Mice used for all experiments were 4 to 8-week-old of mixed sex and housed in a specific-pathogen-free facility managed by the Southern Medical University Division of Laboratory Animal Center. EACBE^-/-, Rag1 ^-/-, Rag2 ^-/-, EACBE^-/- Rag2 ^-/-and EACBE^-/- Rag1 ^-/- mice had been previously characterized (PMID: 32853367, 37534534). All procedures involving mice were conducted in strict compliance with the protocols sanctioned by the Institutional Animal Care and Use Committee at Southern Medical University.

Thymus glands were carefully harvested and homogenized in MACS buffer. Thymocytes were filtered through a 40 μm nylon mesh to obtain a single-cell suspension. For LAM-HTGTS analysis, DN thymocytes (Thy1.2⁺, CD4^-, CD8^-) and DP thymocytes (Thy1.2⁺, CD4⁺, CD8⁺) were sorted from WT or EACBE^-/- mice. Rag-deficient DN thymocytes are directly isolated from Rag1 ^-/- and EACBE^-/- Rag1 ^-/- mice.

Unless otherwise specified, all antibodies were procured from Biolegend. DN and DP cells were sorted through staining with antibodies targeting CD4 (RM4–5), CD8 (53–6.7), and Thy1.2 (53–2.1). The γδ-T cells in the thymus, spleen, and lymph nodes were identified using anti-γδ-T (GL3) and CD3 (145–2C11) antibodies. Data acquisition was performed using a BD FACSCanto II flow cytometer configured for eight-color analysis.

Total thymocytes were lysed by incubation in a buffer containing 10 mM Tris–HCl (pH 8.0), 150 mM NaCl, 10 mM EDTA, 0.4% (wt/vol) SDS, and 0.1 mg/ml proteinase K, maintained overnight at 37°C. Genomic DNA was subsequently isolated using phenol/chloroform extraction followed by ethanol precipitation. The polymerase chain reaction (PCR) was conducted under the following conditions: an initial denaturation at 95°C for 3 minutes; 30 cycles consisting of denaturation at 95°C for 30 seconds, annealing at 60°C for 30 seconds, and extension at 72°C for 1 minute; and a final extension at 72°C for 5 minutes. Following agarose gel electrophoresis and transfer to nylon membranes, PCR products were detected through hybridization with biotin-labeled oligonucleotide probes. The sequences of primers and probes are detailed in Supplementary Table S1 .

3C-HTGTS libraries were constructed using thymocytes isolated from Rag1 ^-/- or EACBE^-/- Rag1 ^-/- mice. For each experiment, three to four mice were utilized. The detailed methodology has been previously outlined (31). The sequences of the nested primer, and adapter-complementary primer are provided in Supplementary Table S2 .

To analyze open chromatin regions, ATAC-seq was conducted utilizing DN thymocytes derived from Rag1 ^-/- or EACBE^-/- Rag1 ^-/- mice. Initially, approximately 5 × 10^4 cell pellets were washed once with cold PBS. Cells were lysed on ice for 3 minutes in 50 μl ice-cold Lysis Buffer, which comprised 10 mM Tris at pH 7.4, 10 mM NaCl, 3 mM MgCl2, 0.1% NP-40, 0.1% TWEEN 20, and 0.01% Digitonin dissolved in DEPC-treated water. Following lysis, the cells were resuspended in 1 ml of ice-cold RBS-Wash buffer containing 10 mM Tris at pH 7.4, 10 mM NaCl, 3 mM MgCl2, and 0.1% TWEEN 20, and then centrifuged at 4°C at 500 × g for 5 minutes to pellet the cellular material. The tagmentation process was executed in 1 × Tagmentation Buffer that included 10 mM Tris at pH 7.4, 5 mM MgCl2, 10% DMF, 33% PBS, 0.1% TWEEN 20, and 0.01% Digitonin, employing 100 nM of Tn5 Transposase for 30 minutes at 37°C. Immediately after tagmentation, the free DNA fragments were purified following the protocol specified by the QIAquick PCR Purification Kit (QIAGEN, 28106, Germany). This step was followed by a final PCR amplification cycle of 10 to 15 rounds using P5 and P7 primers. Post-purification, the prepared libraries were sequenced via the Illumina NovaSeq 6000 sequencing platform to generate 150bp pair-end reads.

