Introduction

A central challenge in evolutionary biology is to understand how a limited number of ancient conserved genes generate tremendous cellular diversity through functional redeployment^1–3. This challenge is particularly evident in the roles of master transcription factors, which must be precisely regulated across different cellular contexts to control specific developmental programs^4–7. Such functional plasticity is thought to arise through evolutionary co-option, the process by which existing regulators are recruited into new regulatory modules to drive lineage-or organelle-specific innovations^8–10. Elucidating the molecular strategies underlying this repurposing is essential for understanding how complex biological systems evolve. The biogenesis of the cilium—an evolutionarily ancient organelle that plays key roles in signaling and motility^11–14, provides an important model for investigating this fundamental question, and its dysfunction can lead to a class of genetic disorders known as ciliopathies¹⁵.

The assembly of the cilium, or ciliogenesis, requires the precise transcriptional coordination of hundreds of genes^16–18. In metazoans, the regulatory factor X (Rfx) and Forkhead box protein J1 (Foxj1) transcription factor families have been identified as key regulators of ciliogenesis^19–22. Rfx factors, such as DAF-19 in C. elegans, regulate the expression of genes essential for building the core ciliary structure²³, while Foxj1 directs the motile ciliogenic program^24,25. However, this well-established regulatory network primarily applies to motile, multiciliated cells. How the upstream, cell-type-specific transcriptional programs are initiated, particularly for the non-motile primary cilium found on most somatic cells, remains poorly understood^12,26.

The conserved transcription factor Xap5 serves as a prime example of this challenge, presenting a compelling biological paradox. Our previous work showed that Xap5 is a critical regulator for the assembly of motile appendages—dynamic structures that generate movement by beating in fluid environments—a role conserved from the cilia of the unicellular alga Chlamydomonas to the flagella of mammalian sperm^27,28. These findings cemented Xap5’s role in an ancient program for motility but, in doing so, created a profound biological enigma: why is this regulator, seemingly dedicated to motility, highly expressed in somatic cells that exclusively form non-motile primary cilia for sensory functions? Whether Xap5 plays a role in primary ciliogenesis in these cells, and the mechanism by which it might be co-opted for this distinct function, remains to be determined²⁹.

In this study, we specifically address these questions by characterizing the function of Xap5 in primary ciliogenesis using NIH/3T3 mouse fibroblasts, a standard model for primary cilium formation upon serum starvation³⁰. Here, we identify the Non-POU domain-containing octamer-binding protein (Nono) as a critical interacting partner of Xap5. Our integrated analyses demonstrate that Xap5 recruits its cofactor Nono to regulate primary cilia formation by modulating a downstream transcriptional axis involving Sox5 and Sox9. While Sox9 has been previously implicated in primary ciliogenesis^31–33, our findings for the first time identify Xap5 as a critical upstream activator of these transcription factors in somatic cells. Together, these findings define the complete Xap5-Nono-Sox5/9 axis for primary ciliogenesis. Moreover, they offer a powerful illustration of evolutionary co-option in action, establishing a mechanistic framework for how ancient transcription factors are functionally repurposed to drive cellular innovation.

Results

Xap5 is essential for primary cilia formation

To investigate the role of Xap5 in primary cilia formation, we first determined its subcellular localization in NIH/3T3 cells. Immunofluorescence staining with a specific anti-Xap5 antibody revealed that endogenous Xap5 is predominantly nuclear (Fig. 1a), consistent with our previous work. We confirmed this finding via ectopic expression of a Flag-tagged Xap5 construct (Fig. 1b). Consistent with its nuclear localization, bioinformatic analysis predicted a putative nuclear localization signal (NLS) within the Xap5 protein (Supplementary Fig. 1a). To functionally validate this sequence, we generated a site-directed mutant (Flag-Xap5-ΔNLS) (Supplementary Fig. 1a). Crucially, this mutation abolished the nuclear import of Xap5, leading to its mislocalization in the cytoplasm and confirming that this NLS is essential for its function (Fig. 1b).

