Deciphering the genetic interactions between Pou4f3, Gfi1, and Rbm24 in maintaining mouse cochlear hair cell survival

Mammals harbor a limited number of sound-receptor hair cells (HCs) that cannot be regenerated after damage. Thus, investigating the underlying molecular mechanisms that maintain HC survival is crucial for preventing hearing impairment. Intriguingly, Pou4f3-/- or Gfi1-/- HCs form initially but then rapidly degenerate, whereas Rbm24-/- HCs degenerate considerably later. However, the transcriptional cascades involving Pou4f3, Gfi1, and Rbm24 remain undescribed. Here, we demonstrate that Rbm24 expression is completely repressed in Pou4f3-/- HCs but unaltered in Gfi1-/- HCs, and further that the expression of both POU4F3 and GFI1 is intact in Rbm24-/- HCs. Moreover, by using in vivo mouse transgenic reporter assays, we identify three Rbm24 enhancers to which POU4F3 binds. Lastly, through in vivo genetic testing of whether Rbm24 restoration alleviates the degeneration of Pou4f3-/- HCs, we show that ectopic Rbm24 alone cannot prevent Pou4f3-/- HCs from degenerating. Collectively, our findings provide new molecular and genetic insights into how HC survival is regulated.

The similar but delayed phenotype of OHC death in Rbm24 mutants relative to that in Pou4f3-and Gfi1-deficient mice prompted us to dissect the potential genetic interactions among them by using an in vivo genetic approach.Our results showed that the onset of RBM24 expression was completely repressed in Pou4f3 -/-HCs, but that, unexpectedly, RBM24 expression was normal in Gfi1 -/-HCs.Moreover, the expression of neither POU4F3 nor GFI1 was altered in Rbm24 -/-HCs.Thus, POU4F3, but not GFI1, is required in Rbm24 expression.Furthermore, we identified three Rbm24 enhancers that were sufficient to drive specific EGFP reporter expression in HCs, and these enhancers are likely bound by POU4F3.Lastly, we found that restoration of RBM24 expression alone cannot alleviate the degeneration of Pou4f3 -/-HCs, which indicates that the expression of additional POU4F3-targeted genes must be restored to enable Pou4f3 -/-HCs to survive.Our study provides new insights into the genetic interactions among Pou4f3, Gfi1, and Rbm24, which hold potential applications in HC protection.

Results
RBM24 expression is completely repressed in Pou4f3 -/-cochlear HCs HC degeneration occurs in both Pou4f3 -/-and Rbm24 -/-mice, with the phenotype appearing earlier and being more severe in Pou4f3 -/-mice (Hertzano et al., 2004;Wang et al., 2021).This led us to speculate that genetic interaction exists between Pou4f3 and Rbm24.We reasoned that if POU4F3 is an upstream positive regulator of Rbm24, RBM24 expression would be downregulated in the absence of POU4F3.To rapidly test this possibility, we exploited our previously established CRISPRstop approach (Zhang et al., 2018).CRISPR-stop allows early stop codons to be introduced without inducing DNA damage through Cas9, which can cause deleterious effects (Kuscu et al., 2017).More importantly, the CRISPR-stop approach can generate Founder 0 (F0) mice carrying homozygous or mosaic homozygous gene mutations, and the F0 mice are thus immediately ready for phenotypic analysis, considerably faster than in the case with traditional gene-targeting methods (Wang et al., 2021;Zhang et al., 2018).Co-injecting one sgRNA (sgRNA-1) against Pou4f3 and base-editor components into one-cell-stage mouse zygotes yielded F0 mice, whose tail DNA was subject to Sanger sequencing (Figure 1-figure supplement 1A).Relative to wild-type (WT) mice (Figure 1-figure supplement 1B), the F0 mice with homozygous premature emergence of the stop codon TAG were defined as Pou4f3 -/-mice and were immediately ready for phenotypic analysis (Figure 1 Triple staining of POU4F3, RBM24, and INSM1 revealed that OHCs in WT mice at E16.5 were POU4F3+/RBM24+/INSM1+, whereas IHCs expressed POU4F3 and RBM24 but not INSM1 (Figure 1-figure supplement 1D-D''').Conversely, POU4F3 expression was absent in Pou4f3 -/- mice (Figure 1-figure supplement 1E-E'''), confirming that POU4F3 translation was blocked, and unlike in WT mice, RBM24 expression was undetectable in the Pou4f3 -/-mice at E16.5 (Figure 1figure supplement 1E-E''').Notably, INSM1 expression appeared normal in Pou4f3 -/-mice (arrows in Figure 1-figure supplement 1E-E''').Moreover, the presence of INSM1+ OHCs in the Pou4f3 -/-mice eliminated the possibility that the absence of RBM24 was due to a secondary effect of OHC death or delayed differentiation, and also agreed with previous reports that initial HC differentiation can occur without POU4F3 (Xiang et al., 1998;Hertzano et al., 2004).Thus, the rapid Pou4f3 loss-offunction analysis by using CRISPR-stop supported our hypothesis that POU4F3 is required for turning on RBM24 expression.

