Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Immunol. Author manuscript; available in PMC 2010 February 1.
Published in final edited form as:
PMCID: PMC2757056

Themis controls thymocyte selection through regulation of T cell receptor-mediated signaling


Themis (Thymocyte expressed molecule involved in selection), a member of a family of proteins with unknown functions, is highly conserved among vertebrates. Here we found that Themis is expressed in high amounts in thymocytes between the pre-T cell receptor (TCR) and positive selection checkpoints, and in low amounts in mature T cells. Themis-deficient thymocytes exhibit defective positive selection, which results in reduced numbers of mature thymocytes. Negative selection is also impaired in Themis-deficient mice. A higher percentage of Themis-deficient T cells exhibit CD4+CD25+Foxp3+ regulatory and CD62LloCD44hi memory phenotypes than in wild-type mice. Supporting a role for Themis in TCR signaling, this protein is phosphorylated quickly after TCR stimulation, and is needed for optimal TCR-driven Ca2+ mobilization and Erk activation.


T cell development in the thymus follows an ordered progression from CD48 (double negative, DN, subdivided into stages DN1-4) through CD4+8+ (double positive, DP) to single positive (SP, CD4+8 or CD8+4) stages that depends on a complex interplay of signaling pathways1,2. In DN3 cells, VDJ rearrangement and expression of the T cell receptor (TCR) β-chain and its association with a pre-TCRα chain results in signaling that drives cells through the β-selection checkpoint into the TCRlo DP stage. Here, Vα-Jα rearrangement occurs and a mature αβ TCR is expressed on the cell surface. Sequential rearrangements that delete previous rearrangements are frequently required before a TCR is produced that has sufficient ability to interact with self MHC-peptide (pMHC) complexes to induce the positive selection differentiation program that leads to SP thymocyte differentiation3. Cells that do not receive this positive selection signal eventually die through lack of stimulation, whereas those cells whose TCR binds too strongly to self pMHC undergo activation induced apoptosis referred to as negative selection1,2.

Signaling through the TCR is tightly regulated during positive selection. Pre-selection DP cells express about 10% of the amount of cell surface TCR on post-selection TCRhi DP and SP cells, yet the pre-selection DP cells are more sensitive to TCR stimulation by weak ligands than are the TCRhi cells4,5. The low TCR expression on pre-selection DP cells is actively maintained by Src-family protein tyrosine kinase (PTK) signaling and ubiquitination-mediated degradation6,7, and these controls enforce TCR α-chain allelic exclusion8,9. Development through the positive selection checkpoint can be disrupted by mutation of elements of the TCR signaling cascade, such as the ZAP-70, Itk, Lck or Fyn kinases or the Vav guanine nucleotide exchange factor10,11. The regulation of signaling through the MAP kinase Erk is extremely important in distinguishing between positive and negative selection1215. Similarly, mobilization of Ca2+ in response to TCR stimulation is both regulated and related to the discrimination between positive or negative selection4,16,17.

Here we identified a novel gene and protein involved in thymocyte positive and negative selection. Two other groups have also identified this gene and protein, and in consensus we refer to it as thymocyte expressed molecule involved in selection (Themis)18,19. It is expressed in a tightly regulated manner during T cell development. It is predominantly expressed in late DN and DP thymocytes, and is down-regulated after positive selection. Themis is tyrosine phosphorylated within seconds after TCR stimulation. It is a member of a small gene family of unknown function, and is highly conserved amongst vertebrates. It has no known conserved domains other than a polyproline region. We showed that Themis plays an important role in regulating thymocyte development through TCR signaling and in particular through regulation of Ca2+ influx and phosphorylation of Erk.


Themis encodes a novel protein

A cDNA subtraction library was constructed using cDNA from Tcra−/− thymus subtracted with cDNA from Rag1−/− thymus20. A clone which exhibited differential expression between Tcra−/− and Rag1−/− thymocyte mRNAs was analyzed further (Fig.1). Full-length cDNA clones were isolated from a mouse thymocyte library and from a human Jurkat cell library. The main open reading frame of the mouse Themis gene encodes a protein of molecular weight 72.8 kDa. The human THEMIS protein is 81% identical (86.4% similarity) to the mouse protein (Fig.1a). The gene is present in public databases (Mouse gene #NM_178666, RIKEN cDNA E430004N04Rik; human NM_001010923), and orthologs are found across the mammals, and in birds (Gallus) and bony fish (Danio) (Supplementary Fig.1).

Figure 1
Themis structure, expression and localization in wild-type mice. (a) Alignment of predicted amino acid sequence of murine (Mm) and human (Hs) Themis proteins. Sequence alignments and comparisons were performed using Multalin46, and rendered using ESPript ...

Although Southern blot experiments showed only a single Themis-hybridizing band in mouse genomic DNA, suggesting a single copy gene with no close relatives (data not shown), BLAST analysis of the mouse genome revealed two paralogs: the “basement-membrane-induced” gene Icb-121, and the hypothetical protein 9130404H23Rik. Themis is more closely related to Icb-1 (32% identity) than to 9130404H23Rik (25% identity) (Supplementary Fig.2). The mouse Themis gene is on chromosome 10, location A4, and human THEMIS is at Chr 6q22.33 (C6orf190). The sequences of Themis, Icb-1 and 9130404H23Rik revealed no identifiable conserved domains by PFAM or SMART searches, but there is a highly conserved proline rich sequence (PRS) forming three overlapping PXXP putative SH3 domain-recognition motifs near the C-terminus (residues 555–563) (Fig.1a; all numbering refers to mouse protein). The PRS is conserved in Icb-1 and 9130404H23Rik (Supplementary Fig.2). A putative bipartite nuclear localization signal lies between residues 330–346.