Fig. 1

To investigate the function of Xap5, we generated Xap5 knockout (KO) NIH/3T3 cell lines using the CRISPR/Cas9 system³⁴ (Supplementary Fig. 1b). The complete depletion of Xap5 protein product was confirmed by western blotting (Fig. 1c). Upon inducing ciliogenesis via serum starvation (Fig. 1d), we first observed a striking defect in ciliary length (Fig. 1e). Although comparable to wild-type (WT) cells before starvation, Xap5 KO cells exhibited a significant shortening of their primary cilia post-starvation (Fig. 1e, f). To determine if this length defect stemmed from an overall failure in ciliogenesis, we next quantified the ciliation frequency. Interestingly, both WT and Xap5 KO cells responded robustly to the starvation stimulus, showing a comparable increase in the percentage of ciliated cells over 48 h (Fig. 1g). These results collectively demonstrate that Xap5 is dispensable for the initiation of ciliogenesis but plays an essential role in the subsequent elongation and/or maintenance of the ciliary axoneme.

Xap5 forms a nuclear regulatory module with its cofactor Nono

Having established the essential role of Xap5 in primary ciliogenesis, we next sought to uncover its underlying regulatory mechanism by identifying its synergistic cofactors. An unbiased immunoprecipitation-mass spectrometry (IP-MS) screen for Xap5 interactors identified several nuclear proteins (Supplementary Fig. 2), among which we prioritized Non-POU domain-containing octamer-binding protein (Nono), given its well-established, pleiotropic roles in transcriptional regulation and RNA processing^35–37.

As a prerequisite for their interaction, we first confirmed that Nono is also a predominantly nuclear protein in NIH/3T3 cells, a localization we validated for both the endogenous protein via immunofluorescence and an ectopically expressed Nono-GFP fusion protein (Fig. 2a). To biochemically validate their physical association, we next performed endogenous co-immunoprecipitation (Co-IP) assays. These results confirmed a robust and specific interaction, as immunoprecipitation of Xap5 specifically co-precipitated Nono (Fig. 2b). Taken together, these localization and co-immunoprecipitation data provide strong evidence that Xap5 and Nono form a bona fide regulatory complex within the nucleus.

Fig. 2

The Xap5-Nono module is functionally required for ciliogenesis

Given that Nono forms a complex with the essential ciliogenesis factor Xap5, we hypothesized that the entire module is functionally indispensable. To test this, we generated Nono KO NIH/3T3 cells using the CRISPR/Cas9 system (Supplementary Fig. 3a). Indeed, upon serum starvation, Nono KO cells phenocopied the Xap5 KO. They exhibited significantly shorter primary cilia compared to wild-type (WT) controls, yet displayed a comparable increase in ciliation frequency (Fig. 3b-d). Notably, in addition to the ciliary defects, we observed a significant increase in nuclear size in Nono KO cells (Fig. 3b; Supplementary Fig. 3b). Together, these findings demonstrate that Nono, like its partner Xap5, is dispensable for ciliary initiation but essential for ciliary elongation and maintenance, and may also have a broader role in maintaining nuclear architecture.

Fig. 3

The Xap5-Nono module co-regulates a downstream ciliogenic transcriptional cascade

To determine if the Xap5-Nono module cooperates at the transcriptional level, we performed RNA-sequencing (RNA-seq) on WT, Xap5 KO, and Nono KO cells following 24 h of serum starvation, a time point where the phenotype was robustly established. This analysis revealed widespread transcriptional changes; Xap5 KO cells showed 1,116 upregulated and 1,071 downregulated genes, while Nono KO cells had 1,425 upregulated and 1,042 downregulated genes relative to WT (fold change ≥ 1.5; Supplementary Fig. 4a, b; Supplementary Data 1, 2).

Given that loss of either Xap5 or Nono impairs ciliogenesis, we hypothesized that they function together as transcriptional activators of a shared ciliogenic program. We therefore focused our subsequent analysis on the genes commonly downregulated in both knockouts. Notably, Gene Ontology (GO) analysis of these genes revealed a highly significant enrichment for ciliary categories (Fig. 4a, b; Supplementary Fig. 4c, d). To investigate this further, we compared this downregulated gene set against a comprehensive ciliary gene list (Supplementary Data 3), which was assembled from the SYSCILIA gold standard (SCGSv2)¹⁸ and manual literature curation. This comparison confirmed a significant overlap between the two gene sets (Fig. 4c). We validated several of these shared targets—including the transcription factors Sox5 and Sox9, and the ciliary component Pkd2—by qPCR (Fig. 4d, e).