Construction of Gfi1-3×HA-P2A-Cre/+ knockin mouse strain
We sought to determine whether POU4F3 regulates Rbm24 expression through GFI1 or independently of GFI1.Because a suitable commercial GFI1 antibody for immunostaining was unavailable, we constructed a knockin mouse strain, Gfi1 3×HA-P2A-Cre/+ (Gfi1 HA-Cre/+ in brief), by using our routine CRISPR/Cas9 approach (Figure 2-figure supplement 1A-C); here, the GFI1 C-terminus was tagged with three hemagglutinin (HA) fragments and Cre expression was under the control of endogenous cis-regulatory elements (CREs) of Gfi1.The obtained WT (Gfi1 +/+ ), heterozygous (Gfi1 HA-Cre/+ ), and homozygous (Gfi1 HA-Cre/HA-Cre ) mice were readily distinguished using tail-DNA PCR (Figure 2-figure supplement 1D).Southern blotting revealed that in addition to being inserted in the Gfi1 locus, the targeting vector (Figure 2-figure supplement 1B) was randomly inserted in an unknown genomic region; however, the random insertion likely occurred in a silent genomic region as indicated by the analysis discussed below.

GFI1 expression is prevented in Pou4f3 -/-cochlear HCs
Because HA faithfully represented GFI1 expression in the Gfi1 HA-Cre/+ strain, we expected to observe repression of HA expression in Pou4f3 -/-HCs, as reported previously (Wallis et al., 2003;Hertzano et al., 2004).We confirmed this by using our CRISPR-stop approach (Zhang et al., 2018).The experimental pipeline was mostly identical to that used for producing the Pou4f3 mutants (Figure 1-figure supplement 1), except that the zygotes here were derived from male Gfi1 HA-Cre/+ mice (Figure 2 In control Gfi1 HA-Cre/+ mice at E16.5, all HCs expressed POU4F3, RBM24, and HA (GFI1), although the HA (GFI1) levels again appeared heterogenous among the OHCs (Figure 2-figure supplement 2D-D''').By contrast, in the Gfi1 HA-Cre/+ ; Pou4f3 -/-(mosaic) mice, we detected HCs that had either lost or maintained POU4F3 expression (Figure 2-figure supplement 2E-E''').Notably, POU4F3+HCs expressed HA (GFI1) and RBM24 (blue arrows in Figure 2-figure supplement 2E-E'''), whereas both HA (GFI1) and RBM24 were absent in HCs that had lost POU4F3 expression (orange arrows in               supplement 2E-E''').This again confirmed that HA is a reliable readout for GFI1 expression and that the expression is sensitive to the loss of POU4F3.Moreover, the results supported the view that POU4F3 regulates Rbm24 in a cell-autonomous manner.