Themis gene expression in T cell development

Themis gene expression was analyzed by Northern blot of different tissues (Fig.1b). Themis is expressed in the thymus and to a lesser extent in the spleen, but was not detectable in non-lymphoid tissues. Of the two major transcripts, the 5.7 kb transcript was the most abundant. It was readily detectable in a number of transformed cell lines of thymic origin as well, but was undetectable in transformed B cell lines (Supplementary Table 1). Reference to a public gene expression atlas database ( confirmed its very restricted expression pattern, in that it was detected in high amounts in thymus, lower amounts in mature T cells, and very low amounts elsewhere. Remarkably, Icb-1 expression is restricted to B cells, macrophages and dendritic cells, and 9130404H23Rik is specifically expressed in the intestine (Supplementary Fig.3).

Real-time RT-PCR analysis of thymocyte subsets showed that Themis expression was low in DN1 and DN2 cells and was upregulated in DN3 cells; expression remaining high in TCRlow immature pre-selection DP thymocytes, and was downregulated in post-positive selection TCRint and TCRhi cells (Fig.1c). Themis expression remained low in mature SP subsets. Positive selection is also marked by the migration of immature thymocytes from the thymus cortex towards the medulla, where only mature SP thymocytes that have survived the selection process are found. In situ hybridization showed that Themis expression was detectable mainly in cortical thymocytes with little expression in medullary thymocytes (Fig.1d). Thus, Themis expression is high in immature cortical thymocytes but low in mature medullary thymocytes.

These data suggested that Themis expression may be downregulated by stimulation through the αβ TCR. To explore this possibility we used the OT-I TCR transgenic (tg) mouse, which expresses a TCR recognizing a peptide derived from ovalbumin (OVA). A set of peptides is available with different affinities for the OT-I TCR, and positive and negative selection of OT-I TCR tg thymocytes is blocked in mice lacking MHC class I expression (e.g. Tap1−/−)15,2325. We therefore injected OT-I Tap1−/− mice (whose thymocytes are blocked at the pre-selection DP stage25) with peptides to stimulate the TCR in vivo26. Injection of antigenic OVA peptide, but not a non-stimulatory VSV control peptide, resulted in down-regulation of Themis expression (Fig.1e). Basal Themis expression in Tap1-expressing thymocytes was lower than in Tap1-deficient cells, presumably due to the Themis down-regulation induced by positive selection. These data indicate that stimulation through the TCR by pMHC induces downregulation of Themis.

Defective positive selection of Themis−/− thymocytes

To determine the function of Themis in T cell development, we generated Themis-deficient mice. A construct targeting the first exon of Themis, replacing it with a neomycin cassette, was prepared and transfected into 129Sv ES cells (Supplementary Fig.4). These ES cells were screened, selected and injected into blastocysts by standard methods. One progeny was >90% chimeric and was bred. Themis+/− offspring mice were backcrossed to B6 for ≥8 generations and were then intercrossed to generate homozygous mutants (Themis−/−). Themis gene disruption resulted in intermediate or complete loss of expression of Themis protein in Themis+/− or Themis−/− mice, respectively (Supplementary Fig. 4c).

Themis-deficient mice were viable, born at Mendelian frequency, and presented no gross abnormalities. Total thymocyte number was similar to wild-type mice (Supplementary Fig. 5). The proportion and number of DP thymocytes was slightly but significantly increased (P < 0.05, Student’s t-test), while the proportion and number of mature SP cells—particularly CD4 SP cells—was decreased (P < 0.01) (Fig. 2a and Supplementary Fig. 5). To identify the developmental block in the Themis−/− mice, we subdivided thymocytes into five populations on the basis of differential CD3 and CD69 expression and analyzed CD4 and CD8 expression (Fig.2b). Population 1, the most immature TCRloCD69lo cells, is composed mostly of DN and DP cells. This population was present in similar proportions in the Themis−/− and Themis+/+ mice. Population 2, which expresses a TCRintCD69lo phenotype and corresponds to pre-selection DP cells, was present in slightly higher proportions in Themis−/− mice, as expected from the larger percentage of DP cells in these mice (Fig.2a). Population 3 partially upregulated CD69 and increased TCR expression relative to population 2, and was present in slightly lower proportions in Themis−/− mice. Population 4 (TCRhiCD69hi; post-positive selection thymocytes) was markedly depleted in the Themis-deficient thymus. In the wild-type thymus, most of this population was CD4 or CD8 SP, whereas in Themis-deficient mice the cells were mostly CD4+CD8int or CD8 SP. Lastly, Themis−/− thymi contained very few TCRhiCD69lo population 5 cells, which correspond to mature SP cells ready for export to the periphery. From these data it is clear that the Themis-deficient cells are mostly blocked at the earliest stage of positive selection during which CD8 downregulation is initiated, and that very few Themis-deficient thymocytes become fully mature in phenotype.

Figure 2
Defective positive selection in Themis−/− mice. (a) Thymocytes from Themis+/+ and Themis−/− mice (intercross from line backcrossed ×10 to B6) were analyzed by FACS. Bar graph shows percentages of different thymocyte ...

To establish if Themis acts in a T cell-intrinsic manner, we performed a bone marrow (BM) reconstitution experiment (Fig.2c). BM cells from Themis+/+ Thy1.1 Ly5b mice and from Themis−/− Thy1.2 Ly5b mice were injected into irradiated Thy1.2 Ly5a recipient mice and allowed to reconstitute the thymus for 6 weeks. Ly5b-expressing donor cells were gated and then divided into Thy1.1+ (Themis+/+) and Thy1.2+ (Themis−/−) populations. This flow cytometric analysis demonstrated that the defect in development from DP to SP cells is intrinsic to the Themis−/− thymocytes as Themis−/− thymocytes showed reduced proportions of CD4 SP cells and increased proportions of DP cells relative to the Themis+/+ cells developing in the same environment.

