An extracellular portion of granulocyte colony‐stimulating factor (G‐CSF) receptor, which contains an immunoglobulin‐like (Ig) domain and cytokine receptor homologous (CRH) region, was secreted into the medium using Trichoplusia ni‐Autographa californica nuclear polyhedrosis virus system. The gene product was purified to homogeneity mainly as a dimer (85 kDa) using G‐CSF affinity column chromatography and gel filtration HPLC, although the product existed as a monomer (45 kDa) in the medium. Scatchard analyses suggested that only the dimer had high affinity ligand binding ( K d = about 100 pM), which is comparable with the K d value of the cell surface receptor. The binding of G‐CSF to Ig‐CRH induced its tetramerization (200–250 kDa). The molecular composition of the tetrameric complex showed a stoichiometry of four ligands bound to four Ig‐CRH. These results suggested that the oligomeric mechanism of the G‐CSF receptor differs from that reported for growth hormone (GH) receptor, although CD spectrum spectroscopy suggested that the Ig‐CRH has a GH receptor‐like structure.
5-fluoro-2'-deoxyuridine (FUdR) inhibits thymidylate synthase. We have been investigated the molecular mechanisms of cell death in mouse mammary tumor FM3A cells, F28-7 strain and its mutant F28-7-A strain, after treated with FUdR. Previously, we have been reported that F28-7 strain induced DNA cleavage into chromosomal sized fragments and subsequently develop necrosis, but F28-7-A strain induced DNA cleavage into oligonucleosomal sized fragments and subsequently develop apoptosis after treated with FUdR. To understand the molecular mechanisms of regulate of two differential cell death necrosis and apoptosis, we identify cell death regulator by using proteome and transcriptome analysis. When compared with the proteome of F28-7 and F28-7-A strain after treated with FUdR, it was found that 5 proteins were up-regulated and 11 proteins were down-regulated in F28-7-A strain. Furthermore, transcriptome analysis shows that 94 genes were up-regulated and 164 genes were downregulated in F28-7-A strain. Identified proteins and genes were involved in various cellular processes such as cell cycle regulation, apoptosis, proliferation, and differentiation. Our results suggested that numerous features indicated the coordinated regulation of molecular networks from various aspects of necrosis or apoptosis at the proteome and transcriptome levels.
Many nucleoside analogs were screened for anti-protozoa activity on Leishmania tropica in an in vitro culture system. 3'-Deoxyinosine and several tubercidin derivatives were found to be potent inhibitors for growth of the promastigote form of L. tropica. EC50 value of 3'-deoxyinosine was 4.43 X 10(-7)M. This compound was remarkably less toxic towards mouse mammary tumor FM3A cells (EC50, 1.25 X 10(-4) M). 3'-Deoxyinosine is metabolized by Leishmania promastigote to give 3'-deoxyinosine-5'-monophosphate, 3'-deoxy-adenosine(cordycepin)-5'-mono, di-, and triphosphates. This means that Leishmania can aminate the 6-position of 3'-deoxyinosine-5'-monophosphate, thereby converting it into a highly toxic compound.
The amino-terminal domain of the cytokine receptor homologous region (BN domain; roughly 100 amino acid residues) in the receptor for murine granulocyte colony-stimulating factor (G-CSF) was secreted as a maltose-binding protein fusion into the Escherichia coli periplasm. The murine BN domain (mBN) was prepared from the fusion protein by restriction protease Factor Xa digestion and purified to homogeneity. The purified BN domain specifically and stoichiometrically bound G-CSF, with an apparent dissociation constant (Kd) of 3-8 x 10(-8) M. The CD spectrum of the mBN domain was similar to that of the extracellular region of the human growth hormone (GH) receptor, which is composed of turns and beta-sheets held together by disulfide bonds. Tertiary folding and the beta-sheet of this small domain was confirmed by NMR spectroscopy. Disulfide bonds determined by peptide mapping were in the following locations: Cys107-Cys118, Cys153-Cys162, and Cys143-Cys194. Among them, the first and the second produce small loops (roughly 10 amino acid residues) as found in the human GH receptor. These results suggested that the mBN domain of the G-CSF receptor expressed by E. coli has a GH receptor-like structure. However, the third disulfide bond varied considerably between the G-CSF and GH receptors. Disruption of these disulfide bonds in the BN domain of the G-CSF receptor suggested that all of them are critical for maintaining a stably folded protein. Our results will facilitate understanding of the biophysical and structural properties of this receptor.
The mechanism of cytotoxic action of 5-fluorodeoxyuridine (FdUrd) in mouse FM3A cells was investigated. We observed the FdUrd-induced imbalance of intracellular deoxyribonucleoside triphosphate (dNTP) pools and subsequent double strand breaks in mature DNA, accompanied by cell death. The imbalance of dNTP pools was maximal at 8 h after 1 microM FdUrd treatment; a depletion of dTTP and dGTP pools and an increase in the dATP pool were observed. The addition of FdUrd in culture medium induced strand breaks in DNA, giving rise to a 90 S peak by alkaline sucrose gradient sedimentation. The loss of cell viability and colony-forming ability occurred at about 10 h. DNA double strand breaks as measured by the neutral elution method were also observed in FdUrd-treated cells about 10 h after the addition. These results lead us to propose that DNA double strand breaks play an important role in the mechanism of FdUrd-mediated cell death. A comparison of the ratio of single and double strand breaks induced by FdUrd to that observed following radiation suggested that FdUrd produced double strand breaks exclusively. Cycloheximide inhibited both the production of DNA double strand breaks and the FdUrd-induced cell death. An activity that can induce DNA double strand breaks was detected in the lysate of FdUrd-treated FM3A cells but not in the untreated cells. This suggests that FdUrd induces the cellular DNA double strand breaking activity. The FdUrd-induced DNA strand breaks and cell death appear to occur in the S phase. Our results indicate that imbalance of the dNTP pools is a trigger for double strand DNA break and cell death.
