Developmental Stage-Specific Regulation of TCR-α-Chain Gene Assembly by Intrinsic Features of the TEA Promoter¹

T cell progenitors pass through a series of developmental stages in the thymus where they must satisfy specific requirements centered around the assembly and expression of TCR genes if they are to be released into the periphery as mature αβ or γδ T cells (1). Assembly of the TCRβ-, δ-, and γ-chain genes is initiated in double negative (DN)³ thymocytes (2, 3, 4). Expression of a γδ TCR leads to progression along the γδ T cell lineage pathway. Generation of a productive TCRβ-chain gene and expression of a TCRβ-chain, as a pre-TCR, promote commitment to the αβ T cell lineage, cellular expansion, and transition to the double positive (DP) stage of thymocyte development (5). Pre-TCR signals also lead to the initiation of TCRα-chain gene assembly, which occurs primarily in DP thymocytes, as well as the cessation of further TCRβ-chain gene rearrangements through the process of allelic exclusion (2, 3, 5, 6, 7).

The assembly of different TCR genes is precisely regulated within the context of thymocyte development. This regulation is mediated, in part, by modulation of the accessibility of gene segments to the RAG-1 and -2 proteins through the activity of cis-acting elements, such as promoters and enhancers, that regulate transcription (8). The regulation of the TCRα- and δ-chain genes poses perhaps the greatest regulatory challenge, as these genes lie in a single complex locus, the TCRα/δ locus (2, 9). The TCRα/δ locus spans ∼1.6 Mb with the Vα and Vδ gene segments clustered over 1.5 Mb in the 5′ region of the locus (Fig. 1) (9, 10). Some V gene segments are used primarily in TCRα-chain gene assembly (Vα gene segments), and others primarily in TCRδ-chain gene assembly (Vδ gene segments). However, most V gene segments are used in both TCRα- and TCRδ-chain gene assembly (Vα/δ gene segments) (10). The V gene segment cluster is followed by two Dδ (Dδ1 and Dδ2) gene segments, two Jδ (Jδ1 and Jδ2) gene segments, and the TCRδ constant region gene (Cδ) (Fig. 1). A single Vδ gene segment, Vδ5, lies 3′ of Cδ and rearranges by inversion (Fig. 1). A cluster of 61 Jα gene segments and the TCRα constant region gene (Cα) lie in the most 3′ region of the locus (Fig. 1).

TCRδ-chain genes are assembled in DN thymocytes, whereas TCRα-chain genes are assembled in DP thymocytes. The developmental stage-specific assembly of TCRα- and δ-chain genes is thought to be regulated primarily by the TCRα and the TCRδ enhancers Eα and Eδ (2). Eδ lies in the intron between Jδ2 and Cδ and is active in DN thymocytes to promote germline transcription and accessibility of TCRδ-chain gene segments (Fig. 1) (2). Eα, which lies 3′ of Cα, is not active in DN thymocytes (Fig. 1) (2). Rather Eα becomes active only after DN thymocytes commit to either the γδ T cell lineage or the αβ T cell lineage (DP thymocytes) (2, 11). In γδ T cells Eα augments transcription from TCRδ promoters (12). In DP thymocytes Eα activates transcription from germline Jα gene segment promoters, such as the TEA promoter, that are required for TCRα-chain gene assembly (2, 11, 12). Eα is also required for TCRα-chain gene expression in mature αβ T cells (2, 12).

TCRα gene assembly does not occur in DN thymocytes due, at a minimum, to the inactivity of the TEA promoter and other Jα gene segment promoters in these cells (13). This may result from the inactivity of Eα in DN thymocytes (11, 14). However, Eδ is active in DN thymocytes and promotes the accessibility of TCRδ gene segments, including Vδ5 (see Results), that lie immediately upstream of the TEA promoter (2, 15). How is it that Eδ is capable of activating the Vδ5 promoter, which lies 10 kb downstream, but not the TEA promoter, which lies 14 kb downstream (Fig. 1)? Deletion of a 2-kb region between Vδ5 and the TEA promoter known as BEAD-1, which exhibits insulator activity in vitro, did not lead to TEA promoter activation in DN thymocytes (16). It is thus possible that other sequences between Vδ5 and the TEA promoter, such as the Vδ5 promoter, prevent Eδ from functioning with the TEA promoter. In addition, because Eδ may function over a shorter distance than Eα it is possible that Eδ may function over the 10-kb distance to Vδ5 but not the 14-kb distance to TEA (17). Lastly, it is possible that intrinsic features of the TEA promoter prohibit its activation in DN thymocytes.

