Tyr¹⁹² Regulates Lymphocyte-Specific Tyrosine Kinase Activity in T Cells

This article is featured in Top Reads, p.999

Tcells are a crucial part of the adaptive immune system. They have highly specific receptors designed to recognize Ags of pathogenic origin. Interaction of a TCR with its cognate Ag (in the form of a peptide bound to a MHC) results in cytoplasmatic exposure of its associated CD3 chains (1). Phosphorylation of exposed CD3 chains by lymphocyte-specific protein tyrosine kinase (Lck) is a crucial step in the TCR signaling pathway, leading to activation of the T cell.

Lck contains three domains: Src homology 2 (SH2), Src homology 3 (SH3), and tyrosine kinase. The activity of Lck is regulated via tyrosine phosphorylation (pTyr). When phosphorylated, C-terminal Tyr⁵⁰⁵ of Lck interacts with its own SH2 domain, resulting in a closed, inactive conformation of Lck (2). Conversely, dephosphorylation of Tyr⁵⁰⁵ opens Lck, allowing it to transphosphorylate Tyr³⁹⁴ in the catalytic site of another Lck molecule (3). pTyr³⁹⁴ stabilizes interaction of Lck with ATP and is required for the fully active conformation of Lck. Key enzymes involved in this process include C-terminal Src kinase (Csk), which phosphorylates Tyr⁵⁰⁵, and the tyrosine phosphatase CD45, which dephosphorylates it (4). Meanwhile, SHP1 (5), PTPN22 (6), and CD45 (7) can dephosphorylate Lck Tyr³⁹⁴.

SH2 domains recognize specific amino acid motifs containing pTyr. Sequence variability in two loops, EF and BG, defines the SH2 domain specificity (8), which is crucial for propagating pTyr dependent signaling cascades, including the TCR signaling pathway. The SH2 domain of Lck may have multiple roles, including capturing substrates for the kinase, and positioning of the kinase at specific locations within the cell (2). The list of established Lck SH2 domain-binding protein partners includes ZAP70, CD3ε, CD45, LIME, TSAd, SHP1, Syk, Pyk2, Itk, Lnk, and Lck itself (9–19). Phosphopeptides derived from many more proteins have been shown to bind to the SH2 domain of Lck (11), but the importance of these peptides for the interaction of Lck with the intact proteins still needs to be validated.

Lck’s Tyr¹⁹² is a conserved residue, found in all Src kinase SH2 domains (i.e., Src Tyr²¹⁵) (11). Phosphorylation of this site in Src (20) and Fyn (21) has been associated with increased kinase activity. Lck Tyr¹⁹² is phosphorylated upon TCR stimulation by Syk, ZAP70 (22, 23), and Itk (11). Subsequently, pTyr¹⁹² changes SH2 domain specificity of Lck (11, 22). When used to reconstitute Lck deficient cells, Lck Glu¹⁹² (mimicking the negative charge of pTyr) was found to be hyperphosphorylated on Tyr⁵⁰⁵, since Lck Glu¹⁹² is unable to interact with CD45 (13). However, the biological consequence of altered Lck SH2 domain function by pTyr¹⁹² is not well understood.

In this study, we aimed to explore functional aspects of endogenously expressed Lck harboring mutations mimicking phosphorylation or dephoshorylation of Lck Tyr¹⁹². To this end, we used CRISPR/Cas9 genome editing to generate Jurkat cell lines expressing Lck Glu¹⁹² or Phe¹⁹². As we found that endogenously expressed Lck Glu¹⁹² is hyperphosphorylated on Lck Tyr⁵⁰⁵, we also generated a set of mutants with a Phe⁵⁰⁵ substitution. We analyzed the binding capacity, activity, and phosphorylation status of mutated Lck by immunoprecipitation (IP) experiments followed by mass spectrometry as well as immunoblotting. Our data show that Tyr¹⁹² is a key regulator of Lck activity by affecting phosphorylation of Lck Tyr³⁹⁴ and Tyr⁵⁰⁵, as well as its SH2 domain binding ability.