Fig. 4

Among this cohort of downregulated genes, two transcription factors, Sox5 and Sox9, stood out. Sox9 is a known regulator of primary cilia in other contexts, and both factors are known to cooperate in developmental programs. This made them compelling candidates for further investigation as potential downstream effectors in our system. Consistent with this hypothesis, western blot analysis confirmed that the protein levels of both Sox5 and Sox9 were markedly reduced in Xap5 and Nono KO cells (Fig. 4f). These data strongly indicate that Xap5 and Nono co-regulate a transcriptional network that includes the downstream effectors Sox5 and Sox9, thereby driving primary cilia formation.

The Xap5-Nono module directly co-occupies the regulatory regions of Sox5 and Sox9

To determine if Xap5 and Nono directly regulate their target genes, we mapped their genome-wide chromatin occupancy using CUT&Tag³⁸ in NIH/3T3 cells (Supplementary Data 4, 5). This analysis revealed a strong enrichment of both factors at promoter regions; approximately 30% of Xap5 peaks and a striking 50% of Nono peaks were located at promoters (Supplementary Fig. 5a, b). Consistent with this, genomic heatmaps demonstrated that both Xap5 and Nono binding signals were highly enriched around the transcription start sites of their target genes (Fig. 5a, b; Supplementary Fig. 5c, d).

Fig. 5

Integrating these binding profiles with our RNA-seq data revealed a strong correlation between genes bound by Xap5 or Nono and those whose expression was altered upon KO (Fig. 5c, d), supporting a model of direct transcriptional regulation. De novo motif analysis revealed that Xap5 and Nono recognize a highly similar DNA-binding motif (Fig. 5e). Importantly, this Xap5 motif is virtually identical to that from our previous work in testicular germ cells (Fig. 5e), demonstrating that Xap5 maintains its DNA-binding specificity across distinct cellular contexts.

Crucially, we found a significant overlap between the Xap5/Nono co-occupied genes and our curated ciliary gene set (Fig. 5f). Among these, we observed direct co-occupancy of Xap5 and Nono at the promoter regions of the key downstream regulators Sox5 and Sox9, a finding we validated by CUT&Tag-qPCR (Fig. 5g). Collectively, these genomic data provide strong evidence that Xap5 and Nono form a complex that directly binds to the regulatory regions of a common set of ciliary genes, including Sox5 and Sox9, to orchestrate primary cilia formation.

Sox5, a key downstream effector of the Xap5-Nono pathway, is essential for ciliogenesis

Our finding that Xap5 and Nono directly co-regulate Sox5 and Sox9 prompted us to investigate their respective roles in ciliogenesis. Given that the essential role of Sox9 in maintaining primary cilia length is well-documented^31,33, we focused our functional analysis on the uncharacterized role of Sox5. To this end, we generated Sox5 KO NIH/3T3 cells using the CRISPR/Cas9 system (Fig. 6a; Supplementary Fig. 6).

Fig. 6

Analysis after serum starvation revealed that Sox5 KO cells precisely phenocopied the upstream knockouts. Sox5 KO cells exhibited significantly shorter primary cilia than WT controls (Fig. 6b, c), yet this occurred despite a normal ciliation response, with both genotypes showing a robust and comparable increase in ciliation frequency (Fig. 6d). These results establish Sox5 as a critical effector responsible for ciliary elongation. Collectively, our data on Sox5, combined with the established literature on Sox9, strongly support a model where the Xap5-Nono complex acts through two essential downstream transcription factors, Sox5 and Sox9, to orchestrate the maintenance of primary cilia.

