Molecular Microbiology (1991) 5(5), 1175-1181 ADONIS 0 9 5 0 3 8 2 X 9 1 0 0 1 3 0 0 A pyruvate-stimuiated adenylate cyclase has a sequence related to the fes/fps oncogenes and to eukaryotic cyclases E. P. Peters,'* A. F. Wilderspin,^ S. P. Wood,* M. J. J. M. Zvelebii,^^ O. Sezer^ and A. Danchin* ' The ICRF Unit tor Structural Molecular Biology, and ^The Laboratory of Molecular Biology. Department of Crystallography, Birbeck College, Malet Street, London WC1EZHX. UK. ^ Laboratory of Biomoteoular Modelling, Imperial Cancer Research Fund, Lincoln's Inn Fields, London WC2A 3PX, UK. ^Unite de Regulation de I'Expression Genetique. Institut Pasteur, rue du Dr Roux, 75724 Paris Cedex 15, France. Summary The pyruvate-stimuiated adenylate cyclase from Brevibacterium liquefaciens produces up to 450 M.M cyclic AMP in the culture medium when the bacterium is grown on glucose and alanine. In this paper we report the cloning, expression and sequencing of the gene for this enzyme. Residues were identified, within the C-terminal domain, which are conserved in adeny- late and guanylate cyclase sequences from euka- ryotes and in the adenylate cyclase of the prokaryote Rhizobium meliloti. We have also identified a sequence of 30 residues near the N-terminus of the protein which is homologous to part of the regulatory domain of the cellular homologues of the oncogenes fes and fps; this sequence is also present in the avian Fujinami sarcoma virus fps gene. Introduction Adenyiate cyclase is part of the signal-transduction path- way in higher eukaryotes, producing the second mess- enger cyclic AMP (Berrldge, 1985). In prokaryotes it is involved in the regulation of sugar metabolism (Ullmann and Danchin, 1983) and it forms part of the toxins of the causative agents of whooping cough and anthrax (Confer and Eaton, 1982; Leppla, 1984). Many bacteria have a Received 5 September. 1990; revised 21 January, 1991. 'Present address: Slomolacular Structure and Modeiling Unit, Biochemistry Department, University College, Gowsf Street, London WC1E 6BT, UK. 'For corre- spondence. Tel. (071) 6316268; Fax (071) 4368918, pyruvate-stimuiated cytoplasmic adenylate cyclase with an unknown function (Ide, 1971). Adenylate cyclases vary greatly in length (from 193 amino acid residues for the Rhizobium meliloti enzyme to 1706 residues for the Bordetella pertussis enzyme) and may exist as monomers or dimers. The pyruvate-stimuiated adenylate cyclase from Brevi- bacterium liquefaciens has been purified to apparent homogeneity and was found to be a dimer with a molecular mass of 46kDa per subunit (Takai etai, 1974). Heterologous expression of the gene for this enzyme in Escherichia coli is required to allow purification of suffi- cient enzyme for structural studies. The B. liquefaciens adenylate cyclase was selected for structural analysis because it is not an intrinsic membrane protein, it has a relatively low molecular weight, and microcrystals of the enzyme have been reported (Takai et al., 1974). In this paper we report the cloning and sequencing of the B. liquefaciens adenytate cyclase and the expression of low levels of enzyme activity in £ coli. We also discuss its relationship with the sequences of adenylate and guanylate cyclases and other known proteins. Results and Discussion Cloning the B. liquefaciens adenylate cyclase in E. coli The gene for adenylate cyclase from B. liquefaciens was obtained by screening a library of genomic B. liquefaciens DNA for complementation of the Cya phenotype in the E. coli strain TP610 on selective media. (Wang et ai, 1981; Aiba etai. 1983; Hedegaard and Danchin, 1985; Glaser ef ai, 1988). The gene was then subcloned in pBR322. The two smallest subclones, pEP5 (8kb insert) and pEP8 (4kb insert), had similar restriction maps in the centres of the fragments and are thought to contain the same gene. Further subcloning of a 3kb Sg/ll/Ps(i fragment of pEP8 Into the vector pGA46 (An and Freisen, 1979) gave a plasmid, pEPIO, which complemented the Cya" phe- notype. To define the position of the gene within the pEP8 insert, sections of the insert were deleted, as shown in Fig. 1. Deletions of fragments up to the Nru\ site did not result in loss of the Cya'^ phenotype, but deletion to the BamHI site or removal of the Nco\-Nco\ fragment did cause loss ot