A novel Bcl-x isoform connected to the T cell receptor regulates apoptosis in T cells

Abstract

We define a novel Bcl-x isoform, Bcl-x gamma, that is generated by alternative splicing and characterized by a unique 47 amino acid C-terminus. Bcl-x gamma is expressed primarily in thymocytes, where it may depend on an interaction between the TCR and host MHC products, and in mature T cells, where its expression is associated with ligation of the T cell receptor. Overexpression of Bcl-x gamma in T cells inhibits activation-induced apoptosis; inhibition of Bcl-x gamma, after stable expression of Bcl-x gamma antisense cDNA, enhances activation-induced apoptosis. In contrast to other Bcl-x isoforms, cells that fail to express Bcl-x gamma after CD3 ligation undergo programmed cell death, while activated T cells that express Bcl-x gamma are spared. Identification of Bcl-x gamma helps provide a molecular explanation of T cell activation and death after antigen engagement.

Figure 1. Molecular Features of Bcl-xγ

(A) cDNA sequence and predicted protein sequence of Bcl-xγ (GenBank accession number U51277). The position of the ORF encoding Bcl-xγ protein is indicated, and the predicted amino acid sequence (numbered to the right of each sequence) is given below the nucleotide sequence. Arrow, the splice site at which the unique Bcl-xγ C-terminus (continuously underlined) begins; long-dashed underline and a short-dashed underline, the conserved BH1 and BH2 domains, respectively. (B) Schematic comparison of murine Bcl-x isoforms: Bcl-xL, Bcl-xS, Bcl-xβ, Bcl-xΔTM, and Bcl-xγ. Murine Bcl-x isoforms share a long N-terminal region (hatched). The BH1/BH2 region between splice sites 1 and 2 of Bcl-xS is shown. Splice site 2, used to generate Bcl-xβ and Bcl-xγ, is compared to the noncanonical splice site used to generate Bcl-xΔTM (located five residues downstream from splice site 2). The distinct C-termini of Bcl-xβ, Bcl-xΔTM, and Bcl-xγ, which show no homology with Bcl-xL and Bcl-xS C-terminal regions, also are shown, along with the positions of conserved BH1/BH2 domains. The position of the oligonucleotide probe used here to screen a mouse thymus cDNA library is indicated. (C) Hydrophobicity plot of Bcl-xL and Bcl-xγ. The hydrophobicity of Bcl-xL and Bcl-xγ was calculated using the GCG program based on Goldman’s (solid line) or Kyte-Doolittle’s (dashed line) algorithm. The amino acid sequence of Bcl-xL and Bcl-xγ is numbered above and below the hydrophobicity plots. Vertical line, the beginning of the isoform-specific C-terminal sequences of Bcl-xγ and Bcl-xL; vertical arrow, the hydrophobic segment representing a putative membrane anchoring domain of Bcl-xL. (D) Alignment of the specific region of mouse Bcl-xγ with the consensus sequence of the ankyrin-like domain sequence. The consensus sequence of ankyrin-like domain is shown. Amino acid symbols: bold and underlined, the six residues in Bcl-xγ that are identical to ankyrin consensus residues; bold, the two highly conserved residues in Bcl-xγ underlined, three residues in Bcl-xγ with homology to the ankyrin consensus residues. Dashed lines, the correct spacing of the ankyrin-like motif.

Figure 2. Analysis of Protein Products of Bcl-xL, Bcl-xS, Bcl-xβ, and Bcl-xγ on 12%SDS-PAGE after In Vitro Transcription and Translation

The apparent size of Bcl-xγ protein after in vitro transcription/translation is consistent with the size predicted from its ORF, since the Bcl-xγ protein product migrates at a position similar to Bcl-xL (233 aa residues) (see Figure 1B) and more slowly than the Bcl-xβ protein (209 aa) and Bcl-xS (170 aa). Molecular weight standards (kilodaltons) are indicated at the left margin.