The extracellular portion of the granulocyte colony-stimulating factor (G-CSF) receptor has a mosaic structure of six domains (each approximately 100 amino acid residues) consisting of an immunoglobulin-like (Ig) domain, a cytokine receptor homologous region subdivided into amino-terminal (BN) and carboxyl-terminal (BC) domains, and three fibronectin type III repeats. In the present study, we expressed the Ig-BN and the BN-BC regions and purified them to homogeneity as monomers using G-CSF affinity column chromatography. Using gel filtration high performance liquid chromatography, we investigated the molecular composition of receptor-ligand complexes formed between G-CSF and purified BN-BC or Ig-BN domains. In contrast to the well characterized example of the human growth hormone (GH) receptor, in which the BN-BC•GH complex shows a 2:1 receptor-ligand complex stoichiometry, the BN-BC domain of the G-CSF receptor formed a 1:1 complex. The isolated Ig-BN domain also formed a 1:1 complex with G-CSF. However, in the presence of both Ig-BN and BN-BC domains, we detected a 1:1:1 Ig-BN•G-CSF•BN-BC complex corresponding to the 2:1 receptor:ligand stoichiometry. These results suggest that 1) the Ig domain and both the BN and the BC domains are required for oligomerization of the G-CSF receptor, 2) G-CSF contains two binding sites for its receptor, and 3) there are two ligand binding sites on the G-CSF receptor, one site on the BN-BC domain and one on the Ig-BN domain. The extracellular portion of the granulocyte colony-stimulating factor (G-CSF) receptor has a mosaic structure of six domains (each approximately 100 amino acid residues) consisting of an immunoglobulin-like (Ig) domain, a cytokine receptor homologous region subdivided into amino-terminal (BN) and carboxyl-terminal (BC) domains, and three fibronectin type III repeats. In the present study, we expressed the Ig-BN and the BN-BC regions and purified them to homogeneity as monomers using G-CSF affinity column chromatography. Using gel filtration high performance liquid chromatography, we investigated the molecular composition of receptor-ligand complexes formed between G-CSF and purified BN-BC or Ig-BN domains. In contrast to the well characterized example of the human growth hormone (GH) receptor, in which the BN-BC•GH complex shows a 2:1 receptor-ligand complex stoichiometry, the BN-BC domain of the G-CSF receptor formed a 1:1 complex. The isolated Ig-BN domain also formed a 1:1 complex with G-CSF. However, in the presence of both Ig-BN and BN-BC domains, we detected a 1:1:1 Ig-BN•G-CSF•BN-BC complex corresponding to the 2:1 receptor:ligand stoichiometry. These results suggest that 1) the Ig domain and both the BN and the BC domains are required for oligomerization of the G-CSF receptor, 2) G-CSF contains two binding sites for its receptor, and 3) there are two ligand binding sites on the G-CSF receptor, one site on the BN-BC domain and one on the Ig-BN domain. INTRODUCTIONLigand-induced receptor oligomerization has been proposed as the key mechanism of signal transduction for some families of single transmembrane receptors, such as cytokine receptors and tyrosine kinase-type receptors (Ullrich and Schlessinger, 1990; Wells et al., 1993; Heldin, 1995). In these models, the oligomerization of the extracellular regions, induced by ligand binding, is followed by the activation of their cytoplasmic regions. The extracellular regions of these receptors generally have a composite structure containing multiple domains (Ullrich and Schlessinger, 1990; Bazan, 1990; Miyajima et al., 1992; Heldin, 1995), which are presumed to play important roles in ligand-induced oligomerization. Investigation of the molecular properties of purified extracellular domains is a prerequisite for understanding mechanisms of receptor oligomerization that culminate in the transductions of external signals.The extracellular region (˜600 amino acid residues) of the granulocyte colony-stimulating factor (G-CSF) 1The abbreviations used are: G-CSFgranulocyte colony-stimulating factorBN domainamino-terminal domain of the cytokine receptor homologous regionBC domaincarboxyl-terminal domain of the cytokine receptor homologous regionCRHcytokine receptor homologousGHgrowth hormoneHPLChigh performance liquid chromatographyIgimmunoglobulin-likePAGEpolyacrylamide gel electrophoresisChaps3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid. receptor is composed of an immunoglobulin-like (Ig) domain, a cytokine receptor homologous (CRH) region, and three fibronectin type III-like domains (Fukunaga et al., 1990b, 1990c). The Ig domain (˜100 amino acid residues) was originally defined by its homology to the immunoglobulin superfamily (Williams and Barclay, 1988). Though a number of receptors, including those for several cytokines, such as G-CSF, interleukin 6, and ciliary neurotrophic factor, contain Ig domains (Fukunaga et al., 1990b, 1990c; Larsen et al., 1990; Yamasaki et al., 1988: Davis et al., 1991), little is known about their function. A deletion derivate of the G-CSF receptor lacking the Ig domain retained ligand binding activity, although the dissociation constant (Kd) value of the mutant was 10-20-fold higher than that of the intact receptor (Fukunaga et al., 1991). The CRH region (˜200 amino acid residues) was originally defined by its striking homology to the predicted ligand binding domains of the receptors for various cytokines, such as interleukins 2-7, granulocyte-macrophage colony-stimulating factor, erythropoietin, growth hormone (GH), and prolactin (Bazan, 1990) receptors. The CRH region consists of an amino-terminal (BN; ˜100 amino acid residues) domain containing four conserved cysteine residues and a carboxyl-terminal (BC; ˜100 amino acid residues) domain containing a “WSXWS” motif (Bazan, 1990). Earlier work has identified high affinity oligomers of the G-CSF receptor on the surface of mouse myeloid leukemia cells (Fukunaga et al., 1990a), and mutational analyses showed that deletion of the BN domain completely abolishes ligand binding