To distinguish between these possibilities, we used a gene-targeting approach to generate mice with a modified TCRα/δ locus (TCRα/δ^5ΔT), in which a 4-kb region encompassing the Vδ5 gene segment and the immediate downstream region to the 5′ border of the TEA promoter has been deleted. Thus, the TEA promoter on the TCRα/δ^5ΔT allele lies in the same location as the Vδ5 gene segment on the wild-type TCRα/δ allele. The TEA promoter on the TCRα/δ^5ΔT allele exhibits normal activity in DP thymocytes. However, analyses of TCRα/δ^5ΔT/5ΔT DN thymocytes failed to reveal a significant increase in TEA-driven Jα gene segment transcription or accessibility. These findings demonstrate that the inactivity of the TEA promoter in DN thymocytes is not due to its relative position in the locus with respect to other cis-acting elements such as Eδ. Rather, our findings suggest that intrinsic features of the TEA promoter limit its activity in DN thymocytes and, as such, have important implications for the developmental stage-specific regulation of TCR gene assembly.

All mice were bred and maintained under specific pathogen-free conditions at the Washington University School of Medicine (St. Louis, MO) and were handled in accordance to the guidelines set forth by the Division of Comparative Medicine of Washington University.

The 5′ homology arm of the pVδ5^TC targeting vector was generated as PCR fragment corresponding to nt 8574–14306 of the GenBank accession no. M64239 sequence. The 5′ end of this fragment is in exon 1 of the Cδ gene segment and the 3′ end is in the spacer of Vδ5 gene segment recombination signal sequences (RSS). The 3′ homology arm is a BglII-NsiI DNA fragment corresponding to nt 18292–25232 of the GenBank accession no. M64239 sequence. The 5′ end of this fragment is immediately upstream of the 5′ border of the TEA promoter. The homology arms were cloned into the pLNTK vector (12). Embryonic stem cell targeting and Cre-loxP mediated deletion was conducted as previously described (12).

Southern and Northern analyses were performed as described (12). The 5′ (probe 5), Vδ5 (probe 7), Cδ (probe 6), TEA (probe 8), Cα, and GAPDH probes have been described (12, 18). The 3′ probe is a PCR fragment corresponding to nt 26554–27388 of the GenBank accession no. M64239 sequence.

VδDδJδ rearrangements were assayed by PCR using the oligonucleotides 5′-TTTTGGTATCGCAAAAGGCC-3′ and 5′-CCCTGCTCCTATGGAGGAGC-3′. Standard PCR buffer conditions with 1 mM MgCl₂ were used. Cycling conditions were 94°C for 5 min followed by 30 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s. This was followed by a final 7 min of incubation at 72°C. PCR products were hybridized to the oligonucleotide probe 5′-TGTGAAGCACAGCAAGGCCA-3′ as described previously (19). The IL-2 gene PCR was described previously (20).

RT-PCR for analysis of Jα gene segment utilization has been described previously (21). The Jα gene segment utilization index was calculated as index = (Jα^5ΔT/Cα^5ΔT)/(Jα⁺/Cα⁺), where Jα^5ΔT and Jα⁺ are the intensities of the Southern blot bands from the hybridization of Jα-specific oligonucleotides to Vα-Cα RT-PCR products from TCRα/δ^5ΔT/5ΔT and wild-type splenocyte cDNA, respectively. Cα^5ΔT and Cα⁺ are the intensities of the Southern blot bands from the hybridization of a Cα-specific oligonucleotide to the same RT-PCR products and serve as the loading control. LMPCR and Southern analyses were conducted with primer A (5′-CCAAGATTCCTGGGACAACC-3′), primer B (5′-GCGATGGGACTGTGACTGAC-3′), primer BW1H, and the oligonucleotide probe P (5′-TCCAAAAGAGGAAAGGAAGGCAGTC-3′) as previously described (19).

To obtain DN CD25⁺ cells, total thymocytes were enriched for DN cells by complement-mediated lysis of CD4⁺ and CD8⁺ cells as described (19). The resulting cells were stained with FITC-conjugated anti-mouse CD25 (BD Pharmingen) followed by sorting for CD25⁺ cells with FACSDiva (Becton Dickinson) to >95% purity.

