A Stretch of Negatively Charged Amino Acids of Linker for Activation of T-Cell Adaptor Has a Dual Role in T-Cell Antigen Receptor Intracellular Signaling

The adaptor protein linker for activation of T cells (LAT) has an essential role transducing activatory intracellular signals coming from the TCR/CD3 complex. Previous reports have shown that upon T-cell activation, LAT interacts with the tyrosine kinase Lck, leading to the inhibition of its kinase activity. LAT–Lck interaction seemed to depend on a stretch of negatively charged amino acids in LAT. Here, we have substituted this segment of LAT between amino acids 113 and 126 with a non-charged segment and expressed the mutant LAT (LAT-NIL) in J.CaM2 cells in order to analyze TCR signaling. Substitution of this segment in LAT prevented the activation-induced interaction with Lck. Moreover, cells expressing this mutant form of LAT showed a statistically significant increase of proximal intracellular signals such as phosphorylation of LAT in tyrosine residues 171 and 191, and also enhanced ZAP70 phosphorylation approaching borderline statistical significance (p = 0.051). Nevertheless, downstream signals such as Ca2+ influx or MAPK pathways were partially inhibited. Overall, our data reveal that LAT–Lck interaction constitutes a key element regulating proximal intracellular signals coming from the TCR/CD3 complex.

The adaptor protein linker for activation of T cells (LAT) has an essential role transducing activatory intracellular signals coming from the TCR/CD3 complex. Previous reports have shown that upon T-cell activation, LAT interacts with the tyrosine kinase Lck, leading to the inhibition of its kinase activity. LAT-Lck interaction seemed to depend on a stretch of negatively charged amino acids in LAT. Here, we have substituted this segment of LAT between amino acids 113 and 126 with a non-charged segment and expressed the mutant LAT (LAT-NIL) in J.CaM2 cells in order to analyze TCR signaling. Substitution of this segment in LAT prevented the activation-induced interaction with Lck. Moreover, cells expressing this mutant form of LAT showed a statistically significant increase of proximal intracellular signals such as phosphorylation of LAT in tyrosine residues 171 and 191, and also enhanced ZAP70 phosphorylation approaching borderline statistical significance (p = 0.051). Nevertheless, downstream signals such as Ca 2+ influx or MAPK pathways were partially inhibited. Overall, our data reveal that LAT-Lck interaction constitutes a key element regulating proximal intracellular signals coming from the TCR/CD3 complex.
Keywords: laT, Tcr, cD3, signaling, lck inTrODUcTiOn T lymphocytes are activated upon the recognition of peptides bound to MHC molecules in the surface of antigen-presenting cells (APCs). Antigen recognition by T cells is performed by the T-cell antigen receptor (TCR) heterodimer, triggering intracellular signals leading eventually to proliferation, production of cytokines, and differentiation of T lymphocytes, which are essential for proper immune responses (1,2). Most of T cells express a TCR consisting of a TCR-αβ heterodimer associated with CD3ε, γ, δ, and ζ polypeptides, constituting the TCR/CD3 complex. The cytoplasmic tails of CD3 polypeptides contain immunoreceptor tyrosine-based activation motifs (ITAMs), which are essential for the transduction of intracellular signaling from the TCR/CD3 complex. Upon TCR recognition of a foreign antigen, the Src family kinase Lck phosphorylates tyrosine residues of CD3 ITAMs, which in turn leads to the recruitment of the cytosolic tyrosine kinase ZAP70. Once there, ZAP70 is phosphorylated and activated by Lck, allowing the phosphorylation of adaptor proteins LAT and SLP-76. Once their tyrosines are phosphorylated, LAT and SLP-76 bind to several cytosolic proteins leading to the formation of multiprotein complexes, denominated signalosomes, which are critical for the amplification and diversification of intracellular downstream signaling pathways leading to activation, proliferation, and differentiation of naive T lymphocytes. The flexibility of the signaling network is thought to account for the ability of T cells to exhibit different physiological responses, each of which requires engagement of the TCR (1,3).
