If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported by grants from the Fonds der Chemischen Industrie (Frankfurt am Main) and the Deutsche Forschungsgemeinschaft (SFB 542). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Signal transduction through cytokine receptors is mediated mainly by non-covalently associated Jak tyrosine kinases. By confocal microscopy, the cytokine receptor gp130 and Jak1, fused with either yellow (YFP) or cyan (CFP) fluorescent protein, were found to be colocalized predominantly at intracellular vesicular structures and at the plasma membrane. Quantitative fluorescence recovery after photobleaching (FRAP) analysis at the plasma membrane revealed equal mobilities for gp130-YFP and Jak1-YFP. Thus, Jak1-YFP diffuses like a transmembrane protein indicating that membrane-bound Jak1 does not exchange rapidly with cytosolic Jaks. Applying a novel dual-color FRAP approach we found that immobilization of gp130-CFP by a pair of monoclonal antibodies led to a corresponding immobilization of co-transfected Jak1-YFP. We conclude from these findings that Jak1, once bound to a gp130 molecule, does not exchange between different receptors at the plasma membrane neither via the cytoplasmic compartment nor via a membrane-associated state.
Among receptors that signal through activation of tyrosine kinases one can discriminate between receptor tyrosine kinases and cytokine receptors. Receptor tyrosine kinases consist of a ligand binding ectodomain, a single transmembrane region, and a cytoplasmic part that contains a tyrosine kinase domain (
According to structural properties, cytokine receptors are subdivided into class I cytokine receptors activated by hematopoietic cytokines and class II cytokine receptors engaged by interferons and cytokines of the IL-10 family (
). Among the class I cytokine receptors, gp130 is the most promiscuous one since it serves as the common signal transducing receptor subunit for the IL-6-type cytokines comprising IL-6, IL-11, LIF, OSM, CNTF, CT-1 (
). IL-6 and IL-11, after binding to their specific α-receptor subunits IL-6Rα and IL-11Rα, respectively, signal via gp130 homodimers. The remaining IL-6-type cytokines induce the formation of a heterodimer of gp130 and a second signal transducing subunit, namely LIFR or OSMR (
). Upon ligand binding the receptor-associated Jaks become activated by transphosphorylation. The activated Jaks phosphorylate tyrosine residues within the cytoplasmic part of gp130. These phosphotyrosine residues serve as docking sites for the transcription factors STAT3 and STAT1. The STATs also become tyrosine phosphorylated, dimerize and translocate into the nucleus to induce target genes (
By mutagenesis studies, the membrane proximal box1/box2 region of cytokine receptors has been identified to be essential for Jak binding. This part is characterized by a proline-rich motif (box1) and a less conserved region dominated by a hydrophobic stretch of amino acids followed by several charged amino acids (box2). For gp130 it has been demonstrated that box1 as well as the interbox1/2 region are most important for Jak binding (
Jaks are relatively large proteins of about 130 kDa that contain a kinase- and a kinase-like domain in their C-terminal region. By mutation of Lys907 to E in the kinase domain a kinase-deficient (kd) mutant (Jak1/kd) is generated that still associates with cytokine receptors (
). Subdomain F1 has a ubiquitin-like β-grasp fold. A mutagenesis study based on a molecular model of the F1 subdomain of Jak1 has highlighted its importance for the interaction with gp130. Exchange of Leu80-Tyr81 to alanines within the FERM domain leads to a Jak1 mutant (Jak1/nb) that does not bind (nb) to cytokine receptors (
Here, the well characterized gp130/Jak1 interaction was used as a model system to study the dynamics of cytokine receptor/Jak association in living cells. For this purpose, gp130 and Jak1 were fused with the yellow (YFP) and cyan (CFP) variants of the green fluorescent protein. Both, fluorescent gp130 and fluorescent Jak1 were found to be predominantly associated with cellular membranes. The mobilities of the fusion proteins at the plasma membrane were measured by the technique of fluorescence recovery after photobleaching (FRAP) (
). The data reveal that both gp130 and Jak1 have similar recovery kinetics that are typical for membrane-bound proteins. Furthermore, we show that gp130 and Jak1 are tightly bound to each other and that Jak1 does not exchange between different receptor proteins.
