Identification of substrates of LYP in T cells

Zhang and coworkers found that the best peptide sequence dephosphorylated by LYP was YGEEpYDDLY using inverse Ala scanning (8). To find LYP substrates, we searched the PhosphositePlus database (17) with a degenerate version of Zhang's peptide in which the phospho-Tyr (p-Tyr) is surrounded by acidic amino acids, either D or E, in positions −2 to + 2. Based on the hits obtained, we chose for further investigation proteins expressed in T cells that contain phosphopeptides flanked by at least two acidic residues that are (Fig. S1A).

Prior to investigate the interaction of these proteins with LYP, we determined which substrate-trapping version of this phosphatase might be more adequate for these assays [17]. As a result, we developed LYP mutants in key catalytic amino acids: D195A (DA), C227S (CS), and the double mutant D195A/C227S (DACS). We tested the interaction of these mutants with SKAP2, which contains the closest peptide (Tyr75) to the best LYP substrate [8]. The interaction was tested in transiently transfected HEK293 cells using immunoprecipitation (IP). After PV treatment for 5 min to induce SKAP2 tyrosine phosphorylation, LYP was immunoprecipitated, and the presence of SKAP2 in the precipitates was detected by Western blot (Fig. S1B). In line with previous findings, the substrate trapping mutant that showed the highest interaction with SKAP2 was LYP-DACS [7]. Next, we tested the ability of the DACS mutant to bind putative substrates. To avoid protein interactions with other regions of LYP, we transfected Jurkat cells with a construct that only contains the LYP phosphatase (LP) domain with the double mutation (LP-DACS). PV was applied for 5 min to Jurkat cells, and the LP-DACS peptide was immunoprecipitated. Proteins present in the precipitate were detected by Western Blot with a phospho-Tyr (pTyr) specific antibody. This assay showed that LP-DACS was able to bind to Tyr phosphorylated proteins (Fig. S1C). Based on these results, we used LP-DACS substrate trapping mutant to test the interaction of LYP with the potential substrates found in the database search.

Following that, we examined the interaction of the potential substrates with LP-DACS in HEK293 cells transiently transfected with plasmids expressing these proteins. After treatment with PV for 5 min, LP-DACS was immunoprecipitated and the presence of the putative substrates detected by Western Blot. In addition to the proteins found from the database search (Fig. S1A), we tested proteins known to participate in signal transduction pathways in immune cells. Altogether, LP-DACS immunoprecipitated the following proteins: ABL, CBL, DOK-1, DOK-2, FYB, HPK1, HS1, LCK, SKAP1, SKAP2, SLP76, and Vav (Fig. 1A). Among the proteins assayed that did not interact with LP-DACS were: CRKL, WASP, CSK, Grb2, ITK, PSTPIP1, p85 subunit of PI3K, LAT, PKA, and the ζ-chain (data not shown and Fig. S2A). To demonstrate the specificity of these assays, we tested the interaction of the substrate trapping mutant LP-DACS with SKAP2, LCK, and PSTPIP1 in the same experiment (Fig. S2A). These results suggest that the proteins detected in the precipitates are likely LYP substrates and warrant further investigation.

Fig. 1

Interaction of the substrate trapping mutant of LYP phosphatase domain, LP-DACS, with several potential substrates. A HEK293 cells were transiently transfected with 3xmyc, 3xHA or 3xFLAG-LP-DACS along with plasmids that express potential substrates, as indicated on the top of each panel. After PV treatment for 5 min, cells were lysed and LP-DACS was immunoprecipitated with specific antibodies against myc, HA or FLAG epitopes bound to sepharose beads. Proteins were transferred to nitrocellulose membranes after SDS-PAGE, and the presence of potential substrates in the precipitates was detected by Western Blot. B Detection of LYP and ERK2, as loading controls, in Jurkat cells and the LYP-deficient Jurkat-cell line (494). C Jurkat cells deficient in LYP (494) were transiently transfected with plasmids that express 3xFLAG-LP, the active phosphatase domain of LYP, and 3xFLAG-LP-DACS. After PV treatment for 5 min, the phosphatase domain of LYP was immunoprecipitated with FLAG antibody bound to Sepharose beads, and proteins present in the precipitates were identified by Western Blot. HC-IgG heavy chain IgG, IP immunoprecipitates, TL total lysates