activity (Fukunaga et al., 1991). As a first step toward studying its molecular properties, we expressed the gene encoding the BN domain as a minimal binding unit, using an Escherichia coli maltose binding protein fusion system (Hiraoka et al., 1994a). This purified, small domain still retained ligand binding activity but did not form oligomers, such as dimers or tetramers (Hiraoka et al., 1994a), suggesting that a larger binding unit is required for G-CSF receptor oligomerization. As a second approach, we expressed a three-domain binding unit consisting of Ig-BN-BC regions (˜300 amino acid residues; indicated as Ig-CRH in Hiraoka et al. (1994b)) using an insect Trichoplusia ni cell-baculovirus Autographa californica nuclear polyhedrosis virus system. The purified Ig-BN-BC protein retained high ligand binding activity and formed dimers and tetramers in the presence of G-CSF (Hiraoka et al., 1994b). These studies indicated the involvement of the Ig or BC regions in receptor oligomerization and suggested that expression of tandem Ig-BN or BN-BC domains might permit functional dissection of the oligomerization process.In the present study, we describe expression of the Ig-BN and the BN-BC domains of the G-CSF receptor in insect cells using a baculovirus system. Isolated Ig-BN and BN-BC domains did not form homo-oligomers in the presence of ligand, although these products did form 1:1 binary complexes with G-CSF. However, the combined Ig-BN and the BN-BC domains formed a 1:1:1 Ig-BN•G-CSF•BN-BC complex, which is consistent with binding of a single molecule of G-CSF ligand by a dimeric receptor.EXPERIMENTAL PROCEDURESProduction of Recombinant Baculoviruses Carrying the Genes Encoding the Ig-BN and the BN-BC Domains of the G-CSF ReceptorThe Ig, BN, and BC domains contain 4, 6, and 4 cysteine residues, respectively (Fukunaga et al., 1990b; Seto et al., 1992). Since an insect baculovirus A. californica nuclear polyhedrosis virus secretion system has recently been developed for the efficient secretion of such cysteine-rich proteins, (Luckow and Summers, 1988; Ota et al., 1991; Hiraoka et al., 1994b), the Ig-BN and the BN-BC proteins (each ˜200 amino acids) were secreted using the recombinant baculoviruses AcIg-BN and AcBN-BC. DNA encoding the murine Ig-BN domain of the G-CSF receptor (Fig. 1A) and its signal sequence was generated by the polymerase chain reaction amplification of plasmid pBLJ17, which bears the cDNA for the murine G-CSF receptor (Fukunaga et al., 1990b). The 5′-sense primer had the sequence (primer 1) 5′-CGGGATCCATGGTAGGGCTGGGAGCCTG-3′, which contains a BamHI site (underlined) at the amino terminus, followed by the coding sequence for the murine G-CSF receptor starting from Met-25. The 3′-antisense primer has the sequence 5′-GCTCTAGATTATTTCACAACATCCATGGGGTCGAG-3′, which corresponds to the sequence for the carboxyl-terminal part of the Ig-BN domain, terminated at Lys202 and followed by a stop codon and an XbaI site (underlined). Thus, the resultant fragment contains the signal sequence (Met-25-Ser-1)and the Ig-BN domain (Cys1-Lys202) of the murine G-CSF receptor. The fragment was digested with BamHI and XbaI and was used for the construction of the transfer vector, pAcIg-BN.A DNA fragment corresponding to the BN-BC portion of the murine G-CSF receptor (Fig. 1A) and its signal sequence was generated from plasmid pBOSdIg, encoding a deletion mutant GΔ(5-84), which lacks Glu5-Ser84 of the Ig domain (Fukunaga et al., 1991). To mutate Cys1 to Ser1 in the residual portion of the Ig domain of the GΔ(5-84) mutant, two polymerase chain reaction fragments (fragment N and fragment C) were initially generated, using pBOSdIg as the template. Fragment N corresponds to the G-CSF receptor from the amino terminus to Val85, including the signal sequence and the deleted part of the Ig domain. Primer 1 above was used for the 5′-primer. The 3′-antisense primer has the sequence 5′-GGACGCCGATGTGTCCAGAGCTCTCCAGACTTCTGGGGAG-3′, which corresponds to the sequence from Leu-7 to Ile4, and the anti-codon Val85, in which the anticodon ACA for Cys1 was replaced with AGA (underlined) for Ser. Fragment C corresponds to the sequence from Leu-3 to Ala309 of the deletion mutant cDNA GΔ(5-84), followed by a stop codon. The 5′-primer has the sequence 5′-CTGGAGAGCTCTGGACACATCGGCGTCCAACTCCTG-3′, which contains the coding sequence from Leu-3 to Ile4 and from Val85 to Ala309, in which the codon TGT for Cys1 was replaced with TCT (underlined) for Ser. The 3′-antisense primer (primer 2) has the sequence 5′-GCTCTAGATTAGGCCTTCATGGTAGGCCTCA-3′, which corresponds to the sequence for the carboxyl-terminal part of the BN-BC domain, terminated at Ala309 and followed by a stop codon and an XbaI site (underlined). To amplify the full-length BN-BC DNA, small amounts of fragments N and C were used as the template for a second round of the polymerase chain reaction, using primer 1 as the 5′-primer and primer 2 as the 3′-primer, respectively. Thus, the resulting fragment contains the signal sequence (Met-25-Ser-1), part of the Ig domain (Ser-Gly2-His3-Ile4 and Val85-Gly96), and the BN-BC domain (Tyr97-Ala309) of the murine G-CSF receptor. After digestion with BamHI and XbaI, the resultant fragment was used for the transfer vector, pAcBN-BC.The transfer vectors pAcIg-BN and pAcBN-BC were obtained by the insertion of these amplified DNA fragments within the BamHI/XbaI sites of plasmid pVL1393 (Invitrogen), which carries part of the genome of the A. californica nuclear polyhedrosis virus. The BamHI/XbaI sites in pVL1393 are downstream of the polyhedrin promoter. The recombinant viruses AcIg-BN and AcBN-BC, carrying the Ig-BN and the BN-BC genes, were produced by in vivo homologous DNA recombination using the transfer vectors pAcIg-BN and pAcBN-BC, as described (Summers and Smith, 1988; Ota et al., 1991).Expression and Purification of the Ig-BN and the BN-BC Domains of the G-CSF ReceptorThe Ig-BN and the BN-BC domains were expressed using the recombinant viruses AcIg-BN and AcBN-BC. T. ni cells (Invitrogen; HIGH FIVE™ cells) were infected with recombinant viruses and were cultured for 9 days at 18°C. The culture fluids were applied to a G-CSF affinity column (1 