Y Box-Binding Factor Promotes Eosinophil Survival by Stabilizing Granulocyte-Macrophage Colony-Stimulating Factor mRNA1 Elizabeth E. Capowski,2 Ste´phane Esnault,2 Saswati Bhattacharya, and James S. Malter3 Short-lived peripheral blood eosinophils are recruited to the lungs of asthmatics after allergen challenge, where they become long-lived effector cells central to disease pathophysiology. GM-CSF is an important cytokine which promotes eosinophil differentiation, function, and survival after transit into the lung. In human eosinophils, GM-CSF production is controlled by regulated mRNA stability mediated by the 3ⴕ untranslated region, AU-rich elements (ARE). We identified human Y box-binding factor 1 (YB-1) as a GM-CSF mRNA ARE-specific binding protein that is capable of enhancing GM-CSF-dependent survival of eosinophils. Using a transfection system that mimics GM-CSF metabolism in eosinophils, we have shown that transduced YB-1 stabilized GM-CSF mRNA in an ARE-dependent mechanism, causing increased GM-CSF production and enhanced in vitro survival. RNA EMSAs indicate that YB-1 interacts with the GM-CSF mRNA through its 3ⴕ untranslated region ARE. In addition, endogenous GM-CSF mRNA coimmunoprecipitates with endogenous YB-1 protein in activated eosinophils but not resting cells. Thus, we propose a model whereby activation of eosinophils leads to YB-1 binding to and stabilization of GM-CSF mRNA, ultimately resulting in GM-CSF release and prolonged eosinophil survival. The Journal of Immunology, 2001, 167: 5970 –5976.
E
osinophils are a major participant in the inflammatory cascades associated with asthma and allergic disease (1). In particular, disease exacerbations are characterized by recruitment of short-lived, quiescent eosinophils from the peripheral blood to the lung where they become long-lived effector cells (2– 4). Inflammatory cytokines such as TNF-␣, IL-5, and GM-CSF have been implicated in both eosinophil recruitment and enhanced pulmonary survival (5–7). These cytokines are produced by airway epithelial cells, activated T cells, macrophages, and eosinophils themselves (8 –10). GM-CSF promotes eosinophil survival both in vivo and in vitro. Bronchoalveolar lavage fluid from symptomatic asthmatics contained elevated amounts of GM-CSF (5, 9) and the in vitro survival of both bronchoalveolar lavage eosinophils (BALeos)4 and activated peripheral blood eosinophils (pbeos) was GM-CSF but not IL-5 dependent (11, 12). GM-CSF production by eosinophils was stimulated by inflammatory cytokines such as TNF-␣ (13) or by fibronectin-mediated activation of integrin signaling pathways (14).
Department of Pathology and Laboratory Medicine, University of Wisconsin Medical School, Madison, WI 53792 Received for publication May 16, 2001. Accepted for publication September 17, 2001. 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. 1 This work was supported by National Institutes of Health Grant P50HL56396 (to J.S.M.). 2 E.E.C. and S.E. contributed equally to this work and should be considered co-first authors. 3 Address correspondence and reprint requests to Dr. James S. Malter, University of Wisconsin Hospital and Clinics, K4/812, Box 8550, 600 Highland Avenue, Madison, WI 53792. E-mail address:
[email protected] 4 Abbreviations used in this paper: BALeos, eosinophils isolated from bronchoalveolar lavage; UTR, untranslated region; ARE, AU-rich element; pbeos, peripheral blood eosinophils; PMGT, particle-mediated gene transfer; YB-1, Y box-binding protein 1; IP, immunoprecipitation; CSD, cold shock domain; ARM, arginine-rich RNA-binding motif; HA, hemagglutinin; -gal, -galactosidase.
Copyright © 2001 by The American Association of Immunologists
In pbeos, up-regulation of GM-CSF involves posttranscriptional control (15). GM-CSF mRNA is short-lived (half-life as short as 8 min) in resting cells but markedly stabilized after activation. As GM-CSF mRNA accumulates, increased amounts of cytokine are produced, directly enhancing survival (15). The half-life of the GM-CSF message is controlled by a 3⬘ untranslated region (UTR) AU-rich element (the ARE) (16). AREs are found in many rapidly degraded cytokine and protooncogene messages and mediate both rapid decay in quiescent cells and stability in activated cells (17). Recently, the ARE of TNF-␣ has also been implicated in nucleocytoplasmic transport (18). Several ARE-binding proteins have been identified that destabilize (AUF-1) or stabilize (HuR, TTP) ARE-containing messages (19 –22). However, the molecular control of GM-CSF mRNA turnover and its effects on the survival of eosinophils remains essentially unknown. Endogenous GM-CSF message levels even in activated pbeos are very low, making this a difficult process to study. However, we have developed a system that models GM-CSF metabolism in eosinophils and used it to show that normal eosinophils derived from the peripheral blood or airways of atopic subjects can be transfected with GM-CSF mRNA using particle-mediated gene transfer (PMGT; Refs. 12 and 15). Exogenous GM-CSF message decayed rapidly in resting pbeos but showed 2- to 4-fold increased stability in cells activated with calcium ionophore or with TNF-␣ plus fibronectin (12, 15). Of note, GM-CSF mRNA transfected into BALeos isolated from donors 48 h after allergen challenge had a prolonged half-life nearly identical to in vitro activated cells (12). Changes in GM-CSF mRNA stability required the presence of an intact ARE. Message stabilization enhanced GM-CSF production, causing a 2- to 10-fold increased in vitro survival. Therefore, we hypothesize that GM-CSF mRNA stabilization is a critical feature that maintains GM-CSF secretion and long-term in vivo survival of activated eosinophils. In this paper, we identify human Y box-binding protein 1 (YB-1) as a GM-CSF ARE-binding protein. Pbeos transduced with YB-1 showed enhanced, GM-CSF-dependent in vitro survival. In 0022-1767/01/$02.00
The Journal of Immunology resting pbeos, transduction of YB-1-stabilized exogenous GMCSF mRNA in an ARE-dependent fashion. Finally, endogenous YB-1 protein bound to endogenous GM-CSF mRNA in activated pbeos and BALeos but not in resting pbeos, suggesting this interaction is required for increasing GM-CSF secretion which contributes to maintaining long-lived, pulmonary-based eosinophils.
