Guanylate Binding Protein 1–Mediated Interaction of T Cell Antigen Receptor Signaling with the Cytoskeleton Free

Guanylate binding protein 1 (GBP-1) is a 67-kDa protein belonging to the family of IFN inducible GTPases, which itself belongs to the dynamin superfamily (1). GBP-1 expression is inducible through the T cell IFN-γ, but also via other inflammatory cytokines like TNF-α, IL-1β, and IL-1α (2, 3). Within the three-dimensional structure of GBP-1, an N-terminal globular domain can be distinguished from a C-terminal helical domain, which mediates an antiproliferative state in endothelial cells (2, 4). In the helical domain, a coiled-coil domain important for protein interactions and typical for structural or motor proteins is located (5). The C terminus contains an isoprenylation site for membrane binding.

Although the tertiary structure of GBP-1 has been solved, there is not much known about the physiologic role of GBP-1 (6, 7). GBP-1 inhibits the invasiveness and tube-forming capability of endothelial cells by inhibiting expression of matrix metalloproteinase-1 (8) and inducing integrin α4 expression (9). In addition, GBP-1 is an important player in cell-autonomous immunity (10). It was shown that loss of function of all GBP family members makes mice more susceptible to bacterial infection (11), most probably because GBPs interfere with autophagosomes. GBP-1 plays also a role in establishing an antiviral state against vesicular stomatitis virus and encephalomyocarditis virus in HeLa cells (12). Moreover, GBP-1 interferes with hepatitis C virus infection through interaction with the viral protein NS5B (13). The only cellular interaction partner of GBP-1 known so far is β-IIITubulin (14). This points to interaction of GBP-1 with structural proteins and suggests that overexpression of GBP-1 in paclitaxel-resistant cells is directly associated to the therapeutic failure of this antimitotic drug in certain tumors (15).

Although CD4⁺ and CD8⁺ T lymphocytes exhibit the highest relative GBP-1 expression in the human body (16), nothing is known about the function of GBP-1 in these cells. In contrast with GBP-1, there is already data available concerning p47 large GTPases and lymphocytes. In particular, Lrg-47 (also known as Ifi1, Irgm1) was found to be connected to IFN-γ–induced cell death and proliferation deficiencies in murine T cells (17). However, the collection of p47 GTPases in humans is limited to three genes, and none of those is induced by IFN-γ (18). The loss of IFN-γ–inducible p47 GTPases in humans warrants the speculation that their functions might be fulfilled by other proteins (19). Therefore, we asked whether GBP-1 could be one of those substitutes and which role it would play in adaptive immunity in humans.

In this study, we report human GBP-1 as a player in T cell activation interfering with the early stage of TCR signaling. We further discovered new cellular binding partners of GBP-1 and show that GBP-1 exerts its function in TCR signaling through interaction with structural proteins like plastin-2 and spectrin β-chain, brain 1 (βII-spectrin).

GFP-tagged GBP-1 (generated in the Erlangen laboratory) was cloned into the retroviral vector pBMN-Z (provided by G. Nolan, Stanford University School of Medicine, Stanford, CA). P-tagged GBP-1 was cloned into the retroviral vector pBMN-IRES-GFP. Empty pBMN-IRES-GFP vector was used as vector control. The short hairpin RNA (shRNA) expression vector pLKOpuro1 (provided by S. Stewart, Washington University School of Medicine, St. Louis, MO) was used to express control and GBP-1–specific shRNA. For the expression of plastin-2 shRNA, the retroviral shRNA expression vector pSM2c (Open Biosystems, Huntsville, AL) was used. The following sequences were used for shRNA-mediated knockdown of gene expression: shGBPp1 sense strand at position 1467: 5′-CCGGTTGAAGTGGAACGTGTGAAATTCAAGAGATTTCACACGTTCCACTTCATTTTTG-3′; shGBPp2 sense strand at position 492: 5′-CCGGTGGTTGAGGATTCAGCTGACTTCAAGAGAGTCAGCTGAATCCTCAACCTTTTTG-3′; shPlastinp1 sense strand at position 1714: 5′- TGCTGTTGACAGTGAGCGCGGCCTTGATTTGGCAGCTAATTAGTGAAGCCACAGATGTAATTAGCTGCCAAATCAAGGCCATGCCTACTGCCTCGGA-3′; shPlastinp2 sense strand at position 2260: 5′- TGCTGTTGACAGTGAGCGACCAAGTAGCCTCTCCTGTATTTAGTGAAGCCACAGATGTAAATACAGGAGAGGCTACTTGGCTGCCTACTGCCTCGGA-3′. Gene-specific shRNA target sequences are indicated in italic. For silencing of βII-spectrin, we purchased two shRNAs from Sigma-Aldrich with the TRC numbers TRCN0000116822 (shSpectrin1) and TRCN0000296592 (shSpectrin2), respectively. As sh control (shCtrl), we used the MISSION Non-Target pLKOpuro1 control vector (Sigma-Aldrich).