GFI1 is dispensable for RBM24 expression
Next, we produced Gfi1 -/-mutants by using the same CRISPR-stop approach (Zhang et al., 2018).Onecell-stage zygotes derived from male Gfi1 HA-Cre/+ mice were injected with base-editor components and four different sgRNAs located in distinct exons of Gfi1 (Figure 2A).The reason for the combined use of four sgRNAs is detailed in the 'Discussion' section.With sgRNA-6, for example, Sanger sequencing of tail DNA revealed that relative to control Gfi1 HA-Cre/+ mice (black arrow in Figure 2B), F0 mice showed premature emergence of the TAG stop codon, which resulted in homozygous Gfi1 inactivation (red arrow in Figure 2C).All HCs expressed HA (GFI1) and RBM24 in control Gfi1 HA-Cre/+ mice (Figure 2D-D''), and, notably, RBM24 expression was maintained in Gfi1 -/-HCs, which showed no HA (GFI1) expression at E16.5 (Figure 2E-E'').This finding suggested that RBM24 expression does not depend on GFI1.
The advantage of this Gfi1 -/-model is that it allows rapid and direct confirmation of the absence of GFI1 expression in F0 mice.However, complete loss of GFI1 expression in all HCs cannot be guaranteed here at single-cell resolution through HA staining because only one allele of Gfi1 was Gfi1 HA-Cre .Thus, to eliminate the possibility that only the HA-tagged Gfi1 allele was mutated, we generated germlinestable Gfi1 mutants in which the majority of the Gfi1 DNA fragment between sgRNA-4 and sgRNA-9 was deleted (Figure 2-figure supplement 3A).The WT (Gfi1 +/+ ), Gfi1 +/-, and Gfi1 -/-mice were readily identified using tail-DNA PCR (Figure 2-figure supplement 3B).In agreement with previous findings (Matern et al., 2020), dual staining of RBM24 and POU4F3 revealed that relative to WT mice (Figure 2-figure supplement 3C-C''), Gfi1 -/-mice (Figure 2-figure supplement 3D-D'') showed severe degeneration of HCs at P1, particularly OHCs; this validated the successful generation of the Gfi1 -/-mice.Notably, the surviving POU4F3+ HCs maintained the expression of RBM24 (orange arrows in Figure 2-figure supplement 3D-D'').By contrast, we observed no significant difference between WT and Gfi1 -/-mice at E16.5: both WT and Gfi1 -/-HCs expressed POU4F3 and RBM24, and both WT and Gfi1 -/-OHCs expressed BCL11B (Figure 2-figure supplement 3E-F''').This suggests that the degeneration of Gfi1 -/-HCs does not begin by E16.5.The presence of POU4F3 in Gfi1 -/-HCs agreed with the previously mentioned notion that Pou4f3 is epistatic to Gfi1.Collectively, the results of analyses of both Gfi1 mutant models support the conclusion that GFI1 is dispensable for RBM24 expression.
The mosaic inactivation of RBM24 was further confirmed through triple staining of RBM24, HA, and POU4F3 (Figure 3D-E''').Relative to the expression in control Gfi1 HA-Cre/+ ; Rbm24 +/+ mice (Figure 3D-D'''), RBM24 expression disappeared in a fraction of cochlear HCs (orange arrows in Figure 3E-E''') but remained normal in other HCs (blue arrows in Figure 3E-E''') of the Gfi1 HA-Cre/+ ; Rbm24 -/-(mosaic) mice at E17. Notably, regardless of whether RBM24 expression was inactivated or not in the cochlear HCs, the expression patterns of both HA (GFI1) and POU4F3 remained intact.This suggests that Pou4f3 is epistatic to Rbm24 and that inactivation of the Rbm24 gene does not affect POU4F3 expression.Moreover, our results suggested that Gfi1 and Rbm24 do not interact genetically: Gfi1 inactivation did not affect Rbm24 expression (Figure 2, Figure 2-figure supplement 3), and vice versa (Figure 3E-E''').