Materials and Methods Construction of fusion protein The coding region of human YB-1 was cloned from the K562 cell line (a human erythroleukemia cell line) by RT-PCR using primers complementary to the start and stop codons. The full-length coding region was cloned in-frame into pTatHA (Ref. 23; a gift from Dr. S. Dowdy, Washington University, St. Louis, MO) and a fusion protein containing from N to C terminus: 6-histidine, Tat translocation sequence, hemagglutinin (HA) epitope, and YB-1 was expressed in DH5␣ cells. Expressing bacterial cultures were dissolved in 8 M urea, 50 mM Tris (pH 8.0), 150 mM NaCl, sonicated, and centrifuged at 15,000 ⫻ g. Cleared lysate was mixed with Ni-NTA resin (Qiagen, Valencia, CA) in the presence of 10 mM imidazole and His-tagged protein was allowed to bind. The resin was washed once with 8 M urea, 50 mM Tris (pH 8.0), 150 mM NaCl, 10 mM imidazole, and four more times with buffer minus urea. Protein was eluted with 1 M imidazole, 50 mM Tris (pH 8.0), 150 mM NaCl, and dialyzed against PBS. Protein was ⬎90% pure by Coomassie-stained SDS-PAGE analysis and was active in binding GM-CSF RNA at concentrations as low as 2 ng/l.
RNA EMSAs In vitro transcription to produce radiolabeled and unlabeled full-length GM-CSF RNA was done as previously described (24 –26). Radiolabeled probes were quantitated by A260 and unlabeled probes were quantitated both by A260 and RNase-free agarose gel electrophoresis. RNA mobility shift assays were performed by mixing 300 ng of recombinant protein with 0.4 ng of radiolabeled RNA in 15 mM HEPES (pH 8.0), 10 mM KCl, 10% glycerol, 1 mM DTT, 1g/l tRNA, and 2 U/reaction recombinant RNasin (Promega, Madison, WI) for 10 min on ice. Protein was mixed with unlabeled competitor RNA for 15 min on ice before adding radiolabeled probe. Twenty-five units of RNase T1 was added and incubation was continued for 30 min at 37°C. Samples were UV cross-linked for 5 min on ice in a Stratalinker (Stratagene, La Jolla CA), loaded on 7% acrylamide gels, and electrophoresed in 0.25⫻ TBE at 100 V. Gels were dried and exposed to Kodak XAR film.
Western blotting SDS-PAGE gels (10%) were transferred to nitrocellulose using a TransBlot SD Semi-Dry Electrophoretic Transfer Cell (Bio-Rad, Richmond, CA) and membranes were blocked with 5% nonfat dry milk in phosphatebuffered Tris plus 0.1% Tween 20. Two anti-YB-1 Abs were used: antiYBC was a gift from Dr. K. Kohno (27) and anti-TatYB-1 is a rabbit polyclonal raised against recombinant TatYB-1 fusion protein (University of Wisconsin Medical School Animal Care Unit Polyclonal Antibody Service, Madison, WI) and purified with the ImmunoPure(G) IgG Purification kit (Pierce, Rockford, IL). Anti-rabbit HRP-conjugated secondary Ab and the ECL Western blotting detection system (Amersham, Pharmacia, Piscataway, NJ) were used to detect primary Abs, except in some cases antiTatYB-1 was biotinylated using the FluoReporter MiniBiotin-XX Protein Labeling kit and visualized with streptavidin-HRP conjugated secondary (Molecular Probes, Eugene, OR).
Tat protein transduction and in vitro survival pbeos were obtained with informed consent according to a protocol approved by the University of Wisconsin Human Subjects Committee from asymptomatic atopic adult donors and purified using a negative immunomagnetic procedure as previously described (28). Preparations were only used if the purity was ⬎99% (based on microscopic examination of DiffQuik-stained cells, Baxter, Miami, FL); the few contaminating cells were either neutrophils or mononuclear cells. Briefly, 106 cells/well were cultured in RPMI 1640 supplemented with 10% FBS and 50 g/ml gentamicin in 96-well plates. PBS with no protein or TatYB-1 protein or Tatgalactosidase (Tat-gal) protein in PBS was added at concentrations of 10 –200 nM at T0 and survival was assessed by trypan blue exclusion as previously described (12) 4 days after the initiation of culture. All experiments were done with a minimum of three different donors. In some cases, anti-GM-CSF or anti-IL-5-neutralizing Ab (R&D Systems, Minneapolis, MN) was added to a final concentration of 5 g/ml at the initiation of culture.
5971 Tat transduction and mRNA transfection PMGT was performed as described previously (15, 25). Briefly, in vitrotranscribed, capped, full-length, polyadenylated GM-CSF message was precipitated onto 1-m gold beads at a concentration of 5 g of RNA/mg of gold beads. Mutant GM-CSF mRNA containing AUGUA repeats in place of the ARE promotes in vitro survival more efficiently than wild type so to detect Tat fusion protein-mediated changes in viability; 5-fold less mutant mRNA was loaded onto gold beads (1 g of RNA/mg of gold beads). Both mutant and wild type were transfected at 5 g/mg concentrations for stability experiments (see Fig. 4 and Table I). Eighty to 95% of the input mRNA was typically loaded onto the beads. Successive transfections of 2 ⫻ 106 cells were pooled, washed twice in culture medium to remove any extracellular RNA, and placed in culture at 1 ⫻ 107 cells/ml. For transduction before transfection, 100 nM Tat fusion protein (TatYB-1 or Tat-gal) was added to 2 ⫻ 106 cells for 2 h at 37°C and then transfected as above. After washing, cells were placed in culture in medium containing 100 nM Tat protein. For RNA stability measurements, time points were taken at T0, 30, and 60 min, whereas for survival experiments, cells were cultured for 1 h in 100 nM protein at 1 ⫻ 107 cells/ml, then diluted 1:10, and cultured for 4 days. Viability was determined as previously described (12). Northern blot analysis, cDNA probes, and image analysis were described previously (15, 25). All experiments were done with pbeos from a minimum of three different donors.