The helper plasmids used for lentiviral expression vectors were psPAX2 and pMD2.G (constructed by D. Trono, Ecole Polytechnique Federal de Lausanne; obtained via Addgene, Cambridge, MA) and for retroviral expression vectors pGagPol (Addgene) and pMD2.G. The following primers were used for quantitative RT-PCR: GBP-1 forward: 5′-GGTCCAGTTGCTGAAAGAGC-3′; GBP-1 reverse: 5′-CTTGGTTAGGGGTGACAGGA-3′. For normalization, the following primers of eukaryotic translation elongation factor 1A (eEF1-a) were used: EEF1-a forward: 5′-GTGCTAACATGCCTTGGTTC-3′; EEF1-a reverse: 5′-AGAACACCAGTCTCCACTCG-3′. Rat anti–GBP-1 mAb (clone 1B1) was generated in the Erlangen laboratory. Rabbit anti-actin Ab was obtained from Sigma-Aldrich; CD3 mAb MEM-57, anti–p-tag mAb H902, CD59 mAb MEM-43, CD43 mAb MEM-59, and CD147 mAb MEM-M6/1 were kind gifts from V. Horejsi (Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Prague, Czech Republic). CD147 mAb MEM-M6/1 was directly conjugated with AF647 succinimidyl ester (Life Technologies, Carlsbad, CA) in our laboratory. Rabbit anti-histone H3.1 Ab was purchased from Signalway Ab (College Park, MD). Rabbit anti-Zap70 (99F2) mAb, rabbit anti–phospho-Y416-Src Ab, rabbit anti–phospho-Y505-Lck Ab, rabbit anti–phospho–Y783–PLC-γ Ab, rabbit anti–phospho-T202/Y204-p44/p42 (D13.14.4E) mAb, rabbit anti–α/β-tubulin Ab, the calnexin-specific mAb (C5C9), and rabbit anti–plastin-2 Ab were purchased from Cell Signaling Technologies (Beverly, MA). Anti-TCR β-chain mAb C305 was kindly provided by A. Weiss (University of California, San Francisco, CA). The mAb OKT3 to CD3 was obtained from Ortho Pharmaceuticals (Raritan, NJ). mAb Leu²⁸ to CD28, phospho-Y319-Zap-70 pAb, and phospho-Y142-CD3ζ (K25-407.69) mAbs were purchased from BD Biosciences (Franklin Lakes, NJ). Anti–plastin-2 mAb LPL4A.1 and the rabbit pAb against βII-spectrin were purchased from Abcam (Cambridge, MA). Allophycocyanin-conjugated CD4 mAb MEM-241 and Pacific orange–conjugated CD45 mAb FN50 were from EXBIO (Prague, Czech Republic). Pacific blue–conjugated mouse anti-human CD69 mAb (FN50) was purchased from Biolegend (San Diego, CA). Goat anti-mouse IgG (H + L) Ab coupled to AF555 was purchased from Invitrogen (Grand Island, NY), anti-phosphotyrosine (p-Tyr) mAb 4G10 and rabbit anti-human linker for activation of T cells (LAT) Ab from Millipore (Billerica, MA), rabbit anti–phospho-Y191-LAT Ab from Biosource/Invitrogen (Grand Island, NY), and rabbit anti-human Lck mAb H95 from Santa Cruz (Santa Cruz, CA). Peroxidase-conjugated goat anti-rat IgG (H + L) Ab was bought from Dianova (Hamburg, Germany). Goat anti-rabbit IgG (H + L) and rabbit anti-mouse IgG (whole-molecule) Abs, both HRP conjugated, were from Bio-Rad (Hercules, CA) and Sigma-Aldrich, respectively. Goat anti-mouse IgG + IgM (H + L) Ab conjugated with allophycocyanin and goat anti-mouse IgG + IgM (H + L) F(ab′)₂ were obtained from Jackson Immunoresearch (West Grove, PA); FITC-conjugated sheep anti-mouse IgG + IgM F(ab′)₂ was kindly provided by An der Grub (Kaumberg, Austria).