LAM-HTGTS was performed using 1 μg DNA of sorted DN cells or 6 μg DNA of sorted DP cells from one WT or EACBE^-/- mice per experiment. DNA was extracted using DNA Isolation Mini Kit (Vazyme, DC102) and sonicated to about 500bp on a Qsonica Bioruptor Sonicator. Sonicated DNA was linearly amplified with a biotinylated primer that anneals to sites of interest. Biotin-labeled single stranded DNA products were enriched with streptavidin C1 beads (65001, Thermo Fisher Scientific), and followed by 3’ end ligation with the bridge adapter. The adapter-ligated products were amplified through nested PCR using a nested primer and an adapter-complementary primer ( Supplementary Table S3 ). The detailed primers used in this study are also listed in Supplementary Information , Supplementary Table S3 . And a final PCR for another 10–12 cycles of amplification with P5 and P7 primers was performed. After purification, libraries were sequenced on an Illumina NovaSeq 6000 platform to obtain 150 bp pair-end reads.

RNA was extracted from DN thymocytes from Rag1 ^-/- or EACBE^-/- Rag1 ^-/- mice employing TRIzol reagent (Invitrogen), adhering strictly to the manufacturer’s protocol. 500ng RNA was used to synthesize cDNA according to the manufacturer’s instructions (Vazyme, R312). Quantitative real-time PCR (qPCR) was then conducted utilizing a Relative Quantification approach. The thermal cycling conditions were set as follows: an initial denaturation at 95°C for 5 minutes, followed by 45 cycles of denaturation at 95°C for 30 seconds, and annealing/extension at 60°C for 1 minute. The primer sequences employed are cataloged in Supplementary Tables S4 or S5 . The relative expression levels of various gene transcripts were computed using the comparative ΔΔCt method, where the ΔΔCt value for each target gene was normalized against that of the housekeeping gene Actb.

For subsequent library construction, 1μg of total RNA was processed. Initially, ribosomal RNA (rRNA) was removed using an rRNA depletion kit, and the remaining mRNA was fragmented into shorter segments (200–300 bp) with the addition of a fragmentation buffer. First-strand cDNA synthesis was initiated using random hexamer primers, while second-strand cDNA was synthesized in the presence of buffer, deoxynucleotide triphosphates (dNTPs including dUTP, dATP, dGTP, and dCTP), RNase H, and DNA polymerase I. The cDNA was subsequently purified using the QiaQuick PCR kit and eluted with EB buffer. Following this, the cDNA underwent end repair, adenylation, and ligation with Illumina adapters. The second cDNA strand containing uracil was specifically degraded by the USER enzyme. Lastly, PCR amplification was performed to enrich for strand-specific cDNA libraries. Post-purification, these libraries were subjected to high-throughput sequencing on the Illumina NovaSeq 6000 platform, generating 150bp paired-end reads.

