A second isoform of pTα, “pTαb,” is derived from the pTα locus by tissue-specific, alternative splicing. pTαb is coexpressed in the thymus with the previously characterized form of pTα (which we term pTαa) and is also expressed in peripheral cells without pTαa. While pTαa acts to retain most TCR β-chains intracellularly, pTαb permits higher levels of cell surface TCRβ expression and facilitates signaling from a CD3-TCRβ complex.
As part of the TCRαβ on peripheral T cells, the TCR β-chain facilitates recognition of pathogen-derived peptides presented on host MHC molecules (1). Additionally, TCRβ is part of the TCRα-independent pre-TCR that facilitates thymocyte maturation (2). Both TCRαβ and the pre-TCR are complexed with CD35 chains that transduce signals following receptor engagement (1). Beyond this, however, pre-TCR structure and function are largely unresolved. For example, surface pre-TCR expression on thymocytes is very low, provoking the idea that rather than engaging a ligand at the cell surface, it may signal from an inner cell compartment (3). Either possibility could accommodate the fact that, to function, the TCR β-chain must be able to escape from the endoplasmic reticulum (4).
A third form of TCRβ expression, which also is not well understood, is the TCRα-independent surface expression of TCRβ on so-called “β-only” cells. Such cells were described in the periphery of TCRα−/− mice (5, 6), but may also exist in normal animals (see below). β-only cells, like other mature T cells, reportedly lack pTα (7). Hence, the nature of any partner chain for TCRβ in these cells is unresolved. Here, we show that cloned β-only cells express a second pTα isoform, pTαb, which is expressed in vivo both by polyclonal β-only cells and by thymocytes. The expression pattern overlaps but is distinct from that of previously described pTα. Interestingly, transfection experiments demonstrate that each isoform has functionally distinct effects on TCR β-chain expression and signaling.
Materials and Methods
Establishment of a β-only T cell clone
Establishment of β-only T cell hybridomas
TCRα−/− splenocytes, stimulated for 3 days with Con A (2 μg/ml) (Sigma, St. Louis, MO) in Click’s medium containing 10% FCS and 5 U/ml of IL-2 (human rIL-2; PharMingen, San Diego, CA), were fused with the TCRα−β− BW5147 cell line (10). Hybrids were selected in hypoxanthine-aminopterin-thymidine (HAT; Life Technologies, Gaithersburg, MD), and cultures derived from single colonies were analyzed by fluorescence-activated cell sorting (FACS) and RT-PCR. Hybrid 1.10 was TCRβ+, TCRα−, TCRδ−.
Cell staining, FACS, and analysis
Previously described methods (9) were used with the following directly conjugated mAbs: phycoerythrin-conjugated anti-TCRβ (H57-597); anti-TCRγδ (GL3); FITC-conjugated anti-CD4 (RM4-5); anti-CD8 (53-6.7) (all from PharMingen); and anti-HA (12CA5) (Boehringer Mannheim, Indianapolis, IN).
Gene expression analyses
Previously described RT-PCR protocols (9) were used with pTα primers (7), hypoxanthine phosphoribosyltransferase (HPRT) primers (9), and the following primers as listed: CD3γ (5′-GTACAAGTGGATGGCAGC-3′ and 5′-TCACTTCTTCCTCAGTTG-3′); CD3δ (5′-ATACCAGCGTCATGCATC-3′ and 5′-GTATCTTCACGATCTCGA-3′); CD3ε (5′-CGATGCCGAGAACATTGA-3′ and 5′-CAGACTGCTCTCTGATTC-3′); CD3ζ and CD3η (5′-CAGAGCTTTGGTCTGCTG-3′, 5′- TCTGCATATGCAGGGCAT-3′, and 5′-CATGGACTCCACAGAGTG-3′); syk (5′-CGGTACTTCTCCATACAC-3′ and 5′-TTCAGGTCCTCAAAGGGT-3′); zap-70 (5′-ACCCTGTGAGCTGTGATA-3′ and 5′-ACACCATAGCATCACGCA-3′); FcεRIγ (5′-TGATCTCAGCCGTGATCT-3′ and 5′-TCAAAGCACAGAGGTGAC-3′), and TCRβ (5′-ATGAGCTGCAGGCTTCTCCTG-3′ and 5′-TTCATAGGAGCTAACCCAGTA-3′).