The T cell line Jurkat E6.1, human PBMCs, B cell line Raji, and HEK293T cells were maintained in RPMI 1640 or DMEM medium supplemented with 100 μg/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine (all from Invitrogen), and 10% heat-inactivated FCS (Sigma-Aldrich). All cells were grown in a humidified atmosphere at 37°C and 5% CO₂, and split every 2–3 d to maintain viability.

Cells (2 × 10⁷ / ml) were lysed for 25 min at 4°C in ice-cold lysis buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, 2 mM EDTA [all from Roth], and complete protease inhibitor tablets [Roche, Basel, Switzerland]) containing 1% detergent Nonidet P-40 (Thermo Scientific, Waltham, MA). The insoluble material was removed by centrifugation at 10,000 × g for 2 min at 4°C, and the supernatant was subjected to standard SDS-PAGE gel electrophoresis (Peqlab, Erlangen, Germany) followed by semidry Western blotting. The blots were developed using HRP substrate (Biozym, Hessisch Oldendorf, Germany) and monitoring light emission in a Fuji LAS-4000 imager (Fuji, Tokyo, Japan). Signals were quantified using ImageJ software.

After stimulation of cells with OKT3 (1 μg/ml) and Leu²⁸ (0.5 μg/ml) or with 10 ng/ml PMA plus 100 ng/ml ionomycin, cDNA was synthesized with the Superscript II first-strand synthesis system of the RT-PCR kit (Invitrogen) using 1 μg TRI-reagent (Sigma) extracted RNA and poly-dT primers supplied with the kit, according to the manufacturer’s protocol. PCR was carried out in a LightCycler instrument with the LightCycler FastStart DNA master SYBR green I kit (all Roche) according to the manufacturer’s protocol.

Membrane, cytosolic/cytoskeletal, and nuclear fractions were isolated from cells as described elsewhere (20) and subjected to Western blot analysis.

Transduction of Jurkat T cells and PBMCs was performed as previously described (21).

Jurkat T cells expressing either GFP-tagged GBP-1 or GFP alone were allowed to adhere on poly-l-lysine–coated adhesion slides for 10 min (Marienfeld, Lauda-Königshofen, Germany). Cells were fixed with 4% paraformaldehyde (Merck, Whitehouse Station, NJ) in PBS for 5 min followed by permeabilization for 10 min with 0.1% Triton X-100 (Thermo Scientific, Waltham, MA). For labeling of the Jurkat T cell membrane AF647-conjugated CD147 mAb, MEM-M6/1 was used. Slides were mounted with Vectashield Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA). Cells were analyzed on a confocal Zeiss LSM 700 microscope (Carl Zeiss Ag, Oberkochen, Germany), equipped with a “Plan-Apochromat” 63×/1.40 Oil DIC M27 objective. Pictures were acquired with the ZEN 2009 software. The distribution of the GFP signal was evaluated using the ImageJ macro EdgeRatio (http://cail.cn/programming.html). The ratio of the mean intensity values of the CD147 staining versus the GFP signal was calculated for the region of the cell exhibiting the highest CD147 signal.

Jurkat T cells stably expressing an IL2-luciferase reporter were stimulated with either 1 μg/ml OKT3/0.5 μg/ml Leu²⁸, or Raji B cells pulsed with 10 ng/ml staphylococcus enterotoxin E (SEE) or 10 ng/ml PMA plus 100 ng/ml ionomycin. Luciferase assays were performed using the luciferase reporter gene assay kit (Roche) according to the manufacturer’s instructions. Relative luminescence was measured either in the Mithras LB 940 multimode plate reader (Berthold Technologies, Bad Wildbach, Germany) or, for Fig. 2B and 2D, in a Lumat LB 9501 device (Berthold Technologies) and normalized for protein content with the Bio-Rad protein assay (Bio-Rad, Hercules, CA).

Samples were analyzed using the Luminex xMAP suspension array technology (Multimetrix, Heidelberg, Germany). Thirty microliters of cell culture supernatants was used. Standard curves were generated using rIL-2 (R&D Systems, Minneapolis, MN).

Cells were seeded in a 96-well flat-bottom plate in 200 μl RPMI 1640 medium supplemented with 1 μCi [³H]thymidine. After stimulation for 18 h with 1 μg/ml OKT3 and 0.5 μg/ml Leu²⁸, the cells were broken up by a freeze/thaw cycle and harvested on a cell harvester (Tomtec, Hamden, CT). Filters were analyzed in a scintillation counter (Wallac).