Native ChIP was performed on DN thymocytes from 3–4 Rag2 ^-/- or EACBE^-/- Rag2 ^-/- mice per experiment. Cells were lysed in 200μl of a buffer containing 80 mM NaCl, 10 mM Tris-HCl pH8.0, 10 mM sodium butyrate, 6 mM MgCl₂, 1 mM CaCl₂, 250 mM sucrose, 0.2% (vol/vol) NP40, 0.1 mM PMSF, and 1×protease inhibitor cocktail, followed by a 5-minute incubation on ice. The lysate was then subjected to centrifugation at 600 × g for 5 minutes at 4°C. The nuclear pellet was subsequently washed once with a buffer composed of 10 mM NaCl, 10 mM Tris-HCl (pH 8.0), 10 mM sodium butyrate, 3 mM MgCl₂, 1 mM CaCl₂, and 250 mM sucrose. To generate predominantly mononucleosomes with a minor fraction of dinucleosomes, the nuclei were digested by incubating them for 5 minutes at 37°C in 200 μl of the same buffer supplemented with 8 units of Micrococcal nuclease (Worthington). The enzymatic reaction was halted by adding 8 μl of a stop solution containing 0.2 M EDTA and 0.2 M EGTA. Following centrifugation at 18,000 × g for 10 minutes, the supernatant was diluted to achieve a final concentration of 16.7 mM Tris (pH 8.0), 1.2 mM EDTA, 167 mM NaCl, 1.1% Triton X-100 (v/v), 0.1 mM PMSF, and 1× protease inhibitor cocktail. The chromatin was then incubated overnight at 4°C with specific antibodies: anti-trimethylated H3K4 (Millipore, 04-745), anti-acetylated H3K27 (Abcam, ab4729), or control rabbit IgG (R&D Systems, ab-105-c). Protein A/G magnetic beads (Pierce, 88802) were added to the mixture and incubated for an additional four hours. Post-incubation, the immunoprecipitates were rigorously washed, and the DNA was purified for subsequent analysis.

Quantitative PCR (qPCR) was performed using a StepOne™ Real-Time PCR System (Thermo Fisher, 4376373) with Hieff™ qPCR SYBR^® Green Master Mix (YEASEN, China). A standard curve was constructed using gradient concentrations of genomic DNA to ensure accurate quantification. Both immunoprecipitated and input DNAs were quantified, and the Actb gene promoter served as a positive control to normalize the bound/input ratios across different samples. Detailed primer sequences are provided in Supplementary Table S6 . The PCR protocol included an initial denaturation step at 95°C for 5 minutes, followed by 45 cycles of 30 seconds at 95°C and 1 minute at 60°C.

Paired-end Illumina sequencing FASTQ data were processed by removing adapters and low-quality reads using Fastp (v0.20.0). Following quality control, trimmed reads were extracted from the sequence files with Cutadapt (v1.18). Paired-end reads containing nested primers or adapter primers were manually merged into single reads using restriction enzyme recognition sequences with PEAR (v0.9.6). Subsequently, the initial digested fragment located behind the viewpoint (VP) was isolated by fragmenting the single reads according to restriction enzyme recognition sequences. The remaining single-end reads were aligned to the enzyme-digested mm10 reference genome using Bowtie2 (v2.4.5, parameters: -p 8 –sensitive). The mouse genome sequence (mm10) was sourced from UCSC (http://hgdownload.cse.ucsc.edu/goldenPath/mm10/bigZips/chromFa.tar.gz), and concordantly exact alignments were extracted using SAMtools (v1.9). Self-ligation and off-target reads were filtered out post-mapping. For visualization purposes, the final BAM files were converted into bedGraph files using Bedtools (v2.29.2). The signal peak bedGraph file was generated through a process of post-comparison filtering, signal statistical analysis, and standardization. We applied the CPM (Counts Per Million in cis) normalization method to the bedGraph files and visualized the results using the IGV genome browser. Differential pairwise interactions were identified using the R package R.4Cker (version 1.0.0, with k=30), employing the near viewpoint Analysis function to delineate interaction domains with the viewpoint. Additionally, DESeq2 (version 1.34.0, with a significance threshold of p < 0.05) was utilized for further analysis (32). Finally, we compiled the results into a comprehensive report and visualized the data using the Bioconductor package ggplot2 (version 3.3.6).

The quality control of the raw data and the fragmentation process based on restriction enzyme sites were conducted in accordance with the previously described pairwise chromatin interaction method. Subsequently, all fragments retrieved from the same read were organized on a single line according to the unique identifier of each read, and continuous fragments were removed. To construct contact matrices, the first two digested fragments following the viewpoint fragment were extracted, or various combinations of three fragments were generated by arranging all fragments from the same read. Raw contact matrices were produced at resolutions of 3 kb, 5 kb, and 10 kb. For the correction of raw contact matrices, these interaction counts were normalized to a total of 1,000,000 interactions at the same resolutions. Like a Hi-C matrix, coverage was represented in a two-dimensional matrix, where each point indicated the number of interactions identified between two bins at a specific resolution. Differential analysis and visualization of local interactions derived from three-way interactions were performed using the R package GENOVA (v1.0.0). Loops observed on the IGV genome browser were identified using fixed-size bin resolutions ranging from 3 kb to 10 kb. Briefly, interaction loops (contact frequencies >= 5) were identified by using raw contact frequencies.

