic that direct targets of transcriptional activators should be downregulated following their KD (Figure 6C3). Enrichment analysis for the full Castro-Mondragon et al., 2022 CORE motif set also demonstrated a nearly exclusive enrichment for the KLF/Sp motif in downregulated DEG promoters ranging from 1.35- to 2.25-fold (p adj.<<0.01, Figure 6—figure supplement 1A). Klf6 and Klf7 expression declines during early postnatal development, so we wondered if putative targets of these TFs change in parallel as was the case for targets of Klf9/13. We examined changes in gene expression at P10 caused by Klf6/7 KD against developmental gene expression changes between P2 and P30. This revealed a strong positive correlation (R=0.54, p<2.2 × 10–16) among putative targets of Klf6/7, suggesting that Klf6 and Klf7 promote the expression of a subset of neonatally expressed genes that are downregulated by a coordinated developmental loss of KLF activators and phenocopied by their acute KD (Figure 6D2). Normalized trend line plots of downregulated DEGs at P2, P7, and P30 also demonstrated significant and progressive decline of putative target expression that parallels reduced levels of Klf6 and Klf7 (Figure 6D1). Since targets of Klf6/7 are developmentally downregulated, we wondered if they also exhibit changes in promoter accessibility between P2 and P30 that coincide with this change. As with targets of Klf9/13, we found a modest change in overall accessibility at promoter elements in Klf6/7 targets, although only a small percentage (14.8%) of targets contained a promoter DAR (Figure 6—figure supplement 1B and C). Globally, the developmental change in promoter accessibility was weakly but positively correlated with the effect of Klf6/7 KD on gene expression (R=0.17), and this effect was most prominent among the most downregulated genes (Figure 6—figure supplement 1C). Together, these data suggest that Klf6 and Klf7 promote neonatal expression of cytoskeletal genes that are repressed in maturing neurons through both chromatin accessibility-dependent and -independent mechanisms. KLF activators and repressors bidirectionally regulate a set of shared targets If the KLF activators and repressors truly act as a transcriptional switch at a subset of targets, they should share the following features: they should be upregulated following Klf9/13 knockdown, downregulated following Klf6/7 knockdown, and developmentally downregulated in accordance with the shift in from KLF activator to repressor drive across early postnatal development. As a group, these genes would be expected to be enriched in genes that regulate axonal development or cytoskeletal function and should contain an overrepresentation of the KLF/Sp motif in their promoter. Inspection of all RNA-seq libraries by principal component analysis revealed a clean separation of all groups sequenced along the first two principal components (PCs) corresponding to age and sgRNA targets respectively. Intriguingly, Klf6/7 KD samples and Klf9/13 KD samples were found at opposite extremes of PC2 (‘KLF Effect’) with scramble controls in the middle, indicating that genes contributing to this PC were bidirectionally affected by KLF KD (Figure 7A). Furthermore, we observed a noticeable shift along PC1 (‘Age Effect’) depending on the sgRNA targets, with Klf9/13 KD libraries shifted left towards the P10 samples, and Klf6/7 KD samples shifted right past P20 libraries (Figure 7A). This aligns with our hypothesis that the KLF family regulates a developmental gene expression program through an activator-repressor balance: depleting the late-expressed repressors Klf9 and Klf13 returns neurons to a younger transcriptional state while knocking down early-expressed activators Klf6 and Klf7 transcriptionally ‘ages’ these same cells. The transition from KLF activators to repressors constitutes a developmental switch at shared targets. (A) Principal component analysis of all RNA-seq libraries shows that age and KLF valence account for >75% of variance (n=4 mice each). (B) GSEA of genes ranked by the bidirectional effect of KLF KD (defined as absolute loadings of PC2 from A). (C) Heatmap of all DEGs shared between Klf9/13 KD and Klf6/7 KD (|log2FoldChange|>1; p.adj<0.05) shows bidirectional regulation by the KLF family and developmental regulation of all shared targets. Data plotted as z-scored, regularized log-transformed counts. (D) Gene expression changes in Klf6/7 KD relative to scramble controls at P10 are significantly negatively correlated with gene expression changes in Klf9/13 KD relative to scramble controls at P20, especially at shared DEGs (p.adj<0.05 by Wald’s test with BH correction).(E) Bidirectional regulation of novel (Dpysl3, Tubb2b), and established (Gap43, Stmn2) targets Klf7. (Data plotted as TPM ± SD, n=4 mice) (F) Representative IGV tracks of RNA-seq data at targets of Klf7 Rac3 (top) and L1cam (bottom) across treatment groups. Arrows indicate the presence of a KLF/Sp motif identified by FIMO (q<0.05). Figure 7—figure supplement 1.Shared KLF targets are similarly expressed and developmentally regulated in Layer 4 and Layer 6 neurons.(A) Heatmap of scored, log-normalized gene expression of 144 shared KLF targets in Layer IV and Layer VI neurons at P2 and P30 (n=4 mice per group). Candidate genes with significant axonal or neuronal roles are highlighted on the right. Shared KLF targets are similarly expressed and developmentally regulated in Layer 4 and Layer 6 neurons. (A) Heatmap of scored, log-normalized gene expression of 144 shared KLF targets in Layer IV and Layer VI neurons at P2 and P30 (n=4 mice per group). Candidate genes with significant axonal or neuronal roles are highlighted on the right. Figure 7—figure supplement 2.The transition from KLF activators to repressors constitutes a developmental switch at shared targets.(A) Shared targets are highly enriched for the KLF/Sp motif. Transcription factor motif enrichment in ATAC-seq peaks overlapping the promoter (TSS –1000/+200 bp) of shared targets for the root KLF/Sp motif (top, Fisher exact test) and the full (Castro-Mondragon et al., 2022) CORE motif set (bottom, Fisher exact tests with BH adjustment). Enrichment is measured relative to promoter peaks in all non-DEGs. (B) Klf6/7 targets are upregulated by Klf9/13 KD (left), and Klf9/13 targets are downregulated by Klf6/7 KD (right). Grey lines are z-scored, averaged regularized log-transformed RNA-seq counts of individual genes, and the dark line represents the average of all genes ± SEM (n=4 mice per group, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test with BH adjustment, *: p<0.01; **p<1e-6; ***: p<2e-16). (C) Loss of chromatin accessibility around at shared targets. RiP-normalized coverage of average ATAC-seq signal at promoter peaks around shared targets and unaffected control genes (left; shaded region = 95% CI; n=4 libraries per group, aggregate of two cell types). The transition from KLF activators to repressors constitutes a developmental switch at shared targets. (A) Shared targets are highly enriched for the KLF/Sp motif. Transcription factor motif enrichment in ATAC-seq peaks overlapping the promoter (TSS –1000/+200 bp) of shared targets for the root KLF/Sp motif (top, Fisher exact test) and the full (Castro-Mondragon et al., 2022) CORE motif set (bottom, Fisher exact tests with BH adjustment). Enrichment is measured relative to promoter peaks in all non-DEGs. (B) Klf6/7 targets are upregulated by Klf9/13 KD (left), and Klf9/13 targets are downregulated by Klf6/7 KD (right). Grey lines are z-scored, averaged regularized log-transformed RNA-seq counts of individual genes, and the dark line represents the average of all genes ± SEM (n=4 mice per group, Kruskal-Wallis test with post-hoc pairwise Mann-Whitney U test with BH adjustment, *: p<0.01; **p<1e-6; ***: p<2e-16). (C) Loss of chromatin accessibility around at shared targets. RiP-normalized coverage of average ATAC-seq signal at promoter peaks around shared targets and unaffected control genes (left; shaded region = 95% CI; n=4 libraries per group, aggregate of two cell types). Figure 7—figure supplement 3.The transition from KLF activators to repressors constitutes a developmental switch at shared targets.(A) Gene expression changes in Klf9/13 KD relative to scramble controls at P18-20 plotted against gene expression changes between P2 and P30 (as in Figure 5B). Red points are DEGs identified in Klf9/13 KD samples. Green points are DEGs uniquely identified in Klf6/7 KD samples at P10. Lines are linear fit ± 95% CI to Klf9/13 (red) or Klf6/7 (green) DEGs. (B) As in A, but changes in Klf6/7 KD relative to scramble controls at P10-12 against gene expression changes between P2 and P30. The transition from KLF activators to repressors constitutes a developmental switch at shared targets. (A) Gene expression changes in Klf9/13 KD relative to scramble controls at P18-20 plotted against gene expression changes between P2 and P30 (as in Figure 5B). Red points are DEGs identified in Klf9/13 KD samples. Green points are DEGs uniquely identified in Klf6/7 KD samples at P10. Lines are linear fit ± 95% CI to Klf9/13 (red) or Klf6/7 (green) DEGs. (B) As in A, but changes in Klf6/7 KD relative to scramble controls at P10-12 against gene expression changes between P2 and P30. To identify shared targets of our KLF activator and repressor pairs, we integrated differential gene expression analysis results from Klf9/13 KD neurons at P20 and Klf6/7 KD neurons at P10 compared to their respective age-matched scramble controls. Separate ages were selected to represent the developmental periods when KLF activator drive (P10) or repressive drive (P20) is higher. We found that 41.7% (156/374 DEGs) of putative Klf6/7 P10 targets are significantly affected by Klf9/13 KD (p adj.