Abs used for immunoblotting were as follows: anti-Lck (clone 3A5, Santa Cruz Biotechnology), anti-Plcγ (clone 1249, Santa Cruz Biotechnology), anti-LAT (clone LAT-01, EXBIO), anti-VAV1 (clone C-14, Santa Cruz Biotechnology), anti-CD3ζ (clone 6B10.2, Santa Cruz Biotechnology), anti-ZAP70 (clone 29, BD Biosciences), anti-p44/42 MAPK (Erk1/2) (no. 9102, Cell Signaling Technology), anti-GST (clone B14, Santa Cruz Biotechnology), anti-pTyr (clone 4G10, Upstate Biotechnology), anti-pTyr⁵⁰⁵ Lck (no. 2751, Cell Signaling Technology), anti-pTyr⁴¹⁶ Src (no. 2101, Cell Signaling Technology), anti-pTyr⁷⁸³ Plcγ (clone D6M9S, Cell Signaling Technology), anti-pThr²⁰²/pTyr²⁰⁴ p44/42 MAPK (Erk1/2) (clone E10, Cell Signaling Technology), anti-pTyr¹⁴² CD3ζ (clone K25-407.69, BD Biosciences), anti-pTyr³¹⁹ ZAP70 (no. 2701, Cell Signaling Technology), anti–non-pTyr⁴¹⁶ Src (clone 7G9, Cell Signaling Technology), and anti-GADPH (clone 6C5, Chemicon). The secondary reagents, conjugated Abs, and staining dyes used were: HRP conjugated Light chain specific goat anti-mouse, HRP conjugated goat anti-mouse (H+L) and HRP conjugated mouse anti-rabbit Abs (Jackson ImmunoResearch), anti-CD69 FITC (clone FN50, ImmunoTools), anti-NFATc1 Alexa Fluor 488 (clone 7A6, BioLegend), Alexa Fluor 647–conjugated goat anti-mouse IgG (H+L) cross-adsorbed (Thermo Fisher Scientific), biotinylated alpaca anti-GFP V_HH nanobodies (Chromotek), DAPI (Invitrogen, Thermo Fisher Scientific), CellTrace Violet (CTV; Thermo Fisher Scientific), Atto 425 NHS ester (Sigma-Aldrich). The following were used for stimulation of cells: anti-CD3ε (clone OKT-3, American Type Culture Collection), anti-TCR (clone C305, a gift from Arthur Weiss), anti-CD28 (clone CD28.2, BD Biosciences), AffiniPure F(ab′)₂ fragments goat anti-mouse IgG (H+L) (Jackson ImmunoResearch Laboratories), PMA, ionomycin (IO), sodium orthovanadate, and hydrogen peroxide (all Sigma-Aldrich). Plasmids used were pX330-U6-chimeric_BB-CBh-hSpCas9 (24) (Addgene), pEGFP-N1 (Clontech), pEF-Bos-CD8-ζ, pEF-HA-Zap-70 (gifts from A. Weiss, Department of Immunology, University of California, San Francisco, CA), pEGFP-N1-Lck (encoding wild type [WT], Lck Y192F, or Lck Y192E) (11).

Plasmids encoding Lck-SH2 WT, Y192F, or Y192E GST fusion proteins were expressed in BL21 codon plus bacteria (Stratagene) and purified on glutathione Sepharose beads (Amersham Biosciences) as described previously (11). Protein purity and concentrations were assessed with SDS-PAGE and Coomassie Brilliant Blue staining, and by the bicinchoninic acid assay (Thermo Fisher Scientific), respectively.

Human embryonal kidney (HEK) 293T cells (American Type Culture Collection), Jurkat JE6.1 (25), Jurkat TAg cells (26), and all CRISPR/Cas9-generated mutated cell lines were cultured in RPMI 1640 supplemented with 10% FCS, 1 mM sodium pyruvate, 1 mM nonessential amino acids, 1 mM HEPES buffer, 100 U/ml penicillin, and 100 µg/ml streptomycin (all from Gibco BRL, Thermo Fisher Scientific) and 50 µM 2-ME (Sigma-Aldrich).

Up to 3 × 10⁶ HEK 293T cells were seeded on 10-cm plates 24 h prior to transfection. 2 µg of plasmid DNA was mixed with 32.25 µg of polyethyleneimine (Polysciences) and added to cell cultures. Cells were harvested 24 h later and immediately used for IP experiments.

The full protocol and methodological considerations of Jurkat TAg cell knock-in mutant preparation are described elsewhere (27). Prior to gene editing, Jurkat TAg cells were subcloned. Clones were assessed for expression of Lck and surface CD3 by immunoblotting and flow cytometry, respectively. Clones, which displayed similar amounts as the median values of bulk Jurkat TAg cells, were chosen for further experiments. Two subcloned Jurkat TAg cell lines were cotransfected with Cas9 coding plasmid (pX330-U6-chimeric-BB-CBh-hSpCas9, Addgene) with guide sequences designed to target the Lck at the codon for Tyr¹⁹², pEGFP-N1 plasmid (Clontech), and ssDNA repair sequences (Lck Y192F, Lck Y192E, or Lck Y505F). Knockout (KO) cell lines (Lck KO) were obtained as byproducts of this procedure. 10 − 15 × 10⁶ cells and 2–5 µg of DNA were resuspended in 400 µl of RPMI 1640 medium without antibiotics supplemented with 5% FCS and transfected by electroporation with a BTX electroporator (Genetronix) at 240 V for 25 ms. After 3 days, transfected cells were cloned by limiting dilution. Three weeks later clones were screened by mutation-specific PCR, and mutations were confirmed by sequencing (GATC Biotech). Lck Y505F, Lck Y192F/Y505F, and Lck Y192E/Y505F cell lines were generated simultaneously, although the Lck Y505F mutation was introduced before Lck Y192F/E. All experiments were repeated in at least two independent clonal cell lines harboring the same mutations.