Two independent panels of γδ T cell hybridomas were generated from TCRα/δ^+B6/5ΔT mice as described previously (12).

To determine whether Eδ is required for Vδ5 gene segment transcription and accessibility in DN thymocytes, we conducted Northern blot analyses on RNA from RAG-2-deficient DN thymocytes that were homozygous for the wild-type TCRα/δ locus (R2^−/−:Eδ^+/+) or a TCRα/δ locus in which Eδ had been deleted (R2^−/−:Eδ^Δ/Δ) (Fig. 2,A). R2^−/−:Eδ^+/+ DN thymocytes have a robust level of Vδ5-hybridizing germline transcripts (Fig. 2,A). In contrast, the level of Vδ5 germline transcripts were dramatically reduced in R2^−/−:Eδ^Δ/Δ DN thymocytes, demonstrating that Eδ activity is required for the optimal level of Vδ5 gene segment transcription in these cells (Fig. 2,A). That Eδ is required for Vδ5 rearrangement is evidenced by the ∼10-fold reduction in complete Vδ5DJδ1 rearrangements in Eδ^Δ/Δ thymocytes as compared with wild-type thymocytes (Fig. 2,B). In agreement with previous reports, TEA-hybridizing transcripts were not observed in DN thymocytes (Fig. 2 C) (13). Together, these findings demonstrate that in DN thymocytes Eδ functions to promote Vδ5 gene segment transcription and accessibility but it is unable to activate the TEA promoter, which lies just 4 kb downstream of Vδ5.

To determine why Eδ is unable to activate the TEA promoter in DN thymocytes, we generated the TCRα/δ^5ΔT allele through gene targeting of embryonic stem cells (Fig. 3). Initially the TCRα/δ^5NT allele was generated by replacing the 4-kb region upstream of the TEA promoter that includes the Vδ5 gene segment with the neomycin resistance gene flanked by loxP sites (Fig. 3, A and B). The TCRα/δ^5ΔT allele was then generated from the TCRα/δ^5NT allele through Cre-loxP mediated deletion of the neomycin resistance gene (Fig. 3). On the TCRα/δ^5ΔT allele the TEA promoter lies at the same distance from Eδ as the Vδ5 gene segment on the wild-type TCRα/δ allele (Fig. 3 A). Furthermore, any sequences between Vδ5 and the TEA promoter that could inhibit Eδ interactions with the TEA promoter, including the Vδ5 promoter, have been deleted on the TCRα/δ^5ΔT allele.

Analysis of TCRα/δ^5ΔT/5ΔT mice revealed no significant differences in the number of DN, DP, and CD4⁺ or CD8⁺ single positive thymocytes or in the number of mature peripheral αβ or γδ T cells as compared with wild-type mice (data not shown).

To determine whether the TEA promoter on the TCRα/δ^5ΔT allele is regulated normally in DP thymocytes, we generated RAG-2-deficient mice expressing the DO11 TCRβ transgene that were homozygous for the wild-type TCRα/δ allele (R2^−/−:β:TCRα/δ^+/+) or the TCRα/δ^5ΔT allele (R2^−/−:β:TCRα/δ^5ΔT/5ΔT) (Fig. 4,A). Northern blot analysis revealed similar levels of germline TEA-hybridizing transcripts in R2^−/−:β:TCRα/δ^+/+ and R2^−/−:β:TCRα/δ^5ΔT/5ΔT DP thymocytes (Fig. 4,A). Analysis of TCRα mRNA from TCRα/δ^5ΔT/5ΔT and TCRα/δ^+/+ αβ T cells revealed that similar fractions of VJα rearrangements use the Jα58 and Jα57 gene segments, which are located 5′ in the Jα gene segment cluster and are dependent on TEA promoter activity for accessibility (Fig. 4 B) (22). Together, these data demonstrate that the TEA promoter is intact and regulates transcription and accessibility normally on the TCRα/δ^5ΔT allele in DP thymocytes.