Expression constructs and transfections
pTαa and pTαb cDNAs were subcloned using BstxI/ApaI sites into the eukaryotic expression vector pcDNA3 (Invitrogen, Carlsbad, CA). pTαa and pTαb cDNAs lacking the leader sequence were generated by PCR (Ref. 7; 5′-CTACCATCAGGCATCGCT-3′ and 5′-CTACCATCAGGGGAATCT-3′) and cloned into pGEM-T (Promega, Madison, WI), from which they were subcloned in-frame using SalI/SacII sites into pDisplay (Invitrogen), which provided a murine Ig κ-chain leader sequence linked to an HA epitope. The TCR β-chain expression construct was previously developed in our laboratory from a diabetogenic CD4(+) αβ T cell. 4G4 cells (107) (a TCRα−β− T hybridoma), maintained in rapid growth phase, were pulsed at 960 μF, 320V with 20 μg of plasmid in Capecchi’s HBS and transferred to 20 ml of Click’s medium + 10% FCS. FACS analysis was undertaken 48 h later. Stable transfectants were selected and maintained in 1.5 mg/ml of G418 (Life Technologies).
Cells (4 × 105) were activated for 24 and 48 h in the presence of purified Abs (anti-CD3 and anti-I-Ad) previously coated to the plates (1 μg/ml, 12 h at 4°C). IL-2 secretion was tested in supernatants by ELISA (9).
Cells (5 × 106) were lysed on ice in 1 ml of 150 mM NaCl, 1 mM MgCl2, 25 mM HEPES, pH 7.5, 1 mM of Pefablock (Boehringer Mannheim, Indianapolis, IN), 10 μg/ml of leupeptin, 10 μg/ml of antipain, and 0.5% Triton X-100. Lysates were cleared for 10 min at 14,000 rpm, and supernatants were incubated with 1 μl of anti-HA Ab (HA.11, Babco, Richmond, CA) for 1 h on ice with occasional shaking. Abs were precipitated with 25 μl of Gamma-Bind beads (Pharmacia, Piscataway, NJ), and washed three times with 1 ml of RIPA buffer. Proteins were eluted off beads by boiling for 5 min in reducing buffer and run on 15% SDS-PAGE gels in parallel with a prestained standard (Broad Range; Bio-Rad, Richmond, CA). HA-tagged pTα was detected by Western blot using HA.11.
Results and Discussion
Isolation of a TCRα-β+ T cell clone from TCRα−/− mice
A small number of peripheral CD3+CD4+CD8− cells of TCRα−/− mice express surface TCRβ (6). A clone of such β-only cells, H4β, was obtained by limiting dilution from TCRα−/− (H-2b) splenocytes. By FACS, H4β was surface-TCRβ+, CD4+, TCRγδ−, CD8− (Fig. 1,A), and CD3+CD69+ (not shown). The derivation of clone H4β took approximately 2 yr, in large part because of a slow growth rate. This necessitated using RT-PCR rather than protein chemistry to assess the components of the H4β TCR. Signals for CD3-γ, -δ, -ε, -ζ, and -η (Fig. 1,B), were detected at levels comparable with those of a control CD3(+) cell line, CTLL (not shown). Signals were likewise detected for Zap-70, FcεRIγ, and syk (Fig. 1 B). The expression by H4β of surface TCRβ, all five CD3 chains, CD69, and of Zap-70 in excess to syk, is typical of peripheral αβ(+) T cells of normal mice.
An alternatively spliced form of pTα
Because H4β expresses surface TCRβ without TCRα, we tested for expression of pTα, the only other known partner for TCRβ. 5′ and 3′ pTα-specific primers amplified a product of ∼300 bp (Fig. 2,A), composed of pTα exon 1 (5′ untranslated (UT) region, leader peptide, and the first three amino acids of the mature protein), exon 3 (connecting peptide that provides the cysteine for dimerization with TCRβ), and exon 4 (transmembrane region, cytoplasmic tail, and 3′ UT region) (13) (Fig. 2,C). The product lacked the 300-bp exon 2 that encodes the major extracellular, Ig-like domain of pTα. We termed the novel isoform pTαb and refer to the previously characterized form as pTαa (Fig. 2 C).