The membrane adaptor LAT has a critical function in the transduction of intracellular signals coming from the TCR/CD3 complex (4,5). LAT is an adaptor protein with a short extracellular N-terminal fragment, a hydrophobic transmembrane segment followed by a long cytoplasmic tail containing nine conserved tyrosine residues (6,7). The essential role of LAT for TCR/CD3 signal transduction was first shown in LAT-deficient T-cell lines, which showed defects in activation-induced intracellular calcium, Ras activation, and IL-2 production, defects that were relieved upon reintroduction of LAT in these cells (8,9). Moreover, LATknockout (LAT-KO) mice revealed its essential role for T-cell development: LAT-KO mice show a total block in thymic maturation at the CD4 − CD8 − double negative (DN) stage and contain no mature T lymphocytes (10). It has also been demonstrated, both in vitro and in vivo, the critical role of LAT tyrosine residues for the proper transduction of activation signals from the TCR/CD3 complex (11)(12)(13). Unexpectedly, a LAT-knockin (LAT-KI) strain of mice in which the sixth tyrosine was mutated to phenylalanine (LAT Y136F ) exhibited a partial block in T-cell maturation, but they later developed a polyclonal lymphoproliferative disorder of CD4 T cells producing massive amounts of T helper type 2 (TH2) cytokines, which caused tissue eosinophilia and maturation of plasma cells secreting huge amounts of IgE and IgG1 immunoglobulins (14,15). In the same way, another KI strain of mice in which the last three carboxy-terminal tyrosine residues were mutated to phenylalanines (LAT 3YF ) showed a total block in αβ T-cell development and a partially impaired γδ T-cell development but developed a lymphoproliferative disorder of γδ T cells that chronically produced TH2 cytokines (16). These paradoxical phenotypes established for the first time that LAT is a crucial regulator of T-cell homeostasis, later confirmed in conditional LAT-KI strains of mice (17).
In a search of non-tyrosine-based motifs in LAT sequence explaining its regulatory functions, our group has analyzed the role of proteolytic cleavage as a terminator of TCR signaling (18,19). Upon proapoptotic stimuli, LAT is cleaved at three aspartic acid residues which are located close to relevant tyrosine residues and, interestingly, LAT phosphorylation inhibits its Fas-mediated cleavage. This cleavage generates a truncated form of LAT unable to transduce intracellular signals, suggesting a mechanism of termination of signals in T cells. Consistently, the related adaptor molecule NTAL (LAB/LAT2) is also subject of proteolytic cleavage after Fas stimulation in B cells and monocytes (20). Interestingly, it has been previously shown that upon T cells' activation LAT is able to bind and negatively regulate Lck kinase and this interaction is mediated by a negatively charged stretch between amino acids 112 and 126 of LAT (21,22). Thus, the truncated form of LAT at aspartic 126 would still be able to bind Lck, constituting a putative new mechanism for negative regulation of the TCR/CD3 signaling cassette. Here, we have generated a new mutant of LAT in which the negatively charged stretch has been substituted by a non-charged amino acid sequence and expressed this mutant form of LAT in J.CaM2 cells. Substitution of this negatively charged segment of LAT prevented the activation-induced interaction with Lck, and we called this mutant LAT-NIL (from Not Interacting with Lck). Furthermore, expression of LAT-NIL increased several early intracellular signals, including phosphorylation of ZAP70 and LAT Tyr residues 171 and 191, although downstream signals such as Ca 2+ influx and MAPK pathways were partially inhibited. Overall, our data suggest that LAT-Lck interaction is crucial for the termination of intracellular signaling triggered upon TCR engagement.

cell culture
The LAT-deficient J.CaM2 cell line was generously provided by Dr. Arthur Weiss, University of California, San Francisco (CA, USA). Cells were grown in complete RPMI 1640 medium (Lonza) supplemented with 10% FCS (Lonza) and 2 mM l-glutamine at 37°C in a humidified atmosphere containing 10% CO2.

Mutagenesis and lentiviral Transduction
LAT cDNA cloning was performed as previously described (18). Site-directed mutagenesis was performed to change the sequence for segment between residues 112 and 126. Coding sequences in the plasmids were verified by sequencing and then subcloned in frame with GFP in the SIN lentiviral transfer plasmid LAT Segment Regulating Intracellular Signals Frontiers in Immunology | www.frontiersin.org February 2018 | Volume 9 | Article 115 pHR'SINcPPT-Blast by means of site-specific recombination (Gateway LR Clonase; Invitrogen). Lentiviral supernatants were generated as previously described (18)

Measurement of intracellular ca 2+ Mobilization
Measurement of intracellular free Ca 2+ was carried out using Indo-1 AM (acetoxymethyl) (2 µM; Molecular Probes, Invitrogen) as previously described (18). Calcium measurements were performed using a Synergy MX Multi-Mode Reader (Biotek) at 37°C. Cells were excited by light at a wavelength of 340 nm, and the fluorescence emitted at 405 and 485 nm was collected alternately per second. Calcium mobilization was evaluated by the ratio of 405/485 nm fluorescence signal.

statistical analysis
Western blots were densitometrically quantified, and statistics was performed with Microsoft Excel using a two-tailed t-test. Levels of significance p < 0.05 and <0.005 are presented as * and **, respectively.