In accordance with the methodology described by Vermeulen et al. (33), our study identified cooperative, random, or competitive multi-way interactions involving the viewpoint (VP) and two additional sites of interest: a second site of interest (SOI) and a third site. This was achieved through an association analysis. Specifically, in cases where the interaction is cooperative among the VP, SOI, and the third site, a subset of reads containing both the VP and SOI should also frequently encompass the third site. To evaluate whether the third site exhibits cooperative, random, or competitive interactions, we compared its frequency in the set of reads containing both the VP and SOI (referred to as the positive set) with its frequency in the set of reads containing the VP but lacking the SOI (referred to as the negative set). To mitigate the effects of technical and sampling variations, we randomly sampled same reads from the negative set equivalent to the number of reads in the positive set. Subsequently, we randomly filtered one fragment from each sampled read in the negative set to substitute for the SOI fragment present in all reads of the positive set. This procedure was iterated 1,000 times to construct an average negative profile, with the mean and standard deviation calculated accordingly. Subsequently, the positive contact profile was compared to the negative profile, and a z-score was computed to assess the significance of cooperative or competitive interactions among the VP, the SOI, and the third partner. A z-score approaching zero suggests a random contact frequency between the SOI and the third partner in the presence of the VP, whereas a positive or negative z-score indicates cooperative or competitive interactions among these three genomic regions, respectively.

The raw sequence reads were initially processed to remove adapter sequences and low-quality reads using fastp (version 0.20.0). Subsequently, the filtered reads were aligned to the mouse genome (mm10) utilizing Bowtie2 (version 2.4.5) with parameters set to -p 8 –sensitive. PCR duplicate fragments were removed using Picard (version 2.22.8). Unmapped, multi-mapped reads, as well as those mapping to chromosome M (chrM), were filtered out. The Fragments Ratio in Peaks (FRiP) value was calculated using Bedtools (version 2.29.2) and awk (version 4.0.2). We employed deepTools (version 3.5.0) to generate bigWig files with CPM normalization, which can be visualized in IGV. SAM files were converted to BAM format using SAMtools (version 1.9) for subsequent peak calling. Peaks were identified using MACS2 (version 2.2.4) with specified parameters (–nomodel –shift -100 –extsize 200 -B –keep-dup all –broad –broad-cutoff 0.1), and annotations were performed using the R package ChIPSeeker (version 1.36.0).

Tcra repertoire sequencing data were obtained from our publicly available resource. The detailed analytical methodology has been previously described (28). To determine the differences in Vα gene usage, the usage of each Vα gene in EACBE^-/- was subtracted from its corresponding usage in the WT.

The initial raw data underwent filtration using fastp (version 0.20.0). Subsequently, trimmed reads, which included nested primers and adapter primers, were removed and extracted from the sequence file following quality control procedures implemented with Cutadapt (version 1.18). Additionally, reads exhibiting contamination or low quality were eliminated. The identification of T-cell receptor alpha and delta chain V, D, and J genes, as well as the extraction of CDR3 sequences from the clean reads, was conducted utilizing MiXCR (version 3.0.11) (available at https://github.com/milaboratory/mixcr). The corresponding germline sequences were aligned with reference sequences obtained from the international ImMunoGeneTics (IMGT) database.

