Germ-line elimination of electric charge on pre–T-cell receptor (TCR) impairs autonomous signaling for β-selection and TCR repertoire formation

Abstract

The pre–T-cell receptor (TCR) is crucial for the early T-cell development, but the ligand for pre-TCR remains unidentified. We recently proposed a model that pre-TCR complexes oligomerize spontaneously through interactions of the pre-TCRα chain. To investigate the mechanism underlying this ligand-independent signaling in vivo, we established knock-in mice that express a pre-TCRα mutant lacking charged amino acids (D²²R²⁴R¹⁰²R¹¹⁷ to A²²A²⁴A¹⁰²A¹¹⁷; 4A). CD4⁺CD8⁺ thymocyte number was significantly reduced in invariant pre-TCRα (pTα^4A/4A) mice, whereas CD4⁻CD8⁻ thymocytes were unaffected. The percentages of double-negative 3 (DN3) cells and γδ T cells were increased in the pTα^4A/4A thymus, indicating that β-selection is impaired in pTα^4A/4A mice. Pre-TCR–mediated tyrosine phosphorylation and clonal expansion into double-positive thymocytes were also defective in the knock-in mice. Pre-TCR was expressed at higher levels on pTα^4A/4A cell surfaces than on those of the wild type, suggesting that the charged residues in pTα are critical for autonomous engagement and subsequent internalization of pre-TCR. Pre-TCR–mediated allelic exclusion of the TCRβ gene was also inhibited in pTα^4A/4A mice, and thereby, dual TCRβs were expressed on pTα^4A/4A T cells. Furthermore, the TCRβ chain variable region (Vβ) repertoire of mature T cells was significantly altered in pTα^4A/4A mice. These results suggest that charged residues of pTα are critical for β-selection, allelic exclusion, and TCRβ repertoire formation.

The pre–T-cell receptor (TCR) is a multimeric complex consisting of variable rearranged TCRβ chains, an invariant pre-TCRα (pTα) chain, and CD3 molecules. Pre-TCR is crucial for T-cell development from the CD4⁻CD8⁻ [double-negative (DN)] to the CD4⁺CD8⁺ [double-positive (DP)] stages, a process known as β-selection. Despite its structural similarity to mature αβTCR, which requires MHC antigen for engagement, it has been proposed that pre-TCR initiates signals in a ligand-independent manner (1, 2). Several mechanisms have been proposed in attempts to explain how pre-TCR mediates autonomous signaling (3–6); however, the precise molecular mechanism underlying the process remains unclear (2).

We have recently proposed that the pre-TCR complex forms oligomers spontaneously. The pTα-erythropoietin receptor (EPOR) chimera indeed triggers a growth signal autonomously through self-oligomerization, and four conserved charged residues (D22, R24, R102, and R117) are critical for the functioning of pTα-EPOR (7).

Part of the difficulty in analyzing the pre-TCR signal in vivo arises because the threshold for the β-selection checkpoint is quite low, and thus, a very precise dosage of the expressed pre-TCR needs to be achieved. One reason for this low threshold is that the DN stage is uniquely sensitive to signals mediated by the immunoreceptor tyrosine-based activation motif (ITAM) (7–10). The high sensitivity of DN cells may arise from (i) the coexpression of spleen tyrosine kinase (Syk) and zeta-associated protein 70 (ZAP-70) in the DN stage (10), (ii) the high level of expression of microRNA (mir181a) in the DN stage, which enhances T-cell responsiveness by suppressing several inhibitory phosphatases (11), (iii) the predominant role of Rap1 rather than Ras downstream of pre-TCR (12), or (iv) the expression of C-X-C chemokine receptor (CXCR)4 as a costimulatory receptor of pre-TCR (13, 14). Therefore, the overexpression of pTα mutants might overcome their original structural requirements. Indeed, although many studies have addressed mutagenesis analysis of pTα by using transgenic or retroviral introduction of mutant pTα, conclusions regarding the definitive functional domain remain a matter of controversy (3–9, 15, 16). These observations suggest that the structure–function relationship of pTα needs to be analyzed with the pTα mutant present at the correct times, amounts, and locations.