Proteins, Chemicals, and Cell Culture—Enzymes were purchased from Roche Applied Science (Mannheim, Germany). DMEM and antibiotics were obtained from Gibco Invitrogen Corporation (Auckland, UK) fetal calf serum was provided by Seromed (Berlin, Germany). [α-32P]DeoxyATP was purchased from Hartmann Analytic (Braunschweig, Germany). Epo (220500 units/mg) was a kind gift from Roche Applied Science. The gp130 mAbs B-P4, B-P8, and B-S12-G7 (Diaclone, Besançon, France) were generated as described elsewhere (
). Secondary antibodies were purchased from Dako (Hamburg, Germany). COS-7 simian monkey kidney cells and Jak1-deficient U4C cells (kindly provided by Dr. I. M. Kerr, Cancer Research UK, London) were cultured in DMEM supplemented with 10% fetal calf serum, 100 mg/liter streptomycin, and 60 mg/liter penicillin. To the medium of U4C cells 400 μg/ml G418 were added. Cells were grown at 37 °C in a water-saturated atmosphere at 5% CO2.
Cloning of Expression Vectors—The vectors pEYFP-N1, pECFP-N1, and pEGFP-N1 (Clontech) were used as templates for the amplification of cDNAs encoding fluorescent proteins. Applying standard PCR and cloning procedures, the cDNA encoding full-length (amino acids 1-918) human gp130 was fused with the cDNA encoding the respective fluorescent protein (amino acids 2-239) and ligated into the pSVL expression vector (Amersham Biosciences) yielding the plasmids pSVLgp130-YFP and pSVLgp130-CFP. The constructs pSVLgp130/id-YFP and pSVLgp130/id-CFP encode a truncated gp130 protein in which Pro668 of gp130 is immediately followed by Val2 of the fluorescent protein. In the construct gp130/nb-YFP the truncated gp130 described above lacks also the box 1 motif (Ile651-Pro658). pSVLEg-CFP was cloned by fusing the cDNA encoding CFP (amino acids 2-239) immediately downstream of the cDNA encoding the earlier described Epo-gp130 chimera (Eg) (
) were constructed by fusing the cDNA encoding full-length human Jak1 (amino acids 1-1142) with the gene encoding the respective fluorescent protein. Due to the cloning strategy Jak1 and the fluorescent protein are separated by a linker of nine amino acids. Applying the same approach, the earlier described Jak1/nb mutant (
) was fused with YFP. All constructs were verified by DNA sequencing.
Characterization of gp130 Fluorescent Fusion Proteins—COS-7 cells were transfected with the expression vectors pSVLgp130-YFP, pSVLgp130/id-YFP, or pEYFP-N1 (Clontech) or empty pSVL vector in addition to pSVLJak1 using FuGene (Roche Applied Science). 48 h after transfection, cells were lysed using a Brij lysis buffer (50 mm Tris/HCl, pH 8.0, 10% glycerol, 0.25% Brij-96, 50 μm Na3VO4, 100 μm EDTA, 1 mm phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml leupeptin) and immunoprecipitations were performed with a goat polyclonal GFP antibody (Rockland). After immunoblotting, Jak1 and gp130 fusion proteins were detected using a rabbit polyclonal Jak1 antibody (HR-785, Santa Cruz Biotechnology) and a monoclonal gp130 antibody (B-P4, Diaclone, Besançon, France), respectively.
For assaying STAT activation mediated by the chimeric receptors, COS-7 cells were transfected with pSVLEg, pSVLEg-CFP, or empty pSVL-vector. 48 h after transfection, cells were stimulated with 7 units/ml Epo for 15 min. Subsequently, nuclear extracts were prepared and STAT DNA binding activity was determined in an EMSA using the m67SIE-probe as described previously (
Characterization of Jak1 Fluorescent Fusion Proteins—COS-7 cells were co-transfected with gp130-CFP and with expression vectors for the various Jak1-derivatives. After 48 h, cells were lysed and immunoprecipitations with the gp130 antibody B-P8 (Diaclone) were performed. The immunoblot after SDS-PAGE was probed with the Jak1 antibody HR-785. After stripping, the membrane was reprobed with gp130 antibody B-P4 (Diaclone) to control for equal loading. Expression controls were performed with lysate blots detected with the same Jak1 and gp130 antibodies. Jak1-phosphorylation was detected using the antibody PY99 (Santa Cruz Biotechnology).