resUlTs lentiviral Transduction of J.caM2 cells
It has been previously shown that activation of T cells induces the interaction between LAT and Lck kinase (21)(22)(23). The segment comprising residues 112-126 of human LAT is required for the interaction with Lck, and this sequence is rich in negatively charged amino acids (22). Interestingly, interaction with LAT negatively regulates kinase activity of Lck, suggesting a negative feedback loop of TCR signaling. Concordantly, in LAT-deficient J.CaM2 cells, Lck shows increased kinase activity with regard to Jurkat cells, and restoration of LAT expression decreases its kinase activity (21). At the same time, upon proapoptotic stimuli LAT is cleaved at aspartic residues 126, 138, and 186, leaving an N-terminal fragment that would be able to bind to Lck (18,22). The amino acid sequence of this segment is evolutionarily conserved, supporting a relevant function ( Table 1). This prompted us to analyze the role of this segment of LAT by substituting it with a sequence of amino acids without negative charge ( Figure 1A). The original sequence in LAT is 14 amino acids long and has a net charge of -10 at pH 7. We have introduced a segment of 12 amino acids with a net charge of + 1. The introduced sequence incorporates amino acids predicted to have at least the same or higher flexibility than the original sequence, in order to prevent any conformational effect (24). Lentiviral vectors were generated including the coding region of WT-LAT or the LAT-NIL (Not Interacting with Lck) mutant, fused to a 6His tag ( Figure 1A). Lentiviral vectors contained an IRES and GFP coding sequence as a reporter of the level of transfection, and J.CaM2 cells lentivirally transduced with those vectors coding for WT-LAT or LAT-NIL mutant expressed similar levels of GFP ( Figure 1C). Analysis of cell lysates by Western blot showed similar expression levels for WT-LAT and LAT-NIL mutants, regardless if the Western blot analysis was performed with anti-LAT or anti-6His antibodies ( Figure 1B). Interestingly, electrophoretic mobility of LAT-NIL mutant was increased with regard to WT-LAT, although the predicted molecular weight of both fusion proteins was very similar (27.1 kDa for WT-LAT and 26.4 kDa for LAT-NIL).
In order to verify the membrane expression of both types of LAT molecules, lentiviral vectors coding for fusion proteins with ZsGreen were generated (Figure 2A), and J.CaM2 cells were transduced with these vectors. Transduced cells were then labeled with PKH26, a membrane-labeling red fluorescent marker (25), in order to analyze membrane expression of WT-LAT-ZsGreen and LAT-NIL-ZsGreen fusion proteins by confocal microscopy. As it can be seen in Figure 2B, the change introduced in the negatively charged stretch of LAT does not seem to alter the LAT expression pattern and suggests that WT and the LAT-NIL mutant are excluded from the nucleus and distributed in a mix of vesicles and plasma membrane. The analysis of green and red fluorescence in WT-LAT-ZsGreen and LAT-NIL-ZsGreen expressing cells was represented graphically, showing a similar pattern in both types of cells ( Figure 2C). Moreover, Pearson's correlation coefficient (PCC) values of green and red fluorescence were 0.9452 ± 0.0182 for cells expressing WT-LAT-ZsGreen and 0.9066 ± 0.0235 for cells expressing LAT-NIL-ZsGreen. Altogether, our results suggest that the mutation of the negatively charged stretch of LAT does not affect its subcellular localization.