Jurkat T cells were cultured in RPMI 1640 medium (lacking arginine and lysine; Invitrogen), 10% FCS, and supplemented either with the heavy (R6K4) or light (R0K0) forms of l-arginine and l-lysine. Labeling efficiency was tested by liquid chromatography–tandem mass spectrometry after 10 d, and at that time the general labeling incorporation was already >98% for all identified proteins.

Cells were lysed in 4% SDS (Roth), 0.1 M Tris-HCl (pH 7.6) with protease and phosphatase inhibitors (Roche). Protein contents were estimated by fluorimetric tryptophan measurement. Equal protein amounts from control or GBP-1–silenced cells were pooled, reduced with 0.1 M DTT (Roth), digested with trypsin and lysC, and rebuffered by filter-aided sample preparation as described previously (22). Sequential phosphopeptide isolation with TiO₂ beads (MZ-Analysentechnik, Mainz, Germany) was done as described previously (23).

Cells were lysed in ice-cold lysis buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, and complete protease inhibitor tablets) containing 1% detergent Brij58 (Thermo Scientific). After removing insoluble material, the supernatant was mixed with 50 μl Sepharose beads (GE Healthcare, Little Chalfont, U.K.) coupled to the p-tag specific mAb H902. After an incubation for 1 h at 4°C with constant rotation, the beads were washed three times with lysis buffer and elution was performed using 1.5× Laemmli buffer. For mass spectrometry analysis, elution was performed with 6 M urea, 2 M thiourea (Sigma-Aldrich) in 10 mM HEPES, pH 8.0 (Biomol GmbH, Hamburg, Germany), and the resulting sample was trypsin (Promega, Fitchburg, WI) and LysC (WAKO, Osaka, Japan) digested overnight. For the precipitation of endogenous GBP-1, beads covalently cross-linked with GBP-1 mAb 1B1 by BS3 (Thermo Fisher Scientific) were used.

Cells were rested in RPMI 1640 medium supplemented with 1% FCS for 1 h at 37°C. Jurkat T cells (1.5 × 10⁶ cells/time point) were stimulated for different time points at 37°C using a 1:100 dilution of the hybridoma supernatant of the TCR mAb C305. Peripheral blood T lymphocytes (3 × 10⁶ cells/time point) were stimulated with 10 μg/ml CD3 mAb OKT3 cross-linked with 5 μg/ml goat anti-mouse IgG + IgM (H + L) F(ab′)₂. The reactions were stopped by the addition of ice-cold washing buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, and 5 mM EDTA). After centrifugation (2 min at 850 × g and 4°C), cells were immediately lysed and subjected to immunoblotting as described previously. The lysis buffer was supplemented with 1 mM sodium orthovanadate (Sigma-Aldrich) and 20 mM NaF (Sigma-Aldrich).

Cells were washed with staining buffer [1% BSA (Roth) and 0.02% NaN₃ (Sigma-Aldrich) in PBS], and incubated for 30 min with 4 μg/ml human Ig (CSL Behring, King of Prussia, PA) on ice. Primary Ab (10 μg/ml) was added and the cells were incubated for 30 min on ice. Cells were washed with staining buffer and, if needed, incubated with conjugated secondary Ab (10 μg/ml) for 30 min on ice. After a final wash, cells were analyzed on an LSRII flow cytometer (BD Biosciences).

The peptides were desalted by RPC18 StageTip columns (Agilent Technologies, Santa Clara, CA) and loaded on Spare part Agilent NanoHPLC (Agilent Technologies) filled with reversed-phase ReproSil-Pur C18-AQ 3 μm resin (Dr. Maisch GmbH, Ammerbuch-Entringen, Germany). The sample was eluted with a gradient ranging from 2 to 40% MeCN (Merck) in 0.5% acetic acid (Roth) over 100 min and a flow rate of 250 nL/min. The eluate was analyzed with an LTQ Orbitrap mass spectrometer (Thermo Electron, Bremen, Germany). Results were evaluated with Max Quant software (24) and Andromeda database search engine (25), and label-free quantification algorithm (26) was used to make several independent experiments comparable.

Epoxy-functionalized glass coverslips were provided by CBL GmbH (Linz, Austria). Silicon incubation chambers (Secure-Seal hybridization chambers; Sigma-Aldrich) were stuck onto the slides and filled with 1 μg/ml OKT3 mAb and 0.5 μg/ml Leu-28 mAb in PBS for 1 h. Cells were stained using an AF647-labeled CD59 mAb, and time-lapse imaging was started directly after the cells were seeded on the mAb-coated slides. The imaging was done on a system based on a Zeiss Axiovert 200/M microscope equipped with a Zeiss α-Fluar 100× objective (NA = 1.45) under total internal reflection settings. Scanning was performed in time delay and integration mode. (For further details concerning the setup, see Ref. 27.) Cell spreading of individual cells was assessed using the ImageJ software.