Immunoprecipitation Pulmonary eosinophils (BALeos) were obtained with informed consent from atopic asthmatic patients 48 h after allergen bronchoprovocation by bronchoalveolar lavage and isolation as described elsewhere (12). pbeos were activated with 10 ng/ml recombinant human TNF-␣ (R&D Systems) and 20 g/ml soluble human cellular fibronectin (Sigma, St. Louis, MO) in fibronectin-coated 96-well plates (BD Biosciences, Bedford, MA) for 5 h at 1 ⫻ 106 cells/ml. BALeos, resting pbeos, or pbeos activated with TNF-␣ plus fibronectin (12) were snap frozen at ⫺80°C and cell pellets were dissolved in 50 mM Tris (pH 7.4), 150 mM NaCl, 1 mM EDTA, 10 mM NaF, 1 mM Na3VO4, 200 g/ml pefabloc, protease inhibitor mixture P8340 (Sigma), 1 mM DTT, 2 U of recombinant RNasin/l (Promega), 0.1% Triton X-100, and 0.1% SDS by passing them through a 29-gauge needle. Cleared lysate was made by centrifugation at ⬎15,000 ⫻ g for 10 min and 12 g of Ab was added for 2 h with rocking at 4°C. Protein G-agarose beads (Sigma) were added and incubation was continued overnight. Pellets were washed five times with 1 ml of lysis buffer (without detergent) and the last wash was split with 60% dissolved in TRI Reagent (Molecular Research Center, Cincinnati, OH) and 40% dissolved in SDSPAGE loading buffer. RNA was isolated according to the manufacturer’s recommendation. RT reactions were primed with random nonamers (Life Technologies, Rockville, MD) and PCR/Southern blotting was as described previously (12, 29), except for immunoprecipitation (IP) pellets from resting cells, 35 cycles of synthesis were used and for pellets from activated cells (TNF-␣ plus fibronectin treated or BALeos) 28 cycles of synthesis were used. Abs used for the immunoprecipitations were either anti-HA mouse monoclonal (Sigma) or anti-TatYB-1 rabbit polyclonal.
Results Recombinant human YB-1 fusion protein binds the GM-CSF ARE in vitro A truncated human cDNA coding for the multifunctional nucleic acid-binding protein YB-1 was identified in a screen for potential ARE-binding proteins (S. Bhattacharya and J. S. Malter, unpublished data). To determine whether YB-1 interacts with GM-CSF mRNA, we cloned the full-length coding sequence from the K562 cell line using primers complementary to the start and stop codons (see Materials and Methods). This full-length clone was fused to the 11-aa acid protein translocation sequence from HIV Tat protein in the vector pTat-HA. Proteins containing Tat tags cross lipid membranes with high efficiency (⬎95%), permitting transduction of normal as well as transformed cells (23). YB-1 fusion protein (TatYB-1) was used in in vitro RNA EMSAs to determine whether it bound GM-CSF RNA. Fig. 1A shows that TatYB-1 bound in vitro-synthesized, radiolabeled, full-length GM-CSF RNA. This binding is due to the YB-1 portion of the fusion protein since it can be blocked by preincubation with an interfering anti-YB-1 Ab but not with preimmune serum (data not shown). In addition, Tat-gal
5972
YB-1 BINDS GM-CSF mRNA IN EOSINOPHILS matically reduce YB-1-GM-CSF mRNA interactions (Fig. 1). This likely reflects the addition of unlabeled competitors RNAs before radiolabeled probe addition. However, neither mutant full-length nor ⌬2-AUGUA eliminated complex B (Fig. 1A, lanes 7–10 and Fig. 1B, lanes 1 and 3). Thus, complex B most likely represents a specific association between TatYB-1 and the GM-CSF ARE. The ⌬2-AUGUA RNA also failed to compete complex A while the full-length mutant GM-CSF RNA eliminated complex A in a dosedependent manner, suggesting that YB-1 recognized another site in the GM-CSF RNA in addition to the ARE. This second site is not sufficient for binding since radiolabeled full-length mutant GMCSF fails to bind to YB-1 even at high protein concentrations (data not shown). This site has not been further characterized. TatYB-1 protein transduction enhances the in vitro survival of resting pbeos
FIGURE 1. Recombinant YB-1 protein binds the 3⬘ UTR ARE of GMCSF RNA. RNA EMSAs showing specificity of binding of TatYB-1 fusion protein with radiolabeled full-length wild-type GM-CSF RNA probe. A, Competition with unlabeled full-length wild-type GM-CSF RNA, lanes 4 – 6, or unlabeled full-length mutant GM-CSF RNA which is identical to wild type except at the ARE (see text), lanes 8 –10. Molar excess of competitor is noted above the lanes. Free nucleotide and two RNA-protein complexes designated A and B are noted with arrowheads. B, Competition with no competitor (lane 1) unlabeled ⌬2 RNA at 10-fold molar excess (lane 2) or unlabeled ⌬2-AUGUA RNA at 10-fold molar excess (lane 3). The probe is full-length wild-type radiolabeled GM-CSF RNA as in A.