We next analyzed the effect of Themis-deficiency on thymocyte development in mice expressing the MHC class II restricted TCR tg AND27, and the class I-restricted OT-I TCR tg24 (Fig.2d,e). The defect in thymocyte development in Themis-deficient TCR tg mice was more pronounced than in non-tg polyclonal Themis-deficient mice. In AND TCR tg thymi, development was blocked at the DP stage, with very few cells proceeding to CD4 SP (Fig.2d). The maturation arrest in the Themis−/− AND thymus was first manifest at the start of the transition from CD4hi8hi to CD4hi8int, with little or no progression to mature SP (Supplementary Fig.6a). In OT-I thymi, the developing class I-restricted cells pass through the CD4hi8int stage prior to upregulating CD8 and down-regulating CD428. Strong reduction of this stage in OT-I Themis−/− thymi (Fig.2e and Supplementary Fig.6b) suggests that, as in AND TCR tg and non-tg mice, Themis−/− OT-I mice show a strong defect at the earliest stage of positive selection. The architecture of the thymus was altered in Themis−/− mice, the medulla being smaller and fragmented compared to wild-type thymi (Fig.2f); this defect is indicative of impaired T cell maturation29.

Themis involvement in negative selection

TCR stimulation of DP and SP cells also induces negative selection. Thus, we investigated the role of Themis in negative selection in a superantigen-mediated deletion model. The C57BL-derived mouse strains express superantigens Mtv-8 and 9, which cause deletion of Vβ5+, Vβ11+ and Vβ12+ TCR clonotypic thymocytes when the MHC molecule I–E is also present30,31. To introduce I-E, which is absent on the B6-backcrossed Themis−/− mice, we bred them to B10.D2 mice which express I-E. Comparison of T cell populations in H-2b/b (lacking I-E) and H-2b/d (expressing I-E) Themis+/− thymi showed superantigen-mediated deletion of T cells expressing Vβ5, Vβ11, and Vβ12 (Fig.3). Vβ6+ (and Vβ7+, not shown) thymocytes do not recognize Mtv-8 and 9, and were not deleted. In fact, their proportion rose slightly in mice where deletion of the other clonotypes had occurred. In Themis−/− mice, superantigen-mediated deletion was significantly less efficient than in Themis+/− H-2b/d mice, although it was clearly partially active.

Figure 3
Negative selection defect in Themis−/− mice. B6 mice (H-2b) bearing wild-type or mutated Themis alleles were backcrossed or not onto B10.D2 (H-2d) mice and the expression of indicated Vβ elements in CD4 or CD8 peripheral blood ...

We also analyzed thymocyte development in mice expressing the MHC class I-restricted HY TCR tg. Here, positive selection in females leads to a large CD8 SP population, but negative selection predominates in males, resulting in few DP or SP cells32,33. HY TCR tg females lacking Themis resembled AND TCR tg and OT-I TCR tg Themis−/− mice in that few thymocytes progressed to become mature CD8 SP (Supplementary Fig.7a). HY TCR tg Themis+/− males showed a much reduced number of thymocytes compared to HY TCR tg females, and a large proportion of these cells were HY-TCR+ clonotypic cells expressing low amounts of CD4 and CD83434. These cells are believed to escape negative selection by loss of the coreceptor. In male HY TCR tg Themis−/− thymi, negative selection was apparently incomplete in that CD4 and CD8 expression was much lower than normal, but not as low as in the Themis+/− males, and the clonotype-negative cells found in the Themis+/− males were not present (Supplementary Fig.7b). Thus both negative and positive selection were reduced in the absence of Themis, suggesting that absence of Themis leads to a general defect in TCR-mediated signaling.

Peripheral T cells in Themis−/− mice

The deficiency in SP thymocytes led to a reduction in peripheral T cell numbers in Themis−/− mice (Fig.4a,b). The percentage of B cells increased due to the loss of T cells (data not shown). The deficiency in peripheral T cells was most evident in the CD4 population, but both CD4 and CD8 cells were significantly reduced in number. Analysis of TCR tg mice confirmed these data, with very few mature peripheral CD4 cells in AND Themis−/− (Fig.4c), and few CD8 cells in OT-I Themis−/− (Fig.4d) TCR tg mice. Within the CD4+ population, cells with a regulatory T (Treg) cell phenotype (CD25+Foxp3+) were increased in frequency from 10% in Themis+/+ to 21% in Themis−/− (Fig.4e), though the absolute number of Themis−/− Treg cells was ~50% of the number in Themis+/+ mice (data not shown).

Figure 4
Peripheral T cell phenotype in Themis−/− mice. (a,b) CD4 and CD8 T cells in the spleen of B6 Themis+/+ and Themis−/− mice of indicated ages were quantified by flow cytometry. Bar graph shows average percentage of each population ...

The expression of CD62L and CD44 in peripheral T cells is used to differentiate naïve from memory T cells. There was a lower percent of CD62LhiCD44lo naïve phenotype cells, and a higher proportion of CD62LloCD44hi memory phenotype cells in Themis−/− mice (Fig.4f). Expression of CD122+ (interleukin 2Rβ) was also higher in Themis−/− mice (data not shown). Thus the few peripheral cells in Themis-deficient mice have a phenotype redolent of memory cells, a frequent observation in partially lymphopenic mice that probably reflects homeostatic T cell population expansion in the periphery35.

We then tested the ability of Themis−/− spleen T cells to respond to anti-CD3 stimulation (Fig.5). CD69 upregulation was strongly reduced in Themis-deficient CD4 and CD8 T cells compared to wild-type T cells. Proliferation of both CD4 and CD8 Themis−/− T cells in response to anti-CD3 was very poor compared to Themis+/+ T cells. Themis+/+ and Themis−/− CD8 T cells expressed similar amounts of surface TCR, but TCR expression was slightly reduced on the surface of Themis−/− CD4 cells (Fig.5a).

