Mistranslation of a TGA termination codon as tryptophan in recombinant platelet-derived growth factor expressed in Escherichia coli.

The mature 109-amino-acid human platelet-derived growth factor B (PDGF-B) peptide is derived by intracellular processing from a 241-amino-acid precursor synthesized in mammalian cells, with removal of 81 N-terminal and 51 C-terminal amino acids. In order to produce directly the mature 109-amino acid PDGF-B peptide as a recombinant protein in Escherichia coli, a CGA codon at position 110 of a DNA sequence encoding the full-length precursor form of PDGF-B was converted into the translation termination codon TGA by in vitro mutagenesis. Expression of this DNA via a plasmid vector in E. coli resulted in production of two distinct PDGF-B proteins having apparent molecular masses of 15 and 19 kDa, with the latter species predominating. Structural characterization employing N- and C-terminal amino acid sequencing and MS analyses indicated that the 15 kDa protein is the expected 109-amino-acid PDGF-B, and that the 19 kDa protein represents a C-terminal extended PDGF-B containing 160 amino acids. Characterization of a unique tryptic peptide derived from the 19 kDa protein revealed that this longer form of PDGF-B results from mistranslation of the introduced TGA termination codon at position 110 as tryptophan, with translation subsequently proceeding to the naturally occurring TAG termination codon at position 161. Owing to the high rate of translation readthrough of TGA codons in this and occasionally other proteins, it appears that the use of TGA as a translation termination codon for proteins to be expressed in E. coli should be avoided when possible.

Formation of mitogenically active PDGF-B dimer does not require interchain disulfide bonds.

Platelet-derived growth factor (PDGF), a major mitogen for mesenchymal cells, is a disulfide-bonded dimer of two subunit polypeptides named A and B. All of the three possible dimeric forms, i.e. AA, BB, and AB, exist in nature. The dimeric structure has been presumed to be necessary for biological activity, since reduction of the dimer results in loss of activity and simultaneous conversion to monomeric form as determined by SDS-gel electrophoresis. However, reduction of the native molecule destroys intrachain, as well as interchain, disulfide bonds, and it is possible that the former rather than the latter are critical for proper conformation of the active protein. We show here that PDGF-B polypeptides in which all 8 cysteines or the 2nd, 4th, 5th, and 8th cysteines have been mutated to serines fail to form covalent dimers and possess dramatically less mitogenic activity than native PDGF-BB. Another mutant, PDGF-B(C2,4S), in which just the 2 cysteines involved in interchain disulfides were converted to serine, ran as a monomer on SDS-polyacrylamide gels as expected. Somewhat unexpectedly, however, the mitogenic activity of the PDGF-B(C2,4S) analog was similar to the activity of wild-type PDGF-BB disulfide-bonded dimer under physiological conditions. The activity of the analog was more sensitive to the effect of low pH than was the activity of wild-type PDGF-BB. Molecular weight analysis utilizing light scattering and sedimentation equilibrium demonstrated that the PDGF-B(C2,4S) analog exists as a noncovalent dimer at pH 4-7 but dissociates to a monomer at pH 2.5. Disulfide analysis of the mutant protein demonstrated that the intrachain disulfide bonds are the same as those formed in wild-type PDGF-BB homodimers. We conclude that proper formation of intrachain disulfide bonds is critical to maintaining the correct conformation of PDGF monomers, but that appropriately folded monomers can associate into active noncovalent dimers in the absence of interchain disulfide bonds. Interchain disulfide bonds thus appear to increase the stability of the PDGF dimer rather than being crucial to its existence.

Hybrid tyrosine kinase/cytokine receptors transmit mitogenic signals in response to ligand.

