Solution structure of an extracellular domain containing the WSxWS motif of the granulocyte colony-stimulating factor receptor and its interaction with ligand.

We have determined the NMR structure of a ligand-binding domain of the granulocyte colony-stimulating factor (G-CSF) receptor, containing the highly conserved WSxWS motif. The domain consists of seven beta-strands with the fibronectin type III-like topology seen in several cytokine receptors. Comparisons between the spectra of the 15N-labelled domain with and without G-CSF indicate that the major ligand-recognition site is on the FG loop just upstream of the WSxWS sequence, and not on the BC loop which is mainly used in the growth hormone system. The WSxWS residues are suggested to contribute to ligand-recognition and to the protein architecture of the G-CSF receptor.

Formation of 1:1 complex of the cytokine receptor homologous region of granulocyte colony-stimulating factor receptor with ligand.

The cytokine receptor homologous (CRH) region of the murine granulocyte colony-stimulating factor (G-CSF) receptor was secreted using a Escherichia coli maltose binding protein (MBP) fusion system. The CRH region was prepared from the periplasmic fraction by G-CSF affinity column chromatography and restriction protease factor Xa digestion, and was purified to homogeneity. The purified CRH region specifically bound G-CSF, with an apparent dissociation constant (kd) of about 1.5 x 10(-9) M. A 1:1 CRH.G-CSF complex was established by gel-filtration high pressure liquid chromatography (HPLC). However, a 2:1 stoichiometric complex was not established, as in the case of the growth hormone (GH) receptor [Recent Prog. Hormone Res., 48, 233-275 (1993)].

Ligand binding characteristics of the carboxyl-terminal domain of the cytokine receptor homologous region of the granulocyte colony-stimulating factor receptor.

The carboxyl-terminal domain (BC domain, roughly 100 amino acid residues) of the cytokine receptor homologous region in the receptor for murine granulocyte colony-stimulating factor was secreted as a maltose binding protein fusion into the Escherichia coli periplasm. The murine BC domain was prepared from the fusion protein by restriction protease factor Xa digestion and was purified to homogeneity. The purified BC domain specifically and stoichiometrically bound granulocyte colony-stimulating factor. This result indicates that the BC domain is also critical for ligand binding, as shown for the amino-terminal domain of the cytokine receptor homologous region (Hiraoka, O., Anaguchi, H., Yamasaki, K., Fukunaga, R., Nagata, S., and Ota, Y. (1994) J. Biol. Chem. 269, 22412-22419). The tertiary folding and the beta-sheet structure of the BC domain were confirmed by NMR spectroscopy. The disulfide bond pattern suggested from peptide mapping was Cys224-Cys271 and Cys242-Cys285. Disruption of the disulfide bonds suggested that both bonds are critical for maintaining the folding of the BC domain, although a BC domain lacking the second bond still retained ligand binding activity. Mutational analysis of the WSXWS sequence conserved in the cytokine receptor family suggested that this motif is critical for protein folding rather than for ligand binding.

Requirement for the immunoglobulin-like domain of granulocyte colony-stimulating factor receptor in formation of a 2:1 receptor-ligand complex.

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.

Evidence for the ligand-induced conversion from a dimer to a tetramer of the granulocyte colony-stimulating factor receptor.

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 (Kd = about 100 pM), which is comparable with the Kd 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.

Ligand binding domain of granulocyte colony-stimulating factor receptor.

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.

Involvement of two sulfur atoms of protein disulfide isomerase and one sulfur atom of the DsbA/PpfA protein in the oxidation of mutant human lysozyme.

Protein disulfide isomerase (PDI) and the DsbA/PpfA protein catalyze the oxidation of mutant human lysozyme, L79CC81A, which has two native disulfide bonds, Cys6-Cys128 and Cys30-Cys116, a non-native Cys79-Cys95, and 2 free cysteine residues at positions 65 and 77. Oxidation of L79CC81A (R-form) yielded two isomers, L79CC81A-a (A-form) with tandem-linked Cys65-Cys77 and Cys79-Cys95, and L79CC81A-b (B-form) with cross-linked Cys65-Cys79 and Cys77-Cys95 (Kanaya, E., Ishihara, K., Tsunasawa, S., Nokihara, K., and Kikuchi, M. (1993) Biochem. J. 292, 469-476). PDI mainly enhanced the formation of the A- form in the absence of oxidized glutathione (GSSG); however, as the concentration of GSSG increased, it markedly accelerated the formation of the B-form. In contrast, the DspA/PpfA protein mainly enhanced the formation of the A-form, regardless of the presence or absence of GSSG. These results and the presumed spatial locations of Cys65, Cys77, and Cys79-Cys95 in the R-form suggest that 1 of the half-cystine residues in the active site of PDI and the DsbA/PpfA protein can react with 1 of the 2 free Cys residues of the R-form. The dependence on GSSG of the B-form formation with PDI can be explained by the formation of two transient intermolecular disulfide bonds between PDI and the R-form and the attack of GSSG by the resultant thiolate anion of Cys79 or Cys95. The independence of the reaction with the DsbA/PpfA protein from GSSG can be explained by the formation of one transient intermolecular disulfide bond. The possible formation of the two transient intermolecular disulfide bonds involving two sulfur atoms of PDI and 2 cysteine or half-cystine residues of the substrate could explain the high isomerase activity of PDI.