failed to bind GM-CSF mRNA (data not shown). Two distinct complexes were resolved on 7% native acrylamide gels at TatYB-1 concentrations ranging from 40 to 600 ng (Fig. 1 and data not shown). We used four types of unlabeled competitor RNAs to determine the specificity and site of binding. The first two were full-length wild-type GM-CSF RNA or a short 80-base synthetic RNA containing four consecutive AUUUA repeats (denoted ⌬2; Ref. 29). The latter RNA mimics the GM-CSF ARE which consists of five AUUUA repeats in an AU-rich region. The final two competitors used were full-length GM-CSF RNA containing a mutated ARE (four of the five AUUUA repeats mutated to AUGUA, which fails to bind ARE-specific proteins, (30)) or an 80-base synthetic RNA identical to ⌬2, except the four AUUUA repeats were mutated to AUGUA (denoted ⌬2-AUGUA; Ref. 29). Competition with either wild type or ⌬2 RNA eliminated both YB-1-GM-CSF RNA complexes (Fig. 1A, lanes 3– 6 and Fig. 1B, lanes 1 and 2). Only 1⫻ molar excess of unlabeled RNA was sufficient to dra-
YB-1 is constitutively present in most cell types including eosinophils as shown by Western blotting in Fig. 2. In general, cell lines express more YB-1 than untransformed primary cells (compare Fig. 2 lane 2 to lane 1 where equal amounts of total protein were loaded). YB-1 levels do not change in activated eosinophils compared with resting pbeos, although there is some variability in basal levels among individuals (data not shown). Cell lysates from resting or activated pbeos contained too many nucleases to perform EMSAs. Thus, to determine directly whether YB-1 affects GM-CSF mRNA metabolism in eosinophils, we transduced pbeos with TatYB-1 fusion protein and assayed in vitro survival. Survival was used as an end point because it is an exquisitely sensitive surrogate marker for increased GM-CSF production. For example, pbeos supernatant with GM-CSF levels below the limit of ELISA detection (⬍1 pg/ml) supported in vitro survival (11). The Tat protein translocation signal has been used successfully to deliver a number of proteins into many different cell types (23). pbeos were incubated with 100 nM TatYB-1 for 2 h and after extensive washing, cell lysates were analyzed by Western blotting. Fig. 2, lane 3, shows that TatYB-1 can be detected in transduced eosinophils. The transduction rate was ⬃95%, at concentrations from 50 to 100 nM and remained detectable by Western blot for at least 2 h after transduction (data not shown). We anticipated that if transduced TatYB-1 affected GM-CSF secretion, we would detect a change in in vitro survival of pbeos. Thus, pbeos were exposed to a range of TatYB-1 or ⌻at-gal concentrations and cultured for 4 days. As shown in Fig. 3A for a single donor, increasing concentrations of TatYB-1 increased survival by up to 3-fold while similar concentrations of Tat-gal had no effect. TatYB-1-mediated survival was abrogated by culturing transduced pbeos with neutralizing Ab to GM-CSF (Fig. 3) but unaffected by anti-IL-5 Ab (data not shown), indicating that TatYB-1 promotes survival by increasing extracellular levels of GM-CSF protein. The general trend of enhanced survival after TatYB-1 transduction was seen in multiple independent donors but the degree of
FIGURE 2. pbeos express YB-1 protein and can be transduced by TatYB-1 fusion protein. Western blots detected with anti-YB-1 Ab. Lane 1, 20 g of total cell lysate from AML14.3D10 an eosinophil-like cell line; lane 2, 20 g of total cell lysate from pbeos from an asthmatic patient 48 h after Ag challenge; lane 3, cell lysate from 2 ⫻ 106 pbeos transduced with 100 nM TatYB-1 for 2 h; lane 4, 100 ng of recombinant TatYB-1 fusion protein.
The Journal of Immunology
5973
FIGURE 4. TatYB-1 protein stabilizes exogenous GM-CSF mRNA in transduced pbeos. A, Northern blot of RNA extracted at T0, 30, or 60 min after GM-CSF mRNA was transfected into pbeos without (lanes 1–3) or after transduction with 100 nM TatYB-1 (lanes 4 – 6). Endogenous -actin mRNA did not decay within the time period assayed and served as a loading control. B, Plot of transfected GM-CSF mRNA decay rates for untransduced and 100 nM TatYB-1-transduced pbeos cultures normalized to -actin. Each point represents the average of three separate experiments.
FIGURE 3. pbeos transduced with TatYB-1 protein show increased GM-CSF dependent survival. A, Survival of pbeos after 4 days in culture without Tat protein (Control) or with 10 –100 nM TatYB-1 or 25–100 nM Tat-gal. Neutralizing Ab to GM-CSF was added to some TatYB-1-transduced cultures (o) at initiation of culture. Data shown are from a single donor. B, Survival after 4 days in culture of pbeos transfected with wildtype GM-CSF mRNA. Control, no protein transduction or mRNA transfection; GM, transfection with wild-type GM-CSF mRNA only, YB-1 ⫹ GM, transduction for 2 h with 100 nM TatYB-1 before transfection with wild-type GM-CSF mRNA, -gal ⫹ GM, transduction for 2 h with 100 nM Tat-gal before transfection with wild-type GM-CSF mRNA. o, Cultures that were treated with neutralizing GM-CSF Ab. Results shown are the average of three experiments with pbeos from three different donors. ⴱ, A significant difference was found for GM-CSF-transfected cells transduced without or with YB-1 (p ⬍ 0.029), as determined by Student’s t test.