Sox5 executes a shared ciliogenic program downstream of the Xap5-Nono complex

To confirm that Sox5 functions transcriptionally downstream of the Xap5-Nono complex, we performed RNA-seq on Sox5 KO and WT cells after 24 h of serum starvation. The analysis identified 2,124 upregulated and 2,105 downregulated genes in Sox5 KO cells (fold change ≥ 1.5; Supplementary Fig. 7a; Supplementary Data 6). To assess the global impact on ciliary pathways, we first performed Gene Set Enrichment Analysis (GSEA)³⁹ on all ranked differentially expressed genes, which revealed a significant enrichment of the “Cilium organization” pathway (Fig. 7a). Consistent with this, GO analysis of the strictly defined downregulated gene set (fold change ≥ 1.5) also highlighted terms related to cilium assembly (Fig. 7b; Supplementary Fig. 7b). Furthermore, direct comparison showed a significant overlap between the full set of genes dysregulated by Sox5 loss and our curated ciliary gene list (Supplementary Fig. 7c), reinforcing the strong link between Sox5 and ciliogenesis.

Fig. 7

Critically, to place Sox5 within the Xap5-Nono pathway, we compared the downregulated gene sets across all three KO lines. A Venn diagram revealed a significant and substantial overlap among them (Fig. 7c). We confirmed this shared transcriptional dependency by qPCR, validating that several key ciliary genes—including Adcy3, Ccdc39, and Pkd2—were indeed downregulated across all three KO models (Fig. 7d). These results provide strong evidence that Sox5 acts as a key transcriptional mediator downstream of Xap5 and Nono, executing a common program of ciliary gene expression required for proper cilia elongation.

Discussion

The assembly of the cilium, a highly conserved organelle, requires a sophisticated and hierarchical transcriptional program that is responsive to extracellular cues and precisely adapted to generate enormous functional diversity^{13,19–22,26}. The transcriptional program’s complexity is evident in its role in regulating flagella in unicellular organisms like Chlamydomonas reinhardtii and orchestrating the synchronous formation of hundreds of motile cilia in the mammalian airway^14,40–43. In mammals, a core ciliogenic program is controlled by the Rfx family, which act as master architects for the core ciliary structure^20,44–46. Built upon this foundation are multi-tiered regulatory cascades, often initiated by morphogenetic signaling pathways, that deploy cell-type-specific “selector cassettes”^14,20. For the highly specialized motile cilia, this cascade involves key initiators like GMNC/MCIDAS (antagonized by GMNN)^47–50, integrates context-specific factors like NOTO^51,52, and converges on downstream effectors TAp73/MYB^53–55, which in turn activate the master regulator Foxj1 to execute the motile program^24,25. While this elegant hierarchy explains the formation of multiciliated cells, how upstream signaling pathways initiate the transcriptional programs for the vast diversity of primary cilia remains a key unanswered question.

Central to this complexity is the ancient, evolutionarily conserved transcription factor Xap5. Xap5 regulates flagellar assembly in Chlamydomonas reinhardtii, and is widely expressed in mammalian tissues^27,56,57. Our recent work revealed that during motile ciliogenesis in spermatogenesis, Xap5 functions upstream of Rfx2 and Foxj1 to drive the sperm flagellar program²⁸. This finding solidified Xap5’s role as a key regulator for motility but also sharpened the central paradox addressed in this study: how does this same factor, which utilizes the established motile cilia pathway in germ cells, function in somatic cells that build exclusively non-motile primary cilia?

In this study, we resolve this fundamental paradox by uncovering a regulatory strategy that enables Xap5’s functional switch. In our somatic cell model, Xap5 is repurposed through recruitment of the nuclear protein Nono, forming an Xap5–Nono regulatory module that promotes primary ciliogenesis by directly engaging downstream Sox transcriptional effectors. This work provides a molecular blueprint for how a deeply conserved master regulator can achieve functional diversity through context-dependent cofactor assembly, while ultimately converging on core organelle biogenesis programs.

Specifically, we find that the Xap5–Nono module directly co-regulates Sox5 and Sox9, positioning them as downstream transcriptional effectors in primary ciliogenesis. The essential role of Sox9 in promoting primary cilia formation is well-established in multiple developmental contexts^31–33. Significantly, the finding that Xap5-Nono directly co-regulates Sox9 and Sox5 nominates Sox5 as a previously unrecognized, critical player in primary ciliogenesis. This organization exemplifies the “selector cassette” concept²⁰, in which context-specific deployment of transcription factors modifies the execution of a conserved program; we therefore propose that the context-dependent deployment of Sox5 and Sox9 downstream of Xap5–Nono functions as such a cassette. This model is supported by evidence that Sox5 can cooperate with Sox9 in other developmental programs (e.g., chondrogenesis⁵⁸) and that Sox5 has been linked to ciliogenic gene regulation in motile contexts^59,60. Together, these findings position Xap5–Nono as an upstream integrator that co-opts two related Sox factors to regulate primary ciliogenesis.