For pervanadate (PV) phosphatase inhibition treatment, cells were suspended in PBS in a concentration of up to 10⁷ cells/ml and prewarmed at 37°C water bath for 5 min. Cells were treated with 0.01 mM sodium orthovanadate (Sigma-Aldrich) and 0.01% hydrogen peroxide (Sigma-Aldrich) for 5 min. Treatment was stopped by sample transfer to 4°C PBS followed by pull-down (PD) experiments (see below). For PMA/IO stimulation, cells were stimulated with 50 ng/ml PMA and 500 ng/ml IO overnight. Cells were washed and rested overnight. For long-term TCR stimulation, cells in RPMI 1640 containing 10% FCS were added to plates or flasks precoated with 1 µg/ml anti-CD3 (OKT3) (or with 1 µg/ml anti-CD3 and 0.5 µg/ml anti-CD28) in PBS and incubated overnight at 37°C. Cells were washed and rested in new plates/flasks overnight. For NFAT translocation experiments, cells were stimulated on anti-CD3–coated plates for only 3 h prior to the experiment. For short-term TCR stimulation, up to 10⁸ cells/ml in PBS were prewarmed at 37°C for 5 min. Cells were stimulated with 5 µg/ml anti-TCR Ab (OKT3 or C305) for the indicated time (2 min, unless specified otherwise). For combined stimulation with 5 µg/ml anti-CD3 and 2 µg/ml anti-CD28, additional crosslinking with 2.5 µg/ml anti-mouse F(ab′)₂ fragments was used. For control purposes, 2.5 µg/ml anti-mouse F(ab′)₂ fragments was added to all samples (including not stimulated). For kinetics experiments, stimulation was stopped by adding 1 ml of cold PBS at the time points 0, 2, 5, 10, 20, 30, and 60 min, and cells were collected for lysis.

Stimulated or treated cells were pelleted and lysed with 0.1% LDS/1% Triton solution containing 0.5% Triton X-100 (VWR International), 50 mM HEPES (Gibco BRL, Thermo Fisher Scientific), 0.05% LDS (Merck), 0.05 M LiCl, 0.5 mM PMSF, 2.5 mM EDTA (pH 8.0), 1 mM sodium vanadate, and 1× SIGMAFAST protease inhibitor mixture (all from Sigma-Aldrich). Cells were lysed for 45 min on ice. Lysates were sonicated briefly to break the DNA. Prior to IP, lysates were preabsorbed using protein G Dynabeads (Invitrogen, Thermo Fisher Scientific). Protein G Dynabeads used for IP, coated with appropriate Ab (anti-Lck or anti-pTyr) at room temperature for 1 h, were washed and incubated with lysates on a rotating wheel at 4°C for 1 h. Glutathione beads (Amersham Biosciences) coated with Lck-SH2 domain GST fusion molecules and MyOne streptavidin T1 Dynabeads (Invitrogen, Thermo Fisher Scientific) coated with anti-GFP V_HH biotinylated nanobodies (Chromotek) (used to eliminate background caused by H chains of standard Abs) were used analogically. After IP, beads were washed three times with lysis buffer, and the appropriate volume of loading buffer (containing 0.35 M Tris HCl, 10% SDS, 6% 2-ME [all Sigma-Aldrich], 30% glycerol [VWR International], and 0.175 mM bromphenol blue [Fluka], pH 6.8) was added. Samples were denatured at 95°C for 10 min, separated by SDS-PAGE, and transferred onto PVDF membranes (Trans-Blot Turbo transfer system [Bio-Rad Laboratories]). Membranes were incubated overnight at 4°C with primary Abs diluted in TBS (pH 7.4), 0.1% Tween 20 (Sigma-Aldrich), and 3% skimmed milk or for anti-pTyr Abs with 3% BSA (Bio-Rad Laboratories), followed by incubation with the appropriate HRP-conjugated secondary Ab for 1 h at room temperature. SuperSignal West Pico stable peroxide solution (Pierce) and a ChemiDoc imaging system (Bio-Rad Laboratories) were used to visualize bands. Images were quantified with ImageJ software. The normalization method for each experiment is given in the Results or in the figure legends. The quantification of immunoblot signals was performed on raw, unsaturated, 16-bit images. The images presented in the figures have adjusted brightness and contrast and are downgraded to 8-bit format, but they accurately represent the data.

Supernatant collected from stimulated cells was subjected to an ELISA MAX human IL-2 standard set (BioLegend) according to the manufacturer’s instructions using Nunc MaxiSorp 96-well plates (Thermo Fisher Scientific). All standards and samples were measured in duplicates. Absorbance measured at 570 nm was subtracted from the absorbance measured at 450 nm.

For surface staining (CD69), cells were barcoded with four different concentrations of CTV (Thermo Fisher Scientific) or Atto 425 NHS ester (Sigma-Aldrich) according to the manufacturers’ protocols. Differentially stimulated samples were stained with varying concentrations of CTV (or with Atto after fixation) and combined respectively. The combined samples were incubated for 20 min on ice with surface Ab and fixed with 2% paraformaldehyde. Fixed cells were analyzed with FACSCanto II flow cytometer (BD Biosciences) and FlowJo software.

For intracellular staining (NFAT), after 3 h of TCR stimulation, cells were fixed with 2% paraformaldehyde at room temperature, permeabilized with 0.1% Triton X-100 for 15 min, and subsequently stained for 1 h at room temperature with a mix of anti-NFATc1 Ab Alexa Fluor 488 and DAPI (1.7 μg/ml). Cells were analyzed with ISX imaging flow cytometer and IDEAS 6.2 software (Amnis) as described previously (11). Briefly, 10,000 cells per sample were collected with a ×40 objective. Cells were gated on a bright field area plotted against an aspect ratio graph to distinguish single cells from cell conglomerates and debris, and on a root mean square gradient to get only cells that were in focus. Dead cells were excluded by gating on the bright field area plotted against dark field side scatter. Data were analyzed with IDEAS 6.2. Similarity score was calculated by IDEAS as previously described (28). Briefly, the program does pixel-by-pixel comparison of the NFAT signal with DAPI signal.