While much is known about the mechanisms by which members of the receptor tyrosine kinase family effect mitogenic signal transduction, much less is known about such mechanisms for members of the hematopoietic cytokine receptor family. In an effort to determine the extent to which the signal transduction mechanisms of these two receptor families may be related, we constructed and tested interfamily hybrid receptors. Two hybrid receptors consisting of the ligand-binding domain of the epidermal growth factor receptor (EGFR) fused to the transmembrane and cytoplasmic domains of the erythropoietin receptor (EPOR), as well as the parental EPOR or EGFR, were introduced into interleukin-3-dependent 32D cells. Part of the EPOR extracellular region containing a conserved WSXWS amino acid motif was present in one of the hybrid receptors but not in the other. Cells expressing EGFR grew only poorly in response to EGF, whereas cells expressing either of the EGFR/EPOR hybrid receptors or EPOR grew robustly in epidermal growth factor or erythropoietin, respectively. This is the first demonstration of a chimera between these two unrelated receptor families that responds to ligand stimulation. The results indicate that the mechanisms by which receptor tyrosine kinases and cytokine receptors propagate mitogenic signals are sufficiently similar to allow interchange of their ligand binding domains and that generation of an efficient mitogenic signal by a cytokine receptor depends primarily on its cytoplasmic and/or transmembrane regions.

mRNA species regulated during the differentiation of HL-60 cells to macrophages and neutrophils.

Using cDNA clone banks from differentiated and undifferentiated HL-60 promyelocytic leukemia cells, we have selected clones for genes which are regulated during this differentiation. Regulation of the corresponding mRNAs in HL-60 cells during both monocytic and neutrophilic differentiation was measured for 21 of these clones. The levels of mRNA hybridizing to some of these clones changed by more than 100-fold during differentiation. Unlike erythropoiesis or myogenesis, in which the synthesis of a few new proteins is synchronously regulated, mRNAs in differentiating HL-60 cells are asynchronously regulated, suggesting a complex series of regulatory events. About half of these regulation-selected clones contained repeat sequences, including both Alu and novel repeat families. Most of the regulated genes are members of extensive gene families.

Expression of biologically active bovine luteinizing hormone in Chinese hamster ovary cells.

Biologically active bovine luteinizing hormone (LH) has been obtained through expression of the alpha- and LH beta-subunit genes in stably transformed clones of DUXB11, a Chinese hamster ovary cell line deficient in dihydrofolate reductase (DHFR). Expression of alpha-and LH beta-subunit mRNAs of the expected sizes (approximately 910 and 770 nucleotides, respectively) were revealed by blot analysis after electrophoresis of total cellular RNA. Furthermore, presence or absence of the gonadotropin mRNAs in several clonal lines was directly correlated with the appearance of one or both bovine LH subunits in the culture medium. Media from three clones secreting significant immunoreactive levels of both subunits also stimulated the release of progesterone in ovine luteal cells, suggesting that the secreted LH was assembled into a biologically active and glycosylated dimer. Immunoprecipitation and NaDodSO4/PAGE of [35S]methionine-labeled proteins secreted from one of the clones, CHODLH20, further confirmed the presence of an alpha/beta dimer with apparent subunit molecular weights of 20,500 and 16,000, only slightly higher than those of pituitary alpha and LH beta subunits.

The gene for the beta subunit of bovine luteinizing hormone encodes a gonadotropin mRNA with an unusually short 5'-untranslated region.

Both cDNA and genomic clones encoding the beta subunit of bovine luteinizing hormone (LH) have been isolated and characterized. The nucleotide sequence was determined for the entire gene and 776 base pairs of 5'-flanking sequence. The mRNA cap site and polyadenylation site were mapped by primer extension and S1 nuclease protection, respectively. The bovine LH beta spans less than 1.1 kilobase pairs and has three exons encoding a 550 nucleotide mRNA (excluding the poly(A) tail). Bovine LH beta is a single-copy gene, in contrast to human LH beta, which is a member of the LH/chorionic gonadotropin beta subunit multigene family. Comparison of the bovine LH beta gene with the human LH beta/chorionic gonadotropin gene family reveals a high degree of nucleotide sequence homology, both within the genes and in the 5'-flanking sequences. Despite this extensive sequence conservation, there is a major difference between the two species in the selection of a promoter site. As a result, the bovine LH beta gene produces an mRNA with an usually short 5'-untranslated region of only 6-11 nucleotides.

The common alpha subunit of bovine glycoprotein hormones: limited formation of native structure by the totally nonglycosylated polypeptide chain.