cm, inner diameter, × 5 cm), and the Ig-BN and the BN-BC proteins were eluted by a glycine-HCl, pH 2.0, buffer. The affinity-purified proteins were then chromatographed by gel filtration high performance liquid chromatography (HPLC) (7.6 mm, inner diameter, x 60 cm; TSKgel G3000SW; TOSO Co., Ltd) as described (Hiraoka et al. 1994b). For further purification, the Ig-BN fraction was applied to a DEAE HPLC column (4.6 mm, inner diameter, × 25 cm; TSKgel DEAE-2SW; TOSO Co., Ltd) equilibrated with 20 mM sodium phosphate buffer, pH 6.0. Ig-BN was eluted with a linear gradient of NaCl from 0 to 0.4 M.G-CSF Binding to the Ig-BN and the BN-BC Domains of the G-CSF ReceptorThe assay used to quantitate G-CSF binding to the BN-BC domain of the G-CSF receptor was essentially that described by Fukunaga et al. (1990a). The BN-BC domain was mixed with various concentrations (10 pM to 1 nM) of 125I-labeled G-CSF (Amersham International plc) in 50 μl of phosphate-buffered saline containing 10% fetal calf serum and 0.1% Chaps in the presence or absence of 500 nM unlabeled G-CSF. The binding reaction proceeded for 90 min at room temperature. After the incubation, the BN-BC portion of the G-CSF receptor•G-CSF complex was precipitated with carrier protein (γ-globulin) and anti-M1 serum specific for the CRH region of the G-CSF receptor (Fukunaga et al., 1991) by 15% (w/v) polyethylene glycol 6000. The radioactivity in the precipitate was measured using a γ-counter (Packard, AUTO-GAMMA model A5002 COBRA). The 125I-G-CSF binding assays used to establish the complexes of the Ig-BN or the BN-BC domains of the G-CSF receptor with G-CSF were performed by means of gel filtration HPLC, as described (Hiraoka et al., 1994b).Production and Purification of the BN-BC Domain of the GH ReceptorThe BN-BC domain of the human GH receptor corresponds to its entire extracellular region (Bazan, 1990). Production and purification of the BN-BC domain of the GH receptor, using the insect baculovirus system, were performed according to the procedure described for the extracellular region of the human GH receptor (Ota et al., 1991). In the present study, the DNA sequence encoding Met-18-Leu233 of the human GH receptor (Leung et al., 1987) followed by a stop codon, TAA, comprised the BN-BC domain sequence of the GH receptor.RESULTSExpression and Purification of the Ig-BN ProteinThe purified Ig-BN protein migrated as a homogeneous 33-kDa band during polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS-PAGE) (Fig. 1B). Western blotting using anti-M1 serum specific for the CRH region of the G-CSF receptor (Fukunaga et al., 1991) identified the same 33-kDa band (Fig. 1B). Furthermore, the amino-terminal amino acid sequence of the purified 33-kDa protein ((Cys)-Gly-His-Ile-Glu-Ile-), identified by direct sequencing on an automated gas phase sequencer (Applied Biosystems 477A sequencer equipped with a model 120A PTH analyzer), corresponded precisely to that of the amino terminus of the G-CSF receptor, as deduced from the murine G-CSF receptor sequence (Fukunaga et al., 1990b; Hiraoka et al., 1994b). These results confirmed that the purified product was the Ig-BN protein. The position of the eluted peak at 35 kDa by gel-filtration HPLC indicated that the purified Ig-BN protein existed as a monomer. Approximately 50 μg of the purified Ig-BN protein was obtained per liter of culture.The molecular mass of the Ig-BN protein, identified by SDS-PAGE and gel-filtration HPLC, is larger than the molecular mass of 23 kDa deduced from the Ig-BN protein sequence (Cys1-Lys202) (Fukunaga et al., 1990b; Seto et al., 1992). The discrepancy between the observed and predicted molecular masses is most likely due to glycosylation. The Ig-BN protein has three potential Asn-linked glycosylation sites, which are conserved between the murine and human G-CSF receptors (Fukunaga et al., 1990b, 1990c). The 33-kDa band identified by Western blotting of the culture fluid was not detected when tunicamycin, a glycosylation inhibitor, was added to the culture medium after infection, supporting the proposed glycosylation of the Ig-BN protein. When Western blotting of the culture fluid was performed under nonreducing conditions, we noted that the 33-kDa band was less abundant and that most anti-M1 reactive material migrated at more than 100 kDa. Presumably, most native Ig-BN molecules are secreted into the culture fluid as large molecular weight aggregates due to the effects of random disulfide bond formation, and only the 33-kDa monomeric form, which bound to the G-CSF affinity column, was purified. The far UV CD spectrum of the Ig-BN protein (Fig. 2A) exhibited positive ellipticity at 230 nm and negative ellipticity around 210 nm. The extracellular region of the human GH receptor (Fig. 2D) (Bass et al., 1991) showed a very similar CD spectrum, suggesting that the purified Ig-BN protein has a GH receptor-like structure. These results are consistent with the spectra of the Ig-BN-BC region (Fig. 2C; Hiraoka et al., 1994b) and the BN domain (Hiraoka et al., 1994a).Figure 2Far-UV CD spectra of the Ig-BN, the BN-BC, and the Ig-BN-BC regions of the G-CSF receptor and the BN-BC portion of the GH receptor. The CD spectrum was measured on a Jasco J-720 spectropolarimeter (Japan Spectroscopic Co., Ltd.) at 20°C with solutions in 10 mM sodium phosphate buffer (pH 6.0) containing 0.093 mg/ml Ig-BN protein (A), 0.11 mg/ml BN-BC protein (B), and 0.12 mg/ml of Ig-BN-BC protein (C) (Hiraoka et al., 1994b) of the G-CSF receptor and 0.10 mg/ml of the BN-BC protein of the GH receptor (D; indicated as hGH-R; Ota et al., 1991). The optical path length was 1 mm for the far ultraviolet CD spectra. The mean residue ellipticity, [⊘]φ, has units of deg cm2 dmol-1. The purified protein concentrations were calculated from the absorption at 280 nm (A2800.1% values of 1.3, Ig-BN; 2.2, BN-BC; and 1.9, Ig-BN-BC domains of the G-CSF receptor, and 2.3, BN-BC portion of the GH receptor). This value was calculated using 1,576 M-1•cm-1 for tyrosine and 5,225 M-1•cm-1 for tryptophan at 280 nm (Goodwin and Morton, 1946).