Pre-TCR–mediated signaling inhibits further rearrangement of TCRβ chain, which is called allelic exclusion (17). Because pre-TCR signal is triggered autonomously and independent of individual TCRβ chain, pre-TCR may also contribute to generate a diversity of TCRβ during β-selection. Therefore, optimal pre-TCR signal might play a crucial role in shaping appropriate TCRβ repertoire in mature T cells. However, this issue has not been clearly addressed, because pTα^−/− mice show severe defects in the development of mature T cells themselves (18).

In the current study, to fulfill those prerequisites, we used a knock-in approach and showed that charged residues in pTα are critical for the initiation of the pre-TCR signal, β-selection, and TCRβ repertoire formation.

Results

Generation of Knock-In Mice Expressing Charge-Less pTα Chain.

To examine whether the charged residues are crucial for pre-TCR–mediated β-selection in vivo, we generated knock-in mice that expressed a pTα mutant in which four residues (D22, R24, R102, and R117) were exchanged for alanine, a neutral residue (Fig. 1A). A targeting construct bearing four-point mutations in two exons was used for homologous recombination, and the Neo cassette was deleted after germ-line transmission (Fig. S1). The knock-in allele was detected by means of primers located on both sides of loxP sites (Fig. 1B). Homozygous pTα^4A/4A mice were born at rates consistent with Mendelian inheritance, and they showed no obvious abnormalities.

Fig. 1.

Generation of charge-less pTα knock-in mice. (A) Locations of point mutations in the pTα mutant. The four charged residues in the extracellular domain of pTα were replaced by alanine. (B) Genomic PCR. Genomic DNAs from WT (pTα^+/+), pTα^4A/+, and pTα^4A/4A mice were amplified with primers that cross over loxP sites, indicated as arrows in Fig. S1. (C) mRNA expression of the charge-less pTα mutant. CD4⁻CD8⁻ (DN) thymocytes from pTα^+/+, pTα^4A/+, pTα^4A/4A, and pTα^−/− mice were analyzed by real-time RT-PCR using common (pan-pTα) and pTα^WT- and pTα^4A-specific primers as described in SI Materials and Methods. Data are represented as mRNA levels relative to that of β-actin. Each value is the mean ± SD for triplicate assays. (D) Surface staining of pre-TCR complex. DN thymocytes from pTα^+/+, pTα^4A/4A, and pTα^−/− mice were stained with anti-pTα, anti–IgG₁-biotin, and streptavidin-allophycocyanin (APC) (Upper) or anti–TCRβ-APC (Lower), as described in Materials and Methods. Nonlabeled mouse IgG₁ mAb and APC-labeled hamster IgG₂ were used as isotype controls for anti-pTα and anti-TCRβ, respectively (thin histograms).

First, we analyzed whether the mutant allele was normally expressed by comparing mRNA levels for pTα^WT and pTα^4A in DN thymocytes in wild-type (WT; pTα^+/+), heterozygous knock-in (pTα^4A/+), homozygous knock-in (pTα^4A/4A), and knock-out (pTα^−/−) mice. We designed common primers for both pTα alleles as well as specific primers that were able to distinguish between the WT and mutant allele, and we confirmed that the total amount of pTα was constant and that mutant pTα was properly expressed from the targeted allele (Fig. 1C).

Next, we confirmed the expression of pTα^4A proteins (Fig. 1D). Pre-TCR was shown to be expressed at very low levels on the surface of WT thymocytes, because the degree of specific staining of anti-pTα or anti-TCRβ was only slightly higher than that of isotype control antibodies (7, 19). This slight shift was significant, because it was absent in pTα^−/− thymocytes. However, surface pre-TCR was clearly detected on pTα^4A/4A thymocytes (Fig. 1D), indicating that the mutant pTα protein was correctly expressed and thus, likely assembled with the other components of the pre-TCR complex.

From these observations, we could confirm the successful establishment of the knock-in mice expressing charge-less pTα.

Impaired Thymocyte Development in pTα^4A/4A Mice.

