The cytoplasmic domain of the human type I IFN receptor chain 2 (IFNAR2c or IFN-αRβL) was used as bait in a yeast two-hybrid system to identify novel proteins interacting with this region of the receptor. We report here a specific interaction between the cytoplasmic domain of IFN-αRβL and a previously identified protein, RACK-1 (receptor for activated C kinase). Using GST fusion proteins encoding different regions of the cytoplasmic domain of IFN-αRβL, the minimum site for RACK-1 binding was mapped to aa 300–346. RACK-1 binding to IFN-αRβL did not require the first 91 aa of RACK-1, which includes two WD domains, WD1 and WD2. The interaction between RACK-1 and IFN-αRβL, but not the human IFN receptor chain 1 (IFNAR1 or IFN-αRα), was also detected in human Daudi cells by coimmunoprecipitation. RACK-1 was shown to be constitutively associated with IFN-αRβL, and this association was not effected by stimulation of Daudi cells with type I IFNs (IFN-β1b). RACK-1 itself did not become tyrosine phosphorylated upon stimulation of Daudi cells with IFN-β1b. However, stimulation of cells with either IFN-β1b or PMA did result in an increase in detectable immunofluorescence and intracellular redistribution of RACK-1.
Type I IFNs induce a variety of cellular responses, including antiviral, antiproliferative, and immunomodulatory effects (1, 2). The mechanism by which IFNs induce such effects is known to require type I IFN binding to its cell surface receptor, which initiates receptor-mediated signaling events leading to gene induction. In recent years the role of STATs in IFN signaling has been well documented (3, 4). In the case of type I IFN, ligand-dependent IFN-αβ receptor heterodimerization leads to activation of two members of the Janus kinase family, Tyk2 and JAK-1, that catalyze receptor tyrosine phosphorylation. All STAT proteins contain Src homology 2 (SH2)4 domains that specifically recognize and bind to phosphorylated tyrosine residues located within the cytoplasmic domain of cytokine receptor chains (4, 5). After recruitment to the receptor complex, STAT proteins become phosphorylated on a specific C-terminal tyrosine, homo- or heterodimerize, translocate to the nucleus, and induce transcription of specific genes. In addition to tyrosine phosphorylation, serine/threonine phosphorylation has been shown to be important in mediating signaling, regulation of biological responses, and STAT activity (6).
It is clear that a number of important protein-protein interactions involve the cytoplasmic domain of various receptors upon ligand binding. The precise regulation of these interactions leads to the formation of activated cytoplasmic transcription factors that are critical for mediating gene activation and subsequent biological effects. The regulation of such interactions is likely to involve proteins with different functional properties such as kinases and phosphatases (7), STATs (8), adapter proteins (9), chaparonins (10), and nuclear translocation factors (11).
RACKs (receptor for activated C kinases) (12, 13) are members of a new group of proteins that may function as adaptors in different systems. It has been suggested that RACKs may be responsible for the intracellular localization of activated protein kinases C (PKCs). RACK-1 specifically associates with the C2 domain of PKCβ (12, 13), while another RACK, the coantomer protein β (βCOP) protein, associates with PKCε (14). A common feature of RACK-1 and conantomer protein β (βCOP) is the presence of tryptophan and aspartic acid (WD) repeats within these proteins, which are thought to mediate protein-protein interactions. RACK-1 has a molecular mass of 36,000 and is composed of seven WD repeats, thus resembling the structure of the β subunit of G proteins (Gβ) (15). The Gβs, and very likely RACK-1, form a rigid seven-blade β-propeller structure that has been proposed to function as an adapter or scaffold motif (15). Recently, RACK-1 has been reported to associate with other proteins, including Src kinase where it inhibits kinase function (16). Also, RACK-1 appears to interact with β integrin subunit cytoplasmic domain, directly linking RACK-1 to integrin function (17). In addition, RACK-1 has been shown to interact with the cAMP-specific phosphodiesterase PDE4D5 isoform, where it is proposed to recruit other proteins to a signaling complex (18).
Using the entire cytoplasmic domain of the long form of IFN-αRβL as bait in a yeast two-hybrid system screen, we identified RACK-1 as an IFN receptor-associated protein. RACK-1 specifically binds to the cytoplasmic domain of IFN-αRβL in a ligand-independent manner as demonstrated by coimmunoprecipitation and GST pull-down experiments. Moreover, treatment of cells with type I IFN results in an increase in detectable RACK-1, suggesting an active role of this WD repeat-containing protein in IFN signaling.