Figure 3. Tissue Expression of Bcl-xγ

(A) Expression of Bcl-xγ in different murine tissues. RT-PCR products were analyzed on 1% agarose gels stained with ethidium bromide. Molecular weight markers are indicated at the right margin of the gels. The amount of amplified Bcl-xγ fragment and amplified β-actin fragment both increased in proportion to the number of thermal cycles, and the PCR amplification efficiency of Bcl-xγ primer pair was similar to that of the β-actin primer pair used as an internal control for RT-PCR for each tissue cDNA. Levels of β-actin from these tissues in the linear phase of PCR before saturation of amplification indicated that the actual amount of cDNA from each tissue used in this analysis was similar. A Bcl-xL fragment was also amplified to determine potential differences in tissue expression of Bcl-x isoforms. As reported previously, Bcl-xL was detected in every tissue examined. In contrast, the Bcl-xγ fragment was detected in thymus (T) and eye (E) but not in brain (B), heart (H), intestine (I), kidney (K), or liver (L). The specificity of the Bcl-xγ fragments from the tissues scoring as positive was confirmed by a Southern blot hybridization using a [³²P]dCTP-labeled Bcl-xγ–specific probe that did not overlap with either primer of Bcl-xγ used in the RT-PCR. (B) Expression of Bcl-xγ in lymph nodes of normal, nu/nu, and Rag2^−/− mice. Analysis of cDNA indicated that Bcl-xγ is amplified in lymph nodes of normal but not nu/nu donors (PCR amplification of β-actin fragment served as control). To measure the linear phase of PCR amplification, aliquots of reaction mix were tested at 27, 31, and 35 cycles for β-actin and at 37, 41, and 45 cycles for Bcl-xγ and were visualized by agarose gel electrophoresis and ethidium bromide staining. The specificity of Bcl-xγ fragments amplified was confirmed by a Southern blot hybridization as described in (A); the blot was overexposed to detect any weak hybridization. (C) Expression of Bcl-xγ in thymocyte subpopulations. Expression of Bcl-xγ was not detected in DN (CD4⁻CD8⁻) thymocytes (either CD44⁺ or CD44⁻) from normal, Rag2^−/−, or TCRβ^−/− donors. Bcl-xγ was detected in normal thymocytes, and DP (90%) and single-positive thymocytes (after depletion of DN cells) according to RT-PCR. PCR analysis of thymocytes from MHC^−/− mice also showed that expression of MHC products are required for expression of Bcl-xγ but not Bcl-xL and Bcl-xS. The specificity of Bcl-xγ fragments amplified was confirmed by Southern blot hybridization, as described (A); the blot was overexposed to detect weak hybridization in PCR lanes that did not show apparent hybridization (negative groups).

Figure 4. Expression of Bcl-x Isoforms in Activated T Cells

(A) Expression of Bcl-x isoforms in activated T cells after CD3 ligation. (Left) After exposure of O3 T cells to plate-bound anti-CD3 antibody for the indicated intervals, total RNA was extracted, and RT-PCR amplification by an IL-2 and IFNγ fragment indicated activation as early as 4 hr. Expression of all of Bcl-x isoforms was up-regulated between 4 and 8 hr after stimulation, but the increase in Bcl-xγ expression was the most dramatic and returned to normalized levels more quickly. Because of variations in the levels of expression among the different Bcl-x isoforms, the number of PCR cycles used to measure the linear phase of PCR amplification varied among the isoforms but was constant at all time points for each given isoform. The specificity of Bcl-xγ fragments amplified was confirmed by a Southern blot hybridization as described in Figure 3A; the blot was overexposed to detect weak hybridization in negative PCRlanes (negative groups). (Right) Agarose gels were scanned and quantitated using an IS-1000 digital imaging system, adjusting for exposure times so that the intensity of DNA fragment signals corresponded to the linear range of densitometric detection. To ensure that comparisons of cDNA levels in different samples were based upon the same amount of cDNA in each sample, the area under the densitometric peak of each sample was divided by the area under the β-actin densitometric peak for the corresponding sample. The ratios of Bcl-x isoform and control (IL-2, IFNγ) cDNA to β-actin cDNA for each sample are shown as relative densitometric units. (B) Expression of Bcl-x isoforms in activated T cells after IL-2 stimulation. (Left) After incubation of O3 T cells with 25 U/ml IL-2 resulting in levels of [³H]TdR incorporation that were similar to that obtained after CD3 ligation, total RNA was extracted. Bcl-xL, Bcl-xβ, and Bcl-xΔTM were up-regulated 4–24 hr after IL-2 treatment, but Bcl-xγ expression was not detectable. (Right) The PCR-amplified fragments on agarose gels were scanned, quantitated using the IS-1000 digital imaging system, and normalized as described above. The results indicate that signaling through IL-2 receptor does not up-regulate the expression of Bcl-xγ. (C) Specificity of a rabbit anti-Bcl-xγ antisera. (Top) Western blot analysis was performed to define the specificity of an antiserum raised in rabbits immunized with a GST-fusion protein containing the Bcl-xγ–specific C-terminus (outlined in bottom panel) (GST-Bcl-xγ-C). After absorption of antisera with a GST column, the IgG fraction reacts with GST-Bcl-xγ-C but not GST (left) and with GST-Bcl-xγ but not GST-Bcl-xL (right) at a 1:2500 final dilution. (Bottom) Schematic representation of three GST fusion proteins used in these studies (see Experimental Procedures and [B] for details). (D) Expression of Bcl-xγ protein in activated T cells after CD3 ligation detected by Western blotting Western blots were performed using the anti-Bcl-xγ IgG characterized in (C) at a final dilution of 1:2500 after 12% SDS-PAGE analysis of 10 mg/each of cell lysates prepared at the indicated times after exposure of RF33.70 cells (left) or splenic T cells (right). A protein that migrated at approximately 34 kDa (consistent with the apparent size of Bcl-xγ in Figure 5A) was detected within 12 hr of activation of a T cell hybridoma RF33.70 and within 24 hr of activation of purified splenic T cells.