To confirm that the interactions detected in HEK293 cells are also found in T cells, we transfected Jurkat cells deficient in the LYP phosphatase (JK 494) with LP-DACS (Fig. 1B). After treatment of Jurkat cells with PV for 5 min, the LP-DACS peptide was immunoprecipitated and the proteins present in the precipitates were detected by Western blot. This way, we found FYB, HPK-1, Vav, SLP76, LCK, HS1, Pyk2, and Syk in LP-DACS precipitates (Fig. 1C). Although these data do not demonstrate a direct association between LYP and these proteins, they indicate that LYP is linked to a number of proteins involved in TCR signaling that share phosphopeptides with similar sequences [18].

Dephosphorylation of putative substrates by LYP

Then, we performed in vitro phosphatase assays with full-length LYP to see whether the proteins that interacted with LP-DACS were dephosphorylated by this phosphatase. LYP was purified from a stable cell line established in HEK293 cells by transfection of 3xFLAG-LYP. Substrates were transiently transfected in HEK293, phosphorylated by co-expression with Tyr kinases or PV treatment for 5 min, immunoprecipitated and incubated in phosphatase buffer for 0, 5, 15 and 30 min with either LYP-R620 (LYP-R) or LYP-W620 (LYP-W), the polymorphic variant associated with autoimmune diseases. Dephosphorylation was detected by Western blot with P-Tyr antibody (4G10) (Fig. 2). The fastest dephosphorylation was detected in SKAP2 and SKAP1, followed by FYB and SLP76. HS1 and Vav were also dephosphorylated, although at a slower rate. On the other hand, we did not detect changes in Tyr phosphorylation in HPK-1, LCK, PSTPIP1, and WASP proteins (Fig. S2B).

Fig. 2

In vitro dephosphorylation assays with LYP-R620 and LYP-W620. Several proteins were tested for dephosphorylation by either LYP-R or LYP-W at the time points indicated on the top of the panel. Phosphatase assays were carried out with a full-length version of LYP and its putative substrates as indicated in Material and Methods. Dephosphorylation was detected by Western blot with p-Tyr (4G10) antibody

Given that some proteins dephosphorylated by LYP contain SH3 domains (SKAP2, SKAP1, HS1, Vav and FYB), and that LYP presents several PRMs, we wonder whether LYP could bind to these proteins through their SH3 domain. First, we compared in the same assay the interaction of LP-DACS with these proteins by immunoprecipitation after PV treatment for 5 min (Fig. S3A). In this assay, the proteins with the highest binding were SKAP2 and LCK, while the proteins with the lowest binding were Vav and HS1. As before, PSTPIP1 did not bind to LP-DACS, in keeping with the data obtained in the phosphatase assay. The interaction of these proteins goes in parallel to the phosphatase activity shown previously (Fig. 2). Then, we tested the interaction of active LYP with these proteins to explore whether they could interact through additional regions other than the PTP domain of LYP (Fig. S3B). SKAP1 and HS1 clearly bound to LYP, in addition to CSK and PSTPIP1, two proteins known to bind to LYP PRMs [4, 19], while SKAP2 showed a weaker binding. However, we did not detect in this assay the binding of Vav, SLP76 and FYB, suggesting that the interaction of LYP with these proteins only occurs through the phosphatase domain, and, therefore, LYP and these substrates are put together by adaptor proteins.