Materials and Methods
Cell lines and reagents
All cell lines were purchased from American Type Tissue Culture (Manassas, VA) and grown at 37°C in 5% CO2. U-266 and Daudi cells were grown in RPMI 1640 (Life Technologies, Gaithersburg, MD) containing 10% (v/v) bovine calf serum, l-glutamine, 5% penicillin, and streptomycin (Life Technologies) and were harvested at 1 × 106 cell/ml as previously described (19). Human dermal microvascular endothelial cells (HMEC) were obtained from Dr. Ades (Centers for Disease Control, Atlanta, GA). The cells were grown to confluence in T75 flasks in MCDB 131 medium containing 0.7% FBS, 10 mM glutamine, 0.5 mg/ml hydrocortisone, 10 ng/ml epidermal growth factor, 100 U/ml penicillin, and 100 mg/ml streptomycin (20), and then subcultured on 12-mm diameter no. 1 glass coverslips in 24-well plates for immunostaining. Human IFN-β1b (sp. act., 2.5 × 107 U/mg) was produced as previously described (21), and IFN-α2 (sp. act., 3.0 × 108 U/mg) was purchased from PeproTech (Rocky Hill, NJ). Anti-phosphotyrosine Abs Ab-2 and 4G10 were purchased from Oncogen Sciences (Cambridge, MA) and Upstate Biotechnologies (Lake Placid, NY). IFNAR1 (IFN-αRα) and IFNAR2c (IFN-αRβL) antisera were prepared as previously described (22). RACK-1 mAb was purchased from Transduction Laboratories (Lexington, KY).
Yeast two-hybrid screen
The IFN-αRβL cytoplasmic domain cDNA (aa 281–515) was generated by PCR, subcloned into the Gal4 DNA binding domain plasmid pAS2-1, and used as bait in a yeast two-hybrid screen of a human B lymphocyte cDNA library constructed in pACT (Clontech, Palo Alto, CA). Approximately 1.6 × 106 colonies were screened for activation of HIS3 and lacZ reporter genes using the host strain CG1945. Primary screen positives were isolated by cycloheximide counterselection against the bait and retested against the IFN-αRβL cytoplasmic domain (pβL281–515) and two other unrelated bait plasmids pLAM 5′-1 (encodes DNA binding domain fusion with human lamin C) and pVA3-1 (encodes DNA binding domain fusion with murine p53). The inserts from positive library plasmids were then amplified by PCR and mapped by AluI digests, and plasmids were sequenced after isolation and bacterial transformation.
Immunoprecipitation and immunoblotting
U-266 or Daudi cells (1 × 107 cells) were lysed in lysis buffer (20 mM Tris-HCl, pH 6.6, containing 1% Nonidet P-40, 50 mM NaCl, 1 mM EDTA, 2.5% glycerol (v/v), 1.0 mM sodium fluoride, 1.0 mM sodium orthovanadate, 1.0 mM PMSF, 0.5 μg/ml leupeptin, and 5.0 μg/ml trypsin inhibitor) for 30 min at 4°C. For immunoprecipitation, the indicated Abs were added to each sample, incubated overnight, mixed with protein G-agarose (Roche, Indianapolis, IN), and resolved by SDS-PAGE (10% gels, Novex, San Diego, CA). Proteins were transferred to polyvinylidene difluoride membranes and then sequentially incubated with an anti-RACK-1 mAb (Transduction Laboratories) and anti-mouse IgM μ-chain coupled to HRP (Pierce, Rockford, IL) for 1 h. Filters were washed three times in blocking buffer and developed using a chemiluminescent detection method (Pierce).
GST fusion constructs
The different GST fusion proteins encoding different regions of the cytoplasmic domain of IFN-αRβL have been described previously (23). Pull-down experiments and immunoblotting were performed using the same procedure described above for immunoprecipitations.
