The generation of Rosa26-loxP-stop-loxP-HEBAlt-HA (ALT-Tg) mice was performed by Ingenious Targeting Laboratories (Ronkonkoma, NY). Hybrid (129/SvEv x C57BL/6) embryonic stem cells (ESCs) were targeted and microinjected into C57BL/6 blastocysts. Resulting chimeras were mated to WT C57BL/6N mice to generate F1 heterozygous offspring, and those with germline integration were backcrossed for six generations with C57BL/6 mice to fix them on the C57BL/6 background. To induce expression in all hematopoietic cells, ALT-Tg mice were crossed to Vav-iCre mice (Jax; 008610)91. All experiments were conducted using 6-8 wk old littermate controls. Mice were maintained at the Sunnybrook Research Institute and all protocols were approved by the Animal Care Committee.

HEBAlt (NM_001253864.1) and HEBCan (NM_011544.3) cDNAs were cloned into pCMV-HA or pCMV-myc expression vectors using the KpnI and SalI cloning sites. Site-directed mutagenesis performed on the pCMV-HEBAlt-HA plasmid to generate the mutant constructs using the Agilent QuikChange Site-Directed Mutagenesis Kit or Q5^® Site-Directed Mutagenesis Kit. Mutations were confirmed by cloning the inserts into Top10 Competent E. coli cells, followed by plasmid DNA extraction and Sanger Sequencing (TCAG; SickKids, Toronto, Canada).

Cells lines were cultured under standard conditions at 37°C with 5% CO2. Adherent cells were trypsinized (0.25% Trypsin) for passaging. Both adherent and suspension cells were centrifuged at 550 g for 5 minutes before re-plating in fresh media. HeLa cells, HEK293T cells, and Baby Mouse Kidney (BMK) cells (deficient for Bax and Bak, kind gift from David Andrews, Sunnybrook) were cultured in DMEM supplemented with 10% FBS and antibiotics (100 mg/mL penicillin and 100 U/mL streptomycin). Jurkat cells were cultured in RPMI supplemented with 10% FBS and antibiotics. SCID.adh cells were cultured in RPMI supplemented with 10% FBS, antibiotics, 1% non-essential amino acids, 1% sodium pyruvate, and 50mM β-mercaptoethanol. OP9-DL4 co-cultures were seeded with LSK (Lin-Sca1+ckit+) cells sorted from bone marrow of WT or ALT-Tg mice, or with GFP+ LSK cells sorted from retrovirally transduced bone marrow, as previously described (35).

Cells were seeded 18 h before transfection, and equal amounts of DNA were transfected into each well using Lipofectamine 3000 Transfection Reagent (Invitrogen; Cat. L3000008) or FuGene HD Reagent (Promega; E2311) in OPTI-MEM. Analysis was performed 24 h post-transfection.

For immunoprecipitation experiments, Jurkat cells were transfected with 5 μg of vector DNA by Neon Transfection System based on manufacturer’s protocols. After washed with PBS, cells were resuspended in R Buffer to reach a concentration of 2 ×10⁷ cells/ml. Using 100 μl Neon tips, cells were electroporated with the parameter of 1350V, 10 ms, three times, in a Neon tube containing E₂ Buffer. After the pulse, cells were quickly transferred into a 6-well-plate with 3 ml RPMI supplemented with 10% FBS and antibiotics and cultured for 2 days.

Transcriptional activity was assessed by performing Dual Luciferase Reporter (DLR) Assays (Promega; Cat. E1910). The following constructs were co-transfected into cells in 24-well plates using FuGene HD transfection reagent (Promega; Cat. E2311): Renilla Luciferase, 8X-E-box Firefly luciferase reporter construct (pGL3/4 vector) (33), and pCMV-HA HEB expression constructs. Cells were lysed 24 h post-transfection using lysis buffer provided by the DLR assay kit, and luminescence was measured using the BioTek Synergy H1 Hybrid Reader. Firefly Luciferase Units were normalized to Renilla Luciferase Units to calculate Relative Luciferase Units (RLU).