Figure 5
Defective activation of Themis-deficient T cells. Themis−/− cells are shown in blue and Themis+/+ cells in red. (a) CD3 expression on gated CD4+ (CD25) and CD8+ spleen cells. (b) Spleen cells were activated by cross-linking with ...

Themis in early TCR-signaling

The role of Themis in the TCR signaling cascade was investigated by immunoprecipitation and immunoblotting (Fig.6). Initially, a proteomic study of global dynamics of TCR-directed tyrosine phosphorylation using stable isotopic labeling of amino acids in cell culture (SILAC), identified a number of novel potential signaling components in Jurkat cells, including C6orf190 (C.B., O.A., M.S., to be published elsewhere), which was named SPoT (for Signaling Phosphoprotein specific of T cell, which we refer to here as THEMIS). Reduction of THEMIS expression by siRNA in Jurkat cells resulted in substantial decrease of interleukin 2 (IL-2) production induced by anti-TCR and anti-CD28 stimulation (C.B., O.A., M.S., unpublished data). SILAC data suggested that THEMIS was phosphorylated on tyrosine as early as 30 seconds after TCR stimulation and that this phosphorylation was reduced at 5 minutes after TCR stimulation; these phosphorylation kinetics are similar to those of other known TCR-proximal signaling proteins (e.g., LAT, SLP-76) (data not shown).

Figure 6
Themis is part of the TCR signal cascade. Jurkat (a) or fresh human PBL T cells (b) were stimulated with anti-CD3 for the indicated times (min, US is unstimulated), lysed and immunoprecipitated with anti-THEMIS serum or anti-p-Tyr mAb. Blots were probed ...

To corroborate these data, Jurkat cells were stimulated with anti-CD3, lysates were immunoprecipitated with anti-phosphotyrosine (pTyr), and immunoprecipitates were incubated with an antiserum against the C-terminal sequence of THEMIS. This treatment revealed a phosphoprotein of the expected size for THEMIS (~72 kD, Fig.6a) that increased in abundance after 30 seconds of stimulation then declined. Immunoprecipitation with anti-THEMIS and immunoblotting with anti-pTyr provided direct evidence that THEMIS was itself phosphorylated on a tyrosine residue, indicating that it is a TCR-dependent PTK substrate. Similar data were obtained with human CD4+ PBL T cells (Fig.6b). In mouse thymocytes, Themis was tyrosine phosphorylated within 30 seconds after TCR crosslinking, and phosphorylation was undetectable after 3 minutes (Fig.6c). Together these data show that in mouse and human cells Themis is an early target of TCR-controlled PTK(s) and suggest that it is a component of the TCR signaling pathway whose function and-or recruitment may depend on tyrosine phosphorylation.

Defective TCR signaling in Themis−/− thymocytes

Because the proline-rich sequence in Themis and related proteins is predicted to be a binding site for SH3 domains, we tested an SH3 domain array to screen for binding to the N-terminal and C-terminal (including the PRS) halves of Themis. The PRS-containing region bound significantly to the phospholipase C-γ1 (PLC-γ1) SH3 domain (data not shown). To confirm a Themis-PLCγ1 interaction in thymocytes, we stimulated B6 thymocytes with anti-CD3, then immunoprecipitated cell lysates with Themis-specific antiserum. Although there was some interaction between Themis and PLC-γ1 prior to stimulation, the amount of PLCγ1 coprecipitated with Themis increased at 2 minutes post-stimulation, but ultimately returned to a weaker interaction (Fig.6d). As Itk is a Tek-family PTK that is required for activation of PLC-γ1 and for TCR-signaling36, we tested the possibility of an interaction between Itk and Themis. Anti-Itk coprecipitated both Themis and PLC-γ1 from freshly isolated, anti-CD3-stimulated thymocytes (Fig.6e). It appeared that that Itk-Themis and Itk-PLC-γ1 interactions increased after CD3 crosslinking. To test this, we rested thymocytes in culture for 2h before activation with anti-CD3 plus anti-CD4. As previously reported37, PLC-γ1 interacted constitutively with Itk in thymocytes; after stimulation the interaction increased slightly before declining after 30 seconds (Fig.6f). Interaction of Themis with Itk was not constitutive, but was induced by 30 seconds after TCR stimulation.

To determine which signaling pathways are defective in Themis−/− thymocytes, we analyzed immunoblots of anti-CD3 plus anti-CD4 activated thymocytes from Themis+/+ and Themis−/− thymocytes (Fig.6g). We found no difference in ZAP70, PLC-γ1 or Itk (or LAT, data not shown) phosphorylation in response to TCR stimulation. However, Erk showed reduced phosphorylation in Themis−/− cells. Erk1 phosphorylation (p-p44) was delayed in Themis−/− compared to Themis+/+ cells, and Erk2 phosphorylation (p-p42) was reduced in Themis−/− cells at 30 and 90 seconds after stimulation (Fig.6g, Supplementary Fig.8a), and was not always detectable at 3 minutes. This experiment was performed three times with similar results (Supplementary Fig.8b). Phosphorylation of Raf, Mek and Jnk was not affected by Themis deficiency (data not shown).

To look at other aspects of TCR signaling, we stimulated pre-selection DP thymocytes from OT-I TCR tg Tap1−/− mice. Stimulation by anti-CD3 plus anti-CD4 or OVA-Kb tetramers (Fig.7a) induced a slightly slower and reduced initial peak of store-operated Ca2+ influx in Themis−/− cells relative to Themis+/+ cells. This difference in Ca2+ influx, while relatively subtle, was very reproducible (a total of 9 Themis−/− versus 5 Themis+/+ in 3 separate experiments). The release of endoplasmic reticulum Ca2+ stores was very low in the thymocytes, and we found no difference between Themis−/− cells relative to Themis+/+ cells.