We first examined thymic cellularity in pTα^4A/4A mice. The total number of thymocytes was significantly reduced in pTα^4A/4A mice compared with pTα^+/+ mice (Fig. 2A). Population analyses revealed that, in pTα^4A/4A mice, the number of CD4⁺CD8⁺ (DP) thymocytes but not CD4⁻CD8⁻ (DN) thymocytes was reduced, showing that differentiation from DN to DP was blocked by the charge-eliminating mutation of pTα (Fig. 2B). The percentage of DP thymocytes was also decreased in pTα^4A/4A mice (Fig. 2C), although the reduction was not as drastic. We, therefore, conducted a precise examination of the efficiency of β-selection in pTα^4A/4A thymocytes using an in vitro differentiation system (7). DN thymocytes from WT mice gave rise to over 20% of DP cells when cultured on stroma cells expressing the Notch1 ligand for 2 d (20, 21). However, the proportion of pTα^4A/4A-derived DP cells was only one-fifth of that of WT mice, offering further confirmation that these differentiation defects are significant in a cell-intrinsic way (Fig. 2D).

Fig. 2.

DN to DP thymocyte development was impaired in pTα^4A/4A mice. (A and B) Thymocyte numbers. Thymocytes from 4-wk-old mice were analyzed for the absolute number of total thymocytes (A), CD4⁻CD8⁻ population (DN), and CD4⁺CD8⁺ population (DP) (B). Each symbol represents an individual mouse. **P < 0.01. (C) Flow cytometric profiles of whole thymocytes from WT (pTα^+/+), pTα^4A/4A, and pTα^−/− mice. The percentages of thymocytes in each quadrant are shown. Data are representative of four independent experiments. (D) In vitro thymocyte differentiation. DN cells negatively sorted by anti-CD8 and anti-CD4 magnetic beads were cultured on Tst-4/delta-like ligand (DLL)1 stromal cells. At day 0 (Upper) and day 2 (Lower), cells were analyzed for DN to DP transition. The percentage of DP population is shown. Data are representative of three independent experiments. (E) Thymocytes were analyzed for the expression of CD25 and CD44 within the CD4⁻CD8⁻ (DN)-gated population, and the percentages of CD44⁻CD25⁺ (DN3) cells among the whole thymocytes are plotted as in A. Each symbol represents an individual 4-wk-old mouse. **P < 0.01. (F) Thymocytes were stained with anti-TCRδ and anti-CD3ε antibodies, and the percentages of γδ T cells in the whole thymocytes are plotted as in E. **P < 0.01. (G) Thymic cellularity of pTα^4A/4A × TCRδ^−/− mice. Total thymocyte numbers and percentages of DN3 population are also shown. Data are representative of three independent experiments.

We further examined the subpopulations of thymocytes that are known to be affected by β-selection. DN thymocytes develop through CD44⁺CD25⁻ (DN1), CD44⁺CD25⁺ (DN2), CD44⁻CD25⁺ (DN3), and CD44⁻CD25⁻ (DN4) stages. Pre-TCR is expressed mainly at the DN3 stage, and there are several reports of demonstrations that the blocking of pre-TCR signaling results in an accumulation of a DN3 population (22–24). Therefore, we examined subpopulations of DN cells in pTα^4A/4A mice and found that the DN3 population accumulated significantly in these mice (Fig. 2E).

It has been shown that impairment of pre-TCR signaling also results in an increase in the ratio of γδ T cells, because the pre-TCR signal is critical for the development of αβ T cells but γδ T cells (18). Indeed, the proportions of γδ T cells were significantly increased in pTα^4A/4A mice (Fig. 2F). The absolute number of thymic γδ T cells was increased slightly in the pTα^4A/4A mice [WT (1.76 ± 0.59 × 10⁵) vs. pTα^4A/4A (3.38 ± 1.62 × 10⁵)]. However, these γδ T cells do not account for the phenotypes of αβ T cells in pTα^4A/4A mice, because impaired β-selection was observed even in the absence of γδ T cells (i.e., in pTα^4A/4A × TCRδ^−/− mice) (Fig. 2G).