Samples for label-free quantitative proteomics were prepared according to the protocol for Lck IP. Afterwards, proteins were reduced, alkylated, and digested directly on beads with ProteaseMAX surfactant, trypsin enhancer (Promega) overnight at 37°C, according to the manufacturer’s protocol. Samples were desalted with the STAGE-TIP method using C18 resin disk (3M Empore). Peptide elution was performed with 0.1% formic acid and 80% acetonitrile. Eluates were concentrated with SpeedVac until a volume of ∼7 µl. Samples were analyzed on an Easy nLC1000 nano-LC system connected to a quadrupole Orbitrap mass spectrometer (QExactivePlus, ThermoElectron, Bremen, Germany) equipped with a nanoelectrospray ion source (EasySpray/Thermo). Mass spectroscopy (MS) analysis was performed exactly as described previously (29). The resulting MS raw files of all biological replicates were submitted to the MaxQuant software for protein identification and phosphorylation site mapping using the Andromeda search engine. The UniProt human database (September 2018) supplemented with mutant Lck sequences was used for database searches. Carbamidomethyl was set as a fixed modification, and protein N-acetylation, methionine oxidation, and phospho-STY were set as variable modifications. A first search peptide tolerance of 20 ppm and a main search error 4.5 ppm were used. Trypsin without proline restriction enzyme option was used, with two allowed miscleavages. The minimal unique + razor peptides number was set to 1, and the allowed false discovery rate was 0.01 (1%) for peptide and protein identification. Label-free quantitation was used with default settings. Both razor and unique peptides, except phospho-STY–modified peptides, were considered for protein group quantification. Lck pY-containing peptides with localization probability lower than 0.75 were filtered out. The MS data are available in the PRIDE database (30) under accession number PXD014470 (https://www.ebi.ac.uk/pride/archive/projects/PXD014470).

GraphPad Prism version 7.04 was used for statistical analysis. Statistical significance of the results was assessed with unpaired, two-tailed t tests. The p values for kinetics experiments were calculated using two-way ANOVA. All statistically significant comparisons are shown on the graphs accordingly (*p < 0.05, **p < 0.01, and ***p < 0.001).

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, A.S. ([email protected]).

Lck Tyr¹⁹² is located in the EF loop of the SH2 domain, and it may thus affect the domain’s ability to bind pTyr ligands (11, 22). We previously suggested that Lck pTyr¹⁹² may fulfill different functions than Lck Tyr¹⁹² (11). In this study, to examine the role of Lck Tyr¹⁹² in intact cells, we used CRISPR/Cas9 genome editing to generate T cell lines harboring either Phe¹⁹² (Lck Y192F) or Glu¹⁹² (Lck Y192E) mutations (Fig. 1A) (27). In addition, Lck KO mutants were obtained from the procedure. While Phe mimics unphosphorylated Tyr, Glu is a phosphomimetic, which resembles pTyr both in size and charge. Lck Phe¹⁹² thus represents stably nonphosphorylated Lck Tyr¹⁹², while Lck Glu¹⁹² represents stable Lck pTyr¹⁹². Neither substitution alters binding of the Lck SH2 domain to Lck pTyr⁵⁰⁵ (11, 13).

We first examined how Lck Tyr¹⁹² affects T cell activation by monitoring expression of the activation marker CD69 in Lck mutants. To this end, cells were stimulated for 17 h, either with anti-TCR Ab or with PMA/IO. The latter stimulation directly activates protein kinase C (PMA action) and induces Ca²⁺ release from the endoplasmic reticulum (IO action) (31) and thus bypasses Lck in the TCR signaling pathway. CD69 surface expression was analyzed by flow cytometry (Fig. 1B). While TCR stimulation of Lck WT and Lck Y192F cells induced expression of similar amounts of CD69, it did not induce CD69 expression in Lck KO or Lck Y192E cells. In contrast, PMA/IO stimulation induced similar amounts of CD69 in all cell lines (Fig. 1C).

Similarly to CD69, secretion of IL-2 is a commonly measured outcome of T cell activation. However, we were unable to determine whether Tyr¹⁹² affects IL-2 expression since, in comparison with the original Jurkat cell line JE6.1 (Supplemental Fig. 1A), the JTAg subline of Jurkat cells used in this study lost its ability of potent IL-2 production (Supplemental Fig. 1B).

Having found that the Lck Y192E mutant was unable to initiate expression of CD69, we next used imaging flow cytometry to examine whether NFAT translocation was affected in these cells. NFAT translocates to the cell nucleus upon dephosphorylation by activated calcineurin (26) and regulates many activation-induced genes in T cells, including CD69. The cells were incubated with anti-TCR Ab for 3 h, followed by staining for NFAT and the nucleus (DAPI) (Fig. 1D). As a measure of nuclear translocation, the similarity between NFAT and DAPI stain was calculated for each cell analyzed (Fig. 1E). In all cell lines tested, except Lck KO, more NFAT translocated to the nucleus in TCR stimulated cells compared with untreated cells (Fig. 1F). Taken together, although Lck Y192E cells were able to initiate NFAT nuclear translocation upon TCR stimulation, this was not sufficient to promote CD69 expression.