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Expression and Purification of the BN-BC ProteinTo purify the BN-BC protein, we used the recombinant baculovirus AcBN-BC, which was constructed from the plasmid pBOSdIg, harboring the mutated murine cDNA GΔ(5-84). GΔ(5-84) encodes the BN-BC portion but lacks most of the Ig domain. The GΔ(5-84) mutant receptor was shown to retain significant ligand binding activity when expressed on the surface of mammalian cells, despite the absence of an intact Ig region (Fukunaga et al., 1991). The GΔ(5-84) mutation removes all but 16 amino acid residues (Cys1-Ile4, Val85-Gly96) of the Ig domain. The Cys1-Ile4, Val85-Gly96 interval was postulated to contain the amino terminus of the adjacent BN-BC domain (Tyr97-Ala309), based on the exon-intron organization of the G-CSF receptor gene (Seto et al., 1992; Hiraoka et al., 1994b). In the construction of the recombinant baculovirus AcBN-BC, we replaced Cys1, which lies within the Cys1-Ile4, Val85-Gly96 interval, with Ser1 in the mutant GΔ(5-84) in order to prevent formation of disulfide bonds in the BN-BC protein. The purified BN-BC protein eluted at a position corresponding to a molecular mass of 27 kDa. The eluted material was analyzed by SDS-PAGE and revealed a nearly homogenous 32-kDa band (Fig. 1C). Western blotting identified the same 32-kDa band (Fig. 1C). These data strongly suggested that the purified product corresponds to a BN-BC protein monomer. The molecular masses of 32 kDa (SDS-PAGE) and 27 kDa (gel filtration) of the purified BN-BC protein appear larger than the molecular mass of 24 kDa deduced from the predicted BN-BC protein sequence (Ala95-Ala309) (Fukunaga et al., 1990b; Seto et al., 1992). However, the BN-BC protein has one potential Asn-linked glycosylation site (Fukunaga et al., 1990b, 1990c) conserved between the murine and the human cDNAs, suggesting, like the Ig-BN unit, that the BN-BC protein is glycosylated in insect cells. Consistent with this notion, tunicamycin treatment resulted in disappearance of the 32-kDa band on Western blots (data not shown). About 2.5 mg of purified BN-BC protein was obtained per liter of culture.A Scatchard analysis of the 125I-G-CSF binding data revealed that the BN-BC protein has a Kd of about 2.5 × 10-9M. The far UV CD spectrum of the BN-BC protein (Fig. 2B) was similar to those of the Ig-BN domain and the human GH receptor (Fig. 2D) (Bass et al., 1991), suggesting that the BN-BC protein also has a GH receptor-like structure. The amino-terminal sequence of the purified BN-BC protein was determined. The predominant amino-terminal residue was Ala, followed by the sequence Gly-Tyr-Pro-Pro. This amino-terminal residue corresponds to Ala95 of the Ser1-Ile4, Val85-Gly96 interval (see underlines: Ser1-Gly2-His3-Ile4 and Val85-Gln-Leu-Leu-Asp89-Gln-Ala-Glu-Leu-His-Ala95-Gly96) derived from the Ig region. A minor (less than 10%, as estimated by gel filtration HPLC), contaminating amino-terminal residue was Asp, followed by the sequence Gln-Ala-Glu-Leu. This amino-terminal residue corresponds to Asp89 of the 16 residual amino acids from the Ig domain. Thus, the expressed BN-BC protein contains the amino terminus expected of the BN-BC domain; however, most of the short Ig-derived sequence was removed during or after the processing of the signal peptide. As a control, we expressed the BN-BC domain from Tyr97 to Ala309 using an E. coli maltose binding protein fusion system, as described (Hiraoka et al. 1994a). The E. coli purified BN-BC domain exhibited almost the same Kd for G-CSF and the same CD spectrum as those of the purified BN-BC domain expressed using the baculovirus system. 2O. Hiraoka, H. Anaguchi, and Y. Ota, manuscript in preparation. These results suggested that the residual 16 amino acids of the Ig domain affect neither the Kd nor the conformation of the BN-BC domain.Formation of the BN-BC•G-CSF and the Ig-BN•G-CSF Binary Complexes and the Ig-BN•G-CSF•BN-BC Ternary ComplexTo analyze the effect of the ligand G-CSF upon oligomerization of purified BN-BC protein, the size of the purified BN-BC•G-CSF (27 and 19 kDa, respectively) complex was established by separating mixtures of the proteins (in ratios of BN-BC:G-CSF; 1:5, 1:1, 1:0.5, 1:0.1, and 1:0) by gel filtration HPLC (Fig. 3). At a 1:1 ratio of BN-BC to G-CSF, virtually all of the protein chromatographed at the position corresponding to a molecular mass of 35 kDa (Fig. 3, line 2, and Table 1). When the ratio of the BN-BC protein to G-CSF was 1:5, an excess of free G-CSF (19 kDa) appeared as a peak (Fig. 3, line 1). The 35-kDa peak did not dissociate, and no peak with a higher molecular mass was detected (Fig. 3, line 1). When the ratio was below 1:1 (ratios of 1:0.5, 1:0.1, and 1:0), a BN-BC (27 kDa) peak was present, whereas free G-CSF was absent (Fig. 3, lines 3-5). Gel filtration of the BN-BC•G-CSF complex using 125I-G-CSF confirmed that most of the 125I-G-CSF chromatographed at the position corresponding to 35 kDa at an extremely low concentration of the ligand (1:0.0004 ratio of BN-BC to G-CSF), whereas most of this 125I-G-CSF migrated at the position corresponding to 19 kDa in the presence of a high concentration of the unlabeled ligand (1:4 ratio of BN-BC to G-CSF) (data not shown). Thus, these data indicated that the BN-BC domain of the G-CSF receptor formed a stable 1:1 complex with G-CSF. The mass of the observed complex, 35 kDa, was somewhat smaller than the 46-kDa mass calculated from the sum of the components (Table 1). However, gel filtration HPLC analysis of the BN-BC domain expressed using an E. coli system also gave a smaller apparent mass for the BN-BC•G-CSF complex.2Figure 3Analysis of BN-BC•G-CSF complexes by gel filtration HPLC. The gel filtration profiles of various ratios of the BN-BC protein to G-CSF, corresponding to 1:5 (line 1), 1:1 (line 2), 1:0.5 (line 3), 1:0.1 (line 4), 1:0 (line 5) (top to bottom tracing) are shown. The protein concentrations of BN-BC (fixed at 1.6 μM) and G-CSF were calculated from the absorption at 280 nm (A2800.1% values of 2.2 and 0.81, respectively) as described in Fig. 2. Protein mixtures were equilibrated for 90 min at 20°C in 20 mM sodium phosphate buffer (pH 7.0) containing 0.2 M NaCl. Samples (200 μl) were applied to a TSKgel G3000SW gel filtration HPLC (7.6 mm, inner diameter, × 60 cm; TOSO Co., Ltd) and were eluted with the same buffer at 0.5 ml/min. Peaks were monitored for absorbance at 280 nm. The elution positions of the BN-BC portion of human GH receptor, the human GH, and their complexes estimated from Table 1 are indicated by arrows.