<0.05), with the vast majority (96.1%, 150/156) being oppositely regulated by KLF pairs. Similarly, we found that 47.4% of Klf9/13 P20 targets exhibit significant changes following Klf6/7 KD, 97.4% of which were bidirectional effects. We refined this to a set of 144 high-confidence shared targets affected by manipulation of both KLF paralog pairs out of 553 total DEGs between them (26.0%; Shared Targets: protein-coding genes with an absolute fold change ≥2 in at least one comparison and ≥1.5 in the other, adj. p-value <0.05 in both). K-means clustering (k=3) demonstrated that 141/144 (97.9%) of genes meeting these criteria were bidirectionally regulated by the KLFs and the vast majority (134, 93.1%) were upregulated by Klf9/13 KD and downregulated by Klf6/7 KD (Figure 7C). In addition, most shared targets were downregulated 1.5-fold between P2 and P30 in both Layers 4 and 6 (72.2 and 69.4%, respectively), indicating that KLF targets are components of the shared developmental gene regulatory program described above (Figure 7—figure supplement 1, Figure 7—figure supplement 3). Notably, knockdown of Klf9/13 returned expression of shared targets to levels more similar to their expression at P10 in scramble controls, while Klf6/7 KD accelerated the developmental decline in these genes to levels closer to those measured at P20 (Figure 7—figure supplement 2B). The fact that KD of Klf9/13 did not fully restore expression of many genes to P10 levels by P20 is consistent with push-pull regulation by developmentally increasing levels of repressors and decreasing levels of activators, although other repressing factors were not ruled out. In further support of the view that these KLF paralog pairs make up a bidirectional transcriptional switch, we detected a significant negative correlation (R=–0.53, p=8.4 × 10–12) between effects of Klf6/7 KD and Klf9/13 KD at shared targets (Figure 7D). Furthermore, the scarcity of targets affected in the same direction by both KD conditions suggests that there are few transcriptional effects attributable to multiplex knockdown alone. As before, we looked for enrichment of the core KLF/Sp motif in accessible promoter regions of shared targets. Reassuringly, we found a significant (2.23-fold, p=1.12 × 10–10) enrichment of the cognate motif, higher than detected for either Klf9/13 or Klf6/7 KD alone. This enrichment was only observed for genes downregulated by Klf6/7 KD and upregulated by Klf9/13 KD. Furthermore, motifs recognized by KLF/Sp family TFs were the only overrepresented motifs identified in promoters of shared targets in when this analysis was conducted using the full (Castro-Mondragon et al., 2022) core vertebrate motif collection (2.01- to 3.55-fold enrichment, p adj.<2.74 × 10–6; Figure 7—figure supplement 2A). The magnitude of this effect may indicate that putative shared targets are more likely to represent genuine direct targets of the KLF family than DEGs identified by either knockdown individually. Next, we looked at changes in the accessibility of promoter peaks upstream of shared targets using our developmental ATAC-seq data. We detected a notable loss of ATAC-seq signal at shared KLF targets that was less apparent in non-DEG control genes with similar developmental expression (Figure 7—figure supplement 2C). Although GO analysis of shared targets failed to identify any enriched terms, we conducted the threshold-free Gene Set Enrichment Analysis (GSEA) using the absolute loadings of genes along the PCA axis that separated Klf9/13 from Klf6/7 KD libraries (PC2). We reasoned that genes contributing to this axis exhibit the strongest bidirectional regulation by the KLFs. GSEA identified enrichment of terms identified among Klf6/7 targets and among shared downregulated genes identified between P2 and P30, including actin filament organization (q-value=0.017) and axon development (q-value=0.037; Figure 7B). Shared targets with known cytoskeletal and axonal roles were among those with the largest bidirectional effect sizes. Two of the most potently regulated genes in either condition were the actin and microtubule-polymerizing factor Dpysl3 and the β-tubulin subunit Tubb2b with respective roles in autism spectrum disorder (ASD) and polymicrogyria (Stottmann et al., 2013; Tsutiya et al., 2017; Figure 7E). Other notable shared targets included the actin regulators Ppp1r18, Prmt2, Rac3, and Ablim3, the tubulin polymerizing protein Tppp3, the PTEN inhibitor Plppr3, axon guidance mediator Plxna3 (PlexinA3), inhibitor of microtubule polymerization Dpysl5, and the microtubule acetylase Atat1. We also identified Gap43, a common growth cone marker, strong promoter of axon regeneration and known target of Klf7, as a shared target (Bomze et al., 2001; Kajimura et al., 2007; Skene et al., 1986; Figure 7E). We also identified significant but subthreshold upregulation (p adj. 0.05; -log2(1.5)95% knockdown of our genes of interest as early as 10 days post-injection, representing an improvement over many reports of knockdown efficiency in vivo or in vitro. This could be a product of the abundance of endogenous dCas9-KRAB in LSL-dCas9-KRAB mice and the high viral penetrance produced by ICV injections, which overcomes any limitations on CRISPRi imposed by insufficient enzyme or sgRNA concentration (Fontana et al., 2018; Li et al., 2016). By using CRISPRi to knock down KLF repressor and activator pairs, we identified specific genes through which the KLF family could mediate its bidirectional effect on developmental gene expression and axon growth (Moore et al., 2009). While this provides insight into the genetic mechanisms underlying the loss of intrinsic axon growth capacity in maturing neurons, it is possible that postnatal repression of this program by KLF repressors is essential for circuit stability and precision in the mature brain. Methods Key resources tableReagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional informationGene (Mus musculus) Klf9 ENSMUSG00000033863Gene (M. musculus) Klf13 ENSMUSG00000052040Gene (M. musculus) Klf7 ENSMUSG00000025959Gene (M. musculus) Klf6 ENSMUSG00000000078Strain, strain background (M. musculus)C57BL/6Jackson LaboratoriesRRID:IMSR_JAX:000664Genetic reagent (M. musculus)Rosa26-LSL-dCas9-KRABJackson LaboratoriesRRID:IMSR_JAX:033066Genetic reagent (M. musculus) Emx1cre/cre Jackson LaboratoriesRRID:IMSR_JAX:005628Genetic reagent (M. musculus)RorbGFP (Rorb1g)PMID:23652001Dr. Douglas Forrest (Laboratory of Endocrinology and Receptor Biology, National Institutes of Health)Genetic reagent (M. musculus)56 LPMID:30977723Cell line (Homo sapiens)HEK293TATCCRRID:CVCL_0063Human embryonic kidney cell lineRecombinant DNA reagentlentiCRISPR v2(plasmid)AddgeneRRID:Addgene_52961Starting construct for gRNA cloningRecombinant DNA reagentpAAV-SYN-DIO-mCherry(plasmid)AddgeneRRID:Addgene_50459Starting construct for multiplexed gRNA AAV vectorRecombinant DNA reagentpAAV-SYN-DIO-EGFP(plasmid)AddgeneRRID:Addgene_50457Starting construct for multiplexed gRNA AAV vectorRecombinant DNA reagentpAdDeltaF6AddgeneRRID:Addgene_112867Used for AAV productionRecombinant DNA reagentpAAV2/9 nAddgeneRRID:Addgene_112865Used for AAV productionChemical compoundStreptavidin, Alexa Fluor 594 ConjugateThermo Fisher ScientificS32356;RRID:AB_2313574From Streptomyces avidinii, 1:500Chemical compoundOpal 620Akoya BiosciencesFP1495001KT1:750 –1:1500Chemical compoundOpal 690Akoya BiosciencesFP1497001KT1:750 –1:1500Commercial assay or kitSMARTer-Seq Stranded Total RNA v3Takara634485Commercial assay or kitSMART-Seq mRNA LPTakara634762Commercial assay or kitRNAscope Fluorescent Multiplex kit v2Advanced Cell Diagnostics323100Sequence-based reagentRNAScope Klf9 ProbeAdvanced Cell Diagnostics488371;488371-C2Sequence-based reagentRNAScope Klf7 ProbeAdvanced Cell Diagnostics570961-C2Sequence-based reagentRNAscope Dpysl3 ProbeAdvanced Cell Diagnostics496611Sequence-based reagentRNAScope Tubb2b ProbeAdvanced Cell Diagnostics1039621-C1OtherrAAV9-Klf9sg1,sg2-SYN-DIO-mCherryThis studyKlf9 KD AAVNelson Lab, Brandeis University, see MethodsOtherrAAV9-Klf13sg1,sg2-SYN-DIO-mCherryThis studyKlf13 KD AAV Nelson Lab, Brandeis University, see MethodsOtherrAAV9-Klf9sg1,sg2- Klf13sg1,sg2-SYN-DIO-mCherryThis studyKlf9/13 KD AAV (mCherry) Nelson Lab, Brandeis University, see MethodsOtherrAAV9-Scrsg1- SYN -DIO-mCherryThis studyScramble control AAV (mCherry) Nelson Lab, Brandeis University, see MethodsOtherrAAV9-Klf9sg1,sg2- Klf13sg1,sg2- SYN -DIO-EGFPThis studyKlf9/13 KD AAV (EGFP) Nelson Lab, Brandeis University, see MethodsOtherrAAV9-Klf6sg1,sg2- Klf7sg1,sg2- SYN -DIO-EGFPThis studyKlf6/7 KD AAV Nelson Lab, Brandeis University, see MethodsOtherrAAV9-Scrsg1- SYN -DIO-EGFPThis studyScramble control AAV (EGFP) Nelson Lab, Brandeis University, see Methods Animal husbandry and transgenic lines All animals were bred, housed, and cared for in Foster Biomedical Research Laboratory at Brandeis University (Waltham, MA, USA). Animals were provided with food and water ad libitum and kept on a 12 hr:12 hr light:dark cycle. Cages were enriched with huts and tubes. All experiments were approved by the Institutional Animal Care and Use Committee of Brandeis University, Waltham, MA, USA (Protocol 22015 A). Both sexes were used for all experiments. RorbGFP (Rorb1g) mice were obtained from Dr. Douglas Forrest (Liu et al., 2013) and maintained as heterozygotes, which labels Layer 4 RORβ+neurons with GFP without disrupting Layer 4 neuron identity or function (Clark et al., 2020). 56 L mice, which label cortico-thalamic Layer 6 neurons with mCitrine, were created through lentiviral enhancer trapping and characterized in Shima et al., 2016. Emx1cre/+ mice (Gorski et al., 2002) were obtained from Jackson Labs (JAX #005628). Rosa26:LSL-dCas9KRAB (CRISPRi) mice (Gemberling et al., 2021) were obtained from Jackson Labs (JAX #033066). To produce cell-type-specific dCas9-KRAB expression in the Emx1-positive lineage, homozygous CRISPRi mice were bred with Emx1cre/cre mice and F1 offspring of both sexes were used for all experiments. Developmental RNA-seq library preparation For RNA-seq of Layer 4 neurons and P2 Layer 6 neurons, libraries were prepared using Ovation Trio RNA-Seq library preparation kit with mouse rRNA depletion (0507–32) according to manufacturer’s specifications and sequenced on the Illumina NextSeq 550/500 platform as published in Clark et al., 2020. For P30 Layer 6 libraries, RNA was reverse transcribed using the NuGEN Ovation RNA-Seq System V2 (#7102; NuGEN) followed by fragmentation to an average of ~200 bp and ligation of Illumina sequencing adaptors with the Encore Rapid Kit (0314; NuGEN). These libraries were sequenced on the Illumina HiSeq 2500 platform as published in Sugino et al., 2019. ATAC-seq library preparation ATAC-seq libraries were prepared as described in Clark et al., 2019. Briefly, tissue from two male and two female mice of the same age and strain were sorted together to collect 30,000–50,000 cells per library. Nuclei were tagmented for 30 min and libraries were prepared in accordance with Corces et al., 2016 followed by paired end (PE) sequencing on the NextSeq Illumina platform. Developmental RNA-seq analysis The adaptor sequence AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC was trimmed from libraries using cutadapt v. 3.4 and libraries were mapped to the UCSC mm10 genome using STAR v. 2.7.10 (Dobin et al., 2013) using the ENCODE