survival advantage was variable. This may reflect differences in the initial levels of GM-CSF mRNA or other components of the signaling pathways contributing to cytokine production. To clarify whether YB-1 could regulate GM-CSF mRNA and how it effected eosinophil survival, we took advantage of a transfection system we have developed that mimics GM-CSF metabolism in eosinophils. We have shown previously that wild-type GM-CSF message trans-
fected into resting eosinophils by PMGT was rapidly degraded (Refs. 12 and 15; also see Fig. 4). Furthermore, the survival of pbeos transfected with wild-type or mutant GM-CSF mRNAs was increased by 3- to 10-fold and was dependent on extracellular GMCSF. Thus, we asked whether TatYB-1 could modulate in vitro survival mediated by transfected GM-CSF mRNA. As shown in Fig. 3B, resting pbeos showed minimal in vitro survival at 4 days which was increased ⬃6-fold by wild-type GM-CSF mRNA transfection. Survival was dependent on GM-CSF protein (Fig. 3B, o) but not IL-5 (data not shown). TatYB-1 transduction doubled the survival of pbeos transfected with GM-CSF mRNA whereas Tatgal had no effect (Fig. 3B). The transfection of consistent amounts of GM-CSF mRNA also minimized variability: Fig. 3B represents the average of three experiments, each done with pbeos from different donors. The effects of TatYB-1 were dependent on the presence of a functional ARE since we observed no difference in survival among pbeos transfected with mutant GM-CSF mRNA that were untransduced, transduced with TatYB-1, or transduced with Tat-gal (data not shown). Thus, enhanced production of GM-CSF leading to increased survival in vitro in transduced and transfected pbeos is mediated by TatYB-1 through the ARE of GM-CSF mRNA. TatYB-1 stabilizes GM-CSF mRNA in resting pbeos Since the ARE is required for stabilization or destabilization of GM-CSF mRNA, we examined how TatYB-1 affected the decay
5974
YB-1 BINDS GM-CSF mRNA IN EOSINOPHILS
rate of GM-CSF mRNA, once again taking advantage of our transfection system. Transfected wild-type GM-CSF mRNA has a short half-life in resting cells but is stabilized upon activation and can be visualized by Northern blot while endogenous levels even after stabilization are below the limits of detection (15). Thus, at various times after transduction with Tat fusion proteins and transfection with GM-CSF mRNA, total RNA was isolated and subjected to Northern blotting with GM-CSF or -actin cDNA probes. The Northern blot in Fig. 4A shows that transfected wild-type mRNA decays rapidly and is barely detectable by 60 min after transfection (lanes 1–3) but 60 min after transfection of TatYB-1-transduced cells, GM-CSF mRNA is still present (lanes 4 – 6). The half-life calculated for wild-type GM-CSF mRNA was increased 2-fold by TatYB-1 transduction (Fig. 4B and Table I). This is equivalent to the enhancement of stability seen in pbeos activated in vitro with TNF-␣ plus fibronectin or in BALeos activated in vivo (12). Table I also shows that GM-CSF mRNA stability was unaffected by Tat-gal, suggesting that the effect was specific to YB-1. TatYB1-mediated stabilization required an intact ARE since transfected mutant GM-CSF mRNA stability was unaffected. These data, along with the in vitro binding data, are consistent with the model that TatYB-1 stabilized GM-CSF mRNA in eosinophils by binding to the ARE. Endogenous YB-1 protein interacts with the GM-CSF message in activated eosinophils We wished to establish that endogenous YB-1 interacts with endogenous GM-CSF mRNA to address the possibility that transduced protein was playing a nonphysiologic role. In addition, it remained possible that the effects we saw were indirect; e.g., via protein-protein interactions with a different ARE-binding protein. Thus, we immunoprecipitated endogenous YB-1 from resting pbeos or those activated in vitro with TNF-␣ plus fibronectin or in vivo by allergen challenge (BALeos) and asked whether endogenous GM-CSF mRNA could be detected in the pellet. Fig. 5A, lanes 4 and 5, show that anti-YB-1 Ab immunoprecipitated endogenous YB-1 protein from pbeos and BALeos. Neither an irrelevant control IgG nor protein G alone precipitated detectable levels of endogenous YB-1 (Fig. 5A, lanes 2 and 3). The presence of GM-CSF mRNA in IP pellets was assayed by RT-PCR and Southern blot. Fig. 5B shows that in activated eosinophils, levels of GM-CSF mRNA were up to 4-fold above control when IP was done with anti-TatYB-1. Some GM-CSF mRNA bound nonspecifically to protein G-agarose used to pellet the Ab (data not shown) but very little GM-CSF mRNA was brought down with an irrelevant Ab (anti-HA mouse monoclonal: control IgG). -Actin mRNA was not detected in IP pellets either with control IgG or anti-YB-1 Ab nor did TNF-␣ mRNA coimmunoprecipitate with endogenous YB-1 from resting or from TNF-␣ plus fibronectinactivated pbeos (data not shown). Cell lysates from resting pbeos contained very low levels of GM-CSF mRNA and message in IP pellets was barely detectable by RT-PCR/Southern blotting. We present the data graphically in Fig. 5C as the ratio of GM-CSF
Table I. GM-CSF mRNA is stabilized by TatYB-1 in human pbeos Transfected mRNA
Wild type Wild type Wild type Mutant Mutant
Transduced Protein
None TatYB-1 Tat-gal None TatYB-1
Half-life (min)
17 34 15 35 26
FIGURE 5. GM-CSF mRNA interacts with YB-1 protein in activated eosinophils. A, Western blot detected with anti-YB-1 Ab. Lane 1, 10 g of cell lysate from AML14.3D10 eosinophil-like cell line; lanes 2– 4, IP pellet from resting pbeos cell lysate immunoprecipitated with no Ab (lane 2), control IgG (lane 3), anti-YB-1 Ab (lane 4). Lane 5, IP pellet from BALeos cell lysate immunoprecipitated with anti-YB-1 Ab. B, Southern blot of RT-PCR of IP pellets from in vitro-activated eosinophils (TNF ⫹ Fn) or in vivo-activated eosinophils (BALeos) probed with GM-CSF cDNA. Lysates were immunoprecipitated with either a control IgG (cIgG) or anti-YB-1 Ab (aYB1). C, Data from Southern blots of RT-PCR of IP pellets expressed as the average of the ratio of signal in anti-YB-1-precipitated pellets to signal in control IgG pellets.
signal detected from anti-YB-1 IPs to that detected in IPs with control IgG. Thus background levels of GM-CSF from nonspecific binding are set to 1. For both in vitro-and in vivo-activated cells, this ratio ⬎3, indicating an interaction between YB-1 and GMCSF mRNA. In resting pbeos, the ratio of 1 or less strongly suggests YB-1 and GM-CSF mRNA do not coassociate in the absence
The Journal of Immunology of an activation signal. Thus, endogenous YB-1 and endogenous GM-CSF mRNA interact specifically in activated eosinophils.