Impaired β-Selection in pTα^4A/4A Mice.

Pre-TCR–induced β-selection is also accompanied by robust cell expansion during the DN3 and DN4 stages. We, therefore, investigated proliferation in these stages by examining the incorporation of BrdU into thymocytes after injection of the bromonucleoside. Around one-half the population of DN4 cells was BrdU-positive in WT mice, but this proportion rapidly decreased in DP cells (Fig. 3A Left), showing that a massive expansion occurs during the DN3 to DN4 stages in which pre-TCR is expressed. The weak incorporation of BrdU before pre-TCR expression (DN2) was of a similar level in WT and pTα^4A/4A mice, but the proportion of BrdU-positive DN4 cells decreased by one-half in the pTα^4A/4A mice (Fig. 3A Right). These results suggest that the pre-TCR–mediated proliferation was still impaired, even in the decreased but still substantial population of DN4 cells in pTα^4A/4A mice.

Fig. 3.

Impaired β-selection and pre-TCR proximal signaling in pTα^4A/4A mice. (A) BrdU incorporation into thymocytes. WT (pTα^+/+) and pTα^4A/4A mice were injected intraperitoneally with BrdU. At 4 h after injection, thymocytes were stained with anti-BrdU, anti-CD4, anti-CD8, anti-CD44, and anti-CD25 mAbs. Percentages of BrdU-positive cells within each population are shown. Data are representative of three independent experiments. (B) Competitive repopulation assay. BM cells from Ly5.1⁺ WT and Ly5.2⁺ pTα^4A/4A mice were mixed in a ratio of 1:1 and injected into irradiated Rag1^−/− mice. Reconstituted thymi and spleens were analyzed 4 wk later for Ly5.1⁺ (WT) and Ly5.2⁺ (pTα^4A/4A) cells. Each sample of mixed BM cells was injected into at least five recipient mice of matched sex and age, and representative data of similar results are shown. (C) Pre-TCR–induced phosphorylation. Freshly isolated DN thymocytes from Rag1^−/−, pTα^+/+, pTα^4A/4A, and pTα^−/− mice were detected by Western blot analysis with antiphosphotyrosine (pY) Ab. The membrane was also blotted with anti-CD3ε as a loading control. Data are representative of four independent experiments.

The impaired β-selection was also reflected in the decreased numbers of peripheral αβ T cells in neonates; however, in adult mice, this was restored in a compensatory manner so that similar numbers were present to those in WT mice (Fig. S2). The defective ability of mutant T cells to develop was confirmed by means of a competitive repopulation assay in which a 1:1 mixture of bone marrow (BM) cells from pTα^4A/4A and WT mice was injected into Rag1-deficient recipients. The pTα^4A/4A-derived BM cells generated a comparable number of DN thymocytes with that of the WT-derived BM cells. However, the pTα^4A/4A DN cells developed into fewer DP, single positive (SP), and peripheral αβ T cells and equivalent γδ T cells under the competition with WT-derived BM cells (Fig. 3B), confirming that the charged moiety of pTα is critical for optimal development of αβ T cells.

Defective Pre-TCR Proximal Signaling in pTα^4A/4A Thymocytes.

We next analyzed proximal signaling of pre-TCR. Freshly isolated DN thymocytes from WT mice contained several constitutive phosphorylated proteins (Fig. 3C, lane 2). This is highly likely to be mediated by pre-TCR, because it was not detected in pre-TCR–deficient Rag1^−/− and pTα^−/− thymocytes (Fig. 3C, lanes 1 and 4). Importantly, this pre-TCR–induced phosphorylation of cellular proteins was impaired in pTα^4A/4A thymocytes (Fig. 3C, lane 3 and Fig. S3).

Autonomous engagement of pre-TCR also induces constitutive internalization of pre-TCR itself, which leads to a low level of surface pre-TCR expression (19). If this low surface expression of pre-TCR is dependent on the charged residues in pTα, surface pre-TCR should be augmented in pTα^4A/4A thymocytes. Indeed, surface expression levels of pTα and TCRβ in pTα^4A/4A mice were higher than those in WT mice (Fig. 1D, Left and Center).