In our previous study, we analyzed the Tcrd repertoire in wild-type (WT) and EACBE^-/- mouse using 5’ rapid amplification of cDNA ends (5’ RACE) with a C_δ-specific primer. The results revealed a significant reduction in the usage of proximal V_δ segments, such as Trdv2–2 and Trdv1, in EACBE-deleted thymocytes, whereas the usage of Trdv5 and distal V_δ segments was increased (28). These findings suggest that EACBE contributes to the regulation of V(D)J recombination of the Tcrd gene. The enhancer E_α was previously shown to be dispensable for Tcrd rearrangement but necessary for maintaining physiological expression levels of mature VDJ_δ transcripts (30). To determine whether EACBE directly regulates Tcrd rearrangement, we performed a PCR-Southern blot assay. The results confirmed that EACBE deletion leads to an increase in Trdv5 rearrangements and a decrease in Trdv2–2 rearrangements ( Supplementary Figure S1A ).

To obtain a more comprehensive view of Tcrd rearrangement dynamics, we employed Trdj1-HTGTS-seq to examine the usage of V_δ genes in sorted DN cells. In WT cells, frequently used V_δ segments included Trdv5, Trdv2-2, Trdv1, Trav21/dv12, Trav15-2/dv6-2, Trav15-1/dv6-1, Trav15n-1, Trav15d-2/dv6d-2, and Trav15d/dv6d-1, with Trdv2–2 being the most prominently used. Compared to WT DN cells, the usage of Trdv5 and distal Trav15 family genes was increased in EACBE-deleted DN cells, alongside reduced usage of proximal V_δ genes (Trdv2–2 and Trdv1) ( Figure 1A ), which is consistent with the 5’ RACE results. These data support a direct regulatory role of EACBE in Tcrd rearrangement.

To assess the functional consequences of altered Tcrd rearrangement, we assessed the γδ T cell populations in the thymus, spleen, and lymph nodes of WT and EACBE^-/- mice. Although γδ T cell proportion were comparable in the thymus and spleen, a slight reduction was observed in the lymph nodes of EACBE^-/- mice ( Supplementary Figure S1B ). These results suggest that EACBE deletion does not markedly impair γδ T cell development.

Previous studies have demonstrated that Tcrd rearrangement increases the usage of V_α segments in the repeat region and enhances the diversity of the Tcra repertoire (26, 34). Therefore, we examined the usage of V_δ genes in sorted DP thymocytes from WT and EACBE^-/- mice using Trdj1-HTGTS-seq. We observed that the V_δ usage profile in WT DP cells mirrored that in WT DN cells. In contrast, EACBE^-/- DP cells exhibited increased rearrangement of Trdv5 and distal V_δ genes and reduced rearrangements of proximal V_δ segments ( Figure 1B ). These findings indicate that EACBE may indirectly regulates Tcra rearrangement by modulating Tcrd rearrangement during the DN stage.

Our recent study demonstrated that EACBE deletion changed V_α usages in Tcra primary rearrangement (28). To assess whether this change is attributable to alterations in Tcrd rearrangement in EACBE-deficient mice, we performed a comprehensive analysis of V_α usage in WT and EACBE^-/- thymocytes, utilizing previous Tcra 5’ RACE data. The analysis revealed a distinctive, repetitive alteration in V_α usage, in which V_α segments could be grouped into four repetitive domains based on three frequently used V_δ segments from the Trav15 family: 1) Trav1 to Trav15d-1/dv6d-1, 2) Trav9d-2 to Trav15n-1, 3) Trav9n-2 to Trav15-1/dv6-1, and 4) Trav9–2 to Trdv2-2 ( Figure 1C ). Notably, the overall V_α usage within these four regions remained unchanged in EACBE^-/- mouse thymocytes compared to WT ( Supplementary Figure S1C ). However, within region 4, the usage of certain proximal V_α genes increased. Furthermore, EACBE deletion resulted in increased usage of several V_α segments just upstream of the Trav15 family members in the other three regions, followed by a subsequent decrease in usage from 3’ to 5’ regions ( Figure 1C , Supplementary Figure S1D ). Consequently, we hypothesized that EACBE modulates V_α gene usage by regulating distal V_δ usage, especially the Trav15 family.

