U‐937 cells, a monocytic line derived from a human histiocytic lymphoma, were induced for human tumor necrosis factor (TNF) secretion into the medium and were used for the preparation of TNF mRNA. Biological activity of the latter was quantified in a Xenopus laevis oocyte injection system. TNF mRNA was enriched by gradient centrifugation and this size‐fractionated mRNA was used for synthesis of cDNA and inserted into the unique PstI site of pAT153. A recombinant plasmid containing human TNF cDNA was selected by colony hybridization using an internal fragment of a mouse TNF cDNA clone [Fransen, L., Mueller, R., Marmenout, A., Tavernier, J., Van der Heyden, J., Kawashima, E., Chollet, A., Tizard, R., Van Heuverswyn, H., Van Vliet, A., Ruysschaert, M. R. & Fiers, W. (1985) Nucleic Acids Res. 13, 4417–4429] as a probe. The sequence of this human TNF cDNA is in agreement with the one published by Pennica et al. [Pennica, D., Nedwin, G. E., Hayflick, J. S., Seeburg, P. H., Derynck, R., Palladino, M. A., Kohr, W. J., Aggarwal, B. B. & Goeddel, D. V. (1984) Nature (Lond.) 312, 724–729]. The 157‐amino‐acid‐long mature sequence is about 80% homologous to mouse TNF and its hydrophilicity plot is also very similar, in spite of the apparent species specificity of TNF. In contrast to mouse TNF, it contains no potential N‐glycosylation site. When compared to other cytokines, like IFN‐β, IFN‐γ, or IL‐2, there is a remarkably high preference for G · C pairs in the third‐letter positions. Expression of the TNF cDNA in monkey COS cells or in Escherichia coli gives rise to a protein having similar biological and serological properties as natural human TNF. A human genomic clone was also identified and sequenced; it was found to be in good agreement with the one recently published by Shirai et al. [Shirai, T., Yamaguchi, H., Ito, H., Todd, C. W. & Wallace, R. B. (1985) Nature (Lond.) 313, 803–806], except for some differences in the introns and 5′‐untranslated region.

INTRODUCTIONThe antitumour effect of endotoxins has been known for over a century because of remarkable observations regarding spontaneous regression of

The biosynthesis of human tumor necrosis factor (hTNF) was studied. The amino-termina1 extension of the hTNF precursor (26 kDa polypeptide) was not cleaved off in a cell-free system supplemented with dog pancreas microsomes. Correct maturation of pre-hTNF was nevertheless not restricted to the macrophage system: in the medium of a TNF-producing, transformed CHO ceil line, a (weak) N 20 kDa, an N 18.5 kDa (doublet) and a 17 kDa TNF poly~ptide, the latter corresponding to mature hTNF, were revealed by specific immunop~~ipitation.Similar results were obtained with Xenopus laevis oocytes, injected with hTNF mRNA, except that the 20 kDa band was lacking. The results are discussed in relation to the secretion mechanism of hTNF.