These results suggest that the proximal events through pre-TCR are dependent on the presence of charge in the pTα chain.

Break of Allelic Exclusion in pTα^4A/4A Mice.

The pre-TCR signal is critical for extinguishing further rearrangement of the TCRβ chain (17), presumably to avoid producing T cells bearing two different TCRβs on a single cell, a behavior that is called allelic exclusion (25). Therefore, we examined the effect of pTα mutation on allelic exclusion by introducing transgene of rearranged TCRβ into pTα^4A/4A mice as an in vivo model of allelic exclusion (26). We confirmed that thymocytes number in pTα^4A/4A mice decreased, even in the presence of the TCRβ transgene (Fig. S4).

Rearrangement of the endogenous Tcrb allele was detected by specific primers encompassing TCR variable (V)βs (Vβ5, Vβ8, and Vβ10), TCR diversity (D)β1, and TCR joining (J)β1 (Fig. 4A). In WT DN thymocytes, the products of V–DJ rearrangement was visualized as five major bands reflecting the usage of Jβ1.1, Jβ1.2, Jβ1.3, Jβ1.4, and Jβ1.5, respectively (Fig. 4B, lanes 1–3). D–J rearrangement could also be detected as well by using a Dβ-specific primer (Fig. 4B, lane 1–3). These are specific, because no product was detected in Rag2^−/− thymocytes, except for germ line-derived product (Fig. S5).

Fig. 4.

Break of allelic exclusion in pTα^4A/4A mice. (A) Schematic diagram of the TCRβ locus and the location of the primers used for detection of rearranged TCRβ genes. Gray box, pseudo gene. (B) Rearrangement of Tcrb allele. Genomic DNA purified from DN thymocytes was analyzed for TCRβ rearrangement using the primers described in A. Transgenic Vβ3-Jβ2 (TCRβ^Tg), pTα^WT, and pTα^4A alleles were also detected as a control for the amount of input DNA as well as for genotypes. Data are representative of three independent experiments.

The introduction of productive TCRβ transgene (2B4 TCRβ; Vβ3–Jβ2) attenuated V–DJ rearrangement of the endogenous TCRβ chain (Fig. 4B, lanes 4–6). This suppressed V–DJ rearrangement was, however, restored in pTα^4A/4A background (Fig. 4B, lanes 10–12). As previously reported, D–J rearrangement was not affected by the transgene (26). The amount of genomic DNA and the genotypes were confirmed by PCR for TCRβ^Tg, pTα^WT, and pTα^4A allele (Fig. 4B). These results suggest that the charged residues in pTα are required for the allelic exclusion of TCRβ.

Given that the allelic exclusion of TCRβ was incomplete in pTα^4A/4A mice, the single TCR on a single T-cell rule might also be broken in these mice. We, therefore, examined whether two different TCRβs are expressed on a single T cell in pTα^4A/4A mice. To address this possibility, T cells were stained with anti-Vβ3 mAb together with a mixture of all available anti-Vβs mAbs except for Vβ3 (Vβ_endo). The anti-Vβ_endo mixture stained 70% of TCRβ expressed in WT T cells (Fig. 5A, Left) . In TCRβ Tg mice, Tg-derived Vβ3⁺ cells lost the expression of endogenous Vβs as a result of allelic exclusion (Fig. 5A). However, in TCRβ Tg × pTα^4A/4A mice, T-cell population expressing dual TCRβ on the surface of each cell (Vβ3⁺Vβ_endo⁺) was significantly increased in both SP thymocytes and splenic T cells (Fig. 5 A and B).

Fig. 5.