Figure 7
Signaling in Themis-deficient thymocytes. (a) Pre-selection DP thymocytes from Themis+/+ or Themis−/− OT-I Tap1−/− mice were stimulated by crosslinking biotinylated anti-CD3 and -CD4 with streptavidin (SAv) (left and center) ...

Organization of the cytoskeleton at the immunological synapse is regulated by TCR signaling38. We therefore imaged actin polymerization in OT-I Tap1−/− thymocytes expressing or lacking Themis as they interacted with OVA peptide presented by EL4 tumor cells (Fig.7b). The actin polymerization pattern in Themis-sufficient thymocytes was similar to that in OT-I lymph node T cells, but Themis-deficient thymocytes showed a significant defect in actin polymerization at the immunological synapse. A useful correlate of stimulation of thymocytes through the TCR by positive or negative selecting ligands is upregulation of CD6915,25,39,40. We found that the Themis−/− OT-I Tap1−/− pre-selection DP thymocytes were less able to upregulate CD69 in response to OVA peptide than their Themis+/+ OT-I Tap1−/− counterparts (Fig.7c). These data suggest that Themis is involved in regulating the TCR-signaling cascade, albeit with relatively small effects on early events in the cascade.


Development of αβ T cells is critically dependent on signaling through the TCR, and can be disrupted by deficiencies in the receptor itself and many signaling molecules and mediators. In this work, we identified a novel protein, Themis, that is required for passage through the positive selection checkpoint.

Themis deficiency results in a defect in development that is stronger in CD4 cells than in CD8 cells. However, this defect was evident in both MHC class I and class II-restricted TCR transgenic lines, wherein thymocyte development was severely blocked at the earliest stages of positive selection. A stronger phenotype in TCR tg thymocytes is a common finding in mice deficient in signaling molecules, as the normal TCR repertoire is plastic enough to select for TCRs with higher or lower affinity, thereby allowing the signal strength to be modulated to permit selection10. The strong developmental block in the TCR tg Themis−/− mice, together with the evidence indicating that negative selection signaling is also deficient in Themis−/− mice, suggested a role for Themis in the TCR signaling cascade rather than in a developmental program per se. Because Themis is expressed in immature thymocytes, and is downregulated by stimulation through the TCR as occurs during positive selection, the signaling role of Themis is likely to be most important in DP thymocytes before or during positive selection. In agreement with this notion, Themis−/− mice pre-selection thymocytes responded poorly compared to Themis+/+ thymocytes to TCR stimulation, as measured by CD69 upregulation. Themis−/− thymocytes also showed reduced actin polymerization at the immunological synapse, as in Itk-deficient and Vav-deficient cells. These results are consistent with a defect in TCR signal transduction3942.

Themis was tyrosine phosphorylated within 30 seconds of TCR stimulation, suggesting that it may play a role early in signal transduction through the TCR. Moreover, it interacted inducibly with Itk and PLC-γ1, although phosphorylation of these proteins was not altered in Themis−/− cells. Deletion of Itk has a profound effect on Ca2+ signaling, but a relatively minor effect on thymocyte development36,41. In contrast, Themis-deficiency had a strong effect on development but a small effect on Ca2+ signals (e.g. slightly slower and weaker Ca2+ flux). The effect of loss of Themis on Erk signaling was also quite subtle, and manifested as a reduction in the strength and duration of phosphorylation. These findings suggest that Themis might negatively regulate phosphatases that turn off Erk signaling. Notably, similarly small differences in p-Erk induction distinguished signals that induce positive and negative selection of thymocytes14,15, and sustained Erk signals are required for positive selection2,12,14. Additionally, there is evidence that slight changes in Ca2+ signaling lead to profound differences in MAP kinase activation in non T cells43, and in pre-selection thymocytes, and that these differences distinguish the quality of TCR stimulation17. The alterations in Ca2+ and Erk signaling between Themis−/− and Themis+/+ pre-selection thymocytes suggest that Themis may be involved in setting the threshold for TCR responses to selecting signals. Such a role for Themis can account for its expression primarily in pre-positive selection thymocytes, and would indicate that its role in TCR-mediated signaling may not be so important in peripheral T cell activation. Study of the normal role of Themis in peripheral T cell signaling will have to await generation of a conditional Themis−/− mouse.

When given the opportunity to produce a full TCR repertoire, some Themis-deficient thymocytes pass through the positive selection checkpoint and generate peripheral T cells. These cells respond poorly to TCR-mediated stimulation. Interestingly, in both CD4 and CD8 subsets, there was an increased prevalence of memory-phenotype (CD44hiCD62LloCD122+) cells and reduced proportions of naïve-phenotype populations. This may be the result of homeostatic population expansion in a lymphopenic environment35, and is similar to the phenotype seen in mice lacking Itk. Memory phenotype T cells in Itk-deficient mice have been referred to as “innate-like” T cells44. This phenotypic similarity to Itk-deficient mice supports the notion that Themis is involved in TCR signal transduction via Itk.

We also noted a strongly increased proportion (though reduced absolute number) of Treg cells in Themis−/− mice. Interestingly the Themis gene was identified as one of the transcripts most strongly downregulated by Foxp3 in response to TCR stimulation45, and its suppression may therefore be part of the Treg cell development program.

Themis, Icb-1 and 9130404H23Rik are members of a small and essentially uncharacterized family. Each has a very restricted cell-type-specific expression and is highly conserved in vertebrates. Other than a PRS, they have no predicted conserved domains that are shared with other proteins. These properties point to the family having important and perhaps novel functions and structure. Here we showed that one of these family members, Themis, is a T cell specific protein with a crucial function in positive selection through transducing TCR signals in thymocytes.

Accession Codes

Mouse Themis: MGI:2443552, Human THEMIS GeneID: 387357.