The negatively charged stretch of laT is needed for activation-induced interaction with lck Next, we sought to verify the role of the negatively charged segment of LAT in the activation-induced interaction with Lck. When it was first cloned, LAT was suggested as a possible molecular partner of Lck, since activated Lck was able to induce low levels of LAT phosphorylation in the absence of ZAP70 (6), and very recently it has been observed that one tyrosine residue of LAT could be phosphorylated by Lck (26). Moreover, an activationinduced interaction of LAT with Lck was previously shown (21,23,27) and this interaction appeared to be mediated by a segment between amino acids 113 and 126 in LAT (22). In order to confirm the relevance of this segment of LAT, we proceeded to immunoprecipitate WT-LAT or the LAT-NIL mutant from stimulated J.CaM2 cells expressing either type of LAT isoforms. To achieve a stimulation level able to induce LAT-Lck interaction, we treated cells for 5 min at 37°C with 5 µM pervanadate or  in Supplementary Material). Consequently, these results support a relevant role for this part of LAT sequence in its stability and/ or turnover. Given that the negatively charged stretch of LAT is needed for the activation-induced interaction with Lck, we sought to evaluate whether the presence of Lck could modify LAT turnover. Therefore, we performed a cycloheximide chase assay with Jurkat cells and J.CaM1.6 cells deficient in Lck expression. As it can be seen in Figure S1B in Supplementary Material,  JCaM1.6 expresses lower levels of LAT with regard to Jurkat cells. Moreover cycloheximide treatment induces a greater loss of LAT expression in J.CaM1.6 cells with regard to Jurkat cells, suggesting that Lck interaction could somehow stabilize LAT. Next, given that the substituted sequence in LAT-NIL isoform included one of the described cleavage points in LAT (18), we decided to verify if the substitution of this amino acid residue altered the cleavage of LAT induced upon Fas engagement. Therefore, we treated J.CaM2 cells expressing WT-LAT or LAT-NIL with anti-Fas antibody for 4 h at 37°C. As it can be seen in Figure 4B (lanes 1-4), cleavage of LAT-NIL also produced two major proteolytic fragments. The calculated molecular weight of proteolytic fragments generated in WT-LAT upon cleavage at Asp186 and Asp138 are 19.9 and 15.1 kDa, respectively, and the fragments that LAT-NIL should produce upon cleavage of LAT-NIL at Asp184 and Asp136 (LAT-NIL is two amino acids shorter than the WT-isoform) are very similar to the WT counterpart (i.e., 19.2 and 14.4 kDa, respectively). Interestingly, the difference in the electrophoretic mobility of proteolytic fragments generated by LAT-NIL with regard to the ones produced by WT-LAT was greater than expected. Thus, in order to explain the origin of the proteolytic bands observed in LAT-NIL, we generated two additional mutants in which aspartic residues 136 and 184 were mutated to alanine (mutants LAT-NIL D136A and LAT-NIL D184A , respectively). Anti-Fas treatment of LAT-NIL D136A expressing cells generated only one proteolytic fragment, the one with lower electrophoretic mobility observed in LAT-NIL expressing cells ( Figure 4B, lanes 7 and 8). Conversely, Fas-mediated cleavage of LAT-NIL D184A mutant generated only the lower band, indicating that the upper band is generated by cleavage at aspartic at position 184 of LAT-NIL, and the lower band is produced upon cleavage at Asp136. Overall these results show that the negatively charged stretch of LAT has a role in the stability of the molecule, without a role in the proteolytic cleavage of this membrane adapter.

analysis of Proximal Tcr intracellular signals in laT-nil expressing cells
Since interaction of Lck with LAT downregulates its kinase activity (21), we next investigated early TCR signaling in LAT-NIL expressing cells. We were not able to directly detect activation-induced phosphorylation of Tyr394 of Lck in J.CaM2 cells expressing WT-LAT or the LAT-NIL mutant (data not shown), although this was not unexpected because it has been previously observed that anti-CD3 stimulation of Jurkat cells or primary human T cells does not induce substantial increase in the phosphorylation of this tyrosine residue (29,30). So, considering that ZAP70 is an Lck kinase substrate (26) and constitutes one of the initial intracellular signals triggered after TCR engagement in T cells, we analyzed ZAP70 phosphorylation at tyrosine residue 319, which is required for its enzymatic activation (31). J.CaM2 cells expressing WT-LAT or LAT-NIL were stimulated with anti-CD3 antibodies, and cell lysates were analyzed by Western blot with a specific antibody for phosphorylated Tyr 319 of ZAP70 ( Figure 5). As expected, CD3 stimulation of WT-LAT expressing cells induced a rapid increase in ZAP70 phosphorylation.
Interestingly, activation-induced ZAP70 phosphorylation was increased in LAT-NIL with regard to WT-LAT expressing cells. Although our results only showed borderline statistical significance (p = 0.051 at 3 min), this result suggests a possible negative regulatory role of LAT-Lck interaction (21,22). Provided that LAT is a ZAP70 substrate, we next analyzed LAT phosphorylation in lysates obtained in CD3-stimulated cells. LAT phosphorylation at tyrosine residues 171 and 191 (Figures 6A,B, respectively)  was enhanced in LAT-NIL expressing cells, at all analyzed time points, and statistically significant differences were observed after 10 and 20 min of CD3 stimulation ( Figure 6A). Also, at 3 and 40 min of stimulation the difference in Tyr171 phosphorylation between WT-LAT and LAT-NIL almost attained statistical significance (p = 0.059 and p = 0.071, respectively). We also analyzed phosphorylation of LAT tyrosine 191, and again we observed an increase in the phosphorylation of this residue in LAT-NIL, with statistical significance at 3 and 10 min time points ( Figure 6B). Strikingly, blots of phospho-Tyr171 and phospho-Tyr191 show rather similar electrophoretic mobility, although the same membranes stripped and blotted with total anti-LAT antibody show the difference in the electrophoretic mobility already observed with anti-LAT and anti-6His antibodies ( Figure 1B).
