Urease is a virulence determinant, a taxonomic and diagnostic marker, and immunogen for Helicobacter pylori, an aetiologic agent of gastritis and peptic ulceration. This enzyme requires Ni2+ ions in the active site for successful hydrolysis of urea. When expressed in Escherichia coli, recombinant urease is only weakly active unless urease structural subunits are overexpressed, exogenous NiCl2 is added, and the host strain is grown in medium that does not chelate free Ni2+. As wild-type H. pylori does not require such conditions for very high levels of urease expression, we reasoned that additional genes were required to accumulate the metal ion. To isolate such genes, E. coli SE5000 (pHP808), which carries the H. pylori urease gene cluster, was complemented with a lambda ZAP-derived plasmid library of the H. pylori chromosome. One of 1000 ampicillin-resistant clones, plated onto urea segregation agar, produced detectable urease. Urease activity of this co-transformant, grown in Luria broth containing 1 microM NiCl2, was 36 mumol NH3 min-1 mg-1 protein. Urease-enhancing activity, which is not directly linked to the urease gene cluster, was localized by subcloning and nucleotide sequencing. The largest open reading frame, designated nixA, predicted a polypeptide of 34,317 Da that displayed characteristics of an integral membrane protein. In vitro transcription-translation of nixA sequences yielded a polypeptide estimated to be 32 kDa in size. An in-frame Bal31 deletion within nixA abolished urease-enhancing activity. At 50 nM NiCl2, E. coli containing the nixA clone transported 1250 +/- 460 pmol Ni2+ min-1 10(-8) cells, whereas the vector control transported only 140 +/- 85 pmol Ni2+ min-1 10(8) cells, i.e. significantly less (P = 0.01). We conclude that NixA confers upon E. coli a high-affinity nickel-transport system (KT = 11.3 +/- 2.4 nM; Vmax = 1750 +/- 220 pmol Ni2+ min-1 10(-8) cells) and is necessary for expression of catalytically active urease, regardless of growth conditions.