Confocal Microscopy—Confocal imaging was carried out on a Zeiss LSM 510 confocal microscope equipped with an argon ion laser (Zeiss, Jena, Germany). The argon ion laser was modulated by an acousto-optical modulator. For excitation of CFP imaging the 458 nm line was used. The beampath for CFP contained a 458 nm main dichroic mirror and a 480/20 nm bandpass filter for detection of the emitted fluorescence. This narrow filter excludes any fluorescence emerging from YFP that is also excited by the 458 nm laser beam. The beampath for YFP detection contained excitation at 514 nm and a main dichroic mirror of 514 nm. The emitted light was monitored with a 530 nm longpass filter. CFP is not excited by the 514 nm laser. For cells co-expressing CFP and YFP the multitrack function of the LSM 510 was used.
Live cell imaging was carried out with COS-7 cells transiently transfected using the DEAE-dextran method. Transfected cells were grown on 18 mm glass coverslips. 48 h after transfection the coverslips were placed in a home built perfusion chamber and flushed with DMEM supplemented with 10% fetal calf serum, 100 mg/liter streptomycin, and 60 mg/liter penicillin. To maintain 37 °C during all image acquisition the perfusion chamber and the objective were thermostatted. For qualitative cell imaging a laser power of 35% of 25 milliwatts and 8% transmission was used. The images shown here represent 1-1.5-μm thick confocal slices of whole cells.
FRAP Experiments—The general optical setup is described under Confocal Microscopy. A 63× water corrected immersion lens (numerical aperture 1.2) was used with a pinhole adjustment resulting in a 2-μm optical slice. After 10 prebleach scans (1 scan/second) of a circular region of interest (ROI 1) with a diameter of 1.3 μm, ROI 1 was bleached. Directly after bleaching the fluorescence recovery was sampled for 70 s once every second. In the antibody immobilization experiments the ROI fluorescence was recorded with 10 pre- and 40 postbleach scans (1 scan/2.2 s). An additional ROI on the plasma membrane (ROI 2) was monitored in parallel to ROI 1 but was not bleached in order to detect cellular movements and fluorescence fluctuations independent of bleaching. A third ROI (ROI 3) was placed outside the cell to measure background fluorescence.
The settings for prebleach and recovery image scans were 100% of 25 milliwatts and 1% transmission for 458 nm (CFP excitation) and 100% of 25 milliwatts and 0.1% transmission for 514 nm (YFP excitation). Bleaching in ROI 1 was carried out with 200 scan iterations (3.7-3.8 s total bleach time) of 100% of 25 milliwatts and 100% transmission for 458 and 514 nm. During all scan and bleach-scan procedures the scan speed was set to maximum (pixel-time, 1.8 μs). The optical zoom factor of 3.5 led to a pixel size of 224 nm within the ROIs. For qualitative FRAP analysis, during the prebleach and recovery time periods the whole cell was scanned.
For dual FRAP experiments in co-transfected cells, CFP and YFP were bleached simultaneously and images of CFP and YFP fluorescence were obtained frame by frame within one scanning procedure using the main dichroic mirror 458/514 for both channels and the multitracking function of the Zeiss LSM 510. Due to the low laser power and the short time of image acquisition the fluorescence intensity was not significantly influenced through 80 acquisitions.
To determine the effects of antibodies against gp130 on the mobilities of the fluorescently labeled proteins, cells were incubated with BP-8 and BS-12-G7 at a concentration of 10 μg/ml for each antibody.
Evaluation of FRAP Data—Average fluorescence intensities within ROIs were measured under the same condition for each data set and exported into Microsoft Excel. The half-life of fluorescence recovery (t½) was determined by curve fitting of experimental data using a mathematical model of fluorescence recovery by diffusion (
) where F∞ is the average fluorescence in the ROI after full recovery, F0 is the fluorescence immediately after the bleach and F- is the average fluorescence before bleaching.