Calcium flux measurements were performed as described previously (28).

If not stated otherwise, unpaired Student t test or two-way ANOVA followed by Bonferroni posttest was performed using GraphPad Prism (GraphPad Software, San Diego, CA).

Basal GBP-1 expression in human peripheral blood T lymphocytes can be increased by stimulation with mAbs to CD3 (OKT3) plus CD28 (Leu²⁸) or PMA plus ionomycin (Fig. 1A). This was due to higher GBP-1 transcription demonstrated by quantitative RT-PCR in primary T lymphocytes, as well as in Jurkat T cells (Fig. 1B, 1C). Because GBP-1 possesses an isoprenylation site for possible membrane binding, we monitored the localization of GBP-1. We therefore transduced Jurkat T cells with either GFP-tagged GBP-1 or GFP alone as a control. After transduction, the cells were sorted for high and equal levels of GFP expression and immunostained for CD147 as a surface marker. Colocalization analysis with CD147 showed that GFP–GBP-1 and control protein GFP were exclusively expressed in the cytosol of Jurkat T cells (Fig. 1D). Cell fractionation experiments confirmed the cytosolic localization of GBP-1, and stimulation of Jurkat T cells with the TCR β-chain mAb C305 for different time points did not lead to membrane binding of GBP-1 (Fig. 1E, Supplemental Fig. 1).

To evaluate a functional role for GBP-1 in T cells, we silenced GBP-1 expression using two different lentiviral delivered shRNA constructs (Fig. 2A). Silencing of GBP-1 with both shRNAs in Jurkat T cells expressing an IL-2 promoter–driven luciferase reporter gene resulted in higher IL-2 promoter activity after stimulation with CD3 and CD28 mAbs or SEE pulsed Raji B cells compared with cells transduced with a control shRNA (Fig. 2B, 2C). In contrast, bypassing early TCR signaling by stimulation of these cells with PMA/ionomycin did not lead to higher IL-2 expression (Fig. 2D). Enhanced IL-2 promoter activity in GBP-1–silenced cells was independent of the time used for stimulation (Supplemental Fig. 2A). We verified our promoter studies by detecting higher IL-2 protein concentrations in the supernatant of GBP-1–silenced Jurkat T cells irrespective of the SEE concentration used (Supplemental Fig. 2B). We further verified our results with primary human peripheral blood T lymphocytes. GBP-1 expression was efficiently silenced with shRNA construct 1 (Fig. 2E). GBP-1–silenced T lymphocytes stimulated with increasing concentrations of CD3/CD28 mAbs or SEE-pulsed Raji B cells showed higher IL-2 concentrations in the culture supernatants (Fig. 2F, 2G). Again, no difference was observed when we used PMA/ionomycin for stimulation (Fig. 2H).

To identify interaction partners, which may explain the effects of GBP-1 on T cell activation, we stably overexpressed GBP-1 in Jurkat T cells. We used a p-tagged GBP-1 in a retroviral pBMN-IRES-GFP expression system. Independent transduction and sorting for high GFP expression led to the generation of two batches of p-tagged GBP-1–expressing Jurkat T cells (Fig. 3A). Both GBP-1–overexpressing cell batches showed reduced proliferation upon stimulation with CD3/CD28 mAbs (Fig. 3B). This is in accordance with results obtained with endothelial cells (6). We used these cell lines to precipitate GBP-1 with a p-tag–specific mAb (clone H902). The precipitate was analyzed by silver stain and GBP-1–specific immunoblotting to estimate precipitation efficiency of GBP-1 and to ensure adequate purity for subsequent mass spectrometry (Figs. 3C, 4B). As a negative control, we used cell lysates derived from Jurkat cells not expressing p-tagged GBP-1. For each mass spectrometry run, the intensities of the precipitated proteins in the negative control were subtracted from the intensities obtained with the p-tag–specific immunoprecipitation. To increase confidence of identification, we analyzed four independent mass spectrometry runs from the two GBP-1–overexpressing Jurkat T cell batches. We judged as potential binding partners of GBP-1 those molecules that were present in each of the four independent experiments. Based on this criterion, 10 proteins were classified as interaction partners of GBP-1 (Fig. 3D).