The folding of the bovine glycoprotein hormone alpha subunit, synthesized in bacteria following insertion of the nucleotide sequence coding for this polypeptide, has been studied to determine the effect that a complete lack of carbohydrate has on this process. The bacterially derived alpha polypeptide (bac-alpha), extracted from E. coli in the presence of reductant and denaturant, had an estimated 0.2% native structure as determined by a conformationally sensitive radioimmunoassay. Upon reduction of disulfide bonds and reoxidation in air, the amount of native structure increased about 18-fold. Approximately 2% of the refolded bac-alpha preparation combines with the beta subunit of human chorionic gonadotropin (hCG beta) to form a complex that binds to the gonadotropin receptor and elicits a biological response. Since the correct folding (by immunological criteria) of bac-alpha (ca 3%) is significantly greater than expected from a random formation of disulfide bonds (0.1%), it appears that correct folding of alpha subunit can occur in the complete absence of carbohydrate, though in very low yield. Native bovine lutropin alpha subunit (LH alpha) and chemically deglycosylated LH alpha (which retains two asparagine-linked N-acetyl glucosamine residues per alpha oligosaccharide) were subjected to the same reduction/reoxidation regimen as the bacterially produced alpha subunit. As has been reported previously [Giudice LC, Pierce, JG, J Biol Chem 251: 6392, 1976] intact LH alpha fully regained its native structure. The partially deglycosylated LH alpha also refolds to a native-like structure in high yield as assessed by immunological assays and by its ability to combine with HCG beta to form a biologically active complex. The data show that carbohydrate, while not obligatory for correct folding, greatly facilitates the formation of functional alpha subunit.

Nucleotide sequence of a cDNA for the common alpha subunit of the bovine pituitary glycoprotein hormones. Conservation of nucleotides in the 3'-untranslated region of bovine and human pre-alpha subunit mRNAs.

A cDNA clone for the pre-alpha subunit of the pituitary glycoprotein hormones has been isolated from a bovine pituitary cDNA library through the use of a pool of synthetic oligodeoxynucleotide probes. This clone, designated pB alpha, contains a 564-base pair insert which includes a portion of the signal sequence, the entire coding sequence of the mature protein, and 224 base pairs of the 3'-untranslated sequence. As expected, the nucleotide and amino acid sequence of the mature bovine alpha subunit was homologous to the sequences reported for humans and rodents, with the most extensive homology occurring between bovine and rodents (85-90%). However, a comparison of the 3'-untranslated regions of pre-alpha subunit mRNA from three different mammalian species indicated that in bovine and rat, or in human and rat, these sequences have rapidly diverged, yielding respective homologies of 21 and 36%. In contrast, the sequence homology observed between the 3'-untranslated regions of bovine and human was 79%, which approaches the level of homology shared by their coding sequences. Thus, the conservation of the 3'-untranslated sequence in bovine and human pre-alpha subunit mRNA may be an indication that this region is functionally significant in these two species.

Construction and characterization of a cDNA clone containing a portion of the bovine prolactin sequence.

Poly(A)-containing RNA from the bovine anterior pituitary has been used as a template for the enzymatic synthesis of double-stranded cDNA. The resulting double-stranded cDNA was inserted into the Pst I site of pBR322 with the oligo(dG)-oligo(dC) tailing technique and subsequently cloned in E. coli chi 1776. Clones containing sequences complementary to prolactin mRNA were identified by colony hybridization with partially purified prolactin cDNA. A 250 base pair sequence from one prolactin positive clone was extensively characterized and shown to contain the coding information for amino acids 119-192 of authentic bovine prolactin. The recombinant DNA from this clone was covalently attached to diazotized aminocellulose and used to purify prolactin mRNA from a mixture of mRNAs.

Presence of 5'-terminal cap structures in virus-specific RNA from feline leukemia virus-infected cells.