GFP (empty vector negative control), HEBAlt (ALT), FFF, EEE, and HEBCan (CAN) retrovirus- producing GP+E cell lines were generated using pMIG-IRES-GFP, pMIG-HEBALT-HA-(WT/FFF/EEE)-IRES-GFP, and pMIG-HEBCAN-HA-IRES-GFP vectors, as previously described (36). BMK cells were co-cultured with the retrovirus-producing cell lines overnight (18h) in media containing Polybrene (10mg/μL). GFP+ cells were sorted by flow cytometry and expanded in culture to create stably expressing cell lines.

SCID.adh cells and HeLa cells, or HEK293T cells transfected with expression constructs using Lipofectamine 2000 one day earlier, were harvested using Pierce lysis buffer supplemented with protease (Thermo Scientific, 78425) and phosphatase inhibitors (EMD Millipore, 4906845001). Lysates were incubated on ice for 30 min, with vigorous vortexing every 10 min, and then pre-washed with Pierce Protein A/G agarose slurry for 1 h at 4°C to block non-specific binding. HA-tagged proteins were subjected to IP at 4°C overnight using antibody-agarose beads (Clontech, 631207) followed by three washes with Pierce lysis buffer or were collected anti-HA-conjugated magnetic beads followed by TBST washes (Thermo Scientific, 88836).

Cells were lysed using RIPA buffer with protease inhibitors (Halt Protease Inhibitor Cocktail (100X); Thermo Scientific; Cat. 78430). Protein concentrations were measured using the Pierce BCA Protein Assay Kit (Thermo Scientific; 23225) and equal amounts were used in SDS-PAGE analysis. Lysates were denatured in SDS loading buffer (containing DTT), heated at 100°C for 5 min, and loaded onto acrylamide gels. Sizes were determined using the PageRuler Plus Prestained Protein ladder (Thermo Scientific, 10 to 250kDa). Samples were transferred from the gel onto a PVDF membrane (Biorad TransBlot Turbo) by the semi-dry transfer method using the TransBlot Turbo Machine (Biorad), and the membrane was blocked overnight at 4°C on a shaker in 1X TBST (Tris-buffered saline with Tween 20 at 0.1% v/v) with 5% skim milk to eliminate non-specific binding. The next day, the blots were probed with primary antibodies in 1X TBST with 5% BSA, washed, and probed with HRP-conjugated secondary antibodies. After washing in TBST, the blot was visualized using the Clarity ECL kit (Biorad) and the Fusion Fx Chemiluminescence and Fluorescence Imager (Vilber Lourmat). Image quantification was performed using ImageJ software. The primary antibodies used in these studies were the HA-Tag polyclonal antibody (Clontech, 631207), pan anti-HEB antibody (Santa Cruz Biotechnology, A-20, sc-357), anti-GAPDH (mouse, EMD Millipore, MAB374), anti-Tubulin (mouse, SCBT; sc-69970), and anti-Alt (in-house). The secondary antibodies were HRP-conjugated Goat anti-Rabbit (Invitrogen, 626120) and Goat anti-Mouse (BioRad, 1706516). The anti-phosphotyrosine antibody (4G10 Platinum) was directly conjugated to HRP (Sigma-Aldrich 16-316).

For the cyclohexamide experiments, cells were treated with 300 μg/mL cyclohexamide (CHX) to block translation elongation 24 h post-transfection. The cells were lysed at 0 h, 0.5 h, 1 h, and 2 h post-CHX-treatment, and lysates were analyzed by Western blotting. Band densities were quantified using ImageJ.

Jurkat cells were co-transfected with ALT, FFF, EEE or pCMV-HA (negative control) and the 8X E-box luciferase reporter and treated with DMSO (no Ruxolitinib) or 1 mM of the pan-JAK inhibitor Ruxolitinib (Invitrogen, tlrl-rux) for 4 h. Dual luciferase assays were conducted, and the data was depicted as relative luciferase units (RLU) normalized to untreated ALT.