With the help of stable isotope labeling by amino acids in cell culture (SILAC)–based mass spectrometry, it is possible to identify quantitative differences in the phosphorylation status of a protein in distinct populations of cells. For this purpose, GBP-1–silenced Jurkat T cells were grown in culture medium containing the light forms of l-arginine and l-lysine, whereas wild type Jurkat cells were grown in culture medium supplemented with the heavy forms of l-arginine and l-lysine. The two populations were combined before sequential phospho-peptide enrichment and liquid chromatography–tandem mass spectrometry analysis. Using this technique, we identified two altered phosphorylation sites in plastin-2, 1 of the 10 potential interaction partners identified previously: serine residues 5 and 7, known to play a role in actin bundling (29) and T cell regulation (30), were constitutively higher phosphorylated in GBP-1–silenced Jurkat T cells (Fig. 4A). To corroborate the mass spectrometry data, we performed coimmunoprecipitation experiments with the Jurkat T cell batches overexpressing p-tagged GBP-1. As shown in Fig. 4B, we were able to coimmunoprecipitate plastin-2 with p-tagged GBP-1. To exclude possible effects of the p-tag or cell line artifacts, we precipitated endogenous GBP-1 of Jurkat T cells and of primary T lymphocytes, left either unstimulated or stimulated with OKT3 alone or in combination with Leu²⁸ for 24 h. Again, we obtained coimmunoprecipitation of GBP-1 and plastin-2 (Fig. 4C). Because plastin-2 controls the assembly of the actin cytoskeleton into bundles (29), we were interested whether GBP-1 was capable of associating with β-actin. Indeed, we found β-actin as a binding partner of GBP-1 in three of the four mass spectrometry runs performed. Further, we coprecipitated endogenous GBP-1 and β-actin from lysates of primary human T lymphocytes (Fig. 4C).

To analyze whether plastin-2 plays a role in the phenotype of shGBP-1 cells, we used Jurkat T cells doubly silenced for GBP-1 and plastin-2. For plastin-2 silencing, we used two retrovirally delivered shRNAs specific for two different positions in the plastin-2 mRNA. The plastin-2 shRNA shPlastinp1 yielded a silencing efficiency of around 80%. The second position, shPlastinp2, did not show a reduction of plastin-2 expression. Therefore, we used the latter construct as a control shRNA for plastin-2 silencing (Fig. 4D). Wabnitz et al. (30) reported that higher plastin-2 activation led to a higher CD69 surface display in T cells that was dependent on the transport of CD69 to the surface. Strikingly, we were able to blunt the higher CD69 expression in GBP-1–silenced Jurkat T cells by cosilencing of plastin-2 (Fig. 4E). In addition, we assessed the IL-2 protein concentration in the supernatants of singly and doubly silenced cells stimulated with SEE pulsed Raji B cells (Fig. 4F). Cosilencing of GBP-1 and plastin-2 did not alter the higher IL-2 expression observed in Jurkat T cells silenced for GBP-1 alone. Moreover, silencing of plastin-2 did not abrogate IL-2 expression in human cells.

In summary, although this set of experiments shows a physical and functional interaction between GBP-1 and plastin-2 in terms of CD69 expression, this interaction is not responsible for the enhanced IL-2 expression upon TCR/CD3 stimulation of GBP-1–silenced T cells. So the question remained which interaction partner of GBP-1 transduced IL-2 inhibitory signals during T cell activation.

Because the actin cytoskeleton is essential to drive T cell spreading to organize the immunological synapse (31), we aimed to prove whether the cytoskeleton contributed to higher activation of GBP-1–silenced T cells. We assessed the spreading of GBP-1–silenced Jurkat T cells on a surface coated with OKT3. GBP-1–silenced cells showed a more pronounced spreading on the CD3 surface than control cells (Fig. 5A). Interestingly, the difference in spreading was less outspoken on a surface cocoated with CD3 and CD28 mAbs (Fig. 5A–C, Supplemental Fig. 3, Supplemental Movies 1–4).