Tcra−/− (B6.129S2-Tcratm1Mom/J), Rag1−/− (B6.129S7-Rag1tm1Mom/J), and Ly5a congenic mice (B6.SJL-Ptprca Pep3b/BoyJ) on the B6 background were obtained from Jackson Labs. Other strains were bred at The Scripps Research Institute (TSRI) and maintained in accordance with TSRI Animal Care and Use Committee.


Antibodies recognizing CD3 (145-2C11), CD4 (RM4.4, RM4.5, GK1.5), CD8α (53–6.7), CD25 (PC61.5), CD62L (MEL14), CD44 (IM7), Itk (2F12), PLC-γ1 (10/PLCg), Foxp3 (FJK-165), Vβ5 (MR9-6), Vβ6 (RR4-7), Vβ11 (CTVB11), and Vβ12 (CTVB12b), were obtained from eBiosciences or BDBiosciences. Anti-p-Tyr (4G10) was from Upstate Biotechnology. Anti-p42/44 (3A7) and p-p42/44 (197G2) and antisera to p-ZAP70 and p-PLC-γ1 were from Cell Signaling Technology. Anti-Themis was raised in rabbits against the N-terminal 294 amino acids, made as a fusion to glutathione-S-transferase (GST), purified and with the GST removed. The antiserum was affinity purified. Anti-THEMIS serum was made by immunizing rabbits with a peptide corresponding to the C-terminal 15 amino acids of the human sequence. Specificity is demonstrated in Supplementary Fig.9.

Cloning of Themis

A subtraction library was constructed from cDNA from Tcra−/− thymus mRNA subtracted with cDNA from Rag1−/− thymus mRNA as described20. Clones randomly picked from the library were sequenced and used to search Genbank to identify potentially novel clones. A full length clone of human THEMIS was isolated by screening a Jurkat cDNA library.

Real-time RT-PCR Analysis

Real-time PCR was performed using SYBERgreen PCR Master Mix and the Prism 7900 instrument and analyzed with SDSv2.1 software (Applied Biosystems). Results were normalized to β-actin expression. Primer sequences are shown in Supplementary Table 2).

Generation of Themis-deficient mice

Exon 1 of Themis was replaced by the 1.6 kb PGK neomycin cassette. The target vector (pKO Scrambler) contained the flanking sequence of exon 1: a 5′ 2.8 kb short arm and a 3′ 3.9 kb long arm sequence flanking the neomycin phosphotransferase gene. Embryonic stem cells (129SvEv strain) were electroporated with 25µg linearized DNA/107 cells, and cultured for 7d in medium containing 400 µg/ml G418. Drug resistant clones were expanded and homologous recombination confirmed by PCR and Southern blot. Positive clones were injected into blastocysts derived from B6 embryo donor mice mated with C6D2F1 stud males (B6 × DBA2/J hybrids) and transplanted into pseudopregnant CD-1 recipients. Chimeras were bred to C57BL/6J females in order to reveal germ-line transmission followed by backcrossing to B6 (10 generations).

Calcium flux

Thymocyte suspensions were prepared and split into separate tubes. Part was loaded with CFSE (20 nM, for 10 min at 37°C) or Cy5 (1µg/ml, 5min RT), or mock treated, followed by washing. CFSE-labeled and-or Cy5-labeled populations were mixed equally with a mock-loaded sample, for example; OT-I Tap1−/−Themis+/+ (mock) plus OT-I Tap1−/−Themis−/− (CFSE or Cy5). In some experiments, the CFSE- and Cy5-treated and untreated pairs were reversed to ensure that the CFSE or Cy5 did not alter the results. Cells were suspended at 107/ml in cRPMI (RPMI supplemented with 10% FCS, 100U/ml Penicillin, 10µg/ml Streptomycin, 292µg/ml Glutamine, 50µM 2-ME, 25mM HEPES) and were incubated with 2µM Indo-1 AM (Molecular Probes) for 30 min at 37°C, 5% CO2. Cells were washed twice with cRPMI. Cells were treated with biotin-anti-CD3 and anti-CD4 as well as PerCPCy5.5-anti-CD8 and PECy7-anti-CD4 on ice for 20min. Cells were washed once with cRPMI and once with Ca2+ free medium (cHBSS: Ca2+ and Mg2+ free HBSS supplemented with 1% FCS, 1mM MgCl2, 1mM EGTA, and 10mM HEPES), and resuspended in cHBSS. Cells were pre-warmed to 37°C before analysis and were kept at 37°C during event collection on a Becton-Dickinson LSRII. To stimulate cells, 10µg/ml Streptavidin (Jackson) was added to cross-link biotinylated antibodies; alternatively, cells were stimulated with tetramers made as described from refolded Kb and β2m15,23, or purchased from Beckman-Coulter (iTag MHC tetramer). 5mM CaCl2 was added during analysis, and maximal Ca2+ flux obtained by adding 500ng/ml of ionomycin (Calbiochem). Mean fluorescence ratio was calculated using FlowJo (TreeStar).

T cell activation and proliferation assay

Lymphocytes from Themis+/+ (Thy1.1, B6.PL) or Themis−/− (Thy1.2, B6) mice were mixed 1:1. Cells were cultured in the presence of soluble anti-CD3 for 5h, stained with anti-CD69 and anti-Thy1 allotypes and analyzed by FACS. For proliferation assays, cells were stained with 0.2µM CFSE at 37°C for 10 min. Staining was stopped by adding an equal volume of FCS. After washing, cells were cultured with or without stimulants (e.g. anti-CD3) in cRPMI for 48h. T cell proliferation was measured by CFSE dilution analyzed by FACS.

Thymocyte CD69 upregulation assay

Thymocytes (2–3×105) from OT-I tg Tap1−/− mice (Themis+/− or −/−) were incubated with 2×105 peptide-pulsed EL4 cells in round-bottom 96-well plates at 37° for 5h. Cells were stained and analyzed by FACS as described39,40. Gates defining CD69 versus CD69+ were determined from thymocytes incubated with non-peptide pulsed EL4 cells.