Considering the proximity of the negatively charged fragment of amino acids substituted in LAT-NIL to tyrosine 132, we next analyzed the effect of this mutation on the phosphorylation of Tyr 132. Contrary to the results obtained with tyrosines 171 and 191, Western blot analysis performed in whole-cell lysates obtained from anti-CD3-stimulated cells expressing either WT-LAT or the LAT-NIL revealed no differences in the intensity or the kinetics of phosphorylation of Tyr 132 (Figure 7A). The same analysis was done for LAT tyrosine 226, and again there were no differences between WT-LAT and LAT-NIL ( Figure 7B). Interestingly, as it happened in the analysis of phosphorylation of LAT tyrosines 171 and 191, we observed again the same behavior in the electrophoretic mobility of WT-LAT and LAT-NIL: membranes probed with phospho-specific antibodies showed a small decrease in the electrophoretic mobility of LAT-NIL with regard to WT-LAT, and after stripping and hybridization with anti-LAT antibody, the difference in electrophoretic mobility was increased. Given the defect observed in LAT-NIL stability after cycloheximide treatment (Figure 4A), we decided to verify LAT expression in cultured cells in order to exclude that the observed effects were not due to a diminished stability of LAT-NIL. Consequently, we performed Western blotting analysis of cell lysates obtained from J.CaM2 cells 1, 2, 3, and 4 weeks after lentiviral transduction with WT-LAT or LAT-NIL vectors. As it can be seen in Figure S2 in Supplementary Material, J.CaM2 cells lentivirally transduced with the LAT-NIL mutant stably expressed similar levels to WT-LAT, either with anti-LAT or anti-6His antibodies. Therefore, these results show that substitution of the negatively charged stretch of LAT has differential effects on the phosphorylation of its four functionally relevant tyrosines, affecting its electrophoretic mobility.
substitution of the negatively charged stretch of laT negatively affects Downstream signals Considering the differential effects induced by the substitution of the negatively charged stretch of LAT in the phosphorylation of its tyrosine residues, we were then interested in the analysis of downstream signals. Previous reports had shown the negative impact of mutation of tyrosines 171 and 191 in TCR downstream signals (6,11,32). Thus, the increased phosphorylation of these residues observed in LAT-NIL should induce an increase in downstream signals. Both tyrosine residues have an identical Grb2-binding motif (YVNV), a small cytosolic adaptor consisting in two SH3 domains and one SH2 domain, which binds to phosphorylated tyrosines 171, 191, and 226 in LAT (6,33). Grb2 constitutively binds through its SH3 domains to a proline-rich region in the Ras guanine nucleotide exchange factor, SOS, and phosphorylation of LAT recruits SOS close to Ras at the cell membrane, allowing the activation of the MAPK pathway (34). Therefore, we analyzed phosphorylation of Erk at Thr202 and Tyr204 in WT-LAT and LAT-NIL expressing cells, since phosphorylation of these two residues is indicative of Erk enzymatic activation (35). As expected, anti-CD3 stimulation of J.CaM2 cells expressing WT-LAT induced a rapid increase in Erk phosphorylation after 3 min (Figure 8A). LAT-NIL expressing cells showed an initial similar increase in activation-induced Erk phosphorylation after 3 min of stimulation, although at longer time points the kinetics were different, with LAT-NIL showing a slightly reduced phosphorylation. However, we did not find statistically significant differences between Erk phosphorylation in WT-LAT and LAT-NIL expressing cells. In order to go deeper into the analysis of the effects of LAT-NIL mutant in the MAPK pathway, we analyzed activation of MEK, the kinase responsible of the phosphorylation and activation of Erk (36). Therefore, we used a specific antibody to phospho-MEK-1/2 to investigate MEK activation. CD3 stimulation of WT-LAT expressing cells rapidly induced phosphorylation of MEK and the same was observed for LAT-NIL expressing cells ( Figure 8B). However, at longer time points, LAT-NIL expressing cells showed a faster decay in the phosphorylation level of MEK, with regard to WT-LAT expressing cells, similar to Erk phosphorylation kinetics. Densitometric quantification of five independent experiments showed that the   We next analyzed the effect of replacing the fragment between residues 112 and 126 of LAT on PLC-γ1 activation. TCR ligation leads to rapid LAT phosphorylation, thereby allowing the recruitment of PLC-γ1 to the cellular membrane, where it is phosphorylated on Tyr783, a hallmark of its enzymatic activation (38). Anti-CD3 stimulation induced phosphorylation of PLC-γ1 in both WT-LAT and LAT-NIL J.CaM2 expressing cells ( Figure 9A). The kinetics were similar in both types of cells, although there was a slightly enhanced increase in PLC-γ phosphorylation in LAT-NIL expressing cells, but without statistically significant differences. This observation was consistent with the similar intensity and kinetics of phosphorylation of LAT tyrosine 132, the principal binding site for PLC-γ1. Consequently, we next analyzed whether the substitution of the negatively