HEK293 cells transfected with ALT, CAN, EEE, FFF, TR, or empty vector, or whole thymocytes from littermate ALT-Tg mice with or without Vav-Cre, were lysed using RIPA lysis buffer and subjected to IP using the Pierce™ HA-Tag Magnetic IP/Co-IP Kit (ThermoFisher Scientific; 88838) and anti-HA antibodies. Samples were trypsin-digested and analyzed by mass spectrometry at the SPARC BioCentre facility using a Thermo Scientific Q Exactive HF-X hybrid quadrupole-Orbitrap mass spectrometer (SickKids, Toronto). Scaffold (version Scaffold_5.0.1, Proteome Software Inc., Portland, OR) was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95.0% probability. Peptide Probabilities from Sequest (XCorr Only) and MS-Amanda Proteome Discoverer were assigned by the Scaffold Local FDR algorithm. Peptide Probabilities from X! Tandem were assigned by the Peptide Prophet algorithm (37) with Scaffold delta-mass correction. Protein probabilities were assigned by the Protein Prophet algorithm (38). A filter of 95% protein identity, 95% peptide identity, and a minimum of 3 spectra were applied to all samples, and then proteins present in the HA negative control were removed.

Scaffold PTM (Proteome Software, Portland, OR) was used to annotate PTM sites derived from MS/MS sequencing results using the site localization algorithm developed by Beausoliel et al. (39). MS/MS spectra identified with modified peptides were identified and AScore values and site localization probabilities were calculated to assess the level of confidence in each PTM localization.

Anti-ALT antibodies were made by Abcam (Cambridge, U.K). A peptide corresponding to the Alt domain coding region was synthesized, conjugated to KHL, and injected into rabbits. After several rounds of boosting, the rabbits were exsanguinated to provide polyclonal stocks. Specificity was confirmed by Western blotting of HEK293T cells transfected with either HEBAlt- or HEBCan- expressing constructs ( Figure S2A ), and Rag2-/- thymocytes, which express abundant HEBAlt protein ( Figure S2B ).

Data was analyzed using Prism software (GraphPad). Statistical significance was determined by comparing two sets of data with the unpaired two-tailed Student’s t test, where a p value of less than 0.05 fulfilled the criteria of statistical significance. The variation between biological replicates of the same group was depicted by graphing the standard error of the mean as error bars. All data shown are reflective of at least two independent experiments. Results of multiple experiments were pooled, normalized, graphed, and shown as individual data points.

We have previously shown that the activity of HEBAlt on an E-box promoter-reporter luciferase construct was lower than HEBCan-induced transcriptional activation (33). This could have been due to the absence of the HEBCan-specific N-terminus, or due to unique properties of the Alt domain. To distinguish between these possibilities, we generated HA-tagged HEBAlt mutant truncation expression constructs using naturally occurring methionines as start codons ( Figures 1C, D ). We removed the first third of the Alt domain (MET1), the first two-thirds, (MET2), the entire Alt domain (TR), or the Alt domain and part of exon 9, which is shared between HEBAlt and HEBCan (MET3) and tested their ability to transactivate the reporter construct as compared with WT (ALT) HEBAlt in HEK293T cells. All constructs were expressed at comparable protein levels ( Figure 1E ). ALT induced luciferase activity was above background levels, in agreement with previous studies (33) ( Figure 1F ). The MET1 and MET2 constructs displayed comparable activity to full length ALT, but the TR and MET3 constructs, which lacked the Alt domain completely, had a ~3-fold decrease in activity. These results indicated that the last third of the Alt domain is required for effective transcriptional activation from a multimerized E box site by HEBAlt.