To confirm this, we detected Tcra primary rearrangements in sorted DP cells from WT and EACBE^-/- mice using Traj61-HTGTS-seq. Although Traj61 is a pseudogene, it is the first J_α gene to undergo rearrangement, and its rearrangement serves as a marker for Tcra primary rearrangement. In WT DP cells, Traj61 predominantly rearranged with proximal V_α genes, particularly Trav21, consistent with previous findings (28, 34). Additionally, we observed frequent rearrangements of Traj61 with V_α segments just upstream of the Trav15 family ( Figure 1D ). In contrast, EACBE deletion in DP cells resulted in a reduction of Traj61 rearrangements with proximal V_α genes, like Trav21, while increasing rearrangements with the Trav12-Trav14 region upstream of the Trav15 family ( Figures 1D, E ). These findings suggest that EACBE not only facilitates proximal V_α-to-J_α rearrangements during the DP stage but also restricts Trav15 family rearrangements during the DN stage, thereby preserving the diversity of the Tcra repertoire at the DP stage. Recent work by Danielle J et al. (27) revealed that the usage of the Trav15-dv6 family in Tcrd recombination enhances Tcra repertoire diversity, further supporting our observations.

Sleckman BP et al. reported that the usage of V_α segments in peripheral T cells of E_α-deficient mice were highly restricted, as the majority of these cells expressed Trav14(V_α2)-related TCRs, compared to 5%–10% of peripheral T cells in WT mice (30). To elucidate the effect of EACBE on Trav14 rearrangement, we analyzed Trav14 family rearrangements in sorted DP cells from WT and EACBE-deficient mice using Trav14-HTGTS sequencing. The results revealed a significant increase in rearrangement between the Trav14 family and 5’ J_α segments (from Traj61 to Traj38), and a significant decrease in rearrangement with 3’ J_α segments (Traj21 to Traj2) ( Figures 2A, B ), consistent with 5’ RACE results.

The Trav14-HTGTS sequencing panel contains nine V_α members of the Trav14 family, located at varying distances from the Tcra gene recombination center, namely Trav14d-1, Trav14d-2, Trav14d-3-dv8, Trav14n-1, Trav14n-2, Trav14n-3, Trav14-1, Trav14-2, and Trav14-3 (2). In WT mice, Trav14d-1, Trav14-1, and Trav14–3 are the most frequently used segments, with Trav14–1 being the most prevalent ( Figure 2C , Supplementary Figure S2A ). The frequencies of segments within the Trav14 family do not align consistently with those observed in the Trav15 family. For instance, Trav14–2 is located upstream of Trav15-2-dv2, which demonstrates the highest rearrangement frequency among Trav15 family members. The rearrangement of the Trav15 family with DJ_δ in DN cells reduces the spatial distance between Trav14 family members and the Tcra recombination center in DP cells. Despite Trav14–2 exhibiting the highest primary rearrangement frequency within the Trav14 family, its overall usage frequency remains relatively low compared to other Trav14 family members ( Figures 1A, E , 2C ). EACBE deletion significantly enhances the rearrangement of Trav14d-1 and Trav14d-2, while significantly reduces the rearrangement of Trav14-1 ( Figure 2C , Supplementary Figure S2A ). Furthermore, we observed a significant increase in the rearrangement of each Trav14 member with 5’ J_α, except for Trav14-3 ( Figures 2D, E ). These results suggest that EACBE plays a role in modulating the diversity of Trav14-related TCRα chain.

The CDR3 region of the antigen receptor is crucial for antigen recognition, with its amino acid composition playing a central role in determining specificity (35). To assess the impact of EACBE on CDR3 diversity, we conducted an analysis of the amino acid sequence characteristics of CDR3 in Trav14-related T-cell receptor α (TCRα). The result revealed that EACBE deletion did not alter the amino acid length or composition of CDR3 within the Trav14 family ( Figure 2F , Supplementary Figure S2B ). However, it did influence the frequency distribution of CDR3 lengths and types among various members of the Trav14 family ( Figures 2G, H , Supplementary Figures S2C, 2D ). Additionally, we observed a slight, albeit statistically insignificant, reduction in the overall CDR3 diversity of the Trav14 family following EACBE deletion ( Supplementary Figures S2E, 2F ). Interestingly, the CDR3 diversity of individual Trav14 family members either increased or decreased, consistent with the rearrangement outcomes ( Figures 2I, J ). These results indicate that EACBE plays a regulatory role in the rearrangement processes and diversity of TCRs associated with the Trav14 family.