charged fragment of LAT had an effect on calcium influx generation. J.CaM2 cells are deficient in TCR-mediated intracellular calcium generation (8), and lentiviral expression of the WT form of LAT restored Ca 2+ influx ( Figure 9B). CD3 stimulation of LAT-NIL expressing cells also restored the increase in cytoplasmic Ca 2+ concentration, with initial kinetics very similar to WT-LAT expressing cells. However, J.CaM2 cells expressing the LAT-NIL mutant had a statistical significant reduction in both the peak and the sustained levels of calcium influx. Therefore, these results show the calcium influx is negatively affected by the substitution of the negatively charged stretch of LAT. Although similar kinetics in PLC-γ1 phosphorylation and reduced Ca 2+ influx seem contradictory, it has been recently observed that phosphorylation of PLC-γ1 is initiated very quickly after TCR stimulation, similar to LAT tyrosine 132; however, 20 and 30 min after stimulation a significant portion PLC-γ1 is still phosphorylated, but phosphorylation of LAT Tyr132 has recovered basal levels (39). These data suggest that the current model of PLC-γ1 activation is incomplete, and although phosphorylation of LAT at Tyr132 is needed for the initial binding of PLC-γ1, this interaction is transient and other signals regulate PLC-γ1 activation at longer time points. Our results support this report, and substitution of sequence between amino acids 112 and 126 in LAT generates diminished calcium mobilization without appreciable differences in LAT-Tyr132 phosphorylation.

DiscUssiOn
The integral membrane adaptor LAT has a crucial role in TCRmediated activation of T cells, and its five C-terminal tyrosine residues are needed for the transduction of signals leading to cell activation (1). However, data obtained from the analysis of several strains of mice expressing different mutant forms of this adaptor have revealed that LAT has an intrinsic ability to regulate T-cell homeostasis, although the molecular basis for this negative role is not completely understood (14)(15)(16)(17)40). A self-explanatory negative feedback loop involving LAT was described in 2004 by Dr Koyasu's group, by which TCR-mediated activation would lead to LAT-dependent activation of Erk, which in turn was able to phosphorylate human LAT at Thr155, thereby reducing the ability of LAT to recruit PLC-γ1 and decreasing Ca 2+ mobilization and MAPK pathway activation (41). This negative feedback loop would prevent TCR signaling with inappropriate duration or intensity, which could be potentially harmful for the immune system, and failures of this regulation mechanism could be at the base of immune disorders. However, this threonine residue is not conserved in other species like mouse or rat, excluding an essential role for Thr155 in the negative functions played by LAT in these species and preventing in vivo analysis of the role played by such feedback loop.
In this context, it has been previously described that upon TCR-mediated activation of T cells, LAT interacts with Lck and this interaction decreases Lck kinase activity (21,23). Very interestingly, by means of expressing truncated forms of LAT, it was shown that a truncated form of LAT at Asp126 was still able to interact with Lck but not an isoform truncated at Asn112 (22). Therefore, LAT-Lck interaction could constitute a model for termination of activatory signals coming from the TCR/ CD3 complex. The fragment between residues 112 and 126 in human LAT is composed by a stretch of negatively charged amino acids, and this sequence is evolutionarily conserved in human, mouse, rat, gorilla, chimpanzee, cow, cat, and other species, supporting an important role for this fragment of LAT for its functions in intracellular signaling coupled to the TCR/CD3 complex (see Table 1). Remarkably, this fragment is preceding the most N-terminal cleavage point of human LAT (18), and Fas-mediated cleavage at this point would generate a LAT fragment still able to bind to Lck and diminishing its kinase activity (21,22). This prompted us to analyze the role of this stretch by means of substituting it with a flexible peptide fragment without negatively charged amino acids. Our results confirm that this sequence of LAT is necessary for the activation-induced interaction of LAT with Lck kinase, since the LAT-NIL mutant did not show the increase of LAT-Lck interaction previously described upon PHA or pervanadate stimulation (22,23,42), contrary to WT-LAT. Moreover, we have shown that LAT-Lck interaction constitutes a regulatory mechanism for the TCR signaling cassette, since the mutation of the negatively charged stretch of amino acids in LAT increases the TCR-mediated phosphorylation of Tyr319 in the interdomain B of ZAP70, required for the activation of ZAP70 function (31). We were not able to directly detect activation-induced phosphorylation of Tyr394 of Lck in J.CaM2 cells expressing WT-LAT or the LAT-NIL mutant (data not shown), although this was not unexpected because it has been previously observed that anti-CD3 stimulation of Jurkat cells or primary human T cells does not induce substantial increase in the phosphorylation of this tyrosine residue (29,30).