Emergence of a dual TCRβ expresser in pTα^4A/4A mice. (A) Detection of T cells expressing dual TCRβ on the surface. SP thymocytes (Upper) and splenic T cells (Lower) from TCRβ Tg mice were stained with phycoerythrin (PE)-labeled anti-Vβ3 together with a mixture of FITC-labeled anti-Vβ2, 4, 5.1/5.2, 6, 7, 8.1/8.2, 8.3, 9, 10, 11, 12, 13, 14, and 17a (Vβ_endo). (B) Frequency of dual TCRβ expresser. Thymocytes (Left) and splenic T cells (Right) were analyzed for the percentage of dual expresser (Vβ3⁺Vβ_endo⁺) in the Vβ3⁺ population. Data are means ± SD of three independent mice. *P < 0.05. (C) Frequency of dual TCRβ expresser in non-TCR Tg mice. SP thymocytes (Left) and splenic T cells (Right) from WT (pTα^+/+), pTα^4A/+, and pTα^4A/4A mice were stained with PE-labeled anti-Vβ3 or -Vβ5 together with a mixture of FITC-labeled mAbs specific for the other Vβs. Percentage of Vβ3⁺Vβ_endo⁺ and Vβ5⁺Vβ_endo⁺ T cells is shown, respectively. Data are means ± SD of three independent mice. *P < 0.05. Data are representative of three independent experiments.

This phenomenon was also observed in non-TCR Tg background, because the frequency of SP thymocytes expressing dual endogenous Vβs increased in pTα^4A/4A mice (Fig. 5C Left), although this was not observed clearly in splenic T cells (at least not for Vβ3⁺ and Vβ5⁺ cells) (Fig. 5C Right).

These results suggest that an optimal pre-TCR signal is critical for preventing the expression of dual TCRβs on a single T cell, which would otherwise lead to autoimmune disorder (27).

TCR Repertoire of T Cells in pTα^4A/4A Mice.

Autonomous pre-TCR signaling may also contribute to permit diverse TCRβ to pass through β-selection. We finally analyzed whether the repertoire of TCRβ is altered in pTα^4A/4A mice by examining the usage of TCRVβ using several mAbs specific for different Vβs. The distribution of TCRVβ was significantly altered in pTα^4A/4A SP thymocytes (Fig. 6A) and peripheral T cells (Fig. 6B). Although peripheral T cells were apparently normal in pTα^4A/4A mice in terms of number (Fig. S2), alterations in the TCR repertoire might result in defective immune responses of peripheral T cells to a variety of antigens. Further extensive analysis is needed to clarify this issue, because simple allo-reactive T-cell responses against B10.D2 (H-2^d) and B10.BR (H-2^k) looked normal in pTα^4A/4A T cells.

Fig. 6.

Altered distribution of TCRVβ in pTα^4A/4A mice. (A) CD4⁺CD8⁻ (Left) and CD4⁻CD8⁺ (Right) thymocytes from WT, pTα^4A/+, and pTα^4A/4A mice were stained with anti-Vβs and analyzed by flow cytometry. (B) Distribution of Vβs in splenic CD4⁺ T cells (Left) and CD8⁺ T cells (Right) from WT, pTα^4A/+, and pTα^4A/4A mice were analyzed as in A. Data are means ± SD for at least three independent mice. Representative results from two independent experiments with similar results are shown. *P < 0.05; **P < 0.01.

Taken together, although pTα is not expressed in mature T cells, the charge on pTα also contributes to shape the subsequent TCR repertoire in the periphery.

Discussion

In this study, we present in vivo data supporting the idea that self-engagement of pre-TCR through charged residues is a significant driving force of β-selection.

The impairment of β-selection in pTα^4A/4A mice was significant but not as severe as that in pTα-deficient mice (Fig. 2 A and B). This might be because charged residues other than D22, R24, R102, and R117 can compensate for the function of these four residues. Alternatively, the intracellular/transmembrane domain of pTα may have some role in pre-TCR–mediated signaling (15). Interestingly, the proline-rich motif in the pTα tail is proposed to interact with signaling protein(s), thereby facilitating β-selection (5).