GATA3 is downregulated and IKZF2 is upregulated in Rbm24-deficient HCs
We next determined whether the expression patterns of two additional genes, Gata3 and Ikzf2, which are involved in OHC development and survival (Li et al., 2023a;Chessum et al., 2018;Bi et al., 2022;Bardhan et al., 2019;Sun et al., 2021), were altered in the absence of RBM24.For analyzing GATA3, mosaic Rbm24 mutant mice were produced using the CRISPR-stop approach (Figure 3figure supplement 1A).Dual immunostaining of GATA3 and RBM24 revealed that GATA3 was evenly expressed in control OHCs and IHCs at P1 (Figure 3-figure supplement 1B-B'').By contrast, the GATA3 level in OHCs lacking RBM24 expression (white arrows in Figure 3-figure supplement 1C-C'') was weaker than that in neighboring OHCs that retained RBM24 expression (yellow arrows in Figure 3-figure supplement 1C-C'').This result suggests that GATA3 is downregulated in Rbm24deficient OHCs.OHCs degenerate in Gata3 +/- (Bardhan et al., 2019), and thus the downregulation of GATA3 might contribute to the cell death of Rbm24-deficient OHCs.

Three Rbm24 enhancers can drive specific EGFP expression in cochlear HCs
After showing that POU4F3, but not GFI1, is indispensable in mediating Rbm24 expression, we determined the mechanism by which POU4F3 controls Rbm24 expression, particularly the role of the CREs of Rbm24.According to our previous high-throughput ATAC-seq (transposase-accessible chromatin sequencing) analysis of neonatal cochlear HCs (Luo et al., 2022), four CREs of Rbm24 were identified: one proximal promoter (arrow in Figure 4A), and three distal potential enhancers that were defined as Eh1, Eh2, and Eh3 (dotted boxes in Figure 4A).Eh1 and Eh2 were located upstream and Eh3 was downstream of the Rbm24 coding region.Moreover, we reanalyzed the results of POU4F3 Cut&Run assays from one previous study (Yu et al., 2021) and found that POU4F3 binds to Eh1, Eh2, and Eh3 but not the Rbm24 promoter (Figure 4A).Whether Eh1, Eh2, and Eh3 are bona fide Rbm24 enhancers has remained unknown.We reasoned that if Eh1, Eh2, and Eh3 were Rbm24 enhancers, one of these, together with the mini-promoter of mouse heat shock protein 68 gene (Hsp68) (Luo et al., 2022;Kothary et al., 1989;Xu et al., 2021;Shu et al., 2022), would be sufficient to drive specific reporter expression in cochlear HCs, and to test this, we established three transgenic mouse strains: Eh1-EGFP+ (Figure 4B-C''), Eh2-EGFP+ (Figure 4D-E''), and Eh3-EGFP+ (Figure 4F-G''); in these strains, EGFP expression would be driven by the mini-promoter of Hsp68 and Eh1 or Eh2 or Eh3, respectively.The mini-promoter of Hsp68 alone is reported to be incapable of driving EGFP expression (Sun et al., 2022), and in contrast to this, we detected strong EGFP expression through whole-mount analysis in all three transgenic lines (Figure 4B, D, and F).Moreover, dual labeling for EGFP and MYO7A in cryosectioned cochleae showed that EGFP was specifically expressed in IHCs and OHCs in all three strains at P1 (Figure 4C-C'', E-E'', and G-G'').Collectively, our transgenic assay results suggest that Eh1, Eh2, and Eh3 are Rbm24 enhancers, and further that POU4F3 regulates Rbm24 expression primarily by binding to the Rbm24 enhancers.
Source data 2. File containing Figure 5B and the original agarose gel analysis with bands and sample labels.