The F-422 line of feline thymus tumor cells, chronically infected with the Rickard strain of feline leukemia virus (R-FeLV), was labeled with 32P, and the total cytoplasmic RNA was isolated. The RNA was centrifuged through sucrose gradients, and R-FeLV virus-specific RNA (vRNA) was located by hybridization of portions of the gradient fractions to R-FeLV complementary DNA. vRNA classes with average sedimentation coefficients of approximately 36S, 28S, 23S, and 15S were identified. Each class of RNA was recovered by hybridized with mercurated R-FeLV complementary DNA, and the hybrids were chromatographed on columns of sulfhydryl-Sepharose to separate them from unhybridized cellular RNA. Although insufficient amount of 36S and 28S vRNA were obtained for further analysis, the 23S and 15S VRNA classes were analyzed to determine the nature of their 5' termini. Each of these vRNA classes was found to contain stoichiometric amounts of cap structures per unit length of RNA, consistent with the presence of one cap per molecule. The structure of the 23S vRNA cap was found to be m7G5'ppp5'GmpAp, whereas that of the 15S vRNA cap was m7G5'ppp5'GmpGp.

Methylation of high-molecular-weight subunit RNA of feline leukemia virus.

The high-molecular-weight subunit RNA of feline leukemia virus (Rickard strain) (FeLV-R) was analyzed for the presence of methyl groups. After purification of native 50-60S FeLV-R RNA on nondenaturing aqueous sucrose density gradients. FeLV-R 28S subunit RNA, doubly labeled with [14C]uridine and [methyl-3H]methionine, was isolated by centrifugation through denaturing sucrose density gradients in dimethyl sulfoxide. As calculated from their respective 3H/14C ratios. FeLV-R 28S RNA was methylated to the same degree as host cell poly(A)+ mRNA. When the 28S FeLV-R RNA was hydrolyzed to completion with RNase T2 or alkali, all of the methyl-3H chromatographed with mononucleotides on Pellionex-WAX, a weak anion exchanger. The methyl-labeled material co-chromatographed with 6-methyladenosine if the mononucleotide fraction obtained by Pellionex-WAX chromatography was hydrolyzed to nucleosides by bacterial alkaline phosphatase or with 6-methyladenine if purine bases were released from the mononucleotides by acid hydrolysis. In another experiment in which FeLV-R 28S RNA uniformly labeled with 32P was hydrolyzed and then analyzed by Pellionex-WAX chromatography, all of the 32P label again co-chromatographed with mononucleotides. Thus FeLV-R 28S RNA does not appear to contain a 5' structure, either methylated or nonmethylated similar to those recently reported for cellular and some animal virus mRNA's.

Properties of feline leukemia virus. III. Analysis of the RNA.

The kinetics of virus labeling was used to study the maturation of viral RNA in the Rickard strain of feline leukemia virus. Viral RNA labeled over differing intervals was characterized by gel electrophoresis and velocity sedimentation in sucrose gradients made up in aqueous buffer and 99% dimethyl sulfoxide. Labeled virus was found within 30 min after adding radioactive uridine to the cells and production of labeled virus reached a maximum at 4 to 5 h after pulse labeling. Native RNA from feline leukemia virus resolved into three size classes when analyzed by electrophoresis on 2.0% polyacrylamide-0.5% agarose gels: a 6.2 x 10(6) to 7.1 x 10(6) mol wt (50 to 60S) class, an 8.7 x 10(4) mol wt (approximately 8S) class, and a 2.5 x 10(4) mol wt (4 to 5S) class. From two experiments during which RNA degradation appeared minimal, these made up to 57 to 76%, 2 to 5%, and 6 to 12%, respectively, of the total RNA. The 8S RNA in feline leukemia virus has not previously been reported. The 50 to 60S RNA from virus harvested after 4 h of labeling electrophoretically migrated faster and sedimented more slowly in sucrose gradients than did the same RNA species harvested after 20 h of labeling. This argues for an intravirion modification of the high-molecular-weight RNA. The large subunits of denatured viral RNA from both 4- and 20-h labeled-viral RNA electrophoretically migrated with an estimated molecular weight of 3.2 x 10(6) but sedimented with 28S ribosomal RNA (1.8 X 10(6) mol wt) when analyzed by velocity sedimentation through 99% dimethyl sulfoxide.