Long Read RNA-seq parameters. Reads mapping to exons of known genes were quantified using featureCounts in the Rsubread package (Liao et al., 2019). Differential expression analysis was conducted using DESeq2 (Love et al., 2014) with model design ~Group, where group is the Age:Layer identity of each sample. Genes with significant changes in expression between P2 and P30 were identified independently for either cortical layer using cutoffs |log2FC|>1, adjusted p-value <0.01, and TPM >10 in at least one sample, and combined as a master list of 5579 unique developmental DEGs for downstream analysis. Adjusted p values for each gene were reported from a Wald test with Benjamini-Hochberg (BH) correction for multiple comparisons, and log2FC values were shrunk using the ashr method. To identify shared and cell-type-specific gene sets and plot heatmaps, a scaled counts matrix was generated by z-scoring the condition-averaged relative log (rlog) expression. The Gap Statistic method from R package cluster was used to identify k=6 as the optimal number of clusters, and genes were clustered using k-means clustering with 100 starts. Assignment of clusters as ‘Shared’ or ‘Layer x Specific’ was manually determined by inspecting plots of normalized gene expression within clusters as in Figure 1 and Figure 1—figure supplement 2. For comparison of Layer 4 and Layer 6 RNA-seq with other cortical cell types from Sugino et al., 2019, batch effects corresponding to sequencing location (Brandeis v. Janelia) were removed using limma’s removeBatchEffect() function prior to log2CPM conversion in edgeR. Bigwig files for IGV visualization were generated from BAM files of two merged replicates using deepTools v. 3.5.4 bamCoverage normalized to counts per million (CPM) mapped reads while excluding reads mapped to the mitochondrial genome. ATAC-seq data analysis ATAC-seq libraries were processed using an abbreviated version of the ENCODE ATAC-seq pipeline. First, Illumina adaptors were trimmed using cutadapt v 3.4 and trimmed reads were mapped to the mm10 genome using bowtie2 v. 2.4.2 (Langmead and Salzberg, 2012). Reads mapping to the mitochondrial genome, unannotated chromosomes, or ENCODE mm10 blacklist regions, or reads where MAPQ <30, were removed using samtools v. 1.11 (Danecek et al., 2021), followed by removal of PCR duplicates with Picard tools v. 2.23.5 (Broad Institute). Reads were shifted and converted to PE BED files in samtools and MACS2 v. 2.2.7 was used to call peaks in each library (Zhang et al., 2008). Only peaks identified in both replicates were retained in the final peak set. Peaks called in all four groups were merged into a non-redundant set of 242,127 peaks using the method described in Corces et al., 2018. DiffBind was used for differential accessibility analysis. First, reads overlapping the non-redundant peak set were counted and filtered to remove peaks containing less than 10 counts in at least 1 library and peaks overlapping mm10 v2 blacklist regions (Amemiya et al., 2019), resulting in 99,995 peaks eligible for differential accessibility. Resulting peaks were then normalized using the TMM method (Reske et al., 2020; Robinson et al., 2010). Peaks with differential accessibility between P2 and P30 were identified independently for either cortical layer using cutoffs |log2FC|>1 and FDR <0.01, and combined as a master list of 38,321 unique developmental DARs for downstream analysis. Peaks were assigned to genes and regulatory elements using the ChIPSeeker package in R using mm10 Ensembl annotations with the TSS region defined as –1000 bp upstream of the TSS to 200 bp downstream. Gene assignments for intronic/exonic peaks were adjusted to reflect the gene that they were within (as opposed to their nearest TSS). For subsequent analysis, the full list of all DARs was subdivided into Promoter Peaks, Intronic Peaks, and Intragenic Peaks according to their ChIPSeeker annotation. To identify shared and cell-type-specific peaks and plot heatmaps, a scaled counts matrix was generated by z-scoring the condition averages of Reads in Peaks-normalized counts. Optimal k was determined using the R package cluster as described above for RNA-seq analysis, resulting in k=2 Promoter Peak sets, k=5 Intronic Peak sets, and k=6 Intragenic Peak sets. Assignment of clusters as ‘Shared’ or ‘Layer x Specific’ was manually determined. ATAC-seq coverage plots were created from RiP-normalized peaks using the dba.plotProfile function in DiffBind. Summit-centered regions of interest were taken by extending peaks in DAR clusters or peaks assigned to DEG sets out 1.5 kb in either direction. For TSS-centered regions, the Ensembl TSS with the highest cumulative coverage across libraries was selected as the ‘primary’ TSS and extended 1.5 kb in either direction. Normalized coverage at each position was averaged across genes within a sample, then across age groups for plotting chromatin changes between P2 and P30. Bigwig files for IGV visualization were generated from BAM files of merged replicates using deepTools v. 3.5.4 bamCoverage normalized to reads per genomic content (RPGC) while excluding reads mapped to the mitochondrial genome or ENCODE blacklist regions. Motif overrepresentation analysis Motif matches for the 836 motifs in the non-redundant Castro-Mondragon et al., 2022 Vertebrate CORE collection (Castro-Mondragon et al., 2022) were identified in the merged, non-redundant peak set using FIMO from the MEME suite v. 5.5.0 (Bailey et al., 2015) with an adjusted q value cutoff of 0.05 using the nucleotide frequency in the full peak set as background. Motif matches passing threshold were counted in the original peak set using a custom R script and used for overrepresentation analysis. Peaks were identified as belonging to a test set based on cluster membership or ChIPSeeker gene assignment for DAR and DEG analysis respectively, and all remaining peaks were identified as the control set. Motif counts were binarized and a 2×2 contingency table was constructed for each motif. The effect size and significance of motif overrepresentation was calculated using Fisher’s exact tests followed by BH adjustment for multiple testing. As above, all overrepresentation analysis was conducted separately for peaks assigned to different regulatory features (promoter, intron, or intragenic). Motif matching and enrichment analysis using the 137 root motifs from the Castro-Mondragon et al., 2022 Vertebrate CORE clusters (Castro-Mondragon et al., 2017) was conducted using identical methods. Gene ontology analysis GO analysis was conducted using the clusterProfiler package in R (Yu et al., 2012). Overrepresentation of terms of interest in DEG clusters or genes assigned to DAR clusters was measured relative to all genes eligible for inclusion in the test group using the enrichGO function with an adjusted p value cutoff of 0.05. For bar plots, redundant GO terms were simplified using a cutoff of 0.5. Gene set enrichment analysis To identify terms enriched among genes bidirectionally regulated by the KLF family by GSEA, principal component analysis was performed on the rlog counts matrix of all KLF KD, which separated samples by Age (PC1) and ‘KLF Effect’ (PC2, as in Figure 7A). The absolute loadings from the second principal component (PC2) representing the magnitude of the ‘KLF Effect’ were used as input to clusterProfiler’s gseGO function with 10,000 permutations. For plotting, redundant GO terms were simplified using a cutoff of 0.5. sgRNA design and cloning Single-guide RNAs (sgRNAs) were cloned into custom AAV transfer vectors using a custom cloning strategy. Briefly, sgRNAs were selected from among the top 5 scoring sgRNAs for the selected target in the mouse CRISPRi v2 library (Horlbeck et al., 2016). Guides were aligned to the target’s TSS region and inspected for overlap with other guides and to peaks identified in the developmental ATAC-seq dataset. Preference was given to sgRNAs with ≥ 50 bp of separation from another sgRNAs and located within an identified region of open chromatin. In the case where other guides overlapped the top scoring guide, an additional sgRNA was predicted using the CRISPick tool (Sanson et al., 2018). All guide RNA sequences are listed in Table 2. Table 2.sgRNA sequences.Guide RNA sequences used in the CRISPRi knock-down viruses.TargetsgRNA sequence (5’ ->3’) Klf9 sg1: GGCGGGAAGAAGTGTCTCGTsg2: GGAGCTGGAGAATGCCGGAG Klf13 sg1: GCGCCTGCCACAATGCACGGsg2: GGTCGAGCGGAGGCGACGTT Klf6 sg1: GCTGCGGGAGGGTCCGGACCsg2: GGGCAGCGGAGCTTTGAATA Klf7 sg1: GACAGCGCTGCAGATATAGAsg2: GGAGTGATTAGAGCCCTGGA*Scramblesg1: CACTACCAGAGCTAACTCAT*CRISPick sgRNA. sgRNA sequences. Guide RNA sequences used in the CRISPRi knock-down viruses. CRISPick sgRNA. Guides were ordered as oligonucleotides from Azenta (Genewiz) with additional bases appended to facilitate seamless Golden Gate Cloning into the Lenti-CRISPR-v2 vector (Addgene #52961) in frame with a U6 promoter sequence and sgRNA scaffold. Following annealing and phosphorylation of gRNA pairs as in Sanjana et al., 2014, guides were cloned into the Lenti-CRISPR-v2 vector by Golden Gate Cloning using BsmBI (NEB; R0739). Competent DH5a cells (Zymo; T3007) were transformed with the reaction product, and plasmid DNA was isolated from resultant colonies and sequenced to verify successful cloning. sgRNA cloned in frame with the U6 promoter and sgRNA scaffold were amplified from the vectors by PCR using primers in Table 3 containing a 5’ recognition element for the type IIS restriction endonuclease SapI and unique overhang sequences to facilitate cloning of tandem arrays. PCR products were purified and cloned in tandem by Golden Gate Cloning using SapI (NEB; R0549) into a custom vector modified from the pAAV-SYN-DIO-EGFP (Addgene #50457) or pAAV-SYN-DIO-mCherry (Addgene #50459) AAV transfer plasmids. Custom sgRNA delivery AAV vectors were generated by inserting a dual-SapI cassette between the MluI and ApaI sites upstream of the hSyn promoter in either plasmid. Plasmids isolated from competent DH5a cells transformed with the reaction product were sequenced to confirm successful cloning of two to four guides. Table 3.sgRNA Golden Gate primers.Primers used for Golden Gate cloning into the Lenti-CRISPR-v2 vector.NamePrimer sequence (5’ ->3’)GG1F: TAGCTCTTCACGTGACCGAGGGCCTATTTCCCATGATTCCR: TAGCTCTTCATAGGGCGAATTCAAAAAAGCACCGACTCGGGG2F: TAGCTCTTCACTAGACCGAGGGCCTATTTCCCATGATTCCR: TAGCTCTTCATGAGGCGAATTCAAAAAAGCACCGACTCGGGG3F: TAGCTCTTCATCAGACCGAGGGCCTATTTCCCATGATTCCR: TAGCTCTTCAATCGGCGAATTCAAAAAAGCACCGACTCGGGG4F: TAGCTCTTCAGATGACCGAGGGCCTATTTCCCATGATTCCR: TAGCTCTTCAGCAGGCGAATTCAAAAAAGCACCGACTCGGGG5F: TAGCTCTTCATGCGACCGAGGGCCTATTTCCCATGATTCCR: TAGCTCTTCACAGGGCGAATTCAAAAAAGCACCGACTCGGGG6F: TAGCTCTTCACTGGACCGAGGGCCTATTTCCCATGATTCCR: TAGCTCTTCACATGGCGAATTCAAAAAAGCACCGACTCGG sgRNA Golden Gate primers. Primers used for Golden Gate cloning into the Lenti-CRISPR-v2 vector. HEK293T cell culture AAV production HEK293T cells were obtained from ATCC (CRL-3216) and were certified to be mycoplasma free at the time of ordering but were not further validated. These cells were maintained at low passage numbers under standard cell culture conditions (37 °C, 5% CO2) in DMEM, high glucose, pyruvate (Gibco; 11–995) supplemented with 10% FBS (Gibco; 10438026) and 1% Penicillin/Streptomycin (Gibco; 15140122). For viral production, HEK293T cells were plated in three to six large 150-mm cell culture plates (Corning; CLS430599) per virus and transfected with 10 µg pAdDeltaF6 (pHelper, Addgene #112867), 6 µg pAAV2/9 n (Rep2/Cap9, Addgene #112865), and 6 µg transfer plasmid (Custom, described above) using PEI MAX (PolySciences Inc; 24765). Transfected cultures were maintained in 5% FBS media for 72 hr until harvesting, and pooled pellets were resuspended in 1 x PBS-MK and stored at –80 °C. rAAV was purified from frozen pellets using the IDX gradient method described in Zolotukhin et al., 1999. Following ultracentrifugation, the viral fraction was repeatedly passed through an Amicon Ultra 100 kDa dialysis column (Millipore; UFC910008) by centrifugation to remove IDX carryover and concentrate the virus in Lactated Ringer’s solution. Resultant viral titer was between 2.0 and 8.0 × 1013 vg/ml as measured by qPCR using an AAV Titer kit (ABM; G931). Intracerebroventricular injections LSL-dCas9-KRAB+/+ mice were housed in harems with Emx1cre/cre mice, and dams were checked twice weekly for pregnancy. Pregnant mice were housed individually with extra bedding and paper huts and monitored daily for new litters. At P0.5, mice were injected intracerebroventricularly with AAV using the protocol established in Kim et al., 2013. Briefly, pups were separated from their mothers in groups of four and anesthetized on ice. Once movement could not be detected, pups were placed on a cooling block and injected bilaterally with 1 µl AAV9-[U6-gRNA]-hSyn-DIO-[EGFP/mCherry] at 2.0–5.0×1013 viral genome copies/ml into their lateral ventricles. Pups were then marked by tail or toe clips and immediately warmed by hand until movement was restored, then placed in soft bedding on a heat pad until the entire cohort had been injected. Pups were returned to their mother once all pups had been injected, and mice were housed overnight on a heating pad. Injected pups were monitored the next day to monitor viability and ensure that maternal care was being provided, and cross-fostered with experienced LSL-dCas9-KRAB+/+ mothers if needed. Fluorescence-activated cell sorting (FACS) Cells were sorted from acute coronal slices as in Clark et al., 2019 and Clark et al., 2020. Briefly, animals were fatally anesthetized with 0.1 ml/10 g of a ketamine (20 mg/ml), xylazine (2.5 mg/ml), and acepromazine (0.5 mg/ml) (KXA) mixture. After loss of toe pinch response, animals P10 and older were transcardially perfused with 10–20 ml of cold choline chloride solution (CS) containing blockers of neuronal activity. Brains were swiftly extracted and sliced on a vibratome (Leica; VT1000S) at 300 µm in cold CS, then slices were transferred to filtered inserts to equilibrate at 37 °C for 15 min. Inserts containing slices were transferred to room temperature artificial cerebrospinal fluid (ACSF) containing blockers of neuronal activity and Protease for 90 min to digest tissue. Relevant cortical layers (for Layer 4/6 experiments) or densely infected regions (for CRISPRi experiments) were microdissected using an 18 G syringe tip (BD Biosciences) under a fluorescent dissection microscope (Zeiss; MZFLIII) and kept on ice in ACSF containing 1% FBS (Gibco; 10438026). Dissected tissue was triturated using progressively smaller flame-polished glass pipettes and filtered through a cell strainer (Falcon; 352235), then centrifuged in a 10% Percoll gradient (Sigma-Aldrich; P1644) to generate a pellet enriched in pyramidal cells. The pellet was resuspended in fresh ACSF +1% FBS and stored on ice prior to isolation by FACS (BD FACSAria Flow Cytometer). Neurons in the top 10–25% of fluorescent cells above the autofluorescence background were collected. 1000–5000 cells were sorted into PicoPure Extraction buffer, incubated at 42 °C for 30 min, and stored at –80 °C until RNA extraction. RNA extraction RNA was isolated from sorted cell extract using PicoPure RNA Isolation Kit (Arcturus; KIT0204) with on-column DNAse digestion step (QIAGEN; 79254), and quantified on Qubit fluorometer using the RNA High Sensitivity reagents (Thermo Fisher; Q32852). cDNA synthesis and qPCR Reverse transcription was performed with 10–20 ng of RNA using the iScript Reverse Transcription Supermix (Bio-Rad; 1708891) according to manufacturer’s instructions. cDNA was diluted to a 0.25 ng/µl equivalent and used as a template for qPCR with SsoAdvanced Universal SYBR Supermix (Bio-Rad; 1725272) according to manufacturer’s instructions on a Corbett RotorGene 6000 thermocycler. All reactions were performed in triplicate. qPCR primers were designed to span 3’ exon-exon junctions using Primer3plus and assessed for homo- and heterodimer formation using IDT’s oligoanalyzer tool. Primers were tested for efficiency and specificity against serial dilutions of whole cortex cDNA, and only primer pairs with 95–105% efficiency were retained. qPCR primer sequences are listed in Table 4. Table 4.qPCR primers.qPCR primers used to measure abundance of named transcripts.NamePrimer sequence (5’ ->3’) Klf9 F: ACAGAGTGCATACAGGTGAACGR: CTTCTCACACAGTGGGCAAC Klf13 F: GAAGCACAAGTGCCACTACGR: CTCACCTGTGTGAGTTCTCAGG Klf6 F: CTCCCACTGTGACAGGTGTTR: AGCCTACAGGATTCGTCCCT Klf7 F: AGTGCTCATGGGAAGGATGCR: CTGGAGAAACACCTGTCGCA GAPDH F: CGTCCCGTAGACAAAATGGTR: TCAATGAAGGGGTCGTTGAT qPCR primers. qPCR primers used to measure abundance of named transcripts. qPCR data was analyzed using the Ct method. To account for technical differences across several qPCR runs, relative expression was calculated using Ct’s obtained from all samples across multiple runs. The control Ct for each measured gene was computed by average Ct’s from the control or baseline group (scramble sgRNA at P10-12 for Klf6/7 data, scramble sgRNA at P18-20 for Klf9/13) and subtracted from the Ct to obtain the Ct. As primer efficiency was ~100% for all genes measured, relative expression was simply calculated as. For Klf9/13 qPCR, significance was measured using a Wilcoxon rank sum test. For Klf6/7 qPCR, significance for the effect of Age and Virus on each measured gene was calculated using a 2-way ANOVA using the formula Relative expression ~Age*Virus, followed by Wilcoxon rank sum tests. Perfusion Animals were fatally anesthetized with KXA mixture. After loss of toe pinch response, mice were transcardially perfused with 10 ml of cold 1 x PBS followed by 10 ml of cold 4% PFA. Brains were carefully removed from the skull and incubated in 4% PFA for 24–48 hours on a rocker at 4 °C before being transferred to 1 x PBS for storage at 4 °C. For P2 brains, wild-type pups were anesthetized on ice until there was no toe pinch response and then swiftly decapitated. The skin of the head was carefully peeled back, and the brain and skull were washed with ice cold 1 x PBS every 30 s while the skull and pia were peeled away with fine forceps. Brains were immediately transferred to 5 ml tubes containing ice cold 1 x PBS and washed twice with cold PBS to remove residual blood, then resuspended in ice cold 4% PFA and incubated overnight at 4 °C on a rocker. Vibratome sectioning and epifluorescence imaging The cerebellum of fixed brains was removed using a scalpel, and brains were imbedded in 4% agarose dissolved 1 x PBS. Brains were mounted on a pedestal and sliced on a vibratome (Leica; VT1000S) in groups of 2–4 at a thickness of 50 µm. Slices were transferred to 24 well plates containing 1 x PBS +0.05% Sodium Azide using a paintbrush and stored protected from light at 4 °C. For imaging of representative slices, selected slices were incubated in 1 µg/ml DAPI (Thermo; 62248) in 1 x PBS for 10 min at room temperature, washed once with PBS for 5 min, and mounted on slides with VectaMount Mounting Medium (SouthernBiotech; 0100–01). Slides were sealed with nail polish and stored at 4 °C prior to imaging. Whole brain slices were imaged on a Keyence BZ-X epifluorescent microscope in a single z plane with exposure settings calibrated to prevent saturation of the brightest infected cells and stitched using Keyence’s native stitching algorithm. As a consequence, infection appears biased towards thick-tufted Layer 5 neurons while the true laminar distribution is much broader. CRISPRi RNA-sequencing Prior to library preparation, RNA quality was measured using High Sensitivity RNA ScreenTape (Agilent; 5067–5579). For experiments using AAV containing mCherry as a reporter, RNA-seq libraries were prepared using the SMARTer Stranded Total RNA-Seq Kit v3 (Takara Bio; 634485) with 5–10 ng of RNA as input according to manufacturer’s instructions. Library fragment size was evaluated using D1000 High Sensitivity ScreenTape (Agilent; 5067–5584). For experiments using EGFP as a reporter, libraries were prepared using the SMART-Seq mRNA LP kit (Takara Bio; 634762) with 5–10 ng of RNA as input followed by 2 ng of cDNA according to manufacturer’s instructions. cDNA concentration and library fragment size distribution were measured using D5000 High Sensitivity ScreenTape (Agilent; 5067–5593). For all libraries, concentration was measured using KAPA qPCR Library quantification kit (Roche; KK4824). Samples were pooled for PE sequencing on the Illumina NovaSeq platform (Novogene) or Illumina NextSeq 550/500 platform. RNA-sequencing analysis Illumina TruSeq adapter sequences were trimmed from all libraries using cutadapt v 3.4, and reads with MAPQ <30 were removed. For v3 libraries containing UMIs, 8 bp UMIs were trimmed from Read 2, and reads were marked using umitools v. 1.1.2, and reads were trimmed to 75 bp for faster mapping using cutadapt. Reads were then mapped to the GENCODE M27 genome using STAR v. 2.7.8 a with ENCODE Long Read parameters, and mapped reads were deduplicated using umitools. For LP libraries, cutadapt was used to trim bases added by the SMART-Seq TSO ATTGCGCAATGNNNNNNNNGGGG from Read 1 prior to mapping to the GENCODE M27 genome by STAR as above. Paired