Molecular mechanisms underlying cochlear HC survival
Mammalian sound-receptor HCs are vulnerable to various genetic mutations, environmental ototoxic factors, and aging.HC degeneration is one of the primary reasons for human sensorineural hearing impairment (Petit et al., 2023), and several genes whose mutations lead to HC degeneration starting at different ages have been identified previously.For example, HC development is severely defective in Atoh1 -/-mutants (Bermingham et al., 1999;Fritzsch et al., 2005), and Atoh1 -/-cochlear sensory cells undergo apoptosis.However, whether the absence of HCs in Atoh1 -/-mice is because of HCs not being produced or dying immediately after initial emergence remains unclear, partly due to the difficulty of unambiguously defining nascent HCs by using molecular markers.
Unlike in Atoh1 -/-mutants, in both Pou4f3 -/-and Gfi1 -/-mice (Xiang et al., 1998;Wallis et al., 2003;Hertzano et al., 2004), initial cochlear HC development is normal but becomes defective at perinatal ages, consistent with the notion that POU4F3 and GFI1 are dispensable for HC fate specification but necessary for subsequent differentiation and survival.Intriguingly, caspase-3 is active during HC death in Pou4f3 -/-mutants, and the antiapoptotic factor z-VAD-fmk exerts a protective effect on Pou4f3 -/- HCs between E14.5 and E16.5 (Atar and Avraham, 2010).Moreover, we noted here milder overall HC degeneration in Gfi1 -/-mice than in Pou4f3 -/-mice at P1 because only OHC degeneration occurred in Gfi1 -/-mice but both IHCs and OHCs were degenerated in Pou4f3 -/-mice (Figure 1 Gfi1 is a recognized target of POU4F3 (Hertzano et al., 2004), but whether forced GFI1 expression can alleviate the degeneration of Pou4f3 -/-HCs is unknown.Similarly, Pou4f3 is one of the target genes regulated by ATOH1 (Yu et al., 2021), but whether forced expression of POU4F3 or GFI1, or both, can mitigate the developmental defects of Atoh1 -/-HCs remains undetermined.Future studies must address these questions to enable comprehensive understanding of the mechanisms underlying the maintenance of cochlear HC survival.

Roles of RBM24 and its regulation during cochlear HC development
Shortly after cochlear HCs emerge, RBM24 expression begins and then is maintained permanently, as revealed by transcriptomic analyses and antibody staining (Bi et al., 2022;Grifone et al., 2018;Sun et al., 2021;Liu et al., 2022a;Ranum et al., 2019).RBM24 is not necessary in the early phase of cochlear HC development because Rbm24 -/-HC development is normal by P1, but RBM24 is required for OHC survival after birth (Wang et al., 2021); Rbm24 -/-OHCs, but not IHCs, are degenerated by P19 (Wang et al., 2021).RBM24 is also involved in mRNA stability and pre-mRNA alternative splicing of several genes, including Cdh23 and Pcdh15, which are crucial for the development of HC stereocilia (Wang et al., 2023;Zheng et al., 2021;Liu et al., 2022b).Our study also revealed that GATA3 expression is decreased and IKZF2 expression is increased in OHCs in the absence of RBM24, which could be a direct or indirect effect of RBM24 loss.Nonetheless, single-cell transcriptomic analysis has not yet been performed on Rbm24 -/-IHCs or OHCs, and future investigation is necessary to clarify the detailed molecular mechanism underlying Rbm24 -/- HC degeneration.
What are the trans-acting factors involved in regulating Rbm24 expression?First, this study has provided strong genetic evidence indicating that POU4F3 is necessary for turning on Rbm24 expression; in the absence of POU4F3, Rbm24 expression was not triggered.Moreover, the normal expression of POU4F3 in Rbm24 -/-HCs confirmed that Pou4f3 is epistatic to Rbm24 and that RBM24 is dispensable for Pou4f3 expression.Besides POU4F3, ATOH1 appears to regulate Rbm24 expression, as per two lines of evidence: (1) RBM24 expression is lost in Atoh1 -/-cochlear HCs (Cai et al., 2015) and (2) Rbm24 is one of the ATOH1-binding targets revealed by the ATOH1 Cut&Run assay (Luo et al., 2022;Yu et al., 2021).Notably, ATOH1 binds to both the Rbm24 promoter and the three Rbm24 enhancers (Eh1, Eh2, and Eh3) (Luo et al., 2022;Yu et al., 2021), whereas POU4F3 only binds to the Rbm24 enhancers (Figure 4).Thus, POU4F3 and ATOH1 likely cooperate to regulate Rbm24 expression, and either Pou4f3 or Atoh1 mutation leads to repression of Rbm24 expression.Future conditional Pou4f3 loss-of-function studies are necessary to determine whether Rbm24 expression in adult HCs requires POU4F3.