Closer inspection of the amino acid residues in the Alt domain revealed an abundance of tyrosine (Y) residues, including a triple tyrosine (YYY) motif. To evaluate whether the Alt domain could respond to signaling through tyrosine kinase-mediated pathways, we generated mutant HEBAlt constructs in which the YYY motif was replaced by EEE, imparting a negative charge and mimicking phosphorylation, or FFF, which cannot be phosphorylated ( Figure 2A ). We transfected these constructs into HEK293T cells and confirmed that EEE and FFF were expressed at similar protein levels to ALT ( Figure 2B ). To evaluate the ability of these constructs to activate transcription from the 8X E box reporter construct, we co-transfected them into BMK (Baby mouse kidney) cells. Consistent with our previous findings, luciferase assays showed that ALT had lower levels of activity than HEBCan (CAN) and E2A (E47) ( Figure 2C ). Strikingly, EEE exhibited a ~10-fold increase in activity over ALT, to levels that surpassed CAN. By contrast, FFF activity was ~2-fold lower than ALT. We also tested the activity of these factors in HEK293T cells on the Ptcra promoter, which drives expression of the gene encoding pTα in DN thymocytes (40) ( Figure 2D ). Under these conditions, ALT and FFF activity were both low, but EEE approximated HEBCan activity. These results suggest that phosphorylation of the YYY motif in the Alt domain might enable HEBAlt to induce transcription more strongly than HEBCan.

To evaluate whether the FFF acted as a dominant negative form of HEB, we assessed whether mixing FFF with ALT, CAN, or E2A would decrease the overall activity when co-transfected into HEK293T cells with the reporter construct. We found that FFF did not interfere with CAN or E2A transcriptional activity but induced a small increase in ALT activity ( Figure 2E ). Next, we asked whether mixing the EEE mutant with each of the wildtype E proteins would impact their activity. No major impact on CAN or E47 activity was observed with the addition or either EEE, suggesting that a negative charge on one E protein within an E protein dimer is sufficient to drive enhanced transcriptional activation ( Figure 2F ).

To formally assess whether ALT proteins can form homodimers, we generated a myc-tagged ALT construct ( Figure 3A ). ALT-HA and ALT-myc were co-transfected into HEK293T cells, and antibodies were used to immunoprecipitate (IP) ALT-myc. Western blot analysis of HA-tagged constructs clearly showed that ALT-myc and ALT-HA can form a complex ( Figure 3B ), consistent with the existence of HEBAlt homodimers. Next, we assessed whether JAK tyrosine kinases could enhance HEBAlt activity. ALT or mutant constructs and 8X-E box reporter constructs were co-transfected into Jurkat T cells, which provide a background of low constitutive JAK signaling (41). ALT, FFF, EEE, or control empty vector (HA) transfected Jurkat cells were cultured with or without the JAK inhibitor Ruxolitinib for two days, followed by luciferase assays ( Figure 3C ). As before, FFF displayed lower activity than ALT. In contrast, EEE exhibited a ~20-fold higher activity than ALT ( Figure 3D ). Importantly, JAK inhibition decreased ALT activity but did not significantly affect the activity of FFF or EEE. These results indicate that modulation of JAK function can regulate HEBAlt activity in a YYY-dependent manner.

To assess whether HEBAlt could be phosphorylated, we used anti-HEB (A20) to IP endogenous HEB factors from HeLa and SCID.adh cells (42) ( Figure 3E ). Anti-HEB was used to IP HEB factors, and the Western blot was probed using anti-phosphotyrosine antibodies (4G10). One band around 60 kDa was detected, consistent with phosphorylation of HEBAlt tyrosine residues. To verify the specificity of the endogenous bands, we transfected HA-tagged ALT, CAN, or FFF constructs into Jurkat cells and cultured them for two days. The transfected Jurkat cells were subjected to IP with anti-HA antibodies ( Figure 3F ) and Western blots were probed with anti-phosphotyrosine antibodies. In the ALT and CAN transfected Jurkat cells, we observed two bands, one at the size of HEBCan (upper band; ~100 kD) and another at the size of HEBAlt (lower band; ~60 kD), whereas the FFF sample lacked bands indicating phosphorylated tyrosine. To assess whether the presence of two bands in the each of the ALT and CAN samples might be due to enrichment of complexes containing both HEBAlt and HEBCan, we performed co-IPs in HEK293T cells with myc-tagged and HA-tagged HEBAlt, HEBCan, E2A, and MyoD ( Figure 3G ). MyoD is a Class II bHLH factor that binds more strongly to E proteins than they do with each other. Our results showed that HEBAlt can form heterodimers with HEBCan, E2A, and MyoD. Therefore, the presence of a phosphorylated CAN band in the ALT sample and a phosphorylated ALT band in the CAN sample suggests that both proteins were precipitated as components of complexes containing HA-tagged proteins. Moreover, these bands suggest that both HEBCan and HEBAlt, whether exogenous or endogenous, can be tyrosine phosphorylated, whereas FFF cannot. Moreover, the lack of bands in the FFF lane suggest that it may not be able to form stable dimers with either HEBAlt or HEBCan, providing a partial explanation for its decreased function.