Besides β-actin and plastin-2, we identified by mass spectrometry three other GBP-1 interacting proteins (Fig. 3D) that interfere with the cytoskeleton and early TCR signaling: stomatin-like protein 2 (STOML2) (32, 33), myosin regulatory L chain MRCL3 variant (MRCL3) (34), and the βII-spectrin. βII-Spectrin was described to be involved in CD3 and CD45 surface display and IL-2 production (35, 36). We found that CD45 and CD3 surface expression was enhanced in GBP-1–silenced Jurkat T cells, whereas the levels of several other surface receptors such as CD11a, CD18, and CD43 were not altered (Fig. 5D). We confirmed the higher surface expression of CD45 and CD3 upon silencing of GBP-1 in primary T lymphocytes (Fig. 5E). Thus, these results suggest that the interaction of GBP-1 with βII-spectrin altered CD3 and CD45 surface expression. To test whether βII-spectrin is involved in the GBP-1–mediated negative regulation of TCR signaling, we cosilenced βII-spectrin and GBP-1 in Jurkat T cells. We used two different shRNA constructs for βII-spectrin silencing, which lead to either 20% (shSpectrin1) or 95% silencing (shSpectrin2; Fig. 5H). Silencing of βII-spectrin by both shRNA constructs evoked decreased expression of CD3 and CD45 in Jurkat T cells (Fig. 5I, 5J). Experiments in Jurkat T cells expressing an IL-2 luciferase reporter construct and stimulated with SEE-pulsed Raji B cells showed that silencing of βII-spectrin decreased IL-2 expression. Moreover, the overshooting expression of IL-2 in GBP-1–silenced cells was partially reversed by cosilencing of βII-spectrin (Fig. 5K).

The alterations in CD3 and CD45 expression, and cytoskeletal modifications found in GBP-1–silenced cells might influence T cell signaling. Therefore, we probed with an anti-phosphotyrosine mAb lysates of GBP-1–silenced and control silenced Jurkat T cells stimulated with the TCR β-chain mAb C305 for up to 30 min. With both shRNA constructs used for GBP-1 silencing, we observed higher tyrosine phosphorylation of several signaling molecules compared with the control cells (Fig. 6A). To investigate this finding in more detail, we used phosphorylation-site–specific Abs against important signaling molecules and stimulated over a broader time range (Fig. 6B, Supplemental Fig. 4). Lck was consistently higher phosphorylated at the activatory position 394 in GBP-1–silenced Jurkat T cells, whereas the inhibitory tyrosine residue at position 505 was not affected (Fig. 6C). All subsequent TCR signaling molecules tested showed higher phosphorylation in GBP-1–silenced cells (Fig. 6B, 6C, Supplemental Fig. 4). To see whether the differences in early TCR signaling were transferred to distal parts of the signaling cascade, we monitored intracellular Ca²⁺ concentrations by flow cytometry. GBP-1–silenced cells stimulated via cross-linking of the TCR exhibited elevated intracellular Ca²⁺ levels. When we used the sarcoplasmic/endoplasmic reticulum calcium ATPase inhibitor, thapsigargin, thereby bypassing early TCR signaling, the difference in the intracellular Ca²⁺ concentration between GBP-1–silenced and control silenced cells disappeared (Fig. 6D). We also detected higher phosphorylation of ERK in GBP-1–silenced Jurkat T cells, proving also higher activation of distal signaling proteins (Fig. 6E, 6F, Supplemental Fig. 4).

Finally, we wanted to corroborate the earlier findings in primary T lymphocytes isolated from human peripheral blood. We observed higher phosphorylation of PLC-γ (Fig. 7A, 7B) and higher intracellular Ca²⁺ concentration upon cross-linking of CD3 (Fig. 7C) in GBP-1–silenced T lymphocytes. Furthermore, LAT was higher phosphorylated in the GBP-1–silenced T lymphocytes. Higher phosphorylation of the MAPK ERK indicated that in GBP-1–silenced primary T lymphocytes, the higher activation of early TCR signaling molecules was transduced to late signaling events, confirming our findings with Jurkat T cells (Fig. 7A, 7B).

Screening of human cells and tissues showed that GBP-1 expression is highest in T lymphocytes (16). We show in this article that this basal expression of GBP-1 can be further increased by stimulation via the TCR. Most likely this is due to the fact that T cells produce IFN-γ upon stimulation that leads to subsequent activation of the GBP-1 promoter (37, 38). Recently, an alternative pathway of promoter activation of GBP-1 was described involving Src and p38 kinases (39). Thus, it is also possible that TCR activation directly influences GBP-1 expression. We detected the highest expression of GBP-1 24 h after onset of T cell stimulation, a typical time point when negative regulators are getting expressed (40). Therefore, this finding was our first hint that GBP-1 may assist in tuning down T cell responses.

The data presented in this article establish GBP-1 as a negative regulator of T cell activation by interfering with early TCR signaling. Early TCR signaling molecules are hyperphosphorylated in GBP-1–silenced cells. Moreover, treatment of T cells with agents bypassing early signaling events do not show any difference between GBP-1–silenced and control silenced T cells: specifically, IL-2 promoter activity was not changed upon stimulation of T cells using the PKC activator PMA plus the Ca²⁺ ionophore ionomycin or by evoking a Ca²⁺ response with the sarcoplasmic/endoplasmic reticulum calcium ATPase inhibitor thapsigargin.