Confluent HMEC on glass coverslips were washed twice and incubated for 2 h at 37°C in serum-free and phenol red-free DMEM containing 25 mM HEPES, pH 7.4. The cells were then stimulated at 37°C for 0–60 min with agonists (20,000 U/ml IFN-β1b or 100 nM PMA) before fixation for 20 min in 4% paraformaldehyde in HBSS supplemented with 25 mM HEPES (HBSS). The coverslips were washed three times for 5 min each time in 100 mM glycine and three times for 10 min each time in HBSS to quench and remove the fixative. The cells were then preincubated in HBSS containing 5% goat serum, 0.2% BSA, 0.01% NaN3, and 0.1% Triton X-100 (blocking solution) for 30 min at room temperature, followed by incubation with anti-RACK-1 mouse monoclonal IgM overnight at 4°C in the same buffer. After washing three times for 10 min each time with HBSS, the coverslips were again preincubated in blocking solution for 30 min before exposure to Alexa 568 goat anti-mouse IgG F(ab′)2 conjugate (Molecular Probes, Eugene, OR) for 2 h at room temperature. Primary and secondary Abs were centrifuged at 14,000 × g for 10 min through Spin-X filters (Amicon, Beverly, MA) and used at concentrations of 1 and 4 μg/ml, respectively. The coverslips were finally washed three times for 10 min each time in HBSS and mounted on a drop of ProLong Antifade mounting medium (Molecular Probes).
The intracellular localization of labeled RACK-1 was assessed in reference to nuclear 4′6′-diamidine-2-phenylindole dihydrochloride (DAPI) staining using a Zeiss 510 laser scanning confocal microscope (New York, NY). One-micron optical sections were scanned individually with 364- and 568-nm laser lines from Enterprise UV (Coherent, Palo Alto, CA) and OmniChrome Ar/Kr lasers (Melles Griot, Carlsbad, CA). The emission filters used to detect DAPI and Alexa 568 were BP385–470 nm and LP590 nm, respectively. Thus, nuclear DAPI staining appeared blue and RACK-1 red. Four frame averages of the brightest 1-μm thick focal plane observed with the ×63 1.4 NA water immersion objective were acquired and later assembled using Adobe Photoshop and Macromedia Freehand image processing software (Adobe Systems, Mountainview, CA).
Yeast two-hybrid screen
A human B cell cDNA library was screened for cDNAs encoding proteins that interact with the cytoplasmic domain of IFN-αRβL. Among a number of positive yeast clones, two different clones, C1-L and C2-N1, encoded the C-terminal 216 aa of the previously described receptor for activated C kinase, RACK-1 (Fig. 1⇓). The C1-L and C2-N1 clones strongly bound the cytoplasmic domain of IFN-αRβL, suggesting that amino acids 91–317 contain the interaction site for IFN-αRβL binding. The first 91 aa of RACK-1 include two WD repeats (WD1 and WD2) that do not appear to be required for IFN-αRβL binding. The interaction between RACK-1 and IFN-αRβL detected in the two-hybrid system was shown to be specific, because both the pACT-C1-L and pACT-C2-N1 clones failed to produce β-galactosidase-positive colonies when either was reintroduced with one of two different control plasmids (pLAM 5′-1 or pVA3-1) in CG1945 cells (Fig. 1⇓). Cotransformation of the control (pLAM 5′-1 or pVA3-1) and bait plasmids (pACT-C1-L or pACT-C2-N1) was confirmed by the observed growth of transformed yeast in tryptophan- and leucine-deficient media (Fig. 1⇓).
Mapping the site in IFN-αRβL required for binding RACK-1
To confirm the interaction between IFN-αRβL and RACK-1, we performed “pull-down” experiments. GST constructs containing the entire cytoplasmic domain of IFN-αRβL bound RACK-1 in IFN-stimulated or unstimulated cell lysates (Fig. 2⇓A). No RACK-1 was observed bound to GST alone. We next sought to determine which region of the cytoplasmic domain of IFN-αRβL was responsible for the interaction with RACK-1. Fig. 2⇓B shows that RACK-1 bound to GST-βL truncated at aa 462, 375, and 346, but not to a GST-βL containing only aa 265–299 or GST control. Therefore, the binding domain within IFN-αRβL required for RACK-1 binding was shown to be between aa 300 and 346 (Fig. 2⇓B). No differences in the interaction between RACK-1 and IFN-αRβL were observed after type I IFN treatment (Fig. 2⇓A). It is worth mentioning that truncations proximal to aa 375 bound less RACK-1 than truncation at aa 462 or full-length IFN-αRβL, raising the possibility that there may be more than one RACK-1 binding site on IFN-αRβL. These data strongly suggest that RACK-1 is constitutively associated with the cytoplasmic domain of IFN-αRβL and probably binds via a mechanism independent of ligand-induced receptor phosphorylation.