end reads mapping to exons of known genes were quantified using featureCounts in the Rsubread package. Differential expression analysis was conducted using DESeq2 with model design ~Batch + Group, where Group is the Age:Virus identity of each sample controlling for batch effects across sequencing runs. DEGs in each KD condition were identified relative to scramble gRNA controls of the same age with cutoffs |log2FC|>1 and adjusted p-value <0.05. Adjusted p values for each gene were reported from a Wald test with BH correction for multiple comparisons, and log2FC values were shrunk using the ashr method. For plotting heatmaps and normalized line plots, scaled counts matrices were generated by z-scoring the condition-averaged relative log (rlog) expression and corrected for batch effects using the removeBatchEffect function from limma (Ritchie et al., 2015). Optimal cluster numbers for heatmaps were determined as above, but implemented in ComplexHeatmap (Cha et al., 2016) using 100 random starts. Motif counting and enrichment in Promoter peaks were conducted using the methods described above. For coverage plots, control genes were chosen among those unaffected by either KLF manipulation with a quantile-matched distribution of log2-FoldChanges between P2 and P30 in order to account for expected effects of developmental expression on chromatin accessibility. For analysis of the effect of Thyroid Hormone on KLF target expression, snRNA-seq data from T3-treated and control mice published in Hochbaum et al., 2024 were downloaded from the CELL x GENE repository (here), converted to a Seurat object in R, and filtered for excitatory neuronal cell type clusters annotated by the authors. Differentially expressed genes were identified using Seurat’s FindMarkers() function comparing T3 and control neurons within each cluster using a Wilcox test with a log fold-change threshold of 0.1. The resulting data was subset by high-confidence shared KLF targets identified as a DEG in at least one cluster for plotting. Electrophysiology Recordings were performed as described in Valakh et al., 2023. Mice were fatally anesthetized with KXA between P18-20 and their brains were swiftly removed and allowed to acclimate in ice-cold oxygenated ACSF containing (in mM) 126 NaCl, 3 KCl, 1 NaH2PO4, 25 NaHCO3, 2 MgCl2, 2 CaCl2, and 10 Glucose. 300 µm acute coronal slices were made using a vibrating microtome (Leica; VT1000S) and slices were incubated at 37 °C in oxygenated ACSF for 30 min to 1 hr. Slices were subsequently maintained in oxygenated room temperature ACSF and recordings were conducted at 34–35°C. Neurons were visualized on an Olympus upright epifluorescence microscope with 40× water immersion objectives and infected cells were identified based on the presence of an mCherry reporter. Whole-cell patch-clamp recordings were made using near-infrared differential interference contrast microscopy. Recording pipettes of 3–5 MΩ resistance contained internal solution containing 20 KCl, 100 K-gluconate, 10 HEPES, 4 Mg-ATP, 0.3 Na-GTP, 10 Na-phosphocreatine, and 0.1% biocytin. Recordings were amplified (Multiclamp 700B, Molecular Devices) and digitized at 4 kHz using a National Instruments Board under control of IGOR Pro (WaveMetrics). For Intrinsic Excitability recordings, synaptic currents were blocked with 25 µM picrotoxin, 25 µM DNQX, and 35 µM APV. Cells were maintained at –70 mV by injecting the required current and steady-state series resistance was compensated. 500 ms currents between 50–700 pA were randomly injected every 10 s (at least two times per cell) with a 25 pA seal test before each trial. To record miniature Excitatory AMPA currents, action potentials and non-AMPA synaptic currents were blocked with 25 µM picrotoxin, 500 nM TTX, and 35 µM APV. Cells were clamped at –70 mV with no compensation for the liquid junction potential and at least 3 10 s sweeps were recorded for each cell with a seal test before each trial. Threshold-based detection of mEPSCs and action potentials was implemented using custom-written R scripts with routines for baseline subtraction, custom filtering, and measurements of intervals, amplitudes, and kinetic properties (https://github.com/rkirk822/Kirk2026_elife copy archived at Kirk and Nelson, 2026). Only cells with a Vrest < –60 mV that maintained Rin and Vrest within 20% of their value between the first and final recording were retained for analysis. Comparisons between sgRNA treatment groups were made by a two-way ANOVA for the effects of current and sgRNA for intrinsic excitability, and a Mann-Whitney U test for mEPSC parameters. For visualization of patched cells by biocytin staining, slices were fixed overnight in 4% PFA at 4 °C followed by 2 1 X PBS washes and 2 hr in blocking buffer (1 x PBS supplemented with 5% normal goat serum [NGS, Gibco], 3% bovine serum albumin [BSA, Thermo Fisher], and 0.5% Triton-X-100 [Sigma]). Slices were then incubated at 4 °C overnight with Alexa-594-Streptavidin diluted 1:500 in blocking buffer. The following day, slices were washed three times in 1 x PBS, stained with 1 µg/ml DAPI (Thermo Fisher; 62248) in 1 x PBS for 10 min at room temperature, washed once with PBS for 5 min, and mounted on slides with VectaMount mounting medium (SouthernBiotech; 0100–01). Slides were sealed with nail polish and stored at 4 °C prior to imaging. Slides were imaged on a Nikon AX-R Confocal Microscope with a 20 x air objective by taking Z-stacks spanning the entire dendritic arbor at a 5 µm step size. Images were converted into maximum intensity projections in Fiji. RNAScope Following overnight incubation in 4% PFA, brains were transferred to 30% sucrose/PBS overnight followed by 15% sucrose/PBS overnight to dehydrate tissue. The olfactory bulb and cerebellum were removed using razor blades in a mouse brain matrix (AgnTho’s; 69-1175-1) and brains were flash frozen in OCT (Thermo Fisher; 23-730-571) molds by submerging them in a 100% ethanol and dry ice bath for 5 min. Frozen tissue blocks were immediately transferred to –80 °C. Before slicing, brains were equilibrated to –20 °C for ~30 min and excess OCT was removed. Brains were mounted onto a freezing microtome sliced at 16 µm with 100 µm between each slice with slices centered around primary somatosensory cortex (S1). Slices were transferred to PBS using a paintbrush and mounted on charged slides (Thermo Fisher; 12-550-15). Slides were dried at 4 °C overnight and transferred to a sealed slide box at –80 °C the following day. For RNAScope of brains infected with AAV, we began the RNAScope Multiplexed Fluorescent Assay v2 protocol (ACD; PN 323110) just before the protease treatment step to avoid denaturing viral EGFP during the Target Retrieval step. Slides with one to two slices each were removed from –80 °C, washed once with PBS, and allowed to air dry before the application of a hydrophobic barrier. Slides were stored overnight at 4 °C and probed following the standard RNAScope protocol, beginning with Protease III treatment. P20 brains were probed with Mm-C2-Klf9 (ACD; 488371-C2) and either Mm-C1-Dpysl3 (ACD; 496611) or Mm-C1-Tubb2b (ACD; 1039621-C1) and stained with Opal 620 Dye (Akoya; FP1495001KT) at a 1:750 dilution for the C2 channel, and Opal 690 (Akoya; FP1497001KT) at 1:750 for C1. P10 brains were probed with Mm-C2-Klf7 (ACD; 570961-C2) and either Mm-C1-Dpysl3 or Mm-C1-Tubb2b and stained with Opal 620 Dye at a 1:1500 dilution for the C1 channel and Opal 690 at 1:750 for C2. For RNAScope of Klf9 and Klf7 expression across development, samples were processed following the RNAScope Multiplexed Fluorescent Assay v2 protocols with 2 exceptions: for P2 and P7 samples, the Target Retrieval step was reduced to 5 minutes and brains were treated with with Protease Plus for 15 minutes to prevent over-digestion of fragile neonatal tissue. All slices were probed with Mm-C1-Klf7 (ACD; 570961) and Mm-C2-Klf9 and stained with Opal 520 Dye (Akoya; FP1487001KT) at a 1:1500 dilution for the C1 channel and Opal 620 at 1:1500 for C2. RNAScope image acquisition and analysis All slides were imaged on a Zeiss 880 confocal microscope. For CRISPRi samples, z-stacks were acquired from at least two locations in each hemisphere centered on cortical Layer 5 (minimum four total) under a ×40 oil-immersion objective. For P2 brains, stitch-stacks spanning the entire height of the cortex were acquired under a ×40 oil-immersion objective and stitched using the native stitching algorithm in Zen (Zeiss). For P7 and P30 brains, stitch-stacks containing all cortical layers were acquired under a ×20 air-immersion objective and stitched as above. Image analysis and quantification was conducted using custom Python scripts (https://github.com/rkirk822/Kirk2026_elife copy archived at Kirk and Nelson, 2026). For RNAScope of Klf9 and Klf7 expression across development, images were first rotated so that the pial surface was parallel to the bottom and top borders of the image, then cropped so images spanned from the bottom of Layer 6 to the epithelial cell layer above Layer 2/3 (dorsal-ventral axis). The nuclear DAPI channel was downscaled as above and used to segment and label nuclei with CellPose v. 3.0.1 (Stringer et al., 2021) in 2.5-D mode using a stitch threshold of 0.3, cell probability threshold of 3, and a diameter of 25 pixels for ×40 images and 12 for ×20 images. Artifacts were removed by excluding labels with a total pixel volume in the bottom 2% of the measured range. RNA puncta were segmented and labeled as above. The number of unique nuclei and RNA labels in each x-by-z slice along the y (dorsal-ventral) axis were counted and exported to R for downstream analysis. To normalize RNA puncta counts for differences in cortical thickness across animals and ages, a binned dorsal-ventral axis was created by binning tables containing counts of nuclei and RNA along the y-axis into n=2000 bins for P7 and P30 samples and n=1500 bins for P2 samples. Finally, RNA puncta counts were normalized by nuclei counts within each bin to create a measure of RNA quantity scaled by cell density. 