Gfi1 and Rbm24 are expressed independently of each other
We initially hypothesized that POU4F3 regulates Rbm24 expression through GFI1.However, this hypothesis was not supported by our observation that RBM24 expression is normal in Gfi1 -/-HCs (Figure 2, Figure 2-figure supplement 3).Moreover, in the absence of RBM24, GFI1 expression was not altered in cochlear HCs.Thus, Rbm24 expression and Gfi1 expression appear to be independent of each other.This might be due to the functional difference between POU4F3 and GFI1 during cochlear HC development.Although both POU4F3 and GFI1 are necessary for promoting the expression of genes involved in HC differentiation (Xiang et al., 1997;Erkman et al., 1996;Hertzano et al., 2004;Matern et al., 2020), GFI1, but not POU4F3, also represses the preceding expression of neural genes in nascent HCs (Matern et al., 2020).
Another observation here relevant to Gfi1 isoforms is noteworthy.It is known that one efficient sgRNA is adequate for inducing homozygous gene inactivation by using CRISPR-stop (Wang et al., 2021).Here, we used four Gfi1 sgRNAs distributed across distinct exons (Figure 2A).Interestingly, we successfully established one Gfi1 mutant by using sgRNA-5 alone, and this, in principle, should effectively pre-stop GFI1 translation in exon 1, which codes for the SNAG repressor domain; the obtained Gfi1 mutant presented the HC degeneration phenotype, although HA (GFI1) remained detectable in HCs.This agrees with the notion that the Gfi1 mutant lacking the SNAG domain is equivalent to the Gfi1-null model (Fiolka et al., 2006).Thus, Gfi1 is likely expressed as multiple unknown isoforms, many of which might be recognized by the HA-tag antibody because HA is tagged to the last exon (exon 6).However, when we used the four Gfi1 sgRNAs together, we were able to both reproduce the HC degeneration phenotype and completely block HA (GFI1) expression.
Why does ectopic RBM24 fail to alleviate the degeneration of Pou4f3 -/- HCs?
Upon confirming the epistatic genetic interaction between Pou4f3 and Rbm24, we predicted that ectopic RBM24 expression should be able to, at least partly, alleviate the degeneration of Pou4f3 - /-HCs, similar to what we recently reported in our study on INSM1 and IKZF2 in cochlear OHCs (Li et al., 2023a).However, RBM24 failed to rescue the HC degeneration in Pou4f3 -/-mice.Two potential explanations for this finding are the following: (1) although Rbm24 expression is modulated by POU4F3, RBM24 is not directly involved in the pathway regulating HC survival; accordingly, cell death occurs later in Rbm24 -/-HCs than in Pou4f3 -/-HCs.(2) Rbm24 is a key POU4F3 downstream target, but forced expression of RBM24 alone cannot compensate for the loss of POU4F3 or other known POU4F3 targets such as orphan thyroid nuclear receptor Nr2f2 and Caprin-1 (Tornari et al., 2014;Towers et al., 2011); CAPRIN-1 is recruited to stress granules in cochlear HCs exposed to ototoxic trauma (Towers et al., 2011).

Potential application of the three Rbm24 enhancers in cochlear HC gene therapy
Gene therapy is a promising strategy for restoring hearing capacity in humans with inherited gene mutations causing hearing impairment (Petit et al., 2023), and a few such therapy examples have been reported to date, including Otoferlin and vGlut3 gene-replacement therapies (Akil et al., 2019;Tang et al., 2023;Al-Moyed et al., 2019;Akil et al., 2012).Currently, therapeutic cDNAs are primarily delivered into HCs by using an adeno-associated virus (AAV) vector.Although several AAVs have been reported (Tan et al., 2019;Wu et al., 2021), the vectors transfect cochlear cells non-selectively and HC-specific AAVs are not yet available.A solution to this problem is the following: instead of the CAG/CMV ubiquitous promoter widely used in current AAVs (Isgrig et al., 2019;Landegger et al., 2017), using any one of the three Rbm24 enhancers together with the Hsp68 mini-promoter (in brief Rbm24-Hsp68) should generate an AAV that would allow HC-specific transfection.Moreover, Rbm24 is permanently expressed in cochlear HCs, and if Rbm24-Hsp68 AAV acts as expected, it would represent a powerful tool for future HC-specific gene therapy at all postnatal ages, with potential applications in treating clinical human deafness.