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Tabled 1View Large Image Figure ViewerDownload Hi-res image Download (PPT) Open table in a new tab Fuh et al.(1990) expressed the BN-BC portion of the GH receptor as a soluble protein using an E. coli secretion system. The molecular composition of the BN-BC portion of the GH receptor•GH complex established by gel filtration HPLC showed a stoichiometry of 2:1 (Cunningham et al., 1991). As a control to test the accuracy of our gel filtration assessment of molecular mass, we also expressed the BN-BC portion of the GH receptor using the insect-baculovirus system and attempted to induce the formation of the 2:1 receptor•ligand complex. At a 1:0.5 ratio of the BN-BC portion of the GH receptor to GH, all of the protein chromatographed as a single peak, which corresponded to an 82-kDa complex composed of two molecules of the BN-BC portion of the GH receptor (30 kDa)/molecule of GH (22 kDa) (Fig. 4, line 3). When the ratio of the GH receptor to GH was decreased, a peak appeared corresponding to a 1:1 complex of 52 kDa, and the 82-kDa peak decreased in size (Fig. 4, lines 1 and 2). These results confirmed that a 2:1 GH receptor to GH complex was formed, and this complex dissociated to the 1:1 complex with an excess of GH, as described (Cunningham et al., 1991; de Vos et al., 1992; Fuh et al., 1992). These results also indicated that the stoichiometry of the receptor BN-BC•ligand complex differs between the G-CSF and the GH receptors, despite the structural similarity between their BN-BC domains.Figure 4Analysis of the BN-BC portion of the GH receptor-GH complexes by gel filtration HPLC. The gel filtration profiles of various ratios of the BN-BC portion of the GH receptor (indicated as hGH-R) to GH, corresponding to 1:5 (line 1), 1:1 (line 2), 1:0.5 (line 3), 1:0.1 (line 4), 1:0 (line 5) (top to bottom tracing) are shown. The protein concentrations of the BN-BC portion of the human GH receptor (fixed at 1.3 μM) and GH (recombinant human GH; Ota et al. 1991) were calculated from the absorption at 280 nm (A2800.1% values of 2.3 and 0.80, respectively) as described in Fig. 2. Protein mixtures were equilibrated for 90 min at 20°C in 20 mM sodium phosphate buffer (pH 7.0) containing 0.2 M NaCl. Samples (200 μl) were applied to the TSKgel G3000SW gel filtration HPLC and were eluted with the same buffer at 0.5 ml/min. Peaks were monitored for absorbance at 280 nm. The elution positions of bovine serum albumin (BSA; 67 kDa) and ovalbumin (ova; 43 kDa) are indicated by arrows.View Large Image Figure ViewerDownload Hi-res image Download (PPT)The Ig-BN (35-kDa) complex with G-CSF (19 kDa) was also established using gel filtration HPLC (Fig. 5). At a 1:1 ratio of Ig-BN to G-CSF, all of the protein chromatographed at 48 kDa (Fig. 5, line 2), corresponding to the size of a 1:1 complex (Table 1). When the ratio of Ig-BN to G-CSF was 1:10, peaks corresponding to the 1:1 complex, as well as to free G-CSF, were present, and no peaks with higher or lower molecular masses were detected (Fig. 5, line 3). These results show that the Ig-BN protein formed a stable 1:1 complex with G-CSF. When the Ig-BN protein was mixed with the BN-BC protein in addition to G-CSF, a peak of 74 kDa appeared, with a corresponding decrease in the 48-kDa peak (in ratios of Ig-BN:G-CSF:BN-BC; 1:1:0, 1:1:0.5, 1:1:3; Fig. 5, lines 2, 4, and 5). This result is consistent with the formation of a 1:1:1 Ig-BN•G-CSF•BN-BC complex, as the observed molecular mass of 74 kDa is quite close to the expected mass of 81 kDa for a 1:1:1 ternary complex (Table 1). Components of the ternary complex were confirmed by analysis of amino-terminal residues released from the 74-kDa peak (data not shown). No Ig-BN•BN-BC complex was formed in the absence of G-CSF (Fig. 5, line 6). The 35-kDa peak that would correspond to release of the Ig-BN domain was not detected by gel filtration HPLC, either upon addition of a small (1:1:0.5 ratio; Fig. 5, line 4) or an excess amount (1:1:3 ratio; Fig. 5, line 4) of the BN-BC domain. G-CSF was detected only in the 74-kDa peak seen upon addition of an excess amount of the BN-BC domain (1:1:3 ratio; Fig. 5, line 5; Western blotting; data not shown). These results suggested that the BN-BC protein did not compete with the Ig-BN protein for binding to G-CSF. The formation of the 74-kDa peak and the retention of the Ig-BN domain in the complex were confirmed by the addition of an excess amount of the BN-BC domain to the purified Ig-BN•G-CSF complex, using gel filtration HPLC (data not shown).Figure 5Analysis of Ig-BN•G-CSF and Ig-BN•G-CSF•BN-BC complexes by gel filtration HPLC. The gel filtration profiles of various ratios of the Ig-BN domain, G-CSF, and the BN-BC domain, corresponding to 1:0:0 (lane 1), 1:1:0 (lane 2), 1:10:0 (lane 3), 1:1:0.5 (lane 4), 1:1:3 (lane 5), and 1:0:1 (lane 6) (top to bottom tracing) are shown. The protein concentrations of the Ig-BN domain (fixed at 0.5 μM), the BN-BC domain, and G-CSF were calculated from the absorption at 280 nm (A2800.1% values of 1.3, 2.2, and 0.81, respectively) as described in Fig. 2. Protein mixtures were equilibrated for 90 min at 20°C in 20 mM sodium phosphate buffer (pH 7.0) containing 0.2 M NaCl. Samples (200 μl) were applied to the TSKgel G3000SW gel filtration HPLC and were eluted with the same buffer at 0.5 ml/min. Peaks were monitored for absorbance at 280 nm. The elution positions of the BN-BC portion of human GH receptor, the human GH, and their complexes estimated from Table 1 are indicated by arrows.View Large Image Figure ViewerDownload Hi-res image Download (PPT)DISCUSSIONThe BN-BC domain of the human GH receptor forms a 2:1 receptor•ligand complex with the GH ligand, with consequent dimerization of the GH receptor (Cunningham et al., 1991; de Vos et al., 1992). We confirmed the existence of this 2:1 complex using GH receptor produced by a baculovirus