Genome sequencing of cellulolytic myxobacterium Sorangium cellulosum reveals many open-reading frames (ORFs) encoding various degradation enzymes with low sequence similarity to those reported, but none of them has been characterized. In this paper, a predicted lipase gene (lipA) was cloned from S. cellulosum strain So0157-2 and characterized. lipA is 981-bp in size, encoding a polypeptide of 326 amino acids that contains the pentapeptide (GHSMG) and catalytic triad residues (Ser114, Asp250 and His284). Searching in the GenBank database shows that the LipA protein has only the 30% maximal identity to a human monoglyceride lipase. The novel lipA gene was expressed in Escherichia coli BL21 and the recombinant protein (r-LipA) was purified using Ni-NTA affinity chromatography. The enzyme hydrolyzed the p-nitrophenyl (pNP) esters of short or medium chain fatty acids (<= C(10)), and the maximal activity was on pNP acetate. The r-LipA is a cold-adapted lipase, with high enzymatic activity in a wide range of temperature and pH values. At 4 degrees C and 30 degrees C, the K(m) values of r-LipA on pNP acetate are 0.037 +/- 0.001 and 0.174 +/- 0.006 mM, respectively. Higher pH and temperature conditions promoted hydrolytic activity toward the pNP esters with longer chain fatty acids. Remarkably, this lipase retained much of its activity in the presence of commercial detergents and organic solvents. The results suggest that the r-LipA protein has some new characteristics potentially promising for industrial applications and S. cellulosum is an intriguing resource for lipase screening.