Mice
The Atoh1 Cre/+ model was kindly provided by Dr. Lin Gan (Augusta University, USA).The Rosa26-loxpstop-loxp-tdTomato (Ai9)/+ strain (Jax#: 007905) was from The Jackson Laboratory.The Ikzf2 V5/+ mouse strain is described in detail in our previous reports (Li et al., 2023a;Bi et al., 2022).All mice were bred and raised in an SPF-level animal room, and all animal procedures were performed according to the guidelines (NA-032-2022) of the Institutional Animal Care and Use Committee (IACUC) of the Institute of Neuroscience (ION), Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences.
One-step generation of homozygous Pou4f3, Gfi1, or Rbm24 mutants by using CRISPR-stop The detailed protocol for using CRISPR-stop to generate homozygous gene mutants is described in our previous reports (Wang et al., 2021;Zhang et al., 2018).Briefly, efficient pre-tested sgRNAs (Supplementary file 1) and hA3ABE3 were co-injected into one-cell-stage mouse zygotes that were then transplanted into pseudopregnant female mice, which gave birth to F0 mice.The F0 mice carrying the expected homozygous mutation (pre-emergence of protein translation stop codon) were identified using Sanger sequencing of tail-DNA PCR samples and were immediately ready for analysis.
Generating Pou4f3-or Gfi1-null mutants harboring large DNA fragment deletions by using CRISPR/Cas9 To construct germline-stable null mutants of either Pou4f3 or Gfi1, Cas9 mRNA, two efficient pretested sgRNAs located at the proximal and distal ends of the targeted gene, and a single-stranded DNA donor (120 bp) (Supplementary file 2) were co-injected into one-cell-stage WT zygotes.Notably, the left and right halves (60 bp each) of the single-stranded DNA donor were homologous to the 5ʹ and 3ʹ ends of the targeted gene, respectively.The post-injected zygotes were transplanted into pseudopregnant females, which gave birth to F0 mice; the F0 mice were subject to tail-DNA PCR screening with the primers listed in Supplementary file 2, and the mice harboring the designed large DNA deletion between the two sgRNAs were identified and further bred with WT mice to establish the germline-stable mutants (F1 or afterward).
Construction of Gfi1 3◊HA-P2A-Cre knockin mouse strain An sgRNA against Gfi1 (5ʹ-ATGG ACTC AAAT GAGT ACCC -3ʹ), Cas9 mRNA, and the targeting vector (Figure 2-figure supplement 1B) were co-injected into one-cell-stage WT mouse zygotes.The targeting vector comprised three portions: the 5ʹ homologous arm (800 bp), the 3ʹ homologous arm (800 bp), and the region between the 5ʹ and 3ʹ arms that contained three HA fragments followed by 2A-Cre.The F0 mice with the potential gene targeting (Figure 2-figure supplement 1C) were screened using tail-DNA PCR and then crossed with WT mice to produce F1 mice; these F1 mice were confirmed using tail-DNA PCR again and further screened using Southern blotting.The detailed Southern blotting protocol is described in our previous report (Li et al., 2018).Tail-DNA PCR was used for routine genotyping, and the knockin (Gfi1 HA-Cre/+ ) and WT alleles were distinguished using the primers F4, R5, and R6 (Supplementary file 3).
Construction of Eh1-EGFP+, Eh2-EGFP+, and Eh3-EGFP+ transgenic reporter lines The three transgenic reporter mouse lines used here, Eh1-EGFP+, Eh2-EGFP+, and Eh3-EGFP+, were produced using the same procedures.The core DNA sequences of each enhancer (Eh1, Eh2, or Eh3 in Figure 4A), the mRNA encoding PiggyBac transposase, and the PiggyBac vector (Figure 4B, D, and F) were co-injected into one-cell-stage WT mouse zygotes.The PiggyBac vector contained the Eh1/Eh2/Eh3 core DNA sequence (Supplementary file 2), the mini-promoter of mouse Hsp68, and the EGFP coding sequence, and the vector was randomly integrated into the mouse genome by the transposase.The F0 mice harboring the PiggyBac vector were screened using tail-DNA PCR and further bred with WT mice to produce germline-stable F1 transgenic reporter strains.The samples analyzed in Figure 4B-G'' were from F1 or later generations.The primers used for genotyping transgenic reporter strains were F8, F9, F10, and the common primer R11 (primer sequences are listed in Supplementary file 3).