Since mutation of Tyr¹⁹² can alter Lck SH2 domain binding specificity, we investigated whether the mutated Lck displayed altered interactions with proximal signaling molecules after TCR triggering. To this end, we immunoprecipitated Lck from Lck Y192 mutated cell lines stimulated with anti-TCR Abs. The samples were immunoblotted for a known Lck SH2 domain binding partner (ZAP70) (9) and for suspected partners (LAT, Plcγ, VAV1) (11, 32–34) (Fig. 1G, 1H). There was no statistically significant difference in binding ability between Lck Tyr¹⁹² and Lck Phe¹⁹² molecules, whereas Lck Glu¹⁹² revealed significantly less coprecipitated targets compared with Lck WT. The only exception was VAV1, which bound equally well to all Lck variants. These results suggest that Lck pTyr¹⁹² influences the interaction of Lck with critical molecules in the TCR signaling pathway.

To further explore how Lck pTyr¹⁹² affects early TCR signaling, Lck Y192-mutated Jurkat cells were stimulated with anti-TCR Abs for 2 min, followed by IP of phosphorylated proteins (pTyr IP). The samples were immunoblotted for pTyr (Fig. 1I) as well as for TCR signaling cascade proteins (Fig. 1J). TCR-stimulated Lck Y192F cells displayed increased amounts of pTyr proteins, whereas Lck Y192E cells displayed a strikingly hyperphosphorylated 50-kDa protein (Fig. 1I, 1K), representing Lck itself (Fig. 1J, 1K). Quantification of specific phospho-proteins detected by immunoblotting (Fig. 1K) showed that compared with Lck WT, Lck Y192F cells displayed an increased amount of phospho-ZAP70 and -CD3ζ, whereas Lck Y192E cells displayed a significantly lower amount of phospho-ZAP70 and -Plcγ. Taken together, this indicates that Lck Glu¹⁹² is unable to sustain sufficient TCR signaling for induction of the activated T cell phenotype.

It was recently reported that Lck-deficient cells reconstituted with Lck Glu¹⁹² displayed hyperphosphorylation of Lck Tyr⁵⁰⁵ (13, 35). This could explain the results observed in (Fig. 1J. We therefore performed MS analysis of Lck molecules immunoprecipitated from TCR-stimulated Lck Y192 mutated cells to directly examine phosphorylation of Lck. MS analysis confirmed that endogenously expressed Lck Glu¹⁹² is hyperphosphorylated on Lck Tyr⁵⁰⁵, but not Lck Tyr³⁹⁴ (Fig. 1L). Approximately 25% of Lck Glu¹⁹² molecules detected by MS contained pTyr⁵⁰⁵, compared with 5% of Lck Tyr¹⁹² and Lck Phe¹⁹² (Fig. 1M), whereas Lck pTyr³⁹⁴ was similar in all three cell lines (Fig. 1M).

To confirm these results, we immunoblotted lysates from the same cells with Abs recognizing pTyr³⁹⁴ and pTyr⁵⁰⁵ (Fig. 1N). Independent of TCR stimulation, Lck WT and Lck Y192E cells displayed similar amounts of Lck pTyr³⁹⁴ (Fig. 1O), whereas Lck Y192E cells displayed a 5-fold higher amount of Lck pTyr⁵⁰⁵, both in resting and activated cells (Fig. 1P). In contrast, both resting and TCR-stimulated Lck Y192F mutant cells displayed a 2-fold increase of Lck pTyr³⁹⁴ compared with Lck WT cells (Fig. 1O). These results show that Lck Tyr¹⁹² influences phosphorylation of Lck both in resting and in TCR-stimulated T cells.

Physiological activation of T cells via the TCR is supported by costimulation with the CD28 coreceptor. However, in contrast to the original Jurkat cell line JE6.1, the JTAg subline used in this study expresses minimal amounts of CD28 on its surface (Supplemental Fig. 1C). Nonetheless, despite expressing CD28 on the surface, costimulation of both CD3 and CD28 receptors in JE6.1 does not influence IL-2 and CD69 expression, nor short-term phosphorylation, in comparison with TCR stimulation alone (Supplemental Fig. 1A, 1D, 1E).

A main aim of this study was to determine whether Lck Glu¹⁹² associates with novel binding partners. The closed conformation of Lck Glu¹⁹² results in reduced activity of Lck and consequently reduced availability of SH2 domain ligands in Lck Y192E cells. This confounds our analysis of Lck Glu¹⁹² binding ability (Fig. 1G, 1H). Thus, to eliminate the influence of Lck pTyr⁵⁰⁵, we used CRISPR/Cas9 to generate cells expressing Lck Phe⁵⁰⁵ (Lck Y505F), as well as double-mutated Lck (Lck Y192F/Y505F and Lck Y192E/Y505F) (Fig. 2A).

It is well established that Lck Phe⁵⁰⁵ is hyperactive compared with Lck Tyr⁵⁰⁵ (36). This phenotype can be observed after short-term TCR stimulation of the mutated cells and is thought to be due to constantly open conformation of Lck. We thus subjected the Lck Y505F doubly mutated cell lines to analysis of phosphoproteins in resting and TCR-stimulated cells (Fig. 2B, 2C). In cells where Lck Phe⁵⁰⁵ also carried Lck Phe¹⁹² or Lck Glu¹⁹² mutations, TCR-induced protein phosphorylation was even more pronounced than that observed in Lck Y505F cells (Fig. 2D). In analogically prepared Lck IPs, Lck Phe¹⁹²/Phe⁵⁰⁵ and Lck Glu¹⁹²/Phe⁵⁰⁵ bound to proximal signaling molecules to the same extent (Fig. 2E, 2F).