Discussion We have demonstrated that human YB-1 bound the ARE of GMCSF RNA in vitro (Fig. 1) and interacted with GM-CSF mRNA in activated eosinophils (Fig. 5). The result of binding was increased GM-CSF release and markedly enhanced eosinophil survival (Fig. 3). These studies describe a novel role for human YB-1 and help elucidate how activated eosinophils regulate GM-CSF production. YB-1 is a multifunctional nucleic acid-binding protein that has been implicated in such diverse functions as transcriptional activation and repression, both cellular and viral, specific and nonspecific mRNA binding, mRNA stabilization, and translational control (31–36). YB-1 is a member of the highly conserved cold shock domain (CSD) family of proteins and consists of three domains: an N-terminal stretch of undetermined function, a CSD which mediates nucleic acid binding and contains a consensus RNP-1 RNAbinding motif and a C-terminal domain of basic/acidic repeats which has been implicated in RNA binding as well as proteinprotein interactions (33, 34, 37). CSD proteins have been linked to GM-CSF transcriptional control in cell lines and epithelial cells (38, 39). However, recent CSD transcriptional control data suggest that nuclear translocation and subsequent transcriptional activity of CSD proteins requires cleavage of their COOH termini (32, 40). We have not detected smaller protein species either in resting or activated eosinophils (data not shown). It remains possible that there is some transcriptional contribution to increased GM-CSF production in these cells, especially when transduced since the TatYB-1 fusion protein requires no modification to cross the nuclear membrane. However, our data suggest that alterations in mRNA stability play a major role in GM-CSF production after activation of eosinophils (Fig. 4 and Refs. 12 and 15). YB-1 is also an RNA-binding protein. It has not previously been shown to bind ARE sequences but was shown to bind the 5⬘ UTR of IL-2 mRNA and to mediate increased message stability in T lymphocytes in response to activation by the c-Jun terminal kinase signaling pathway. Stabilization required uncharacterized 3⬘ UTR sequences which, however, were not themselves sufficient for stabilization (36). The IL-2 5⬘ RNA sequence identified as the YB-1 binding site is CU rich and without substantial homology to any part of the GM-CSF mRNA. In addition, c-Jun terminal kinase signaling does not appear to be up-regulated during eosinophil activation (41). Therefore, the mechanism of GM-CSF mRNA stabilization by YB-1 in eosinophils is likely to be distinct from that described for IL-2 in lymphocytes. In vitro, YB-1 can bind a number of ARE-containing cytokines including TNF-␣ (E. E. Capowski, unpublished data). However, endogenous TNF-␣ mRNA did not coimmunoprecipitate with YB-1 from eosinophils, suggesting that YB-1 mRNA interactions in vivo may be restricted to a subset of cytokine messages or indeed may be specific for GMCSF mRNA in pbeos. Five RNA recognition motifs have been identified in YB-1 as necessary for both its specific and nonspecific interactions with RNA (33, 34, 42). These include a consensus RNP-1 sequence in the CSD and four regions of alternating acidic and basic residues with homology to the arginine-rich RNA-binding motif (ARM) used by several RNA-binding proteins to recognize specific RNA hairpins (43). The RNP-1 sequence and the first ARM motif, located within the CSD, have been implicated in sequence-specific RNA recognition in the Xenopus YB-1 homologue FRGY-2, which also interacts nonspecifically with RNA through its COOH tail (34). ARM-1 and ARM-2 of human DNA-binding protein
5975 B/YB-1 have been implicated in specific RNA binding, possibly to stem-loop structures, in HIV TAR RNA and the insertion signal for the addition of the uncommon amino acid selenocysteine into cellular proteins (33, 42). All of the other ARE-binding proteins characterized to date also contain multiple RNA recognition motifs that define specificity in different contexts (for example, see HuR (44), HuD (45), TTP (46), or AUF-1 (47). HuR and HuD both have consensus RNP-1 motifs that are identical to each other and similar to that found in YB-1. However, AUF-1 and TTP show little commonality of mechanism at the amino acid sequence level with any of the other ARE-binding proteins. The majority of the RNA binding roles ascribed to YB-1 and CSD proteins are associated with translational control (34, 35, 42). We present data in Fig. 4 and Table I that strongly suggest that YB-1 has a role in posttranscriptional mRNA stability as well. Because our studies have been performed in primary cells of limited numbers, attempts to study GM-CSF translation have so far been impractical. The addition of YB-1 protein to rabbit reticulocyte lysate had no effect on GM-CSF mRNA translation (S. Esnault, unpublished data). However, the contribution of translational control to YB-1-mediated GM-CSF production remains to be elucidated. Fig. 5 shows that activation of pbeos is necessary to promote GM-CSF binding to YB-1, suggesting that YB-1 may be a target of signal transduction cascades initiated by extracellular proinflammatory factors. Interestingly, transcriptional activation of YB1/DNA-binding protein B has recently been linked to the activity of a protein tyrosine phosphatase (32), suggesting dephosphorylation might be involved in activating the protein. Our unphosphorylated bacterially expressed protein, which bound GM-CSF mRNA in vitro, was able to stimulate GM-CSF production and enhance survival when transduced into resting pbeos, providing support for such a mechanism. Ongoing studies should further elucidate the relationship among signal transduction, YB-1 phosphorylation state, and GM-CSF binding in eosinophils. Another possible control for YB-1-mediated RNA association is through a coassociated protein or complex. YB-1 partners a multitude of other proteins both for DNA and RNA binding, including the transcription factor CTCF, hnRNPK, FMRP, and nucleolin (36, 37, 48, 49). Unraveling the relevant control mechanisms in pbeos may well provide novel insight into the pathways that lead to inflammation and pathology.
Acknowledgments We thank Dr. S. Dowdy for the pTat-HA plasmid, Dr. K. Kohno for antiYBC Ab, Dr S. R. Soltaninassab for technical assistance, and members of the University of Wisconsin SCOR-asthma research group, particularly Julie B. Sedgwick for providing pbeos and Nizar N. Jarjour and E. A. B. Kelly for BALeos.
References 1. Bousquet, J., P. Chanez, J. Y. Lacoste, G. Barneon, N. Ghavanian, I. Enander, P. Venge, S. Ahlstedt, J. Simony-Lafontaine, P. Godard, and F. B. Michel. 1990. Eosinophilic inflammation in asthma. N. Engl. J. Med. 323:1033. 2. Bousquet, J., P. Chanez, J. Y. Lacoste, I. Enander, P. Venge, C. Peterson, S. Ahlstedt, F. B. Michel, and P. Godard. 1991. Indirect evidence of bronchial inflammation assessed by titration of inflammatory mediators in BAL fluid of patients with asthma. J. Allergy Clin. Immunol. 88:649. 3. Aalbers, R., H. F. Kauffman, B. Vrugt, G. H. Koeter, and J. G. deMonchy. 1993. Allergen-induced recruitment of inflammatory cells in lavage 3 and 24 hours after challenge in allergic asthmatic lungs. Chest 103:1178. 4. Sedgwick, J. B., W. J. Calhoun, R. F. Vrtis, M. E. Bates, P. K. McAllister, and W. W. Busse. 1992. Comparison of airway and blood eosinophil function after in vivo antigen challenge. J. Immunol. 149:3710. 5. Woolley, K. L., E. Adelroth, M. J. Woolley, R. Ellis, M. Jordana, and P. M. O’Byrne. 1995. Effects of allergen challenge on eosinophils, eosinophil cationic protein, and granulocyte-macrophage colony-stimulating factor in mild asthma. Am. J. Respir. Crit. Care Med. 151:1915.