Expression and Characterization of Fluorescent gp130 and Jak1 Fusion Proteins—In order to examine the subcellular localization and mobilites of gp130 and Jak1 in living cells by confocal microscopy and FRAP, fluorescent fusion proteins were generated. YFP and CFP were fused to the C terminus of full-length human gp130 resulting in the constructs gp130-YFP and gp130-CFP. Additionally, a truncated form of gp130 (amino acids 1-668) was C-terminally fused with fluorescent proteins leading to the constructs gp130/id-YFP and gp130/id-CFP. These constructs are internalization deficient (id) since they lack the dileucine motif (Leu786-Leu787) (
The confocal images of two transfected COS-7 cells in Fig. 1A show the typical expression patterns of the fusion proteins. As expected, gp130-YFP appears as a membrane-bound protein enriched at intracellular membranes and vesicles. Staining of the plasma membrane is also clearly visible. Gp130/id-YFP shows a similar distribution with a more pronounced staining of the plasma membrane due to reduced internalization of the receptor.
The receptor constructs were analyzed with respect to Jak association and ligand-induced STAT activation. COS-7 cells were co-transfected with expression vectors encoding Jak1 and either gp130-YFP, gp130/id-YFP or YFP. Immunoprecipitation of the receptor fusion proteins with a GFP antibody led to the co-precipitation of Jak1 (Fig. 1B, upper panel, lanes 1 and 2). In cells transfected with YFP and Jak1 or Jak1 alone, no co-precipitation of Jak1 was observed (lanes 3 and 4). Therefore, we conclude that both receptor constructs are associated with Jak1. Immunodetection of the precipitated gp130 fusion proteins revealed apparent molecular masses that are in agreement with the calculated molecular masses of 155 kDa and 125 kDa for gp130-YFP and gp130/id-YFP, respectively (Fig. 1B, lower panel).
Next, the capability of fluorescent full-length gp130 to induce STAT activation upon cytokine stimulation was investigated. Since COS-7 cells endogenously express low levels of gp130 wild type the extracellular portion of gp130-CFP was replaced by the extracellular region of the EpoR. The resulting chimeric receptor Eg-CFP readily activates STAT1 in response to Epo stimulation (Fig. 1C, lanes 5 and 6) similarly to the well established Epo-gp130 chimera (
) (lanes 3 and 4) indicating that signaling is largely unaffected by the CFP fusion. Non-transfected COS-7 cells do not respond to Epo stimulation since they do not express the EpoR endogenously (lanes 1 and 2).
Jak1-YFP, Jak1-CFP, Jak1/nb-YFP, and Jak1/kd-YFP were cloned by fusing the fluorescent proteins C-terminal to the kinase domain of Jak1 via a linker of nine amino acids. Jak1/nb is the L80AY81A mutant of Jak1 that does not bind (nb) to gp130 anymore (
). To demonstrate that fluorescent Jak1 has the ability to associate with the cytoplasmic region of gp130, co-precipitation experiments with gp130-CFP were performed. After expression in COS-7 cells both Jak1/kd-YFP and Jak1-YFP were co-precipitated with the receptor (Fig. 1D, lanes 1 and 2). As expected, Jak1/nb-YFP does not bind to the receptor (Fig. 1D, lane 3). In the loading and expression control panels fairly equivalent amounts of Jak proteins and receptors can be detected. Immunoprecipitation from cellular lysates reveals that overexpressed Jak1-YFP is constitutively phosphorylated (Fig. 1E, lane 3) while the kinase-deficient Jak1/kd-YFP is not (lane 2).
Lateral Mobility of gp130-YFP Determined by FRAP—To determine the mobility of gp130 fusion proteins, small plasma membrane areas (regions of interest, ROIs) of living COS-7 cells expressing gp130-YFP were bleached using the 514 nm laser of the confocal microscope (Fig. 2A, upper images). Usually, bleaching down to 30-20% of the initial fluorescence intensity was achieved. Subsequently, the return of fluorescence into the bleached area was measured. In a typical experiment presented in the diagram in Fig. 2A, recovery of about 85% was observed. It is a well-known phenomenon of FRAP studies that no full recovery occurs (
). The remainder of 15% is designated as “immobile fraction” since it is believed that it represents the portion of bleached gp130-YFP that is unable to diffuse freely and therefore cannot be replaced by non-bleached gp130-YFP. If the immobile fraction is generated by the energy input of the laser beam one would expect that consecutive bleaching leads to an increase in the immobile fraction. We observed that after two (Fig. 2B) or more rounds of bleaching and recovery of the same ROI, the immobile and mobile fractions remained constant. Therefore, the immobile fraction seems to be of biological significance rather than an artifact of the bleaching procedure.