95% confidence intervals for both mRNA species were estimated by bootstrapping with 1000 random resamples using the R package boot. Significance of the difference between Klf9 and Klf7 abundance for each age group at each position was measured by a Mann-Whitney U test with BH correction. For CRISPRi RNAScope, images were split into their composite channels and the EGFP channel was downscaled by 4 x and a Gaussian blur was applied with sigma = 1.3. The downscaled image was used to segment and label cells with CellPose v. 3.0.1 in 2.5-D mode using the cyto2 model with a stitch threshold of 0.5 and a diameter of 40 pixels. Presumptive false positives were removed by excluding labels with a summed fluorescence signal in the bottom 10% of the measured range or with more than 5% of their pixel volume along the edges of the image (including top and bottom). Remaining labels were upscaled 4 x to the dimensions of the original EGFP image. The raw image intensities from the two RNA channels were reciprocally subtracted from one another to eliminate bleed-through artifacts, and RNA puncta were labeled and segmented using pyclesperanto’s veronoi_otsu_labeling with a spot sigma of 2 and outline sigma of 1. The overlap (in pixels) of every RNA puncta with every cell label was calculated and exported to R for downstream analysis. Puncta below 20 pixels in size or puncta overlapping cells by less than 75% of their pixel volume were excluded. The number of unique RNA puncta overlapping each cell label was quantified and scaled by the area of the cell label in pixels ×10–6 to produce area scaled puncta counts. Significance for the difference between each gene across treatment groups was measured by an independent samples t-test. Materials availability statement All AAV transfer plasmids generated in this study are available upon request by contacting nelson@brandeis.edu. This is an important study utilizing innovative CRISPR based approaches demonstrating the role of the KLF family of transcription factors in the post-natal maturation of cortical projection neurons. The strength of evidence overall is compelling, and the study is well executed. The screen data presented provides a number of interesting candidates for future analyses into the mechanism of action of KLF family members in neuronal maturation. [Editors' note: this version has been assessed by the Reviewing Editor without further input from the original reviewers. The original reviews were generally positive. In their revision, the authors partially address reviewers concerns, but mention that a number of requested experiments are for future studies.] Summary: This is an interesting manuscript by Kirk and colleagues describing a highly valuable knock-down system that leverages CRISPRi in order to further elucidate the role of the Kruppel-Like Factor (KLF) transcription factor family in regulating the maturation of postnatal cortical projection neurons. The authors firstly use RNA-Seq and ATAC-Seq data in order to identify the KLF TF family as a potential regulator of cortical neuron maturation in the postnatal brain and subsequently knock down four KLF family members; KLF9, KL13, KLF6 and KLF7, in order to ascertain the functions of specific KLF genes in the developing cortex. The described CRISPRi knock down strategy is highly robust and penetrant as evidenced by a KD efficiency > 95% (assessed by both qPCR and single molecule FISH) and demonstrates that KLF6 and KLF7 play an activating role in driving the expression of target genes relating to axonal growth whereas KLF9 and 13 play a repressive role that inhibits the expression of overlapping gene targets. Together, the authors propose a model where the KLF TF family acts as a regulatory "switch" from activation to repression in the postnatal cortex as a mechanism to control a shift in projection neuron function from axonal growth to circuit refinement. The findings and conclusions of the manuscript offer a valuable contribution to the field of postnatal cortical development and further our understanding of the regulatory mechanisms that govern neuron maturation. Summary: Kirk et al. use RNA-Seq and CRISPRi to provide evidence that KLF family transcription factors regulate postnatal neuronal maturation of pyramidal neurons. The genetic programs regulating postnatal neuronal maturation are not well understood. The authors first analyzed chromatin accessibility and gene expression data from layer 4 and 6 pyramidal neurons and found that KLF TFs are predicted regulators of postnatal neuronal maturation. They then use CRISPRi knockdown and find that KLF activators first activate genes and then this is followed by KLF repressors repressing genes. Interestingly, some genes, such as those with cytoskeletal functions, are shared targets of KLF activators and repressors. Strengths: The study is well-executed and the paper is well-written. A major strength of this study is the application of state-of-the-art transgenic approaches. The CRISPRi approach used to knock down multiple KLFs is compelling. The genomic data generated appears to be high quality and is carefully analyzed. The presented findings provide important insights into the genetic programs that regulate postnatal maturation in cortical pyramidal neurons. The discovery that KLF family activators/repressors regulate gene expression changes during this critical step of neuronal development fills an important gap in the field. Weaknesses: While beyond the scope of the current study, future studies should investigate the contributions of KLFs on postnatal morphological and physiological changes. Summary: In their manuscript "Multiplexed CRISPRi Reveals a Transcriptional Switch Between KLF Activators and Repressors in the Maturing Neocortex", Kirk and colleagues seek to dissect the developmentally regulated pan-neuronal gene programs that control the postnatal maturation of cortical neurons. For this, the authors analyzed newly generated and existing RNA-seq and ATAC-seq of Layer 4 and Layer 6 cortical pyramidal neurons at postnatal day 2 (P2) and day 30 (P30) and identified thousands of shared developmentally regulated genes and genomic (promoter) regions, including genes involved in axon growth (tend to be downregulated) and synaptic function (tend to be upregulated). Motif enrichment analysis of promoters of differentially regulated genes revealed a strong presence of KLF/Sp family binding motifs, pointing to Krüppel-Like Factors (KLFs) as key transcriptional regulators of cortical maturation. Expression profiling showed a developmental switch from activating KLFs (Klf6, Klf7) expressed neonatally to repressive KLFs (Klf9, Klf13) upregulated during maturation. Using an elegant in vivo multiplexed CRISPR interference (CRISPRi) system, the authors achieved efficient, cell-type-specific knockdown of these TFs and showed that Klf9 and Klf13 repress a set of genes that includes cytoskeletal regulators such as Tubb2b, Dpysl3, and Rac3. Conversely, Klf6 and Klf7 promoted the expression of these same genes in the early postnatal period, and their knockdown led to reduced expression of these genes, particularly at P10 when their activating influence is strongest. Since promoters of shared KLF targets were enriched for KLF/Sp motifs but showed little change in chromatin accessibility, the authors propose a model in which distinct KLF family members function either as transcriptional repressors and activators that compete at constitutively accessible promoters and thereby act as a developmental transcriptional switch that coordinates the downregulation of axon growth programs and upregulation of synaptic maturation genes during cortical development. Strengths: The study addresses an interesting question and advances our understanding of the transcriptional regulation underlying postnatal cortical development. A major strength of the study lies in the innovative use of in vivo multiplexed CRISPR interference (CRISPRi), which allows for cell-type-specific, combinatorial knockdown of redundant TFs - this an elegant solution to a long-standing challenge in transcription factor research and should be useful also for other neuroscience studies that require local and cell-type-specific gene loss-of-function. Also, the integration of RNA-seq and ATAC-seq across developmental time points provides a robust foundation for identifying direct targets of the KLF family, and the findings are reinforced by cross-species conservation and the identification of targets with clear neurodevelopmental relevance. Weaknesses: The major weakness of the study lies in its relatively narrow scope: the study focuses primarily on transcriptional mechanisms and largely lacks functional validation of the neuronal phenotypes that are predicted by the gene expression data (e.g. axonal morphology). Despite these limitations, the paper offers an interesting model for a transcriptional switch during neuronal maturation in the cortex and establishes a powerful methodological framework for dissecting redundant gene networks in vivo. The following is the authors’ response to the original reviews. Public Reviews: Reviewer #1 (Public review): Summary: This is an interesting manuscript by Kirk and colleagues describing a highly valuable knock-down system that leverages CRISPRi in order to further elucidate the role of the Kruppel-Like Factor (KLF) transcription factor family in regulating the maturation of postnatal cortical projection neurons. The authors firstly use RNA-Seq and ATAC-Seq data in order to identify the KLF TF family as a potential regulator of cortical neuron maturation in the postnatal brain and subsequently knock down four KLF family members; KLF9, KL13, KLF6 and KLF7, in order to ascertain the functions of specific KLF genes in the developing cortex. The described CRISPRi knock down strategy is highly robust and penetrant as evidenced by a KD efficiency > 95% (assessed by both qPCR and single molecule FISH) and demonstrates that KLF6 and KLF7 play an activating role in driving the expression of target genes relating to axonal growth whereas KLF9 and 13 play a repressive role that inhibits the expression of overlapping gene targets. Together, the authors propose a model where the KLF TF family acts as a regulatory "switch" from activation to repression in the postnatal cortex as a mechanism to control a shift in projection neuron function from axonal growth to circuit refinement. The findings and conclusions of the manuscript offer a valuable contribution to the field of postnatal cortical development and further our understanding of the regulatory mechanisms that govern neuron maturation. The conclusions of this manuscript are generally supported by the data, but some aspects of the data collection and analysis require some further clarification. Specifically: (1) The authors comprehensively assess the molecular effects of KLF TF knock-down, however, the authors do not deeply address the cellular effects of these knock-downs. The authors conclude that knockdown of KLF6/7 and KLF9/13 cause downregulation and upregulation, respectively, of a common set of genes