Sites targeted by LYP in putative substrates

Afterwards, to demonstrate the Tyr targeted by LYP in the proteins here studied, we tested the interaction of LP-DACS with Tyr to Phe mutants in the residues identified in the initial database search. As before, immunoprecipitation after PV treatment for 5 min was used to examine these interactions in transiently transfected HEK293 cells. First, we looked at SKAP2, which contains Y75 (DAEDGEEyDDPFAGP), i.e., the protein sequence most similar to the peptide best dephosphorylated by LYP [8]. In addition to Tyr75, we mutated to Phe Y237 (YDERGELyDDVDHPL) and Y261 (QPIDDEIyEELPEEE), both of which are embedded in an acidic environment. LP-DACS immunoprecipitation of these mutants revealed that Y237F had the lowest binding to LP-DACS, followed by Y261F, and Y75F (Fig. 3A, B). To confirm the Y75 result, which was unexpected, we created double mutants of the three Tyr studied in SKAP2 to test the interaction of LP-DACS with any single Tyr of the three. As previously stated, Y237 had the highest binding, followed by Y261, and finally Y75 (Fig. S4A). Next, we focused on SKAP1, a close protein to SKAP2 that presents Y232 and Y271, which align with SKAP2 Y237 and Y261, but lacks a Tyr equivalent to SKAP2 Y75. Mutation Y232F or Y271F in SKAP1 reduced the interaction with LP-DACS, which was mostly abolished when both Tyr were mutated to Phe (Fig. 3C). We followed our analysis with HS1 Tyr 378 and 397 (Fig. 3D). In HS1, both Tyr are required for the interaction with LP-DACS. In the case of Vav, mutants of several Tyr known to be phosphorylated were used in this assay [20]; however, only Y160 and Y174 were critical for the interaction with LP-DACS (Fig. 3G).

Fig. 3

Identification of the Tyr targeted by LYP in several potential substrates. A HEK293 cells were transiently transfected with 3xHA-LP-DACS along with expression plasmids of SKAP2 wild type and mutated to Phe in the Tyr indicated. After PV treatment for 5 min, cells were lysed, and LP-DACS was immunoprecipitated with HA Ab bound to Sepharose beads. Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. The presence of SKAP2 in the precipitates was detected by WB, as indicated. B Graph representing the mean data obtained by densitometry in at least three experiments (n = 4 for single Y to F mutant and n = 3 for 3YF) like the one shown in (A). Densitometry values of the myc blot on the top, expressed as percentage of the WT are indicated below. C As before, SKAP1 wild type and the mutants in Tyr232 and Tyr271 were assayed for binding to LP-DACS. D Interaction of HS1 YF mutants with 3xHA-LP-DACS, as before. E Interaction of Vav YF, as indicated, with 3xHA-LP-DACS. F Graph representing the mean data obtained by densitometry in three experiments like the one shown in (E). Densitometry values of the myc blot on the top, expressed as percentage of the WT are indicated below. G Interaction of FYB wild type or mutated in five of the six tyrosines tested, as indicated, with 3xHA-LP-DACS. H As before, interaction of SLP76 YF full length mutants with 3xHA-LP-DACS. NS nonsignificant, *P < 0.05,**P < 0.01 and ***P < 0.001 for comparison of proteins mutated in Tyr with the wild type protein

Following that, FYB was examined. This protein contains several phosphorylated Tyr [21]. We chose six of them (Y462, Y571, Y595, Y625, Y651 and Y771) based on their similarity to the peptide best dephosphorylated by LYP [8]. Initially, we tested the interaction of LP-DACS with two fragments of FYB, N-terminal FYB (1–450), and C-terminal FYB (401–783). This assay showed that LP-DACS interacts with the C terminal half of FYB (Fig. S4B) that includes the aforementioned 6 Tyr. To further demonstrate this point, we mutated to Phe the 6 Tyr, and tested its interaction with LP-DACS. Mutation of the 6 Tyr to Phe blocked the interaction between LP-DACS and FYB (Fig. S4C). However, when we mutated one by one any of these Tyr, the interaction did not change (Fig. S4D). Then, to see whether there could be differences in the interaction of each Tyr with LP-DACS, we generated mutants that contained only one of the 6 Tyr studied in FYB. Assays with these mutants showed that the least critical Tyr for the association is Y571 followed by Y771 and Y625, while the other 3 Tyr (595, 651 and 771), which have been shown to bind to the SH2 domain of SLP76 [22], presented a similar binding to LP-DACS (Fig. 3E, F).