Cell quantification and statistical analysis
Before immunofluorescence staining, each cochlear sample was grossly separated into three portions of distinct lengths, and each portion of the same cochlea was initially scanned using a confocal microscope at low magnification (×10 lens).After calculating the total length of each cochlea, the cochlear sample was precisely divided into basal, middle, and apical turns of equal length.Subsequently, for cell counting in experiments (Figure 1-figure supplement 2D), an ~200 μm stretch of the sensory epithelium in each turn was scanned using a confocal microscope at high magnification (×60 lens) and the number of HCs was determined.In Pou4f3 -/-cochleae, the percentage of surviving HCs was calculated by normalizing the number of remaining HCs against their counterparts in WT mice.All cell numbers are presented as means ± SD.For statistical analyses, we used GraphPad Prism 6.0 software and performed Student's t-tests with Bonferroni correction.

Institutional review board statement
The animal study protocol (NA-032-2022) was approved by the Institutional Animal Care and Use Committee (IACUC) of the Institute of Neuroscience (ION), Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences.

Figure supplement 2
Figure supplement 2-source data 1.File containing raw counts of the cochlear hair cells reported in the graph of Figure 1-figure supplement 2D.

Figure 2 .
Figure 2. GFI1 is dispensable for RBM24 expression.(A) Simple cartoon depicting how Gfi1 was inactivated in one-cell-stage zygotes derived from male Gfi1 HA-Cre/+ mice by using the CRISPR-stop approach.(B, C) Using sgRNA-6 as an example, Sanger sequencing chromatograms of control Gfi1 HA- Cre/+ (B) and F0 Gfi1 -/-(C) samples are presented.The base 'C' (black arrow in B) was converted to 'T' (red arrow in C), resulting in pre-emergence of the translation stop codon TAG.The red 'T' appears as a single peak, indicating that the stop codon pre-emerged in both alleles.(D-E'') Dual staining of HA (GFI1) and RBM24 in cochleae from control Gfi1 HA-Cre/+ (D-D'', n = 3) and F0 Gfi1 -/-(E-E'') mice (n = 4) at E16.5.RBM24 expression (E') was not altered in the absence of HA (GFI1) expression (E).OHC: outer hair cell; IHC: inner hair cell.Scale bar: 20 μm (E'').The online version of this article includes the following source data and figure supplement(s) for figure 2:

Figure supplement 1
Figure supplement 1-source data 2. File containing Figure 2-figure supplement 1D and the original agarose gel analysis with highlighted bands and sample labels.

Figure supplement 3 .
Figure supplement 3. RBM24 expression is normal in the absence of Gfi1.

Figure
Figure supplement 3-source data 2. File containing Figure 2-figure supplement 3B and the original agarose gel analysis with highlighted bands and sample labels.

Figure
Figure2-figuresupplement 2E-E''').This again confirmed that HA is a reliable readout for GFI1 expression and that the expression is sensitive to the loss of POU4F3.Moreover, the results supported the view that POU4F3 regulates Rbm24 in a cell-autonomous manner.