For quantitative FRAP measurements, two different ROIs on the plasma membrane with a diameter of 1.3 μm each were defined. One of these two ROIs was bleached and recovery of fluorescence was monitored. The other ROI remained unbleached and was used to monitor movements of the plasma membrane. Experiments in which changes in this control ROI occurred were not used for further evaluation. From the increase of fluorescence in the bleached ROI the half-life of fluorescence recovery was calculated by curve fitting (red graph in Fig. 2A) (
The confocal images in Fig. 3 represent the distribution pattern of the co-expressed fusion proteins used in FRAP experiments. COS-7 cells were co-transfected either with gp130-YFP and Jak1-CFP (Fig. 3A), with gp130-CFP and Jak1-YFP (Fig. 3B) or with gp130-CFP and Jak1/kd-YFP (Fig. 3C). Images were taken from transfected living cells by confocal microscopy using the YFP- and CFP channels equipped in a way (see “Experimental Procedures”) that no cross-bleeding between the channels occurs. The intracellular distribution of Jak1-YFP (Fig. 3B) and Jak1/kd-YFP (Fig. 3C) is similar to the one observed for gp130-YFP (Fig. 3A); Jak1-YFP and Jak1/kd-YFP are localized at the plasma membrane and at intracellular membrane structures while there is little cytoplasmic staining.
The diagram in Fig. 4A shows the half-lives for YFP fluorescence recovery obtained from COS-7 cells expressing gp130-YFP and Jak1-CFP (7 cells). From each cell 3-6 FRAP curves were obtained. The average half-lives range from 13.2-20.2 s for gp130-YFP. The overall mean value is 16.1 ± 2.1 s (Table I).
Table ISummary of YFP fusion protein fluorescence recovery at 37 °C
Since transiently transfected COS-7 cells were used for the experiments, expression levels of fluorescent fusion proteins varied from cell to cell. All experiments were performed under identical microscope settings for the laser energy input and the fluorescence detection. Therefore, the initial fluorescence intensities are a good measure for the expression levels of the fluorescent proteins. In Fig. 4B, the calculated half-lives of gp130-YFP fluorescence recovery from individual experiments are depicted as a function of fluorescence intensity of the ROIs before bleaching. No dependence on fluorescence intensity is observed indicating that the variable expression of the fusion proteins does not have any influence on the actual half-lives of fluorescence recovery.
From each FRAP curve, the mobile and immobile fractions were calculated. From the chart in Fig. 4C it becomes evident that the mobile fraction is largely independent of the initial fluorescence intensity of the bleach ROI. Thus, the actual expression level of the fluorescent gp130 fusion proteins does not determine the extent of the mobile fraction. Most interestingly, there is a strong correlation between diffusion and the mobile fraction (Fig. 4D). A large mobile fraction of fluorescent gp130 is accompanied with a low half-life of fluorescence recovery.
The Mobility of Jak1-YFP Resembles the One of gp130-YFP—Next, the half-lives of fluorescence recovery for Jak1-YFP and Jak1/kd-YFP were determined as a measure of their mobility in a quantitative manner as described for the fluorescent gp130 fusion protein. From the appearance of Jak1-YFP and Jak1/kd-YFP in the microscopic images, it could not be excluded that the membrane-bound Jaks might be in equilibrium with a small pool of cytosolic Jaks. Since cytosolic proteins diffuse at least 100-fold faster than membrane-bound proteins, a rapid exchange of membrane-bound Jaks with a cytosolic pool should become evident by a markedly decreased half-life of fluorescence recovery compared with gp130.
The overall mean values of half-lives of fluorescence recovery are 14.2 ± 3.0 s and 15.7 ± 4.0 s for Jak1-YFP and Jak1/kd-YFP, respectively (Table I, Fig. 4E). Not only the half-lives of Jak1-YFP, Jak1/kd-YFP, and gp130-YFP (Fig. 4E), but also the mobile fractions (Table I) are similar (about 77%). The measured half-lives of Jak1-YFP and Jak1/kd-YFP are independent of the actual expression levels of the fluorescent protein. As for fluorescent gp130, we observed no dependence of the mobile fraction on the expression level of Jak1-YFP or Jak1/kd-YFP.