SLP76 was the next protein we studied. In this case we mutated the three Tyr in its N-terminal acidic domain, Y113, Y128 and Y145, which are known to be crucial for its function in T cells through the interaction with NCK, Vav, and ITK [23]. A triple mutant of these Tyr (3YF) abrogated the interaction of SLP76 with LP-DACS (Fig. 3H). The mutation that produced the higher reduction in the interaction of SLP76 with LP-DACS was Y128F, followed by Y113F and Y145F. In line with this, the Tyr that generated the higher interaction was Tyr128 followed by Tyr113 and Tyr145 (Fig. S4E). These findings suggest that LYP targets pTyr in several proteins involved in TCR signaling that share a similar acidic amino acid sequence.

LYP regulates TCR signalling downstream of LCK and ZAP70

To determine whether LYP regulates substrates downstream of LCK and ZAP70, like the proteins studied in this work, we examined T cell activation in cells overexpressing Vav, which is recruited by SLP76 [23]. Jurkat cells were transfected with a plasmid expressing Vav as well as a reporter plasmid expressing luciferase driven by the IL-2 promoter [4]. Indeed, LYP expression blocked the activation caused by Vav (Fig. 4A). Then, we tested whether LYP could prevent the activation induced by racQL, a dominant active mutant of the small GTPase rac that is activated by Vav in T cells [24]. In this case, LYP did not prevent luciferase induction (Fig. 4B). Next, we tested whether LYP could affect T-cell activation via SLP76. Transfection of SLP76 increased luciferase induction in response to CD3 plus CD28 stimulation, whereas LYP transfection reduced IL-2 activation induced by SLP76 in stimulated JK cells (Fig. 4C). When these plasmids were transfected in LYP-deficient Jurkat cells (494), SLP76 induced a higher expression of luciferase in comparison with wild type cells. Co-expression of LYP in these cells reduced luciferase induction to the control levels, indicating that the increase in T cell activation in these cells is due to the absence of LYP. These assays collectively revealed that LYP also regulates proteins downstream of LCK and Zap70.

Fig. 4

Regulation of TCR signaling by LYP downstream of ZAP70. A Activation of a luciferase reporter gene driven by the IL-2 minimal promoter in Jurkat cells transfected with Vav and stimulated with CD3 plus CD28 antibodies for 6 h, as indicated. The insert shows the IB of the Vav and LYP proteins expressed. *P < 0.05 and **P < 0.01 for comparison of cells transfected with different plasmids and cells transfected with empty vector (pEF). B Activation of a luciferase reporter gene driven by the IL-2 minimal promoter in Jurkat cells transfected with rac-QL, a dominant active mutant of rac, and LYP and and stimulated with CD3 plus CD28 antibodies for 6 h. The insert shows the IB of LYP. *P < 0.05 and **P < 0.01 for comparison of cells transfected with different plasmids and cells transfected with empty vector (pEF) C Activation of a luciferase reporter gene driven by the IL-2 minimal promoter in wild type and 494 Jurkat cells transfected with LYP and SLP76 as indicated. The cells were left untreated or stimulated with CD3/CD28 antibodies for for 6 h. The insert shows the IB of the SLP76 and LYP proteins expressed. *P < 0.05 **P < 0.01 and ***P < 0.001 for comparison of cells transfected with different plasmids and cells transfected with empty vector (pEF). D Wild type and 494 JK cells were stimulated with CD3/CD28 antibodies during the indicated periods of time. Phosphorylation of Y128 of SLP76 in each condition was measured in cell lysates by IB. Similarly, phosphorylation of LCK in Y394 and Y493 are shown. E, Wild type Jurkat cells and cells deficient in LYP (JK 494) transfected with LYP were stimulated with CD3/CD28 antibodies for 15 min and the phosphorylation of Y128 of SLP76, and Y191 of LAT were detected in each condition by IB

Then, we wanted to test whether LYP deficiency increased SLP76 phosphorylation in cells, using the LYP-deficient Jurkat cell line (494). For this purpose, celIs were stimulated with CD3 and CD28 Abs, and the phosphorylation of Y128 was analysed by Western blot. In these experiments, LYP deficiency increased Y128 phosphorylation in SLP76 (Fig. 4D). In line with previous findings, we found increased phosphorylation in Tyr493 of ZAP70 [25]. Nevertheless, in the case of LCK Tyr394, we found an increase of LCK phosphorylation in resting cells, though the effect upon CD3 stimulation is less clear. We reintroduced LYP phosphatase into 494 cells to confirm that the increase in phosphorylation is LYP specific. In these cells, LYP expression reduced phosphorylation of this protein close to Jurkat parental cells (Fig. 4E). However, when we detected the phosphorylation of Y191 in LAT, we did not see an increase in its phosphorylation in LYP deficient cells. These findings support that LYP targets SLP76 in T cells.