To directly identify phosphorylated residues on HEBAlt, we transfected HEK293T cells with HA-tagged ALT, FFF, EEE, CAN, and TR. After 24 h, protein lysates were generated in the presence of phosphatase inhibitors and subjected to IP with anti-HA. The precipitated proteins were subjected to mass spectrometry sequencing, and ScaffoldPTM was used to identify phosphorylated residues on ALT, FFF, EEE, CAN, and TR ( Figure 4A ). Five unique peptides were recovered from this set of samples, all of which were located downstream of the ALT/CAN junction ( Figures 4B, C ). Spectra consistent with phosphorylation were detected on two serines, two threonines, and one tyrosine. These were sparsely present among the peptides and were differentially represented in each sample ( Figure 4C ). However, we also detected a paralog of HEBAlt, E2-2Alt (43), likely due to co-immunoprecipitation with HEBAlt. The Alt domains of E2-2 and HEB differ by two amino acids ( Figure 4D ), allowing unequivocal identification. This spectrum showed a +80 shift on the last tyrosine of the YYY motif indicating the presence of a phosphate group ( Figure 4E ). It should be noted that E2-2Alt spectra represented an endogenous protein that is not overexpressed, further supporting the capacity of the YYY motif in E protein ALT variants to be directly phosphorylated on at least one tyrosine residue.

Given the lack of reagents available for studying HEBAlt, we generated several new tools to move our studies into the context of T cell development. The first was an antibody that detects the Alt domain ( Figure S2 ). These antibodies (anti-ALT) worked well for detection of HEBAlt using Western blots but not for immunoprecipitations. Therefore, we designed a new mouse model with inducible expression of HEBAlt. To enhance the versatility of these mice, we inserted an HA-tagged HEBAlt construct into a loxP-stop-loxP cassette driven by the Rosa26 constitutively expressed promoter to generate ALT-Tg mice ( Figure 5A ). In these mice, the HEBAlt-HA transgene is silent until the stop cassette is removed by Cre-mediated excision, enabling induction of HEBAlt-HA expression in a stage- and lineage-specific manner. We bred these mice to Vav-iCre mice to induce deletion in all hematopoietic cells, and generated ALT-Tg^Vav-iCre (ALT-Tg) mice, which were compared with ALT-Tg (WT) littermates containing the transgene in the absence of Cre.

The ALT-Tg mice had no obvious defects at the level of gross morphology. The adult thymus had normal cellularity ( Figure 5B ) and undisturbed proportions of DN, DP, and SP thymocytes at steady state ( Figures 5C, G ), as well as a normal distribution of DN subsets ( Figures 5D, H ). We also evaluated the percentages of γδ T cells ( Figures 5E, I ), and the distribution of immature versus mature cells within the γδ T cell subset according to expression of CD24 ( Figures 5F, I ). No significant differences were observed. We next undertook a more in-depth analysis of thymocyte subsets that normally do not express HEBAlt at the mRNA level ( Figure 6 ). Gating on the CD8+CD4- subset ( Figure 6A ) revealed that the proportion of immature (CD24+ TCRβ-) cells to mature (CD24-low TCRβ+) cells was higher in ALT-Tg mice than in WT mice, with a corresponding decrease in the percentage of mature CD8+ SPs in the ALT-Tg mice ( Figures 6B, G ). This was not accompanied by a decrease in the percentage of DP cells, suggesting the increase in CD8+ ISP did not indicate a developmental block at the DN to DP transition. To assess the ratio of mature CD4 to CD8 cells, we first gated on TCRβ+ CD3+ cells ( Figures 6C, H ), and then on CD24- cells within that population ( Figure 6H ). These two populations were present at similar frequencies in WT and ALT-Tg. We next evaluated the frequencies of CD4 and CD8 cells within the mature (TCRβ+ CD3+ CD24-) population and found no differences between WT and ALT-Tg mice ( Figures 6D, I ). To assess whether positive selection was impaired in ALT-Tg mice, we examined CD24 and CD69 expression within the TCRβ+ CD8+ ( Figures 6E, J ) and the TCRβ+ CD4+ ( Figures 6F, K ) populations. Thymocytes from ALT-Tg mice exhibited transient upregulation of CD69 accompanied by CD24 downregulation comparable to WT thymocytes in both the CD4+ and CD8+ SP subsets. Therefore, we found no evidence for alterations in positive selection in the ALT-Tg mice on the C57Bl/6 background with a polyclonal repertoire.