involved in cytoskeletal or axon regulation such as Tubb2 and Dpysl3. How is the morphology of the cells affected by these knockdowns? For example, does KLF9/13 knockdown cause neurite/axonal outgrowth? The authors should perform some basic experiments to assess changes in cell morphology following KLF TF KD. This is the one key point that needs addressing, in my opinion. We appreciate this comment and agree that the cellular effects of KLF activator and/or repressor KD are not addressed by the experiments in this manuscript. However, the effects of KLF9, KLF13, and KLF9/13 knockdown on neurite outgrowth have been previously examined in vitro (Avci et al., 2012; Avila-Mendoza et al., 2020) and in vivo (Apara et al., 2017). The collective findings of these papers (enhanced neurite outgrowth and axon regeneration following KLF repressor KD) align with the predicted outcome of upregulating the set of cytoskeletal remodeling genes identified as putative KLF family targets in this manuscript. Furthermore, the in vitro results demonstrate that the partial redundancy between KLF9 and KLF13 we identified at the transcriptional level is relevant for their roles in repressing neurite outgrowth. Our results suggest that these previous findings are likely to hold true in cortical neurons in vivo while offering a molecular explanation for these effects at the level of gene expression. Other groups have examined the effect of KLF6 or KLF7 overexpression on corticospinal axon regeneration in vivo and found that these transcription factors can individually promote axon regrowth after injury (Blackmore et al., 2012; Wang et al., 2018). Similarly, this earlier work did not identify transcriptional mechanisms underlying the observed effect so our results also offer a plausible set of targets that could mediate the link between KLF activator overexpression and axon regeneration. (2) The authors identify 374 DEGs in P10 Klf6/7 KD neurons and 115 DEGs at P20 (figure 6B). Have the authors looked to see what proportion of these DEGs are upregulated in the KLF9/13 KDs in order to get a more global understanding of the degree of overlap in the genes regulated by the KLF family members? [MOU2] Along similar lines, the authors later indicate that there are 144 shared targets between the KLF activator and repressor pairs (Figure 7C). What percentage does this represent of the total number of DEGs between the KLF pairs. This could further illustrate the degree to which the KLF pairs regulate the same set of genes. If it is already indicated in the manuscript, it should be made a bit more clear to the reader. We thank the reviewer for the suggestion to include more comprehensive quantitative measures of the degree of overlap between targets of KLF activator and repressor pairs. We have included the exact number and percentage of P10 and P20 KLF6/7 targets that are differentially expressed (adj. p-value <= 0.05) in KLF9/13 KD neurons and vice versa in the text in the section associated with Figure 7, which is devoted to describing this class of overlapping targets. Furthermore, figures 7D and S7.2B both show how the full set of KLF6/7 targets are affected by Klf9/13 KD and vice versa. Collectively, this demonstrates that both KLF activator and repressor targets trend towards opposite regulation by the opposing pair. (3) Figures 5B and 6D2 are very interesting as they relate the changes in gene expression over time in neurons from P2 to P30 to the functions of KLF9/13 and KLF6/7, respectively. I would be curious to see how these two forms of analyses overlap with one another. For example, in Figure 6D2, where would the KLF9/13 upregulated genes fall on the plot shown in Figure 6D2? And would those overlapping genes fit a similar correlation? We agree with the reviewer that this is a powerful way to visually demonstrate the overlap between KLF6/7 and KLF9/13 targets in a more unbiased way and within a developmental context. The suggested analysis has been included in the supplement to Figure 7 (Fig S7.3). (4) Figure 7E shows expression levels of shared KLF TF targets in control or KD conditions. Interestingly, the expression of Tubb2b, shows higher expression in ScrGFP P10 when compared to KLF9/13 P20, suggesting that derepression of KLF9/13 does not fully restore the expression level of Tubb2b seen at P10. This may suggest that other repressive regulators may be involved in the downregulation of Tubb2b from P10 to P20[MOU4] . Can the authors further comment on this, perhaps in the discussion, and speculate if there are other regulatory factors at play that may be controlling some of the shared targets by KLF6/7 and KLF9/13? We thank the reviewer for this keen observation. We have included a comment on this within the text associated with Figure 7. While other regulatory factors are plausible, this is easily explained by low expression of Klf6 and Klf7 in the P20 cortex, which cannot drive transcription of Tubb2b and other targets to the same level as what is observed at P10 when the KLF activators are still relatively abundant, even after repression by Klf9/13 is removed. Within this framework, overexpression of KLF activators on a KLF repressor background would be the only way to restore expression of Tubb2b to its P10 expression levels. This is wholly compatible with our model of ‘push-pull’ regulation of KLF targets outlined in Figure 9. The finding could also be affected by the addition of repressive histone marks by the KLF9/13-associated SID complex in early neonatal development that may persist following the knockdown of these repressors, preventing complete restoration of neonatal expression patterns. Reviewer #2 (Public review): Summary: Kirk et al. use RNA-Seq and CRISPRi to provide evidence that KLF family transcription factors regulate postnatal neuronal maturation of pyramidal neurons. The genetic programs regulating postnatal neuronal maturation are not well understood. The authors first analyzed chromatin accessibility and gene expression data from layer 4 and 6 pyramidal neurons and found that KLF TFs are predicted regulators of postnatal neuronal maturation. They then use CRISPRi knockdown and find that KLF activators first activate genes and then this is followed by KLF repressors repressing genes. Interestingly, some genes, such as those with cytoskeletal functions, are shared targets of KLF activators and repressors. Strengths: The study is well-executed and the paper is well-written. A major strength of this study is the application of state-of-the-art transgenic approaches. The CRISPRi approach used to knock down multiple KLFs is compelling. The genomic data generated appears to be high quality and is carefully analyzed. The presented findings provide important insights into the genetic programs that regulate postnatal maturation in cortical pyramidal neurons. The discovery that KLF family activators/repressors regulate gene expression changes during this critical step of neuronal development fills an important gap in the field. Weaknesses: A limitation of the current study is that the functional importance of KLF for postnatal neuronal maturation is unclear. Although the authors find that KLFs regulate some of the gene expression changes during postnatal neuronal maturation, it is still unclear whether such gene expression changes mediate the postnatal changes in morphology and physiology. While beyond the scope of the current study, future studies should investigate the contributions of KLFs on postnatal morphological and physiological changes. We thank the reviewer for their helpful comments on this manuscript. We agree that the effects of KLF knockdown identified in this study are primarily descriptive, but – as noted - a detailed analysis of the morphological and physiological consequences of KLF knockdown are beyond the scope of this paper. However, we believe that a more mechanistic model of KLF function during neuronal maturation can be obtained by considering our findings on the bidirectional regulation of core cytoskeletal genes by KLF activators and repressors alongside published in vivo and in vitro data on the opposing roles of KLF family members on axon outgrowth/regrowth (Apara et al., 2017; Avila-Mendoza et al., 2020; Blackmore et al., 2012; Moore et al., 2009; Wang et al., 2018). Thus, we offer a set of transcriptional targets that likely mediate these opposing effects and a developmental context within which they might operate. Reviewer #3 (Public review): Summary: In their manuscript "Multiplexed CRISPRi Reveals a Transcriptional Switch Between KLF Activators and Repressors in the Maturing Neocortex", Kirk and colleagues seek to dissect the developmentally regulated pan-neuronal gene programs that control the postnatal maturation of cortical neurons. For this, the authors analyzed newly generated and existing RNA-seq and ATAC-seq of Layer 4 and Layer 6 cortical pyramidal neurons at postnatal day 2 (P2) and day 30 (P30), and identified thousands of shared developmentally regulated genes and genomic (promoter) regions, including genes involved in axon growth (tend to be downregulated) and synaptic function (tend to be upregulated). Motif enrichment analysis of promoters of differentially regulated genes revealed a strong presence of KLF/Sp family binding motifs, pointing to Krüppel-Like Factors (KLFs) as key transcriptional regulators of cortical maturation. Expression profiling showed a developmental switch from activating KLFs (Klf6, Klf7) expressed neonatally to repressive KLFs (Klf9, Klf13) upregulated during maturation. Using an elegant in vivo multiplexed CRISPR interference (CRISPRi) system, the authors achieved efficient, cell-type-specific knockdown of these TFs and showed that Klf9 and Klf13 repress a set of genes that includes cytoskeletal regulators such as Tubb2b, Dpysl3, and Rac3. Conversely, Klf6 and Klf7 promoted the expression of these same genes in the early postnatal period, and their knockdown led to reduced expression of these genes, particularly at P10 when their activating influence is strongest. Since promoters of shared KLF targets were enriched for KLF/Sp motifs but showed little change in chromatin accessibility, the authors propose a model in which distinct KLF family members function either as transcriptional repressors and activators that compete at constitutively accessible promoters and thereby act as a developmental transcriptional switch that coordinates the downregulation of axon growth programs and upregulation of synaptic maturation genes during cortical development. Strengths: The study addresses an