Figure 5. The Effect of Overexpression of Bcl-xγ Sense or Antisense cDNA on T Cell Apoptosis

(A) Overexpression of Bcl-xγ protein in stable Bcl-xγ transfectants. Western blot analysis (right) using a rabbit anti-Bcl-xcommon region (1:1000 dilution) on blots of cell lysate (10 µg) from pRC/RSV (vector control) or pRC/Bcl-xγ CTLL-2 transfectants after 12% SDS-PAGE. Autoradiographed (left) in vitro–translated Bcl-xL and Bcl-xγ alone or together on the same 12% SDS-PAGE were used as migration markers for cellular Bcl-xL and Bcl-xγ. pRC/Bcl-xγ transfectants but not vector control transfectants display a band that reacts with anti-Bcl-x and comigrates with translated Bcl-xγ at approximately 34 kDa. (B) Influence of Bcl-xγ overexpression on T cell apoptosis. The cell cycle profile represents CTLL-2 cells that stably express the indicated constructs pRC/RSV (vector control), pRC/Bcl-xL, and antipRC/γC4 (Bcl-xγ) after activation by plate-bound anti-CD3. The distribution of cells between the G1, S, and G2/M phases of the cell cycle are shown; the abscissa indicates the relative cell number and the ordinate indicates DNA content based on PI staining. The numbers at top left represent the percentage of cells that display apparent DNA contents of less than diploid (subdiploid), corresponding to the subpopulation of apoptotic cells. These stable transfectants expressed similar levels of CD3 according to immunofluoresence, and all had similar baseline levels of apoptosis (4%– 10%; data not shown). Each panel represents one of triplicate wells; results are representative of three experiments. (C) Comparison of apoptosis of Bcl-xγ antisense (pRC/ γAS.A3) transfectants with vector control (pRC/RSV) transfectants. In experiments in which relatively low levels of apoptosis were obtained after incubation of transfectants on anti-CD3 coated plates, the degree of apoptosis was substantially elevated in activated pRCγAS.A3 antisense transfectants, compared to vector-transfected controls (pRC/RSV). Each panel represents one of triplicate wells; results are representative of three experiments.

Figure 6. Expression of Bcl-x Isoformsin O3 T Cell Clone Stimulated by Plate-Bound Anti-CD3 Antibody for 5 hr and Sorted by Flow Cytometry

(Left) After stimulation of the O3 T cell clone by plate-bound anti-CD3 antibody for 5 hr, cells were subjected to staining with Hoechst 33342 dye and PI. After dead cells were gated out, activated cells were sorted into two subpopulations by flow cytometry, Hoechst-high (apoptotic) and Hoechst-low. Total RNA from both Hoechst-positive and Hoechst-negative subpopulations was extracted as described in Experimental Procedures. RT-PCR amplification of IFNγ fragment and IL-2 fragment after stimulation by anti-CD3 in both Hoechst-high and Hoechst-low cells confirmed activation of O3. Bcl-xγ was selectively expressed in Hoechst-low cells but not in Hoechst-high (apoptotic) cells, while all other Bcl-x isoforms are expressed in both forms. Failure to detect Bcl-xγ in Hoechst-high cells did not result from degraded preparations of total RNA or cDNA, since β-actin and other Bcl-x isoforms were detected in these samples. (Right) The PCR-amplified fragments analyzed on agarose gel were scanned and quantitated using an IS-1000 digital imaging system, followed by normalization, as described above.