Immunoprecipitation and immunoblotting

Cells were lysed in buffer containing 20mM Tris pH7.5, 150mM NaCl, 0.5% n-Dodecyl-β-D-maltoside or NP40, 20mM NaF, 1mM Na3VO4, and protease inhibitor cocktail (Roche). Immunoprecipitation was performed on cell lysates (0.5–1 mg of protein) overnight at 4°C, then protein-A sepharose beads (Amersham) were added for 1h. After washing, samples were resolved on 4–12% SDS-PAGE and blotted. The membrane was blocked 1h with PBS, 1% TWEEN, 5% dry milk, incubated with primary Ab for 3h at room temperature or overnight at 4°C, washed and incubated with the secondary Horseradish peroxidase-antibody for 1.5h followed by enhanced chemiluminescence detection of signal (Pierce).

In situ hybridization

Thymi from 3–4 week-old mice were fixed, embedded in paraffin and 5µm sections probed with antisense 35S-UTP-labeled Themis RNA. Sections were RNAase A-treated, washed (0.2X SSC, 42°C), dried, coated with Kodak NTB2 emulsion, exposed for 1–2 weeks, developed, and counterstained with methyl green. Controls included hybridizations with sense and unrelated probes.

Peptide stimulation

The peptides OVA (SIINFEKL) or VSV (RGYVYQGL) were injected i.v. into OT-I Tap1−/− mice as described26. After 1h, the thymus was removed, RNA extracted, and subjected to real time RT-PCR.

Statistical testing

Statistical differences were calculated using the mean difference hypothesis of Student’s two tailed t-test assuming different variances and confidence level of 95%. Calculations performed using GraphPad Prism.S

Supplementary Material


This work was supported by NIH grants GM048002 and AI073870 to NRJG. SV was supported by the Concern Foundation for Cancer Research. MVM and JAHH were supported by T32 AI07244. JAHH was supported by the Irving S. Sigal Fellowship. OA, CB and MS were supported by Wellcome Trust program grant GR076558MA and by funds from the Sir William Dunn School of Pathology. CB received the Sloane Robinson Graduate Award from Lincoln College, Oxford. We thank James Bradley, Alex Field, Yugarshi Mondal, Samuel Rose, Wayne Sakati, Vandana Sharma, Jovan Shepherd, and Jacob Vasquez for technical help. This is manuscript 19524 from The Scripps Research Institute.


The authors have no conflicting financial interests.