To confirm that the combination of Lck Glu¹⁹² and Phe⁵⁰⁵ rescues the inhibited Lck Glu¹⁹² phenotype, Lck double mutants were activated for 17 h, with anti-TCR Ab (in the absence or presence of anti-CD28) or with PMA/IO. The activated Lck Y505F mutants displayed similar levels of CD69 (Supplemental Fig. 2A) as well as IL-2 production (Supplemental Fig. 2B). Taken together, combined mutations of Lck Tyr¹⁹² and Tyr⁵⁰⁵ resulted in hyperactive Lck, where Lck Phe¹⁹²/Phe⁵⁰⁵ and Lck Glu¹⁹²/Phe⁵⁰⁵ displayed a similar ability to interact with known Lck-binding partners.

We next analyzed the phosphorylation status of double-mutated Lck. MS analysis of Lck IP from Lck Y505F, Lck Y192F/Y505F, and Lck Y192E/Y505F cell lines that had been stimulated with anti-TCR Ab showed that all three mutants displayed high amounts of Lck pTyr³⁹⁴ (Fig. 2G). Approximately 60% of Lck Phe⁵⁰⁵ molecules captured by MS contained pTyr³⁹⁴, compared with 80% of the Lck Phe¹⁹²/Phe⁵⁰⁵ and Lck Glu¹⁹²/Phe⁵⁰⁵ molecules (Fig. 2H). Lck Glu¹⁹²/Phe⁵⁰⁵ also showed an elevated amount of Lck pSer¹⁹⁴ and pSer²¹³ (Fig. 2G). Both residues are located in the Lck SH2 domain (data not shown). The elevated amount of Lck pTyr³⁹⁴ observed in MS was confirmed by immunoblotting lysates of the corresponding mutated cell lines (Fig. 2I). Compared to Lck Y505F, both Lck Y192F/Y505F and Lck Y192E/Y505F cells harbored elevated amounts of Lck phosphorylated on Tyr³⁹⁴ in the absence of stimulation, and this was further increased upon TCR stimulation (Fig. 2J). TCR induced Lck pTyr³⁹⁴ was significantly more pronounced in Lck Y192E/Y505F cells compared with Lck Y192F/Y505F cells.

The observation that the elevated amount of pTyr³⁹⁴ in the Lck Y192E/Y505F mutant cells could be further increased upon TCR stimulation showed that only a fraction of Lck molecules carried pTyr³⁹⁴ in the steady state. Lck Glu¹⁹² has been found to be resistant to CD45 dephosphorylation (13). However, other phosphatases may still dephosphorylate Lck. We thus investigated the kinetics of Lck pTyr³⁹⁴ upon TCR stimulation in double-mutated Lck cell lines. We first subjected the cells to long-term TCR stimulation followed by 24 h of rest. In our experience, protein phosphorylation, including Lck pTyr³⁹⁴, is increased in such preactivated cells (data not shown), making kinetic analysis of dephosphorylation after TCR stimulation more robust. These preactivated cells were restimulated with anti-TCR Ab for 1–60 min. Samples were subjected to immunoblotting with Lck pTyr³⁹⁴ Ab (Fig. 2K). Of the three mutants, Lck Glu¹⁹²/Phe⁵⁰⁵ displayed the strongest Tyr³⁹⁴ signal (Fig. 2L). However, 20 min after the start of TCR triggering, the Lck pTyr³⁹⁴ level of Lck Glu¹⁹²/Phe⁵⁰⁵ dropped to initial values, and after 60 min the signal disappeared completely. The same kinetics was observed for the two other Lck mutants (Fig. 2M). A similar kinetics of phosphorylation was observed for Plcγ pTyr⁷⁸³, which is downstream of TCR triggering (Supplemental Fig. 2C–E). Taken together, these findings strongly indicate that Lck pTyr¹⁹² does not affect TCR-triggered dephosphorylation of Lck.

Considering that Lck Glu¹⁹² is known to affect Lck SH2 domain specificity (11, 22), the Lck Y192F/Y505F and Lck Y192E/Y505F cell lines had surprisingly similar phenotypes (Fig. 2E, 2F). However, it is not well established whether Phe¹⁹² influences Lck SH2 interactions. To address this, we aimed to isolate the effect of pTyr¹⁹² on the SH2 domain’s specificity from the effect on Lck’s activity (Lck Tyr³⁹⁴ and Tyr⁵⁰⁵ phosphorylation). To this end, we took advantage of the HEK 293T cell line. These embryonic kidney cells do not possess proteins involved in the TCR signaling pathway, including the CD45 phosphatase, providing a clean system to investigate protein–protein interactions. We transfected HEK 293T cells with plasmids encoding GFP-tagged Lck (WT, Y192F, Y192E) or GFP alone, together with plasmids encoding either ZAP70 or CD3ζ (i.e., CD8ζ chimeric receptor, where the extracellular part of CD8 is fused to the intracellular part of CD3ζ). Lysed cells were subjected to IP using anti-GFP nanobodies. Samples were immunoblotted and results were quantified. No difference in binding ability between Lck mutants to CD8ζ (Fig. 3A, 3B) or ZAP70 (Fig. 3C, 3D) was observed. As HEK 293T cells do not possess all the regulatory pathways controlling phosphotyrosine signaling of T cells, it is possible that the interactions between Lck and CD3ζ or ZAP70 were not correctly recreated. Both the SH2 domain, as well as other parts of Lck, may contribute to the binding capacity of Lck toward these two targets (for example, the Src homology 3 domain or the kinase domain itself). We therefore also directly assessed the effect of Tyr¹⁹² mutation on the SH2 domain binding ability by performing PD experiments with recombinant Lck SH2 domains.