To confirm expression of the HEBAlt transgene in ALT-Tg mice, we measured HEBAlt mRNA and protein levels ( Figure 7 ). Although HEBAlt mRNA decreased at the DN3 to DN4 transition in the WT mice, as expected, it was sustained up to the DP stage in ALT-Tg thymocytes, as detected by anti-HA ( Figure 7A ). Western blots of sorted thymocyte subsets showed that the transgenic HEBAlt protein was also expressed up to the DP stage, as detected by anti-HA ( Figure 7B ). Surprisingly, we also found that endogenous HEBAlt protein, as detected by the anti-ALT antibody, was present in WT thymocytes subsets that had downregulated HEBAlt mRNA ( Figure 7C ). These observations suggest that HEBAlt protein stability may be enhanced across the β-selection checkpoint.

Given the discordance between HEBAlt mRNA and HEBAlt protein expression, we evaluated HEBAlt protein stability by performing cycloheximide chase experiments. BMK cells were stably transduced with HA-tagged HEBAlt or HEBCan to provide a non-T cell context for protein stability studies, and these were compared with HA-tagged ALT-Tg thymocytes. Cycloheximide was added to the cells to stop de novo protein synthesis, and then washed out (chase). Samples were taken at 0, 0.5, 1, and 2 h after chase, and protein levels were evaluated using Western blots probed with anti-HA and anti-GAPDH. In BMK cells, HEBAlt and HEBCan were expressed at comparable amounts at t=0, but HEBAlt protein levels dropped more precipitously than HEBCan levels, indicative of decreased stability ( Figure 7D ). By contrast, HEBAlt protein appeared more stable in thymocytes than in BMK cells ( Figures 7F, G ), and the levels were statistically indistinguishable from HEBCan in BMK cells after 2 hrs ( Figure 7H ). EEE showed only a minor improvement in HEBAlt stability in BMK cells ( Figures 7E, H ), suggesting that factors other than YYY phosphorylation were at least partially responsible for this feature of HEBAlt biology.

We have shown in previous studies that forced expression of HEBAlt was inhibitory to the development of non-T cell lineages in vitro (33). We therefore analyzed the proportions of myeloid cells (CD11b⁺), B cells (CD19⁺), and the ratio of immature (B220^int CD19 ^int) to mature B cells (B220^hi CD19^hi) in the bone marrow ( Figure 8A ). We found no significant differences between the WT and ALT-Tg mice. Absolute numbers of bone marrow cells and splenocytes were also indistinguishable ( Figure 8B ). We also analyzed the percentages of B cells, the ratios of T to B cells, and the ratios of CD4 to CD8 T cells within the CD3⁺TCRβ⁺ population in the spleen. There was a slight decrease in the ratio of T cells to B cells in the ALT-Tg spleen, but otherwise no major perturbations were observed ( Figures 8A, C ). Given our observation of differential HEBAlt protein stability in transduced BMK cells versus ALT-Tg thymocytes, we next analyzed the expression of the HEBAlt transgene in spleen and bone marrow subsets from ALT-Tg mice. Thymocytes, RBC-depleted bone marrow cells and splenocytes that had been sorted to obtain T (CD3+), B (CD19+), and myeloid CD11b+ (myeloid)-enriched samples, and these samples were subjected to Western blot analysis with anti-HA. Transgenic HA-tagged HEBAlt protein was expressed strongly in ALT-Tg thymocytes but was undetectable in ALT-Tg bone marrow ( Figure 8D ). Splenic B and T cells had similar levels of HEBAlt-HA protein expression, whereas none was apparent in the myeloid fraction. However, qRT-PCR showed that ALT-Tg cells had higher overall levels of HEBAlt mRNA than their WT littermates in all three tissues, confirming expression of the transgene at the RNA level ( Figure 8E ). Interestingly, HEBAlt-HA protein levels were much stronger in thymocytes than they were in the splenic T or B cells. These results suggest that the thymus provides a protective environment for HEBAlt stability, that this protection fades in mature T cells, and that it is absent in myeloid cells.