The accessibility and germline transcription of antigen receptor genes are crucial for regulating V(D)J recombination (36). To investigate whether EACBE modulates Tcrd rearrangement by influencing the chromatin activity of the Tcrd locus, we assessed accessibility, active histone modifications, and germline transcription at the Tcra-Tcrd locus in DN cells. ATAC-seq analysis revealed a marked decrease in the accessibility of Trdv5, Trdd2, Trdv2-2, and upstream region of Trav17 ( Figure 3A ). Chromatin markers indicative of active regions, such as H3K27 acetylation (H3K27ac) (37) and H3K4 trimethylation (H3K4me3) (38–40), are integral to V(D)J recombination. ChIP-qPCR assays showed no statistically significant alterations in these active chromatin marks across most regions of the Tcra-Tcrd locus following EACBE deletion, including Trav21, Trdd1, Trdj1, Trdj2, Trdv5, TEAp, and E_α. Specifically, at the Trdv2–2 promoter, we noted slight reductions in H3K4me3 and H3K27ac levels ( Figures 3B, C ). These results suggest a potential impact on the transcriptional activity of Trdv2-2.

To further explore the effects of EACBE deletion on Tcrd transcriptional activity, we conducted GT-RNA-Seq using DN thymocytes derived from EACBE^+/+ × Rag1^−/− and EACBE^−/− × Rag1^−/− mice. In WT DN cells, the highest transcriptional activity was observed at the Tcrd recombination center, followed by a region proximal to Trav17. Weak transcriptional activity was also detected in the region between Trdv1 and Trdv2-2 ( Figure 3D ). These transcriptional patterns align with the rearrangement activity of Tcrd in DN cells. EACBE deletion resulted in a reduction of forward transcription at Trdv2-2, while reverse transcription experienced a slight increase ( Figure 3E ). Furthermore, transcription at D_δ-J_δ segments were modestly decreased in EACBE-deficient mice ( Figure 3F ).

To corroborate these findings, RT-qPCR experiments were conducted to assess germline transcription of Tcrd gene. Although the results did not reach statistical significance, EACBE deletion was associated with a reduction in the germline transcription of these segments, including Trav17, Trdv2-2, Trdd2, E_δ, and Trdc, with Trdv2–2 exhibiting the most pronounced decrease ( Figure 3G ). These findings suggest that EACBE deletion attenuated chromatin activity at the Trdv2–2 promoter, leading to a subsequent decrease in its rearrangement.

We previously reported that EACBE deletion reduced interactions between the proximal V_α and proximal J_α regions in DP thymocytes (28). To further explore the effect of EACBE deletion on interactions involving Trdv2-2, Trdv5, and Trdd2, we conducted a 3C-HTGTS assay using these segments as viewpoints. In DN cells from Rag1-deficient mice, Trdv2–2 exhibited substantial interactions with sequences extending from upstream Trav21 to downstream INTs ( Figures 4A, B ). Notably, EACBE deletion resulted in a significant reduction in interactions between Trdv2–2 and sequences from INTs to E_α, including the D_δ-J_δ-C_δ region ( Figure 4B ). Additionally, we observed a significant increase in interactions between Trdv2–2 and the downstream region of EACBE, which may be attributed to the EACBE deletion weakening the insulation at the TAD boundary in which it is situated ( Supplementary Figures S3A , S3B ).