Our results suggest an evolutionary framework in which the DNA-recognition properties of Xap5 function as a conserved genomic anchor, while biological outputs diversify through context-dependent regulatory logic. Intriguingly, the GC-rich Xap5 binding motif identified here is nearly reverse-complementary to the CT-rich C-box core (5’-CCTCC-3’) that functions as a DAF-19/RFX-dependent enhancer in C. elegans ciliome promoters^23,45,61. Although direct binding of Xap5 family members to C-box elements remains to be demonstrated, this sequence-level complementarity, together with the deep conservation of the Xap5 family^27,56,57, supports the hypothesis that Xap5 could contribute to C-box–like regulatory architectures in distinct lineages. If so, it would be consistent with a broader model of evolutionary co-option^1–3,8–10, in which the core DNA-binding interface of Xap5 remains relatively constrained while regulatory outcomes may be reprogrammed via partner switching—for example, potential integration into RFX-driven X-box/C-box promoter modules in nematodes^23,45,61, versus association with Nono in mammalian somatic cells. This “rigid anchor, flexible cofactor” model provides a plausible route to expand regulatory complexity while preserving constraints imposed by DNA-binding specificity.

Collectively, the discovery of this adaptable, Xap5-driven regulatory network offers new perspectives on human ciliopathies¹⁵. Given that Sox9 dysfunction is linked to cystic diseases with ciliary defects^31,32, disruption of the upstream Xap5–Nono module represents a plausible, yet unexplored, pathogenic mechanism. As upstream regulators of primary ciliogenesis continue to emerge—including the metazoan-specific SP5/SP8 factors^62,63—our work highlights a distinct regulatory layer rooted in an evolutionarily ancient transcription factor^27,56,57. Notably, systemic loss of Xap5 causes embryonic lethality prior to E7.5²⁸, indicating essential functions that likely extend beyond ciliary regulation alone. One possibility is that Xap5 establishes a permissive transcriptional state that enables (or constrains) the execution of ciliogenic gene programs, upon which later-evolved, developmental regulators such as SP5/SP8 impose tissue- and stage-specific logic⁶². Future work should define the upstream cues that govern Xap5–Nono assembly and test how alternative cofactors reprogram Xap5 outputs across tissues. Mapping the context-dependent partner repertoire of Xap5 will be key to explaining how an ancient regulator is repurposed to drive distinct transcriptional outputs across cell types, with direct implications for development and ciliopathy mechanisms.

Materials and methods

Antibodies

The following primary antibodies were used in this study: from ABclonal Biotechnology, anti-Flag (IF 1:1,000; AE005), anti-Gapdh (IB 1:10,000; AC002), anti-Nono (IB 1:1,000, IF 1:500; A3800), and anti-Sox5 (IB 1:1,000; A23127); from Sigma-Aldrich, anti-Acetyl-α-Tubulin (IF 1:500; T7451); from Proteintech, anti-Arl13b (IF 1:500; 17711-1-AP); from HUABIO, anti-Sox9 (IB 1:1,000; ET1611-56); and from Prospertech, anti-Xap5 (IB 1:1,000, IF 1:500; HU-412003).

The following secondary antibodies were used: Alexa Fluor 488- and 594-conjugated goat anti-mouse/rabbit IgG (IF 1:500; ThermoFisher Scientific); HRP-conjugated goat anti-mouse IgG (IB 1:2,000; Bio-Rad, #1706516) and HRP-conjugated goat anti-rabbit IgG (IB 1:2,000; Bio-Rad, #1706515).

Cell culture

The NIH/3T3 mouse embryonic fibroblast cell line (ATCC, CRL-1658) was cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% calf serum (CS) and 1% penicillin-streptomycin (all from Life Technologies). Cells were maintained in a humidified incubator at 37°C and 5% CO.