Coimmunoprecipation of RACK-1 with IFN-αRβL
To determine whether the interaction between RACK-1 and IFN-αRβ can be detected in vivo, we performed coimmunoprecipitation experiments. RACK-1 could be detected directly in Daudi cell lysates by immunoblotting as a 36-kDa protein (Fig. 3⇓, lane 1). This molecular mass corresponds to the previously reported m.w. of RACK-1 (12, 13). A protein that comigrated with RACK-1 could be shown to coimmunoprecipitate with IFN-αRβL in IFN-β1b-stimulated and unstimulated Daudi cells (Fig. 3⇓, lanes 5 and 6). We estimate that the fraction of cellular RACK-1 that coimmunoprecipitates with IFN-αRβL is within the range of 1–2%. No RACK-1 coimmunoprecipitation was observed using a nonspecific Ab as a negative control (Fig. 3⇓, lane 2) or anti-IFN-αRα Abs before or after IFN-β1b stimulation (Fig. 3⇓, lanes 3 and 4). However, it should be pointed out that we cannot completely rule out the presence of low affinity interactions between RACK-1 and IFN-αRα that cannot be detected by coimmunoprecipitation.
Because tyrosine phosphorylation is an important event in IFN signaling, we next investigated whether RACK-1 was tyrosine phosphorylated after IFN-β1b stimulation. Fig. 4⇓ shows that RACK-1 was not tyrosine phosphorylated either before (lanes 2 and 4) or after (lanes 1 and 3) stimulation of Daudi cells with IFN-β1b, demonstrating that RACK-1 is not regulated by this mechanism.
Intracellular localization of RACK-1 after type I IFN stimulation
HMEC were used to localize the distribution of RACK-1 within the cell before and after treatment with either PMA (Fig. 5⇓, c, e, and g) or IFN-β (Fig. 5⇓, b, d, and f). In unstimulated cells (Fig. 5⇓a) RACK-1 was detected throughout the cytoplasm. Interestingly, upon treatment of HMEC with IFN-β1b, a dramatic increase in RACK-1 immunostaining became evident (Fig. 5⇓, c, e, and g). This increase was maximal at 60 min after stimulation with IFN (Fig. 5⇓g) and was similar to the increase in immunostaining induced by PMA (Fig. 5⇓, b, d, and f). Treatment with IFN-β1b and PMA also appeared to induce intracellular redistribution of RACK-1 toward the perinuclear area (Fig. 5⇓, d and e). This observation is consistent with the idea that RACK-1 distribution within the cell is influenced by ligand-induced activation of the type I IFN receptor. However, it is not clear whether IFN stimulation of cells results in the release of sequestered RACK-1 or induces a conformational change in the protein resulting in the observed increase in immunostaining. Furthermore, it is important to note that the increase in RACK-1 immunostaining is observable within 15 min after stimulation of cells with IFN-β1b, suggesting that this increase is probably not dependent on new protein synthesis.
IFNs have been implicated in regulating important functions, such as resistance to virus infection, control of cell proliferation, and modulation of the immune response to infection. Therefore, responses to IFNs may be expected to be mediated by a diverse set of intracellular signals. In support of this concept, a variety of both early and late stage signaling events occur in response to IFN treatment of cells (1, 2). The most studied intracellular signaling pathway used by IFNs is that of STAT proteins (3, 4). In addition, a number of other intracellular mediators involved in IFN-responsive gene induction have also be been suggested and these include activation of MAPK (24), CrkL (25), phosphoinositol 3-kinase (26), PKC (27), and even Lyn and ZAP70 in human T cells (28). However, additional efforts will be needed to precisely determine the roles of these signaling proteins in the antiviral and antiproliferative effects of type I IFNs.
In this study we found an additional protein, RACK-1, (12, 13), associated with a region of the cytoplasmic domain of IFN-αRβL that was previously reported in murine cells to be required for signaling by type I IFNs (29). This association is not dependent on ligand-induced assembly of the receptor or tyrosine phosphorylation (30). RACK-1 is therefore constitutively associated with IFN-αRβL. In addition, the minimum binding site of RACK-1 on IFN-αRβL was mapped using GST-βL constructs to aa 300–346, close to the proposed JAK1 binding site.
RACK-1 was originally identified and cloned from a rat brain cDNA expression library as a protein that specifically bound activated Ca2+/phospholipid-dependent PKCβ (12). Among other functions, RACK-1 is believed to be a member of a family of receptors for PKC, where it is proposed to direct and localize activated PKCs to their respective sites of action (31). RACK-1 is a 317-aa-long protein containing seven WD motifs present in the cytoplasm of a variety of cells. The WD motif has previously been identified as playing a role in mediating protein-protein interactions and is defined as a motif present in intracellular adaptor proteins (32). The WD repeats in RACK-1 are conserved from Chlamydomonas ssp. to humans, suggesting an important function through evolution (32).