interesting question and advances our understanding of the transcriptional regulation underlying postnatal cortical development. A major strength of the study lies in the innovative use of in vivo multiplexed CRISPR interference (CRISPRi), which allows for cell-type-specific, combinatorial knockdown of redundant TFs - this an elegant solution to a long-standing challenge in transcription factor research, and should be useful also for other neuroscience studies that require local and cell-type-specific gene loss-of-function. Also, the integration of RNA-seq and ATAC-seq across developmental time points provides a robust foundation for identifying direct targets of the KLF family, and the findings are reinforced by cross-species conservation and the identification of targets with clear neurodevelopmental relevance. Weaknesses: The major weakness of the study lies in its relatively narrow scope: the study focuses primarily on transcriptional mechanisms and largely lacks functional validation of the neuronal phenotypes that are predicted by the gene expression data (e.g. axonal morphology). For example, the authors analyzed the effects of KLF9/13 KD on the neurons' excitability and excitatory inputs, but did not assess the effects on inhibitory inputs and E/I-ratio or morphological parameters such as axonal length and axonal target fields - the manuscript would be strengthened considerably by such analyses (axonal projections could be analyzed e.g. via local injections of the gRNA AAVs and subsequent immunolabeling of brain sections). Similarly, the chromatin-based mechanisms underlying KLF activity remain relatively speculative, and the transcriptional mechanisms upstream of the KLFs remain unexplored (this could be addressed by analyzing existing datasets; see "Additional Point 1" below). Finally, the manuscript is too long (e.g., nearly five pages in the Discussion section are devoted to discussing various misregulated genes) and would benefit from presenting the Results and Discussion sections more concisely. However, despite these limitations, the paper offers an interesting model for a transcriptional switch during neuronal maturation in the cortex and establishes a powerful methodological framework for dissecting redundant gene networks in vivo. Shorten discussion (possibly results also). We appreciate the reviewer’s feedback and agree that linking transcriptional perturbations to cellular phenotypes of KLF activator and/or repressor KD would significantly strengthen this manuscript. However, we believe this is beyond the scope of this current manuscript and the effects of individual KLF activators and repressors on axon outgrowth/regrowth are known from prior in vitro and in vivo studies (Apara et al., 2017; Avila-Mendoza et al., 2020; Blackmore et al., 2012; Moore et al., 2009; Wang et al., 2018). Since there was no detectable effect of KLF repressor KD on excitatory synaptic transmission (Fig. S4.2), we did not believe it was likely that our excitatory neuron-specific knockdown would affect inhibitory synapses under our model where KLF targets have primarily axonal functions. In the absence of ATAC-seq from Klf9/13 KD neurons, the relationship between chromatin accessibility and KLF binding is limited to descriptions of chromatin accessibility around KLF targets in neonatal and mature neurons. However, we do offer two testable hypotheses in our Discussion of early developmental transitions driving the KLF switch (Thyroid Hormone and structured activity input). Finally, we recognize that the manuscript is too long and have edited or removed parts of the Discussion. Recommendations for the authors: Summary of Recommendations from Reviewing Editor: While the screen data is important and novel, there is consensus among the reviewers that the current study would greatly benefit by additional analyses as to the consequences of KLF knockdown on neuronal morphology and neurite outgrowth. It is recommended that this major open question be experimentally addressed. It is also recommended that additional comments/points below are addressed with changes to the figures/text as appropriate. Reviewer #1 (Recommendations for the authors): Minor points: (1) In Figure 2C2, it can be assumed that green represents KLF7 and red represents KLF9 but a legend indicating this would be helpful here. The Y axis is labeled as "Binned D-V Axis". Can the authors further clarify what this means? Is it referring to the different layers of the cortex? It looks like the expression pattern of KLF7 and 19 is not consistent throughout the entire dorsal-ventral axis at P7 and there is a region where KLF19 expression is higher than KLF7. Could this suggest that KLF family members may act on different types of neurons in distinct layers at different timepoints? Could the authors expand on this at all? We thank the reviewer for pointing this out and the appropriate legend has been added. A detailed description of our method for binning counts in the Dorsal-Ventral axis (y-axis) to normalize for differences in cortical thickness across ages is included in the methods section (see RNAScope Image Acquisition and Analysis), but text has been added to the results to clarify this. The non-uniform distribution of Klf7 across cortical layers at P2 and P7 likely indicates that that KLF switch, while conserved, occurs with variable kinetics across cortical cell types which could reflect their distinct rates of maturation in vivo (Gao et al., 2025). Minor edits have been made to the text associated with Figure 2 to call attention to this detail. (2) Figure 6C1 describes three clusters of DEGs. If I understand correctly, cluster 1 represents downregulated DEGs by KLF6/7 that are also developmentally downregulated between P10 and P20 and cluster 2 represents downregulated DEGs by KLF6/7 that do not change between P10 and P20. It could be interesting to see how these gene sets compare with one another. For example, are the genes in cluster 1 that decrease from P10 to P20 more related to axon regulation/cytoskeleton when compared to cluster 2 which consist of genes that do not change from P10 to P20 and perhaps may reflect other regulatory functions of KLF6/7. We appreciate this suggestion and we have updated Figure 6C1 to reflect this observation by highlighting genes that are part of significant enriched GO terms to allow the reader to see which cluster they belong to. Separate GO analyses performed for each downregulated gene cluster did not yield significant results and were therefore not included. Reviewer #3 (Recommendations for the authors): See above Additional points: (1) Upstream transcriptional regulation of KLFs: The idea that KLFs are regulated by Thyroid hormone (T3) is interesting, and the manuscript would be strengthened by exploring this idea further. This could be done, e.g. by analyzing existing snRNA-seq on genes regulated in cortical neurons by T3 (see PMID: 39178853). Similarly, KLFs were suggested to act together with AP1, and many KLFs were previously found to be activity-regulated - hence, the manuscript would be strengthened by analyzing the connection between KLFs and AP1. This could be done e.g. by re-analyzing transcriptomic and proteomic data generated e.g. by the Greenberg laboratory. We agree with the reviewer that the role of T3 in driving the KLF switch is intriguing and is an active area of investigation in our lab. However, these experiments are ongoing and are beyond the scope of this current manuscript. A preliminary analysis of the T3-treated snRNA-seq dataset present in Hochbaum et al., 2024 found that all major excitatory cortical cell types upregulate Klf9 upon T3 treatment while the expression of selected shared targets including Dpysl3 and Gap43 are downregulated in several cell types, supporting our hypothesis. These findings have been included in Figure S9. It was not our intention to suggest that there may be cooperation between AP1 and the KLF family, as AP1 motif enrichment was detected in shared upregulated genes while the KLF/Sp motif was found in shared downregulated genes. Furthermore, DEGs from both KLF knockdown experiments had exclusive enrichment for promoter KLF/Sp motifs, making it unlikely that there is widespread co-regulation of these genes by the AP1 family. (2) Figure 5D and Page 27, regarding Rac3: The authors state that " no change in accessibility in motif-positive peaks upstream of Plppr1 and Rac3 (Figures 5D and S5C)." This statement seems incorrect for Rac3 in L6 neurons where the ATAC signal is strongly reduced at P30 relative to P2. The authors might want to explain this or choose a better example. We acknowledge that Rac3 was a poor choice to represent a developmentally regulated Klf9/13 target with no change in promoter accessibility, and have updated Figure 5 with Atat1, a tubulin acetylase and shared KLF target, as our exemplar for this category of transcript. (3) Supplemental Figure 2: There seems to be a discrepancy between the expression levels in Panels A and C: the data in Panel A were generated from adult mouse cortex, and the levels of six KLFs are indicated as rather high - however, in Panel C, the adult levels of KLF6, 7 and 13 are rather low. How can this be explained? We thank the reviewer for detecting this apparent discrepancy. Most of this can be attributed to the log scale used in Figure S2A, which flattens expression differences in the moderate to high expression range. We elected to use a log scale in Figure S2A to highlight KLF family members with higher cortical expression relative to those with low or no expression. In addition, the counts displayed in this figure were batch-corrected to account for significant batch effects between libraries from additional excitatory cortical cell types included in Figure S2A, which were prepared and sequenced at different locations (Brandeis University v. Janelia Research Campus, as in Sugino et al., 2019). The additional processing step has been included in our methods section. (4) IGV plots in Figure 5D and Supplemental Figure 5B: It might be worth highlighting/shading the promoter region to see the position of the peaks relative to the promoters and TSSs. The TSS is indicated on all IGV plots by a dark arrow indicating the direction of transcription, so the promoter can be inferred to be immediately upstream. (5) Middle of page 20: Remove "saw"; this seems like a typo. (6) Page 27, reference to the plot for Plppr1: The plot for Plppr1 is in Supplemental Figure 5B2, not in S5C. (7) Page 28: There seems to be a typo/omission before "...general applicability of CRISPRi"