Transcription from the TEA promoter in DN thymocytes was assayed by the Northern blotting of RNA isolated from RAG-2^−/−:TCRα/δ^+/+ (R2^−/−:TCRα/δ^+/+) and RAG-2^−/−:TCRα/δ^5ΔT/5ΔT (R2^−/−:TCRα/δ^5ΔT/5ΔT) thymocytes (Fig. 5). As shown previously, no TEA- or Cα-hybridizing transcripts were detectable in R2^−/−:TCRα/δ^+/+ (Fig. 5). Although there were slightly more TEA-hybridizing transcripts in DN thymocyte RNA from R2^−/−:TCRα/δ^5ΔT/5ΔT as compared with R2^−/−:TCRα/δ^+/+ mice, the level was significantly lower than what was observed in R2^−/−:β:TCRα/δ^5ΔT/5ΔT DP thymocytes where the TEA promoter is active (Fig. 5). Furthermore, although TEA transcripts are frequently spliced to Cα, Cα hybridizing transcripts were not detected in R2^−/−:TCRα/δ^5ΔT/5ΔT DN thymocytes (Fig. 5). Germline TCRδ transcripts from the TCRα/δ⁺ and TCRα/δ^5ΔT alleles were found at similar levels in DN thymocytes (Fig. 5). Together, these data demonstrate that although Eδ is functional and promotes normal level of germline TCRδ transcripts on the TCRα/δ^5ΔT allele in DN thymocytes (Fig. 5), it is unable to promote a significant level of TEA transcripts in these cells.

To determine whether Jα gene segments are accessible for RAG-mediated cleavage, we assayed for Jα61 signal ends (SEs) in thymocytes at different stages of development by LMPCR. To this end, thymocytes from mice heterozygous for the wild-type B6 and 129 TCRα/δ alleles (TCRα/δ^+B6/+129) or those hemizygous for the wild-type B6 TCRα/δ allele and the TCRα/δ^5ΔT allele (TCRα/δ^+B6/5ΔT) were assayed. The TCRα/δ^5ΔT allele was generated through targeting of the 129 TCRα/δ allele. LMPCR products generated by TCRα/δ⁺¹²⁹ or TCRα/δ^5ΔT Jα61 SEs can thus be distinguished from TCRα/δ^+B6 Jα61 SEs by a DdeI restriction site polymorphism (Fig. 6 A).

LMPCR analyses revealed robust level of Jα61 SEs in total TCRα/δ^+B6/+129 thymocytes, which are composed predominantly of DP thymocytes with ongoing TCRα-chain gene assembly (Fig. 6,B). Digestion of LMPCR products with DdeI revealed similar levels of Jα61 SEs from the TCRα/δ^+B6 and TCRα/δ⁺¹²⁹ alleles, indicating that there are no intrinsic differences in RAG-mediated cleavage at the wild-type TCRα/δ locus in these two alleles (Fig. 6,C). TCRα/δ^+B6/5ΔT thymocytes had level of Jα61 SEs similar to that observed in TCRα/δ^+B6/+129 thymocytes (Fig. 6,B). Furthermore, digestion of these products with DdeI demonstrated that a similar fraction originated from cleavage at the TCRα/δ^+B6 and TCRα/δ^5ΔT alleles (Fig. 6 C). Thus, in total thymocytes the level of Jα61 cleavage on the TCRα/δ^5ΔT allele is similar to that observed on the wild-type TCRα/δ⁺ allele.

To determine whether the TEA promoter on the TCRα/δ^5ΔT allele can promote Jα gene segment accessibility in DN thymocytes, CD25⁺ DN thymocytes were purified from TCRα/δ^+B6/5ΔT and TCRα/δ^+B6/+129 mice by flow cytometric cell sorting. CD25⁺ DN thymocytes from TCRα/δ^+B6/5ΔT and TCRα/δ^+B6/+129 mice had similar levels of Jα61 SEs, which were considerably lower than the levels observed in DP thymocytes (Fig. 6,B). Furthermore, digestion of the LMPCR products with DdeI revealed that there were similar levels of Jα61 SEs from the TCRα/δ^+B6 allele and TCRα/δ^5ΔT allele in TCRα/δ^+B6/5ΔT DN thymocytes (Fig. 6 D). Together, these findings demonstrate that there is a significantly lower level of RAG-mediated cleavage at the Jα61 gene segment on the TCRα/δ^5ΔT allele in DN thymocytes as compared with DP thymocytes. Furthermore, the level of Jα61 gene segment cleavage on the TCRα/δ^5ΔT allele is similar to that observed for the wild-type TCRα/δ allele in DN thymocytes. Thus, the TEA promoter does not promote increased Jα gene segment accessibility on the TCRα/δ^5ΔT allele in DN thymocytes.