RT-PCR also detected pTαb expression, in the absence of either pTαa or TCRα, in primary, polyclonal, β-only cells from TCRα−/− mice (Fig. 2, A and B) and in a hybrid, 1.10, derived by fusion of TCRβ(+)α− splenocytes (not shown). By contrast, RT-PCR detected both pTαb and pTαa in the thymus (Fig. 2,A). Sequencing the thymic products (Fig. 2 A) confirmed the 600-bp band to be pTαa; the 300-bp band to be pTαb; and the 500-bp band to be an artifactual hybrid of single strands of pTαa and pTαb, respectively.
pTα mRNA expression
A small pTα isoform was previously detected in the thymus by RT-PCR (7), but was reportedly not detectable by Northern blot or RNase protection and hence was considered a possible PCR artifact. By contrast, pTαb could be detected by both methods (Fig. 3, A and B). For RNase protection, a radioactively labeled pTα RNA probe was generated in which exons 1 and 3 were contiguous. Three hundred nucleotides of this probe should be protected by pTαb mRNA in which exons 1 and 3 are linked, but not by pTαa mRNA, in which a single-stranded, RNase-sensitive gap would be created by failure of the pTαb probe to bind to pTα exon 2. At the same time, the pTαb probe contained at its 5′ end 40 bases of vector sequences that would not be protected by pTαb mRNA, allowing the 300-nt protected band to be distinguished from undigested probe (340 nt) (Fig. 3 A, lane P). The 300-nt protected pTαb-specific band was seen with thymus RNA (lane T), and the β-only hybridoma, 1.10 RNA (lane F).
By Northern blot (Fig. 3,B) pTαa expression was ∼10-fold greater than that of pTαb in thymus, whereas polyclonal β-only cells expressed only pTαb. These quantitative data were highly consistent with RT-PCR analyses of pTαa and pTαb in different tissues (Fig. 2): hence, further RT-PCR analyses of pTα were performed. Again, thymus expressed more pTαa than pTαb (Fig. 3 C), while peripheral CD4(+) TCRγδ− cells from TCRα−/− mice (a subset that would contain β-only cells) again expressed pTαb but no pTαa. Interestingly, peripheral CD4(+) cells from normal mice also expressed more pTαb than pTαa. These cells may include murine counterparts of human CD4(+)CD3− progenitors that reportedly express pTαa (14). Additionally, the strong pTαb signal may reflect the presence of β-only cells in normal mice. Other peripheral subsets from normal or TCRα−/− mice expressed neither pTαa nor pTαb.
Analysis of pTα isoforms
Currently, the only known function of pTα is to facilitate β-selection of thymocytes (15, 16), in which process pTα is hypothesized to stabilize surface TCRβ (4, 17). We have detected expression of both pTα isoforms in thymocyte subsets undergoing β-selection (N. Douglas, D. F. Barber, and A. C. Hayday, unpublished observations). Therefore, a transfection experiment was undertaken to test whether pTαa and pTαb had equivalent effects on TCRβ expression. (Although pTαb lacks the Ig-like extracellular domain, it retains the connecting peptide and within it the cysteine that allows dimerization with TCRβ).
To detect expression in transfected cells, pTαa and pTαb were tagged with HA epitope before each was individually cotransfected with TCRβ into the TCR-deficient T cell line, 4G4. In parallel, 4G4 cells were transfected with empty vector or TCRβ alone. Then, 48 h later, cells were examined for the expression of both surface and intracellular TCRβ and HA-tagged pTα (Fig. 4). A significant percentage of 4G4 cells transfected with TCRβ alone expressed moderate but measurable levels of surface TCRβ (Fig. 4,A). Invariably, when cells were cotransfected with pTαa, surface TCRβ expression was reduced (Fig. 4,B). It was difficult to trace redistribution of surface TCRβ to the cytoplasm, because cells transfected with TCRβ alone or TCRβ + pTαa both expressed intracellular TCRβ (Fig. 4, J and K). However, unlike pTαa, cotransfection with pTαb did not measurably reduce surface TCRβ expression (Fig. 4,C). The expression of both forms of pTα was confirmed by anti-HA reactivity, predominantly of intracellular protein (Fig. 4, H and I), and by Western blot (Fig. 4 M).