Plasmid construction, transfection and generation of knockout cell lines

For protein expression, the coding sequences of mouse Xap5 and Nono were amplified from NIH/3T3 cDNA and cloned into the pECMV-MCS-Flag and pCMV-C-GFP vectors, respectively, to generate constructs encoding C-terminally tagged proteins. Site-directed mutagenesis to delete the Xap5 nuclear localization signal (NLS) was performed using the Mut Express II Fast Mutagenesis Kit V2 (Vazyme, C214). For CRISPR/Cas9-mediated gene knockout, the all-in-one plasmid pU6-sgRosa26-1_CBh-Cas9-T2A-BFP (a gift from Ralf Kuehn; Addgene plasmid #64216) was used as the backbone. Gene-specific sgRNA sequences were inserted by replacing the original sgRosa26-1 guide RNA sequence via site-directed mutagenesis. The integrity of all constructs was confirmed by Sanger sequencing. All primer and sgRNA sequences are listed in Supplementary Tables 1 and 2, respectively.

For transient expression experiments, cells were transfected with the indicated DNA constructs using Lipofectamine 3000 (Life Technologies) according to the manufacturer’s protocol.

Stable knockout cell lines were generated by electroporation. Briefly, 1×10⁶ NIH/3T3 cells were resuspended in 250 µL of Opti-MEM I Reduced Serum Medium (Gibco) containing 20 µg of the appropriate all-in-one CRISPR-Cas9 plasmids. The cell suspension was transferred to a 0.4 cm gap electroporation cuvette (Bio-Rad) and electroporated using a Gene Pulser Xcell Electroporation System (Bio-Rad) with a single exponential decay pulse (250 V, 950 µF, ∞ Ω resistance). Twenty-four hours post-electroporation, single cells exhibiting bright BFP fluorescence were identified by fluorescence microscopy and manually isolated. Individual cells were seeded into separate wells of a 96-well plate to grow into monoclonal colonies, which were subsequently expanded and validated for successful gene knockout.

Immunofluorescence and microscopy

To induce primary cilia formation, NIH/3T3 cells were seeded at low density (< 30% confluency) onto sterile, poly-D-lysine-coated (Gibco) glass coverslips (CITOTEST). Following 24 h of growth in standard medium, ciliogenesis was induced by switching to serum-free DMEM for an additional 24–48 h.

Cells on coverslips were washed three times with PBS and fixed with 4% paraformaldehyde (PFA) in PBS for 10 min at room temperature. Following fixation, residual PFA was quenched with 50 mM NH₄Cl in PBS for 5 min. Cells were then permeabilized with 0.2% Triton X-100 for 15 min and blocked with 3% bovine serum albumin (BSA) in PBS for 1 h. Primary antibody incubations were performed for 1 h at room temperature in a humidified chamber. After washing, cells were incubated with Alexa Fluor-conjugated secondary antibodies for 1 h, followed by a 5 min counterstain with DAPI (Sigma-Aldrich). Finally, coverslips were washed and mounted onto glass slides using Fluoromount-G mounting medium (SouthernBiotech). This protocol was used for all cilia detection and Xap5 ectopic expression experiments. For the detection of ectopically expressed Nono-GFP, cells were fixed and washed as above, then directly counterstained with DAPI before mounting, omitting the permeabilization and antibody incubation steps.

General fluorescence for confirming ectopic protein expression was assessed using a Zeiss Axio Vert. A1 inverted microscope. High-resolution imaging of primary cilia was performed on a Leica SP8 confocal microscope using a 63× oil-immersion objective. The percentage of ciliated cells and cilia length were quantified from confocal images using the Automated Cilia Detection and Counter (ACDC) software (v0.93)⁶⁴. The nuclear area was quantified from DAPI-stained images using Fiji software.