It is often argued that substitution of amino acids critical for protein structure may destabilize the protein and thereby, impair its function. On the basis of a molecular modeling approach (7), we believe that it does not apply to the four mutations (D22A, R24A, R102A, and R117A) for the following reasons. First, although the substitution of a small residue by bulky residue can sometimes destabilize the structure of a protein, this is not the case for the substitutions by alanine that we used. Second, none of these residues are located within a highly fluctuated region, such as a flexible loop, in which a local conformational change occurs; instead, they are located in a stable segment within a turn region (D22, R24, and R102) or next to a disulfide-bonded cysteine (R117). Third, all these residues are located on the surface of the pTα–TCRβ complex rather than in an internal region of the complex, and therefore, the possibility of substitution changing the folded structure is reduced; indeed, molecular modeling suggests that the pTα^WT–TCRβ dimer and the pTα^4A–TCRβ dimer are expressed with quite similar conformations (7). Finally, our finding that the surface pre-TCR complex was clearly detected in pTα^4A/4A mice by anti-pTα as well as anti-TCRβ showed that the pTα mutant protein was correctly expressed and thus, likely assembled with the other components of the pre-TCR complex (Fig. 1D). The characteristics of pTα^4A protein were also confirmed by microscopic and biochemical analyses (Fig. S6).

Initially, we assumed that, if the charge-mediated interaction of pre-TCR is critical for β-selection, impaired development should be observed to some degree in heterozygous pTα^4A/+ mice as well, because the quantity of pre-TCR signals would be reduced by pre-TCR^WT–pre-TCR^4A interactions. There was, however, no significant difference in the total numbers of thymocytes (or any fractions thereof) between WT (pTα^+/+) and heterozygous (pTα^4A/+) mice (Fig. 2 and Fig. S2). One possible explanation for this is that, even in heterozygous (pTα^4A/+) mice, pre-TCR^WT–pre-TCR^WT interactions should still occur, and the probability of this interaction can be stochastically calculated to be one-quarter of that in WT pTα^+/+ mice. It is, therefore, possible that the lower but nevertheless substantial frequency of pre-TCR^WT–pre-TCR^WT interactions could be sufficient to drive β-selection in terms of cell number, particularly in the uniquely sensitive DN context (9). It should be noted that the frequencies of some Vβs were significantly altered, even in heterozygous (pTα^4A/+) mice (Fig. 6), implying that partial reduction of pre-TCR signal might essentially affect β-selection in terms of repertoire formation. The analysis on the dynamics and stoichiometry of the self-engagement of pre-TCR may clarify this issue.

The teleological objective of β-selection would be to generate diverse TCRβ chains that subsequently lead to variety in the αβTCR repertoire. Considering the accumulating evidence for β-selection, pre-TCR may validate the quality of TCRβ by these minimum criteria: (i) the capacity to assemble with CD3- and TCRα-like protein (i.e., pTα) and (ii) the possession of proper intracellular signaling pathway. This apparently permissive selection may be necessary to secure diversity of TCRβ but sufficient as a prerequisite selection to confirm productive rearrangement of TCRβ before positive/negative selection, in which the specificity of αβ TCR is strictly verified.

In view of this, it would be tempting to make extensive comparisons of the variety of individual TCRβ sequences in pTα^4A/4A and WT thymocytes at the preselected DP stages to clarify whether charge-mediated self-oligomerization of pre-TCR indeed contributes to the acquisition of diverse TCRβ chains, which should be critical for the host defense against various foreign antigens. In line with this hypothesis, the frequencies of some particular Vβs were found to significantly decrease in BrdU⁺-proliferating DN4-DP thymocytes (Fig. S7) as well as mature T cells (Fig. 6) in pTα^4A/4A mice, thus implying that a sufficient strength of pre-TCR signal may, therefore, be required to select diverse Vβ chains during β-selection.

Materials and Methods

Mice.

