Secretion of human epidermal growth factor from Saccharomyces cerevisiae using synthetic leader sequences.

We have investigated different leader sequences for their ability to direct the efficient secretion of human epidermal growth factor (hEGF) from Saccharomyces cerevisiae. We designed a consensus signal sequence which directs secretion of hEGF from yeast as efficiently as the yeast invertase signal sequence. However, secretion is increased over fivefold by the introduction, after the signal sequence, of a synthetic 19-amino acid (aa) pro-sequence containing a cleavage recognition site for the KEX2 protease. Even in the absence of an Asn-linked glycosylation site, secretion of hEGF using the synthetic prepro-leader was as efficient as that directed by the alpha-factor leader. The role of the KEX2 protease cleavage site was investigated by mutation of the yeast alpha-factor KEX2 site (cleavage after Lys-Arg). Cleavage was obtained with the following order of efficiency, Lys-Arg greater than Pro-Arg greater than Asp-Arg, although the sequence context was also found to affect efficiency.

Novel modified beta-interferons: gene cloning, expression, and biological activity in bacterial extracts.

A series of novel, modified interferons based on the structure of human beta-interferon have been expressed in Escherichia coli. Modified interferon genes were constructed from sequences derived from the natural beta-interferon gene, a synthetic beta-interferon gene, or a specific combination of the two. A total of 23 out of the 25 novel interferons exhibited antiviral (AV) and antiproliferative (AP) activity which varied from 3 to 230% and 8 to 490% of the values for beta-interferon, respectively. None of the novel interferons had only AV or AP activity, although one had a much reduced ratio of AV/AP activity compared with beta-interferon. Substitution of beta-interferon amino acids 2-7 or 28-46 resulted in interferons with significantly increased AP activity on Daudi lymphoblastoid cells (four- to fivefold). All the novel interferons except two with modifications in the 82-105 region reacted with a neutralizing beta-interferon monoclonal antibody.

Biological properties of human interferon beta 1 synthesized in recombinant bacteria.

Human fibroblast interferon, designated IFN-beta 1, has been produced in E. coli by direct expression of the cloned cDNA coding for the mature polypeptide. Bacterial lysates from recombinant cultures contain a polypeptide with an apparent molecular weight of 17,500 that corresponds in size to the unglycosylated IFN-beta 1 molecule. The latter could be specifically immunoprecipitated by antibodies to purified natural IFN-beta and could inhibit the replication of Herpes simplex virus types 1 and 2 in many different cell lines. Like the natural fibroblast IFN-beta, the bacterial IFN-beta 1 was active in many human cell lines, less active in a monkey cell line and inactive in rabbit and mouse fibroblasts. The antibody titre required to neutralise the anti-herpes activity of both IFN preparations was similar suggesting that they have the same specific activities. Similarly, the bacterial IFN-beta 1 was equally active in inhibiting the proliferation of Daudi cells grown in culture. Bacterial IFN-beta 1 was also capable of enhancing natural killer cell activity and antibody-dependent cellular cytotoxicity in vitro. Thus, IFN-beta 1 produced in recombinant bacteria displays a large range of biological properties ascribed to the natural fibroblast IFN-beta molecule.

The absence of introns within a human fibroblast interferon gene.

Experiments in which immobilised restriction fragments of genomic DNA were hybridised with a cloned human fibroblast interferon cDNA indicate that the homologous chromosomal genes exist in only one basic arrangement. This is in marked contrast to recent studies by Nagata et al. (1) showing that there are at least eight gene arrangements for human leukocyte interferon. Having isolated a chromosomal human fibroblast interferon gene from a gene bank, we conclude from nucleotide sequencing studies that there is a complete absence of introns within the RNA-coding region. In view of a similar observation recently made for a human leukocyte interferon gene (1), it would appear as if interferon genes in general are unlike the vast majority of eukaryote genes in this respect.

The complete amino acid sequence of human fibroblast interferon as deduced using synthetic oligodeoxyribonucleotide primers of reverse transcriptase.

Using synthetic oligodeoxyribonucleotides to prime the transcription of interferon mRNA and cDNA, we recently determined the mRNA sequence coding for the 47 amino-terminal amino acids of mature human fibroblast interferon (1). From this sequence, we have now synthesised an oligodeoxyribonucleotide that is homologous with the mRNA sequence coding for amino acids 42-45 and used it as a primer to selectively transcribe an interferon cDNA template. The sequence of the newly synthesised DNA predicted the sequence of amino acids 48-109 in the interferon polypeptide. By repeating this process with one more primer, we have determined the complete amino acid sequence of mature human fibroblast interferon, a polypeptide of 166 amino acids.

Influenza antigenic determinants are expressed from haemagglutinin genes cloned in Escherichia coli.

A gene sequence for the fowl plague virus (FPV) haemagglutinin molecule has been inserted into a bacterial plasmid such that its transcription is under the control of a promoter derived from the tryptophan operon. Such plasmids direct the synthesis of a protein that reacts specifically with antisera to FPV haemagglutinin. Evidence is also presented that in some cases DNA inserted at the HindIII site of pBR322 is expressed.

Polyadenylation and reverse transcription of influenza viral RNA.

The polyadenylation of Fowl Plague Viral RNA and of Influenza A/Victoria Viral RNA using E. coli poly (A) polymerase and the subsequent reverse transcription of the polyadenylated species is reported. We have shown that all 8 genome fragments are adenylated and that an average of 25--30 adenylic acid residues per molecule is sufficient for maximal transcription with reverse transcriptase. The cDNA product is 95% sensitive to Sl-nuclease and hybridisation analysis against viral RNA reveals it to be a faithful copy of the RNA. Amongst the transcription products are long, discrete copies of genes 1--8, the lengths of which are comparable with those of the vRNA determined by electrophoresis on formamide acrylamide gels. These single-stranded cDNAs have been further transcribed to form double-stranded products with hair-pin structures at one end. Analysis of this material on native acrylamide gels revealed some DNA bands corresponding to the predicted sizes for genes 4--8.