system. Similar 2:1 receptor-ligand complexes have also been obtained using the receptors for interferon-γ and prolactin (Greenlund et al., 1993; Hooper et al., 1993). Among the members of the cytokine receptor family, these receptors, as well as the G-CSF receptor, seem to function as homo-oligomers (Nagata and Fukunaga, 1991; Hiraoka et al. 1994b), supporting the view that the BN-BC domain of the G-CSF receptor may form a 2:1 receptor-ligand complex. Surprisingly, however, we found that the BN-BC domain of the G-CSF receptor forms a 1:1 binary complex with its ligand G-CSF, suggesting that the BN-BC domain of the G-CSF receptor alone is not sufficient for G-CSF-induced oligomerization. Notably, extracellular portions of the receptors for GH, interferon-γ, and prolactin compose only the BN-BC region (Bazan, 1990), while the G-CSF receptor contains Ig and fibronectin type III domains in addition to the BN-BC functional unit. In a previous study, we demonstrated that the purified Ig-BN-BC protein existed as dimeric or tetrameric forms in the presence of G-CSF (Hiraoka et al. 1994b). These results suggested the possibility that the Ig domain might play an important role in oligomerization of the G-CSF receptor.Cunningham et al. indicated that the GH receptor contains two overlapping binding sites that interact with two distinct sites on the GH protein, thereby producing the 2:1 receptor-ligand complex (Cunningham et al., 1991). In contrast, our results suggest that the BN-BC region of the G-CSF receptor contains a single binding site for G-CSF, while the Ig domain may provide a second ligand binding site. The purified Ig-BN protein formed a stable 1:1 binary complex, indicating that the Ig-BN region is also able to contribute to stable binding of G-CSF. Furthermore, we detected the formation of a 1:1:1 Ig-BN•G-CSF•BN-BC ternary complex upon addition of BN-BC protein to the 1:1 Ig-BN•G-CSF complex. These results suggested that two binding sites for G-CSF are found within the Ig-BN-BC region and that, unlike the GH receptor paradigm, two different functional units of the receptor molecule (both BN-BC and Ig-BN) are involved in interactions with two distinct binding sites on the G-CSF ligand.The apparent Kd of the BN-BC domain for G-CSF in the present study is about 2.5 × 10-9M. This value is similar to the Kd of about 3.7 × 10-9M exhibited by the deletion mutant of the G-CSF receptor expressed on the surface of a murine myeloid cells that contains both the BN and the BC domains but lacks the Ig domain (Fukunaga et al., 1991). We could not measure the Kd of the Ig-BN domain for G-CSF, because of the low purification yield and weak affinity of the anti-G-CSF receptor CRH antibody (anti-M1 serum; Fukunaga et al., 1991) for the Ig-BN protein. Fukunaga et al. reported a Kd value of about 1.1 × 10-8M for a mutant G-CSF receptor expressing the Ig and BN domains in addition to the fibronectin type III domains but lacking the BC domain at the cell surface (Fukunaga et al., 1991). Thus, it is likely that the affinity of the BN-BC domain for G-CSF is higher than that of the Ig-BN domain. Interestingly, in our study the Ig-BN domain was not released when the BN-BC domain was incubated with the Ig-BN•G-CSF complex. This finding is consistent with the idea that the binding site on the Ig-BN domain is distinct from that of the BN-BC domain. We have reported previously that the expressed Ig-BN-BC protein was purified primarily as a dimer, using G-CSF affinity column chromatography, and that only the dimeric Ig-BN-BC protein retained affinity for the ligand, termed “high” affinity (Kd = about 10-10M), while the monomeric Ig-BN-BC protein showed reduced affinity, which was termed “low” (Kd = about 2.5 × 10-9M) (Hiraoka et al., 1994b). These data supported the proposal that the Ig-BN-BC protein has two binding sites and suggested that the two sites provided by each dimeric Ig-BN-BC chain are required for high affinity binding of G-CSF. The apparent Kd of the purified BN-BC domain (about 2.5 × 10-9M) is nearly the same as that exhibited by the monomeric Ig-BN-BC protein and corresponds to low affinity ligand binding (Hiraoka et al., 1994b).In our previous work, we indicated that the Ig-BN-BC•G-CSF complex was composed of a dimeric Ig-BN-BC protein which forms at an extremely low concentration of G-CSF, and that this dimer converted to a tetramer with an increase in the concentration of ligand (Hiraoka et al., 1994b). The stoichiometry of the tetrameric complex was determined to be 4:4 Ig-BN-BC•G-CSF (Hiraoka et al., 1994b). We could not estimate the stoichiometry of the dimeric complex in that study because of the extremely low yield. In the present study, we did not obtain any receptor-ligand complexes containing four molecules of receptor (such as a 2:2:2 or 2:4:2 Ig-BN•G-CSF•BN-BC complexes, which would correspond to a 4:2 or 4:4 receptor-ligand stoichiometry), even at a high concentration of G-CSF (data not shown). These data suggest that a minimal Ig-BN-BC functional unit is required for the tetramerization. Fuh et al. expressed chimeric GH and G-CSF receptor cDNAs in a leukemia cell line and showed that the 2:1 GH receptor-GH complex, induced at a low concentration of GH, was the active form for signal transduction, whereas the 1:1 complex, induced at a high concentration of GH, was an inactive form (Fuh et al., 1992). Thus, it is likely that the receptor-ligand complex identified in the present study was “frozen” during purification as 2:1 G-CSF receptor-G-CSF stoichiometry, which appeared as a 1:1:1 Ig-BN-BC•G-CSF•BN-BC ternary complex (Fig. 6). The possibility remains that the G-CSF receptor exists on the cell surface as a dimer and that higher molecular weight aggregates, e.g. tetramers, constitute the activated form of the receptor induced by G-CSF stimulation (Fukunaga et al., 1990a; Hiraoka et al., 1994b). However, identification of a 1:1:1 Ig-BN•G-CSF•BN-BC ternary complex, formed by two discrete ligand binding sites in separate functional domains, suggests that the formation of a 2:1 G-CSF receptor-ligand complex may play an important role in the function of the G-CSF receptor, even if the dimerization of the G-CSF receptor