To investigate how LYP can regulate the interaction of SLP76 with its ligands NCK, vav, and ITK via the C-terminal tyrosines 113, 128 and 145, we transfected HEK293 cells and tested by immunoprecipitation if LP-DACS could block this interaction (Fig. S5A). In this assay, we found that LP-DACS inhibited the interaction of NCK, which is primarily mediated by Tyr128, but did not affect the interaction with vav, mediated by Tyr113, and diminished the association with Itk via Tyr 145 [23], which is consistent with previous data that showed that Tyr128 is the tyrosine mainly targeted by LYP in SLP76 (Fig. 3F and Fig. S4E). Furthermore, in a different assay, expression of LYP prevented NCK-SLP76 interaction (Fig. S5B). We further noticed that LYP expression could inhibit the interaction between SLP76 and FYB, but this effect was not seen when an inactive version of LYP (3xFLAG-R or W-DACS) was used in this assay (Fig. S5C). In summary, these assays show that LYP alters the interaction of SLP76 and FYB with other proteins involved in T cell activation.

LYP regulates the formation of microclusters in T cells

SLP76 is found in microclusters, which are protein complexes organized in T cells to transduce signals upon TCR engagement [22]. Because LYP is a potential negative regulator of SLP76, we studied the association of LYP with SLP76 in these complexes. To this end, we activated Jurkat T cells by placing them on CD3 antibody coated coverslips for different periods of time. Cells were then processed for immunofluorescence to detect the localization of LYP and SLP76. Microclusters began to form in these assays after 3 min of stimulation and were present in all cells after 5 min (data not shown). We found that LYP colocalized with SLP76 in microclusters (Fig. 5A). Then, we investigated whether LYP localization occurred in PBL microclusters and confirmed the results obtained in Jurkat cells (Fig. 5B). These findings suggest that LYP and SLP76 co-localize in microclusters. Next, we wondered whether LYP could regulate the assembly of these complexes upon T cell activation. We use the 494 JK cell line to investigate how LYP deficiency affects microcluster formation. When we stimulated these cells with CD3 antibody coated coverslips, we found that microclusters were present even before proper activation by CD3 antibodies and there was no increase in microcluster formation at 5 min (Fig. 5C). In comparison to 494 cells, there was no aggregation of SLP76 in microclusters in resting JK cells, which was not detected until later (5 min) (Fig. 5C). Then, we explored the role of LYP in microcluster formation in PBL. LYP levels were low in resting PBLs but significantly increased after T-cell activation (Fig. 5D). Thus, in resting PBLs, with low amounts of LYP, there were more microclusters than in PBLs stimulated with PHA for 72 h, which expressed higher amounts of LYP (Fig. 5D). Overall, these findings suggest that LYP associates with SLP76 in T cell microclusters, and regulates their formation, most likely by targeting key Tyr involved in protein interaction.

Fig. 5

LYP co-localizes with SLP76 in T cell microclusters. Jurkat cells (A) or PBL (B) were plated on coverslips covered with antibody for CD3 for the indicated periods of time. Then, cells were fixed and stained with specific antibodies for SLP76 and LYP, as indicated. Images were taken with a confocal microscope and representative images are shown. C, Wild type and 494 Jurkat cell lines were plated on coverslips covered with CD3 antibody for the indicated periods of time. Then, cells were fixed and stained with specific antibodies for SLP76. D, PBL cells were left untreated or stimulated with PHA for 72 h to induce the expression of LYP. As before, they were plated on stimulatory coverslips and processed to detect LYP and SLP76 by immunofluorescence with a confocal microscope. E, LYP expression in PLB treated with PHA. Scale bar represents 5 μm