Interactions involving Trdd2 were confined to the region between INTs and TEAp ( Figures 4A, C ), consistent with previous observation (26). As anticipated, the deletion of EACBE resulted in increased interactions of Trdd2 with sequences located downstream of EACBE ( Supplementary Figures S3A , S3C ). The deletion also increased interactions between Trdd2 and sequences between INTs and TEAp, including Trdv5, while leaving interactions with sequences upstream of INTs unaffected ( Figure 4C ). Furthermore, Trdv5 exhibited slightly increased interactions with sequences between INTs and TEA ( Figures 4A, D ), consistent with Trdd2 3C-HTGTS data. Trdv5 also demonstrated increased interactions with sequences downstream of EACBE ( Supplementary Figures S3A , S3D ). Our previous research demonstrated that the EACBE deletion would affect the expression of its downstream genes in the thymocyte cells (28). However, the EACBE deletion did not exhibit a similar impact on downstream gene expression in DN cells ( Supplementary Figure S3E ).

In summary, these results indicate that EACBE establishes a TAD boundary at the DN stage to restrict the interaction between the Tcra-Tcrd locus and its downstream regions. We also observed that EACBE deletion not only enhanced interactions within the region from INTs to TEAp, but also weakened interactions between Trdv2–2 and the region from INTs to E_α. This suggests that EACBE deletion enhance the insulation of INTs. In brief, EACBE is involved in regulating the spatial organization of the Tcrα-Tcrd locus at the DN stage, facilitating the normal rearrangement of the Tcrα-Tcrd locus.

To investigate the impact of EACBE on the higher-order chromatin structure of Tcra-Tcrd at the DN stage, we did a three-way interaction analysis recently developed in our laboratory. This method has previously been used to examine the higher-order chromatin architecture of the Tcra-Tcrd locus in DP thymocytes, as well as the cooperative interactions among Vα, Jα, and Eα (31). In this study, we applied this method to analyze the higher-order chromatin structure of the locus in DN cells. The Trdd2 three-way contact heatmap showed that the deletion of EACBE resulted in a marked increase in interactions between INTs and TEAp with Trdd2 in DN cells ( Figure 5A ). Furthermore, the KO – WT subtraction heatmap demonstrates a significant enhancement in three-way interactions between INTs and Dad1 ( Figure 5B ), a pattern also observed in the Eα three-way contact heatmap ( Supplementary Figures S4A, S4B ). Examining the Trdv2–2 three-way contact heatmap, we observed that co-occurring interaction pairs are confined in a small region surrounding Trdv2–2 in WT DN cells and the EACBE deletion does not influence the three-way interaction from the viewpoint of Trdv2-2 ( Supplementary Figure S4D ). However, the sequences interacting with the Eα-INT2 combination were significantly reduced, including the three-way contact involving Eα-INT2-Trdv2-2 ( Supplementary Figures S4A, S4B ). These results indicate that EACBE deletion increases the insulation of INTs.

To elucidate the relationship between EACBE and INTs in DN cells, we performed a method developed by Allahyar et al. to analyze specific three-way contacts. This method employs a second Site of Interest (SOI) to distinguish between preferred and random or disfavored three-way contacts (31, 33). First, we examined the co-occurrence frequency of third sequences across the Tcra-Tcrd locus when Trdd2 interacts with INT2 as an SOI. Most sequences located between INTs and TEAp, such as E_δ, are disfavored in three-way contacts with the Trdd2-INT2 combination ( Figure 5C ). However, some sequences upstream of INTs and surrounding TEAp are favored in three-way contacts with the Trdd2-INT2 combination ( Figure 5C ). Notably, EACBE deletion leads to a reduction in synergistic interactions of upstream sequences of INTs with the Trdd2-INT2 combination ( Figure 5C ). When Trdd2 and Trdv5 are used as the viewpoint-SOI combination, the coordinated behavior of sequences upstream of INTs diminishes, accompanied by a shift in sequences between INTs and TEAp ( Figure 5D ). Furthermore, the viewpoint-SOI analysis reveals that EACBE deletion also facilitates the synergistic interaction of the sequences from INTs to TEAp with the Eα-Trdd2 or Eα-INT2 combinations ( Figure 5E , Supplementary Figure S4C ). These findings suggest that EACBE reduces the insulation of INTs, thereby facilitating the rearrangement of proximal V_δ segments.