Rapid Method for Cloning a Ca2+ -Dependent Metalloprotease From a Gram-Positive Bacterium in Escherichia coli

Cloning the genes of Ca2+-dependent proteases from different genera of gram-positive bacteria such as Bacillus (1,10,11) or Staphylococcus (8) in E. coli is very often lethal for the recipient, and that is why these genes had to be cloned by propagating different flanks separately in E. coli or by choosing another cloning host (1,8–11). The nprM gene of Bacillus megaterium coding for the neutral protease also has a lethal effect on E. coli (9). However, by applying the mentioned strategies, it has been cloned and sequenced for different strains of B. megaterium (3,4). Here we describe a method that allows the cloning of this protease gene directly in E. coli. The method was developed for B. megaterium DSM 319, which is known to produce two extracellular proteases, a minor serine protease (9) and a neutral metalloprotease (5,6). It is central to the method and necessary that EDTA would complex the metal cofactor, eventually resulting in an enzymatically inactive protein. We used the expression vector pUCnprM carrying, in addition to the ampicillin-resistance gene, the nprM gene governed by its own promoter (9). Cells of the strain E. coli HB101 were transformed following the Hanahan method (7), with the following modifications: a Ca2+-free transformation buffer (10 mM 2-N-morpholinoethane sulfonic acid, pH 6.3, 45 mM MnCl2 • 4 H2O, 10 mM MgCl2 • 6 H2O, 100 mM KCl, 3 mM hexamine cobalt trichloride) was used. Furthermore, SOBmedium (2% Bacto-tryptone, 0.5% Bacto-yeast-extract, 10 mM NaCl, 2.5 mM KCl, pH 7.0) that was used to grow cells intended for an immediate transformation was made 20 mM in MgCl2 (out of a 2 M MgCl2 stock). To protect cells from neutral protease activity, transformations were performed in the presence of different concentrations of EDTA in the regeneration medium (liquid SOB medium without MgCl2 with 20 mM glucose by adding a 1 M glucose stock) and/or the SOB plates (SOB medium solidified with 1.5% agar and containing 100 μg/mL ampicillin) on which transformants were selected. As SOB medium is <0.1 mM in Ca2+ and <0.5 mM in Mg2+ (2), which can both interact with EDTA, we tested concentrations ranging from 0 mM EDTA up to 0.5 mM EDTA (Figure 1). The use of a modified transformation buffer allows transformation of E. coli HB101 with pUCnprM, although transformation efficiency remains considerably low, i.e., 4 colony-forming units (cfu) per μg plasmid-DNA. Selection of transformants on SOB medium containing EDTA allowed increasing transformation efficiencies up to 18-fold (0.3 mM EDTA). Higher transformation efficiencies were achieved when the regeneration medium also contained EDTA. We obtained the best transformation results when using 0.1 mM EDTA in the regeneration medium and 0.3 mM EDTA in the selective medium (5.625 × 103 cfu/μg DNA; i.e., 140-fold higher when compared to a procedure without EDTA in regeneration and selective medium). In the regeneration medium, 0.1 mM EDTA proved to give the best results and should not be exceeded, as higher concentrations led to a decrease in transformation efficiencies (up to 6fold, when compared to the highest transformation efficiency). Selective medium containing 0.4 or 0.5 mM EDTA also results in a decrease in transformation efficiencies (remaining below 103 cfu/μg DNA under all tested conditions), thus the concentration of EDTA in selective media should not exceed 0.3 mM. Once the nprM-bearing plasmid is established in E. coli, cells can be propagated without addition of EDTA. Transformants grown on Ca2+-caseinate agar (Merck, Darmstadt, Germany) containing 100 μg/mL ampicillin were found to form clear halos, indicating protease activity. Furthermore, the B. megaterium strain MS941, carrying an inactivated copy of the nprM gene (9), was transformed with pUCnprM isolated from transformed E. coli cells. The formation of clear halos on Ca2+-caseinate agar by these transformants proved that pUCnprM from E. coli carries a functional copy of the nprM gene. This result was confirmed by the determination of protease activities (data to be published elsewhere). We suggest that the effect of the neutral