Prior to PD, Jurkat cells were stimulated with anti-TCR Ab or with PV for physiological and maximal levels of pTyr proteins, respectively. Samples were immunoblotted with pTyr-Ab (Supplemental Fig. 3A) or Abs against specific proteins downstream of TCR signaling (Fig. 3E). Although the amount of PD proteins varied depending on the stimulation (Supplemental Fig. 3B), there were no statistically significant differences between the amounts of proteins binding to Lck-SH2 WT and Y192F. In contrast, Lck-SH2 Y192E displayed significantly less binding to most of the chosen targets (Fig. 3F, 3G, Supplemental Fig. 3C). In summary, these results indicate that Lck containing Phe¹⁹² mutation has a prototypic Lck SH2 domain specificity, which is not altered by phosphorylation at position 192.

Lck pTyr¹⁹² was first associated with alteration of Lck downstream signaling more than 20 years ago (22). It is, however, still unclear whether this effect of pLck¹⁹² is due to altered specificity of the Lck SH2 domain (11, 22) or, as recently discovered, via altered association with CD45 (13). In this study, we present data from mutated Jurkat cell lines that lend support to both mechanisms. Our observation that Tyr¹⁹² influences the amount of Lck pTyr³⁹⁴ and pTyr⁵⁰⁵, both at steady state as well as upon TCR triggering (schematically depicted in (Fig. 4), strongly supports the notion that Tyr¹⁹² is a critical regulator of Lck activation.

We used CRISPR/Cas9-mediated gene editing in the human T cell line Jurkat TAg cells (27) to investigate whether endogenous Lck harboring Glu¹⁹² or Phe¹⁹² mutations displayed functional changes due to altered Lck SH2 domain binding capacity. While Jurkat cells cannot recapitulate all aspects of T cell activation due to mutations in critical signaling pathways (37), demonstrated by low IL-2 expression (Supplemental Fig. 1B) and lack of CD4 (38) and CD28 (Supplemental Fig. 1C), these cells are still highly useful for dissection of proximal signaling events downstream of the TCR. In support of this notion, part of the results reported in this study were recently confirmed in the mouse model (35). With more efficient methods for gene editing of primary human T cells (39), it should be possible to determine whether the hyperactivity of Lck Tyr¹⁹² mutations combined with Phe⁵⁰⁵, as seen in this study, is also present in primary cells with an intact signaling apparatus.

The strong effect of Lck Tyr¹⁹² on the overall activity of Lck (the current study and Refs. 13 and 35) was unexpected and made our aim to better understand the role of Tyr¹⁹² in Lck SH2 domain function particularly challenging. Hyperphosphorylation of Tyr⁵⁰⁵ in Lck Y192E cells reduced the availability of potential Lck SH2 domain ligands for Lck Glu¹⁹². When we sought to overcome that problem by additionally mutating Lck Tyr⁵⁰⁵ and creating combined Lck Y192F/Y505F and Lck Y192E/Y505F mutants, we were surprisingly faced with the opposite challenge. The signaling activity in double mutants increased by several fold, as measured by Lck pTyr³⁹⁴ levels and amount of TCR-induced pTyr proteins, compared with that observed in cells expressing the Lck Phe⁵⁰⁵ mutation only.

Lck Phe¹⁹² displayed an increased amount of pTyr³⁹⁴ both in the presence of Tyr⁵⁰⁵ and Phe⁵⁰⁵, whereas Lck Glu¹⁹² displayed an increased amount of pTyr³⁹⁴ only in the presence of Phe⁵⁰⁵. The striking increase of pTyr⁵⁰⁵ in Lck Glu¹⁹², both in resting and activated T cells, which was not observed in Lck Phe¹⁹² molecules, indicates that the mechanism for increased Lck pTyr³⁹⁴ in the two different Lck Y192 mutants differs.

During the course of this study it was reported that Lck Glu¹⁹² is unable to interact with CD45, which may explain why the amount of Lck pTyr⁵⁰⁵ was constitutively elevated in cells expressing this mutation (13). A similar increased amount of Lck pTyr⁵⁰⁵ is seen in T cells from CD45 KO mice (7). Our data agree with the notion that the closed form of Lck (Lck pTyr⁵⁰⁵) is controlled by CD45 dephosphorylation (40). Furthermore, the increased amount of Lck pTyr³⁹⁴ in resting Lck Y192E/Y505F cells suggests that CD45 is the primary phosphatase mediating Lck dephosphorylation in the steady state. The possibility that other phosphatases (5, 6) have indirect control, for instance over Lck pTyr¹⁹² (41) or CD45 itself (42), can, however, not be excluded.