Eδ may only function with the TEA promoter on the TCRα/δ^5ΔT allele in DN thymocytes that have committed to the γδ T cell lineage and promote V-Jα rearrangements in these cells. To investigate this possibility, we generated a panel of 81 γδ T cell hybridomas from splenic TCRα/δ^+B6/5ΔT γδ T cells. That these T cell hybridomas were all derived from γδ T cells was evidenced by flow cytometric analyses demonstrating cell surface expression of TCRδ- but not TCRβ-chains (data not shown).

All Vα to Jα rearrangements occur by deletion and result in deletion of the TEA exon from the TCRα/δ^+B6 or TCRα/δ^5ΔT allele in TCRα/δ^+B6/5ΔT γδ T cells. Genomic DNA isolated from the TCRα/δ^+B6/5ΔT γδ T cell hybridomas was digested with BglII and subjected to Southern blot analyses using the TEA exon probe (Fig. 7). A BglII polymorphism, generated by the targeting, permits the distinction of TEA hybridization on the TCRα/δ^+B6 or TCRα/δ^5ΔT alleles in TCRα/δ^+B6/5ΔT γδ T cells (Fig. 3,A). These analyses revealed the retention of TEA hybridization on both the TCRα/δ^+B6 and TCRα/δ^5ΔT alleles in all 81 of the TCRα/δ^+B6/5ΔT γδ T cell hybridomas (Fig. 7 and data not shown). That these hybridomas represented a diverse population of γδ T cells was evidenced by hybridization with a probe 3′ of the Jδ1 gene segment, which revealed many different size-hybridizing fragments that represent heterogeneous TCRδ-chain gene rearrangements in the TCRα/δ^+B6/5ΔT γδ T cell hybridomas (Fig. 7). Thus, Eδ is not capable of promoting V to Jα gene segment rearrangement on the TCRα/δ^5ΔT allele in γδ T cell precursors.

In this study we show that the TEA promoter is not active in DN thymocytes even when it is in the same location as the Vδ5 gene segment promoter, which is normally active in these cells. Thus, TEA promoter inactivity in DN thymocytes is not due to its physical distance from enhancer elements, such as Eδ, that promote TCRδ gene rearrangement and expression in DN thymocytes. Furthermore, the absence of TEA transcripts in these DN thymocyte also suggests that there are no cis-acting elements, including the Vδ5 promoter, between Vδ5 and TEA that are necessary to prevent Eδ from activating the TEA promoter in DN thymocytes. Rather, our findings suggest that TEA promoter activity in DN thymocytes may be limited by intrinsic properties of the promoter itself.

The TEA promoter is factor-loaded in DN thymocytes but is unable to function in these cells (23). It is possible that transcription factors bound to the TEA promoter in DN thymocytes are repressive in nature. However, it is also possible that some additional trans-acting factors are required for TEA activity but are only expressed in DP thymocytes. In this regard, it is notable that the expression of Ets family transcription factors, which are predicted to bind to the TEA promoter, are up-regulated as cells transit from the DN to the DP stage of thymocyte development (24). It is also possible that the TEA promoter has a 5′ region with insulator activity.

Because many V gene segments in the TCRα/δ locus are accessible and participate in TCRδ gene assembly in DN thymocytes, it is likely that V to Jα rearrangements are prevented in these cells primarily due to the inaccessibility of the Jα gene segments. Our findings suggest that this inaccessibility is due, in part, to intrinsic features of the TEA promoter that prohibit it from being activated by cis-acting elements, such as Eδ, that function to promote TCRδ-chain gene rearrangements in DN thymocytes.

Activation of V to Jα rearrangements in DN thymocytes could have a detrimental effect on both αβ and γδ T cell development and lineage commitment. All V to Jα rearrangements occur by deletion, leading to the excision of the TCRδ-chain genes from the chromosome and, thus, limit the ability of DN thymocytes to generate a γδ TCR (25). Expression of an αβ TCR or a pre-TCR by DN thymocytes leads to signals that induce transition to the DP stage of thymocyte development (5). However, the pre-TCR is more efficient at mediating this transition than the αβ TCR (26, 27). The TCRα-chain can compete successfully with the pre-Tα protein for dimerization with the TCRβ-chain (28). Thus, a productive VJα rearrangement and the expression of a TCRα-chain in DN thymocytes would favor formation of an αβ TCR instead of pre-TCR and, as a result, promote a less efficient transition to the DP stage of thymocyte development.

The authors have no financial conflict of interest.