To further corroborate that Jak1 mobility is not influenced by the high expression levels achieved in COS-7 cells, Jak1-deficient U4C cells were stably transfected with Jak1 or Jak1-YFP. In the transfectants, Jak1 and Jak1-YFP are expressed at low levels. Thus, phosphorylation of the Jaks is not constitutive but can be induced by IL-6 stimulation (not shown). As shown in Fig. 3E, fluorescence recovery curves of Jak1-YFP obtained in U4C cells do not differ from the one measured in transiently transfected COS-7 cells.
The receptor-binding deficient Jak1-L80AY81A mutant fused with YFP (Jak1/nb-YFP) exhibits a clearly different subcellular localization compared with wildtype Jak1-YFP and Jak1/kd-YFP (Fig. 3D). Jak1/nb-YFP is evenly distributed throughout the cytoplasm and is excluded from the nucleus. As typical for a freely diffusing cytoplasmic protein bleaching of Jak1/nb-YFP results in a dramatically faster recovery compared with wild-type Jak1-YFP or Jak1/kd-YFP (Fig. 3E). We conclude that the mobility of Jak1-YFP and the kinase deficient Jak1/kd-YFP resembles the one of a membrane protein. There is no evidence for a rapid exchange of membrane-bound Jak1 with a cytosolic pool of Jaks.
Influence of gp130 Immobilization on Jak1 Mobility—The data presented so far suggest that Jak1 is constitutively associated with membranes and binds to available receptors. With the following approach we examined to which extent a dynamic equilibrium at the plasma membrane exists between receptor-bound and freely lateral diffusing membrane-bound Jaks. Such an equilibrium would enable Jaks to exchange between different receptor molecules. We therefore immobilized gp130 by treatment of cells with a pair of monoclonal antibodies against the gp130 extracellular region. A bleached ROI at the plasma membrane of COS-7 cells transfected with gp130-YFP almost completely recovers within 2 min (Fig. 5A, upper panel). Incubation of the cells with the monoclonal gp130 antibodies B-P8 and B-S12-G7, however, leads to a strongly reduced recovery (Fig. 5A, lower panel) with a concomitant drastic decrease of the mobile fraction. This effect was used to determine to which extent the mobilities of Jak1 and gp130 are linked.
For this purpose COS-7 cells co-transfected with gp130-CFP and Jak1-YFP were analyzed by bleaching both fluorophores simultaneously in plasma membrane ROIs using the 458 and 514 nm laser of the confocal microscope. Due to the narrow bandpass filter used for exclusive CFP detection, the CFP signal is relatively weak. To achieve a signal intensity required for reproducible FRAP experiments, cells were transfected with gp130/id-CFP or gp130/nb-CFP constructs which results in increased cell surface expression (compare Fig. 1A).
Without antibody treatment gp130/id-CFP and Jak1-YFP display normal recovery kinetics (Fig. 5B, left diagram) with a mobile fraction of about 50%. On the confocal images of the bleached membrane region no gap of fluorescence is visible after recording of the fluorescence recovery (Fig. 5B, right images). However, after treatment of cells with the antibodies B-P8 and B-S12-G7, gp130/id-CFP as well as Jak1-YFP do not recover. The bleached region at the plasma membrane is still visible after data recording (Fig. 5C). This demonstrates the high extent of gp130/id-CFP and concomitant Jak1-YFP immobilization by the monoclonal antibodies. To prove whether Jak1 binding to gp130 is required for Jak1 co-immobilization, a non-binding gp130 mutant lacking almost the complete cytoplasmic part was fused to CFP (gp130/nb-CFP). COS-7 cells were co-transfected with gp130/nb-CFP and Jak1-YFP. Here, the FRAP curves and the images (Fig. 5D) show a recovery of Jak1-YFP in the bleached region independent of the gp130 immobilization by the antibody pair.
In further control experiments (not shown), immobilization of the fluorescent proteins was not observed when cells were treated only with a single monoclonal antibody or had been transfected with the chimeric receptor Eg-CFP. From these findings we draw the important conclusion that Jak1 is linked to gp130 in a way that does not allow exchange of Jak1 between different gp130 molecules because otherwise, an increased mobility of Jak1-YFP compared with immobilized gp130/id-CFP would have been observed.
The aim of this study was to investigate the nature of the interaction of cytokine receptors with Janus kinases. Upon cytokine stimulation, the non-covalently associated Jaks of the receptor dimer become activated most probably by transphosphorylation. By subsequent tyrosine phosphorylation of substrates, the activated Jaks trigger all downstream signaling events. While the stimulation-independent association of Jaks with cytokine receptors is well established, the strength and dynamics of this interaction have not been characterized yet. To study the gp130/Jak1 interaction by FRAP in living cells, variants of GFP were fused to the C termini of Jak1 and gp130.