Our observation that also Lck Phe¹⁹² affected phosphorylation of Lck pTyr³⁹⁴ both in resting and activated T cells is in contrast to Courtney et al. (13) who reported that although Lck Phe¹⁹² interacted less well with CD45, the signaling capacity of this mutant was not affected. The increased TCR-induced phosphotyrosine activity and elevated amount of Lck pTyr³⁹⁴ in Lck Y192F cell lines were particularly apparent in the Lck Phe¹⁹²/Phe⁵⁰⁵ double mutants. Our PD experiments with isolated Lck SH2 domains indicated that the binding ability of Lck-SH2 Y192F is similar to that of Lck-SH2 WT. Because the SH2 domain of Lck Phe¹⁹² cannot be transiently modified by phosphorylation of Tyr¹⁹², our data support the hypothesis that the ability to dynamically change the specificity of the SH2 domain may influence activation of Lck. This is potentially important since clustering of Lck molecules and switching of Lck between open and closed conformation are critical steps of the TCR signaling cascade (43, 44). SH2 domains with specificity that cannot be dynamically changed may promote clustering of Lck by stably locating one Lck molecule in the vicinity of other Lck molecules to allow its transphosphorylation on Lck Tyr³⁹⁴ (Fig. 4A, 4B). Such a phenomenon might possibly cause Lck overactivation, as observed in the Lck Phe¹⁹² phenotype.

At steady state, phosphorylation of Lck Tyr¹⁹² is controlled by ZAP70 (23). Simultaneously, the resistance of Lck pTyr¹⁹² to CD45 suggests that in the steady state, Lck pTyr¹⁹² may preserve the phosphorylation status of other Lck phosphosites as well, including Lck pTyr⁵⁰⁵ and pTyr³⁹⁴ (Fig. 4C). It is thus an intriguing possibility that Lck pTyr¹⁹² serves as a mechanism to maintain a small but stable amount of Lck pTyr³⁹⁴, priming the T cell for rapid initiation of the TCR signaling cascade (44).

Despite the high amount of Lck pTyr³⁹⁴ in unstimulated Lck Y192E/Y505F mutated cells, phosphorylation could be further increased by TCR activation. However, the kinetics with which Lck pTyr³⁹⁴ decreased to below-detectable levels was similar in all three Lck Y505F mutant cells lines. This strongly suggests that the downregulation of Lck pTyr³⁹⁴ subsequent to TCR stimulation is not affected by Lck pTyr¹⁹² and thus most likely does not critically depend on CD45 (Fig. 4D). Instead, SHP1 and PTPN22 phosphatases could be responsible for this dephosphorylation, as both have been reported to selectively regulate Lck pTyr³⁹⁴ (5, 6).

Taken together, our data contribute to an ongoing debate in the signaling field on the amount of Lck pTyr³⁹⁴ in the not activated versus activated T cells. It has been suggested that an abundant pool of preactivated Lck phosphorylated on Tyr³⁹⁴ exists in resting cells, which is not further increased upon TCR stimulation (45). These observations have since been criticized for the methodological approach that neglected the possibility that kinase activity can be sustained in the lysates (46). While additional evidence favors the model where the level of Lck pTyr³⁹⁴ is rising upon TCR stimulation (44), the existence of preactivated Lck pTyr³⁹⁴ prior to TCR triggering has not been refuted.

Furthermore, our results both confirm that Lck pTyr³⁹⁴ increases upon TCR stimulation, and they also show that the amount of preexisting (as well as stimulated) Lck pTyr³⁹⁴ was previously overestimated (45, 46). Lck Phe⁵⁰⁵ with its open conformation was already known to be the most active mutant of Lck (36). However, when the Lck Phe⁵⁰⁵ mutation was combined with additional mutation of Tyr¹⁹², the level of Lck pTyr³⁹⁴ raised several fold, suggesting that even Lck Phe⁵⁰⁵ is activated only to a limited extent in intact T cells.

Lastly, we have noticed that our Lck-mutated Jurkat cell lines could be particularly useful in studies designed to explore Lck-dependent signaling pathways. Hyperactive Lck mutants can be especially useful in TCR signaling studies, as they generate a much more pronounced response, while being stimulated through the physiological pathway (in contrast to unspecific PV treatment). To our knowledge, Lck Glu¹⁹² is the least, whereas Lck Glu¹⁹²/Phe⁵⁰⁵ is the most, active form of Lck, where the kinase domain itself is not directly manipulated by mutation. If the specificity of the Lck SH2 domain is of concern, the Lck Y192F/Y505F mutant can be used, as Lck Phe¹⁹²/Phe⁵⁰⁵ is still at least twice as active as Lck Phe⁵⁰⁵, while maintaining a prototypic Lck SH2 domain specificity.

In summary, our results point to a crucial role for Lck pTyr¹⁹² in regulating Lck activity in the steady state, both via Lck SH2 domain specificity and via Lck interaction with phosphatases. As phosphorylation is a dynamic process, Lck Tyr¹⁹² can serve as a quick molecular switch, maintaining a basal level of Lck phosphorylation. Unphosphorylated Tyr¹⁹² promotes Lck clustering (43) and transphosphorylation of Lck pTyr³⁹⁴, whereas phosphorylated Tyr¹⁹² preserves the current phosphorylation status of Lck. Therefore, Lck Tyr¹⁹² can be an overriding regulatory site that governs both Tyr³⁹⁴ and Tyr⁵⁰⁵.

The authors have no financial conflicts of interest.