To get an insight into the molecular mechanism of GBP-1 in cells in general and in T cells in particular, we screened for interaction partners using mass spectrometry. Currently, little is known about cellular partners of GBP-1. It is known that GBP-1 binds (besides homodimerization and heterodimerization with GBP family members) (41, 42) to the cytoskeletal proteins β-IIITubulin and PIM1 (14). Five among the 10 potential GBP-1 binding partners identified by us by mass spectrometry (multisynthetase complex auxiliary component p43, plastin-2, STOML2, MRCL3, and βII-spectrin) are described to be involved in the remodeling of the (actin-) cytoskeleton (33, 36, 43, 44) and 4 of these 5 (plastin-2, STOML2, MRCL3, and βII-spectrin) to influence T cell activation (30, 32–36, 44–46). Based on the fact that IL-2 expression is abrogated in plastin-2 knockout mice, we suspected plastin-2 to be involved in the GBP-1–mediated IL-2 regulation. However, we could not confirm the results obtained in mice with human T cells. We found that plastin-2 silencing neither blunted the higher IL-2 expression observed in GBP-1–silenced cells nor decreased IL-2 expression on its own, pointing to a plastin-2–independent effect of GBP-1 on human IL-2 regulation and to a different function of plastin-2 in mouse versus human T cells. This finding was at first glance hard to understand, because our pull-down and SILAC mass spectrometry experiments further showed that GBP-1 regulated the phosphorylation status of plastin-2 at Ser⁵, which was shown to increase its actin-bundling activity (29, 47), and moreover, GBP-1 itself bound to β-actin (Fig. 4A–C). Strikingly, we found that this physical and functional interaction of GBP-1 with plastin-2 regulates CD69 surface expression (Fig. 4E). Finally, double-silencing experiments identified βII-spectrin as one of the molecular cooperation partners of GBP-1 regulating IL-2 activation.

It was described that the βII-spectrin cytoskeleton influences IL-2 production in T cells by mediating CD3 and CD45 surface expression (35, 36). We show in this article that GBP-1 interferes with CD3 and CD45 surface display (Fig. 5D–G). Thus, we speculated that the interaction of GBP-1 with actin and βII-spectrin would control the mobility and compartmentalization of TCR/CD3 and its signaling molecules. Higher TCR/CD3 expression on the cell surface makes more triggering molecules for Ag and more cytoplasmic binding sites (ITAMs) for juxtamembrane signaling molecules available. Higher CD45 expression should lead to lower phosphorylation at the inhibitory tyrosine residue 505 of Lck, and thus higher activity of this central kinase of TCR/CD3 signaling. In this respect, it is at first glance a paradox that Lck is not lower phosphorylated at the inhibitory tyrosine position 505 in GBP-1–silenced cells. Indeed, it has been reported that low expression levels of CD45 lead to hypoactivation, and small changes in CD45 expression have a big effect in T cell activation (48). However, Lck is higher phosphorylated at the activatory tyrosine position 394, and more important, the ratio of phosphorylation at 394 to 505 is higher in GBP-1–silenced cells. Thus, we speculate that GBP-1, via changing expression of CD45 and CD3 through interaction with βII-spectrin, shifts the balance of inactivated Lck-pools, which are either unphosphorylated on both the activatory tyrosine 394 and the inhibitory tyrosine 505 or phosphorylated on tyrosine 505 only, to the activated Lck pools, which are either phosphorylated on tyrosine 394 or on both tyrosine residues (49, 50). Nonetheless, nearly complete βII-spectrin silencing did not completely blunt the overshooting IL-2 expression in GBP-1–silenced T cells, which indicates involvement of additional proteins and processes in the GBP-1–mediated effect on TCR signaling.

In summary, we established the IFN-γ–inducible GBP-1 as a new negative regulator of early TCR signaling. We identified previously unknown cellular binding partners of GBP-1, which together with GBP-1 modify T cell activation. However, some of the newly identified binding partners of GBP-1 could also provide, apart from T cell stimulation, an explanation for the molecular basis of the diverse physiological functions of GBP-1 described so far. Because most of these proteins regulate cytoskeletal processes, we suggest that T cells can regulate via IFN-γ cytoskeleton-dependent cellular functions at several levels in an autocrine and paracrine manner through GBP-1.

We thank Eva Steinhuber and Margarethe Merio for excellent technical assistance. We are grateful to the imaging core facility of the Medical University of Vienna.

The authors have no financial conflicts of interest.