Recent observations have demonstrated that overexpression of RACK-1 in NIH-3T3 cells suppresses cell proliferation and demonstrated the direct interaction between RACK-1 and the SH2 domain of Src (16). Such an interaction suggests the possibility that RACK-1 may act not only as a receptor for activated PKC, but also as an adapter protein for proteins containing SH2 domains. Furthermore, it has been shown that RACK-1 binds directly to the integrin β subunit and that this interaction is dependent on the prior treatment of cells with phorbol esters such as PMA (12, 13). Such an interaction links RACK-1 directly to integrins and perhaps regulation of integrin function. In addition, RACK-1 was recently demonstrated to associate with the cAMP-dependent phosphodiesterase isoform, PDE4D5 (18), in a PMA-independent manner, where it is proposed to bring regulatory proteins to the enzyme complex. This further supports the suggestion that RACK-1 may play a more general role in intracellular signaling by mediating important protein-protein interactions. Therefore, RACK-1 is likely to play a role in regulating signaling different from that simply involving PKC (31). In light of this, RACK-1/PKC interactions appear to first require the activation of PKC, by PMA, before its association with RACK-1; however, interaction of RACK-1 with IFN-αRβL, Src kinase, and integrin β subunit is constitutive and does not appear to require any activation step. Therefore, different mechanisms of RACK-1 interactions exist. However, both PMA and IFN stimulation of HMECs can similarly influence RACK-1 levels, in that both treatments resulted in a dramatic increase in the levels of RACK-1 within the cell as measured by immunofluorescence. Such a rapid increase in immunofluorescent signal for RACK-1 suggests a change in conformation of RACK-1 after type I IFN treatment rather than an increase in steady state level.
It is interesting to note that RACK-1 is constitutively associated with IFN-αRβL in the absence of ligand. Such an observation suggests that the role of RACK-1 in type I IFN signaling is different from that resulting from the direct activation of PKC by type I IFNs themselves. However, it is possible that PKC activated by other receptors may sequester, via RACK-1, to the IFN receptor. Such a process may lead to sensitive cross-talk between different receptors and stimuli.
Currently, there are only a few reports suggesting a role for PKC (33) and serine/threonine phosphorylation (34) in cytokine signaling and STAT regulation (6). However, recent efforts have demonstrated a direct effect of IFN-β in reversing the rapid nuclear translocation of protein kinase C-δ in human T cells (35). Such an effect demonstrates a link between type I IFNs and upstream signaling events surrounding T cell apoptosis (35). It is possible, however, that RACK-1 is acting not only as a receptor for activated PKC but also as a general adaptor protein, helping to bring proteins involved in signaling to the receptor (36). In this role, RACK-1 may function as an adapter protein recognizing other kinases or transcription factors with SH2 domain motifs, as suggested by the observation that RACK-1 binds to the SH2 domain of Src (16). Because many adapter proteins interact with several different proteins (36), it is also tempting to speculate that the association of RACK-1 with IFN-αRβL is not unique to the type I IFN receptor and that RACK-1 may link other cytokine receptors with different downstream signaling molecules. Indeed, recent evidence has demonstrated an interaction of RACK-1 with another cytokine receptor subunit, the common β-chain of the IL-5/IL-3 and GM-CSF receptor (37). Finally, the involvement of PKC-δ in the regulation of STAT3 function has been recently linked to IL-6 stimulation. In this case, PKC-δ was shown to bind to and phosphorylate STAT3 on Ser727 (38).
We thank T. Charis Wagner, Dean Russell-Harde, and Dr. John Parkinson for helpful discussions.
↵1 This work was supported by National Institutes of Health Grants GM54709 and CA55079 (to O.C.).
↵2 The laboratories of E.C. and O.C. contributed equally to this work.
↵3 Address correspondence and reprint requests to Dr. Ed Croze, Department of Immunology, Berlex Biosciences, 15049 San Pablo Avenue, Richmond, CA 94806. E-mail address:
↵4 Abbreviations used in this paper: SH2, Src homology 2; HMEC, human dermal microvascular endothelial cells; IFNAR1, human type I IFN receptor chain 1; IFNAR2c, full-length human type I IFN receptor chain 2; RACK-1, receptor for activated protein kinase C; PKC, protein kinase C; DAPI, 4′6′- diamidine-2-phenylindole dihydrocholoride.
- Received March 24, 2000.
- Accepted August 9, 2000.
- Copyright © 2000 by The American Association of Immunologists