In order to prove that fusion of the fluorescent tag does not interfere with the biological activities of the fusion proteins, the constructs were functionally characterized. Both gp130-YFP and gp130/id-YFP associate with co-expressed Jak1 while only gp130-YFP is able to activate STAT. The gp130/id-YFP construct is fused downstream of P668 to the fluorescent protein resulting in deletion of all tyrosine residues required for STAT recruitment and activation. From mutagenesis studies it is known that Jak1 binds to the membrane-proximal region of gp130. Most critical for this interaction are box1 and a tryptophan residue (Trp666) in the interbox1/2 region (
As expected, Jak1-YFP associates with the cytoplasmic part of gp130 as shown by co-immunoprecipitation. Also the kinase-deficient mutant of Jak1 (Jak1/kd-YFP) co-precipitates with gp130 confirming that the association between gp130 and Jak1 is phosphorylation-independent. Fluorescently labeled Jak1 and Jak1/kd show the same subcellular distribution as gp130. The finding that amino acid substitutions in the F1 subdomain of the FERM domain of Jak1 (Leu80-Tyr81/Ala-Ala) lead to an exclusively cytosolic localization indicates that the integrity of the FERM domain is important for membrane association. At present it is not clear whether the observed membrane staining of Jaks is mediated only via their association with receptors. It is conceivable that Jaks might also have an intrinsic ability to interact with membrane structures, such as other proteins with FERM domains (
). There is a reciprocal relationship between the half-life of fluorescence recovery and the diffusion constant of membrane proteins. The half-life of fluorescence recovery determined for gp130-YFP (16.1 ± 2.1 s) is within the range of values of other transmembrane receptors fused to YFP (
). Thus, there is no evidence for anchoring of gp130 at the cytoskeleton that would lead to its immobilization. Recently, it has been shown that Jak/STAT signal transduction is not inhibited in the presence of agents that destroy the cytoskeleton (
). Therefore, one can conclude that the Jak/STAT pathway relies on freely diffusing proteins or protein complexes including the receptors at the plasma membrane.
How to think about Jaks? According to their amino acid sequence, Jaks are regarded as cytoplasmic proteins since they neither contain a signal peptide for translation at ER-bound ribosomes nor a predicted transmembrane helix typical for transmembrane proteins (
). On the other hand, it is known that Jaks associate with the cytoplasmic part of cytokine receptors. Therefore, Jaks could be regarded as cytoplasmic proteins that are recruited to membranes by binding to cytokine receptors. Dependent on the strength of this Jak/cytokine receptor interaction one would expect a distribution of Jaks between a cytosolic and a membrane-bound pool. From evaluation of our microscopic images there is no evidence for a considerable pool of cytosolic Jak proteins. Cytosolic proteins diffuse by a factor of 102 faster than membrane-bound proteins (
). Thus, if a dynamic equilibrium between membrane-bound Jaks and a small pool of cytosolic Jaks existed, one would expect half-lives of fluorescence recovery of Jak1-YFP to be significantly shorter than those of gp130-YFP. We observed a virtually identical diffusion of fluorescent Jak1 and gp130 indicating that membrane-bound Jaks do not exchange with a cytosolic Jak pool.
Overexpression of fluorescent proteins is a prerequisite for FRAP studies. Since Jaks are autophosphorylated upon overexpression (see Fig. 1E and Ref.
), most experiments were performed with a kinase deficient fusion protein (Jak1/kd-YFP) in addition to Jak1-YFP. Fluorescence of Jak1-YFP and Jak1/kd-YFP recovers indistinguishably, indicating that Jak autophosphorylation does not interfere with diffusion. Most FRAP experiments were performed with transiently transfected COS-7 cells. Therefore, expression levels varied by a factor of about six as measured by the initial fluorescence intensity of the bleach ROI (see Fig. 4, B and C). We found no influence of the actual expression levels of the fluorescent fusion protein on the FRAP half-lives or the mobile fractions. Thus, there is no evidence for any artificial aggregation of the proteins caused by overexpression. Otherwise, a decrease in diffusion constants and mobile fractions with increasing expression levels would have been observed.