5976 6. Yamaguchi, Y., T. Suda, S. Ohta, K. Tominaga, Y. Miura, and T. Kasahara. 1991. Analysis of the survival of mature human eosinophils: interleukin-5 prevents apoptosis in mature human eosinophils. Blood 78:2542. 7. Virchow, J. C., C. Walker, D. Hafner, C. Kortsik, P. Werner, H. Matthys, and C. Kroegel. 1995. T cells and cytokines in bronchoalveolar lavage fluid after segmental allergen provocation in atopic asthma. Am. J. Respir. Crit. Care Med. 151:960. 8. Cromwell, O., Q. Hamid, C. J. Corrigan, J. Barkans, Q. Meng, P. D. Collins, and A. B. Kay. 1992. Expression and generation of interleukin-8, IL-6 and granulocyte-macrophage colony-stimulating factor by bronchial epithelial cells and enhancement by IL-1 and tumour necrosis factor-␣. Immunology 77:330. 9. Broide, D. H., and G. S. Firestein. 1991. Endobronchial allergen challenge in asthma: demonstration of cellular source of granulocyte macrophage colonystimulating factor by in situ hybridization. J. Clin. Invest. 88:1048. 10. Broide, D. H., M. M. Paine, and G. S. Firestein. 1992. Eosinophils express interleukin 5 and granulocyte macrophage-colony-stimulating factor mRNA at sites of allergic inflammation in asthmatics. J. Clin. Invest. 90:1414. 11. Esnault, S., and J. S. Malter. 2001. Minute quantities of GM-CSF prolong eosinophil survival. J. Interferon Cytokine Res. 21:117. 12. Esnault, S., and J. S. Malter. 2001. GM-CSF mRNA is stabilized in airway eosinophils and peripheral blood eosinophils activated by TNF-␣ plus fibronectin. J. Immunol. 166:4658. 13. Levi-Schaffer, F., V. Temkin, V. Malamud, S. Field, and Y. Zilberman. 1998. Mast cells enhance eosinophil survival in vitro: role of TNF-␣ and granulocytemacrophage colony-stimulating factor. J. Immunol. 160:5554. 14. Meerschaert, J., R. F. Vrtis, Y. Shikama, J. B. Sedgwick, W. W. Busse, and D. F. Mosher. 1999. Engagement of ␣47 integrins by monoclonal antibodies or ligands enhances survival of human eosinophils in vitro. J. Immunol. 163:6217. 15. Esnault, S., and J. S. Malter. 1999. Primary peripheral blood eosinophils rapidly degrade transfected granulocyte-macrophage colony-stimulating factor mRNA. J. Immunol. 163:5228. 16. Shaw, G., and R. Kamen. 1986. A conserved AU sequence from the 3⬘ untranslated region of GM-CSF mRNA mediates selective mRNA degradation. Cell 46:659. 17. Chen, C. Y., and A. B. Shyu. 1995. AU-rich elements: characterization and importance in mRNA degradation. Trends Biochem. Sci. 20:465. 18. Dumitru, C. D., J. D. Ceci, C. Tsatsanis, D. Kontoyiannis, K. Stamatakis, J.-H. Lin, C. Patriotis, N. A. Jenkins, N. G. Copeland, G. Kollias, and P. N. Tsichlis. 2000. TNF-␣ induction by LPS is regulated posttranscriptionally via a Tpl2/ ERK-dependent pathway. Cell 103:1071. 19. Brewer, G. 1991. An A ⫹ U-rich element RNA-binding factor regulates c-myc mRNA stability in vitro. Mol. Cell. Biol. 11:2460. 20. Carballo, E., W. S. Lai, and P. J. Blackshear. 1998. Feedback inhibition of macrophage tumor necrosis factor-␣ production by tristetraprolin. Science 281:1001. 21. Fan, X. C., and J. A. Steitz. 1998. Overexpression of HuR, a nuclear-cytoplasmic shuttling protein, increases the in vivo stability of ARE-containing mRNAs. EMBO J. 17:3448. 22. Rodriquez-Pascual, F., M. Hausding, I. Ihrig-Biedert, H. Furneaux, A. P. Levy, U. Forstermann, and H. Kleinert. 2000. Complex contribution of the 3⬘-untranslated region to the expressional regulation of the human inducible nitric-oxide synthase gene. Involvement of the RNA-binding protein HuR. J. Biol. Chem. 275:26040. 23. Vocero-Akbani, A., N. A. Lissy, and S. F. Dowdy. 2000. Transduction of fulllength tat fusion proteins directly into mammalian cells: analysis of T cell receptor activation-induced cell death. Methods Enzymol. 322:508. 24. Ruth, J. H., S. Esnault, J. A. Jarzembowski, and J. S. Malter. 1999. Calcium ionophore upregulation of AUUUA-specific binding protein activity is contemporaneous with granulocyte macrophage-colony stimulating factor messenger RNA stabilization in AML14.3D10 cells. Am. J. Respir. Cell Mol. Biol. 21:621. 25. Rajagopalan, L. E., and J. S. Malter. 1996. Turnover and translation of in vitro synthesized messenger RNAs in transfected, normal cells. J. Biol. Chem. 271: 19871. 26. Malter, J. S. 1989. Identification of an AUUUA-specific messenger RNA binding protein. Science 246:664. 27. Ohga, T., K. Koike, M. Ono, Y. Makino, Y. Itagaki, M. Tanimoto, M. Kuwano, and K. Kohno. 1996. Role of the human Y box-binding protein YB-1 in cellular sensitivity to the DNA-damaging agents cisplatin, mitomycin C and ultraviolet light. Cancer Res. 56:4224. 28. Hansel, T. T., I. J. M. deVries, T. Iff, S. Rihs, M. Wandzilak, S. Betz, K. Blaser, and C. Walker. 1991. An improved immunomagnetic procedure for the isolation of highly purified human blood eosinophils. J. Immunol. Methods 145:105.