Isolation and expression of the full-length cDNA encoding CD59 antigen of human lymphocytes.

To identify the primary structure of CD59 antigen and to elucidate its function, a full-length cDNA clone of CD59 was isolated. The cDNA sequence contained an open reading frame that encodes an 128-amino-acid peptide. The amino-terminal 25 amino acids represented a typical signal peptide sequence and the carboxy-terminal hydrophobic amino acids were characteristic for phosphatidylinositol-anchored proteins. The predicted mature protein sequence showed 35% homology with murine Ly-6C.1 and 31% with Ly-6A.2. The number and the distribution of cysteine residues were conserved, implying that the CD59 represented a human homologue of murine Ly-6. RNA blot hybridization analysis revealed the expression of CD59 mRNA in placental, lung, and pancreatic tissues. The mRNA was not only expressed in T-cell lines but in some of monocytic, myeloid, and B-cell lines. In all of these tissues and cell lines, at least four mRNA species were detected. DNA blot hybridization analysis revealed a rather simple genomic structure, which suggested a single gene as compared with the complex multigene family of murine Ly-6.

'Salivary peptide P-C' of human pancreatic B-cells shares only partly immunoreactivity with salivary peptide P-C indicating a new B-cell protein which is different from insulin.

Salivary peptide P-C like immunoreactivity, originally isolated from human whole saliva has later been found in the human pancreatic B-cells. In the present work an indirect immunofluorescence technique using monoclonal antibodies against isolated salivary peptide P-C was applied to Bouin fixed pancreas and parotid glands to study the possible identity of the two substances. Positive P-C immunofluorescence was found in the serous cells of parotid glands but not in pancreatic B-cells, suggesting that pancreatic P-C substance is not salivary peptide P-C itself, but a substance sharing the common antigenic site with salivary peptide P-C. To examine this, an indirect immunofluorescence technique using polyclonal P-C antisera pre-absorbed with six kinds of synthetic fragments (1-22, 23-44, 23-29, 30-44, 30-38 and 38-44) of salivary peptide P-C was applied to the human pancreas. The result showed that pancreatic P-C substance was a substance which shares the common antigenic site with the 38-44 amino acid residue of salivary peptide P-C. Western blot analysis using extracts of human pancreata further showed that pancreatic P-C substance is not a precursor of insulin but a protein with molecular weight of 11,500 dalton, indicating the presence of a new protein in the insulin secretory granules of human pancreatic B-cells.

Two developmental genes encoding sigma factor homologs are arranged in tandem in Bacillus subtilis.

The sporulation-essential gene spoIIG of the Gram-positive bacterium Bacillus subtilis encodes the sporulation-specific sigma factor sigma 29(sigma E). We report here the initial characterization of a gene, referred to as ORF3, located immediately downstream of the spoIIG gene. The results indicate that ORF3 encodes a sigma homolog, whose expression is highly regulated during development. Analysis of the ORF3 nucleotide sequence reveals an open reading frame encoding a polypeptide of 260 amino acid residues (molecular mass of 30.1 kDa). Its predicted amino acid sequence shows significant similarity to that of other RNA polymerase sigma factor sequences. S1 nuclease mapping experiments indicate that ORF3 is initially cotranscribed with spoIIG from about 1 to 4 hr into the sporulation process and that later on ORF3 is transcribed independently from a new site located between spoIIG and ORF3. The role of ORF3 was investigated by constructing a deletion mutation in its structural gene. The mutant exhibits normal growth but is unable to produce heat-resistant spores. We propose that the ORF3 gene product is a sigma factor or a related peptide essential for sporulation at a late stage of development.

Revised restriction maps of Bacillus subtilis bacteriophage phi 105 DNA.

Physical mapping of Bacillus subtilis temperate phage phi 105 DNA was carried out by using restriction endonucleases EcoRI, SmaI, and KpnI, and a new revised EcoRI cleavage map is presented. In addition, the EcoRI cleavage maps of six specialized transducing phages carrying sporulation genes of B. subtilis were revised.

Cloning of sporulation gene spoIIC in Bacillus subtilis.

Specialized transducing phages rho 11spoIIC and phi 105spoIIC, carrying the Bacillus subtilis sporulation gene spoIIC, were constructed by the prophage transformation method. An EcoRI fragment (2.4 MDal) carrying the spoIIC gene was isolated from the phi 105spoIIC genome and recloned into the EcoRI site of plasmid pUB110. The recombinant plasmids corrected the sporulation defect of a Spo- Rec- host, but slightly inhibited the sporulation of a Spo+ Rec- host.