Considering these observations, we reasoned that the dynamics of steady state T cell development might mask changes caused by elevated levels of HEBAlt in thymocytes in the ALT-Tg mice. We therefore sorted bone marrow Lin-Sca1+Kit+ (LSK) progenitors from WT and ALT-Tg mice and placed them in OP9-DL4 co-cultures ( Figure 9A ), which allowed us follow T cell differentiation over time ( Figures 9B, C ). We found that ALT-Tg cells were subject to developmental delays compared to their WT counterparts, both at early and at later timepoints ( Figure 9C ). No major changes in cellularity were noted ( Figure 9D ), suggesting that these changes were not due to defects in cell growth or survival. Instead, these results indicate that disrupting the normal balance of HEBAlt in T cell precursors partially inhibits T cell development.

We next evaluated the impact of forced expression of the EEE mutant on T cell development. WT bone marrow precursors were transduced with MIGR1-based retroviral vectors encoding HEBAlt (ALT), EEE, or empty vector. GFP+ LSK cells were sorted and placed in OP9-DL4 cell co-cultures for varying amounts of time ( Figure 9E ). Interestingly, progression of ALT-transduced cells from the DN2 to DN3 stages was similar to that seen in WT cultures, whereas the EEE cells exhibited a strong block at this transition ( Figure 9F ). However, both ALT and EEE cultures exhibited a partial block in the DN3 to DN4 transition and failed to progress to the DP stage. Therefore, dysregulation of HEBAlt expression and/or activity can interfere with T cell development.

We next set out to determine whether the HEBAlt-HA protein was tyrosine phosphorylated in ALT-Tg thymocytes ( Figure 10 ). Unfractionated thymocytes from ALT-Tg or WT littermates were isolated and subjected to co-IP with anti-HA antibodies, followed by Western blot analysis of tyrosine phosphorylation ( Figure 10A ). These results clearly showed that transgene-derived HEBAlt can be tyrosine phosphorylated in ex vivo thymocytes from ALT-Tg mice. To identify putative protein partners of HEBAlt in this context, we subjected WT and ALT-Tg thymocytes to IP using anti-HA antibodies. Two independent experiments were performed, each of which contained biological duplicates ( Figure 10B ). The IP fractions were subjected to trypsin-mediated digestion, followed by mass spectrometry sequencing. The data was analyzed using Scaffold5, with a cut-off of 95% protein identity and 95% peptide identity. Spectra represented in the littermate controls were used to screen out background hits. Twenty-two proteins represented by at least one spectrum all four replicates were identified ( Figure 10C ). HEBAlt was represented in all four samples for a total of 10 spectra but was not found at all in the negative control samples, consistent with IP-mediated enrichment of HEBAlt-HA. The two most enriched proteins were STAT1, a transcription factor that is phosphorylated and activated by JAK, and Xpo1, which is involved in nuclear export. These results suggest that HEBAlt, like STAT1, can be shuttled between the nucleus and the cytoplasm by Xpo1 (44) in response to cytokine signaling, and that HEBAlt and STAT1 may be coordinately regulated.