YB-1 BINDS GM-CSF mRNA IN EOSINOPHILS 29. Esnault, S., J. Jarzembowski, and J. S. Malter. 1998. Stabilization of granulocytemacrophage colony-stimulating factor RNA in a human eosinophil-like cell line requires the AUUUA motifs. Proc. Assoc. Am. Physicians 110:575. 30. Rajagopalan, L. E., J. K. Burkholder, J. Turner, J. Culp, N. S. Yang, and J. S. Malter. 1995. Granulocyte-macrophage colony-stimulating factor mRNA stabilization enhances transgenic expression in normal cells and tissues. Blood 86:2551. 31. Didier, D. K., J. Schiffenbauer, S. L. Woulfe, M. Zaceis, and B. D. Schwartz. 1988. Characterization of the cDNA encoding a protein binding to the major histocompatibility complex class II Y box. Proc. Natl. Acad. Sci. USA 85:7322. 32. Stenina, O. I., E. J. Poptic, and P. E. DiCorleto. 2000. Thrombin activates a Y box-binding protein (DNA-binding protein B) in endothelial cells. J. Clin. Invest. 106:579. 33. Ansari, S. A., M. Safak, G. L. Gallia, B. E. Sawaya, S. Amini, and K. Khalili. 1999. Interaction of YB-1 with human immunodeficiency virus type 1 Tat and TAR RNA modulates viral promoter activity. J. Gen. Virol. 80:2629. 34. Bouvett, P., K. Matsumoto, and A. P. Wolffe. 1995. Sequence-specific RNA recognition by the Xenopus Y-box proteins: an essential role for the cold shock domain. J. Biol. Chem. 47:28297. 35. Ranjan, M., S. R. Tafuri, and A. P. Wolffe. 1993. Masking mRNA from translation in somatic cells. Genes Dev. 7:1725. 36. Chen, C-Y., R. Gherzi, J. S. Andersen, G. Gaietta, K. Ju¨ rchott, H.-D. Royer, M. Mann, and M. Karin. 2000. Nucleolin and YB-1 are required for JNK-mediated interleukin-2 mRNA stabilization during T-cell activation. Genes Dev. 14: 1236. 37. Shnyreva, M., D. S. Schullery, H. Suzuki, Y. Higaki, and K. Bomsztyk. 2000. Interaction of two multifunctional proteins: heterogeneous nuclear ribonucleoprotein K and Y-box-binding protein. J. Biol. Chem. 275:15498. 38. Coles, L. S., P. Diamond, F. Occhiodoro, M. A.Vadas, and M. F. Shannon. 1996. Cold shock domain proteins repress transcription from the GM-CSF promoter. Nucleic Acids Res. 24:2311. 39. Coles, L. S., P. Diamond, F. Occhidoro, M. A. Vadas, and M. F. Shannon. 2000. An ordered array of cold shock domain repressor elements across tumor necrosis factor-responsive elements of the granulocyte-macrophage colony-stimulating factor promoter. J. Biol. Chem. 275:14482. 40. Koike, K., T. Uchiumi, T. Ohga, S. Toh, M. Wada, K. Kohno, and M. Kuwano. 1997. Nuclear translocation of the Y-box binding protein by ultraviolet irradiation. FEBS Lett. 417:390. 41. Kampen, G. T., S. Stafford, T. Adachi, T. Jinquan, S. Quan, J. A. Grant, P. S. Skov, L. Poulsen and R. Alam. 2000. Eotaxin induces degranulation and chemotaxis of eosinophils through the activation of ERK2 and p38 mitogenactivated protein kinases. Blood 95:1911. 42. Shen, Q., R. Wu, J. L. Leonard, and P. E. Newburger. 1998. Identification and molecular cloning of a human selenocysteine insertion sequence-binding protein, a bifunctional role for DNA-binding protein B. J. Biol. Chem. 273:5443. 43. Tan, R., and A. D. Frankel. 1995. Structural variety of arginine-rich RNA-binding peptides. Proc. Natl. Acad. Sci. USA 92:5282. 44. Ma, W-J., S. Cheng, C. Campbell, A. Wright, and H. Furneaux. 1996. Cloning and characterization of HuR, a ubiquitously expressed Elav-like protein. J. Biol. Chem. 271:8144. 45. Park, S., D. G. Myszka, M. Yu, S. J. Littler, and I. A. Laird-Offringa. 2000. HuD RNA recognition motifs play distinct roles in the formation of stable complex with AU-rich RNA. Mol. Cell. Biol. 20:4765. 46. Lai, W. S., E. Carballo, J. M. Thorn, E. A. Kennington, and P. J. Blackshear. 2000. Interactions of CCCH zinc finger proteins with mRNA, binding of tristetraprolin-related zinc finger proteins to AU-rich elements and destabilization of mRNA. J. Biol. Chem. 275:17827. 47. Zhang, W., B. J. Wagner, K. Ehrenman, A. W. Schaefer, C. T. DeMaria, D. Crater, K. DeHaven, L. Long, and G. Brewer. 1993. Purification, characterization, and cDNA cloning of an AU-rich element RNA-binding protein, AUF1. Mol. Cell. Biol. 13:7652. 48. Chernukhin, I. V., S. Shamsuddin, A. F. Robinson, A. F. Carnes, A. Paul, A. I. El-Kady, V. L. Lobanenkov, and E. M. Klenova. 2000. Physical and functional interaction between two pluripotent proteins, the Y-box DNA/RNA-binding factor, YB-1, and the multivalent zinc finger factor, CTCF. J. Biol. Chem. 275:29915. 49. Ceman, S., R. Nelson, and S. T. Warren. 2000. Identification of mouse YB1/p50 as a component of the FMRP-associated mRNP particle. Biochem. Biophys. Res. Commun. 279:904.