1. Starr TK, Jameson SC, Hogquist KA. Positive and negative selection of T cells. Annu. Rev. Immunol. 2003;21:139–176. [PubMed]
2. Werlen G, Hausmann B, Naeher D, Palmer E. Signaling life and death in the thymus: timing is everything. Science. 2003;299:1859–1863. [PubMed]
3. Malissen B, Malissen M. Allelic exclusion of T cell antigen receptor genes. In: Bell JI, Owen MJ, Simpson E, editors. T cell receptors. Oxford: Oxford University Press; 1995. pp. 352–368.
4. Davey GM. Preselection thymocytes are more sensitive to T cell receptor stimulation than mature T cells. J. Exp. Med. 1998;188:1867–1874. [PMC free article] [PubMed]
5. Lucas B, Stefanova I, Yasutomo K, Dautigny N, Germain RN. Divergent changes in the sensitivity of maturing T cells to structurally related ligands underlies formation of a useful T cell repertoire. Immunity. 1999;10:367–376. [PubMed]
6. D’Oro U, Vacchio MS, Weissman AM, Ashwell JD. Activation of the Lck tyrosine kinase targets cell surface T cell antigen receptors for lysosomal degradation. Immunity. 1997;7:619–628. [PubMed]
7. Naramura M, Kole HK, Hu R-J, Gu H. Altered thymic positive selection and intracellular signals in Cbl-deficient mice. Proc. Natl. Acad. Sci. USA. 1998;95:15547–15552. [PubMed]
8. Alam SM, Gascoigne NRJ. Post-translational regulation of TCR Va allelic exclusion during T cell differentiation. J. Immunol. 1998;160:3883–3890. [PubMed]
9. Niederberger N, et al. Allelic exclusion of the TCR a-chain is an active process requiring TCR-mediated signaling and c-Cbl; J. Immunol. 2003;170:4557–4563. [PubMed]
10. Fischer A, Malissen B. Natural and engineered disorders of lymphocyte development. Science. 1998;280:237–243. [PubMed]
11. Zamoyska R, et al. The influence of the src-family kinases, Lck and Fyn, on T cell differentiation, survival and activation. Immunol. Rev. 2003;191:107–118. [PubMed]
12. Mariathasan S, et al. Duration and strength of extracellular signal-regulated kinase signals are altered during positive versus negative thymocyte selection. J. Immunol. 2001;167:4966–4973. [PubMed]
13. Fischer AM, Katayama CD, Pages G, Pouyssegur J, Hedrick SM. The role of erk1 and erk2 in multiple stages of T cell development. Immunity. 2005;23:431–443. [PubMed]
14. McNeil LK, Starr TK, Hogquist KA. A requirement for sustained ERK signaling during thymocyte positive selection in vivo. Proc. Natl. Acad. Sci. USA. 2005;102:13574–13579. [PubMed]
15. Daniels MA, et al. Thymic selection threshold defined by compartmentalization of Ras/MAPK signalling. Nature. 2006;444:724–729. [PubMed]
16. Kane LP, Hedrick SM. A role for calcium influx in setting the threshold for CD4+CD8+ thymocyte negative selection. J.Immunol. 1996;156:4594–4601. [PubMed]
17. Freedman BD, Liu QH, Somersan S, Kotlikoff MI, Punt JA. Receptor avidity and costimulation specify the intracellular Ca2+ signaling pattern in CD4+CD8+thymocytes. J. Exp. Med. 1999;190:943–952. [PMC free article] [PubMed]
18. Schwartz RH. THEMIS PAPER. Nature Immunol. 2009
19. Love P. THEMIS PAPER. Nature Immunol. 2009
20. McGuire MV, Suthipinijtham P, Gascoigne NRJ. The mouse Supt16h/Fact140 gene, encoding part of the FACT chromatin transcription complex, maps close to Tcra and is highly expressed in thymus. Mamm. Genome. 2001;12:664–667. [PubMed]
21. Treeck O, Strunck E, Vollmer G. A novel basement membrane-induced gene identified in the human endometrial adenocarcinoma cell line HEC1B. FEBS Lett. 1998;425:426–430. [PubMed]
22. Su AI, et al. A gene atlas of the mouse and human protein-encoding transcriptomes. Proc. Natl. Acad. Sci. USA. 2004;101:6062–6067. [PubMed]
23. Alam SM, et al. T cell receptor affinity and thymocyte positive selection. Nature. 1996;381:616–620. [PubMed]
24. Hogquist KA, et al. T cell receptor antagonist peptides induce positive selection. Cell. 1994;76:17–27. [PubMed]
25. Hogquist KA, et al. Identification of a naturally occurring ligand for positive selection. Immunity. 1997;6:389–399. [PubMed]
26. Fiorini E, et al. Peptide-induced negative selection of thymocytes activates transcription of an NF-kB inhibitor; Mol. Cell. 2002;9:637–648. [PubMed]
27. Kaye J, et al. Selective development of CD4+ T cells in transgenic mice expressing a class II MHC-restricted antigen receptor. Nature. 1989;341:746–749. [PubMed]
28. Gallegos AM, Bevan MJ. Central tolerance to tissue-specific antigens mediated by direct and indirect antigen presentation. J. Exp. Med. 2004;200:1039–1049. [PMC free article] [PubMed]
29. van Ewijk W, Hollander G, Terhorst C, Wang B. Stepwise development of thymic microenvironments in vivo is regulated by thymocyte subsets. Development. 2000;127:1583–1591. [PubMed]
30. Irwin MJ, Gascoigne NRJ. Interplay between superantigens and the immune system. J. Leukocyte Biol. 1993;54:495–503. [PubMed]
31. Scherer MT, Ignatowicz L, Winslow GM, Kappler JW, Marrack P. Superantigens: Bacterial and viral proteins that manipulate the immune system. Annu. Rev. Cell Biol. 1993;9:101–128. [PubMed]
32. Kisielow P, Bluthmann H, Staerz UD, Steinmetz M, von Boehmer H. Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4+8+ thymocytes. Nature. 1988;333:742–746. [PubMed]
33. Teh HS, et al. Thymic major histocompatibility complex antigens and the aβ T-cell receptor determine the CD4/CD8 phenotype of T cells. Nature. 1988;335:229–233. [PubMed]
34. Teh HS, Kishi H, von Boehmer H. Deletion of autospecific T cells in T cell receptor (TCR) transgenic mice spares cells with normal TCR levels and low levels of CD8 molecules. J. Exp. Med. 1989;169:795–800. [PMC free article] [PubMed]
35. Ge Q, Hu H, Eisen HN, Chen J. Different contributions of thymopoiesis and homeostasis-driven proliferation to the reconstitution of naive and memory T cell compartments. Proc. Natl. Acad. Sci. USA. 2002;99:2989–2994. [PubMed]
36. Berg LJ, Finkelstein LD, Lucas JA, Schwartzberg PL. Tec family kinases in T lymphocyte development and function. Annu. Rev. Immunol. 2005;23:549–600. [PubMed]
37. Huang YH, et al. Positive regulation of Itk PH domain function by soluble IP4. Science. 2007;316:886–889. [PubMed]
38. Burkhardt JK, Carrizosa E, Shaffer MH. The actin cytoskeleton in T cell activation. Annu. Rev. Immunol. 2008;26:233–259. [PubMed]
39. Yachi PP, Ampudia J, Gascoigne NRJ, Zal T. Nonstimulatory peptides contribute to antigen-induced CD8-T cell receptor interaction at the immunological synapse. Nat. Immunol. 2005;6:785–792. [PMC free article] [PubMed]
40. Yachi PP, Lotz C, Ampudia J, Gascoigne NRJ. T cell activation enhancement by endogenous pMHC acts for both weak and strong agonists but varies with differentiation state. J Exp Med. 2007;204:2747–2757. [PMC free article] [PubMed]
41. Liu KQ, Bunnell SC, Gurniak CB, Berg LJ. T cell receptor-initiated calcium release is uncoupled from capacitative calcium entry in Itk-deficient T cells. J. Exp. Med. 1998;187:1721–1727. [PMC free article] [PubMed]
42. Labno CM, et al. Itk functions to control actin polymerization at the immune synapse through localized activation of Cdc42 and WASP. Curr. Biol. 2003;13:1619–1624. [PMC free article] [PubMed]
43. Kupzig S, Walker SA, Cullen PJ. The frequencies of calcium oscillations are optimized for efficient calcium-mediated activation of Ras and the ERK/MAPK cascade. Proc. Natl. Acad. Sci. USA. 2005;102:7577–7582. [PubMed]
44. Berg LJ. Signalling through TEC kinases regulates conventional versus innate CD8+ T-cell development. Nat. Rev. Immunol. 2007;7:479–485. [PubMed]
45. Marson A, et al. Foxp3 occupancy and regulation of key target genes during T-cell stimulation. Nature. 2007;445:931–935. [PMC free article] [PubMed]
46. Corpet F. Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res. 1988;16:10881–10890. [PMC free article] [PubMed]
47. Gouet P, Courcelle E, Stuart DI, Metoz F. ESPript: analysis of multiple sequence alignments in PostScript. Bioinformatics. 1999;15:305–308. [PubMed]