RNA isolation and RT-qPCR

Total RNA was isolated from cells using the FastPure Cell/Tissue Total RNA Isolation Kit (Vazyme, RC112) according to the manufacturer’s protocol. For each sample, 500 ng of total RNA was reverse-transcribed into cDNA using the HiScript II 1st Strand cDNA Synthesis Kit (Vazyme, R212). RT-qPCR was then performed with ChamQ SYBR qPCR Master Mix (Vazyme, Q311) on an Applied Biosystems StepOnePlus Real-Time PCR Systems. Relative gene expression was calculated using the 2^−ΔΔCt method, with Gapdh serving as the internal normalization control. All primer sequences are listed in Supplementary Table 1.

Western blotting

Cells were lysed on ice in RIPA buffer (Beyotime, P0013B) supplemented with 1× protease inhibitor cocktail (Roche, 11873580001) and 1 mM PMSF (Sigma, 93482). The resulting lysates were resolved by SDS-PAGE and transferred to polyvinylidene fluoride membranes (Millipore, IPVH00010) using a wet transfer system (Tanon). Membranes were blocked with 5% non-fat milk in TBST (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween-20) for 1 h at room temperature, followed by incubation with primary antibodies overnight at 4 °C. After washing, membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Chemiluminescence signals were developed using the ECL Super Kit (ABClonal, RM02867) and visualized with a Tanon 5200 Chemiluminescence Imaging System.

Co-immunoprecipitation (Co-IP) and mass spectrometry

For Co-IP and IP-MS experiments, cultured cells were washed twice with cold 1× PBS and harvested by scraping. Total proteins were extracted on ice using a Co-IP buffer containing 50 mM HEPES (pH 7.5), 200 mM KCl, 1 mM EGTA, 1 mM MgCl₂, 10% (v/v) glycerol, 0.5 mM DTT, 0.3% (v/v) NP-40, and 1× protease inhibitor cocktail (Roche). For immunoprecipitation, antibodies were first crosslinked to Protein A/G magnetic beads (Invitrogen, 80105G) using 5 mM BS³. The resulting antibody-conjugated beads were then incubated with the cell lysates for 2 h at 4°C with gentle rotation. After extensive washing, bound proteins were eluted with 0.1 M Glycine-HCl (pH 2.5). For western blot analysis, the eluates and input samples were neutralized, mixed with SDS loading buffer, boiled for 5 min, and subsequently analyzed by immunoblotting. For protein identification, the eluates were sent for analysis by liquid chromatography with tandem mass spectrometry (LC-MS/MS) at Novogene (Beijing, China).

RNA-sequencing and analysis

For transcriptomic analysis, cells were plated for 24 h and subsequently serum-starved for 24 h. At the end of the starvation period, cells were immediately washed three times with 1× PBS and then lysed directly in TRIzol Reagent (Invitrogen) for total RNA extraction. Library construction and paired-end (2 × 150 bp) sequencing on an Illumina NovaSeq 6000 platform were performed by BioMarker Technologies (Beijing, China).

The raw sequencing reads were mapped to the Mus musculus reference genome (GRCm39/mm39) using HISAT2⁶⁵ (v2.1.0). StringTie⁶⁶ (v1.3.3b) was subsequently used to assemble the mapped reads. Gene expression levels were quantified as fragments per kilobase of transcript per million fragments mapped (FPKM). Differential expression analysis between experimental groups was performed using DESeq2⁶⁷ (v1.12.4). Genes with a false discovery rate (FDR) < 0.01 and an absolute fold change ≥ 1.5 were considered significantly differentially expressed. Gene Ontology (GO) enrichment analysis was performed using the DAVID database. Gene Set Enrichment Analysis (GSEA) was conducted using the OECloud platform (https://cloud.oebiotech.com).

CUT&Tag-sequencing and analysis

For CUT&Tag analysis, NIH/3T3 cells were plated for 24 h and subsequently serum-starved for 24 h. Following harvesting, 100,000 cells were collected for each sample. Library construction was performed using the Hyperactive Universal CUT&Tag Assay Kit for Illumina Pro (Vazyme, TD904) according to the manufacturer’s protocol. The resulting libraries were sequenced on an Illumina NovaSeq platform (150 bp paired-end reads) at Novogene Co., Ltd. (Beijing, China).