To confirm that the different effects of pTαa and pTαb on TCRβ expression were not due to the epitope tag, the experiment was repeated with nontagged pTα isoforms. Again, surface TCRβ expression was invariably reduced by coexpression with pTαa but not pTαb. These different capacities of pTα isoforms and TCRα to regulate surface TCRβ expression appear consistent with the expression of TCRβ in vivo. Thus, in double-negative thymocytes, pTαa is in excess and surface TCRβ expression is barely detectable (17); in β-only cells, pTαb is in excess and surface TCRβ expression is measurable but low (Figs. 1 and 2); and in mature T cells, TCRβ pairs with TCRα, and surface TCRβ expression is high. Although surface pTα expression in cotransfected cells was difficult to detect by FACS (Fig. 4, E and F), the significance of this is unclear, since the accessibility of the HA epitope in any of the surface complexes is unknown.
To test further whether pTαa and pTαb had distinct biologic effects, we examined TCR-mediated signaling in a panel of 12 cell lines, stably transfected with combinations of TCRβ and pTα (Table I). None of the cell lines showed significant levels of surface CD3-TCR expression, but the expression of various components was readily detected by RT-PCR (Table I). Anti-CD3ε monoclonal 2C.11, which strongly activates cells expressing stable TCR complexes, provoked significant IL-2 release from only 3 cell lines (β/300.2; T6 and T10), each of which expressed high amounts of both TCRβ and pTαb RNA (Table I). No cells expressing TCRβ alone (β8, β9); TCRβ with pTαa (β/600.12; T2); pTαa alone (β/600.3) or pTαb with little or no TCRβ (β/300.3; β/300.5; T3; T9) responded strongly to anti-CD3 stimulation. Hence, in the absence of TCRα, strong signaling from the CD3 complex depended on TCRβ and pTαb.
|Transfectant .||pTαb .||pTαc .||TCRβ .||IL-2 (ng/ml)c .|
|Transfectant .||pTαb .||pTαc .||TCRβ .||IL-2 (ng/ml)c .|
Shading indicates functionally responsive cell lines.
+, indicates bright ethidium-stained band after one round of PCR; low, very faint band after one round of PCR (>10-fold less than +); −, undetectable.
IL-2 was measured at 48 h poststimulation.
In sum, our data define biochemical and functional differences between pTαa and pTαb. pTαa is strongly expressed in the thymus and more weakly in some peripheral cells; it retains significant amounts of TCRβ intracellularly, consistent with which it is poor at establishing CD3-associated cell surface signaling. pTαb is expressed in the thymus but more strongly in some peripheral cells; it does not obviously retain TCRβ intracellularly, but rather enhances the capacity of transfected TCRβ to form signaling-competent, CD3-associated complexes. Based on these data, one should consider the possibility that more than one type of pre-TCR complex exists, with either opposing or complementary effects on the cells that express them.
The suppression of surface TCRβ expression by pTαa might seem to support the hypothesis that the pre-TCR transmits signals from an intracellular compartment (3, 18). Alternatively, a role of pTαa may be to limit the expression of active, cell surface, pre-TCR complexes that may contain pTαb. This proposal is rooted in the evidence that low/moderate avidity interactions mediated by the TCR can activate thymocytes, while high avidity interactions can induce apoptosis.
An indication that pTαb can regulate thymocyte development in vivo is provided by two gene-targeted mutations of pTα. One, generated by deletion of the transmembrane domain and the pairing residue for TCRβ, inhibits essentially all β-selection of cells (15); but the other, in which only pTα exon 2 was disrupted, is leaky with regard to β-selection and allelic exclusion (19). It seems possible that pTαb expression was retained in the latter animal and that this promoted thymocyte progression, albeit inefficiently.
Finally, the different properties of pTαa and pTαb are reminiscent of the distinct biologic effects of the products of regulated alternative splicing at the IgCμ locus at different stages of B cell maturation. One product facilitates IgM functioning as a signaling-competent surface receptor that promotes B cell maturation, while the other acts as part of a secreted Ag-binding complex (20).
We thank T. Taylor, E. Hoffman, N. Douglas, W. Pao, R. McCord, and J. Silas.
Support was provided by National Institutes of Health Grant GM37759 (A.C.H.), and by a fellowship (Ministerio de Educación y Ciencia) from the Spanish government to D.F.B.
Abbreviations used in this paper: CD3, cluster of differentiation Ags 3; FACS, fluorescence-activated cell sorter; nt, nucleotide; bp, base pairs; HA, haemagglutinin.