An interesting observation is the strong correlation of the mobile fraction with the half-life of fluorescence recovery supporting current views on diffusional behavior of membrane proteins. During recent years, the classical fluid mosaic model of biological membranes has been modified leading to the domain model that includes constrained diffusion of macromolecules in the lipid bilayer (
). This model takes into account that due to diffusion barriers the lateral diffusion of proteins can be restricted to submicroscopic membrane domains. Although membrane proteins might diffuse freely within these domains, the long-range diffusion as measured over a larger surface area by FRAP is governed by the much slower “hopping” between microdomains. These diffusion constraints contribute to the immobile fraction. When in a particular membrane segment many of these microdomains exist, the diffusion of the proteins outside these domains is hindered. Therefore, a high microdomain density leads to an increased immobile fraction and prolonged FRAP half-lives.
The question remaining is whether Jak1 bound to gp130 is able to exchange between receptors. In other words: Is the affinity of the Jak1/gp130 interaction sufficiently low to allow dissociation from the receptor and subsequent diffusion of Jak1 in a receptor-independent membrane-bound state to become associated with another receptor? To answer this question we took advantage of our observation that gp130 can be efficiently immobilized on the cell surface by treatment with pairs of antibodies against the gp130 ectodomain. This finding was combined with a novel approach in which the mobile fractions of gp130-CFP and Jak1-YFP were measured simultaneously. Indeed, immobilization of gp130-CFP (not shown) or gp130/id-CFP led to a concomitant immobilization of the associated Jak1-YFP and Jak1/kd-YFP. The co-immobilization of Jak1 is due to the interaction between gp130 and Jak1, because it is abrogated by deletion of the gp130 box1 motif, which is crucial for Jak binding (
). Thus, we conclude that Jak1 neither diffuses independently from gp130 nor does it exchange between different receptors to an extent measurable by our FRAP experiments. That Jaks faithfully stick to one receptor guarantees specificity of the signal so that Jaks only phosphorylate the receptor at which they were activated by ligand and not other receptors. The latter would be the consequence if Jaks could freely diffuse between receptors.
Similar approaches were used to analyze the interaction of proteins at erythrocyte membranes (
). In the latter study, the T-cell receptor ζ-chain (TCRζ) and the tyrosine kinase ZAP-70 were fused with GFP in order to investigate the recruitment of ZAP-70 to the phosphorylated immunoreceptor tyrosine-based activation motifs (ITAMs) of TCRζ. Comparison of mobilities of the transmembrane protein TCRζ-GFP and the membrane-recruited ZAP-70-GFP by FRAP indicated that the kinase moved about 20-fold faster than the receptor suggesting a rapid exchange of ZAP-70 molecules at the receptor. Thus, the interaction of the ZAP-70 SH2-domains with the phosphorylated ITAMs of TCRζ is of considerable lower affinity than the Jak/cytokine receptor interaction. This discrepancy makes sense in respect to the different biological functions of the interactions: The ZAP-70/TCRζ association is stimulation-dependent and therefore has to be transient to be sensitive to regulation by phosphorylation and dephosphorylation of the ITAMs. The Jak/cytokine receptor interaction, however, seems not to be regulated and is of constitutive nature.
Recent findings suggest that Jak binding to cytokine receptors occurs very early at ER or Golgi membranes and is a prerequisite for efficient cytokine receptor surface expression (
). Our studies suggest that these cytokine receptor/Jak complexes formed upon biosynthesis do not dissociate until proteolytic degradation of the proteins occurs. This view is supported by the finding that Jaks and gp130 have very similar protein turnover rates of about 3 h (
). Thus, cytokine receptor/Jak complexes have closest analogy to receptor tyrosine kinases.
We thank Dr. John Wijdenes (DIACLONE, Besançon, France) for the generous gift of the gp130 antibodies B-P4, B-P8, and B-S12-G7. We are most grateful to Dr. Neil Emans (Fraunhofer Institute for Molecular Biotechnology, Aachen, Germany) for his help in establishing live cell imaging at the confocal microscope. We would like to thank Dr. Ian Kerr (Cancer Research UK, London) for helpful discussions. We also thank Waraporn Komyod and Tanya Smyczek for their help in establishing stably Jak1-transfected cell lines and Hildegard Schmitz-van de Leur for excellent technical assistance.