For the generation of pTα^4A/4A mice, C57BL/6-derived Bruce4 ES cells (provided by T. Kurosaki, RIKEN, Yokohama, Japan) were transfected with a linearized targeting vector. Neomycin-resistant clones were subjected to genomic PCR to identify homologous recombinants. Chimeric mice were bred with C57BL/6 mice to obtain heterozygotes, which were then crossed with chicken β actin (CAG)-Cre Tg mice (28) with a C57BL/6 background to delete the Neo cassette. The pTα^−/− mice were provided by H. von Boehmer (Harvard Medical School, Boston) (18). 2B4 TCRβ Tg mice were provided by M. M. Davis (Harvard Medical School, Boston) (29). TCRδ^−/− mice were provided by S. Itohara by the courtesy of Y. Yoshikai (Kyushu University, Fukuoka, Japan) (30). All mice were maintained in a filtered-air, laminar-flow enclosure and given standard laboratory food and water ad libitum. Animal protocols were approved by the committee of Ethics on Animal Experiment, Faculty of Medical Sciences, Kyushu University and Research Center for Allergy and Immunology, RIKEN.

Reagents.

Anti-pTα (2F5), anti-TCRδ (GL3), mouse IgG₁ (MOPC), anti-BrdU, anti-mouse Vβs, and BrdU were obtained from BD Biosciences. Anti-CD4 (GK1.5), anti-CD8 (53.6.7), anti-CD25 (PC61), anti-CD44 (IM7), anti-CD3ε (2C11), and anti-TCRβ (H57) were obtained from eBioscience. Anti-CD4 and anti-CD8 microbeads were obtained from Miltenyi Biotec. Anti-phosphotyrosine (4G10) was from Cell Signaling. Anti-CD3ε was purchased from Santa Cruz Biotechnology.

Flow Cytometry.

For pre-TCR staining, CD4⁻CD8⁻ DN thymocytes were sorted by magnetic bead cell sorting (Miltenyi Biotec) and stained with anti-pTα, anti–mIgG-biotin, and streptavidin-allophycocyanin (APC) in the presence of anti–CD4-CyChrome, anti–CD8-CyChrome, anti–CD44-phycoerythrin (PE), and anti–CD25-FITC. The DN3 to DN4 transitional population (CD4⁻CD8⁻CD44⁻CD25^int) was gated and analyzed for surface pTα staining. Control mouse IgG₁ was used as a control mAb for anti-pTα.

In Vitro Thymocyte Differentiation.

Thymocytes from 4-wk-old mice were cultured on Tst4 cells expressing delta-like ligand (DLL)1-internal ribosome entry site (IRES)-GFP (31). At day 3 of culture, cells were analyzed for CD4 and CD8 expression after gating out of stroma cells using the forward scatter (FSC)–side scatter (SSC) and fluorescence channel 1 (FL1) gate.

Competitive Repopulation Assay.

BM cells (5 × 10⁶) from WT and pTα^4A/4A mice were mixed at ratio of 1:1 and injected i.v. into irradiated (4Gy) Rag1^−/− mice. Reconstituted thymi and spleens were analyzed 4 wk later. Every mix of BM cells was injected into at least five recipient mice with sex and age matched.

BrdU Incorporation.

Mice were injected intraperitoneally with 1 mg BrdU and then, killed 4 h later. After surface staining with anti-CD8, anti-CD4, anti-CD25, and anti-CD44, the thymocytes were fixed and permeabilized with Cytofix/Cytoperm buffer followed by treatment with DNase. Cells were stained with APC-conjugated anti-BrdU Ab (BD Biosciences).

Western Blot Analysis.

DN thymocytes were freshly isolated by negative sorting using anti-CD8 and anti-CD4 magnetic beads. Cells were lysed in lysis buffer containing 1% Nonidet P-40 and analyzed for phosphorylated proteins using anti-phosphotyrosine (pY).

Statistics.

An unpaired two-tailed Student t test was used for all of the statistical analysis.

Detection of TCRβ Gene Rearrangement.

DN thymocytes were sorted and lysed in lysis buffer containing 1% Nonidet P-40. After centrifugation, the supernatant was removed, and genomic DNA was prepared from the nuclear pellet by the addition of proteinase K-containing buffer. Threefold serially diluted DNAs were used as templates. Rearrangement of the Tcrb gene was detected by PCR as described previously (32) using Blend Taq plus DNA polymerase (TOYOBO) and the primers described in SI Materials and Methods.

Note Added in Proof.

At the proof stage of this publication, the structure analysis of pre-TCR was reported (33).