Consistent with the increase in ZAP70 activation, anti-CD3 stimulation of LAT-NIL expressing cells induced augmented phosphorylation of LAT tyrosines 171 and 191 with regard to WT-LAT expressing cells. However, tyrosine residues 132 and 226 showed a different behavior, with similar levels of phosphorylation in the basal state and at different time points after TCR stimulation. This was not completely unexpected, since there are multiple evidences of functional specialization of the last five C-terminal tyrosine residues of LAT (11,15,32,43). Moreover, two independent groups have previously demonstrated that tyrosine residues 132 and 191 have different kinetics of phosphorylation, with a delay in the phosphorylation of Tyr132 with regard to Tyr191 (39,44). Strikingly, we have observed an increase of phosphorylation of both Tyr171 and Tyr191 in LAT and Tyr319 of ZAP70 in LAT-NIL expressing cells. Increase in phosphorylation of Tyr171 as a consequence of augmented activity of ZAP70 in LAT-NIL expressing cells is also consistent with a recent report by Kuriyan and co-workers (26). Kuriyan's group has used an in vitro approach to identify peptide substrates for ZAP70 and Lck, showing that tyrosine 191 in LAT is a relatively poor substrate for ZAP70, which is not in agreement with the results presented in this work. Nevertheless, we cannot discard that phosphorylation efficiency in the whole LAT molecule, attached to the membrane, can be somewhat different to the phosphorylation of peptides bound to the surface of bacteria. Moreover, it has been recently shown that LFA crosslinking induces phosphorylation of LAT at tyrosine 171 by Focal Adhesion Kinase 1 (FAK1), and this event leads to a reduction in proliferation and adhesion of T cells to dendritic cells (45). This is consistent with our results, since the mutation introduced in LAT-NIL increased phosphorylation of Tyr171, leading to a reduction of downstream intracellular signals. More work is needed to clarify the molecular mechanisms underlying this observation.
In contrast to the increase in early intracellular signals, more distal signals such as the MAPK pathway activation or the increase in the calcium concentration are negatively affected by the substitution of the negatively charged stretch in LAT. LAT-NIL expressing cells show a statistically significant reduction in the phosphorylation of Thr202 and Tyr204 in Erk, which is indicative of its enzymatic activation (35). This result was in agreement with the reduction of phosphorylation of MEK, the kinase responsible of Erk phosphorylation. Indeed we have observed a strikingly similar kinetics for both Erk and MEK phosphorylation, with a rapid and very similar increase in phosphorylation at short time points in WT-LAT and LAT-NIL expressing cells (3 min), and a reduction in phosphorylation at longer time points in LAT-NIL expressing cells (more than 10 min). In the same way, although LAT-NIL expressing cells rapidly increased calcium concentration after CD3 stimulation, they showed a significant reduction in both the peak and the sustained levels of calcium influx with respect to WT-LAT expressing cells. However, PLC-γ1 phosphorylation at Tyr783, indicative of its enzymatic activation, was similar in both types of cells (with a slightly enhanced increase in LAT-NIL expressing cells). We have also seen no differences in the phosphorylation of LAT tyrosine 132, the binding point of PLC-γ1. Although both observations seem inconsistent, it has been recently reported that even if initial activation of PLC-γ1 requires LAT-Y132 phosphorylation, once triggered, PLC-γ1 activation is independent of LAT and LAT and activated PLC-γ are located in different membrane clusters (39), suggesting that although phosphorylation of LAT at Tyr132 is needed for the initial binding and activation of PLC-γ1, this interaction is transient and other signals are needed to keep PLC-γ1 activated at longer time points. Our results are in agreement with these observations, and although the substitution of sequence between amino acids 112 and 126 in LAT does not appreciably modify phosphorylation of Tyr132, a statistically significant decrease in calcium mobilization is observed a long times ( Figure 9B). One possible explanation for these results could be that the imbalance generated by the increased tyrosine phosphorylation of Tyr171 and Tyr191 could lead to diminished activation of the MAPK pathway, calcium mobilization or CD69 expression, as it happens in two LAT-KI strains of mice (14)(15)(16). Another possible explanation for the decrease in downstream signaling is the reduced ability of Grb2 recruitment of LAT-NIL mutant, which is in agreement with a previous report showing that Grb2 deficient cells have reduced MAP kinase activation and calcium flux (46).