is insufficient to activate the receptor. Interestingly, a mutant G-CSF receptor (GΔ(5-84)) containing the BN-BC and fibronectin type III domains but lacking the Ig domain still retains weak signal transduction activity (Fukunaga et al., 1991). Thus, if the activated form of the G-CSF receptor is indeed a dimer, it is likely that high concentrations of G-CSF can drive the GΔ(5-84) mutant receptor into a dimeric form, at least to some extent, even in the absence of the Ig domain. The mechanism and the meaning of the conversion from dimeric to tetrameric forms of the G-CSF receptor are still open questions. Further studies, including both mutational and structural analyses, are required to understand the precise mechanism of oligomerization and to correlate the observed oligomerization in solution to the mitogenicity of G-CSF.Figure 6Schematic representation of complexes of the Ig-BN-BC, the BN-BC, and the Ig-BN domains of the G-CSF receptor with G-CSF. A, each Ig-BN and BN-BC domain formed 1:1 binary complexes with G-CSF. These Ig-BN and BN-BC regions form a 1:1:1 ternary complex with G-CSF in the presence of both the Ig-BN and the BN-BC domains. B, thus, it is conceivable that the 2:1 Ig-BN-BC•G-CSF complex may be formed at a low concentration of G-CSF, and this 2:1 complex is converted to a 4:4 tetrameric complex with an increasing concentration of G-CSF (see “Discussion”; Hiraoka et al., 1994b).View Large Image Figure ViewerDownload Hi-res image Download (PPT) INTRODUCTIONLigand-induced receptor oligomerization has been proposed as the key mechanism of signal transduction for some families of single transmembrane receptors, such as cytokine receptors and tyrosine kinase-type receptors (Ullrich and Schlessinger, 1990; Wells et al., 1993; Heldin, 1995). In these models, the oligomerization of the extracellular regions, induced by ligand binding, is followed by the activation of their cytoplasmic regions. The extracellular regions of these receptors generally have a composite structure containing multiple domains (Ullrich and Schlessinger, 1990; Bazan, 1990; Miyajima et al., 1992; Heldin, 1995), which are presumed to play important roles in ligand-induced oligomerization. Investigation of the molecular properties of purified extracellular domains is a prerequisite for understanding mechanisms of receptor oligomerization that culminate in the transductions of external signals.The extracellular region (˜600 amino acid residues) of the granulocyte colony-stimulating factor (G-CSF) 1The abbreviations used are: G-CSFgranulocyte colony-stimulating factorBN domainamino-terminal domain of the cytokine receptor homologous regionBC domaincarboxyl-terminal domain of the cytokine receptor homologous regionCRHcytokine receptor homologousGHgrowth hormoneHPLChigh performance liquid chromatographyIgimmunoglobulin-likePAGEpolyacrylamide gel electrophoresisChaps3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid. receptor is composed of an immunoglobulin-like (Ig) domain, a cytokine receptor homologous (CRH) region, and three fibronectin type III-like domains (Fukunaga et al., 1990b, 1990c). The Ig domain (˜100 amino acid residues) was originally defined by its homology to the immunoglobulin superfamily (Williams and Barclay, 1988). Though a number of receptors, including those for several cytokines, such as G-CSF, interleukin 6, and ciliary neurotrophic factor, contain Ig domains (Fukunaga et al., 1990b, 1990c; Larsen et al., 1990; Yamasaki et al., 1988: Davis et al., 1991), little is known about their function. A deletion derivate of the G-CSF receptor lacking the Ig domain retained ligand binding activity, although the dissociation constant (Kd) value of the mutant was 10-20-fold higher than that of the intact receptor (Fukunaga et al., 1991). The CRH region (˜200 amino acid residues) was originally defined by its striking homology to the predicted ligand binding domains of the receptors for various cytokines, such as interleukins 2-7, granulocyte-macrophage colony-stimulating factor, erythropoietin, growth hormone (GH), and prolactin (Bazan, 1990) receptors. The CRH region consists of an amino-terminal (BN; ˜100 amino acid residues) domain containing four conserved cysteine residues and a carboxyl-terminal (BC; ˜100 amino acid residues) domain containing a “WSXWS” motif (Bazan, 1990). Earlier work has identified high affinity oligomers of the G-CSF receptor on the surface of mouse myeloid leukemia cells (Fukunaga et al., 1990a), and mutational analyses showed that deletion of the BN domain completely abolishes ligand binding activity (Fukunaga et al., 1991). As a first step toward studying its molecular properties, we expressed the gene encoding the BN domain as a minimal binding unit, using an Escherichia coli maltose binding protein fusion system (Hiraoka et al., 1994a). This purified, small domain still retained ligand binding activity but did not form oligomers, such as dimers or tetramers (Hiraoka et al., 1994a), suggesting that a larger binding unit is required for G-CSF receptor oligomerization. As a second approach, we expressed a three-domain binding unit consisting of Ig-BN-BC regions (˜300 amino acid residues; indicated as Ig-CRH in Hiraoka et al. (1994b)) using an insect Trichoplusia ni cell-baculovirus Autographa californica nuclear polyhedrosis virus system. The purified Ig-BN-BC protein retained high ligand binding activity and formed dimers and tetramers in the presence of G-CSF (Hiraoka et al., 1994b). These studies indicated the involvement of the Ig or BC regions in receptor oligomerization and suggested that expression of tandem Ig-BN or BN-BC domains might permit functional dissection of the oligomerization process.In the present study, we describe expression of the Ig-BN and the BN-BC domains of the G-CSF receptor in insect cells using a baculovirus system. Isolated Ig-BN and BN-BC domains did not form homo-oligomers in the presence of ligand, although these products did form 1:1 binary complexes with G-CSF. However, the combined Ig-BN and the BN-BC domains formed a 1:1:1 Ig-BN•G-CSF•BN-BC complex, which is consistent with binding of a single molecule of G-CSF ligand by a dimeric receptor.