For bioinformatic processing, raw reads were assessed for quality with FastQC⁶⁸ (v0.11.9) and trimmed with fastp⁶⁹ (v0.20.0; parameters: --length_required 15, --n_base_limit 6). Clean reads were aligned to the Mus musculus reference genome (mm39) using BWA-MEM (v0.7.12; parameters: -k 32, -T 30, -M). Peak calling was performed using MACS2⁷⁰ (v2.1.0) with a q-value threshold of 0.05 and the following parameters: -f AUTO --call-summits --nomodel --shift -100 --extsize 200 --keep-dup all. For de novo motif discovery, peak summits were extended to 500 bp and analyzed using the HOMER⁷¹ (v4.9.1) findMotifsGenome.pl script. Peak annotation and association with the nearest genes were performed using the ChIPseeker⁷² R package. Genome browser tracks were generated using the Integrative Genomics Viewer (IGV, v2.18.1)⁷³.

CUT&Tag-qPCR

For targeted validation, DNA from 100,000 cells was prepared in parallel using the same CUT&Tag kit according to the manufacturer’s protocol. qPCR was performed using the resulting DNA. Immunoprecipitation with IgG sera served as a negative control, and an exogenous DNA spike-in was used as an internal normalization control. Enrichment at specific loci was calculated relative to the spike-in DNA control, and the 2^−ΔΔCt method was used to compare experimental samples. All primer sequences used for CUT&Tag-qPCR are listed in Supplementary Table 1.

Statistical analysis

Statistical analysis was performed using GraphPad Prism v6.01 or v8.0.2. All data were presented as mean ± SEM. Values of P < 0.05 were considered statistically significant. Statistical significance between two groups was calculated using an unpaired, parametric, two-sided Student’s t-test.

Data availability

The RNA-sequencing data generated in this study have been deposited in the Gene Expression Omnibus (GEO) database under accession code GSE305355. The CUT&Tag-sequencing data reported in this paper are available in the GEO database under accession code GSE304980. All other data supporting the findings of this study are available within the article and its Supplementary Information files. Source data are provided with this paper.

Supplementary figures

Supplementary Fig. 1

Supplementary Fig. 2

Supplementary Fig. 3

Supplementary Fig. 4

Supplementary Fig. 5

Supplementary Fig. 6

Supplementary Fig. 7

Supplementary Table 1.

Supplementary Table 2.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (grant number: 32170702), the Natural Science Foundation for Distinguished Young Scholars of Hubei Province (grant number: 2025AFA060) and the Wuhan Municipal Education Bureau’s Program for the Integration of Research and Education (grant number: 2025KCJ03). We thank Dr. Peter Swoboda (Karolinska Institute) for his constructive comments on the preprint version of this manuscript, specifically for pointing out the sequence similarity between the Xap5 motif and the C. elegans C-box.

Additional information

Author Contributions

W.W., X.Z., Y.Q., X.M., G.T. and Z.H. conceived and designed the research. W.W., X.Z., Y.Q., X.M., S.C., Y.C., S.L., W.C., J.Y., X.Y., H.L., J.X., C.X. and H.J. performed the research. W.W., G.T. and Z.H. contributed new reagents/analytic tools. W.W., H.W., G.T. and Z.H. analyzed the data. W.W. and Z.H. wrote the manuscript. All authors reviewed the manuscript and approved the final version.

Funding

MOST | National Natural Science Foundation of China (NSFC) (32170702)

• Zhangfeng Hu

湖北省科技厅 | Natural Science Foundation of Hubei Province (湖北省自然科学基金) (2025AFA060)

• Zhangfeng Hu

Wuhan Municipal Education Bureau's Program (2025KCJ03)

• Zhangfeng Hu

Additional files

Supplementary Data 1. Differentially expressed genes between WT and Xap5 KO identified by RNA-seq.

Supplementary Data 2. Differentially expressed genes between WT and Nono KO identified by RNA-seq.

Supplementary Data 3. Curated list of cilia-associated genes used in the study.

Supplementary Data 4. List of Xap5-bound genes identified by CUT&Tag analysis.

Supplementary Data 5. List of Nono-bound genes identified by CUT&Tag analysis.

Supplementary Data 6. Differentially expressed genes between WT and Sox5 KO identified by RNA-seq.

source data