It is also of interest the reduction observed in the stability of LAT-NIL mutant with regard to the WT isoform. This result is consistent with our previous observation of a LAT isoform resulting from an intron 6 retention event, which generates an in-frame splice denoted as LATi6 (28). The amino acid sequence introduced as a consequence of the retention of intron 6 is located just after Glu113, at the beginning of the negatively charged stretch of LAT needed for the interaction with Lck, confirming a relevant role for this part of LAT in its stability and/or turnover. Interestingly, it seems that the decreased interaction with Lck could be at the base of this reduction of LAT-NIL stability, since Lck deficient J.CaM1.6 cells express basal lower levels of LAT with regard to Jurkat cells (from which they are derived) (47). Moreover, the stability of LAT in J.CaM1.6 cells after cycloheximide treatment is clearly reduced with regard to LAT in Jurkat cells. Although it has been previously shown that the lack of LAT expression in J.CaM2 cells decreases Lck enzymatic activity (21), to our knowledge this is the first time that a reduction of LAT expression in Lck-deficient cells has been reported. Therefore, our data support a functional relationship between LAT and Lck, with Lck regulating LAT expression and LAT regulating Lck activity. Additional work should be done in order to reach a deeper understanding about this functional relationship.
Additionally, we have observed a change in electrophoretic mobility when the negatively charged stretch of LAT is substituted by a sequence of 12 amino acids with a net charge of +1, two residues shorter. We decided to introduce a shorter sequence in order to easily discriminate between the two isoforms, but in our view the difference in electrophoretic mobility is greater than expected. Moreover, Western blots with anti-LAT phosphospecific antibodies showed smaller differences in electrophoretic mobility than anti-total LAT or anti-6His antibodies in the same stripped membranes. It has been demonstrated that the p36/38 doublet is not due to alternative splicing, but probably to posttranslational modifications of unknown nature. Although we have no explanation for this change in the mobility of LAT-NIL isoform, we think that phosphorylation differences are not at the basis of this event, since the basal phosphorylation state of tyrosine residues 132, 171, 191, and 226 is similar between WT and the LAT-NIL mutant, and calf intestinal phosphatase treatment did not modify electrophoretic mobility of either WT-LAT or LAT-NIL isoforms (data not shown).
In summary, the results presented extend previous reports about the role of LAT as an intrinsic regulator or terminator of intracellular signals associated to the TCR/CD3 complex. CD95 ligation in naive T cells interferes with proximal intracellular signals, probably by means of proteolytic cleavage of essential proteins in the signaling cassette (18,20,(48)(49)(50). In this way, it is conceivable a scenario in which activated T cells expressing Fas interact with FasL expressing APCs and this interaction would induce LAT cleavage, preventing the transduction of high intensity or very prolonged activation signals. However, the TCR/ CD3 complex is able to transduce intracellular signals in a LATindependent way (51), and LAT cleavage would constitute an incomplete terminator mechanism. Although our study has the limitations due to the use of a cell line rather than primary T cells, our data suggest that Fas-mediated cleavage of LAT generates a membrane attached LAT isoform, still able to bind to Lck, inhibiting its kinase activity for a more complete termination of signals, with the negatively charged stretch of LAT having an essential role in this regulatory interaction.

eThics sTaTeMenT
This study was carried out in accordance with the recommendations of the University of Cádiz and Puerto Real University Hospital ethics committees.

aUThOr cOnTriBUTiOns
All authors contributed to discussions of experimental design and data analysis. MA-E, IN-S, and EA designed and performed most experiments and interpreted results. MA-E and EA wrote the manuscript. CF-P and IV-B provided technical assistance with transfections, cell culture and confocal microscopy. MD-R and IV-B provided assistance with Western blotting. MD, MR-Y, FG-C, MD-R, and AM provided input on study design and helped with manuscript writing. EA directed the study.

acKnOWleDgMenTs
The authors would like to thank Antonio Trujillo-Aguilera, Emilio de la Orden, and Consuelo Rivera for excellent technical assistance. We are also grateful to Dr Weiss for donating J.CaM2 cells.

FUnDing
This work was supported by Consejería de Salud de Andalucía, Junta de Andalucía (grants PI-0365-2013 and PI-0055-2017 to EA) and Instituto de Salud Carlos III (grant PI16-00784 from the "Plan Estatal de I+D+I 2013-2016/FEDER" to MD-R and FG-C).